Page 1


Evaluation of outcomes in dogs treated for pyothorax: 46 cases (1983–2001) Harry W. Boothe, dvm, ms, dacvs; Lisa M. Howe, dvm, phd, dacvs; Dawn M. Boothe, dvm, phd, dacvim, dacvcp; Loren A. Reynolds, dvm; Mark Carpenter, phd

Objective—To determine the effect of treatment approach on outcome and the appropriateness of initial empirical antimicrobial treatment in dogs with pyothorax. Design—Retrospective case series. Animals—46 dogs with pyothorax confirmed by either (n = 15) or both (31) of the following: intracellular bacteria in pleural fluid or tissue (41) and bacteria recovered via culture of pleural fluid (36). Procedures—Medical records of dogs treated for pyothorax from 1983 through 2001 were reviewed. Data on signalment, history, clinical signs, and treatment and results of diagnostic imaging and cytologic and microbiological evaluations were obtained. Follow-up was performed via reexamination (n = 15) and contact with referring veterinarians (26) and owners (24). Results—46 dogs were treated with at least 1 antimicrobial and thoracocentesis (n = 7; noninvasive group), a thoracostomy tube (26; invasive group) with or without pleural lavage and heparin, or a thoracotomy (13; surgical group) and thoracostomy tube with or without pleural lavage and heparin. Pyothorax recurred in 7 dogs, and 5 of the 7 died or were euthanatized. In the respective groups, the short-term survival rate was 29%, 77%, and 92% and the longterm survival rate was 29%, 71%, and 70%. Pleural lavage and heparin treatment increased the likelihood of short- and long-term survival. Results of antimicrobial susceptibility testing suggested empirical antimicrobial selection was associated with a 35% risk of inefficacy. Conclusions and Clinical Relevance—In the dogs with pyothorax in this study, favorable treatment effects were achieved with surgery (for short-term survival) and pleural lavage and heparin treatment (for short- and long-term survival). Findings failed to support the hypothesis that invasive (surgical) versus noninvasive treatment of pyothorax in dogs leads to a better long-term outcome. (J Am Vet Med Assoc 2010;236:657–663)


yothorax, defined as septic inflammation of the pleural cavity, is a potentially recurrent condition that results in systemic illness. Clinical signs of pyothorax commonly include dyspnea, tachypnea, anorexia, pyrexia, and exercise intolerance. Pleural effusion is diagnosed on the basis of results from radiography or ultrasonography, whereas pyothorax is diagnosed on the basis of results from cytologic or microbiological evaluation. Treatment is intended to eliminate infection, improve ventilation, and minimize recurrence. Although pyothorax in dogs has been investigated, associated studies1–9 have differed in their approach to evaluation and assessment of treatment effects, including the appropriateness of initial empirical antimicrobial selection. From the Departments of Small Animal Medicine and Surgery (HW Boothe, Howe, Reynolds) and Veterinary Physiology and Pharmacology (DM Boothe), College of Veterinary Medicine, Texas A&M University, College Station, TX 77843; and the Department of Mathematics and Statistics, College of Sciences and Mathematics, Auburn University, Auburn, AL 36849 (Carpenter). Dr. Harry Boothe’s present address is the Department of Clinical Sciences, College of Veterinary Medicine, Auburn University, Auburn, AL 36849. Dr. Dawn Boothe’s present address is the Department of Anatomy, Physiology, and Pharmacology, College of Veterinary Medicine, Auburn University, Auburn, AL 36849. Dr. Reynolds’ present address is Los Gatos Dog and Cat Hospital, 17480 Shelburne Ave, Los Gatos, CA 95030. Presented as a poster at the American College of Veterinary Surgeons Veterinary Symposium, Orlando, Fla, October 1997. Address correspondence to Dr. Harry Boothe ( JAVMA, Vol 236, No. 6, March 15, 2010

The purpose of the study reported here was to determine short- and long-term survival rates in dogs with pyothorax. In addition, we sought to examine the effect of treatment type on outcome (survival rate) and to evaluate the appropriateness of initial empirical antimicrobial selection. Surgical treatment was hypothesized to result in a better long-term outcome than less invasive treatment. Materials and Methods Case selection—Consecutive medical records were reviewed for dogs with pyothorax admitted to the Texas Veterinary Medical Teaching Hospital between June 1983 and December 2001. Inclusion criteria were cytologic evidence of septic inflammation in pleural fluid or tissue or microbiological evidence of pleural infection. Dogs were separated into 3 treatment groups: those treated with thoracocentesis (noninvasive group), those treated with a thoracostomy tube or tubes (invasive group) with (n = 22) or without (4) pleural lavage, and those treated with surgical debridement (surgical group) via a thoracotomy and thoracostomy tube or tubes with (10) or without (3) pleural lavage. All dogs received systemically administered antimicrobials. Medical records review—Information collected from medical records included dog signalment and history, clinical signs, diagnostic imaging findings, cytologic and microbiological findings, and treatment. Scientific Reports



Microbiological testing—Microbiological culture and antimicrobial susceptibility testing was performed at the Clinical Microbiology Laboratory of the Veterinary Medical Teaching Hospital, which followed the guidelines of the Clinical Laboratory and Standards Institute.10 Treatment response—Short-term survival rate (number of dogs that left the hospital alive divided by the number of dogs in the study) and long-term survival rate (number of dogs alive at 12 months after discharge divided by the number of noncensored dogs) were calculated. Follow-up information was obtained via reexamination at the teaching hospital (n = 15 dogs) and questionnaire or telephone contact with the referring veterinarian (26) and owner (24). Whether pyothorax had recurred was determined by reexamination or questionnaire. Statistical analysis—Data were summarized among the 3 treatment groups (noninvasive, invasive, and surgical) by use of descriptive statistics. For categorical variables (short-term or long-term survival among groups; breed distribution between the teaching hospital and study populations), proportions were compared with the Fisher exact test. Survival data were analyzed by use of Kaplan-Meier survivor function estimation and log-rank tests. Dogs that were lost to follow-up or that died of unrelated causes were censored. All analyses were performed by use of commercially available softwarea; values of P < 0.05 were considered significant. Results Dogs—The medical record review revealed 46 dogs for inclusion in the study. Median age was 3 years (range, 4 months to 14 years), and median body weight was 22.3 kg (50 lb; range, 4 to 41 kg [9 to 90 lb]). There were more males (n = 29) than females (17; Table 1). Seventeen breed types were represented, the most common of which were Pointer (n = 12), hound-type (7), Labrador Retriever (7), and mixed-breed dog (3). None of Table 1—Characteristics of 46 dogs treated for pyothorax by noninvasive (n = 7), invasive (26), or surgical (13) means at a veterinary teaching hospital from 1983 through 2001. Characteristic




Signalment Age (y) 4 (0.8−6.0) 3 (0.3−13.7) 3 (1.2−8.5) Body weight (kg) 27.3 (4.0−37.3) 21.1 (5.5−41.0) 23 (6.8−32.0) Male 4 19 6 Female 3 7 7 Clinical signs Dyspnea 3 23 11 Anorexia or weight loss 4 15 9 Fever 3 15 6 Exercise intolerance 4 11 5 Radiographic findings Unilateral fluid 1 0 2 Bilateral fluid 6 26 11 Large fluid volume 1 14 6 Medium fluid volume 5 8 5 Small fluid volume 1 4 2 Cytologic findings for pleural fluid Sulfur granules 0 12 3 Bacteria 7 23 9 Values in parentheses represent ranges. Pleural fluid volume was subjectively assessed.


Scientific Reports

the breeds in the study population were overrepresented in comparison with the hospital population (P = 0.26). Dyspnea (n = 37 [80%]) was the most common clinical sign for which dogs were initially evaluated, and anorexia (28 [61%]), pyrexia (24 [52%]), and exercise intolerance (20 [43%]) were also common. A presumed initiating event of pyothorax could be identified in 10 dogs. Such events included foreign body (n = 5; mediastinal or pulmonary [3], thoracic wall [1], or esophageal [1]), thoracic trauma (4; penetrating wounds [2], gunshot injury [1], or blunt trauma [1]), and aspiration pneumonia (1). Thoracic radiographs were obtained for all 46 dogs, with thoracic ultrasonography performed in 18 (39%). Radiographic and ultrasonographic findings were consistent with pleural fluid accumulation. Bilateral pleural effusion was observed commonly (n = 43 [93%]). Volume of pleural fluid present at initial radiologic evaluation was assessed subjectively by the radiologist as large (n = 21 [46%]), moderate (18 [39%]), or small (7 [15%]). Nine of 19 dogs in which positional (horizontal beam), radiographic (n = 6), or ultrasonographic (3) evaluations were performed had findings consistent with organized (compartmentalized) pyothorax; 7 of the 9 dogs were in the surgical group. Ten dogs were assessed as having freely moveable pleural fluid: 2 in the noninvasive group, 6 in the invasive group, and 2 in the surgical group. Cytologic evaluation—Eighty-six samples of pleural fluid and 22 samples of pleural tissue collected from 45 dogs were cytologically examined. Ninety-five samples were collected during hospitalization, and 13 were obtained at necropsy. Findings of all cytologic analyses were consistent with an exudate. Samples from 41 (91%) dogs were assessed as indicative of sepsis, samples from 4 (9%) were assessed as indicative of no sepsis, and 1 (2%) dog did not undergo a cytologic evaluation. Intracellular bacteria were observed cytologically in fluid (n = 39), tissue (14), or both (12) in these 41 dogs. Microbiological culture of pleural-fluid samples yielded bacterial growth for all 4 dogs assessed as nonseptic and the 1 dog with no cytologic evaluation. Cytologic findings did not match microbiological findings in 13 of 44 (30%) dogs. Samples from 9 dogs with cytologic evidence of intracellular bacteria yielded no growth on microbiological culture, and samples from 4 dogs with no cytologic evidence of bacteria yielded bacterial growth. Sulfur granules were detected in pleural fluid samples from 15 (33%) dogs. A filamentous, branching organism was detected cytologically in 18 dogs, 11 of which had sulfur granules. Bacteria isolated from these 18 dogs included Prevotella spp (n = 5), Clostridium spp (4), Bacteroides spp (3), Corynebacterium spp (3), Staphylococcus spp (3), Actinomyces spp (2), Enterobacter spp (2), Fusobacterium spp (2), Pasteurella multocida (2), and Nocardia sp (1). Microbiological evaluation—Pleural-fluid samples for microbiological culture were obtained from 45 dogs by thoracocentesis (n = 29), at surgery (12), or at necropsy (4). The number of isolates recovered from pleural fluid ranged from 0 to 5 isolates/dog (n = 9 and 2 dogs, respectively), with a distribution of 1 (10), 2 (9), 3 (11), or 4 (4) isolates/dog. Thirty-six (80%) dogs had at least 1 bacterial isolate, whereas 26 (58%) dogs had 2 or more isolates. Eleven dogs had both aerobic and obligate anaerobic isolates, whereas 8 other dogs had at least 1 facultative anaerobic isolate. JAVMA, Vol 236, No. 6, March 15, 2010

and gram-positive) were most likely to be susceptible to the following antimicrobials: imipenem (14/15 [93%] isolates), cefotaxime (11/12 [92%]), ceftazidime (11/12 [92%]), gentamicin (38/44 [86%]), enrofloxacin (17/20 [85%]), amikacin (19/23 [83%]), and chloramphenicol (50/61 [82%]). Such isolates were most likely to be resistant to the following antimicrobials: cefazolin (5/10 [50%] isolates), tetracycline (19/44 [43%]), ampicillin (19/46 [41%]), and trimethoprimsulfonamide (16/45 [36%]). Laboratory interpretation suggested aerobic isolates had an overall susceptibility rate of 72% to antimicrobials tested, whereas gram-negative and gram-positive isolates had 67% and 78% susceptibility rates, respectively. Anaerobic isolates had an overall susceptibility rate of 97% to the 6 antimicrobials tested. Treatment—Treatment was noninvasive in 7 (15%) dogs, invasive in 26 (57%) dogs, and surgical in 13 (28%) dogs. All dogs received systemically administered antimicrobials. The reason for choosing noninvasive treatment (thoracocentesis) was usually financial. Reasons for choosing surgical treatment were radiographic or ultrasonographic findings of trapped (compartmentalized) fluid (n = 7), ultrasonographic evidence of an intrathoracic mass (2), insufficient response to medical treatment (2), presence of a thoracic wall lesion (1), or a perforated esophagus (1). Median duration of hospitalization for all dogs was 8 days (range, 1 to 24 days). Initial choice of antimicrobial was empirical in all dogs. The most commonly used systemically administered antimicrobials in the study dogs were trimethoprimsulfonamide (n = 30), enrofloxacin (24), amoxicillin (23), ampicillin (20), and penicillin (10). Forty-one (89%) dogs were systemically treated with multiple antimicrobials in

Table 2—Antimicrobial susceptibility of bacterial isolates recovered from pleural fluid of 46 dogs with pyothorax. Antimicrobial Amikacin Amoxicillin−clavulanic acid Ampicillin Carbenicillin Cefazolin Cefotaxime Cefoxitin Ceftazidime Ceftiofur Cefuroxime Cephalothin Chloramphenicol Clindamycin Enrofloxacin Erythromycin Gentamicin Imipenem Kanamycin Metronidazole Penicillin Piperacillin Tetracycline Ticarcillin Trimethoprim-sulfonamide Overall

All aerobic isolates 19/23 (83) 25/33 (76) 27/46 (59) 7/12 (58) 5/10 (50) 11/12 (92) 7/10 (70) 11/12 (92) 9/13 (69) 8/10 (80) 28/42 (67) 35/45 (78) 15/23 (65) 17/20 (85) 18/24 (75) 38/44 (86) 12/13 (92) 8/9 (89) — 15/22 (68) 10/13 (77) 25/44 (57) 14/19 (74) 29/45 (64) 393/544 (72)

Gram-positive isolates 5/7 (71) 14/16 (88) 19/24 (79) — 2/2 (100) 2/2 (100) 2/2 (100) 2/2 (100) 5/6 (83) 2/2 (100) 18/22 (82) 19/22 (86) 15/23 (65) 11/12 (92) 18/24 (75) 20/23 (87) 5/5 (100) — — 15/22 (68) 2/2 (100) 15/22 (68) 3/3 (100) 14/24 (58) 208/267 (78)

Gram-negative isolates

All obligate anaerobic isolates

14/16 (88) 11/17 (65) 8/22 (36) 7/12 (58) 3/8 (38) 9/10 (90) 5/8 (63) 9/10 (90) 4/7 (57) 6/8 (75) 10/20 (50) 16/23 (70) — 6/8 (75) — 18/21 (86) 7/8 (88) 8/9 (89) — — 8/11 (73) 10/22 (45) 11/16 (69) 15/21 (71) 185/277 (67)

— 6/6 (100) — — — — — — — — — 15/16 (94) 16/16 (100) — — — 2/2 (100) — 15/16 (94) 9/9 (100) — — — — 63/65 (97)

Data are reported as No. susceptible/No. tested (percentage). — = Not applicable. JAVMA, Vol 236, No. 6, March 15, 2010

Scientific Reports



Twenty-seven bacterial genera were identified (plus 1 unidentified isolate) for a total of 87 (45 gram-positive and 42 gram-negative) bacterial isolates. Sixty aerobic isolates were obtained from 27 dogs. Gram-positive aerobic isolates included Corynebacterium spp (n = 10 isolates, including 3 facultative anaerobes), Staphylococcus spp (9, including 2 facultative anaerobes), Streptococcus spp (5, including 2 facultative anaerobes), Bacillus spp (2), Nocardia spp (2, including 1 facultative anaerobe), Arcanobacterium pyogenes (1), Enterococcus sp (1), Streptomyces sp (1), and 1 unidentified filamentous organism. Gram-negative aerobic isolates included Escherichia coli (n = 9, including 4 facultative anaerobes), Klebsiella spp (5, including 2 facultative anaerobes), Acinetobacter spp (3), Pasteurella spp (3), Aeromonas hydrophila (1 facultative anaerobe), Alcaligenes xylosoxidans (1), Capnocytophaga sp (1 facultative anaerobe), Enterobacter sp (1 facultative anaerobe), Kluyvera sp (1), Pseudomonas sp (1), Serratia marcescens (1 facultative anaerobe), and Xanthomonas maltophilia (1). Twenty-seven obligate anaerobic isolates were obtained from 16 dogs. Gram-positive obligate anaerobic isolates included Clostridium spp (n = 5), Peptostreptococcus spp (4), Actinomyces spp (3), and Propionibacterium acnes (1). Gramnegative obligate anaerobic isolates included Prevotella spp (n = 7), Bacteroides spp (4), and Fusobacterium spp (3). Results of antimicrobial susceptibility testing were available for 65 of 87 (75%) isolates. Data were quantitative, providing mean inhibitory concentrations (n = 36; 19 aerobic and 17 anaerobic isolates), or qualitative, as obtained via the agar disk diffusion method (29; 28 aerobic and 1 anaerobic isolates). Results of antimicrobial susceptibility testing were summarized (Table 2). Bacterial isolates (both gram-negative


combination (n = 37) or in sequence (ie, monotherapy that changed over time; 4). Five (11%) dogs received only 1 antimicrobial agent. A total of 2 antimicrobials were used in 13 dogs, 3 in 5 dogs, 4 in 11 dogs, 5 in 10 dogs, and 7 in 2 dogs, in various combinations. The most commonly used antimicrobial combinations were trimethoprim-sulfonamide and amoxicillin (or ampicillin or amoxicillin−clavulanic acid; n = 20), enrofloxacin and amoxicillin (or ampicillin, amoxicillin−clavulanic acid, or cefazolin; 18), and gentamicin (or amikacin) and ampicillin (or amoxicillin−clavulanic acid; 6). Seven dogs received 2 combinations. Antimicrobials used most commonly as single agents were trimethoprim-sulfonamide (n = 2) or penicillin (2). Local (pleural) antimicrobial administration was used in 9 (20%) dogs (7 dogs in the invasive group and 2 dogs in the surgical group). Antimicrobials used were penicillin (n = 5), ampicillin (3), and trimethoprim-sulfonamide used in combination with ampicillin (1). Microbial culture and antimicrobial susceptibility testing results prompted a change in antimicrobial treatment in 12 of the 34 (35%) dogs that survived hospitalization, which was prompted by reported resistance to the empirically selected antimicrobial or isolation of an organism not within the expected spectrum of the selected antimicrobial (eg, an anaerobe). Combination antimicrobial treatment based on susceptibility testing results was used in these 12 dogs. When thoracocentesis was performed, it was in nonanesthetized dogs and a hypodermic needle or over-the-needle catheter was used. In those situations, multiple attempts at pleural evacuation were usually necessary. For dogs in the invasive and surgical groups (n = 39), when a thoracostomy tube was placed, it was in a sedated or anesthetized dog. A trocar-type tube was placed unilaterally in 22 dogs (13 on the right, 7 on the left, and 2 on alternate sides) and bilaterally in 17 dogs. Overall median duration of thoracostomy tube placement was 8 days (range, 1 to 18 days; Table 3). Pleural lavage was performed in 32 of the 39 (82%) dogs with a thoracostomy tube by instillation of warm lac-

tated Ringer’s solution (n = 17) or warm saline (0.9% NaCl) solution (15) at a dose of 10 to 20 mL/kg via a thoracostomy tube, with retrieval of the liquid approximately 30 minutes later. Heparin (10 U/mL) was added to the warm crystalloid solution immediately prior to lavage in 25 dogs (Table 3). Volume of retrieved fluid was usually slightly less than that infused for each lavage; however, because of the pleural effusion, only 2 dogs had less fluid retrieved than infused over the hospitalized period. Frequency of pleural lavage ranged from 2 to 5 times/d, as determined by the attending clinician. Overall median duration of pleural lavage treatment was 6 days (range, 2 to 16 days). Thoracotomy enabled diagnostic assessment (volume and character of fluid and detection of proliferative tissue, fibrosis, and foreign material) and therapeutic intervention (removal of pleural fluid [n = 13], debridement of proliferative tissue [13], intraoperative pleural cavity flushing [13], partial or complete pulmonary lobectomy [3], removal of foreign material [1], and closure of an esophageal perforation [1]). Ventral (median sternotomy; n = 8) or lateral (7) thoracotomy was used. Two dogs had 2 surgical procedures separated by 1 month (lateral thoracotomy and median sternotomy) or 8 months (median sternotomies). Outcome—Three dogs were lost to follow-up, and 2 that died of causes unrelated to pyothorax were censored. Overall short-term (ie, discharged alive) survival rate was 74% (34 dogs), and the difference in rates among the 3 groups was significant (P = 0.009), with the noninvasive group having the lowest rate (29%; Table 3). Twelve dogs died while hospitalized, and all deaths were related to the pyothorax. Seven of these dogs died at various stages of treatment, including 3 dogs with complications associated with anesthesia and thoracostomy tube placement; the other 5 dogs were euthanatized because of a perceived poor prognosis. Necropsy was performed in 11 dogs, and findings confirmed the diagnosis of pyothorax. A higher short- and long-term (12-month) survival rate was associated with pleural lavage (vs no pleural lavage;

Table 3—Features of treatment and outcome in 46 dogs treated for pyothorax by noninvasive (n = 7), invasive (26), or surgical (13) means at a veterinary teaching hospital from 1983 through 2001. Feature




Median duration of thoracostomy tube placement (d) ND 6 (1−17) No. of antimicrobials used/dog 2 (1−5) 3 (1−7) No. treated with heparin ND 15 No. treated with pleural lavage ND 22 Median duration of pleural lavage (d) ND 5 (2−16) Median duration of hospitalization (d) 1 (1−22) 8 (1−20)

9 (3−18) 4 (2−7) 10 10

No. (%) discharged alive*

12 (92)

2 (29)

20 (77)

No. (%) alive at 12 mo 2 (29) 17 (71)† Median duration of follow-up (mo) 64 (60−68) 44 (1−129) No. (%) alive at follow-up 1 (14) 12 (46) No. (%) dead at follow-up because of pyothorax 0 2 (8) No. (%) dead at follow-up because of other reasons 1 (14) 6 (23)

8 (3−12) 11 (4−24)

7 (70)‡ 18 (1−83) 7 (54) 2 (15) 3 (23)

Values in parentheses are ranges unless stated otherwise. *Value is significantly (P  0.05) different among the groups. †Two dogs were censored. ‡Three dogs were censored. ND = Not done. The noninvasive group included dogs treated with thoracocentesis, the invasive group included those treated with a thoracostomy tube or tubes with (n = 22) or without (4) pleural lavage, and the surgical group included those treated via a thoracotomy and thoracostomy tube or tubes with (10) or without (3) pleural lavage.


Scientific Reports

JAVMA, Vol 236, No. 6, March 15, 2010

Discussion In the present study, outcome associated with various treatments for pyothorax in dogs was evaluated. The signalment of the study dogs was typical of that in other studies,3,4,6–8 with male dogs of performance breeds predominating. Although a wide range of ages was evident, median age at initial evaluation also corresponded to that of other studies.6–8 CliniJAVMA, Vol 236, No. 6, March 15, 2010

cal signs at admission to the hospital were reflective of respiratory compromise and systemic infectious disease. The point at which onset of pyothorax occurred was often difficult to identify because initial clinical signs were commonly nonspecific to pleural cavity disease. An initiating event for development of pyothorax was identified in 22% of dogs in the present study, with foreign body within or near the pleural cavity and trauma being the most common events. This finding was similar to that of other investigators.3 Aspiration pneumonia was identified as a cause of pyothorax in 1 dog during hospitalization by means of serial thoracic radiographs obtained before and after the aspiration event. To the authors’ knowledge, such a cause of pyothorax has not been reported previously in dogs. Fewer dogs in the present study (78%) had bacteria isolated from pleural fluid than did dogs in other studies9,11 (100% and 92%). The distributions of aerobic and anaerobic bacterial isolates recovered here had similarities to those of other studies.3,6–8,11–13 Aerobic isolates predominated, with somewhat fewer anaerobic organisms isolated than reported elsewhere.6,7,11 Pleural fluid samples from 35% of the study dogs contained obligate anaerobic bacteria, compared with 60% of samples in another study.11 Similarly, only 14% of study dogs with positive microbial culture results had a mixture of obligate and facultative anaerobic isolates, compared with 36% in that other study.11 Recovery of anaerobic isolates in the present study appeared more common in later medical records, perhaps reflecting improved microbiological laboratory isolation methods. No impact of Gram stain results on short-term survival rate was identified; however, dogs that had pleural fluid containing a combination of aerobic and anaerobic isolates did better long term. Perhaps this reflected the effect of combination antimicrobial treatment or pleural lavage. Only 13% of dogs in the present study had pleural fluid from which Nocardia spp, Actinomyces spp, or an unidentified filamentous organism was isolated, despite detection of filamentous bacteria during cytologic evaluations of pleural fluid or tissue in 39% of dogs. Plant material was histologically evident in pleural tissue from 1 of 3 dogs from which Actinomyces spp were recovered, which has been reported elsewhere.9,11 The distributions of various types of isolates from dogs in our study differed from those in another study,11 with E coli, Staphylococcus spp, and Corynebacterium spp detected more commonly and Bacteroides spp, Peptostreptococcus spp, and Porphyromonas spp detected less commonly in our group of dogs. Bacteria were detected either cytologically or microbiologically, but not both, in 30% of dogs in the present study. Timing differences between collection of cytologic and microbiological samples may partly explain this observation. Antimicrobial treatment prior to collection of samples for microbiological culture also may have been an influence. Antimicrobial susceptibility testing results revealed a rather resistant bacterial population. Aerobic isolates had an overall susceptibility rate of 72% to antimicrobials tested. On the basis of the results of this study, empirical selection of a single antimicrobial for use in a dog with pyothorax had a risk of inefficacy of 35%. Because of the potential for recovery of multiple isolates from a dog, combination antimicrobial treatment in dogs with pyothorax seems prudent. Although the role of antimicrobial resistance in the long-term outcome was not evaluated in the present study, a lower degree of therapeutic success would have been expected when antimicrobials to which isolates were resistant were used. Imipenem was the only antimicrobial for which susceptibility was tested in at least 15 isolates, and > 90% of the tested isolates were Scientific Reports



P = 0.003 and P = 0.002, respectively) and heparin treatment (vs no heparin treatment; P < 0.001 and P = 0.004, respectively). Dogs for which heparin was included in the lavage had a higher short-term survival rate, compared with dogs that did not have heparin in the lavage (P = 0.025). Short-term survival rates were not different among dogs from which gram-negative (56%), gram-positive (91%), mixed (69%), or no (78%) bacteria were isolated (P = 0.307) or among dogs from which aerobic (63%), anaerobic (67%), or a combination (91%) of isolates were recovered (P = 0.283) from pleural fluid. Shortterm survival rates also were not different for dogs that had large (76%), moderate (72%), or small (71%) volumes of pleural fluid at initial radiographic evaluation (P > 0.999). Follow-up information was available for all 34 dogs that survived hospitalization. Median duration of follow-up for the 34 dogs was 38 months (range, 1 to 129 months). Twentyseven dogs had at least a 12-month follow-up. Of the 7 dogs that were followed up for < 12 months after discharge from the hospital, 3 dogs were alive at last contact, 3 dogs died of causes related to pyothorax 1 to 10 months after hospitalization, and 1 dog died of an unrelated cause at 10 months. Overall long-term survival rate was 63% (26/41 uncensored dogs), and rates did not differ among the groups (P = 0.139; Table 3). Long-term survival rates did not differ among dogs from which gram-negative (44%), gram-positive (70%), mixed (64%), or no (75%) bacteria were recovered from pleural fluid (P = 0.618). Dogs with a combination of anaerobic and aerobic isolates (90%) had a higher longer-term survival rate than did those with aerobic (44%) or anaerobic (60%) isolates (P = 0.048). Long-term survival rates were not different for dogs with large (67%), moderate (59%), or small (67%) volumes of pleural fluid at initial radiographic evaluation (P = 0.282). Recurrence of pyothorax was assessed in 35 dogs, 34 of which survived hospitalization and 1 of which was the dog that had been euthanatized in the hospital because of recurrent pyothorax. Dogs with no recurrence and a follow-up of < 12 months were censored (n = 5). Twenty-three dogs had no recurrence, and 7 dogs had a recurrence. Six dogs had 1 recurrence at 1 (n = 2), 6 (1), 10 (1), 37 (1), or 69 months (1) after discharge from the hospital, and 1 dog had 2 recurrences at 1 and 8 months after discharge. Existence of recurrent pyothorax was determined by the referring veterinarian (n = 3), the teaching hospital (2), or owner survey (2). Of the 7 dogs with recurrence, 3 dogs were in the invasive group and 4 were in the surgical group. Microbiological data at recurrence were available for only 3 of 7 dogs; hence, ability to distinguish between reinfection and relapse was limited. One dog had the same type of isolate at recurrence at 1 month, and 2 dogs had different types of isolates at recurrence at 1, 8, or 10 months. Five of 7 dogs that had a recurrence died or were euthanatized as a result of the recurrence. Of the 2 dogs that survived, 1 was still alive 9 months afterward (78 months after discharge from the hospital) and 1 died of unrelated causes 83 months after the second recurrence.


susceptible to the drug. Two of the most commonly used antimicrobials in our study (trimethoprim-sulfonamide and ampicillin) were associated with a low proportion of susceptible isolates (64% and 59%, respectively). Pleural lavage in the dogs of this report was well tolerated, and it provided therapeutic benefits. Potential beneficial effects of pleural lavage include modification of pleural pH and reduction of pleural fluid viscosity, particulate matter, and bacterial numbers. Although frequency of pleural lavage was variable among the study dogs, it was continued until gross and cytologic characteristics of the pleural fluid improved. Inability to recover the entire volume of pleural lavage fluid occurred often. Potential reasons for pleural lavage fluid retention include resorption and compartmentalization. It is estimated that approximately 25% of pleural lavage fluid is absorbed,14,15 but less absorption of lavage fluid appeared to occur in dogs of the present study. No adverse effect of retention of pleural lavage fluid was noticed. Despite the beneficial effects of pleural lavage, unresolved issues concerning the process include the appropriate number of thoracostomy tubes to use and the frequency and duration of pleural lavage. The successful use of drainage via 1 thoracostomy tube and systemic antimicrobial use in 15 dogs has been reported.5 No beneficial effect attributable to use of 2 thoracostomy tubes was identified in our study, in part, because of the small number of dogs evaluated. Although the use of continuous pleural evacuation with systemically administered antimicrobials to treat pyothorax in dogs has been variably described as ideal or acceptable,4,8 intermittent pleural evacuation was used in dogs of the present study. In another study,8 use of continuous pleural evacuation in 15 dogs was associated with a long-term survival rate (60%) similar to that in our study. Whereas dogs treated with systemically administered antimicrobials and thoracocentesis alone (noninvasive group) had the lowest shortand long-term survival rates, 60% (3/5 dogs) of the dogs in this group that did not survive hospitalization were euthanatized because of a perceived poor prognosis or financial concerns. The benefits of intrapleural heparin administration revealed in the present study have not been reported before for dogs with pyothorax. Heparin interrupts pleural adhesion formation in rabbits when used immediately after tetracycline administration.16 Heparin use in naturally occurring pyothorax has been reported thus far only in cats17–19; however, heparin use appears to be of benefit when used to treat experimentally induced peritonitis in dogs.20 The heparin dose used intrapleurally in the present study was identical to that used in heparinized saline solutions for flushing IV catheters. No adverse effects associated with heparin administration were recorded, although blood coagulation profiles were not assessed. Pyothorax recurrence was evident in < 20% of dogs at risk in our study; however, recurrence was associated with a high mortality rate, given that 5 of 7 dogs with recurrence died or were euthanatized. Approximately half of the recurrences took place within 6 months after initial treatment; however, the latest recurrence took place at 69 months. Two dogs that had a recurrence were identified via owner questionnaire, so no supporting data were available. No specific risk factors for recurrence were identified in our study. Although the short-term survival rate was better in dogs that underwent surgery, the long-term survival rate did not appear to be affected by treatment type (noninvasive, invasive, or surgical). Criteria for selecting surgery as a mode of treatment and the timing of surgery varied during the study period; however, thoracotomy appeared to be performed later rather than earlier in the course of the disease. Results of subjective assessment of pleural fluid volume at initial radio662

Scientific Reports

graphic evaluation were not associated with treatment type, although evidence of fluid compartmentalization was. Selection of surgical approach appeared to be influenced by the presence of unilateral pleural fluid or a thoracic wall abscess. Because pyothorax was usually evident bilaterally, simultaneous exposure of both pleural cavities via a median sternotomy may be preferable to a lateral thoracotomy. Also, because proliferative mediastinal tissue was consistently present in the ventral (subphrenic) region in dogs undergoing thoracotomy in our study, access was facilitated by a median sternotomy rather than a lateral thoracotomy. Generally, selection of therapeutic modalities and sequences should be based on defined criteria. Treatment should be instituted early and should be specific to the stage of the disease process.21 Such criteria have yet to be established for pyothorax in dogs. Additional clarification of stages of pyothorax in dogs should help establish criteria for therapeutic decisions. The degree of compartmentalization of fluid within the pleural cavity may be particularly important when selecting a therapeutic approach. Methods to assess degree of pleural cavity compartmentalization in humans include positional (horizontal beam) radiography, ultrasonography, thoracoscopy, computed tomography, and magnetic resonance imaging.21 Only positional radiography, ultrasonography, and thoracoscopy were used in selected dogs of our study. Ultrasonography is useful for fluid detection, volume estimation, and characterization.22 Computed tomography has been used to assess experimentally induced acute pleural effusion in dogs.23 The usefulness of thoracoscopy in the treatment of pediatric pyothorax has been reported24,25; however, its suitability for assessing dogs with pyothorax requires investigation. Thoracoscopy may hold promise for determining the status of the pleural cavity and for assisting therapeutic decision making.26 The main limitations of the present study involved its retrospective nature, long study period, lack of well-defined criteria for treatment selection, and limited data available at recurrence of pyothorax. The effect of study period length was evident in the microbial culture and antimicrobial susceptibility data, with quantitative susceptibility data becoming available in the later medical records. Therapeutic decision making also appeared to evolve over the study period, with 2 thoracostomy tubes used more commonly in later years and surgery performed more commonly in the mid years. Nonetheless, findings suggested that favorable treatment effects were achieved with surgery (for short-term survival) and pleural lavage and heparin treatment (for shortand long-term survival). There appeared to be > 1 effective therapeutic approach for long-term resolution of pyothorax in dogs. More compartmentalized effusions may require more invasive treatment. Findings failed to support the hypothesis that invasive (surgical) versus noninvasive treatment of pyothorax in the dog leads to a better long-term outcome. a.

Statistix, version 7, Analytical Software, Tallahassee, Fla.

References 1. 2. 3. 4.

Marino DJ, Jaggy A. Nocardiosis. A literature review with selected case reports in two dogs. J Vet Intern Med 1993;7:4–11. Steyn PF, Wittum TE. Radiographic, epidemiologic, and clinical aspects of simultaneous pleural and peritoneal effusions in dogs and cats: 48 cases (1982–1991). J Am Vet Med Assoc 1993;202:307–312. Demetriou JL, Foale RD, Ladlow J, et al. Canine and feline pyothorax: a retrospective study of 50 cases in the UK and Ireland. J Small Anim Pract 2002;43:388–394. Bauer T. Pyothorax. In: Kirk RW, ed. Current veterinary therapy IX. Philadelphia: WB Saunders Co, 1986;292–295. JAVMA, Vol 236, No. 6, March 15, 2010


7. 8. 9. 10.

11. 12. 13. 14.

15. Tomlinson J. Review of pyothorax in the feline. Feline Pract 1980;10(6):26–32. 16. Strange C, Baumann MH, Sahn SA, et al. Effects of intrapleural heparin or urokinase on the extent of tetracycline-induced pleural disease. Am J Respir Crit Care Med 1995;151:508–515. 17. Sherding RG. Pyothorax in the cat. Compend Contin Educ Pract Vet 1979;1:247–252. 18. Crane SW. Surgical management of feline pyothorax. Feline Pract 1976;6(2):13–19. 19. Forrester SD. Pleural effusion in the dog and cat. Vet Annu 1980; 30:283–297. 20. Hau T, Simmons RL. Heparin in the treatment of experimental peritonitis. Ann Surg 1978;187:294–298. 21. Stovroff M, Teague G, Heiss KF, et al. Thoracoscopy in the management of pediatric empyema. J Pediatr Surg 1995;30:1211–1215. 22. Stowater JL, Lamb CR. Ultrasonography of noncardiac thoracic diseases in small animals. J Am Vet Med Assoc 1989;195:514–520. 23. Dechman G, Mishima M, Bates JHT. Assessment of acute pleural effusion in dogs by computed tomography. J Appl Physiol 1994;76:1993–1998. 24. Campbell PW III. New developments in pediatric pneumonia and empyema. Curr Opin Pediatr 1995;7:278–282. 25. Poe RH, Marin MG, Israel RH, et al. Utility of pleural fluid analysis in predicting tube thoracostomy/decortication in parapneumonic effusions. Chest 1991;100:963–967. 26. Kovak JR, Ludwig LL, Bergman PJ, et al. Use of thoracoscopy to determine the etiology of pleural effusion in dogs and cats: 18 cases (1998– 2001). J Am Vet Med Assoc 2002;221:990–994.

From this month’s AJVR

Effect of dietary fats with odd or even numbers of carbon atoms on metabolic response and muscle damage with exercise in Quarter Horse–type horses with type 1 polysaccharide storage myopathy Lisa A. Borgia et al Objective—To evaluate effects of fats with odd and even numbers of carbon atoms on muscle metabolism in exercising horses with polysaccharide storage myopathy (PSSM). Animals—8 horses with PSSM (6 females and 2 males; mean ± SD age, 6.3 ± 3.9 years). Procedures—Isocaloric diets (grain, triheptanoin, corn oil, and high-fat, low-starch [HFLS] feed) were fed for 3 weeks each; horses performed daily treadmill exercise. Grain was fed to establish an exercise target, and HFLS feed was fed as a negative control diet. Daily plasma samples were obtained. For each diet, a 15-minute exercise test was performed, and gluteus medius muscle specimens and blood samples were obtained before and after exercise. Results—Feeding triheptanoin, compared with the corn oil diet, resulted in exercise intolerance; higher plasma creatine kinase (CK) activity and concentrations of C3:0- and C7:0-acylcarnitine and daily insulin; and lower concentrations of nonesterified fatty acids (NEFA) and C16:0-, C18:1-, and C18:2-acylcarnitine, without changes in concentrations of plasma glucose or resting muscle substrates and metabolites. Feeding grain induced higher CK activity and insulin concentrations and lower NEFA concentrations than did corn oil or HFLS feed. Feeding grain induced higher glucose concentrations than did triheptanoin and corn oil. In muscle, feeding grain resulted in lower glucose-6-phosphate, higher citrate, and higher postexercise lactate concentrations than did the other diets. Conclusions and Clinical Relevance—Triheptanoin had detrimental effects, reflecting decreased availability of NEFA, increased insulin stimulation of glycogen synthesis, and potential inhibition of lipid oxidation. Long-chain fats are the best dietetic for PSSM. (Am J Vet Res 2010;71:326–336)

JAVMA, Vol 236, No. 6, March 15, 2010

March 2010

See the midmonth issues of JAVMA for the expanded table of contents for the AJVR or log on to for access to all the abstracts.

Scientific Reports




Johnson MS, Martin MWS. Successful medical treatment of 15 dogs with pyothorax. J Small Anim Pract 2007;48:12–16. Frendin J. Pyogranulomatous pleuritis with empyema in hunting dogs. Zentralbl Veterinarmed A 1997;44:167–178. Robertson SA, Stoddart ME, Evans RJ, et al. Thoracic empyema in the dog; a report of twenty-two cases. J Small Anim Pract 1983;24:103–119. Turner WD, Breznock EM. Continuous suction drainage for management of canine pyothorax—a retrospective study. J Am Anim Hosp Assoc 1988;24:485–494. Rooney MB, Monnet E. Medical and surgical treatment of pyothorax in dogs: 26 cases (1991–2001). J Am Vet Med Assoc 2002;221:86–92. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standard. NCCLS M31–A2. 2nd ed. Wayne, Pa: National Committee for Clinical Laboratory Standards, 2003. Walker AL, Jang SS, Hirsh DC. Bacteria associated with pyothorax of dogs and cats: 98 cases (1989–1998). J Am Vet Med Assoc 2000;216:359–363. Jang SS, Breher JE, Dabaco LA, et al. Organisms isolated from dogs and cats with anaerobic infections and susceptibility to selected antimicrobial agents. J Am Vet Med Assoc 1997;210:1610–1614. Hirsh DC, Jang SS. Antimicrobial susceptibility of selected infectious bacterial agents obtained from dogs. J Am Anim Hosp Assoc 1994;30:487–494. Holmberg DL. Management of pyothorax. Vet Clin North Am Small Anim Pract 1979;9:357–362.

Evaluation of outcomes in dogs treated for pyothorax_46 cases