(140) Impact of Cardiovascular Comorbidities on Bevacizumab Effectiveness and Safety in Older Patients with Metastatic Colorectal Cancer

Publication Number: 814

Background: Older patients with cardiovascular (CV) comorbidities were under-represented in clinical trials evaluating bevacizumab in metastatic colorectal cancer (mCRC). Yet, CV comorbidities are not formal contraindications.

Objectives: To evaluate the impact of CV comorbidities on overall survival (OS) and CV safety in older mCRC patients treated with bevacizumab as a first-line therapy.

Methods: A 2009-2015 cohort of mCRC patients ≥65 years initiating bevacizumab as a first-line therapy was extracted from the French healthcare insurance system claims database (Système National des Données de Santé). At baseline, heart failure, hypertension, myocardial infarction, stroke, pulmonary embolism, and venous/arterial thrombosis have been identified through existing algorithms. 3-year OS has been described using the Kaplan-Meier method, and the impact of baseline CV comorbidities on 3-year OS evaluated using a time-dependent multivariable Cox proportional hazards model. 3-year cumulative incidence of the same CV events and the impact of baseline CV comorbidities on their 3-year occurrence have been evaluated using Fine & Gray models considering death as competing risk.

Results: 9222 patients were included: 56.4% male, median age 73 years [IQR 68-78], hypertension 63.5%, heart failure 3.6%, pulmonary embolism 1.5%, stroke 1.3%, myocardial infarction 1.3%, and venous/arterial thrombosis 8.4%. At 3 years, 71.4% had died: median OS (months) was respectively 20.4 [95%CI 19.9; 21.0] and 21.8 [21.1; 22.6]; p=0.003) in patients with and without CV comorbidities. Heart failure was the comorbidity that shortened the most the median OS (-4.3 months). In multivariable analyses, none of the CV comorbidities was associated with a higher risk of death unlike age ≥75, Activities of Daily Living (ADL)-dependency, radiotherapy and initiation of another targeted therapy during follow-up. At 3 years, CV events had occured respectively in 60.2% [58.9; 61.4] and 44.1% [42.3; 45.9] of patients with and without CV comorbidities. Venous thrombosis at baseline was associated with a higher risk of CV events as well as female gender, ≥3 CV medications at baseline, initiation of another targeted therapy and ≥7 bevacizumab injections during follow-up. When excluding hypertension from CV events, heart failure, and arterial thrombosis at baseline were also associated with a higher risk of CV events.

Conclusions: CV comorbidities before bevacizumab initiation in older mCRC patients had a clear impact on the occurrence of CV events but few on OS. If judged beneficial, older patients with CV comorbidities should be treated with bevacizumab under a close monitoring. Attention should be paid to CV comorbidities impacting the patients' autonomy, especially heart failure.