Effect of a one-time sensitive screening intervention on prostate cancer mortality: The STHLM3 randomized trial.

Abstract

312 Background: Prostate cancer (PCa) screening remains controversial and further results on the long-term benefits and harms are needed. The population-based STHLM3 trial (ISRCTN84445406) invited men to undergo a one-time screening using both prostate-specific antigen (PSA) and Stockholm3 (a predictive model using clinical variables, blood biomarkers, and polygenic risk). Methods: Between 2012 and 2014, men aged 50–69 years without diagnosis of PCa and residing in Stockholm County, Sweden, were randomized (7:3) to receive invitation for PCa screening or no invitation. Participants provided blood for PSA; those with PSA ≥1 ng/mL also had Stockholm3 performed. Men with PSA ≥3 ng/mL or Stockholm3 risk ≥10 were referred for systematic biopsy. We calculated cumulative risks and risk ratios (RRs) for PCa mortality (primary outcome), all-cause mortality, PCa incidence, and cumulative mean number of PSA tests and biopsies per man after 11.6 years of follow-up, comparing invited and non-invited men (intention to screen, ITS). An instrumental variable analysis was used to estimate the effect of screening participation on PCa mortality. An alpha level of 0.035 was used for the primary outcome. Results: We randomized 240,494 men, 169,621 invited and 70,873 non-invited. Of the invited men, 59,088 (35%) participated. After 11.6 years, 352 PCa deaths occurred in the invited group and 161 in the non-invited group (RR 0.96, 96.5% CI 0.77–1.20). After adjusting for participation, the RR of PCa death with one-time screening was 0.41 (95% CI 0.18–0.91) at 6 years and 0.85 (95% CI 0.39–1.88) at 11.6 years (Table 1). PCa incidence was similar among invited and non-invited men at 11.6 years (risk 73.1 vs. 72.3 per 1000; RR 1.01, 95% CI 0.96–1.07), as was all-cause mortality (risk 115.8 vs. 118.1 per 1000; RR 0.98, 95% CI 0.93–1.04). On average, men took more than 3 PSA tests in both groups (mean ratio 1.03, 95% CI 1.02-1.05); 11.5% and 10.6% of men underwent a prostate biopsy in the invited and non-invited groups, respectively (RR 1.08, 95% CI 1.05-1.11). Conclusions: In a population with high opportunistic testing rates, a single invitation to a sensitive PCa screening did not significantly reduce PCa mortality compared to no invitation after 11.6 years, however effects of mortality improvement from one-time screening were seen at 6 years. Clinical trial information: ISRCTN84445406 . Prostate cancer mortality in the intention-to-screen (ITS) and instrumental variable (IV) analysis. Invited (n = 169,621) Non-invited (n = 70,873) Risk Ratio (95% CI) Analysis Events Risk* (95% CI) Events Risk* (95% CI) P value End of follow-up (11.6 years) ITS 352 2.50 (2.21-2.82) 161 2.60 (2.19-3.07) 0.96 (0.77-1.20) † 0.711 IV - - - - 0.85 (0.39-1.88) 6 years ITS 122 0.73 (0.61-0.87) 66 0.94 (0.74-1.19) 0.77 (0.57-1.04) IV - - - - 0.41 (0.18-0.91) 10 years ITS 327 2.02 (1.81-2.25) 154 2.28 (1.94-2.66) 0.89 (0.73-1.07) IV - - - 0.65 (0.35-1.18) *Per 1000 men. †96.5% Confidence Interval (α = 0.035).