A prospective randomized trial was conducted at Srinagarind and Khon Kaen hospitals. Ceftazidime (100 mg/kg of body weight per day) and co-trimoxazole (trimethoprim, 8 mg/kg/day; sulfamethoxazole, 40 mg/kg/day) therapy was compared with conventional therapy (chloramphenicol, 100 mg/kg/day; doxycycline, 4 mg/kg/day; trimethoprim, 8 mg/kg/day; sulfamethoxazole, 40 mg/kg/day) in the treatment of 64 patients with bacteriologically confirmed cases of severe melioidosis who were admitted during September 1986 to January 1989. Of 61 evaluable patients (3 were excluded because of severe drug allergies), 42 were septicemic, and 31 of these patients had the most severe form, disseminated septicemic melioidosis. Their cumulative mortalities on day 7 were compared. There were significantly lower overall mortalities from melioidosis, septicemic melioidosis, and disseminated septicemic melioidosis in the group receiving the new treatment compared with those in the group receiving the conventional treatment (47 versus 18.5% [P = 0.039], 57.7 versus 25% [P = 0.039], and 82.3 versus 30.7% [P = 0.006], respectively); but the differences could have been influenced by the greater severity of illness, e.g., shock at initial presentation, in the patients who received the conventional treatment. Among patients with disseminated septicemia and initial shock, there was no significant difference in mortality between the regimens. Both regimens effectively eradicated bacteria from the circulation within 24 h (97 versus 96%, respectively). We recommend ceftazidime and co-trimoxazole as the drugs of choice for treatment of severe melioidosis, especially in those patients with disseminated septicemia.