Impact of fluoroquinolone prophylaxis on reduced infection-related mortality among patients with neutropenia and hematologic malignancies.

BACKGROUND Fluoroquinolone prophylaxis during neutropenia in patients with cancer has been associated with decreased incidence of gram-negative bacteremia. Bacterial antimicrobial resistance is likely to cause a progressive lack of efficacy of fluoroquinolones, but no convincing evidence from clinicoepidemiologic observations has proved this hypothesis. METHODS This prospective observational study assessed the impact of discontinuing fluoroquinolone prophylaxis on the incidences of fever and bacteremia and on mortality among patients with neutropenia, after chemotherapy for hematologic malignancies. RESULTS After a 12-month baseline period of levofloxacin prophylaxis, a period of discontinuation of fluoroquinolone prophylaxis was planned but was stopped prematurely after 9 neutropenic episodes over 3 weeks, because the mortality rate (33.3%) was higher than that with routine fluoroquinolone prophylaxis (2.9%) (odds ratio [OR], 16.6; 95% confidence interval [CI], 3.6-77.2). Fewer patients had gram-negative bacteremia during the baseline period (4.8%; n=15) than during the discontinuation period (44.4%; n=4) (OR, 16.9; 95% CI, 4.1-70.0). After levofloxacin therapy was reintroduced, the incidence of gram-negative bacteremia and the mortality rate were comparable to those during the first period. Escherichia coli isolated during the discontinuation period was susceptible to levofloxacin in vitro, whereas all E. coli isolates isolated during both prophylaxis periods were resistant. Bloodstream infections were caused by a single agent when the patient had received levofloxacin prophylaxis, whereas most cases of gram-negative bacteremia were polymicrobial after discontinuation. CONCLUSIONS These findings suggest that, despite increasing rates of antimicrobial resistance, levofloxacin prophylaxis during neutropenia may have a beneficial impact on morbidity and infection-related mortality. Continued monitoring of the rate of gram-negative bacteremia is warranted for timely detection of the loss of efficacy of fluoroquinolone prophylaxis.