Because hypoglycaemia is common in severe malaria, intravenous glucose is often given empirically to patients on admission to hospital. To investigate the metabolic response to rapid glucose injection in acute malaria, 50 ml of 50% w/v (25 g) dextrose was given over 5 min to 10 adult patients (7 males, 3 females; mean age 30 years) with acute falciparum malaria. Five patients with severe infections were studied between doses of intravenous quinine; 5 cases were uncomplicated and previously untreated. The patients with severe malaria had lower pre-injection plasma glucose concentrations than patients with uncomplicated infections (mean +/- standard deviation, 4.2 +/- 0.9 vs 5.8 +/- 1.1 mmol/litre, 2P less than 0.015). However, peak glucose concentrations (18.6 +/- 4.8 vs 17.0 +/- 2.4 mmol/litre) and integrated responses (AUC0-245 min) were similar in the groups (2P greater than 0.1 in each case), and pre- and post-injection plasma insulin concentrations and AUC0-245 min values were also not significantly different (2P greater than 0.05 in each case). No 'rebound' hypoglycaemia was observed. The patients with severe malaria had higher peak plasma lactate concentrations than the uncomplicated patients (2.5 +/- 0.7 vs 1.5 +/- 0.9 mmol/litre, 2P less than 0.05), but the highest plasma lactate achieved and the greatest maximum post-injection rise were only 3.8 and 0.8 mmol/litre respectively. The average maximum reduction in plasma potassium after injection was 0.2 mmol/litre at 35 min. These data suggest that injections of hypertonic dextrose given empirically in conventional doses to non-acidotic patients with acute, severe malaria are not harmful, but the metabolic response in patients with an established acidosis remains unknown.