Background: The advancement in implant dentistry has allowed shortened treatment time by restoring the implants earlier. Whether the timing of restoration has an impact on implant marginal bone level has not been systematically analyzed. The aim of this study is to compare marginal bone loss (MBL) between implants that were restored with the following protocols: 1) immediate restoration/loading (IR/L); 2) early loading (EL); and 3) conventional loading (CL).
Methods: An electronic literature search from three databases (until November 2011) and a hand search in implant-related journals were conducted. Clinical human studies in English language that had reported a comparison of MBL between implants with IR/L, EL, or CL with ≥12-month follow-up were included. In addition, the minimal number of implants had to be 10 for each group. Implants with both immediate placement (IP) and delayed placement (DP) were included and analyzed separately. An assessment of the publication bias for the included randomized clinical trials (RCTs) was performed.
Results: The initial search resulted in 1,640 articles, of which 27 articles in full text were further evaluated for eligibility. Finally, 11 studies (eight RCTs, two controlled clinical trials, and one retrospective study) were qualified and classified into four groups: 1) IR/L + DP versus CL + DP (n = 6 articles); 2) IR + DP versus EL + DP (n = 2 articles); 3) EL + DP versus CL + DP (n = 1 article); and (4) IL + IP versus CL + IP (n = 2 articles). A meta-analysis performed for group 1 showed 0.09 mm (95% confidence interval = -0.27 to 0.09 mm) difference in the mean MBL, favoring the IR/L protocol but without significant difference (P = 0.33). No significant difference in MBL was found for groups 2 through 4 after adjusting for the implant placement level. The eight RCTs were determined to be at moderate-to-high risk of publication bias.
Conclusions: This meta-analysis does not show an effect of the timing of restorations on implant MBL. The selection of restoration protocols should be based on factors other than MBL.