BackgroundRadical cure of Plasmodium vivax malaria with 8-aminoquinolines (primaquine or tafenoquine) is complicated by haemolysis in individuals with glucose-6-phosphate dehydrogenase (G6PD) deficiency. G6PD heterozygous females, because of individual variation in the pattern of X-chromosome inactivation (Lyonisation) in erythroid cells, may have low G6PD activity in the majority of their erythrocytes, yet are usually reported as G6PD “normal” by current phenotypic screening tests. Their haemolytic risk when treated with 8-aminoquinolines has not been well characterized.Methods and FindingsIn a cohort study nested within a randomised clinical trial that compared different treatment regimens for P. vivax malaria, patients with a normal standard NADPH fluorescent spot test result (≳30%–40% of normal G6PD activity) were randomised to receive 3 d of chloroquine or dihydroartemisinin-piperaquine in combination with primaquine, either the standard high dose of 0.5 mg base/kg/day for 14 d or a higher dose of 1 mg base/kg/d for 7 d. Patterns of haemolysis were compared between G6PD wild-type and G6PD heterozygous female participants. Between 21 February 2012 and 04 July 2014, 241 female participants were enrolled, of whom 34 were heterozygous for the G6PD Mahidol variant. Haemolysis was substantially greater and a larger proportion of participants reached the threshold of clinically significant haemolysis (fractional haematocrit reduction >25%) in G6PD heterozygotes taking the higher (7 d) primaquine dose (9/17 [53%]) compared with G6PD heterozygotes taking the standard high (14 d) dose (2/16 [13%]; p = 0.022). In heterozygotes, the mean fractional haematocrit reductions were correspondingly greater with the higher primaquine dose (7-d regimen): −20.4% (95% CI −26.0% to −14.8%) (nadir on day 5) compared with the standard high (14 d) dose: −13.1% (95% CI −17.6% to −8.6%) (nadir day 6). Two heterozygotes taking the higher (7 d) primaquine dose required blood transfusion. In wild-type participants, mean haematocrit reductions were clinically insignificant and similar with both doses: −5.8 (95% CI −7.2% to −4.4%) (nadir day 3) compared with −5.5% (95% CI −7.4% to −3.7%) (nadir day 4), respectively. Limitations to this nested cohort study are that the primary objective of the trial was designed to measure efficacy and not haemolysis in relation to G6PD genotype and that the heterozygote groups were small.ConclusionHigher daily doses of primaquine have the potential to cause clinically significant haemolysis in G6PD heterozygous females who are reported as phenotypically normal with current point of care tests.Trial RegistrationClinicalTrials.gov NCT01640574.

Mutations in the glucose-6-phosphate dehydrogenase (G6PD) gene result in red blood cells with increased susceptibility to oxidative damage. Significant haemolysis can be caused by primaquine and other 8-aminoquinoline antimalarials used for the radical treatment of Plasmodium vivax malaria. The distribution and phenotypes of mutations causing G6PD deficiency in the male population of migrants and refugees in a malaria endemic region on the Thailand-Myanmar border were characterized. Blood samples for G6PD fluorescent spot test (FST), G6PD genotyping, and malaria testing were taken from 504 unrelated males of Karen and Burman ethnicities presenting to the outpatient clinics. The overall frequency of G6PD deficiency by the FST was 13.7%. Among the deficient subjects, almost 90% had the Mahidol variant (487G>A) genotype. The remaining subjects had Chinese-4 (392G>T), Viangchan (871G>A), Açores (595A>G), Seattle (844G>C) and Mediterranean (563C>T) variants. Quantification of G6PD activity was performed using a modification of the standard spectrophotometric assay on a subset of 24 samples with Mahidol, Viangchan, Seattle and Chinese-4 mutations; all samples showed a residual enzymatic activity below 10% of normal and were diagnosed correctly by the FST. Further studies are needed to characterise the haemolytic risk of using 8-aminoquinolines in patients with these genotypes.