Risk factors for premature death in middle aged men

Abstract

The causes of premature death and the associated risk factors were analysed in a cohort of 7935 middle aged men participating in a preventive population programme in Malmö. They were screened when aged 46-48 and then followed up for 3½-8 years. Two hundred and eighteen died, of whom 181 (83%) underwent necropsy. Three major causes of death were established: cancer in 61 (28%), deaths related to consumption of alcohol in 55 (25%), and coronary heart disease in 50 (23%). Distinctly different patterns of risk factors were found to be associated with each of the three main causes of premature death. In death due to coronary heart disease smoking (p=0·0062), serum cholesterol concentration (p=0·00014), serum triglyceride concentration (p=0·00013), systolic blood pressure (p=0·000012), and diastolic blood pressure (p=0·0021) were the strongest single determinants but diastolic blood pressure ceased to be a predictive factor in a multivariate analysis whereas all the other variables could be combined in a highly predictive logistic model. In death related to consumption of alcohol equal or even stronger associations were found for serum γ glutamyltransferase activity (p<0·0001), points scored in a questionnaire screening for alcoholism (p<0·0001), and, inversely, serum cholesterol (p=0·0046) and serum creatinine (p<0·0001) concentrations both when applied independently and when combined in a logistic model. In death due to cancer significant associations were found for serum urate concentration (p=0·023) and, inversely, serum cholesterol concentration (p=0·056-0·031). Malignant diseases and diseases related to consumption of alcohol were at least as prominent as cardiovascular disorders in causing premature death in the cohort of men studied. All three types of conditions are potentially avoidable and seem to be associated with significant and distinctive patterns of risk factors. These patterns should be used, as blood pressure and serum lipid concentrations already are, to predict the risk of premature death and indicate preventive measures.