Evaluating risk of hospitalization with G-CSF use in real-world oncology practice

Abstract

6626 Background: Limited published data exist on how granulocyte colony-stimulating factor (G-CSF) treatment patterns affect risk of neutropenia-related hospitalizations. This study examines prophylactic vs. delayed use of G-CSF and compares the effectiveness of prophylaxis with filgrastim (FIL) vs. pegfilgrastim (PEG). Methods: A retrospective analysis of administrative claims from U.S. commercial health plans identified adult patients with non-Hodgkin's lymphoma, breast, or lung cancer, treated with chemotherapy between July 2004 and January 2008. For these patients, the first course of chemotherapy and each unique cycle with use of G-CSF (FIL or PEG) was identified and designated ‘prophylaxis’ if used within the first 5 days of each cycle, or ‘delayed', if after day 5. The risk of neutropenia-related hospitalization was then evaluated on a cycle basis. A generalized estimating equation (GEE) was used to adjust for baseline demographics and clinical characteristics. Results: Among 5,571 patient-cycles identified: 87.4% were prophylactic and 12.6% were delayed G-CSF. PEG use was primarily prophylactic (93.6%) in contrast to use of FIL which was delayed in 62.5% of patient-cycles. The rate of neutropenic hospitalization was 1.2% for prophylaxis (n=59) and 3.7% for delayed G-CSF (n=26) (P < 0.001). Among prophylactic patient-cycles, the neutropenic hospitalization rate was lower with PEG than FIL (1.1% (n=51) vs. 3.5% (n=8), p = 0.001). Multivariate analyses using GEE model showed consistent trends (Table). Conclusions: Prophylactic G-CSF use was associated with lower neutropenic hospitalization risk than delayed use. Among prophylactic use in real-world oncology practice, PEG was associated with about a two-thirds lower risk of neutropenic hospitalization compared with FIL. [Table: see text] [Table: see text]