Industry sponsorship bias in clinical trials in implant dentistry: Systematic review and meta-regression.

Abstract

Industry sponsorship might distort the conduct and findings of studies in a large range of medical disciplines. The objective of this study was to assess whether industry sponsorship bias is present in randomized controlled trials (RCTs) on dental implants. Two databases were searched (MEDLINE; Web of Science) to identify RCTs published between 1996 and 2016 assessing different implant systems, components or techniques, such as implant-abutment connections, geometries, surfaces, loading protocols or regions of placement. Studies' sponsorship status was classified as unclear, non-sponsored or sponsored. Our outcome was marginal bone loss per year (MBL/year) of follow-up. Random-effects meta-analysis of MBL/year with subgroup analysis according to sponsorship status was performed. Moreover, multivariable stepwise-selection meta-regression was performed to assess whether sponsorship status, among other covariates, was associated with MBL/year. One hundred and two RCTs (4,775 patients, 8,806 implants) were included. Overall mean (95% confidence interval) MBL/year was 0.74mm (95% CI 0.67/0.82). There was no significant difference in MBL/year among sponsorship categories; unclear: 0.64 (95% CI 0.37/0.91); non-sponsored: 0.65 (095% CI 0.55/0.75); and sponsored: 0.82 (95% CI 0.71/0.94). Meta-regression did not demonstrate a significant association of MBL/year with sponsorship status or other covariates was found. We did not detect significant sponsorship bias in RCTs on dental implants.