Abstract
Mental health conditions (MHCs) affect both psychological health and biological systems and have also been linked to cancer risk. However, evidence from epidemiological studies regarding this link remains inconsistent. We conducted a population-based prospective cohort study involving 402,255 UK Biobank participants to investigate the associations of five MHCs (depressive disorders [DD], anxiety disorders [AD], bipolar disorder [BD], schizophrenia [SZ] and post-traumatic stress disorder [PTSD]) with overall and site-specific cancer risk. Cox proportional hazard models were used, adjusting for sociodemographic, health-related and lifestyle confounders. Over a median follow-up of 13.4 years, 68,065 (17%) incident cancer cases were recorded. DD (HR 1.18; 95% CI 1.13-1.23), AD (HR: 1.17, 95% CI: 1.11-1.24) and BD (HR: 1.29, 95% CI: 1.11-1.51) were associated with increased overall cancer risk. No significant association was found for SZ and PTSD. The associations of DD (HR: 1.27, 95% CI: 1.18-1.35) and BD (HR: 1.54, 95% CI: 1.26-1.88) were only significant in men. AD and DD were positively associated with lung, blood and liver cancers, while AD was also associated with prostate cancer. A dose-response relationship was observed between depressive symptom severity and cancer risk. While we cannot assume causality, our finding suggests that diagnoses of MHCs could be useful for cancer risk stratification and prevention.