Safety outcomes in advanced non-small-cell lung cancer patients treated with first-line platinum-based regimens in the United States

Abstract

We analyzed the treatment patterns and safety outcomes of the most common first-line platinum-based regimens initiated on or after non-small cell lung cancer (NSCLC) diagnosis in a real-world setting. Based on a United States oncology electronic medical record (EMR) database, patients treated with first-line platinum-based regimens after advanced NSCLC diagnosis from September 2008 to November 2014 were analyzed. Baseline characteristics and selected adverse events during treatment [incidence proportions and incidence rates (IRs)] were described by regimen. Propensity score stratification was used to adjust for baseline characteristics differences. Hazard ratios (HRs) were estimated using Cox proportional hazards model, with paclitaxel (Pac)/carboplatin (Carbo) as reference. Subgroup analysis was conducted for elderly patients (≥70 years old). The most common five regimens for the eligible patients were as follows: Pac/Carbo (n=3,009), pemetrexed (Pem)/Carbo (n=1,625), Pem/Carbo/bevacizumab (Bev) (n=735), Pac/Carbo/Bev (n=531), Pem/cisplatin (Cis) (n=357), and docetaxel (Doc)/Carbo (n=355). Highest IRs were reported for anemia, neutropenia, nausea, and vomiting across these regimens in patients of all ages. After propensity score stratification, compared with Pac/Carbo, risk of anemia was significantly lower with Pac/Carbo/Bev (HR =0.67), Pem/Cis (HR =0.68), and Pem/Carbo/Bev (HR =0.82); risk of neutropenia was comparable among all regimens except Doc/Carbo (significantly lower risk; HR =0.72); and risk of nausea (HR =1.45) and vomiting (HR =1.50) was significantly higher with Pem/Cis. Safety outcomes in elderly patients were consistent with the overall population. While EMR data have limitations, the real-world safety outcome with individual chemotherapy regimen could be considered for the better selection of platinum-based therapies in NSCLC.