Serum hormones and prostate cancer incidence and mortality in UK Biobank


Session type:

Ruth Travis1,Eleanor Watts1,Georgina Fensom1,Aurora Perez-Cornago1,Anika Knuppel1,Naomi Allen1,Marc Gunter2,Richard Martin3,Karl Smith Byrne2,Neil Murphy2,Kostas Tsilidis4,Tim Key1
1University of Oxford,2International Agency for Research on Cancer,3University of Bristol,4University of Ioannina



Modifiable risk factors for prostate cancer are not well established. Insulin-like growth factor-I (IGF-I) and testosterone have been implicated in prostate cancer development and progression, but these associations are not fully understood. We investigated the associations of circulating concentrations of IGF-I, free (biologically active testosterone) and total testosterone, and sex hormone binding globulin (SHBG) with prostate cancer incidence and mortality in a large British cohort.


We studied up to 200,452 male UK Biobank participants who were free from cancer and not receiving hormone therapy at recruitment, and had hormone measurements on baseline blood samples. Free testosterone was calculated from measured total testosterone and binding protein concentrations. Follow-up for cancer incidence and vital status was via data linkage to national registries. Cox proportional hazards models yielded multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for prostate cancer incidence and mortality by serum hormone concentrations. HRs were corrected for regression dilution using repeat hormone measurements from a subsample of up to 7794 men.


After a mean follow-up of 6.9 years, 5412 men were diagnosed with prostate cancer and 296 had died from the disease. Higher circulating IGF-I was associated with an elevated risk of prostate cancer diagnosis (HR per 1 SD increase = 1.11, 95%CI 1.07-1.15) and prostate cancer mortality (1.17, 1.01-1.36). Higher free testosterone was associated with an elevated risk of incident prostate cancer (1.14, 1.08-1.20), whereas higher SHBG was associated with a lower risk (0.93, 0.90-0.96), but neither was associated with prostate cancer mortality. Total testosterone concentration was not associated with prostate cancer incidence or mortality.


The findings from this large prospective cohort study of circulating hormone concentrations support the hypothesised roles of IGF-I and free testosterone in prostate cancer development. Future research will examine associations by tumour characteristics.