Editor's Note: Commentary based on Abraham NS, Singh S, Alexander GC, et al. Comparative risk of gastrointestinal bleeding with dabigatran, rivaroxaban, and warfarin: a population based cohort study. BMJ 2015;350:h1587 and Chang H, Zhou M, Tang W, et al. Risk of gastrointestinal bleeding associated with oral anticoagulants: a population based retrospective cohort study. BMJ 2015;350:h1585.

Background

Clinical trial data suggest that novel oral anticoagulants (NOACs) are non-inferior to warfarin for prevention of stroke in atrial fibrillation (AF), but increase the risk of gastrointestinal (GI) bleeding.^1,2 Yet, this increased risk of GI bleeding is poorly understood outside of clinical trial contexts.

Methods

Two recent large, retrospective cohort analyses provide further data on the risk of GI bleeding in patients with AF being treated with NOACs in real-world clinical settings.^3,4 Both were retrospective analyses of claims-based administrative data sets from geographically diverse, commercially-insured, adult populations in the U.S. In Abraham et al., the study population included enrollees with an incident prescription of warfarin, dabigatran, or rivaroxaban between November 1, 2010 and September 30, 2013, with follow-up through December 31, 2013.³ In Chang et al., the final study population included enrollees with an incident prescription of warfarin, dabigatran, or rivaroxaban between October 1, 2010 and March 31, 2012, with follow-up through March 31, 2012.⁴ In both studies, the primary exposure was the study drug; the covariates of interest were baseline demographics, co-morbidities, and pharmacologic risk factors for bleeding; and the primary outcome measure was the incidence of GI bleeding as defined by the international classification of diseases, ninth revision, clinical modification (ICD-9-CM) code. Both study designs employed propensity score matching and Cox proportional hazards models.

Results

Results are summarized in Table 1.^3,4 In both cohorts, the majority of enrollees (>70%) were prescribed warfarin. In Abraham et al., there was no significant difference in the risk of GI bleeding for enrollees prescribed dabigatran relative to warfarin (adjusted hazard ratio [HR] 0.79; 95% confidence interval [CI], 0.61 to 1.03), or rivaroxaban relative to warfarin (adjusted HR 0.93, 95% CI, 0.69 to 1.25). The risk of GI bleeding increased significantly for both dabigatran and rivaroxaban, relative to warfarin, for patients over age 75 years. In Chang et al., there was no significant difference in the risk of GI bleeding for enrollees prescribed dabigatran relative to warfarin (adjusted HR 1.21; 95% CI, 0.96-1.53) or rivaroxaban relative to warfarin (adjusted HR 0.98, 95% CI, 0.36-2.69). There was a borderline significant increase in risk of GI bleeding among patients older than 65 years taking dabigatran relative to warfarin (adjusted HR 1.33; 95% CI, 0.98 to 1.83).

Table 1. Risk of GI Bleeding in Two Retrospective Cohort Studies of U.S. Adults With AF Prescribed Oral Anticoagulants

	Abraham et al.		Chang et al.
Cohort size (enrollees)	92,816		46,163
Prescribed warfarin (%)	67,985 (73.2%)		39,607 (85.8%)
Prescribed dabigatran (%)	16,253 (17.5%)		4,905 (10.6%)
Prescribed rivaroxaban (%)	8,578 (9.2%)		1,649 (3.6%)
Incidence of GI bleeding per 100 person years
Prescribed warfarin (%)	2.87 (2.41-3.41)		7.02 (NA)
Prescribed dabigatran (%)	2.29 (1.88-2.79)		9.01 (NA)
Prescribed rivaroxaban (%)	2.84 (2.30-3.52)		3.41 (NA)
Crude hazard ratio for GI bleeding (95% CI)
Dabigatran relative to warfarin	Not reported		1.20 (0.96-1.52)
Rivaroxaban relative to warfarin	Not reported		0.95 (0.31-2.94)
Adjusted hazard ratio for GI bleeding (95% CI)
Dabigatran relative to warfarin	0.79 (0.61-1.03)		1.21 (0.96-1.53)
Rivaroxaban relative to warfarin	0.93 (0.69-1.25)		0.98 (0.36-2.69)

Conclusion

These two retrospective cohort analyses suggest that the risk of GI bleeding among patients with AF is similar between patients taking NOACs and warfarin,^3,4 a finding that contrasts with prior clinical trial data suggesting a higher risk of GI for patients with AF taking NOACs relative to warfarin.^1,2

Commentary/Perspective

The NOACs represent a convenient alternative to warfarin, with fixed dose regimens and no need for laboratory monitoring. Clinical trial data suggest that the NOACs are non-inferior to warfarin for prevention of stroke in AF but also increase the risk of GI bleeding by 25%.¹ This finding is worrisome because GI bleeding is a major cause of morbidity and medical costs in the U.S., accounting for more than 1 million clinic visits, and 140,000 hospital admissions in 2009.⁵ Yet, the risk of GI bleeding with NOACs is poorly understood outside of clinical trial contexts. Results from these two cohort studies suggest that the risk of GI bleeding associated with use of NOACs may be lower in real-world use relative to clinical trial settings^3,4, and supports prior findings that the increased risk of bleeding associated with the NOACs relative to warfarin may be greatest in older patients.^6,7 Differences in outcomes between these retrospective cohort analyses and prior pivotal clinical trials may be due, in part, to the greater mean age of patients enrolled in the clinical trials relative to the retrospective cohort analysis.

Results from these two studies should be interpreted with caution in light of the following limitations. First, as retrospective cohort studies, they could not account for unmeasured confounding between exposed groups. Second, as claims-based studies, they were vulnerable to variation in coding practices. Third, they assumed that prescribed medications were taken, and that over-the-counter medications that could influence bleeding risk (i.e., aspirin, proton pump inhibitors, etc.) were not used at different rates among study populations.

For clinicians, these studies provide further insight into the risk of GI bleeding in patients with AF taking NOACs and suggest that age may play an important role in risk stratification for GI bleeding in this population.

References

1. Abraham NS, Singh S, Alexander GC, et al. Comparative risk of gastrointestinal bleeding with dabigatran, rivaroxaban, and warfarin: a population based cohort study. BMJ 2015;350:h1587.
2. Chang H, Zhou M, Tang W, et al. Risk of gastrointestinal bleeding associated with oral anticoagulants: a population based retrospective cohort study. BMJ 2015;350:h1585.
3. Ruff CT, Giugliano RP, Braunwald E, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials. Lancet 2014;383:955-62.
4. Holster IL, Valkhoff VE, Kuipers EJ, et al. New oral anticoagulants increase risk for gastrointestinal bleeding: a systematic review and meta-analysis. J Gastroenterol 2013;145:105-12.
5. Peery AF, Dellon ES, Lund J, et al. Burden of gastrointestinal disease in the United States: 2012 update. J Gastroenterol 2012;143:1179-87.
6. Graham DJ, Reichman ME, Wernecke M, et al. Cardiovascular bleeding, and morality risks in elderly Medicare patients treated with dabigatran or warfarin for non-valvular atrial fibrillation. Circulation 2015;131:157-64.
7. Hernandez I, Baik SH, Pinera A, et al. Risk of bleeding with dabigatran in atrial fibrillation. JAMA Intern Med 2015;175:18-24.

Keywords: Ambulatory Care, Anticoagulants, Aspirin, Atrial Fibrillation, Benzimidazoles, International Classification of Diseases, Morpholines, Outcome Assessment, Health Care, Propensity Score, Proportional Hazards Models, Proton Pump Inhibitors, Retrospective Studies, Risk Factors, Stroke, Thiophenes, Warfarin, beta-Alanine

< Back to Listings