BACKGROUND: Fludarabine is often used as an important drug in reduced toxicity conditioning regimens prior to hematopoietic cell transplantation (HCT). As no definitive pharmacokinetic (PK) basis for HCT dosing for the wide age and weight range in HCT is available, linear body surface area (BSA)-based dosing is still used. OBJECTIVE: We sought to describe the population PK of fludarabine in HCT recipients of all ages. METHODS: From 258 HCT recipients aged 0.3-74 years, 2605 samples were acquired on days 1 (42%), 2 (17%), 3 (4%) and 4 (37%) of conditioning. Herein, the circulating metabolite of fludarabine was quantified, and derived concentration-time data were used to build a population PK model using non-linear mixed-effects modelling. RESULTS: Variability was extensive where area under the curve ranged from 10 to 66 mg*h/L. A three-compartment model with first-order kinetics best described the data. Actual body weight (BW) with standard allometric scaling was found to be the best body-size descriptor for all PK parameters. Estimated glomerular filtration rate (eGFR) was included as a descriptor of renal function. Thus, clearance was differentiated into a non-renal (3.24+/-20% L/h/70 kg) and renal (eGFR 0.782+/-11% L/h/70 kg) component. The typical volumes of distribution of the central (V1), peripheral (V2), and second peripheral (V3) compartments were 39+/-8%, 20+/-11%, and 50+/-9% L/70 kg respectively. Intercompartmental clearances between V1 and V2, and V1 and V3, were 8.6+/-8% and 3.8+/-13% L/h/70 kg, respectively. CONCLUSION: BW and eGFR are important predictors of fludarabine PK. Therefore, current linear BSA-based dosing leads to highly variable exposure, which may lead to variable treatment outcome.