Diabetes mellitus and risk of prostate cancer - A focus on shared risk factors and survivor bias

Prostate cancer (PCa) is a heterogenous disease including very aggressive lethal tumors to incidentally discovered clinically insignificant tumors. Age, ethnicity, and family history are the only established risk factors. Type 2 diabetes mellitus (T2D) has been found to associate with increased overall cancer risk but with lower risk of specifically PCa. T2D is initially associated with increased blood glucose, insulin and insulin-like growth factor-1 levels. However, after long disease duration, insulin production diminishes and among men testosterone levels decrease. In healthy men, blood glucose and insulin levels are affected by various lifestyle factors, including diet and physical activity. Since lifestyle factors and T2D are associated with a range of health outcomes other than PCa, so called competing events, estimation of their association with PCa may be biased due to longer survival among men with healthy lifestyles and non-diabetics. Further, there is evidence to suggest that T2D and cancer may share some genetic risk factors. The effect of the strongest T2D risk variant to date, TCF7L2 rs7903146, has been shown to be modified by dietary and lifestyle factors. This doctoral project consisted of four epidemiological studies in the population-based, prospective Malmö Diet and Cancer Study. The cohort includes 11,063 men with information on dietary habits, socioeconomic and lifestyle factors, disease history, anthropometric measurements, and blood samples. Among this population (aged 45-73 at baseline), 10,578 men were free of cancer at baseline, and 1,016 men were subsequently diagnosed with PCa until 31 December 2009. We found that high intakes of refined carbohydrates (primarily sugar-sweetened beverages) may increase the risk of incident PCa. In a nested case-control study we found that a biomarker for whole-grain intake, plasma alkylresorcinol metabolites, was positively correlated with whole-grain intake but not with lower risk of PCa. Men with prevalent diabetes at study baseline had a lower risk of PCa diagnosis, partly attributed to a higher risk of non-cancer deaths. In contrast, we found that incident T2D may increase the risk of high-risk PCa. The TCF7L2 T2D genetic risk variant was not significantly associated with PCa, however, carriers of the risk allele who were hyperglycemic at baseline or reported unhealthy lifestyles had a higher risk of incident PCa. In conclusion, this project showed that there may be an association between diet, T2D, and risk of PCa that is potentially mediated by metabolic changes and modified by genetic factors. Survivor bias may be of particular concern in epidemiological studies on PCa