The public health importance of vitamin D: determining causal effects of intrauterine and lifetime exposure to vitamin D

This aim of this grant is to determine whether low levels of vitamin D during fetal development and in childhood increase the risk for high blood pressure, high cholesterol and high glucose and insulin (markers of diabetes) in early adulthood. Vitamin D is generated by exposure of the skin to sunlight. Recent public health campaigns to reduce skin exposure to sunlight have been successful in reducing the occurrence of skin cancer, but there is some concern that they have also caused low levels of vitamin D in many people. Pregnant women and children in particular are likely to have inadequate levels. Low levels of vitamin D could have detrimental effects. There is increasing evidence that individuals who have lower levels of vitamin D have higher risk factors for diabetes and heart disease, such as higher fasting blood sugar levels and higher cholesterol and blood pressure. We will examine the association of vitamin D (measured in blood samples taken from their mothers during pregnancy and themselves when aged 7-9 years) with measures of blood cholesterol, glucose and insulin and blood pressure in a large study of individuals who are now aged 17-18 years.

The success of public health campaigns to decrease UVB exposure has contributed to population level declines in vitamin D. A high proportion of pregnant women and children in developed countries have sub-optimal vitamin D levels. There is increasing evidence that lower levels of vitamin D are associated with increased risk of type 1 and 2 diabetes, insulin resistance, dyslipidaemia and cardiovascular disease. There is also some evidence that low levels of vitamin D during pregnancy could increase the risk of adverse metabolic and cardiovascular traits in the next generation. However, whether any of these associations are causal is unclear. Most studies to date have been cross-sectional, unable to control for potentially important confounders and of small sample size. There is a need to determine whether these associations are causal and also the magnitude of any causal associations. Randomised controlled trials provide the gold-standard for determining causality but are expensive and often have problems with generalisability. A better understanding of whether vitamin D is causally related to metabolic and vascular outcomes, and if so the magnitude of this causal effect, could be obtained by better designed epidemiological studies (prospective studies, with large numbers and adequate detail on important confounders), and by the use of genetic variants that are robustly associated with vitamin D levels as instrumental variables to provide causal inference. Such approaches within an established cohort would provide timely and resource efficient results. We propose to examine the association of intrauterine vitamin D (maternal pregnancy levels of 25(OH)D) and childhood vitamin D (own levels of 25(OH)D assessed at mean age 10 years) with fasting glucose, insulin and lipid levels and blood pressure at age 17 (N=7,000). Further, we propose to identify genetic variants that are robustly associated with variation in 25(OH)D (heritability 40-80%) in genome-wide association analyses of 8,200 individuals, with replication of any variants identified in the genome-wide approach in over 22,000 individuals. Identified variants will then be used in our main cohort and collaborating cohorts (numbers totalling 15,000 for continuous traits and 7,000 cases for pre-eclampsia) as instrumental variables to determine the unbiased and unconfounded association of life-time exposure to different levels of vitamin D with fasting glucose, insulin and lipid levels, blood pressure and pre-eclampsia. We will also combine maternal and offspring genotype to determine the unbiased and unconfounded association of variation in intrauterine vitamin D with later metabolic and cardiovascular risk factors.