Chloroquine induces a severe generalized pruritus, in predisposed Black African patients, during treatment of malaria fever, and also in some Caucasian patients treated for rheumatological diseases. We have previously shown that chloroquine may release endogenous opioids and/or interact with micro-opiate receptors in rats, and that both histamine and malaria parasite blood density, contribute to the itching severity in malaria fever in humans. The aim of our present study was to assess and compare the antipruritic efficacy of the micro-opiate receptor antagonist, naltrexone, and the antihistamine, promethazine, in chloroquine treated patients with malaria fever. A double-blind, randomized, parallel group comparison of the chloroquine-induced pruritus intensity and time profile in patients with parasitologically proven malaria fever, who were pretreated with a single dose of either naltrexone 50 mg or promethazine 25 mg orally (six patients each). All patients had an established history of severe pruritus following chloroquine treatment of malaria fever. A self-assessed itching severity score was undertaken at 0, 6, 12, 24, 48 and 72 h after initial chloroquine dosing, and the areas under the pruritus-intensity time curve AUCP0-72 h was determined in each patient and correlated to the malaria parasite density in blood. Both naltrexone and promethazine subjectively reduced itching severity compared with prior historical experience. One patient on naltrexone and two on promethazine never experienced any itching. There was no statistically significant treatment effect, but a significant time effect (P = 0.001, F = 4.77 d.f. 5) by two-way repeated measures ANOVA. The AUCP for naltrexone was 82 +/- 25 units/h, and 57 +/- 34 units/h for promethazine [95% confidence interval for the difference being -73 to 123]. However, the malaria parasite density in the naltrexone group (740 +/- 178 microl(-1)) tended to be higher than in the promethazine group 314 +/- 69 microl(-1) (P = 0.056, 95% confidence interval for the difference being -15 to 866 microl(-1)). Correction of the AUCP for malaria parasite density (parasite pruritogenic index, AUCP. units/h/parasites/microl blood) tended to be lower with naltrexone 9.1 +/- 2.6 than with promethazine 12.2 +/- 7.0 There was a highly significant and positive correlation between the malaria parasite density and the AUCP0-72 h, on naltrexone (r2 = 0.78, P = 0.040) and promethazine (r2 = 0.93, P = 0.008). However, comparison of regressions revealed that the slope of the regression was significantly steeper with promethazine 0.48 than naltrexone 0.12 (P = 0.006, t = 4.2), with the intercepts showing a trend to a difference (P = 0.086). Naltrexone exerted an antipruritic action, at least to a similar extent to promethazine in patients with chloroquine-induced itching in malaria fever. However, the relationship between parasite density and resultant pruritus was significantly different between naltrexone and promethazine. Thus, micro-opiate receptors/and or endogenous opioids may contribute to chloroquine itching in malaria fever, in humans, in accord with animal experimental findings. Malaria parasite density in blood is a strong determinant of itching severity in patients predisposed to chloroquine-induced pruritus.