Renal Function Estimates and Dosing of Direct Oral Anticoagulants in Stroke Patients with Atrial Fibrillation: An Observational Study.

Abstract

Purpose: Appropriate dosing of direct oral anticoagulants (DOACs) requires consideration of renal function. Discordance between commonly used estimated glomerular filtration rate (eGFR) and creatinine clearance (CrCl) might affect the dosing appropriateness in stroke patients with atrial fibrillation (AF). We aimed to explore the effect of renal function estimates on the dosing patterns in a real-world setting.

Methods: Using a hospital-based stroke registry, we identified consecutive patients between 2014 and 2017 who were hospitalized for acute stroke, had AF, and started DOACs within 90 days after stroke. We compared the difference between eGFR and CrCl in assessing appropriateness of dosage. Effectiveness and safety outcomes were verified by chart review, and event rates were presented as per 100 person-years.

Results: Of the156 patients with mean age 74±11 years, 72 (46%) were prescribed dabigatran and 84 (54%) rivaroxaban. Substituting eGFR for CrCl would have 55% (37/67) of patients with CrCl less than 50 mL/min and 89% (8/9) of patients with CrCl less than 30 mL/min not correctly classified, and potentially lead to overdosing. The misclassification would cause underdosing in 6% (5/89) of patients with CrCl ≥50 mL/min and 1% (1/147) of patients with CrCl ≥30 mL/min. In reality, the substitution resulted in reduction of overdosing from 10% to 4% for dabigatran and from 2% to 1% for rivaroxaban; underdosing increased from 17% to 26% for rivaroxaban. After median follow-up of 17 months, 33 patients developed outcomes including 21 major bleedings. The event rate was 6.9% per year (95% CI, 4.1%-11.4%) for effectiveness, and 9.6% per year (95% CI, 6.3%-14.8%) for safety.

Conclusion: Although substituting eGFR for CrCl carries potential risks of DOAC overdosing in patients with AF, the effect might be offset by clinicians' predilection for lower dosage in this stroke cohort.