Background and objectives: The alcohol-attributable mortality is a positive function of two quantities: the relative risk, which has a biological specificity, and the proportion of exposed, which has a temporal and geographic specificity. Unfortunately, only little knowledge is available on these quantities.
Methods: To estimate alcohol-attributable mortality we approached the problem by estimating: i) the prevalence of drinkers to various amounts of alcohol in the Italian population from 1970 to 1993; ii) the dose-response relation between alcohol consumption and risk of each alcohol-related condition through a meta-analytic approach; iii) the alcohol-attributable risks from 1970 to 1993 by using the drinkers prevalence estimates and relative risks evaluated previously; iv) the proportion of deaths related to alcohol consumption in Italy for the same period.
Results: Decreasing averages of per capita alcohol consumption (-44%) were observed from 1970 to 1993. In the same period was observed a reduction of -80%, -51% e -15%, respectively for heavy drinkers'prevalence (>100 g/die), moderate drinkers'prevalence (more than 50 g/die) and for drinkers'prevalence (any consumption). The greatest alcohol-attributable risks were observed for hepatic cirrhosis and for upper digestive and respiratory tract cancers. Applying the alcohol-attributable fractions to all deaths of 1993 about 44000 (corresponding to the 8% of overall mortality) were attributable to alcohol and about 32000 deaths were attributable to moderate intake (< or =100 g/die).
Conclusions: The main suggestion from this study is that the best strategy in preventing alcohol-related problems should consider as target the whole population.