Assessing the causal association between 25-hydroxyvitamin D and the risk of oral and oropharyngeal cancer using Mendelian randomization

Abstract

Circulating 25-hydroxyvitamin D (25OHD) is an appealing potential intervention for cancer risk and has been associated with oral and oropharyngeal cancer risk but evidence is inconsistent. The availability of genetic variants, uncorrelated with known confounders, but predictive of 25OHD and genetic data in a large oral and oropharyngeal cancer collaboration aids causal inference when assessing this association. A total of 5,133 oral and oropharyngeal cancer cases and 5,984 controls with genetic data were included in the study. Participants were based in Europe, North America and South America and were part of the Genetic Associations and Mechanisms in Oncology (GAME-ON) Network. Five genetic variants reliably associated with circulating 25OHD were used to create a relative genetic measure of 25OHD. In the absence of measured 25OHD, two-sample Mendelian randomization using individual level outcome data were used to estimate causal odds ratios (OR) for cancer case status per standard deviation increase in log25OHD. Analyses were replicated in an independent population-based cohort (UK Biobank). In the GAME-ON study, there was little evidence of a causal association between circulating 25OHD and oral cancer (OR = 0.86 [0.68;1.09], p = 0.22), oropharyngeal cancer (OR = 1.28 [0.72;2.26], p = 0.40) or when sites were combined (OR = 1.01 [0.74;1.40], p = 0.93). Replication in UK Biobank and pooled estimates produced similar results. Our study suggests that a clinically relevant protective effect of 25OHD on oral and oropharyngeal cancer risk is unlikely and supplementation of the general population with 25OHD is unlikely to be beneficial in preventing these cancers.