Comparison of prostate cancer survival in Germany and the USA : can differences be attributed to differences in stage distributions?

OBJECTIVES: To better understand the influence of prostate-specific antigen (PSA) screening and other health system determinants on prognosis of prostate cancer, up-to-date relative survival (RS), stage distributions, and trends in survival and incidence in Germany were evaluated and compared with the United States of America (USA). PATIENTS AND METHODS: Incidence and mortality rates for Germany and the USA for the period 1999-2010 were obtained from the Centre for Cancer Registry Data at the Robert Koch Institute and the USA Surveillance Epidemiology and End Results (SEER) database. For analyses on stage and survival, data from 12 population-based cancer registries in Germany and from the SEER-13 database were analysed. Patients (aged ≥ 15 years) diagnosed with prostate cancer (1997-2010) and mortality follow-up to December 2010 were included. The 5- and 10-year RS and survival trends (2002-2010) were calculated using standard and model-based period analysis. RESULTS: Between 1999 and 2010, prostate cancer incidence decreased in the USA but increased in Germany. Nevertheless, incidence remained higher in the USA throughout the study period (99.8 vs 76.0 per 100,000 in 2010). The proportion of localised disease significantly increased from 51.9% (1998-2000) to 69.6% (2007-2010) in Germany and from 80.5% (1998-2000) to 82.6% (2007-2010) in the USA. Mortality slightly decreased in both countries (1999-2010). Overall, 5- and 10-year RS was lower in Germany (93.3%; 90.7%) than in the USA (99.4%; 99.6%) but comparable after adjustment for stage. The same patterns were seen in age-specific analyses. Improvements seen in prostate cancer survival between 2002-2004 and 2008-2010 (5-year RS: 87.4% and 91.2%; +3.8% units) in Germany disappeared after adjustment for stage (P = 0.8). CONCLUSION: The survival increase in Germany and the survival advantage in the USA might be explained by differences in incidence and stage distributions over time and across countries. Effects of early detection or a lead-time bias due to the more widespread utilisation and earlier introduction of PSA testing in the USA are likely to explain the observed patterns.