Association between caffeine intake and cognitive function in adults: data from National Health and Nutrition Examination Survey (NHANES)2013-2014

Background and Aim: There was considerable controversy about the relationship between caffeine consumption and cognitive function. In some studies, there was an interaction between caffeine and sex in relation to cognitive function. Thus, we aimed to evaluate the association between caffeine intake and cognitive function in a large, nationally representative sample of US adults. We also aimed to evaluate whether there is an interactive effect of sex. Methods: We used data from the 2013–2014 NHANES. To conduct a cross-sectional analysis, our research subjects were older adults aged 60 years or above (N = 1440). The individual s cognitive functions were evaluated using the Consortium to Establish a Registry for Alzheimer ’s disease (CERAD) Word List Learning Test, the CERAD Word List Recall Test, the Animal Fluency test, and the Digit Symbol Substitution Test (DSST). Higher scores represent better cognitive function for all tests. The collected 24-h dietary recall data in the survey cycle were used to estimate the food intakes in this study. The amount of the consumed caffeine was estimated from all caffeine-containing foods and beverages. Participants were categorized based on the quartiles of caffeine intake. SPSS version 18 was used for analysis. After preliminary analyses, in each dimension of cognitive, we calculated and used median value as cut-off point and assessed the association between each dimension (binary) and caffeine intake using logistic regression analysis in different models. Adjustment was done for potential covariates; age, sex, family PIR, education, marital status, history of disease, sleep disorders, thyroid problems, physical activity, social support, smoking, and some nutrients. P < 0.05 was considered statistically significant. In all analyses, we considered sampling design complexity by specifying primary sampling units (PSUs), strata, and weights using the complex samples module in SPSS. Results: In all dimensions of the cognitive function, just the highest quartile of caffeine intakes was associated with the cognitive function in the crude model. The trend also existed just in the crude model (P trend<0.05). After adjusting for potential confounders in further models, the association was marginally significant in CERAD Word List Recall Test, but was not significant in other dimensions of cognitive function. In subgroup analyses, a statistically significant interaction was noted between caffeine intake and gender in relation to the CERAD Word List Recall Test (p=0.02).