Objective To assess the extent to which the increasing incidence of prostate cancer in Scotland can be explained by increased detection, particularly through transurethral resection of the prostate (TURP) and use of the prostate-specific antigen (PSA) test.
Subjects and methods This population-based study was confined to men resident in Scotland and aged 50 years. Temporal trends were examined in age-specific and age-standardized incidence, mortality and TURP rates, and PSA testing rates during 1981–1996. Also analysed were the geographical variations in age-standardized incidence and mortality rates during two distinct periods, 1984–1986 (before PSA testing) and 1994–1996 (after PSA testing). Finally, incidence rates and relative survival at 5 years were calculated by age group and 5-year periods of diagnosis during 1968–1992.
Results The incidence of prostate cancer in men aged 50 years increased from an age-standardized rate of 142.0 per 100 000 in 1981 to 240.9 in 1996, with the steepest increase occurring between 1992 and 1993. The mortality rate increased similarly until 1993, but was relatively stable thereafter, falling slightly in 1996. In 1981–1988, incidence rates were closely correlated with TURP rates (r = 0.98, P < 0.001). In 1989–1996, incidence was closely correlated with PSA testing rates (r = 0.98, P < 0.001). By 1994–1996, incidence rates varied substantially between Scottish mainland health boards (range 167.7–303.0 per 100 000), with much less variation in mortality rates (90.7–110.0). Relative survival has increased recently in all age groups although, in the era before PSA testing, survival was reasonably stable despite increasing incidence.
Conclusion Although there may have been a true increase in risk, much of the observed increase in the incidence of prostate cancer in Scotland between 1981 and 1996 has been caused by increased detection, leading recently to considerable variation among different areas of the country. The extent to which this represents the early diagnosis of tumours which would eventually cause symptoms or be life-threatening, or detection of latent disease which would never have become symptomatic, is not clear. There is no evidence so far that the increased incidence is associated with any substantial reduction in mortality.