Research published in the European Journal of Epidemiology by Costanzo S, Di Castelnuovo de Gaetano G et al has sought to separate the effects of wine, beer or spirit drinking in relation to fatal and non-fatal cardiovascular events. The Italian authors carried out an updated meta-analysis on the relationship between wine, beer or spirit consumption and cardiovascular outcomes, using state-of-the-art statistical techniques.

From 16 studies, results confirmed a J-shaped relationship between wine intake and reduced vascular risk, with maximal protection - an average 31% (95% confidence interval (CI): 19%) was observed at 21 g/day of alcohol. Similarly, from 13 studies a J-shaped relationship was apparent for beer (maximal protection: 42% (95% CI: 19%) at 43 g/day of alcohol). From 12 studies reporting separate data on wine or beer consumption, two closely overlapping dose-response curves were obtained suggesting maximal protection of 33% at 25 g/day of alcohol approximately (2 drinks/day by US standards and 3 units for the UK) for vascular diseases. A statistically significant association between spirits intake and vascular disease was not found.

A major problem with all meta-analyses is the inability to control for variables that were not included in the original reports. While there were adequate data to adjust for most of the usual confounders, there was no way to evaluate effects of the pattern of drinking (frequency, binge drinking, etc.) on the cardiovascular outcomes.

Limited data were available about the association of spirits intake and cardiovascular risk. While the trend was for a decrease in such risk with increasing spirits consumption, there was not a statistically significant relation in the meta-analysis based on 10 independent relationships using random models.

The key result of this meta-analysis is the finding of a very similar inverse association between the consumption of beer and the consumption of wine in relation to cardiovascular outcomes. While a similar association was not seen for spirits consumption, the data presented do not permit the conclusion that the key effects on cardiovascular disease are primarily due to the polyphenols in beer and wine. Similarly, the results do not permit the conclusion that the effect on cardiovascular disease is due primarily to the alcohol in these beverages. The lack of a similar J-shaped association for spirits may have been due to different drinking patterns (e.g., more binge drinking among consumers of spirits), as the pattern of drinking was not included as a confounder in the analyses.