Dietary Quality and the Association Between Tooth Loss and Cognitive Performance in Older Adults: A Causal Mediation Analysis of NHANES 2011–2014

OBJECTIVES: To quantify how much of the association between tooth loss and cognitive performance is statistically attributable to dietary quality. METHODS: Cross-sectional causal mediation analysis of 2600 adults aged 60-80 years (NHANES 2011-2014). Functional tooth loss (< 20 natural teeth) was the exposure, dietary quality (Healthy Eating Index 2020; HEI-2020) the mediator, and a composite cognitive z-score (processing speed, episodic memory, semantic fluency) the outcome. Survey-weighted models estimated the total effect; mediation quantities were estimated from unweighted models with 1000 bootstrap replicates. RESULTS: Severe tooth loss was associated with lower cognitive z-scores (survey-weighted β = -0.167 SD; 95% CI: -0.226, -0.107) and lower dietary quality (-2.8 HEI-2020 points; 95% CI: -3.7, -2.0). Dietary quality accounted for 8.9% of the association (95% CI: 3.8%-16.9%). The mediated effect was small in absolute terms (-0.012 SD, ≈0.2 Digit Symbol points) and fragile: it would be nullified by unmeasured mediator-outcome confounding only 1.5 times that implied by race/ethnicity, and was of borderline significance after periodontal adjustment in the dentate subsample (7.7%; 95% CI: 0.1%-32.6%; p = 0.047). CONCLUSIONS: Dietary quality accounted for a small share of the cross-sectional tooth loss-cognition association. Because exposure, mediator, and outcome were measured contemporaneously and the estimate was sensitive to plausible confounding, these findings are consistent with-but do not establish-a dietary mediating pathway. The robust association between tooth loss and poorer diet nonetheless supports nutritional screening in dental care of older adults.