SCRIP: Scholarly Research In Progress 2020

Page 1

Volume 4 • October 2020

Scholarly Research In Progress

Table of contents

54 The Neuroscience of Resilience


57 Chronic Traumatic Encephalopathy: Investigating Blood-Brain Barrier Disruptions and Neuroimaging Techniques

Rise and Regional Variations in Schedule II Stimulant Use in the United States Sneha M. Vaddadi1†, Nicholas J. Czelatka1†, Belsy D. Gutierrez2, Bhumika C. Maddineni3, Carlos D. Torres-Teran4, Daniel N. Tron2, Kenneth L. McCall5, and Brian J. Piper1,6


Hip Dysplasia and Osteogenesis Imperfecta: Case Review and Management Recommendations Mark Mandel1†, Kaitlin Saloky1†, William Mirenda2, Andrea Seeley2, and Mark Seeley2


Adult Vasoactive-Inotropic Score Predicts Mortality in Adult Patients with Shock Alex McGeough1†, Timothy Farrell1†, Navindra Tajeshwar1†, Katarina Smigoc1†, and Niraj J, Vyas1†


Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States Kathy-Ann Cadet1*, Jasmine Nkwocha1*, and Myiah Smothers1*


A Critical Examination of the Mechanism of Action of Buprenorphine: Not Just a Mu Partial Agonist? Leana J. Pande1*, Stephanie D. Nichols2, and Brian J. Piper1,3

30 Genotypic and Phenotypic Abnormalities Linking 22q11.2 Deletion Syndrome and Psychiatric Disorders Cecilia Leng1*, Khine S.Y. Win1*, Ben Truong1*, Josely Frias1*, and Brian J. Piper1,2

34 Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018 John M. Boyle1†, Holly E. Funk1†, Susannah E. Pitt1†, Alison T. Varano1†, Sneha Vaddadi1†, Brian J. Piper1, and Kenneth L. McCall2


Pharmacological Progress of Biased Agonism of the μ Opioid Receptor Laura A. Christman * and Courtney J. Merwin1* 1

48 Brain Death: Investigating the Electrophysiological Events at Death in Animal and Human Research

Kelsey N. Plummer1*, Trent M. Filter1*, and Anvith Chidananda1*

Yasmín Mamani1*, Payel Girwarr1*, and Sayrah Rauf1*

63 A Bath a Day Keeps the Germs Away: Chlorhexidine Gluconate Bathing Expansion Quality Improvement Project Mahrukh Choudhary1†, Kenneth Lam1†, Connor M. Magura1†, Jasmine Santos1†, Julia Shamis1†, Navindra Tajeshwar1†, Hina Farrukh2, and Joseph Santora2

67 Methadone Distribution Trends from 2017–2018 Across the United States John A. Furst1†, Nicholas J. Mynarski1†, Jessica M. DeAngelis1†, Viraj Kothari1*, Jonique Depina1*, Kenneth L. McCall2, and Brian J. Piper1,3

72 Novel Treatments for Multiple Sclerosis Christine Rittenhouse1*, Jhamal Wallace1*, and Rebecca Kane1*

78 Systematic Review of Viral mRNA Vaccine Formulations, Modifications, and Adjuvants for Enhanced Safety and Efficacy Christian Pardo1*

85 Investigating Trends in Elder-Elder Caregiving in the United States: 1997 – 2014 SooYoung H. VanDeMark * 1

90 The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions Cecilia Leng1*‡, Leah E. Thomas1*‡, Khine S.Y. Win1*‡, Brian J. Piper1, Joseph B. Fraiman2, and Alex Hodkinson3

98 Oxalate Nephropathy after Bariatric Surgery: Incidence and Case Report of Successful Kidney Transplant after Conversion from Roux-en-Y Gastric Bypass to Gastric Sleeve Amandeep Kaur1†, G. Craig Wood2, Dan Bucaloiu3, Maria C. Bermudez3, Jason George3, and Alex R. Chang3

101 Radiotherapy Treatment Plan Evaluation Software Implementation and Breast Cancer Radiotherapy Plan Quality Kelley Chan1†, Daniel P. Talenti2, and Thomas J. Gergel2

107 Analysis of Opioid Distribution Before and After Recreational Marijuana Legalization in California Michelle N. Anyaehie1*, Elijah J. Johnson1*, and Christian Pardo1*

112 Emerging Modalities that Support Differential Diagnoses of Creutzfeldt-Jakob Disease and other Neurodegenerative Disorders Yasmin R. Mamani1*

116 Severe Pertussis Infection with Hyperleukocytosis Greater Than 200,000 103/uL in a 10-month-old Unvaccinated Amish Female: A Case Report and Review of the Literature Stephen Long1† and R. Blake Lowe2

119 Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats Albena Gesheva1†

126 Exploring the Role of APOE4 Genotype in Chemo Brain or Chemotherapy-Induced Cognitive Impairment Tracy L. Chofor1*

131 Perceived Barriers to Physical Activity and Nutrition Among Geisinger Students Rebecca Kane1*, Peace Nwankwo1*, and Catherine Freeland1

136 Analysis of Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities Laura A. Christman1*, Rachel A. Simon1*, Amy Hoang1*, Kayla M. Day1*, Kelley Chan1†, Jesse T. Clayton1†, and William A. McLaughlin1

143 Colonic Lavage in Treatment of Refractory Clostridium difficile Infection Adaptation of the Pittsburgh Protocol Maura Morgan1†, Timothy Farrell1†, Gordian U. Ndubizu2, and Timothy J. Farrell3

145 Medical Research Honors Program

Regan A. Gallagher1*, Brooke A. Goldstein1*, Jonah A. Joffe1*, Kyle A. Kidd1*, and SooYoung H. VanDeMark1*

Scholarly Research In Progress

A message from the editor-in-chief As the Journal of Scholarly Research in Progress (SCRIP) enters its fourth year of publication, I would like to offer thanks to our readers, our contributors, our faculty reviewers, and our student editors for their continued support of the journal and its mission: to promote and disseminate student scholarly activity at Geisinger Commonwealth School of Medicine. In the last four years, SCRIP has seen continual growth in the form of submissions and accepted published works. This year was no different, as we received well over 40 quality submissions from our students which included literature reviews, case reports, and original research manuscripts on topics ranging from oxalate nephropathy after bariatric surgery to investigating trends in elder-elder caregiving throughout the United States. In addition, this year’s cover features the beautiful artwork of Leana Pande, a student of Geisinger Commonwealth’s Master of Biomedical Sciences Program. I am heartened by the level of student interest in disseminating their research and scholarly work. In large part, this pursuit is due to the commitment of faculty throughout Geisinger who serve as mentors to our students and who recognize the importance of research and scholarship for their education and career development. Those submitting authors who have had their work accepted should be proud of their achievement. As always, I take the responsibility of sustaining and building upon the SCRIP’s quality and success very seriously. Suggestions from our contributors and readers to further develop and/or improve the journal are more than welcome; if you would like to share your thoughts, please email me at Lastly, I would like to invite students who have an interest in being involved in the editorial work of the journal. The role of the student editor is to assist in the writing, peer review, publication, and editing of content for the journal. It is a great way to build your academic scholarship portfolio, and I believe that this would help to ensure the journal’s growth and sustainability. Potential candidates may send their updated CV (that includes all relevant research and/or creative scholarship experience; all relevant writing, editing, or peer critique experience) to with the subject "Application for Student Editor."

Student editors Julia Schroer, MD Class of 2023 Christian Pardo, MBS Class of 2020

Acknowledgments The SCRIP would not be possible without the contributions of faculty and student volunteers committed to the review and assessment of submitted articles. Their feedback provides student authors with an opportunity to strengthen their writing and to respond to critiques. We gratefully acknowledge the following faculty members for their support in providing peer review. Kathleen Doane, PhD Eric Wright, PharmD, MPH Kimberly Miller, PharmD Brian Wilcox, MD, PhD Gabi Waite, PhD Vicki T. Sapp, PhD Michael Gionfriddo, PharmD, PhD Elizabeth Kuchinski, MPH John A. Arnott, PhD Mushfiq Tarafder, MPH, MBBS Brian Piper, PhD Cyamatare Felix Rwabukwisi, MD, MPH Christian Carbe, PhD Youssef Soliman, MD, PhD William McLaughlin, PhD Cathy Wilcox, PhD Youngjin Cho, PhD Carmine Cerra, MD Thomas M. Churilla, MD Jacob Parrick, MD Michael Sulzinski, PhD Jennifer K. Wagner, JD, PhD Angela Slampak-Cindric, PharmD, BCPS, BCCCP

Office of Research & Scholarship MSB, Suite 2024, Second Floor West 570-504-9662 Sonia Lobo, PhD, RYT Associate Dean for Research & Scholarship Professor of Biochemistry Michele Lemoncelli Administrative Assistant to the Associate Dean for Research & Scholarship Laura E. Mayeski MT(ASCP), MHA Manager, Research Compliance Thomas Majernick, MS Manager, Research Education Resources

Volume 4 • October 2020

Scholarly Research In Progress

Sincerely, Sonia Lobo, PhD Editor-in-Chief

On the cover: The cover image of the brain is done in watercolor by Leana Pande, a student in the Master of Biomedical Sciences Program.


Scholarly Research In Progress • Vol. 4, October 2020

Rise and Regional Variations in Schedule II Stimulant Use in the United States Sneha M. Vaddadi1†, Nicholas J. Czelatka1†, Belsy D. Gutierrez2, Bhumika C. Maddineni3, Carlos D. Torres-Teran4, Daniel N. Tron2, Kenneth L. McCall5, and Brian J. Piper1,6

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 The University of Scranton, Scranton, PA 18510 3 University of Texas Southwestern Medical Center, Dallas, TX 75390 4 Misericordia University, Dallas, PA 18612 5 University of New England, Portland, ME 04103 6 Center for Pharmacy Innovation and Outcomes, Geisinger Precision Health Center, Forty Fort, PA 18704 †Doctor of Medicine Program Correspondence: 1


Abstract Background: There is a need to better understand recent trends in stimulant usage. This report compares the pharmacoepidemiology of three Schedule II stimulants in the United States from 2010 through 2017. Methods: Drug weights were extracted from the Automated Reports and Consolidated Ordering Systems (ARCOS) for amphetamine, methylphenidate, and lisdexamfetamine. Total grams per drug were averaged across all states and compared from 2010 through 2017. Median stimulant daily dosage per patient user was determined from electronic medical records for a regional analysis. Results: There was a rise in amphetamine (+67.5%) and lisdexamfetamine (+76.7%) use from 2010 through 2017. The change in methylphenidate (-3.0%) was modest. Regional analysis indicated that persons/day usage of stimulants in the west was lower than that of other U.S. regions from 2014 through 2017. There was a negative correlation (r(48) = -0.43 to -0.65, p < .05) between the percent Hispanic population per state and the daily dose/population per stimulant. Conclusion: The increasing use of amphetamine and lisdexamfetamine, but not methylphenidate, may be explained by a rise in adult ADHD diagnoses and treatment. Regional analysis indicates that the use of stimulants in the west may be distinct from that in other regions. The lower stimulant use in areas with greater Hispanic population may reflect socioeconomic factors. Further research is needed on social factors impacting stimulant use and reasons for the pronounced regional variance.

Introduction Stimulants are used in treatment for Attention Deficit Hyperactivity Disorder (ADHD) — a disorder characterized by hyperactivity, inattention, executive function deficits, and emotional dysregulation (1, 2). ADHD is one of the more common biopsychosocial disorders, with a growth of 3.4% per year from 1998 through 2008 and a prevalence of 11% in children in 2011 (3). Based on data from the National Survey of Children’s Health, the percent of medicated ADHD nationally in children age 4 to 17 rose from 4.8% to 6.1% from 2007


through 2011 (3). ADHD diagnosis and treatment extends beyond childhood. With a revised ADHD criterion released in 2013 in the DSM-5, ADHD diagnoses have been more inclusive of adolescents and adults. As of 2018, approximately 4% of U.S. adults are afflicted with ADHD and two-thirds of children continue to experience at least one ADHD symptom throughout their lives (4). Despite the growth in ADHD, there is variation within cultural communities as well. Despite being one of the largest ethnic minorities in the United States, Hispanic youths have less ADHD diagnoses and stimulant use (5, 6). The common ADHD stimulants include amphetamine, lisdexamfetamine, and methylphenidate. Stimulant use has been associated with mild adverse effects, such as appetite and sleep disturbances, that impact quality of life, but the long-term adverse effects of these substances are not well established (7). A meta-analysis of cross-sectional positron emission tomography investigations showed that long-term blockade of the dopamine transporter caused neuroadaptive striatal elevations in this protein (8). This finding was subsequently confirmed in a longitudinal report (9). Patients prescribed stimulants had a nine-fold elevated risk of developing basal ganglia and cerebellar disorders (10). Nonmedical use of stimulants is also appreciable. The Monitoring the Future survey of recreational drug use determined that 4.6% of 12th graders misused Adderall in 2018 (11). With increasing prevalence of medicated ADHD in children and adults, there is a greater need to understand the extent of stimulant use nationally. This report utilized the U.S. Drug Enforcement Administration’s Automated Reports and Consolidated Ordering Systems (ARCOS) comprehensive database to evaluate changes in use of amphetamine, methylphenidate, and lisdexamfetamine nationally from prior years to 2017. We extended upon past research by investigating the overall change of stimulant use from 2010 through 2017 (12). We then calculated daily dose values to investigate the change in use from 2016 and 2017. We also explored variations in use in the Hispanic population and geographical regions. Investigation of these databases will allow for a greater understanding of the most recent pattern of use in treatment of ADHD and recent fluctuations of use within the United States.

Rise and Regional Variations in Schedule II Stimulant Use in the United States

Methods Data sources Stimulant data were extracted from the DEA’s Automated Reports and Consolidated Ordering Systems (ARCOS). This national database contains a yearly updated report of retail drug distribution by zip code and by state submitted by manufacturers and distributors to the DEA. Extracted data included total grams of stimulant use per drug per state from 2010 to 2017. Three Schedule II stimulants were examined: amphetamine, methylphenidate and lisdexamfetamine. Methamphetamine was not examined here due to low values. This database has been frequently used in prior pharmacoepidemiology reports (12, 13). Our goal was to examine the change in stimulant use and change in number of patients utilizing Schedule II stimulants, but ARCOS data is limited to the total quantity of Schedule II stimulants distributed in a geographical location. To determine the median stimulant daily dosage per patient user, we used de-identified data from the electronic health record (EHR) of Geisinger, an integrated health delivery system in central and northeastern Pennsylvania. With Geisinger EHR data from 2018 (n=88,202), we were able to estimate the median doses as 20 mg/day/person for methylphenidate and amphetamine and 40 mg/day/person for lisdexamfetamine. These values were used for daily dose calculations, which was used for a 2016 to 2017 and a regional comparison. An analysis was completed with state-specific Hispanic population data. The percent Hispanic population per state was obtained from the demographic profiles from the Pew Research Center. Institutional Review Board approval was obtained from the University of New England and Geisinger.

Data analysis Analysis by stimulant weight was completed by averaging state data per year for each stimulant and comparing use from 2010 through 2017. ARCOS aggregate data was divided by median values for daily dosage (mg/person/day) per stimulant that was determined from Geisinger EMR data. These values were named “Persons/day” and were used for a regional analysis by averaging state data within each region (Midwest, Northeast, South, West) for all three stimulants and compared from 2014 through 2017. Data was deemed to be significant (p < 0.05) after an unpaired t-test. A population-corrected analysis termed “Daily Dose/Population” was done by further dividing the ARCOS mg/daily dose by population per state for each drug. The percent change in these values per state per drug from 2016 to 2017 were calculated and compared. Values were deemed significant as 1.96 standard deviations above and below the mean. For all above calculations, outliers were determined through a Grubbs analysis and significant values were excluded. A heat map was constructed with Excel using the sum of ARCOS mg/daily dose value for all three stimulants divided by population for 2017. A similar calculation was done for 2016 and the percent change values were also depicted. A linear regression analysis was done with daily dose/population values and Hispanic population data per state in 2016 and 2017. Variance was reported as the SEM.

Results Stimulant use by weight increased 67.5% for amphetamine and 76.7% for lisdexamfetamine from 2010 through 2017 on average across all 50 states. In contrast, methylphenidate use decreased slightly (3.0%). For amphetamine and lisdexamfetamine, there was a significant increase in the total stimulant use compared to 2010 starting from 2014 (Figure 1A).

Figure 1. (A) Weight per stimulant per state showed a 67.5% and 76.7% increase in amphetamine and lisdexamfetamine (*p < 0.05 in comparison to 2010) and a 3% decrease in methylphenidate. 3

Rise and Regional Variations in Schedule II Stimulant Use in the United States

Further investigation into the change from 2016 to 2017 was completed with a daily-dose and population-corrected analysis. The percent change in daily dose/population across 50 states (with New Mexico excluded as an outlier) was +4.6% for amphetamine, +2.3% for lisdexamfetamine, and -1.4% for methylphenidate (Figure 1B, C). The preponderance

(85.0%) of states increased their amphetamine and over two-thirds (72.0%) increased their lisdexamfetamine use. In contrast, 86.0% of states decreased their methylphenidate use. Wisconsin, South Dakota, and West Virginia were all significantly greater than the mean for amphetamine. Hawaii had a significantly greater value, while Nevada and South

Figure 1. (B) Average percent change of 50 states data in Daily Dose/Person from 2016 to 2017 for amphetamine (+4.6%), lisdexamfetamine (+2.3%), and methylphenidate (-1.4%). Lisdexamfetamine was significantly different from amphetamine (ap <0.05) and methylphenidate (mp < 0.05). (C) Heat map of United States depicting percent change in total stimulant daily dose/population per state from 2016 to 2017. (D) Percent change in daily dose and population adjusted analysis per state for amphetamine, methylphenidate, and lisdexamfetamine reveals 85.0% of states increased their amphetamine, 72.0% increased their lisdexamfetamine, and 86.0% states decreased their methylphenidate use. Significant states were marked if 1.96*SD greater or less than the mean for each stimulant. The percent change for New Mexico was excluded as an outlier in all graphs (-17.45% for amphetamine, -19.98% for methylphenidate, and -36.63% for lisdexamfetamine). 4

Rise and Regional Variations in Schedule II Stimulant Use in the United States

Dakota had lower values compared to the national average for methylphenidate. For lisdexamfetamine, South Dakota had a significantly greater value, while Wisconsin had a lower value relative to the mean (Figure 1D). The daily dose/population values for 2017 are depicted in the heat map in Figure 2A, indicating pronounced regional variance. There was a six-fold difference between values for the highest (12.2) and lowest (2.9) states. Six states with the lowest values of daily dose/population were all in the

western region of the United States. A regional analysis was completed to further investigate geographic-based variability. A comparison of persons/day per region from 2014-2017 revealed a significant difference in the values for the Midwest, Northeast, and South compared to the West consistently from 2014 through 2017 (Figure 2B). Analysis with the percent Hispanic population and Daily dose/population identified a negative correlation for each stimulant in 2016. These findings were replicated for 2017 (Figure 3). States with a greater portion of Hispanic populations used less stimulants.

Figure 2. (A) Heat map for Daily Dose/Population per state for 2017 (B) Person/Day per Region from 2014-2017 indicated a significant difference with the West compared to the South, Midwest and Northeast from 2014 through 2017 (*p < 0.05). Time points were slightly offset for display purposes.

Figure 3. (A) Negative correlation between total Daily Dose/Population value and the percent Hispanic per state for 2016. (amphetamine: r(48) = -0.43, p = 0.0017; methylphenidate: r(48) = -0.64, p < 0.0001; lisdexamfetamine: r(48) = -0.49, p < 0.0001). (B) Negative correlations between total Daily Dose/Population value and percent Hispanic per state for 2017 (amphetamine r(48) = -0.43, p = 0.0017; methylphenidate: r(48) = -0.65, p < 0.0001; lisdexamfetamine r(48) = -0.52, p < 0.0001). 5

Rise and Regional Variations in Schedule II Stimulant Use in the United States

Discussion This study investigated trends in stimulant use in the United States in 2017 relative to prior years. Our data was indicative of an overall increase of Schedule II stimulant use, which is consistent with accounts of rising number of ADHD diagnoses (1). An analysis for total grams from 2010 through 2017 and a daily dose and population-considered analysis from 2016 to 2017 indicated a rise in amphetamine and lisdexamfetamine and no appreciable change in methylphenidate. This pattern of change in stimulant use may be explained by a rise in adult ADHD diagnoses and treatment. The revision of ADHD criteria in the DSM-5 is more inclusive of adult ADHD and has led to adults meeting more requirements for diagnosis than that for the DSM-IV (16). A 2018 study looking at ADHD treatment in privately-insured women age 15 to 44 found a similar pattern in these three stimulants as this study, with the largest change in stimulant use being in the age range of 25 to 29 years (15). Lisdexamfetamine and mixed amphetamine salts were found to be effective in treating adult ADHD, causing a significant improvement in adult ADHD symptoms without symptoms rebound after ceasing medication (18, 24). In addition, despite being considered a first-line treatment for child and adolescent ADHD, the long-term (>12 months) efficacy of methylphenidate is not well established (19). There are also drug-drug interactions with many medications, including monoamine oxidase inhibitors, vasopressors, and anticoagulants (2, 23). Alternate uses for stimulants outside of AHDH treatment may also contribute to these patterns in stimulant use. Lisdexamfetamine is a well-tolerated treatment for moderate to severe binge eating disorder (19). Literature reviews have also explored the role of amphetamine and methylphenidate in treatment of apathy in Alzheimer’s patients and other neuropsychiatric conditions in the elderly (20). Further exploration is needed on how the expansion of stimulant use in other neuropsychiatric conditions or obesity impacts usage trends. The percent Hispanic population had a negative correlation with stimulant use per state for 2016 and 2017. Other studies have also indicated a lower stimulant use by Hispanic children compared to their non-Hispanic peers (14). Young Hispanic adults and children have a significantly lower use of outpatient mental health services for mental health and substance abuse care (21). The low rate of stimulant use among the Hispanic communities may be due to difficulties with access to health care. Prior to the implementation to the ACA in 2014, 30% of Hispanics reported no health insurance compared to 11% of non-Hispanic whites (6). Along with social factors such as language barriers, cultural factors such as a perceived difference in the need for outpatient mental health care may also explain differences in resource utilization (22). The cumulative effect of these factors may lead to individuals being unable or hesitant to seek medical attention for ADHD symptoms. Finally, our regional analysis with data controlled for daily dosage found that the West has a significantly lower schedule II stimulant use compared to the South, Northeast, and Midwest. This pattern was seen in other studies spanning from 1998 through 2018 focusing on both child and adult ADHD,


with the West having the lowest ADHD prevalence or change in stimulant use (12, 25, 26). Though widely noted, this pattern has little explanation and may be due to various factors. A 2015 report suggests that with many states of higher altitude located primarily in the West, the altitude may serve as a protective factor against ADHD by increasing dopamine levels (25). Cultural diversity may also play a role. In California, almost 40% of youths are Hispanic, an ethnicity that has significantly lower stimulant use (5). In conclusion, this report identified increases in use in amphetamine and lisdexamfetamine in the United States. Further investigation is needed to better understand the sociocultural or economic factors mediating the pronounced regional variance observed.

Acknowledgments The generation of heat maps was done with the help of Daniel Kaufman. This project was also completed with the technical assistance of Iris Johnston. NJC, BDG, CDTT, DNT, and BJP were supported by the Center of Excellence, Health Resources Services Administration (D34HP31025).

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer. The other authors have no conflicts of interest to declare.

References 1.

Bădescu GM, Fîlfan M, Sandu RE, Surugiu R, Ciobanu O, Popa-Wagner A. Molecular mechanisms underlying neurodevelopmental disorders, ADHD and autism. Rom J Morphol Embryol. 2016;57(2):361-6.

2. ADHD: Clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit/hyperactivity disorder in children and adolescents. Pediatrics. 2011;128(5):1007. 3. Visser SN, Danielson ML, Bitsko RH, Holbrook JR, Kogan MD, Ghandour RM, Perou R, Blumberg SJ. Trends in the parent-report of health care provider-diagnosed and medicated attention-deficit/hyperactivity disorder: United States, 2003–2011. J Am Acad Child Adoles Psychiatry. 2014;53(1):34-46. 4. Wei YJ, Zhu Y, Liu W, Bussing R, Winterstein AG. Prevalence of and factors associated with long-term concurrent use of stimulants and opioids among adults with Attention-Deficit/Hyperactivity disorder. JAMA Netw Open. 2018;1(4):e181152-. 5. Pennap D, Burcu M, Safer DJ, Zito JM. Hispanic residential isolation, ADHD diagnosis and stimulant treatment among Medicaid-insured youth. Ethn Dis. 2017;27(2):85. 6. Velasco-Mondragon E, Jimenez A, Palladino-Davis AG, Davis D, Escamilla-Cejudo JA. Hispanic health in the USA: a scoping review of the literature. Public Health Rev. 2016;37(1):31.

Rise and Regional Variations in Schedule II Stimulant Use in the United States


Groenman AP, Schweren LJ, Dietrich A, Hoekstra PJ. An update on the safety of psychostimulants for the treatment of attention-deficit/hyperactivity disorder. Expert Opin Drug Safety. 2017;16(4):455-64.

8. Fusar-Poli P, Rubia K, Rossi G, Sartori G, Balottin U. Striatal dopamine transporter alterations in ADHD: Pathophysiology or adaptation to psychostimulants? A meta-analysis. Am J Psychiatry. 2012 Mar;169(3):264-72. 9. Wang GJ, Volkow ND, Wigal T, Kollins SH, Newcorn JH, Telang F, Logan J, Jayne M, Wong CT, Han H, Fowler JS. Long-term stimulant treatment affects brain dopamine transporter level in patients with attention deficit hyperactive disorder. PloS One. 2013;8(5): e63023. 10. Curtin K, Fleckenstein AE, Keeshin BR, Yurgelun-Todd DA, Renshaw PF, Smith KR, Hanson GR. Increased risk of diseases of the basal ganglia and cerebellum in patients with a history of attention-deficit/hyperactivity disorder. Neuropsychopharmacology. 2018;43(13):2548-55. 11. Monitoring the Future 2018 Survey Results [Internet]. National Institute on Drug Abuse; National Institutes of Health; U.S. Department of Health and Human Services.; 2018 [cited 2020Mar15]. Available from: https:// infographics/monitoring-future-2018-survey-results 12. Piper BJ, Ogden CL, Simoyan OM, Chung DY, Caggiano JF, Nichols SD, McCall KL. Trends in use of prescription stimulants in the United States and Territories, 2006 to 2016. PloS One. 2018;13(11): e0206100. 13. Simpson KJ, Moran MT, Foster ML, Shah DT, Chung DY, Nichols SD, McCall KL, Piper BJ. Descriptive, observational study of pharmaceutical and nonpharmaceutical arrests, use, and overdoses in Maine. BMJ Ppen. 2019;9(4):e027117. 14. Davis DW, Feygin Y, Creel L, Williams PG, Lohr WD, Jones VF, Le J, Pasquenza N, Ghosal S, Jawad K, Yan X. Longitudinal trends in the diagnosis of Attention-Deficit/ Hyperactivity Disorder and stimulant use in preschool children on Medicaid. J Pediatr. 2019;207:185-91.

18. Stevens JR, Wilens TE, Stern TA. Using stimulants for attention-deficit/hyperactivity disorder: clinical approaches and challenges. Prim Care Companion CNS Disord. 2013;15(2) pii: PCC.12f01472. 19. Heo YA, Duggan ST. Lisdexamfetamine: A review in binge eating disorder. CNS Drugs. 2017;31(11):1015-22. 20. Dolder CR, Nicole Davis L, McKinsey J. Use of psychostimulants in patients with dementia. Ann Pharmacotherapy. 2010;44(10):1624-32. 21. Marrast L, Himmelstein DU, Woolhandler S. Racial and ethnic disparities in mental health care for children and young adults: A national study. Int J Health Serv. 2016;46(4):810-24. 22. Alegría M, Canino G, Ríos R, Vera M, Calderón J, Rusch D, Ortega AN. Mental health care for Latinos: Inequalities in use of specialty mental health services among Latinos, African Americans, and non-Latino Whites. Psychiatr Serv. 2002;53(12):1547-55. 23. Groenman AP, Schweren LJ, Dietrich A, Hoekstra PJ. An update on the safety of psychostimulants for the treatment of attention-deficit/hyperactivity disorder. Expert Opinion on Drug Safety. 2017 Apr 3;16(4):455-64. 24. Buoli M, Serati M, Cahn W. Alternative pharmacological strategies for adult ADHD treatment: a systematic review. Expert Review of Neurotherapeutics. 2016 Feb 1;16(2):13144. 25. Huber RS, Kim TS, Kim N, Kuykendall MD, Sherwood SN, Renshaw PF, Kondo DG. Association between altitude and regional variation of ADHD in youth. Journal of Attention Disorders. 2018 Dec;22(14):1299-306. 26. Xu G, Strathearn L, Liu B, Yang B, Bao W. Twenty-year trends in diagnosed attention-deficit/hyperactivity disorder among US children and adolescents, 1997-2016. JAMA network open. 2018 Aug 3;1(4):e181471-.

15. Anderson KN, Ailes EC, Danielson M, Lind JN, Farr SL, Broussard CS, Tinker SC. Attention-deficit/hyperactivity disorder medication prescription claims among privately insured women aged 15–44 years—United States, 2003–2015. Morbidity and Mortality Weekly Report. 2018 Jan 19;67(2):66. 16. Epstein JN, Loren RE. Changes in the definition of ADHD in DSM-5: subtle but important. Neuropsychiatry. 2013;3(5):455. 17. Punja S, Shamseer L, Hartling L, Urichuk L, Vandermeer B, Nikles J, Vohra S. Amphetamines for attention deficit hyperactivity disorder (ADHD) in children and adolescents. Cochrane Database Syst Rev. 2016(2): CD009996.


Scholarly Research In Progress • Vol. 4, October 2020

Hip Dysplasia and Osteogenesis Imperfecta: Case Review and Management Recommendations Mark Mandel1†, Kaitlin Saloky1†, William Mirenda2, Andrea Seeley2, and Mark Seeley2

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Geisinger Medical Center, Danville, PA 17822 †Doctor of Medicine Program Correspondence: 1


Abstract A 1-week old female presented with bilateral dislocated hips and was subsequently treated in a Pavlik harness. Harness treatment failed, requiring a closed reduction and spica cast application. In the Post-anesthesia Care Unit, the patient was found to have a right humerus fracture. Six weeks after cast application the patient was found to have nondisplaced bilateral femur fractures, prompting a genetics evaluation. The patient was subsequently found to have osteogenesis imperfecta Type 3. Perioperative fractures in pediatric patients should raise suspicion for osteogenesis imperfecta. Early diagnosis can improve the management of hip dysplasia and allow for early bisphosphonate therapy.

Introduction Osteogenesis imperfecta (OI) is a group of inherited disorders characterized by brittle bones. The disorder can manifest as easily fractured bones, hearing loss, a curved spine, brittle teeth, or weak muscles (1). The severity of symptoms ranges from mild to severe, and the combination of symptoms can vary. This disorder is commonly associated with pathogenic variants in type 1 collagen that lead to decreased bone quality and quantity (2). The most common pathogenic variants are located in either the COL1A1 or the COL1A2 gene, which encode for the alpha chains of collagen type 1 (3). Hip dysplasia in infants affects 1–3% of newborns (4). Factors such as a positive family history, perinatal breech positioning, and ligamentous laxity put a newborn at increased risk for developing hip dysplasia (4, 5). Testing infants with known OI for hip dysplasia can pose challenges due to the concern for fractures while performing the Barlow or Ortolani maneuvers, or the false negatives that occur with pre-existing femoral deformities (6). This places patients at risk for a delay in diagnosis and future care (7). In subtler OI phenotypes, the diagnosis may not be established at the time of hip dysplasia treatment as management begins early in infancy. Due to the low frequency of the condition, treatments are poorly described in literature and are often performed under the surgeon's discretion. The purpose of this case report is to describe the presentation and treatment plan of a child born with OI and bilateral hip dysplasia. The patient’s family was informed that data concerning her case would be submitted for publication, and they provided consent.

Case Presentation A female child presented at 1 week of age with bilateral dislocated hips. She was the product of an uncomplicated,


term, vaginal delivery. She was the third of four children in the family. All other children were healthy. The patient’s initial evaluation demonstrated a right hip dislocation (positive Ortolani) and left hip subluxation (positive Barlow) on exam. She had symmetric hip abduction with a combined abduction angle of 150 degrees with an internal rotation of 60 degrees, and external rotation of 50 degrees. The patient was initially treated with a Pavlik harness. At 5 weeks old the patient had an ultrasound that demonstrated bilateral unstable dislocated hips. At 8 weeks of age, the patient was taken to the operating room for an arthrogram and closed reduction (Figures 1A-C). There were no intraoperative complications; however, in the Post-anesthesia Care Unit (PACU) she was found to have a right humerus fracture (Figure 2A). This was determined to be secondary to positioning at the time of spica cast placement. She was placed into a sling for two-and-a-half weeks (Figure 2B). Per routine protocol, the patient was taken back to the operating room at 6 weeks after the initial cast placement for bilateral hip examination and application of a new spica cast. This allowed for a skin check and an assessment of hip stability. A



Figure 1. A) Intraoperative arthrogram at 8 weeks demonstrating displacement of both hips. B) Intraoperative bilateral hip arthrograms showing the reduction of the femoral heads into the acetabulum. C) This post-operative X-ray demonstrates the successful relocation of both hips.

Hip Dysplasia and Osteogenesis Imperfecta


Figure 3. Bilateral hip arthrograms taken at 6 weeks after the first application of the first spica cast demonstrating bilateral congruent hips (rose thorn sign; yellow arrows).


Figure 2. A) This X-ray of the patient at 8 weeks old shows the humerus fracture that was discovered while the patient was in the PACU. B) This X-ray was taken 2 weeks after the identification of the humerus fracture and shows callus formation and periosteal reaction.

Figure 4. AP Pelvis X-ray conducted after the application of the second spica cast demonstrating bilateral congruent hips and a right proximal femoral deformity (yellow arrow).

The patient’s hips were found to be reduced and in the safe zone of 60 degrees (Figure 3). Due to scheduling conflicts, the second cast was removed at 5 weeks instead of 6 weeks. At that time, the patient had concentrically reduced hips with good safe zones. Intraoperatively she was found to have periosteal reaction in both femurs (Figure 4). Rather than placing her in another spica cast, it was decided to place the patient in a Rhino abduction brace for nights only. Approximately 2 months after her second spica removal, she sustained a right proximal femur fracture while being held by her mother and was treated with another spica cast for 6 weeks (Figure 5). The patient was referred to genetics and subsequently diagnosed with OI Type III. At 10 months of age she was started on pamidronate therapy that was continued for 15 cycles over a several-year period. Despite this treatment, she experienced several fractures in her upper and lower extremity that were treated with non-operative management. Follow-up examination over the ensuing 9 years revealed continued acetabular development and no signs of dysplasia (Figure 6 and 7).

Figure 5. This intraoperative arthrogram was taken at the time of the patient’s second spica removal when the decision was made to transition to a Rhino brace. The right hip remains located in this image and periosteal reaction in the femur is noted (rose thorn sign; yellow arrow).

Discussion The first case report describing the association of hip dysplasia and OI was in 1969. Unfortunately, since then, there has been sparse investigation on the topic, with the majority being limited to case reports/series (8). The initial paper reported on two cases of congenital hip dysplasia in male siblings with osteogenesis imperfecta and noted that their mother had blue sclera and joint hypermobility without a diagnosis of OI. Importantly, the paper discussed joint hypermobility as a common etiological factor in hip dysplasia and demonstrated that the mother and both of the patients had joint laxity (8). Fassier et al. retrospectively reviewed 687 patients with OI, 8 hips in 5 patients were diagnosed with hip dysplasia. In 4 of the 5 patients (80%), there was an association with collagen type-1 C pro-peptide variant and the development of hip dysplasia (8, 9). In the same study it was observed that these patients received delayed diagnoses of developmental dysplasia of the hip (DDH) and delayed treatment potentially due to the fear of causing fracture during clinical testing or the difficulty of performing the examination due to hyperlaxity (9). The results of this study also indicated that the incidence of DDH in patients with OI is higher than in the general population (9).

Figure 7. Further follow-up imaging at 8 years of age demonstrating normalized acetabular indices and no significant proximal femoral deformities.

Figure 6. This X-ray of the patient at 4 years old shows that her hips are located and maturing normally.


Hip Dysplasia and Osteogenesis Imperfecta

Establishing the diagnosis of DDH in a patient with OI poses additional challenges and risks, as conducting the Barlow and Ortolani screening maneuvers on these patients has been shown to cause femur fractures (6). With regard to our patient’s case, it is possible that the femur fracture identified intraoperatively during the second spica cast application may have been caused by these screening maneuvers or manipulation of the child’s legs in the operating room, further supporting the previous research showing that physical exam maneuvers may place these children at risk. For this reason, it is suggested that any patient with a family history of OI or clinical features suggestive of OI should undergo alternative methods for evaluation of DDH, such as diagnostic imaging with X-ray or ultrasound (6). Pavlik harness treatment in the management of DDH has an approximately 70% success rate in young infants (10). In the largest series published, Fassier et al. used a Pavlik harness in two patients (three hips) with hip dysplasia and OI. This treatment was ineffective in both patients and resulted in avascular necrosis in one hip. In our case the Pavlik harness also failed to stabilize the patient’s hips, thus requiring closed reduction and spica cast application (11). In addition to the known complications of Pavlik harness treatment, there are also several risk factors that are associated with Pavlik harness treatment failure, which include the severity of the dislocation determined by imaging and age greater than 3 months at the start of treatment (12). However, more likely in our case, the failure of the Pavlik harness could be due to ligamentous laxity or the presence of fractures. This theory is based on studies of children with Marfan’s syndrome, in which the researchers found that their generalized joint laxity lead to the failure of Pavlik harness treatment (13). Although not specifically documented in our patient’s case, joint hypermobility due to ligamentous laxity is a common feature among several variants of OI. Similarly, Kishta et al. recognized that their patients with OI and DDH also had joint hyperlaxity (9, 13). Literature surrounding operative management of hip dysplasia in OI shows that treatment goals may differ between age groups. In regard to the older age groups (18 months and older), treatment is generally focused on correcting femoral deformities (9, 14). However, review of the case series of patients who were treated surgically for hip dysplasia in the newborn age group (under 18 months) showed that half were managed with closed reduction and half had an open reduction with or without pelvic osteotomies (9). Reduction was successful after surgical treatment in all but one of these 8 hips (five patients). The open reduction was indicated in children who were older than 18 months. A pelvic osteotomy was necessary in one patient due to significant acetabular dysplasia (9). Our patient was treated with closed reduction and spica cast application, which successfully reduced her hips. In children younger than 18 months without significant acetabular dysplasia, a closed reduction is a successful treatment option for hip dysplasia in OI patients.

Conclusion Perioperative fractures in pediatric patients, specifically those that occur from positioning in the operating room, should raise suspicion for OI. Similarly, femur fractures that occur 10

while undergoing Pavlik harness treatment are very rare and patients should be referred to genetics for further workup. Prompt diagnosis will help guide management of the hip dysplasia and allow for early discussion of bisphosphonate intervention.

References 1.

Devogelaer J-P, Coppin C. Osteogenesis Imperfecta. Treatments in Endocrinology. 2006;5(4):229–42.

2. Forlino A, Cabral WA, Barnes AM, Marini JC. New perspectives on osteogenesis imperfecta. Nature reviews Endocrinology. 2011;7(9):540–57. 3. Sillence DO, Senn A, Danks DM. Genetic heterogeneity in osteogenesis imperfecta. Journal of Medical Genetics. 1979;16(2):101–16. 4. Agarwal A, Gupta N. Risk factors and diagnosis of developmental dysplasia of hip in children. Journal of Clinical Orthopaedics and Trauma. 2012;3(1):10–4. 5. Wilkinson J. Persistent Joint Laxity of and the Congenital. The Journal of Bone and Joint Surgery. 1964;46-B(1):40–5. 6. Paterson CR, Beal RJ, Dent JA. Osteogenesis imperfecta: fractures of the femur when testing for congenital dislocation of the hip. BMJ. 2009;305(6851):464–6. 7.

Shorter, D., Hong, T., Osborn, D. A. Screening programmes for developmental dysplasia of the hip in newborn infants. Cochrane Database of Systematic Reviews. 2013;54(9):11-54.

8. du Toit S, Weiss C. Congenital dislocation of hips associated with osteogenesis imperfecta in male siblings. A case report. Bull Hosp Joint Dis. 1969;30(2):164–70. 9. Kishta W, Abduljabbar FH, Gdalevitch M, Rauch F, Hamdy R, Fassier F. Hip Dysplasia in Children With Osteogenesis Imperfecta. Journal of Pediatric Orthopaedics. 2017;37(7):479–83. 10. Ömeroğlu H, Köse N, Akceylan A. Success of Pavlik Harness Treatment Decreases in Patients ≥ 4 Months and in Ultrasonographically Dislocated Hips in Developmental Dysplasia of the Hip. Clin Orthop Relat Res. 2016;474(5):1146-1152. 11. Atalar H, Sayli U, Yavuz OY, Uraş I, Dogruel H. Indicators of successful use of the Pavlik harness in infants with developmental dysplasia of the hip. International Orthopaedics (SICOT). 2007 Mar 28;31(2):145–50. 12. Cashman J.P, Round J, Taylor G. Clarke, N.M.P. The natural history of developmental dysplasia of the hip after early supervised treatment in the Pavlik harness: a prospective, longitudinal follow-up. The Journal of Bone and Joint Surgery. 2002; 84 (3): 418-425 13. Kerrigan A, Ayeni OR, Kishta W. Developmental Dysplasia of the Hip in Patient with Connective Tissue Disorders. JBJS Reviews. 2019; 7(4):e5. 14. Gillingham BL, Sanchez AA, Wenger DR. Pelvic osteotomies for the treatment of hip dysplasia in children and young adults. JAAOS-Journal of the American Academy of Orthopaedic Surgeons. 1999 Sep 1;7(5):32537.

Scholarly Research In Progress • Vol. 4, October 2020

Adult Vasoactive-Inotropic Score Predicts Mortality in Adult Patients with Shock Alex McGeough1†, Timothy Farrell1†, Navindra Tajeshwar1†, Katarina Smigoc1†, and Niraj J. Vyas1†

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 †Doctor of Medicine Program Correspondence: 1

Abstract Background: Several mortality-predicting scores for hypotensive patients have been developed. However, they have not been thoroughly evaluated in the adult population. Currently, the best-known score for similar purposes is the vasoactive inotropic score (VIS), which is rarely used in clinical practice and was only designed for pediatric patients following cardiac surgery. There are no mortality prediction scores for circulatory shock currently in use for adult patients. Our aim is to develop and assess an effective mortality prediction score for circulatory shock to use in the adult population. Methods: We used the Multiparameter Intelligent Monitoring in Intensive Care III (MIMIC III) database to retrospectively evaluate data for patients admitted to an intensive care unit with a diagnosis of shock undergoing treatment with vasopressor or inotropic agents. We developed an approximated mortality score using expert opinion and derivation of previously published shock mortality prediction data. The mortality score was calculated for each patient in the last 24 hours and then evaluated for the primary outcome of mortality during their stay. Patients were divided into two groups based on mortality, and a binary linear regression was performed. Results: Mortality score calculated for 720 patients using linear analysis was statistically significant at a p-value of 1.83 x 10^-12. Conclusion: This score assists in the evaluation of complex circulatory shock patients undergoing treatment using a variety of vasopressor and/or inotropic agents. The primary implication in clinical practice is to help guide resource allocation to the patients who are most in need based on a predictive model of prognosis. This has an advantage over the current VIS of no longer being limited to pediatrics. The ability for clinicians to rapidly assess the condition of patients in shock will allow for more effective care administration to the patients in the most emergent condition.

Introduction Patients admitted to the intensive care unit (ICU) are either in critical condition or suffer acute disturbances resulting from surgery (1). Several postoperative diagnostic scoring systems have been developed to predict the clinical course of these patients and the necessary pharmacologic intervention. A scoring system tested by Wernovsky et al. used an inotropic scoring system to measure the needed pharmacologic intervention in post-heart-surgery neonates and infants (2). Although designed to apply to neonates/infants, the Wernovsky scoring system has been used in clinical practice in pediatric populations (3). The implementation of a clinically

significant predictor scoring system could assist physicians and staff in directing patient care in ways that could improve outcomes. The unknown efficacy of an inotropic scoring system in pediatric populations was explored by Gais et al. by the creation of the vasoactive inotropic score (VIS) (3). The VIS scoring system expanded upon the Wernovsky score by incorporating medications commonly used in clinical practice with the finding that a maximum VIS score in the first 24 hours had a significant correlation with increased morbidity and mortality. These findings were elucidated in a previous study by Gais on the association of the VIS scoring system with the clinical outcomes in children <6 months of age undergoing cardiac surgery with cardiopulmonary bypass using a multicenter analysis of data reported to the Pediatric Cardiac Critical Care Consortium (4). The results of the previous studies indicate that the VIS scoring system is an effective predictor of clinical outcome in pediatric populations. Although the VIS scoring system is currently the best-known inotrope scoring system available, its applicability has been limited to pediatric patients following cardiac surgery and has not been thoroughly evaluated in adult patients. A study by Yamazaki et al. assessed VIS as a predictor of morbidity and mortality in adult cardiac surgery (5). They concluded that a high VIS score was associated with a poor outcome, including longer ICU stay and ventilation time. The study was, however, limited to patients who underwent cardiopulmonary bypass surgery. Na et al. also investigated VIS in nonsurgical cardiogenic shock patients and found an association between high levels of vasoactive inotropic support and in-hospital mortality (6). While the use of the VIS scoring system looks promising in adults, further evaluation is needed. This study aims to assess the association between the VIS scores and clinical outcomes in adult ICU patients undergoing both surgical and non-surgical hypotensive shock.

Methods Study design and setting This was a retrospective data analysis using data gathered through the Multiparameter Intelligent Monitoring in Intensive Care III (MIMIC III) database of adult patients admitted to the ICU at Beth Israel Deaconess Medical Center in Boston, MA. Data from MIMIC III was stored and extracted using PostgreSQL (7). An approximated mortality score was made using expert opinion and derivation of previously published shock mortality scores. Patients were excluded from the analysis if one of the following criteria were met: 1) Mean arterial pressure (MAP) was not recorded; 2) Vasopressors were not used in the treatment of the patient; 3) Patient was


Adult Vasoactive-Inotropic Score Predicts Mortality in Adult Patients with Shock

not treated in the ICU. After filtering the patient data based on these criteria, the data of 720 patients was used. Clinical endpoints and outcome variables The clinical endpoint was defined as patient mortality which marked the point at which data collection for that patient was terminated. Statistical analysis Shock score was calculated using Equation 1 shown below. In order to use the equation, a recording of MAP was required. Prior to the analysis, if MAP recordings were not taken, the entirety of the data set was removed. Shock predictor scores were divided into quartiles and patient mortality per quartile was measured. Using R (v3.5.2, The R Foundation)8 a Pearson's chi-square analysis was completed on the cumulative average shock predictor score per patient, as well as the average shock predictor score in the last 12, and 24 hours in their respective tables. The Pearson’s Chi-squared test was used to determine if a correlation existed between shock predictor score in these three groups (cumulative average, last 12 hours, and last 24 hours) and patient mortality. A logistic binomial regression was then performed on these values to define the relationship between shock predictor score and patient mortality. A logistic binomial regression was completed on the cumulative average shock predictor score per patient. The shock scores of the last 12 hours per patient were averaged and then a logistic binomial regression was run to determine if there was an association between the average

Figure 1: Frequency of AVIS scores from included patients 12

scores in the last 12 hours and mortality. This process was then repeated for the last 24 hours as well. Equation 1: 20*(rate at which norepinephrine was given in mic/min + the rate at which epinephrine was given in mic/min + 2*(MAP - 65))

Results The distribution of the AVIS score is approximately uniform (Figure 1); however, as this analysis is the first to be conducted on this assessment tool, the data was divided into equal quartiles in order to reduce bias and the complex intricacies of the data (Table 1). A chi-square test of homogeneity was performed at Îą = 0.05, which indicated a significant difference in the rate of patient mortality among the 4 quartiles (p < 0.001). The results indicate lower AVIS scores are associated with a higher rate of patient mortality, specifically in the lowest quartile which demonstrated an approximate 23% mortality

Table 1: AVIS score quartiles compared with patient mortality

Adult Vasoactive-Inotropic Score Predicts Mortality in Adult Patients with Shock

among the cohort. To capture the extent of this association within the continuity of the AVIS score, an odds ratio analysis with a 95% confidence interval indicated a 5% reduction in the odds of mortality as the AVIS score increased by one point (OR = 0.949).

Discussion The goal of this study was to assess the utility of employing a modified VIS in order to determine an increased risk of mortality in the setting of an ICU. Using the framework of a previously developed score to evaluate the need for pharmacologic intervention in pediatric postoperative cardiac cases, we applied a similar set of parameters to quantify the risk of mortality in shock patients. Using the AVIS equation, AVIS = norepinephrine mic/min + epinephrine mic/min + (MAP 65) x 2, we retrospectively measured the use of these guidelines with lower scores indicating a worse prognosis. With this methodology, patients with a lower AVIS score had a six-fold higher likelihood of death within the first 24 hours of admission to the ICU with shock. This demonstrates that this score can aid physicians in rapidly considering the needs of a patient in the ICU or their probability of dying once admitted. A review of current critical care literature illustrates a large gap in objective scoring systems to assess the risk of mortality in patients admitted to the ICU. The transparency and comprehensive documentation permitted with electronic medical records (EMR) will likely drive the advancement of further scoring systems like AVIS. Current clinical assessment tools include APACHE (Acute Physiology and Chronic Health Evaluation), MPM (mortality probability model), and SAPS (simplified acute physiology score) (9). While these tools are useful guidelines, they provide only general information regarding the length of stay and risk of mortality and subtleties between patient populations and data lead to different efficacy in the clinical space. It is also documented that using scoring systems improves the efficiency of clinical practice. ICUs are inundated with increasing patient volumes and limited resources. The United States spends approximately $80 billion per year in the ICU alone. With increasing demands, intensivists are pressed to make difficult decisions in allocating care. Utilizing a scoring system can make difficult subjective decisions on care plans objective and efficient. Studies have demonstrated that employing scoring systems such as APACHE improve the efficiency of a hospital setting (10). However, comprehensive meta-analysis studies assessing the accuracy and usefulness of these scoring systems or creating new ones are limited. There are data from Castella et al. suggesting that newer parameters being assessed in the APACHE, SAPS, and MPM scoring systems leads to improved discrimination inadequately assessing a patient’s risk of mortality within the first 24 hours of a hospital stay (11). The lack of updated data over the past 20 years regarding this facet of intensive care emphasizes the need to create new scoring systems, and comprehensive studies need to be investigated to improve the care of our ICU patients.

Furthermore, none of these studies directly assess the risk of mortality in patients experiencing shock. Septic shock is the admitting diagnosis in 10% of patients to the ICU with an additional 8% developing septic shock during their stay (11). David Berg et al. demonstrated that of the admissions to the cardiac ICU, 22% of patients met criteria for shock (12). These figures have increased dramatically over the past 20 years, showing that rapid assessment of a patient’s status will be crucial now and in the future. There is very little data regarding objective assessment of the severity of a patient’s condition regarding shock. This illustrates the need for further studies like ours to assess the value of scoring systems and to begin implementation so as to better care for critically ill patients. This study was intended to be a preliminary investigation into the application of the AVIS score for adult patients. As this is a preliminary study applied retrospectively to patients, there is a need to evaluate these data further in a clinical study employed in an ICU setting. Future developments of this study also necessitate establishing stricter scoring boundaries to apply to AVIS system. It is important to determine clinically significant restrictions in using these scoring systems to refine and maximize their use in the hospital. As in the study, the cutoff guidelines were made arbitrarily. Furthermore, shock is often one problem amid a storm of medical problems that plague patients upon entry into the ICU. While the AVIS yielded promising data regarding its value in assessing the risk of mortality in shock patients, it does not factor in other medical maladies or treatment plans being employed on the same patients. Therefore, further studies must be conducted to evaluate the utility of the AVIS scoring system when a patient is experiencing multiple comorbid conditions.

Conclusion In this study, we investigated the use of a previously developed scoring system to assess the mortality of pediatric patients after cardiac surgery. Using a modified version of this tool, AVIS, we retrospectively applied this formulation and scoring system to ICU patients in order to discover if lower scores were indicative of a higher risk of mortality. Our data demonstrated that lower AVIS scores inversely correlated with a higher risk of mortality in admitted ICU patients. This study illustrates that future studies such as these can assist clinicians in rapidly and objectively assessing the severity of a patient’s condition and allocated resources accordingly or adjust treatments as indicated.

Acknowledgments All data were collected by Khaled Sorour, MD, with affiliations with Harvard Medical School and Deaconess Medical Center.

References 1.

Pittet D, Rangel-Frausto S, Li N, Tarara D, Costigan M, Rempe L, Jebson P, Wenzel RP. Systemic inflammatory response syndrome, sepsis, severe sepsis and septic shock: incidence, morbidities and outcomes in surgical ICU patients. Intensive Care Med. 1995; 21(4): 302-309.


Adult Vasoactive-Inotropic Score Predicts Mortality in Adult Patients with Shock

2. Wernovsky G, Wypij D, Jonas RA, Mayer JE Jr., Hanley FL, Hickey PR, Walsh AZ, Chang AC, Castaneda AR, Newburger JW, et al. Postoperative course and hemodynamic profile after the arterial switch operation in neonates and infants. A comparison of low-flow cardiopulmonary bypass and circulatory arrest. Circulation. 1995; 92(8): 2226-2235. 3. Gaies MG, Jeffries HW, Niebler RA, Pasquali SK, Donohue JE, Yu S, Gall C, Rice TB, Thiagarajan RR. VasoactiveInotropic Score (VIS) is associated with outcome after infant cardiac surgery: an analysis from the pediatric cardiac critical care consortium (PC4) and virtual PICU system registries. Pediatr Crit Care Med. 2014; 15(6): 529537. 4. Gaies MG, Gurney JG, Yen AH, Napoli ML, Gajarski RJ, Ohye RG, Charpie JR, Hirsch JC. Vasoactive-inotropic score as a predictor of morbidity and mortality in infants after cardiopulmonary bypass. Pediatr Crit Care Med. 2010; 11(2): 234-238. 5. Yamazaki Y, Oba K, Matsui Y, Morimoto Y. Vasoactiveinotropic score as a predictor of morbidity and mortality in adults after cardiac surgery with cardiopulmonary bypass. J Anesth. 2018; 32(2): 167-173. 6. Na SJ, Chung CR, Cho YH, Suh GY, Ahn JH, Carriere KC, Park TK, Lee GY, Lee JM, Song YB, et al. Vasoactive inotropic score as a predictor of mortality in adult patients with cardiogenic shock: medical therapy versus ECMO. Rev Esp Cardiol. 2018. 7.

MIMIC-III, a freely accessible critical care database. Johnson AEW, Pollard TJ, Shen L, Lehman L, Feng M, Ghassemi M, Moody B, Szolovits P, Celi LA, and Mark RG. Scientific Data (2016). DOI: 10.1038/sdata.2016.35. Available at:

8. R, A Language and Environment for Statistical Computing. R Development Core Team. 2019. ISBN: 3-900051-07-0. Version 3.5.2. Available at: 9. Balkan B, Essay P, Subbian V. (2018). Evaluating ICU Clinical Severity Scoring Systems and Machine Learning Applications: APACHE IV/IVa Case Study. 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC). doi: 10.1109/ embc.2018.8513324 10. Castella X, Artigas A, Bion J, Kari A. (1995). A comparison of severity of illness scoring systems for intensive care unit patients. Critical Care Medicine, 23(8), 1327–1335. doi: 10.1097/00003246-199508000-00005 11. Vincent J-L, Jones G, David S, Olariu E, Cadwell KK. (2019). Frequency and mortality of septic shock in Europe and North America: a systematic review and metaanalysis. Critical Care, 23(1). doi: 10.1186/s13054-019-24786 12. Berg DD. (2019). Epidemiology of Shock in Contemporary Cardiac Intensive Care Units Data From the Critical Care Cardiology Trials Network Registry. American Heart Association. doi: 10.1161/CIRCOUTCOMES.119.005618.


Scholarly Research In Progress • Vol. 4, October 2020

Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States Kathy-Ann Cadet1*, Jasmine Nkwocha1*, and Myiah Smothers1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract In the United States, racial disparities in health care are evident throughout history. The rising of the opioid epidemic results in the misuse of prescription opioid medications causing addiction and even drug overdoses; a deeper look into the racial disparities in the opioid epidemic was of interest. The census was used to find the total population in all states in 2011 and the population of African Americans in these states. Automated Reports and Consolidated Ordering System (ARCOS) data from the Drug Enforcement Act (DEA) was used to find the amount of the top 5 opioids that were purchased by pharmacies in 2011. Heatmaps and correlational studies were then created to analyze the data. No significance was found in the linear correlation between percent African American population versus oxycodone weight per population. When comparing the percent African American heatmap with the oxycodone drug weight heatmap, there were significant differences in saturation found. Discrepancies were found when comparing the heatmaps, thus more detailed research needs to be done to determine why. Data analysis showed a nonsignificant data correlation, although the heatmaps and literature review proved otherwise.

Introduction In 2002, Congress requested that the Institute of Medicine (IOM) convene a committee to evaluate racial and ethnic disparities within the United States health care system (1). The committee disclosed that there were apparent, significant sources of racial and ethnic disparities in healthcare received by minorities (1–3). One of the committee’s indispensable findings was that when patients' economic status, insurance status, age, and health conditions were similar, racial bias and ethnic disparities persisted (1–3). Prior studies have shown that racial discrimination in healthcare began long before 2002 and has been occurring since African slaves were forced to leave their various countries and come to this country over 400 years ago (2). In Four Hundred Years Since Jamestown: An American Journal of Public Health (AJPH) Dossier, Theodore M. Brown explains to readers how the era of slavery had an impact on U.S. medicine and public health (4). Africans were believed to be of a different nature or species, they did not have the same human makeup as their Caucasian counterparts and were thought to be medically and biologically different (4). This theory was sustained by slave owners, physicians, and scientists to justify why the slaves were property and could be used as such (4). During the time of the yellow fever, John Lining, a physician from Charleston, SC, would insist that African Americans were immune to yellow fever because the mortality rate was thought to be low

(5–6). Because Lining was a credible physician, this provoked other physicians to not only support this claim but to motivate African Americans to help during the epidemic, assuring them that they could not contract the disease, which was completely false (5–6). These claims fueled racial disparities in the United States health care system, leaving African Americans susceptible to unethical experimentations and inadequate care (4, 7–9). Tests and experiments executed using African Americans were callous and brutal. An example of this is the Tuskegee Syphilis Study, where African American men were injected with syphilis and left untreated from 1932 to 1972, or the experimentation of Thomas Hamilton, a physician from Georgia, who would test medications for heatstroke by placing slaves in extremely hot temperatures until they collapsed and then proceed to test his remedies on the slaves (4, 7–9). As explicated by Rana Asali Hogarth (5) in the article From the 1793 Yellow Fever Epidemic in Philadelphia, Pennsylvania, it is critical to note that these ideas do not only emanate from the time of slavery and these ideas did not dissipate as the time went on. These assumptions and postulations are entrenched in the medical knowledge of health care professionals today (5). In previous studies, scholars have scrutinized the earlier events of racial bias and compared them to situations that transpired more recently (4). Their findings propose that there is profound, unconscious racial bias that accounts for the finding that medical students, residents, and practicing physicians believe that African Americans feel less pain than Caucasian people do and are less likely to receive the proper treatment or correct dosage of a drug, if necessary (4). Many studies examine how likely African American patients are to receive analgesics compared to their Caucasian counterparts. For example, a study analyzed whether African American patients with isolated long-bone fractures were less likely to receive any pain medication compared to Caucasian patients, and found that Caucasian patients were 17% more likely to receive the needed pain medications (9, 10, 14). In another study, dealing with opioid prescription to children, it was found that Caucasian children are more likely to receive opioid prescriptions when compared to minorities who are possibly comparable in terms of medical needs (11). Pain is a subjective phenomenon and there is no accurate way to assess a patient's pain level (12). Physicians are forced to rely on what the patient is telling them and what they can observe, thus leaving pain susceptible to social psychological influences such as explicit or implicit biases — and leaving African American patients susceptible to not getting proper treatment for their condition (12). Multiple studies analyzed whether these racial biases, if any, are explicit or implicit,


Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States

examining cognitive indications from medical students, residents, and laypeople (7, 12–13). Most of the articles found approached the topic of racial disparities in opioid prescription with medical professionals directly, creating tests to rate how they viewed pain on different subjects throughout the study. The studies mentioned support the idea that physicians are not prescribing African Americans the same amount of opioids they prescribe Caucasians because they believe that African Americans feel less pain, which leads to inadequate care for those African American patients. In our study, we wanted to examine the differential distribution of opioids for any health condition at the year opioid prescriptions reached their peak, which was in 2011. We wanted to evaluate which states had the most prescriptions of these drugs and compare it to the states that had the highest African-American populations. The goal of this investigation was to identify if the states with the highest African American populations had the lowest opioid prescriptions, supporting our claim that physicians prescribe African American individuals opioids less than they do Caucasians. Research engines such as PubMed helped to guide the framework for this research paper. Key words such as African Americans or African Americans and pain or opioid or prescription proved to yield the best results in terms of background information. A number of the articles found have previously investigated the racial bias associated with pain in different communities especially in African American communities. Mathur et al. explain that the racial differences seen in pain treatment are a direct result of physicians

providing less treatment to African Americans when compared to other races based on the notion that they have a higher pain tolerance (13). All of the articles helped to frame the purpose of this paper — were African Americans being prescribed the same amount of opioid drugs compared to other races during the peak opioid year of 2011?

Methods Procedures The total population of individuals was collected in every state including the District of Columbia in 2011 using the United States Census. The total number of African Americans was then looked at in each state in 2011 using the U.S. Census data. This number was then divided by the total population of all individuals to produce what we will refer to as our control data. These data would then be compared the top 5 drugs in 2011, including the peak drug during that year as oxycodone. The drug data was found using the Automated Reports and Consolidated Ordering System (ARCOS), which is a data collection system where pharmacies and hospitals and other distributors report the number of controlled substances they have to the Drug Enhancement Administration (DEA). The 2011 Retail Drug Summary ARCOS Report 5: Statistical Summary for Retail Drug Purchases was used. This project examined the distribution (in total grams) of oxycodone, buprenorphine, hydrocodone, and codeine across pharmacies in all of the United States and the District of Columbia. Lisdexamfetamine

Figure 1. (A) Heatmap of percent African American population in the United States. (B) Heatmap of oxycodone distribution in the United States. (C) Linear correlation for percent African American versus the oxycodone weight per person (r= (1,49), 0.7663).


Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States

was used as a non-opioid drug control. This number was recorded and then divided by the total population of individuals in the state. This data was then compiled into 5 different heatmaps, with oxycodone, buprenorphine, hydrocodone, and codeine all being compared to the control lisdexamfetamine. Data analysis A simple linear regression analysis is used to compare the relationship between the top 5 drugs to the control lisdexamfetamine using Prism 8 software database. This analysis helped to determine the correlation between percent African Americans in each state and the total grams of oxycodone, buprenorphine, hydrocodone, codeine, and lisdexamfetamine per person in each state. A simple linear regression analysis encompasses best fit values and confidence intervals for all states per drug used. Microsoft Excel was then used to create the heatmaps for each of the 5 drugs.

Results The peak year of opioid use was found to be 2011 (14). In addition to the District of Columbia, the states with the highest percent of African Americans in the year 2011 were Mississippi, Louisiana, Georgia, and Mississippi. The percent of African Americans heatmap Figure 1A was compared with the heat maps of drug weight per person of oxycodone, Figure 1B, along with buprenorphine, hydrocodone, codeine, and lisdexamfetamine. The oxycodone map showed lower opioid

distribution for the top percent population African American states of Louisiana, Mississippi, Alabama, and Georgia. A simple linear regression was calculated to predict oxycodone use based on African American population. No significant regression equation was found (F(1,49)= 0.1367, p= 0.7135), with an R2 of .002. Figure 1C. For the remaining opioids examined (buprenorphine, codeine, hydrocodone), no regression equation was found with the percentage of African American population (F(1,49)= 0.0002), p= 0.9881, with an R2 of 0.000, F(1,49)= 0.3714, p= 0.4214), with an R2 of 0.01324, F(1,49)= 0.1953, p= 0.4216), with an R2 of 0.01323). The heatmaps of buprenorphine, codeine, and hydrocodone varied in retail distribution throughout the United States, when compared to the percent African American heatmap, Figure 2. Codeine had the least retail distribution in Louisiana, Mississippi, Alabama, and Georgia. For the control drug lisdexamfetamine, a significant regression equation was found, (F(1,49)= 5.034, p= .0294), with an R2 of .09316, Figure 3A. Visual comparisons were made when observing the lisdexamfetamine and percent African-American heatmaps (Figure 3B and 3C). The percent African American heatmap was compared with a heatmap of percent African American uninsured in the United States, Figure 4.

Discussion Our data is inconsistent with multiple published reports showing significant racial disparities in opioid prescriptions.

Figure 2. (A) Heatmap of percent African American population in the United States. (B) Heatmap of codeine distribution in the United States. (C) Heatmap of hydrocodone distribution in the United States. (D) Heatmap of buprenorphine distribution in the United States. 17

Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States

Figure 3. (A) Linear correlation for percent African American versus the lisdexamfetamine distribution in the United States. (B) Heatmap of percent African American population in the United States. (C) Heatmap of lisdexamfetamine distribution in the United States.

Studies with these findings saw significant differences in opioid and non-opioid prescribing based on children’s race and ethnic status and significant racial and ethnic disparities in opioid prescription and administration at emergency department visits for back pain and abdominal pain (11, 15). Unlike our research, both studies used secondary analyses of data from Medical Expenditure Panel Surveys and the National Hospital Ambulatory Medical Care Survey (11, 15). This data provided information on what prescription drugs were given to patients, having the ability to show if African Americans were prescribed opioids (11, 15). Other studies focused on implicit and explicit biases, suggesting that known disparities in pain treatment could be caused by automatic rather than deliberate biases (12, 13). Conflicting with our study, these findings focused on the origin of racial biases within medical professionals. Our focus was to investigate whether medical professionals prescribe fewer opioids per capita to states with a larger population percentage of African Americans. When comparing the various opioids (oxycodone, buprenorphine, hydrocodone, and codeine) and the nonopioid control (lisdexamfetamine) heatmaps to the percent of African American population in 2011 heatmap, visual differences were apparent, warranting further analysis. Comparing Figure 1A to Figure 1B heat maps show a decrease in color saturation in Louisiana, Mississippi, Alabama, and Georgia — the states that had the highest percent African American population percentage. This meant that in the states with the highest percent of African American population, there was a decrease in oxycodone retail distribution. Yet, when the simple linear regression was run, significance was not found


Figure 4. Percent of the African American population that is uninsured in the United States; data from the U.S. Census.

(p= 0.7135). Interestingly, a significant regression equation was found between the percentage African American population and the control drug lisdexamfetamine, a drug that specifically treats attention-deficit/hyperactivity disorder (ADHD) and is used for binge-eating disorder symptoms. ADHD is classified by the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) as a disorder most common in children and adolescents who show the inability to remain attentive paired with hyperactivity (16). When discussing racial differences in ADHD diagnosis, Morgan et al. found that African American and Hispanic children were less likely to receive an ADHD diagnosis and treatment compared to

Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States

Caucasian children (17). Figure 3C is the heatmap of the ADHD control drug lisdexamfetamine. The most saturated state for lisdexamfetamine retail distribution, Louisiana, has also one of the highest percent African American populations, resulting in a significant linear correlation. The U.S. Centers for Disease Control and Prevention (CDC) found that 15.8% of children in Louisiana vs. the national level of 11% were diagnosed with ADHD in 2011, though when looking specifically at race, it was Caucasian males who received the highest amount of ADHD prescriptions (16). The most common age for prescriptions was received at 10 years old, with African American males receiving 21.8% of the prescriptions while Caucasian males received 35.8% (16). Kumar and Morgan relate these disparities to genetic, prenatal, and environmental factors that have longterm negative effects on child development (16).


Another opioid that showed a similar change in color saturation was codeine, Figure 2B. This figure shows the heatmap of codeine drug weight per state having a decrease in color in Louisiana, Mississippi, Alabama, and Georgia. Figure 2D (buprenorphine) shows gradual increase in color in states while Figure 2C (hydrocodone) is even more saturated in color in these same areas and other regions throughout the United States. The change in distribution in the states with the highest percent African American population could be due to differing levels of insurance coverage among the residents of these states, as shown in Figure 4. A study using data from the 2006–2013 National Health Interview Survey found that compared to whites, African Americans were more likely to choose to be without medical care; this trend increased posteconomic-recession between June 2010 and December 2013 (18). Since there are more African Americans who lack health insurance compared to other races, that could explain the vast differences in the heat maps. When looking into reasons why this disparity exists, Levesque et al. found that lack of access to health care was the main component in reasons why individuals may opt out (19).

We thank Dr. Brian Piper for his assistance throughout our research, specifically with our preparation and data analysis, Elizabeth Kuchinski for helpful feedback on this manuscript, and Iris Johnston for providing access to articles we have referenced.

When comparing these heatmaps, the focus states (Louisiana, Mississippi, Alabama, Georgia) were deeper in saturation representing more uninsured African Americans in Figure 4, as anticipated. Surprisingly, Montana and Wyoming have a large African American uninsured population. As previously stated, the studies that focused on the amount of opioids being distributed to patients falling into different racial or ethnic groups did find significance to support their claims (11, 15). In these studies, researchers used data from the National Hospital Ambulatory Medical Care Survey and the Medical Expenditure Panel Surveys (11, 15). These databases provided information on what prescription drugs were prescribed to each patient, showing the authors clearly if African Americans are getting opioid prescriptions or not. We did not use a survey database to give us those details explicitly; instead we gathered data regarding the total grams of an opioid in a specific state and compared it to the percent of African Americans in that state. Limitations for this study include the U.S. Census only reporting every 10 years. The peak year of oxycodone use was 2011, therefore the population data that we used was an estimation and is most likely underestimated.

In conclusion, though the simple linear regression done through Prism 8 yielded no significant results, our literature review and heat map comparisons demonstrated a different result. As mentioned earlier, a number of articles highlighted the differences seen in different races and how their physicians handled their pain level. Our next steps would be to look more closely at the insurance data and determine if there is a correlation between prescription of opioids and the lack of insurance in African American communities. A deeper look into Louisiana and the distribution of ADHD medication is also a potential area for more research.


References 1.

Nelson A. Unequal treatment: confronting racial and ethnic disparities in health care. J Natl Med Assoc. 2002; 94(8): 666–668.

2. Baicker K, Chandra A, Skinner JS. Geographic variation in health care and the problem of measuring racial disparities. Perspectives in Biology and Medicine. 2005; 48(1), S42-S53. 3. Baicker K, Chandra A, Skinner JS, Wennberg JE. Who you are and where you live: how race and geography affect the treatment of medicare beneficiaries. Health Aff (Millwood). 2004; 23(2). 4. Brown TM. Four hundred years since Jamestown: an AJPH dossier. Am J Public Health. 2019; 109(10),1309. 5. Hogarth RA. The myth of innate racial differences between white and black people's bodies: lessons from the 1793 yellow fever epidemic in Philadelphia, Pennsylvania. Am J Public Health. 2019; 109(10), 1339-1341. 6. Hogarth RA. Medicalizing blackness: making racial difference in the Atlantic world, 1780-1840. Chapel Hill: The University of North Carolina Press; 2017. 7.

Hoffman KM, Trawalter S, Axt JR, Oliver MN. Racial bias in pain assessment and treatment recommendations, and false beliefs about biological differences between blacks and whites. Proc Natl Acad Sci U S A. 2016; 113(16), 4296–4301.

8. Reverby SM. Invoking "Tuskegee": problems in health disparities, genetic assumptions, and history. J Health Care Poor Underserved. 2010; 21(3), 26-34. 9. Spigner C. Medical Apartheid: The dark history of medical experimentation on black americans from colonial times to the present. J Natl Med Assoc. 2007; 99(9), 1074-1075. 10. Todd KH, Deaton C, D’Adamo AP, Goe L. Ethnicity and analgesic practice. Ann Emerg Med. 2000; 35(1), 11-6.


Comparing the Racial Disparities in the Prescription of Opioid Drugs in the United States

11. Groenewald CB, Rabbitts JA, Hansen EE, Palermo TM. Racial differences in opioid prescribing for children in the United States. Pain. 2018; 159(10), 2050-2057. 12. Tait RC, Chibnall JT. Racial/ethnic disparities in the assessment and treatment of pain: psychosocial perspectives. Am Psychol. 2014; 69(2), 131-41. 13. Mathur VA, Richeson JA, Paice JA, Muzyka M, Chiao JY. Racial bias in pain perception and response: experimental examination of automatic and deliberate processes. J Pain. 2014; 15(5), 476-484. 14. Piper BJ, Shah DT, Simoyan OM, McCall KL, Nichols SD. Trends in Medical Use of Opioids in the U.S., 2006–2016. American Journal of Preventive Medicine. 2018;54(5). 15. Singhal A, Tien Y-Y, Hsia RY. Racial-ethnic disparities in opioid prescriptions at emergency department visits for conditions commonly associated with prescription drug Abuse. PloS one. 2016; 11(8) 16. Kumar R, Gleason MM. Pediatric Attention-Deficit/ Hyperactivity Disorder in Louisiana: trends, challenges, and opportunities for enhanced quality of care. Ochsner J. 2019;19(4):357-368. 17. Morgan PL, Staff J, Hillemeier MM, Farkas G, Maczuga S. Racial and ethnic disparities in ADHD diagnosis from kindergarten to eighth grade. Pediatrics. 2013;132(1):85-93. 18. Travers JL, Cohen CC, Dick AW, Stone PW. The great American recession and forgone healthcare: do widened disparities between African-Americans and whites remain? PLoS One. 2017;12(12). 19. Levesque JF, Harris MF, Russell G. Patient-centred access to health care: conceptualising access at the interface of health systems and populations. Int J Equity Health. 2013;12(18).


Scholarly Research In Progress • Vol. 4, October 2020

A Critical Examination of the Mechanism of Action of Buprenorphine: Not Just a Mu Partial Agonist? Leana J. Pande1*, Stephanie D. Nichols2, and Brian J. Piper1,3

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 University of New England, Biddeford, ME 04005 3 Center for Pharmacy Innovation and Outcomes, Geisinger Precision Health Center, Forty-Fort, PA 18704 *Master of Biomedical Sciences Program Correspondence: 1


Abstract Buprenorphine, an analogue of thebaine, is a Schedule III opioid used for opioid-use disorder and as an analgesic. Buprenorphine is generally described as a partial mu-opioid receptor agonist with limited activity and a decreased response at the mu-receptor relative to full agonists. The muopioid receptor remains important clinically in defining efficacy in analgesic potential. In patients who are opioid naïve, the drug’s efficacy as an analgesic is found to be equivalent to a full mu-opioid receptor agonist, despite decreased receptor occupancy and the “ceiling effect” produced from larger doses of buprenorphine. Buprenorphine’s respiratory depressant effects, while less than many other opioids, are increased by benzodiazepines or alcohol. There have also been 11,000 reports involving buprenorphine and minors (age <19) to U.S. poison control centers, the preponderance (89.2%) with children. Contemporary research shows the traditionally taught pharmacology of buprenorphine does not take into account changes to receptor theory, pharmacological terminology, as well as consideration for the drug’s route of administration and biologically active major metabolites.

Methods We reviewed articles through GoogleScholar, PubMed, and Ovid databases to find relevant information regarding buprenorphine, buprenorphine’s binding capability, buprenorphine’s metabolites, respiratory depression, and analgesic properties. General concepts were also reviewed on UptoDate and GoogleScholar to establish definitions for receptor theory, agonists, and antagonists. We read articles that characterized buprenorphine as early as the 1970s and 1980s, all the way up to papers published in 2020. The searches were done from August 2019 through May 2020 using terms including “buprenorphine,” “norbuprenorphine,” “buprenorphine-3-glucuronide,” or norbuprenorphine3-glucuronide” in combination with the terms/ phrases “pharmacokinetics,” “pharmacology,” “receptor theory,” “receptor activity,” “efficacy,” “respiratory depression,” “ceiling effect”, “analgesia,” “antinociception,” “pain,” “metabolite,” and/ or “partial agonist.”

Introduction Buprenorphine is a derivative of thebaine, which can be found in the poppy of Papaver somniferum. During the mid to late 1900s, buprenorphine was considered a part of the solution to what was known as the “opium problem” (1). In the 1920s, the Committee on Drug Addiction (CDA) primarily focused

on morphine. At the time, the CDA looked at its multitude of uses, without its addictive side effects (1). Over 40 years later, buprenorphine was discovered in 1966, and in 1972 its agonistantagonist pharmacological character was presented by John Lewis to the College on Problems of Drug Dependence (1), and thought to be a potential addiction treatment, first recognized in 1979 by Don Jasinski (2). Although marketed for analgesia and addiction treatment, most studies at the time found this was the “most reinforcing drug they had ever used” (1). By 1985, it was available in 29 countries (1). Buprenorphine was originally considered a Schedule V narcotic, until 2002 when it was rescheduled to Schedule III, after the Drug Enforcement Agency made three attempts to reschedule it (1). Buprenorphine is employed, with or without naloxone, for the treatment of opioid use disorder. From 2008-2019, buprenorphine distribution increased seven-fold (476.8 to 3179.9 kg) and five-fold (18.6 to 97.6 kg) to pharmacies and hospitals, respectively (3). The U.S. Medicaid program spent $1 billion on buprenorphine in 2017 alone (3). Buprenorphine has activity at the mu, delta, and kappa, as well as the opioid receptor-like (ORL-1) also known as nociceptin, opioid receptors. There are four main opioid receptors, mu (MOR), delta (DOR), and kappa (KOR), identified in the 1960s, and the opioid receptor-like (ORL) or nociceptin (NOP), discovered in the 1990s (4). The NOP is considered an atypical, low-affinity receptor for opioid peptides (4). The mu-opioid receptor is primarily responsible for analgesic effects as well as euphoria, miosis, constipation, and respiratory depression (8). It may have a greater impact at spinal MOR relative to the brain receptors (5). Delta receptors have minimal antinociceptive effects relative to the MOR, but have more activity in chronic pain than acute pain. The DOR also participates in analgesic tolerance and physical dependence (8). The KOR has been seen to have analgesic and proalgesic effects to opioids, while also contributing to miosis and sedation (8). Buprenorphine’s active metabolite, norbuprenorphine, is a potent and major metabolite that attenuates the typical analgesic effects of buprenorphine due to binding of the ORL-1 receptor. The NOP is also responsible for the respiratory depressant effects. Buprenorphine is a unique opioid as a result of its receptor activity at the MOR (5). Buprenorphine dissociates from the muopioid receptor slowly, resulting in a slow duration of action (2). While most opioids show activity at the mu, delta, and kappa receptors, buprenorphine is an antagonist for the delta and kappa opioid receptors, with high affinity (6). Buprenorphine is potent at MOR and DOR, with efficacy at MOR, DOR, and the KOR, in order of descending efficacy (7).


A Critical Examination of the Mechanism of Action of Buprenorphine

Discussion Buprenorphine and receptor theory All opioids have activity at the mu receptor (9). Opioids have previously been classified as “weak” or “strong” based on their affinity for the receptor. “Weak” opioids are considered less likely to lead to addiction and adverse effects, such as respiratory depression, “Strong” opioids have greater analgesic effect and greater risk for addiction (8). Buprenorphine has antinociceptive effects that are considered primarily the result of activity at MOR (10). Traditionally, buprenorphine was described as a partial mu agonist with analgesic effects, developed to limited respiratory depression and addiction (11, 12). Since buprenorphine’s initial classification, the meaning of the terms “agonist” and “antagonist” (8, 13, 14) have been more fully elucidated. Initially, it was believed that all agonists for a receptor will result in different degrees of the same intracellular response (13,14). The transduction pathways of a drug activated by an agonist do not act identically for each receptor (4). Partial agonists are known for lack of intrinsic efficacy (16). The antinociceptive effect ascribed to buprenorphine is considered mainly mediated by mu opioid receptors (18). Bell-shaped dose-response curves of buprenorphine in the 1980s and 1990s showed there is an optimal range in concentrations for a maximum analgesic effect, with a decrease in activity below or above this range (19). The perception of buprenorphine’s clinical usages may depend on the correct application or interpretation of terms from concepts in receptor theory, such as efficacy and agonist (20). In recent years, it has become clearer that different ligands for the same receptor can cause different responses, contrary to traditional receptor theory (16). For receptor theory models to be useful, it must aid in determining the extent in which drug effects can be interpreted and applied to predict future effects (14). The term “ligand bias” has been used to describe opioid analgesic drugs which elicit a different intracellular response; therefore, their effects are not only the result of receptor binding affinity (7). Buprenorphine differentiates itself from other opioids with mu-receptor activity with its slow dissociation from the receptor (19). Buprenorphine alone is not responsible for its antagonistic effects, but its varying metabolite concentration through different forms of drug administration may alter the efficacy of the drug. The acute toxicity (LD50) of buprenorphine varies based on the method of drug administration (See Table 1, from reference 21, 22). Studies have suggested that differed opioid agonists have

different downstream effects in the cell, while still binding and activating the same receptor. Therefore, different opioids cannot be considered equivalent by changing the dose (8). It can no longer be assumed that any ligand activating a receptor will produce relatively the same response, with differences attributed to the agonists’ efficacies (4). Ligands for a receptor can alter the downstream activity in a pathway, known as biased agonism, ligand-directed signaling, and functional selectivity (23). Reservations regarding buprenorphine’s clinical use were due to misconceptions about an analgesic “ceiling effect” (15). Data that shows a bell-shaped dose response curve displaying a ceiling effect is typically derived from animal pain models that use high doses of buprenorphine beyond what would be used clinically for humans. In humans, these curves are produced through the extrapolation of existing data (15). Until recently, agonists like buprenorphine have been known for limited intrinsic activity and inability to produce as large a response at a receptor (11, 16). Distinctions between “weak” and “strong” opioids or “full” and “partial” agonists may be needed to account for “weak” opioids like buprenorphine having characteristics considered “strong” (8). Opioids that are pure agonists such as morphine or fentanyl produce stronger analgesic effects than drugs like codeine that have decreased receptor binding (17). However, factors such as affinity and efficacy, as well as variables like metabolite binding and concurrent receptor binding may alter the perceived effects and receptor activity of buprenorphine. Buprenorphine metabolites: receptor activity and effects Buprenorphine’s metabolism supports its analgesic effects (5). The hepatic cytochrome P (CYP) 450 (CYP P450-3A4) system metabolizes buprenorphine to norbuprenorphine through N-dealkylation of the cyclopropylmethyl group (5, 24, 25). This step allows for blood-brain barrier transport of the drug (5). Additional active metabolites are produced through the formation of conjugates with glucuronic acid with UDP glucuronosyl transferase, to produce buprenorphine3-glucuronide and norbuprenorphine-3-glucuronide, from buprenorphine and norbuprenorphine, respectively. (5, 24, 26). Norbuprenorphine-3-glucuronide has a sedative effect and norbuprenorphine-3-glucuronide is an analgesic with low-potency (See Table 2 for a summary of buprenorphine’s metabolite effects) (5). Glucuronide metabolites of buprenorphine are biologically active, and have heterogenous binding affinity at opioid receptors. Binding affinity, the ability a drug has to bind to a receptor, is measured by the equilibrium inhibitory constant (Ki) (5). Buprenorphine has a high binding affinity at the MOR and KOR, with debatable effects (5). Studies show that the KOR receptor activity might be characterized as partial agonist (28, 70), antagonist (71), and is even thought to have no activity (21, 71). Buprenorphine-3-glucuronide had high affinity for MOR (Ki = 4.9 ± 2.7 μM), and NOR (Ki = 36 ± 0.3 μM), receptors. Norbuprenorphine-3-glucoronide had an affinity for NOR (Ki = 18 ± 0.2 μM), but not

Table 1. LD50 (acute toxicity) of buprenorphine based on the method of administration 22

A Critical Examination of the Mechanism of Action of Buprenorphine

MOR (27) (See Table 3 for full list of inhibition constants and Figures 1–5 for visual summary of receptor-metabolite binding). While norbuprenorphine has a greater efficacy, it is considered a less potent partial agonist than buprenorphine at MOR (28). All metabolites except norbuprenophine-3-glucuronide have analgesic properties (27, 29).

Table 2. Summary of the effects of buprenorphine metabolites

Table 3. Receptor affinity shown as apparent Ki (inhibition constant) of buprenorphine and its metabolites, as determined by Brown et al. 2011

Figure 1. Visual summary of buprenorphine and buprenorphine metabolites’ receptor activity

Norbuprenorphine is the only wellresearched metabolite, compared to others which need to be studied in greater detail for a greater understanding of their clinical effects. Norbuprenorphine derives from buprenorphine as a result of N-dealkylation catalyzed by cytochrome P450 in the liver (24, 25). At the MOR, both norbuprenorphine and buprenorphine are potent partial agonists, with norbuprenorphine having moderate efficacy and buprenorphine having low efficacy. At the NOP receptor, norbuprenorphine has moderate efficacy and buprenorphine having low efficacy, with both metabolites having low affinity for the receptor. This information was determined using ligand binding experiments and cAMP assay (28). In rats, the LD50 of buprenorphine through intravenous administration was 1,149.5 and 234.6 mg kg-1, and was found to have a norbuprenorphine-to-buprenorphine LD50 ratio of 1/16-1/23 (30). Norbuprenorphine was 50-fold less potent than buprenorphine through intravenous administration and 4-fold less potent after intracerebroventricular (ICV) administration in in vivo animal studies. This decrease in potency may be due to poor penetration across the bloodbrain barrier compared to buprenorphine (31). The intraventricular administration of buprenorphine and norbuprenorphine showed norbuprenorphine’s analgesic activity was 25% that of buprenorphine (32). Norbuprenorphine is the only wellresearched metabolite, compared to others which need to be studied in greater detail for a greater understanding of their clinical effects. Norbuprenorphine derives from buprenorphine as a result of N-dealkylation catalyzed by cytochrome P450 in the liver (24, 25). At the MOR, both norbuprenorphine and buprenorphine are potent partial agonists, with norbuprenorphine having moderate efficacy and buprenorphine having low efficacy. At the NOP receptor, norbuprenorphine has moderate efficacy and buprenorphine having low efficacy, with both metabolites having low affinity for the receptor. This information was determined using ligand binding experiments and cAMP assay (28). In rats, the LD50 of buprenorphine through intravenous administration was 1,149.5 and 234.6 mg kg-1, and was found 23

A Critical Examination of the Mechanism of Action of Buprenorphine

to have a norbuprenorphine-to-buprenorphine LD50 ratio of 1/16-1/23 (30). Norbuprenorphine was 50-fold less potent than buprenorphine through intravenous administration and 4-fold less potent after intracerebroventricular (ICV) administration in in vivo animal studies. This decrease in potency may be due to poor penetration across the blood brain barrier compared to buprenorphine (31). The intraventricular administration of buprenorphine and norbuprenorphine showed norbuprenorphine’s analgesic activity was 25% of buprenorphine (32). While buprenorphine has a low risk for respiratory depression and is rarely considered clinically relevant in that respect, (34, 35) norbuprenorphine is a potent respiratory depressant (29). Buprenorphine’s active metabolite, norbuprenorphine, was 10 times more potent than the parent drug (33). Respiratory depression can be induced by norbuprenorphine and mediated by MOR (33). Buprenorphine was found to be protective against norbuprenorphine’s effect of respiratory depression, both preventing and reversing these effects. Binding experiments show DOR and, primarily, MOR as responsible for buprenorphine protecting against the norbuprenorphine-induced respiratory depression (36). Respiratory depression with the use of buprenorphine varies depending upon method of drug administration and possibly age. In a study with healthy volunteers, intramuscular buprenorphine (0.15–1.2 mg) increased the risk of respiratory depression linearly, but the effect was not clinically significant (37). With sublingual buprenorphine (1–31 mg), patients reached respiratory depression at doses 8 mg or more (38). A study on 50 postoperative patients with intravenous buprenorphine (0.4–7.0 mg) showed no signs of respiratory depression for a 24-hour period (42). Healthy volunteers with intravenous buprenorphine (0.1 mg/70 kg body weight) demonstrated a ceiling in respiratory depression, but not in analgesic efficacy (35, 39). Animal experiments show that the respiratory ceiling occurs at a lower dose (>0.2 mg/kg) than the analgesic effect ceiling, which will only occur in doses beyond the therapeutic dose range (35, 39). Experimental and clinical data show that there is a limit on buprenorphine’s maximum depressant effect (15). Buprenorphine may be protective against respiratory depression, but does not account for drug interactions that can result in buprenorphine overdose or mixed route of drug administration that could increase norbuprenorphine levels. Buprenorphine reports to poison control centers, especially involving minors, are concerning. More research needs to be done to address the potential for norbuprenorphine presence to result in overdose in the presence of other drugs. In terms of norbuprenorphine’s analgesic ability, when combining results of animal and biochemical studies, norbuprenorphine and buprenorphine are considered by some to be, partial agonists at the mu receptor (31). The co-activation of the NOP receptors by buprenorphine modulates the antinociceptive effect of buprenorphine at opioid receptors (40). Additionally, the mu-opioid receptor may be responsible for counteracting the hyperalgesic effect from NOR. If mu receptors are blocked, NOR produces hyperalgesia (41). Norbuprenorphine had a high binding affinity for the mu receptor and low affinity for the NOR and presented as a potent analgesic with an efficacy equal to buprenorphine in a writhing test (28). Buprenorphine’s agonistic effect at NOR is believed to counter antinociception by buprenorphine and norbuprenorphine on opioid receptors, producing the bell24

shaped curves in nociceptive assays (28). Preclinical reports show NOR agonism contributes to decreased analgesia at high concentrations. However, buprenorphine’s affinity for the NOR is approximately 50 times lower than its affinity for the MOR and NOR activation causing a pronociceptive effect has not been validated in clinical settings (27, 42, 43). Analgesic effect and route of buprenorphine administration Buprenorphine’s properties including low molecular weight, high lipophilicity, and high potency (5) influence its perceived effects. Buprenorphine is 96% protein bound after absorption (5). Oral absorption is considered to be poor due to buprenorphine’s “first pass metabolism” (5). Transdermal absorption is limited, Sublingual administration is considered effective Some studies consider buccal formulations to be the most efficient with the highest non-intravenous bioavailability (5). Because of the options in different methods of drug administration (5), buprenorphine’s analgesic ability does not appear to be limited and shows promise for pain treatment in patients who are opioid naïve (29, 43, 44). Preclinical studies have shown the effectiveness of buprenorphine in various pain conditions (45). In conscious rats, buprenorphine was even considered 100 times more potent than morphine (equipotent 0.03–3.0 mg/kg SC.) in paw pressure tests, but buprenorphine produced a bell-shaped dose response curve in hot plate tests. The antinociceptive effects of buprenorphine and morphine were equipotent in both paw pressure and hot plate tests when administered intrathecally at 10 micrograms (46). The paw-pressure test with subcutaneous administration showed buprenorphine was more potent than morphine (47). The method of administration of buprenorphine has significant implications for the efficacy and benefits or detriments associated with it (29). Buprenorphine is considered a potent analgesic when administered intravenously, intramuscularly, buccally, and sublingually, from moderate to severe pain levels (44). Buprenorphine’s slow onset time decreases its effectiveness for acute pain (44). However, based on the formulation and method of application, buprenorphine can be approximately 25 to 100 times more potent than morphine (2, 47, 48). Intrathecal injections of buprenorphine and morphine showed similar antinociceptive potencies after their peak, but with a shallower dose-response curve for buprenorphine. Similar results were shown through in the hot plate test, a test used for measuring acute and subcutaneous pain (46). For thermal pain, intrathecal buprenorphine was found to be 17 times more effective than hydromorphone (18). Buccal administration of buprenorphine was effective and tolerable in opioid naïve patients with moderate to severe low back pain (49, 50) and general “round-the-clock” chronic pain (51). Thirty-three clinical studies showed efficacy in buprenorphine for pain relief with 88% using transdermal buprenorphine and 12% using buprenorphine buccal film. Pergolizzi and Raffa considered buprenorphine to have the efficacy of a Schedule II buccal film and similar efficacy and tolerance to the transdermal formulation (42, 52). Buprenorphine buccal film (150–900 μg/12 h) had similar efficacy results as Schedule II hydromorphone hydrochloride (12–64 mg/12 h) (52). Sublingual buprenorphine in the tablet form was 15 times more potent than intramuscular morphine. Sublingual buprenorphine is also active longer than morphine (53) and was effective

A Critical Examination of the Mechanism of Action of Buprenorphine

postoperative analgesic (54–57). In postoperative cancer patients, it was found the relative potencies of intramuscular to sublingual buprenorphine is 2:1 (53). For postoperative pain by the intramuscular route, buprenorphine was found to be 30 times more potent than morphine (58–60) and have a longer duration of action than morphine in cancer patients (61). The analgesic potency of buprenorphine (23, 47) and its lipophilicity and low molecular weight make buprenorphine ideal for transdermal delivery (15). Lower doses of transdermal buprenorphine were required to produce the same equipotency as transdermal fentanyl (47). In two case studies, buprenorphine gave a positive response where transdermal fentanyl had failed (47). Transdermal administration of buprenorphine in chronic non-cancer, neuropathic, and cancer-related pain did not antagonize analgesia and demonstrated efficacy and safety as well as reduced negative effects like withdrawal. Transdermal buprenorphine has advantages for chronic pain treatment (62–63). Compared with placebo, the initiation of transdermal buprenorphine in patients with chronic non-cancer, neuropathic, and cancerrelated pain resulted in effected analgesia and showed beneficial efficacy. It was also safe, well-tolerated, and did not cause opioid withdrawal symptoms (62–63). Transdermal administration of buprenorphine was found to be efficacious and well tolerated in moderate to severe chronic low back pain (64) and long-term control of chronic pain in cancer patients (65, 45). Transdermal buprenorphine was seen to be effective for longer term chronic cancer and noncancer pain, with at least satisfactory analgesic effects reported in the preponderance (90%) of 215 patients (62). Patients with moderate to very severe chronic pain, both cancer- and noncancer-related, slept longer uninterrupted by pain and of the 239 patients participating, 90% found satisfactory pain relief and 95% tolerated the patch well (66).

curve of buprenorphine, however this team noted that dose comparisons between partial and full mu agonist would be made cautiously since extrapolation does not accurately estimate potency (38). Opioids rarely bind to a single receptor and will have difference in affinities to others. Buprenorphine co-activates other receptors that may play a role in its efficacy. Some of buprenorphine’s negative effects such as respiratory depression and abuse can be attributed to peripheral DOR (68). While opioid analgesics like buprenorphine often bind to the mu-opioid receptor, there is a variation in their affinity for this receptor as well as their affinity ratio for other receptors, such as the previously mentioned ORL-1 receptor, in addition to kappa and delta-opioid receptor (8). Buprenorphine has antagonistic activity at the KOR that causes antihyperalgesic effects to some extent (15). The antihyperalgesic effects of buprenorphine have successfully treated neuropathic pain (48, 62, 69), which may show neuropathic pain may be more susceptible to buprenorphine than other opioids (15). Antagonism from the KOR activation leads to predictions that drugs with lower affinity for the KOR relative to MOR will be effective in producing MOR-related effects (8). However, buprenorphine’s KOR activity is controversial as this agent is characterized as a partial agonist (28, 70), antagonist (71), and even thought to have no activity (21, 71). Buprenorphine is even contested as an antagonist or inverse agonist at KOR (See Table 4 for summary of buprenorphine’s analgesic efficacy) (5). Therefore, this receptor should have little role for consideration in buprenorphine’s activity considering these mixed results (30). Additionally, sigma-1 receptors modulate the analgesic effects of opioids. Ablation of TRPV+ cells with high densities of sigma-1 receptors, an orphan receptor whose endogenous neuropeptide ligand is unknown, did not alter IB4+ neurons with high amounts sigma-1 receptors, mechanical nociception, or sigma-1 antagonism on morphine antinociception. It did impair response to heat stimuli and morphine’s effect on heat nociception (73). Additionally, dimers of the receptors can arise as homo- or hetero-conformations, that may have distinct signals (74). MOR-DOR and DORKOR specific agonists have different signal, outcomes, and antinociceptive results (8, 75, 76).

With respect to analgesia, there was no observed ceiling effect within the therapeutic dose range of buprenorphine for pain (48). Buprenorphine was a potent analgesic with full efficacy in mouse models or acute somatic and visceral pain. As a result, the analgesic efficacy of buprenorphine in those who are opioid naïve is not limited by its categorization as a partial agonist or previous reports of the bell-shaped dose response curve, with a maximal efficacy of the compound was maintained at almost 100% maximum possible effect (10). In clinical studies, no ceiling has been found with buprenorphine’s analgesic effect (34, 45). In a report in humans with acute pain, ascending intravenous doses did not show any ceiling effect up to 0.6 mg of buprenorphine (roughly equivalent to 10–20 mg of intravenous morphine (39). In earlier papers classifying buprenorphine, mention of the ceiling effect seen with the MOR used dose ranges that were not relatively equivalent to the potency of other drugs it was tested against such as morphine. Morphine has been found to be 25–50 times more potent than buprenorphine as an analgesic (38, 67). Table 4. Summary of receptor activity associated with buprenorphine’s analgesic efficacy A plateau was reported in the dose-effect


A Critical Examination of the Mechanism of Action of Buprenorphine

Discussion The classification as a partial agonist comes in part from the reduced efficacy in morphine and other mu-opioid receptor agonists analgesics when first exposed to buprenorphine. The “antagonist profile” was a conclusion drawn from reduced efficacy if buprenorphine was injected before morphine. Buprenorphine is still a more potent analgesic than morphine and pentazocine in rat tail pressure tests, and marginally more potent than morphine in mouse and rat tail flick tests (66). Buprenorphine’s pharmacology allows for it to be combined with other mu-opioid receptor agonists for an additive analgesic effect, but only when the full agonist is added to buprenorphine and not in the reverse order. The reverse causes a precipitation of acute withdrawal (45, 76). Administering intrathecal morphine and IV buprenorphine simultaneously alleviates pain with decreased sedation and other side effects than either drug alone (79). Additionally, switching between buprenorphine and full mu-agonists is possible without the loss of analgesic efficacy and without refractory period when switching from buprenorphine to new mu-opioid treatment (44, 16). Overall, clinical practice guidelines state the importance of patients self-reporting effective analgesics as pain is considered a personal experience that varies based on individual threshold and tolerance (21). Partial antagonism is seen in animals with high dose ranges that are clinically irrelevant. Buprenorphine has a mu-agonistic profile of high potency and efficacy, as well as reversibility and no lag time for action, making it ideal for long-term pain treatment, if an opioid is absolutely necessary (44). In a partial agonist, the less than full effect should remain the same even with full receptor saturation (77). PET technology shows that buprenorphine can produce analgesia at less than full receptor occupancy Therefore, other receptors beyond the mu-opioid receptor may be responsible than analgesia (20). Buprenorphine has a high affinity for MOR, but occupies fewer receptors for analgesia. Buprenorphine increases mu-opioid receptor expression so that other mu agonists can interact with the receptors (44). Additionally, buprenorphine’s activation at the MOR occurs at lower levels of receptor phosphorylation (5). When administering buprenorphine, receptors are available for full agonism at MOR for the treatment of acute pain (5). The classification of a receptor depends on the ability of a drug to function in the environment it is presented in (5, 78). While different factors like temperature influence the agonist or antagonist activity perceived, it was determined that buprenorphine can be considered a full agonist in a clinical setting when used in people who are opioid naïve (78). Some in vitro assays have shown morphine to act as an antagonist, despite morphine being considered a fullagonist clinically (5). The bioavailability of certain metabolites in plasma, like norbuprenorphine, requires more research as this has implications on medications that can be coadministered with buprenorphine. Intravenous administration has a 100% bioavailability, buccal buprenorphine has 46–65% bioavailability, sublingual has a 28–51% bioavailability, and transdermal has a 15% bioavailability (5). Some studies show there is a significant amount of norbuprenorphine remaining in the plasma following buprenorphine’s administration (80), 26

contrary to others (32). Buprenorphine overdoses reported in the mid-2000s can be related to varied norbuprenorphine plasma concentrations (81, 82), which can be related to method of administration (36, 83). Buprenorphine’s clearance in anesthetized patients was seen to be lower than individuals not under anesthesia, as well as in patients with reduced hepatic blood flow as a result of another administered anesthetic (57). Buprenorphine as a tablet has a bioavailability that is 50–60% that of a buprenorphine solution (84–85). Intranasal buprenorphine is 50% bioavailable in humans in a polyethylene glycol 300 and 5% dextrose vehicle, with a maximum concentration at 30 minutes. In sheep, buprenorphine’s intranasal bioavailability was 70% with a polyethylene glycol 300 vehicle and 89% with a dextrose vehicle (86). The full extent of buprenorphine’s pharmacokinetics’ is important clinically. Opioids were previous classified as “weak” or “strong” based on their affinity for the mu-opioid receptor. This is ineffective if used clinically, as “weak” opioids are considered less likely to lead to addiction and adverse side effects. Classifying buprenorphine as a weak opioid is harmful because the drug can cause physiologic dependence and has the potential for analgesic benefits, if an opioid is deemed necessary for treatment. Without taking into consideration of factors such as method of administration or distinguishing myth from fact, this can lead to incorrect assumptions in the efficacy of the opioid prescribed. Additionally, without taking into account the full effects of the metabolites’ transduction based on method of administration, there can be side effects as a result of residual effects of buprenorphine’s metabolism. Continued efforts to better understand the complex pharmacodynamics and pharmacokinetics of buprenorphine and metabolites will result in a better appreciation of the risks and benefits of this ubiquitous opioid.

References 1.

Campbell ND, Lovell AM. The history of the development of buprenorphine as an addiction therapeutic. Ann NY Acad Sci. 2012; 1248:124-139.

2. Jasinski DR, Pevnick JS, Griffith JD. Human pharmacology and abuse potential of the analgesic buprenorphine: a potential agent for treating narcotic addiction. Arch Gen Psychiatry. 1978; 35(4): 501–516. 3. Pashmineh AA, Cruz-Mullane A, Podd JC, Lam WS, Kaleem SS, Lockard LB, et al. Rise and Regional Disparities in Buprenorphine Utilization in the United States. Pharmacoepidemiology Drug Saf. 2019; 29(6): 1-8. 4. Cox BM, Christie MJ, Devi L, Toll L, Traynor JR. Challenges for opioid receptor nomenclature: IUPHAR Review 9. Br. J. 2014; 172(2): 317-323. 5. Gudin J, Fudin J. A narrative pharmacological review of buprenorphine: A unique opioid for the treatment of chronic pain. 2020; Pain Ther. 2020; 9: 41-54. 6. Khroyan TV, Wu J, Polgar WE, Cami-Kobeci G, Fotaki N, Husbands SM, et al. BU08073 a buprenorphine analogue with partial agonist activity at mu-receptors in vitro but long-lasting opioid antagonist activity in vivo in mice. Br J Pharmacol. 2015;172(2): 668–80.

A Critical Examination of the Mechanism of Action of Buprenorphine


Kuo A. Magiera J, Rethwan N, Andersson A, Lam AL, Wyse B, Meutermans W, Lewis R, Smith M. in vitro profiling of opioid ligands using the cAMP formation assay and the beta-arrestin2 recruitment assay: No two ligands have the same profile. Eur J Pharmacol. 2020; 872:1-10.

8. Emery MA, Eitan S. Members of the same pharmacological family are not alike: Different opioids, different consequences, hope for the opioid crisis? Prog Neuropsychopharmacol Biol Psychiatry. 2019; 92:428– 449. 9. Pathan H, Williams J. Basic opioid pharmacology: an update. Br J Pain. 2012;6(1):11–16. 10. Kögel B, Christoph T, Strassburger W, Friderichs E. Interaction of mu-opioid receptor agonists and antagonists with the analgesic effect of buprenorphine in mice. Eur J Pain. 2005; 9(5):599–611. 11. Rivera SM, Gilman AG. Access Medicine. Goodman & Gilman's: The Pharmacological Basis of Therapeutics13e New York: McGraw-Hill, 2015. 12. Ni Q, Xu H, Partilla JS, de Costa BR, Rice KC, Kayakiri H, et al. Opioid peptide receptor studies. 3. Interaction of opioid peptides and other drugs with four subtypes of the κ2 receptor in guinea pig brain Synapse. 1995; 16(6):6064, 13. De Lean A, Stadel JM, Lefkowitz RJ. A ternary complex model explains the agonist-specific binding properties of the adenylate cyclase-coupled beta-adrenergic receptor. J Biol Chem. 1980; 255(15): 7108–7117. 14. Kenakin T. Principles: receptor theory in pharmacology. Trends Pharmacol Sci. 2004; 25(4):186–192. 15. Kress HG. Clinical update on the pharmacology, efficacy and safety of transdermal buprenorphine. Eur J Pain. 2009; 13(3):219–230. 16. Katzung BG. Access Medicine. Introduction: The nature of drugs & drug development & regulation. In: Katzung BG. eds. Basic & Clinical Pharmacology New York: McGrawHill Education 2018. 17. Gladson B, Myslinski M, Streifer M. Access Medicine. Principles of pharmacology and selective agents in physical medicine and rehabilitation. Mitra R. eds. Principles of Rehabilitation Medicine New York: McGrawHill Education 2018. 18. Tejwani GA, Rattan AK. The role of spinal opioid receptors in antinociceptive effects produced by intrathecal administration of hydromorphone and buprenorphine in the rat. Anesth Analg. 2002; 94(6): 1542-6. 19. Dum JE, Herz A. In vivo receptor binding of the opiate partial agonist, buprenorphine, correlated with its agonistic and antagonistic actions. Br J Pharmacol. 1981; 74(3):627–633. 20. Ding Z, Raffa RB. Identification of an additional supraspinal component to the analgesic mechanism of action of buprenorphine. Br J Pharmacol. 2009; 157(5): 831–843. 21. Johnson RE, Fudala PJ, Payne R. Buprenorphine: considerations for pain management. J Pain Symptom Manage. 2005; 29(3):297–326.

22. Gueye PN, Borron SW, Risede P, Monier C, Buneaux F, Debray M, et al. Lack of effect of single high doses of buprenorphine on arterial blood gases in the rat. Toxicol Sci. 2001; 62(1):148–54. 23. Galandrin S, Oligny-Longpré G, Bouvier M. The evasive nature of drug efficacy: implications for drug discovery. Trends Pharmacol Sci. 2007; 28(8):423–430. 24. Iribarne C, Picart D, Dréano Y, Bail JP, Berthou F. Involvement of cytochrome P450 3A4 in N-dealkylation of buprenorphine in human liver microsomes. Life Sci. 1997; 60(22):1953–1964. 25. Kobayashi K, Yamamoto T, Chiba K, Tani M, Shimada N, Ishizaki T, et al. Human buprenorphine N-dealkylation is catalyzed by cytochrome P450 3A4. Drug Metab Dispos. 1998; 26(8):818–821. 26. Mistry M, Houston JB. Glucuronidation in vitro and in vivo. Comparison of intestinal and hepatic conjugation of morphine, naloxone, and buprenorphine. Drug Metab Dispos. 1987; 15(5):710–717. 27. Brown SM, Holtzman M, Kim T, Kharasch ED. Buprenorphine metabolites, buprenorphine-3-glucuronide and norbuprenorphine-3-glucuronide, are biologically active. Anesthesiology. 2011; 115(6):1251–1260. 28. Huang P, Kehner GB, Cowan A, Liu-Chen LY. Comparison of pharmacological activities of buprenorphine and norbuprenorphine: norbuprenorphine is a potent opioid agonist. J Pharmacol Exp Ther. 2001; 297(2):688–695. 29. Butler S. Buprenorphine—Clinically useful but often misunderstood. Scan J Pain 2013; 4(3),148-152. 30. Mégarbane B, Hreiche R, Pirnay S, Marie N, Baud FJ. Does high-dose buprenorphine cause respiratory depression?: possible mechanisms and therapeutic consequences. Toxicol Rev 2006; 25(2):79–85. 31. Coller JK, Christrup LL, Somogyi AA. Role of active metabolites in the use of opioids. Eur J Clin Pharmacol. 2009; 65(2):121–139. 32. Ohtani M, Kotaki H, Sawada Y, Iga T. Comparative analysis of buprenorphine- and norbuprenorphineinduced analgesic effects based on pharmacokineticpharmacodynamic modeling. J Pharmacol Exp Ther. 1995; 272(2):505–510. 33. Ohtani M, Kotaki H, Nishitateno K, Sawada Y, Iga T. Kinetics of respiratory depression in rats induced by buprenorphine and its metabolite, norbuprenorphine. J Pharmacol Exp Ther. 1997; 281(1):428–433. 34. Dahan A, Yassen A, Bijl H, Rohmberg R, E. Sarton, L. Teppema, et al. Comparison of the respiratory effects of intravenous buprenorphine and fentanyl in humans and rats. Br J Anaesth. 2005; 94(6):825–834. 35. Dahan A. Opioid-induced respiratory effects: new data on buprenorphine. Palliat Med. 2006; 20(8); 3-8. 36. Mégarbane B, Hreiche R, Pirnay S, Marie N, Baud FJ. Does high-dose buprenorphine cause respiratory depression?: possible mechanisms and therapeutic consequences. Toxicol Rev. 2006; 25(2):79–85.


A Critical Examination of the Mechanism of Action of Buprenorphine

37. Orwin JM, Orwin J, Price M. A double-blind comparison of buprenorphine and morphine in conscious subjects following administration by the intramuscular route. Acta Anaesthesiol Belg. 1976; 27(3):171–181. 38. Walsh SL, Preston KL, Stitzer ML, Cone EJ, Bigelow GE. Clinical pharmacology of buprenorphine: Ceiling effects at high doses. Clin Pharmacol Ther. 1994; 55(5):569–580. doi:10.1038/clpt.1994.71 39. Yassen A, Kan J, Olofsen E, Suidgeest E, Dahan A, Danhof M. Mechanism-based pharmacokinetic-pharmacodynamic modeling of the respiratory-depressant effect of buprenorphine and fentanyl in rats. J Pharmacol Exp Ther. 2006; 319(2):682–692. 40. Lutfy K, Eitan S, Bryant CD, Yang YC, Saliminejad N, Walwyn W, et al. Buprenorphine-induced antinociception is mediated by mu-opioid receptors and compromised by concomitant activation of opioid receptor-like receptors. J Neurosci. 2003; 23, 10331–10337. 41. Lutfy K, Hossain SM, Khaliq I, Maidment NT. Orphanin FQ/nociceptin attenuates the development of morphine tolerance in rats. Br J Pharmacol. [Internet]. 2009; 134(3):529–534. 42. Pergolizzi J, Böger RH, Budd K, Dahan A, Erdine S, Hans G, et al. Opioids and the management of chronic severe pain in the elderly: consensus statement of an International Expert Panel with focus on the six clinically most often used World Health Organization Step III opioids (buprenorphine, fentanyl, hydromorphone, methadone, morphine, oxycodone). Pain Pract. 2008; 8(4):287–313. 43. Davis MP, Pasternak G, Behm B. Treating chronic pain: An overview of clinical studies centered on the buprenorphine option. Drugs. 2018; 78(12):1211–1228. 44. Christoph T, Kögel B, Schiene K, Méen M, De Vry J, Friderichs E. Broad analgesic profile of buprenorphine in rodent models of acute and chronic pain. Eur J Pharmacol. 2005; 507(1-3):87–98. 45. Khanna IK, Pillarisetti S. Buprenorphine - an attractive opioid with underutilized potential in treatment of chronic pain. J Pain Res. 2015; 8:859–870. 46. Bryant RM, Olley JE, Tyers MB. Antinociceptive actions of morphine and buprenorphine given intrathecally in the conscious rat. Br J Pharmacol. 1983; 78(4):659–663. 47. Sorge J, Sittl R. Transdermal buprenorphine in the treatment of chronic pain: results of a phase III, multicenter, randomized, double-blind, placebo-controlled study. Clin Ther. 2004; 26(11):1808–1820. 48. Sittl R, Likar R, Nautrup BP. Equipotent doses of transdermal fentanyl and transdermal buprenorphine in patients with cancer and noncancer pain: results of a retrospective cohort study. Clin Ther. 2005; 27(2):225– 237. 49. Rauck RL, Potts J, Xiang Q, Tzanis E, Finn A. Efficacy and tolerability of buccal buprenorphine in opioid-naive patients with moderate to severe chronic low back pain. Postgrad Med. 2015; 128(1):1–11.


50. Gimbel J, Spierings EL, Katz N, Xiang Q, Tzanis E, Finn A. Efficacy and tolerability of buccal buprenorphine in opioid-experienced patients with moderate to severe chronic low back pain: results of a phase 3, enriched enrollment, randomized withdrawal study [published correction appears in Pain. 2017 Feb;158(2):366]. Pain. 2016; 157(11):2517–2526. 51. Hale M, Urdaneta V, Kirby MT, Xiang Q, Rauck R. Longterm safety and analgesic efficacy of buprenorphine buccal film in patients with moderate-to-severe chronic pain requiring around-the-clock opioids. J Pain Res. 2017; 10:233–240. 52. Pergolizzi JV Jr., Raffa RB. Safety and efficacy of the unique opioid buprenorphine for the treatment of chronic pain. J Pain Res. 2019 12:3299–3317. doi:10.2147/JPR. S231948 53. Wallenstein SL, Kaiko RF, Rogers AG, Houde RW. Crossover trials in clinical analgesic assays: Studies of buprenorphine and morphine. Pharmacotherapy. 1986; 6: 228-235. 54. Edge WG, Cooper GM, Morgan M. Analgesic effects of sublingual buprenorphine. Anaesthesia. 1979; 34(5):463– 467. 55. Fry EN. Relief of pain after surgery. A comparison of sublingual buprenorphine and intramuscular papaveretum. Anaesthesia. 1979; 34(6):549–551. 56. Risbo A, Jøosrgensen BC, Kolby P, Pedersen J. Sublingual buprenorphine for premedication and postoperative pain relief in orthopaedic surgery. Acta Anaesthesiol Scand. 1985; 29(2):180-182. 57. Bullingham R, McQuay H, Moore R, Weir L, An oral buprenorphine and paracetamol combination compared with paracetamol alone: A single dose double-blind postoperative study. Brit J Clin Pharmacol. 1981; 12: 863867. 58. Wang RI, Johnson RP, Robinson N, Waite E. The study of analgesics following single and repeated doses. J Clinical Pharmacol. 1981; 21: 121-125. 59. Tigerstedt I, Tammisto T., Double-blind, multipledose comparison of buprenorphine and morphine in postoperative pain. Acta Anaesthesiol Scand. 1981; 24: 462-468. 60. Mok MS, Lippmann M, Steen SN. Multidose/ observational, comparative clinical analgetic evaluation of buprenorphine. J Clinical Pharmacol. 1981; 21(7): 323-329. 61. Kjaer M, Henriksen H, Knudsen EJ. Intramuskulaer buprenorfin og morfin i behandlingen af cancersmerter. En kontrolleret undersøgelse [Intramuscular buprenorphine and morphine in the treatment of cancer pain. A controlled study]. Ugeskr Laeger. 1982 ;144(18):1306–1309. 62. Likar R. Transdermal buprenorphine in the management of persistent pain - safety aspects. Ther Clin Risk Manag. 2006; 2(1):115–125.

A Critical Examination of the Mechanism of Action of Buprenorphine

63. Gordon A, Callaghan D, Spink D, Cloutier C, Dzongowski P, O’ Mahony et al. Buprenorphine transdermal system in adults with chronic low back pain: a randomized, doubleblind, placebo-controlled crossover study, followed by an open-label extension phase. Clin Ther. 2010; 32(5):844– 860. 64. Steiner D, Munera C, Hale M, Ripa S, Landau C. Efficacy and safety of buprenorphine transdermal system (BTDS) for chronic moderate to severe low back pain: a randomized, double-blind study. J Pain. 2011;12(11):1163– 1173. 65. Pace MC, Passavanti MB, Grella E, Mazzariello L, Maisto M, Barbarisi M, et al. Buprenorphine in long-term control of chronic pain in cancer patients. Front Biosci. Pain. 2007;12:1291–1299. 66. Radbruch L, Vielvoye-Kerkmeer A. Buprenorphine TDS: The clinical development rationale and results. Int J Clin Pract Suppl Pain. 2003; (133):15–24 67. Cowan A, Doxey JC, Harry EJ. The animal pharmacology of buprenorphine, an oripavine analgesic agent. Br J Pharmacol. Pain. 1977; 60(4):547–554. 68. Gaveriaux-Ruff C, Nozaki C, Nadal X, Henver CH, Weibel R, Matifas A, et al. Genetic ablation of delta opioid receptors in nociceptive sensory neurons increases chronic pain and abolishes opioid analgesia. Pain. 2011; 152(6):1238–1248. 69. Louis F. Transdermal buprenorphine in pain management – experiences from clinical practice: five case studies. Int J Clin Pract. 2006; 60 (10): 1330-1334. Available from: doi:10.1111/j.1742-1241.2006.01109.x

75. Yekkirala AS, Banks ML, Lunzer MM, Negus SS, Rice KC, Portoghese PS. Clinically employed opioid analgesics produce antinociception via μ-δ opioid receptor heteromers in rhesus monkeys. ACS Chem Neurosci. 2012; 3(9):720–727. 76. Tröster A, Ihmsen H, Singler B, Filitz J, Koppert W. Interaction of fentanyl and buprenorphine in an experimental model of pain and central sensitization in human volunteers. Clin J Pain. 2012; 28(8):705–711. 77. Katzung BG, Kruidering-Hall M, Trevor AJ. Pharmacodynamics. Katzung & Trevor's Pharmacology: Examination & Board Review, 12e New York: McGraw Hill; 2019. 78. Greenwald MK, Johanson CE, Moody DE, Woods JH, Kilbourn MR, Koeppe et al. Effects of buprenorphine maintenance dose on mu-opioid receptor availability, plasma concentrations, and antagonist blockade in heroin-dependent volunteers. Neuropsychopharmacology. 2003; 28(11):2000–2009. 79. Beltrutti D, Niv D, Ben-Abraham R, Di Santo S, Weinbroum AA. Late antinociception and lower untoward effects of concomitant intrathecal morphine and intravenous buprenorphine in humans. J Clin Anesth. 2002; 14(6):441– 446. 80. Gopal S, Tzeng TB, Cowan A. Characterization of the pharmacokinetics of buprenorphine and norbuprenorphine in rats after intravenous bolus administration of buprenorphine. Eur J Pharm Sci. 2002; 15(3):287–293. 81. Kintz P. A new series of 13 buprenorphine-related deaths. Clin Biochem. 2002; 35(7):513–516.

70. Zhu J, Luo LY, Li JG, Chen C, Liu-Chen LY. Activation of the cloned human kappa opioid receptor by agonists enhances [35S]GTPgammaS binding to membranes: determination of potencies and efficacies of ligands. J Pharmacol Exp Ther Pain. 1997; 282(2):676–684.

82. Pirnay S, Borron SW, Giudicelli CP, Tourneau J, Baud FJ, Ricordel I. A critical review of the causes of death among post-mortem toxicological investigations: analysis of 34 buprenorphine-associated and 35 methadone-associated deaths. Addiction. 2004; 99(8):978–988.

71. Romero DV, Partilla JS, Zheng QX, Heyliger SO, Ni Q, Rice KC, Lai J., Rothman RB. Opioid peptide receptor studies. 12. Buprenorphine is a potent and selective μ/κ antagonist in the [35S]-GTP-γ-S functional binding assay. Synapse. 1999; 34(2): 83-94.

83. Kuhlman JJ Jr, Lalani S, Magluilo J Jr., Levine B, Darwin WD, Johnson RE, Cone EJ. Human pharmacokinetics of intravenous, sublingual, and buccal buprenorphine. J Anal Toxicol. 1996; 20(6):369–378.

72. Toll L, Berzetei-Gurske IP, Polgar WE, S R Brandt, I D Adapa, L Rodriguez, et al. Standard binding and functional assays related to medications development division testing for potential cocaine and opiate narcotic treatment medications. NIDA Res Monogr. 1998; 178:440–466. 73. Montilla-García Á, Tejada MÁ, Ruiz-Cantero MC, BravoCaparrós I, Yeste S, Zamanillo D. et al. Modulation by sigma-1 receptor of morphine analgesia and tolerance: Nociceptive pain, tactile allodynia and grip strength deficits during joint inflammation. Front Pharmacol. 2019; 10:136.

84. Nath RP, Upton RA, Everhart ET, Cheung P, J.E., Shwonek P, Jones RT, et al. Buprenorphine pharmacokinetics: Relative bioavailability of sublingual tablet and liquid formulations. J Clinical Pharmacol. 2013; 39(6): 619-623. 85. Schuh KJ, Johanson CE. Pharmacokinetic comparison of the buprenorphine sublingual liquid and tablet. Drug Alcohol Depend. 1999;56(1):55–60. 86. Lindhardt K, Ravn C, Gizurarson S, Bechgaard E. Intranasal absorption of buprenorphine--in vivo bioavailability study in sheep. Int J Pharm. 2000; 205(1-2):159–163.

74. Rozenfeld R, Devi LA. Exploring a role for heteromerization in GPCR signalling specificity. Biochem J. 2011; 433(1):11–18.


Scholarly Research In Progress • Vol. 4, October 2020

Genotypic and Phenotypic Abnormalities Linking 22q11.2 Deletion Syndrome and Psychiatric Disorders Cecilia Leng1*, Khine S.Y. Win1*, Ben Truong1*, Josely Frias1*, and Brian J. Piper1,2

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Center for Pharmacy Innovation and Outcomes, Geisinger Precision Health Center, Forty Fort, PA 18704 *Master of Biomedical Sciences Program Correspondence: 1


Abstract 22q11.2 deletion syndrome (22q11.2DS), otherwise known as DiGeorge syndrome, is a disease that causes many physical deformities as well as neurological abnormalities. Of all the diseases linked to microdeletions, 22q11.2DS has the highest prevalence due to its large unstable low copy number repeat blocks (LCR). Previous studies have shown associations between the 22q11.2DS and psychiatric illnesses, particularly schizophrenia, with potential connections to significant changes in white and grey matter in the brain. It is difficult to diagnose this disease due to the various symptomatologies; currently, the only way to confirm this disease is through genetic testing. In this review, the genotypic and phenotypic differences in addition to the clinical symptomatology and diagnostic tools of 22q11.2DS and their links to psychiatric illnesses are examined.

Introduction Chromosome 22q11.2 deletion syndrome, also known as DiGeorge syndrome, is most often a result of a microdeletion of chromosome 22. It was first described in 1965 by Dr. Angelo DiGeorge, who studied a group of infants that exhibited conotruncal cardiac anomalies, hypoplastic thymus, and hypocalcemia (preceding parathyroid hypoplasia), otherwise known as the clinical triad. Molecularly, four blocks of low copy number repeats (LCRs) found in this DNA region of chromosome 22 undergo unequal meiotic recombination due to mispairing of the DNA repeats, leading to the chromosomal deletions (1). The LCRs on chromosome 22 contain more DNA than other LCRs and are highly unstable, leading to an increased number of patients with 22q11.2 deletion syndrome in comparison to other chromosomal deletions. This can occur by autosomal inheritance of mutated DNA from parents with low penetrance (2). Some clinical characteristics include autoimmune disease, hearing loss, learning difficulty, abnormal facial morphology, congenital heart disease, and psychiatric illness (3). Patients with chromosome 22q11.2 deletion syndrome are at a 25-fold higher risk for developing schizophrenia or a schizoaffective disorder (4). A majority of the clinical features are indistinguishable from schizophrenia or related schizoaffective disorders and can be genetically verified using molecular testing by fluorescence in situ hybridization (FISH) and a standard fluorescently-labeled DNA probe from the deleted area (4). It has been found that patients with schizophrenia and 22q11.2DS have a lower IQ and may differ in auxiliary clinical features such as myofascial dysmorphic features, palatal defects, and velocardiofacial defects (5).


In conjunction with major schizophrenic symptoms, such as delusions, hallucinations, movement difficulty, reduced speaking, and reduced feelings (20), the median age of disease onset in genetically predisposed children ranges from 12 to 26 years, and those with parents who did not transmit the deletion commonly exhibit less severe symptoms (5). Unique to the 22q11.2 deletion syndrome, some patients show signs of neurobehavioral characteristics different to the standard schizophrenic symptomatology, such as increased levels of excitement and impulsive reactions, which are linked to short temper or emotional outbursts (5). The exact mechanism leading to these characteristics is currently unknown, and there are many theories as to why it may occur. Researchers speculate that the deletion region could be linked to a unique locus sensitive to psychiatric illnesses, or that it could be a step in allowing more mutations to occur which would lead to psychiatric disorders (4, 6). There are various names that apply to the phenotypic features of the syndrome, such as cardiofacial syndrome, velocardiofacial, or conotruncal anomaly face syndrome, in addition to DiGeorge syndrome or 22q11.2 deletion syndrome. The multitude of names often leads to diagnostic confusion when treating these patients. Other patients may also present with clinical characteristics associated with this syndrome, but not in conjunction with chromosomal deletion, generating further uncertainty. This leads to confusion between other disorders such as such as coloboma-heart-atresia-retardedgenital-ear (CHARGE) syndrome, Smith-Lemli-Optiz syndrome, Kabuki syndrome, and Goldenhar syndrome. As such, it is standard to only diagnose patients presenting with the chromosomal 22q11.2DS and the clinical presentations associated with the syndrome with DiGeorge syndrome, and not consider other cases without 22q11.2 chromosome deletion as 22q11.2DS (1). Due to the issues pertaining to the diagnosis of this syndrome as well as the prevalence of syndrome, it is important to review the potential link that the genetic background can contribute to this phenomenon.

Methods Literature on chromosomal abnormalities as well as psychiatric symptomatology related to the 22q11.2DS were reviewed from a variety of sources, such as PubMed and Google Scholar. Articles were found using the search strategies of: “22q11.2 deletion syndrome” or “DiGeorge Syndrome” in addition to “schizophrenia,” “brain,” and “psychiatric disorders.” Articles referencing psychiatric illness alongside the deletion syndrome or DiGeorge syndrome were examined, and those

Genotypic and Phenotypic Abnormalities Linking 22q11.2 Deletion Syndrome and Psychiatric Disorders

including other significant psychiatric illnesses were included. Importance was weighted based on the mechanism proposed by the articles and the link between the syndrome and the disorder. Case studies connecting psychiatric disorders and the chromosomal deletion were included.

Discussion 22q11.2 deletion syndrome 22q11.2 deletion syndrome (22q11.2DS) is one of the most common microdeletion syndromes, also known as DiGeorge syndrome. It was originally used to diagnose children who presented with the symptomatology of immunodeficiency (~77%), hypoparathyroidism (hypocalcemia 49%), and congenital heart disease (~77%) (1). These children would often also present with thymic aplasia, or the lack of a fully functional thymus. Other symptoms were later discovered to be part of the syndrome and usually present alongside the clinical triad listed above; these symptoms include but are not limited to congenital anomalies (cleft palate 11%, polyhydramnios 16%, polydactyly 6%), autoimmune diseases (~10%), variable cognitive delays (nonverbal learning disability 65%, cognitive disability 30%, borderline cognitive disability 32%, low average 20%), behavioral presentations (anxiety 17–29%, phobia 23–61%, ADHD 3–46%, autism 14–45%), and later-onset psychiatric disorders (schizophrenia 10-30%) (1, 8). However, these symptoms vary case by case and are part of the reason for diagnostic difficulty. Deletions of LCR blocks 22q11.2DS consists of a chromosomal deletion in a region surrounded by four distinct blocks of LCRs (LCR A-D), each composed of multiple repeats with LCR A being the most proximal (1). It is known that these LCR blocks are larger and more complex than any other LCR blocks associated with human chromosomal deletion syndromes, and the commonality of their deletion underlies the theory that these blocks are inherently unstable. The LCRs are >96% identical, leading to a higher chance of meiotic error (9). It has been found that LCR A and LCR D are the largest segmental repeats and surround a 3-Mb 22q11.2 region hemizygous in most patients (8). This deletion is seen in 70–80% of patients (1) and occurs due to asynchronous replication at the 22q11.2 locus, leading to unequal meiotic recombination (10). This mechanism also results in proximal and distal nested deletions. Patients having proximal (centromeric) nested deletions, including LCR22A-LCR22B and LCR22A-LCR22C, show similar features to the 5–10% of patients that have the LCR22A-LCR22D deletion. The distal nested deletions include LCR22B-LCR22D and LCR22C-LCR22D deletions, which are associated with synchronous yet less penetrant presentations in comparison to individuals with LCR22A-LCR22D deletions. These deletions are more frequently inherited and show symptomatology such as palatal defects and developmental differences, along with a few others (8). TBX1, COMT, and PRODH genes Deletions of certain genes in 22q11.2DS have begun to be identified with various symptoms associated with the microdeletion syndrome. It is thought that T-box1 (TBX1)

gene deletion leads to cardiac and craniofacial features, while the catechol-O-methyl-transferase (COMT) and proline dehydrogenase (oxidase) 1 (PRODH) gene deletions lead to psychiatric presentations (1, 18). COMT gene encodes the enzyme involved in the breakdown of dopamine mainly in the prefrontal cortex. Hemizygosity leads to low activity, allowing dopamine to aggregate and potentially increase the risk for psychiatric illnesses. In the prefrontal cortex, dopamine plays a role in prepulse and latent inhibition (21), which can affect an individual’s ability to focus and avoid distractions (7). PRODH gene encodes the mitochondrial enzyme proline dehydrogenase, a catalyst used during the conversion of proline to glutamate. The low conversion levels observed due to PRODH deletion result in increased plasma proline levels, which affects N-methyl-D-Aspartate (NMDA) receptor signaling and may lead to damaged brain tissue. Type I hyperprolinemia (HPI) is an autosomal recessive disorder correlated with seizures, intellectual disability, and psychiatric symptoms, all of which are also associated with 22q11.2DS. HPI is caused by an inherited deficiency of PRODH, of which the polymorphic PRODH gene has had several SNPs studied for an association with schizophrenia, supporting the observations made by de Kong et al. in their study of the PRODH rs450046 polymorphism (18). Connection to psychiatric illnesses Patients with the 22q11.2DS have presented with higher rates of psychiatric illnesses. While it is uncommon for young children to be diagnosed, young adults and older individuals with this chromosomal deletion have shown increasing symptomology for depression and psychiatric illnesses. Patients diagnosed as psychotic and non-psychotic with velocardiofacial syndrome (VCFS) both show a negative correlation with respect to cognitive testing (i.e., the higher the patient’s age, the lower the patient's cognitive scores) (12). While there are associations between mood and anxiety disorders with the microdeletion syndrome, there is a significant increase in the comorbidity between schizophrenia or schizoaffective disorders and 22q11.2DS (13). Studies done on patients with the 22q11.2DS reveal a higher prevalence of midline brain abnormalities, such as the presence of a cava septum pellucidum et vergae, which is also a common feature of other genetic and neurodevelopmental disorders (14). Cava septum pellucidum, also known as the cave of septum pellucidum, is a cavity in the septum pellucidum present in fetuses. It usually fuses during infancy. If this cavity is not closed during infancy, cognitive and developmental problems can occur during childhood (19). In addition, it has been found that affected patients with the microdeletion can have significant increases or decreases of their grey matter up to 27%, as well as white matter decreases up to 33% (16). The reductions in the grey matter of the cerebellum, medial temporal-occipital lobe, and posterior cingulate cortex suggest that the loss of grey matter plays an important role in the pathways of neural processing in comparison to the frontal cortex (16). White matter anomalies predominate in regions specific to connectivity, mostly projection fibers and tracts (17). Behavioral features such as schizotypal traits may also result from the anatomical differences in brain regions such as the hippocampus,


Genotypic and Phenotypic Abnormalities Linking 22q11.2 Deletion Syndrome and Psychiatric Disorders

cerebral ventricles, and temporal lobes (15). Previous studies have shown abnormal brain development in autistic patients exhibiting social behavior problems (22, 23), and an increase of the grey matter has been thought to have a positive correlation with schizophrenia in addition to social behavioral issues (15). Diagnosis, treatment, and management Most of the symptoms associated with chromosome 22 microdeletions are variable, and the differing severity of symptoms can lead physicians to misdiagnose 22q11.2DS under different names, including VCFS syndrome (ShprintzenGoldberg Syndrome), conotruncal anomaly face syndrome, and Optiz G/BBB syndrome, and Cayler cardiofacial syndrome (24). Thus, molecular methods like fluorescent in situ hybridization (FISH) have been used to diagnose the disease in cytogenetics laboratories since 1992. FISH specifically targets chromosome 22 and identifies the absence of commonly deleted regions. If a region is missing, DiGeorge syndrome is confirmed. Even though FISH is a common method used to identify the syndrome, there are two potential complications that can arise. The results can take up to 3–14 days to receive, leaving the patient’s family anxious. Also, FISH lacks the precision necessary to detect atypical deletions. Two emerging methods, PCR-based technology and SNP arrays, provide quicker results. However, they still are similar to FISH in the sense that they can only detect typical deletions. Also, these methods are expensive compared to FISH. Most patients are diagnosed in utero using prenatal ultrasound. If cardiac anomalies are found on the ultrasound, chromosome 22q11.2 deletion tests are ordered and performed (1). Treatment of 22q11.2DS includes medications to treat associated symptoms, genetic counseling, and patientcentered therapy. While there is no cure for the syndrome, symptoms can be managed if carefully monitored and addressed as needed. If the patient is functioning normally, regular follow-up is needed to make sure no new health concerns arise. Physicians only consider prescribing medication if the patient has severe immunodeficiency stopping the body from preventing secondary infections. Since many drugs are still in the developmental and testing phase, adverse side effects can occur in patients with pre-existing autoimmune conditions. These side effects are currently not fully understood because there is insufficient research determining the harm of these drugs on immunocompromised patients. Because of the uncertainties surrounding the medications, physicians utilize genetic counseling in order to detect possible genetic abnormalities in utero, as well as inform parents of the risk of having additional affected children. Siblings are therefore also tested for the possibility of being carriers. Various forms of therapy dependent on the patient’s direct needs are also used to help people with DiGeorge syndrome smoothly navigate their everyday lives (1, 25) These findings suggest areas for future research to optimize patient care and advance diagnostic tools used to detect the syndrome early (1).


Conclusion 22q11.2DS is a debilitating disease resulting in variable clinical features, leading to diagnostic confusion with other diseases. It affects an estimated 1 in 4,000 people, but researchers suspect it is underdiagnosed due to the variable associated symptoms (24). It is also the most common deletion syndrome due to the unstable large LCR blocks with the deletion of the TBX1, PRODH, and COMT genes, and has a possible correlation to psychiatric illnesses. However, there is currently no existing data supporting the exact links between the genetic abnormalities and the psychiatric symptomatology. While there are many characteristics to this disease, the defining features influencing the brain are the variable changes in white and grey matter. Connections between the abnormal anatomy of the brain and schizotypal presentations are still not fully understood, and future studies are needed to further extend our knowledge on the exact mechanisms behind the progression of psychiatric disorders and brain anomalies in patients with 22q11.2DS.

Acknowledgments This work was supported by the Geisinger Commonwealth School of Medicine. The authors thank Dr. Brian Piper for assistance with the manuscript.

References 1.

McDonald-McGinn DM, Sullivan KE. Chromosome 22q11.2 deletion syndrome (DiGeorge syndrome/velocardiofacial syndrome). Medicine (Baltimore). 2011 Jan;90(1):1–18.

2. Shaikh TH, O’Connor RJ, Pierpont ME, McGrath J, Hacker AM, Nimmakayalu M, et al. Low copy repeats mediate distal chromosome 22q11.2 deletions: sequence analysis predicts breakpoint mechanisms. Genome Res. 2007 Apr;17(4):482–91. 3. Emanuel BS, McDonald-McGinn D, Saitta SC, Zackai EH. The 22q11.2 deletion syndrome. Adv Pediatr. 2001;48:39– 73. 4. Bassett AS, Chow EW, Weksberg R. Chromosomal abnormalities and schizophrenia. Am J Med Genet. 2000;97(1):45–51. 5. Bassett AS, Chow EWC. Schizophrenia and 22q11.2 deletion syndrome. Curr Psychiatry Rep. 2008 Apr;10(2):148–57. 6. Bassett AS. Chromosomal aberrations and schizophrenia. Autosomes. Br J Psychiatry. 1992 Sep;161:323–34. 7.

van Schouwenburg M, Aarts E, Cools R. Dopaminergic modulation of cognitive control: distinct roles for the prefrontal cortex and the basal ganglia. Curr Pharm Des. 2010;16(18):2026–32.

8. McDonald-McGinn DM, Sullivan KE, Marino B, Philip N, Swillen A, Vorstman JAS, et al. 22q11.2 deletion syndrome. Nat Rev Dis Primers. 2015 19;1:15071.

Genotypic and Phenotypic Abnormalities Linking 22q11.2 Deletion Syndrome and Psychiatric Disorders

9. Babcock M, Pavlicek A, Spiteri E, Kashork CD, Ioshikhes I, Shaffer LG, et al. Shuffling of genes within low-copy repeats on 22q11 (LCR22) by Alu-mediated recombination events during evolution. Genome Res. 2003 Dec;13(12):2519–32.

20. National Institute of Mental Health. Schizophrenia. National Institute of Mental Health. U.S. Department of Health and Human Services. 2012. Available from: https:// shtml#part_145430

10. Baumer A, Riegel M, Schinzel A. Non-random asynchronous replication at 22q11.2 favours unequal meiotic crossovers leading to the human 22q11.2 deletion. J Med Genet. 2004 Jun;41(6):413–20.

21. Ellenbroek BA, Budde S, Cools AR. Prepulse inhibition and latent inhibition: the role of dopamine in the medial prefrontal cortex. Neuroscience. 1996 Nov;75(2):535–42.

11. Edelmann L, Pandita RK, Morrow BE. Low-copy repeats mediate the common 3-Mb deletion in patients with velo-cardio-facial syndrome. Am J Hum Genet. 1999 Apr;64(4):1076–86. 12. Green T, Gothelf D, Glaser B, Debbane M, Frisch A, Kotler M, et al. Psychiatric disorders and intellectual functioning throughout development in velocardiofacial (22q11.2 deletion) syndrome. J Am Acad Child Adolesc Psychiatry. 2009 Nov;48(11):1060–8. 13. Schneider M, Debbané M, Bassett AS, Chow EWC, Fung WLA, van den Bree M, et al. Psychiatric disorders from childhood to adulthood in 22q11.2 deletion syndrome: results from the International Consortium on Brain and Behavior in 22q11.2 Deletion Syndrome. Am J Psychiatry. 2014 Jun;171(6):627–39.

22. McAlonan GM, Daly E, Kumari V, Critchley HD, van Amelsvoort T, Suckling J, et al. Brain anatomy and sensorimotor gating in Asperger’s syndrome. Brain. 2002 Jul;125(Pt 7):1594–606. 23. Murphy DGM, Critchley HD, Schmitz N, McAlonan G, Van Amelsvoort T, Robertson D, et al. Asperger syndrome: a proton magnetic resonance spectroscopy study of brain. Arch Gen Psychiatry. 2002 Oct;59(10):885–91. 24. 22q11.2 deletion syndrome - Genetics Home Reference NIH [Internet]. U.S. National Library of Medicine. National Institutes of Health; Available from: condition/22q112-deletion-syndrome#diagnosis 25. Purswani P, Walter JE, Sriaroon P. Treatment Challenges of Refractory Thrombocytopenia in DiGeorge Syndrome. Journal of Allergy and Clinical Immunology. 2019 Feb 1;143(2):AB112.

14. Kasai K, McCarley RW, Salisbury DF, Onitsuka T, Demeo S, Yurgelun-Todd D, et al. Cavum septi pellucidi in first-episode schizophrenia and first-episode affective psychosis: an MRI study. Schizophr Res. 2004 Nov 1;71(1):65–76. 15. Campbell LE, Daly E, Toal F, Stevens A, Azuma R, Catani M, et al. Brain and behaviour in children with 22q11.2 deletion syndrome: a volumetric and voxel-based morphometry MRI study. Brain. 2006 May;129(Pt 5):1218– 28. 16. van Amelsvoort T, Daly E, Robertson D, Suckling J, Ng V, Critchley H, et al. Structural brain abnormalities associated with deletion at chromosome 22q11: quantitative neuroimaging study of adults with velo-cardio-facial syndrome. Br J Psychiatry. 2001 May;178:412–9. 17. Barnea-Goraly N, Menon V, Krasnow B, Ko A, Reiss A, Eliez S. Investigation of white matter structure in velocardiofacial syndrome: a diffusion tensor imaging study. Am J Psychiatry. 2003 Oct;160(10):1863–9. 18. de Koning MB, van Duin EDA, Boot E, Bloemen OJN, Bakker JA, Abel KM, et al. PRODH rs450046 and proline x COMT Val158 Met interaction effects on intelligence and startle in adults with 22q11 deletion syndrome. Psychopharmacology (Berl). 2015 Sep;232(17):3111–22. 19. Beaton EA, Qin Y, Nguyen V, Johnson J, Pinter JD, Simon TJ. Increased incidence and size of cavum septum pellucidum in children with chromosome 22q11.2 deletion syndrome. Psychiatry Res. 2010 Feb 28;181(2):108–13.


Scholarly Research In Progress • Vol. 4, October 2020

Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018 John M. Boyle1†, Holly E. Funk1†, Susannah E. Pitt1†, Alison T. Varano1†, Sneha Vaddadi1†, Brian J. Piper1, and Kenneth L. McCall2

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 University of New England, Portland, ME, 04103 †Doctor of Medicine Program Correspondence: 1


Abstract Background: Over the past two decades, diagnosis of attention-deficit/hyperactivity disorder (ADHD) has significantly increased in the United States. ADHD is a neurodevelopmental disorder treated with amphetamine, methylphenidate, and lisdexamfetamine. In addition to pharmacotherapeutic use, stimulants have a high misuse potential. The purpose of this study was to analyze pharmacoepidemiological trends in both prescription and illicit stimulant distribution in the United States from 2014 through 2018. Methods: Drug weight data was obtained from the Automation of Reports and Consolidated Orders System (ARCOS) retail drug summary reports and in all 50 states for the following drugs: amphetamine, methylphenidate, and lisdexamfetamine. Mass of the stimulants in grams was corrected from population estimates from the U.S. Census Bureau, Population Division. The median values (20 mg/day/ person for methylphenidate and amphetamine and 40 mg/ day/person for lisdexamfetamine) were used for daily dose calculations. Results: Average prescription distribution increased by 8.2% and 15.2% for lisdexamfetamine and amphetamine, respectively, from 2015 through 2018. Methylphenidate prescription distribution decreased by 5.5% during this period. Lisdexamfetamine average daily dose for all 50 states increased from 0.74 mg/day/person to 0.83 mg/day/ person over the 4 years. Amphetamine daily dose increased from 2.92 mg/day/person to 3.62 mg/day/person. In contrast, methylphenidate daily dose decreased from 2014 through 2017, from 3.47 mg/day/person in 2014 to 3.23 mg/day/ person in 2017. Diverted lisdexamfetamine, amphetamine, and methylphenidate distribution changed by an average of +0.24%, -0.04% and +0.35% across all 50 states, respectively. Conclusion: From 2014 through 2018, the average prescription distribution and daily dose of amphetamine and lisdexamfetamine have increased throughout the United States, while methylphenidate prescription distribution and daily dose have decreased. Diverted stimulant distribution underwent modest changes. Efficacy, safety, and availability likely play an underlying role in these trends.

Introduction The diagnosis of attention-deficit/hyperactivity disorder (ADHD) is increasing in the United States. ADHD is a neurodevelopmental disorder which presents in childhood with symptoms of hyperactivity, impulsivity, and/or inattention. 34

Between 1997 and 2016, the prevalence of ADHD rose from 6.1% to 10.2% (1). The reason behind this significant increase remains unclear. One hypothesis is that there has been a change in physician and parental attitudes toward diagnosing attention disorders and pursuing treatment through medication (2). Stimulant based pharmacological therapy is indicated in the first-line treatment of ADHD in the U.S., along with behavioral therapy in younger children (3). Commonly used medications are amphetamine, methylphenidate, or lisdexamfetamine, and less commonly methamphetamine (4). The overall usage of stimulants has doubled in the last decade, with amphetamine distribution increasing 2.5-fold (5). The U.S. is somewhat unusual in that amphetamine is approved for use in 3- to 5-year-olds with ADHD. A report from the CDC documented instances of prescription stimulant use among 2-year-olds (6). A 2012 analysis of outpatient pediatric prescription drug trends found an increase in ADHD medication from 2002 through 2010, with methylphenidate being the most frequently prescribed treatment. The second most frequently prescribed treatment was amphetamine/ dexamphetamine, although there began a slight decrease resulting in lower prescriptions in 2010 compared with 2002 (7). Methylphenidate usage peaked in 2012, subsequently declining (5). Analysis of lisdexamfetamine began in 2007 after FDA approval for treatment, thus resulting in an increase in prescriptions from 2007 through 2010 (7). Alterations to catecholamine neurotransmitters are indicated in the pharmacology of ADHD. In general, stimulants prescribed to treat ADHD aim to increase dopamine and norepinephrine levels in regions of the brain with attentional function. While there are several elements to the mechanism of amphetamine, the drug works primarily to increase levels of extracellular dopamine and norepinephrine. This is accomplished by inhibiting reuptake in the synaptic cleft, particularly through inhibition of dopamine transporter and norepinephrine transporter. Methylphenidate directly inhibits these two transporters as well, while additionally agonizing serotonin receptors (8). Lisdexamfetamine is a prodrug that is converted in the body to d-amphetamine, which also inhibits the reuptake of dopamine and norepinephrine (9). While these medications benefit 70% of those diagnosed with ADHD, adverse side effects may include appetite loss, abdominal pain, headaches, and insomnia (3,10). However, the evidence base supporting the use of stimulants, and particularly methylphenidate, for ADHD has been characterized as “very weak” due to overreliance on short-term (<12 weeks) studies, underestimation of the risk of harms, particularly in industry-supported trials,

Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018

and inadequate efforts to be blinded when the comparison group receives an inactive placebo (11). A well-powered (N = 190,559) epidemiological report from the laboratory of the former director of the National Institute of Drug Abuse (NIDA) determined that there was a nine-fold increase in basal ganglia and cerebellar disorders among those with a history of prescription stimulant use (12). Interestingly, the current NIDA director and colleagues determined that one-year of prescription methylphenidate (daily dose = 1 mg per kg of body weight) use resulted in an increase in dopamine transporter binding in the caudate and putamen (13) but it is too early to even speculate if this could contribute to neurodegenerative disease processes. Stimulants are used in the treatment of other conditions as well, such as narcolepsy and binge eating disorder, and can often be diverted for nonmedical use (14,15). These pharmacotherapies are associated with significant abuse potential (16). Amphetamine, methylphenidate, and lisdexamfetamine are classified as Schedule II Controlled Substances, defined as having a high potential for abuse and possible adverse effects (17). Prescription drug diversion is defined as “the unlawful channeling of regulated pharmaceuticals from legal sources in the illicit market” (18). Stimulants, specifically methylphenidate, have been known to be commonly diverted drugs (19). While it is possible to track the distribution of prescription medications used in the treatment of ADHD through national databases, it is more difficult to monitor the illicit use of these commonly abused medications. In order to gain a more complete understanding of the use of ADHD stimulants, our study further looked at the illicit distribution of stimulants. The Researched Abuse, Diversion and AddictionRelated Surveillance (RADARS®) System uses publicly reported submissions to surveil the diversion of prescription medications. This allows for clearer understanding of the street price of prescription and illicit drugs, the geographical distribution of such drugs, and the trends in misuse. The RADARS System’s website,, allows the public to anonymously submit information regarding which drugs users bought, the amount of that drug, as well as the date and location of their transaction (20). also allows the public to access and rate this information, so that health officials can better understand the illegal market that exists for prescription drugs with the hopes of providing necessary outreach where needed. The purpose of this study was to examine and analyze pharmacoepidemiologic trends in both prescribed and illicit stimulant usage in the United States from 2014 through 2018. We sought to specifically analyze the mass distribution per drug and daily dose for amphetamine, methylphenidate, and lisdexamfetamine. Although previous studies have analyzed the trends of stimulant usage (21), we aimed to further analyze these trends with the incorporation of data from drug diversion to better understand the usage patterns.

Methods Procedures Drug weight data was obtained from the Automation of Reports and Consolidated Orders System (ARCOS) retail drug

summary reports for the years 2014 through 2017 in all 50 states for the following drugs: amphetamine, methylphenidate, and lisdexamfetamine. 2018 drug summary reports were obtained for lisdexamfetamine and amphetamine as well. The amphetamine and methylphenidate prescription distribution data reflect drugs of all enantiomers reported. The 2018 data have been temporarily excluded from the statistical reporting for methylphenidate and thus were not included in this analysis. Additional drug weight data for illicit distribution was obtained from through a Data Use Agreement between Geisinger Commonwealth School of Medicine and the Rocky Mountain Poison and Drug Safety department of the Denver Health and Hospital Authority. The date of submission, location of submission, generic name, and raw dose from 2015 through the end of 2018 was used for the following drugs: amphetamine, methylphenidate, and lisdexamfetamine. This study was approved by the Institutional Review Board of the University of New England. Data analysis Using ARCOS data, mass of the stimulants in grams was corrected from population estimates from the U.S. Census Bureau, Population Division for each year and state. Percent changes between consecutive pairs of years were calculated for the years 2015-2018. Outliers were identified using Grubbs’ Test on GraphPad Prism. Average daily dose calculations were completed for amphetamine, methylphenidate, and lisdexamfetamine for all 50 states for the years 2014 through 2018, with the exception of methylphenidate daily dose for 2018. These calculations used the median values of 20 mg/ day/person for methylphenidate and amphetamine and 40 mg/day/person for lisdexamfetamine (21). Figures were created using GraphPad Prism to plot the mean prescription distribution with standard error of mean (SEM) for each drug during their respective years. Heat maps were created using Excel to map the percent changes in total prescription distribution of amphetamine and lisdexamfetamine from 2015 through 2018 and from 2015 through 2017 for methylphenidate. Using data from, the raw dose in milligrams was totaled for each drug according to year and state to determine the total quantity in milligrams of amphetamine, methylphenidate, and lisdexamfetamine during the years of 2015 through 2018. The mass of the stimulants in milligrams was similarly corrected from population estimates from the U.S. Census Bureau, Population Division for each state. Percent changes between consecutive pairs of years were calculated for each drug. Figures were created using GraphPad Prism to plot the mean illicit distribution with standard error of mean (SEM) for each drug from 2015 through 2018. Paired t-tests were completed using GraphPad Prism to analyze the statistical significance of yearly changes. For each drug’s ARCOS data, a paired t-test was conducted to compare the daily dose value of each year, 2015 through 2018, relative to 2014 values. Paired t-tests were also conducted using the total distributed mass of each drug, comparing the 2015 through 2018 values to the 2014 value. The exception was methylphenidate, for which the final year tested was 2017. For each drug’s StreetRx data, paired t-tests were employed


Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018

Figure 1. Total prescription distribution in grams of lisdexamfetamine (A), amphetamine (B), and methylphenidate (C) for all 50 states from 2014 through 2017/18 as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System.

Figure 2. Percent change in lisdexamfetamine distribution from 2015 through 2018. The total mass in grams of lisdexamfetamine distributed as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System has been corrected for population.

Figure 3. Percent change in amphetamine distribution from 2015 through 2018. The total mass in grams of amphetamine distributed as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System has been corrected for population.

for analysis of the total diverted drug mass, comparing the 2016 through 2018 values to the 2015 value. Additionally, the percent change calculation comparing the 2015 value to the 2018 value was done using both the ARCOS and the StreetRx data for each drug.

Results Total mass distribution and daily dosage (ARCOS data set)

Figure 4. Percent change in methylphenidate distribution from 2015 through 2017. The total mass in grams of methylphenidate distributed as reported by the Drug Enforcement Administration’s Automated Reports and Consolidated Orders System has been corrected for population. 36

Lisdexamfetamine prescription distribution increased by +8.2% on average from 2015 through 2018 (Figure 1A). Wisconsin had the largest increase in prescription distribution, +39.6% (Figure 2). Amphetamine prescription distribution increased by an average per state of +15.2% from 2015 through 2018 across all 50 states (Figure 1B). Arizona saw the largest increase in prescription distribution, +27.8% (Figure 3). Methylphenidate prescription distribution decreased by -5.5% from 2015 through 2017 (Figure 1C). New Mexico had the largest decrease, -24.3% (Figure 4). New Mexico consistently showed

Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018

Figure 5. Total diverted distribution as reported by StreetRx in milligrams of lisdexamfetamine (A), amphetamine (B), and methylphenidate (C) for all 50 states from 2015-2018.

the largest decrease in prescription stimulant distribution during the time periods studied. New Mexico prescription distribution decreased by -17.3%, -42.9%, and -24.3% for amphetamine, lisdexamfetamine, and methylphenidate respectively, and was considered a statistically significant outlier (Grubbs’ p <0.05). Lisdexamfetamine daily dose increased from 0.7410 mg/ day/person in 2014 to 0.8289 mg/day/person in 2018. This reflects the mild increase in prescription distribution of lisdexamfetamine. Amphetamine daily dose increased from 2014 to 2018, with an average across all 50 states of 2.9249 mg/day/person in 2014, compared to 3.6168 mg/day/person in 2018. In contrast, methylphenidate daily dose decreased from 2014-2017, with an average of 3.4667 mg/day/person in 2014, compared to 3.2299 mg/day/person in 2017. Total mass distribution (diverted, StreetRx) Diverted lisdexamfetamine distribution decreased by an average of -0.2376% across all 50 states (Figure 5A). Montana showed the largest decrease of -0.9375%, while Michigan showed the smallest decrease of -0.01872%. Rhode Island contrastingly showed the largest increase of +1.328%. Diverted amphetamine distribution increased by an average of +0.0417% from 2015-2018 across all 50 states (Figure 5B). Iowa showed the largest increase of +0.4779% and Connecticut showed the smallest increase of +0.0047%. In contrast, Alaska showed the largest decrease of -0.2095%. Diverted methylphenidate distribution decreased by an average of -0.3453% across all 50 states (Figure 5C). New Mexico showed the largest decrease of -0.8372%, while Kentucky showed the smallest decrease of -0.0608%. In contrast, Alaska showed the largest increase of +0.6883%. Paired t-tests comparing the total mass of illicit distribution for prescription

stimulants showed that the change in mass was statistically significant for every year compared to 2014 (p <0.0001). This included the years 2015 through 2018 for amphetamine and lisdexamfetamine and 2015 through 2017 for methylphenidate. Paired t-tests comparing the daily dose values showed that the change in daily dose was statistically significant for every year compared to 2014 (p <0.0001). This included the years 2015 through 2018 for amphetamine and lisdexamfetamine and 2015 through 2017 for methylphenidate. Paired t-tests comparing the total mass of illicit distribution for stimulants showed varying significance through the years 2015 through 2018. For amphetamine, the change in total mass of illicit distribution was statistically significant in 2016 and 2017 compared to 2015, but not in 2018. For methylphenidate, the change in total mass of illicit distribution was statistically significant in 2017 and 2018 compared to 2015, but not in 2016. For lisdexamfetamine, the change in total mass of illicit distribution was statistically significant for every

Table 1. T-Test results comparing ARCOS daily dose drug values between years 37

Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018

distribution and daily dose (mg/day/person) compared with lisdexamfetamine across 2014 through 2018. The etiology of these trends remains unclear, yet the efficacy, safety, and availability of these different pharmacotherapies may play a role. Both amphetamine and methylphenidate work by increasing dopamine and norepinephrine synaptic availability in cortical and subcortical regions which are associated with ADHD pathology (8). A possible explanation for why amphetamine-based therapy is expanding could be due to superior efficacy. A 2010 meta-analysis showed that amphetaminebased treatments for ADHD proved to be significantly more efficacious than methylphenidate-based treatments (22). When asked to evaluate the potential of Table 2. T-Test results comparing ARCOS total prescription distribution between years methylphenidate to be added to the World Health Organization’s List of Essential Medications, uncertainties about the benefit-to-harm ratio resulted in a negative decision which was unanimous (11). Furthermore, lisdexamfetamine prescription distribution may be increasing because it is potentially a safer version of amphetaminebased therapy. Lisdexamfetamine is a prodrug which must be processed in the liver to transform into the therapeutically active metabolite, d-amphetamine (9). Many parameters determine the abuse potential of a substance. One of these parameters is the time to reach peak plasma concentration, Tmax (23). It is commonly known that drugs of abuse are intentionally administered via routes that decrease Tmax (e.g., insufflation, intravenous). Lisdexamfetamine cannot be administered Table 3. T-Test results comparing StreetRx total Illicit distribution between years via a route that bypasses first-pass hepatic metabolism, as this step is necessary to transform it into its active metabolite. year, 2016 through 2018, compared to 2015. Results and This greatly decreases the possibility of using certain routes statistical analysis for ARCOS daily dose values and total illicit of administration that would decrease Tmax. In theory, a distribution are summarized in Tables 1 and 2, respectively. misuser could attempt to convert the lisdexamfetamine to Statistical analysis for StreetRx total illicit distribution is d-amphetamine via incubation with a proteolytic enzyme like summarized in Table 3. trypsin that cleaves at the carboxy terminal of lysine residues. This would prepare the drug for administration via routes Discussion which sidestep first pass hepatic metabolism. Our group has not found any reports of this sort of at-home chemistry The average amphetamine prescription distribution taking place. However, misuse by taking multiple oral doses increased across all 50 states from 2014 through 2018. continues to be reported (19). When compared with instant Lisdexamfetamine prescription distribution increased from release d-amphetamine, lisdexamfetamine has been shown to 2014 through 2017; however, it showed a very slight decrease have a lower score on the Drug Rating Questionnaire-Subject in 2018. In contrast, methylphenidate prescription distribution Liking Scale in patients with a history of stimulant abuse (24). decreased throughout all years. Methylphenidate daily dose Our group has shown that stimulant diversion increased from was higher than amphetamine daily dose in 2014, but this 2015 through 2018. Increasing diversion and stimulant abuse is trend reversed by 2017. Amphetamine daily dose was higher cause for concern. Clinicians may be more likely to prescribe than methylphenidate daily dose in 2017. Amphetamine a drug with a lower misuse profile such as lisdexamfetamine, and methylphenidate both had a higher total prescription


Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018

which may explain the increase seen in lisdexamfetamine prescription distribution, or this may be due to increased usage for binge eating disorder. The patterns of illicit stimulant use per the StreetRx data are vaguer and more modest compared to the trends in prescription stimulant distribution obtained from ARCOS. The trends in amphetamine and methylphenidate diversion paralleled one another, with an increase from 2015 through 2016, followed by a subsequent decline through 2018. Lisdexamfetamine, however, showed a consistent decline from 2015 through 2017, followed by an increase in 2018. The average mass of diverted amphetamine was greater in 2018 compared to 2015. Conversely, the average mass of diverted methylphenidate and lisdexamfetamine were smaller in 2018 versus 2015. This indicates a decrease in the misuse of methylphenidate and lisdexamfetamine. The decrease in methylphenidate misuse could potentially be explained by overall phasing out of methylphenidate due to lower efficacy (17). Although the United States federal government defines all three stimulants as Schedule II Controlled Substances, it is known that lisdexamfetamine carries a lower abuse profile (19). Additionally, lisdexamfetamine has been shown to have a lower score on the Drug Rating QuestionnaireSubject Liking Scale (24). This could explain the downward trend in lisdexamfetamine misuse. The superior efficacy of amphetamine could be the major reason why both amphetamine prescription distribution and illicit distribution increased overall from 2015 through 2018. Additionally, the total milligrams per population values for amphetamine are much larger than the values for methylphenidate and lisdexamfetamine. It is difficult to understand illicit drug use to the same extent at which prescription drug use can be understood, but it is vital to a community’s health and well-being to attempt analysis and to hypothesize possible etiologies. The trends found in illicit stimulant use in this study can, in a way, be explained by the trends in prescription stimulant use. While the trends do not directly mirror one another, inferences can be made as to why these variations are seen. Limitations One limitation in this study was the lack of complete 2018 drug summary report data for the drugs analyzed. As mentioned previously, 2018 data was unavailable from ARCOS reporting for methylphenidate. Having a lack of data for methylphenidate made it difficult to compare trends among the drugs for 2018, the most current data set available. Further studies will need to be done comparing amphetamine, methylphenidate, and lisdexamfetamine with complete data for 2018. The ARCOS dataset does not provide the indication for use. Other data sources will be needed to determine whether there were changes in use for specific indications or among demographic groups, including for obesity or binge eating disorder, for adult ADHD, including during pregnancy, or among 2- to 5-year-olds (6). Another limitation is that is user-submitted data, therefore, potential trends may not have been elucidated due to lack of self-reporting or inaccurate submissions.

Socioeconomic status and user accessibility likely influence the amount of submissions received from various U.S. regions. It should also be taken into consideration that is only able to collect the data that the public is willing to submit. It may seem unusual to some to disclose information regarding one’s own participation in illegal activity.

Conclusion In recent years, the average prescription distribution and daily dose of lisdexamfetamine and amphetamine have increased throughout the United States. The prescription distribution and daily dose of methylphenidate contrastingly decreased during this time period. States and regions of the country varied greatly in the degree of total prescription distribution change of each stimulant medication. New Mexico showed the largest decrease in all three stimulants studied and proved to be an outlier upon statistical analysis. We hypothesize that amphetamine prescription distribution may have increased due to its superior efficacy and lisdexamfetamine prescription distribution may have increased because it is potentially a safer version of amphetamine-based therapy. We also hypothesize that amphetamine illicit distribution may have increased due to effectiveness reasons, while methylphenidate and lisdexamfetamine illicit distribution may have decreased due to decreased availability and less abuse potential, respectively. Further research may unveil possible additional reasons for the trends found in this analysis.

Acknowledgments We would like to thank the Biomedical Research Club at Geisinger Commonwealth School of Medicine for their guidance and assistance throughout the development of this project.

Disclosures BJP is part of osteoarthritis research team funded by Pfizer. The other authors have no disclosures.

References 1.

Xu G, Strathearn L, Liu B, Yang B, Bao W. Twenty-year trends in diagnosed attention-deficit/hyperactivity disorder among US children and adolescents, 1997-2016. JAMA Netw Open. 2018;1(4):e181471-. Available from: doi:10.1001/ jamanetworkopen.2018.1471

2. Davidovitch M, Koren G, Fund N, Shrem M, Porath A. Challenges in defining the rates of ADHD diagnosis and treatment: trends over the last decade. BMC Pediatr. 2017;17(1):218. Available from: DOI 10.1186/s12887-017-09710 3. ADHD: clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit/hyperactivity disorder in children and adolescents. Pediatrics. 2011;128(5):1007. Available from: doi:10.1542/peds.20112654 4. De Sousa A, Kalra G. Drug therapy of attention deficit hyperactivity disorder: Current trends. Mens Sana Monogr. 2012;10(1):45. Available from: doi:10.4103/0973-1229.87261


Changes in Prescription and Illicit Stimulant Use in the United States, 2014 to 2018

5. Piper BJ, Ogden CL, Simoyan OM, Chung DY, Caggiano JF, Nichols SD, et al. Trends in use of prescription stimulants in the United States and Territories, 2006 to 2016. PloS One. 2018;13(11). Available from: doi: 10.1371/ journal.pone.0206100

16. Huskinson SL, Naylor JE, Rowlett JK, Freeman KB. Predicting abuse potential of stimulants and other dopaminergic drugs: Overview and recommendations. Neuropharmacology. 2014;87:66-80. Available from:

6. Schwarz A. Thousands of Toddlers Are Medicated for A.D.H.D., Report Finds, Raising Worries [Internet]. The New York Times. The New York Times; 2014 [cited 2020May13]. Available from: among-experts-scrutiny-of-attention-disorder-diagnosesin-2-and-3-year-olds.html

17. United States Drug Enforcement Administration. Drug Scheduling [Internet]. U.S. Department of Justice; 2020 [updated 2020 April 20]. Available from: https://www.dea. gov/drug-scheduling


Chai G, Governale L, McMahon AW, Trinidad JP, Staffa J, Murphy D. Trends of outpatient prescription drug utilization in US children, 2002–2010. Pediatrics. 2012;130(1):23-31. Available from: DOI: https://doi. org/10.1542/peds.2011-2879

8. Faraone SV. The pharmacology of amphetamine and methylphenidate: relevance to the neurobiology of attention-deficit/hyperactivity disorder and other psychiatric comorbidities. Neurosci Biobehav Rev. 2018;87:255-70. Available from: neubiorev.2018.02.001 9. Goodman DW. Lisdexamfetamine dimesylate (Vyvanse), a prodrug stimulant for attention-deficit/hyperactivity disorder. P&T. 2010;35(5):273-287. 10. Jensen PS, Hinshaw SP, Swanson JM, Greenhill LL, Conners CK, Arnold LE, et al. Findings from the NIMH Multimodal Treatment Study of ADHD (MTA): implications and applications for primary care providers. J Dev Behav Pediatr. 2001;22(1):60-73. 11. Storebø OJ, Gluud C. Methylphenidate for ADHD rejected from the WHO Essential Medicines List due to uncertainties in benefit-harm profile. BMJ Evid Based Med. 2020; pii: bmjebm-2019-111328. Available at https:// doi: 10.1136/bmjebm-2019-111328. 12. Curtin K, Fleckenstein AE, Keeshin BR, Yurgelun-Todd DA, Renshaw PF, Smith KR, et al. Increased risk of diseases of the basal ganglia and cerebellum in patients with a history of attention-deficit/hyperactivity disorder. Neuropsychopharmacology. 2018; 43(13):2548-2555. Available from https://doi: 10.1038/s41386-018-0207-5. 13. Wang GJ, Volkow ND, Wigal T, Kollins SH, Newcorn JH, Telang F, et al. Long-term stimulant treatment affects brain dopamine transporter level in patients with attention deficit hyperactive disorder. PloS One. 2013;8(5): e63023. Available from https:// doi: 10.1371/journal.pone.0063023. 14. Bhattarai J, Sumerall S. Current and future treatment options for narcolepsy: a review. Sleep Science. 2017;10(1):19. Available from: doi: 10.5935/19840063.20170004 15. Lakhan SE, Kirchgessner A. Prescription stimulants in individuals with and without attention deficit hyperactivity disorder: misuse, cognitive impact, and adverse effects. Brain Behav. 2012;2(5):661-77.Available from: doi: 10.1002/ brb3.78


18. Inciardi JA, Surratt HL, Kurtz SP, Burke JJ. The diversion of prescription drugs by health care workers in Cincinnati, Ohio. Subst Use Misuse. 2006;41(2):25564. Available from: https://doi-org.gcsom.idm.oclc. org/10.1080/10826080500391829 19. Wood D. Drug diversion. Aust Prescr. 2015 Oct;38(5):164. Available from: doi: 10.18773/austprescr.2015.058 20. Rocky Mountain Poison and Drug Safety. RADARS® System. Denver, CO: Denver Health; 2020. Available from: 21. Vaddadi SM, Czelatka NJ, Guitierrez BD, Maddineni BC, Torres-Teran CD, Tron DN, et al. Rise and Regional Variations in Schedule II Stimulant Use in the United States. medRxiv. 2020. 22. Faraone SV, Buitelaar J. Comparing the efficacy of stimulants for ADHD in children and adolescents using meta-analysis. Eur Child Adolesc Psychiatry. 2010;19(4):353-64. Available from: DOI 10.1007/s00787009-0054-3 23. Assessment of Abuse Potential of Drugs: Guidance for Industry. Silver Spring, MD: U.S. Department of Health and Human Services Food and Drug Administration, Center for Drug Evaluation and Research (CDER); 2017. 34. Available from: 24. Jasinski DR, Krishnan S. Abuse liability and safety of oral lisdexamfetamine dimesylate in individuals with a history of stimulant abuse. J Psychopharmacology. 2009;23(4):419-27. Available from: https://doi-org.gcsom.

Scholarly Research In Progress • Vol. 4, October 2020

Pharmacological Progress of Biased Agonism of the μ Opioid Receptor Laura A. Christman1* and Courtney J. Merwin1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Traditional μ opioid receptor agonists have been heavily researched, largely due to the negative side effects observed in opioid drugs such as morphine and oxycodone. The opioid epidemic continues to plague the United States and without a solution, progress will be unable to proceed. Recently, biased agonism has come to the forefront of research as an alternative method of opioid development to optimize pain relief and minimize addictive properties with other negative side effects. This review evaluates recent evidence of the development of opioid pharmaceuticals demonstrating biased agonism. These include herkinorin, mitragynine, PZM21, and TRV130 or oliceridine. Biased agonism is a promising approach for drug development; however, it is still unclear if it will truly improve the therapeutic index of the analgesic effects while minimizing adverse effects in vivo.

Introduction Opioid medications have been available for centuries to treat a variety of maladies; however, they remain in clinical practice due to their unmatched effects of pain relief (1–2). Over time, the adverse effects of this medication class became apparent through abuse, dependency, and addiction arising in patients seeking to ease acute or chronic pain. In the United States, this has led to a nationwide epidemic of opioid substance abuse. According to the 2015 National Survey on Drug Use and Health, an estimated 37.8% of the United States population self-reported using prescription opioids that year, 4.7% of the population self-reported misusing prescription opioids, and

0.8% of the population reported an opioid misuse disorder (2). As of 2019, these rates have been rising exponentially because it has not yet been possible to achieve the same level of analgesia opioids provide without adverse effects (1). The prevalence of opioid use and misuse in addition to opioid-related harm and death have generated great concern throughout the country. This has led to various interventions by the justice system, social services, public health officials, and novel developments in neuropharmacological research. Opioid pharmaceuticals already in clinical practice and new drugs in development are being analyzed to enhance understanding and potentially optimize their physiological action on a molecular level. The International Union of Basic and Clinical Pharmacology (IUPHAR) defines the receptors that interact with endogenous and exogenous opioids as opioid peptide receptors (OP) (4). The first of these receptors to be discovered was defined as the μ opioid peptide receptor, abbreviated MOP or MOR, and named for its affinity for morphine (5–6). MOR is distributed throughout the central nervous system, most notably in the periaqueductal gray, raphe magnus nuclei, thalamus (brain areas associated with the sensation and processing of painful stimuli – locus coeruleus) and in the superficial laminae of the dorsal horn of the spinal cord (3, 7–8). Opioid consumption triggers dopaminergic neurons in the nucleus accumbens, which activates feelings of reward and motivation, creating conditions susceptible to opioid abuse, withdrawal, dependency, and addiction (9). Over time, more evidence is being discovered to no longer classify opioid receptor agonist

Figure 1. This diagram shows an example of biased agonism signaling activating either G protein- or arrestin-dependent signaling (50).


Pharmacological Progress of Biased Agonism of the μ Opioid Receptor

pharmaceuticals by their “strength,” but by the different molecular pathways that are activated when the drug binds to the receptor. A concept being investigated recently with the goal of producing analgesia while reducing opioid-related side effects like respiratory depression is biased agonism. Even though it has been known by many different names since its application to opioid receptors in 2007 (10), biased agonism has shown potential to be a novel pharmacological treatment to aid in the opioid epidemic. Biased agonism describes when a ligand preferentially binds and activates one of multiple signaling pathways over another (11). An example of biased agonism involves the activation of the G-protein MOP and its altered recruitment of β-arrestin (Figure 1). Initially, the rationale for biased agonism stemmed from experiments that used β-arrestin knockout mice in which MOP activation produced analgesia, but lacked some of the negative associated effects of opioid administration (12). Studies show a range in bias of G-protein and β-arrestin recruitment, from partial to full agonism of MOP, with a strong correlation of a large therapeutic window for G-protein biased ligands (13). It is difficult to produce a specific calculation of bias; however, an approach that generates an array of biases in a single scaffold is useful for translation of studies from in vitro to in vivo (12). Ultimately, biased agonism is a promising approach for analgesic drug discovery, but more evidence needs to be gathered that analyzes the pharmacokinetics, antinociception, and side effects before any biased agonists can be incorporated into the treatment of pain. Like all medications, opioids have a wide array of adverse reactions. The major side effects in question are the severe gastrointestinal distress and respiratory depression; however, many individuals experience other mild adverse side effects. Another side effect to note is hyperalgesia, which was initially discovered when studying morphine in 1987 (14). The β-arrestin signaling, downstream of the μ opioid receptor is linked to both respiratory depression and constipation but is not related to the addictive properties of opioids (15). The dopamine reward pathway is modulated via the recruitment of β-arrestin, therefore it should be thought to decrease addictive properties. In vitro studies support β-arrestin’s involvement with dopamine receptors one and two; however, studies in vivo have given a more complex picture (16). In knock-out mice, lack of β-arrestin results in reduced or minimal behavioral responses to certain drugs that affect the dopamine reward system, meaning that it is potentially involved in other cellular processes (16). Despite completion of studies to evaluate the involvement of the dopaminergic pathway, more refined models are needed to further investigate the involvement of dopaminergic pathway and MOP signaling in addition to biased agonism (17). Research provides substantial data that displays a correlation between biased agonism of MOP and decreased opioidrelated adverse effects. Recent developments in MOR biased agonists have produced four potential candidates to enter into pharmaceuticals. Herkinorin, mitragynine, PZM21, and TRV130 (oliceridine) provide promising research in animal models, but more research will need to be completed in clinical trials for each before they are considered for integration into medical practice.


Methods A literature review was conducted using Google Scholar and PubMed using combinations of the following terms: biased agonism, bias factor, mu opioid receptor, herkinorin, mitragynine, TRV130, oliceridine, PZM21, beta-arrestin, kratom (the plant from which mitragynine is derived). These therapies were included based on their actions as biased agonists and their significant exclusivity or notable selectivity for the μ opioid peptide receptor. If potential therapies did not meet these criteria or did not have sufficient evidence available, they were not included in this review.

Discussion Herkinorin Salvia divinorum is a plant that originates from the sage family and has been used in spiritual practice by the Mazteca Indians of Oaxaca, Mexico, due to its hallucinogenic experiences (18). In 2005, when analyzing the main component of S. divinorum, Salvinorin A, its analog herkinorin was discovered (18). Herkinorin is a neoclerodane diterpene that is the result of de-acetylation of salvinorin A. Herkinorin has been the target of numerous studies during the last decade to analyze its analgesic effects. A 2012 experiment used the formalin test in rats in order to assess acute and inflammatory pain (19). Herkinorin was found to possess antinociceptive properties that are reversible with the antagonist naloxone, and activity is relatively localized to injection sites (19). Additionally, tolerance testing has been performed in rats to analyze if herkinorin would produce similar tolerance that is seen in traditional opioids like fentanyl or morphine. Traditional opioids resist tolerance after repeated paw injections due to chronic inflammation, so in order to adequately assess the tolerance potential of herkinorin, a secondary approach was taken using a systemic pump (20). Following systemic treatment, morphine tolerance was induced in the animals whereas herkinorin maintained antinociceptive properties and effectively reduced formalininduced flinch responses (20). The formalin test functions as a measure of tonic inflammatory pain, with an increased flinch response displaying that the animal has developed tolerance to a previously effective dose of the drug in question (21). Primarily, herkinorin causes antinociception ipsilaterally, making it unclear whether there is selective activity in the periphery, or if it is unable to cross the blood-brain barrier, meaning that more studies involving its pharmacokinetics are necessary (20). Herkinorin does not cause recruitment of β-arrestin-2 to the μ opioid receptor and does not result in internalization, leading to the question of its antinociceptive effects in vivo. More recently, a 2019 literature review from Frontiers in Psychiatry has noted the significance of herkinorin’s ability to promote the phosphorylation of ERK1/2 without recruitment and signaling of β-arrestin-2 (21). This article notes that the lack of internalization of the μ opioid receptor could potentially improve the adverse drug effects that are seen in traditional opioids like morphine, making it a potential analgesic (21). While other Salvinorin A derivatives are more selective for the μ opioid receptor than herkinorin, it is important to note that

Pharmacological Progress of Biased Agonism of the μ Opioid Receptor

herkinorin has the highest analgesic potential for application (21). It should also be recognized that while herkinorin is in fact a biased agonist at the μ opioid receptor, it is a full agonist at the κ opioid receptor (15). In the future, more work needs to be done in the studies of herkinorin to ensure that it does not induce physical dependence. In 2018, the FDA publicized the potential of herkinorin as a potential analgesic, but it has not yet gone through clinical trials (22). In order to further understand the potential for pharmacological use, herkinorin will need to undergo various clinical trials that analyze the potential for abuse and dependence in humans. So far, herkinorin is a promising alternative for opioid therapy in the future. Mitragynine Mitragynine has been used for over 100 years in traditional medicine as it originates from the evergreen tree Mitragyna speciosa in the jungles of Southeast Asia (23). Mitragynine is a terpenoid indole alkaloid that has primarily been used in the illicit market; however, recent studies prove hopeful for its abilities to become a recognized pharmacological drug. While the exact metabolic mechanism is unknown, there are multiple possibilities to explain its metabolism. Through oxidation, [bis(trifluoroacetoxy)iodo]benzene acts on mitragynine to produce 7-OH (23). It is also hypothesized that CYP3A is able to function in this conversion as well, and the varying degrees of conversion explain some of the variance in the literature. After it was discovered that CYP3A was the primary metabolic pathway for conversion in enzyme preparations, mitragynine was incubated in human liver microsomes (HLM) alone and in the presence of ketoconazole, a CYP3A inhibitor (23). The presence of the ketoconazole inhibited the degradation of mitragynine to 7-OH in both the

human and mouse liver microsomes. Additionally, it was found that while there was significant conversion of mitragynine to 7-OH in the HLM, mitragynine had low stability in the HLM (23). This contradicts a 2014 analysis that found mitragynine to be metabolically stable in HLM and S9 fractions. It was found to be unstable in simulated gastric fluid, but stable in simulated intestinal fluid in this study (23). The researchers concluded that the data could indicate the possibility of drug interactions if administered with drugs that are P-glycoprotein substrates (24). The conversion of mitragynine to 7-OH has also been examined in-vivo in mice (23, Figure 2). It was found that both mitragynine and 7-OH were found in plasma and the brain, but 7-OH was only a minor metabolite (23). More evidence needs to be collected to determine the pharmacokinetics of mitragynine, especially to determine the role of 7-OH as a potent analgesic in humans. Mitragynine and its analog, 7-hydroxymitragynine (7-OH), are both biased agonists of the μ opioid receptor and potentially have a larger therapeutic window between analgesia and the adverse side effects of traditional opioids (23). It is suspected that the severe respiratory depression that is seen in other opioids would not occur in mitragynine, as the 7-OH metabolite is seen to be a partial G-protein biased agonist at the μ opioid receptor (23). Mitragynine functions as a partial agonist of the human μ opioid receptor, and a competitive antagonist at the κ and δ opioid receptors (25). A lack of β-arrestin recruitment after receptor activation allows for a potentially improved therapeutic profile. This is due to a binding position that differs from classic opioids, creating a safer in vivo profile (11). Promising research portrays the potential for pharmacological use of mitragynine; however, controlled clinical trials will still be necessary to determine the safety profile of the drug in vivo.

Figure 2. Molecular structures and metabolic transformations of mitragynine (23).


Pharmacological Progress of Biased Agonism of the μ Opioid Receptor

PZM21 Developed in 2016 via structure-based optimization, PZM21 is a compound that exhibits biased agonism at the μ opioid receptor (15). This experiment used structure-based docking to the μ opioid receptor which visualized the bonding between the inactive receptor and compounds under investigation (15). It was found that PZM21 has a double hydrogen bond coordination not seen in previously studied opioid ligands, and was able to activate the receptor with low levels of β-arrestin 2 recruitment (15). PZM21 is selective for MOP after comparison to 316 other GPCRs despite its structural dissimilarities to traditional opioid drugs (15, 26). Based upon ligand bias calculation, PZM21 has undetectable β-arrestin recruitment to the μ opioid receptor (15, 25). The PZM21 compound has undergone various alterations to its chemical structure to determine which components deplete or enhance its activity as a biased MOR agonist. Following its creation, it was determined that the benzodioxolane group within the structure of PZM21 is required to prevent β-arrestin recruitment and μ opioid receptor selectivity (28). However, in a subsequent study (29), it was determined that replacing the hydroxyl group of a phenol with naphthalene in PZM21 would further decrease β-arrestin recruitment while maintaining selective agonist activity at MOP (Figure 3).

transporters in addition to inhibition of the hERG ion channel (15). By inhibiting neurotransmitter transit, the reward pathway is inhibited, thus preventing opioid-induced reinforcement. A conditioned place preference response test was administered to confirm inhibition of reward and reinforcement activation in rat behavior (15, 30). hERG channel inhibition is an increasingly important safety component in development of novel pharmaceuticals (31). hERG activation is associated with potassium channel opening in the heart, causing dangerous arrhythmias that can be visualized as prolonged QT segments on an electrocardiogram (31). No adverse effects of respiratory depression were reported in the rats following PZM21 administration (15). In comparison to morphine, PZM21 was considered to be a more effective analgesic and produced negligible side effects upon initial analysis (15). Despite promising results from its inception, there is a lack of evidence in initial studies regarding the chronic effects of PZM21 administration. One subsequent analysis (32) associated systematic PZM21 with low doses inducing hyperalgesia and both high and low dosages causing the development of chronic pain over time, called hyperalgesic priming. However, hyperalgesia at doses much lower than therapeutic doses is a common phenomenon associated with opioids, initially discovered by studying morphine in 1987 (14). Based on chemical analysis of PZM21 (28), the future of biased μ opioid receptor agonists is dependent upon compounds which contain benzodioxolane and alkyl benzene groups. These modifications have limited, but promising evidence towards production of maximized analgesic effects and enhanced ligand bias. As of 2020, there are no active clinical trials for PZM21 being completed through the NIH. TRV130 (Oliceridine)

Figure 3. Structure of PZM21 before and after naphthalene replaces the hydroxyl group on ring A (29).

The effects of PZM21 have not been tested in humans; however, several studies have been conducted to determine its functions in rats. PZM21, like other biased μ opioid receptor agonists, was developed with the intention to decrease opioidrelated adverse effects including respiratory depression, constipation, hyperalgesia, dependency, and addiction. PZM21 is able to cross the blood-brain barrier and produces analgesia without altering spinal reflexes as determined by a tail-flick assay in rats (15). Upon molecular analysis, PZM21 inhibited serotonin, dopamine, and norepinephrine neurotransmitter


TRV130 was first introduced by pharmaceutical company Trevena in 2013 to minimize respiratory and gastrointestinal side effects of opioids for the treatment of postoperative pain (33). When compared to morphine, TRV130 had 86% lower β-arrestin recruitment to MOP (33). TRV130 is thought to be MOP selective based on a stabilizing interaction with asparagine at position 2.63 (34). κ and δ opioid receptors contain valine and lysine in the same position respectively and do not produce the same stabilization (34). This provides additional context regarding ligand binding interactions and allosteric modulation of MOP. TRV130 progressed into further clinical trials due to improved opioid-related side effects such as respiratory distress and decreased gastrointestinal motility, compared to the effects of morphine observed in mice and rats (33). An additional experiment was completed using rats to further explore the therapeutic index of analgesia and respiratory suppression, adequate hERG inhibition, and to refine the chemical structure of TRV130 (35). TRV130 was first administered to humans in 2013 for initial determination of dosage, tolerability, pharmacokinetics, and pharmacodynamics (36). CYP2D6 was identified as a cytochrome complex with variability which may affect the clearance of TRV130 from the body (36). In a subsequent analysis involving humans, TRV130 was determined to have quicker, more efficacious analgesic effects as compared to morphine or placebo at doses of 3 milligrams and 4.5 milligrams (37). The decreased engagement and internalization of β-arrestin along with decreased receptor phosphorylation was found to result in reduced

Pharmacological Progress of Biased Agonism of the μ Opioid Receptor

respiratory depression and gastrointestinal dysfunction when initially studied in rodents (37). When analyzing respiration in humans, TRV130 was determined to produce a transient reduction in respiratory drive, but was shorter in duration and lower in intensity as compared to morphine (37). This was a randomized, double-blinded study with 29 participants who completed the course of treatment in 2014 which included a thermal assay called a cold pain test to measure analgesia. It was as early as this publication that participants displayed dependence behaviors measured by a likability scale, demonstrating an abuse potential for TRV130 (38–40). Throughout the clinical trials, the effects of TRV130 were often compared to the effects of morphine, but not hydrocodone or oxycodone. Oxycodone has a greater potential for abuse and has a greater annual consumption in the United States as opposed to morphine (1, 41). In one analysis, the analgesic action of TRV130 was compared to that of oxycodone in rats exposed to thermal stimuli which showed similar patterns of significantly diminished pain sensation (38). This study also included comparable self-administration behaviors, which presents evidence that TRV130 may have a similar abuse potential to oxycodone (38, 39). Conversely, TRV130 administration in rats did not cause conditioned place preference, indicating a decreased activation of the dopamine-induced reward pathway and reducing opioidinduced reinforcement behavior (21, 30). It is important to note that the authors of TRV130 clinical trials have disclosed financial conflicts of interest including but not limited to consulting, employment, research funding, and owning stocks in pharmaceutical companies such as Trevena, Inc. and Lotus Clinical Research LLC (33, 35-37, 42-48). In February 2016, the FDA approved TRV130 for breakthrough status as oliceridine for intravenous administration (49). In January 2018, the FDA approved a drug application for oliceridine as a schedule II narcotic for use in a hospital or monitored clinical setting (49). FDA has declined to approve oliceridine due to concerns of its efficacy and safety which did not show improved therapeutic efficacy as compared to morphine (28, 29).


β-arrestin recruitment for pharmaceuticals ranges from 0.4-fold to 100-fold differences (13, 17). TRV130 provides an example of the risks that this lack of definition can pose to patients, thus future pharmacologic development should proceed with caution. Some suggested approaches for future research in biased agonism include improvements in methods to evaluate bias factor and β-arrestin recruitment in vivo and in methods to better predict the ratio of therapeutic benefits to adverse effects (50). The current evidence shows promise for the future of biased agonists of MOP and downregulation of β-arrestin recruitment, however this should be considered to be only a potential contributor to solving the opioid epidemic currently plaguing the United States.

Acknowledgments The authors would like to thank Brian Piper, MS, PhD for guidance and assistance throughout the review process.

Disclosures The authors have no financial conflicts of interest to disclose.

References 1.

Emery MA, Eitan S. Members of the same pharmacological family are not alike: Different opioids, different consequences, hope for the opioid crisis? Prog Neuropsychopharmacol Biol Psychiatry. 2019;92:428– 449.

2. Han B, Compton WM, Blanco C, Crane E, Lee J, Jones CM. Prescription opioid use, misuse, and use disorders in U.S. adults: 2015 National Survey on Drug Use and Health. Ann Intern Med. 2017;167(5):293–301. 3. Pasternak GW, Pan Y-X. Mu opioids and their receptors: Evolution of a concept. Pharmacol Rev. 2013;65(4):1257– 1317. 4. Borsodi A, Bruchas M, Caló G, Chavkin C, Christie MJ, Civelli O, et al. Opioid receptors (version 2019.4) in the IUPHAR/BPS Guide to Pharmacology Database. IUPHAR/BPS Guide to Pharmacology CITE. 2019; 2019(4). Available from:

Based upon the available evidence, it is unclear whether biased agonism of the μ opioid peptide receptor is correlated with optimal analgesia and minimal opioid-related adverse effects. The research completed has enhanced understanding of current and newly developed pharmacotherapies in addition to gaining knowledge about the μ opioid peptide receptor. The therapies mentioned are being tested for the treatment of acute pain and are often completed in animal models, if not cellular models. This poses a problem for implementing treatment in humans because MOP is expressed in varying tissue types and even cellular types throughout the body, therefore MOP signaling and β-arrestin involvement will change throughout the body due to varying cellular and tissue contexts.

5. Cox BM, Christie MJ, Devi L, Toll L, Traynor JR. Challenges for opioid receptor nomenclature: IUPHAR Review 9. Br J Pharmacol. 2015;172(2):317–323.

While the current evidence surrounding biased agonism of β-arrestin recruitment of μ opioid receptor is promising, more evidence regarding the definition and specific action of biased agonists is needed. Degrees of bias for MOP without

9. Laviolette SR, Nader K, van der Kooy D. Motivational state determines the functional role of the mesolimbic dopamine system in the mediation of opiate reward processes. Behav Brain Res. 2002;129(1):17–29.

6. Martin WR, Eades CG, Thompson JA, Huppler RE, Gilbert PE. The effects of morphine- and nalorphine- like drugs in the nondependent and morphine-dependent chronic spinal dog. J Pharmacol Exp Ther. 1976;197(3):517–532. 7.

Arvidsson U, Riedl M, Chakrabarti S, Lee JH, Nakano AH, Dado RJ, et al. Distribution and targeting of a mu-opioid receptor (MOR1) in brain and spinal cord. J Neurosci. 1995;15(5):3328–3341.

8. Siegel A, Sapru HN. Essential Neuroscience. 4th ed. Philadelphia: Wolters Kluwer; 2019:107–144.


Pharmacological Progress of Biased Agonism of the μ Opioid Receptor

10. Galandrin S, Oligny-Longpré G, Bouvier M. The evasive nature of drug efficacy: Implications for drug discovery. Trends Pharmacol Sci. 2007;28(8):423–430. 11. Michel MC, Charlton SJ. Biased agonism in drug discovery—Is it too soon to choose a path? Mol Pharmacol. 2018;93(4):259–265. 12. Stanczyk MA, Kandasamy R. Biased agonism: The quest for the analgesic holy grail. Pain Rep. 2018;3(3). Available from: PMC5999417/ 13. Schmid CL, Kennedy NM, Ross NC, Lovell KM, Yue Z, Morgenweck J, et al. Bias factor and therapeutic window correlate to predict safer opioid analgesics. Cell. 2017;171(5):1165-1175.e13. 14. Kayser V, Besson JM, Guilbaud G. Paradoxical hyperalgesic effect of exceedingly low doses of systemic morphine in an animal model of persistent pain (Freund’s adjuvant-induced arthritic rats). Brain Research. 1987;414(1):155–157. 15. Manglik A, Lin H, Aryal DK, McCorvy JD, Dengler D, Corder G, et al. Structure–based discovery of opioid analgesics with reduced side effects. Nature. 2016;537(7619):185–190. 16. Del’Guidice T, Lemasson M, Beaulieu J-M. Role of Beta-Arrestin 2 Downstream of Dopamine Receptors in the Basal Ganglia. Front Neuroanat. 2011;5. Available from: fnana.2011.00058/full 17. Harding WW, Tidgewell K, Byrd N, Cobb H, Dersch CM, Butelman ER, et al. Neoclerodane diterpenes as a novel scaffold for μ opioid receptor ligands. J Med Chem. 2005;48(15):4765–4771. 18. Grim TW, Acevedo-Canabal A, Bohn LM. Toward directing opioid receptor signaling to refine opioid therapeutics. Biol Psych. 2020;87(1):15-21.

23. Kruegel AC, Uprety R, Grinnell SG, Langreck C, Pekarskaya EA, Le Rouzic V, et al. 7-hydroxymitragynine is an active metabolite of mitragynine and a key mediator of its analgesic effects. ACS Cent Sci. 2019;5(6):992–1001. 24. Manda VK, Avula B, Ali Z, Khan IA, Walker LA, Khan SI. Evaluation of in vitro absorption, distribution, metabolism, and excretion (ADME) properties of mitragynine, 7-hydroxymitragynine, and mitraphylline. Planta Med. 2014;80(7):568–576. 25. Kruegel AC, Gassaway MM, Kapoor A, Váradi A, Majumdar S, Filizola M, et al. Synthetic and receptor signaling explorations of the mitragyna alkaloids: Mitragynine as an atypical molecular framework for opioid receptor modulators. J Am Chem Soc. 2016;138(21):6754–6764. 26. Kroeze WK, Sassano MF, Huang X-P, Lansu K, McCorvy JD, Giguère PM, et al. PRESTO-Tango as an open-source resource for interrogation of the druggable human GPCRome. Nat Struct Mol Biol. 2015;22(5):362–369. 27. Kenakin T, Christopoulos A. Signalling bias in new drug discovery: Detection, quantification and therapeutic impact. Nat Rev Drug Discov. 2013;12(3):205–216. 28. Ma M, Li X, Tong K, Cheng J, Yu Z, Ren F, et al. Discovery of biased mu-opioid receptor agonists for the treatment of pain. ChemMedChem. 2020;15(1):155–161. 29. Ma M, Sun J, Li M, Yu Z, Cheng J, Zhong B, et al. Synthesis and evaluation of novel biased μ-opioid-receptor (μOR) agonists. Molecules. 2019;24(2):259. 30. Tzschentke TM. Measuring reward with the conditioned place preference paradigm: A comprehensive review of drug effects, recent progress and new issues. Progress in Neurobiology. 1998;56(6):613–72. 31. Garrido A, Lepailleur A, Mignani SM, Dallemagne P, Rochais C. hERG toxicity assessment: Useful guidelines for drug design. Eur J Med Chem. 2020;195:112290.

19. Coffeen U, Pellicer F. Salvia divinorum: From recreational hallucinogenic use to analgesic and anti-inflammatory action. J Pain Res. 2019;12:1069–1076.

32. Araldi D, Ferrari LF, Levine JD. Mu-opioid receptor (MOR) biased agonists induce biphasic dose-dependent hyperalgesia and analgesia, and hyperalgesic priming in the rat. Neuroscience. 2018;394:60–71.

20. Lamb K, Tidgewell K, Simpson DS, Bohn LM, Prisinzano TE. Antinociceptive effects of herkinorin, a MOP receptor agonist derived from salvinorin A in the formalin test in rats: new concepts in mu opioid receptor pharmacology: From a symposium on new concepts in mu-opioid pharmacology. Drug Alcohol Depend. 2012;121(3):181–188.

33. DeWire SM, Yamashita DS, Rominger DH, Liu G, Cowan CL, Graczyk TM, et al. A G protein-biased ligand at the μ-opioid receptor is potently analgesic with reduced gastrointestinal and respiratory dysfunction compared with morphine. J Pharmacol Exp Ther. 2013;344(3):708– 717.

21. Zjawiony JK, Machado AS, Menegatti R, Ghedini PC, Costa EA, Pedrino GR, et al. Cutting-edge search for safer opioid pain relief: Retrospective review of salvinorin A and its analogs. Front Psychiatry [Internet]. 2019;10. Available from: fpsyt.2019.00157/full

34. Schneider S, Provasi D, Filizola M. How oliceridine (TRV-130) Binds and stabilizes a μ-opioid receptor conformational state that selectively triggers G protein signaling pathways. Biochemistry. 2016;55(46):6456– 6466.

22. FDA Anesthetic and Analgesic Drug Products Advisory Committee (AADPAC) briefing document: Oliceridine (NDA 210730), 2018. download


35. Chen X-T, Pitis P, Liu G, Yuan C, Gotchev D, Cowan CL, et al. Structure–activity relationships and discovery of a G protein biased μ opioid receptor ligand, [(3-Methoxythiophen-2-yl)methyl]({2-[(9R)-9-(pyridin-2yl)-6-oxaspiro-[4.5]decan-9-yl]ethyl})amine (TRV130), for the treatment of acute severe pain. J Med Chem. 2013;56(20):8019–8031.

Pharmacological Progress of Biased Agonism of the μ Opioid Receptor

36. Soergel DG, Subach RA, Sadler B, Connell J, Marion AS, Cowan CL, et al. First clinical experience with TRV130: Pharmacokinetics and pharmacodynamics in healthy volunteers. J Clin Pharm. 2014;54(3):351–357. 37. Soergel DG, Subach RA, Burnham N, Lark MW, James IE, Sadler BM, et al. Biased agonism of the μ-opioid receptor by TRV130 increases analgesia and reduces on-target adverse effects versus morphine: A randomized, doubleblind, placebo-controlled, crossover study in healthy volunteers. PAIN®. 2014;155(9):1829–1835. 38. Austin Zamarripa C, Edwards SR, Qureshi HN, Yi JN, Blough BE, Freeman KB. The G-protein biased mu-opioid agonist, TRV130, produces reinforcing and antinociceptive effects that are comparable to oxycodone in rats. Drug and Alcohol Dependence. 2018;192:158–162. 39. Altarifi AA, David B, Muchhala KH, Blough BE, Akbarali H, Negus SS. Effects of acute and repeated treatment with the biased mu opioid receptor agonist TRV130 (oliceridine) on measures of antinociception, gastrointestinal function, and abuse liability in rodents. J Psychopharmacol. 2017;31(6):730–739. 40. Altarifi AA, Negus SS. Some determinants of morphine effects on intracranial self-stimulation in rats: Dose, pretreatment time, repeated treatment and ratedependence. Behav Pharmacol. 2011;22(7):663–673.

46. Singla N, Minkowitz HS, Soergel DG, Burt DA, Subach RA, Salamea MY, et al. A randomized, Phase IIb study investigating oliceridine (TRV130), a novel µ-receptor G-protein pathway selective (μ-GPS) modulator, for the management of moderate to severe acute pain following abdominoplasty. J Pain Res. 2017;10:2413–24. 47. Nafziger AN, Arscott KA, Cochrane K, Skobieranda F, Burt DA, Fossler MJ. The influence of renal or hepatic impairment on the pharmacokinetics, safety, and tolerability of oliceridine. Clin Pharmacol Drug Devel [Internet]. Available from: https://accp1.onlinelibrary.wiley. com/doi/abs/10.1002/cpdd.750 48. Bergese SD, Brzezinski M, Hammer GB, Beard TL, Pan PH, Mace SE, et al. ATHENA: A Phase 3, open-label study of the safety And effectiveness of oliceridine (TRV130), a G-protein selective agonist at the µ-opioid receptor, in patients with moderate to severe acute pain requiring parenteral opioid therapy. J Pain Res. 2019;12:3113–3126. 49. Kaye AD, Cornett EM, Patil SS, Gennuso SA, Colontonio MM, Latimer DR, et al. New opioid receptor modulators and agonists. Best Pract Res Clin Anaesthesiol. 2018;32(2):125–136. 50. Conibear AE, Kelly E. A biased view of μ-opioid receptors?. Molec Pharmacol. 2019;96(5):542-549.

41. Pedersen MF, Wróbel TM, Märcher-Rørsted E, Pedersen DS, Møller TC, Gabriele F, et al. Biased agonism of clinically approved μ-opioid receptor agonists and TRV130 is not controlled by binding and signaling kinetics. Neuropharmacol. 2020;166:107718. 42. Wightman R, Perrone J, Portelli I, Nelson L. Likeability and abuse liability of commonly prescribed opioids. J Med Toxicol. 2012;8(4):335–340. 43. Viscusi ER, Skobieranda F, Soergel DG, Cook E, Burt DA, Singla N. APOLLO-1: A randomized placebo and active-controlled phase III study investigating oliceridine (TRV130), a G protein-biased ligand at the µ-opioid receptor, for management of moderate-to-severe acute pain following bunionectomy. J Pain Res. 2019;12:927– 943. 44. Viscusi ER, Webster L, Kuss M, Daniels S, Bolognese JA, Zuckerman S, et al. A randomized, phase 2 study investigating TRV130, a biased ligand of the μ-opioid receptor, for the intravenous treatment of acute pain. PAIN. 2016;157(1):264–272. 45. Singla NK, Skobieranda F, Soergel DG, Salamea M, Burt DA, Demitrack MA, et al. APOLLO-2: A randomized, placebo and active-controlled Phase III study investigating oliceridine (TRV130), a G protein–biased ligand at the μ-opioid receptor, for management of moderate to severe acute pain following abdominoplasty. Pain Practice. 2019;19(7):715–731.


Scholarly Research In Progress • Vol. 4, October 2020

Brain Death: Investigating the Electrophysiological Events at Death in Animal and Human Research Regan A. Gallagher1*, Brooke A. Goldstein1*, Jonah A. Joffe1*, Kyle A. Kidd1*, and SooYoung H. VanDeMark1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Brain death is an understudied and often misunderstood process. Identifying when death occurs in the clinical setting is not always the biological start point of when death actually occurs. Once circulation and perfusion of neuronal tissue ceases, a cascade of many processes ensues leading to eventual iso-electricity and true death. Identifying processes in the brain during active death are mostly studied in animal models, but recent studies have examined terminal patients removed from life-sustaining care. Measurement of brain electrical activity provides insight to the function of neuronal tissue, and the spreading and nonspreading of electrical depression has been studied in several patients. Among the various research articles examined in this paper, there is consensus that death is a transitional process, not a singular point in time, and that the brain is still active for a limited time after electrical silencing.

Introduction Brain death has been an elusive area of medical science throughout the ages and is only beginning to be understood in the modern era. The development of new medical practices and scanning techniques have shed light on previous misconceptions in the study of brain function and dysfunction. As our collective scientific knowledge on pre- and postneurological death improves, new definitions and clinical practices for patients’ rights need to be determined. The act of declaring brain death is a culturally unique phenomenon across the globe that is often determined by institutional and legal procedures. However, it varies greatly in its execution depending on a country’s economic status as low-, middle-, or high-income, and the existence of a national organ transplant network (1). Similarly, death has been redefined over time. Historically, the classification of death (and subsequently brain death) fell into one of three categories: decapitation or decay, non-response to painful stimuli, or the absence of respiration and pulse (2). Before life-sustaining technologies such as ventilators were developed, the practice of declaring death was much simpler than it is today. Due to our present ability to prolong life with medical machinery, it has become increasingly difficult to define when brain death has specifically occurred. The current diagnosis of brain death is multifaceted; focusing on the irreversible loss of brain function resulting in coma, absence of brainstem reflexes, and apnea (2, 3). The clinical process for determining and certifying brain death in humans involves three measures: identifying a potential cause, ruling out confounding conditions, and performing a complete physical and neurological examination (2). Diagnosis 48

begins with identifying a potential cause of brain death, such as severe head injury, hemorrhage, hypoxia, ischemia, infection, or other terminal conditions (2). A physician must also rule out any condition that may mimic or confound neurological death, such as shock, hypotension, hypothermia, and injection or ingestion of neurologic substances (e.g., drugs, toxins, or other neuro depressants). Guillain-Barré syndrome and other diseases that cause neuromuscular paralysis should also be investigated prior to a diagnosis of brain death (4). After identifying possible causes and ruling out confounders, a physical examination is performed. Identification of outside stimuli; pupillary, gag, and oculovestibular reflexes; reaction of heart rate to atropine; and respiratory movements would contradict a diagnosis of brain death (2). An apnea test is then performed following the brainstem reflex exam. A patient must first be hemodynamically stable and ventilated to normal levels. Once stable, the patient is removed from ventilation to begin the apnea test. If no spontaneous respirations are present, an arterial blood gas is drawn after 8 to 10 minutes. If the returned PCO2 value is >60 mmHg or 20 mmHg over the baseline value, it supports the diagnosis of brain death (3). Testing respiratory reflexes is generally the last test performed. The precise clinical examination performed to determine brain death was found to vary globally, specifically when conducting the apnea test (1). Some institutional protocols recommend that the apnea test should only be carried out after two failed brain stem exams, as described above, separated by at least a 6-hour period (2). The American Academy of Neurology was unable to recommend a minimally acceptable period of time for observation prior to declaring brain death (3). Determining neurologic death is a precarious event for patients, families, clinicians, and others including the organ transplant community. However, this determination is not guided by evidence-based methods but rather opinion-based ones (3) due to the lack of evidence and limitations of research on this topic. In this review, we examine the prevailing science on the dying brain immediately pre- and post-mortem within animal and human models in order to illuminate the complexity of defining and determining brain death even with the advancement of new technologies.

Methods Findings from several peer-reviewed journal articles were examined to create a cohesive understanding of the classification and current understanding of the dying brain. Journal articles were obtained from a variety of reputable sources, utilizing database searches of Google Scholar, PubMed, Cochrane Reviews, and the National Center for

Brain Death

Biotechnology Information (NCBI). Search terms included: brain death, dying brain, perfusion, neuronal death, hypoxia, ischemia, and clinical death. After the initial round of articles were collected, additional source material was found within the references of said articles.

Discussion Small animal models Given the difficulty of examining the dying brain in humans due to the delicate nature of the topic and various ethical constraints, much of what is known about this topic comes from studying animal models. In an experiment by Borjigin et al., 9 rats had electrodes placed directly onto the bilateral frontal, parietal, and occipital cortices. After a period of recovery, the rats were lethally injected to induce cardiac arrest. Electrical brain activity of the rats was recorded during wakefulness, sedation, and post-cardiac arrest using an electroencephalogram (EEG). [Note: This article refers to the use of an EEG, but based on its description, electrocorticography (ECoG) or intracranial electroencephalogram would be the more accurate term.] Electrical activity was seen for approximately half a minute after the last heartbeat in all 9 subjects. Electrical silence occurred after this period, as EEG recordings fell to under 10 µV. The collected data and analysis of the gamma, theta, and delta waves; the global coherence across the brain of activity; and the feedback-feedforward connectivity found, led the researchers to suggest that a high degree of information processing may occur during clinical death (5). They interpreted this activity as a possible scientific and physiological explanation for human near-death experiences. However, an alternative explanation for the electrical activity points to a cellular interpretation, positing that the cessation of blood flow into the brain causes an influx of calcium into cells, leading to cell damage that the EEG was recording as electrical brain activity (6). Rodent brains have also been studied at the cellular level to explore the activity of injured neurons. Typically, neurons do not divide after birth in the mammalian brain. However, new evidence based on the presence of cell cycle associated proteins has shown that some neurons can reenter the cell cycle after brain injury (7–9), passing the G1/S-phase checkpoint and resuming DNA synthesis. The ability for damaged cells to pass this checkpoint and synthesize new DNA is dependent upon how long the apoptotic cells can survive (10, 11). As brain tissue becomes hypoxic and dies, glutamate is released by neurons triggering the release of inflammatory cytokines and producing a small-scale positive feedback loop (12). Cytokine upregulation intensifies at the injured area. Hypoxic conditions activate anaerobic metabolism and cause a decline in synaptic signaling (13) as well as the reentry by neurons into the cell cycle (14). Adult mice were anesthetized, and there was a permanent blockage of both carotid arteries. Following sedation, the mice were placed in a controlled atmospheric chamber that was infused with 7.5% oxygen and 92.5% nitrogen for 4 hours inside a hypoxic chamber (14). Blockage of only one carotid

artery or an experience of hypoxia alone did not cause brain damage. Only with both insults, the hypoxic environment and the blockage of the carotid arteries, was enough to induce brain damage. Three days after the brain damage had ensued, the hippocampus was populated with pyknotic nuclei, which showed characteristic features of apoptosis, such as the condensing of chromatin and the preservation of the nuclear membrane. Bromodeoxyuridine/5-bromo-2'deoxyuridine (BrdU) assays were used to identify proliferating cells on days 2, 6, 10, and 14, after the hypoxic-ischemic conditions were induced. BrdU-labeled cells were present in the damaged area of the hippocampus, and not present in any non-damaged region of the brain (14). The presence of the BrdU-labeled cells in the region of brain damage indicates that hypoxic-ischemic conditions may be able to induce apoptotic neurons to synthesize DNA and reenter the cell cycle (14). However, it is important to note that there was a failure to detect phosphorylated histone, H3, after cerebral hypoxiaischemia. H3 being used as a late G2/M phase marker. This suggests that neurons that incorporate BrdU were unable to complete the cell cycle division (14). Nonetheless, the discovery of neurons entering the G1/S phase of the cell cycle triggered by hypoxia-ischemia is novel. In the future, this finding could be used for potential treatment of neurodegenerative diseases — once, if ever, completion of the cell cycle does occur. Injured neurons can undergo DNA synthesis in parts of the genome as long as these neurons do not die immediately after injury or disease. Under hypoxic and ischemic combination, neurons from adult rodents were induced to resume DNA synthesis. Reentry into the cell cycle could be an important, yet not fully understood, mechanism of treatment for Alzheimer’s disease, stroke and other neurological conditions and diseases. While there may be concern surrounding damaged neurons reentering the cell cycle, when these neuronal cells do reenter the cell cycle, they typically resume apoptoticassociated DNA synthesis (14). Large animal models Moving from rodents to larger mammal models (dogs, pigs, and primates) has provided additional insights into the processes that occur when the brain is dying and dies. In a study by Chen et al., 10 adult mongrel dogs underwent brain death using an inflated subdural balloon. Following induced brain death, all 10 subjects presented Cushing reflex, a hyperdynamic response, and diabetes insipidus (15). Cushing reflex describes an increase in both systolic and diastolic blood pressures (16). After brain death, blood pressure increased 321% and 296% from baseline values (15). A hyperdynamic response is characterized by an increase in cardiac output, contractility, and heart rate (17). This response lasted on average 13.3 minutes. Cardiac output increased the longest, returning to baseline after 90 minutes, while contractility returned to baseline values 45 minutes post-neurological death (15). Diabetes insipidus is a hormonal abnormality that causes an imbalance of fluids (18). Catecholamine levels (dopamine, epinephrine and norepinephrine) were increased in the 15 minutes immediately post-brain death and then tapered off with time. Additionally, adrenocorticotropic hormone, cortisol, and vasopressin 49

Brain Death

values were significantly decreased over the course of the experiment accounting for the fluid imbalance (15). Overall, the results in this study were consistent with previous findings of the process of brain death in animal models; and indicates animal brain death models are comparable to the human model of neurological death (15). In a separate study, scientists worked with freshly slaughtered pigs in an attempt to reperfuse dead brain tissue. Their goal was to determine if any signs of activity could be measured post-mortem (19). The swine brain tissues (n=300) were connected to a special instrument the researchers created called BrainEx. BrainEx was a custom, “pulsatile-perfusion” machine; requiring new surgical techniques in order to connect the tissue and machine for reperfusion in large animals (19). Data analysis found restoration of cellular activity and function at the 6- and 10-hour marks post-mortem (19). Six hours post-mortem, upon administering nimodipine to increase blood flow, blood vessels within the brain were able to dilate and constrict, allowing reperfusion. At that same time point, the amount of microglial and astroglial cells were at normal levels compared to control groups at 1 hour post-mortem. Hemoglobin levels at a normal concentration similar to that of live tissue were detected 10-hours post-mortem with perfusion. While sentience was in no way returned to the dead brain tissue, the results of the BrainEx experiment advances our knowledge on the functional and structural integrity of dead brain tissue in larger animals many hours after death (19). In 1974, 25 rhesus monkeys underwent complete cerebral ischemia, a lack of blood flow to the brain, caused by intrathoracic clamping of the brachiocephalic and left subclavian arteries for a period of 60 minutes (20). After an hour, the clamps were released, and the brain was reperfused for 19 subjects; resulting in a recovered group and a nonrecovered reperfusion group. A variety of physiological measures were taken over the course of 45 minutes up to 24 hours. For the monkeys who experienced reperfusion, the experiment found significant results including an increase in cerebral blood flow, mean arterial blood pressure, arterial PO2, serum potassium, thrombin time, and a decrease in cortical pH, as compared to the controls’ (before ischemia) values. For the monkeys that did not recover, they had a significantly decreased cerebral blood flow, arterial PO2, cortical pH and increased serum potassium values compared to the reperfused monkeys. Interestingly, the non-recovered monkeys showed a significant increase in blood glucose levels as compared to the control group, but not to the recovered group. Neurological activity was recorded using ECoG with electrodes placed 3 millimeters apart over the dura mater of the sensory cortex. Researchers activated the pyramidal tract by electrically stimulating the motor cortex. During ischemia, electrical silencing occurred quickly in seconds to a few minutes. After reperfusion, 11 of 19 subjects were able to regain pyramidal response and electrical activity; suggesting the return of neuronal function (20). Ultimately, the research by Hossmann et al. found that no major neurological damage resulted if complete blood flow was restored after the ischemic event, which was consistent with findings on cats and dogs.


In another experiment involving 8 rhesus monkeys, the whole head of one monkey was detached from the cervical spine and then attached to the spinal cord of a different decapitated monkey to investigate the continuation of cerebral functioning. The researchers posited that prior brain transplants were unsuccessful because the brain was unable to regenerate in the host body to receive and send sensory input. However, in this experiment, the key was keeping the cranial nerves and the vasculature as intact as possible, so the brain within the head could remain functional (21). Before and after the procedure on both the receiving and giving monkey, levels of arterial blood gas, carotid pressure, pH, and body temperature, as well as EEG levels, were recorded. A drop in blood pressure was expected due to surgery, and the researchers gave a dose of norepinephrine to prevent hypotensive shock in the body. The specimens were also put on mechanical ventilators to keep them breathing. Before the head could be put onto the new body, researchers cauterized the carotid and jugular arteries to ensure no blood loss. These two blood pathways were previously found to sustain adequate blood supply to a monkey’s brain upon dissection (21). All 8 monkeys survived for periods ranging from 6 to 36 hours after cephalic transplantation. To test neurocognitive ability, food was placed into the subject’s mouth and attempts to chew and swallow food were made. Similarly, the monkeys were able to track objects in front of their face with their eyes. These actions suggest awareness of their environment and intact neurological functioning of the cranial nerves postsurgical procedure (21). The results from the EEG showed patterns characteristic of wakefulness such as small amplitude waves with occasional saw tooth peaks when arterial pressure remained above 40 mmHg (21). When mean arterial pressure decreased in the body, the EEG showed a corresponding decrease with reduced peaks and slower waves, but nevertheless electrical activity persisted (21). No subjects survived past 36 hours due to blood loss at the surgical site. Post-mortem, all brain areas showed adequate perfusion and uptake (confirmed with Evans Blue vital dye). Under light microscopy, there was no cellular damage to the cerebrum, cerebellum and brainstem (21). As monkeys are our closest relatives, these results have many implications on human brain death, transplantation, and future studies. Another primate experiment induced brain death via intracranial hypertension in 11 chacma baboons to analyze the mechanism of pulmonary edema after death. Eight baboons received a midline sternotomy for insertion of two electromagnetic instruments to measure flow. The other 3 baboons had Swan-Ganz catheters placed to monitor pulmonary artery pressure and pulmonary capillary wedge pressure (22). A multichannel recording system was used to simultaneously record all pressure values in order to calculate systemic vascular resistance (SVR) and pulmonary vascular resistance (PVR) (22). Baseline measurements were taken for both systemic and pulmonary hemodynamics. Subjects were sedated with ketamine, and normal saline was introduced into the subdural space to induce intracranial hypertension and subsequently brain death. Measurements were recorded

Brain Death

every 15 minutes for the first hour and then every 30 minutes for hours 2 through 7. Once finished, the lungs and heart were histologically examined (22). Most notably, an increase in circulating catecholamines occurred in the first 5 minutes after induced brain death followed by a return to baseline levels after 15 minutes and continuous reduction. Specifically, in the first 5-minute time period, epinephrine levels increased 11 times the normal amount, while norepinephrine increased threefold, and dopamine levels doubled (22). This is indicative of increased sympathetic activity (the “fight or flight” response), which is believed to be the primary reason for peripheral vasoconstriction. A cascade of events followed vasoconstriction, eventually leading to the pulmonary edema associated with brain death. Under normal conditions, the left side of the heart contains 4% of the total blood volume while the lungs contain about 20% (23). After brain death, a rise in SVR by 537% led to decreased arterial blood flow, and thus, an increase in both pulmonary and left ventricular blood volumes to about 72% of the total blood volume (22). This, combined with the low PVR in the lungs, allowed blood to pool in the lungs without being pumped back out into systemic circulation. This study identified the mechanism of post-brain-death pulmonary edema, which could be important in decreasing risks associated with organ transplantations. Using animal models to explore the dying brain has provided a repository of invaluable information regarding cellular viability, reperfusion capability, and external stimuli responses; however, the question remains: what does the human brain experience during death? Human research With such fraught legal and ethical consideration, it is not unexpected that much of our knowledge on brain death has been gained from animal models. However, while limited, it has been possible to study neurological death in humans. Understanding the various processes which occur leading up to and during active brain death can widen our understanding of neuronal complexity. The brain is highly dependent on adequate oxygen perfusion and a constant supply of nutrients in order to remain functional. Loss of oxygen leads to a state of acute hypoxia, and eventually, death of neuronal tissue (12). Glutamate is released by neurons as hypoxic tissue dies. Additionally, inflammatory cytokines such as tumor necrosis factor alpha, nuclear factor kappa B (NFκB), and interleukin 1-beta (IL-1β) are released by microglia, astrocytes, as well as neurons. The release of cytokines is initiated by hypoxiainducible factor 1-alpha (HIF-1α) binding to hormone response element (HRE) like binding sites; leading to the upregulation of cytokine release. These inflammatory agents exacerbate an already injurious situation. Oxygen sensors switch from aerobic to anaerobic metabolism due to the acute hypoxia and synaptic signaling decreases (14). Within less than 5 minutes, as much as 90% of adenosine triphosphate (ATP) in brain tissue is depleted (12). The existence of a hypoxic condition activates the signaling cascade of NFκB. Reactive oxygen species can either activate or deactivate NFκB, and expression of NFκB increases during hypoxia. Ultimately, it is the HRE binding

that directly upregulates NFκB. Once upregulated, NFκB acts on transcription pathways for inflammatory genes and HIF proteins (24). The upregulation of IL-1β cytokines leads to the activation of tyrosine kinases. Activated tyrosine kinases allow an excess flow of calcium ions and the hyperexcitability of neuronal tissue and ultimately tissue damage (25). As systemic ischemia of neuronal tissue increases, the silencing of spontaneous neuronal electrical activity has been found to happen simultaneously in both humans and animal models. This simultaneous depression of electrical activity has been termed nonspreading depression (26). Global ischemia begins a process of hyperpolarization in oxygen-starved neurons. This process happens after only 20 to 40 seconds of ischemia in affected tissue and signals the electrical shutdown process across all hypoxic areas simultaneously. This nonspreading depression gets its namesake because of the simultaneous nature of electrical silencing (26). Following 1 to 5 minutes after nonspreading depression, neurons have reached the end of their ATP reserve, starving sodium-potassium pumps of energy. The ATP starved pumps fail to restore resting membrane potential, and ultimately neurons lose their ability to maintain polarization and become iso-electric (26). As one neuron becomes iso-electric, other surrounding neurons follow suit in a domino pattern. The result is a self-propagating wave of neuron depolarization; originating either from a single point or from many. Like ripples in a pond, the depolarization spreads, giving rise to the term spreading depolarization. It is this spreading depolarization that begins the process of toxic changes, cytotoxic edema, and extracellular rise in glutamate (26). As discussed above, one of the most important factors to consider when determining brain death is cerebral ischemia (27). Computer tomography perfusion (CTP) used in conjunction with radionuclide scans were found to be the most useful in determining brain death. Researchers compared CTP, computer tomography angiogram (CTA), radionuclide scans, cerebral angiograms, magnetic resonance imaging (MRI), and nonenhanced CT (NECT) for diagnostic accuracy when investigating suspected cases of brain death. These scanning techniques were compared against the accepted standard of a clinical brain death diagnosis across 74 patients. The study found all scanning techniques except NECT to have 100% specificity; meaning the tests correctly avoided flagging a non-brain-dead individual as brain dead (aka a false positive). Among the tests performed, CTP and radionuclide scan were found to have the highest sensitivity (96.7% and 92.9% respectively) and interrater reliability (κ=1 for both techniques); correctly identifying individuals who were clinically diagnosed as brain dead (aka true positives) (27). Another study conducted a literature review of Medline and Embase from 1996 to 2009 and explored the capabilities of advanced imaging tests, including MRI, magnetic resonance angiogram, and CTA, in determining brain death. Insufficient evidence was found that these tests accurately assessed brain function cessation (3). CTP and radionuclide scanning techniques were not discussed within that paper.


Brain Death

Conclusion As technological advances have been made, the concept of brain death has become more obtuse. Advances, such as new medical practices and scanning techniques have eliminated false conclusions that have been previously made. While medical advances have been able to prolong life, they have at times made it difficult to define brain death in some patients. As previously mentioned, the clinical process for determining neurological death in humans is to identify a potential cause of brain death and rule out confounding conditions, and then perform a complete physical and neurological examination (2). It is difficult to examine the dying brain in humans due to ethical considerations, so animal models have provided much of the information that establishes our knowledge of the dying brain. Through gathering information from peer-reviewed journal articles, we were able to analyze and compare animal and human models of the dying brain. Electrical activity on an EEG showed rats having neurological activity for approximately 30 seconds after cardiac arrest. The activity and coherence in the brain was interpreted to potentially be a high degree of internal information processing occurring moments before complete neurological death (5). Although, there is controversy about how to interpret these findings, they provide one plausible scientific and physiologic explanation for human near-death experiences. Neuronal death in Alzheimer’s disease has shown that in some cases, damaged neurons are able to reenter the cell cycle. By understanding both the oxygen and blood perfusions of the dying human brain in conjunction with neuronal death found in Alzheimer’s disease and brain injury, additional research and future treatments should be considered (13). Using the various techniques that were examined in this paper, accurate diagnosis of true positive individuals enables the further exploration of Alzheimer’s disease and brain injury at the level of the dying neurons. The accurate diagnosis from CTP, CTA, radionuclide scans, cerebral angiograms, MRI, and NECT provide much needed information about an individual and the state of the human brain. It is extremely important that any diagnostic test be accurate in declaring brain dead patients from non-brain-dead patients (27). Additionally, the speed of this determination is vitally important so that organ retrieval and transplantation may occur. The study performed on mongrel dogs took a deeper dive into the hemodynamics and hormonal changes that follow immediately after brain death (15). This study found a Cushing’s reflex, followed by a hyperdynamic response and diabetes insipidus. The Cushing’s reflex was shown by an extremely large increase in both systolic and diastolic pressures. A hyperdynamic response was shown by an increase in cardiac output, heart contractility, and heart rate. Diabetes insipidus was shown with hormonal abnormalities leading to an imbalance of fluids (15). These findings explain how the body, at large, reacts to brain death. Post-mortem pig brains were tested for cellular viability and restoration (19). It was also found that the BrainEx perfused tissue showed normal pyramidal cells compared to the non-perfused samples which showed only swollen tissue. Additionally, the researchers found normal levels of astrocytes, 52

microglia, and hemoglobin in both the 6- and 10-hour postmortem tissue with the BrainEx samples compared to an in vivo brain. In these animals, when ischemic conditions were induced, electrical silencing occurred shortly afterwards. The animals were reperfused and nearly half of the subjects showed electrical activity, suggesting the return of neuronal function (19). The researchers were in fact successful in restoring both the molecular and cellular processes. While these were significant findings, they are not completely generalizable, as these findings are not associated with the same cognition of a living brain. As shown with both the small and large animals, ischemic conditions may enable nervous system cells to restore function. A study on monkeys examined complete cephalic transplantation at the cervical vertebrae (21). All monkeys survived, and they showed awareness of their environment and intact neurological functioning of the cranial nerves postsurgical procedure. Another monkey study focused specifically on cerebral perfusion and found an unexpected relationship between complete ischemia and degree of recovery (20), which potentially harkens back to the rat study on neuronal regeneration in hypoxic-ischemic conditions. Intracranial hypertension in chacma baboons showed interesting results such as an increase in circulating catecholamines that were shown shortly after brain death (22). This was followed by a return to baseline levels after about 15 minutes. Although these findings occurred, most notable was the study’s identification of the mechanism of post-braindeath pulmonary edema (22). This finding may be important in decreasing risks that are linked to organ transplants. In both animal and human models, systemic ischemia caused silencing of neuronal electrical activity. Once this occurs, the neurons will have depleted their ATP reserves, preventing the sodium-potassium pumps from functioning. As the sodiumpotassium pumps break down, neurons are forced into a state of iso-electricity and maintain polarization. This creates a ripple effect, causing the neighboring neurons to also be forced into an iso-electric state, creating the spreading depolarization phenomenon (26). Among the various research articles examined in this paper, there is consensus that death is a transitional process, not a singular point in time, and that the brain is still active for a limited time after electrical silencing. These conclusions are based on our examinations of the various research articles in this paper. With new technologies hijacking and bypassing a person’s autonomous ability to breathe and circulate blood, the definition of neurological death will need to be continually reexamined. As a society, we are far from bringing an “offline” brain back online. Eventually, there may be a time in the future where we are able to restore life to someone who previously would have met the criteria of brain death.

Acknowledgments The authors would like to thank Brian Piper, PhD, professor of Foundations of Neuroscience, and Iris Johnston in Library Services at Geisinger Commonwealth School of Medicine for their support in this endeavor.

Brain Death

Disclosures None.

References 1.

Wahlster S, Wijdicks EF, Patel PV, Greer DM, Hemphill JC, Carone M, et al. Brain death declaration: practices and perceptions worldwide. Neurology. 2015; 84(18):1870-9.

2. Goila AK, Pawar M. The diagnosis of brain death. Indian J Crit Care. 2009; 13(1):7. 3. Wijdicks EF, Varelas PN, Gronseth GS, Greer DM. Evidence-based guideline update: determining brain death in adults: report of the quality standards subcommittee of the American Academy of Neurology. Neurology. 2010; 74(23):1911-8. 4. Pritchard J. Guillain-Barre syndrome. Clin Med (Lond). 2010; 10(4):399. 5. Borjigin J, Lee U, Liu T, Pal D, Huff S, Klarr D, et al. Surge of neurophysiological coherence and connectivity in the dying brain. Pro Natl Acad Sci. 2013; 110(35):14432-7. 6. Yong E. In dying brains, signs of heightened consciousness. National Geographic. 2013. Available from: phenomena/2013/08/12/in-dying-brains-signs-ofheightened-consciousness/ [Accessed 22 January 2020] 7.

Copani A, Sortino MA, Nicoletti F, Bruno V, Ubertia D, Memo M. Activation of cell-cycle-associated proteins in neuronal death: a mandatory or dispensable path? Trends Neurosci. 2001; 24(1):25-31.

16. Cushing H. Some experimental and clinical observations concerning states of increased intracranial tension. Am J Med Sci. 1902; 124(3):375. 17. Chiari P, Hadour G, Michel P, Piriou V, Rodriguez C, Budat C, et al. Biphasic response after brain death induction: prominent part of catecholamines release in this phenomenon. J Heart Lung Transplant. 2000; 19(7):67582. 18. Fiser DH, Jimenez JF, Wrape VI, Woody RO. Diabetes insipidus in children with brain death. Crit Care Med. 1987; 15(6):551-3. 19. Vrselja Z, Daniele SG, Silbereis J, Talpo F, Morozov YM, Sousa AM, et al. Restoration of brain circulation and cellular functions hours post-mortem. Nature. 2019; 568(7752):336. 20. Hossmann KA, Zimmermann V. Resuscitation of the monkey brain after 1 h complete ischemia. I. Physiological and morphological observations. Brain Res. 1974; 81(1):5974. 21. White RJ, Wolin LR, Massopust LC, Taslitz N, Verdura J. Cephalic exchange transplantation in the monkey. Surgery. 1971; 70(1):135-9. 22. Novitzky D, Wicomb WN, Rose AG, Cooper DK, Reichart B. Pathophysiology of pulmonary edema following experimental brain death in the chacma baboon. Ann Thorac Surg. 1987; 43(3):288-94. 23. Matzen S, Perko G, Groth S, Friedman DB, Secher NH. Blood volume distribution during head‐up tilt induced central hypovolaemia in man. Clin Physiol. 1991; 11(5):411-22

8. Liu DX, Greene LA. Neuronal apoptosis at the G1/S cell cycle checkpoint. Cell Tissue Res. 2001; 305(2):217-28.

24. Eltzschig HK, Carmeliet P. Hypoxia and inflammation. N Engl J Med. 2011; 364(7):656-65.

9. Herrup K, Arendt T. Re-expression of cell cycle proteins induces neuronal cell death during Alzheimer's disease. J Alzheimers Dis. 2002; 4(3):243-7.

25. Viviani B, Bartesaghi S, Gardoni F, Vezzani A, Behrens MM, Bartfai T, et al. Interleukin-1β enhances NMDA receptor-mediated intracellular calcium increase through activation of the Src family of kinases. J Neurosci. 2003; 23(25):8692-700.

10. Yang Y, Geldmacher DS, Herrup K. DNA replication precedes neuronal cell death in Alzheimer's disease. J Neurosci. 2001; 21(8):2661-8. 11. Kuan CY, Whitmarsh AJ, Yang DD, Liao G, Schloemer AJ, Dong C, et al. A critical role of neural-specific JNK3 for ischemic apoptosis. PNAS. 2003; 100(25):15184-9.

26. Dreier JP, Major S, Foreman B, Winkler MK, Kang EJ, Milakara D, et al. Terminal spreading depolarization and electrical silence in death of human cerebral cortex. Ann Neurol. 2018; 83(2):295-310.

12. Mukandala G, Tynan R, Lanigan S, O’Connor JJ. The effects of hypoxia and inflammation on synaptic signaling in the CNS. Brain Sci. 2016; 6(1):6.

27. MacDonald D, Stewart-Perrin B, Shankar JJ. The role of neuroimaging in the determination of brain death. J Neuroimaging. 2018; 28(4):374-9.

13. Kuan CY, Schloemer AJ, Lu A, Burns KA, Weng WL, Williams MT, et al. Hypoxia-ischemia induces DNA synthesis without cell proliferation in dying neurons in adult rodent brain. J Neurosci. 2004; 24(47):10763-72. 14. Sharp FR, Bernaudin M. HIF1 and oxygen sensing in the brain. Nat Rev Neurosci. 2004; 5(6):437-48. 15. Chen EP, Bittner HB, Kendall SW, Van Trigt P. Hormonal and hemodynamic changes in a validated animal model of brain death. Crit Care Med. 1996; 24(8):1352-9.


Scholarly Research In Progress • Vol. 4, October 2020

The Neuroscience of Resilience Kelsey N. Plummer1*, Trent M. Filter1*, and Anvith Chidananda1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Resilience, a concept influenced by genetic, epigenetic, and neurophysiological factors, may offer biological protection against the diagnosis of a mental health disorder. Research identifying neurophysiological factors related to resilience has important implications in better understanding human behavior, determining therapeutic targets, and developing medical interventions. Such therapeutics may aid vulnerable individuals in combating their susceptibility to mental health disorders by optimizing their resilience.

Introduction The concept of resilience refers to one’s ability to overcome challenges and thrive in the face of adversity. Individuals who experience the same event may respond differently based on their capacity to be resilient. While some are able to learn from a trauma and grow stronger, others may experience an inability to cope, potentially resulting in the development of a mental health disorder if the incident is severe enough. This spectrum of potential reactions that varies between individuals is a result of combined genetic, epigenetic, and neurophysiological influences (1, 2). Understanding these factors and their role in determining whether one succumbs to stress or is resilient to it provides insight into human behavior. In addressing the neuroscience of resilience, progress can be made toward identifying neurophysiological factors that offer protection against mental health disorders (3). Investigating this topic may illuminate approaches to promoting resilience through medical interventions that would confer benefit to those with a greater susceptibility to stress (4).

Methods Systematic procedures were implemented to ensure a thorough, high-quality review of the scientific literature on the neuroscience of resilience. In performing a search for relevant sources, databases including Google Scholar and PubMed were consulted to find recently published peerreviewed journal articles that reference key topics such as the neuroscience of resilience and resilience versus vulnerability. Particular focus was given to studies that aimed to explain the relationship between resilience and mental health as well as to those that examined the neurophysiological factors, rather than genetic or epigenetic influences.

Discussion Anatomy and physiology Resilience involves coping, a type of behavioral adaptation that involves dynamic neurophysiological processing within the regions of the brain linked to stress, emotion, reasoning, and memory (5). These regions of the brain include the 54

hypothalamic-pituitary-adrenal (HPA) axis, the prefrontal cortex, and the amygdala. The HPA axis is an interconnected network between the hypothalamus, pituitary gland, and adrenal gland that serves in responding to stress and regulating emotion (6). The prefrontal cortex, specifically the ventromedial and orbitofrontal regions, is involved in regulating emotions, evaluating risk and fear, and making decisions. The amygdala functions in processing fear and other emotions, making decisions, and forming memories. One case report offers insight into a connection between the substantia nigra and transient acute depression (7). Identifying linkages such as these can offer targets for medically intervening to bolster resilience. In this case, a patient with Parkinson’s disease and no previous history of mental illness was being treated for her condition with deep brain stimulation of the bilateral subthalamic nuclei. Four electrodes were implanted in her brain using stereotactic guidance to place them in the appropriate region; however, one of the electrodes was inadvertently placed below the left subthalamic nucleus in the left substantia nigra. Upon high-frequency stimulation through this particular electrode, the patient experienced an episode of transient acute depression. In a matter of minutes, she cycled through an array of severely depressive emotions, even expressing a desire to stop living. The symptoms lasted only while the electrode was being stimulated and ceased less than 2 minutes after stimulation was discontinued. This case report offers unexpected findings that may be useful in the search for mechanistic pathologies related to depression; however, it would be premature to apply these findings across the board. A study aimed specifically at measuring the response of high-frequency deep brain stimulation of the substantia nigra would be beneficial in demonstrating that these findings are consistent among a larger group. While other regions of the brain have been more consistently tied to mechanistic pathologies related to mental health disorders, the importance of this case report lies in its demonstration that neurophysiological pathways are directly tied to emotion and behavior and that an alteration of these pathways can have profound effects. Targeting these pathways and identifying possible pathologies within may offer further insight into the treatment of mental health disorders through medical interventions that aim to increase resilience. Neuroplasticity Neuroplasticity refers to the brain’s ability to continually reorganize itself by creating new neural pathways to meet the needs of new functions (8). This restructuring occurs in response to stimuli in order to reinforce learning as well as in response to neural damage in order to mechanistically cope (8). The brain demonstrates learning by exhibiting neuroplasticity in response to emotional stimuli, forming positive memories as a result of positive stimuli and enhancing

The Neuroscience of Resilience

preparedness as a result of negative stimuli (9). Thanks to experience-dependent neuroplasticity, the brain can learn from fear and more readily identify threatening situations in the future (9). The remodeling of neural circuitry in one region of the brain can have direct and indirect implications on activity in other regions. Due to the interconnectedness between the HPA axis, the prefrontal cortex, and the amygdala, neuroplasticity within these regions influences one another with regard to emotion and behavior (10). Neurophysiological pathologies, as they relate to mental health disorders, have been studied to provide a better understanding of the pathways involved so that therapeutic targets can be identified. Neuroplasticity may contribute to correcting for certain pathologies, as in the case of emotional remodeling (11,12). Recalling positive memories or ruminating on negative ones can enhance the effects of those original emotions. Symptoms of post-traumatic stress disorder (PTSD) can be triggered by intrusive negative memories. As these negative memories are continually called upon, the neural circuitry used to do so is strengthened and so are the feelings of trauma surrounding the memory. This neural pathway offers opportunity for therapeutic emotional remodeling to reverse these effects, alleviating negative emotions and replacing them with positive ones. Reminiscing about positive memories reinforces positive emotions and may have a direct impact on strengthening resilience. In the case of major depressive disorder (MDD), many individuals with this diagnosis have difficulty both recalling positive memories and sustaining positive emotions as well as a diminished capacity to engage in reward processing. This indicates that there is value in the recall of positive memories, both in increasing one’s capacity to experience positive emotions and in benefiting from this action by being rewarded with increased resilience. The decreased capacity of individuals with MDD to perform this act of recall may be linked to their decreased resilience. This pathological neural pathway in those with MDD presents another opportunity for therapeutic emotional remodeling. Utilizing emotional remodeling to downregulate rumination of negative emotions and upregulate the recall of positive emotions enables those with mental health disorders to reap the benefits of increased resilience.

Immediately following this inhibitory avoidance task, the rats were subjected to an SPS. This involved restraining the rats in a plexiglass device that restricted their breathing for 2 hours and then immediately placing them in a tank where they were forced to swim for 20 minutes without rest. Next, they were given the opportunity to rest for 15 minutes before being placed in a carbon dioxide chamber until loss of consciousness was induced. After the SPS, the rats were given an opportunity to recuperate. During the entire SPS procedure, a tone was sounded and an odor was released in order to achieve specific auditory and olfactory associations with this trauma that could later be repeated to remind the rats of the trauma. A control group of rats was exposed to the same auditory and olfactory cues, but was not subjected to the trauma. Two weeks after the rats were subjected to the trauma, they were divided into groups based on anxiety profiling and were labeled as either resilient or vulnerable. Their anxiety profile was measured based on their activity in an elevated plus maze. The rats with a greater ability to explore open spaces and a decreased acoustic startle response (ASR) were placed in the resilient group. The rats with a hindered ability to explore open spaces and an increased ASR were described as having PTSD-like symptoms and were placed in the vulnerable group. On day 21 of the study, the rats were implanted with a cannula device in the right lateral ventricle so that oxytocin could be administered to the rats. On days 29 through 31, within the experimental group of rats, one subgroup received a dose of oxytocin, while another subgroup was given a saline solution. Forty minutes after the oxytocin or saline infusion, the rats were subjected to stimuli to remind them of their previous trauma and their responses were recorded.


On day 54, the rats were euthanized and their brains were removed for examination. Two markers, immunofluorescence of c-fos-labeled cells and dendritic length extension, indicated resilience versus vulnerability in the rats. These two markers were measured in both the prelimbic (PL) region and the basolateral nucleus of the amygdala (BLA). Greater immunofluorescence of c-fos-labeled cells is indicative of greater neuronal activity, and longer dendritic extensions are indicative of greater neuronal complexity and connectivity.

A recent experiment examined the effects of oxytocin on emotional remodeling in rats (11). In this study, male rats, identified as having either a resilient or vulnerable baseline, were administered oxytocin, and the resulting degree to which their resilience increased was measured using specific neurophysiological markers. The premise of this experiment is that oxytocin’s ability to reduce responsiveness to fear decreases symptomatology of PTSD and instead, promotes resilience.

Results of the study showed that the difference in the immunofluorescence of c-fos-labeled cells between the vulnerable group that received the saline solution and the vulnerable group that received the oxytocin infusion was significant (0.05 > p > 0.01) in both the PL region and the BLA. In addition, the difference in dendritic length between these same two groups was also significant (0.05 > p > 0.01). These findings indicate that oxytocin has a significant effect on increasing the resilience of vulnerable rats.

To determine which rats had a natural baseline of resilience or vulnerability, the rats were subjected to adverse stimuli to induce a stress response. Trauma was induced in 50 rats by subjecting them to an inhibitory avoidance task and a single prolonged stress (SPS). The inhibitory avoidance task involved one light compartment and one dark compartment; while the rat was safe in the light compartment, once the rat entered the dark compartment, he received two shocks.

Oxytocin, synthesized in the hypothalamus and excreted by the posterior pituitary gland, modulates emotions, reducing those associated with fear, potentially by inhibiting processing within the amygdala and instead, enhancing feelings of social trust and attachment. The ability of oxytocin to increase resilience in rats with PTSD-like symptoms is a valuable concept worth exploring in human studies. This type of experimentation, in which oxytocin is administered


The Neuroscience of Resilience

and its effects on resilience are measured, has not yet been performed in human clinical trials; however, the findings of this animal model experiment are encouraging. Inevitably, limitations may arise in translating the study from an animal model to a human clinical trial due to the fact that rat and human brains are not completely homologous.

research to include not only neurophysiological factors but genetic and epigenetic influences as well, may bring to light an even better understanding of the biological components that contribute to resilience.



A human study investigated the relationship between cortisol and resilience (13). The study evaluated stress-induced cortisol levels of 120 healthy male participants between the ages of 18 and 30. Due to the fact that cortisol levels have a normal fluctuation pattern associated with the diurnal rhythm, the study was careful to measure cortisol levels consistently in order to eliminate possible confounding factors. The participants were exposed to either a stressful or neutral movie clip and then were immediately shown images of fearful and happy faces. They were then asked to identify the emotions present in the images, while being monitored by functional magnetic resonance imaging. Results of the study showed that the participants who watched the stressful movie clip were in a stress-induced state when viewing the images and as a result, demonstrated a slower response time in identifying the emotions of the facial expressions shown in the images when compared to the participants who watched the neutral movie clip. This indicates that stress may have an effect on the ability to concentrate and perform mental tasks. Interestingly, apart from the differences between the experimental and control groups, there were also differences between individuals within the experimental group, indicating that an underlying mechanism may cause certain individuals to be more resilient to stress and others to be more vulnerable to it.

Conclusion Investigation into the neuroscience of resilience is of great importance because of the connection between resilience and mental health. While optimal resilience may serve as a protective factor, decreased resilience has been linked to individuals with mental health disorders (1–3). This concept is especially relevant in light of the predicted rise in mental health disorders as a result of the COVID-19 pandemic. Current events are inducing worldwide hardship, and those individuals that are least resilient are most at risk for developing a mental health disorder. By determining distinguishing factors between resilient and vulnerable individuals, therapeutic targets can be identified and medical interventions can be developed to optimize resilience and combat mental health disorders. Current research offers clues as to the neurophysiological factors involved in conferring resilience. Emotional remodeling in humans, which relies on the brain’s neuroplastic nature, has been shown to successfully reconfigure triggering memories by tailoring the emotions associated with them (12). Administration of oxytocin has been shown to bolster resilience in rats (11). Experiments monitoring cognition in humans during stress-induced situations reveal individuals to be either resilient or vulnerable (13). While these findings are non-competing, there are still gaps of knowledge in explaining how these findings relate to one another, and the concept of resilience remains largely abstract. Expanding this field of 56

References Han MH, Nestler EJ. Neural substrates of depression and resilience. Neurotherapeutics. 2017; 14(3):677-86.

2. Osório C, Probert T, Jones E, Young AH, Robbins I. Adapting to stress: understanding the neurobiology of resilience. J Behav Med. 2017; 43(4):307-22. 3. Holz NE, Tost H, Meyer-Lindenberg A. Resilience and the brain: a key role for regulatory circuits linked to social stress and support. Mol Psychiatry. 2019; 18:1-8. 4. Spencer-Segal JL, Akil H. Glucocorticoids and resilience. Horm Behav. 2019; 111:131-4. 5. Sinha R, Lacadie CM, Constable RT, Seo D. Dynamic neural activity during stress signals resilient coping. Proc Natl Acad Sci USA. 2016; 113(31):8837-42. 6. Siegel A, Sapru HN. Essential Neuroscience. Fourth Edition. Philadelphia: Wolters Kluwer; 2019. 7.

Bejjani BP, Damier P, Arnulf I, Thivard L, Bonnet AM, Dormont D, et al. Transient acute depression induced by high-frequency deep-brain stimulation. N Engl J Med. 1999; 340(19):1476-80.

8. Mateos-Aparicio P, Rodríguez-Moreno A. The impact of studying brain plasticity. Front Cell Neurosci. 2019; 13:66. 9. Pignatelli M, Umanah GK, Ribeiro SP, Chen R, Karuppagounder SS, Yau HJ, et al. Synaptic plasticity onto dopamine neurons shapes fear learning. Neuron. 2017; 93(2):425-40. 10. McEwen BS, Morrison JH. The brain on stress: vulnerability and plasticity of the prefrontal cortex over the life course. Neuron. 2013; 79(1):16-29. 11. Le Dorze C, Borreca A, Pignataro A, Ammassari-Teule M, Gisquet-Verrier P. Emotional remodeling with oxytocin durably rescues trauma-induced behavioral and neuromorphological changes in rats: a promising treatment for PTSD. Transl Psychiatry. 2020; 10(1):1-3. 12. Speer ME, Bhanji JP, Delgado MR. Savoring the past: positive memories evoke value representations in the striatum. Neuron. 2014; 84(4):847-56. 13. Henckens MJ, Klumpers F, Everaerd D, Kooijman SC, Van Wingen GA, Fernández G. Interindividual differences in stress sensitivity: basal and stress-induced cortisol levels differentially predict neural vigilance processing under stress. Soc Cogn Affect Neurosci. 2016; 11(4):663-73.

Scholarly Research In Progress • Vol. 4, October 2020

Chronic Traumatic Encephalopathy: Investigating Blood-Brain Barrier Disruptions and Neuroimaging Techniques YasmĂ­n Mamani1*, Payel Girwarr1*, and Sayrah Rauf1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Chronic traumatic encephalopathy (CTE) is a neurodegenerative tauopathy resulting from repeated cranial trauma that is often associated with contact sports, military duty, and vehicle collisions. Individuals afflicted with CTE experience a multitude of symptoms, such as depression, impaired judgement and decision-making, suicidality, disorientation, and dementia. On a molecular level, CTE is caused by the aggregation of insoluble tau protein isoforms, creating neurofibrillary tangles. Neurofibrillary tangles are known to result in the death of peripheral neurons and glial cells, although the direct mechanism is unclear. This article will discuss how disruption of the blood-brain barrier via repetitive traumatic brain injury influences the development of CTE. Furthermore, this article will also review the current imaging modalities used to support a probable diagnosis of CTE in a living patient. Studies indicate that exposure to repeated mild traumatic brain injury correlates with an increase in phosphorylated tau deposition of the perivascular space that parallels local disruption of the blood-brain barrier. Immunoreactivity patterns of examined CTE-positive neural tissue indicate blood vessels with high p-tau deposition lack many of the cellular components constituting the blood-brain barrier. Moreover, white matter degeneration is associated with delayed microvascular damage incurred through traumatic brain injury. Blood brain barrier integrity is dependent upon several transmembrane proteins belonging to the claudin and occludin superfamilies, and studies indicate that absence of these components correlates to blood-brain barrier degradation. Microvascular pathology observed in CTEpositive neural tissue contains decreased levels of claudin-5, occludin-1/zona occludens, suggesting that dysfunctional tight junctions contribute to the neuropathology of CTE. As with many proteopathies, CTE cannot be positively diagnosed antemortem. There are many imaging techniques that healthcare providers and researchers utilize to analyze the brain topography of a patient. Positron emission tomography scans permit clinicians to visualize the physiological state of a bodily compartment, such as the central nervous system.

Introduction Chronic traumatic encephalopathy (CTE) is a tau-protein-linked neurodegenerative condition resulting from recurrent brain trauma as illustrated in contact sports and blast injuries (1). The hyperphosphorylated tau protein (p-tau) slowly accumulates throughout the brain, leading to the death of brain cells (13). The pathophysiology of this condition is not fully understood, but there has been a correlation made between consistent head trauma and the development of CTE (1). Limited autopsy

findings of individuals with CTE determined the presence of a thinned corpus callosum and atrophied frontal lobes, as well as additional impacts to the medial temporal lobe, mammillary bodies, hippocampus, and substantia nigra (6). These brain regions account for the symptoms associated with CTE including altered mood and behavior, impaired cognition and executive function, and suicidal ideation (14). Additional findings through gross examination of the brain showed the presence of enlarged lateral and third ventricles, which has become a common indicator of CTE (6). An issue that has arisen with CTE studies is the lack of nonrandomized studies regarding such autopsy cases where selection bias is present (6). In accordance with a study conducted by the Boston University Center for the Study of Traumatic Encephalopathy, of the 321 known American football player deaths that occurred between February 2008 and June 2010, only 12 of the brains had undergone postmortem neuropathological examination (6). The brains that were examined provided evidence of CTE and has suggested an estimated lifetime prevalence of 3.7% (6). This correlation is a public health concern that has been brought to the attention over the last decade. Microscopic observation demonstrates the abundance of neurofibrillary tangles along with glial tangles and neuropil thread with microtubule-associated protein tau being the main protein component of these tangles (11). This protein is known to play a role in the stabilization of microtubules (14). Tau protein pathology in CTE has been found in superficial cortical laminae (I and III) (6). It is non-uniform and is restricted to the foci of the frontal, insular, and temporal cortices (6). These neurofibrillary tangles had increased density at the level of the cortical sulci and were typically perivascular or positioned around the blood vessels (11). With this distinction, there is a possibility that bloodbrain barrier disruptions that occurred during traumatic brain injury (TBI) may play a significant role in neurofibrillary tangle formation (4). The blood-brain barrier (BBB) refers to the properties of the microvasculature of the CNS that serve to protect the brain from exposure to circulating pathogens (5). This is performed through the tight regulation of the BBB in order to maintain the CNS homeostasis for proper neuronal functioning (5). Studies have shown the partial or complete loss of the BBB functioning associated with several neurological disorders may be a major contributor to the pathology of such conditions (5). BBB breakdown permits the entry of pathogens and other damaging substances, resulting in the development of neurodegenerative diseases. For example, neuroimaging studies performed for conditions like Alzheimer’s disease (AD) have shown a BBB dysfunction (16).


Chronic Traumatic Encephalopathy

In relation to CTE, a study found BBB breakdown present in regions containing increased perivascular p-tau protein. The study utilized immunoreactivity patterns of claudin-5 and occluden-1, BBB associated tight junction components (5). It was determined that both tight junction components were either discontinued or absent in perivascular p-tau deposited regions (5). This resulted in the presumption that BBB dysfunction could serve as a correlate or neural dysfunction in individuals suspected of having CTE (5). More research needs to be conducted in order to further comprehend the pathway and pathophysiology of CTE. A definitive diagnosis of CTE can only be administered through postmortem neuropathological analysis of tau protein patterns in the brain (17). There is no test that can determine if a living individual has this condition, as this is a fairly new area of research in the medical field. A definitive diagnosis of CTE can only be administered through postmortem neuropathological analysis (17). Currently there is no method of utilizing an CT, MRI, or any brain imaging technique to diagnose CTE aside from autopsy. Physicians will administer neurological exams as well as mental status testing to patients displaying symptoms of CTE. Research on developing diagnostic techniques that can be performed on live patients is currently underway. Treatment options are limited and often arrive in the form of behavioral therapy, pain management regimes, palliative care, and memory exercises. Behavioral therapy may include the prescription of medications, such as serotonin reuptake inhibitors (SSRI), to combat depression and anxiety. Pain management includes the aforementioned medications as some individuals experiencing CTE symptoms have reported a slight reduction in mood swings, improved memory and decreased pain (17).

Methods Databases used for research included Pubmed, sourced through the Geisinger Commonwealth School of Medicine library collection, and Google Scholar. Several keywords and phrases aided the research process: chronic traumatic encephalopathy, traumatic brain injury, proteopathy, blood brain barrier, tight junctions, microvascular damage, neurotoxicity, and endothelial cells. Articles were carefully chosen to avoid research older than a decade, with the oldest utilized paper having been published in 2009.

Discussion Neuroimaging via positron emission tomography Currently, there is no antemortem examination that can positively diagnose a patient suspected of CTE. Since many individuals with a neurodegenerative disease present with sudden mood changes and behavioral disorders later in their lifetime, discriminating CTE from other pathologies is challenging (18). Having a reliable diagnostic tool to detect CTE antemortem would enable medical intervention to occur at an earlier age, which is crucial for an individual with a history of repeated TBIs. Concussions may be classified as TBIs but current literature describes a TBI as an irreparable alteration to brain function caused by an external force that results in


detrimental long-lasting effects (3, 10). Changes in behavior, attitude, and mood after a history of recurrent TBIs could suggest CTE has developed (10). Traditionally, neuroimaging has not been used to confirm a conclusive CTE diagnosis (18). Often, imaging results appear similar between AD and CTEassociated AD, which is why any distinct clinical characteristics must be identified prior to imaging to avoid a misdiagnosis. PET scans are used to study different organs and tissues by injecting and tracking a radioactive drug, or tracer that is known to detect certain diseases or physiologic states. An example of one such tracer is [F-18] FDDNP, which was utilized by Barrio et al. as a tracer due to its sensitivity for tau (2). The investigative team performed PET scans on both living former football players and two control groups, which consisted of individuals who had no history of TBI and AD patients, respectively. In comparing images from different brain regions, Barrio et al. observed that the limbic, cortical, and subcortical areas contained the highest amounts of tracer; even higher concentrations were detected in the amygdala (2). These findings indicate that the aforementioned regions possessed markedly increased concentrations of tau. The investigation also compared the patterns identified via PET scans using FDDNP to actual brain tissue sampled from CTE-positive patients. The regions of the brain identified via the PET scan as having excess amounts of tau protein were similar to regions that had immunohistochemically confirmed tau deposition in CTE-positive brain tissue samples (2). However, it is interesting to note that advanced stages of CTE can be confused for advanced-stage AD, even with the use of PET imaging. As the degradation of key brain structures advances, neuroimaging results between AD and CTE begin to mimic each other, often resulting in a misdiagnosis. Finally, when comparing neural tissue from an individual with probable CTE to samples derived from an AD case, there was a statistically significant difference observed, suggesting that the CTE-positive patient suffered from advanced stages of CTE instead of AD (2). Stern et al. investigated levels of tau in the brains of living former National Football League players with suspected CTE by using flortaucipir as a tracer in PET scans and 18F-florbetapir to detect any amyloid-beta. To test how much of the marker was being taken up by tissues in the brain, researchers measured cerebellar florbetapir standard uptake value ratio (SUVR) and found that in the group comprised of former NFL players, there was a significantly higher SUVR compared to the controls (asymptomatic individuals who had no history of TBI) (18). Tracer concentrations were highest in the bilateral superior frontal area of the brain, the bilateral medial temporal portion of the brain, and left parietal area; each is known to be affected by CTE when diagnosed after death (18). Utilization of immunohistochemistry postmortem also highlighted significant amounts of tau deposition in the amygdala and midbrain in older and younger players which suggests that damage to these structures coincide with CTE rather than AD due to the mechanics of TBIs. To rule out if any of the behavioral changes that the NFL players were experiencing were related to a tauopathy other than CTE, measurements of the amount of amyloid-beta were taken via PET scan. Subsequent findings were not significant, indicating that the symptoms the players felt were not due to early onset AD, but rather were more in

Chronic Traumatic Encephalopathy

line with early presentations of CTE (18). This again highlights the importance of having proper diagnostic tools so that misdiagnosis does not occur and proper intervention can take place. The use of [F-18] FDDNP as a radioactive ligand for PET scans can be considered controversial due to FDDNP having the ability to not only bind to tau aggregates but also to amyloidbeta plaques, which are a clinical indicator used to diagnose AD (2, 13). However, that is where postmortem analysis of confirmed CTE cases is a useful tool in identifying patterns that differ markedly from AD. In a study conducted by Omalu et al., a 59-year-old football player was recruited in order to support the use of PET imaging in identifying CTE during his lifetime with the goals of confirming the condition and further supporting the utilization of such neuroimaging techniques after death (13). During his life, the player underwent numerous PET scans with [F-18] FDDNP as a tracer and upon death, postmortem tissue analysis was used to confirm a diagnosis of CTE (13). Antemortem scans utilizing [F-18] FDDNP highlighted statistically significant tau deposition with insignificant amounts of amyloid beta accumulation in the frontal and temporal lobes (13). Areas of the brain with predominant tau deposition were identified as being in the midbrain and amygdala and were confirmed postmortem via immunohistochemistry. The amygdala and midbrain are structures of the brain that tend to get damaged first during a TBI due to the rapid acceleration and sudden deceleration of the head (13, 18). These structures exhibited dense areas of tau deposition that accumulated early in life after sustaining a TBI and persisted well into older age (13, 18). This pattern of tau deposition within the amygdala and midbrain and lack of beta-amyloid highlights some of the key findings that indicate a diagnosis of CTE rather than AD (Figure 1). Once brain structures begin to develop tau neurofibrillary tangles, they can continue to spread and clinical presentation of late stage CTE can be confused for late stage

AD, therefore, being able to utilize imaging techniques like [F-18] FDDNP and identifying patterns that separate CTE from AD is important. Even with a history of TBIs and reported new onset of abnormal behavioral and cognitive abilities, there is no diagnostic test currently screening patients suspected of having CTE. However, repeated claims have been made and supported that utilizing PET scans along with a radioactively labeled tracer for a known protein of interest involved in CTE tauopathy, such as tau, can be applied antemortem to identify the onset of CTE. Thus, early intervention before advanced CTE and neurodegeneration occur may help a patient and their family adjust to a new diagnosis and remain proactive for the remainder of the prognosis. Blood-brain barrier Components and general structure The BBB is responsible for maintaining separate environments between systemic circulation and the microenvironment of the central nervous system. By ensuring limited contact of the two compartments, the BBB prevents neurotoxic agents, such as antibodies, macromolecules, and certain ions, from crossing into the delicate cerebral environment. No neuron is more than 20 Âľm from a capillary, so BBB integrity is essential to proper neurological performance. Preserving homeostasis in the central nervous system requires a controlled microenvironment (5).

The core components of the BBB are pericytes, astrocytes, and the endothelial cells lining the cerebellar capillaries; each of these elements aids in forming a tight seal around the blood vessel(s). Endothelial cells of the central nervous system differ in structure and function from those found in the peripheral nervous system. With low rates of transcytosis and low expression of leukocyte adhesion molecules (LAMs), endothelial cells of the brain exhibit reduced trafficking and inflammation. Most importantly, these endothelial cells possess continuous intercellular tight junctions that prohibit paracellular transport of various molecules from the blood to the perivascular space (12). The perivascular space refers to the local region surrounding the microvasculature in the brain (5). Cerebral endothelial tight junctions depend on two transmembrane proteins, occludin and claudin, for their contributions to BBB stability and barrier function, respectively. Claudin-5 expression of brain endothelium predominates, with claudin-3 comparatively expressed. Studies involving mice genetically altered to lack claudin-5 demonstrated BBB breakdown, as did similar investigations utilizing human endothelial cell cultures. Occludin breakdown/absence has been correlated with an increased permeability of the BBB, but not breakdown of the structure, in analyses involving human endothelial cell Figure 1. Usage of [F-18]FDDNP in PET to evaluate tau deposition in the midbrain cultures (9). As discussed in later sections, and amygdala (2).


Chronic Traumatic Encephalopathy

immunohistochemical examinations indicate that claudin-5 and occludin are implicated with BBB breakdown involving p-tau (5). Blood-brain barrier disruption and p-tau Disruption to the BBB has long been theorized to be a leading factor in the development of many neurodegenerative diseases, including CTE. Despite many years of clinical research, the role of BBB disruption in neurodegenerative pathogenesis remains inconclusive. Specifically, researchers and clinicians seek to determine if BBB degeneration occurs at the inception of neuronal decay, or if it is a subsequent event (12). Although the timeline of BBB breakdown has yet to be understood, investigations concerning the changes to BBB structure and function have determined that claudin-5 and occludin-1, subunits forming the zona occludens, or tight junctions, are heavily implicated in the compromised BBB (5). As mentioned earlier, continuous tight junctions in the endothelial cells of brain capillaries prevent paracellular movement of neurotoxic agents (9, 12). Several factors influence the tight junctions in endothelial cells, including local concentrations of p-tau. One case report describes an investigation that sought to ascertain whether perivascular p-tau deposition promotes the degradation of the BBB. A 56-year-old man presented to the emergency room with sudden confusion and disorientation. Further investigation revealed two facts; the patient had a 5-year history of declining cognitive faculties and he was a former rugby player. Physicians diagnosed the patient with progressive supranuclear palsy (PSP), and he passed away one year after his hospitalization. PSP is a neurodegenerative disease involving a slow decrease in brain volume, resulting in a gradual loss of cognitive and physical abilities. Given the clinical similarities between PSP and CTE, it is not surprising that the patient was misdiagnosed. The study aimed to examine claudin-5 expression in a brain clinically confirmed to be positive for CTE; antibodies targeting p-tau affirmed the presence of the insoluble protein. Investigators wanted to compare the deceased patient’s brain to those of individuals who had been diagnosed with PSP, but tested negative for p-tau and CTE. PSP-positive and control brains were donated from the Dublin Brain Bank. In a double immunofluorescence model, sectioned slices of the patient in question, the PSP brains, and the controls were exposed to both rabbit anti-claudin-5 antibody and mouse anti-phosphotau. Immunoreactivity patterns of the control brains clearly revealed high levels of claudin-5 and occludin-1, but minimally detectable p-tau, indicating that the BBB was intact and there was no significant level of p-tau. Immunoreactivity of the CTEpositive brain sections exhibited a stark contrast; areas of high perivascular p-tau deposition presented absent claudin-5 and occludin-1 immunopositivity, indicating a disrupted BBB. In order to determine the additive capacity of PSP to a CTE case, researchers compared immunoreactivity patterns between the CTE/PSP brain and a PSP only brain. Claudin-5 levels of the PSP did not appear to correlate with presence of p-tau, suggesting that BBB degradation might be unique to CTE and PSP. Regions with high p-tau deposition observed no claudin-5, or other tight junction elements. Immunoreactivity patterns


showed that p-tau deposition was highest in the perivascular space, depths of sulci, and in the subpial space. Retrospective inquiry revealed that the patient had played rugby professionally for approximately 40 years, and had sustained innumerable head concussions during his career (5). Acute versus chronic vascular changes observed with TBI CTE pathology remains nebulous not only at the molecular level, but also in a temporal sense; how long after a TBI does CTE pathogenesis take before noticeable structural changes occur? Furthermore, how long do these changes last after the initial head trauma? Many studies compare BBB and microvascular damage through observations made during the acute and chronic time periods post-impact injury. It would appear that BBB disruptions may linger in an elevated proportion of late survivors when compared to a lower proportion of survivors with BBB damage who are in the acute stage of recovery (8). These findings are interesting, but complicate the timeline of CTE development and recognizability. For approximately a century, neuroscientists and physicians alike have utilized the controlled cortical impact (CCI) model to elucidate the influence of a TBI on the microvasculature of the brain. Animal models permit researchers to mimic and observe the pathology that ensues after a simulated brain injury and gain invaluable insight in the process. Typically, CCI procedures combine a craniotomy with an impact injury that imitates the type of TBI that one might sustain during high contact sports and vehicle collisions, for example. A craniotomy is a small opening made between the bregma and lambda sutures of the parietal bones in rats (or other small rodents), exposing the brain. Next, a stereotaxic impactor apparatus inflicts an impact injury on the exposed portion of the brain. One such study adopted this model to assess the microvascular irregularities that followed during the acute and chronic time periods post-impact injury (7). Male rats sustained either a bilateral CCI (bCCI), or a unilateral CCI (uCCI), of two different magnitudes, following a 5 mm craniotomy. A Benchmark Stereotaxic Impactor apparatus with a 4-mm impactor tip inflicted an impact injury on the exposed brain with a velocity of 3.5 m/second, dwell time of 200 ms, and a compression distance measuring 200 mm. Acute stages of post-impact injury were defined as 24 hours and 1 week. Control rats received a craniotomy, but no impact injury. Chronic stages of post-impact injury were defined at 1 month, 2 months, and 3 months. After reaching the designated time period of examination, the respective rats were euthanized and their corpora callosa analyzed for alteration. Investigators observed a number of different changes associated with the impact injury, including endothelial damage, BBB degradation, macrophage mediated-inflammation, astrogliosis, and evolving white matter degradation. Each of these conditions has a damage marker that indicates their degree of disturbance or immunoreactivity. It was apparent that significant white matter degeneration ensued as a consequence of TBI; uCCI histological images revealed a gradual increase in the density and diameter of white matter lesions found in the sectioned corpora callosa of rats measured from 24 hours to 3 months post-impact injury. These findings confirm the demyelination of neurons local to the region of impact. While white matter

Chronic Traumatic Encephalopathy

damage predominated the findings, there were other types of brain damage present that indicated BBB disruption (7). Microbleeding, or microvascular hemorrhaging of the corpus callosum, was not detected until 1 week post-impact injury, but the number of individual microbleeds continued to increase up to 3 months post-impact injury. Ranging from 2 to 20 µm in diameter, microbleeds were observed in both bCCI and uCCI rats, while control rats did not present with a significant increase in microbleeds. It is interesting to note that the uCCI rats with the ipsilateral impact injury demonstrated both ipsilateral and contralateral microbleeds, indicating that microbleeding can occur contralaterally to the site of impact. It can be assumed that microbleeding occurs due to damaged BBB (7). As discussed previously, endothelial cells are perhaps the most essential element to the BBB because they contain tight junctions that prevent paracellular transport (5). Researchers assessed endothelial damage by examining the immunoreactivity patterns of Intracellular adhesion molecule-1 (ICAM-1), a transmembrane protein present in endothelial cells. Results demonstrated significantly elevated levels of ICAM-1 immunoreactivity at 1 and 3 months, which is biphasic. This marked increase in free ICAM-1 concentrations correlates, temporally, with the gradual increase in microbleeds, which also peaked at 3 months. In order to evaluate BBB damage as a whole, researchers exposed primary antibodies in the corpora callosa to rat Immunoglobulin G (IgG). Immunohistochemistry revealed conspicuous, scattered IgG immunoreactivity surrounding the sites of impact injury. Maximal IgG staining occurred at 24 hours, but numerous punctate stainings were detected at 1 month. Furthermore, these punctate stainings were temporally colocalized with microbleeds, suggesting that BBB breakdown and microbleeds occurred in the same spot. Finally, gliosis of the astrocytes remained elevated at all measured time points, meaning that the astrocytes were continuing to proliferate and undergo hypertrophy. Glial fibrillary acidic protein staining of the corpus callosum measured immunoreactivity of the astrocytes. These results might suggest that as the damage to the microvasculature persisted, astrocytes continued to support the neuronal cells by attempting to repair damage, form scar tissue, and attenuate any CNS damage sustained (7).

Conclusion As the molecular mechanisms behind CTE remain cryptic, new evidence supports previous research stating that BBB disruption is central to the pathology of this disease. Simulated TBI results in significant delayed microvascular damage surrounding the regions of impact, implicating the BBB as a foci of damage. From a clinical standpoint, the lack of an antemortem diagnostic examination complicates matters of routine for both healthcare professionals and concerned patients. There is no subsequent cure, and even the etiologic factors of CTE remain shrouded in mystery. When thinking of CTE and individuals who may be exposed to recurrent TBIs, football players often come to mind. However, they are not the only individuals prone to developing CTE. Some populations at greater risk to sustain a head injury can include individuals who engage in various high-contact sports like boxing and rugby, victims of domestic and child abuse, and

military personnel. Therefore, having a way to detect the onset of CTE is vital for individuals under these circumstances so interventions can occur earlier in pathogenesis, possibly mitigating the impact that behavioral changes would have on daily life. Neuroimaging modalities refined for CTE detection have proven themselves to be a promising avenue for living patients possibly faced with a CTE diagnosis. PET scans utilized with probe sensitivity for tau may allow for further investigation of suspected CTE in living individuals who have prior history of TBI.

Acknowledgments First and foremost, we wish to thank Iris Johnston for helping our group conduct research by obtaining access to articles that we otherwise would not have been able to utilize. We would also like to thank Dr. Brian J. Piper for helping us develop the topic of our paper and encouraging us to submit our research for publication while also providing us with invaluable guidance during the writing process. Lastly, we would like to thank Geisinger Commonwealth School of Medicine for giving us the tools necessary to conduct such thorough research.

References 1.

Asken BM, Sullan MJ, Snyder AR, Houck ZM, Bryant VE, Hizel LP, McLaren ME, Dede DE, Jaffee MS, DeKosky ST, Bauer RM. Factors influencing clinical correlates of chronic traumatic encephalopathy (CTE): a review. Neuropsychol Rev. 2016;26(4):340-63.

2. Barrio JR, Small GW, Wong K-P, Huang S-C, Liu J, Merrill DA, et al. In vivo characterization of chronic traumatic encephalopathy using [F-18]FDDNP PET brain imaging. Proc Natl Acad Sci U S A. 2015;112(16):E2039-E2047 3. Concussion | Brain Injury Association of America [Internet]. Brain Injury Association of America. 2020 [cited 15 April 2020]. Available from: about-brain-injury/concussion 4. Daneman R, Prat A. The blood-brain barrier. Cold Spring Harb Perspect Biol. 2015;7(1): a020412. 5. Doherty CP, O'Keefe E, Wallace E, et al. Blood-Brain Barrier Dysfunction as a Hallmark Pathology in Chronic Traumatic Encephalopathy. J Neuropathol Exp Neurol. 2016;75(7):656–662. 6. Gavett BE, Stern RA, McKee AC. Chronic traumatic encephalopathy: a potential late effect of sport-related concussive and subconcussive head trauma. Clin Sports Med. 2011;30(1):179–xi. 7.

Glushakova OY, Johnson D, Hayes RL. Delayed increases in microvascular pathology after experimental Traumatic Brain Injury are associated with Prolonged Inflammation, Blood–Brain Barrier Disruption, and Progressive White Matter Damage. J Neurotrauma. 2014;31(13):1180–93.

8. Hay JR, Johnson VE, Young AM, Smith DH, Stewart W. Blood-Brain Barrier Disruption is an early event that may persist for many years after traumatic brain injury in humans. J Neuropathol Exp Neurol. 2015;74(12):1147–57.


Chronic Traumatic Encephalopathy

9. Luissint A-C, Artus C, Glacial F, Ganeshamoorthy K, Couraud P-O. Tight junctions at the blood brain barrier: physiological architecture and disease-associated dysregulation. Fluids Barriers CNS. 2012;9(1):23. 10. McAllister TW. Neurobiological consequences of traumatic brain injury. Dialogues Clin Neurosci. 2011;13(3):287. 11. McKee AC, Cantu RC, Nowinski CJ, Hedley-Whyte ET, Gavett BE, Budson AE, Santini VE, Lee HS, Kubilus CA, Stern RA. Chronic traumatic encephalopathy in athletes: progressive tauopathy after repetitive head injury. J Neuropathol Exp Neurol. 2009;68(7):709-35. 12. Obermeier B, Daneman R, Ransohoff RM. Development, maintenance and disruption of the blood-brain barrier. Nat Med. 2013;19(12):1584–96. 13. Omalu B, Small GW, Bailes J, Ercoli LM, Merrill DA, Wong KP, Huang SC, Satyamurthy N, Hammers JL, Lee J, Fitzsimmons RP. Postmortem autopsy-confirmation of antemortem [F-18] FDDNP-PET scans in football player with chronic traumatic encephalopathy. Neurosurgery. 2018;82(2):237-46 14. Omalu BI, Hamilton RL, Kamboh MI, DeKosky ST, Bailes J. Chronic traumatic encephalopathy (CTE) in a National Football League player: Case report and emerging medicolegal practice questions. J Forensic Nurs. 2010;6(1):40-6. 15. Provenzano FA, Jordan B, Tikofsky RS, Saxena C, Heertum RLV, Ichise M. F-18 FDG PET imaging of chronic traumatic brain injury in boxers. Nucl Med Commun. 2010;31(11):952–7. 16. Sweeney MD, Sagare AP, Zlokovic BV. Blood–brain barrier breakdown in Alzheimer disease and other neurodegenerative disorders. Nat Rev Neurol. 2018;14(3):133. 17. Shura RD, Taber KH, Brenner LA, Wortzel HS. Chronic traumatic encephalopathy and traumatic brain injury: bridging pathology, function, and prognosis. Curr Phys Med Rehabili Rep. 2015;3(2):106-14. 18. Stern RA, Adler CH, Chen K, Navitsky M, Luo J, Dodick DW, et al. Tau Positron-Emission Tomography in Former National Football League Players. N Engl J Med. 2019;380(18):1716–25.


Scholarly Research In Progress • Vol. 4, October 2020

A Bath a Day Keeps the Germs Away: Chlorhexidine Gluconate Bathing Expansion Quality Improvement Project Mahrukh Choudhary1†, Kenneth Lam1†, Connor M. Magura1†, Jasmine Santos1†, Julia Shamis1†, Navindra Tajeshwar1†, Hina Farrukh2, and Joseph Santora2

Geisinger Commonwealth School of Medicine, 525 Pine St, Scranton, PA 18509 Atlanticare Regional Medical Center, 1925 Pacific Ave, Atlantic City, NJ 08401 †Doctor of Medicine Program Correspondence: 1


Abstract Health-care-associated infections (HAI) among patients admitted to the hospital are on the rise, particularly for patients in the intensive care unit (ICU) (1). These infections are difficult to treat and often require potent antibiotics. Infection control protocols put out by The Centers for Disease Control and Prevention (CDC) include routine chlorhexidine gluconate (CHG) baths for high-risk patients. Daily bathing with CHG has been proven, in several prior studies, to reduce the risk of infection in this patient population. To lower the high possibility of these patients developing resistant infections, a CHG protocol was added to the appropriate order sets at two AtlantiCare Regional Medical Centers, and the corresponding staff were trained accordingly and compliance was measured by ICU vs PCU units and Mainland vs Atlantic City locations. The purpose of this quality improvement project was to look at these implementations and assess compliance and completion rates after the intervention. Results showed significant differences between average percentage of eligible days CHG ordered, average percentage of completed CHG protocol, and average percentage of completed CHG protocol per days ordered between Mainland PCU and Mainland ICU (p < 0.0001), City PCU and City ICU (p < 0.0001), and Mainland PCU and City PCU (p < 0.01). There were no significant differences between Mainland ICU and City ICU. In the future, utilizing this data we hope to target and improve the lowcompliance areas of the hospital via an integrated education plan.

Introduction Multidrug-resistant organisms (MDROs), like methicillinresistant Staphylococcus aureus (MRSA) and vancomycinresistant enterococcus (VRE), have become endemic to hospitals, particularly in higher-acuity care units. Such infections are associated with significant morbidity and mortality, particularly in intensive care units (ICUs), and linked with the use of medical instrumentation like central lines and venous catheters (1). Furthermore, these infections have been shown to significantly extend the length of stay for patients, and subsequently result in high attributable costs per infection ranging from $25,000 to $56,000 (2). Many of these costs, in some cases, are no longer reimbursed by third-party payers like the Centers for Medicare and Medicaid Services (3). Reduction of health-care-associated infections is thereby a top priority for hospitals. The Center for Disease Control and Prevention (CDC) has created a variety of strategies, such as hand hygiene and the use of isolation precautions in order to limit the spread of MDROs among patients, but such strategies

require consistent adherence from all members of the healthcare team during frequent patient encounters, and can thus be difficult to sustain (4). Targeted interventions, like the use of antiseptic agents for patient bathing, can significantly decrease the risk of hospital-acquired bloodstream infections. Chlorhexidine gluconate (CHG) is an antiseptic agent with broad-spectrum activity against many organs, particularly MDROs, and has residual antibacterial activity that can decrease both the microbial colonization on patients’ skin and prevent secondary contamination from the health care environment (5). There have been many studies that found significant rate reductions in the acquisition of both VRE and MRSA in ICU and central line patients following the implementation of daily CHG bathing protocols, with reductions ranging from 36–74% (5-9). While compliance was variable throughout these studies, the effects of implementation were still considerable, significantly impacting morbidity and mortality of patients, hospital costs, and length of stay (5–9). ICUs are generally reserved for the most acutely ill patients, those that are unstable, in critical condition, and who require very intensive care and monitoring, and thus more concentrated staffing for each individual patient. Progressive care units (PCUs) are considered to be intermediate or transitional care units, essentially implying that these patients are an intermediary step between the ICU and the medical or surgical floor. They can be used for either providing a higher level of care for patients deteriorating on a ward, or rather a lower level of care for those transitioning out of intensive care (10). AtlantiCare Regional Medical Center has two divisions; the City Division is located in Atlantic City, NJ, and the Mainland Division is located in Pomona, NJ. The City Division is a level II Trauma Center that serves a large, diverse population in an urban environment. It has a total of 26 ICU beds and 181 acute-care beds. The Mainland Division is a Primary Stroke and Advanced Cardiac Surgery Center that serves a suburban area surrounded by many skilled nursing facilities. It has a total of 22 ICU beds, and 255 acute-care beds. The purpose of this quality improvement project was to expand daily CHG bathing of hospitalized patients at both AtlantiCare Regional Medical Center City Division and Mainland Division to include patients with invasive devices (central lines, chest tubes, nephrostomy tubes), those who are going to surgery, and all patients in the ICU and PCU settings for the betterment of infection control. To implement this change, the order sets and protocols for these groups of patients were changed, all members of the health care team 63

A Bath a Day Keeps the Germs Away

were educated in these changes and the importance of daily CHG bathing, and adherence was audited and assessed throughout the various hospital units to ensure an increase following the intervention.

different department data stratified by each patient visit. These results were then compared and analyzed using a t-test found on where appropriate.


Mainland eligibility for the intervention was noted at 1,868 indicated days. Data showed that 704 (37.69%) of these occasions were ordered appropriately and 414 (22.16%) of these indications were completed. Of the number of orders placed 58.81% were fulfilled. From Mainland PCU there were 1,248 (66.81%) days where CHG orders were eligible, 350 (28.04%) where eligible CHG orders were placed, and 193 days (15.46%) where eligible CHG orders were completed. Of the number of orders placed, 55.14% were fulfilled. From Mainland ICU, there were 620 days (33.19%) eligible for CHG orders, 354 days (57.10%) where eligible CHG orders were placed, and 221 days (35.64%) where eligible CHG orders were completed. Of the number of orders placed, 62.43% were fulfilled.

The CHG project sought to assess the level of compliance with CHG washing of patients in the selected clinical settings and determine their baseline before performing a programmatic initiative to increase compliance. Our pre-intervention data included patients eligible from January 1, 2019, to January 31, 2019. Our various patient cohorts included PCU and ICU patients who required CHG bathing at our city and mainland hospitals. The patients’ data was extracted by reviewing their charts and included: their hospital, their unit, their floor, their days of eligibility for CHG bathing, their early discontinuation, the ordering provider, and a reason for termination of bath if applicable. Data analysis Compliance data was analyzed retrospectively using chart review. De-identified data was then stored using Google Drive. All other data such as compliance statistics were derived from this initial chart review and compiled using Microsoft Excel. We then stratified our data by patient location in PCU or ICU, and their hospital. Our main outcomes were to look at compliance in three different ways. First, compliance of completing washes when they are ordered. Second, having orders entered for washing on eligible days. Third, for completing washes for all eligible days. We were able to compare the different departments for their integrated performance for each eligible day of each encounter and we were also able to look at


Examination of Mainland data showed that the average percentage of eligible days where CHG was ordered was 38.21% and the average percentage of eligible days where CHG orders were completed was 22.61%. The average percentage of fulfilled orders per day was 33.36%. From Mainland PCU, the average percentage of eligible days where CHG was ordered was 22.16% and the average percentage of eligible days where CHG orders were completed was 12.68%. The average percentage of fulfilled orders per day was 21.54%. From Mainland ICU, the average percentage of eligible days where CHG was ordered was 62.54% and the average percentage of eligible days where CHG orders were completed was 41.03%. The average percentage of fulfilled orders per day was 54.88%.

Figure 1. Average patient encounter compliance of chlorhexidine gluconate (CHG) ordering and washing, stratified by department. 64

A Bath a Day Keeps the Germs Away

Analysis of Mainland data showed a significant difference in the average percentage of eligible days CHG ordered per encounter between Mainland PCU and Mainland ICU (24.16% vs 62.54% [p <0.0001]), average percentage of completed CHG protocol per encounter (12.68% vs 41.03% [p < 0.0001]), and average percentage of completed CHG protocol per days ordered (21.54% vs 54.88% [p < 0.0001]), as shown in Figure 1. At City, there were a total of 2,144 days where CHG orders were eligible, 808 days (37.69%) where eligible CHG orders were placed, and 519 days (24.21%) where eligible CHG orders were completed. Of the number of orders placed, 64.23% were fulfilled. From City PCU, there were 1,052 days (49.06%) where CHG orders were eligible, 152 days (14.44%) where eligible CHG orders were placed, and 80 days (7.06%) where eligible CHG orders were completed. Of the number of orders placed, 52.63% were fulfilled. From City ICU, there were 1,092 days (50.93%) where CHG orders were eligible, 656 days (60.07%) where eligible CHG orders were placed, and 439 days (40.2%) where eligible CHG orders were completed. Of the number of orders placed, 66.92% were fulfilled. In examining City data at an individualized level, the average percentage of eligible days where CHG was ordered was 33.52% and the average percentage of eligible days where CHG orders were completed was 21.61%. The average percentage of fulfilled orders per day was 31.93%. From City PCU, the average percentage of eligible days where CHG was ordered was 9.21% and the average percentage of eligible days where CHG orders were completed was 4.79%. The average percentage of fulfilled orders per day was 10.91%. From City ICU, the average percentage of eligible days where CHG was ordered was 62.26% and the average percentage of eligible days where CHG orders were completed was 41.49%. The average percentage of fulfilled orders per day was 56.80%. Analysis of City data showed a significant difference in the average percentage of eligible days CHG ordered per encounter between City PCU and City ICU (9.21% vs 62.26% [p <0.0001]), average percentage of completed CHG protocol per encounter (4.79% vs 41.49% [p < 0.0001]), and average percentage of completed CHG protocol per days ordered (10.91% vs 56.80% [p < 0.0001]), as shown in Figure 1. There were significant differences between Mainland PCU and City PCU in the average percentage of eligible days CHG ordered per encounter (24.16% vs 9.21% [p <0.0001]), average percentage of completed CHG protocol per encounter (12.68% vs 4.79% [p < 0.0002]), and average percentage of completed CHG protocol per days ordered (21.54% vs 10.91% [p < 0.0055]). There were no significant differences between Mainland ICU and City ICU, as shown in Figure 1.

Discussion Research has shown that daily CHG bathing is linked to significant reduction in hospital-acquired infections among patients (5). Adherence to CHG bathing protocol is vital in order to ensure infection risk is minimized for our patients. As with any change in protocol or methodology, the shift towards CHG baths and the use of CHG poses several concerns to be further evaluated, including effectiveness of the new CHG

protocol, provider compliance with the new protocol, and financial implications for the hospital systems. The effectiveness of CHG baths has been proven to result in a statistically significant reduction in infections in multiple studies. A study at McGill University Health Center compared central-line-associated blood infections (CLABSI) with and without CHG usage in the neonatal intensive care unit, with findings that showed that the rate of infections decreased in the experimental (i.e., CHG usage) but remained the same in the control group (no CHG usage) (11). These findings were also reproduced in an adult population. A study comparing the efficacy of 0.5% CHG, 1% CHG and 10% aqueous povidineiodine (PVI) in prevention of CLABSI showed more than double the rate of infections in the PVI group. There were no significant differences in the rate of infections when the 0.5% and 1% CHG groups were compared (12). Another study comparing CHG with PVI also concluded that CHG was more effective in prevention of CLABSI, particularly for infections caused by gram-positive bacteria (13). There have been many studies conducted to assess the effectiveness of CHG usage in CLABSI prevention. Most of these have shown it to be superior to other antiseptics, such as PVI (11, 12, 13). These findings suggest that CHG usage should be used in patients, particularly those at high risk of developing an infection due to a central line insertion or a surgical procedure. Implementation of any protocol requires thorough evaluation of financial implications that may result from the intervention. A meta-analysis evaluating the costs associated with CHG implementation showed a less than $5 increase to make the switch from a non-medicated bathing cloth to a 2% CHG bathing cloth. While this may seem high when considered in a larger scale of a hospital system, the same study reports that the cost associated with a single CLABSI is more than 10 times the cost to switch to a medicated CHG bathing cloth (14). Another study performed within a neurosurgery department evaluated the implementation of a new protocol along with a provider intervention where each provider was informed of their individual infection rates among spinal surgery patients. They were also informed of the success of the new protocol and the financial implications associated with surgical infections. The authors found not only a greater than 40% reduction in the number of infections, but also found that this intervention resulted in an estimated annual cost savings of almost $300,000 (15). Although we were unable to consider and find the cost savings associated with CHG use within our own hospital systems, the findings from the studies above suggest that proper CHG and infection prevention protocol can result in improvement in overall costs within a department or a hospital. Although the effectiveness of CHG has been evaluated by many researchers and there are many studies that have shown CLABSI reduction with CHG usage, provider compliance continues to be an area of concern. The goal of this study was to evaluate compliance rates between two different hospitals as well as between two different units within the hospitals, the PCU and ICU. Overall, within the 4 units that were assessed in this study, the provider compliance average was below 33%. The completion average was below 25%. The findings did show a higher compliance and completion rate among


A Bath a Day Keeps the Germs Away

providers in the ICU when compared to the PCU. This finding could be attributed to two different reasons. Firstly, the ICU protocol includes a CHG bath when a patient is admitted to the unit, whereas the PCU requires that a provider manually order the CHG bath. Additionally, we believe that the higher acuity of patients in the ICU may have also played a role in the higher provider compliance rates. During our investigation, we uncovered a multitude of reasons why CHG baths were not completed. Limited understanding of CHG protocol and utility of EHR order sets may have played a role in provider ordering compliance during this timeframe. We also discovered various reasons why a CHG bath may have been ordered but not given to a patient at bedside, which included findings such as: medication not available, patient downgraded, not tolerated, not appropriate, provider canceled, and patient refused. Overall, gaining an understanding of CHG bathing protocol compliance in January 2019 helps us to better target provider and patient education, streamline EHR ordering, and improve access to the medication for eligible patients. This information will be vital to improve future CHG bathing compliance within AtlantiCare hospitals and minimize future hospital infection risk for these vulnerable ICU/PCU patients. Limitations of our data include selection and sampling error. Patient lists were generated by pulling data from the electronic health records. Our selection of patients was not randomized and sample size limited to eligible ICU/PCU patients with the January 2019 timeframe. In addition, although we aimed to standardize the data collection process among investigators with the creation of a data collection protocol, there exists an inherent amount of measurement error given the limitations of our data collection tools and collection process. Future direction for our project will aim to address the drastic differences in the compliance and completion rates between the PCU and ICU at both hospitals. Implementing a new PCU order set, which will include a CHG protocol, similar to those found in the ICU, then reassessing after the intervention can elucidate how technology affects compliance and completion rates. From there, further steps can focus on provider and patient education to assess their impact on compliance and completion rates in both hospitals’ PCUs and ICUs.

References 1.

Vincent JL, Rello J, Marshall J, et al. International study of the prevalence and outcomes of infection in intensive care units. JAMA. 2009;302(21):2323-9. doi: 10.1001/ jama.2009.1754.

2. Dixon JM. The effects of daily 2% CHG cloth bathing on catheter-related bloodstream infection rates in a surgical intensive care unit. Am J Infect Control. 2009;37(5):E57. 3. Stone PW, Glied SA, McNair PD, et al. CMS changes in reimbursement for HAIs: setting a research agenda. Med Care. 2010;48:433-9. doi: 10.1097/MLR.0b013e3181d5fb3f. 4. Siegel JD, Rhinehart E, Jackson M, Chiarello L. 2007 Guideline for Isolation Precautions: preventing transmission of infectious agents in health care settings. Am J Infect Control. 2007;(35):S65- S164. doi: 10.1016/j. ajic.2007.10.007.


5. Climo MW, Yokoe DS, Warren DK, et al. Effect of daily chlorhexidine bathing on hospital-acquired infection. N Engl J Med. 2013;368(6):533–542. doi:10.1056/ NEJMoa1113849 6. Vernon MO, Hayden MK, Trick WE, Hayes RA, Blom DW, Weinstein RA. Chlorhexidine gluconate to cleanse patients in a medical intensive care unit: the effectiveness of source control to reduce the bioburden of vancomycinresistant enterococci. Arch Intern Med. 2006;166:306-12. doi: 10.1001/archinte.166.3.306. 7.

Bleasdale SC, Trick WE, Gonzalez IM, Lyles RD, Hayden MK, Weinstein RA. Effectiveness of chlorhexidine bathing to reduce catheter-associated bloodstream infections in medical intensive care unit patients. Arch Intern Med. 2007;167:2073-9. doi: 10.1001/archinte.167.19.2073.

8. Milstone AM, Elward A, Song X, et al: Daily chlorhexidine bathing to reduce bacteraemia in critically ill children: a multicentre, cluster-randomised, crossover trial. Lancet. 2013; 381:1099-1106. doi: 10.1016/S0140-6736(12)61687-0. 9. Septimus E, Hickok J, Moody J, et al: Closing the translation gap: toolkit-based implementation of universal decolonization in adult intensive care units reduces central line-associated bloodstream infections in 95 community hospitals. Clin Infect Dis. 2016;63:172-177. doi: 10.1093/cid/ciw282. 10. Prin M, Wunsch H. The role of stepdown beds in hospital care. Am J Respir Crit Care Med. 2014;190(11):1210-1216. doi:10.1164/rccm.201406-1117PP 11. Quach C, Milstone AM, Perpête C, Bonenfant M, Moore DL, Perreault T. Chlorhexidine bathing in a tertiary care neonatal intensive care unit: impact on central lineassociated bloodstream infections. Infect Control Hosp Epidemiol. 2014;35(2):158-163. doi:10.1086/674862 12. Yasuda H, Sanui M, Abe T, et al. Comparison of the efficacy of three topical antiseptic solutions for the prevention of catheter colonization: a multicenter randomized controlled study. Crit Care. 2017;21(1):320. Published 2017 Dec 21. doi:10.1186/s13054-017-1890-z 13. Mimoz O, Pieroni L, Lawrence C, et al. Prospective, randomized trial of two antiseptic solutions for prevention of central venous or arterial catheter colonization and infection in intensive care unit patients. Crit Care Med. 1996;24(11):1818-1823. doi:10.1097/00003246-19961100000010 14. Shah HN, Schwartz JL, Luna G, Cullen DL. Bathing with 2% chlorhexidine gluconate: evidence and costs associated with central line-associated bloodstream infections. Crit Care Nurs Q. 2016;39(1):42-50. doi:10.1097/ CNQ.0000000000000096 15. Agarwal N, Agarwal P, Querry A, et al. Implementation of an infection prevention bundle and increased physician awareness improves surgical outcomes and reduces costs associated with spine surgery. J Neurosurg Spine. 2018;29(1):108-114. doi:10.3171/2017.11.SPINE17436

Scholarly Research In Progress • Vol. 4, October 2020

Methadone Distribution Trends from 2017–2018 Across the United States John A. Furst1†, Nicholas J. Mynarski1†, Jessica M. DeAngelis1†, Viraj Kothari1*, Jonique Depina1*, Kenneth L. McCall2, and Brian J. Piper1,3

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 University of New England, Portland, ME 04103 3 Center for Pharmacy Innovation and Outcomes, Geisinger Precision Health Center, Forty Fort, PA 18704 †Doctor of Medicine Program *Master of Biomedical Sciences Program Correspondence: 1


Abstract Background: The primary objective of this study is to explore methadone distribution patterns between the years of 2017 and 2018 across the United States (U.S.). This study builds upon previous literature that has analyzed prior years of U.S. distribution patterns, and further outlines state specific methadone trends. Methods: The Drug Enforcement Administration’s Automation of Reports and Consolidated Orders System was used to acquire methadone distribution amounts (grams) and number of narcotic treatment programs (NTPs). State population estimates were obtained from the 2010 Census published by the U.S. Census Bureau, Population Division. Methadone distribution weights and number of NTPs were compared from 2017 to 2018. Results: From 2017 to 2018, there was a national decrease in methadone production quota (6.00%), total distribution (6.34%), and methadone distribution for both opioid use disorder (OUD) and pain (2.77% and 20.14%, respectively). Distribution for pain management per 100K population decreased in every state, while considerable variation was seen across states in regard to distribution for OUD per 100K population. Analysis of percent change in the number of NTPs also showed broad variation across states. Conclusion: Total U.S. methadone distribution, distribution for OUD, and distribution for pain management all decreased between the years 2017 and 2018. State-specific analysis revealed an overall decrease in methadone distribution for the use of pain management across all 50 U.S. states and change in distribution for OUD showed a majority of states are decreasing in this regard. These trends lead us to question whether methadone’s therapeutic role as an analgesic and as a treatment for OUD has decreased. Further analysis of federal and state legislation, as well as comparisons with other OUD pharmacotherapy distribution trends, may help to explain the observed results.

Introduction Opioids interact with stereospecific and saturable binding sites within the nervous system and other tissues where their actions alter the release of various neurotransmitters including acetylcholine, norepinephrine, substance P, and dopamine (1–3). By mimicking endogenous enkephalins and endorphins, opioids modulate neurotransmission to produce global analgesic effects dependent upon opioid binding

specificity and receptor saturability (1). Methadone, a synthetic opioid, is considered a full agonist of the μ-opioid receptor and is known to have a long half-life relative to other opioids (4). After being identified as a potent pharmacotherapy option for those suffering from opioid use disorder (OUD), methadone has been demonstrated to decrease illicit opioid use, decrease HIV-associated risk behavior and reduce drugrelated criminal behaviors (5, 6). In addition to its use as a treatment for opioid substance use disorders, methadone has been used clinically in the management of significant, chronic pain; no significant difference in pain relief between morphine and methadone in cancer patients has been noted (4, 7). Consequently, methadone is currently classified as a schedule II-controlled substance in the United States (U.S.) and is Food and Drug Administration (FDA)-approved for the following uses: (1) detoxification treatment of opioid addiction; (2) pain management in patients nonresponsive to non-narcotic analgesics (4, 6, 7). Despite methadone’s clinical uses, current research has shown there is a positive association between distribution rate and methadone overdose death rate (8). Both methadone distribution and methadone-associated deaths in the U.S. increased between the years 2002 and 2006, an average of 25% and 22% per year, respectively (9). Following the years of increasing distribution and associated deaths, a steady decline in methadone distribution occurred between 2007 and 2013 and a significant decrease in methadone volume distribution occurred between 2015 and 2016 (1, 10). Despite published methadone distribution trends being clearly outlined for this earlier time frame, data on more recent trends and state-specific trends is currently lacking. With significant uses in opioid recovery and pain management, as well as a link between methadone and overdose deaths, further research into the distribution of methadone in the U.S. could serve to benefit both the individual and public health sector. This study utilized the United States’ Drug Enforcement Administration’s Automated Reports and Consolidated Ordering Systems (ARCOS) comprehensive database to evaluate changes in the distribution of methadone in recent years. The years 2017 and 2018 are the most recent, complete datasets published. Investigation of methadone distribution trends using ARCOS will allow a greater understanding of the most up-to-date pattern of use in treatment of OUD and recent geographic fluctuations of methadone use within the United States.


Methadone Distribution Trends

Methods Data sources Distributed by the U.S. Drug Enforcement Administration (DEA), methadone yearly U.S. production quotas were obtained from final adjusted aggregate production quotas for Schedule I and II controlled substances for 2017 and 2018 via the Federal Registrar. State population estimates were obtained from the 2010 Census published by the U.S. Census Bureau, Population Division, to normalize data between states. Data related to methadone distribution amount (in grams) was extracted from the DEA’s Automated Reports and Consolidated Ordering System (ARCOS) yearly drug summary data reports for 2017 and 2018. This federal program, mandated by the 1970 Controlled Substances Act, is a comprehensive drug reporting system that tracks the distribution of controlled substances, in grams, to pharmacies, hospitals, clinical providers, and Narcotic Treatment Programs (NTPs), and has been used in previous pharmacological trend studies (8, 10, 11). Methadone distributed to NTPs was classified as treatment of OUD. Methadone distribution to non-NTP locations, e.g. hospitals and pharmacies, was considered to be for pain management. Procedures were approved by the Institutional Review Board from the University of New England and Geisinger. Data analysis The following analyses were completed: (1) U.S. total methadone distribution to NTPs for management of OUD and distribution of the drug to other clinical locations for pain management (e.g., hospitals, pharmacies, practitioners) was calculated and converted to kg for each year (2017 and 2018); (2) percent change of methadone distribution from 2017 to 2018 for OUD and pain management for each state per 100K population quotient; (3) percent change in number of methadone NTPs from 2017 to 2018 for each state per 100K population quotient; (4) number of NTPs in 2018 for each state per 100K population quotient. Data analysis and figures were completed with GraphPad Prism, version 8.4.2 and Microsoft Excel.

largest percent increase of 74.84%, while Minnesota showed the largest percent decrease of -17.11% (Figure 4). Additionally, analysis of the number of NTPs per 100K in 2018 alone, showed Rhode Island had the largest number of NTPs per 100K, which contrasts many other states that had less than 1 NTP per 100K (44 states). Additionally, all states with more than one NTP per 100K are states found within the Northeast region (RI, VT, DE, MD, CT, MA). Wyoming was notably the only state without an NTP in 2018 (Figure 5). From a national standpoint, the number of methadone buyers for NTPs increased from 1,483 in 2017 to 1,572 in 2018, a 6.00% change.

Conclusion From 2017 to 2018, the ARCOS database demonstrated an overall decrease in the total weight of methadone distributed in the United States. Every state demonstrated a decrease in percent change of methadone use for pain management, possibly suggesting a loss of demand for the drug’s use as an analgesic. This is consistent with a recent ARCOS-based publication that shows decreases in the distribution of other opioids, including oxycodone, hydrocodone, and morphine, from 2013 to 2016 (10). Additionally, when comparing values on a national scale, an overall decrease in methadone distribution for OUD was documented. Although a minority of states (18) demonstrated a percent increase in methadone use for OUD, an overwhelming 31 states, plus the District of Columbia, demonstrated a percent decrease. This is excluding Wyoming, which had no methadone NTPs as of 2018. Interestingly, this national decrease in methadone use for OUD was accompanied by an increase of NTPs buyers in the United States. Many states in the Northeast that had the largest number of NTPs per 100K population (RI, PA, VT, MD, CT, MA) have shown decreases in the amount of distribution of methadone. Despite many states seeing decreases in methadone distribution during this time frame, there was a 173.63% increase in methadone distribution per 100K in North Dakota, however, there was no increase in the number of NTPs in this state. This may suggest an increasing demand

Results Both DEA production quota and total methadone distribution amount decreased between 2017 and 2018, 6.00% and 6.34%, respectively. However, production quotas were markedly higher than U.S. total distribution amount for both years. Additionally, distribution of methadone for OUD decreased by 2.77%, while use for pain decreased by 20.14% (Figure 1). Figure 2 demonstrates variability across individual states regarding percent change in methadone distribution amount for OUD. These changes range from a 173.63% increase, seen in North Dakota, to a 15.18% decrease seen in Maine. Overall, the majority of states saw decreases in methadone distribution for OUD between the years 2017 and 2018 (Figure 2). In contrast to distribution for OUD, distribution of methadone for pain management decreased in all 50 states, as well as the District of Columbia. These decreases ranged from a 41.58% decrease in Mississippi to a 5.9% decrease in the state of New York (Figure 3). The majority of states saw an increase in the number of methadone NTPs per 100K population. Iowa showed the 68

Figure 1: United States methadone distribution amounts and production quotas in kilograms for 2017 and 2018.

Methadone Distribution Trends

Figure 2: Percent change in methadone NTP distribution amount per 100K population from 2017 to 2018. States were ranked according to percent change and subsequently placed into groups of 10 for representation on the heat map.

within the state of North Dakota and therefore, a possible need for additional NTPs. Further exploration of state specific legislation and policy may help to explain these findings.

Figure 3: Percent change in methadone pain distribution amount per 100K population from 2017 to 2018.

Moreover, total production quota for methadone decreased from 2017 through 2018. Despite this decrease, methadone production quotas remained noticeably higher than the total U.S. distribution for both 2017 and 2018. It may be possible that methadone produced in the U.S. is being exported to other countries; however, this would not explain the decrease in distribution of methadone within the U.S. These observed trends raise the question, is methadone becoming a less popular choice for treatment of opioid use disorder? When compared to an alternative OUD medication, buprenorphine (partial Îź-agonist), methadone (full Îź-agonist) demonstrates increased risk of death during treatment induction but is more successful at patient retention (12). Although difficult to weigh the risks and benefits between these two treatments, it must be noted that methadone use by weight from 2017 to 2018 dropped by 6.34%, while buprenorphine distribution by weight increased by 13.26% according to ARCOS data. Whether these differences are due to perceived benefits of buprenorphine, possible price differences or restricted access due to varying state legislature is unknown and of interest to explore in further research. Additionally, in 2016, the Comprehensive Addiction and Recovery Act was put into effect, which expanded the ability to prescribe buprenorphine to qualified nurse practitioners and physician assistants and also increased the maximum number of patients that these providers can treat with buprenorphine, or other schedule III, IV, or V narcotic for the treatment of OUD, from 100 to 275 patients (13). This federal legislation change may help to explain the opposing 69

Methadone Distribution Trends

Figure 4: Percent change in number of NTPs per 100K population from 2017 to 2018. States were ranked according to percent change and subsequently placed into groups of 10 for representation on the heat map.

distribution trends of methadone and buprenorphine and warrants further exploration. In an effort to further analyze methadone trends across the U.S., incorporation of the 2019 methadone data set from ARCOS will be merged with this data set upon its publication. A more in depth look into federal and specific state legislation, as well as any recent changes to these laws, may help to explain the different trends uncovered. Furthermore, comparison of ARCOS methadone distribution trends to Medicaid enrollee information in matching years is also desired to assess any correlations. In the future, analysis of opioid use per state could be compared to these results to explore a possible correlation with the distribution patterns of methadone within each state. Zip codes provided by the ARCOS database can also be utilized for a county analysis within each state. Particularly, a state-specific approach of analysis for Pennsylvania is of relevant interest as Geisinger Commonwealth School of Medicine’s place of residence.

Acknowledgments Iris Johnston provided technical support. A special thanks to Molly Kropp-Lopez of the Biomedical Research Club and Dr. Stephanie Nichols of the University of New England.

Disclosures BJP is part of an osteoarthritis research team supported by Pfizer. The other authors have no conflicts of interest to declare.


Figure 5: Number of methadone NTPs per 100K population in 2018.

Methadone Distribution Trends

References 1.

Snyder SH. The opiate receptor and morphinelike peptides in the brain. Am J Psychiatry. 1978 Jun;135(6):645-52.

2. Terenius L. Endogenous peptides and analgesia. Annu Rev Pharmacol Toxicol. 1978;18:189-204. 3. Beaumont A, Hughes J. Biology of opioid peptides. Annu Rev Pharmacol Toxicol. 1979;19:245-67. 4. Anderson IB, Kearney TE. Use of methadone. West J Med. 2000 Jan;172(1):43-6. 5. Salsitz E, Wiegand T. Pharmacotherapy of Opioid Addiction: "Putting a Real Face on a False Demon." J Med Toxicol. 2016 Mar;12(1):58-63. 6. Marsch LA. The efficacy of methadone maintenance interventions in reducing illicit opiate use, HIV risk behavior and criminality: a meta-analysis. Addiction. 1998 Apr;93(4):515-32. 7.

Mercadante S, Bruera E. Methadone as a First-Line Opioid in Cancer Pain Management: A Systematic Review. J Pain Symptom Manage. 2018 Mar;55(3):998-1003.

8. Jones CM, Baldwin GT, Manocchio T, White JO, Mack KA. Trends in Methadone Distribution for Pain Treatment, Methadone Diversion, and Overdose Deaths - United States, 2002-2014. MMWR Morb Mortal Wkly Rep. 2016 Jul 8;65(26):667-71. 9. Toce MS, Chai PR, Burns MM, Boyer EW. Pharmacologic Treatment of Opioid Use Disorder: a Review of Pharmacotherapy, Adjuncts, and Toxicity. J Med Toxicol. 2018 Dec;14(4):306-322. 10. Piper BJ, Shah DT, Simoyan OM, McCall KL, Nichols SD. Trends in Medical Use of Opioids in the U.S., 2006-2016. Am J Prev Med. 2018 May;54(5):652-660. 11. Atluri S, Sudarshan G, Manchikanti L. Assessment of the trends in medical use and misuse of opioid analgesics from 2004 to 2011. Pain Physician. 2014 MarApr;17(2):E119-28. 12. Hser YI, Evans E, Huang D, Weiss R, Saxon A, Carroll KM, Woody G, Liu D, Wakim P, Matthews AG, Hatch-Maillette M, Jelstrom E, Wiest K, McLaughlin P, Ling W. Long-term outcomes after randomization to buprenorphine/naloxone versus methadone in a multi-site trial. Addiction. 2016 Apr;111(4):695-705. 13. Patterson R. Implementations of the provision of the Comprehensive Addiction and Recovery Act of 2016 relating to the dispensing of narcotic drugs for opioid use disorder. Federal Registrar. 2018 Jan;83(15):3071-3075.


Scholarly Research In Progress • Vol. 4, October 2020

Novel Treatments for Multiple Sclerosis Christine Rittenhouse1*, Jhamal Wallace1*, and Rebecca Kane1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Multiple sclerosis (MS) is an autoimmune disease of the central nervous system characterized by demyelination of the axons in the brain and spinal cord. MS affects approximately 400, 000 people in the United States alone. Although this is an acquired disease there is no cure. Current treatment options focus on managing MS symptoms and slowing the progression of the disease. In this review article, we examine new treatment modalities, ozanimod, siponimod, and leukemia inhibitory factor nanoparticle (LIF-NP). Ozanimod, an S1P modulator, was shown to be more effective in reducing annualized relapse rates, T2 MRI brain lesions, and T1 gadolinium-enhanced MRI lesions compared to the traditional IFβ-1a treatment option when treating relapsing-remitting multiple sclerosis. Siponimod, also an S1P modulator, produced a reduction in confirmed disability progression when treating secondary progressive multiple sclerosis. The LIF-NP, a recent discovery by Cambridge scientists, has the potential to not only slow the progression of the MS but promote neuronal protection and remyelination in the central nervous system. This discovery could lead to a possible cure for MS. These treatments for MS aim to provide patients with the ability to delay disease progression and possibly completely resolve all symptoms.

Introduction Multiple sclerosis (MS) is an immune-mediated disease that causes damage to the central nervous system (CNS) as a result of demyelination thus leading to neurodegenerative deterioration (1, 2). MS affects approximately 2.3 million people worldwide and is the leading cause of disability in young adults in the U.S. (1, 3). The symptoms observed in patients with MS are due to the disruption of the myelinated axons tracts in the CNS (4). The most prevalent symptoms include fatigue, weakness, vision problems, spasticity, gait difficulties, and numbness and tingling all over the body (2, 4). As MS progresses patients may also experience memory loss, slower information processing, depression, and bladder and bowel dysfunction (2, 4). The wide range of symptoms seen in patients with MS reflects multifocal lesions in their CNS (4). Although there is extensive ongoing research to better understand the pathogenesis of MS, scientists have been unable to confirm its cause. What scientists do know is that environmental and genetic factors play a role (1, 4). Researchers have determined that the inflammatory aspect of MS is controlled primarily by immune cells including CD8+, CD4+, and Th17 T cells (1–3, 5, 6). When T cells attack myelin, they trigger an immune response that involves cytokine secretion. The cytokine secretion activates an inflammatory cascade that recruits other immune cells that cause further inflammation and degradation which leads to neuroinflammation and axonal damage (1–3, 5, 6). B cells also play a role in MS inflammation by producing antibodies against 72

myelin that phagocytose myelin and oligodendrocytes (1, 3, 5). This immune response that takes place in the CNS causes an increase in the permeability of the blood-brain barrier (BBB) allowing the movement of more immune cells toward the axons of the CNS resulting in the formation of sclerotic plaques or lesions, demyelination, and neurodegeneration (3, 6). There is no definitive diagnostic test for MS. Diagnosing MS relies on clinical presentation, neuroimaging, and cerebrospinal fluid (CSF) analysis. Clinicians look for manifestations of signs and symptoms of MS while ruling out other conditions that may resemble MS (4, 6). An MRI scan is a critical tool in diagnosing MS and determining the progression of the disease (4, 6). Newer MRI techniques are able to detect both white and gray matter damage in the CNS (4). MRI scans can also be used to detect MS lesions that appear throughout the CNS. These lesions are most easily recognized in the white matter as focal areas of demyelination and inflammation (7). There are four major classifications of MS include relapsing-remitting (RRMS), primary progressive (PPMS), progressive relapsing (PRMS), and secondary progressive (SPMS) (6). This review article examines three novel treatments for MS ozanimod, siponimod, and LIF-NP — and offers insight into which of these is most promising as an effective MS therapy.

Methods The database used to collect articles for this paper was PubMed. PubMed was used because it is publicly available and contains important biomedical research. The initial search criteria were the keywords “multiple sclerosis.” This initial result provided over 90,000 articles. Because we were looking for new treatments, we narrowed down the result by excluding any articles prior to 2018 and due to the scarcity of MS in children, any articles involving children. A number of the background articles were located in the reference list of relevant articles. While gathering background information, we noticed an increase in IL-6 for patients affected by MS and proceeded to look for any treatments that affected that biological pathway. In order to see developing treatments, we filtered the results using a randomized clinical trial as inclusion criteria as well as IL-6 modifiers. The data in the articles were analyzed based on the type of MS that was treated and if the new therapies were compared to current treatments or placebo groups.

Discussion Current treatment modalities for MS Since there is no cure for MS, current treatment options only slow the progression of the disease and help manage the associated symptoms. The different types of therapeutic

Novel Treatments for Multiple Sclerosis

treatments for MS include injectables, oral administrations, and infusions (IV) (5, 6). There are now 14 medications that have been approved by various regulatory agencies around the world to treat MS (6). Some of them include interferon beta, glatiramer acetate, teriflunomide, dimethyl fumarate, fingolimod, natalizumab, alemtuzumab, ocrelizumab, and mitoxantrone (8). The majority of these treatments are used to help stop the body’s immune attack on the myelin sheath, thus limiting the relapse rate (3, 5, 6). While the bulk of MS treatments do help to repress the immune attack on the myelin they also suppress the body’s immune system leaving patients vulnerable to illness and infection (3, 6). In the realm of treatment options for MS, there is an unmet need to provide neuroprotection while also repairing previously damaged nerve fibers. Ozanimod RRMS is the most prevalent type of MS. It is characterized by attacks of worsening neurological function called relapses that are followed by partial or complete remission (10, 11). Approximately 85% of individuals who are initially diagnosed with MS present with RRMS (10, 11). While complete physical recovery from a relapse can occur, about half of all relapses are associated with lingering neurological deficits resulting in a persistent increase in disability (11). Even though there is currently no cure for MS, treatment options do exist for RRMS. One novel drug to treat RRMS is known as ozanimod (Zeposia®). Ozanimod is a sphingosine 1-phosphate (S1P) receptor antagonist. S1P is a phospholipid involved in lymphocyte migration and other physiological processes (9). Ozanimod selectively binds to 2 of the 5 S1P receptors: S1PR1 and S1PR5 (1, 3). Functional antagonism of S1PR1 inhibits the release of immune cells such as B cells and T cells from lymph nodes, thereby reducing the number of lymphocytes in the peripheral blood (9, 11). The modulation mechanism of S1PR5 is not known, but it may involve the reduction of lymphocyte migration into the CNS, thus being neuroprotective (9, 11). Currently, ozanimod is under review by the U.S. Food and Drug Administration (FDA) for the treatment of RRMS. The clinical efficacy, safety, and tolerability of ozanimod for the treatment of patients with RRMS have been demonstrated in the Phase III RADIANCE and SUNBEAM clinical trials (11). The RADIANCE and SUNBEAM trails were both multicenter, randomized, double-blind, double-dummy, active-controlled trials evaluating the efficacy, safety, and tolerability of two doses of oral ozanimod (0.5 mg and 1 mg) versus interferonbeta-1a (Avonex®), a modestly efficacious weekly RRMS intramuscular treatment (12). Interferon-beta-1a (IFβ-1a) is a cytokine that modulates the immune system by reducing antigen presentation and T-cell proliferation (13). The RADIANCE clinical trial was conducted over a 24-month treatment period and included 1,320 people living with RRMS; the SUNBEAM trial was conducted over a 12-month treatment period that included 1,346 people living with RRMS (10). The goals of both trials were to assess annualize relapse rates (ARR) during the treatment period and evaluate the number of new or enlarged hyperintense T2-weighted brain MRI lesions, the number of gadolinium-enhanced brain MRI lesions and the percent change from baseline in brain volume (10, 12). In the SUNBEAM trial, the researchers found a significant

decrease in ARR for ozanimod 0.5 mg (ARR = 0.24; p = 0.0013) and 1 mg (ARR = 0.18; p < 0.0001) compared to IFβ-1a (Avonex) (ARR = 0.35). They also saw a significant reduction in new or enlarging T2 lesions for ozanimod 0.5 mg (25%; p = 0.0032) and 1 mg (48%; p < 0.0001) compared to IFβ-1a. There was also a decrease in gadolinium-enhanced MRI lesions for ozanimod 0.5 mg (34%; p = 0.0182) and 1 mg (63%; p <0.0001) compared to IFβ-1a (12). Results from the RADIANCE trial were very similar to those observed in the SUNBEAM trial. A significant decrease in ARR for ozanimod 0.5 mg (ARR = 0.22; p = 0.0167) and for 1 mg (ARR= 0.17; p < 0.0001) compared to IFβ-1a (ARR = 0.28) was observed. They also found a significant reduction in new or enlarging T2 lesions for ozanimod 0.5 mg (34%; p = 0.0001) and for 1 mg (42%; p < 0.0001) compared to IFβ-1a. In addition, there was also a significant decrease in gadolinium-enhanced MRI lesions for ozanimod 0.5 mg (47%; p = 0.0030) and for 1 mg (53%; p = 0.0006) compared to IFβ-1a (12). Both clinical trials demonstrated ozanimod to be more effective than IFβ-1a in treating patients with RRMS. These results can be explained by the processing rates of the two drugs. According to a post hoc analysis, ozanimod was metabolized quicker than IFβ-1a (14). Therefore, since ozanimod can be processed faster in the body, it is able to work sooner than IFβ-1a. In addition to being effective, ozanimod was well tolerated by most participants. However, there were a few participants who suffered from adverse effects (AE). About 4.9% of participants experienced an increase in alanine aminotransferase (ALT) that was above three times the upper limit of normal. Most of the participants with elevated ALT levels did not require discontinuation of ozanimod (9). A total of 12 participants reported experiencing ≥1 serious AE. Two serious AE that were reported in these studies were pancytopenia, low counts for all three types of blood cells, and acute myocardial infarction. No serious AE occurred in more than one participant, for that reason researchers do not consider them related to ozanimod. (9). In a separate study, researchers compared the safety and efficacy of ozanimod to fingolimod by using patient data from the ozanimod RADIANCE and SUNBEAM trials, and the fingolimod FREEDOMS trials (11). Fingolimod was the first oral disease-modifying therapy approved for the treatment of RRMS, and like ozanimod, it is an S1P receptor modulator. Unlike ozanimod, fingolimod binds to 4 of the 5 S1P receptors: S1PR1, S1PR3, S1PR4, S1PR5 (9, 11). The results of this study showed that ozanimod and fingolimod were comparable in terms of efficacy outcomes on the ARR and confirmed disability progression, but ozanimod was associated with a significantly lower risk of any AE (3). In the first-year outcomes, ozanimod had higher absolute mean lymphocyte count (difference in means: 0.4 × 109/l) and lower risk of ALT elevations compared to fingolimod (11). In the second-year outcomes, ozanimod had significantly lower risks of AE leading to discontinuation, herpetic infection, basal cell carcinoma, bradycardia, and ALT elevations compared to fingolimod (all p < 0.05) (11). One possible reason why ozanimod was associated with a significantly lower risk of any AE is due to the fact that ozanimod is more selective than its


Novel Treatments for Multiple Sclerosis

predecessor, fingolimod. As previously stated, ozanimod binds to S1PR1 and S1PR5, whereas, fingolimod binds to S1PR1, S1PR3, S1PR4, S1PR5. The lack of selectivity for fingolimod may be the reason why there are more AE associated with it. Ozanimod has shown great promise as a relapsing multiple sclerosis treatment through the trails previously mentioned. In the RADIANCE and SUNBEAM trials, ozanimod showed to be an effective drug by demonstrating significant reductions in ARR, in new and enlarging T2 lesions, and in gadoliniumenhanced MRI lesions (12). In the comparative study with fingolimod, ozanimod proved to be associated with a more favorable benefit-risk profile than fingolimod when considering the cardiac monitoring, safety outcomes, and efficacy outcomes (11). Based on this information, ozanimod has the potential to provide RRMS patients and their physicians an additional option for treating relapsing multiple sclerosis (11). Siponimod Another drug in the same family as ozanimod is siponimod. This drug, by Novartis Pharmaceuticals, is also a selective modulator for the S1PR1 and S1PR5 receptors (7, 15). Additional research has found that these S1P receptors are located abundantly on lymphocytes, oligodendrocytes, astrocytes, and erythrocytes (7). Siponimod is indicated for the treatment of SPMS. SPMS is a form of MS that progresses from RRMS, despite being treated (16). Siponimod is a lipophilic drug that can cross the blood-brain barrier, making it a viable drug to decrease the inflammation that could occur from the disease. Ginatile et al. investigated this potential benefit using mice that had experimental autoimmune encephalitis (EAE) and injecting the drug directly into the CSF. Their results demonstrated a dose-dependent effect, with a high dose of 4.5 mg/day completely inhibiting EAE. A lower dose of 0.45 mg/day improved disease severity (7). This study also noted a significant reduction in astrogliosis, which is an increase in the number of astrocytes resulting from the loss of nearby neurons from the CNS due to trauma or disease, and an increase in gamma-aminobutyric acid (GABA) signal in the striatum of siponimod-treated mice (7). When it comes to treating neuroinflammation, siponimod was shown to decrease oligodendrocyte and axonal loss (7). This suggests that it might be able to protect axons during both the acute and the chronic demyelinating phases of MS (7). These studies are important because even a minor change in expanded disability status scale (EDSS) for patients with SPMS, has a significant change in brain function and daily activities (15). EDSS is a scale to evaluate disability in MS and monitoring changes in the level of disability over time which range. The scale ranges from zero to 10 in 0.5 increments. Siponimod’s primary function was analyzed in a double-blind phase-III clinical trial by Kappos et al. which evaluated the effectiveness and safety of the drug in patients suffering from SPMS. The inclusion criteria were patients who were 18 to 60 years old who had a diagnosis of SPMS and documented moderate-to-advanced disability indicated by an EDSS score of 3.0 to 6.5 at screening, a history of RRMS, documented EDSS progression in the 2 years before the study. Eligible participants also had to have no evidence of relapse in the 3 months before randomization (15). The participants were randomized in a 2:1 manner, with the experimental groups receiving 2 mg of siponimod or a placebo. A full neurological 74

assessment was initiated at baseline to assess walking range, functional systems, and an EDSS score. This assessment was done every 3 months. MRI scans were also part of the assessment and were taken 12, 24, and 36 months from baseline. The primary endpoint was the start of the study to 3-month confirmed disability progression (CDP). CDP is defined as a 1-point increase in EDSS if the baseline score was 3.0 to 5.0, or a 0.5-point increase if the baseline score was 5.5 to 6.5, which would be confirmed at a visit at least 3 months later. The study also analyzed the data looking for secondary endpoints. These included the time to 3 months where the patient had a 20% decrease from baseline, when performing the 25-foot walk test (T25FW), ARR, and a change in T2 lesion volume. There was a total of 1,651 patients randomly assigned — 1,105 to the siponimod group and 546 to the placebo group. The duration of the study had a median time of 21 months with exposure to the study drug at 18 months (15). The characteristics of each group were similar. The study results showed 26% of the siponimod group versus 32% of the placebo had a 3-month CDP, with a hazard ratio of 0.79 which demonstrated a risk reduction of 21%, with a p-value of 0.0131. When the data is divided into smaller subgroups, the results favor siponimod over the placebo. Evaluation of CDP done at 6 months showed a reduction by siponimod with a hazard ratio of 0.74. This translated to a 26% risk reduction with a p-value of less than 0.05. ARR was decreased with siponimod when compared with placebo, from baseline to confirmed first relapse, with a risk reduction of 46% with a p-value of less than 0.0001 (HR=0.54). Lesion volume when compared to baseline, was lower with siponimod than with placebo, with a group difference of -695.3 mm3 and a p-value of less than 0.001. Brain volume decreased at a lower rate with siponimod than with placebo. Unfortunately, there wasn't any significant reduction in time to 3 months where the patient had a 20% reduction from baseline when performing the T25FW for the entire population of the study. Even though the clinical benefits of using siponimod are apparent, there are also AE. The highest frequency of AE includes headache, nasopharyngitis, urinary tract infection, and falls, which were reported in more than 10% of patients in both treatment groups (15). The more serious AE affected a low percentage of the population in the treatment groups. These included increased liver transaminase concentrations, basal cell carcinoma, concussion, depression, urinary tract infection, suicide attempts, gait disturbances, multiple sclerosis relapse, and paraparesis. There were also 4 deaths for each treatment group caused by metastatic gastrointestinal melanoma within 4 months of using siponimod. Patients who were in the siponimod treatment groups had a slight increase, proportionally, in adverse effect than the placebo group. This is due to the immune modulation of the drug. Siponimod also has an adverse effect on the cardiovascular system due to its mechanism of action. In comparison patients receiving this S1P modulation were more proportionally affected by bradycardia, hypertension, lymphopenia, and macular edema at the start of treatment. Convulsions were also more common with siponimod than with placebo. Summarily, the EXPAND clinical trial has shown a sizable reduction in 3-month CDP compared with placebo. This means

Novel Treatments for Multiple Sclerosis

it can be used as an effective disease-modifying treatment with those suffering from SPMS. LIFnano Scientists at Cambridge University have discovered a promising regenerative treatment known as leukemia inhibitory factor (LIF). LIF is a stem cell growth factor that has the potential to slow the progression of MS while also repairing damage to the CNS seen in diseases like MS (18). LIF is a four-helix bundle protein and is a part of the IL-6 cytokine family expressed in the hypothalamus and anterior pituitary (18–20). It has been shown to promote neuroprotection while also possessing therapeutic benefits for treating MS (18). LIF is known to act on cell surface receptors, gp190, and has anti-inflammatory effects (Figure 1). Interleukin-6 (IL-6), a proinflammatory cytokine suppresses the transcription of the LIF receptor (18, 19). This suppression makes the cells unable to contact LIF; this, in theory, suppresses the anti-inflammatory processes of LIF (18). IL-6 is shown to be elevated in the plasma of patients with MS (19). LIF is naturally occurring in the CNS and also required for myelination and remyelination (18). LIF-knockout mice have been shown to have a delay during development (22). It has been recognized that LIF regulates

the lineage of regulatory T cells. This is vital in treating MS because MS is known to be a T-cell-mediated autoimmune disease. LIF has also recently been discovered to be a promyelination factor (22). Despite all of its benefits, LIF has a short half-life (20 minutes). LIF’s rapid degradation in vivo warranted further research as a potentially effective treatment for MS (18). Researchers used nanoparticles to overcome the short half-life of LIF and determine whether it could be effectively delivered to the CNS. Nanoparticles provide a sustained-release delivery system through the blood-brain barrier and are proven to be biodegradable and biocompatible (18, 19). The nanoparticles were used to target the delivery of LIF to oligodendrocyte precursor cells (OPC) within the brain. OPCs help areas of damage and promote differentiation into mature oligodendrocytes which then are able to repair myelin (22). Researchers conducted an experiment to identify if improved OPC biological function was due to the signaling induced by LIF-nanoparticle (LIF-NP) treatment. In vivo, the goal was to demonstrate OPCs to survival, proliferation, and differentiation for remyelination to occur (22). Researchers assigned six treatment groups, Group 1: untreated controls,

Figure 1. Schematic of the leukemia inhibitory factor (LIF) pathway (20). 75

Novel Treatments for Multiple Sclerosis

Group 2: recombinant LIF (positive control) (100 U/ml), Group 3: NG2- targeted LIF-NP (300 mg/ml), Group 4: non-targeted LIF-NP (300 mg/ml), Group 5: NG2-targeted empty-NP (300 mg/ml), Group 6: non-targeted empty-NP (300 mg/ml). LIFNP treatment was given to each group and survival and proliferation were measured after being treated for 24 hours. Myelin basic protein (MBP) was assessed for differentiation, and the results showed increased MBP in the group treated with NG2-targeted LIF-NP (19). The group that received the NG2-targeted LIF-NP treatment was seen to have a significantly greater number of myelin basic proteins and also oligodendrocytes compared to the targeted empty-NP group. This solidifies the requirement for LIF delivery and targeting specific NG2 cells to accomplish enhanced OPC maturation (19). The next step was to examine the LIF-NP in vivo. Myelin toxin lysophosphatidylcholine (LPC) was injected into the corpus callosum of mice to simulate a model of CNS demyelination. The three treatment groups for the in vivo study consisted of, NG2-targeted LIF-NP, non-targeted LIF-NP, NG2-targeted empty-NP. The mice were then injected with the LIF-NP in the exact same location 8 days following the LPC injection. Remyelination in the mice was then examined by electron microscopy 10 and 17 days following the LIF-NP treatment. The electron micrographs show increased remyelination in mice treated with targeted LIF-NP compared to non-targeted LIF-NP and the targeted empty-NP. It was identified that the group that received NG2-targeted LIF-NP had the highest percentage of myelinated fibers and the lowest mean G ratio. This indicated the NG2-targeted treatment group had the thickest myelin and also the thickest axon diameter when compared to the other groups. There were no AE observed in any of the treated mice. This study confirmed that in vitro NG2-targeted LIF-NP were able to bind to OPCs and promote differentiation, while in vivo the NG2-targeted LIF-NP improved remyelination within the CNS (22). This research revealed how LIF-NP can contribute to neuroprotection while also remyelinating within the CNS. This technology will hopefully be able to halt the neuro-axonal damage while also promoting remyelination in the CNS. Clinical trials with this therapeutic technology will begin in 2020 (18).

Discussion As of today, MS remains one of the most prevalent debilitating diseases seen in young adults. Current pharmacotherapies on the market for treating MS focus on slowing down the progression of the disease while only managing symptoms. Ozanimod is an S1P modulator that has shown to be effective in reducing ARR, while also decreasing the size of T2 MRI brain lesions and T1 gadolinium-enhanced MRI lesions when compared to the traditional IFβ-1a treatment options. Siponimod, also an S1P modulator, was proven to be an effective treatment for SPMS. These results show that ozanimod was better at reducing the disease progression when compared to IFβ-1a, which is an approved treatment for RRMS. The AE profile of ozanimod is similar to currently approved treatments on the market. As of March 2020, ozanimod has been FDA approved for the treatment of multiple forms of relapsing MS. 76

Patients treated with siponimod showed a larger reduction in CDP. This study is important, because this drug has consistently proven to reduce CDP in SPMS patients. This reduction corresponds to small changes in EDSS which can present as significant changes in neurological functions and day-to-day activities. When comparing siponimod with other S1P receptor modulators, like fingolimod, the safety profile is similar and produces less bradycardia. Patients that use this drug should be under medical supervision until the bradycardia is gone, which occurs after about 7 days. This study corrected this with a dose titration but if this adverse effect could be rectified, siponimod can be very beneficial to slow the progression of SPMS. This study could have tested patients who have relapsed recently and determine if the drug has any effect on active SPMS. The duration of the study was not long enough to determine long term AE associated with continued drug use. Siponimod should be evaluated for how selective it is for S1PR1 and S1PR5 and if genetics play a role in its metabolism. With these AE, siponimod has currently been approved for the treatment of SPMS in both the European Union and the U.S. Ozanimod and siponimod demonstrated success in slowing the progression of the disease with less frequent AE when compared to current MS treatment options. These drugs did not reverse the damage already caused by MS. LIF-NP has been proven safe and effective in animal models of MS. Since LIF-NP uses a molecule that is naturally occurring in the human body, the presence of AE is very limited compared to other MS treatment options. LIF-NP not only displayed neuroprotective activity in the CNS but it also assisted in the repair and remyelination of axons by promoting the differentiation of OPCs. This suggests that LIF-NP can prevent the progression of neurodegenerative diseases like MS while also promoting tissue repair in previously damaged areas of the CNS. While further clinical research is needed to be conducted in humans to test for safety and efficacy, the promising data seen in the animal models show that LIF-NP has the potential to be a curative treatment for MS.

References 1.

Lemus H, Warrington A, Rodriguez W. Multiple Sclerosis: Mechanisms of Disease and Strategies for Myelin and Axonal Repair. Neurol Clin. 2018;36(1): 1-11.

2. Lassmann H. Multiple Sclerosis Pathology. Cold Spring Harb Perspect Med. 2018;8(1). 3. Garg N, Smith T. An Update on Immunopathogenesis, Diagnosis, and Treatment of Multiple Sclerosis. Brain Behav. 2015;5(1). 4. Hunter S. Overview and Diagnosis of Multiple Sclerosis. Am J Manag Care. 2016;22(6): S141-S150. 5. Hart F, Bainbridge J. Current and Emerging Treatment of Multiple Sclerosis. Am J Manag Care. 2016;22(1): S159-S170. 6. Dargahi N, Katsara M, Tselios T. Multiple Sclerosis: Immunopathology and Treatment Update. Brain Sci. 2017;7(1): 1-27. 7.

Reich D, Lucchinetti C, Calabresi P. Multiple Sclerosis. N Engl J Med. 2018;378(2): 169-180.

Novel Treatments for Multiple Sclerosis

8. Goodman A, Anadani N, Gerwitz L. Siponimod in the Treatment of Multiple Sclerosis. Expert Opin Investig Drugs. 2019;28(12): 1051-1057.

21. Steckler T, Kalin NH, Reul JMHM. Handbook of Stress and the Brain Part 2. (1st ed.). Amsterdam: Elsevier Science; 2005.

9. Cohen J, Comi G, Arnold D, Bar-Or A, Selmaj K, Steinman L et al. Efficacy and safety of ozanimod in multiple sclerosis: Dose-blinded extension of a randomized phase II study. Multiple Sclerosis Journal. 2018;25(9):1255-1262.

22. Rittchen S, Boyd A, Burns A. Myelin Repair in Vivo Is Increased by Targeting Oligodendrocyte Precursor Cells With Nanoparticles Encapsulating Leukaemia Inhibitory Factor (LIF). Biomaterials. 2015;56(2): 78-85.

10. U.S. Food and Drug Administration Approves Bristol Myers Squibb’s ZEPOSIA® (ozanimod), a New Oral Treatment for Relapsing Forms of Multiple Sclerosis | BMS Newsroom [Internet]. 2020 [cited 21 May 2020]. Available from: press-release/corporatefinancial-news/us-food-and-drugadministration-approves-bristol-myers-squibbs 11. Swallow E, Patterson-Lomba O, Yin L, Mehta R, Pelletier C, Kao D et al. Comparative safety and efficacy of ozanimod versus fingolimod for relapsing multiple sclerosis. Journal of Comparative Effectiveness Research. 2020;9(4):275285. 12. Koscielny V. Phase III SUNBEAM and RADIANCE PART B trials for ozanimod in relapsing multiple sclerosis demonstrate superiority versus interferon-β-1a (Avonex®) in reducing annualized relapse rates and MRI brain lesions. Neurodegenerative Disease Management. 2018;8(3):141-142. 13. Markowitz C. Interferon-beta: Mechanism of action and dosing issues. Neurology. 2007;68(Issue 24, Supplement 4):S8-S11. 14. Celgene's Ozanimod Versus Biogen's Avonex: Celgene Has the Edge | BioSpace [Internet]. BioSpace. 2020 [cited 2020 May 21]. Available from: article/celgene-s-ozanimod-versus-biogen-s-avonexcelgene-has-the-edge/ 15. Kappos L. Siponimod Versus Placebo in Secondary Progressive Multiple Sclerosis (EXPAND): A Double-Blind, Randomised, Phase 3 Study. Lancet. 2018;391(1): 12631273. 16. Lassmann H, Horrssen J, Mahad D. Progressive Multiple Sclerosis: Pathology and Pathogenesis. Nat Rev Neurol. 2012;8(11): 647-56. 17. Costantino C, Baecher-Allan C, Hafler D. Multiple Sclerosis and Regulatory T Cells. J Clin Immunol. 2008;28(6): 697706. 18. Metcalfe S. LIF and Multiple Sclerosis: One Protein With Two Healing Properties. Mult Scler Relat Disord. 2018;20(1): 223-227. 19. Metcalfe S, Strom T, Williams A, Fahmy T. Multiple Sclerosis and the LIF/IL-6 Axis: Use of Nanotechnology to Harness the Tolerogenic and Reparative Properties of LIF. Nanobiomedicine. 2015;5(2). 20. Graf, Urs & Casanova, Elisa & Cinelli, Paolo. (2011). The Role of the Leukemia Inhibitory Factor (LIF) — Pathway in Derivation and Maintenance of Murine Pluripotent Stem Cells. Genes. 2. 280-97. 10.3390/genes2010280.


Scholarly Research In Progress • Vol. 4, October 2020

Systematic Review of Viral mRNA Vaccine Formulations, Modifications, and Adjuvants for Enhanced Safety and Efficacy Christian Pardo1*

Geisinger Commonwealth School of Medicine, 525 Pine St., Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Traditional viral vaccines and adjuvants are associated with safety and efficacy concerns and have proven to be an ineffective tool in immediate response to outbreaks and pandemics. Validating studies and clinical trials have shown mRNA to be a nonimmunogenic, safe, and effective therapy for immunization. mRNA vaccines are inexpensive, easy to mass produce, and can potentially be used in response to uncontained viral outbreaks. Encoded in the mRNA is the selected epitope sequence to be translated and presented by antigen presenting cells. It is reported that adjuvating activation molecules such as cytokines may also be encoded in the mRNA strand. The mRNA strand itself must be designed in a way that eukaryotic cells can translate it. As such, more success has been reported with the inclusion of a proper 5’ cap, 3’ and 5’ UTRs, and a poly-A tail. Furthermore, select base modifications, particularly 1-methylpseudouridine, have shown increased expression and as a result, lower doses may present the same benefits. Finally, various delivery methods have been described. While naked mRNA has shown the ability to be internalized and expressed, various nanoparticle technologies are reported to expedite and facilitate this process by reducing immunogenicity and protecting the mRNA from degradation. With initial reports of cellular and humoral responses equal to or greater than traditional vaccines along with low numbers of adverse events and traditional vaccine associated pathologies, mRNA vaccines present a new era of inoculation.

Introduction Traditional vaccine technology has successfully eradicated viral diseases such as smallpox and rinderpest (1, 2). Fifteen vaccines have been licensed for use in humans, but developed of new vaccines since 2006 has been limited (3). Additionally, live animal, egg and cell culture, and molecular biology techniques are time-consuming and expensive in response to outbreaks and pandemics (3). This is highlighted by the difficulties containing outbreaks and high illness and mortality rates observed. Coronaviruses are associated with respiratory illness in humans ranging from the common cold to severe acute respiratory syndrome and Middle East respiratory syndrome (4). The most recent coronavirus disease-2019 outbreak has over 2 000 000 confirmed cases and 140 000 confirmed deaths (5). Influenza A, coronavirus, and hanta virus present respiratory viral outbreak and pandemic threats (6). These respiratory viruses, norovirus, hepatitis C virus (HCV), Ebola virus, and many others regularly emerge and reemerge as threats to human health and prove challenging to contain. Improvements in surveillance and outbreak response protocols will help but that does not negate the need for improvements in vaccinations (7, 8). 78

In addition to the time-consuming, expensive, and impractical nature in current vaccine manufacturing techniques for outbreak response, issues regarding efficacy, safety, and immunogenicity need to be addressed. Current vaccines have been associated with pathologies and adverse reactions to the vaccines, adjuvants, and injection site adverse reactions. One study on herpes zoster vaccinations in older adults found that vaccinated groups had higher incidences of adverse reactions and injections site adverse reactions (9). The authors noted that although the adverse events were rated mild to moderate, subjects receiving the vaccination were at higher risk of not returning for following doses compared to placebo. Respiratory syncytial virus is one of the leading causes of respiratory illness in toddlers and young children, but live attenuated vaccines have been associated with vaccine enhanced respiratory disease, which is worse than the illness caused by the virus. It is therefore a danger that protein-based vaccinations carry this possibility as well (10). As mentioned before, there is a very limited number of available viral vaccines. Dengue, cytomegalovirus (CMV), human immunodeficiency virus (HIV), HCV, Ebola virus, and RSV are among just a few viruses with no available vaccinations using current platforms, yet they recurrently pose threats to human health. Viral vaccine research has seen limited advancement in the last century. While better formulations of subunit-based, attenuated, and inactivated vaccines and improvements in adjuvant technology have occurred, they have not been enough to overcome the stated challenges. A novel platform is needed for a more precise design that can produced less expensively, much quicker, and will also address safety and efficacy concerns of protein based vaccines. Nucleic acid therapy provides this opportunity. Although DNA had not yet been identified, natural translation of exogenous genetic material in non-expressing organisms was first proposed in 1928 as non-virulent bacteria were made virulent after exposure to heat-killed virulent strains termed the “transforming principle” (11). Transformation in a laboratory using electroporation was done in the 1980s (12). Attempts to introduce DNA for therapeutic application such as protein replacement in protein deficiency diseases was met with great difficulties. The vectors used were inefficient and dangerous and genomic DNA integration had dangers of its own along with ethical considerations. mRNA and viral RNAs were translated in frog oocytes, but the idea of RNA therapeutics was not initially well entertained due to its unstable structure (13, 14). Despite this, experiments showed mRNA’s transient nature did not pose much of a barrier and in fact provided advantages over DNA therapy as delivery is much easier and it does not pose risks of genomic integration, and envelopment in lipid nanoparticles (LNPs) was a simple and

Systematic Review of Viral mRNA Vaccine

effective solution to protect it from degradation and facilitate intracellular delivery (15). With the validation of mRNA as a strong vaccine platform candidate, a lot of research has been done to make the technology better. To date, it has been shown that structural elements mimicking human mRNA such as 5’ caps, 3’ and 5’ untranslated regions (UTRs), and poly-A tails, reduce immunogenicity and improve translation. Base pair modifications not commonly seen in endogenous mRNA have also show similar benefits. As previously mentioned, LNP delivery methods allowed for more efficient cellular uptake, but alternative nanoparticle technologies, such as polymer nanoparticles, and formulations, such as antigen-coated nanoparticles, may be more effective than LNP technologies. Most importantly, the coding sequence can be precisely chosen to match not only the correct epitope but can be mutated to match epitope conformations easier for cells and antibodies to recognize and be constructed to contain costimulatory molecules and cytokines for an improved immune response. This paper explores research done to date and elucidate what combinations may provide the best benefits.

Methods Databases including PubMed, Cochrane Library, Ovid, Wiley Online Library, and SAGE Journals were searched for articles mentioning mRNA vaccinations. Various iterations of the search terms antiviral vaccine, mRNA vaccine, mRNA vaccine adjuvants, and mRNA vaccine design were searched. The articles collected were vetted by title, abstract, and discussion to identify relevant articles. Generally, the articles included were regarding research towards viral mRNA vaccines in animal models and humans. Studies aimed at validating exogenous mRNA translation in vivo, protein replacement therapies, and cancer research were generally excluded unless aspects of the study were relevant to the development of viral vaccines, such as adjuvants and modifications that are applicable in both cases.

Discussion Internalization and presentation Extracellular nucleic acids have been postulated to be internalized using scavenger receptor A (16). Evidence for scavenger receptor mediated internalization of naked mRNA through caveolae/lipid raft rich membrane domains has also been found in smooth muscle cells, keratinocytes, fibroblasts, and dendritic cells (17). The non-APCs most likely presented antigens on class I MHCs, but the mRNA translation mechanism after internalization has not been fully characterized. A study of the same year found that naked mRNA injected into lymph nodes was selectively up-taken through micropinocytosis by dendritic cells (DCs) (18). As will be discussed later, LNP and CNP capsules have been used as vehicles for delivery and their internalization mechanism has not been described but can be assumed to be through vesicle fusion and an unknown endosome escape mechanism (19, 10). Escape and sorting mechanisms within cellular compartments have not been fully elucidated for APCs either, but the mRNA

is presumably translated on free ribosomes in the cytosol. The intracellular origin of the antigen initially hints at class I MHC presentation in APCs as well, and while possible, APCs generally present on class II MHCs (20). Understanding the escape, translation, and presentation mechanisms in nonAPCs and APCs will allow better formulations with appropriate signaling and costimulatory molecules to be designed. Vaccine design As the translating and presenting system is a eukaryotic cell, viral genetic material, whether positive or negative sense, single- or double-stranded, must first be transcribed to a single-stranded mRNA molecule. Eukaryotic translation is a tightly regulated process with various stringent components. The mRNA molecule must be designed to resemble eukaryotic mRNA in order to facilitate the translation process. This includes the incorporation of a modified guanosine 5’ cap, 5’ and 3’ untranslated regions (UTRs), and a 200 nucleotide long poly-A tail (21). Shorter poly-A tails of 120 nucleotides have also been reported to independently enhanced mRNA stability (22). There has been debate regarding the UTR sequences. Tanguay and Gallie reported that shorter UTRs (19 bases) yielded higher expression than longer UTRs (156 bases) and this difference in expression was length and not sequence dependent (23). However, a 13 nucleotide sequence deletion in the 3’ UTR of the β- globin gene in two individuals with β -thalassemia suggest UTR sequence is important for proper expression of mRNA as well (24). α and β- globin 5’ and 3’ UTRs are well accepted sequences in mRNA design as β -globin 5’ and 3’ UTRs improve translation and α- globin 3’ UTRs improve stability (25). There is less debate on optimal 5’ sequence than 3’ sequence. Three independent studies agreed on 5’ UTR sequence, but one used a much shorter 3’ UTR and the other two had 3’ UTRs of the same length but differed in sequence (shown below). H9N2 immunization in chickens (19): 5' UTR: G​G​GA ​ ​A​A​U​A​A​GA ​ ​G​A​G​A​A​A​A​G​A​A​G​A​G​U​A​A​GA ​ ​A​G​A​A​A​U​A​ U​A​A​G​A​G​C​C​A​C​C​ 3` UTR: UGAUAAUAGGCUGGAGCCUCGGUGGCCAU POWV immunization in mice (26): 5′ UTR: G​G​G​A​A​AU ​ ​AA ​ ​G​AG ​ ​A​G​A​AA ​ ​AG ​ ​A​A​G​A​G​U​A​A​G​A​A​G​A​A​AA ​​ U​A​A​G​A​G​C​C​A​C​C​ 3′ UTR: G​CU ​ ​G​G​A​G​C​C​U​C​G​G​U​G​G​C​C​U​A​GC ​ U ​ U ​ C ​ ​U​UG ​ ​CC ​ C ​ ​C​UG ​ ​ G​C​CU ​ ​C​C​CC ​ ​C​C​A​G​C​C​C​C​U​C​C​U​C​C​C​C​U​U​C​C​U​G​CA ​ ​C​C​C​G​U​A​C​C​ C​C​CG ​ ​U​G​GU ​ ​C​U​UU ​ ​GA ​ ​AU ​ ​A​A​AG ​ ​U​C​U​GA ​ ​G​U​G​G​G​C​G​G​C​ ZIKV immunization in mice (27): 5’ UTR: G​G​G​A​A​A​U​A​A​G​A​G​AG ​ ​A​A​A​AG ​ ​A​AG ​ ​A​G​U​A​A​G​A​A​G​A​A​AU ​ ​ A​U​A​A​G​A​G​C​C​A​C​C​ 3’ UTR: U​G​A​UA ​ ​A​U​A​G​G​C​U​G​G​A​G​C​C​U​C​G​G​U​G​G​C​C​A​U​G​C​U​U​C​ U​U​GC ​ ​C​C​CU ​ ​U​G​GG ​ ​C​C​ ​U​C​C​C​C​C​C​A​G​C​C​C​C​U​C​C​U​C​C​C​C​U​U​C​C​ U​G​C​A​C​C​C​G​U​A​C​C​C​C​C​G​U​G​G​U​C​U​U​U​G​A​A​U​A​A​A​G​U​C​U​GA ​ ​ These sequences all enhanced stability and translation independently, but no studies comparing these sequences directly were found. Characterization of a more universally accepted 3’ UTR could be of clinical use to identify the most potent sequence and better regulate production for more efficacious vaccines. 79

Systematic Review of Viral mRNA Vaccine

Various base modifications have been proposed to reduce immunogenicity and improve stability and translation as well. Replacement of uridine with 5-methyluridine showed a nearly 2-fold increase in translation (21), postulated to occur by inhibition of protein kinase R mediated translation inhibition which generally occurs during viral attack (26). Inclusion of a 5-methylcytidine modification also showed a promising 4-fold translation enhancement in cultured mouse cells (21). Combined 1-methylpseudouridine/5-methylcytidine modifications was more recently found to increase protein luciferase expression up to 44 times greater than combined pseudouridine/5-methylcytidine (27). The animal and human studies examined used unmodified mRNA or single 1-methylpseudouridine modifications. Considering the above mentioned results, characterization of double modified mRNA induced antigen responses may be of interest.

and others. While naked mRNA has shown the ability for internalization and translation, various limitations have hindered advancement in naked mRNA immunization including immunogenicity and rapid degradation, as studies employing naked mRNA, especially in human trials, have shown increased AEs and less robust immune responses (18). Lipid nanoparticles (LNPs) are the most extensively used vehicles as they are the simplest to produce, are non-immunogenic, and facilitate endocytosis. Various polymer nanoparticles have also been studied in animal models (Table 1), but regulatory guidelines have greatly limited research in humans and no human applications of polymer nanoparticles were found. This is certainly an area that warrants further development as chitosan nanoparticles (CNPs) have provided greatly enhanced immune response and decreased AEs and a diverse range of drug delivering nanoparticles are being developed (19).

The most foundational element in mRNA vaccine design is the antigen coding sequence. Viruses have relatively few genes coding for non-structural and structural elements from polymerases to spike and fusion proteins. Chosen antigen sequences vary by virus, but are usually surface protein subunits of naturally targeted epitopes. Common targets for influenza strains, RSV, and ZIKV include hemagglutinin (HA) and matrix protein 2 (M2e), fusion protein (F), and premembrane (prM) and envelope protein (E) respectively (19, 28, 10, 29). Mutations can also be introduced to stabilize more favorable antigen conformations for immune recognition, particularly for antigens that undergo major conformational changes, such as the RSV F protein, but not so much for certain glycoproteins. Espeseth et al. highlighted this by creating a pre-fusion stabilized F antigen encoding mRNA that elicited higher responses than other conformational stabilized variants (10).

Safety The abovementioned considerations of mRNA vaccine design allow for protection against degradation and facilitate internalization and translation processes, but also help further regulate inherent immunogenicity and allow for safer vaccinations with less AEs. Animal studies regularly report successful immunizations without reporting AEs at injection sites, mRNA related reactions, adjuvant related reactions, or nanoparticle induced toxicities (10, 19, 36, 29). Animal models

mRNA vaccines may include costimulatory molecules such as traditional adjuvants or mRNA encoded molecules such as cytokines. TriMix encodes CD40L (T cell proliferation), TLR-4 variant (T-cell survival), and CD70 (induction of dendritic cell maturation). This is the most extensively used cytokine and allows for a more robust and specific cell-mediated response (30). Finally, various nanoparticle technologies have been proposed for mRNA delivery including lipids and lipid derivatives, polymers, Table 1. List of various lipid and polymer based nanoparticle delivery methods in animal models.


Systematic Review of Viral mRNA Vaccine

have also reported no incidences of vaccine exacerbated pathologies. RSV vaccinations were discontinued because they could result in more serious infections, such as vaccine enhanced respiratory disease, instead of attenuating the virus, but this phenomenon was not observed in RSV mRNA vaccinated cotton rats (10). Investigations in non-human primates monitoring for injection site reactions, body weight, rectal temperature, hematologic variables, and clinical chemistry variables did not report adverse effects in response to self-amplifying HIV-1 mRNA intramuscular injections (36). These studies provided preclinical evidence validating the safety of mRNA vaccines for advancement to clinical trials. Few mRNA vaccine clinical trials have been completed to assess safety and AEs, but initial results have shown mRNA vaccines to be well tolerated. A randomized placebocontrolled double-blind phase 1 trial using a chemicallymodified LNP-encapsulated H10N8 HA mRNA vaccine on a two-dose schedule 3 weeks apart (intramuscular; 25, 50, 75, 100, and 400 μg and intradermal; 25 and 50 μg) and H7N9 HA mRNA vaccine on a two-dose schedule six months apart (intramuscular; 10, 25, and 50 μg) did not report any vaccine related severe AEs (N=357) (28). Two grade three solicited AEs (injection site erythema and headache) were reported at the 400 μg intradermal dose and this treatment arm was discontinued. A randomized placebo-controlled double-blind clinical trial delivering intradermal immunization with HIV-1 Gag and Nef mRNA transfected DCs reported no grade three or four AEs (N=10) (37). A phase 1 HIV-1 naked mRNA vaccine (iHIVRNA) study in 21 males delivering intranodal TriMix-300 g with 300, 600, and 900 g T-cell immunogen (HTI), an mRNA sequence targeting more vulnerable sites than traditional Gag and Nef targeting sequences, showed 31 grade 1/2 adverse effects, of which two had a definite relationship to the vaccine, and one grade 3 AE (30). A phase 2a randomized placebocontrolled double-blind study further evaluating internodal naked iHIVRNA vaccination in 40 HIV patients is currently being conducted (38). Based on the large sample size and low AE rates of the influenza investigation, NP mRNA delivery appears to be correlated with lower rates of AEs. The HIV studies all had much smaller sample sizes but more reports of AEs. Due to the small sample sizes (N<30), this observation may not be significant. Additionally, although HIV participants were included on the basis of successful ART and acceptable immune system function parameters, there may be a predisposition to AEs based on a compromised immune system. Animal investigations more readily employ nanoparticle delivery techniques, which may not be as easy in human trials due to regulatory concerns. Regardless, a phase I clinical trial assessing the safety of the LNP delivered mRNA-1273 vaccine, encoding the SARS-CoV-2 spike protein and showed success and is being advanced to stage II and III clinical trials. Nanoparticle delivery has shown a large degree of preclinical success and may warrant advancement of this research to clinical phases. Efficacy In animals, the efficacy of mRNA vaccines has been greatly supported. An LNP-encapsulated base-modified Powassan virus (POWV) mRNA vaccine encoding prM and E proteins

was given to mice in one or two intramuscular doses of 10 μg. After the first and second immunization, the 50% maximal effective inhibitory concentration titers were approximately 2-fold and 3-fold higher than placebo, respectively. In response to challenge, 100% of placebo-treated mice died, and all POWV immunized mice survived (39). HA2 and M2e peptide antigen-coated CNPs delivering H9N2 HA2 and M2e 1-methylpseudouridine-modified mRNA (CNP+RNA+Pr), empty antigen coated CNPs (CNP+Pr), or uncoated empty CNP were intranasally delivered to chickens and IgG, IgA, CD4+, and CD8+responses (19). HA2 specific IgG antibodies were significantly higher in CNP+RNA+Pr treated mice than CNP+Pr and CNP treated mice while M2e specific IgG antibodies were not statistically different between CNP+RNA+Pr and CNP+Pr. Both HA2 and M2e specific IgA antibodies were significantly higher for CNP+RNA+Pr than CNP+Pr treated mice. Virus neutralization titers were also significantly higher in CNP+RNA+Pr treated mice than CNP+Pr treated mice. CD4+ responses were reported at a higher degree than CD8+ responses in all mice, but CNP+RNA+Pr showed a significantly greater amount of CD4+ and CD8+ cell responses. The viral load in CNP and CNP+Pr mice was approximately 27 and 3 times higher than CNP+RNA+Pr mice respectively. LNP encapsulated 1-methylpsudouridine modified RSV F protein coding mRNA was intramuscularly delivered to mice in two 10 μg doses 3 weeks apart and compared to the RSV DS-Cav1 protein antigen with an aluminum phosphate-based adjuvant (10). Various mRNA sequences corresponding to conformation stabilized versions of F protein were included in the study to identify the most immunogenic conformation. They all elicited comparable or greater serum neutralization titers than the control. Control mice did not have detectable CD4+ or CD8+ responses while mRNA treated mice had significantly higher CD4+ and CD8+ responses as observed by increased IL-2, IFN- γ, and TNF- α. HIV-1 gp140 encoding self-replicating mRNA formulated with cationic nanoemulsion was used to intramuscularly immunize rhesus macaques with 50 μg mRNA. mRNA gave anti-gp140 titers comparable to gp140 protein immunized controls and antigen-specific B-cells levels of 1,500 per 106 peripheral blood monocytes versus 1,400 per 106 PBMCs observed in controls. mRNA immunization also provided greater cellular immune response as seen by IFN-g levels approximately 3 times greater than in protein immunized mice. Animal models have repeatedly shown comparable or greater specific systemic and mucosal humoral and cellular immune responses to mRNA produced antigens as observed by IgG, IgA, neutralizing antibodies, and IFN- γ, TNF- α, and interleukin producing CD4+ and CD8+ cells. Based on the safety and efficacy demonstrated in animal models, clinical trials evaluating safety as primary endpoints and efficacy as secondary/exploratory endpoints are currently underway. Completed and current clinical trials for viral mRNA vaccines are mostly at phase 1, one study found was at phase 2, and none found were at phase 3. The phase 2 study is in progress and aims to evaluate the SARS-CoV-2 mRNA-1273 vaccine by Moderna and is expected to be completed August 2021 (40). Exploratory endpoints for all trials have revealed mixed results. As seen in Table 2, the clinical trials included induced some degree of humoral response but failed to elicit protective cell-mediated responses. Naked delivery of HIV targeting mRNA to DCs in vitro and in vivo did not elicit 81

Systematic Review of Viral mRNA Vaccine

Table 2. Completed clinical trials evaluating safety as the primary endpoint and assessing efficacy as a secondary/exploratory endpoint.

strong immune responses. As previously stated, naked mRNA delivery has limitations that can be overcome with nanoparticle delivery. In addition, Gag is a particularly difficult immunization target due to high mutation rates which constantly change the epitope. Needle-free administration of rabies vaccination did elicit humoral responses, but syringe administration did not. This highlights another variable to consider in the development of mRNA vaccines.

Acknowledgments I would like to thank Dr. Kimberly Miller and Dr. Darina Lazarova for their guidance and the opportunity to prepare this manuscript. I would also like to thank the reviewers for their comments and suggestions.

Disclosures The author has no relevant disclosures.

Conclusion mRNA vaccines can elicit robust humoral and cell-mediated immune responses as validated in countless animal models. Strong mRNA vaccine candidates include classical eukaryotic mRNA structures including a 5’ cap, 5’ and 3’ UTRs, and a poly-A tail. Several UTR sequences of varying lengths have independently enhanced translation. Unlike most eukaryotic mRNA however, select base modifications improve stability and translation and decrease immunogenicity of the mRNA molecule. 1-methylpseudouridine and 5-mythylcytidine have provided the most promising results but further investigation into the combined use of these is warranted. Co-administration with T-cell and DC proliferation and survival cytokines (TriMix) has supported stronger immune responses. While naked mRNA injections have been validated to produce protein products, LNPs and protein coated CNPs have shown enhanced uptake for translation and presentation.


Safety in clinical trials has been supported as low numbers of AEs are reported. Analysis of various biomarkers in preclinical investigations, including IFN-g, TNF-a, IgG, IgA, CD4+, and CD8+, have shown efficacy equal to or greater than traditional subunit or whole virus-based vaccines in animal models, but clinical trials have not shown protective cell-mediated immune responses. Continued research into nanoparticle delivery and characterizing the most efficacious 5’ and 3’ UTRs and epitope sequences will provide the greatest advancements to this platform. Although initial clinical trials have not met efficacy expectations set by animal models, initial successes of developing immunity in animals against viruses with no current vaccinations still makes mRNA vaccines a thrilling prospect and the available technology makes these challenges exciting to face and overcome.

6. Roberts H, Lim WS. Viral lung infections and the potential for a human pandemic. Medicine. 2008;36(6):291–94.



Normile D. Driven to Extinction. Science. 2008: 21;319(5870):1606.

2. Fenner F, Henderson D, Arita I, Jezek Z, Ladnyi I. Smallpox and its Eradication. Geneva: World Health Organization; 1988. 3. Graham BS. Advances in antiviral vaccine development. Immunol Rev. 2013;255(1):230–42. 4. Maier HJ, Bickerton E, Britton P, editors. Coronaviruses: Methods and Protocols [Internet]. New York, NY: Springer New York; 2015. (Methods in Molecular Biology; vol. 1282). 5. Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real time. Lancet Infect Dis. 2020; S1473309920301201.


de Lusignan S, Lopez Bernal J, Zambon M, Akinyemi O, Amirthalingam G, Andrews N, et al. Emergence of a Novel Coronavirus (COVID-19): Protocol for Extending Surveillance Used by the Royal College of General Practitioners Research and Surveillance Centre and Public Health England. JMIR Public Health Surveill. 2020; 2;6(2):e18606.

8. Sims S, Longmire AG, Campo DS, Ramachandran S, Medrzycki M, Ganova-Raeva L, et al. Automated quality control for a molecular surveillance system. BMC Bioinformatics. 2018; 19(S11):358.

Systematic Review of Viral mRNA Vaccine

9. Gagliardi AM, Andriolo BN, Torloni MR, Soares BG, de Oliveira Gomes J, Andriolo RB, et al. Vaccines for preventing herpes zoster in older adults. Cochrane Acute Respiratory Infections Group, editor. Cochrane Database Syst Rev. 2019.

22. Holtkamp S, Kreiter S, Selmi A, Simon P, Koslowski M, Huber C, et al. Modification of antigen-encoding RNA increases stability, translational efficacy, and T-cell stimulatory capacity of dendritic cells. Blood. 2006; 108(13):4009–17.

10. Espeseth AS, Cejas PJ, Citron MP, Wang D, DiStefano DJ, Callahan C, et al. Modified mRNA/lipid nanoparticlebased vaccines expressing respiratory syncytial virus F protein variants are immunogenic and protective in rodent models of RSV infection. Npj Vaccines. 2020;5(1):16.

23. Tanguay RL, Gallie DR. Translational efficiency is regulated by the length of the 3’ untranslated region. Mol Cell Biol. 1996; 16(1):146–56.

11. Griffith F. The Significance of Pneumococcal Types. J Hyg. 1928;27(2):113–59. 12. Wirth R, Friesenegger A, Fiedler S. Transformation of various species of gram-negative bacteria belonging to 11 different genera by electroporation. Mol Gen Genet MGG. 1989; 216(1):175–7. 13. Gurdon JB, Lane CD, Woodland HR, Marbaix G. Use of Frog Eggs and Oocytes for the Study of Messenger RNA and its Translation in Living Cells. Nature. 1971; 233(5316):177–82. 14. Laskey RA, Gurdon JB, Crawford LV. Translation of Encephalomyocarditis Viral RNA in Oocytes of Xenopus laevis. Proc Natl Acad Sci. 1972; 69(12):3665–9. 15. Martinon F, Krishnan S, Lenzen G, Magné R, Gomard E, Guillet J-G, et al. Induction of virus-specific cytotoxic T lymphocytesin vivo by liposome-entrapped mRNA. Eur J Immunol. 1993; 23(7):1719–22. 16. Kawabata K, Takakura Y, Hashida M. The Fate of Plasmid DNA After Intravenous Injection in Mice: Involvement of Scavenger Receptors in Its Hepatic Uptake. Pharm Res. 1995;12(6):825–30. 17. Lorenz C, Fotin-Mleczek M, Roth G, Becker C, Dam TC, Verdurmen WPR, et al. Protein expression from exogenous mRNA: Uptake by receptor-mediated endocytosis and trafficking via the lysosomal pathway. RNA Biol. 2011; 8(4):627–36. 18. Diken M, Kreiter S, Selmi A, Britten CM, Huber C, Türeci Ö, et al. Selective uptake of naked vaccine RNA by dendritic cells is driven by macropinocytosis and abrogated upon DC maturation. Gene Ther. 2011; 18(7):702–08. 19. Hajam IA, Senevirathne A, Hewawaduge C, Kim J, Lee JH. Intranasally administered protein coated chitosan nanoparticles encapsulating influenza H9N2 HA2 and M2e mRNA molecules elicit protective immunity against avian influenza viruses in chickens. Vet Res. 2020; 51(1):37. 20. Hewitt EW. The MHC class I antigen presentation pathway: strategies for viral immune evasion. Immunology. 2003; 110(2):163–9. 21. Karikó K, Muramatsu H, Ludwig J, Weissman D. Generating the optimal mRNA for therapy: HPLC purification eliminates immune activation and improves translation of nucleoside-modified, protein-encoding mRNA. Nucleic Acids Res. 2011; 39(21): 142.

24. Bilenoglu O, Basak AN, Russell JE. A 3’UTR mutation affects beta-globin expression without altering the stability of its fully processed mRNA. Br J Haematol. 2002; 119(4):1106–14. 25. McNamara MA, Nair SK, Holl EK. RNA-Based Vaccines in Cancer Immunotherapy. J Immunol Res. 2015; 2015:1–9. 26. Anderson BR, Muramatsu H, Nallagatla SR, Bevilacqua PC, Sansing LH, Weissman D, et al. Incorporation of pseudouridine into mRNA enhances translation by diminishing PKR activation. Nucleic Acids Res. 2010; 38(17):5884–92. 27. Andries O, Mc Cafferty S, De Smedt SC, Weiss R, Sanders NN, Kitada T. N1-methylpseudouridine-incorporated mRNA outperforms pseudouridine-incorporated mRNA by providing enhanced protein expression and reduced immunogenicity in mammalian cell lines and mice. J Controlled Release. 2015; 217:337–44. 28. Feldman RA, Fuhr R, Smolenov I, (Mick) Ribeiro A, Panther L, Watson M, et al. mRNA vaccines against H10N8 and H7N9 influenza viruses of pandemic potential are immunogenic and well tolerated in healthy adults in phase 1 randomized clinical trials. Vaccine. 2019; 37(25):3326–34. 29. Richner JM, Himansu S, Dowd KA, Butler SL, Salazar V, Fox JM, et al. Modified mRNA Vaccines Protect against Zika Virus Infection. Cell. 2017; 168(6):1114-1125.e10. 30. Leal L, Guardo AC, Morón-López S, Salgado M, Mothe B, Heirman C, et al. Phase I clinical trial of an intranodally administered mRNA-based therapeutic vaccine against HIV-1 infection. AIDS. 2018; 32(17):2533–45. 31. Zhuang X, Qi Y, Wang M, Yu N, Nan F, Zhang H, et al. mRNA Vaccines Encoding the HA Protein of Influenza A H1N1 Virus Delivered by Cationic Lipid Nanoparticles Induce Protective Immune Responses in Mice. Vaccines. 2020; 8(1):123. 32. Pardi N, Secreto AJ, Shan X, Debonera F, Glover J, Yi Y, et al. Administration of nucleoside-modified mRNA encoding broadly neutralizing antibody protects humanized mice from HIV-1 challenge. Nat Commun. 2017; 8(1):14630. 33. Pardi N, Hogan MJ, Pelc RS, Muramatsu H, Andersen H, DeMaso CR, et al. Zika virus protection by a single lowdose nucleoside-modified mRNA vaccination. Nature. 2017; 543(7644):248–51. 34. Zhao M, Li M, Zhang Z, Gong T, Sun X. Induction of HIV-1 gag specific immune responses by cationic micelles mediated delivery of gag mRNA. Drug Deliv. 2015; 29:1–12.


Systematic Review of Viral mRNA Vaccine

35. Chahal JS, Fang T, Woodham AW, Khan OF, Ling J, Anderson DG, et al. An RNA nanoparticle vaccine against Zika virus elicits antibody and CD8+ T cell responses in a mouse model. Sci Rep. 2017; 7(1):252. 36. Bogers WM, Oostermeijer H, Mooij P, Koopman G, Verschoor EJ, Davis D, et al. Potent Immune Responses in Rhesus Macaques Induced by Nonviral Delivery of a Self-amplifying RNA Vaccine Expressing HIV Type 1 Envelope With a Cationic Nanoemulsion. J Infect Dis. 2015; 211(6):947–55. 37. Gandhi RT, Kwon DS, Macklin EA, Shopis JR, McLean AP, McBrine N, et al. Immunization of HIV-1-Infected Persons With Autologous Dendritic Cells Transfected With mRNA Encoding HIV-1 Gag and Nef: Results of a Randomized, Placebo-Controlled Clinical Trial. JAIDS J Acquir Immune Defic Syndr. 2016; 71(3):246–53. 38. de Jong W, Aerts J, Allard S, Brander C, Buyze J, et al. iHIVARNA phase IIa, a randomized, placebocontrolled, double-blinded trial to evaluate the safety and immunogenicity of iHIVARNA-01 in chronically HIVinfected patients under stable combined antiretroviral therapy. Trials. 2019; 20(1):361. 39. VanBlargan LA, Himansu S, Foreman BM, Ebel GD, Pierson TC, Diamond MS. An mRNA Vaccine Protects Mice against Multiple Tick-Transmitted Flavivirus Infections. Cell Rep. 2018; 25(12):3382-3392.e3. 40. Alberer M, Gnad-Vogt U, Hong HS, Mehr KT, Backert L, Finak G, et al. Safety and immunogenicity of a mRNA rabies vaccine in healthy adults: an open-label, nonrandomised, prospective, first-in-human phase 1 clinical trial. The Lancet. 2017; 390(10101):1511–20. 41. Bethesda, MD: National Library of Medicine. 2000 Feb 29. Identifier NCT04405076, DoseConfirmation Study to Evaluate the Safety, Reactogenicity, and Immunogenicity of mRNA-1273 COVID-19 Vaccine in Adults Aged 18 Years and Older; May 2020.


Scholarly Research In Progress • Vol. 4, October 2020

Investigating Trends in Elder-Elder Caregiving in the United States: 1997 – 2014 SooYoung H. VanDeMark1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Background: Caregiving is a difficult field to study due to the subjective nature of the data, and some of the research findings have been contradictory. Little research exists on the specific situation of elder-elder caregiving (where the caregiver is a person 64 years old and older caring for a care recipient who is also 64 years old and older.) Given the aging American population, this paper explores these elder relationships to better understand caregiving from their perspective. Methods: ANOVA tests and a Pearson Correlation Matrix were performed in Prism 8 on publicly available data sets from the National Alliance for Caregiving. This survey data was from a random sample of caregivers collected via telephone and internet in 1997, 2004, 2009, and 2014. Results: The ANOVA tests found a significant statistical difference (p < 0.0001) between the mean age of elder-elder caregivers. The post-hoc Games-Hollow test found statistically significant differences (p < 0.0001) for 1997 and 2009 when each was compared to 2014. No statistical difference was found between the mean age of elder-elder care recipients. No correlation was found between elder-elder caregiver age and level of burden experienced due to caregiving. The average length of time of caregiving for elder-elder caregivers was approximately 5 years for each year of survey administration. Conclusion: The lack of prior investigation on the elderelder caregiving population may be concealing the needs of this specific population. Further research could help society prioritize education and inform action plans that assist elderelder caregivers, so that a) they may have a high quality of life near the end of life, and b) that their caregiving workload does not shift to stressed institutional health care systems.

Introduction The term “caregiving” may not have much significance unless you have actually been or known a caregiver (CG) or care recipient (CR). However, there are over 44 million American CGs, and over 17 million of those CGs are caregiving for someone over the age of 64 (1, 2). Caregiving generally refers to the informal and unpaid work performed by a CG to assist a CR. In 2013, the economic value of caregiving in the United States was estimated to be $470 billion (3). Caregiving may include but is not limited to emotional support, physical assistance with personal care and household tasks, and more recently, financial support, home medical care and management, decision-making, and the navigation of complex

medical systems such as hospitals and insurance companies (1, 2). The field of caregiving is a stark reminder for health professionals that they are not the main providers of patient care (3). It is the eldest in society who are likely to have the most significant physical and cognitive limitations; and thus, be in need of caregiving (2). As the U.S. population ages, it is likely that the informal and unpaid work of caregiving will also increase. In 2010, there were 40.2 million people over the age of 65 living in the United States (4). The youngest of the Baby Boom Generation, composed of people born between 1946 and 1964, will be in their mid-60s by 2030 — that is roughly 20% of the country’s population that will be over the age of 64 (4,5). By 2050, that number will more than double to 88.5 million (4). The eldest of the elders, those over 85 years old, will represent 2.3% of the population in 2030, and are projected to nearly double by 2050 (4). Although the baby boomers will also be facing mortality declines by and after 2050, their presence will be an unprecedented increase in the total number of elders within the U.S. population. It is not only the CRs who are aging but also the CGs. This study explores the specific relationship of elder CGs who are 64 years old and older that care for elder CRs who are also 64 years old and older. This is referred to herein as elder-elder (E-E) caregiving. Caregiving is a difficult field of study, largely because most of the research is opinion-based and completed through surveys. The collected data is self-reported and difficult to objectively verify. Similarly, unrelated surveys ask questions and record answers in slightly different ways, such as using a 1-to-5 scale versus a yes/no answer format. These discrepancies make comparing research problematic. Unsurprisingly, findings between studies sometimes seemed contradictory. For example, there is no clear consensus around the impact of CG burden amongst those caring for older CRs on mortality, depression, and mental stress (2,6,7). The current literature did underscore the clinical need and real-world application for this type of research given that people suffering with a chronic illness receive the majority of their care from lay caregivers at home (1). The primary aim of this study was to investigate if the E-E caregiving population has also been aging from 1997 to 2014, given that the general population was and is aging? What is the sex of these CGs and their relationship to the CRs? And lastly, is there a correlation between E-E CGs age and level of burden experienced due to caregiving?


Investigating Trends in Elder-Elder Caregiving

Methods This paper used the National Alliance for Caregiving’s open data sets (8,9). The following two data sets were explored: 1) Caregiving in the U.S. 2015 and 2) Caregiving in the U.S. 2009. The 2015 files contained surveys completed in 2014 and will here on out be referred to by that year. Advantageous for this project, the data set within Caregiving in the U.S. 2009 also included raw data from the years 1997 and 2004. Inclusion of these years were not indicated in the file title, and thus, not expected. The raw data was separated by year into separate files for statistical analysis. Both data sets included codebooks explaining the data’s nomenclature. The 2009 files included the final survey questions in English and Spanish, and a report of their methodology. All 4 years of data were gathered from a random sample of CGs via telephone and online surveys, and over 200 variables were collected through the survey questions. Variables analyzed in this report included age and sex for both CG and CR; the length of time the CG had been caregiving; the relationship of the CR to the CG as defined by the CG; the CG’s health status; and the level of burden experienced by the CG due to caregiving. Data exclusion criteria included any incomplete surveys. Additionally, the survey responses were further whittled to the population of interest to include only CGs aged 64 and over who care for CRs aged 64 and over. Survey respondents who withheld or did not know their own age or the age of their care recipient were also excluded. This resulted in an E-E caregiving sample of n = 147, 115, 282, and 445 for years 1997, 2004, 2009, and 2014. Data analysis IBM’s SPSS Statistics Version 26 was used to organize the data and Prism 8 to run the statistical tests and create the figures. Summary statistics for age, sex, length of time caregiving, relationship of CG to CR, and health status of CG were calculated for each data set by year. ANOVA tests was performed to compare the mean age of CG and CR respectively in 1997, 2004, 2009 and 2014.

The one-way Brown-Forsythe and Welch ANOVA tests compared the mean elder-elder CG ages from the four specified years (alpha=0.05). These versions of the test were chosen because the standard deviation for each group were unequal. The post-hoc test for significance was the GamesHollow multiple comparison test (alpha=0.05), which was chosen because the sample sizes were over 50. A one-tail Pearson Correlation matrix was performed to see if there was an association between caregiver age and the level of burden experienced during any one year. Results The median, minimum, and maximum ages of E-E CGs and CRs from 1997 to 2014 is shown in Figure 1. The main variation in CG age occurred from 2004 to 2014, increasing from 69 years old to 75 years old. However, the median age of CRs showed minimal variation during the 17-year time frame. In 2004, the maximum age for CGs (99 years) was older than the maximum age for CRs (97 years). The two 99-year-old CGs in 2004 were outliers, as no other year saw CG ages above 90 years old. The maximum age of CRs (107 years old) occurred in 2014, the most recent year of data collection, and is another outlier, as no other year saw CR ages above 99 years old. Both the Brown-Forsythe ANOVA (F(3,541.1) = 16.69, p <0.0001) and Welch ANOVA (F(3, 350.7) = 18.05, p <0.0001) tests found a significant statistical difference in the mean age of E-E CGs between the 4 specified years. The post-hoc test found statistically significant differences for the mean age of E-E CGs for each year when compared to 2014 (Table 1). The most significant difference in mean age was found between 1997 and 2014. The ANOVA tests found no statistical difference between the mean age of E-E CRs; thus, no follow-up testing was done on that group. The sex breakdown among those involved in E-E caregiving is shown in Table 2. The majority of CGs each year were female, with an uptick in male CGs in 2014. CR sex is also predominantly female with the most equal sex division being in 2004. The relationship between the E-E CG and the CR as defined by the CG is shown in Figure 2. The identifier shown

Figure 1. Summary information for E-E CGs (left) and E-E CRs (right) including median and min-max ages in 1997, 2004, 2009, and 2014. 86

Investigating Trends in Elder-Elder Caregiving

Table 2. Breakdown of E-E CG and CR Sexes in 1997, 2004, 2009, and 2014. Note: M = Male, F = Female, UK = Unknown, DK = Don’t know. Table 1. Results of the Games-Howell multiple comparison test comparing mean ages in 1997, 2004, 2009, and 2014. The asterisk indicates a statistically significant p-value.

(e.g. mother) refers to the CR. The relationship of spouse, mother, or friend/non-relative/neighbor dominated the top three spots, accounting for at least 65% of the relationships in any given year. Within the catchall percentage of “other” are CRs who are children and grandchildren of the CG. These CRs of the next generation accounted for 0.13% in 2017, 5.7% in 2004, 7.6% in 2009, and 6.0% in 2014. The average length of time for E-E caregiving is approximately 5 years with a standard deviation ranging approximately from 5 to 10 years in any given year (Figure 3). The Pearson Correlation matrix found no correlation found between E-E CG age and the level of burden experienced, with r = 0.04, 0.112, 0.014, and 0.022 for 1997, 2004, 2009, and 2014 respectively. Relatedly, when asked, “How has caregiving affected your health status?” most E-E CGs responded that caregiving had not affected their health (Figure 4).

Discussion The results thus far support the idea that the E-E caregiving population is aging, reflecting U.S population trends, with the median age of caregivers increasing from 69 to 75 years old; although, the median age of care recipients did not significantly change. While the traditional view of E-E caregiving might be of two 70 y/o spouses caring for each other or a 70-year-old child taking care of her 95-year-old mother, those are not the only relationships present in E-E caregiving. The percentage of CGs caring for a family member of the next generation (i.e. child, grandchild) rose to a high of 7.6% in 2009. Another interesting trend is the decrease in caregiving of nonrelative friends (29.9% to 18.4%) with the increase in spousal caregiving (22.5% to 38.9%) from 1997 to 2014. It is possible that this shift in relationship between E-E CGs and CRs may be partly due to a more accepting cultural shift, where nonheterosexual partners were more comfortable labeling their partner as spouse instead of non-relative friend.

Figure 2. Relationships of E-E CGs to their CRs for 1997, 2004, 2009, and 2014. The identifying relationship is of the CR. In 2014, six relationships are shown as Brother and Sister were tied for fifth place at 3%.


Investigating Trends in Elder-Elder Caregiving

Figure 3. Mean Length of Time in years (with standard deviation) that E-E CGs report caregiving in 1997, 2004, 2009, and 2014.

Prior research found that 70% of CGs who are caring for an older adult do so for 2 to 7 years (2). My data supports the notion that CG is not a short-term responsibility—the average length of time a CG is caregiving was 5 years, which was unchanged from 1997 to 2014 — even when the CGs themselves are elderly. The larger standard deviations in 1997 (10.4 years) and 2014 (10.0 years) compared to the other two years may be due to the survey’s limitations. When CGs were asked how long they had been caregiving, some CGs may have responded by stating, the CR’s whole life. In such cases, interviewers had been instructed to enter the CR’s age as the answer; thus, potentially resulting in skewed data. The primary limitation on this research stems from the method of data collection: surveys. Although the surveys were thorough (over 200 variables), allowed for some open-ended answers, and gave interviewers cues on how to “probe” a respondent, the fact remains that subjective survey data can often have less than optimal accuracy. The original study design did not include any additional follow-up or verification of the information. The Level of Burden score used in the correlation test was a consolidated score meaning the study makers asked multiple questions regarding hours of care and number of daily activities performed in the survey, assigned values to those answers, and then consolidated a respondent’s answers into a Level of Burden scale (9). More interesting results could come from performing correlation tests between the specific variables asked (regarding financial, emotional and physical stress) and an E-E CGs age.

Conclusion As CGs are as group get older, it is imperative that health care professionals ensure the needs of CGs are being sufficiently addressed. The lack of thorough investigation on the E-E caregiving population may be concealing the needs of this specific population. Follow-up studies should explore caregiving subset populations such as generation spanning E-E caregiving, particularly when the CG is the older participant, or non-relative social E-E caregiving, because such relationships likely have interpersonal dynamics and complexities that are not being addressed elsewhere. Similarly, additional research should investigate how best to shorten the average length of time that E-E CGs undergo


Figure 4. Histogram of E-E CG responses on impact of caregiving on self-reported Health Status for 1997, 2004, 2009, and 2014.

such responsibilities. Likewise, with no clear consensus in the field of caregiving on the positive and negative impacts of caregiving on CGs, particularly E-E CGs, new studies need to explore potential benefits and burdens. Such research could help society prioritize education and inform action plans to assist caregivers, ensuring that a) they have a high quality of life near the end of life, and b) that their workload does not shift to stressed institutional health care systems.

Acknowledgments I would like to thank Elizabeth Kuchinski, MPH, and Brian Piper, PhD, professors of Community Health Research at Geisinger Commonwealth School of Medicine for their support on this endeavor.

Disclosures There is no financial relationship between this paper’s author and any institution mentioned herein.

References 1.

Wittenberg E, Prosser LA. Health as a family affair. N Engl J of Med. 2016; 374(19):1804–6.

2. National Academies of Sciences, Engineering, and Medicine. Families caring for an aging America. 2016. DOI: 3. Boyle DA. The caregiving quandary. Clin J Oncol Nurs. 2017; 21(2):139. 4. Vincent GK, Velkoff VA. The next four decades: The older population in the United States: 2010 to 2050: US Department of Commerce. Economics and Statistics Administration, US Census Bureau. 2010; 25-1138 5. Colby SL, Ortman JM. The baby boom cohort in the United States: 2012 to 2060. Current Population Reports. U.S. Census Bureau. 2014; 25-1141. 6. Schulz R, Beach SR. Caregiving as a risk factor for mortality: The caregiver health effects study. JAMA. 1999; 282(23):2215-9.

Investigating Trends in Elder-Elder Caregiving


Anderson LA, Edwards VJ, Pearson WS, Talley RC, McGuire LC, Andresen EM. Adult caregivers in the United States: Characteristics and differences in well-being, by caregiver age and caregiving status. Prev Chronic Dis. 2013; 10:130090. DOI: pcd10.130090

8. National Alliance for Caregiving. Caregiving in the US 2015. NAC and the AARP Public Institute. Washington DC: Greenwald & Associates. 2015. https://www.caregiving. org/research/open-data/ 9. National Alliance for Caregiving. Caregiving in the US 2009. NAC and the AARP Public Institute. Washington DC: Greenwald & Associates. 2009. https://www.caregiving. org/research/open-data/


Scholarly Research In Progress • Vol. 4, October 2020

The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions Cecilia Leng1*‡, Leah E. Thomas1*‡, Khine S.Y. Win1*‡, Brian J. Piper1, Joseph B. Fraiman2, and Alex Hodkinson3

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Slidell Memorial Hospital, Slidell, LA 70458 3 Centre for Primary Care, Manchester, UK *Master of Biomedical Sciences Program ‡Authors contributed equally Correspondence: 1


Abstract Background: Medical interventions come with many hidden risks that are not fully understood in healthcare. Currently, clinical guidelines are systematically biased towards these interventions. The objective of this descriptive study was to systematically review previous studies which have analyzed the adequacy of harms reporting in RCTs using the CONSORTharms extension guidelines. Methods: Articles were searched using search methodology provided from the previous study. An article met the inclusion criteria if it was a systematic review with trials using all ten CONSORT harms-extension recommendations. To analyze the risk of bias(RoB), a risk of bias assessment tool, ROBIS, was utilized. Results: Articles (N = 3,000) were retrieved from Pubmed search, and 0.3% (N = 9) of the articles were included based on pre-set criteria. A total of 787 trials reported in nine articles met the inclusion criteria. Analysis showed the majority of the articles had a high concern for risk of bias (66.67%, or N=6). About half of the articles (44.44%, N=4)) included studies that did not report on the CONSORT 9 guidelines, and some articles (33.33%, N=3) only included published studies. Many did not mention any method of attempting to reduce bias in their research (55.56%, N=5). Conclusions: The findings have clinical implications for the healthcare system. Most research focuses disproportionately on the benefits compared to the adverse effects of drugs. The results aligned with those of previous studies and explained how clinical guidelines are biased towards pharmaceutical interventions.

Introduction The Consolidated Standards of Reporting Trials (CONSORT) is a guideline designed to improve the quality of reporting of results within randomized controlled trials (RTCs) (1). It has been used across multiple journals as a guideline to help improve quality of articles when used by authors and peer reviewers. However, the CONSORT standards developed in 2001 are focused on the reporting of the favorable effects of the medical intervention(s) with only one recommendation (recommendation 19) out of 22 recommendations that focused on the unintended effects of the intervention (2). In 2004, due to increased evidence of an imbalance between benefit-harms reporting, an update of the CONSORT statement was released, and an additional 10 harms-extension recommendations were included to address the harms section 90

of medical interventions (3). A revision of the CONSORT standards was released in 2010, where an additional harmextension recommendation was added, however it is uncertain how many authors have adopted the usage of this update (1). For this reason, we have chosen to review studies that utilized the 2004 CONSORT guidelines. In 2013, a systematic review attempted to determine the standard of reporting harms-related data in RCTs. The study used the CONSORT extensions to analyze the harms (4). This systematic review identified nine studies meeting their inclusion criteria, and found harm to be generally under-reported as per the CONSORT guidelines. Since the publication of this systematic review, multiple new studies have been published and an update appeared warranted to see if there was any improvement in reporting harm. In addition, a risk of bias tool called ROBIS was developed to assess the risk of bias. ROBIS was specifically developed to assess the risk of bias in systematic reviews, and so would be inapplicable in other research articles other than systematic reviews, such as primary research articles (5). The aim of this paper was to systematically review previous studies which have analyzed the adequacy of harms reporting in RCTs using the CONSORT-harms extension guidelines.

Methods Identification of studies A protocol search method for this systematic review was provided by Alex Hodkinson, PhD. The search strategy was adapted from the prior systematic review (4) and applied multiple databases including: the Cochrane Database of Systematic Reviews, Database of Abstracts of Reviews of Effects (DARE), Ovid MEDLINE, Scopus and ISI Web of Knowledge Conference Proceedings Citation Indexes (CPCI-S or CPCI-SSH), and the Zetoc database (4). Only reviews published after 2012 were of interest, as all studies previous to this time were included in the prior systematic review. Articles obtained by the search methodology were screened by their titles and abstracts. In order to determine eligibility, two reviewers analyzed full articles for potentially suitable studies. Study inclusion criteria We included published research that assessed the quality of the harms reporting in RCTs against the CONSORT harms extension. Articles that focused on medications were included. Excluded studies included those that assessed harms reporting using assessments other than CONSORT guidelines.

The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

There were no language limitations. After narrowing down the articles, nine articles were qualified. The detailed identification and characteristics of these articles are described in Table 1 and Table 2, respectively. Quality assessment Two reviewers independently assessed each study using the ROBIS as a standard to determine the methodological quality. Following the ROBIS (5) tool guidelines, questions were split into 3 different phases. Phase one was optional and was chosen to be omitted due to criteria being inapplicable to the review. Phase two was additionally split into 4 domains: study eligibility criteria, identification and selection of studies, data collection and study appraisal, and synthesis and findings. Criteria were graded as yes (Y), probably yes (PY), probably no (PN), no (N), no information (NI). Phase three included the determination of the high or low risk of each article based on the answers from the previous phase.

Table 1 – Included review articles PubMed ID Title and Publication Year (PMID) 30653255 “A systematic review of adverse events in randomized trials assessing immune checkpoint inhibitors.” (2019) 23127360 “Adverse event reporting in randomised controlled trials of neuropathic pain: considerations for future practice.” (2012) 28866634 “Quality of reporting of harms in randomised controlled trials of pharmacological interventions for rheumatoid arthritis: a systematic review.” (2017) 26991481 “Adverse event assessment and reporting in trials of newer treatments for post-operative pain.” (2016) 29737504 “Inadequate harms reporting in randomized control trials of antibiotics for pediatric acute otitis media: a systematic review.” (2018) 26482955 “Adverse event methods were heterogeneous and insufficiently reported in randomized trials on persistent depressive disorder.” (2015) 24515989 “Reporting of harms by randomised controlled trials in ophthalmology.” (2014) 22985899


Author(s) Arnaud-Coffin P, Maillet D, Gan HK, Stelmes JJ, You B, Dalle S, et al. Cornelius VR, Sauzet O, Williams JE, Ayis S, FarquharSmith P, Ross JR, et al. Hadi MA, McHugh GA, Conaghan PG. Hoffer D, Smith SM, Parlow J, Allard R, Gilron I. Hum SW, Golder S, Shaikh N.

Meister R, von Wolff A, Mohr H, Nestoriuc Y, Härter M, Hölzel L, et al. O’Day R, Walton R, Blennerhassett R, Gillies MC, Barthelmes D. “Adherence to CONSORT harms-reporting Smith SM, Chang RD, Pereira recommendations in publication of recent analgesic A, Shah N, Gilron I, Katz NP, et clinical trials: an ACTTION systematic review.” al. (2012) “Adverse Event Reporting in Clinical Trials of Williams MR, McKeown A, Intravenous and Invasive Pain Treatments: An Pressman Z, Hunsinger M, Lee ACTTION Systematic Review.” (2016) K, Coplan P, et al.

Table 1. Included review articles

Table 2. Characteristics of included reviews


The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

Table 3. Quality of harms reporting in Randomized Controlled Trials (RCTs) using 2004 extended CONSORT-harms recommendations

Data collection and extraction


Reviewers were trained in the screening and data extraction of articles. Two reviewers independently analyzed each article and any discrepancies were resolved through a discussion with a third reviewer. The following data was collected for each article: (1) Study characteristics (inclusion criteria including drug classes and medications, databases searched within the study, and restrictions on the published date of the article), (2) Sample size (defined by the number of RCT articles analyzed for reporting quality), (3) Reporting quality (the number of the 10 recommendations from the 2004 CONSORT-harms extension checklist included in each article) (Table 3). Data was recorded using a spreadsheet on Microsoft Excel.

Using the search strategy stated above, 3,000 potential studies were identified. Through the screening process, 25 studies were identified and further assessed for eligibility in this study. Nine studies were included that collectively evaluated harm reporting of 787 RCTs (Figure 1). The included reviews were assessed for bias using the ROBIS tool (Table 4); six out of the nine studies were found to be at high risk of bias, while the remaining three studies were found to have a low risk of bias (Figure 2, Table 5). One study resulted in disagreement between the two reviewers, and an agreement was reached with the addition of the third reviewer (Table 6). Several studies were marked as high risk for concern during phase 3 of the ROBIS assessment. The reasons for concern ranged from inadequate range of databases searched to lack of formal assessment of bias and calculation of the level of concern was determined by answers from phase 2 (Table 7). Wide variability was noted in the adherence to various CONSORT guideline items. Four of the articles found that most

Figure 1. Flow chart of selection process, adjusted from Hodkinson et al. 2013 92

Figure 2. Number of studies vs. level of concern for risk of bias

The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

Table 4. ROBIS: Tool to assess risk of bias in systematic reviews


The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

Table 5. Risk of bias in the review  


The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

Table 6. ROBIS Phase 2: Identifying concerns with the review process

of their RCTs evaluated did not report on the CONSORT 9 guideline, and five of the articles made no mention of trying to reduce bias in their study (Table 1). Three of the articles only included published articles in their study (Table 2). Adherence to overall CONSORT recommendations specific to harm ranged from 0.0% to 100% across reviews, with item 8 being the most consistently reported, with an average compliance of 77.3%, and item 9 being the least consistently reported, with an average compliance of 3.6%. (Table 8)

Discussion This study was a continuation of a previous systematic review assessing the quality of reporting data based on the CONSORT-harms guidelines. It has been 16 years since the release of the 2004 CONSORT guidelines, allowing adequate time for authors to adapt to the guidelines. The data obtained in this report indicates a pronounced inconsistency of harms reporting which can be seen in the percentage of RCTs adhering to each individual CONSORT recommendation. Heterogeneity among the different studies is found between each CONSORT guideline. These results supported the hypothesis that harms reporting remains inadequate following the CONSORT harms extension of 2004. Most of the articles showed a lack of reporting all 10 of the CONSORT guidelines, especially in regards to the 9th guideline. Many of the articles showed a lack of attention to reducing the risk of bias in their own studies, such as by not including a formal assessment of bias, as well as utilizing only one reviewer instead of multiple, leading to potentially skewed results. The possible presence of bias, may lead to a potential for human error to occur, which may unconsciously select for expected data results (Table 1). Our study aligns with results from previously published literature showing the lack of quality in the reporting of harms due to a preference of reporting benefits (4). These results were expected and followed the trend that has been previously observed, with not much noticeable improvement in the average reporting compliance over the years.

This study was strengthened by the use of the ROBIS tool to the quality of the included reviews. Future research could be done to include more articles through the usage of other databases to analyze their quality. Further studies that demonstrate the benefits of the use of a tool such as CONSORT may provide strength to the recommendation of making the use of CONSORT a requirement for those carrying out RCTs. A universally accepted guideline would provide a standardized norm in order to better correlate data and reconcile data reporting between different databases.

Conclusion This analysis was an update of the previous study to perform a systematic review of empirical studies in order to evaluate harms-reporting quality in adherence to CONSORT-harms guidelines. Nine studies were included in this review which collectively reported on 787 RCTs. This provided an updated report on the continued inadequacy in harms-reporting. An updated measure in this study was the use of the ROBIS tool to assess for the risk of bias. Proper use of the CONSORT checklist items has been associated with an improvement in the quality of harm-reporting in RCTs (15). Comparison of harmreporting in RCTs prior to the implementation of the CONSORT statement to RCTs published in the years following the release of the guidelines suggested improvement in quality of reports. This study focuses on reporting of harms, which is primarily and robustly determined through evaluation according to CONSORT guidelines. This review shows a continued inconsistency in reporting, even following the CONSORT guideline update in 2010. Assessing the quality of the studies included in this review through the use of the ROBIS tool reduced the risk of bias in assessment.

Acknowledgements The authors would like to thank the library staff at Geisinger Commonwealth School of Medicine for assisting with the article search. 95

The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

Table 7. ROBIS Phase 3: Judging risk of bias


The Under-Reporting of Harm in Randomized Controlled Trials of Medical Interventions

Table 8. CONSORT harms criteria reported across included reviews

References 1.

Schulz KF, Altman DG, Moher D. CONSORT 2010 statement: Updated guidelines for reporting parallel group randomised trials. J Pharmacol Pharmacother. 2010;1(2):100–7.

2. Moher D, Schulz KF, Altman DG, CONSORT GROUP (Consolidated Standards of Reporting Trials). The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. Ann Intern Med. 2001;134(8):657–62.

9. Hum SW, Golder S, Shaikh N. Inadequate harms reporting in randomized control trials of antibiotics for pediatric acute otitis media: a systematic review. Drug Saf. 2018;41(10):933–8. 10. O’Day R, Walton R, Blennerhassett R, Gillies MC, Barthelmes D. Reporting of harms by randomised controlled trials in ophthalmology. Br J Ophthalmol. 2014;98(8):1003–8.

3. Campbell MK, Elbourne DR, Altman DG, CONSORT group. CONSORT statement: extension to cluster randomised trials. BMJ. 2004;328(7441):702–8.

11. Williams MR, McKeown A, Pressman Z, Hunsinger M, Lee K, Coplan P, et al. Adverse Event Reporting in Clinical Trials of Intravenous and Invasive Pain Treatments: An ACTTION Systematic Review. J Pain. 2016;17(11):1137–49.

4. Hodkinson A, Kirkham JJ, Tudur-Smith C, Gamble C. Reporting of harms data in RCTs: a systematic review of empirical assessments against the CONSORT harms extension. BMJ Open. 2013;3(9):e003436.

12. Arnaud-Coffin P, Maillet D, Gan HK, Stelmes J-J, You B, Dalle S, et al. A systematic review of adverse events in randomized trials assessing immune checkpoint inhibitors. Int J Cancer. 2019;145(3):639–48.

5. Whiting P, Savović J, Higgins JPT, Caldwell DM, Reeves BC, Shea B, et al. [ROBIS: a new tool to assess risk of bias in systematic reviews was developed. Recenti Prog Med. 2018;109(9):421–31.

13. Meister R, von Wolff A, Mohr H, Nestoriuc Y, Härter M, Hölzel L, et al. Adverse event methods were heterogeneous and insufficiently reported in randomized trials on persistent depressive disorder. J Clin Epidemiol. 2016;71:97–108.

6. Cornelius VR, Sauzet O, Williams JE, Ayis S, FarquharSmith P, Ross JR, et al. Adverse event reporting in randomised controlled trials of neuropathic pain: considerations for future practice. Pain. 2013;154(2):213– 20. 7.

Hadi MA, McHugh GA, Conaghan PG. Quality of reporting of harms in randomised controlled trials of pharmacological interventions for rheumatoid arthritis: a systematic review. Evid Based Med. 2017;22(5):170–7.

8. Hoffer D, Smith SM, Parlow J, Allard R, Gilron I. Adverse event assessment and reporting in trials of newer treatments for post-operative pain. Acta Anaesthesiol Scand. 2016;60(7):842–51.

14. Smith SM, Chang RD, Pereira A, Shah N, Gilron I, Katz NP, et al. Adherence to CONSORT harms-reporting recommendations in publications of recent analgesic clinical trials: an ACTTION systematic review. Pain. 2012;153(12):2415–21. 15. Ghimire S, Kyung E, Kang W, Kim E. Assessment of adherence to the CONSORT statement for quality of reports on randomized controlled trial abstracts from four high-impact general medical journals. Trials. 2012 Dec 1;13(1):77.


Scholarly Research In Progress • Vol. 4, October 2020

Oxalate Nephropathy after Bariatric Surgery: Incidence and Case Report of Successful Kidney Transplant after Conversion from Roux-en-Y Gastric Bypass to Gastric Sleeve Amandeep Kaur1†, G. Craig Wood2, Dan Bucaloiu3, Maria C. Bermudez3, Jason George3, and Alex R. Chang3

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Obesity Institute, Geisinger, Danville, PA 17822 3 Department of Nephrology, Geisinger Center of Health Research, Danville, PA 17822 †Doctor of Medicine Program Correspondence: 1


Abstract Oxalate nephropathy is characterized by tubular crystalline deposits of calcium oxalate leading to acute and chronic tubular injury, interstitial fibrosis, and progressive renal insufficiency. Conditions characterized by fat malabsorption can result in binding of fatty acids to calcium, leading to hyperabsorption of oxalate and sometimes kidney injury since calcium normally binds to oxalate. Oxalate nephropathy has been reported after bariatric surgery although the incidence is unclear. Reported cases have been in the setting of malabsorptive bariatric surgeries such as jejunoileal bypass, and to a lesser extent, Roux-en-Y gastric bypass (RYGB). No cases in the literature have been reported after gastric sleeve. Out of 4,554 patients in our bariatric surgery research registry, we identified 2 cases of biopsy-confirmed oxalate nephropathy in patients who underwent RYGB surgery. We report a unique case of a patient who had oxalate nephropathy associated with RYGB, then underwent reversal of RYGB and conversion to gastric sleeve, and eventually received a kidney transplant.

Introduction Oxalate is naturally found in certain foods and is also a metabolic end-product that is excreted in the urine after absorption in the gastrointestinal tract. In most individuals, the impact of dietary oxalate is on urinary oxalate excretion is small (1). However, in the presence of fat malabsorption, excessive oxalate absorption may increase the risk for oxalate nephropathy (2). Oxalate nephropathy is characterized by tubular crystalline deposits of calcium oxalate leading to acute and chronic tubular injury, interstitial fibrosis, and progressive renal insufficiency (3). In the normal state, calcium and oxalate within the lumen of the intestine combine to form insoluble calcium oxalate complexes that are excreted in feces. In the setting of fat malabsorption and enteric hyperoxaluria, excessive intraluminal free fatty acids bind to and saponify calcium within the intestine, thereby inhibiting the formation of calcium oxalate. As a result, greater quantities of soluble free oxalate are absorbed by the colonic mucosa, which can lead to kidney injury. Initial cases were reported after jejunoileal bypass surgery, a purely malabsorptive surgery that is no longer performed due to the high incidence of oxalosis (4). More recently, cases of oxalate nephropathy have been reported after malabsorptive bariatric surgery Roux-en-Y gastric bypass (RYGB), although the incidence is unclear (5). In contrast, no 98

case reports of oxalate nephropathy after restrictive bariatric surgery procedures such as gastric sleeve or gastric banding have been reported. Proximal RYGB is often referred to as a combination restriction–malabsorption procedure. It involves stapling of the stomach to create a small (≤30.0 ml) upper gastric pouch. The small intestine is then divided at the mid jejunum, and the distal portion (Roux limb) is anastomosed to the gastric pouch. Incidence of oxalate nephropathy We examined data from 4,554 patients in Geisinger’s bariatric surgery research registry to determine the incidence of biopsy-confirmed oxalate nephropathy after bariatric surgery. We employed the following strategy. First, we identified 123 patients who had either “oxalate nephropathy” in EHR progress notes using a text mining tool, a diagnosis of ESKD, or eGFR <15 ml/min/1.73m2. Second, we searched our kidney biopsy pathology databases for “oxalate nephropathy” and did not identify any additional patients in the bariatric surgery research registry. Lastly, chart review on the 123 patients meeting the above criteria were reviewed by an attending nephrologist to determine if patients had evidence of a kidney biopsy or biopsy-confirmed oxalate nephropathy. Out of a total of 3 patients in the bariatric surgery cohort who had kidney biopsy data, 2 patients had biopsy-confirmed oxalate nephropathy.

Case Presentation Case presentation of a patient who had oxalate nephropathy associated with RYGB, then underwent reversal of RYGB and conversion to gastric sleeve, and eventually received a kidney transplant. A 58-year-old female with hypertension, Type 2 diabetes, chronic kidney disease stage 3, underwent RYGB (150 cm or less) for morbid obesity in 2012. She had a long-standing history of diabetes (2003) requiring insulin (2008), diabetic retinopathy, diabetic neuropathy, and proteinuria. At the time of surgery, she had stage 4 chronic kidney disease. About 11 months after RYGB, she was hospitalized for acute kidney injury, which was initially thought to be due to dehydration. She received intravenous fluid resuscitation. However, her kidney function continued to worsen. Other workup was negative, including complement C4, anti-nuclear antibody, anti-neutrophil cytoplasmic antibody, cryoglobulin panel, rheumatic fever, hepatitis profile, and renal panel. A kidney biopsy was performed, demonstrating oxalate

Oxalate Nephropathy after Bariatric Surgery

Figure 1. Left kidney, renal biopsy confirmed renal oxalosis with significant tubulointerstitial disease and interstitial fibrosis. Microscopic examination of the tubules demonstrates large crystalline deposits in tubular lumina with reactive tubular epithelial changes.

Figure 3. Allograft kidney, needle biopsy, light microscopy. There was no evidence of oxalate nephropathy on transplant kidney biopsy.

Figure 2. Kidney biopsy, trichrome stain. Examination of the interstitium with trichrome stain shows moderate interstitial fibrosis of 26 to 50% of cortical area, mild edema and interstitial inflammation of 26% to 50% of the cortical parenchyma consisting predominantly of mononuclear inflammatory cells and a remarkable number (>10% of inflammation) of eosinophils and neutrophils.

Figure 4. Allograft kidney, needle biopsy, light microscopy. There was no evidence of oxalate nephropathy on transplant kidney biopsy.

nephropathy (Figure 1). In an attempt to reverse the acute kidney injury, maintain weight loss to preserve candidacy for the kidney transplant list, the patient had laparoscopic reversal of RYGB and conversion to gastric sleeve approximately 4 months later. Despite the operation, her kidney function did not recover, and she went on to develop end-stage kidney disease and started continuous cyclic peritoneal dialysis. She eventually received a deceased donor kidney transplant. No evidence of oxalate nephropathy was seen on transplant kidney biopsies that were done for delayed graft function (Figure 2 and Figure 3).

Discussion Medical management of oxalate nephropathy after RYGB focuses on lowering intake of dietary oxalate and fat and measures to bind oxalate such as calcium supplementation (3). However, there is little evidence demonstrating whether interventions to reduce enteric hyperoxaluria after bariatric surgery can stabilize kidney function. Unlike a prior case report, (6) her chronic kidney disease continued to progress despite reversal of the RYGB and conversion to gastric sleeve (Figure 4). In our patient’s case, the continued progression may have been a result of her underlying diabetic kidney disease and significant renal fibrosis seen on kidney biopsy (Figure 5). However, she was able to maintain her weight so that she was able to qualify for eligibility on the transplant waitlist and eventually receive a kidney transplant. Interestingly, on kidney biopsies of her transplanted kidney, no oxalate


Oxalate Nephropathy after Bariatric Surgery

Figure 5. Changes in serum creatinine (mg/dL) in the patient

crystals were observed. There are some limitations to note. First, kidney biopsy data was only available on 3 patients after bariatric surgery, and it is possible that oxalate nephropathy is underrecognized. In addition, we had incomplete data available on 24-hour urine oxalate levels post-RYGB surgery when kidney function was worsening.

Conclusion In conclusion, we identified 2 cases of biopsy-confirmed oxalate nephropathy in a cohort of 4,554 bariatric surgery patients. However, additional studies are needed to determine the exact incidence of oxalate nephropathy since kidney biopsies and 24-hour urine oxalate are not completed routinely after bariatric surgery. Prompt measurement of 24-hour urine oxalate levels and kidney biopsy should be considered in the event of unexplained acute kidney injury or chronic kidney disease progression in patients post-RYGB surgery. Conversion to gastric sleeve surgery to maintain weight loss after oxalate nephropathy may be helpful for patients with oxalate nephropathy to preserve kidney transplant candidacy.

Acknowledgments The author acknowledges Dr. Alex Chang for assistance with case report conceptualization and for supervision on the case, and the patient for her willingness to participate in the data collection for this case report.

References 1.


Taylor EN and Curhan GC. Determinants of 24-hour urinary oxalate excretion. Clinical Journal of the American Society of Nephrology. 2008;3(5):1453-1460. doi: 10.2215/ CJN.01410308

2. Lumlertgul N, Siribamrungwong M, Jaber BL, Susantitaphong P. Secondary Oxalate Nephropathy: A Systemic Review. Kidney Int Rep. 2018;3(6):1363-1372. doi: 10.1016/j.ekir.2018.07.020 3. Nasr SH, Agati VD, Said SM, Stokes MB, Largoza MV, Radhakrishnan J, et al. Oxalate Nephropathy Complicating Roux-en-Y Gastric Bypass: An underrecognized cause of irreversible renal failure. Clinical Journal of American Society of Nephrology. 2008;3(6):1676-1683. doi: 10.2215/ CJN.02940608 4. Stauffer QJ. Hyperoxaluria and calcium oxalate nephrolithiasis after jejunoileal bypass. American Journal of Clinical Nutrition.1977;30:64-71. ajcn/30.1.64 5. Lieske CJ, Mehta RA, Milliner DS, Rule AD, Bergstralh EJ, Sarr MG. Kidney stones are common after bariatric surgery. International Society of Nephrology. 2014. doi: 10.1038/ki.2014.352 6. Agarwal V, Wilfong JB, Rich CE, Gibson PC. Reversal of Gastric Bypass Resolves Hyperoxaluria and Improves Oxalate Nephropathy Secondary to Roux-en-Y Gastric Bypass. Case Rep Nephrol Dial. 2016;6:114-119. doi: 10.1159/000449128

Scholarly Research In Progress • Vol. 4, October 2020

Radiotherapy Treatment Plan Evaluation Software Implementation and Breast Cancer Radiotherapy Plan Quality Kelley Chan1†, Daniel P. Talenti2, and Thomas J. Gergel2

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Guthrie Clinic, Department of Radiation Oncology, Sayre, PA 18840 †Doctor of Medicine Program Correspondence: 1


Abstract Background: Evaluation of radiotherapy plans has historically been a tedious and iterative manual process requiring assimilation of many data points from dose volume histograms. Radiotherapy treatment plan evaluation software allows for specification of dosimetric goals for target coverage and critical structure avoidance, with a visually appealing presentation of compliance with those goals for a particular radiotherapy plan. One version of this software (ClearCheck by RADformation) was introduced at the Guthrie Clinic in April 2018 primarily for standardization of dosimetric parameters across multiple radiotherapy clinic sites and for ease of evaluating treatment plans. This study evaluated the effect of radiotherapy treatment plan evaluation software implementation on radiotherapy treatment plan quality for breast cancer.

(≤2 cc), body V107% (≤10 cc), and ipsilateral lung V1600cGy (≤10%) of 21.8%, 23.8%, 20.9%, and 26.7% respectively (p≤0.05). Conclusions: Treatment plan evaluation software implementation incidentally improved whole breast radiotherapy plan quality. Most notable improvements were evident in PTV coverage as well as heart and lung avoidance. These improvements may be clinically relevant for certain patients.


Methods: Consecutive patients with pathological early-stage breast cancer (stage 0 or 1) status post breast conserving surgery treated with radiotherapy from October 2017 to February 2019 at a single geographic site of an integrated health care system were evaluated. The Radiation Therapy Oncology Group (RTOG) and NRG Oncology guidelines for defining target volumes and critical structures provided standardization. All patients received 4256 cGy in 16 fractions to the entire breast only via tangent fields with field-in-field technique. Critical structures defined included body, contralateral breast, heart, left anterior descending (LAD) vessels, and ipsilateral and contralateral lungs. Dosimetric goals were unchanged before and after software implementation. Certain dosimetric parameters (such as PTV_WB_EVA V95%, body V107%, and heart Dmean) were specifically tracked by the software at the time of treatment plan generation and approval. Other dosimetric parameters (such as PTV_WB_EVA D99%, heart D5%, heart D0.03cc, LAD vessels Dmean, and ipsilateral lung V1600cGy and V800cGy) were retrospectively evaluated.

Following the introduction of breast-conserving therapy, local excision with or without adjuvant radiotherapy has been used to treat early-stage invasive breast cancer. Breast-conserving surgery removes detectable macroscopic disease, while radiotherapy provides treatment for microscopic tumor foci in the conserved breast, which if left untreated can lead to locoregional recurrence or life-threatening metastasis, or both (1). Randomized clinical trials have shown benefits of adjuvant radiotherapy in terms of both locoregional control and breast cancer mortality (1, 2). Analysis of randomized clinical trials with follow-up data since 1985 was performed by the Early Breast Cancer Trialist’s Collaborative Group (EBCTCG). In their analysis of over 3,700 women, adjuvant radiotherapy for DCIS reduced the 10-year risk of any ipsilateral breast event from 28.1% to 12.9% (absolute risk reduction 15.2%) (2). The benefit of radiotherapy in reducing ipsilateral breast events was seen in patients regardless of various absolute risks including their age at diagnosis, extent of breast-conserving surgery, margin status, focality and grade. In a meta-analysis of over 10,800 women with invasive breast cancer by the EBCTCG, radiotherapy reduced the absolute 10-year risk of any first recurrence, whether locoregional or distant, by 15.7% (1). Additionally, adjuvant radiotherapy reduced the 15-year absolute risk of breast cancer mortality by 3.8%, suggesting adjuvant radiotherapy on average avoids one breast cancer death by year 15 for every 4 recurrences avoided by year 10 (1).

Results: Fifty-two consecutive breast cancer patients were evaluated. The first 25 patients were prior to treatment plan evaluation software implementation and the second 27 patients were after implementation. Thirty-three patients had left breast cancer. PTV_WB_EVA volume ranged from 237.4 to 1760.2 cc. Relevant dosimetric parameter changes after software implementation include statistically significant improvements in PTV_WB_EVA V95% (%) and body V107% (cc) of 1.38% and 8.03 cc respectively (p≤0.05). Differences in the likelihood of achieving dosimetric parameter ideal values before and after treatment plan evaluation software implementation include statistically significant improvements in compliance rates for PTV_WB_EVA V95% (≥93%), body V107%

Cardiovascular-related morbidity and mortality has been one of the most significant concerns associated with adjuvant radiotherapy in left breast cancer. Early radiotherapy trials noted an appreciably increased mortality from heart disease upwards of two decades after treatment (4). Since then, the doses of radiation the heart is exposed to are generally lower at approximately 1 to 5 Gy. In a population-based casecontrol study of major coronary events, defined as myocardial infarction, coronary vascularization, or death from ischemic heart disease, the overall average mean of doses to the whole heart was 4.9 Gy (5). Exposure of the heart to radiotherapy increased the rate of major coronary events linearly by 7.4% per Gy, with no apparent threshold (5). In radiation induced 101

Radiotherapy Treatment Plan Evaluation

heart disease, angiopathy of coronary vessels leads to coronary artery disease, and eventually ischemic heart disease. Specifically, in irradiated left breast cancer, there is a four- to seven-fold increase in stenosis of the mid and distal left anterior descending (LAD) and distal diagonal arteries, when compared to irradiated right breast cancer (3). These are the vessels that can be incidentally radiated with typical radiotherapy fields. The significant increase in coronary artery stenosis in these hotspot areas suggests that care should be taken to avoid radiation dose to these critical structures. An effective radiotherapy treatment plan for breast cancer should specify dosimetric goals for planning target coverage as well as critical structure avoidance. Evaluation of radiotherapy plans has historically been a tedious and iterative manual process requiring the assimilation of many data points from dose volume histograms. The implementation of radiotherapy treatment plan evaluation software has allowed for the specification of dosimetric goals for target coverage and critical structure avoidance, with a visually appealing presentation of compliance with these goals for a given radiotherapy plan. One version of this software (ClearCheck by RADformation Inc.), which was introduced at the Guthrie Clinic in April 2018, provides a means for standardization of dosimetric parameters across multiple radiotherapy clinic sites and ease of evaluating treatment plans. In this study, we evaluated the effect of software implementation on radiotherapy treatment plan quality for breast cancer.

Dosimetric parameter evaluation Dose-volume histograms (DVHs), which plot radiotherapy dose versus tissue volume, were used to summarize a threedimensional radiotherapy dose distribution in two dimensions (Figure 2). Specific parameters were extracted from DVHs to see how well targets were being covered and critical normal tissues were being avoided by a particular radiotherapy plan. Treatment plan evaluation software (such as ClearCheck) automatically extracts the specified dosimetric parameters from the DVHs and presents them in a visually appealing format for review by the physician. As the radiotherapy plan is adjusted with iterative planning, the treatment plan evaluation software automatically updates the specified dosimetric parameters for review. The target volume for radiotherapy was a structure called PTV_WB_EVA, or the evaluation planning target volume (PTV) for the whole breast. PTV_WB_EVA coverage by the prescribed radiotherapy dose was assessed by V95% (volume of the target receiving 95% or more of the prescribed radiotherapy dose as an index of uniform coverage), V105% (volume of the target receiving 105% or more of the prescribed radiotherapy dose as an index of “hot spots�), V107% (volume

Methods Inclusion criteria Consecutive breast cancer patients treated with radiotherapy from October 2017 to February 2019 were identified. All patients had pathological early stage breast cancer (stage 0 or I) status post breast-conserving surgery. All patients sought care at a single geographic site of an integrated health care system. The RTOG and NRG Oncology guidelines for defining radiotherapy target volumes and critical structures provided standardization. All patients received 4256 cGy in 16 fractions to the entire breast only via tangent fields with field-in-field planning technique (Figure 1).

Figure 1. Example left breast radiotherapy isodose distribution (with heart and left anterior descending vessels contoured)

Figure 2. Example left breast radiotherapy plan dose-volume histogram (DVH) 102

Radiotherapy Treatment Plan Evaluation

of the target receiving 107% or more of the prescribed radiotherapy dose as an index of “hot spots”), and D99% (minimum dose to the hottest 99% of the target as an index of “cold spots”). Critical normal structures defined on the radiotherapy planning CT scan included the body, contralateral breast, heart, left anterior descending (LAD) vessels, and ipsilateral and contralateral lungs. Critical normal structure avoidance was assessed by ipsilateral lung Dmean (mean dose to the ipsilateral lung), ipsilateral lung V1600cGy (percentage of ipsilateral lung volume receiving 1600 cGy or more), heart

Dmean (mean dose to the heart), heart D5% (minimum dose to the hottest 5% of the heart), and heart D0.03cc (minimum dose to the hottest 0.03 cc of the heart), amongst other parameters. Certain dosimetric parameters had ideal and acceptable levels associated with them. For example, with PTV_WB_EVA, V95% was classified as acceptable if at least 93% of the PTV_WB_EVA received 95% of the prescribed dose and as ideal if at least 95% of the PTV_WB_EVA received 95% of the prescribed dose. For these parameters, in addition to assessing mean values, the percentage of patients achieving acceptable or ideal classifications was also tracked. The

Figure 3. Protocol left breast radiotherapy plan ClearCheck 103

Radiotherapy Treatment Plan Evaluation

ClearCheck protocol specifically utilized for extracting data for this study is shown in Figure 3 (with ranges for ideal and acceptable values if appropriate). Dosimetric goals were unchanged before and after treatment plan evaluation software implementation. Certain dosimetric parameters (such as PTV_WB_EVA V95%, body V107%, and heart Dmean) were specifically tracked by the software at the time of treatment plan generation and approval. Other dosimetric parameters (such as PTV_WB_EVA D99%, heart D5%, and ipsilateral lung V1600cGy) were retrospectively evaluated.

had left breast cancer and 19 patients had right breast cancer. PTV_WB_EVA volume ranged from 237.4 to 1760.2 cc, with an average of 1039.8 cc. Other relevant patient parameters are summarized in Table 1. Differences in dosimetric parameter mean values before and after treatment planning evaluation software implementation are shown in Table 2. After software implementation, there were statistically significant improvements in PTV_WB_

Endpoints and statistical analysis The differences in the means of the various dosimetric parameters before and after evaluation software implementation were compared via 2-tailed Wilcoxon rank-sum test. The differences in the rates of compliance with the pre-specified goals (ideal or acceptable) of the various dosimetric parameters before and after evaluation software implementation were compared via Chi-squared test.

Results Fifty-two consecutive breast cancer patients were evaluated. The first 25 patients were prior to treatment plan evaluation software implementation and the second 27 patients were after implementation. Thirty-three patients

Table 1. Relevant patient parameters

Table 2. Differences in dosimetric parameter mean values before and after ClearCheck implementation (p-values are per 2-tailed Wilcoxon ranksum test) 104

Radiotherapy Treatment Plan Evaluation

Table 3. Differences in dosimetric parameter compliance before and after ClearCheck implementation (p-values are per Chi-squared test)

EVA V95% (%) and body V107% (cc) of 1.38% and 8.03 cc respectively (p≤0.05). Improvements were also seen in PTV_WB_EVA V105% (%), ipsilateral lung Dmean (cGy), and ipsilateral lung V1600cGy (%) after software implementation. Differences in heart dosimetric parameters (including heart D5% (cGy), D0.03cc (cGy), and Dmean (cGy)) before and after software implementation were not statistically significant. Differences in the likelihood of achieving dosimetric parameter ideal values before and after treatment plan evaluation software implementation are shown in Table 3. There were statistically significant improvements in compliance for PTV_WB_EVA V95% (≥93%), body V107% (≤2 cc), body V107% (≤10 cc), and ipsilateral lung V1600cGy (≤10%) after software implementation.


any reduction in mean heart dose may provide clinical benefit to patients with left breast cancer who are receiving adjuvant radiotherapy. A limitation of our study was the small sample size, especially left breast cancer patients. Out of the total 52 patients included in the study, 33 were left breast cancer patients (16 presoftware implementation and 17 post-software implementation). Anatomical variations in patients play a large role in how much radiotherapy dose is received by the heart and other critical structures. Thus, given our limited patient number, patients with anatomy that is not conducive to heart avoidance may disproportionately skew the mean dose to the heart. However, we have standardized multiple aspects of the study including using the same dosimetrist and standardizing the drawing of target volumes and critical structures according to the RTOG and NRG Oncology guidelines.

The overall findings from our study show that the implementation of treatment plan evaluation software incidentally improved whole breast radiotherapy quality with improvements in both PTV coverage and critical structure avoidance. Mean PTV coverage (V95%) of patients with left or right breast cancer improved post-software implementation by 1.38% with a corresponding improvement of the PTV coverage meeting acceptable criteria (V95% ≥ 93%) by 21.8%. Additionally, improvements were seen in critical structure avoidance in regard to the body and ipsilateral lung in patients with right or left breast cancer, as well as to the heart in patients with left breast cancer. For instance, there was a 9.68 cGy reduction in heart mean dose, a 131.04 cGy reduction in heart D5%, and a 140.63 cGy reduction in heart D0.03cc in patients with left breast cancer post-software implementation.

Treatment plan evaluation software implementation improves whole breast radiotherapy quality. The automation and standardization of dosimetric goals for target coverage and critical structure avoidance has shown notable improvements in PTV coverage and heart and lung avoidance. Clinicians may wish to consider the routine use of radiotherapy treatment plan evaluation software to improve care for patients with earlystage invasive breast cancer.

Among the 33 women in this study with left breast cancer, the mean dose to the heart ranged from 61.9 cGy to 356.4 cGy, with an overall average of the mean doses of 110.26 cGy. There was a 9.68 cGy reduction in mean dose to the heart in patients with left breast cancer post-software implementation. These findings are clinically significant because the risk of major coronary events increases linearly with mean heart dose, with no threshold below which there is no risk (5). Thus,


Acknowledgments Omar Yumen, MD, was the attending radiation oncologist for the majority of the patients on this study. Anton Vatnitsky, CMD, did the radiotherapy planning for all of the patients in this study.


Early Breast Cancer Trialists' Collaborative Group (EBCTCG), Darby S, McGale P, et al. Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet. 2011;378(9804):1707-1716. 105

Radiotherapy Treatment Plan Evaluation

2. Early Breast Cancer Trialists' Collaborative Group (EBCTCG), Correa C, McGale P, et al. Overview of the randomized trials of radiotherapy in ductal carcinoma in situ of the breast. J Natl Cancer Inst Monogr. 2010;2010(41):162-177. 3. Nilsson G, Holmberg L, Garmo H, et al. Distribution of coronary artery stenosis after radiation for breast cancer. J Clin Oncol. 2012;30(4):380-386. 4. Darby SC, McGale P, Taylor CW, Peto R. Long-term mortality from heart disease and lung cancer after radiotherapy for early breast cancer: prospective cohort study of about 300,000 women in US SEER cancer registries. Lancet Oncol. 2005;6(8):557-565. 5. Darby SC, Ewertz M, McGale P, et al. Risk of ischemic heart disease in women after radiotherapy for breast cancer. N Engl J Med. 2013;368(11):987-998.


Scholarly Research In Progress • Vol. 4, October 2020

Analysis of Opioid Distribution Before and After Recreational Marijuana Legalization in California Michelle N. Anyaehie1*, Elijah J. Johnson1*, and Christian Pardo1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract The opioid epidemic escalated to an all-time high this past decade. States have searched for the right harm reduction and overdose prevention policies to no avail. Due to its controversial history, marijuana has not been regarded as an analgesic therapy. However, recent investigations have shown promising data highlighting the efficacy of marijuana as an analgesic and the use of marijuana for pain reduction could decrease opioid usage for pain mitigation. The distribution of codeine, fentanyl, hydrocodone, morphine, and oxycodone was examined per capita and by zip codes for the years 2014 to 2018 in California, which legalized recreational marijuana in 2016, and was compared to Texas, where medicinal and recreational marijuana is prohibited. Drug weights were obtained from the U.S Automation of Reports and Consolidated Orders System (ARCOS) and converted to their oral morphine milligram equivalent (MME) to facilitate analyses. Heat maps were constructed from distribution by zip code data to compare gross changes from 2014 to 2018. Opioid distribution per 100K people was analyzed to observe trends accounting for population differences. Both California and Texas showed statistically significant reductions in opioid distribution from 2014 to 2018, but a 43.7% decrease was seen in California versus 27.3% in Texas. Opioid distribution per 100K people in California decreased 38.9% over this time while distribution in Texas also decreased by 26.4%. Analyses showed that during this time period, opioid distribution decreased in larger areas of California than in Texas. The evidence supports marijuana legalization as a mitigating factor to the opioid epidemic. Furthermore, it highlights the idea that various legislative measures may help alleviate this crisis and a resolution to this issue is attainable. Continued studies on safer pain management alternatives will help identify measures that have a positive impact on the opioid epidemic and can be further developed to eradicate this issue.

Introduction The Controlled Substance Act of 1970 has classified marijuana as a Schedule I drug along with heroin and ecstasy. Under this classification, marijuana is regarded as having high abuse potential and no medical efficacy. Various opioids are listed as schedule II drugs, which are regarded as having limited medical use with a high abuse potential. Certain opioid combination drugs are listed as Schedule III, such as acetaminophen with codeine. Here, the opioid distributions of fentanyl, oxycodone, morphine, hydrocodone, and codeine in California after legalization of recreational marijuana have been analyzed. The opioids previously mentioned are listed from most to least potent (1) and have been chosen on a basis of familiarity to the general population.

Since the passing of the Controlled Substances Act, marijuana has traditionally been disregarded as an option for analgesic treatment, but recent studies may show an alternative view. A daily regimen of cannabis was shown to be an efficacious and safe alternative to conventional treatments for chronic pain (2). Other studies have also risen showing the efficacy of marijuana in treating pain associated with neuropathy (3), fibromyalgia (4) and inflammatory bowel disease (5). With such promising data illustrating efficacy in pain management, a separate study showed a 64% decrease in opioid use and a 45% increase in improved quality of life in patients that used marijuana as an alternative treatment for chronic pain (6). Along with these promising findings is increased interest in attempting such therapy among patients. A survey of patients with gynecological malignancies reported 60 patients out of 225 began using non-prescription cannabis as therapy for symptoms related to cancer therapy and 80 out of 225 are interested in attempting forms of cannabis-related therapy (7). Among the 60 patients using marijuana for symptom management, 45% reported decrease in use of narcotic prescriptions (7). The United States has suffered from a severe opioid epidemic dating back to the turn of the century. The opioid epidemic was characterized by an increase in opioid-related deaths following a steep increase in opioid and opioid combination prescription rates in the early 1990s. Opioid prescriptions quadrupled from 1998 to 2008 and resulted in a steady increase in medically related overdoses leading to deaths (8). With the Centers for Disease Control (CDC) reporting nationwide opioid prescription rates steadily increasing from 2006 through 2012 but showing a steady decrease from 2013 through 2018 (9), this shows there has been an effort to alleviate the crisis. These measures include removing pain as the fifth vital sign and the implementation of proper drug disposal programs, that also act in taking back unused medication (10). However, these measures, though imperative, have had little to no effect on the rate of distribution and control of opioid-related deaths. In fact, there were approximately 65,000 deaths in 2016, an increase of 21% from the year prior (8). California reported having 6 opioid-related deaths per 100K people in the year 2018 (11). Comparably, Texas reported 5 opioid-related deaths per 100K people in the year 2018 (12). The prescription rate was 35.1 prescriptions per 100 persons in California (11) and 47.2 prescriptions per 100 persons in Texas (12) during 2018. For our study, we chose to compare the states California and Texas. California has legislation in place which legalized cannabis use medicinally in 1996 and recreationally in 2016. Texas currently prohibits cannabis use medicinally and recreationally. Due to the controversial history of marijuana, there has been little research deciphering the effects of adult use of marijuana 107

Analysis of Opioid Distribution Before and After Recreational Marijuana Legalization in California

in a recreationally legalized state on opioid distribution trends. However, there have been recent changes with California legalizing medicinal marijuana in 1996 and recreational marijuana in November of 2016. With this change, recreational dispensaries began opening to individuals who were at least 21 years of age in January 2018. Accompanied by previous data concerning its effects as an alternative analgesic (2, 3, 4, 5, 6, 7), we can now begin to analyze the effects the change in legislation could have on the distribution of opioids. Our study analyzed the medical opioid distribution rates in California, where medical marijuana and recreational marijuana use is legalized, by zip code and per capita as compared to Texas, which has not legalized recreational or medical marijuana. We hypothesize that there will be a greater decrease in opioid distribution in California following this change in legislation than in Texas.

Texas had a distribution of approximately 100 000 or less while most zip codes in California ranged from 150 000 to 650 000, visualized in Figure 1A. However, in 2018, the largest distributions in California dropped to approximately 720 000 MME in the Bay Area and 525 000 MME in Southern California, decreases of approximately 50% and 40% respectively (Figure 1B). The surrounding Dallas area saw a decrease of 25%, while most of Texas did not show a comparable decrease relative to California as depicted in Figure 1C. Both California and Texas showed statistically significant decreases in opioid distribution (p<0.0001, p<0.0001), but California had a greater mean difference (162,476 versus 62,808) over this 5-year span. This represents an approximately 27.3% decrease in Texas and

Methods Procedure The Drug Enforcement Agency’s ARCOS Database is used to track controlled substances from manufacture, through distribution channels to point-of-sale at the level of hospitals, pharmacies, practitioners, and teaching institutions. The data is organized in several reports summarizing drug distribution by zip code, per 100K population gathered and includes summaries of purchases. Report 1, which reports drug distribution by zip code, and Report 3, which reports quarterly drug distribution by state per 100 000 population by gram were selected for analysis (13, 14). Data analysis The weight of each opioid was converted to oral Morphine Milligram Equivalents (MME) using the following factors: codeine 0.15, fentanyl base 75, oxycodone 1.5, morphine 1, and hydrocodone 1 (15). Heatmaps and graphs were made using Microsoft Excel and GraphPad Prism 8. Twotailed paired equal variance t-tests were performed using GraphPad Prism 8 software on zip code distribution data to determine statistical differences with p<0.05 considered significant. Outliers were tested for using Grubbs’ Test on GraphPad Prism.

Results Opioid distribution by zip code In both 2014 and 2018, California showed much more widespread and substantial cumulative opioid distribution as seen in Figure 1A and 1B. In 2014, portions of California surrounding the Bay Area had cumulative opioid distributions as high as 1 388 000 MME and 900 000 MME in portions of Southern California. In contrast, the highest distribution in Texas was 790 000 MME in the surrounding Dallas area. Most of 108

Figure 1. (A) 2014 cumulative opioid distribution in California and Texas by zip code (MME). (B) 2018 cumulative opioid distribution in California and Texas by zip code (MME). (C) Comparison of the cumulative change in MME of the selected opioids with darker green showing a greater decrease and darker red showing a greater increase.

Analysis of Opioid Distribution Before and After Recreational Marijuana Legalization in California


Figure 2. Net opioid distribution by zip code in California and Texas for the years 2014 and 2018.

43.7% decrease in California (Figure 2). There was an increase of approximately 6,000 MME in a 16-zip-code area of southern California near Glendale and Burbank. In this area, fentanyl, morphine, and oxycodone saw approximately 1.5, 2.4, and 1.8fold increases respectively, while codeine and hydrocodone did not undergo appreciable changes (Figure 1C). Opioid distribution per capita Cumulative opioid distribution per capita decreased in both California and Texas from 2014 to 2018 (38.9% and 26.4% percent respectively). The change in California was more linear for all drugs while Texas experienced more variable change (Figure 3). Individual opioid distribution changes in Texas showed much more variability with hydrocodone, fentanyl, and morphine distribution decreasing (42.3, 36.7, and 27.8 percent), oxycodone remaining essentially unchanged, and codeine distribution showing a slight upward trajectory (Figure 3). Distribution of all 5 opioids in California decreased to a greater extent than their Texan counterparts (hydrocodone 65.4%, fentanyl 50.3%, morphine 42.7%, oxycodone 32.4%, and codeine 19.1%) (Figure 3).

Evidence from our research showing retail distributions by zip code and per capita in California and Texas supports recreational marijuana legalization as a possible alternative to chronic pain in order to help reduce opioid distribution. The years analyzed show a greater decrease in California than in Texas (Figure 2). Opioid distribution per 100K people in California decreased 38% (Figure 1A) from 2014 through 2018, while distribution in Texas decreased by only 26% (Figure 1B). Analysis of our heat maps in (Figure 1C) illustrate a greater decrease in opioid distribution over large areas of California compared to Texas pre/post recreational legalization years. These findings are consistent with previous investigations demonstrating fewer opioid prescriptions written in states with access to medical marijuana (7). A similar study showed a significant correlation between marijuana use as an alternative treatment for chronic pain and a reduction in opioid use and medication side effects indicating a potential health benefit of replacing opioids with cannabis (6). Results from our MME per 100K population in Texas seen in (Figure 3), shows a drop in distribution for hydrocodone and an increase in codeine and oxycodone distributions. This is consistent with previous data showing that oxycodone and codeine increased by weight distribution in 2014 to 2017 specifically in Texas (12). This is possible due to 2014’s reclassification of hydrocodone combination products being moved from a Schedule III to a Schedule II classification by the Drug Enforcement Administration (12, 16). This reclassification controlled the use and misuse on hydrocodone, however potentially room for less-regulated Schedule III drugs like oxycodone and codeine to increase. Previously mentioned from 2014 to 2018 Figure 3 shows California’s steady decline across all 5 opioids assessed in our study. California’s trends may be explained by their approaches to mitigating pain and the opioid crisis by legalizing marijuana both medicinally and recreationally. To support this idea, the most recent study on women with gynecologic malignancies and the use of cannabis products specifically state their increased interest in guidance from physicians in the use of cannabis derivatives, since recreational marijuana has been legalized, and that significant numbers of these patients may already be using nonprescription cannabis products to manage pain (7).

Figure 3. Distribution of selected opioids in California and Texas from 2014 to 2018


Analysis of Opioid Distribution Before and After Recreational Marijuana Legalization in California

Harm reduction and medication-assisted treatments for opioid dependence have increased and become more available to help combat the staggering rate of overdose, like methadone and buprenorphine used as effective medications in the treatment of opioid dependence (17). Preparations should be made before individuals get to the level of dependence and overdose by promoting more natural preventive ways to initially treat pain. To strengthen the findings from our research, a recent study shows that states that enact medical cannabis laws have a 25% reduction in opioid overdose mortality, with this association strengthening each year after implementation (18). The goal moving forward with this research is to promote more natural, less addictive alternatives for analgesic relief, and to destigmatize labels placed on the safety and efficacy of marijuana, encouraging legalization in all states. There were a variety of limitations to our study. Due to the limited research on recreational marijuana for analgesic effects makes it difficult to distinguish the purpose for recreational use. Most findings on cannabis and chronic pain adhere to medical cannabis. Surveys taken for the use of recreational marijuana on dispensary records, although time-consuming for buyers, would aid in research for pain management data correlation, helping to differentiate between medical and recreational marijuana use for chronic pain (19). Secondly as we factored in ways to use our 3-digit zip code and per capita data for comparison during the years 2014 through 2018. Across our two states, we took into consideration that ARCOS only accounts for the annual estimation of the resident population for the United States. Being that Texas and California are two states with very rapid population growth, the population numbers that we used are most likely underestimated for the years 2014 through 2018. Expanding on our study by including more control comparison states with a population size closer to California’s population would be advantageous, showing multiple comparisons in distribution changes (20). Lastly although zip code data was accounted and extracted from ARCOS Report 1, this does not account for mail-order pharmacies and internet pharmacies that could ship different prescription opioids across state lines, including the 5 opioids used in our study (11).

Author Contributions Michelle N. Anyaehie conceived and designed the study, performed literature search and review, authored drafts of the paper, and approved the final manuscript. Elijah J. Johnson conceived and designed the study, performed literature search and review, authored drafts of the paper, and approved the final manuscript. Christian Pardo conceived and designed the study, analyzed the data, prepared figures, authored drafts of the paper, and approved the final manuscript.

Disclosures The authors do not report any conflicts of interest.

Acknowledgments We would like to thank Brian Piper, PhD, MS, Elizabeth Kuchinski, and Dan Kaufman at Geisinger Commonwealth School of Medicine for their support and guidance throughout this study. We would also like to thank Kathy-Ann Cadet, Jasmine Nkwocha, and Myiah Smothers for their feedback on the manuscript.

References 1.

Written by the Addiction Resource Editorial Staff. List of Opioids from Strongest to Weakest [Internet]. Addiction Resource. 2020. Available from: https://www.


Ware MA, Wang T, Shapiro S, Collet JP, Boulanger A, Esdaile JM, Gordon A, Lynch M, Moulin DE, O'Connell C. Cannabis for the management of pain: assessment of safety study (COMPASS). J Pain. 2015; 16(12):1233-42.


Lee G, Grovey B, Furnish T, Wallace M. Medical cannabis for neuropathic pain. Curr pain and headache reports. 2018;22(1):8.


Sagy I, Bar-Lev Schleider L, Abu-Shakra M, Novack V. Safety and Efficacy of Medical Cannabis in Fibromyalgia. J Clinical Med. 2019; 8(6):807.


Kienzl M, Storr M, Schicho R. Cannabinoids and Opioids in the Treatment of Inflammatory Bowel Diseases. Clin Transl Gastroenterol. 2020; 11(1).


Boehnke KF, Litinas E, Clauw DJ. Medical cannabis use is associated with decreased opiate medication use in a retrospective cross-sectional survey of patients with chronic pain. J Pain. 2016 Jun 1;17(6):739-44.


Blake EA, Ross M, Ihenacho U, Figueroa L, Silverstein E, Flink D, et al. Non-prescription cannabis use for symptom management amongst women with gynecologic malignancies. Gynecol Oncol Rrep. 2019; 30:100497.


Jones GH, Bruera E, Abdi S, Kantarjian HM. The opioid epidemic in the United States—Overview, origins, and potential solutions. Cancer. 2018; 124(22):4279-86.

Conclusion In our study, we discovered that the opioid distribution in the experimental state, California, showed a significant decrease of 43.7% compared to the control state of Texas (27.3%). These findings support our hypothesis by suggesting a correlation between California’s legislation legalizing recreational marijuana and the decreased opioid distribution when compared to a state that has legislation prohibiting medicinal or recreational marijuana use. However, these preliminary results should be followed with additional testing, including comparable states and legislation to further define the relationship. From our research, we hope to promote the research of natural, less addictive alternatives for analgesic relief to aid in alleviating the opioid epidemic. Also, we are advocates of providing accurate information and public availability of marijuana as an alternative analgesic therapy.


Analysis of Opioid Distribution Before and After Recreational Marijuana Legalization in California

9. U.S. Opioid Prescribing Rate Maps [Internet]. Centers for Disease Control and Prevention. Centers for Disease Control and Prevention; 2020 [cited 2020]. Available from: 10. Cabrera FF, Gamarra ER, Garcia TE, Littlejohn AD, Chinga PA, Pinentel-Morillo LD, et al. Opioid distribution trends (2006–2017) in the US Territories. Peer J. 2019; 7:e6272.

22. Schultz, S, et al. Analgesic Utilization before and after Rescheduling of Hydrocodone in a Large Academic Level 1 Trauma Center. J Opioid Manag. 2016; 12(2). 23. Bachhuber MA, Saloner B, Cunningham CO, Barry CL. Medical cannabis laws and opioid analgesic overdose mortality in the United States, 1999-2010. JAMA Int Med. 2014 Oct 1;174(10):1668-73.

11. National Institute on Drug Abuse. California: OpioidInvolved Deaths and Related Harms [Internet]. NIDA. 2020. Available from: 12. National Institute on Drug Abuse. Texas: Opioid-Involved Deaths and Related Harms [Internet]. NIDA. 2020. Available from: 13. Ighodaro EO, McCall KL, Chung DY, Nichols SD, Piper BJ. Dynamic changes in prescription opioids from 2006 to 2017 in Texas. Peer J. 2019; 7:e8108. 14. Drug Enforcement Administration. Automation of Reports and Consolidated Orders System (ARCOS) Retail Summary Reports. Office of Diversion Control. 2008; 2011. 15. Piper BJ, Shah DT, Simoyan OM, McCall KL, Nichols SD. Trends in medical use of opioids in the US, 2006–2016. Am J Prev Med. 2018;54(5):652-60. 16. Bohnert AS, Guy GP, Losby JL. Opioid prescribing in the United States before and after the Centers for Disease Control and Prevention's 2016 opioid guideline. Ann Intern Med. 2018; 169(6):367-75. 17. Tran S, Lavitas P, Stevens K, Greenwood BC, Clements K, Alper CJ, et al. The effect of a federal controlled substance act schedule change on hydrocodone combination products claims in a Medicaid population. J Manag Care Spec Pharm. 2017; 23(5):532-9. 18. Mattick RP, Breen C, Kimber J, Davoli M. Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database Syst Rev. 2014(2). 19. Piper BJ, DeKeuster RM, Beals ML, Cobb CM, Burchman CA, Perkinson L, Lynn ST, Nichols SD, Abess AT. Substitution of medical cannabis for pharmaceutical agents for pain, anxiety, and sleep. Journal of Psychopharmacology. 2017 31(5):569-75. 20. Andreae MH, Rhodes E, Bourgoise T, Carter GM, White RS, Indyk D, Sacks H, Rhodes R. An ethical exploration of barriers to research on controlled drugs. The American Journal of Bioethics. 2016; 16(4):36-47. 21. Steckler TJ, Mosher HJ, Desloover-Koch Y, Lund BC. Impact of hydrocodone reclassification on analgesic prescribing in the Veterans Health Administration. Am J Health Syst Pharm. 2019; 76(Supplement_2):S61-7.


Scholarly Research In Progress • Vol. 4, October 2020

Emerging Modalities that Support Differential Diagnoses of Creutzfeldt-Jakob Disease and other Neurodegenerative Disorders Yasmin R. Mamani1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Creutzfeldt-Jakob disease (CJD) is a transmissible spongiform encephalopathy (TSE) that afflicts humans. Like all TSEs, CJD belongs to a family of incurable neurodegenerative disorders characterized by the accumulation of misfolded prion proteins in the central nervous system. Due to the similarities between CJD and other neurodegenerative diseases, such as an extremely low rate of incidence and older age of onset, CJD is frequently misdiagnosed as a rapidly progressive dementia disorder. This review examines several different case reports that highlight the challenges in differentiating CJD from another neurodegenerative diseases. Secondly, as no test exists that can reliably diagnose a living individual suspected of CJD, this review investigates techniques that are currently being developed for CJD diagnosis that are ante-mortem. Recent studies indicate that new accurate and sensitive approaches, such as neurofilament biomarkers and prion seeding assays, introduce a promising outlook into the diagnostic capacity for CJD. Indeed, properly discriminating CJD and its variants from other diseases is important for a patient’s subsequent treatment plan and the family’s peace of mind. Recent studies indicate that neurofilament light chains are a promising surrogate biomarker candidate in the field of TSE diagnostics.

Introduction As with all transmissible spongiform encephalopathies (TSEs), CJD is caused through the accumulation of misfolded prion proteins in the central nervous system. Scrapie prion isoforms (PrPSc) are the pathogenic form of normal prion proteins (PrPC) and induce a conformational change in PrPC alpha-helical structure to that of beta-pleated sheets, resulting in the formation of indigestible amyloid plaques. Seeding between cells produces a highly augmented conversion of PrPC to PrPSc, although the mechanism remains nebulous. PrPC may sporadically misfold, leading to a prion disease (1). According to the Centers for Disease Control and Prevention, sporadic CJD (sCJD) has a disease rate of approximately 1 case per 1 million people of the total population per year (2). Even more terrifying, 85% of all cases are sporadic, meaning that PrPC misfolds without ingestion of infected meat, familial history, or contaminated medical instruments. Many symptoms of CJD are similar to those of other neurodegenerative diseases. For example, some patients with CJD experience a shuffling gait, where the patient never fully lifts his/her feet off the ground to ambulate (3). A shuffled gait is one of the most common presentations of Parkinson’s disease (PD). Other shared manifestations 112

of neurodegenerative disease include ataxia, severe cognitive decline, a personality change, amnesia, myoclonic jerking, difficulty swallowing/speaking, coma, and eventual death (4). Discerning between other neurodegenerative disorders and CJD on a molecular level without a reliable test is difficult enough, but the heterogeneity of symptoms for all aforementioned pathologies only compounds the diagnosis. The challenge for health care providers lies in the discriminating CJD from other neurodegenerative diseases with the same symptoms. A screening technique that can detect CJD antemortem is essential. Researchers and clinicians attempt to identify CJD in two phases. First, they extract blood serum or cerebrospinal fluid (CSF) from potential CJD patients to look for a possible surrogate biomarker, such as neurofilament light chains (NFL). Secondly, standard assays, like RT-QuIC, amplify any detected amount of PrPSc found in either the CSF or blood serum samples to obtain a final diagnosis. If there is PrPSc is present, even at miniscule quantities, then amplification should confirm its presence (5). This article examines several studies that assess the application of NFL as a surrogate biomarker, as well as the application of RT-QuIC assays to facilitate the conversion of PrPC to PrPSc. Biomarkers are measurable biomolecules found in bodily fluids or tissues that can be signs of normal or abnormal biological conditions (6). They are an integral part of medicine, with applications in clinical and research settings and must be reliably reproduced for objective measurement. Unfortunately, prion diseases are one group of disorders that remains without an officially accepted biomarker. Ascertaining a clinical end point for patients suffering from CJD is difficult without a measurable outcome, leaving patients and their families in a state of uncertainty. Abu-Rumeilah et al. claims that the insufficient availability of reproducible biomarkers for CJD significantly limits the ability of clinicians to apply a broad screening program (7). Currently, 14-3-3 protein, t-tau, and NFL are some of the most researched biomarker candidates for CJD diagnosis. Moreover, 14-3-3 protein is the only candidate blood biomarker authorized by the World Health Organization (WHO) (5). As auspicious as these candidate biomarkers appear, most are still combined with a secondary assay to confirm prion activity. Research continues to explore the clinical and research utility of NFL as a biomarker for CJD. This review aims to provide a critical assessment of current literature that discusses emerging diagnostic modalities for CJD. An ideal examination would be sensitive and specific for CJD and able to distinguish the proteopathy from other neurodegenerative diseases with comparative presentations.

Emerging Modalities that Support Differential Diagnoses of Creutzfeldt-Jakob Disease

Methods The databases utilized for this research included PubMed, sourced exclusively through the Geisinger Commonwealth School of Medicine library collection, and Google Scholar. The following keywords and terms were used to facilitate the literature search: Creutzfeldt-Jakob disease, proteopathy, neurofilament light chain, biomarker, and transmissible spongiform encephalopathy. Care was taken to ensure that the articles chosen were recent and relevant to the topic, with the oldest source utilized having been published in 2009.

Discussion Discriminating CJD from other proteopathies CJD shares many symptoms with other more common degenerative diseases, like Alzheimer’s disease (AD), amyotrophic lateral sclerosis (ALS), and dementia disorders. Patients experience a personality change, cognitive decline, ataxia, loss of gross motor faculties, and eventual death. The changes associated with CJD, however, are more pronounced and sudden when compared to AD, ALS, and other disorders (8). This review investigated multiple case reports that display the challenges of neurodegenerative diagnosis. Case reports often cite the complications in identifying CJD from a pool of non-specific symptoms that may apply to multiple neurodegenerative conditions. One report discussed a case of CJD in an otherwise healthy 76-year-old male who, following a 6-month period of rapidly progressive dementia, presented a wide range of neurodegenerative symptoms including: ataxia, aphasia, myoclonus, akinetic mutism, hallucinations, and progressive dementia. The patient expired within 8 weeks of hospitalization, and an autopsy was performed. Postmortem tissue analysis confirmed that he had CJD. Pathologists reported micro-spongiosis, the classic hallmark of TSEs, as well as neuronal death, and gliosis of the basal ganglia, hippocampus, cerebellum, and cortex. This patient had no family history of dementia or CJD, as was the situation with two other case reports (9). These involved patients who possessed no known family history of dementia, CJD, or malignancy. Both investigative teams relied on supplementary tests to support a probable CJD diagnosis for each patient, although none of these assessments officially confirm the presence of a prion disease (10 ,11). Diffusion-weighted magnetic resonance imaging (DW-MRI) can reveal much about the pathological changes of a patient’s brain and is regarded as one of the most valuable techniques for investigating a possible CJD case. Different variants of CJD often exhibit subtle but distinct characteristics that may show up on a DW-MRI. DW-MRI combines routine MRI imaging with the diffusion of water molecules over the brain; movement/localization of the water molecules contrasts with the original MRI, exposing the brain’s topography. Regions of high neuronal loss, as is seen in CJD, will reflect intense regions of water perfusion (12). Kwon et al. utilized DW-MRI using gadolinium for assessing a 76-year-old male patient who experienced an abrupt decline in cognitive and physical functioning over a 2-week period. The patient passed away only 5 weeks after symptom onset. A DW-MRI found bilateral hyperintensities of the cortical gyri, which is a common characteristic of sCJD (10). Tomizawa et al. encountered an

extremely unique case in which a patient with genetic CJD (gCJD) mimicked dementia with Lewy bodies (DLB). A 75-yearold female patient deteriorated over the course of 3 years, experiencing dementia, parkinsonism, depression, rigidity, bradykinesia, and a resting hand tremor. Although she initially scored a 15 on the Mini Mental State Examination (MMSE), her disorientation and cognitive dysfunction continued to decline, and her score plummeted to 4 when tested again at age 77. Genetic counseling concluded that a valine-toisoleucine point mutation at codon 180 of PRNP had resulted in gCJD. Her case is unusual due to the unique nature of her DW-MRI results. In cases where sCJD has been reported to mimic DLB, hyperintensities had consistently shown up early during surveillance of a patient. This patient, however, exhibited radiological data accordant with patients who have suffered from DLB for almost 3 years. Cortical hyperintensities were not found until 39 months after she had presented at the hospital. These findings make sense, as gCJD has been known to have a slower pathogenic course than other CJD variants. Information obtained from DW-MRI findings could help discriminate between the various forms of CJD (11). Although DW-MRI is heralded as an excellent diagnostic tool for detecting some characteristics associated with CJD, it is just one of many ancillary assessments that are performed to extract supportive data for a probable CJD case in question. None of these preliminary tests provide data that is clinically accepted as criteria for a positive CJD diagnosis. A sensitive and specific test for CJD has yet to be created, but there are several emerging procedures reported in the literature that have had promising experimental results. Neurofilament light chains NFL is a subunit of a heteropolymer known as a neuronal intermediate filament (NIF), which is a major constituent of the axon architecture and the most plentiful feature of mature neuronal cytoskeletons. NFL are highly localized to axons, where they support axonal strength and growth. Elevated NFL concentrations may indicate the presence of axonal and white matter degeneration. Several studies have examined the potential for NFL to act as a biomarker for CJD when extracted from blood and CSF samples. In a study comparing NFL concentrations among two independent cohorts composed of patients suspected of sCJD, NFL levels were significantly increased in individuals confirmed to be positive for sCJD. Each cohort contained non-sCJD patients who, although initially suspected of sCJD, were excluded because their supplementary test results did not meet the diagnostic criteria for a probable prion disease. Investigators noted that when compared to all other neurodegenerative disorders, NFL concentrations were highest in cases of sCJD. NFL might be a candidate biomarker for the differential diagnosis of sCJD (13). Wang et al. conducted a meta-analysis of 36 studies that examined NFL efficacy as a biomarker in blood and CSF. Axonal degeneration in the CNS resulting from neurodegenerative disorders disperses NFL into the interstitial fluid surrounding the neurons, which eventually migrates into the blood and CSF. Because the populations differed across a variety of laboratory locations, investigators utilized a ratio of means method to approximate the difference in NFL concentration between controls and cases. NFL levels


Emerging Modalities that Support Differential Diagnoses of Creutzfeldt-Jakob Disease

were found to be significantly elevated among cases of dementia, ALS, Huntington’s disease, and CJD. However, NFL concentrations were not significantly increased in cases of PD, but they were increased in PD-related disorders. The investigation concluded that NFL may be a promising biomarker for a differential diagnosis between certain neurodegenerative disorders, but not without the supplementary assessments or presentation of clinical symptoms (14). These findings are congruent with results from another investigation that compared the diagnostic potency of 3 different biomarkers: 14-3-3, t-tau, and NFL. Abu-Rumeileh et al. found that NFL performance was highly specific to distinguishing between the different forms of CJD. It was evident that the degree of sub-cortical pathology strongly affected the concentrations of NFL, possibly explaining why NFL was so able to differentiate between different types of CJD. However, NFL did not perform to the same degree of sensitivity as t-tau, which had an overall sensitivity of 91.3% for all CJD types. Summarized, studies show that NFL biomarkers have the potential to contribute to a differential diagnosis of CJD but may not be appropriate to alone definitively confirm a prion disorder (7). As mentioned, axonal degradation disseminates NFL particles in the CSF and blood (13). Countless diseases and pathological phenomena can be screened using a blood sample containing some measurable biomarker. Utilizing a blood biomarker for CJD diagnosis might prove more pragmatic and less invasive than having to extract CSF for analysis, which involves a lumbar puncture (18). Currently, the WHO has yet to endorse a blood biomarker capable of detecting CJD. If NFL fragments from axonal degeneration leaked into the interstitial fluid and migrated to the CSF and blood, then the concentrations of both compartments should be comparable. Thus, a blood sample could be representative of prion activity in the central nervous system (15). Steinacker et al. looked at the difference in NFL levels between blood serum and CSF samples. The retrospective study called for CSF and serum samples to be taken from patients 103 patients in total: 60 controls, both with and without dementia, and 43 cases of CJD. It should be noted that of the 43 cases, 4 had been labeled as probable, but not validated neuropathologically. Of those individuals neuropathologically authenticated to be CJD positive, increases in serum NFL corresponded to measurements of increased NFL found in the CSF. Perhaps blood serum could provide an alternative to CSF extraction (5). Real-time quaking induced conversion assay One case report describes a living patient who was diagnosed with sCJD using an experimental RT-QuIC assay. This technique is interesting because RT-QuIC does not detect a surrogate biomarker, such as NFL or 14-3-3, but the PrPSc itself. The assay is quickly becoming one of the most specific and sensitive modalities to detect CJD antemortem using a CSF sample. Unfortunately, the patient passed away before a follow-up could ensue and any postmortem biopsy of her neural tissue could not be found, so sCJD could not be precisely diagnosed. However, an RT-QuIC assay of her CSF sample proved positive for sCJD, as confirmed by the National Prion Disease Pathology Surveillance Center. Investigators utilized the standard procedure for CJD diagnosis via a RTQuIC assay. 114

The 58-year-old female patient presented to the hospital after a 6-week history of serious decline. Her daughter disclosed that her mother had experienced an unsteady gait and general confusion for the past 2 months. Within the last week, the patient’s walk worsened to an ataxic “shuffled” gait, and she displayed dysdiadochokinesia, the inability to execute quick and alternating gestures. The investigative team performed a lumbar puncture to extract CSF for the study. The patient displayed myoclonic jerks, bruxism, and a total loss of orientation with regards to time, space, and self over the course of the study. RT-QuIC assays exploit the seeding mechanism of PrPSc by exposing patient-derived samples of CSF, assumed to contain PrPSc, to recombinant PrP (rPrP), a surrogate substrate that imitates PrPC found naturally in the brain. If PrPSc is present, then it should amplify the rapid conversion of rPrP to PrPSc, eventually resulting in aggregations of amyloid fibrils. The protocol involves distributing rPrP substrate across a 96well plate already seeded with rPrPSc samples; experimental RT-QuIC usually utilizes rPrPSc to seed as well (3). It should be noted that rPrP can only be sourced from truncated recombinant protein, from the vole rodent, if investigators plan to seed the recombinant protein with PrPSc (16). The aforementioned investigative team most likely seeded each well with the patient’s CSF sample. Applied heat and vigorous shaking of the wells causes the amyloid plaque formation and degradation to occur at a faster rate. PrPSc seeds continue to convert rPrP, generating even more seeds at an exponential rate. Amyloid plaques are identified using thioflavin T dye (ThT), which contains fluorophores that bind to amyloid substrates and generate a measurable fluorescence spectrum that can be interpreted in real time. However, the study did report that limitations were imposed by several factors. First, not possessing autopsy results of the patient’s neural tissue made it impossible to verify the RT-QuIC results. Second, there are many possible factors that could contribute to variable results in future RT-QuIC analyses, such as speed and temperature of thermal mixer appliances, human error in distribution of seeds across a well-plate, pH of the solutions, and intervals of shaking (3). Timed protein seeding assay The process behind a RT-QuIC amplification assay is similar to a timed protein seeding assay, which also aims to quantify the conversion of PrPC to PrPSc in real time. Quantification is measured and interpreted via a fluorescence spectrum graph. Gray et al. sought to investigate the in vitro infectivity of prions in elk neural tissue; chronic wasting disease is a prion disorder that afflicts cervids. This study is novel because amyloid fibril formation has not yet been timed in non-human animals. As the paper mentions, RT-QuIC assays and amyloid seeding assays both utilize rPrP as a substrate for conversion by PrPSc. Researchers combined a sample of rPrP with varying brain homogenate solutions that were either positive or negative for PrPSc. Should the sample homogenate be positive for PrPSc, then the conversion rates would be identified on the ThT signal times on the fluorescence spectrum. The study concluded that the ThT signal times are the specific diagnostic component of the assay (17).

Emerging Modalities that Support Differential Diagnoses of Creutzfeldt-Jakob Disease

Conclusion In summary, there is still a significant amount of research to be conducted before a sensitive and specific CJD diagnostic test can be employed for routine use. Scientists have made impressive strides in the field of prion research, exploring different biomolecules and their capacity to serve as potential biomarkers. Although NFL have demonstrated a role as a potential biomarker for the differential diagnosis of CJD subtypes, such as sCJD and gCJD, they do not perform to the same caliber as other neural substrates, such as t-tau or 14-3-3. NFL might be a good tool to combine with the latter substrates to maximize the precision of a CJD diagnosis. Although other substrates have proven more effective at diagnosing CJD outright, such as protein 14-3-3 and t-tau, NFL has proven the most specific to the subtype of CJD. Additionally, researchers have continued to refine commonplace laboratory techniques for the application of prion detection. RT-QuIC assays are used to amplify the concentration of PrPSc from a patient sample by exposure to a concentration of PrPC, indicating the absence or presence of prions. A diagnostic test is urgently needed to provide patients and their families an answer antemortem. Furthermore, a proper examination procedure would allow health care providers to provide their patients with the best care possible, as well as develop new treatments.

Acknowledgments Firstly, I would like to thank Iris Johnston for helping me acquire the necessary articles that were essential to this project; without her help, I would not have been able to conduct such thorough research. I also wish to thank Dr. Brian J. Piper for serving as such an excellent mentor to my topic. His expertise in neuroscientific research provided me with all the guidance to construct this review. Finally, I want to thank Geisinger Commonwealth School of Medicine for providing me the education, tools, and opportunity to embark on this project.

References 1.

Cobb NJ, Surewicz WK. Prion Diseases and Their Biochemical Mechanisms. Biochemistry. 2009 48(12): 2574–2585. doi: 10.1021/bi900108v

2. Creutzfeldt-Jakob Disease. Classic CJD [Internet]. 2018 Oct 9. Available from: index.html 3. Reddy VD, Hamed A, Settipalle N, et al. Real-time Quaking-induced Conversion Assay for the Diagnosis of Sporadic Creutzfeldt-Jakob Disease in a Living Patient. Infect Dis (Auckl) 2019 12. doi: 10.1177/1178633719874797 4. Creutzfeldt-Jakob disease. Symptoms & Causes [Internet]. 2018 Oct 4. Available from: diseases-conditions/creutzfeldt-jakob-disease/symptomscauses/syc-20371226 5. Steinacker P, Blennow K, Halbgebauer S, et al. Neurofilaments in blood and CSF for diagnosis and prediction of onset in Creutzfeldt-Jakob disease. Sci Rep. 2016 6(1). doi: 10.1038/srep38737

6. Strimbu K, Tavel JA. What are biomarkers?. Curr Opin HIV AIDS. 2010 5(6): 463–466. COH.0b013e32833ed177 7.

Abu-Rumeileh S, Baiardi S, Polischi B, et al. Diagnostic value of surrogate CSF biomarkers for Creutzfeldt–Jakob disease in the era of RT-QuIC. J Neurol. 2019 266(12): 3136–3143. doi: 10.1007/s00415-019-09537-0

8. Erdtmann R, Sivitz L. Advancing Prion Science. 2004 doi: 10.17226/10862 9. Parmar P, Cooper CL, Kobewka D. An Evaluation of Rapidly Progressive Dementia Culminating in a Diagnosis of Creutzfeldt–Jakob Disease. Case Rep Infect Dis. 2018, 1–4. doi: 10.1155/2018/2374179 10. Kwon GT, Kwon MS. Diagnostic challenge of rapidly progressing sporadic Creutzfeldt-Jakob disease. BMJ Case Rep. 2019 12(9). doi: 10.1136/bcr-2019-230535 11. Tomizawa Y, Taniguchi D, Furukawa Y. Genetic Creutzfeldt-Jakob disease mimicking dementia with Lewy bodies: Clinical and radiological findings. J Neuro Sci. 2020 409. doi: 10.1016/j.jns.2019.116604 12. Abdulmassih R, Min Z. An ominous radiographic feature: cortical ribbon sign. Intern Emerg Med. 2015 11(2): 281–283. doi: 10.1007/s11739-015-1287-4 13. Kanata E, Golanska E, Villar-Piqué A, et al. Cerebrospinal fluid neurofilament light in suspected sporadic CreutzfeldtJakob disease. J Clin Neurosci. 2019 60: 124–127. doi: 10.1016/j.jocn.2018.09.031 14. Wang SY, Chen W, Xu W, et al. Neurofilament Light Chain in Cerebrospinal Fluid and Blood as a Biomarker for Neurodegenerative Diseases: A Systematic Review and Meta-Analysis. J Alzheimer’s Dis. 2019 72(4): 1353–1361. doi: 10.3233/jad-190615 15. Bacioglu M, Maia LF, Preische O, et al. Neurofilament Light Chain in blood and CSF as marker of disease progression in mouse models and in neurodegenerative diseases. Neuron. 2016 91(2): 494–496. doi: 10.1016/j. neuron.2016.07.007 16. Kaelber N, Bett C, Asher DM, et al. Quaking-induced conversion of prion protein on a thermal mixer accelerates detection in brains infected with transmissible spongiform encephalopathy agents. Plos One. 2019 14(12). doi: 10.1371/journal.pone.0225904 17. Gray JG, Graham C, Dudas S, et al. Defining and Assessing Analytical Performance Criteria for Transmissible Spongiform Encephalopathy–Detecting Amyloid Seeding Assays. J Diagn. 2016 18(3): 454–467. doi: 10.1016/j.jmoldx.2016.01.005 18. Teunissen CE, Petzold A, Bennett JL, et al. A consensus protocol for the standardization of cerebrospinal fluid collection and biobanking. Neurology. 2009 73(22): 1914–1922. doi: 10.1212/wnl.0b013e3181c47cc2


Scholarly Research In Progress • Vol. 4, October 2020

Severe Pertussis Infection with Hyperleukocytosis Greater Than 200,000 103/uL in a 10-month-old Unvaccinated Amish Female: A Case Report and Review of the Literature Stephen Long1† and R. Blake Lowe2

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Internal Medicine-Pediatrics, Geisinger Medical Center, Danville, PA 17822 †Doctor of Medicine Program Correspondence: 1


Abstract Bordetella pertussis commonly infects individuals of all ages, but pertussis, the disease caused by B. pertussis infection, is most severe in young infants. Severe pertussis — defined by the presence of refractory hypoxemia, cardiogenic shock, pneumonia, and hyperleukocytosis — is associated with significant morbidity and mortality. Both hyperleukocytosis and pulmonary hypertension have been found to be predictive of mortality in young infants, and as such, leukoreductive strategies such as leukapheresis and exchange transfusion have been employed to treat these complications. Pulmonary hypertension is thought to be a result of aggregation of white blood cells in pulmonary vasculature; however, studies have suggested that the mechanism of pulmonary hypertension is multifactorial. We report a case of a 10-month-old unvaccinated Amish female with pertussis complicated by an initial hyperleukocytosis of 204,900 103/uL successfully treated with leukapheresis in our pediatric intensive care unit. This infant never showed signs of pulmonary hypertension, which is often associated with hyperleukocytosis in severe or fatal cases of pertussis in infants and neonates. To our knowledge, this is the most significant degree of hyperleukocytosis reported in pertussis. The findings in this case support the clinical utility of leukoreductive therapy in severe pertussis and provide some evidence that the mechanism of pulmonary hypertension in these patients is multifactorial.


We report a case of a 10-month-old unvaccinated Amish female with severe B. pertussis infection complicated by an initial WBC of 204 900 103/uL treated successfully with hyperhydration and leukoreduction via leukapheresis.

Case Presentation


A 10-month-old, previously healthy Amish female presented to a local emergency department for evaluation of 1 week of worsening cough, several episodes of post-tussive emesis, reduced urine output, and increased work of breathing. The child was born at home and received no vaccinations. On examination, she was found to be febrile at 38.3°C with a respiratory rate of 38 breaths/minute, oxygen saturations of 88% on room air, tachycardic with a heart rate of 180 BPM, and hypertensive with a blood pressure of 130/90. Chest examination demonstrated clear lung sounds bilaterally but accessory muscle use. Complete blood count showed marked hyperleukocytosis with white blood cell count (WBC) of 204.90 103/uL with 44% neutrophils and 46% lymphocytes and platelet count of 780 103/uL. Peripheral saturation improved with supplemental oxygen via nasal canula. She received intravenous azithromycin and ceftriaxone and was transferred to the pediatric intensive care unit (PICU) for further management. Respiratory PCR of a nasopharyngeal swab was found to be positive for Bordetella pertussis. Flow cytometry of a blood sample showed no evidence of a neoplastic process. A bone marrow biopsy confirmed the flow cytometry findings.

Bordetella pertussis infection is the most poorly controlled vaccine-preventable illness and has increased in incidence in the United States over the past two decades (1). Pertussis in neonates and young infants is frequently associated with hyperleukocytosis, pulmonary hypertension, and hypoxemia (2–7). Hyperleukocytosis, with peak WBC counts often surpassing 100 000 103/uL, is a frequently reported phenomenon of pertussis with many retrospective observational studies linking its presence to adverse outcomes and mortality (2, 4, 8–17). Hyperleukocytosis typically occurs in cases of severe pertussis, which is defined by pneumonia with refractory hypoxemia, cardiogenic shock, and severe hyperleukocytosis (9). The development of pulmonary hypertension in these patients is thought to be a result of leukocyte-rich thrombi becoming lodged in the pulmonary arterial system (16, 18). Hyperhydration, exchange transfusion, and leukapheresis have been employed successfully to prevent and treat pulmonary hypertension in cases of pertussis complicated by hyperleukocytosis (8–14).

Aggressive intravenous fluid hydration was initiated. A transesophageal echocardiogram did not reveal any evidence of pulmonary hypertension on day one of hospitalization. Due to the continued high risk of developing hyperviscosityrelated complications and small vessel occlusion, the decision was made to proceed with double-volume leukoreduction via apheresis with goal WBC of less than 50,000 103/uL. She underwent the first round of leukapheresis on day one of hospitalization, which resulted in a reduction of her WBC to 98.45 103/uL with 25% lymphocytes and 64% neutrophils. Unfortunately, this was complicated by worsening interstitial edema thought to be secondary to the rapid change in oncotic pressure produced by leukapheresis. Bilevel positive airway pressure ventilation was initiated. A second leukoreduction was carried out on the day two of hospitalization without significant improvement. A third leukoreduction targeting granulocytes was conducted on the hospital day three and her WBC was reduced to 50.62 103/uL before rebounding to 70 103/uL 12 hours later. She did not tolerate attempts to

Severe Pertussis Infection with Hyperleukocytosis

wean respiratory support; therefore, the decision was made to proceed with single-volume apheresis to reduce the circulating volume of pertussis toxin (PT). Afterward, she made very slow improvements and was stable for discharge from the hospital without supplemental oxygen after 21 days.

Discussion Pertussis remains a common infectious disease in individuals of all ages worldwide and can produce a wide spectrum of clinical manifestations, ranging from upper respiratory symptoms to hypoxia, hypotension, and shock. The severity of disease in young infants is influenced by several factors, including maternal vaccination during pregnancy, infant age and weight, vaccination status, and breastfeeding (4). The most significant risk factor for neonatal infection appears to be a large family size (4). This 10-month-old unvaccinated female is a member of a self-contained Amish community. Although the Amish beliefs do not universally reject all vaccinations, rates of vaccination in these communities are significantly lower than national averages (19). Children in similar Amish and Mennonite communities have been found to be at increased risk of hospitalization due to vaccine-preventable illnesses (20). Hyperleukocytosis with lymphocytic predominance is a well-documented occurrence in severe pertussis, with cases reported throughout the 1900s until recent years (8–14). This phenomenon occurs more frequently in young infants than older children and adults (5). Several retrospective studies have established hyperleukocytosis as an indicator of increased mortality in young infants and an independent predictor of mortality in all patients (2, 4, 15–17). One prospective cohort study found that infants with an initial WBC greater than 50,000 103/uL had a relative risk of death of 9.8 (2). Hyperleukocytosis in pertussis is thought to be caused by pertussis toxin (PT), which functions as an inhibitor of G-protein coupled receptor (GPCR) signaling (22). In animal models, PT has been shown to reduce expression of several cellsurface markers — including L-selectin and leukocyte function antigen-1 (LFA-1) — needed for lymphocyte extravasation into lymphoid tissue. Thus, impaired extravasation causes an increase in the number of circulating lymphocytes within the vasculature (23). Additionally, PT is theorized to cause widespread insult to cardiopulmonary function by inhibiting G-protein signaling in the heart and lungs (14). Infants with higher peak WBC counts are at an increased risk of developing pulmonary hypertension which is associated with increased mortality (2, 21). Hyperleukocytosis leading to intravascular leukocyte-rich thrombi in the pulmonary vasculature was one proposed mechanism (7). This has been supported by post-mortem histologic analyses of pulmonary tissue from infants who succumbed to pertussis (7, 18). It is important to note that other similar studies from postmortem tissue analysis have found the presence of thrombi to be inconsistent, which suggests the mechanism is likely to be multifactorial (16, 18). Hyperleukocytosis and pulmonary hypertension appear to be most common in infants <120 days old, suggesting that PT may induce age-specific effects (22). Expression and function of several GPCRs have been found to change throughout

development. One such GPCR, the angiotensin-II receptor AT2, has been shown to be highly expressed in fetal tissue before being balanced by similar levels of AT1 receptors in childhood (24). The AT2 receptor functions to help regulate smooth muscle cell production in vasculature. Inhibition of this receptor by pertussis toxin could result in unopposed smooth muscle cell proliferation, leading to pulmonary arterial thickening, vasoconstriction, and pulmonary hypertension (22). A histopathologic study by Winter and Harriman demonstrated medial thickening of pulmonary arteries in 4 out of 5 infants after fatal pertussis complicated by pulmonary hypertension (14). Due to the well-established relationship between hyperleukocytosis and mortality, several strategies have been employed to reduce high WBC counts and prevent microvascular complications. Case reports have described successful use of leukoreduction by either exchange transfusion or leukapheresis. One study provided clear indication that leukoreduction may be associated with ICU survival (18). There is currently no evidence favoring the use of leukapheresis over exchange transfusion or vice versa. This clinical decision depends in large part on resource and staff availability. To our knowledge, this is the highest degree of hyperleukocytosis secondary to pertussis reported in the literature. Moreover, most cases of hyperleukocytosis occurred in infants less than 120 days old, while this infant was 10 months old at the time of admission. This child did not develop echocardiographic signs of pulmonary hypertension despite her marked leukocytosis of greater than >200,000 103/uL. Given the studies suggesting that higher peak WBC counts are correlated with — and may even be causative of — the development of pulmonary hypertension, the lack of this finding in this case suggests that hyperleukocytosis is one aspect of a multifactorial mechanism or that older infants may be less susceptible to the development of pulmonary hypertension (2). While there are no clear guidelines on when in the clinical course to perform leukoreductive therapy in severe pertussis, this infant underwent consecutive rounds of leukopharesis on the first 3 days of hospitalization. Prompt initiation of leukoreductive therapy in this child may have reduced the risk of developing pulmonary hypertension later in the clinical course.

Conclusion We report a case of dramatic hyperleukocytosis in a 10-month-old unvaccinated female successfully treated with leukapheresis. This case underscores the need for improved vaccination rates among target communities and provides some insight into the complex interaction between hyperleukocytosis and pulmonary hypertension. Several studies have proposed guidelines for the use of leukoreduction in cases of severe pertussis; however, this evidence is still evolving and no formal, society-backed recommendations exist. Even scarcer are the data available for older infants with extreme hyperleukocytosis. We believe the outcome of this case supports the viability of leukoreduction via leukapheresis in severe pertussis. This treatment likely provides benefit by reducing both the number of circulating


Severe Pertussis Infection with Hyperleukocytosis

leukocytes and the amount of circulating pertussis toxin and may help prevent or decrease the risk for developing pulmonary hypertension leading to decreased mortality. Further studies are needed to gain insight into the preferred timing, method, and duration of leukoreductive therapy in pertussis.

References 1.

Skoff TH, Hadler S, Hariri S. The Epidemiology of Nationally Reported Pertussis in the United States, 2000– 2016. Clinical Infectious Diseases. 2018;68(10):1634-1640. doi:10.1093/cid/ciy757

2. Berger JT, Carcillo JA, Shanley TP, et al. Critical Pertussis Illness in Children. Pediatric Critical Care Medicine. 2013;14(4):356-365. doi:10.1097/pcc.0b013e31828a70fe.

13. Oñoro G, Salido AG, Martínez IM, Cabeza B, Gillén M, Azagra AMD. Leukoreduction in Patients With Severe Pertussis With Hyperleukocytosis. The Pediatric Infectious Disease Journal. 2012;31(8):873-876. doi:10.1097/ inf.0b013e31825ba6cf. 14. Winter K, Zipprich J, Harriman K, et al. Risk Factors Associated With Infant Deaths From Pertussis: A Case-Control Study. Clinical Infectious Diseases. 2015;61(7):1099-1106. doi:10.1093/cid/civ472. 15. Mikelova LK, Halperin SA, Scheifele D, et al. Predictors of death in infants hospitalized with pertussis: a case-control study of 16 pertussis deaths in Canada. The Journal of Pediatrics. 2003;143(5):576-581. doi:10.1067/s00223476(03)00365-2.

3. Cherry JD. Pertussis in Young Infants Throughout the World. Clinical Infectious Diseases. 2016;63(suppl 4). doi:10.1093/cid/ciw550.

16. Palvo F, Fabro AT, Cervi MC, Aragon DC, Ramalho FS, Ana Paula De Carvalho Panzeri Carlotti. Severe pertussis infection. Medicine. 2017;96(48). doi:10.1097/ md.0000000000008823.

4. Cherry JD, Wendorf K, Bregman B, et al. An Observational Study of Severe Pertussis in 100 Infants ≤120 Days of Age. The Pediatric Infectious Disease Journal. 2018;37(3):202205. doi:10.1097/inf.0000000000001710.

17. Pierce C, Klein N, Peters M. Is leukocytosis a predictor of mortality in severe pertussis infection? Intensive Care Medicine. 2000;26(10):1512-1514. doi:10.1007/ s001340000587.

5. Desjardins M, Iachimov D, Mousseau S, et al. Clinical characteristics of pediatric pertussis cases, Quebec 2015–2017. Canada Communicable Disease Report. 2018;44(9):190-195. doi:10.14745/ccdr.44i09a01.

18. Rowlands H, Harrington K, Karimova A, et al. Impact of Rapid Leukodepletion on the Outcome of Severe Clinical Pertussis in Young Infants. Pediatrics. 2010;127(2). doi:10.1542/peds.2009-2860d.

6. Nieves DJ, Singh J, Ashouri N, Mcguire T, Adler-Shohet FC, Arrieta AC. Clinical and Laboratory Features of Pertussis in Infants at the Onset of a California Epidemic. The Journal of Pediatrics. 2011;159(6):1044-1046. doi:10.1016/j.jpeds.2011.08.010.

19. Centers for Disease Control and Prevention (CDC). Pertussis Outbreak in an Amish Community—Kent County, Delaware, September 2004–February 2005. JAMA. 2006;296(16):1960. doi:10.1001/jama.296.16.1960.


Paddock CD, Sanden GN, Cherry JD, Gal AA. Pathology and Pathogenesis of Fatal Bordetella pertussis Infection in Infants. Clinical Infectious Diseases. 2008;47:328-338.

8. Assy J, Séguéla P-E, Guillet E, Mauriat P. Severe Neonatal Pertussis Treated by Leukodepletion and Early Extra Corporeal Membrane Oxygenation. The Pediatric Infectious Disease Journal. 2015;34(9):1029-1030. doi:10.1097/inf.0000000000000781. 9. Carbonetti NH. Pertussis leukocytosis: mechanisms, clinical relevance and treatment. Pathogens and Disease. 2016;74(7). doi:10.1093/femspd/ftw087. 10. Kuperman A, Hoffmann Y, Glikman D, Dabbah H, Zonis Z. Severe pertussis and hyperleukocytosis: is it time to change for exchange? Transfusion. 2013;54(6):1630-1633. 11. Lashkari HP, Karuppaswamy S, Khalifa K. Pertussis-Related Hyperleukocytosis: Role of Hyperhydration and Exchange Transfusion. Clinical Pediatrics. 2011;51(10):987-990. doi:10.1177/0009922811410971. 12. Martinez M, Rochat I, Corbelli R, Tissières P, Rimensberger PC, Barazzone-Argiroffo C. Early blood exchange transfusion in malignant pertussis: A case report. Pediatric Critical Care Medicine. 2011;12(2). doi:10.1097/ pcc.0b013e3181f3a189.


20. Williamson G, Ahmed B, Kumar PS, Ostrov BE, Ericson JE. Vaccine-Preventable Diseases Requiring Hospitalization. Pediatrics. 2017;140(3). doi:10.1542/peds.2017-0298. 21. Vitek CR, Pascual FB, Baughman AL, Murphy TV. Increase in deaths from pertussis among young infants in the United States in the 1990s. The Pediatric Infectious Disease Journal. 2003;22(7):628-635. doi:10.1097/01. inf.0000073266.30728.0e 22. Scanlon K, Skerry C, Carbonetti N. Association of Pertussis Toxin with Severe Pertussis Disease. Toxins. 2019;11(7):373. doi:10.3390/toxins11070373. 23. Schenkel AR, Pauza CD. Pertussis toxin treatment in vivo reduces surface expression of the adhesion integrin leukocyte function antigen-1 (LFA-1). Cell Adhesion and Communication. 1999;7(3):183-193. doi:10.3109/15419069909010801 24. Viswanathan M, Selby DM, Ray PE. Expression of renal and vascular angiotensin II receptor subtypes in children. Pediatric Nephrology. 2000;14(10):1030. doi:10.1007/ s004670050067

Scholarly Research In Progress • Vol. 4, October 2020

Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats Albena Gesheva1â€

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 †Doctor of Medicine Program Correspondence: 1

Abstract Echolocation is the dominant sensory modality for threedimensional orientation in bats. However, bats also integrate different additional sensory modes such as olfaction and vision during flight, and especially during foraging. Studies in nectar-feeding bats have shown that they integrate olfaction to improve maneuvering in flight. But the extent to which bats use vision as a compensatory mechanism to echolocate is unknown. The objective of this study is to determine how behavior changes when both visual and echolocation information is available for orientation. The bat species Glossophaga soricina and Leptonycteris yerbabuenae were chosen for this study due to their distinct foraging behaviors. The difference in their dependency on vision and echolocation makes them ideal models. The prediction was that with lower light intensity, echolocation calls will be more frequent during approach towards an object and with more light available, the call sequences should be less frequent, indicating more reliance on visual cues. Since echolocation is an active and energy-consuming sense, using passive visual perception when available would be a more cost-effective approach. Although it seems that more light availability and visual cues encourage the bats to forage, operating in light environments jeopardizes their survival from natural predators. Navigating between more energy-effective foraging and less protection from predators is a challenge and no obvious approach prevails.

Introduction New world leaf-nosed bats (Phyllostomidae) are found in Central and South America. They have a large variety of food sources, comprising of insects, fruits, nectar, and small vertebrates. A recent study on nectar-feeding bats showed that different modalities like olfaction and echolocation are integrated to improve maneuvering in flight (1). But how echolocation and vision work together under different light conditions has not been previously studied. Phyllostomid bats are generally narrow-space gleaners and use a multitude of sensory cues while foraging. Their passive sense of olfaction is usually used for long distance orientation, for example to locate a flowering or fruiting tree. Their active sense of echolocation is used for collision avoidance and close distance orientation. For example, the acoustic properties of flower structures can provide more detailed information about the flower position on the plant for nectarivorous bats (1). Muchhala & Serrano show that nectarivorous bats rely on certain sensory methods during foraging depending on the complexity of the background (2). They presented two nectar bats, Anoura caudifer and A. geoffroyi, with scented and unscented flowers in simple

and complex environments. They found that in a simple environment, the bats chose flowers indiscriminately. However, in a complex environment, the scented flowers were located first. This suggests that for exposed flowers, vision and echolocation are sufficient for detection, but when they are concealed, the bats rely on olfaction to locate them. Another example of a multimodal sensory system is found in Leptonycteris yerbabuenae. Here a combination of olfaction and echolocation cues elicited a higher approach response than each individual cue, suggesting that nectar bats integrate various sensory modes to detect and precisely localize open flowers (1). On the other hand, insectivorous bats may utilize vision to locate large objects, such as vegetation, and echolocation to target small prey. Vision distinguishes faraway landscapes while echolocation can classify and precisely locate an object, making it important for closerange localization of targets (3). This creates a multimodal sensory system of echolocation (active hearing), scent, passive hearing, and vision, which contributes both to orientation and to the detection and localization of food items. Experiments in Old World bats show that they use echolocation in both light and dark conditions (4). Individuals of the migratory flower-visiting bat species Leptonycteris yerbabuenae often fly in open habitats with relatively high light availability from moonlight (5). By contrast, individuals of Glossophaga soricina forage for flowers in narrow spaces inside the forest where vegetation eliminates most light (6). Interestingly, these bats are color-blind but capable of ultraviolet light perception, possibly as an adaptive function to enhance the contrast of ultraviolet-light-reflecting flowers during low light, thereby increasing flower search efficiency (7). Echolocation was refined to detect small targets in conjunction with using vision, revealing a nocturnal niche for bats and inciting diversification in echolocation and foraging tactics (3). Because echolocation is a strategy for close-distance orientation and not suitable for long-distance orientation, individuals of the migrating L. yerbabuenae presumably rely more on vision than other phyllostomid species (such as G. soricina), which could result in a better flight performance in more complex environments when light is available. Echolocation is the dominant sensory modality for threedimensional orientation in bats. However, bats also integrate different additional sensory modes such as olfaction and vision during flight and especially during foraging. Echolocation is constrained by the attenuation of high-frequency sounds and echoes from the background, called clutter. Therefore, echolocating bats must rely on other sensory cues in many situations where more clutter is present. Some Old-World fruit bats have practically lost echolocation as an orienting sense, mainly relying on vision for orientation 119

Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats

and foraging (8). Microchiropteran bats are known to use vision for long-range orientation, when echolocation is not very effective. Past experiments with the brown long-eared bat have shown that they preferred putting more importance on visual information rather than sonar cues when both sonar cues and visual cues are available (9). Recent experiments also show that bats usually do not echolocate when leaving their roost but rather use vision, avoiding obstacles like mist nets during highly moonlit nights (10). Study aims Bats play a critical role in pollen distribution, seed dispersal, and pest control, but it is not clear how they integrate visual information in their foraging behaviors. Bats primarily orient themselves by echolocation, but evidence suggests that they also utilize vision when light is available, regulating their biosonar based on the availability of visual information as a method of energy conservation (11). The extent to which bats use vision as a compensatory mechanism to echolocation is unknown. In their natural habitat, the primary light source is the moon, and light intensities vary depending on the moon cycle and vegetation cover. Understanding how light availability and intensity alters the bats’ behavior can give insight into how different parameters of the environment are integrated into a multimodal sensory system. The objective of this study was to determine how behavior changes when both visual and echolocation information is available for orientation. To investigate whether vision affects echolocation, I tested the hypothesis that with enough light available, nectar-feeding bats will rely both on vision and echolocation to improve foraging efficiency. Since vision is a passive form of orientation, it will reduce the effort spent on the active (energy-consuming) echolocation sense, changing the amount of echolocation calls per time unit. When maneuvering in a more complex arrangement, light availability could influence flight performance. The bat species Glossophaga soricina and Leptonycteris yerbabuenae were chosen for this study due to their distinct foraging behaviors. The difference in their dependency on vision and echolocation makes them ideal models. I tested their ability to orient to their environment under various simulated moonlight intensities and in total darkness. From the acoustic data, I calculated temporal call parameters. The prediction was that with lower light intensity, echolocation calls during approach will be more frequent; with more light available, the call sequence should be less frequent. The results will determine whether bats’ vision influences echolocation behavior and facilitate a more in-depth understanding of their multimodal sensory system.

of the mezcal plant (Agave angustifolia) in Mexico (13). Bats involved in this study were held in a species-appropriate husbandry at Ulm University (Ulm, Germany) and fed a honeywater and pollen diet. Typical day and night rhythm were simulated in the husbandry, using a time switch to adjust their activity rhythm to more common working hours. Experimental setup To test the bats’ echolocation behavior during foraging under different light conditions, individuals were isolated from the colony, three or four at a time, and placed in a climate chamber measuring 4 m x 2.4 m x 2 m including resting places (two inversely hanging buckets with a rough cloth) and feeders. Room conditions were adjusted to about 24°C and 80–90% relative humidity. In total, 20 bats were tested: 5 adult females and 5 adult males of the species G. soricina and 5 adult females and 5 adult males of the species L. yerbabuenae. One additional female G. soricina was included in the study but performed only the simple environment experiments. After the animals were captured, they were transferred in a cotton bag into the climate chamber and given 24 hours to habituate prior to experiments. The chamber was completely dark. To apply different light conditions, 6 small 10-watt halogen lamps (Goobay: 10W/12V, G4, 2000h,140 lumens, 8x31, 2650k warm white) in parallel circuit were connected to a voltmeter that allowed voltages of 0 V, 10 V, 15 V, and 20 V to be set. Due to these three voltages, various light conditions during dusk or dawn could be simulated. A feeder on a stand offered a solution of Alete® milk powder and honey water (with a sugar content of 18%) to L. yerbabuenae and a Nektar-Plus (17% sugar concentration, Nekton®) solution to G. soricina. Respectively, these were their preferred diet in the husbandry. The bats could feed ad libitum. To record the echolocation calls, a custom-made ultrasound microphone with a flat frequency response (±3dB between 18 and 100 kHz) was connected to an AvisoftUltraSoundGate. Recordings of a minimum of 10 seconds were made using a pre-trigger of 5 seconds and a hold time of 5 seconds, and a sampling rate of 480 kHz. The ultrasound microphone was fixed to a tripod above the feeder as parallel as possible to the direction of the feeder. This captured the calls sent out when approaching the feeder. To avoid spatial learning, the feeder tripod was presented in four different positions in random sequence (Figure 1). The positions were

Methods Study species The small nectar-feeding bat Glossophaga soricina ranges from northern Mexico to Paraguay and northern Argentina. It lives in a wide variety of habitats, including tropical rainforests and savannas. This species feeds mainly on nectar but can include insects, fruits, and pollen in its diet (12). The mediumsized Leptonycteris yerbabuenae, another nectar-feeding bat, ranges from southern Arizona and New Mexico to Honduras and El Salvador. They live in thorn scrub and deciduous forests, and their range corresponds closely to the distribution 120

Figure 1. Simple environment setup.

Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats

changed so that the height of the feeder above ground was consistent and kept the same distance to the halogen lamps. The distance between the microphone and the feeder remained unchanged. An infrared-sensitive video camera (SONY Handycam, Japan) was set on a tripod directed at the feeder to monitor the bat approaching, hovering, or drinking. Cables of the camera and the microphone were led outside the room through a hole in one wall which was clogged by a big piece of cloth to prevent the diffusion of external light. To simulate a cluttered environment, two plastic garbage bags were cut into 30 strips of 5cm wide strips, roughly 50 cm long, and hung parallel to the corner 10 cm apart from each other and offset by 5 cm (Figure 2).

connected to the SONY camera, inserting its snout into the feeder. This was categorized as a drinking event. At least 5 successful feeding events were recorded for every light condition, environment type, and feeder position for each bat. The sequences were divided into orientation and search events before drinking. The decrease of echolocation calls per time unit and their decrease in intensity were defined as search sequences. The less organized sequences of the call were classified as orientation. The more organized and patterned sequences were classified as search. Approach calls of different light conditions were compared among each other. Data analysis Measuring sequence parameters Recordings were analyzed with the software Avisoft SASLAB PRO v. 5.2.07 (R. Specht, Avisoft Bioacoustics, Glienicke, Germany) using Window Hamming, FFT 256, y-scale enlargement 2, frame % 100, resolution 1953 Hz (Figure 3).

Figure 3. FB CALL2017-12-15_17-21-47_545 20V: Call example traversing the simple environment.

Figure 2. Cluttered environment setup.

Data sampling Video recordings Videos were taken alongside audio recordings and ranged from 10 minutes to over an hour, depending on the bat’s motivation. The visual information from the cameras verified that an approach was successful (when the bat’s tongue or snout was inserted into the feeder), giving the cue to manually trigger the audio. The videos themselves were not analyzed for data. Audio recordings To observe the bats’ echolocation behavior, an audio recording was triggered when the bat was seen on a monitor

All temporal parameters were measured from the oscillogram to improve accuracy and precision, and checked using the spectrogram with a magnification window of about 5 seconds. The call sequences were analyzed starting from the end of the sequence (last call) toward the beginning to avoid echoes and increase measurement accuracy and precision. Measurements were taken from the beginning of the first pulse of a group to the beginning of the first pulse of the following group. Any pulse or group number is counted from the terminal pulse or group as the starting point. Measurements were taken from the audio files recorded using the parameters below: • Approach was defined as a successful drinking attempt towards the feeder. • Sequence Start Time (seconds) was defined as the apparent start of the approach toward the feeder, distinguished from the waveform, either by the appearance of pulses or, if connected to a longer burst of calls, to the first clear pulse shown on the spectrogram that looks as if it relates to the other calls in an orientation effort. • Orientation Sequences were defined as when the distance of the groups is visually different from the search calls and usually consistently spaced with each other. • Search Sequences were defined from when the group becomes distinguished from previous groups by spacing or grouping (G9). 121

Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats

• Time Spent Orienting (seconds) was defined as the distance from the start of sequence to where the pulse splits (sequence from left to Pulse 27). • Search Sequence Time (seconds) was defined from where the pulse split appears to the end of the sequence (Pulse 27–1) • Pulse Interval (seconds) was defined as the distance from the previous pulse. • The Group Interval (seconds) was defined as the distance from the previous group. • Pulse Split demarcates the beginning of the search phase and was defined as the point at which the group becomes distinguished from previous groups by a few characteristics varying between individuals: whenever a hooked call appears (90% of sequences), the second harmonic elongates and becomes wavy, the harmonics overlap more, or there is a single isolated pulse (Pulse 27/Group 9). This term was also interpreted as the bat’s intent to attempt a drinking event. • Pulse ID was the number of the pulse within the group. • Number of Pulses is defined as the total amount of pulses between where the pulse split occurs and the end of the sequence (Pulse 27-1). • Group ID was the number of the group starting from the terminal group. • Number of Groups was counted from the terminal group to where the pulse split appears. • Feeding Buzz (Figure 4) was defined as any sequence that ended with a group containing 6 or more pulses with pulse intervals below 015s.

Figure 4. FB CALL2017-12-13_18-32-12_486 20V: Call example traversing the cluttered environment

Statistical analysis The program R was used to perform statistical analysis. The first 7 or 8 approaches were used to determine statistical significance, excluding any feeding buzzes to create consistency and improve accuracy. I used a paired t-test to determine a significant p-value and a Welch’s two sample t-test to confirm significance and account for any variance in the population.

Results To study whether L. yerbabuenae and G. soricina change their echolocation behavior while foraging, calls of the feeding 122

Figure 5. Sequences (N~100 sequences) between two light conditions (0 V, shaded gray, and 20 V) and two environments (left simple, right/ patterned cluttered) were compared from the same 10 individuals (5 male, 5 female) of L. yerbabuenae. The data was collected from 05/12/17 to05/02/18 at Ulm University. Three parameters were measured: Duration of Search Sequence, Location of the group containing the Pulse Split, and number of groups. Paired T-ttst determined significance: A) Simple (t = 3.04, df = 32, p-value = 0.0047) versus Cluttered (t = 4.02, df = 32, p-value = 0.0003). B) Simple (t = 3.0, df = 31, p-value = 0.005) versus Cluttered (t = 3.44, df = 32, p-value = 0.0017). C) Simple (t = 2.22, df = 31, p-value = 0.03354) versus Cluttered (t = 3.90, df = 32, p-value = 0.0005).

events were recorded under different light conditions. Several variables suggest that when L. yerbabuenae and G. soricina can use vision in foraging, they adjust their echolocation both in time and composition. Both species take significantly longer to search for the feeder when approaching in the dark versus the lighter environment (Figure 5A, Figure 7A). L. yerbabuenae show a statistical difference of a t-value of 3.04 with a degree of freedom of

Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats


Figure 6. Five female and 5 male L. yerbabuenae were tested from 05/12/17 to 05/02/18 at Ulm University under two conditions (0V and 20V). The types of groups and their occurrence from the first 5 approaches were compared (30,34 sequences under 0V and 38,36 sequences under 20V) between the simple condition (t = 1.2451, df = 8, p-value = 0.2483) and the cluttered condition (t = 1.7802, df = 9, p-value = 0.1087).

32 and a p-value of 0.0047 in the simple environment and a t-value of 4.02 with a degree of freedom of 32 and a p-value of 0.0003 in the cluttered environment. G. soricina show a statistical difference of a t value of 4.8013 with a degree of freedom of 52 and a p-value of 0.00001375 in the simple environment and a t-value of 4.6392 with degree of freedom of 50 and p-value of 0.00002551 in the cluttered environment. Both species produce more groups during their approaches in the dark, and significantly less groups when there is more light available (Figure 5B, Figure 7B). This is true both in the simple and cluttered environment. In the simple environment, L. yerbabuenae has a t-value of 3 with a degree of freedom of 31 and a p-value of 0.005, and in the cluttered environment has a t-value of 3.44 with a degree of freedom of 32 and a p-value of 0.0017. G. soricina show a statistical difference of a t-value of 6.9223 with a degree of freedom of 52 and a p-value of 0.000000006583 in the simple environment and a t-value of 3.3061 with a degree of freedom of 50 and a p-value of 0.001756 in the cluttered environment. The sizes of the groups do not seem to show significance and remain the same (Figure 6, Figure 8). This suggests that when the bat can use more of its vision, it does not need to search as much with echolocation, at least initially. Both species begin their search call echolocation (or change their call types) significantly earlier in the approach during a dark environment versus a lighter environment (Figure 5C, Figure 7C). L. yerbabuenae show a statistical difference of a t-value of 2.22 with a degree of freedom of 31 and a p-value of 0.03354 in the simple environment and a t-value of 3.90 with a degree of freedom of 32 and a p-value of 0.0005 in the cluttered environment. G. soricina show a statistical difference of a t-value of 5.2764 with a degree of freedom of 49 and a p-value of 0.000002978 in the simple environment and a t-value of 4.1611 with degree of freedom of 49 and p-value of 0.0001274 in the cluttered environment.

Significant differences in the search sequence time, the time at which the bat starts searching (signified by the Pulse Split), and the number of groups in the search sequence show that the echolocation behavior during foraging in both L. yerbabuenae and G. soricina changes during an environment with light (20V) when compared to total darkness (0V), supporting the hypothesis that these bats rely more on vision when there is more light available while foraging. These results imply that visual cues during dusk or dawn might help the bats to localize flowers more quickly than with echolocation alone. Since echolocation is an active and energy-consuming sense, using passive visual perception when available would be a more cost-effective approach. For many animals, energy is a limiting factor. This is especially true for bats, as they do not build up fat tissue as energy reserve and their foraging flights are very energy-consuming.

Statistical significance found in the simple environment was often amplified in the cluttered environment. L. yerbabuenae exhibit a significant difference in the time spent during search flight both in the simple environment (p-value = 0.0047) and in the cluttered environment (p-value = 0.0003). Comparison in G. soricina also shows similar statistical significance in the simple environment (p-value = 0.00001375) and in the cluttered environment (p-value = 0.00002551). This demonstrates that the bats spent significantly more time in their search flight in a dark environment than in a light environment. The difference suggests that the bats were getting the required spatial information more quickly when visual cues were available. The group in which the pulse type changes (Pulse Split) was compared showed statistical significance in L. yerbabuenae both in the simple environment (p-value = 0.005) and in the cluttered environment (p-value = 0.0017). Experiments with G. soricina also showed similar statistical significance both in the simple environment (p-value = 0.000000006583) and in the cluttered environment (p-value = 0.001756). The temporal location where this one pulse type splits into another pulse type was interpreted as where the bat had sufficient information to attempt to approach the target. The change in pulse type occurred significantly earlier when visual cues were available as opposed to when they were not. This could mean that the bats were sure of their target much earlier in the light environment, if this change in pulse type is interpreted as the start of a decisive approach. The number of pulse groups that the bats produced during the search phase of their foraging flight shows a statistical significance in L. yerbabuenae both in the simple environment (p-value = 0.03354) and in the cluttered environment (p-value = 0.0005). Experiments with G. soricina also showed similar statistical significance in the simple environment (p-value =


Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats

Figure 8. Five female and 5 male G. soricina were tested from 01/02/2018 to 01/24/2018 at Ulm University under two conditions (0V and 20V). The types of groups and their occurrence from the first 5 approaches were compared (30,34 sequences under 0V and 38,36 sequences under 20V) between the simple condition (t = 0.8353, df = 5, p-value = 0.4416) and the cluttered condition (t = 1.2585, df = 7, p-value = 0.2486).

Figure 7. Sequences (N=100) between two light conditions (0 V, shaded gray, and 20 V) and two environments (left simple, right/ patterned cluttered) were compared from the same 10 individuals (5 male, 5 female) of G. soricina (*6 females for the cluttered environment). The data was collected from 05/12/17 to05/02/18 at Ulm University. Three parameters were measured: Duration of Search Sequence, Location of the group containing the Pulse Split, and number of groups. Paired t-test determined significance: A) Simple (t = 4.8013, df = 52, p-value = 0.00001375) versus Cluttered (t = 4.6392, df = 50, p-value = 0.00002551). B) Simple (t = 6.9223, df = 52, p-value = 0.000000006583) versus Cluttered (t = 3.3061, df = 50, p-value = 0.001756). C) Simple (t = 5.2764, df = 49, p-value = 0.000002978) versus Cluttered (t = 4.1611, df = 49, p-value = 0.0001274).

0.000002978) and in the cluttered environment (p-value = 0.0001274). Their inclination to produce less groups and less ultrasound calls to locate a potential food source suggests that, even from a distance, they are acquiring enough visual cues to distinguish the location accurately. This accounts for their ability to spot obstacles like trees as potential food sources from a distance.


Although it seems that more light availability and visual cues encourage the bats to forage, operating in light environments jeopardizes their survival from natural predators. The bats exhibited significant individual differences in their echolocation and foraging behavior, as exemplified by the large range in the data. Although significant results were shown despite this, a larger sample size of bats and species would illustrate a better picture of the data and behavior. To obtain distinct calls, further experiments should eliminate more variables such as directionality toward the feeder and microphone, since the subjects were untrained and allowed to fly and forage ad libitum. To improve precision and simplify data analysis, future experiments should record the approaches with a high-speed infrared camera. Additionally, studies involving the neurology of these bats would provide insight into their visual and echolocation perceptions. Light intensities of 10 V and 15 V were also tested but preliminary data analysis was not promising, and further analysis was not pursued. These lower light intensities might be under the neuronal potential threshold and not sufficient to elicit a response in the photoreceptors. Therefore, eliminating more variables could show differences even in lower intensity light.

Acknowledgments I would like to thank my supervisor apl. Proffessor Dr. Marco Tschapka for his support throughout and constructive suggestions. I would like to thank Gloria GlĂźcksklee for her guidance and unwavering patience throughout this project. I would also like to thank the staff, students, and educators at Ulm University Institute of Evolutionary Ecology and Conservation Genomics for welcoming me with warmth and allowing me to use their facilities.

Effect of Light Intensity on Echolocation Behavior in Phyllostomid Bats

References 1.

Gonzalez-Terrazas T, Martel C, Milet-Pinheiro P, Ayasse M, Kalko E, Tschapka M. Finding flowers in the dark: nectar-feeding bats integrate olfaction and echolocation while foraging for nectar. Royal Society Open Science. 2016;3(8):160199.

2. Muchhala N, Serrano D. The Complexity of Background Clutter Affects Nectar Bat Use of Flower Odor and Shape Cues. PLOS ONE. 2015;10(10):e0136657. 3. Boonman A, Bar-On Y, Yovel Y, Cvikel N, 2013. It's not black or white—on the range of vision and echolocation in echolocating bats. Frontiers in Physiology, 4. 4. Waters DA, Vollrath C. Echolocation Performance and Call Structure in the Megachiropteran Fruit-Bat Rousettus aegyptiacus. Acta Chiropterologica. 2003;5(2):209. 5. Kalko EKV, Handley CO, Handley D. Long-term Studies in Vertebrate Communities. Los Angeles, CA: Academic Press; 1996. 6. Denzinger A, Schnitzler H. Bat guilds, a concept to classify the highly diverse foraging and echolocation behaviors of microchiropteran bats. Frontiers in Physiology. 2013;4. 7.

Winter Y, López J, von Helversen O. Ultraviolet vision in a bat. Nature. 2003;425(6958):612-614.

8. Kulzer E. Flughunde erzeugen Orientierungslaute durch Zungenschlag. Die Naturwissenschaften. 1956;43(5):117118. 9. Rydell J, Eklöf J. Vision complements echolocation in an aerial-hawking bat. Naturwissenschaften. 2003;90(10):481483. 10. Wang D, Oakley T, Mower J, Shimmin L, Yim S, Honeycutt R, Tsao H, Li WH. Molecular Evolution of Bat Color Vision Genes. Molecular Biology and Evolution. 2003;21(2):295302. 11. Danilovich S, Krishnan A, Lee W, Borrisov I, Eitan O, Kosa G, et al. Bats regulate biosonar based on the availability of visual information. Current Biology. 2015;25(23):R1124-R1125. 12. Fox D. Glossophaga soricina (Pallas's long-tongued bat) [Internet]. Animal Diversity Web. 1999 [cited 20 May 2018]. Available from: Glossophaga_soricina/ 13. Arita H. Spatial Segregation in Long-Nosed Bats, Leptonycteris nivalis and Leptonycteris curasoae, in Mexico. Journal of Mammalogy. 1991;72(4):706-714


Scholarly Research In Progress • Vol. 4, October 2020

Exploring the Role of APOE4 Genotype in Chemo Brain or Chemotherapy-Induced Cognitive Impairment Tracy L. Chofor1*

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Chemo brain, also known as chemo fog or chemotherapyinduced cognitive impairment, refers to the neuropsychological decline that about 75% patients experience during cancer treatment. It has become an increased area of research interest, because as more patients survive cancer, it is increasingly evident that available treatment methods affect their cognitive function, and subsequently their quality of life. While the biological mechanisms for this phenomenon are still under investigation, there is increasing evidence that the apolipoprotein genotype, specifically the APOE ε4 allele, may be a potential risk factor. APOE ε4 has been identified as the major genetic risk factor in the pathogenesis of Alzheimer’s disease, which is also characterized by decline in memory and other cognitive function. Recent research suggests that the APOE ε4 allele, when exposed to chemotherapeutic agents, can impair cognitive behavior and alter brain structure. Particularly, impairment in spatial memory and reduction of hippocampal volume was observed. This review examines and critically analyzes pre-clinical and clinical studies that test the theory that the APOE ε4 allele increases susceptibility to chemotherapy-induced cognitive impairment. Better understanding the role of APOE ε4 in cognitive decline will allow for further research to be conducted on different chemotherapeutic agents as well as the development of preventive measures for APOE ε4 carriers.

Introduction Chemotherapy-induced cognitive impairment (CICI), also termed “chemobrain” or “chemo fog” by cancer patients, describes the cognitive dysfunction that results from chemotherapeutic agents often used to treat cancer (1). As cancer-related survival rates increase, the prevalence of CICI has increased as well (2), with up to 75% of cancer patients experiencing CICI during treatment, and 35% still reporting symptoms post treatment (3). Some of the symptoms that are commonly observed include a deficit in cognitive domains related to visual-spatial ability, processing and motor speed, memory, concentration and executive function (1). Subsequently, changes are also seen in the brain regions that modulate these cognitive skills, such as the prefrontal cortex, frontal and temporal lobes, and the hippocampal and cortical regions (1, 4). Assessing and diagnosing CICI has proven to be challenging due to the sometimes transient nature of the symptoms, as well as the extensive number of ways in which cognitive function can be measured (3). While there are no official guidelines to clinically assess CICI (1), the International Cognition and Cancer Task Force (ICCTF) recommends a neuropsychological assessment battery consisting of the 126

following measures: the Trail making Test (TMT) to assess the psychomotor speed and executive function, the Hopkins Verbal Learning Test-Revised (HVLT-R) to assess verbal memory and delayed recall, and the Controlled Oral Word Association (COWA) Test to assess speeded lexical fluency and executive function (3, 5, 6). Other supplemental measures include the Wechsler Adult Intelligence Scale (WAIS) and the Paced Auditory Serial Addition Test (PASAT) (5, 6). The etiology of CICI is still under investigation, but several characteristics including demographic variables, comorbidities and biological factors, all play a role in the pathophysiology of the disease (1, 3, 5). Age is a known risk factor for cognitive decline, which can also be exacerbated by premorbid psychological issues such as anxiety and depression. Inflammation, altered blood-brain barrier, and cytokine levels have also been associated with CICI (1, 5). Studies investigating the apolipoprotein E (APOE) and catechol-o-methyltransferase (COMT) genes showed a correlation between these genes and a higher susceptibility to cognitive deficit, especially in patients who received chemotherapy treatment (3, 5). The APOE ε4 allele in particular has been linked to several diseases characterized by cognitive dysfunction, as well as poor cognitive outcomes in other neurological conditions like stroke (3). APOE ε4 genotype APOE is a glycoprotein that regulates the metabolism of cholesterol and fatty acids and is made up of 299 amino acids (7). APOE levels are typically highest in the liver and the brain, where it is synthesized by astrocytes and sometimes neurons (7). The APOE gene consists of different polymorphisms with three main alleles — APOE ε2, APOE ε3, and APOE ε4 (7) — which can allow for up to six genotype combinations (8). Research has shown the APOE ε4 genotype to be a risk factor for the development of Alzheimer’s (1, 7, 8), traumatic brain injury (3, 8) and cerebral amyloid angiopathy (CAA) (7). Other studies conclude that APOE ε4 has neurotoxic effects that can manifest as behavioral deficits in animals (9). The exact mechanism of the APOE ε4 allele is still under investigation, but researchers hypothesize that the ε4 allele lacks in its ability to conduct neuronal repair compared to the other alleles (10). Several human and mouse studies have been conducted to understand the role APOE ε4 plays in CICI, but further research is needed to establish the genotype as a definite risk factor for CICI.

Methods A literature review was performed using Google Scholar, PubMed and ClinicalKey. Search terms included chemobrain, chemofog, chemotherapy-related cognitive impairment,

Exploring the Role of APOE4 Genotype in Chemo Brain

chemotherapy-induced cognitive impairment, APOE, apolipoprotein E, CRCI and CICI. Articles were selected based on impact factor, investigation relevant to CICI and APOE, and a publication date within the last 20 years.

Discussion APOE ε4 in clinical studies Over the years, the acute and long-term cognitive changes reported by cancer survivors treated with chemotherapy were attributed to fatigue and other psychological factors (10). However, after controlling for the aforementioned factors, recent literature presents evidence suggesting that these changes are in fact linked to chemotherapy treatment (10, 11, 12, 13). These studies also show that other risk factors like age, pretreatment cognitive reserve and the APOE ε4 allele play a role in the extent of cognitive decline (12, 11, 14). The longitudinal studies conducted assessed visual memory, spatial ability and psychomotor functioning (8), verbal fluency and memory (13), learning and memory domain and cognitive function (12), as well as processing speed and working memory (11, 14). While examining the detrimental effects of chemotherapy on cognitive function of cancer survivors, Ahles et al. concluded that APOE ε4 carriers are more likely to experience more severe levels of change in cognition for longer periods of time (8, 10). After assessing different neuropsychological domains, individuals with the APOE ε4 allele performed significantly worse than those lacking the ε4 allele in the visual memory and spatial ability domains, with p values of p<0.03 and p<0.05 respectively (8). There was also a substantial difference in psychomotor functioning across the two APOE statuses, although not statistically significant (8). Findings by Mandelblatt et al. also demonstrated a larger decline in the memory domain of APOE ε4 carriers versus other treatment groups (12). It also took twice as long for APOE ε4 carriers to show improvements in these domains compared to other genotypes (12). In the attention, processing and executive function domain, the interaction of age and APOE status on chemotherapy was evident as the survivors with lowest scores in this domain were older survivors with the APOE ε4 allele (12). Of the limited number of clinical studies examining the effects of APOE ε4, most show its association to cognitive decline in the context of age and specific chemotherapeutic regimens, but the current evidence is not sufficient to establish a direct correlation between the APOE ε4 allele and worse CICI outcomes. Other chemotherapeutic agents need to be studied individually as well as in conjunction with other agents or as part of common regimens. Additionally, patient reports and neuropsychological batteries provide information of cognitive effects but are limited in the inability to suggest actual biological processes and mechanisms of action. Thus, animal studies are necessary to analyze these mechanisms and structural changes associated with APOE ε4 in CICI. Effect of APOE ε4 in mouse models Human APOE knock-in mouse model and CICI Data supporting APOE as a risk factor for CICI and the

lack of reports establishing that chemotherapy-related cognitive decline is directly due to chemotherapeutic agents administered to cancer patients, led researchers to develop a human APOE knock-in mouse model to understand and define the mechanism of APOE in CICI (4). Previous APOE knock-in mouse models have shown that APOE ε4 mice exhibit higher levels of inflammatory cytokines than APOE ε3 mice, as well as deficits in dendritic spine density (4, 15). Biochemical differences in APOE ε4 and APOE ε3 mice are evident, and researchers suggest that this may be due to differences in post-translational modifications (15). However, Speidell et al. sought to explain the how these genotypes interact in the presence of cancer, as well as during the administration of a therapeutic agent. Thirty-nine human knock-in C57BL/6J mice with APOE ε4 (n=24) and APOE ε3 (n=18) were selected and maintained according to the ethical and national/international guidelines for animal welfare (4). The experimental group of mice was treated with doxorubicin, a common chemotherapeutic agent used for breast cancer. The control group remained untreated. Both groups were assessed with the use of behavioral assays and their brains analyzed via magnetic resonance imaging (MRI). After the behavioral assay, the mice were euthanized and their brains extracted to conduct an MRI study. This study was completed on 12 control animals with an even APOE ε3 and APOE ε4 distribution as well as 5 APOE ε3 and 6 APOE ε4 treated mice (4). Voxel-based morphometry (VBM) was then used to evaluate the discrepancies in brain anatomy between the treated group versus the untreated group. These anatomical changes were assessed with a T2weighted sequence. Data acquired via the MRI study enabled researchers to analyze difference in grey matter maps across both groups (4). When both genotypes were compared with respect to distance traveled in the open field, rearing events, and amount of time spent in the open arm of the maze, there was no substantial difference between the control APOE ε4 and APOE ε3 mice (4). This suggests that inherently, there were no major differences in locomotion, exploration, and anxiety between mice with APOE ε4 versus APOE ε3. These three parameters were also tested at baseline in mice that were treated with doxorubicin a week prior to doxorubicin administration. A decline was evident 1 week post-treatment as the distance traveled and number of rearing events in the open field dropped to almost half that of baseline, in both treatment groups (4). This demonstrates the negative effect of a single dose of doxorubicin on exploration. After the second dose of doxorubicin at the 6-week mark, APOE ε3 show improved exploratory levels, almost similar to baseline. APOE ε4 mice on the other hand, exhibited even worse levels compared to baseline (4). No substantial differences were observed in anxiety levels when compared across both genotypes. The most notable differences between APOE ε4 and APOE ε3 mice were found in the cognitive behavior trials. Spatial learning and memory were assessed in the treatment group 5 weeks after doxorubicin administration, and APOE ε4 mice showed a much longer latency period than APOE ε3 mice (4). Although the APOE ε4 latency period decreased as the training days progressed, there was still an almost 10-fold 127

Exploring the Role of APOE4 Genotype in Chemo Brain

difference in the latency period (100 secs) compared to APOE ε3 (13 secs) with a p value of <0.001 (4). This variation was also evident in the 72-hour probe trial, suggesting that the poor performance noted during the training days affected their ability to learn and retain information on how to locate the target hole. It is evident that doxorubicin treated APOE ε4 mice showed the longest latency periods to locate the escape hole of all groups; however, because the control APOE ε4 mice also showed deficits in spatial learning abilities compared to control APOE ε3 mice, this suggests that the APOE ε4 genotype is a risk factor for spatial learning and memory. MRI results showed smaller grey matter (GM) volumes in control APOE ε4 mice compared to control APOE ε3 mice. The difference in GM volumes was visible especially on the hippocampus and certain regions of the striatum (4). In the doxorubicin treatment group, imaging showed decreased GM in even more regions of the brain, including the lateral geniculate nucleus and the piriform cortex (4). VBM analyses of the MRI studies comparing the untreated versus the doxorubicin-treated brains revealed that even in APOE ε3 mice there were substantial differences and alterations in GM within the frontal cortex and hippocampus (4). These patterns were also visible in the APOE ε4 mice, but on a larger scale (4). These data indicate that although both the APOE ε4 and APOE ε3 mice showed learning impairment post-treatment, APOE ε4 mice had a significantly higher magnitude of impairment. Additionally, structural changes such as reduced GM in areas like the hippocampus were evident in both the doxorubicin treated APOE ε4 and APOE ε3 mice but more so in the APOE ε4 mice. This model, if further studied, can be replicated to take into account other factors that could affect the mechanism of APOE ε4 on CICI, including but not limited to age, administration of chemotherapy regimens as seen in clinical practice, and eventually longitudinal clinical trials in cancer patients. Importantly, this study did not take into account the implications of the administration of doxorubicin in clinical settings, which may be administered at varying doses over a long period of time or concurrently with other therapies. This is important because other factors have been shown to exacerbate CICI, such as hormone therapy (1, 4), or potentially improve cognitive function, like smoking (11). Additionally, non-oncogenic mice were used as the model in this study, and while it was successful in establishing an association between APOE ε4 and CICI, replicating the study in oncogenic mice would be more representative of the interactions and mechanisms between cancer and APOE ε4. Finally, the data in the report were presented using the standard error of the mean (SEM), which does not necessarily depict the variability of the data points and does not accurately establish statistical significance. Using standard deviation (SD) would be more appropriate as a descriptive statistic, as it is more accurate in showing the variability of the sample (16). Human APOE knock-in mouse model and CICI in the context of aging As stated above, it would be important to use the APOE knock-in mouse model to study APOE ε4 together with aging as risk factors for CICI. The study was in fact replicated by the same group of researchers to account for age in the breast


cancer survivor population by using aged mice and assessing other cognitive areas. The C57BL/6J human APOE knock-in mouse model was used with a bigger sample size of 61 that included 30 APOE ε3 and 31 APOE ε4 mice (17). The two APOE ε4 and APOE ε3 control groups received saline, and the APOE ε4 and APOE ε3 treatment groups received doxorubicin (17). Some of the behavioral assessments used for this experiment were similar to those performed previously, but the set-up of the assays differed. These tests included the open field test, elevated zero maze, and the Barnes maze. Also, two supplemental assays included pre-pulse inhibition (PPI) and fear conditioning. After about 31 to 35 weeks post-treatment, the mice were euthanized so that tissue and blood samples could be collected from healthy mice. The brains that were perfused with phosphate buffered saline (PBS) were then prepared and processed for immunohistochemistry and staining with antibodies (17). Those perfused with formalin were used for the MRI study, which was performed and assessed with a T2weighted sequence. VBM was performed as well to assess GM and white matter (17). Results from this experiment indicated weight loss in the doxorubicin treatment groups but showed no substantial difference across genotypes (17). At baseline, APOE ε3 mice spent 50% more time at the open zones of the open field and elevated zero maze than APOE ε4 mice, pointing to less levels of anxiety in the APOE ε3 groups compared to APOE ε4. Two weeks after treatment, both doxorubicintreated and saline control groups traveled the same distance in both assessments that evaluated anxiety and exploration (17). Fifteen weeks post-treatment, all groups showed similar distance travelled in both apparatuses as well. Also, all groups showed a reduction in the amount of time spent in the exposed zones of the elevated zero maze post treatment (17); therefore, according to this study, the chemotherapeutic agent has minimal effect on locomotion and anxiety. The PPI data showed that there was no difference in the degree of PPI between both genotypes at a baseline of 3 decibels. (17). The doxorubicin group did not experience any significant differences in PPI compared to the saline control group (17). Both genotypes in the treated and untreated group initially explored the Barnes maze at very similar paces (17). However, the treated APOE ε3 was much slower than the untreated APOE ε4 group (p = 0.004) (17). Also, the control APOE ε3 mice were faster at escaping the hole than the control APOE ε4 mice, especially on training days (TD) 1 & 2 (17). This shows that without any chemotherapeutic agents, spatial learning is better in APOE ε3 mice, but an untreated APOE ε4 mouse will perform better than a doxorubicintreated APOE ε3 mouse. Doxorubicin treatment showed no diminishing effect on the latency of APOE ε3 mice throughout all four TDs, meanwhile significant latency was observed in treated APOE ε4 mice (17). Treated APOE ε4 mice also had a harder time escaping the hole in the Barnes maze without some guidance (17). This suggests that the treatment impairs the spatial learning ability of mice with the APOE ε4 genotype but does not affect those with APOE ε3. The results of the probe trial conducted 72 hours after the Barnes maze tests indicated that doxorubicin treatment in both genotypes correlated with increased latency in the ability of the mice

Exploring the Role of APOE4 Genotype in Chemo Brain

to enter the target hole (17). Some of the treated mice failed to complete this task at all, alluding to poor retention and deficiency in spatial memory. VBM analysis was performed from MRI data, and there was no evidence of significant morphological alterations in brain areas of doxorubicin-treated APOE ε3 mice. The APOE ε4 mice, however, showed some GM atrophy in the frontal cortex, when compared to the control group (17). Although these findings were not statistically significant, when all four experimental groups were compared, APOE ε4 had lower GM volumes in the cerebral cortex and striatum, but no differences in the hippocampus. This replication study supported some of the conclusions made in the previous study, but also provides an alternate view for others. Both reports demonstrated decreased performance on spatial learning in APOE ε4 mice that received chemotherapy, which aligns with clinical symptoms that breast cancer patients have reported during and post chemotherapeutic treatment. This study was successful in using a new preclinical model to present data on the intersection of age and APOE ε4 in the pathogenesis of CICI. It also revealed that the structural changes seen in patients with CICI or in mouse models may be transient and not long lasting. While this seems promising, more research is necessary to consider other variables like premorbid conditions and regimens with multiple chemotherapeutic agents. The methodology of neither study allowed for an extensive look at the effects of the chemotherapeutic at every single point from the moment treatment was administered to the end of the 35 weeks for which assessments were conducted. Furthermore, both studies focused on homozygous genotypes for APOE ε4 and APOE ε3 only and did not account for other polymorphisms and their possible effect on cognitive function.

Conclusion The development of better chemotherapeutic agents and cancer treatment regimens has resulted in improved outcomes for cancer patients. Research shows that while some of these treatments allow patients to live longer, they can also have debilitating side effects that greatly impact quality of life. Breast cancer chemotherapeutic agents, in particular, have been shown to cause cognitive dysfunction in patients during and after treatment. The etiology of this chemotherapyinduced cognitive impairment has not yet been established, but there are several associated mechanisms that have been suggested. Some studies indicate that neuronal injury and a lack of plasticity (10), elevated levels of inflammation and circulating cytokines (19), and genetic factors such as APOE ε4 and COMT (3, 5) may play a role. APOE ε4 has been implicated both in preclinical studies and human clinical trials as a potential risk factor for cognitive decline in general, especially in relation to the cognitive changes seen with aging and Alzheimer’s disease. Research has also shown that APOE ε4 is associated with increased susceptibility of CICI for long-term cancer survivors (8, 11, 20). Studies by Speidell et al. using the APOE knock-in mouse model suggests that the APOE ε4 genotype predisposes patients to cognitive dysfunction, especially in response to chemotherapeutics such as doxorubicin (4, 17, 21, 22). These

data show that doxorubicin-treated APOE ε4 mice exhibited significant impaired spatial learning compared to the other experimental groups, as well as poor memory retention (4). Changes in brain structure were also evident as seen in GM volume in the hippocampal regions and the striatum (4). There was compelling evidence suggesting that the APOE ε4 on its own shows decreased cognitive function, which is exacerbated when chemotherapy is administered. The study by Demby et al. supplements the findings by Speidell using the APOE knock-in mouse model. The APOE ε4 also showed higher levels of cognitive impairment in relation to the APOE ε3 genotype. This deficit in spatial learning worsened with doxorubicin treatment. However, one major discrepancy between the studies is the lack of structural changes in the brain observed in the second study. The researchers attribute this to the fact that mice were studied over a long period of time, and there is a possibility that the cerebral alterations seen are just short term (17). Also, the former experiment noted significant GM reductions in the hippocampus region, which was more pronounced in APOE ε4 mice (4). However, the researchers who replicated the study with the APOE knock-in mouse model reported no findings that showed differences in hippocampal volumes between both groups. The discrepancy could be attributed to the fact that the one experiment was conducted over a long period of time or to a slight difference in the behavioral assays. In summary, development of this APOE knock-in C57BL/6J mouse model and the data that has resulted from its use will pave the way for further research. Future studies that are conducted over a longer period of time, that account for premorbid conditions, and that utilize oncogenic mice will potentially yield supplemental information that can be used to consider establishing APOE ε4 as a risk factor for chemo brain.

References 1.

Argyriou A, Assimakopoulos K, Iconomou G, Giannakopoulou F, Kalofonos H. Either called “Chemobrain” or “Chemofog,” the long-term chemotherapy-induced cognitive decline in cancer survivors is real. J Pain Symptom Manage. 2011;41(1):126139.

2. Henderson FM, Cross AJ, Baraniak AR. 'A new normal with chemobrain': Experiences of the impact of chemotherapyrelated cognitive deficits in long-term breast cancer survivors. Health Psychol Open. 2019;6(1):2055 3. Janelsins MC, Kesler SR, Ahles TA, Morrow GR. Prevalence, mechanisms, and management of cancerrelated cognitive impairment. Int Rev Psychiatry. 2014;26(1):102–113. 4. Speidell AP, Demby T, Lee Y, Rodriguez O, Albanese C, Mandelblatt J, Rebeck GW. Development of a human APOE knock-in mouse model for study of cognitive function after cancer chemotherapy. Neurotox Res. 2019;35(2):291-303. 5. Lange M, Joly F, Vardy J, Ahles T, Dubois M, Tron L, Winocur G, De Ruiter M, Castel H. Cancer-related cognitive impairment: an update on state of the art, detection, and management strategies in cancer survivors. Ann Oncol. 2019;30(12):1925-1940. 129

Exploring the Role of APOE4 Genotype in Chemo Brain

6. Winocur G, Johnston I, Castel H. Chemotherapy and cognition: International cognition and cancer task force recommendations for harmonising preclinical research. Cancer Treat Rev. 2018;69:72-83.

19. Henneghan AM, Palesh O, Harrison M, Kesler SR. Identifying cytokine predictors of cognitive functioning in breast cancer survivors up to 10 years post chemotherapy using machine learning. J Neuroimmunol. 2018;320:38-47.


20. Ahles TA, Saykin AJ, McDonald BC, Furstenberg CT, Cole BF, Hanscom BS, Mulrooney TJ, Schwartz GN, Kaufman PA. Cognitive function in breast cancer patients prior to adjuvant treatment. Breast Cancer Res Treat. 2008;110(1):143-52.

Kim J, Basak JM, Holtzman DM. The role of apolipoprotein E in Alzheimer's disease. Neuron. 2009;63(3):287-303.

8. Ahles T, Saykin A, Noll W, Furstenberg C, Guerin S, Cole B et al. The relationship of APOE genotype to neuropsychological performance in long-term cancer survivors treated with standard dose chemotherapy. Psychooncology. 2003;12(6):612-619. 9. Chang S, ran Ma T, Miranda RD, Balestra ME, Mahley RW, Huang Y. Lipid- and receptor-binding regions of apolipoprotein E4 fragments act in concert to cause mitochondrial dysfunction and neurotoxicity. Proc Natl Acad Sci U S A. 2005;102(51):18694-9. 10. Ahles TA, Saykin AJ. Candidate mechanisms for chemotherapy-induced cognitive changes. Nat Rev Cancer. 2007;7(3):192-201. 11. Ahles TA, Li Y, McDonald BC, Schwartz GN, Kaufman PA, Tsongalis GJ, Moore JH, Saykin AJ. Longitudinal assessment of cognitive changes associated with adjuvant treatment for breast cancer: the impact of APOE and smoking. Psycho-oncology. 2014 Apr 29;23(12):13821390. 12. Mandelblatt JS, Small BJ, Luta G, Hurria A, Jim H, McDonald BC, et al. Cancer-related cognitive outcomes among older breast cancer survivors in the Thinking and Living with Cancer Study. J Clin Oncol. 2018;36(32): JCO1800140 13. Quesnel C, Savard J, Ivers H. Cognitive impairments associated with breast cancer treatments: results from a longitudinal study. Breast Cancer Res Treat. 2009 Jul;116(1):113-23. 14. Ahles TA, Saykin AJ, McDonald BC, Li Y, Furstenberg CT, Hanscom BS, Mulrooney TJ, Schwartz GN, Kaufman PA. Longitudinal assessment of cognitive changes associated with adjuvant treatment for breast cancer: impact of age and cognitive reserve. J Clin Oncol. 2010 Oct 10;28(29):4434-40. 15. DiBattista AM, Dumanis SB, Newman J, Rebeck GW. Identification and modification of amyloid-independent phenotypes of APOE4 mice. Exp Neurol. 2016;280:97-105. 16. Jaykaran. "Mean ± SEM" or "Mean (SD)"? Indian J Pharmacol. 2010;42(5):329. 17. Demby T, Rodriguez O, McCarthy C, Lee Y, Albanese C, Mandelblatt J, et al. A mouse model of chemotherapyrelated cognitive impairments integrating the risk factors of aging and APOE4 genotype. Behav. Brain Res. 2020;384:112534. 18. Ahles TA, Saykin AJ. Candidate mechanisms for chemotherapy-induced cognitive changes. Nat Rev Cancer. 2007;7(3):192-201.


21. Salas-Ramirez KY, Bagnall C, Frias L, Abdali SA, Ahles TA, Hubbard K. Doxorubicin and cyclophosphamide induce cognitive dysfunction and activate the ERK and AKT signaling pathways. Behav Brain Res. 2015;292:133–141. 22. Seigers R, Loos M, Van Tellingen O, Boogerd W, Smit AB, Schagen SB. Cognitive impact of cytotoxic agents in mice. Psychopharmacology (Berl). 2015;232(1):17–37.

Scholarly Research In Progress • Vol. 4, October 2020

Perceived Barriers to Physical Activity and Nutrition Among Geisinger Students Rebecca Kane1*, Peace Nwankwo1*, and Catherine Freeland1

Geisinger Commonwealth School of Medicine, Doylestown Campus, Doylestown, PA 18902 *Master of Biomedical Sciences Program Correspondence: 1

Abstract Physical inactivity and obesity are major public health problems in the United States. A cross-sectional study was carried out to evaluate the perceived barriers to physical activity and nutrition among a sample of medical and graduate students at Geisinger Commonwealth School of Medicine (GCSoM) on Doylestown and Scranton campuses. A total of 93 students (32 males and 61 females) were recruited to complete an online questionnaire that included questions related to their eating habits, physical activity, and lifestyle. Overall, 74.2% thought exercise was extremely/very important. Over half (62.4%) of the students performed vigorous and moderate physical activities that met the World Health Organization (WHO) criteria of health-enhancing physical activities. The percentage of students who thought that proper nutrition was extremely/very important was 77.4%. Based on the American Heart Association recommendations, only 40.9% of students consumed the recommended amount of protein per week, and ≤6.5% consumed the recommended servings for all other food groups. Students reported several barriers, noting time as the most significant barrier to getting the recommended physical activity and diet as a student.

Introduction It is crucial to meet daily physical activity and nutritional needs for one’s body to function properly and to maintain optimum health (1, 2). For substantial health benefits, it is recommended that adults do at least 150 minutes to 300 minutes a week of moderate-intensity, or 75 minutes to 150 minutes a week of vigorous-intensity aerobic physical activity, or an equivalent combination of moderate and vigorous-intensity aerobic activity (1). The current dietary guideline from the American Heart Association recommends that adults eat five servings of vegetables, four servings of whole fruits, and six servings of whole grains per day and eight to nine servings of proteinrich foods per week. It suggests consuming three servings of low-fat or fat-free milk a day, and to eat and drink less sodium, saturated fat, and added sugars (3). Most nutritional recommendations such as protein, carbohydrates, fats, vitamins, and minerals intake can be obtained through a wellbalanced diet (2). Despite the well-documented benefits of physical activity and healthy diets, previous studies note the majority of postsecondary education students around the globe do not meet the weekly physical activity and dietary recommendations and had less than optimal health (1, 2, 4, 5, 6). In the United States, it was found that only 42% of college students participated in vigorous activities that met World Health Organization (WHO) criteria, while merely an additional 20% participated in moderate activity (4). It was also found that only a small

percentage of college students ate the recommended amounts of fruits and vegetables (4). Lack of adequate physical activity and dietary habits can have significant negative health consequences on the human body (1). Physical inactivity and poor dietary habits have been proven to be significant risk factors for many chronic diseases such as heart disease, diabetes mellitus, and hypertension (5). These chronic diseases have been noted to be major causes of death and disability in the U.S. and among the most preventable (6). Students who are entering postsecondary education are a group of people who are at a greater risk of making poor health decisions (e.g., poor dietary habits and inactivity) despite the strong emphasis on meeting nutritional and activity requirements (6). These individuals are beginning to live independently and the transition from home to college life often leads to many complex and multifaceted factors that pose a challenge for many students to make healthy lifestyle choices (2). Some of these factors may include adapting to new social and environmental changes, acknowledging new financial responsibilities, building different social networks, and experiencing different time schedules (2). The goal of this study is to identify common barriers to meeting the recommended physical activity and dietary guidelines for graduate and medical students at Geisinger Commonwealth School of Medicine (GCSoM) in Doylestown and Scranton, Pennsylvania. It is important to identify these barriers and to discover how they impact students because today’s students will be tomorrow's medical professionals, who can play a major role in health awareness, promotion, and dissemination of healthy lifestyles to the public (4).

Methods Literature search A literature review was used to evaluate previous research on the perceived barriers of physical activity and dietary habits that post-secondary students experience. To find this information, various phrases and keywords were used in PubMed and Google Scholar to find relevant articles. Inclusion and exclusion requirements were included to help choose articles most applicable to this study. For example, the articles must have been written in English, published in the last five years and related to physical activity, nutrition, food insecurity, and academic performances. A study on the integration of physical activity and nutrition showed that physical activity has a significant impact on learning procedure among students, and that regular exercise leads to better mental health (8). Weekly physical activity and dietary recommendations are not being met by a large percentage of students in postsecondary education (10). When researchers surveyed 131

Perceived Barriers to Physical Activity and Nutrition Among Geisinger Students

student populations about their physical activity and dietary habits, researchers were able to identify various external and internal barriers that students encountered that kept them from becoming more active and eating healthier options (10). Some of the physical activity barriers that researchers were able to identify include: time limitations, having other important priorities, the stress and tiredness incited by the work or study overload, not being interested in sports, lack of motivation, not knowing how to work out properly, lack of accessible and suitable workout facilities as well as their high cost, and lack of social support (9). Researchers have also noted that students face similar barriers to healthy food choices such as: not having local or accessible food options, the cost of nutrient-dense food, time limitations, and social support (10). Unfortunately, studies have noted that students who face barriers to physical activity and healthy diet options have lower probabilities of becoming active and eating healthier options (9). The lower probabilities of becoming active and eating a healthy diet can possibly be attributed to the fact that students perceive the benefit/ barrier ratio not sufficient enough to motivate them to choose a healthier lifestyle (10).

Table 1. The demographics of the 93 students who participated in this study consists of sex, age, program and employment status.

Questionnaire A questionnaire was developed to assess the barriers that Geisinger students face with physical activity and proper dietary nutrition. The questionnaire included items related to students' physical activity and barriers that prevent students from exercising regularly. Nutritional and food insecurity questions were included to determine whether students are facing insecurities related to food access. Prior to launching the survey, a pilot analysis was conducted to determine quality and readability of the survey instrument. Survey items were reassessed after piloting to improve the quality of the survey. The adjusted survey was distributed by email to graduate and medical students at GCSoM. The survey was open for 4 weeks, from January 29 to February 26 of 2020.

Figure 1. The percentage of students who deemed exercise as extremely, very, moderately, slightly, and not important. Approximately 74.2% of the students considered exercise highly important.

Data analysis Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS®) software. Results were expressed as means ± standard deviation. All reported p-values were two-sided, and p-values less than 0.05 were considered statistically significant.

Results This study had 93 participants, with 61 being female (Table 1). Overall, 58.1% of the respondents were medical students, and 42% identified as graduate students. Out of the 42% of the graduate students, 25.8% of the students are from the Scranton campus, 14% from the Doylestown campus, and 2.2% from the online campus. Approximately 77% of students did not hold any form of current employment. Additionally, 74.2% of students reported exercise as “extremely important” and “very important” (Figure 1), while 1.1% of the students reported exercise as “not important.” According to Table 2, we found that 62.4% of students participated in


Table 2. shows the percentage of participants who participates in moderate or vigorous activity. A majority (62.4%) of students participate in either vigorous or moderate activity. Moderate physical activity is defined as brisk walking, jogging, using an elliptical trainer, swimming leisurely, and water aerobics and bicycling for under 10 miles per hour for at least 2.5 to 5 hours a week. Vigorous activity is defined as running, swimming laps, aerobic dancing, jumping rope, hiking uphill for at least 1.25 to 2.5 hours a week.

Perceived Barriers to Physical Activity and Nutrition Among Geisinger Students

access to fast food (Figure 4). Table 4 shows questions we A

Figure 2. Shows the barriers which prevents students from being physically active. Lack of time, motivation and energy are the most common barriers reported.


Figure 3. Shows the importance of healthy diet among medical and graduate students. Approximately 37% of students reported that exercise is extremely important, 39.8% reported exercise as being very important, 18.3% considered exercise as moderately important and 4.3% reported slightly important. None of the students reported that exercise is not important.

moderate and vigorous physical activity. Moderate physical activity is defined as brisk walking, jogging, using an elliptical trainer, swimming leisurely, water aerobics and bicycling for under 10 miles per hour for at least 2.5 to 5 hours a week, while vigorous activity is defined as running, swimming laps, aerobic dancing, jumping rope, hiking uphill for at least 1.25 to 2.5 hours a week. (1) Furthermore, 37.6% of students participated in vigorous and moderate physical activity. Among the different barriers investigated for exercise (lack of time, lack of motivation, lack of energy, lack of enjoyment, and ease of access to facility, cost, not knowing how to exercise properly, and injury), energy, motivation and the ease of access to the facility was the most common barriers students face with exercise (Figure 2). Students reported the importance of nutrition and results demonstrated that 77.4% of the students reported nutrition as “extremely important” and “very important” (Figure 3). 18.3% of the students considered nutrition as being “moderately important.” Table 3A and 3B shows a series of questions asked to obtain data on servings of fruits, vegetables, whole grain, fat-free milk, and protein. The barriers observed in proper diet consumption are lack of time, cost, and ease of

Table 3A and B. Current dietary guideline recommends 5 servings of vegetables per day but only 6.5% of participants ate the recommended amount. 2.2% reported eating 4 servings of fruits per day, 6.4% ate 6 servings of whole grains per day, 0% consumed 3 servings of low-fat or fat free milk a day, and 40.9% consumed 8 to 9 servings of protein-rich foods per week.

asked to investigate the presence of food insecurity and food assistance programs on campus. Healthy People 2020 defined food insecurity as the disruption of food intake or eating pattern because of lack of money and other resources. Currently, the Doylestown campus does not have a food assistance program. The number of students who reported they were not suffering from food insecurity was 81%, while 17% reported they have experienced food insecurity. In addition, the results showed 65% of the students have not utilized any food assistance program on campus, mainly the Scranton students, while 32% have. Thirty-nine percent reported having skipped 1 to 2 meals, while 22.5% reported skipping 3 to 5 meals per day (Figure 5).

Discussion Previous studies have demonstrated that physical activity and diet play an important role in preventing the onset of chronic non-communicable diseases and are crucial for living a healthy 133

Perceived Barriers to Physical Activity and Nutrition Among Geisinger Students

Figure 4. The barriers GCSoM students face with obtaining a healthy diet. We see lack of time, cost, and ease of access to fast food as the common barriers students face.

Table 4. Responses obtained when asked about food assistance program availability on campus and potential food insecurities. Most (81%) of the students have not experienced food insecurity, while 17% have not experienced food insecurity. Sixty-five percent of the students have utilized any food assistance program.

Figure 5. The number of meals students skipped per week. About 39% of the students reported skipping 1 to 2 meals per week.

life (4). Similarly, to the results found in the literature review, this study has found that a large proportion of the students at GCSoM do not meet the recommended amount of physical activity per week or the nutritional intake guidelines despite the known health benefits. At GCSoM, 62.4% of graduate and medical students reported obtaining the recommended 134

amount of moderate and vigorous physical activity laid out by WHO even though most students thought that exercise was very or extremely important. A possible explanation of why students are not meeting the recommended amount of exercise per week can be attributed to the barriers that students face. According to the study, about 86% of students at GCSoM face some sort of barrier that prevents them from meeting these guidelines. The significant barriers reported by physically inactive students were: lack of time, lack of motivation, lack of energy, and lack of accessible workout facilities. We recommend that there is a need to promote physical activity among medical and graduate students. Both individual and population approaches are needed. For an individual approach, developing social support from friends and families may be a key tactic to increase activity levels among students. For a population approach, the medical school policymakers can take suitable action to highlight the importance of fellow students’ and families’ involvement in promoting physical activity. To help students overcome these barriers and see them meet the recommended amount of physical activity, specific interventions that will target these barriers must be carried out. For example, GCSoM can partner with the local fitness centers to provide monthly memberships at a low cost for students. Similarly, the majority of students (77.4%) thought that nutrition was very or extremely important. Based on the American Heart Association’s recommendations, less than half of the students at GCSoM ate the recommended amount of servings from each food group. Only 40.9% of students ate the recommended amount of protein per week, and ≤6.5% of students ate the recommended amount of servings for all other food groups. One possible explanation of why students are not obtaining the recommended amount of servings can be due to various barriers. According to the data, 79.5% of students experience at least one barrier that prevents them from eating the recommended amount of servings. The significant barriers reported by students were: lack of time, cost, and ease of access to fast food. A second explanation could be due to diet fads and trends like the keto diet, which is a low-carbohydrate diet that limits foods that can be eaten, or intermittent fasting, which can limit the amount of food eaten per day. Intermittent fasting could also explain why some students are not meeting the American Heart Association’s recommendations and also the number of skipped meals reported. A third explanation could be attributed to food insecurity found on campus. Approximately 17% of students reported experiencing food insecurity and 32% have reported using the school food pantry. Furthermore, the dining options currently on campus are not very nutritional and are very expensive. Due to the cost and lack of healthy food options, students do not regularly obtain food there. One major limitation of this study was that we did not have direct access to the SPSS software, which made it difficult to calculate correlations. A second limitation was that some survey questions were not specific enough. For example, the survey did not specify what a serving of milk was (8 ounces) or include alternative types of milk, such as soy or almond milk. Since we did not specify that the serving is 8 ounces or milk alternatives, the number of servings reported might not represent the actual amount of servings consumed by

Perceived Barriers to Physical Activity and Nutrition Among Geisinger Students

students. We also did not specifically ask if meal skipping was due to food insecurity. Students may be skipping meals due to different diets, religion practices and lack of cultural and type of foods offered on campus. In relation to that, the third limitation to this study was that we assume that one diet works for all (Figure 6). Furthermore, other demographics information pertaining to race/ethnicity, religion/spirituality necessary to understanding the dietary needs of the participants were not obtained. Even though the American Heart Association has recommendations for how many servings of each food group should be consumed, that does not mean that these recommendations are what is best for each individual.

will allow students to obtain nutritious foods if they are unable to purchase them on their own. Introducing these recommendations at student orientation or open houses will ensure that all students are aware of all the assistance programs available to them. Similarly, handing out fact sheets of local affordable grocery stores, farms, or community food pantries will help tackle the nutrition problem too.

Acknowledgments We would like to thank Professor Catherine Freeland for her help and guidance throughout the entirety of this research. We also appreciate all the students who participated in the survey.

References 1.

Piercy K, Troiano R, Ballard R, Carlson S, Fulton J, Galuska D et al. The Physical Activity Guidelines for Americans. JAMA. 2018;320(19):2020.

2. Abraham S, Noriega BR, Shin J. College students eating habits and knowledge of nutritional requirements. Journal of Nutrition and Human Health. 2018;02(01). 3. Suggested Servings from Each Food Group [Internet]. 2020 [cited 20 May 2020]. Available from: eat-smart/nutrition-basics/suggested-servings-from-eachfood-group Figure 6. Students’ knowledge regarding nutrition

Conclusions Future studies should determine which barriers are significant for the MD as opposed to the MBS program and how the different campuses' locations impact the overall result of the study. Looking at the programs separately will allow for the implementation of specific interventions that work best for each program. In addition to this, future studies should investigate whether physical activity and nutrition barriers affect academic success. One of the questions on this study’s survey asked the participants about their grade point average. However, not enough data were collected to make a correlation since most of the participants only had pass/fail grades. Furthermore, questions related to ethnicity, religion, and socioeconomic status should also be asked to determine if they impact diet or the amount of physical activity per week. Since the most significant barriers reported by students are lack of time, ease of access to fast and cost, it would be imperative to set up programs that will effectively target these barriers. For example, including a time management course that is available to students multiple times a year would potentially help address the lack-of-time barrier, since it may teach students techniques that can help them learn how to manage their time effectively. A second assistance program that would be beneficial to the students would be a food pantry. Approximately 32% of students at the Scranton campus reported using a food pantry while studying at GCSoM, but there are no school pantries currently available to students at the Doylestown campus. Adding a food pantry

4. Awadalla N, Aboelyazed A, Hassanein M, Khalil S, Aftab R, Gaballa I, et al. Assessment of physical inactivity and perceived barriers to physical activity among health college students, south-western Saudi Arabia. Eastern Mediterranean Health Journal. 2014;20(10):596-604. 5. Downes L. Physical Activity and Dietary Habits of College Students. The Journal for Nurse Practitioners. 2015;11(2):192-198.e2. 6. Assessment of Lifestyle and Eating Habits Among Undergraduate Students in Northern Italy. Annali dell'Istituto superiore di sanita [Internet]. 2020 [cited 20 May 2020];51(2):154-161. Available from: https://pubmed. 7.

Jyoti D, Frongillo A, Jones S. Food Insecurity Affects School’s Children’s Academic Performance, Weight gain and Social Skills. Divisions of Nutritional Science. 2005.

8. Wulf B, Niels L, Agnes P, Aro A, Fogelholm M, Phorsdottir I, Alexander J, Anderssen S, Meltzer H, Pedersen J. Nordic Nutrition Recommendations 2004-integrating nutrition and physical activity, Scandinavian Journal of Nutrition. 2004; 48(4): 178-187. 9. Gomez-Lopez M. Perceived Barriers by University Students in the Practice of Physical Activities. Journal of Sports Science and Medicine. 2010; 7(3): 784-798. doi:10.3390/ijerph7030784 10. Lupi S. Assessment of lifestyle and eating habits among undergraduate students in northern Italy. Annali dell'Istituto Superiore di Sanita. 2015; 51(2): 154-161. doi:10.4415/ANN_15_02_14


Scholarly Research In Progress • Vol. 4, October 2020

Analysis of Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities Laura A. Christman1*, Rachel A. Simon1*, Amy Hoang1*, Kayla M. Day1*, Kelley Chan1†, Jesse T. Clayton1†, and William A. McLaughlin1

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 *Master of Biomedical Sciences Program †Doctor of Medicine Program Correspondence: 1

Abstract Background: Comorbidities of Alzheimer’s disease may be explained in part by shared genetic components. Our aim was to characterize statistically significant co-occurring proteins and their functions in comorbid disease pairs. These were established from population-based inpatient data and overlap of involved proteins. The results are used to further describe the pathology of the comorbidities of Alzheimer’s disease. Methods: Statistically significant comorbid disease pairs of Alzheimer’s disease were extracted from Medicare hospitalization data. To identify significant comorbidities based on protein overlap, mappings between diseases and proteins were first retrieved through keyword searches of Disease Ontology terms across the PubMed abstracts mapped to proteins in UniProt. DisGeNET provided additional disease to protein mappings. Quantification of the statistical significance of comorbid diseases was determined through a phi coefficient analogous to what was used in a study by Hidalgo et al. Results: Of the 34 significant comorbidities of Alzheimer’s disease based on a study by Hidalgo et al., 18 were also significant based on protein overlap. Proteins involved in these 18 comorbidities are considered for how gain or loss of functions may contribute. An example of a statistically significant overlapping protein involved in both Alzheimer’s and Parkinson’s disease is tyrosine-protein kinase Fyn (Fyn). A gain of function of Fyn may contribute to the pathology of both diseases and ultimately to the comorbidity via an increase in a proinflammatory response. Conclusion: Understanding genetic associations between disease comorbidities may have important implications for understanding the disease mechanisms. By finding overlapping proteins and elucidating the function of these proteins in comorbid disease states, we further describe the underlying pathology of comorbid disease pairs of Alzheimer’s disease, with a focus on inferred gain or loss of functions that may contribute.

Introduction Comorbidities are thought to have complex genotypic and phenotypic relationships; these relationships often cannot be defined directly by a single gene (1), as cellular functions compose an interconnected cascade of processes working in tandem to sustain normal functioning (2). When there is an alteration in this cascade, a vast array of physiological malfunctions can result as comorbidities (2). The underlying etiology of comorbidities can be influenced by a multitude 136

of determinants from the genome, proteome, metabolome, and the environment (3). The involvement of cellular networks in comorbidities have been studied previously using health care and insurance data to define genotypic and phenotypic overlap between diseases via a network approach (1, 4). Recognition of genetic overlap, or shared genetic components between comorbidities, can enhance understanding of complex phenotypes and their connections to other diseases (1). Genotypic and phenotypic relationships in comorbidities have been previously analyzed through utilization of population-based health care data (1, 2, 4, 5). Once analyzed, this data can be used to create vast disease networks which can quantify comorbid relationships and determine statistical significance (1, 4–6). A common problem encountered throughout these studies is having a multitude of data with a limited scope; for example, protein-protein interactions can only draw conclusions about protein interactions and not genetic information (7). Accessibility of health care data has also proven to be a challenge, but a previous study has made their populationbased health care data readily accessible to the public (4). To support analyses of disease to protein mappings, DisGeNET integrates data from a multitude of sources to include genotype-phenotype relationships of varying genetic or environmental origins which contribute to human disease (8). Consider that Alzheimer’s disease (AD) is a neurodegenerative disorder which may initially manifest as memory loss and difficulty speaking and can progress to more severe clinical symptoms as more neurons become damaged over time (9). It is estimated that 5.8 million people aged 65 years or older are living with AD in 2020 in the United States (10). However, as the general population ages and life spans increase, this number is expected to more than double to 13.8 million people in the United States by the year 2050 (10). As the mean age of the population increases, the percentage of those living with AD increases. Of individuals aged 65 to 74, 3% have been diagnosed with AD, 17% of individuals aged 75 to 84 and 32% of individuals aged 85 years or older share the same diagnosis (9). AD places a great physical burden on those suffering from the disease, but also places significant social and economic burdens on their families who often become their caregivers as their disease progresses. It is estimated that individuals with dementia receive $321,780 of health care and caregiving costs throughout the course of their disease, which is $184,500 greater than their cost of care if they did not have the disease (11). More evidence needs to be collected regarding the progression of AD and how the disease develops over an individual’s lifespan.

Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities

Many modifiable factors have been identified that may increase the risk of developing AD, including comorbidities; however the underlying biological mechanisms contributing to these processes require further elucidation and characterization. Based on systematic analysis (12), Table 1 lists the defined risk factors for AD with other conditions with a positive association with risk of developing AD. Patients diagnosed with AD and other comorbid diseases tend to have decreased cognitive function, decreased daily functioning in daily activities, and more psychiatric symptoms, increasing their health burden (13). The burden of comorbidities leads to the symptoms of AD worsening in a shorter time span due to a dynamic relationship between various disease-causing factors, not a cause-and-effect relationship (13). Patients with AD have significantly more comorbid medical conditions as documented by a greater number of emergency room visits and hospitalizations (14).

Table 1. Risk factors of AD based on systematic analysis (12). Terms denoted with * have a positive association with risk of developing AD.

The greatest contributing risk factors for AD-related dementia include age, family history, and genetics, but AD is not a normal process of aging. Regarding family history, individuals have a greater likelihood of developing AD if they have one or more first-degree relatives with the condition (9). With regard to genetics, it has been estimated that 1% or less of AD cases are linked to mutations involved in the genes for amyloid precursor protein (APP), presenilin 1, and presenilin 2 (9). Individuals with Down syndrome are at a greater risk for developing AD, as the APP gene is encoded on chromosome 21 (9). More research is needed to enhance understanding of other genetic mechanisms involved in the development of AD. We infer that there is a need to further quantify and describe the overlap of genes between AD and its comorbidities. Describing the function of proteins encoded by these overlapping genes may provide valuable evidence toward understanding the underlying disease mechanisms that contribute to each statistically significant comorbidity of AD. Through analysis of population-based data (4) and the use of comorbidity etiology data from DisGeNET and based on identified mapping from UniProt, co-occurring proteins can be identified between AD and others which can improve understanding of the disease phenotype (4, 8). This study sought to determine statistically significant relationships between co-occurring proteins and their functions within associated comorbid disease pairs of AD as determined by population level data and the overlap of the involved proteins. We describe the inferred protein functional changes, for example, gain or loss of functions, that are implicated in each identified comorbidity.

Methods Study population

Figure 1. Flowchart detailing methodology

Medicare claims data were extracted from a previous study by Hidalgo et al (4), which consisted of patients over the age of 65. The dataset included hospitalizations from 1990 through 1993, where 90.1% of the demographic were classified as white (4). The data utilized ICD-9 diagnostic codes, which was the primary coding system available at the time the record was extracted. Of the 32,341,347 inpatient claims retrieved from



! "



Pi is theof number of patients in the sample with disease i. Pj is the number of Protein Function Overlap and Disease Mechanisms Alzheimer’s Comorbidities patients in the sample with disease j. Cij is defined as the number of patients with proteins from both diseases, and N is the total number of patients (4). Others (2) utilized

Pi is the number of patients in the sample with disease i. Pj the study, only 13,039,347 participants were considered (4, similarly; of using prevalence of the disease is the number of however, patients instead in the sample with disease j. Cij is (P), 15, 16). The number of participants in the study only includedthe phi correlation defined as the number of patients with proteins from both hospitalized patients. The remaining were individuals who they replaced this variable with I for incidence (2). The desired result is a φ value diseases, and N is the total number of patients (4). Others (2) were not hospitalized during this period (4). Additionally, utilized phi correlation similarly; however, instead of using the study focused only on hospitalization claims due to its greater than zero, the indicating a greater co-occurrence than would be expected by chance prevalence of the disease (P), they replaced this variable with consistent nature in the diagnosis of diseases (4). Despite I for incidence (2). The desired result is a φ value greater being approximately 30 years old, this data set was used (2). than indicating a greater co-occurrence thanofwould be because it is the most complete and most readily available The phi zero, correlation provides an analogous tool for the study overlapping expected by chance (2). dataset containing population-based healthcare. Disease to protein mappings, UniProt

proteins between comorbid diseases. The variables are modified accordingly.

The phi correlation provides an analogous tool for the study

of overlapping betweenbycomorbid diseases. The Modification of the formulaproteins was accomplished defining Cij as proteins related to both

To determine significant Alzheimer’s comorbidities, the study variables are modified accordingly. Modification of the formula diseases, N as number of proteins, and Pi and Pj as proteins in diseases i or j, depended on retrospective data. A visual interpretation of the was accomplished by defining Cij as proteins related to both methods is shown in Figure 1. diseases, N as number of proteins, and Pi and Pj as proteins in respectively. diseases i or j, respectively. An automated extraction of 11,013,767 PubMed abstracts

A t-test was used to determine statistical significance of the calculated phi values

mapping proteins mapped to 10,353,916 UniProt IDs was done A t-test was used to determine statistical significance of the based on the corresponding reference section in the UniProt using thecalculated following equation: phi values using the following equation: flat files (17). Based on the PubMed IDs, text fields including đ?œ‘đ?œ‘√đ?‘›đ?‘› − 2 titles, abstracts, and medical subject headings (MeSH) were đ?‘Ąđ?‘Ą = &1 − đ?œ‘đ?œ‘$ retrieved from the National Library of Medicine. Keyword We then created a programmed extraction that yielded a total of 494 disease searches using Disease Ontology (DO) terms (18) mapped to We then created a programmed extraction that yielded a the ICD9 codes represented in the Hidalgo study (4) were pairs through Python 2.7 andthrough 3.7.7. The file was2.7 a table that totalthe of use 494ofdisease pairs thegenerated use of Python used as text queries in a MySQL database tables containing and 3.7.7. The generated file was a table that included the PubMed abstract text fields mapped to UniProt accession included the overlapping proteins associated with these pairs, allowing for determination overlapping proteins associated with these pairs, allowing codes. The mappings between disease ontology and ICD-9 for determination of AD comorbidities along with associated codes are available via OBO file (18). The text search for each significant overlapping proteins. From this, the disease pairs disease term thereby provided a list of UniProt accession that included AD and the shared proteins associated with codes associated with the disease. the pair were documented in a separate file for subsequent literature review and analysis. Disease to protein mappings, DisGeNET DisGeNET aids as an extensive resource in providing a collection of genes and variants associated with human diseases (8). We utilized three different tab separated files (19): “curated gene-disease associations,â€? “UniProt Downloads,â€? and “UMLS (united medical language system) CUI (concept unique identifiers) to several disease vocabulariesâ€?. “Curated gene-disease associationsâ€? is a file compiling associations from UniProt and CUI IDs (listed as “diseaseIdâ€?) to various other databases; however, for the purpose of this research, we extracted only UniProt data and CUI IDs and disregarded the other databases and their data. The “UniProt Downloadsâ€? file includes UniProt accession codes corresponding to its respective DisGeNET unique gene identifiers. “UMLS CUI to several disease vocabulariesâ€? contains mappings of DisGeNET genes (gene identifiers) to UniProt entries and UMLS CUIs to various disease vocabularies, including ICD-9-CM and DO (8). We utilized these files to generate disease ontology to UniProt accession codes mappings which added to those generated by analyses described above. Data analyses Quantitative analyses of protein and functional overlap The formula for the phi correlation utilized for analysis of patient data is described in the Hidalgo study (4). The phi correlation is Pearson’s correlation for binary variables, and a value greater than 0 is indicative of a co-occurrence that is more frequent than would be expected by chance alone (2, 4, 20). The calculation for the phi correlation is the following: đ?œ‘đ?œ‘!" =


đ??śđ??ś!" đ?‘ đ?‘ − đ?‘ƒđ?‘ƒ! đ?‘ƒđ?‘ƒ"

&đ?‘ƒđ?‘ƒ! đ?‘ƒđ?‘ƒ" (đ?‘ đ?‘ − đ?‘ƒđ?‘ƒ! )(đ?‘ đ?‘ − đ?‘ƒđ?‘ƒ" )

With the determination of which proteins have a statistically significant overlap, it is also possible to determine the statistical significance of the pathological functions of these proteins as seen in their respective diseases. We were able to redefine the variables in order to apply the phi correlation to functional overlap. In this case, the variables were defined as: Pi and Pj for how many proteins in each disease that had a selected function. Cij is the number of total proteins with that one function that were involved in both diseases. N is therefore the total number of proteins with the selected function. The functional overlap assessment is ongoing and not yet completed. Qualitative analyses of functional overlap using literature reviews Two procedures were followed after the identification of the statistically significant proteins identified for each disease pair. First, utilization of extracted UniProt protein accession codes and the UniProt online database allowed for determination of the protein name. Second, a comprehensive Google Scholar search was completed based on the protein name, followed by analysis of appropriate journals to further understand its pathophysiological functions in AD and its comorbid disease pair. For example, consider one comorbid disease pair, AD and Parkinson’s disease. The first shared protein of interest was P06241, which the UniProt online database identified as tyrosine-protein kinase Fyn, also known as kinase Fyn (21). The next protein we examined was Q16620, the BDNF/NT-3 growth factors receptor or TrkB (22). Lastly, we also gathered

Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities

Table 3. Application of qualitative analysis to describe the functions of protein etiological factors found to be in common between Alzheimer’s disease and Parkinson’s disease.

Table 2. Diseases which were found to have statistically significant overlap based on both comorbidity and shared proteins with AD.

data. There were 374 disease pairs determined to be not statistically significant by patient data, but significant by protein data. Conversely, 1,145 disease pairs were determined to be statistically significant by patient data, but not by protein data. There were 435 disease pairs which had statistically significant overlap based on both comorbidity and shared proteins.

information on P02511, alpha-crystallin B chain, also known as CRYAB (23). The literature review conducted utilizing Google Scholar and PubMed included the following terms: Alzheimer’s, Parkinson’s, comorbid, comorbidity, P06241, tyrosine-protein kinase Fyn, Fyn, kinase Fyn, tyrosine-protein kinase Fyn AND Alzheimer’s, tyrosine-protein kinase Fyn AND Parkinson’s, Q16620, BDNF/NT-3 growth factors receptor, TrkB, P02511, BDNF/NT-3 AND Alzheimer’s, BDNF/NT-3 AND Parkinson’s, TrkB AND Parkinson’s, alpha-crystallin B chain, TrkB AND Alzheimer’s, alpha-crystallin B chain AND Alzheimer’s, CYRAB AND Alzheimer’s, CRYAB, and alpha-crystallin B chain AND Parkinson’s.

Out of 435 disease pairs, there were 18 disease pairs which had statistically significant overlap based on both comorbidity and shared proteins with AD. These diseases are listed in Table 2. From the quantitative analysis conducted, AD and Parkinson’s disease were found to be a significant comorbid pair. The quantitative results revealed several proteins that were shared between the two diseases. Of those, literature analysis confirmed the association of the following proteins: tyrosine-protein kinase Fyn (Fyn kinase), BDNF/NT3 growth factors receptor, and alpha-crystallin B chain (CRYAB). Table 3 provides a brief overview of the analysis of the function of these proteins as seen in Parkinson’s disease and AD.

The selected literature was chosen based on relevance to the disease pair and discussion of the function of the protein as seen in the diseases of interest. We sought evidence for whether there was a gain or loss of function as it contributes to the pathophysiology of the diseases.

Tyrosine-protein kinase Fyn (Fyn), UniProt accession P06241, was an overlapping protein in the comorbid disease pair. Fyn is a non-receptor tyrosine kinase that plays many different roles, including processes such as cell growth and survival, immune response, cell signaling, cytoskeletal remodeling, and axon guidance (21). Fyn kinase is currently being explored as a possible therapeutic target for AD (24). Instead of focusing on the accumulation of amyloid beta plaques for treatment, Fyn is a downstream effector in the cascade. Soluble forms of amyloid beta protein bind to cellular prion protein on the surface of neurons, which then begins a cascade that ends at Fyn kinase (25). Fyn is also known to be associated with microtubule-associated protein tau, a protein known to be involved in AD pathology (25, 26). There are also studies exploring Fyn kinase’s role in Parkinson’s disease, one of which found that Fyn is a regulator of proinflammatory signaling (27). The study showed that Fyn is needed in order for proinflammatory responses, such as cytokine release, and that it may be upregulated in chronic inflammation. Overall, the study had demonstrated that, as an upstream signaling regulator, Fyn mediates microglial neuroinflammatory processes in Parkinson’s disease (27).

Results Quantitative results were received from running the programs that analyzed the phi correlation. As previously mentioned, a t-test was used to determine statistical significance of the calculated phi values. The t-test threshold was set to 2.576 with a confidence interval of 99%. A 2-by-2 contingency table was created to analyze significant comorbidities and significant protein overlap with 1 degree of freedom. The value N, which was the total UniProt accession codes mapped to disease terms, was equal to 1,560,091. This analysis yielded a chisquare statistic of 8,438 and the t-statistic yielded a p-value less than 0.001. Out of 235,302 identified disease pairs, 60,639 disease pairs were linked to DOID terms, ICD-9-CM codes, and PubMed abstracts. Out of these 60,639 disease pairs, 58,685 disease pairs were not statistically significant by both patient data and protein


Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities

Another shared protein between Parkinson’s and AD was the BDNF/NT3 growth factors receptor (TrkB), UniProt Accession Q16620. This is a receptor tyrosine kinase; the BDNF/NT3 growth factors receptor functions as a regulator of neuron survival, growth, and differentiation, as well as synapse formation and plasticity. It may also have a part in signaling and communication between neurons and glia (22). BDNF signaling is known to have effects on gene expression and protein synthesis and to modulate neurotransmission (28). Alpha-synuclein aggregation is characteristic in Parkinson’s disease; alpha-synuclein has been found to selectively interact with TrkB and inhibit its signaling, leading to neuronal cell death and further contributing to the pathology of the disease (29). Furthermore, in animal models of Parkinson’s disease, it was found that BDNF prevents degeneration of dopaminergic neurons and can improve both motor performance and neurotransmission (30). An increase in BDNF and TrkB was shown to inhibit neurodegeneration and was also associated with gene expression related to neuronal proliferation and survival (30). Studies have shown that BDNF and TrkB both significantly decrease in AD (31). One study suggests that a reduction in TrkB can cause signaling dysfunctions in early AD (31). The mechanism at play may be linked to the deprivation of TrkB, causing an increase in inflammatory cytokines (32). Commonly, AD and Parkinson’s disease involve alpha-crystallin B chain, UniProt accession P02511. This heat-shock protein acts like a chaperone and is noted to prevent proteins from aggregating in stress conditions (23). The gene for this protein is known as CRYAB; a study has found that knockdown of CYRAB may be useful as a therapy against proteinopathies, with Parkinson’s disease as an example (33). Alpha-crystallin B chain has also been found to have increased expression in the substantia nigra in Parkinson’s disease, and may play a role in glial pathology during dopaminergic neuron degeneration (34). Another study found an association between alpha-crystallin B chain with the deposition of amyloid plaques in AD (35). It is one of many small heat shock proteins that have been noted to be linked to pathological lesions found in AD (36).

Dicussion Characterization of disease comorbidities often involves genetic components between diseases and neglects the functional overlap of proteins that are involved (37). Considering that various risk factors may increase prevalence of AD, studying its functional characterization can aid in identifying its molecular mechanisms (9). Contrasted with previous studies, our work proposes that functional measures could provide additional information that could possibly be missed if only genetic overlap was considered (37). Subsequent to the identification of statistically significant comorbidities of AD using the quantitative analyses of the protein overlap, a discussion of the functions of the overlapping proteins was considered using reviews of the literature. We contrast the function of the overlapping proteins as describing the non-diseased state, its putative role in AD, its putative role in the disease comorbid with AD, and its putative role specificity in the comorbid condition. Comparison of the role of the protein was conducted to determine how its potential change in function, either a gain or loss of function,


contributed to the comorbidity. For example, loss of function or gain of function mutations associated with a protein may result in an increased risk for each disease separately and for the comorbid condition. We focused on one comorbid disease pair: AD and Parkinson’s disease. Three overlapping proteins were identified: tyrosine-protein kinase Fyn, BDNF/NT3 growth factors receptor (TrkB), and alpha-crystallin B chain. Based on the quantitative results when comparing the two diseases, the overlapping proteins indicate that there is an association between the two conditions. The qualitative review of the literature describing the protein factors was found to support the quantitative results. Insight is formed regarding the molecular etiology of each neurodegenerative disorder as well as for the comorbidity, which happens in a significant number of individuals.

Conclusion As we continue with the research, we will utilize a more comprehensive set of comorbidities of AD based on the population studies which is provided in tabular form and accessible via a web interface. We aim to compile more extensive data by including recent protein analysis from UniProt and better-defined diagnosis of AD, which may be achieved through ICD-10 diagnostic codes. Although our study represents an extensive analysis of functional overlap between co-occurring proteins of AD comorbidities, there are a few limitations that should be noted. First, the study population was restricted to only Medicare patients (4). This limited the dataset, as the age range was bound to individuals 65 and older, which may exclude early-onset AD cases. Second, the dataset only considered inpatient diagnostic codes and disregarded outpatient codes. The lack of attention to outpatient claims could indicate that the claims data only recognizes moderate to severe cases, thus excluding the milder cases. Third, the extracted dataset adopted ICD-9 as its primary diagnostic code, because the dataset was generated prior to the implementation of ICD-10. The ICD-10 coding system classifies diseases that are consistent with current clinical practices and medical advances (38). There is a greater level of detail enabling greater specificity that is aligned with ICD-10 diagnostic codes (39). Fourth, the study population predominantly included white patients only, making up about 90% of the entire study. The lack of racial diversity limits the appearance of different comorbidity pairs and protein function overlap within those pairs. Fifth, the Medicare data included records from multiple hospitals. Removal of personal identifiers of the patients made it difficult to define how AD was diagnosed in these individuals. For example, why were these patients diagnosed with AD as opposed to dementia, vascular dementia, or frontotemporal dementia? Additionally, were these patients diagnosed based on postmortem autopsy? Lastly, prevalence of AD in the United States’ general population was taken from the 2010 Census Bureau. With the 2020 Census currently taking place, there may be inaccuracies included in the numbers, thus not being representative in the future.

Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities

Acknowledgments Laura Christman, Kayla Day, Amy Hoang, Rachel Simon, and William McLaughlin wrote the manuscript. Laura Christman, Kayla Day, Amy Hoang, Rachel Simon, William McLaughlin, Paul J. DePietro, Jesse Clayton, and Kelley Chan helped develop the computer programs. We thank Clinton Sonier for helping to organize the data and programs. Funding was provided in part by the Marquardt Foundation for Alzheimer’s Research.

Disclosures The authors and other contributors to this study have no conflicts of interest.

References 1.

Rzhetsky A, Wajngurt D, Park N, Zheng T. Probing genetic overlap among complex human phenotypes. PNAS. 2007;104(28):11694–11699.

2. Park J, Lee D-S, Christakis NA, Barabási A-L. The impact of cellular networks on disease comorbidity. Mol Sys Biol. 2009;5(1):262.

12. Xu W, Tan L, Wang H-F, Jiang T, Tan M-S, Tan L, et al. Meta-analysis of modifiable risk factors for Alzheimer’s disease. J Neurol Neurosurg Psychiatry. 2015;86(12):1299–1306. 13. Haaksma ML, Vilela LR, Marengoni A, Calderón-Larrañaga A, Leoutsakos J-MS, Olde Rikkert MGM, et al. Comorbidity and progression of late onset Alzheimer’s disease: A systematic review. PLoS One. 2017;12(5). Available from: 14. Zhao Y, Kuo T-C, Weir S, Kramer MS, Ash AS. Healthcare costs and utilization for Medicare beneficiaries with Alzheimer’s. BMC Health Serv Res. 2008;8:108. 15. Lauderdale DS, Furner SE, Miles TP, Goldberg J. Epidemiologic Uses of Medicare Data. Epidemiol Rev. 1993;15(2):319–327. 16. Mitchell JB, Bubolz T, Paul JE, Pashos CL, Escarce JJ, Muhlbaier LH, et al. Using Medicare claims for outcomes research. Medical Care. 1994;32(7):JS38–51. 17. UniProt Consortium. UniProt: a worldwide hub of protein knowledge. Nucleic Acids Res. 2019;47(D1):D506-515.)

3. Loscalzo J, Kohane I, Barabasi A-L. Human disease classification in the postgenomic era: A complex systems approach to human pathobiology. Mol Sys Biol. 2007;3(1):124.

18. Schriml LM, Mitraka E, Munro J, Tauber B, Schor M, Nickle L, et al. Human Disease Ontology 2018 update: classification, content and workflow expansion. Nucleic Acids Res. 2019;47(Database issue):D955–962.

4. Hidalgo CA, Blumm N, Barabási A-L, Christakis NA. A Dynamic Network Approach for the Study of Human Phenotypes. PLoS Comput Biol. 2009;5(4). Available from:

19. DisGeNET Downloads [Internet]. DisGeNET. Available from:

5. Kim JH, Son KY, Shin DW, Kim SH, Yun JW, Shin JH, et al. Network analysis of human diseases using Korean nationwide claims data. J Biomed Inform. 2016;61:276– 282.

21. Tyrosine-protein kinase Fyn [Internet]. UniProt Consortium: European Bioinformatics Institute - Protein Information Resource. 2020. Available from: uniprot/P06241

6. Goh K-I, Cusick ME, Valle D, Childs B, Vidal M, Barabási A-L. The human disease network. PNAS. 2007;104(21):8685–8690.

22. BDNF/NT-3 growth factors receptor [Internet]. UniProt Consortium: European Bioinformatics Institute - Protein Information Resource. 2020. Available from: https://www.


Qiu Y-Q, Zhang S, Zhang X-S, Chen L. Detecting disease associated modules and prioritizing active genes based on high throughput data. BMC Bioinformatics. 2010;11(1):26.

8. Piñero J, Bravo À, Queralt-Rosinach N, GutiérrezSacristán A, Deu-Pons J, Centeno E, et al. DisGeNET: a comprehensive platform integrating information on human disease-associated genes and variants. Nucleic Acids Res. 2017;45(D1):D833–839.

20. Cramér H. Mathematical methods of statistics. Princeton University Press; 1999: 43.

23. Alpha-crystallin B chain [Internet]. UniProt Consortium: European Bioinformatics Institute - Protein Information Resource. 2020. Available from: uniprot/P02511 24. Nygaard HB, van Dyck CH, Strittmatter SM. Fyn kinase inhibition as a novel therapy for Alzheimer’s disease. Alzheimers Res Ther. 2014;6(1):8.

9. Alzheimer’s Association. 2020 Alzheimer’s disease facts and figures. Alzheimer’s & Dementia. 2020;16(3):391–460.

25. Nygaard HB. Targeting Fyn kinase in Alzheimer’s disease. Biol psychiatry. 2018;83(4):369-376.

10. Hebert LE, Weuve J, Scherr PA, Evans DA. Alzheimer disease in the United States (2010–2050) estimated using the 2010 census. Neurology. 2013;80(19):1778–1783.

26. Shirazi SK, Wood JG. The protein tyrosine kinase, fyn, in Alzheimer's disease pathology. Neuroreport. 1993;4(4):435-437.

11. Jutkowitz E, Kane RL, Gaugler JE, MacLehose RF, Dowd B, Kuntz KM. Societal and Family Lifetime Cost of Dementia: Implications for Policy. J Am Geriatr Soc. 2017;65(10):2169–2175.

27. Panicker N, Saminathan H, Jin H, Neal M, Harischandra DS, Gordon R, Kanthasamy K, et al. Fyn kinase regulates microglial neuroinflammatory responses in cell culture and animal models of Parkinson's disease. J Neurosci. 2015;35(27):10058-10077.


Protein Function Overlap and Disease Mechanisms of Alzheimer’s Comorbidities

28. Mercado NM, Collier TJ, Sortwell CE, Steece-Collier K. BDNF in the aged brain: translational implications for Parkinson’s disease. Austin Neurol Neurosci. 2017;2(2). 29. Kang SS, Zhang Z, Liu X, Manfredsson FP, Benskey MJ, Cao X, et al. TrkB neurotrophic activities are blocked by α-synuclein, triggering dopaminergic cell death in Parkinson’s disease. Proc Natl Acad Sci USA. 2017;114(40):10773-10778. 30. Palasz E, Wysocka A, Gasiorowska A, Chalimoniuk M, Niewiadomski W, Niewiadomska G. BDNF as a promising therapeutic agent in Parkinson’s disease. Int J Mol Sci. 2020;21(3):1170. 31. Devi L, Ohno M. TrkB reduction exacerbates Alzheimer’s disease-like signaling aberrations and memory deficits without affecting β-amyloidosis in 5XFAD mice. Transl Psychiatry. 2015;5(5):e562. 32. Wang ZH, Xiang J, Liu X, Yu SP, Manfredsson FP, Sandoval IM, et al. Deficiency in BDNF/TrkB Neurotrophic Activity Stimulates δ-Secretase by Upregulating C/EBPβ in Alzheimer’s Disease. Cell Rep. 2019;28(3):655-669. 33. Lu SZ, Guo YS, Liang PZ, Zhang SZ, Yin S, Yin YQ, et al. Suppression of astrocytic autophagy by αB-crystallin contributes to α-synuclein inclusion formation. Transl Neurodegener. 2019;8(1):3. 34. Liu Y, Zhou Q, Tang M, Fu N, Shao W, Zhang S, et al. Upregulation of alphaB-crystallin expression in the substantia nigra of patients with Parkinson's disease. Neurobiol Aging. 2015;36(4):1686-1691. 35. Renkawek K, Voorter CE, Bosman GJ, Van Workum FP, De Jong WW. Expression of αB-crystallin in Alzheimer's disease. Acta Neuropathol. 1994;87(2):155-160. 36. Wilhelmus MM, Boelens WC, Otte-Höller I, Kamps B, Kusters B, Maat-Schieman ML, et al. Small heat shock protein HspB8: its distribution in Alzheimer’s disease brains and its inhibition of amyloid-β protein aggregation and cerebrovascular amyloid-β toxicity. Acta Neuropathol. 2006;111(2):139-149. 37. Rubio-Perez C, Guney E, Aguilar D, Piñero J, Garcia-Garcia J, Iadarola B, et al. Genetic and functional characterization of disease associations explains comorbidity. Scientific Reports. 2017;7(1):1–14. 38. Cartwright DJ. ICD-9-CM to ICD-10-CM Codes: What? Why? How?. Adv Wound Care (New Rochelle). 2013;2(10):588–592. 39. ICD - ICD-10-CM - International Classification of Diseases,(ICD-10-CM/PCS transition [Internet]. 2019. Available from: pcs_background.htm


Scholarly Research In Progress • Vol. 4, October 2020

Colonic Lavage in Treatment of Refractory Clostridium difficile Infection Adaptation of the Pittsburgh Protocol Maura Morgan1†, Timothy Farrell1†, Gordian U. Ndubizu2, and Timothy J. Farrell3

Geisinger Commonwealth School of Medicine, Scranton, PA 18509 Geisinger Northeast General Surgery Residency, Wilkes-Barre, PA 18711 3 Geisinger Community Medical Center, Scranton, PA 18510 †Doctor of Medicine Program Correspondence:; 1


Abstract Clostridium difficile infection (CDI) is a common nosocomial sequela in patients treated with antibiotics. Surgical intervention is indicated in fulminant cases. However, the mortality associated with total colectomy and end ileostomy is high. Previous reports have indicated that surgical intervention for severe complicated CDI with formation of a loop ileostomy leading to the diversion of fecal stream followed by colonic lavage can be beneficial in treating severe CDI. This procedure is known as the Pittsburgh Protocol and has been reported to decrease the mortality and the need for a total colectomy in patients with severe complicated CDI. In this case, we present a 75-year-old female with refractory CDI. In her treatment, we adapted the Pittsburgh Protocol and utilized a 20 French MIC gastrostomy tube to recreate the ileocecal valve and control the colonic lavage without retrograde flow.

Introduction Clostridium difficile, the culprit of antibiotic-associated pseudomembranous colitis, is an anaerobic Gram positive cytotoxin-producing bacterium (1). C. difficile infection (CDI) is the leading cause of health-care-associated infections in the United States. As many as 500,000 infections occur annually. Manifestations vary from non-severe to fulminant disease (2). Up to 10% of patients have disease progression despite appropriate antimicrobial therapy. In patients who progress to fulminant disease, total abdominal colectomy with end ileostomy is the standard of care. However, this has been called into question since McCreery et al. published their results with fecal diversion and antegrade colonic lavage the so-called “Pittsburgh Protocol.” Standard surgical intervention in those with fulminant disease is associated with a 30day mortality of up to 71% (3). It is for this reason that many advocate for early surgical intervention with total abdominal colectomy as opposed to waiting for the development of fulminant disease. Although it is clear that earlier surgical intervention leads to improved outcomes, it remains unclear what constitutes the optimal time for surgical intervention. Currently accepted indications for surgery in those with CDI include hemodynamic instability, megacolon, perforation, or need for intensive care unit admission (4). Furthermore, Klobuka et al. established the following as parameters for early surgical intervention: age greater than 65, peritoneal signs on physical examination, abdominal distension, end-organ failure, systolic blood pressure less than 90 mm Hg, heart rate greater than 100 beats per minute, need for vasopressor, leukocytosis of greater than 16 × 109/μL, serum lactate of greater than 2.2

mmol/L, and radiographic findings of pancolitis, ascites, megacolon, or colonic perforation (5). The advent of the Pittsburgh Protocol in 2017 in conjunction with an emphasis on earlier surgical intervention has led to a re-evaluation of the use of total abdominal colectomy in the treatment of severe CDI. Fecal diversion with diverting ileostomy disrupts the fecal stream, deprives the intestinal flora of nutrition, and decreases the colonic toxin burden. In the pilot study by McCreery et al., individuals that met the diagnostic criteria for severe or fulminant CDI and were surgical candidates underwent a diagnostic laparoscopy and creation of a loop ileostomy. Subsequently, 8 L warmed polyethylene glycol 3350/electrolyte solution was introduced through the distal end of the ileostomy and collected via a rectal drainage tube. Postoperatively, the patient received antegrade vancomycin flushes, 500 mg in 500 mL lactated ringers every 8 hours, via a Malecot catheter (24 French) left in the efferent limb of the ileostomy, and 500 mg IV metronidazole every 6 hours for 10 days. The reported mortality for these patients was 19% in the immediate postoperative period. An additional 14% died outside of this time frame, of which none of the deaths were related to CDI. Of note, 3 of the 43 patients who underwent diversion required conversion to total abdominal colectomy (6). Here we describe a case of severe CDI managed successfully with slight modification of the original protocol.

Case Presentation A 75-year-old female presented to the hospital exactly 1 month following a prior hospital admission in which methicillinresistant Staphylococcus aureus pneumonia was diagnosed and treated with IV vancomycin for 9 days. Four days prior to readmission, she was diagnosed with CDI and had symptoms of fever, nausea, vomiting, and diarrhea. Her WBC was 22,860 cells per cubic millimeter of blood and PO vancomycin and IV metronidazole were initiated upon admission. Within 2 days, she became increasingly hemodynamically unstable with a leukocytosis of 38,680 cells per cubic millimeter. Due to severe CDI on appropriate medical therapy, surgical intervention was deemed necessary. A laparotomy was performed and the large and small bowel deemed viable. Subsequently, a loop ileostomy was fashioned and a 20-French MIC gastrostomy tube advanced through the efferent limb with the balloon inflated just beyond the ileocecal valve. The colon was lavaged with GoLYTELY solution as per the initial Pittsburgh protocol and the patient transitioned to the intensive care unit following abdominal wall closure. 143

Colonic Lavage in Treatment of Refractory Clostridium difficile Infection

On postoperative day zero antegrade vancomycin lavage, 500 mg in 500 mL lactated ringers every 8 hours, was performed through the ostomy. Lavages were continued until postoperative day 12. Though the Pittsburgh Protocol recommends the duration of vancomycin to be continued for only 10 days, we continued the treatment until the patient was both clinically and grossly improving. Leukocytosis resolved on the fourth postoperative day. The patient was deemed stable for discharge by postoperative day 15. Three months later the patient underwent uneventful reversal of the ileostomy in a hand-sewn manner and is currently awaiting repair of an incisional hernia at her previous stoma site.

2. McDonald CL, Gerding DN, Johnson S, Bakken JS, Carroll KC, Coffin S, E et al. “Clinical Practice Guidelines for Clostridium Difficile Infection in Adults and Children: 2017 Update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA).” Clinical Infectious Diseases, 2018, doi:10.1093/cid/cix1085.


4. Seltman AK. Surgical Management of Clostridium difficile Colitis. Clin Colon Rectal Surg. 2012. doi:10.1055/s-0032-1329390

Antibiotic-associated pseudomembranous colitis due to C. difficile is a common nosocomial infection in the setting of recent antibiotic use. Even with medical therapy, 10% of cases progress, leading to eventual hemodynamic instability and death if not appropriately treated. It is paramount to have protocols in place for refractory cases. The Pittsburgh Protocol presents an alternative to the current gold standard in surgical intervention in this patient population. One significant modification from the original protocol was performed in our case study. A MIC gastrostomy tube was placed in the efferent limb of the ostomy instead of a Malecot drain. Antegrade vancomycin lavage was performed through this route and led to resolution of the patient’s leukocytosis in 4 days. We preferred use of the 20-French MIC gastrostomy tube, as it allows for recreation of the ileocecal valve. Furthermore, balloon inflation ensured a water-tight seal such that all administered enteric/ colonic vancomycin remained within the gastrointestinal tract. No antibiotics were lost due to spillage or retrograde flow of colonic contents. This is of importance in those with pseudocolonic obstruction caused by severe or fulminant colitis. Additionally, this route of administration is more efficient than that described in the original proposal.

Conclusion The high mortality rate associated with total abdominal colectomy in the treatment of severe CDI necessitates improvement in our current treatment protocols. This case highlights the use of colonic diversion and lavage as an alternative to the current gold standard. Continued review is required to formulate the optimal treatment algorithm and optimize patient outcomes.

Disclosures Nothing to disclose.

References 1.


Peikin, SR, Galdibini J, Bartlett JG. Role of Clostridium Difficile in a Case of Nonantibiotic-Associated Pseudomembranous Colitis. Gastroenterology, 1980, doi:10.1016/0016-5085(80)90457-6.

3. McCreery G, Jones PM, Kidane B, Demelo V, MeleT. Polyethylene Glycol Intestinal Lavage in Addition to Usual Antibiotic Treatment for Severe Clostridium Difficile Colitis: A Randomised Controlled Pilot Study. BMJ Open, 2017, doi:10.1136/bmjopen-2017-016803.

5. Klobuka AJ, Markelov A. Current status of surgical treatment for fulminant clostridium difficile colitis. World J Gastrointest Surg. 2013;5(6):167–172. doi:10.4240/wjgs. v5.i6.167 6. Neal MD, Alverdy JC, Hall DE, Simmons RL, Zuckerbraun BS. Diverting Loop Ileostomy and Colonic Lavage: An Alternative to Total Abdominal Colectomy for the Treatment of Severe, Complicated Clostridium difficile Associated Disease. Ann Surg. 2011;254(3):423-429. doi:10.1097/SLA.0b013e31822ade48.

Medical Research Honors Program If you’re a current first-year medical student, consider the Medical Research Honors Program (MRHP). You’ll drive this long-term, thesis-focused research experience with guidance from a mentor. Complete the requirements while remaining in good academic standing, and you’ll graduate with an honors distinction.

Through the MRHP, you will: • Advance fundamental scientific knowledge • Set yourself apart in competitive residency application fields • Refine your scholarly communication • Gain a mindset of continual growth and learning To complete this 4-year program, you must submit a research project proposal, write a thesis and deliver an oral defense. You’ll also write abstracts, present posters and publish findings while building toward your thesis defense. Your research experience is guided by a research mentor, a thesis advisory committee and the program director. We encourage you to participate in the Summer Research Immersion Program as well.

Application packet must include: • MRHP application form • Letter of support from your research mentor • Your CV • Acknowledgment of your mentor’s expectations • Project proposal: Title, specific aims, hypothesis, background, preliminary data (if available)

Questions about the MRHP program or mentoring? Email Sonia Lobo, PhD, associate dean for research and scholarship (slobo01@som. or Thomas J. Majernick, MS, MRHP program director (tmajernick01@som.

Application deadline: Monday, March 29, 2021

Be a mentor Submit a mentor data form to StudentResearch@som. indicating your willingness to commit time, facilities and resources to a student as needed throughout the program.

Geisinger Commonwealth School of Medicine is committed to nondiscrimination in all employment and educational opportunities.


525 Pine St. Scranton, PA 18509



Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.