Does a higher glycemic level lead to a higher rate of dental implant failure?: A meta-analysis

Abstract Background Owing to limited evidence, it is unclear whether diabetes that is not well controlled would lead to a higher rate of dental implant failure. The authors of this meta-analysis evaluated whether the failure rate for patients with diabetes that was not well controlled was higher than the failure rate for patients with well-controlled diabetes. Types of Studies Reviewed The authors searched PubMed, the Cochrane Library, and ClinicalTrials.gov without limitations for studies whose investigators compared the dental implant failure rates between patients with well-controlled diabetes and diabetes that was not well controlled. The authors pooled the relative risk (RR) and 95% confidence interval (CI) values to estimate the relative effect of the glycemic level on dental implant failures. The authors used a subgroup analysis to identify the association between the implant failure rate and the stage at which the failure occurred. Results The authors included 7 studies in this meta-analysis, including a total of 252 patients and 587 dental implants. The results of the pooled analysis did not indicate a direct association between the glycemic level in patients with diabetes and the dental implant failure rate (RR, 0.620; 95% CI, 0.225-1.705). The pooled RR in the subgroup of patients who experienced early implant failure was 0.817 (95% CI, 0.096-6.927), whereas in the subgroup of patients who experienced late implant failure, the pooled RR was 0.572 (95% CI, 0.206-1.586). Conclusions and Practical Implications On the basis of the evidence, the results of this meta-analysis failed to show a difference in the failure rates for dental implants between patients with well-controlled diabetes and patients with diabetes that was not well controlled. However, considering the limitations associated with this meta-analysis, the authors determined that future studies that are well designed and provide adequate controls for confounding factors are required.