Vitamin D and overall cancer risk and cancer mortality: a Mendelian randomization study

There is considerable debate regarding the role that 25-hydroxyvitamin D [25(OH)D] concentrations play in cancer risk or mortality, with earlier studies drawing mixed conclusions. Using data from the UK Biobank (UKB), we evaluate whether genetically predicted 25(OH)D concentrations are associated with overall cancer susceptibility and cancer mortality using five 25(OH)D genetic markers. Data comprised 438 870 white British UKB participants aged 37-73, including 46 155 cancer cases and 6998 cancer-specific deaths. Participants with keratinocyte cancers and/or benign tumors were excluded from the analysis. Odds ratios were calculated per 20 nmol/L increase in genetically predicted 25(OH)D for cancer risk and cancer mortality. For individual cancer risks, estimates were meta-analyzed with publicly available data using a fixed-effect inverse-variance-weighted model. We demonstrated that genetically low plasma 25(OH)D concentrations were not associated with increased cancer risk nor cancer mortality. Stratification by sex or cancer types did not reveal any meaningful differences albeit wider confidence intervals. Fixed-effect meta-analysis of our individual cancer risk estimates with those derived from publicly available cancer consortia data and previous studies further reinforced our null Mendelian randomization findings on prostate, lung, colorectal and breast cancers with tight confidence intervals; for ovarian and pancreatic cancers, our estimates were less precise despite being not statistically significant. Taken altogether, our results provide no genetic evidence for an association between vitamin D and overall cancer outcomes, with tight confidence intervals to exclude all but very small effect sizes.