Prevalence and risk indicators of peri-implantitis after 8 to 10 years of function

de sobrevivência; peri-implantite; fatores de risco. Abstract Introduction: The investigation of peri-implant diseases risk indicators helps to prevent and target treatment techniques. Objective: The aim of this cross-sectional study was to determine the occurrence of peri-implantitis and its potential risk indicator factors, besides to assess the long-term success and survival rates of dental implants after to 10 years of function. Material and method: For this, fifty individuals who had received their implant-supported rehabilitation between 2003 and 2005 were included. Data regarding demographics, medical and dental history were collected and a complete clinical examination was performed. Multivariate analysis was used to identify potential risk indicator factors related to the occurrence of peri-implantitis. Overall, 211 implants had been placed; 197 were in function, 9 were still and had lost. Result: Success and survival rates were 81.5% and Peri-implant mucositis 77.1% diagnosed individuals with osteoporosis = on = with higher plaque (OR (OR peri-implantitis. Conclusion: Through these results, our study suggests that osteoporosis and generalized periodontal/peri-implant mucosa inflammation increase the likelihood of peri-implantitis. Descriptors: Dental implants; survival rate; peri-implantitis; risk factors.


INTRODUCTION
The use of dental implants for oral rehabilitation of patients partially or completely edentulous has become a widely accepted treatment modality. Despite the high success and survival rates of dental implants, early and late failures may occur. Early failures have been associated with inadequate surgical technique, impaired healing, and inadequate occlusal load distribution. Late failures have been linked to peri-implant infection and occlusal overload 1 .
The epidemiology of peri-implantitis is largely unknown, and it has been greatly impacted by the absence of an established diagnostic criteria 2 . Recently, a new classification scheme was performed for periodontal and peri-implant diseases and conditions 3 . This was important to define the diagnostic parameters of these diseases, reducing the results bias of future studies and facilitating disease identification and treatment. Peri-implant diseases are characterized by inflammation of the surrounding peri-implant tissues and subsequent progressive loss of supporting bone. This condition has been associated with patients with a history of periodontal disease and poor plaque control 3 . Peri-implant mucositis is a reversible inflammatory condition, which occurs in the soft tissues around implants. There is strong evidence that plaque induce peri-implant mucositis. This disease is characterized by bleending on probing and visual signs of inflammation 3 . In a recent meta-analysis 4 forty-seven studies were selected and prevalences of peri-implant diseases were analyzed. Results showed weighted mean implant-based and subject-based peri-implantitis prevalences of 9.25% and 19.83% respectively. Regarding periimplant mucositis, weighted mean implant-based and subjectbased were 29.48% and 46.83% respectively, clearly underscoring the magnitude of the problem.
Several environmental, behavioral, systemic and oral health factors have been associated with the establishment and progression of peri-implantis, including smoking, history of periodontitis, poor oral hygien, and diabetes 4,5 . Other factors such as genetic factors, osteoporosis, occlusal overload, lack of keratinized mucosa, and implant surface roughness have also been investigated [6][7][8][9] . Whereas the amount of evidence linking some of these factors to peri-implantitis has recently increased, proper long-term risk assessment is still scarce in the literature and presents very heterogeneous results.
The aim of this study was to determine the prevalence and risk indicators for peri-implantitis 8 to 10 years after loading. In addition, the long-term success and survival rates of the osseointegrated dental implants were assessed.

Study Design and Sample
A total of 495 partially-edentulous individuals treated with dental implants at the School of Dentistry at Araraquara (UNESP) were invited to participate in this cross-sectional study. They received oral rehabilitation between 2003 and 2005, which allowed the dental implants to be 8 to 10 years in function. Individuals who received at least one implant were eligible to be included in the study. The exclusion criteria were individuals that received periodontal treatment and/or antibiotics/anti-inflammatory therapy within the last three months. The protocol of the present study was approved by the Ethics in Human Research Committees of the School of Dentistry at Araraquara (Protocol number CAAE 07513812.3.0000.5416). Written informed consent was obtained from each participant. This study was in accordance with the Helsinki Declaration of 1975, as revised 2004. STROBE guidelines were followed.
Eligible participants were invited to a follow-up visit by telephone calls. Subjects who failed to respond to three consecutive calls in separated occasions were considered uncontactable and were not considered.

Interview and Clinical Examination
At the 8-10-year follow-up, information such as age, medical history, medications and smoking habits was obtained during an interview with patients using a structured questionnaire. In addition, the number of installed implants and region, characteristics of implants and prosthetic rehabilitation aspects were collected from the patients` records.
The clinical examination was carried out by a single trained and calibrated examiner (L.G.N.). Duplicate exams were done 48 hours apart in 10% of the sample, and the intra-examiner correlation was 0.80 (Wilcoxon test p> 0.05; Spearman correlation r = 0.81). The following parameters were recorded during the clinical exam: plaque index (PI); gingival index (GI); probing pocket depth (PD); bleeding on probing (BOP); clinical attachment level (CAL), suppuration. PI, GI, BOP and suppuration were registered as presence or absence. These parameters were assessed in four sites per implant or tooth and the other parameters in six sites. A North-Carolina periodontal probe (Hu-friedy®, Chicago, IL, USA) was used for teeth, and a Colorvue® plastic periodontal probe (Hu-friedy®) was used for implants.
A panoramic radiography (Dentsply Sirona®, Bonn, Rhineland, Germany) was obtained for each participant for general oral health assessment. Digital periapical radiographs (Saevo®, Ribeirão Preto, São Paulo, Brazil) using a parallel and long cone technique were taken for dental implants. All radiographs were performed at the same oral radiology center.

Case Definitions
The case definition of periodontitis was determined as the presence of four or more teeth with at least one site with PD ≥ 4 mm, CAL ≥ 3 mm, and BOP. Implant success was defined as absence of signs of inflammation and bleending on probing 3 . Besides that, no peri-implant infection, mobility, persistent pain or dysesthesia, and continuous radiolucency around the implant were accepted 10 . Peri-implant health can exist around implants with normal or reduced bone support 3 . Implant failure was defined as implants that had been lost or removed, or implants that were broken or with mobility. The case definition of peri-implantitis was determined as presence of BOP and/or suppuration, PD ≥ 5 mm, and radiographic bone loss ≥ 2 mm 11 . Peri-implant mucositis was defined as presence of BOP and signs of inflammation 3 with no radiographic active bone loss regardless of the PD value.
The case definitions of diabetes, osteoporosis, thyroid problems, cardiovascular disease and smoking were performed by a questionnaire and the patients should present a medical diagnosis and treatment. For thyroid problems were considered patients with hypo or hyperthyroidism or patients with any dysfunction on thyroid that interfered with hormone production and required drug therapy. For cardiovascular disease were included patients with hypertension, acute myocardial infarction, angina pectoris, heart valve disease, congenital heart disease, endocarditis, cardiac arrhythmias, myocarditis and tumors in the heart. For smoking, patients were divided in patients who never smoked and smoking or ex-smoking patients.

Data Analysis
For data analysis, the STATA software (Stata for Mac, version 13, Apple Inc., Cupertino, CA, USA) was used. Preliminary analyzes of the association between peri-implantitis and predictive variables were performed using the chi-square test with Fisher correction for analysis at the level of individuals and the chi-square test adjusted for observations of clusters within individuals for analysis at the level of the implants. Generalized Estimating Equations were used to model the association taking into account the clustering of teeth in individuals. The implant was used as the unit of analysis and exchangeable working correlation, logit link and semi-robust standard errors were used to estimate the "odds ratio" and their respective 95% confidence intervals. Adjusted and unadjusted analyzes were performed. Statistical significance was set at 5%.
A purposeful selection of variables was used for the statistical modeling 12 . Initially, a preliminary analysis was performed using univariate models, and all the variables associated with p <0.25 were included in the multivariate model. The variables that did not contribute significantly to the multivariate model were assessed for confounding before being eliminated. Multivariate models with and without the potential confounding factors were compared and a change >25% in the coefficients of the other variables was used to define a variable as a confounding factor. Only age was considered as a confounder and was retained in the final model.

RESULT
Of all the 495 individuals who received implant surgery between 2003 and 2005, 357 (72%) patients were not available or could not be located, 61 (12%) patients declined to participate in the study, and 3 patients passed away. A total of 74 patients agreed to participate, however, only 51 (10%) patients were available for a clinical and radiographic examination, and 1 of these patients was not included in agreement with the exclusion criteria of the study. Thus, the final sample consisted of 50 partially edentulous patients (19 males and 31 females) who had received 211 implants with acid etched surface (Porous® Conexão Sistemas de Próteses, Arujá, Brazil). Overall, 206 implants could be evaluated, and they were categorized as follows: 172 successful, 25 failing, 9 submerged, and 5 implants were lost. Thus, the survival rate was 97.6% and the success rate was 81.5%. Four patients (8%) had lost implants: one patient lost two implants and the others lost one implant. Of these two patients had lost all their implants. Sample and implant characteristics are described in Table 1. From 197 implants in 48 subjects clinically evaluated, 103 (52.3%) implants were diagnosed with mucositis and 25 (12.7%) with peri-implantitis. At individual level analysis, 37 (77.1%) patients had mucositis and 14 (29.2%) had peri-implantitis. The risk indicator factors evaluation for the periimplantitis occurrence was based on the data of 48 patients and 197 implants clinically assessed ( Table 2). Peri-implantitis was significantly more prevalent in individuals with BOP (full month) in 30% or more of the evaluated sites (including teeth and implants). At implant level analysis, peri-implantitis was significantly more frequent in individuals with osteoporosis and BOP in 30% or more of the evaluated sites.   In the univariate analysis (Table 3), osteoporosis (OR = 2.48) and generalized BOP (OR = 6.40) were significantly associated with peri-implantitis. A borderline significant association was observed between peri-implantitis and PD ≥ 4 mm (p = 0.08). No significant associations were observed for other variables. In the multivariate analysis (Table 4), significant associations between peri-implantitis and osteoporosis (OR = 2.84), diabetes (OR = 0.22), generalized BOP (OR = 8.03), and periodontitis (OR = 0.24) were observed at the patient level analysis. At implant level, peri-implantitis was associated with the presence of visible plaque (OR = 4.45) and PD ≥ 4 mm (OR = 4.47). Age was retained in the final model because it interacted with other predictors.

DISCUSSION
In the present study, the reported implant success and survival rates of 81.5% and 97.6% respectively, after a period of 8 to 10 years of function, should be considered a very satisfactory result. According to Albrektsson et al. 13 , 1986, an implant system could be considered effective if a minimum success rate of 80% was achieved after 10 years of function. Similar survival rates were obtained in a retrospective study, whose analysis resulted in a 10-year implant survival rate of 98.8% and a success rate of 97% 11 .
A high prevalence of mucositis was observed in this study, 77.1% and 52.3% at individual and implant level respectively. These results are in accordance with the study of Lindhe, Meyle 14 , that identified the presence of mucositis in 80% of subjects and 50% of sites restored with implants. However, a systematic review of 2015 15 evaluated fifteen articles and estimated weighted mean prevalences of peri-implant mucositis of 43%. The concern related with the high prevalence of mucositis is due to the risk of untreated sites developing into periimplantitis 16 , especially if they are not included in a preventive maintenance program. While mucositis can be treated successfully with non-surgical mechanical debridement, this treatment modality has limited efficacy for peri-implantitis. Therefore, the treatment of early sign of inflammation is essential to prevent or limit marginal bone loss 17 .
Peri-implantitis has been considered one of the major reason for late implant loss. However, its prevalence value varies considerable in the literature. According to the systematic review of Derks, Tomasi 15 , the prevalence of peri-implantitis was 22% (range, 1 to 47%). The prevalence of peri-implantitis in our sample amounted 29.2% at individual level and 12.7% at implant level. Similar values were reported in other studies. Atieh et al. 11 reported the occurrence of periimplantitis in 18.8% of participants and 9.6% of implants.
For adequate diagnosis and prevention of peri-implantitis, annual radiographic exams are requested from the moment of prosthetic crown placement 3 . However, as this study is a retrospective study and the patients were not in follow-up after implant and prosthetic crown placement, anterior periapical radiographs were not obtained. Therefore, it was not possible to perform a comparative of activity and extension of bone loss caused by the peri-implantitis of each patient. This is a limitation of the study, where we are based on the current clinical and radiographic findings of patients to configure them with periimplant diseases or periimplant health.
Regarding the risk of peri-implant diseases, several authors have reported that subjects with a history of periodontal disease may be at greater risk for implant failure. Schwarz et al. 18 in a narrative review, showed that there is an increased risk of developing peri-implantitis in patients who have a history of chronic periodontitis, poor plaque control skills and no regular maintenance care after implants therapy. Surprisingly, we identified a negative relationship in the multivariate analysis between the presence of periodontitis and peri-implantitis prevalence (OR = 0.24). This unexpected result may be justified by the parameters used in the present study to define the presence of periodontitis and the comparatively small sample size. On the other hand, our data demonstrate that the presence of sites with PD ≥ 4 mm (OR = 4.47) and the presence of visible plaque (OR = 4.45) around the implant were significantly associated with a higher occurrence of peri-implantitis. Moreover, individuals with bleeding on probing in more than or equal 30% of sites (OR = 8.03) was at a significantly higher risk of developping periimplantitis. The systematic review by Zangrando et al. 19 showed that patients diagnosed with periodontitis undergoing appropriate therapy and regular maintenance could be successfully treated with dental implants (implant survival rate of 92.1% after 10 years of follow-up). Residual pockets, noncompliance with a regular maintenance program, and smoking were described as negative factors for long-term implant outcomes.
Although previous studies have found a positive association between smoking habits and periimplantitis prevalence 4,20 this correlation is contradictory 21 . In the present study, individuals that smoked 10 cigarettes or more per day were considered smokers. Only 2 patients were classified as smokers and 13 patients were former smokers. Thus, smokers and former smokers were included in a single category to enable the statistical analysis. According to our results, the smoking habit was not associated with an increased occurrence of peri-implantitis. Ata-Ali et al. 21 in their prospective cross-sectional study involving heavy smokers and non-smokers may conclude that smoking, when analyzed alone, does not influence immunological and microbiological parameters in dental implants. Although smoking habit is not considered an absolute contraindication for dental implants treatment, it is a relevant factor for the initiation and progression of peri-implantitis and with that, peri-implant mariginal bone loss 17 .
Guobis et al. 22 in a systematic review verified that osteoporosis cannot be considered a significant a risk factor for implant success. On the other hand, Alsaadi et al. 23 reported significant correlation between the osteoporosis and implant failure. The osteoporosis presence was significantly associated with higher peri-implantitis prevalence at individual level in our study (OR = 2.84). Despite our limited sample size, the association between osteoporosis and periimplantitis observed in the present study suggests that this relation needs to be more investigated under other methodological approaches including laboratorial hormonal examinations.
Naujokat et al. 24 in a systematic review of 22 clinical studies and 20 publications in the aggregate literature concluded that patients with poorly controlled diabetes have impaired osseointegration, a high risk of periimplantitis, and a higher level of implant failure. However, Dowell et al. 25 , evaluating 50 implants in 35 individuals stratified by glycated hemoglobin levels, observed that there was no evidence of clinical failure and healing changes associated with dental implant therapy in individuals with type 2 diabetes mellitus, controlled based on glycated hemoglobin levels. In the present study, the presence of type-2 diabetes mellitus had a protective effect for peri-implantitis (OR = 0.22). However, laboratory tests were not conducted to check the blood sugar levels in the patients considered to have type-2 diabetes. Moreover, there were only six patients with diagnose of diabetes in our sample. Consequently, these limitations could have influenced the results obtained in the present study.
All patients included in this study received implants from the same manufacturer. Therefore, the surface roughness and other implant characteristics could not be evaluated as risk indicators for peri-implantitis. The high patient's evasion observed could be related to the retrospective character of this study where in the recall evaluation was performed after a relatively long period from the prosthetic rehabilitation. These patients were not included in a regular maintenance program after the treatment with osseointegrated dental implants. Most of them (72%) were not available or could not be located for a clinical and radiographic examination of the implants after 8-10 years of function. Since conflicting results could be observed in the literature regarding risk indicators for peri-implantitis, prospective studies are required in order to determine true risk factors for peri-implantitis.
The success and survival rates of osseointegrated implants after 8-10 years of function were 82% and 97.6%, respectively. The prevalence of mucositis and peri-implantitis was 52.3% and 12.7% for implants and 77.1% and 29.2% for patients. Subjects with osteoporosis and with fullmouth BOP scores ≥ 30% of the sites as well as implants with visible plaque and sites with PD ≥ 4 mm were more susceptible to peri-implantitis. Therefore, our results, notwithstanding the limitations of the study regarding sample size, suggest that osteoporosis, diabetes and generalized periodontal/peri-implant mucosa inflammation increase the likelihood of periimplantiitis.