The squares represent the point estimates of risk ratio of individual trials, and the lines extending from them represent 95% CIs. The diamond represents the meta-analytic risk ratio and its 95% CI.
Sipahi I, Celik S, Tozun N. A Comparison of Results of the US Food and Drug Administration’s Mini-Sentinel Program With Randomized Clinical TrialsThe Case of Gastrointestinal Tract Bleeding With Dabigatran. JAMA Intern Med. 2014;174(1):150-151. doi:10.1001/jamainternmed.2013.12217
The US Food and Drug Administration (FDA) recently started an initiative called “Mini-Sentinel Program” to assess medical product safety using administrative-observational databases. One of the first utilizations of this program was to examine the bleeding risk of dabigatran in response to the unexpectedly high number of postmarketing reports of bleeding.1 However, there are major limitations of observational studies, and therefore the reliability of the Mini-Sentinel Program is unknown. Our objective was to compare the results of this program regarding the gastrointestinal (GI) tract bleeding risk of dabigatran vs warfarin with the results of randomized clinical trials (RCTs).
To obtain the results of RCTs regarding GI tract bleeding, a literature search was performed using MEDLINE through July 2013 with the search term “dabigatran AND warfarin” limited to RCTs. This search was supplemented with examination of the FDA website for additional data, as well as a search of the clinical trial registry website maintained by dabigatran’s manufacturer. The RCTs directly comparing dabigatran to warfarin that reported incident GI tract bleeding were then included in a meta-analysis. The meta-analytic risk ratio (RR) of dabigatran vs warfarin for GI tract bleeding was calculated using a fixed-effect model. The results of this meta-analysis were then compared with the results of the Mini-Sentinel Program and its interpretation by the FDA.
Twenty-seven articles were identified using the MEDLINE search. Three articles provided data on incident GI tract bleeding with dabigatran vs warfarin.2- 4 A search of the FDA website provided additional data on GI tract bleeding for one of these clinical trials5 and search of the clinical trials registry of manufacturer provided data for another clinical trial.6 Therefore, a total of 4 RCTs (enrolling 26 076 patients) were included. On meta-analysis, dabigatran significantly increased the risk of GI tract bleeding, compared with warfarin (I2 = 0; RR = 1.41 [95% CI, 1.28-1.55]; P < .001) (Figure). The results remained the same with the random-effects model. Exclusion of any single clinical trial (including the RE-LY Trial2 did not change the statistically significant increased risk. On the contrary, using the Mini-Sentinel Database, the FDA obtained a GI tract bleeding rate of 1.6 with dabigatran and 3.5 with warfarin (per 100 000 days at risk).1 With this analysis, the agency concluded that GI tract bleeding rates are not higher (and indeed lower) with dabigatran, attributed the postmarketing reports of bleeding to “stimulated reporting” and released a reassuring statement about the bleeding risks of this drug.
This analysis shows that the RCTs and Mini-Sentinel Program show completely opposite results regarding the GI tract bleeding risk of dabigatran compared with warfarin. The meta-analytic results of the RCTs have very narrow confidence intervals and no heterogeneity, demonstrating the increased risk of GI tract bleeding with dabigatran (vs warfarin) unequivocally. However, the Mini-Sentinel Program reports a greater than 50% decrease in incident GI tract bleeding with dabigatran compared with warfarin.
Observational studies like the Mini-Sentinel Program are inherently problematic owing to several sources of biases. Because of their limitations, the approval process of drugs relies solely on RCTs. Nevertheless, observational studies are still performed, especially in the postmarketing phase, and may be meaningful to examine issues of drug safety not examined during the approval process. The Mini-Sentinel Program is a new initiative of the FDA that aims to examine medical product safety in the post-marketing period. Our examination of the example of GI tract bleeding risk with dabigatran shows that the results generated by this program may contradict the gold-standard clinical evidence from RCTs. Examination of the reasons behind this contradiction by the regulatory agency may help to improve the reliability of this new program.
Corresponding Author: Ilke Sipahi, MD, Department of Cardiology, Acibadem University Medical School, Acibadem Maslak Hospital, Buyukdere Cad 40, 34457 Istanbul, Turkey (email@example.com).
Published Online: November 18, 2013. doi:10.1001/jamainternmed.2013.12217.
Author Contributions: Dr Sipahi had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Sipahi.
Acquisition of data: Sipahi.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Sipahi.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Sipahi.
Administrative, technical, or material support: Sipahi.
Study supervision: Tozun.
Conflict of Interest Disclosures: None reported.