Search

During a 2-year period, all children with cancer, neutropenia, and fever who were admitted to Hospital de Ninos Luis Calvo Mackenna (Santiago, Chile) were enrolled in a study of the safety of stopping antibiotic therapy on day 3 of treatment. Children who met predefined criteria for nonbacterial fever were randomized on day 3 to stop (group A) or continue (group B) antibiotic therapy. A total of 220 children with cancer had 238 episodes of fever and neutropenia; 68 children with 75 episodes met entry criteria for nonbacterial fever (group A, 36; group B, 39). Both groups were comparable in terms of age, gender, oncological disease, chemotherapy status, and initial neutrophil count. Resolution of symptoms occurred in 34 of 36 episodes in group A and 36 of 39 episodes in group B (P > .05). No deaths occurred, and bacterial superinfections were uncommon. For children with cancer as well as episodes of fever and neutropenia without an identifiable bacterial etiology at admission, stopping antibiotic therapy on day 3 was safe and not associated with a higher risk of bacterial superinfections.

The usefulness of determining serum levels of C-reactive protein (CRP) for the identification of bacterial infections in febrile neutropenic patients with cancer was evaluated. Two hundred children with cancer were monitored prospectively for the occurrence of neutropenia and fever; serum was collected from these children for determining baseline levels of CRP. Of these 200 children, 75 had 85 febrile neutropenic episodes; serum was collected daily from these 75 children for CRP analysis by nephelometry. Children were included into one of the three following groups by physicians blinded to results of CRP analysis: group I, demonstrated bacterial infection (24 episodes); group II, probable bacterial infection (31 episodes); and group III, viral infection or no infection (30 episodes). Baseline CRP values were low (mean, 9 mg/L; range, 0-35 mg/L) irrespective of tumor type or stage of therapy. Mean CRP values on day 1 for children in groups I and II (194 and 143 mg/L, respectively) were higher than those for children in group III (29 mg/L) (P.001). A CRP value of40 mg/L discriminated children with a demonstrated bacterial infection (sensitivity, 100%; specificity, 76.6%). Children with an unfavorable outcome had persistently high levels of serum CRP. For children with cancer, neutropenia, and fever, determination of the serum CRP level is useful for early diagnosis of bacterial infections and for monitoring the course of infection.