Population pharmacokinetics of intravenous and oral panobinostat in patients with hematologic and solid tumors

Abstract

Purpose: The study aimed to characterize the population pharmacokinetics of panobinostat, a pan-deacetylase inhibitor that has demonstrated efficacy in combination with bortezomib and dexamethasone in patients with multiple myeloma.

Methods: A nonlinear mixed-effect model was used to fit plasma panobinostat concentration-time data collected from patients across 14 phase 1 and phase 2 trials following either oral or intravenous (IV) administration. The model was used to estimate bioavailabilities of the two oral formulations and the effects of demographic and clinical covariates on the central volume of distribution and clearance of panobinostat.

Results: A total of 7834 samples from 581 patients were analyzed. Panobinostat pharmacokinetic parameters were best characterized by a three-compartment model with first-order absorption and elimination. Bioavailability was 21.4 %. Median clearance was 33.1 L/h. Interindividual variability in clearance was 74 %. For Caucasian patients of median age 61 years, area under the curve (AUC) decreased from 104 to 88 ng · h/mL as body surface area (BSA) increased from the first to third quartiles, 1.8 to 2.1 m(2). For Caucasian patients of median BSA 1.9 m(2), AUC decreased from 102 to 95 ng · h/mL as age increased from the first to third quartiles, 51 to 70 years. For patients of median BSA and median age, AUC ranged across the four race categories from 80 to 116 ng · h/mL. Covariate analysis showed no impact on panobinostat clearance and volume by patients' sex, tumor type, kidney function, liver markers, or coadministered medications. However, separate analyses of dedicated studies have demonstrated effects of liver impairment and CYP3A4 inhibition.

Conclusions: Although covariate analyses revealed significant effects of body size, age, and race on panobinostat pharmacokinetics, these effects were minor compared to the interindividual variability and therefore not clinically relevant when dosing panobinostat in populations similar to those studied.

Fig. 1

Histograms and summary statistics of baseline covariates for the 581 patients in the analysis data set

Fig. 2

Arithmetic mean (SD) plasma panobinostat concentration-time plots following day 1 for intravenous panobinostat at doses of 1.2, 2.4, 4.8, 7.2, and 9 mg/m² (a) and oral panobinostat at doses of 20, 30, and 40 mg for all schedules (b). Semilogarithmic view (data on file)

Fig. 3

Simulated concentration time curves for the two final models, for a 61-year-old Caucasian patient with BSA 1.9 m² (first final model) or weight 76.4 kg (second final model) receiving 20 mg of the FMI on days 1, 3, and 5

Fig. 4

a–b Visual predictive checks. Data for non-Japanese patients are indicated by gray plus signs. Data for Japanese patients are indicated by black squares. The dashed lines connect the medians and the 10th and 90th percentiles of the raw data in the defined bins. The solid lines connect the medians and the 10th and 90th percentiles of the simulations. Dosing was 20 mg on days 1, 3, 5, and 8