Age, Weight, and CYP2D6 Genotype Are Major Determinants of Primaquine Pharmacokinetics in African Children

Abstract

Low-dose primaquine is recommended to prevent Plasmodium falciparum malaria transmission in areas threatened by artemisinin resistance and areas aiming for malaria elimination. Community treatment campaigns with artemisinin-based combination therapy in combination with the gametocytocidal primaquine dose target all age groups, but no studies thus far have assessed the pharmacokinetics of this gametocytocidal drug in African children. We recruited 40 children participating in a primaquine efficacy trial in Burkina Faso to study primaquine pharmacokinetics. These children received artemether-lumefantrine and either a 0.25- or a 0.40-mg/kg primaquine dose. Seven blood samples were collected from each participant for primaquine and carboxy-primaquine plasma levels determinations: one sample was collected before primaquine administration and six after primaquine administration according to partially overlapping sampling schedules. Physiological population pharmacokinetic modeling was used to assess the impact of weight, age, and CYP2D6 genotype on primaquine and carboxy-primaquine pharmacokinetics. Despite linear weight normalized dosing, the areas under the plasma concentration-time curves and the peak concentrations for both primaquine and carboxy-primaquine increased with age and body weight. Children who were CYP2D6 poor metabolizers had higher levels of the parent compound, indicating a lower primaquine CYP2D6-mediated metabolism. Our data indicate that primaquine and carboxy-primaquine pharmacokinetics are influenced by age, weight, and CYP2D6 genotype and suggest that dosing strategies may have to be reconsidered to maximize the transmission-blocking properties of primaquine. (This study has been registered at ClinicalTrials.gov under registration no. NCT01935882.).

Keywords: CYP2D6; Plasmodium falciparum; pharmacokinetics; primaquine.

Copyright © 2017 Gonçalves et al.

Figures

FIG 1

PQ and C-PQ plasma levels (y axes) after PQ administration (x axes). In panels A and C, PQ levels are presented for participants who received the 0.25- and 0.40-mg/kg PQ doses, respectively. In panels B and D, the C-PQ levels are presented for the 0.25- and 0.40-mg/kg PQ study arms, respectively. Assay results for all samples collected after PQ administration, including those with PQ or C-PQ levels below the limit of detection (i.e., with assigned level of 0 ng/ml), are presented.

FIG 2

Schematic representation of the model. The mass transport of this model can be described with the following rate constants: k12 = 3/MAT, k23 = 3/MAT, k34 = 3/MAT, k40 = CLH,CYP2D6/VL, k45 = CLH,MAO/VL, k50 = CLCPQ/VCPQ, k46 = [QH (1 − EH)]/VL, and k64 = QH/VPQ.

FIG 3

Prediction-corrected visual predictive check of observed data. The left and right panels depict the prediction-corrected visual predictive checks for PQ and C-PQ, respectively, based on 1,000 simulations. Prediction-corrected simulated (shaded areas) and observed (circles and lines) PQ and C-PQ concentrations are presented over time (h; y axes). The thick red line connects the observed median values per bin. The dotted red lines connect the 5th and 95th percentiles of the observations. The light blue areas are the 95% confidence interval of the 5th and 95th percentiles, and the light red area indicates the confidence interval of the median.

FIG 4

Area under the plasma concentration-time curves for PQ (left panel) and C-PQ (right panel) by weight (x axes).

FIG 5

Effect of age and CYP2D6 activity score (AS) on PQ and C-PQ plasma concentrations over time after a 0.25-mg/kg single dose of PQ. Panels A and B, respectively, show the effects of age on PQ and C-PQ concentrations over time in two hypothetical children of different ages (2-year-old, 12-kg body weight; 14-year-old, 40-kg body weight), with a CYP2D6 AS = 1.0. Panels C and D, respectively, show the effects of genetically determined CYP2D6 AS in four hypothetical 14-year-old children on the PQ and C-PQ concentration over time. The values used to construct curves are model derived.