Survival and success of implants in a private periodontal practice: a 10 year retrospective study

All patients were treated in a private practice from 2003 to 2006 by implant placement of at least one XIVE S Cellplus or Frialit-2-Synchro Implant (Fa. Dentsply Sirona Implants, Mannheim, Germany) by one surgeon (GK). All patients received at least one session of individual hygiene before implants were placed. When a periodontitis was detected a complete active periodontal therapy (APT) was conducted before. Smokers received implant treatment if the maximum daily dose did not exceed 10 cigarettes/day. After implant placement all patients were incorporated into a recall program. This contained re-instruction and re-motivation to an effective individual plaque control, professional mechanical plaque removal and once a year obtaining a dental and periodontal status. Sites exhibiting periodontal pocket depth (PPD) of 4 mm plus bleeding on probing (BOP) and sites with PPD ≥ 5 mm were scaled subgingivally [20, 21].

All treated patients were invited to a re-examination 10 years ±6 months after implant placement until 100 patients could be included when fulfilling the following inclusion criteria:

Data on time for healing after extraction was recorded. The surgical procedure included bone augmentation, if necessary. One week after implant insertion the sutures were removed. In cases without bone augmentation or only minor augmentations loading of the implants was conduced about 3 months after implantation, in cases of major augmentation (e.g. external sinus lift, block augmentation) loading time was about 4–5 months. Fixed supra-structures were mostly cemented with a temporary cement (TempBond®). In case of decementation, Rely X® was taken as cement. Just few supra-structures were screwed.

Complete clinical re-examinations were performed by one independent and calibrated examiner (AB) from November 2013 to May 2016. They included:

Retrospective evaluation of patients’ charts was accomplished by two examiners (ST, AB) independently and included:

Bone loss was calculated by comparing baseline radiographs with radiographs at re-examination. The following distances were measured at the mesial and distal aspect of the implant by using a computer program (VixWin Platinum Version 1.4, Fa. Gendex, Hatfiels, USA) under standardized conditions in a darkened room by two examiner (ST, AB):

The implant length reported by the manufacturer was used for the calibration of the distances. The largest value was taken as the extent of bone loss. As proposed a measurement error of 0.5 mm was included [26]).

According to Derks et al. (2016) mucositis was defined as presence of BOP/suppuration but no detectable bone loss. Peri-implantitis was defined as presence of bone loss of > 0.5 mm with/without BOP/SUP and a moderate/severe peri-implantitis was stated when bone loss reached > 2.0 mm. Implant survival was given when the implant was not lost during the last 10 years.

All data were entered into two separate data files (Excel version 2003, Microsoft Corporation, Redmond, WA, USA) by two investigators (ST, AB). These data were compared thereafter. All differing entries were double-checked by means of comparison with the original patients’ charts.

Primary outcome was survival of implants, secondary outcome implant success (no mucositis, no peri-implantitis, no technical failures).

Descriptive statistical analysis with rates for qualitative characteristics and with mean, standard deviation, median, minimum and maximum for quantitative outcome was performed using R 3.2.2 (R Foundation for Statistical Computing, Vienna, Austria, www.​r-project.​com). This software was also used for logistic regressions with either implant loss or mucositis or peri-implantitis or severe peri-implantitis as dependent variable on both patient and implant level. As independent factors pre-defined variables were included without variable selection and the significance level was chosen to be 0.05, which was not adjusted for multiple testing, also because of the descriptive nature of this study. A χ²-test was conducted to test for a difference between smokers and non-smokers according to smokerlyzer. Spearman correlation was used to correlate the width of keratinized gingiva with peri-implantitis, mucositis, BOP, mPII and mGI.

210 patients received an implant treatment during 2003 and 2006 and fulfilled the inclusion criteria, 103 of them could be reexamined (responder rate of 49.1%). In the analysis 100 patients aged 28–86 years (mean 63.8 years, SD 10.3 years) were included 10 years (range 9.5–10.7 years, SD 0.31 years) after implant setting. The reason for exclusion is given in Fig. 1.

A total of 242 implants could be included. 212 Xive S Cellplus implants (87.6%) in 88 patients (88.0%), 30 Frialit-2-Synchro implants in 12, respectively. Number of implants per patient is given in Table 1. Time of implant placement, form of augmentation, healing as well as length and diameter of implants is presented in Table 2. After implant placement 8.7% of all implants needed peri-implant therapy (5.0% non-surgical, 3.7% surgical). At re-examination data regarding bone loss, PPD, width of keratinized gingiva, mPII, mGI and BOP were collected and are given in Table 2.

Most implants were supported by single crowns (55.6%). 25.6% of implants served as bridge anchor for implant-supported fixed dental prostheses (FDP), 17.5% as anchor for implant-supported removable prostheses, 1.3% as bridge anchor for combined tooth-implant-supported FDPs. Most implant-supported crowns were cemented (97.4%).

Technical complications were found in 47 implants (19.4%). Abutment/screw loosening occurred in 13 cases and was the most common complication (5.3%) (Table 5). No implant fracture took place.

The observed survival rate over 10 years of 94.0% (97.7% for XiVE-implants) on patient-level presents comparable data to another long-term study by Degidi et al. (2016) of 93.0%. Similar survival rates can be found for other implant systems over the same follow-up period [3, 4, 27]. Most implants (8 of 9) were lost in the early healing phase prior the implant-supported restoration was connected (5.0% of patients, 3.3% of implants), just one in the late phase due to peri-implantitis (1.0% of patients, 0.4% of implants). Also, mostly Frialit-2-Synchro implants were lost. In a large study on early and late implant loss Derks et al. (2015) detected comparable rates of early loss in 4.4% of patients (1.4% of implants), but a higher late implant loss in 4.2% of patients (1.4% of implants) 9 years after implant placement. Overall, implant survival rates were in accordance with the long-term data over a follow-up period of 10 years published so far.

In our analysis ‘implant type’ (Frialit-2-Synchro) was the sole statistically significant influencing factor for implant loss as described by Derks et al. (2015), who detected ‘implant brand’ as an influencing factor. All other potential influencing factors such as smoking or history of periodontitis had no impact, as supported by several other studies [2, 4, 5, 16]. This might be due to the low number of lost implants, wherefore detection of influencing factors on implant-level was not possible. Another reason could be that all periodontally compromised patients had received a periodontal treatment of high quality in the private periodontal practice prior to implant placement and afterwards showed a high compliance to the supportive implant therapy.

The high implant loss rate of Frialit implants might be biased due to the relatively small number of observed implants and that solely one patient lost 4 of those implants in the early healing phase.

For comparison of success rates the definition of mucositits (BOP/suppuration without bone loss) and peri-implantitis (BOP/suppuration and bone loss > 0.5 mm; moderate/severe peri-implantitis: BOP/suppuration and bone loss > 2.0 mm, respectively) were chosen from a large study from Sweden [15]. Interestingly, our incidences for moderate/severe peri-implantitis (16.3% of patients, 10.3% of implants) were quite in accordance with the data published by Derks et al. (2016) for moderate/severe peri-implantitis (14.5% of patients, 8.0% of implants). Also, our results are reflected in the data published by Mombelli et al. (2012), which provided an incidence of peri-implantitis of 20% on patient-level and 10% of implant-level as well as by the one by Roos-Jansaker et al. (2006) of 16%. The number of implants per patient could be established as an influencing factor for moderate/severe peri-implantitis, which Derks et al. (2016) also found in their study. Furthermore, on implant-level a wider implant diameter was detected as influencing factor for peri-implantitis. All other risk factors proposed in many other studies such as smoking, history of periodontitis, plaque accumulation or compliance to recall [4, 15, 16, 18, 27] reached no statistically significant influence in our analysis. Also, the often discussed factor ‘width of keratinized gingiva’ could not been detected as influencing factor for peri-implantitis in regression analyses, even if there was a slight correlation between width of keratinized gingiva and higher mod. PI at the implant. Also, higher mod. PI as well as a wider implant diameter was associated with peri-implant mucositis. That plaque accumulation results in peri-implant mucositis was often stated by different authors [17, 27‐29] and is underlined by our data.

There still seems to exist a substantial lack of well-performed longitudinal reports on implant-supported restoration over an observation period of ten or more years [30]. In our studies, technical complications occurred in about 20% of all implants, comparable to the 16-years results by Simonis et al. (2010) with 31.1%. The most common complication over this 10 years follow-up was abutment/screw loosening, which occurred in 5.3% of all implants. Chipping could be observed in 4.0% of all implant-supported restorations and is comparable with chipping of the veneering material of fixed dental prosthesis (4.1%) in an actual review assessed over 5- and 10-years [31].

This analysis has some limitations (retrospective design, limited number of patients and implants with few events regarding implant loss) and could potentially be biased because of a relatively homogenous clientele of mostly highly educated patients, a high quality standard of dental/periodontal care in a specialized periodontal practice and the form of recruitment (few implants of the brand Frialit-2-Synchro compared to XiVE implants). On the other hand, the enrolled patients showed a high heterogeneity regarding age (between 28 and 86 years), smoking status and history of periodontitis. Also, a wide heterogeneity regarding form and extent of augmentation, point of time of implantation after extraction or variations in type of implant loading could have influenced the results. Likewise, the inconsistent distribution of insertion of the two different implant types (XiVE or Frialit-2 Synchro) could have biased the data. Also, the responder rate of about 50% could lead to a selection bias (e.g. patients were not willing to participate possibly due to dissatisfaction/problems with their implants).