Abstract
Colorectal cancer (CRC) remains a leading cause of cancer-related mortality, with alcohol consumption implicated in its etiology. However, alcohol's prognostic impact on CRC survival is unclear, and self-reported intake is limited by bias. This population-based cohort study evaluated blood DNA methylation-based alcohol scores as objective prognostic tools in 2,129 CRC patients from Germany's DACHS study (2003-2021; median follow-up: 10 years). Participants were recruited from 22 hospitals in Southwest Germany, including non-metastatic (n = 1757) and metastatic (n = 372) patients with complete methylation and alcohol data. All three assessed methylation scores (3-CpG, 450-CpG, 144-CpG) correlated with self-reported lifetime/recent alcohol intake (Spearman's r: 0.29-0.36; p < 0.0001), particularly recent consumption. In non-metastatic patients, self-reported alcohol consumption showed a J-shaped mortality risk, with elevated risks in heavy drinkers and abstainers. A similar dose-response pattern was observed for the 3-CpG methylation score, which showed consistent and robust associations with increased overall mortality (adjusted hazard ratio [aHR] per standard deviation increase: 1.18, 95% CI: 1.11-1.25), non-CRC-related mortality (1.22, 1.13-1.32), and CRC-specific mortality (1.12, 1.00-1.25). The 450-CpG score was associated with overall mortality (1.07, 1.00-1.15), non-CRC-related mortality (1.14, 1.05-1.23), and alcohol consumption-related mortality (1.59, 1.17-2.16). These findings highlight the potential utility of DNA methylation-based alcohol scores, especially the 3-CpG and the 450-CpG scores, as prognostic tools for CRC outcomes. Such biomarkers may provide a more objective measure of alcohol exposure and complement self-reported data in risk stratification and clinical decision-making, though further validation is warranted before clinical implementation.