An Interim Evaluation of Systemic F-Ara-a Exposure in Pediatric Allogeneic Hematopoietic Cell Transplant (alloHCT) Recipients Using An Optimal Sampling Design

Fludarabine is a purine analogue used in the preparative regimens of pediatric alloHCT to enhance stem cell engraftment. Administered intravenously as a prodrug, fludarabine (f-ara-AMP) undergoes rapid dephosphorylation in the plasma to the systemically circulating compound, f-ara-a. Despite widespread use, there are no published pharmacokinetic-pharmacodynamic (PK-PD) studies of fludarabine in children undergoing alloHCT. Using an optimal sampling strategy (OSS), we designed a prospective study to evaluate the PK-PD of fludarabine in pediatric alloHCT recipients. We report the year-1 interim PK analysis of this 3-year exposure-response study.

Utilizing prior f-ara-a PK data available in adults and D-optimal sampling methods (PFIM software), we designed an OSS for f-ara-a in children. Based on the OSS, the relative standard errors (RSE), representing the precision of estimated PK parameters, were predicted to be less than 20% in a total of 45 children. An interim analysis was planned after year 1 to ensure the sample collection times selected by the OSS were sufficiently informative. Patients were eligible to participate in PK sampling if they were between 0 to 17 years of age, met protocol specific criteria for alloHCT, and would be receiving fludarabine as part of their preparative regimen. All patients underwent PK sampling with dose 1 of fludarabine. Fludarabine was infused per protocol over 30–60 minutes and 1 mL of whole blood was obtained at 2, 4, 8, and 24 h after the start of infusion. PK sampling was repeated following a subsequent dose of fludarabine (dose 2, 3, 4 or 5) at 2 and 24 h. Plasma samples were analyzed by LC-tandem MS and the assay was linear in the range of 5–500 ng/mL.

PK model development using f-ara-a concentration-time data was carried out using standard population PK methodologies (NONMEM 7.2 software). Further development of a 2-compartment open model was based on exploratory analysis, diagnostic plots and changes in objective function value (OFV). The addition of allometric scaling, with weight built into the base model scaled to a reference patient having the median weight of the population, resulted in a significant drop in the OFV. No other covariates were tested based on exploratory analysis and plots. The model was parameterized in terms of clearance (CL), volume of distribution-central compartment (V_c), volume of distribution-peripheral compartment (V_p), and inter-compartmental clearance (Q). Residual unexplained variability was modeled as being proportional to the predicted concentrations. Area-under-the-curve (AUC) of f-ara-a was derived from the empirical Bayes estimates of individual CL.

A total of 94 quantifiable concentrations from 16 subjects (10 male, 6 female) were available for interim PK modeling. Most patients received fludarabine 30–40mg/m² daily over 3 to 5 days (n=13). In the 3 smallest children (<10kg), fludarabine was dosed at 1.33mg/kg/day for 3 to 4 days. Median age and weight of subjects was 6.5 years (range, 0.3–17) and 23.4kg (6.8-82.3), respectively. Markers for renal function were within normal age limits for all subjects.

A 2-compartment model with linear elimination well described the PK of f-ara-a. The population PK estimates for CL, V_c, V_p, Q, and their RSE (%) were 9.0 (6.3%), 30 (8.9%), 34 (6.4%), and 7.7 (11%), respectively. The final model of this interim analysis estimated f-ara-a CL (L/h) = 9.0 _* (WT/23.4)^0.67. This model predicts f-ara-a CL (%CV) to be lower for children < 10kg (n=3), 3.8 L/h (11.3%) compared to those >10kg, 12.4 L/h (42%). Correspondingly, dose-normalized AUC was predicted to be approximately 2.8 times higher in patients < 10kg. Between-patient variability of CL was estimated to be 23% and the residual variability of concentrations 25%.

The optimal sampling strategy based on adult prior data allows for accurate estimation of f-ara-a population PK parameters in our study of 16 pediatric alloHCT recipients. These interim results suggest body weight may be used to predict f-ara-a clearance, as well as suggest the need for close evaluation of weight-based dosing to prevent over-exposure in very small children. Over the next 2 years we will continue to enroll children in this PK-PD study to confirm the interim PK results and identify exposure-response relationships to inform optimal dosing of fludarabine in pediatric alloHCT.

Off Label Use: Fludarabine (Fludara) has no offical FDA indication for use in children.