There is growing interest in developing criteria that will allow efficient prospective discrimination between cancer patients at high and at low risk for complex fever and neutropenia. The objective of this study was to determine whether there were differences in patterns of documented infections and outcome of episodes of fever and neutropenia in pediatric patients with leukemia and those with solid tumors, a potential risk factor. A total of 283 febrile neutropenia episodes in pediatric cancer patients at a single center were retrospectively reviewed; 38% of the patients concerned had leukemia and 62% had solid tumors. Fever of unexplained origin was seen in 73% and 74% of episodes in patients with leukemia and solid tumor, respectively. Bacteremia occurred in 18% and 16% of patients in these respective groups. There was no difference in the type of microorganisms that were isolated in the groups, with gram positives predominating in both. The median duration of fever was 2 days in both groups. The depth of neutropenia was similar, with 75% of leukemia patients and 70% of solid tumor patients presenting with ANC of 100 cells/microl or lower. The median duration of neutropenia was 9 days in patients with leukemia and 6 days in solid tumor patients. The median duration of antibiotic treatment was 9 days and 7.5 days in the same respective groups. Antibiotic modification occurred in 25% episodes of febrile neutropenia in leukemia patients and in 11% of episodes in solid tumor patients. No deaths occurred in either group. Subgroup analysis of leukemic patients suggested that patients in the induction phase of therapy have a higher rate of bacteremia and pneumonia. No substantial difference in course or outcome was seen between the leukemia and solid tumor groups, possibly because of the intensive treatment administered to pediatric patients with solid tumors. Risk assessment strategies based on chemotherapy dose intensity and patient comorbidities rather than underlying malignancy should be prospectively studied.