Abuse of alcohol in sudden out-of-hospital deaths in Finland.

Alcoholism is known to be greatly underdiagnosed in death certificates, a fact that biases in estimates of alcohol-related mortality. An autopsy series of 1658 cases (920 with natural cause of death and 738 nonnatural) was reviewed to evaluate the extent of this bias, and also to see how well different sources of information served as indicators of alcoholism when alcohol-related disease diagnosed at autopsy was considered as a gold standard. A stepwise logistic regression model adjusted by age and sex showed police reports of individual's alcohol usage and blood alcohol concentration (BAC) of > 2.9/1000 at autopsy to be the two most significant predictors of chronic alcohol abuse (p < 0.0001). The specificities of these two parameters as predictors of chronic alcohol abuse were generally high and sensitivities low. The authors see high BAC (> 2.9/1000), due to its high specificity, as particularly suggestive of chronic heavy drinking. However, it is wise to use these parameters only as an aid in decision-making, not as sole indicators of alcoholism. Deaths associated with chronic heavy drinking were frequent, 50.5% of the total series (male 56.4%, female 37.1%). For all but one age-group (male 45-64 years), however, death certificates mentioned alcohol-related diseases in less than half of these cases. Especially evident underdiagnosis was found for female and males 65 years and older. These results indicate that alcoholism is frequent in such a highly selected population as a series of forensic autopsies and suggest that estimates of prevalence of alcoholism based only on review of death certificates are to be considered with great caution.(ABSTRACT TRUNCATED AT 250 WORDS)