#3207 KIDNEY FUNCTION ESTIMATORS FOR DRUG DOSE ADJUSTMENT OF DIRECT ORAL ANTICOAGULANTS IN OLDER ADULTS WITH ATRIAL FIBRILLATION

Abstract Background and Aims The Food and Drug Administration and the European Medicines Agency (EMA) recommend using the Cockcroft-Gault equation (CG) for drug dose adjustment of direct oral anticoagulant drugs (DOACs) in non-valvular atrial fibrillation (NVAF), whereas Cardiology guidelines recommend using estimated glomerular filtration rate (eGFR). This issue is of great importance in older adults, which are more prone to adverse drug reactions and have higher prevalence of atrial fibrillation as compared to younger adults, and in whom CG has worse diagnostic performance than most eGFR equations. We aimed to assess if prognosis was similar according to drug dose status using different kidney function estimators based on creatinine and/or cystatin C in older adults with NVAF. Method We used data from the Berlin Initiative Study (BIS): a population-based prospective cohort study initiated in 2009, with five biennial study visits. Participants with a history of NVAF (based on ICD-10 codes) and a dispensed prescription of DOAC four months prior to baseline or a follow-up visit according to claims data were included. Drug dose status was defined according to the EMA guidelines. CG, deindexed (in ml/min) creatinine and/or cystatin C-based CKD-EPI, and deindexed BIS 1 (creatinine-based) and BIS 2 (creatinine and cystatin C-based) equations were included. Associations between dosing status and mortality, stroke or systemic embolism, and bleeding events were assessed using marginal structural Cox models with time-varying variables and taking account of various confounders. Subgroup analyses were performed in rivaroxaban and apixaban users, but not in dabigatran and edoxaban users because of too small sample sizes. Results Two hundred twenty four patients treated with DOACs were included in this analysis (median age 87 years, median eGFRCKD-EPIcr 56 ml/min/1.73m², median follow-up length 40 months for the mortality analysis). Of those, 99 (44%) were taking rivaroxaban, 86 (38%) apixaban, 21 (9%) edoxaban, and 18 (8%) dabigatran. Using CG, 154 (69%) had appropriate DOAC dose at baseline, 52 (23%) were underdosed, and 18 (8%) were overdosed. Discrepancies were found by comparing dosing status according to CG and eGFR equations (Figure 1). During the follow-up period, 109 (14.9/100 person-years) participants died, 25 (3.6/100 person-years) experienced a stroke or systemic embolism, and 60 (9.9/100 person-years) experienced a bleeding event. Drug dose status was not significantly associated with mortality and the occurrence of stroke or systemic embolism, whatever equation was used. Underdose status was associated with a significantly lower risk of bleeding events with all the equations but overdose status was not associated with a higher risk of bleeding events (Figure 2). In subgroup analyses, drug dose status was not associated with mortality and bleeding event in apixaban users, whatever equation was used. In rivaroxaban users, underdose status was associated with a significant higher risk of death by using CG, eGFRCKD-EPIcys, and BIS 2, and a lower risk of bleeding event by using eGFRCKD-EPIcr. Conclusion In this population of very old adults with NVAF, drug dose status of DOAC was not associated with mortality or the occurrence of stroke or systemic embolism for any of the studied equations. However, underdose status was associated with bleeding events occurrence regardless of the equation used. These associations differed according to the used drug, but this should be interpreted with cautious due to small sample sizes. Our results do not allow us to provide any guidance which equation to use in this context. A study including a larger group of patients with discrepancy in dose status according to the used equations would be of great interest.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.