Hemorrhagic effects of oral anticoagulants: a comparative study between vitamin K antagonists (VKA) and direct oral anticoagulants (DOA).

Abstract

Clinical trials described differences in anatomical localizations of bleedings between vitamin K antagonists (VKA) and direct oral anticoagulants (DOA) [1]. RE-LY trial did not evidence differences in major bleedings but found more gastrointestinal and less intracranial hemorrhages with dabigatran [2]. However, clinical trials are conducted in standardized conditions far from the real world [3]. In order to verify these conclusions in real life, we compared anatomical localizations of bleedings reported with VKA and DOA as spontaneous notifications in a pharmacovigilance center. The study was performed into the French Pharmacovigilance Database [previously described [4]] that registers adverse drug reactions (ADRs) reported by health professionals or patients. All ADRs registered inMidi-Pyrenees area (SouthWest France, 3 million inhabitants) with VKA or DOA (Bsuspect^ drugs) between January 2009 (first marketing of dabigatran) and August 2014 were included. Cases with more than one anticoagulant were excluded. Comparisons used chi-square and Z tests and adjusted logistic regression models. A multivariate logistic regression was used to compare risk of bleedings. Adjustment variables were gender and age (4 classes: 220 mg daily, 1 apixaban), and 503 (75.1%) with VKA (364 fluindione, 126 warfarin, 13 acenocoumarol). There was no significant difference in mean age and sex ratio. BSerious^ ADRs were significantly more frequent with VKA than with DOA, without any difference in the number of reported deaths. There were more muscular bleedings with VKA than with DOA without any difference in other sites (Table 1). When analysis of bleedings was restricted to atrial fibrillation indication, similar results were found (not shown). Adjusted analysis on age and gender did not find any difference in total or digestive bleedings (not shown). We cannot conclude on delay of occurrence since VKA are marketed for much longer time than DOA. In contrast to clinical trials, this study was unable to describe any difference in main anatomical sites of bleedings between the two classes of oral anticoagulants. Ruff’s metaanalysis of randomized trials found fewer strokes or systemic embolic events (and more gastrointestinal bleedings) [5]. However, these differences in clinical trials come from analyses of secondary and not primary endpoints. There are relatively few data about hemorrhagic ADRs during post-marketing use. Hernandez reported more bleedings (whatever the anatomical site), especially gastrointestinal ones with less intracranial hemorrhage with dabigatran [6]. In patients ≥65 years with non-valvular atrial fibrillation, there was less ischemic stroke and intracranial hemorrhage and more major gastrointestinal bleedings with dabigatran than with warfarin [7]. Since the risk of gastrointestinal bleedings was only significant in patients ≥75 years, we investigated in an adjusted analysis the role of age and gender and were * J. L. Montastruc jean-louis.montastruc@univ-tlse3.fr