Abstract / Summary
Abstract Background. Glucose-6-phosphate dehydrogenase deficiency (G6PDd), the most common human enzymopathy, limits the safe use of 8-aminoquinolines used for the radical cure of Plasmodium vivax malaria and for reducing P. falciparum transmission. Chocó, on the Colombian Pacific coast, carries one of the country's highest malaria burdens and ethnically diverse communities, yet the molecular epidemiology of G6PD variants remains poorly characterized and primaquine is administered without routine G6PD testing. Methods. We conducted a multi-site cross-sectional study in Chocó using two sampling strategies. A primary cohort consecutively enrolled individuals with fever or symptoms suggestive of malaria at diagnostic posts in Quibdó (2025–2026; n = 914 after excluding 4 participants with haemoglobin ≤ 7 g/dL). A complementary, enriched-for-deficiency cohort sampled patients with P. vivax or a previous G6PDd diagnosis at four sentinel sites in a 2023 pilot phase (n = 331). G6PD activity was measured with a quantitative point-of-care biosensor and classified phenotypically. Reduced-activity samples with sufficient DNA were genotyped by PCR–RFLP for the three variants most relevant to the Americas (A−, A + and Mediterranean). Results. In the primary cohort, reduced G6PD activity was found in 95 participants (10.4%; 95% CI 8.6–12.5), including 50 (5.5%) with deficient activity. Reduced activity was more frequent in women (13.0%) than men (7.6%), and higher in Afro-descendant (12.6%) than in mestizo (6.2%) or Indigenous (0.7%) participants. Of the 95 reduced-activity participants, 59 were genotyped; 36 could not be genotyped owing to insufficient DNA. Because genotyping was conditional on reduced activity, these proportions are not population allele frequencies. Among genotyped samples, A−-compatible genotypes predominated (40/59; 67.8%), followed by A+ alone (9/59; 15.3%); 10/59 (16.9%) were unexplained, and the Mediterranean variant was not detected. The complementary cohort (85 genotyped samples) is reported separately for exploratory characterization. Conclusions. Targeted three-variant screening left a substantial fraction of reduced-activity samples molecularly unexplained (16.9% in the primary cohort), underscoring the need for broader variant characterization. These findings support routine quantitative G6PD testing before radical-cure regimens carrying relevant haemolytic risk in this region, regardless of self-reported ethnicity, while recognizing the imprecision of estimates in the smaller Indigenous and mestizo strata.