Gastrointestinal perforation associated with bevacizumab use in metastatic colorectal cancer: results from a large treatment observational cohort study.

BACKGROUND: Bevacizumab prolongs overall and progression-free survival when added to fluorouracil-based chemotherapy in patients with metastatic colorectal cancer in randomised controlled trials (RCTs). However, gastrointestinal perforation (GIP) occurs in 1-2% of treated patients. We sought to describe the incidence, temporal pattern, outcomes and potential risk factors for GIP in a large, community-based observational cohort study of patients treated with bevacizumab. PATIENTS AND METHODS: Baseline patient and tumour characteristics, including potential GIP risk factors, were collected at study entry. Treatment, targeted adverse events, progression events and survival data were recorded every 3 months. Detailed clinical information was collected for all patients experiencing a GIP event. Effects of baseline risk factors on GIP risk were investigated using Cox proportional hazards regression. RESULTS: Of 1953 evaluable patients followed for a median of 20.1 months, 37 (1.9%) experienced GIP. Most GIP events were surgically managed with successful outcomes; four events were fatal. The majority of GIP events (26/37) occurred ≤6 months after starting bevacizumab (median, 3.35 months). Univariate and multivariate analyses showed that age ≥65 years was significantly associated with lower GIP risk. In multivariate analyses, intact primary tumour and prior adjuvant radiotherapy were significantly associated with increased risk of GIP within 6 months after starting bevacizumab. A regression analysis that assessed the risk of GIP over time showed no cumulative risk associated with bevacizumab exposure. CONCLUSION: The observed rate of GIP in this large, community-based experience was consistent with rates reported in RCTs. Most events were successfully managed with surgical intervention.