The associations of dietary patterns with all-cause mortality and other lifestyle factors in the elderly : An age-specific prospective cohort study

Background & aims: The association between dietary pattern and mortality has been well studied in the general population; however, few studies have focused on the elderly. We aimed to examine the association of dietary pattern with subsequent overall mortality in elderly Japanese, and demonstrate the modifiable effect of lifestyle factors on this association. Methods: Totally 2949 Japanese community-dwelling residents aged 64 or 65 years were included in the NISSIN Project in 1996-2005. A validated food frequency questionnaire was adopted to collect dietary information and factor analysis was used to extract dietary patterns. Unadjusted and adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated through the Cox proportional hazard regression model. Results: Over 31,233 person-years, 253 persons died. Three different dietary patterns were identified: meat-fat, healthy, and dairy-bread pattern. Increased risk for all-cause mortality for meat-fat pattern was observed among those who never smoked (HR, 2.81; 95% Cl, 1.37-5.79); this association for dairy-bread pattern was observed among the never smokers (HR. 2.21; 95% CI, 1.20-4.06) and occasional drinkers (HR, 1.62; 95% CI, 1.09-2.39). For healthy pattern, decreased overall mortality risk was observed among never smokers (HR, 0.44; 95% CI, 0.24-0.80), occasional drinkers (HR, 0.63; 95% CI, 0.42-0.93), and those who walked >= 1 h/day (HR, 0.47; 95% CI, 0.28-0.77). Conclusions: We found that tobacco use, alcohol consumption, and daily walking duration could modify the associations of three patterns with overall mortality. Healthy eating along with other healthy lifestyle factor among elderly populations can decrease the overall mortality risk