Most Real-World (Medicare) Bleeding Rates With Pradaxa Higher Than Expected

A new Medicare database analysis, published in the JAMA Internal Medicine by Hernandez et al, indicates that most bleeding rates associated with the use of dabigatran (Pradaxa, Boehringer Ingelheim) are higher than expected (as reported in the pivotal RE-LY trial) and surpass those of warfarin. Although a direct apples-to-apples comparison of data is not readily available, here is a tabulation of the relative risks and hazard ratios of bleeding identified in each study.

Event	RE-LY Trial	Medicare Database
	Dabigatran 150 mg vs Warfarin,Relative Risk (annual rates)	Dabigatran vs Warfarin, Hazard Ratio (adjusted incidence rates)
Any bleeding (major or minor)	0.91 (16.42% vs 18.15%)	1.30 (n/a)
Major bleeding	0.93 (3.11% vs 3.36%)	1.58 (9.0% vs 5.9%)
Intracranial bleeding	0.40 (0.30% vs 0.74%)	0.32 (n/a)
GI bleeding	1.50 (1.51% vs 1.02%)	1.85 (n/a)

Despite the fact that any bleeding and major bleeding were significantly higher with dabigatran in the Hernandez analysis (N = 9404), in contradistinction to the RE-LY trial (n = 12,098 [warfarin + dabigatran 150 mg]), other findings were consistent. Specifically intracranial bleeding remained significantly lower with dabigatran, and GI bleeding was lower with warfarin use in both studies.

In addition, a parsing of the Medicare data by Hernandez et al indicates significantly greater major bleeding risks among subgroups of patients: namely, blacks (vs whites); patients taking antiplatelet drugs; patients with chronic kidney disease; and older patients (ie, >74 years of age).

In an accompanying editorial, Dr. Rita F. Redburg writes, “The haste to approve novel drugs[*] places an increasing importance on postapproval data to help better understand risks and benefits.” She notes that the Medicare analysis of Hernandez et al conflicts with the Mini-Sentinal analysis of Medicare data performed by the FDA (see also here). Hernandez et al charge that the FDA failed to adjust for differences in patient characteristics.

Nevertheless, there are the consistent findings of a lower intracranial bleeding risk with dabigatran and a lower GI bleeding risk with warfarin across all studies. Overall the postmarketing findings will help guide clinicians on the use of warfarin or dabigatran for the prevention of stroke in subgroups of patients with atrial fibrillation. Tailoring therapy = good.

Another, curious finding in the FDA postmarketing study: “Most patients treated with dabigatran 75 mg twice daily appeared not to have severe renal impairment, the intended population for this dose.” This observation may indicate a relatively risk-averse clinician population when selecting a new anticoagulant, like dabigatran, for stroke prevention in at-risk patients.

* Dabigatran was approved by the FDA under an accelerated approval process–ie, 6 months. The FDA only approved 2 dosages of the drug for the prevention of stroke in patients with nonvalvular atrial fibrillation: 150 mg bid or 75 mg bid (the latter for patients with renal insufficiency). This was done despite the fact that the RE-LY trial showed comparable efficacies for 150 and 110 mg bid.