Clinical trials of endosseous implants: issues in analysis and interpretation

The majority of contemporary endosseous dental implant systems are based on designs and materials that, over the last three decades, have proved to be predictably reliable. With proper surgical and prosthetic protocols, rates of implant loss have been held to 15% or less over a 5-year period. This information was obtained largely through longitudinal descriptive studies, primarily aimed at obtaining implant survival rates under ideal clinical conditions, with strict inclusion and exclusion criteria for admitting patients into the studies. It is important to emphasize that under conditions of routine clinical practice, where patient selection may be more relaxed than in clinical trials and clinicians attempt to stretch the limits of current technology, the survival rates may not necessarily match those reported in the literature. Since "surviving" implants may exhibit characteristics likely to lead to eventual loss of the implant, for example severe osseous defects, such implants may not necessarily be considered successful. Successful implants should fulfill a list of other criteria considered essential for long-term survival. Differences in implant design preclude some of these criteria from being uniformly applied to all systems. There is a need to identify criteria for success that can be applied to the majority of implant systems. Implants that fail to meet these criteria should be considered failures. Since failure rates may include "failed" as well as "failing" ("ailing") implants, the two categories should be listed separately. From a practical standpoint, implant failures can be grouped into "early" failures, primarily the result of surgical and/or postoperative complications, and "late" failures that arise during and following the restorative phase. The ability of individual systems to achieve excellent success rates, despite some major differences in their design from other systems, suggests that some requirements, initially considered essential for success, may not be as critical as originally believed. Examples include the need for submerging implants during initial wound healing or the need for stress breaking devices. On the other hand, a basic requirement for implant success, such as primary stability at the time of insertion and following loading of the implant, may be the unifying principle behind the need for adequate bone volume and density, longer or wider implants, and the 3 to 6-month delay recommended before implants are placed in function. With relatively low failure rates, a large number of patients may have to be included in long-term clinical trials before a statistically significant association can be established between failure rates and potential contributing factors. For the same reasons, and to avoid type 2 errors, large populations may be needed to show that two systems have comparable success rates. Proving the superiority of one system over another may require fewer subjects. Given the overall low failure rate and the tendency of failures to cluster in individual subjects, failure rates could be markedly affected by the attrition of a few critical subjects. Additional research is needed to validate methods in current use for the clinical determination of osseointegration, and the diagnosis and treatment of occlusal trauma and microbial infections around implants. Also, more reliable methods are needed for the identification of the primary cause(s) of implant morbidity; i.e., infection or occlusal factors.