Serum and urine metabolite correlates of vitamin D supplementation in the Atherosclerosis Risk in Communities (ARIC) study

Abstract

Background & aims

Vitamin D regulates calcium and phosphorus homeostasis, skeletal health, and potentially other aspects of health. There are limitations of existing vitamin D biomarkers. We aimed to discover novel vitamin D biomarkers by investigating serum and urine metabolites associated with vitamin D supplementation.

Methods

We examined cross-sectional associations between vitamin D supplementation and serum and urine metabolites in Atherosclerosis Risk in Communities Study participants at visit 5 (2011–2013). Untargeted metabolomic profiling of serum and spot urine samples was performed by Metabolon, Inc. We analyzed associations between vitamin D supplementation and log₂-transformed metabolites using linear regression models adjusted for demographic, lifestyle, and health covariates.

Results

Of 5225 participants with serum metabolites analyzed (mean age 76 [SD 5] years, 57 % female, 20 % Black), 45 % reported taking vitamin D supplements. Eighty-two of 933 serum metabolites were associated with vitamin D supplementation (P < 0.05/933). Most were lipids (n = 36). Of 1565 participants with urine metabolites analyzed, one-third (37 %) used vitamin D. Nineteen of 946 urine metabolites were associated with vitamin D supplementation (P < 0.05/946). Most were cofactors and vitamins (n = 12). After adjusting for other supplement use (multivitamin/mineral, omega-3, B and C vitamins), 5 serum metabolites (pro-hydroxy-pro, pyroglutamine, sulfate, creatine, and 2-hydroxypalmitate) and no urine metabolites were significantly associated with vitamin D supplementation.

Conclusions

Many serum and urine metabolites were associated with vitamin D supplementation. Five serum metabolites remained associated with vitamin D after adjustment for other dietary supplements, including metabolites of bone collagen degradation, glutathione metabolism, and sphingolipid metabolism. These metabolites may reflect physiological activities of vitamin D and, thus, improve assessment of vitamin D adequacy to achieve functional outcomes. These merit further investigation as potential vitamin D biomarkers.