The Impact of Baseline Risk Factors on the Incidence of Febrile Neutropenia in Breast Cancer Patients Receiving Chemotherapy with Pegfilgrastim Prophylaxis: A Real-World Data Analysis

Abstract

**Background:** There are sparse data addressing whether standard risk factors for febrile neutropenia (FN) are relevant in patients receiving myelosuppressive chemotherapy and primary prophylaxis for FN, which would have implications for variables to consider during real-world comparative analyses of FN incidence. **Objective:** To assess the impact of baseline patient-specific risk factors and regimen risk on the incidence of FN in patients receiving pegfilgrastim primary prophylaxis. **Methods:** This was a retrospective observational study in patients with breast cancer (BC) who received myelosuppressive chemotherapy and prophylactic pegfilgrastim identified January 1, 2017-May 31, 2018 from MarketScan® research databases. The outcomes were defined as incidence of FN in the first cycle and among all cycles of chemotherapy using three different definitions for FN. Logistic regression and generalized estimating equations models were used to compare outcomes among patients with and without patient-specific risk factors and among those receiving regimens categorized as high-, intermediate-, or other-risk for FN (low-risk or undefinable by clinical practice guidelines). **Results:** A total of 4460 patients were identified. In the first cycle of therapy, patients receiving intermediate-risk regimens were at up to 2 times higher risk for FN across all definitions than those receiving high-risk regimens (_P_<0.01). The odds ratio for main FN among patients with ≥4 versus 0 risk factors was 15.8 (95% confidence interval [CI]: 1.5, 169.4; _P_<0.01). Patients with ≥3 FN risk factors had significantly greater risks for FN across all cycles of treatment than those with no risk factors; this was true for all FN definitions. **Discussion:** The choice of FN definition significantly changed the impact of risk factors on the FN outcomes in our study, demonstrating the importance of evaluating all proxies for true FN events in a database study. This is particularly important during real-world study planning where potential missteps may lead to bias or confounding effects that render a study meaningless. **Conclusions:** In patients with BC receiving chemotherapy with pegfilgrastim prophylaxis, patient-specific risk factors and regimen risk levels are determinants of FN risk. In real-world studies evaluating FN incidence, it is imperative to consider and control for these risk factors when conducting comparative analyses.