Twenty-year prospective cohort study of the association between a Japanese dietary pattern and incident dementia: the NILS-LSA project

PURPOSE: Evidence has suggested that adherence to a Japanese diet may be beneficial for health. However, its association with incident dementia remains unclear. The aim was to explore this association in older Japanese community-dwellers, taking apoprotein E genotype into consideration. METHODS: A 20-year follow-up cohort study involving 1504 dementia-free older Japanese community-dwellers (aged 65–82 years) living in Aichi Prefecture, Japan, was conducted. Based on a previous study, a 9-component-weighted Japanese Diet Index (wJDI9) score (range − 1 to 12) was calculated using 3-day dietary record data and used as an indicator of adherence to a Japanese diet. Incident dementia was confirmed by the Long-term Care Insurance System certificate, and dementia events occurring within the first 5 years of follow-up were excluded. A multivariate-adjusted Cox proportional hazards model was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for incident dementia, and Laplace regression was used to estimate percentile differences (PDs) and 95% CIs (expressed in months) in age at incident dementia (i.e., dementia-free duration differences), according to tertiles (T1–T3) of wJDI9 scores. RESULTS: The median (IQR) follow-up duration was 11.4 (7.8–15.1) years. During the follow-up period, 225 (15.0%) cases of incident dementia were identified. Because the smallest prevalence of incident dementia was 10.7% for the T3 group of wJDI9 scores, to avoid inaccurately estimating the dementia-free duration of participants in the T3 group, the 11th PDs in age at incident dementia between the T1 and T3 groups of wJDI9 scores were estimated. A higher wJDI9 score was associated with a lower risk of incident dementia and a longer dementia-free duration difference. The multivariate-adjusted HR (95% CI) and 11th PDs (95% CI) in age at incident dementia for participants in the T1 vs. T3 group were 1.00 (reference) vs. 0.58 (0.40, 0.86), and 0 (reference) vs. 36.7 (9.9, 63.4) months, respectively. Each 1-point increase of the wJDI9 score was associated with a 5% lower risk of incident dementia (P value = 0.033) and 3.9 (0.3, 7.6) additional months of dementia-free duration (P value = 0.035). No differences were seen in sex or smoking status (current smoker vs. non-current smoker) at baseline. CONCLUSION: These findings suggest that adherence to a Japanese diet defined by wJDI9 is associated with a lower risk of incident dementia in older Japanese community-dwellers, suggesting the benefit of the Japanese diet for dementia prevention. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s00394-023-03107-x.