Serum 25-hydroxyvitamin D, systemic immune-inflammation index, and atherogenic index of plasma in statin-untreated type 2 diabetes: a cross-sectional mediation analysis

Abstract

This cross-sectional study investigated the associations between serum 25-hydroxyvitamin D [25(OH)D] levels, systemic inflammatory burden, and atherogenic lipid profile in statin-untreated adults with type 2 diabetes, with particular emphasis on the contribution of the systemic immune-inflammation index (SII). A total of 619 statin-untreated patients with type 2 diabetes were included. Participants were stratified according to serum 25(OH)D concentrations into deficient (<20 ng/mL), insufficient (20–29.9 ng/mL), and sufficient (≥30 ng/mL) groups. Multivariable linear regression, BCa bootstrap mediation analysis with 5,000 resamples, and restricted cubic spline models were used. In fully adjusted models, each 10 ng/mL decrease in serum 25(OH)D was independently associated with higher SII (β=0.064, 95% CI 0.023–0.105; p  = 0.002) and higher atherogenic index of plasma (AIP; β=0.031, 95% CI 0.009–0.053; p  = 0.006). The indirect association through SII accounted for 26.5% of the total association between 25(OH)D and AIP (bootstrap 95% CI 12.8–43.1). Indirect associations were weaker for the neutrophil-to-lymphocyte ratio and non-significant for the monocyte-to-lymphocyte and platelet-to-lymphocyte ratios. Restricted cubic spline analyses demonstrated significant non-linearity for both SII and AIP, with points of maximal curvature at 18.7 ng/mL and 19.6 ng/mL, respectively. Estimates were similar in overlapping laboratory subgroups, but the indirect association was not significant in men (19.2%, 95% CI −7.6 to 39.4) or after exclusion of patients with severe deficiency (<10 ng/mL), although formal testing showed no significant interaction with sex. In a sensitivity analysis for unmeasured mediator–outcome confounding, the indirect association was reduced to zero at a residual correlation of only ρ =0.20, indicating limited robustness. The findings are compatible with a statistical indirect relationship involving systemic inflammation between vitamin D status and atherogenic dyslipidaemia, but the cross-sectional design precludes causal inference. The non-linear pattern near 20 ng/mL suggests that the inflammatory–atherogenic phenotype may become more pronounced below this threshold.