Performance of the Implant Disease Risk Assessment in Predicting Peri‐Implantitis: A Retrospective Study

Abstract

ObjectiveThis university‐based retrospective study aimed to assess the performance of the implant disease risk assessment (IDRA) in predicting peri‐implantitis.Material and MethodsPatients with implants loaded for at least 1 year were included. Peri‐implantitis development was the outcome, while the IDRA score and its eight vectors were the predictors. The IDRA score was calculated using an online tool. Data were analyzed using Cox proportional hazards models and ROC curve (AUC).ResultsAmong 480 implants in 235 patients, 7.9% of implants and 9.4% of patients developed peri‐implantitis. Implants at high risk for the “number of sites with PD ≥ 5 mm” vector had an increased risk (HR = 9.8, p = 0.004) of peri‐implantitis, compared to those at low risk for this parameter. Implants at moderate (HR = 4.8, p = 0.04) and high (HR = 10.0, p = 0.01) risk for the “distance from the restorative margin (RM) to bone crest (BC)” vector exhibited a higher risk of peri‐implantitis than implants at low risk for this parameter. The IDRA tool demonstrated an AUC of 0.66 (sensitivity = 0.80; specificity = 0.24) when estimated at implant level and an AUC of 0.61 (sensitivity = 0.91; specificity = 0.32) when calculated at patient level. The mixed‐effects Cox model did not reveal a significant association between the overall IDRA score and the development of peri‐implantitis (HR = 7.2, p = 0.18).ConclusionIDRA demonstrates good sensitivity but low specificity and suboptimal discriminatory capacity in predicting peri‐implantitis. The “number of sites with PD ≥ 5 mm” and “distance from RM to BC” emerged as the most effective predictors for peri‐implantitis.