Background and Objective <p>Apixaban is increasingly being used&#xa0;for stroke prevention in patients with end-stage kidney disease with atrial fibrillation undergoing haemodialysis, but no pharmacostatistical model is available for dosage adjustment. This study aimed to develop a population pharmacokinetic model of apixaban in these patients to characterise its dialytic clearance and determine optimal dosing regimens and discontinuation timing before surgery.</p> Methods <p>Patients received 2.5 mg of apixaban twice daily for 9 days, followed by 5 mg twice daily for 8 days after a 5-day washout period (NCT02672709). Apixaban concentrations were measured on and off dialysis. A population pharmacokinetic model was developed using parametric and non-parametric methods. Simulations were performed to assess plasmatic exposure and the time to reach clinically relevant apixaban concentrations after treatment discontinuation for seven dosing regimens and 13 dialysis schedules.</p> Results <p>A total of 289 apixaban concentrations were measured, including 85 during haemodialysis. The best model was a two-compartment model with first-order elimination. Dialytic clearance was estimated at 1.20 L/h with high inter-individual variability. Apixaban daily exposure was proportional to the total daily dose, independent of dosing frequency and dialysis timing. The standard discontinuation period of 48–72 h before surgery was insufficient to achieve clinically negligible concentrations in patients undergoing haemodialysis.</p> Conclusions <p>We propose the first pharmacokinetic model to characterise apixaban clearance in patients with end-stage kidney disease with atrial fibrillation undergoing haemodialysis. Simulations suggest that dialysis timing is not critical for monitoring apixaban, and the discontinuation period before surgery should be extended beyond current recommendations.</p> Graphical Abstract <p></p>