BACKGROUND: Two antibiotic regimens are used commonly in Thailand for the initial treatment of severe melioidosis: ceftazidime in combination with trimethoprim-sulfamethoxazole (TMP-SMX) and ceftazidime monotherapy. It is not known whether TMP-SMX provides an additional benefit. METHODS: Two prospective, randomized trials that compared these regimens for patients presenting with acute severe melioidosis were started independently at tertiary care hospitals in Ubon Ratchathani and Khon Kaen (in northeastern Thailand), and the results were analyzed together as a prospective, individual-patient data meta-analysis. The primary end point was in-hospital mortality rate. RESULTS: The in-hospital mortality rate among all enrolled patients (n=449) was not significantly different between those randomized to ceftazidime alone (25.1%; 56 of 223 subjects) and those randomized to ceftazidime with TMP-SMX (26.6%; 60 of 226 subjects; odds ratio [OR], 1.08; 95% confidence interval [CI], 0.7-1.7; stratified P=.73). Of the 241 patients with culture-confirmed melioidosis, 51 (21.2%) died. Of these 241 patients, 31 (12.9%) died > or =48 h after the time of study entry. Among patients with melioidosis, there was no difference in death rate between the 2 treatment groups for either all deaths (OR, 0.88; 95% CI, 0.48-1.6; stratified P=.70) or for deaths that occurred > or =48 h after hospital admission (OR, 0.88; 95% CI, 0.41-1.9; stratified P=.73). Conditional logistic regression analysis revealed that bacteremia, respiratory failure, and renal failure were independently associated with death and treatment failure. Drug regimens were not associated with death or treatment failure in this model. CONCLUSION: We conclude that the addition of TMP-SMX to ceftazidime therapy during initial treatment of severe melioidosis does not reduce the acute mortality rate.