BACKGROUND: It is largely unknown whether prostate-specific antigen (PSA) level at first date of testing predicts long-term risk of prostate cancer (PCa) incidence and mortality in the general population. OBJECTIVE: Determine whether baseline PSA levels predict long-term risk of PCa incidence and mortality. DESIGN, SETTING, AND PARTICIPANTS: We examined 4383 men aged 20-94 yr from the Danish general population in the prospective Copenhagen City Heart Study. PSA was measured in plasma samples obtained in 1981-1983. MEASUREMENTS: PCa incidence and mortality as a function of baseline PSA was assessed using Kaplan-Meier plots of cumulative incidence and competing risk subhazard ratios. RESULTS AND LIMITATIONS: During 28 yr of follow-up, 170 men developed PCa, and 94 men died from PCa. Median follow-up was 18 yr (range: 0.5-28 yr). For PCa incidence, the subhazard ratio was 3.0 (95% confidence interval [CI], 1.9-4.6) for a PSA level of 1.01-2.00 ng/ml, 6.8 (95% CI, 4.2-11) for PSA 2.01-3.00 ng/ml, 6.6 (95% CI, 3.4-13) for PSA 3.01-4.00 ng/ml, 16 (95% CI, 10.4-25) for PSA 4.01-10.00 ng/ml, and 57 (95% CI, 32-104) for PSA >10.00 ng/ml versus 0.01-1.00 ng/ml. For PCa mortality, corresponding subhazard ratios were 2.2 (95% CI, 1.3-3.9), 5.1 (95% CI, 2.8-9.0), 4.2 (95% CI, 1.8-10), 7.0 (95% CI, 3.8-14), and 14 (95% CI, 6.0-32). For men with PSA levels of 0.01-1.00 ng/ml, the absolute 10-yr risk of PCa was 0.6% for ages <45 yr, 0.7% for ages 45-49 yr, 1.1% for ages 50-54 yr, 1.2% for ages 55-59 yr, 1.3% for ages 60-64 yr, 1.1% for ages 65-69 yr, 1.3% for ages 70-74 yr, and 1.5% for ages ≥75 yr; corresponding values for PSA levels >10.00 ng/ml were 35%, 41%, 63%, 71%, 77%, 69%, 75%, and 88%, respectively. CONCLUSIONS: Stepwise increases in PSA at first date of testing predicted a 3-57-fold increased risk of PCa, a 2-16-fold increased risk of PCa mortality, and a 35-88% absolute 10-yr risk of PCa in men with PSA levels >10.00 ng/ml. Equally important, the absolute 10-yr risk of PCa in men with PSA levels 0.01-1.00 ng/ml was only 0.6-1.5%.
BACKGROUND: It is largely unknown whether prostate-specific antigen (PSA) level at first date of testing predicts long-term risk of prostate cancer (PCa) incidence and mortality in the general population. OBJECTIVE: Determine whether baseline PSA levels predict long-term risk of PCa incidence and mortality. DESIGN, SETTING, AND PARTICIPANTS: We examined 4383 men aged 20-94 yr from the Danish general population in the prospective Copenhagen City Heart Study. PSA was measured in plasma samples obtained in 1981-1983. MEASUREMENTS: PCa incidence and mortality as a function of baseline PSA was assessed using Kaplan-Meier plots of cumulative incidence and competing risk subhazard ratios. RESULTS AND LIMITATIONS: During 28 yr of follow-up, 170 men developed PCa, and 94 men died from PCa. Median follow-up was 18 yr (range: 0.5-28 yr). For PCa incidence, the subhazard ratio was 3.0 (95% confidence interval [CI], 1.9-4.6) for a PSA level of 1.01-2.00 ng/ml, 6.8 (95% CI, 4.2-11) for PSA 2.01-3.00 ng/ml, 6.6 (95% CI, 3.4-13) for PSA 3.01-4.00 ng/ml, 16 (95% CI, 10.4-25) for PSA 4.01-10.00 ng/ml, and 57 (95% CI, 32-104) for PSA >10.00 ng/ml versus 0.01-1.00 ng/ml. For PCa mortality, corresponding subhazard ratios were 2.2 (95% CI, 1.3-3.9), 5.1 (95% CI, 2.8-9.0), 4.2 (95% CI, 1.8-10), 7.0 (95% CI, 3.8-14), and 14 (95% CI, 6.0-32). For men with PSA levels of 0.01-1.00 ng/ml, the absolute 10-yr risk of PCa was 0.6% for ages <45 yr, 0.7% for ages 45-49 yr, 1.1% for ages 50-54 yr, 1.2% for ages 55-59 yr, 1.3% for ages 60-64 yr, 1.1% for ages 65-69 yr, 1.3% for ages 70-74 yr, and 1.5% for ages ≥75 yr; corresponding values for PSA levels >10.00 ng/ml were 35%, 41%, 63%, 71%, 77%, 69%, 75%, and 88%, respectively. CONCLUSIONS: Stepwise increases in PSA at first date of testing predicted a 3-57-fold increased risk of PCa, a 2-16-fold increased risk of PCa mortality, and a 35-88% absolute 10-yr risk of PCa in men with PSA levels >10.00 ng/ml. Equally important, the absolute 10-yr risk of PCa in men with PSA levels 0.01-1.00 ng/ml was only 0.6-1.5%.
Authors: Signe Benzon Larsen; Klaus Brasso; Peter Iversen; Jane Christensen; Michael Christiansen; Sigrid Carlsson; Hans Lilja; Søren Friis; Anne Tjønneland; Susanne Oksbjerg Dalton Journal: Eur J Cancer Date: 2013-05-17 Impact factor: 9.162