Potential Effect of Substituting Estimated Glomerular Filtration Rate for Estimated Creatinine Clearance for Dosing of Direct Oral Anticoagulants

Abstract

To determine the potential effect of substituting glomerular filtration rate (GFR) estimates for renal clearance estimated using the Cockcroft-Gault method (CrCL-CG) to calculate direct oral anticoagulant (DOAC) dosing. Simulation and retrospective data analysis. Community, academic institution, nursing home. Noninstitutionalized individuals aged 19 to 80 from the National Health and Nutrition Examination Survey (NHANES) (2011/12) (n = 4,687) and medically stable research participants aged 25 to 105 (n = 208). Age, height, weight, sex, race, serum creatinine, CrCL-CG, and GFR (according to the Modification of Diet in Renal Disease and Chronic Kidney Disease Epidemiology Collaboration equations). Outcome measures were dosing errors if GFR were to be substituted for CrCL-CG. Renal clearance estimates according to all methods were highly correlated (P < .001), although at lower clearances, substitution of GFR estimates for CrCL-CG resulted in failure to recognize needs for dose reductions of rivaroxaban or edoxaban in 28% of NHANES subjects and 47% to 56% of research subjects. At a CrCL-CG of less than 30 mL/min, GFR estimates missed indicated dosage reductions for dabigatran in 18% to 21% of NHANES subjects and 57% to 86% of research subjects. Age and weight contributed to differences between renal clearance estimates (P < .001), but correction of GFR for body surface area (BSA) did not reduce dosing errors. At a CrCL-CG greater than 95 mL/min, edoxaban is not recommended, and GFR esimates misclassified 24% of NHANES and 39% of research subjects. Correction for BSA reduced misclassification to 7% for NHANES and 14% in research subjects. Substitution of GFR estimates for estimated CrCl can lead to failure to recognize indications for reducing DOAC dose and potentially higher bleeding rates than in randomized trials.