Melissa J. Assel, Axel Gerdtsson, Daniel L.J. Thorek, Sigrid V. Carlsson, Johan Malm, Peter T. Scardino, Andrew Vickers, Hans Lilja, David Ulmert

To evaluate whether anthropometric parameters add to PSA measurements in middle-aged men for risk assessment of prostate cancer (PCa) diagnosis and death. After adjusting for PSA, both BMI and weight were significantly associated with an increased risk of PCa death with the odds of a death corresponding to a 10 kg/m2 or 10 kg increase being 1.58 (95% CI 1.10, 2.28;p= 0.013) and 1.14 (95% CI 1.02, 1.26;p= 0.016) times greater, respectively. AUCs did not meaningfully increase with the addition of weight or BMI to prediction models including PSA. In 1974 to 1986, 22,444 Swedish men aged 44 to 50 enrolled in Malmö Preventive Project, Sweden, and provided blood samples and anthropometric data. Rates of PSA screening in the cohort were very low. Documentation of PCa diagnosis and disease-specific death up to 2014 was retrieved through national registries. Among men with anthropometric measurements available at baseline, a total of 1692 men diagnosed with PCa were matched to 4190 controls, and 464 men who died of disease were matched to 1390 controls. Multivariable conditional logistic regression was used to determine whether diagnosis or death from PCa were associated with weight and body mass index (BMI) at adulthood after adjusting for PSA. Men with higher BMI and weight at early middle age have an increased risk of PCa diagnosis and death after adjusting for PSA. However, in a multi-variable numerical statistical model, BMI and weight do not importantly improve the predictive accuracy of PSA. Risk-stratification of screening should be based on PSA without reference to anthropometrics.