Page 1

J Periodontol • December 2007

Removable Prostheses May Enhance Marginal Bone Loss Around Dental Implants: A Long-Term Retrospective Analysis Moshik Tandlich,* Jakob Ekstein,* Pini Reisman,† and Lior Shapira*

Background: The aim of the study was to retrospectively evaluate marginal bone loss (MBL) around rough-surface dental implants, placed in a private clinic, and to construct a multivariate model based on formerly proposed prognostic variables. Methods: Records of patients who were treated previously with dental implants were reviewed. The patients’ latest annual clinical examinations and radiograms were used for data collection and the calculation of MBL. A patient-based multivariate model was constructed based on two successive iterations of statistical analysis. Results: Eighty-two patients and 265 implants with ‡30 months of follow-up were evaluated. The overall survival rate was 95.8% (2.6% early loss and 1.5% late loss). By evaluating the data with the single implant as a unit of analysis, MBL was correlated with time. Higher MBL values were found in smokers and around implants supporting removable prostheses. In the patient-based analysis, only smoking and the presence of a removable prosthesis correlated with higher values of MBL. Odds ratios for higher rates of MBL were 1.95 and 2.57 for smokers and around removable prostheses, respectively. Neither time nor any of the other suspected variables correlated with higher MBL. Conclusions: The present study corroborated the notion that smoking correlates with higher MBL and implied that implants supporting removable prostheses tend to display more bone loss. Further studies are needed to elucidate the latter finding. J Periodontol 2007;78:2253-2259. KEY WORDS Alveolar bone loss; dental implant; multivariate analysis; partial denture; retrospective study.

* Department of Periodontology, Hadassah Medical Center, Jerusalem, Israel. † Private practice, Hurdegaryp, The Netherlands.


he use of dental implants is a welldocumented treatment modality with a well-known prognosis. Survival rates ranging from 92.4% to 100% have been reported,1,2 depending on the type of prostheses used, their placement schedule, and the follow-up time. Great care is taken to distinguish between the definitions for implant loss (‘‘implant failure’’), implant survival, and implant success,3 which denote different measures for evaluating the dental implant performance. In a systematic review4 of wellcontrolled prospective studies, Berglundh et al. pointed out that the prevalence of implant loss may be anticipated as 5% of cumulative failure rate: 2.5% ‘‘early implant loss’’ and 2% to 3% ‘‘late implant loss.’’ Early implant loss usually is explained by lack of integration due to surgical trauma, anatomic conditions, lack of primary stability, infection, or the operator’s surgical experience.5-7 Late implant loss usually is believed to be caused by high occlusal strain or infection, i.e., periimplantitis,8-10 which manifests itself as a progressive loss of supporting bone around previously integrated implants. The annual rate of marginal bone loss (MBL) was proposed by Albrektsson et al.11 as a predictor of implant failure. They proposed bone loss of 1.2 mm in the first year of service and another 0.2

doi: 10.1902/jop.2007.070113


Removable Dentures May Enhance MBL Around Dental Implants

mm in each succeeding year as thresholds for implant success. Although these thresholds are accepted widely in defining a successful implant,2 there is no consensus regarding the rate of bone loss necessary to define the failure of an implant.3 Nevertheless, it is accepted to deem an implant as ‘‘failed’’ only when it has to be removed from the mouth. Although MBL around dental implants is well documented,12-14 it is not fully understood.15 The process of MBL can be described in two phases: ‘‘early MBL’’ that is related to the connection of the healing abutment or prosthetic unit, which occurs around the first year of service;15 and ‘‘late MBL’’ that may occur over time with the function of the implant in the mouth, years after its prosthetic loading. ‘‘Late MBL’’ seems to be quite common, particularly in patients with no supportive periodontal treatment.16 Different variables emerged in the literature and were proposed as effectors of ‘‘late MBL’’: duration of implant use,2,13,17 type of implant,2,13 the configuration of the prosthetic head and the connected abutments,1,15,18 smoking habits,19 periodontal background,20,21 load of occlusal forces,9,10,22 and bone width at the implant site.14 The aim of the present study was to retrospectively evaluate long-term MBL around rough-surface implants, performed in a private practice setting, and to construct a multivariate explanatory model based on formerly proposed variables, e.g., time, type of prosthetic neck, and prosthetic configuration. MATERIALS AND METHODS The study design was reviewed and approved by the Institutional Review Board of the Hadassah Medical Organization for research on human subjects. We retrospectively evaluated the files of patients who were treated with dental implants‡ and underwent a routine follow-up examination between November 2002 and November 2005. We reviewed the patients’ records for treatment history, prosthetic appliance state, health condition, and habits. We included only patients who had completed ‡30 months of followup since their implant procedure. The latest available radiogram (panoramic or parallel periapical) of the related implants was used for the MBL calculations. We included only patients who had dental implants placed in a standard protocol: all of the implants were placed in a crestal approach, i.e., implant shoulder was flush with the bone level. Implants placed in sites of augmented bone (generated by guided bone regeneration) were not included. The radiographic bone loss around each implant was measured with a 5· scaled magnifying loupe on the approximal implant side presenting the maximal amount of bone loss. Each implant’s actual bone loss 2254

Volume 78 • Number 12

was calculated based on the measured radiographic bone loss and the implant’s known dimensions (radiographic measured bone loss · known implant length/ implant’s radiographic length). The following data were collected: patient-based data, including follow-up time, age, gender, and smoking habits (packs per day, years of smoking, smoking rate in pack-years, and years of smoking cessation); implant-based data, including implant site, implant dimensions, type of prosthetic platform, loading protocol, mobility, inflammation or suppuration, measured radiographic bone loss, number of exposed threads, and type of connectors; and prosthesis-based data comprising type of denture (fixed or removable), lifespan, tooth splinting, cross-arch configuration, cantilevered section, number of crowns per supporting implant, and crown length/implant length ratio. Implants that were removed for any reason (failed to integrate or failed sometime during their use) were recorded as ‘‘lost implants.’’ ‘‘Early implant loss’’ was defined as an implant that failed to integrate prior to restoration, whereas ‘‘late implant loss’’ included implants that were removed after restoration.4 All other implants, provided that they had no sign of inflammation or suppuration, were deemed ‘‘surviving implants.’’4 Data were collected using a database-driven application§ specifically programmed for this study. All statistical analyses were done using a standard personal computer statistical application.i All surgical procedures were performed by a single operator for each type of implant: internal hex or external hex. The clinical data collection was carried out by an independent surveyor, and a second one performed the radiographic measurements and data compilation. Statistical Analysis Descriptive statistics were used to report the demographics of the treated population and the implants’ survival rates. To achieve a multivariate explanatory model based on the patient as the statistical unit of analysis, two successive iterations of data analysis were used: an implant-based statistical analysis to identify putative significant predictors and a patientbased analysis based on these predictors. Initially, an implant-based analysis of variance (ANOVA) was performed. Each implant’s MBL served as the dependent variable of interest for the analysis, using each suspected variable as an independent one (random effect). Continuous variables were evaluated separately with a linear regression analysis. In the second iteration, a patient-based analysis was performed: each patient’s mean MBL was fitted ‡ Biocom, MIS Implants, Shlomi, Israel. § MS Access, Microsoft, Redmond, WA. i Statistica v6, StatSoft, Tulsa, OK.

Tandlich, Ekstein, Reisman, Shapira

J Periodontol • December 2007

into a general regression model, based on the variables found to be significant in the preceding data analysis iteration. Finally, a logistic regression analysis was performed to evaluate the risk factors for a high level of bone loss, based on each patient’s average bone loss: an arbitrary average MBL rate was calculated for every patient (mean MBL/follow-up time). The logistic regression model was used to evaluate the odds ratios (ORs) for extreme rates of MBL based on the significant variables identified previously and the upper and lower quartiles of MBL rates. We assumed a normal distribution of MBL for all statistical models, and this assumption was verified using Kolmogorov-Smirnov testing. RESULTS Demographics Seventy percent of patients responded to our annual recall regimen and completed ‡30 months of followup; therefore, all 82 patient files, with 265 implants, were evaluated. The treated population included a similar number of men and women (40 and 42, respectively) with an average age of 60 years. All patients were healthy or had a stable, treated, chronic disease that did not have contraindications to dental implant placement (Table 1).

The implants were restored by 63 single-unit crowns, 52 multiple-unit fixed partial dentures, and 22 removable dentures (ball-retained overdentures). The fixed prosthetic appliances were connected to the implants by custom-made cast-work abutments (114) or by machine-made abutments (137). Seventeen patients were smokers (Table 2), and 10 other patients had stopped smoking ‡4 years prior to the implant procedure. Seven implants failed to integrate (five internal and two external hex implants) and were regarded as ‘‘early loss,’’ presenting an overall 2.6% failure rate. Four other implants were lost over time, after being restored, and were regarded as ‘‘late loss’’ (Table 3). Three implants never were restored, and thus were termed ‘‘sleeping implants.’’ These were regarded as surviving implants because there were no signs of inflammation or discomfort related to them.3 Implant-Based Analysis By evaluating the data for each implant’s MBL as the variable of interest, only smoking and the type of prosthesis were found to be statistically significant modifiers of bone loss. No variable had an effect on the amount of MBL around the implants (Table 4), Table 2.

Smokers’ Demographics Smoking Status

Table 1.

Demographic Description of the Study Population Category Follow-up period (months) Average Minimum Maximum

N 56.2 30 112

Gender Male Female

40 42

Chronic disease condition Osteoporosis Diabetes Other condition

2 2 17

Prostheses FPD (single units) RPD Type of abutments Cast abutment Milled abutment

115 (63) 22 114 137

FPD = fixed partial denture; RPD = removable partial denture.


Current smokers (cigarettes/day) <5 5 to 9 10 to 14 15 to 20

6 4 3 4

Non-smokers Past smokers Never smokers

10 55

Table 3.

Cumulative Implant Survival Rates in the Study Population Implant Category



Early loss



Late loss



Sleeping implants



Functional implants






4.2% loss rate

95.8% survival rate


Removable Dentures May Enhance MBL Around Dental Implants

Volume 78 • Number 12

Table 4.

Statistical Models for MBL Around Implants Based on the Implant as Unit of Analysis (ANOVA) and the Patient’s Mean MBL as the Unit of Analysis (general regression model) Implant Based

Patient Based


Degrees of Freedom


P Value


P Value







Follow-up time



Sextant in the mouth




Internal or external hex




Cast or milled abutment










Removable or FPD






Smoking and RPD




Cantilevered prosthesis




Cross-arched prosthesis




Implant to tooth splinting prosthesis





– = not applicable. This variable did not enter into the relevant statistical model. FPD = fixed partial denture; RPD = removable partial denture.

Figure 1.

Figure 2.

Linear regression analysis between MBL and follow-up time, using the single implant as the unit of analysis. A statistically significant correlation (P = 0.003) was found between time and MBL around the implants.

Linear regression analysis between each patient’s mean MBL and follow-up time, using the patient as the unit of analysis. No statistically significant correlation was found between these variables.

including the type of implant (internal or external hex), the type of abutment (custom made or machined), the presence of a cantilevered unit over the implant, or any other suspected variable. Figure 1 shows the linear regression analysis between the MBL found around the implants and the time of function. Although the model was statistically significant (P <0.05), the correlation was weak (r2 = 0.034). 2256

Patient-Based Analysis Each patient’s mean MBL was fitted into a general regression model as the dependent variable, with the mean follow-up time, smoking status, and type of prosthesis as the dependent variables. The patient’s mean follow-up time failed to show any statistically significant correlation with MBL (Table 4, Fig. 2). Again, in this analysis, smoking and the presence of a removable prosthesis correlated with greater MBL (Table 4). Using logistic regression analysis, smoking

J Periodontol • December 2007

Tandlich, Ekstein, Reisman, Shapira

and the presence of a removable prosthesis were associated with higher mean rates of MBL (ORs of 1.95 and 2.57, respectively) (Table 5, Fig. 3).

1.2 mm in the first year and a maximum of an additional 0.2 mm in every succeeding year were proposed, different values have emerged in the literature.2 We found no correlation between bone loss and the duraDISCUSSION tion of service when we used the patient’s mean MBL In the present retrospective analysis, we examined the as the unit of analysis. The inconsistent reports of the long-term survival of implants that were placed ‡30 annual rate of MBL may be explained by the study demonths earlier and the degree of MBL found around sign (prospective versus retrospective) with all of its them. Different variables related to the patients’ medimplications: the examined population, the type of ical history, smoking habits, prostheses, and surgical statistical methods, and controlled and confounded procedures were collected and analyzed, using statisvariables. It is conceivable that MBL may progress tics based on the patient as the unit of analysis.23,24 with time; however, its pattern may not display a continuous regression,2 a fact that may explain our The analysis of implant performance in such a design allowed the investigation of a diverse group of varifindings. ables that, according to the literature,13,15,17 may inWe demonstrated that smoking was a statistically significant predictor of MBL. This is in accordance fluence the extent of MBL and the success rate of the with other studies25,26 that demonstrated similar corimplants. Smoking and the presence of a removable prostherelation with higher rates of MBL. sis were the only two variables found to be statistically In an attempt to define the ‘‘population at risk,’’ we significant modifiers of MBL. Neither duration of use identified the patients with high or low rates of MBL nor any of the other suspected variables influenced based on the higher and lower quartile of the average the amount of MBL. MBL of the study population (>0.07 or <0.03 mm per Previous studies2,11,13,17 described bone loss around month). These thresholds were used for the statistical analysis to try to find differences in the study populaimplants that progresses over time. Although rates of tion with high or low rates of bone loss. In our study Table 5. population, we found an OR of 1.95 for the relationship between smoking and a high risk for having a Logistic Regression Between the Presence high rate of bone loss. of Removable Prosthesis, Smoking, and A new finding of the present study was that MBL Outer Quartiles of MBL Rates around implants that retained overdentures was higher than around implants that supported fixed parOR (95% confidence tial dentures. We demonstrated a statistically signifiinterval) P Value cant correlation between higher values of MBL and the presence of a removable prosthesis in the implantSmoker 1.95 (0.05 to 1.29) 0.04 and patient-based statistics. Furthermore, an OR of Presence of RPD 2.57 (0.33 to 1.55) 0.002 2.57 was found for the relationship between higher rates of MBL and the presence of a removable prostheRPD = removable partial denture. sis. In a prospective case series of implant-supported removable partial dentures, Merickske-Stern et al.27 reported an overall high survival rate of implants presenting with a low MBL rate of 0.7 mm. In another retrospective analysis28 and in a long-term prospective analysis,29 there was no difference in MBL around implants that retained removable prostheses with barsplinted or non-splinted implants, i.e., Dolder-bar versus ball attachment denture retainers, in terms Figure 3. of the reported bone loss or the imBox plot of the calculated bone loss rate around implants in smokers and non-smokers (left) and around implants supporting removable and fixed partial dentures (right). Notice the comparable plant success rates. Neither study impact of both parameters on MBL rate. BLRate = bone loss rate in millimeters per month; could elucidate the difference in boxes = SE; whiskers = 95% confidence interval; RPD = removable partial denture; FPD = fixed MBL around implants retaining repartial denture. movable partial dentures compared 2257

Removable Dentures May Enhance MBL Around Dental Implants

to implants supporting fixed appliances.2,17,30-32 Further research is needed to clarify the impact of a removable prosthesis on the rate of MBL compared to the rate around implants retaining fixed partial dentures. CONCLUSIONS Within the limitations of this retrospective analysis, smoking correlated with higher rates of MBL. Implants that retained removable prostheses tended to display more MBL compared to implants that supported fixed prostheses. Further controlled prospective studies are needed to elucidate the impact of removable prostheses and their retentive elements on the amount of MBL around dental implants. ACKNOWLEDGMENT This study was supported by a research grant from MIS Implants, Shlomi, Israel. The authors report no conflicts of interest related to this study. REFERENCES 1. Chen ST, Wilson TG Jr., Hammerle CH. Immediate or early placement of implants following tooth extraction: Review of biologic basis, clinical procedures, and outcomes. Int J Oral Maxillofac Implants 2004;19(Suppl.): 12-25. 2. Schwartz-Arad D, Kidron N, Dolev E. A long-term study of implants supporting overdentures as a model for implant success. J Periodontol 2005;76:1431-1435. 3. Lang NP, Berglundh T, Heitz-Mayfield LJ, Pjetursson BE, Salvi GE, Sanz M. Consensus statements and recommended clinical procedures regarding implant survival and complications. Int J Oral Maxillofac Implants 2004;19(Suppl.):150-154. 4. Berglundh T, Persson L, Klinge B. A systematic review of the incidence of biological and technical complications in implant dentistry reported in prospective longitudinal studies of at least 5 years. J Clin Periodontol 2002;29(Suppl. 3):197-212. 5. Kronstrom M, Svenson B, Hellman M, Persson GR. Early implant failures in patients treated with Branemark System titanium dental implants: A retrospective study. Int J Oral Maxillofac Implants 2001;16:201-207. 6. Esposito M, Hirsch JM, Lekholm U, Thomsen P. Biological factors contributing to failures of osseointegrated oral implants. (I). Success criteria and epidemiology. Eur J Oral Sci 1998;106:527-551. 7. Esposito M, Hirsch JM, Lekholm U, Thomsen P. Biological factors contributing to failures of osseointegrated oral implants. (II). Etiopathogenesis. Eur J Oral Sci 1998;106:721-764. 8. Quirynen M, De Soete M, van Steenberghe D. Infectious risks for oral implants: A review of the literature. Clin Oral Implants Res 2002;13:1-19. 9. Isidor F. Loss of osseointegration caused by occlusal load of oral implants. A clinical and radiographic study in monkeys. Clin Oral Implants Res 1996;7:143-152. 10. Kitamura E, Stegaroiu R, Nomura S, Miyakawa O. Biomechanical aspects of marginal bone resorption around osseointegrated implants: Considerations based on a three-dimensional finite element analysis. Clin Oral Implants Res 2004;15:401-412. 2258

Volume 78 â&#x20AC;˘ Number 12

11. Albrektsson T, Zarb G, Worthington P, Eriksson AR. The long-term efficacy of currently used dental implants: A review and proposed criteria of success. Int J Oral Maxillofac Implants 1986;1:11-25. 12. van Steenberghe D, Klinge B, Linden U, Quirynen M, Herrmann I, Garpland C. Periodontal indices around natural and titanium abutments: A longitudinal multicenter study. J Periodontol 1993;64:538-541. 13. Manz MC. Factors associated with radiographic vertical bone loss around implants placed in a clinical study. Ann Periodontol 2000;5:137-151. 14. Spray JR, Black CG, Morris HF, Ochi S. The influence of bone thickness on facial marginal bone response: Stage 1 placement through stage 2 uncovering. Ann Periodontol 2000;5:119-128. 15. Oh TJ, Yoon J, Misch CE, Wang HL. The causes of early implant bone loss: Myth or science? J Periodontol 2002;73:322-333. 16. Roos-Jansaker AM, Lindahl C, Renvert H, Renvert S. Nine- to fourteen-year follow-up of implant treatment. Part II: Presence of peri-implant lesions. J Clin Periodontol 2006;33:290-295. 17. van Steenberghe D, Quirynen M, Naert I, Maffei G, Jacobs R. Marginal bone loss around implants retaining hinging mandibular overdentures, at 4-, 8- and 12-years follow-up. J Clin Periodontol 2001;28:628-633. 18. Hartman GA, Cochran DL. Initial implant position determines the magnitude of crestal bone remodeling. J Periodontol 2004;75:572-577. 19. Wennstrom J, Zurdo J, Karlsson S, Ekestubbe A, Grondahl K, Lindhe J. Bone level change at implantsupported fixed partial dentures with and without cantilever extension after 5 years in function. J Clin Periodontol 2004;31:1077-1083. 20. Roos-Jansaker AM, Lindahl C, Renvert H, Renvert S. Nine- to fourteen-year follow-up of implant treatment. Part I: Implant loss and associations to various factors. J Clin Periodontol 2006;33:283-289. 21. Karoussis IK, Salvi GE, Heitz-Mayfield LJ, Bragger U, Hammerle CH, Lang NP. Long-term implant prognosis in patients with and without a history of chronic periodontitis: A 10-year prospective cohort study of the ITI Dental Implant System. Clin Oral Implants Res 2003;14:329-339. 22. Duyck J, Ronold HJ, Van Oosterwyck H, Naert I, Vander Sloten J, Ellingsen JE. The influence of static and dynamic loading on marginal bone reactions around osseointegrated implants: An animal experimental study. Clin Oral Implants Res 2001;12:207-218. 23. Petrie A, Bulman JS, Osborn JF. Further statistics in dentistry. Part 2: Research designs 2. Br Dent J 2002; 193:435-440. 24. Petrie A, Bulman JS, Osborn JF. Further statistics in dentistry. Part 7: Repeated measures. Br Dent J 2003; 194:17-21. 25. Roos-Jansaker AM, Renvert H, Lindahl C, Renvert S. Nine- to fourteen-year follow-up of implant treatment. Part III: Factors associated with peri-implant lesions. J Clin Periodontol 2006;33:296-301. 26. Doundoulakis JH, Eckert SE, Lindquist CC, Jeffcoat MK. The implant-supported overdenture as an alternative to the complete mandibular denture. J Am Dent Assoc 2003;134:1455-1458. 27. Mericske-Stern R, Oetterli M, Kiener P, Mericske E. A follow-up study of maxillary implants supporting an overdenture: Clinical and radiographic results. Int J Oral Maxillofac Implants 2002;17:678-686.

J Periodontol â&#x20AC;˘ December 2007

28. Narhi TO, Hevinga M, Voorsmit RA, Kalk W. Maxillary overdentures retained by splinted and unsplinted implants: A retrospective study. Int J Oral Maxillofac Implants 2001;16:259-266. 29. Naert I, Alsaadi G, van Steenberghe D, Quirynen M. A 10-year randomized clinical trial on the influence of splinted and unsplinted oral implants retaining mandibular overdentures: Peri-implant outcome. Int J Oral Maxillofac Implants 2004;19:695-702. 30. Attard NJ, Zarb GA. Long-term treatment outcomes in edentulous patients with implant overdentures: The Toronto study. Int J Prosthodont 2004;17: 425-433. 31. Attard NJ, Zarb GA. Long-term treatment outcomes in edentulous patients with implant-fixed prostheses:

Tandlich, Ekstein, Reisman, Shapira

The Toronto study. Int J Prosthodont 2004;17: 417-424. 32. Palmqvist S, Owall B, Schou S. A prospective randomized clinical study comparing implant-supported fixed prostheses and overdentures in the edentulous mandible: Prosthodontic production time and costs. Int J Prosthodont 2004;17:231-235. Correspondence: Dr. Lior Shapira, Department of Periodontology, Faculty of Dental Medicine, Hadassah Medical Center, P.O. Box 12272, Jerusalem, Israel. Fax: 972-26438705; e-mail: Submitted February 26, 2007; accepted for publication July 5, 2007.


shapira tandlich 2007 perio biocom long term  
shapira tandlich 2007 perio biocom long term  

Removable Prostheses May Enhance Marginal Bone Loss Around Dental Implants: A Long-Term Retrospective Analysis