What are the real-world studies comparing different Direct Oral Anticoagulants (DOACs)?

Real-World Head-to-Head DOAC Comparisons

The most robust real-world head-to-head DOAC comparison comes from a 2022 multinational study of 527,226 patients with atrial fibrillation, which found apixaban was associated with significantly lower gastrointestinal bleeding risk compared to dabigatran, edoxaban, and rivaroxaban, while showing similar rates of stroke, intracranial hemorrhage, and mortality across all DOACs. 1

Key Real-World Evidence

Largest Multinational Study (2022)

The most comprehensive real-world comparison analyzed data from five standardized electronic health care databases covering 221 million people across France, Germany, the United Kingdom, and the United States from 2010-2019. 1

Study population breakdown:

• Apixaban: 281,320 patients
• Rivaroxaban: 172,176 patients
• Dabigatran: 61,008 patients
• Edoxaban: 12,722 patients 1

Primary Findings on Gastrointestinal Bleeding

Apixaban demonstrated consistently lower GIB risk compared to all other DOACs:

• vs. Dabigatran: HR 0.81 (95% CI 0.70-0.94)
• vs. Edoxaban: HR 0.77 (95% CI 0.66-0.91)
• vs. Rivaroxaban: HR 0.72 (95% CI 0.66-0.79) 1

This finding remained consistent across multiple subgroups:

• Standard dose: HR 0.72 (95% CI 0.64-0.82) vs. rivaroxaban
• Reduced dose: HR 0.68 (95% CI 0.61-0.77) vs. rivaroxaban
• Chronic kidney disease: HR 0.68 (95% CI 0.59-0.77) vs. rivaroxaban
• Patients ≥80 years old: consistent associations maintained 1

Efficacy and Mortality Outcomes

No substantial differences were observed between DOACs for:

• Ischemic stroke or systemic embolism
• Intracranial hemorrhage
• All-cause mortality 1

This finding aligns with guideline statements that no systematic reviews or randomized trials comparing different DOACs head-to-head have been identified, with certainty in comparative evidence judged "very low" for all relevant outcomes. 2, 3

Absence of Randomized Head-to-Head Trials

Critical limitation: The American Society of Hematology 2020 guidelines explicitly state that no systematic reviews or randomized controlled trials comparing different DOACs head-to-head were found. 2, 3 Subgroup analyses found no interaction between specific DOAC agents and risk of mortality, PE, symptomatic DVT, or major bleeding when compared indirectly against vitamin K antagonists. 2

Clinical Decision-Making Framework

Given the lack of randomized head-to-head data, DOAC selection should be based on:

Patient-Specific Factors

Renal function considerations:

• Dabigatran: ~80% renal clearance (highest)
• Edoxaban: ~50% renal clearance
• Rivaroxaban: ~33% renal clearance
• Apixaban: ~25% renal clearance (lowest) 2, 3

Hepatic function:

• Dabigatran is least reliant on hepatic clearance 2, 3

Bleeding risk profile:

• Apixaban demonstrates the lowest bleeding risk among all DOACs based on real-world evidence 4, 1
• For patients with prior GI bleeding, apixaban or dabigatran 110 mg bid are preferable as they are not associated with increased GI bleeding compared to warfarin 4

Practical Administration Differences

Lead-in parenteral anticoagulation requirement:

• Dabigatran and edoxaban: require LMWH lead-in for 5-10 days
• Rivaroxaban and apixaban: no lead-in required 2, 3

Dosing frequency:

• Once daily: rivaroxaban, edoxaban
• Twice daily: dabigatran, apixaban 2, 5

Food requirements:

• Rivaroxaban: must be taken with food
• Others: no strict food requirement 5

Special Populations

Cancer-Associated Thrombosis

Only edoxaban and rivaroxaban have been compared with LMWH in randomized controlled trials in cancer populations. 2 Apixaban and dabigatran lack published data comparing them to LMWH in the therapeutic cancer setting. 2

Elderly and Chronic Kidney Disease

The real-world multinational study specifically demonstrated consistent safety and efficacy findings for patients aged 80 years or older and those with chronic kidney disease, populations often underrepresented in clinical trials. 1

Important Caveats

Drug-drug interactions: All DOACs are substrates of P-glycoprotein, with rivaroxaban and apixaban also metabolized by CYP3A4. 2, 6 Dabigatran, edoxaban, and betrixaban have the least drug interactions compared to apixaban and rivaroxaban. 5

Residual confounding: Real-world observational studies cannot eliminate the possibility of residual confounding despite propensity score matching. 1

Patient selection bias: Results from observational studies do not inform about a priori patient selection, which may impact outcomes and relative risks. 2