Skip to navigation Skip to main content Skip to footer

Approved research

Vitamin D pathway and major disease outcomes: Observational associations and causal inference by Mendelian randomisation.

Principal Investigator: Dr Stephen Kaptoge
Approved Research ID: 11833
Approval date: May 1st 2016

Lay summary

Lower circulating levels of 25-hydroxyvitamin D [25(OH)D], the principal clinical biomarker of vitamin D status, have been associated with increased risk of many disease outcomes, including: skeletal, cardiovascular, cancers, and mortality. However, causality remains uncertain, and several features of the observational associations have not been well characterised. We aim to conduct powerful analyses to assess the relevance of vitamin D to major disease outcomes by assessing its: (i) correlates, (ii) shape of dose-response relationship with outcomes (primarily CHD, stroke, fractures, and cause-specific mortality), (iii) precise magnitude of associations, (iv) subgroup differences, and (v) causal relevance. This research is in the public interest considering the ready availability of over-the-counter vitamin D supplements, controversies surrounding sunlight exposure, and variable guidelines on intake, meaning precise knowledge of the nature of potential benefits (or harms) should have key implications on public health guidelines. Findings should also inform the design of randomised trials to test further hypothesis. We will initially assess correlates of vitamin D supplement use and its association with outcomes. Subsequently, when 25(OH)D biomarker levels and genetic data become available, we will assess: (i) The physical, lifestyle, and genetic correlates of 25(OH)D. (ii) The shape of dose-response relationships of 25(OH)D and outcomes (primarily CHD, stroke, fractures, and cause-specific mortality). (iii) The precise magnitude of the preceding associations with consistent adjustment for confounders and correction for regression dilution. (iv) Potential differences in magnitude of associations according to key characteristics (e.g. age, sex, ethnicity, season, physical activity). The full cohort data will be required.