Different glomerular filtration rate estimating formula for prescribing DOACs in oldest patients: appropriate dosage and bleeding risk. Post hoc analysis of a prospective cohort.

Background: Direct oral anticoagulants (DOACs) pharmacokinetics depends on estimated glomerular filtration rate (eGFR), whose estimation is crucial for optimal risk/benefit balance.
Aims: To assess the concordance among different eGFR formulas and the potential impact on DOACs prescription appropriateness and bleeding risk in oldest hospitalized patients.
Methods: Post hoc analysis of a single-centre prospective cohort study. eGFR was calculated by creatinine-based (MDRD, CKD-EPI _Cr , BIS ₁ ) and creatinine-cystatin-C-based (CKD-EPI _Comb and BIS ₂ ) formulas. Patients were stratified according to eGFR [severely depressed (SD) 15-29; moderately depressed (MD) 30-49; preserved/mildly depressed (PMD): ≥ 50 ml/min/1.73 m ² ]. Concordance between the different equations was assessed by Cohen's kappa coefficient.
Results: Among AF patients, 841 (59.2% women, mean age 85.9 ± 6.5 years) received DOACs. By CKD-EPI _Cr equation, 135 patients were allocated in the SD, 255 in the MD and 451 in the PMD group. The concordance was excellent only between BIS 2 and CKD-EPI _Comb and MDRD and CKD-EPI _Cr , while was worse (from good to poor) between the other formulas. Indeed, by adding cystatin-C almost over 1/3 of the patients were reallocated to a worse eGFR class. Bleeding prevalence increased by 2-3% in patients with discordant eGFR between formulas, reallocated to a worse chronic kidney disease (CKD) stage, although without reaching statistical significance. CKD-EPI _Comb resulted the best predictor of bleeding events (AUROC 0.71, p = 0.03).
Discussion: This study highlights the variability in CKD staging according to different eGFR formulas, potentially determining inappropriate DOACs dosing. Although the cystatin-C derived CKDEPI _Comb equation is the most accurate for stratifying patients, BIS ₁ may represent a reliable alternative.
(© 2021. The Author(s).)