Many patients with heart failure (HF) are treated with warfarin or non-vitamin K oral anticoagulants (NOACs). Randomized outcome-driven comparisons of different anticoagulant strategies in HF are lacking. Data from international, government-mandated registries may be useful in understanding the real-life use of various anticoagulants and how they are linked to outcomes.
To assess 2015 annual all-cause mortality, myocardial infarction, and stroke rates co-reported for warfarin and NOACs in subjects with and without HF in the US Food and Drug Administration Adverse Event Reporting System (FAERS) database.
We extracted and examined outcome cases in subjects with HF and on warfarin, dabigatran, rivaroxaban, apixaban, or edoxaban and stratified these according to anticoagulants.
Annual all-cause mortality, myocardial infarction, and stroke in FAERS.
Odds ratio (OR) and χ2
for oral anticoagulants from FAERS with and without HF among complete primary reports issued in 2015.
FAERS reported 137,026 HF cases, with death co-reported in 42,942 (31.3%). In total, 11,278 (8.2%) HF patients were treated with anticoagulants, with more prescribed warfarin (n = 8260) than all NOACs combined (n = 3018). Very few reports for edoxaban were available. Warfarin consistently displayed a signal for excess adverse events compared to NOACs: OR (95% confidence interval) for the composite of mortality, myocardial infarction, and stroke were 1.91 (1.76–2.07) versus apixaban, 1.92 (1.81–2.03) versus dabigatran, 4.09 (3.38–4.37) versus rivaroxaban, and 2.64 (2.53–2.76) versus all NOACs combined (all P < 0.001). Warfarin, compared to all NOACs combined, demonstrated higher rates of all-cause mortality [OR = 2.69 (95% confidence interval, 2.49–2.90)], myocardial infarction [5.30 (4.17–6.74)], stroke [OR = 8.85 (6.61–11.84)], and ischemic stroke [OR = 12.73 (8.87–18.27); all P < 0.001].
Annual 2015 FAERS profiles in HF patients reveal that warfarin was numerically dominant. Warfarin was associated with higher risk of death, myocardial infarction, and stroke compared to NOACs. These observational data provide real-world insight into a potential safety benefit of NOACs over warfarin in the setting of HF.
1Department of Cardiology, Oslo University Hospital Ullevål, and University of Oslo, Oslo, Norway;
2Center for Heart Failure Research, University of Oslo, Oslo, Norway;
3Centre of Cardiovascular Research and Education in Therapeutics, Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Australia;
4Division of Cardiology, Århus University Hospital Skeiby, Århus, Denmark;
5Institute of Cardiovascular Sciences, University of Birmingham, Birmingham, United Kingdom;
6Duke Clinical Research Institute and Division of Cardiology, Duke Medical Center, Durham, NC;
7Dong-A University, Busan, Korea; and
8Johns Hopkins University, Baltimore, MD.
Address for correspondence: Johns Hopkins University, 14110 Rover Mill Road, West Friendship, MD 21794. E-mail: email@example.com
HeartDrug Research LLC (Wilmington, Delaware) supported Food and Drug Administration Adverse Event Reporting System data mining and statistical work. V. L. Serebruany received research grants from the dabigatran, rivaroxaban, and apixaban manufacturers; lecture fees from the dabigatran, apixaban, and edoxaban manufacturers; and consultant fees from all 4 NOACs manufacturers. D. Atar reports lecture honoraria and advisory fees from Boeringer-Ingelheim, BMS/Pfizer, Merck/MSD, and Bayer. The remaining authors have no conflicts of interest to declare.
S. Agewall, D. Atar, I. Hopper, D. Kotecha, R. J. Mentz, and V. L. Serebruany conceived the study and planned the analytic approach. T. G. von Lueder, J. K. Jensen, I. Hopper, D. Kotecha, R. J. Mentz, and V. L. Serebruany interpreted the results. T. G. von Lueder, I. Hopper, D. Atar, R. J. Mentz, and V. L. Serebruany drafted the manuscript. J. K. Jensen, I. Hopper, D. Kotecha, R. J. Mentz, and S. Agewall commented on and edited further drafts. V. L. Serebruany and D. Atar produced the final manuscript. All authors approved the final version.