Estimated numbers of deaths from coronary heart disease "caused" and "prevented" by alcohol: an example from Finland

ENGLISH :
OBJECTIVE: Epidemiological studies show that moderate alcohol consumption rather than abstention is associated with a lower risk of coronary heart disease (CHD) mortality. Our objective was to adjust established methods for calculating attributable fractions to a situation where the risk function is J-shaped and to estimate the number of CHD deaths "caused" and "prevented" by alcohol in Finland. METHOD: Point estimates of relative risk were obtained by a meta-analysis. They were pooled by fitting a nonparametric cubic smoothing spline to the data. Alcohol consumption distribution was estimated from survey data (N = 4,818; 2,488 women). The consequences of various assumptions about changes in alcohol consumption distribution on CHD mortality were estimated. The most detailed analyses are presented for men aged 30-69. The results for the men and women aged 30-79 are summarized. RESULTS: Among men aged 30-69, the beneficial effects of light to moderate alcohol consumption "prevent" some 400 CHD deaths each year which corresponds to 12-14% of the observed CHD deaths. Around 20 CHD deaths are "caused" by alcohol consumption exceeding the estimated optimum level. Among men aged 70-79 and women aged 30-79, the numbers of CHD deaths "prevented" by alcohol consumption were approximately 200 and 100, respectively, whereas there were only a few CHD deaths "caused" by alcohol. CONCLUSIONS: Our best estimates suggest that approximately one-tenth of the observed number of CHD deaths among middle-aged men in Finland is "prevented" by alcohol, while the relative effect is considerably smaller among older men and all women. (Review' s abstract)