Risk assessment in cancer patients with fever and neutropenia: a prospective, two-center validation of a prediction rule.

Abstract

PURPOSE The study was undertaken to validate a clinical model for predicting the medical risk of cancer patients with fever and neutropenia. PATIENTS AND METHODS A consecutive sample of 444 cancer patients with fever and neutropenia (granulocyte count less than 500/microL) at two hospitals, a specialized cancer referral center and a university-affiliated general medical hospital, was studied to identify clinical characteristics in the first 24 hours that predict subsequent serious medical complications during the hospital stay. To control for bias, major risk factors and complications were subject to blinded review. RESULTS Serious medical complications occurred in 34% of patients with risk factors identified in a prior study, including prior inpatient status (group I), outpatients with a serious independent comorbidity (group II), or uncontrolled cancer (group III), compared with 5% of the remaining patients (group IV) (P less than .000001). Two of the complications in group IV patients were transient asymptomatic hypotension, and the remaining three complications occurred after at least 1 week of progressive medical deterioration. These risk groups were independently significant in stepwise logistic regression analysis. Multiple complications (17%) and death (10%) were common among patients in groups I through III but did not occur in group IV patients. CONCLUSIONS This risk assessment model accurately stratified the medical risk of these patients using only clinical information available on the first day of their course. Low-risk patients are an appropriate population in which to study less intensive treatment strategies.