Risk of chemotherapy-induced febrile neutropenia with early discontinuation of pegfilgrastim prophylaxis based on real-world data from 2010 to 2015.

OBJECTIVE: Evidence suggests that not all cancer chemotherapy patients who receive first-cycle pegfilgrastim prophylaxis (PP) continue to receive it in later cycles, and that these patients may be subsequently at higher risk of febrile neutropenia (FN). Available evidence, however, may not be reflective of current clinical practice. We undertook an evaluation to estimate the odds of FN, beginning with second chemotherapy cycle, among patients who received PP in that cycle and all previous cycles versus those who received PP in all previous cycles only, using recent real-world data.

METHODS: A matched-cohort design and data from two US healthcare claims repositories (2010-2015) were employed. The source population comprised cancer patients who received intermediate/high-risk chemotherapy and first-cycle PP. From the source population, beginning with the second cycle, all patients who received PP in all previous cycles were identified. From this subset, patients who did not receive PP in the cycle of interest ("comparison patients") were matched to those who received PP in that cycle ("PP patients"); the same process was repeated for subsequent cycles. Odds ratios (ORs) for FN (broad and narrow definitions) were estimated using generalized estimating equations.

RESULTS: Among 47,254 patients in the source population, 9% did not receive second-cycle PP and were matched to those who did. FN odds in cycle 2 were significantly higher among comparison patients versus PP patients (OR [broad definition]: 1.7, p < .001); OR [narrow definition]: 4.3, p < .001). Results for subsequent cycles and for the last cycle, respectively, were comparable (OR [range, broad definition]: 1.6 to 3.1, p < .001 for all; OR [range, narrow definition]: 2.7 to 11.8, p < .001 for all).

CONCLUSIONS: In this real-world evaluation of cancer chemotherapy patients who received first-cycle PP, FN risk was substantially higher among patients who did not receive PP in subsequent cycles versus those who continued PP.