Country differences in past trends in alcohol-attributable mortality in Europe

Abstract

Background Both levels and trends in alcohol-attributable mortality have been heterogeneous across European countries, but detailed studies are lacking. We study country differences in past alcohol-attributable mortality trends in Europe by period and cohort, and assess whether differences become smaller (convergence) or bigger (divergence). Methods Age-standardized liver cirrhosis (from 1950 onward for 8 selected countries) and alcohol-attributable mortality (from 2000 onward for 20 European countries) were estimated using population level cause-of-death data. We assessed convergence by applying various dispersion measures and applied ageperiod- cohort analysis to examine country differences in cohort effects and cohort patterns. Results Levels and trends of alcohol-attributable mortality clearly differed across European countries and between sexes, especially for men. Western European countries had lower and more stable alcoholattributable mortality rates, whereas countries from Central and Eastern Europe showed higher rates and more dissimilar trends across the countries. Birth cohorts made significant contributions to liver cirrhosis mortality (P < 0.001), but the patterns were different between most of the country pairs (P < 0.001). Birth cohorts at higher risk of liver cirrhosis mortality were the older cohorts in Nordic countries, as compared to the younger cohorts in Eastern and Southern countries. Divergence in alcoholattributable mortality rates showed up until 2007 and convergence thereafter, for both men and women. Conclusions Clear differences between European countries in levels, time trends, and cohort patterns of alcohol-related mortality exist. Due to declines in alcohol-related mortality in Central and Eastern European countries, alcohol-related mortality is converging since 2007 in Europe. Key messages: Alcohol-attributable mortality in Europe is recently turning from divergence to convergence. Birth cohort effects are crucial to understand alcohol-related mortality trends.