Effect of previous implant failure on the prognosis of subsequent implants: A retrospective study.

Abstract

To compare the long-term survival of dental implants placed in patients with and without a history of implant failure. Within a retrospective analysis, an experimental group was selected consisting of 59 patients with 137 implants placed after previous failure. The control group included 1,072 patients with 2,664 implants without previous failure. Kaplan-Meier curves were used to describe the group-specific long-term implant survival. Mixed-effects Cox regression models were applied to examine the effects of patient- and site-specific risk factors. To take multiple implants into account, a random intercept model was applied. During the observation period of up to 15years, 11 implants (8%) failed in the experimental group and 74 implants (2.5%) in the control group (p<.001). Five-year cumulative survival was 96.8% (95% CI 0.96-0.98) in the control group and 91.5% (95% CI 0.86-0.97) in the experimental group. The variables group assignment and simultaneous augmentation had a significant effect on survival, but this effect was lost in the random intercept model. The effect of implant location remained, whereby the risk of failure was five times lower for mandible implants, irrespective of group (p=.013; 95% CI 0.103-0.767; HR: 0.281). Long-term implant survival was lower in the experimental group than in the control group. The effect of previous failure was negligible. However, a patient-specific "clustering effect" was observed. Irrespective of previous implant failure, the risk of long-term failure is two times higher for maxillary implants than for mandibular implants.