Fever in the Pediatric and Young Adult Patient with Cancer

Abstract

Cytotoxic chemotherapy has improved the survival of cancer patients but has also increased the risk for serious infectious complications. To define a rational framework for antibiotic therapy and prophylaxis, prospective evaluations were made of 1001 febrile episodes in 324 patients during a 5 yr study period. The underlying malignancies included leukemia (40.5%), lymphoma (17%) and solid tumors (42.5%) in patients ranging in age from 1-49 yr (mean, 15.7). Fever occurred in 39.8% of patients receiving cytotoxic chemotherapy. Once febrile (F+) (defined as an oral temperature > 38.degree. C 3 times over a 24-h period or as a single temperature spike > 38.5.degree. C), patients were evaluated and then stratified according to whether they were granulocytopenic (G+) (defined as < 500 polymorphonuclear leukocytes mm3). Of the F+ episodes, 20.8% occurred when patients were nongranulocytopenic, of which only 17.3% were associated with infection. At least 75% of the F+G+ episodes were due to a defined or occult infection. In 51.8% of these, an infectious etiology could be diagnosed initially or within a 7-day follow-up evaluation period. Infections of the respiratory tract or blood stream accounted for 62.5% of all documented infections, and 85% of all microbiologically-defined infections were due to bacteria. There was no difference in the types of infections or infectious organisms in patients with leukemia, lymphoma or solid tumors, suggesting that once granulocytopenic, these patients share a comparable risk for serious infections. An infectious etiology could not be initially defined in 48.2% of the 793 F+G+ episodes. Of these, 55.5% were associated with granulocyte recovery within 1 wk and none of these low-risk fever of unknown origin (FUO) episodes were associated with subsequent infectious complications. However, infectious complications occurred in half of the high-risk FUO episodes (i.e., G+ > 7 days) when antibiotics were discontinued after 7 days of empiric therapy, suggesting that undetected infections are important in patients with protracted granulocytopenia. To delineate features which might differentiate patients with serious infections (e.g., bacteremias) from those without (e.g., low-risk FUO), historical and clinical features of 140 such patients were comprehensively compared; no one feature was consistently and reliably predictive of infection. Hence, for the present, regardless of their underlying malignancy or clinical presentation, initiation of empiric antibiotics in any G+ patient who becomes febrile continues to be necessary.