Selective antimicrobial modulation as prophylaxis against infection during granulocytopenia: trimethoprim-sulfamethoxazole vs. nalidixic acid.

Sixty-two profoundly granulocytopenic patients with acute leukemia undergoing induction chemotherapy were prospectively randomized to receive either trimethoprim-sulfamethoxazole plus nystatin or nalidixic acid plus nystatin for prevention of infection. Patients given trimethoprim-sulfamethoxazole plus nystatin during initial remission induction experienced an increased duration (22.6 vs. 13.6 days) of profound granulocytopenia (less than 100 granulocytes/mm3; P = 0.007). Acquisition of gram-negative bacilli was more frequent among patients treated with nalidixic acid plus nystatin while filamentous fungi were acquired more frequently by patients receiving trimethoprim-sulfamethoxazole plus nystatin (P = 0.05). The median duration of on-study time prior to documentation of first infection was longer for patients receiving trimethoprim-sulfamethoxazole plus nystatin (17 days) than for those receiving nalidixic acid plus nystatin (eight days) (P = 0.0002). Three infection-related deaths occurred among patients receiving nalidixic acid; seven occurred among patients receiving trimethoprim-sulfamethoxazole, five of which were secondary to pneumonia due to Aspergillus flavus. Both of these methods of selective antimicrobial modulation have apparent advantages, but each has disadvantages serious enough to limit their routine use.