Kidney function estimators for drug dose adjustment of direct oral anticoagulants in older adults with atrial fibrillation

Abstract

Background: The Cockcroft-Gault equation (CrCl_C-G) is recommended for dose adjustment of direct oral anticoagulant drugs (DOACs) to kidney function. We aimed to assess whether defining DOAC dose appropriateness according to various kidney function estimators changed the associations between dose appropriateness and adverse events in older adults with atrial fibrillation (AF).

Methods: Participants of the Berlin Initiative Study with AF and treated with DOACs were included. We investigated CrCl_C-G and estimated glomerular filtration rate (eGFR) using the Chronic Kidney Disease Epidemiology Collaboration and European Kidney Function Consortium equations based on creatinine and/or cystatin C. Marginal structural Cox models yielded confounder-adjusted hazard ratios for the risk of mortality, thromboembolism and bleeding associated with dose status.

Results: A total of 224 patients were included in the analysis (median age 87 years). Using CrCl_C-G, 154 (69%) had an appropriate dose of DOACs, 52 (23%) were underdosed and 18 (8%) were overdosed. During a 39-month median follow-up period, 109 (14.9/100 person-years) participants died, 25 (3.6/100 person-years) experienced thromboembolism and 60 (9.8/100 person-years) experienced bleeding. Dose status was not associated with mortality and thromboembolism, independent of the equation. Underdose status was associated with a lower risk of bleeding with all the equations compared with the appropriate dose group. In participants with discrepancies in dose status using CrCl_C-G and eGFR equations, the occurrence of endpoints did not differ between participants having an appropriate dose using CrCl_C-G or eGFR.

Conclusion: In older adults with AF, the association of DOAC dose status with adverse events did not differ when using CrCl_C-G or eGFR. Our results suggest that eGFR equations are not inferior to CrCl_C-G within this context.

Keywords: atrial fibrillation; creatinine clearance; glomerular filtration rate; older adults; oral anticoagulant drug.

Graphical Abstract

Figure 1:

Dose status at baseline using various kidney function estimators (A) and Cockcroft–Gault according to drugs (B). C-G: creatinine clearance estimated by the Cockcroft–Gault equation; CKD-EPI and EKFC: GFR estimation using the Chronic Kidney Disease Epidemiology Collaboration and European Kidney Function Consortium equations, based on serum creatinine (cr) and/or serum cystatin (cys).

Figure 2:

Association between dose status and (A) mortality, (B) thromboembolism and (C) bleeding according to various kidney function estimators. Sums of events and sample size can slightly differ between estimators due to the multiple imputation process. C-G: creatinine clearance estimated by the Cockcroft–Gault equation; CKD-EPI and EKFC: GFR estimation using the Chronic Kidney Disease Epidemiology Collaboration and European Kidney Function Consortium equations, based on serum creatinine (cr) and/or serum cystatin (cys).

Figure 3:

Comparison of endpoint occurrence in patients with discrepancies in dose status according to the Cockcroft–Gault equation and eGFR. These participants are represented by the striped bars in Fig. 1A. P-values were obtained using univariable logistic regression models; CKD-EPI and EKFC: GFR estimation using the Chronic Kidney Disease Epidemiology Collaboration and European Kidney Function Consortium equations, based on serum creatinine (cr) and/or serum cystatin (cys).