Background: Nutritional factors can abet or protect against systemic chronic inflammation, which plays an important role in the development and progression of dementia. We evaluated whether higher (i.e. pro-inflammatory) Dietary Inflammatory Index (DII) scores were associated with cognitive decline in the community-based Offspring Framingham Heart Study (FHS).
Method: 1,614 older adults (mean age 61 years [standard deviation (SD)], 9;55% women]) completed validated 126-item Food Frequency Questionnaires (FFQ), administered at FHS examination cycle 7 (1998-2001) and examination cycle 5 (1991-1995) and/or 6 (1995-1998). We created a DII score (based on the published method by Shivappa et al. 2014) for each available exam; a cumulative DII score was calculated by averaging exam-specific scores across the two or three exams. Cognitive testing was completed at call back sessions following examination cycles 7 and 8 (2005-2008). Exam 7 was considered as study baseline and participants were followed over a mean time of seven years (SD 1). We excluded participants with prevalent dementia at baseline and those with no cognitive testing data at either/or exams 7 and 8. We examined associations between the cumulative DII score and cognitive test scores over time using annualized change adjusting for age, sex, education (model 1) and additionally for exam 7 measures of body mass index, total energy intake, total cholesterol: high-density lipoprotein ratio, smoking and anti-cholesterol medication (model 2).
Results: Higher DII scores were significantly associated with (less) decline in performance on Similarities (verbal comprehension/reasoning) and Global cognition, following adjustments for model 1 covariates (Model 1:β and SE, 0.017, 0.008,p=0.03; 0.004,0.002,p=0.02, respectively). The effect sizes remained similar after additional adjustment for Model 2 covariates (0.018, 0.009,p=0.06; 0.004, 0.002,p=0.03, respectively). Additionally, we found that higher DII scores associated with accelerated decline in performance on Trail Making B-A (processing speed and executive function) (Model 2: -0.010, 0.004,p=0.03). We observed no relationships between higher DII scores and other neuropsychological tests. Further, stratified analyses revealed a linear relationship between higher DII scores and (less) decline in performance on Hooper visual organization among men, but not women (Model 2: 0.022, 0.010,p=0.02; -0.011, 0.009,p=0.23, respectively).
Conclusion: Higher DII scores were associated with (less) cognitive decline. We take our unexpected findings with caution as we previous have seen a relationship between higher DII scores and increased risk for dementia. To date, such studies have been very limited, most studies that found a relationship were cross-sectional and have used less sensitive testing. Future longitudinal studies with sensitive neuropsychological test measures are encouraged to elucidate whether a longitudinal relationship between higher DII scores and age-related cognitive decline exists.