Aim: To test the hypothesis that alcohol consumption is inversely related to ischaemic heart disease (IHD) mortality at the population level. Most individual-level studies find a reduced risk of IHD with a moderate level of alcohol consumption, but it is as yet unknown whether this association also exists at the aggregate level.
Measurements: The study period was approximately 1950 to 1995; 14 EU countries and Norway were included. Time series analyses on different data were utilized, and age-standardized IHD mortality for men and women in the age groups 30-44, 45-59, 60-74 and 30-74 years was measured. The effects of alcohol (sales per capita) were controlled for a weighted lag of per capita sales of cigarettes.
Findings: There was a random distribution of insignificant negative and positive alcohol effect estimates. A slight indication of a cardioprotective effect of alcohol among 30- to 44-year-old women in high consumption countries could be observed (significant for Italy). Mean alcohol effect estimates were nearly exactly zero (absent alcohol effect) among men and weakly positive among women. Because changes in cigarette consumption were often significantly and positively related to subsequent changes in IHD mortality, poor validity in the IHD time series cannot explain the unsystematic findings. Including a 6-year weighted lag of alcohol consumption changed the weak positive effect among women to an absent alcohol effect. A brief analysis of abstinence rates indicated no particular relationship to IHD mortality.
Conclusion: The alleged cardioprotective alcohol effect is absent at the population level, and great caution should be taken concerning alcohol policies for cardioprotective purposes.