Excessive alcohol use and increased sudden cardiac death risk: findings from 28 years of follow-up in the Copenhagen city heart study
S Isozaki, P E Warming, T Skjelbred, J Mujkanovic, J Tfelt-HansenAbstract
Background
Excessive alcohol consumption is a well-established cardiovascular risk factor, yet its specific association with sudden cardiac death (SCD) remains insufficiently characterized.
Purpose
To examine whether high levels of alcohol intake are associated with increased risk of SCD using repeated exposure assessments and time-dependent analyses.
Methods
We analyzed 10,100 participants (mean age 60.8 years; 56% women) from the Copenhagen City Heart Study, followed from 1993 to 2021. Alcohol intake was assessed twice, approximately 10 years apart, and analyzed as both a continuous and categorical variable: None (0), Low (1–5), Moderate (6–10), High (11–16), and Very High (>16 units/week). Cause-specific Cox models with time-updated exposure were used, and missing covariates were imputed using multiple imputation. SCD events were adjudicated independently reviewing all death certificates by two physicians. Sensitivity analyses included baseline-only and complete-case models. Two-sided p-values <0.05 were considered statistically significant.
Results
During a median follow-up of 28.6 years, 897 SCD events occurred. Very high alcohol consumption was associated with significantly greater SCD risk compared with moderate intake (p = 0.032) (Figure 1). Time-dependent continuous analyses demonstrated a progressive increase in SCD risk with higher alcohol consumption (p < 0.001), with no significant deviation from linearity when compared with spline models (p = 0.316). Baseline-only and complete-case models tended to overestimate risk compared with time-updated and multiple imputation analysis.
Conclusion
Our findings demonstrate a dose–response relationship between alcohol consumption and SCD risk, with progressively higher intake associated with increasing risk. Baseline-only analyses appeared to overestimate risk at the highest intake levels, underscoring the importance of time-dependent models and multiple imputation to obtain more accurate risk estimates.Figure 1