Sex-specific associations between Mediterranean Diet Score and long-term risk of mortality, dementia and other chronic diseases. 20-year evidence from a Swedish population study of 70-year old individuals

OBJECTIVES: To investigate the association between adherence to the Mediterranean Diet Score (MDS) and the risk of all-cause mortality, cardiovascular disease (CVD), cancer, and dementia morbidity over 10- and 20-year follow-up periods in a population-based cohort of older adults. DESIGN: A prospective population-based cohort study. SETTING: Gothenburg, Sweden. PARTICIPANTS: A total of 1298 participants (56.2% women) aged 70years at baseline from the general population and recruited from the Gothenburg H70 and Population Study of Women in Gothenburg (PSWG) birth cohorts. MEASUREMENTS: Dietary intake was assessed via face-to-face, semi-structured diet history interviews. The MDS was calculated based on eight food groups (excluding alcohol). Outcomes were identified through linkage with national health and death registers. Cox proportional hazard models were used, adjusting for BMI, smoking, marital status, physical activity, and education. RESULTS: A higher MDS was associated with a lower risk of 20-year dementia morbidity in men (multivariable-adjusted HR=0.83; 95% CI: 0.71-0.96), but not in women. This association did not remain statistically significant after correction for multiple testing, and the sex-by-MDS interaction (p=0.048) was nominally significant, but not robust to multiple testing. No statistically significant associations were found for CVD or cancer morbidity in either sex over the available follow-up (i.e., 10 and 20years, and up to 18years for cancer; 20-year CVD follow-up in men: HR=0.98; 95% CI: 0.88-1.08). For all-cause mortality, an initial association in men at 10years (HR=0.90; 95% CI: 0.82-0.99) was no longer significant after full adjustment (HR=0.92; 95% CI: 0.84-1.01). CONCLUSIONS: Higher MDS levels were associated with lower dementia risk in men, but this finding was not robust to multiple testing correction and requires confirmation in larger studies. No association was found with all-cause mortality, CVD and cancer morbidity. A limitation of these findings is that dietary assessment was conducted at a single time point (age 70), and that dementia ascertainment relied on registers which are estimated to miss > 50% of cases, particularly among home-dwellers.