Abstract
<jats:p>Background & aims Food- and nutrient-based inflammatory scores may capture different dietary features. We compared the Food Inflammation Score for Individuals (FISI) and Dietary Inflammatory Index (DII) in relation to cognitive performance and incremental discrimination. Methods This cross-sectional analysis included 2,455 adults aged ≥60 years in NHANES 2011-2014 with four cognitive tests and two reliable 24-h recalls. Daily FISI summed [(food FII per 100 g) × grams/100] across foods; two days were averaged. DII used 33 parameters averaged across the same days. Survey-weighted linear and logistic models incorporated strata, primary sampling units, and WTDR2D/2. Discrimination used survey-weighted AUROCs and 500 stratified primary-sampling-unit bootstrap replicates. Results After full adjustment, a 1-SD higher DII was associated with a 0.091 lower global cognitive Z-score (β = -0.091; 95% CI, -0.143 to -0.040; P = 0.005), whereas FISI was not (β = -0.019; 95% CI, -0.064 to 0.026; P = 0.341). The DII quartile trend was significant (P = 0.007), but nominal domain associations did not survive false-discovery-rate correction. Neither score was associated with low cognitive performance. The base AUROC was 0.8242; adding FISI changed it by <0.0001 (95% bootstrap CI, -0.0002 to 0.0007), and adding DII changed it by 0.0005 (-0.0003 to 0.0021). Conclusions Higher DII, but not FISI, was associated with lower global cognitive performance. Neither score materially improved discrimination for low cognitive performance. The two scores should not be assumed equivalent or interchangeable.</jats:p>