Predicting fluoropyrimidine cardiotoxicity using a focus beyond routine cardiovascular risk assessment.

Abstract

12030 Background: Fluoropyrimidines remain backbone of systemic therapy for colorectal cancer (CRC) but are associated with diverse cardiovascular (CV) toxicities ranging from hypertension to heart failure. Current CV risk stratification tools have uncertain predictive value in cancer populations. We aimed to identify clinically actionable predictors of CV complications during fluoropyrimidine-based therapy. Methods: Prospective single-center observational registry included 171 consecutive stage I-IV CRC patients initiating fluoropyrimidine-based chemotherapy between (median follow-up 6.5 months, IQR 3.0-8.0). All patients underwent structured baseline cardio-oncology assessment including CV risk stratification per ESC guidelines, 12-lead ECG, transthoracic echocardiography with global longitudinal strain measurement, and serum biomarkers (NT-proBNP, high-sensitivity troponin I). Primary endpoint was incidence of ≥1 prespecified CV complication. Missing data ( < 5% for laboratory parameters) were addressed via complete-case analysis. Statistical analysis included χ² tests, Mann-Whitney U tests, and multivariable logistic regression with backward elimination (SPSS v28.0, significance p < 0.05). Results: Overall CV complication incidence was 33%. Event spectrum: Blood pressure destabilization (9%), thromboembolism (pulmonary embolism 5%, deep vein thrombosis 3%), arrhythmias 4,3%, coronary ischemia (angina 2.3%), ischemic stroke 0.6%, left ventricular systolic dysfunction (asymptomatic EF decline 1.8%, symptomatic heart failure 0.6%), moderate hydropericardium (2.3%), and isolated biomarker elevation (2.3%). Univariate analysis identified significant associations with CV complications for: diabetes mellitus (DM) (26.8% vs 10.4%, p = 0.006, OR 3.2), obesity (41.1% vs 26.1%, p = 0.047, OR 2.0), age (median 66.3 vs 63.1 years, p = 0.041), and high/very-high baseline CV risk (37.5% vs 14.3%, p = 0.009, OR 3.6). In multivariable analysis adjusting for age, sex, obesity, hypertension, and baseline CV risk category, only DM retained independent predictive value (adjusted OR 3.14, 95% CI 1.35-7.28, p = 0.008). Baseline NT-proBNP elevation correlated with increased all-cause mortality (58.3% vs 22.0%, p = 0.010) despite absence of CV deaths (all 12 deaths from cancer progression). Conclusions: Diabetes mellitus is the strongest independent predictor of CV toxicity in CRC patients receiving fluoropyrimidine chemotherapy, associated with 3-fold increased risk. While baseline CV risk stratification showed univariate association, it lost significance in adjusted analysis, highlighting DM's primacy. Systematic cardio-oncology monitoring appears effective given zero CV mortality. These data support prioritizing diabetic patients for intensified CV surveillance during fluoropyrimidine therapy.