Investigating the causal effect of vitamin D on serum adiponectin using a Mendelian randomization approach.

ABSTRACT

BACKGROUND/OBJECTIVES: The aim was to examine the causal effect of vitamin D on serum adiponectin using a multiple instrument Mendelian randomization approach. SUBJECTS/METHODS: Serum 25-hydroxy vitamin D (25(OH)D) and serum total or high molecular weight (HMW) adiponectin were measured in two Danish population-based studies the Inter99 study (6405 adults, 30-60 years) conducted in 1999-2001, and the MONICA10 study (2656 adults, 41-71 years) conducted in 1993-1994. RESULTS: In the Inter99 study, serum 25(OH)D was positively associated with total adiponectin (the effect estimate in % per doubling of 25(OH)D was 4.78, 95% CI 1.96, 7.68, P<0.001). Using variations in the vitamin D-binding protein gene and the filaggrin gene as instrumental variables, the causal effect in % was estimated to 61.46, 95% CI 17.51, 120.28, P=0.003 higher adiponectin per doubling of 25(OH)D. In the MONICA10 cohort, no significant association was observed between the serum concentrations of 25(OH)D and HMW adiponectin (the effect estimate in % per doubling of 25(OH)D was -1.51, 95% CI -5.80, 2.98, P=0.50), although the instrumental variables analysis to some extent supported a positive causal association (the effect estimate in % per doubling of 25(OH)D was 37.13, 95% CI -3.67, 95.20, P=0.080). CONCLUSIONS: The results indicate a possible causal association between serum 25(OH)D and total adiponectin. However, the association was not replicated for HMW adiponectin. Thus, further studies are needed to confirm a causal relationship.