What is the strength of evidence available on the risk of GI bleeding with direct oral anticoagulants?

Introduction

The major randomized controlled trials (RCTs) evaluating use of direct oral anticoagulants (DOACs) for stroke prevention in patients with atrial fibrillation (AF) have demonstrated significant reduction in mortality and intracranial hemorrhage compared to warfarin.1,2 However, the risk of gastrointestinal (GI) bleeding compared to warfarin has shown varying results with individual DOACs.1-6 As shown in Table 1, rivaroxaban and higher doses of dabigatran and edoxaban demonstrated an increased risk of major GI bleed compared to warfarin in pivotal stroke prevention RCTs.1,3-6 Important limitations of comparing these risks from individual trials include varying definitions of major GI bleed and the difference in CHADS2 (congestive heart failure, hypertension, age ≥ 75 years, diabetes, and previous ischemic event) score of included patients in each trial. The CHADS2 score aids in identifying patients as low (0 to 1), medium (2), or high risk (≥ 3) for stroke.1 A higher score is also associated with an increased risk of major bleeding as similar factors of hypertension, older age, and history of stroke are used to determine the HAS-BLED score – an assessment of bleeding risk- in patients receiving warfarin.1,2  Patients in the ROCKET-AF trials had an average CHADS2 score of 3.5 compared to 2.1 in RE-LY and ARISTOTLE and 2.8 in ENGAGE-TIMI-48.1,3-6

Table 1. Risk of major GI bleeding in pivotal stroke prevention DOAC clinical trials.3-6
StudyHazard ratio (95% CI)
for major GI bleeding vs warfarin
RE-LYDabigatran 110 mg twice daily: 1.10 (0.86 to 1.41)
Dabigatran 150 mg twice daily: 1.50 (1.19 to 1.89)
ROCKET-AFRivaroxaban 20 mg once daily: 1.66 (1.34 to 2.05)
ARISTOTLEApixaban 5 mg twice daily: 0.89 (0.70 to 1.15)
ENGAGE AF-TIMI 48Edoxaban 30 mg once daily: 0.67 (0.53 to 0.83)
Edoxaban 60 mg once daily: 1.23 (1.02 to 1.50)
Abbreviations: CI, confidence interval; DOAC, direct oral anticoagulant; GI, gastrointestinal.

Randomized controlled trials specifically comparing the risk of GI bleed as a primary outcome between individual DOACs or to warfarin are not available.1,2 The current evidence is comprised of safety outcomes from the major RCTs (as shown in Table 1), observational studies, and indirect comparisons from meta-analyses and systematic reviews.1,2,7-20 The purpose of this review is to provide a summary of published evidence within the past year on the risk of GI bleeding associated with DOACs. A summary of observational/retrospective studies is provided followed by a table summary of recent meta-analyses (Table 2).

Literature Review

Observational/Retrospective studies
In a retrospective cohort study of 50,000 patients with venous thromboembolism (VTE) identified from a health care insurance database, the incidence of GI bleeding in new users of apixaban and rivaroxaban was determined.7 The primary safety outcome was a composite of GI and intracranial bleeding. Diagnosis of bleeding was based on International Classification of Diseases (ICD)-9 and -10 revision codes for bleeding requiring hospitalization. The incidence (per 100 patient-years) of GI bleeding was 7.2% with apixaban and 10.6% with rivaroxaban (hazard ratio [HR] 0.60, 95% confidence interval [CI] 0.53 to 0.69).

The number of total reports of hemorrhage reported from Jan 2014 to Dec 2019 to the Food and Drug Administration Adverse Event Reporting System (FAERS) were 53,085 for rivaroxaban, 13,151 for apixaban, and 14,100 for dabigatran.8 The reporting odds ratio (ROR) was used to compare the hemorrhagic risk of apixaban and rivaroxaban to dabigatran. This ratio adjusted for differences in the proportion of total reports for each agent. An ROR ≥ 2 was considered an indicator of high risk and between 1 and 2 was considered a moderate risk. For GI hemorrhage, the ROR for rivaroxaban compared to dabigatran was 1.38 (95% CI, 1.34 to 1.42) and for apixaban compared to dabigatran was 0.28 (95% CI, 0.27 to 0.29).

A cohort study of new users of apixaban (n=2157), dabigatran (n=494) and rivaroxaban (n=3217) between 2014 and 2019 compared the incidence and risk of GI bleed.9 The number of events per 100 person-years for overall GI bleed with rivaroxaban compared to apixaban was 3.2 versus 2.5 (HR 1.42, 95% CI 1.04 to 1.93) and for major GI bleed was 1.9 versus 1.4 (HR 1.50; 95% CI 1.00 to 2.24).  Although differences were seen in the number of overall and major GI bleeding events with rivaroxaban compared to dabigatran, the HR did not indicate a statistically significantly higher risk with rivaroxaban. A subgroup analysis of patients with AF demonstrated higher overall (but not major) GI bleed risk with rivaroxaban compared to apixaban and dabigatran.

Using information from European and Canadian health care databases, another retrospective cohort study evaluated the risk of major bleeding with vitamin K antagonists (VKAs) and DOACs in over 400,000 adult patients with AF.10  In the analysis of risk of GI bleeding with individual DOACs compared to VKAs, the risk was higher with rivaroxaban (HR 1.28, 95% CI 1.18 to 1.38) and dabigatran (HR 1.21, 95% CI 1.07 to 1.37) and lower with apixaban (HR 0.77, 95% CI 0.67 to 0.87).

In a smaller retrospective study of approximately 1200 patients, a statistically significantly higher risk of GI bleeding was not found between patients who received VKA compared with  rivaroxaban or dabigatran.11 Factors associated with a higher risk of GI bleed included age > 75 years, chronic kidney disease (CKD), concomitant treatment with clopidogrel, and a history of previous GI bleed. A HAS-BLED score of < 3 was associated with a lower risk of GI bleed.

An observational, prospective cohort study evaluated the 1-year risk of major bleeding in 908 patients 80 years or older who received either rivaroxaban or a VKA for AF.12 No significant difference in the rate or risk of GI bleeding was observed between the groups when controlled for confounding factors. Factors associated with a higher risk of major bleeding included advanced age, male sex, reduced renal function, and anemia.

The GI bleeding risk was not found to be significantly different between rivaroxaban and warfarin (HR 0.97, 95% CI 0.87 to 1.09) or dabigatran and warfarin (HR 0.88, 95% CI 0.71 to 1.08) in a retrospective propensity score-matched cohort study of over 56,000 adults with valvular AF.13 Apixaban was associated with a lower risk of GI bleed compared to warfarin (HR 0.56, 95% CI 0.50 to 0.63).

Meta-analyses
Table 2 provides a summary of meta-analyses of RCTs and observational studies evaluating GI bleed risk between DOACs and VKAs and between individual DOACs.

Table 2. Meta-analyses of risk of GI bleed with DOACs.14-20 
Study/Population
Intervention/ Definition of GI bleed
Outcome
Randomized controlled trials
Chen 2022 et al14
MA
31 RCTs
N=76,641 patients receiving VTE prevention or treatment

DOACs (apixaban, dabigatran, edoxaban, rivaroxaban)

VKA

LMWH

Major bleeding defined according to ISTH
VTE treatment
OR (95% CI) for GI bleeding:
Rivaroxaban vs dabigatran:  0.40 (0.17 to 0.95)
Rivaroxaban vs apixaban: 1.71 (0.87 to 3.39)
Rivaroxaban vs edoxaban: 0.65 (0.38 to 1.10)
VKAs vs. rivaroxaban: 1.82 (1.24 to 2.68)

Safety SUCRA ranking (lowest risk to highest) for GI bleeding:
Apixaban
Rivaroxaban
Edoxaban
VKAs
Dabigatran

VTE prevention
No significant difference between groups for GI bleeding

Safety SUCRA ranking (lowest risk to highest) for GI bleeding:
Apixaban
LMWH
Dabigatran
Edoxaban
Rivaroxaban
Radadiya 2021 et al15

NMA
28 RCTs		N=139,587 patients on anticoagulation for various indications
DOACs (apixaban, dabigatran, edoxaban, rivaroxaban)

Warfarin or enoxaparin

Major bleeding defined according to ISTH		Risk of major GI bleed (compared to warfarin or enoxaparin):
No difference in risk with individual DOACs overall compared to warfarin or enoxaparin

Dose-related risk compared to warfarin:
Rivaroxaban 20 mg: 47% higher
Dabigatran 300 mg: 40% higher
Edoxaban 60 mg: 22% higher
Edoxaban 30 mg: 33% lower
Apixaban 2.5 or 5 mg: no difference
Lower doses of rivaroxaban and dabigatran: no difference

OR (95% CI) for major GI bleed between DOACs (statistically significant results shown):
Apixaban 5 mg twice daily vs dabigatran 300 mg daily: 0.63(0.44 to 0.88)
Apixaban 5 mg twice daily vs rivaroxaban 20 mg daily: 0.6 (0.43 to 0.83)
Edoxaban 30 mg once daily vs rivaroxaban 20 mg daily: 0.45 (0.33 to 0.62)
Edoxaban 30 mg once daily vs 60 mg once daily: 0.55 (0.41 to 0.74)
Rivaroxaban 15 mg once daily vs 20 mg once daily: 0.31 (0.12 to 0.79)
Aloysisus 2021 et al16

SR, MA
14 RCTs		N=87,407 patients on anticoagulation for various indications

DOACs (apixaban, dabigatran, edoxaban, rivaroxaban)

Warfarin

No specific definition of GI bleed		RR of any GI bleed
DOACs vs warfarin:1.04, 95% CI 0.85 to 1.27
Apixaban vs warfarin: 1.04, 95% CI 0.72 to 1.51
Dabigatran vs warfarin: 1.23, 95% CI 1.03 to 1.48
Edoxaban vs warfarin: 0.48, 95% CI 0.19 to 1.17
Rivaroxaban vs warfarin: 1.38, 95% CI 1.12 to 1.71

RR of fatal GI bleed:
DOACs vs warfarin: 0.36, 95% CI 0.15 to 0.82
No increased risk with any individual DOAC vs warfarin
Xu 2021 et al17

SR, NMA
23 RCTs		N=87,616 patients with AF

DOACs (apixaban, dabigatran, edoxaban, rivaroxaban)

VKA

Major bleeding defined according to ISTH		OR (95% CI) for GI bleeding:
No statistically significant difference between VKA and DOACs
No statistically significant difference between individual DOACs

Safety SUCRA ranking (lowest risk to highest) for GI bleeding:
Apixaban
VKA
Edoxaban
Rivaroxaban
Dabigatran
Observational/retrospective studies
Hu 2022 et al18

NMA
10 observational cohort studies		N=59,244 patients with AF and prior GI bleed on anticoagulation
DOACs (n=27,793)
VKA (n=24,635)
No anticoagulation (n=6816)

Average HAS-BLEDa score between 2 and 4.2

Average CHA2DS2-VAScb score between 3 and 4.3

No standard definition of GI bleed across included studies (could include use of ICD-9 codes or specific objective criteria)		HR (95% CI) for recurrent GI bleed (statistically significant results shown):
Warfarin vs no resumption of anticoagulation: 1.33 (1.06 to 1.7)
Rivaroxaban vs no resumption of anticoagulation: 1.67 (1.16 to 2.65)
Rivaroxaban vs dabigatran: 1.88 (1.06 to 3.75)

No significant differences between resuming warfarin vs individual DOACs
Mamas 2022 et al 19

SR, MA
8 retrospective studies		N=592,772 patients with NVAF

Apixaban
Rivaroxaban

Average HAS-BLEDa score between 1.8 and 3.2

Average CHA2DS2-VAScb score between 2.4 and 4.6

No standard GI bleed definition stated		HR (95% CI) for GI bleeding:
Apixaban vs rivaroxaban: 0.57 (0.5 to 0.64)
Jiang 2021 et al20

SR, MA
33 observational studies		N=1,391,923 with NVAF

Rivaroxaban
Warfarin

No standard GI bleed definition stated		Frequency and risk of GI bleed in overall population:
Rivaroxaban: 2.63%
Warfarin: 2.48%;
OR (rivaroxaban vs warfarin) 1.06, 95% CI 0.96 to 1.17, p=0.26

Frequency and risk of GI bleed in CKD/dialysis population:
Rivaroxaban: 3.69%
Warfarin: 8.09%
OR (rivaroxaban vs. warfarin) 0.70, 95% CI 0.54 to 0.90, p=0.005

aHAS-BLED score assesses bleeding risk in patients taking warfarin. A score of 3 or greater indicates high risk. bCHA2DS2-VASc score is similar to the CHADS2 with additional components to more accurately assess the risk of stroke in patients with atrial fibrillation.
Abbreviations: AF, atrial fibrillation; CI, confidence interval, CKD, chronic kidney disease; DOAC, direct oral anticoagulant; GI, gastrointestinal; HR, hazard ratio; ICD-9, International Classification of Diseases, 9th edition; ISTH, International Society of Thrombosis and Haemostasis; LMWH, low molecular weight heparin; MA, meta-analysis; NMA, network meta-analysis; NVAF, non-valvular atrial fibrillation; OR, odds ratio; RCT, randomized clinical trial; SR, systematic review; SUCRA, surface under the cumulative ranking curves; VKA, vitamin K antagonist; VTE, venous thromboembolism.   

Literature Summary

Observational studies

Results of recent observational studies indicate a higher incidence and risk of GI bleed with rivaroxaban compared to apixaban, a higher risk GI hemorrhages with rivaroxaban compared to dabigatran, a similar to slightly higher risk with rivaroxaban and dabigatran compared to VKA, and a lower risk with apixaban compared to VKA.7-10,13 Small retrospective studies have not demonstrated a difference in GI bleeding between rivaroxaban or dabigatran when compared to a VKA, but did identify patient factors associated with a higher risk of GI bleeding including age over 75 years, male sex, CKD, concomitant use of clopidogrel, anemia, and a previous GI bleed.11,12

Meta-analyses

Meta analyses of safety outcomes from RCTs of DOACs and VKAs have demonstrated mixed results.14-17 For rivaroxaban, meta-analyses have shown conflicting results when compared to VKAs demonstrating either no difference 15,17or a higher risk of GI bleed.16 Compared to rivaroxaban, VKAs demonstrated a higher risk in one analysis.14 When compared to other DOACs, rivaroxaban did not demonstrate a significantly higher risk of GI bleed in most analyses14,17 and was shown to have a lower risk compared to dabigatran in 1 analysis.14 Similarly, the risk of GI bleed with dabigatran compared to VKAs was observed to be similar or higher.15,16  One meta-analysis demonstrated a higher GI bleed risk with higher doses of rivaroxaban 20 mg, edoxaban 60 mg, and dabigatran 300 mg compared to VKA or apixaban.15 One consistent finding across the meta-analyses was the lack of a higher risk of GI bleed with apixaban compared to either VKA or other DOACs; results demonstrated either no difference or a lower risk.14-17 Two meta-analyses ranked apixaban as the safest among DOACs or VKAs with respect to the risk of GI bleeding.14,17

Three recent meta-analyses of observational studies have evaluated studies comparing individual DOACs or DOACs to VKAs.18-20 The risk of recurrent GI bleed was higher with resumption of warfarin or rivaroxaban when compared to no anticoagulation in a meta-analysis of 10 observational cohort studies that evaluated patients with a prior GI bleed on anticoagulation.18 Patients who resumed anticoagulation with rivaroxaban had a higher risk of GI bleed compared to those who received dabigatran. No significant differences were observed between resuming warfarin compared to individual DOACs. Another meta-analysis of 33 observational studies did not find a significant difference in GI bleed risk between rivaroxaban and warfarin; however, in a subgroup of patients with CKD or on dialysis, rivaroxaban demonstrated a lower risk of GI bleed.19 In another analysis of 8 retrospective studies, the risk of GI bleeding was significantly lower in patients who received apixaban compared to those who received rivaroxaban.20

Conclusion

Although the risk of intracranial bleeding risk has been demonstrated to be lower with DOACs compared to VKAs, the risk of GI bleeding has not been lower with all DOACs compared to VKAs. The results of recent observational studies and meta-analyses indicate that apixaban has a similar or lower risk of GI bleeding compared to VKAs and either the same or lower risk compared to other DOACs. The recommended VTE treatment doses of rivaroxaban, dabigatran, and edoxaban may cause a higher risk of GI bleeding compared to VKAs or apixaban. According to the studies reviewed, factors to consider when assessing GI bleed risk and choosing a VKA over a DOAC or between individual DOACs include: indication for use of anticoagulation (with respect to dose and whether the agent is approved for the indication), age, male sex, presence of CKD, history of GI bleed, anemia, and concomitant clopidogrel use. Additional unanswered questions include the differences between agents in the risk of minor, major and fatal GI bleeding, the occurrence of upper or lower GI bleeding, and differences in GI bleeding treatment outcomes.

References

  1. Benamouzig R, Guenoun M, Deutsch D, Fauchier L. Review Article: Gastrointestinal bleeding risk with direct oral anticoagulants. Cardiovasc Drugs Ther. 2021 Jun 18. doi: 10.1007/s10557-021-07211-0.
  2. Zappulla P, Calvi V. Gastrointestinal bleeding and direct oral anticoagulants among patients with atrial fibrillation: risk, prevention, management, and quality of life. TH Open. 2021 Jun 16;5(2):e200-e210. doi: 10.1055/s-0041-1730035.
  3. Connolly SJ, Ezekowitz MD, Yusuf S, et al; RE-LY Steering Committee and Investigators. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361(12):1139-51. doi: 10.1056/NEJMoa0905561.
  4. Patel MR, Mahaffey KW, Garg J, et al; ROCKET AF Investigators. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365(10):883-91. doi: 10.1056/NEJMoa1009638.
  5. Granger CB, Alexander JH, McMurray JJ, et al; ARISTOTLE Committees and Investigators. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365(11):981-92. doi: 10.1056/NEJMoa1107039.
  6. Giugliano RP, Ruff CT, Braunwald E, et al; ENGAGE AF-TIMI 48 Investigators. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2013;369(22):2093-104. doi: 10.1056/NEJMoa1310907.
  7. Dawwas GK, Leonard CE, Lewis JD, Cuker A. Risk for recurrent venous thromboembolism and bleeding with apixaban compared with rivaroxaban: an analysis of real-world data. Ann Intern Med. 2022;175(1):20-28. doi: 10.7326/M21-0717.
  8. Guo M, Thai S, Zhou J, et al. Evaluation of rivaroxaban-, apixaban- and dabigatran-associated hemorrhagic events using the FDA-Adverse event reporting system (FAERS) database. Int J Clin Pharm. 2021;43(6):1508-1515. doi: 10.1007/s11096-021-01273-8.
  9. Ingason AB, Hreinsson JP, Ágústsson AS, et al. Rivaroxaban is associated with higher rates of gastrointestinal bleeding than other direct oral anticoagulants : a nationwide propensity score-weighted study. Ann Intern Med. 2021;174(11):1493-1502. doi: 10.7326/M21-1474.
  10. van den Ham HA, Souverein PC, Klungel OH, et al. Major bleeding in users of direct oral anticoagulants in atrial fibrillation: A pooled analysis of results from multiple population-based cohort studies. Pharmacoepidemiol Drug Saf.  2021;30(10):1339-1352. doi: 10.1002/pds.5317.
  11. Agudo-Fernández S, Castaño Milla C, González Blanco A, Olmos Jerez JA, Calvo Morillas I, Sancho Del Val L. RHEDAR study: Determination of the risk of gastrointestinal hemorrhage in treatment with dabigatran, acenocoumarol and rivaroxaban. J Gastroenterol Hepatol. 2021;36(10):2794-2802. doi: 10.1111/jgh.15547.
  12. Hanon O, Vidal JS, Pisica-Donose G, et al; SAFIR study group. Bleeding risk with rivaroxaban compared with vitamin K antagonists in patients aged 80 years or older with atrial fibrillation. Heart. 2021;107(17):1376-1382. doi: 10.1136/heartjnl-2020-317923.
  13. Dawwas GK, Dietrich E, Cuker A, Barnes GD, Leonard CE, Lewis JD. Effectiveness and safety of direct oral anticoagulants versus warfarin in patients with valvular atrial fibrillation : a population-based cohort study. Ann Intern Med. 2021;174(7):910-919. doi: 10.7326/M20-6194.
  14. Chen J, Lv M, Wu S, et al. Severe bleeding risks of direct oral anticoagulants in the prevention and treatment of venous thromboembolism: a network meta-analysis of randomised controlled trials. Eur J Vasc Endovasc Surg. 2022;63(3):465-474. doi: 10.1016/j.ejvs.2021.10.054.
  15. Radadiya D, Devani K, Brahmbhatt B, Reddy C. Major gastrointestinal bleeding risk with direct oral anticoagulants: Does type and dose matter? – A systematic review and network meta-analysis. Eur J Gastroenterol Hepatol. 2021;33(1S Suppl 1):e50-e58. doi: 10.1097/MEG.0000000000002035.
  16. Aloysius MM, Perisetti A, Goyal H, et al. Direct-acting oral anticoagulants versus warfarin in relation to risk of gastrointestinal bleeding: a systematic review and meta-analysis of randomized controlled trials. Ann Gastroenterol. 2021;34(5):651-659. doi: 10.20524/aog.2021.0658.
  17. Xu W, Lv M, Wu S, et al. Severe bleeding risk of direct oral anticoagulants versus vitamin K antagonists for stroke prevention and treatment in patients with atrial fibrillation: a systematic review and network meta-analysis. Cardiovasc Drugs Ther. 2021 Aug 26. doi: 10.1007/s10557-021-07232-9.
  18. Hu W, Cai H, Zhang J. Direct oral anticoagulants versus warfarin in nonvalvular atrial fibrillation patients with prior gastrointestinal bleeding: a network meta-analysis of real-world data. Eur J Clin Pharmacol. 2022 Mar 16. doi: 10.1007/s00228-022-03300-7.
  19. Mamas MA, Batson S, Pollock KG, et al. Meta-analysis comparing apixaban versus rivaroxaban for management of patients with nonvalvular atrial fibrillation. Am J Cardiol. 2022;166:58-64. doi: 10.1016/j.amjcard.2021.11.021.
  20. Jiang H, Jiang Y, Ma H, Zeng H, Lv J. Effects of rivaroxaban and warfarin on the risk of gastrointestinal bleeding and intracranial hemorrhage in patients with atrial fibrillation: systematic review and meta-analysis. Clin Cardiol. 2021;44(9):1208-1215. doi: 10.1002/clc.23690.

Prepared by:
Rita Soni, PharmD, BCPS
Clinical Assistant Professor, Drug Information Specialist
University of Illinois at Chicago College of Pharmacy

May 2022

The information presented is current as of April 11, 2022. This information is intended as an educational piece and should not be used as the sole source for clinical decision-making.