Accuracy of Clinical Parameters in Predicting/Diagnosing Peri-Implant Bone Loss.

AIM: To determine whether clinical parameters can serve as (i) predictive tools (before occurrence) and (ii) diagnostic tools (after occurrence) of peri-implant bone loss. MATERIALS AND METHODS: A representative cohort of 72 patients with 298 implants was evaluated at baseline and after a mean follow-up period of 3.9 years. Peri-implant bone loss > 1 mm between the two examinations represented the reference standard. The accuracy of the following clinical parameters in predicting (at baseline) or diagnosing (at follow-up) peri-implant bone loss was assessed: presence of bleeding (BoP) or suppuration (SoP) on probing, visual signs of redness or swelling, BoP extent (number of sites with BoP) and severity (modified Bleeding Index-mBI), probing pocket depth (PPD) at various cut-offs, peri-implant soft-tissue dehiscence (PISTD) and changes in PPD/PISTD over time. Predictive/diagnostic performance was evaluated using mixed model logistic regression analyses and reporting sensitivity, specificity, positive/negative predictive values and area under the curve (AUC) values. RESULTS: Bone loss > 1 mm was observed in 9.4% of implants and was frequently preceded by BoP (sensitivity = 96.4%; specificity = 7.4%). At follow-up, bone loss was always associated with the concomitant presence of BoP (sensitivity = 100.0%; specificity = 14.4%). In predicting the future occurrence of peri-implant bone loss, high sensitivity (94.4%) was also noted for visual redness at baseline, although its specificity was low (25.9%). Conversely, high specificity but low sensitivity was observed for BoP at 6 sites (sensitivity = 25.0%; specificity = 88.1%) and SoP (sensitivity = 14.3%; specificity = 91.5%). For diagnosing recent peri-implant bone loss, high specificity was noted for SoP (100.0%), profuse bleeding (91.9%), BoP at 6 sites (87.0%), PPD >/= 6 mm (81.9%), changes in PPD (95.9%) and changes in PISTD (91.5%). However, all these parameters showed limited sensitivity. The best diagnostic accuracy was achieved using a combined criterion of site-specific PPD or PISTD increases > 1 mm over time (sensitivity = 82.1%; specificity = 70.0%; AUC = 0.76). CONCLUSIONS: Clinical signs considered indicative of peri-implant mucositis (presence of BoP, visual redness) usually precede peri-implant bone loss. Implants with a recent history of bone loss always present with concomitant BoP. However, the predictive/diagnostic value of detecting one or two spots of BoP is limited by its low specificity. Implants with BoP at six sites or SoP are more likely to exhibit bone loss over time. During follow-up, BoP at six sites, profuse bleeding, SoP, PPD >/= 6 mm, or increases in PPD/PISTD over time have high specificity for diagnosis of recent peri-implant bone loss.