RESUMO
PURPOSE: The clinical index of stable febrile neutropenia (CISNE) can contribute to patient safety without increasing the complexity of decision-making. However, febrile neutropenia (FN) is a diverse syndrome. The aim of this analysis is to assess the performance of CISNE according to the type of tumor and infection and to characterize these patients. METHODS: We prospectively recruited 1383 FN episodes in situations of apparent clinical stability. Bonferroni-adjusted z tests of proportions were used to assess the association between the infections suspected at the time of onset and the type of tumor with the risk of serious complications and mortality. The performance of CISNE was appraised in each category using the Breslow-Day test for homogeneity of odds ratios and Forest Plots. RESULTS: 171 patients had a serious complication (12.3 %, 95 % confidence interval 10.7-14.2 %). The most common initial assumptive diagnoses were: fever without focus (34.5 %), upper respiratory infection (14.9 %), enteritis (12.7 %), stomatitis (11.8 %), and acute bronchitis (10.7 %). Lung and breast were the most common tumors, accounting for approximately 56 % of the series. The distribution of complications, mortality, and bacteremia varies for each of these categories. However, Breslow-Day tests indicate homogeneity of the odds ratio of the dichotomized CISNE score to predict complications in all infection and tumor subtypes. CONCLUSION: Despite FN's clinical and microbiological heterogeneity, the CISNE score was seen to be consistent and robust in spite of these variations. Hence, it appears to be a safe tool in seemingly stable FN.