The possible effect of prostate-specific antigen (PSA) testing on prostate cancer mortality has remained controversial, despite the test's widespread application. We examined age-specific mortality trends for prostate cancer in Austria before and after the introduction of (opportunistic) PSA testing, to ask whether PSA screening reduces prostate cancer mortality in a uniform cohort of men with equal access to health care. Prostate cancer mortality data covering all 9 federal states of Austria were analysed from 1970 to 2002. PSA testing became widely available in Austria not before 1989. Tyrol, one of the nine federal states of Austria, independently launched a mass prostate cancer prevention project in 1993. We applied join-point regression models to identify changes in the slope of age-specific mortality trends in selected age groups (50-59, 60-69, 70-79, and 80-89 years) and calculated the annual percent change (APC) in mortality between 1970 and 2002 for Tyrol and the rest of Austria separately. After 12 years of follow-up, we were not able to observe a significant reduction in prostate cancer mortality since the introduction of the PSA test in the age groups of 50-59, 60-69, and 80-89 years. A significant decrease was found in the age group of 70-79 (Austria without Tyrol 1989 through 2002: APC, -2.36; 95% CI, -3.38 to -1.34; Tyrol 1991 through 2002: APC, -6.42; 95% CI, -8.92 to -3.86). In this age group the join points 1989 and 1991 cannot be related to PSA testing. PSA screening does not appear to reduce prostate cancer mortality in a uniform cohort of men with equal access to health care. However, given the long lead-time for prostate cancer, even longer follow-up may still be needed to detect any important trends.

The aim of the study was to assess the impact of prostate-specific antigen (PSA) testing on prostate cancer mortality in Austria. A join-point regression model and permutation tests were used to identify changes in the slope of age-specific trends respectively calculating the annual percentage change (APC). Age-adjusted incidence increased (P < 0.01) between 1983 and 1997 by 79% from 52.2 to 93.6 cases per 100 000 men/year. Incidence in localized/regional stage disease increased in all ages by 143% from 25.7 to 62.4 cases per 100 000 men/year. Incidence in distant disease decreased (P < 0.01) between 1983 and 1997 in all ages by 38% from 9.5 to 5.9 cases per 100 000 men/year. Incidence in unstaged disease increased (P < 0.01) between 1983 and 1997 in all ages by 300% from 4.5 to 18 cases per 100 000 men/year. Age-adjusted mortality increased (P < 0.05) by 13% from 26.8 in 1983 to 30.3 deaths per 100 000 men/year in 1999. No significant changes of trends in mortality rates were detected in the age groups 50-59 years. In the age group 70-79 years the trend changed (P < 0.05) direction in 1991 and in 1994; 1983 through 1991 APC = 3.52 (95% CI 1.37, 5.72), 1991 through 1994 APC = -10.27 (95% CI -26.20, 9.1) and 1994 through 1999 APC = -0.25 (95% CI -4.55, 4.24). PSA testing increased incidence but no impact on mortality in the target population can be observed so far.

The observed changes in cancer mortality are primarily related to changes in incidence and, in the last decade, to improved treatment and early detection, but neither of these contributions can be quantified. The strengthening of prevention research (primary and secondary prevention) and further implementation of preventive knowledge is urgently needed.

Oncologists should involve family physicians in disclosing bad news to patients. There are considerable deficiencies regarding information-exchange in cancer care in Austria.

No abstract