Association of Impaired Cytochrome P450 2D6 Activity Genotype and Phenotype With Therapeutic Efficacy of Primaquine Treatment for Latent Plasmodium vivax Malaria

Key Points Question How is natural variation in cytochrome P450 2D6 activity associated with therapeutic efficacy of primaquine phosphate against latent Plasmodium vivax malaria? Findings In this nested case-control study of 57 patients who had participated in a clinical trial of primaquine for radical cure of acute P vivax malaria, exposure to low levels of cytochrome P450 2D6 activity determined by genotype or measured by dextromethorphan metabolism phenotype was associated with a significantly increased likelihood of relapse of malaria in the year after directly observed high-dose primaquine therapy. Meaning Impaired cytochrome P450 2D6 activity was significantly associated with high risk of therapeutic failure of primaquine, and this finding suggests cytochrome P450 2D6 involvement in producing a therapeutically active metabolite.


Primaquine Plasma Concentration.
Primaquine in plasma were quantified using HPLC-Photodiode Array Detector (Alliance e2695 HPLC system, PDA 2998 detector, Waters, USA) according to the method of Ward [e1], with slight modification. Detection wavelength was set at 263 nm. Mobile phase used consisted of paired ion chromatography (PIC) B5 solution 10%: water (HPLC grade) 45%: methanol 20%: acetonitrile 25%. The separation of primaquine in plasma were done on Phenomenex Gemini C18 3 m 4.6 x 100 mm analytical column and Phenomenex Gemini C18 4.6 x 30 mm guard column. Column temperature was set at 35 0 C. To a 200-µL plasma sample, 5 g/mL 8-/3-amino-1methylpropylamino/-6-methoxyquinoline diphosphate (internal standard) was added and the extraction was done using ethylacetate:hexane (1:9). The supernatant was transferred to another tube and evaporated under a stream of nitrogen gas until dry. Then 150 µL of mobile phase were added to dissolve the residue and 30 µL were injected automatically onto the HPLC system for analysis.
Data were analyzed using nonlinear mixed effect modeling method. Sparse plasma sample was drawn at eight time slots relative to dosing (pre-dose, hour 0-2, h 2-6, h 6-24, day 9-10, day-11, day-14, day-22). Modeling and simulation were performed using NONMEM v.7.4.1 (ICON Development Solutions) on Windows 8.1 with G95 Fortran Compiler. The results were processed using Pirana v.2.9.4. and R statistics v.0.99.903 [e2]. We investigated possible compartment dispositions, absorption models, inter-individual variability implemented as exponential on all parameters, and used additive error on log transformed drug concentration as residual variability. Body weight was tested as an allometric function on clearance and volume distribution parameters, and concomitant antimalarial to build the covariate model. Objective function values were used to compare two hierarchical models (p = 0.05) [e3].

Urinary Dextramethorphan/Dextrorphan Ratio.
Dextramethorphan (DXM) and dextrorphan (DX) were quantified using HPLC-fluorescence detection (Alliance e2695 HPLC system, with 2475 FLR Detector, both from Waters, USA) according to the method of Lam [e4], with slight modification. Emission filter was 280nm and excitation wavelength was 305nm. Mobile phase was 30% acetonitrile and 70% 0.01M heptane sulfonic in 0.01M KH 2 PO 4 . The separation of DXM and DX in urine was on X-Bridge Phenyl analytical column (150 mm x 4.6mm) with a particle size of 5-µm (Waters, USA). Column temperature was 40 0 C. Extraction of DXM and DX in urine was by the method of Daali [e5], with modification. To 500 µL urine, 75 µL of HCl 10% was added, followed by 50-μl aliquot of internal standard solution (codeine, 100 µg/mL). 2000 U of ßglucuronidase was then added and vortexed for 15 seconds, and incubated at 37 0 C for 18 hours. The hydrolyzed urine samples were then alkalinized with KOH 5N 100 µL and 500 µL Na 2 CO 3 1M was added and vortexed, followed by the addition of 4 mL hexane:ethylacetate (1:1). This solution was vigorously mixed for 15 minutes and then centrifuged (3000rpm; 10min). The upper organic layer was discarded and the lower aqueous phase kept in the dark of 20 minutes to ensure complete elimination of organic solvents. Forty microliters were injected into the HPLC instrument by an automated system.

eTable 1. Primaquine Pharmacokinetics in Study Population vs. Clinical Trial Population
Primaquine pharmacokinetic characteristics among clinical trial population (n=174) compared to subjects in the current study (n=57: 36 controls and 21 cases).

. Population Pharmacokinetic Areas Under the Curve (μg h/mL) for Cases vs. Controls
Population pharmacokinetic area under the curve for 14 days primaquine daily dosing of 0.5mg/kg among clinical trial population (n=174; 26 relapsers and 148 non-relapsers) and the current study population (n=57; 21 relapsing cases and 36 non-relapsing controls).

Subject
Case (