Impact of implant interdependency when evaluating success rates: a statistical analysis of multicenter results.

PURPOSE: When evaluating the outcome of oral implant treatment, a statistically important aspect to consider is whether any dependency exists among implants placed within the same patient/jaw. If one implant fails, will the risk of subsequent failures increase, i.e., will any of the remaining implants also fail? In an attempt to study this question, the aims of the present study were to statistically determine if the suggested hypothesis was valid, i.e., whether dependency exists among implants in the same patient/jaw, and to determine how failure lifetable analysis should be calculated. MATERIALS AND METHODS: In the present study, multicenter trial material consisting of 1.738 oral implants in 487 patients from 4 separate studies was used. First, any dependency among implants within the same patient/jaw was determined; thereafter, cumulative success rates were calculated for one implant per patient (chosen randomly), and the range of variations in the randomized success rates was also calculated. This was then compared with the cumulative success rates based on the entire material. RESULTS: The statistical analysis showed that a dependency among the implants existed prior to their functional loading, i.e., the risk for an implant failure among the remaining implants in the same patient/jaw increased after the first failure had occurred. CONCLUSION: Both study design and statistical analysis are of importance when comparing success rates from various investigations, since dependency among implants in the same patient/jaw does exist and may influence the success rates. Consequently, it is suggested that only one randomly selected implant from each patient should be considered when calculating implant success rates.