Children with severe malaria often present with lactic acidosis and hypoglycemia. Although both complications independently predict mortality, mechanisms underlying their development are poorly understood. To study these metabolic derangements we sequentially allocated 21 children with falciparum malaria and capillary lactate concentrations of 5 mmol/L or more to receive either quinine or artesunate as antimalarial therapy, and dichloroacetate or saline placebo for lactic acidosis. We then administered a primed infusion (90 min) of L-[3-13C1]sodium lactate and D-[6,6-D2]glucose to determine the kinetics of these substrates. The mean (SD) glucose disposal rate in all patients was 56 (16) micromol/kg x min, and the geometric mean (range) lactate disposal rate was 100 (66-177) micromol/kg x min. Glucose and lactate disposal rates were positively correlated (r = 0.62; P = 0.005). Artesunate was associated with faster parasite clearance, lower insulin/glucose ratios, and higher glucose disposal rates than quinine. Lactate disposal was positively correlated with plasma lactate concentrations (r = 0.66; P = 0.002) and time to recovery from coma (r = 0.82; P < 0.001; n = 15). Basal lactate disposal rates increased with dichloroacetate treatment. Elevated glucose turnover in severe malaria mainly results from enhanced anaerobic glycolysis. Quinine differs from artesunate in its effects on glucose kinetics. Increased lactate production is the most important determinant of lactic acidosis.