As the implants used in the current study were either platform shift or 1‐stage design <0.5 mm early MBL was expected to occur so a ≥1 mm MBL threshold was used as the MBL threshold. This is comparable to the threshold level of 1.2 mm loss beyond smooth‐rough interface with implants of similar design. Rodrigo et al. also evaluated MBL and recorded the highest score from mesial or distal sites, as was the method used in the current study.

Taking both soft‐tissue and MBL thresholds into account, the prevalence of peri‐implantitis determined in this retrospective study, using a criteria of a ≥1 mm MBL and a “strict” BOP score, was found to be 4.7% after 6 to 7 years (Table 2), while the more “relaxed” soft‐tissue threshold (IMI≥2), excluding minor bleeding, resulted in a rate of 3.6% at 6 to 7 years (Table 3). A study of similar design also found a similar rate of peri‐implantitis and mucositis. In their study the prevalence of mucositis, as determined by probing depth >3 mm and BOP+ but no concomitant bone loss was 48% of implants. This is similar in that about half of the cases in the current study had some bleeding but not necessarily MBL. In the Roos‐Jansåker et al. study peri‐implantitis was defined as exposure of ≥3 threads (1.8‐mm MBL) with BOP or suppuration and they revealed a fairly comparable prevalence of peri‐implantitis at 7% for implant level.

One limitation in this study is that probing depths were not used, however these vary with soft tissue thickness, abutment and prosthetic design so are not easily compared between studies or between patients. The main limitation of this study is its retrospective nature with greater potential for missing data. Nevertheless, the high number of implants and the long follow‐up provides important insights into the clinical outcomes that one can expect in private practice. Another limitation of this study is the number of patients lost during follow up, which limits the ability to draw conclusions beyond 4 to 5 years. Furthermore, no intra‐examiner calibrations were done and all measurements were carried out by the clinician who placed the implants, thereby introducing a potential bias.