Predicting Hydroxyurea Responses in Children with Sickle Cell Anemia

Over the past 15 years, numerous prospective and cross-sectional clinical trials have demonstrated that hydroxyurea has both laboratory and clinical efficacy for pediatric patients with sickle cell anemia (SCA). In infants, toddlers, children, and adolescents, hydroxyurea administered at maximum tolerated dose (MTD) leads to significant increases in hemoglobin (Hb) concentration, MCV, and %HbF along with simultaneous decreases in neutrophils, reticulocytes, total bilirubin, and serum lactate dehydrogenase. Clinical benefits include reduction in acute vaso-occlusive events (pain and acute chest syndrome), and emerging data suggest protection against chronic organ damage with a low risk of genotoxicity. For individual patients, however, the %HbF response to hydroxyurea and the MTD dose itself are highly variable. Currently there are no accurate predictors of the final %HbF or the MTD dose.

To address the phenotypic variability, the Hydroxyurea Study of Long-Term Effects (HUSTLE, NCT00305175) was designed to capture prospectively first-dose (20 mg/kg) pharmacokinetics (PK) of hydroxyurea including maximum serum concentration (Cmax), area under the concentration curve (AUC), apparent oral clearance (CL/F), half-life (t1/2), and apparent volume of distribution (V/F). A consistent dose escalation schedule was then used to achieve a stable MTD, at which time PK studies were repeated and hydroxyurea responses (%HbF and MTD dose) were recorded along with other pharmacodynamics parameters.

After written informed consent, a total of 98 pediatric patients commenced hydroxyurea, 65 of whom reached MTD with complete paired PK information. In the majority of patients (39 of 65), a ‘fast/slow’ absorption phenotype identified at first-dose PK analysis was retained at MTD, supporting the concept of a genetic basis for this variation. Inter-individual PK variability was substantially greater than intra-individual variability; for example, the coefficient of variation (%CV) for CL/F was 44.4% on day 1 and 42.3% at MTD among all subjects, but averaged only 12.8% within the same subject. The CL/F was partly dependent on the rate of absorption but was most strongly influenced by subject weight and serum creatinine. At MTD, the average AUC was 114 ± 22.3 mg·h/mL, and the %CV of AUC was reduced from 24.8% on first-dose PK to 19.5% at MTD, likely reflecting dose titration toward a common degree of myelosuppression. In an attempt to predict the MTD dose toward this target AUC using data available at baseline, ‘best-fit’ equations were developed such as the following with or without PK data:

Predicted MTD (mg/kg/day) = 31.9 - [17.1*Baseline Creatinine] - [0.14*Baseline BMI] + [0.0036*Baseline ARC]

Predicted MTD (mg/kg/day) = 32.6 - [15.1*Baseline Creatinine] - [0.16*Baseline BMI] - [0.56*First-dose half-life] + [0.0034*Baseline ARC] - [0.48*PK phenotype (fast=0, slow=1)]

A simpler equation for predicting MTD dose using only baseline creatinine and weight was then developed: Predicted MTD (mg) = 400 - [1000*Baseline Creatinine] + [21*weight]

The relatively small intra-individual PK variability at hydroxyurea MTD compared to baseline PK studies, coupled with a similar degree of drug exposure when patients achieve a stable MTD, supports future investigation to predict the optimal hydroxyurea dose prior to the first dosing. Prediction equations of the MTD should be tested prospectively against standard dose-escalation schedules.

Off Label Use: Hydroxyurea for children with sickle cell disease.