Empiric use of vancomycin during prolonged treatment-induced granulocytopenia. Randomized, double-blind, placebo-controlled clinical trial in patients with acute leukemia.

Because gram-positive infections cause morbidity following intensive antileukemic chemotherapy, the effects of vancomycin versus placebo were evaluated in a randomized, double-blind, placebo-controlled trial in 60 adult patients with acute leukemia and first infectious fever during prolonged (mean of 32 days) granulocytopenia. Gram-positive sepsis was associated with first fever in 17 (28 percent) of the 60 patients. None of 31 patients randomly assigned to receive vancomycin demonstrated gram-positive infection, whereas 16 of 22 patients randomly assigned to receive placebo subsequently had gram-positive infection (seven had sepsis, and nine had local infections; p less than 0.005). All patients with gram-positive infection were then given vancomycin, and all showed prompt clinical responses. The predominant gram-positive organism causing infection was beta-lactam-resistant Staphylococcus epidermis (19 of 44 isolates). Patients randomly assigned to receive vancomycin had more rapid resolution of first infectious fever and fewer total febrile days during the granulocytopenic course than did patients randomly assigned to receive placebo. Although vancomycin had no effect on the presence or absence of documented fungal infection, patients treated with vancomycin received empiric amphotericin B for recurrent or persistent fever later (mean of 14 days after initial antibiotic coverage was begun) than did patients receiving placebo (mean of 9.9 days; p less than 0.005), and thus received fewer total days of empiric amphotericin B therapy (mean of 16.3 days) than did patients given placebo (mean of 24.6 days; p less than 0.01). These data demonstrate that empiric use of vancomycin reduces the morbidity of gram-positive infections following intensive antileukemic therapy and decreases the need for empiric use of toxic amphotericin B.