Risk of mortality in patients with cancer experiencing febrile neutropenia

Abstract

9561 Background: Febrile neutropenia (FN) is a serious and potentially life-threatening condition that may develop in patients with cancer treated with chemotherapy. The risk of mortality from FN is not well characterized in clinical practice. Methods: Patients with cancer receiving chemotherapy in clinical practice were identified from the HealthCore Integrated Research Database, a geographically diverse spectrum of fully adjudicated longitudinal claims data from 13 health plans with over 20 million US lives. Patient enrollment data, full medical care, prescription drug use and mortality (confirmed using the National Death Index) were examined for each eligible patient. Patients experiencing FN were compared to propensity score-matched patients not experiencing FN within tumor types of Non-Hodgkin lymphoma (NHL), breast, lung, colorectal and ovarian cancer. Study endpoints included overall mortality (anytime during follow-up) and early mortality (during a chemotherapy course). Logistic regression was used to calculate odds ratios (OR), adjusted for demographics, comorbidities, and other covariates. Results: Matched FN and control groups each contained 5,176 patients. Crude incidence rates of overall and early mortality were significantly higher for patients in FN group compared to controls for all tumor types [7.9/1000 person-months (PM) (95% confidence interval [CI] = 7.3, 8.5) vs. 5.6/1000 PM (CI = 5.1, 6.1), P < 0.0001; and 3.4/1000 PM (CI = 3.1, 3.9) vs. 2.4/1000 PM (CI = 2.1, 2.8), P = 0.0001, respectively]. Within tumor strata, lung cancer had the highest mortality rates and breast cancer had the lowest; NHL had the largest magnitude of difference between FN and controls (overall mortality: 7.2/1000 PM [CI = 5.3, 9.7] and 3.3/1000 PM [CI = 2.1, 5.1], P = 0.0035; early mortality: 3.0/1000 PM [CI = 1.8, 4.8] and 1.5/1000 PM [CI = 0.8, 3.0], P = 0.1103). Logistic regression demonstrated a significant increase in risk of overall and early mortality in patients with FN compared to controls (OR = 1.77 [1.53, 2.05] and OR = 1.73 [1.41, 2.11]) respectively. Conclusions: The crude and adjusted risk of mortality in patients experiencing FN is almost two-fold higher than comparably matched patients without FN. Prevention of FN is of high importance and should be carefully considered in clinical practice. [Table: see text]