Quantification of alcohol-related mortality in Sweden

The main aim of the present study was to estimate total alcohol-related mor tality in Sweden. For natural deaths, a metaanalysis carried out in Austral ia was updated to the end of March 1998, and pooled estimates of the relati ve risks were calculated for different diseases based on data from scientif ic studies that have been published in the international literature. The pr oportion of current alcohol drinkers from recent Swedish surveys, and the p ooled relative risk estimates were used to estimate disease-specific alcoho l-attributable fractions. Natural deaths 'caused' or 'prevented' by alcohol were estimated for the period 1992-1996. For unnatural deaths, all cases f rom 1992 through 1996 in Sweden were analysed (n = 23 132). Alcohol was reg arded to attribute to the death: if the deceased was a 'known alcoholic'; i f the underlying or contributing cause of death was alcohol-related; if the deceased had an alcohol-related inpatient diagnosis during a 3-year period prior to death; if the case tested positive for blood alcohol. Person year s of life lost/gained (<70 years) due to alcohol were also assessed. The as sumptions underlying the attributable risk methods used to analyse alcohol- related mortality due to natural causes need to be borne in mind when inter preting the results on natural deaths. Moreover, the preventive effect of a lcohol on coronary heart disease and stroke is still controversial. The fin dings of alcohol-related mortality due to unnatural causes were much more r eliable. About 3.5% of deaths were attributed to alcohol; alcohol involveme nt was more than twice as common in deaths of males (4.8%) than in those of females (2.0%). About one-quarter of the deaths in those aged below 50 yea rs were attributed to alcohol. In those (<less than or equal to>69 years, a lcohol had a net harmful effect in that it 'caused' more deaths than it 'pr evented'; 7% of deaths were in net 'caused'. More person years of life were in net lost than were saved by alcohol, suggesting that alcohol has a nega tive effect oil premature mortality. We conclude that alcohol accounted for about 3.5% of deaths in all ages, and 25% of deaths in those aged below 50 years, and about 10% of person years of life lost in Sweden. The balance o f harm and benefit does not weigh in favour of making a recommendation to t he public to drink in order to prevent death.