Duration of empiric antibiotic therapy in granulocytopenic patients with cancer

Philip A. Pizzo, Kay J. Robichaud, Fred A. Gill, Frank G. Witebsky, Arthur S. Levine, Albert B. Deisseroth, Daniel L. Glaubiger, James D. MacLowry, Ian T. Magrath, David G. Poplack, Richard M. Simon

Research output: Contribution to journalArticlepeer-review

183 Scopus citations


Early initiation of empiric antibiotic therapy in febrile cancer patients has become established practice, but the appropriate duration of antibiotic therapy when no infectious source can be identified is unknown. The complications of broad-spectrum antibiotics argue for brief treatment, but the risk of an inadequately treated infection in the granulocytopenic patient favors longer therapy. We prospectively studied 306 episodes of fever and granulocytopenia in 143 patients with leukemia or solid tumor (age one to 33 years) with respect to the duration of empiric antibiotic treatment. Eligible patients (fever > 38 °C three times/24 hours or > 38.5 °C once, plus polymorphonuclear leukocytes < 500/mm3) had an extensive diagnostic evaluation, including at least two preantibiotic blood cultures, and therapy was then started with a broad-spectrum antibiotic regimen- Keflin®, gentamicin and carbenicillin (KGC). Initial evaluation failed to identify an infectious etiology for the fever in 142 of 306 (46 per cent) episodes. Fifty-six of 142 (39 per cent) of these fevers of unknown origin were associated with persistent granulocytopenia for more than seven days; in 33 of these, defervescence occurred while the patients received KGC. After seven days of empiric KGC therapy, the 33 patients with fevers of unknown origin who had become afebrile with empiric antibiotics but whose polymorphonuclear leukocytes remained less than 500/mm3 were randomized to either continue or discontinue ( d c) to receive KGC. The patients who continued to receive KGC until their polymorphonuclear leukocytes were more than 500/mm3 had no infectious sequelae. However, in seven of 17 (41 per cent) of the patients randomized to d c KGC infectious sequelae developed (p = 0.007) within a median of two days of discontinuing KGC (two with fever which again responded to KGC therapy, and five with a documented infection [two ultimately fatal]). In none of the patients did a resistant microbial flora or superinfection develop. These data suggest that the patient with a fever of unknown origin who becomes afebrile during empiric antibiotic therapy may profit from continued therapy while granulocytopenia persists.

Original languageEnglish (US)
Pages (from-to)194-200
Number of pages7
JournalThe American Journal of Medicine
Issue number2
StatePublished - Aug 1979
Externally publishedYes

ASJC Scopus subject areas

  • General Medicine


Dive into the research topics of 'Duration of empiric antibiotic therapy in granulocytopenic patients with cancer'. Together they form a unique fingerprint.

Cite this