The 'Mini-Sentinel Program' is an initiative recently introduced by the FDA which assesses medical product safety using administrative-observational databases. It was used to examine the bleeding risks associated with dabigatran and the purpose of the current research was to compare the results from this (for GI bleeding) with those of randomised clinical trials.
RCTs directly comparing dabigatran to warfarin that reported incident GI tract bleeding were included in a meta-analysis. Using a fixed-effects model, the risk ratio for dabigatran versus warfarin for GI tract bleeding was calculated as 1.41 (95% CI 1.28-1.55; P<0.001; no heterogeneity). This contrasts to the results of an FDA analysis of the Mini-Sentinel Database, where the GI tract bleeding rate was 1.6 with dabigatran and 3.5 with warfarin (per 100,000 days at risk). The results of this led the agency to conclude that GI tract bleeding rates are not higher with dabigatran, despite the unexpectedly high number of postmarketing reports that were received (see Perspective article published in the New England Journal of Medicine) which they put down to "stimulated reporting".
Due the major limitations of observational studies, the reliability of the Mini-Sentinel Program is unknown. The authors of this research conclude that the evidence generated from the Mini-Sentinel Program may contradict 'gold-standard' evidence from RCTs, and that examination of the reasons behind this may help the Agency to improve the reliability of this program.