Abstract
BACKGROUND: Observational studies have shown that being taller is associated with greater cancer risk. However, the interpretation of such studies can be hampered by important issues such as confounding and reporting bias. METHODS: We used the UK Biobank resource to develop genetic predictors of height and applied these in a Mendelian randomisation framework to estimate the causal relationship between height and cancer. Up to 438,870 UK Biobank participants were considered in our analysis. We addressed two primary cancer outcomes, cancer incidence by age ~60 and cancer mortality by age ~60 (where age ~60 is the typical age of UK Biobank participants). RESULTS: We found that each genetically predicted 9 cm increase in height conferred an odds ratio of 1.10 (95% confidence interval 1.07-1.13) and 1.09 (1.02-1.16) for diagnosis of any cancer and death from any cancer, respectively. For both risk and mortality, the effect was larger in females than in males. CONCLUSIONS: Height increases the risk of being diagnosed with and dying from cancer. These findings from Mendelian randomisation analyses agree with observational studies and provide evidence that they were not likely to have been strongly affected by confounding or reporting bias.