Diet quality and mortality: a population-based prospective study of men

Abstract

To study quality of diet in relation to all-cause mortality, cardiovascular disease (CVD) and cancer mortality. The population-based prospective Cohort of Swedish Men (COSM) included 40 837 men, 45-79 years of age, who filled in a FFQ (96 food items) and were CVD- and cancer-free at baseline. Quality of diet was assessed by Recommended Food Score (RFS) based on 36 items and Non-Recommended Food Score (Non-RFS) based on 16 items. Cox's proportional hazards regression models were used to estimate the hazard ratios (HRs) of mortality and 95% confidence intervals (CIs). Multivariate HRs for RFS and Non-RFS were adjusted for age, education, physical activity, martial status, self-perceived health status, smoking status, dietary supplements use, WHR, alcohol use, intake of energy and mutually adjusted. Between 1998 and 2005, 4501 deaths from all-causes were registered. Between 1998 and 2003, there were 1394 CVD and 759 cancer deaths. High RFS (> or =28) compared with low (< or =20) was associated with lower risk of all-cause mortality (HR: 0.81; 95% CI: 0.71-0.91; P-value for trend<0.0001) and CVD mortality (HR: 0.71; 95% CI: 0.54-0.93; P-value for trend=0.003). In contrast, men with high Non-RFS (> or =5) had higher risk of all-cause (HR: 1.21; 95% CI: 1.09-1.34; P-value for trend=0.001) and CVD mortality (HR: 1.27; 95% CI: 1.05-1.54; P-value for trend=0.07) compared to those with low Non-RFS (< or =2 items). No significant associations with cancer mortality were observed. Both measures of diet quality, RFS and Non-RFS, showed statistically significant associations with all-cause and CVD mortality (recommended foods inversely while nonrecommended foods positively), but not with cancer mortality.