LAUSR

AIMS We investigated the cross-sectional associations of Plasmodium infection (PI) with fasting glucose (FG) and glycated hemoglobin (HbA1c) in malaria-endemic south-central Côte d'Ivoire. METHODS We studied 979 participants (non-pregnant; no treated diabetes; 51% males; 18-87 years) of the Côte d'Ivoire Dual Burden of Disease study. Fasting venous blood was obtained for PI, FG, and HbA1c assessment. We defined PI as a positive malaria rapid diagnostic test (RDT) or microscopic identification of Plasmodium species. We applied multivariable linear regressions to assess beta coefficients (β) and 95% confidence intervals (CIs) of PI positivity for FG and HbA1c independent of diabetes risk factors. RESULTS Prevalence of PI was 10.1% (5.5% microscopy; 9.7% RDT) without clinical fever. Prevalence of FG-based prediabetes (45.8%) and diabetes (3.6%) were considerably higher than HbA1c-based values (2.7% and 0.7%, respectively). PI was independently associated with FG among participants with higher body temperature (β 0.34, 95% CI 0.06-0.63, pheterogeneity=0.028), or family history of diabetes (β 0.88, 95% CI 0.28-1.47, pheterogeneity=0.009). Similar patterns observed with HbA1c were obliterated on accounting for FG. We also observed consistent associations with parasite density. CONCLUSIONS FG-based diabetes diagnosis in the presence of asymptomatic PI may misclassify or overestimate diabetes burden in malaria-endemic settings. Longitudinal studies are needed to confirm these findings and determine the risk for diabetes.