Determining the mediation role of CRP and IL-6 for associations between carotenoids, vitamin A, and vitamin E and insulin resistance in elderly

Background: This study aimed to investigate whether C-reactive protein (CRP) and interleukin-6 (IL-6) mediate the relationship between various antioxidants—specifically vitamin A, vitamin E isoforms, and carotenoids—and insulin resistance (IR). Methods: We conducted a bias-corrected bootstrapping cross-sectional mediation analysis using biomarker data from the Midlife in the United States (MIDUS3) study (2017–2022). Antioxidants evaluated included total lycopene, β-cryptoxanthin, retinol, and α- and γ-tocopherol. IR was assessed using the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR). CRP and IL-6 were included as mediators. Analyses were adjusted for age, sex, waist–hip ratio (WHR), smoking status, physical activity, dyslipidemia (assessed via total to high-density lipoprotein [HDL] cholesterol ratio), and dietary quality, with all continuous covariates mean-centered. Results: The analytic sample comprised 747 participants (56% female; mean age, 66 years; standard deviation [SD] = 9.6). After adjustment, higher total lycopene and β-cryptoxanthin levels were significantly associated with lower HOMA-IR, with no evidence of mediation by CRP or IL-6. In contrast, γ-tocopherol was associated with higher HOMA-IR, an effect partially mediated by elevated CRP (percent change per 10-unit increase: 7.79%; 95% confidence interval [CI]: 2.12–18.53). Retinol showed significant indirect mediating associations through reduced CRP (−31.48%; 95% CI: −50.24 to −13.50) and IL-6 (−15.46%; 95% CI: −33.57 to −1.49), though these effects did not yield a significant total association. However, α-tocopherol did not demonstrate a significant indirect or total association. Conclusions: Among older adults, γ-tocopherol was positively associated with IR, partially mediated by CRP, whereas β-cryptoxanthin and lycopene were inversely associated with HOMA-IR without a significant mediation for inflammatory markers. Although retinol reduced IR via reduced inflammatory markers, its overall association with IR was not significant.