Bevacizumab in metastatic colorectal cancer in a real-life setting – toxicity profile, survival outcomes, and impact of tumor sidedness

Introduction
Targeting angiogenesis in metastatic colorectal cancer (mCRC) using bevacizumab is a standard of care. The addition of this targeted biological agent to first-line infusional fluoropyrimidine-based chemotherapy was associated with superior overall survival (OS) in several randomized and controlled studies for CRC patients in the metastatic setting. However, access to this therapy in countries with limited resources is challenging. In Morocco, bevacizumab was introduced for this indication after considerable efforts of the Ministry of Health and Lalla Salma Foundation to support cancer patients with a limited income. In this report, the real-world efficacy and safety of the combination of bevacizumab with chemotherapy in mCRC are reported based on a retrospective cohort in Eastern Morocco.

Material and methods
The archives of the medical records of 98 mCRC patients treated with first-line bevacizumab at the Hassan II Regional Cancer Center (Oujda, Morocco) were sampled from 1st January 2014 to 31st December 2019 and analyzed using descriptive statistics, Kaplan-Meier estimation, and a multivariable Cox regression model for a time-to-event study.

Results
The toxicity profile was dominated by grade I–II proteinuria (10%), bleeding events (10%), thromboembolic events (9%), grade I–III hypertension (3%), and other rare events such as delayed healing of the stoma, scar dehiscence, intestinal perforation, and heart failure deterioration. In terms of survival, median OS and progression-free survival in the whole cohort were 22 and 13 months respectively. Patients who benefited from a metastasectomy after bevacizumab treatment had 31 months of median OS as compared to 14 months in the matched cohort with non-resectable liver metastasis. Notably, we demonstrated that tumor sidedness is a predictive factor of OS [hazard ratio (HR) = 2.452; 95% CI: 1.434–4.191, p = 0.001]. Moreover, the median OS for patients who received between 10 and 20 or more than 20 bevacizumab administrations was 24 and 33 months respectively as compared to those who received less than 10 cures (17 months) (log rank p < 0.0001). These markedly improved outcomes were also confirmed in multivariate Cox regression. A highly significant association of bevacizumab use and OS was found after adjusting for covariates (HR = 0.518, 95% CI: 0.374–0.717; p < 0.0001).

Conclusions
The current study confirmed the important place of this therapeutic strategy in mCRC. Additional studies with prospective enrollment are awaited to validate these findings.