Dietary provitamin A and non-provitamin A carotenoid in relation to cognitive function among middle-aged and older...

Dietary carotenoids have been associated with better cognitive function, yet evidence regarding the role of different subgroups of carotenoid is inconclusive. This study aimed to investigate the relation of specific carotenoid subgroups to cognitive function.

We conducted a prospective study among US adults aged 50 years and older based on the Health and Retirement Study from 2013 to 2020. Average daily intakes of dietary provitamin A carotenoid (α-carotene, β-carotene and β-cryptoxanthin), and non-provitamin A carotenoid (lutein-zeaxanthin and lycopene) were assessed by a 163-item semi-quantitative food frequency questionnaire (FFQ) in 2013. In 2016, inflammatory biomarkers of major serum cytokines were assayed. Cognitive function was measured using a composite test score of global cognitive function (range 0-27) every 2 years from 2014 to 2020. Beta coefficients (β) and the 95% confidence intervals (CI) were estimated by linear mixed-effects models to evaluate the associations of energy-adjusted provitamin A and non-provitamin A carotenoid with cognitive function. We also performed mediation analysis to assess the potential mediating role of those serum cytokines.

Among 6015 participants (mean age of 67.8 ± 9.8 years), the median daily intake of provitamin A carotenoid and non-provitamin A carotenoid were 3.2 mg/d (IQR, 2.0-5.2) and 6.0 mg/d (IQR, 4.3-8.6), respectively. We observed a significant association between non-provitamin A carotenoid intake and cognitive function. The multi-variable adjusted mean difference in cognitive function score was 0.288 (95% CI:0.071, 0.504) comparing top to bottom quintile (median intake 12.6 mg/d vs. 3.0 mg/d) of non-provitamin A carotenoid. The corresponding association of the non-provitamin A carotenoid with cognitive function was significantly mediated by serum level of sTNFR-1 (3.97%, 95% CI:0.70-12.00, P < 0.05). No significant associations with cognitive function were observed for provitamin A carotenoid (β = 0.033, 95% CI: -0.188, 0.255) in the overall participants, however the association became significant among individuals with higher intake level of retinol (Q4 vs. Q1: β = 0.408, 95% CI: 0.099, 0.717) (P-interaction > 0.05).

The study demonstrated heterogeneous associations of different carotenoid subgroups with cognition in U.S. adults. Further studies are warranted to confirm the study findings and explore the potential mechanisms.