Nonlinear associations between serum vitamin mixtures and cardiovascular disease risk: insights from a national cross-sectional analysis.

Although the association did not reach statistical significance in the fully adjusted model, the observed trend may be driven by reverse causality, increased supplement use, or other uncontrolled confounding, and should not be interpreted as evidence of a direct causal relationship.

also in 111,027 other papers

Pointwise tests along the z grid with FDR adjustment ( α = 0.05) showed that the proportions of the exposure range reaching statistical significance were: vitamin D 62% (minimum FDR p = 0.0052), vitamin E 54% (0.0036), vitamin B6 32% (1.1 × 10− 5 ), and vitamin C 24% (0.0020); vitamins A and B12 had no significant segments (both 0%; minimum FDR p = 0.255 and 0.884, respectively).The curve shapes indicated nonlinear effects of varying magnitude across vitamins: vitamins D and E displayed overall inverse trends; vitamin B6 showed a positive trend; vitamin C exhibited a weak-to-moderate inverse association primarily over part of its concentration range; and vitamins A and B12 showed no significant effects.

also in 11,005 other papers