Dr A. Majeed, School of Public Policy, University College London, 29–30 Tavistock Square, London WC1H 9EZ, UK. E-mail: email@example.com
Objectives To examine trends in prostate cancer incidence and mortality in England and Wales between 1971 and 1998, using a newly developed and validated national cancer database and the national mortality database.
Methods Age-standardized incidence and death rates were calculated directly and trends in relative survival rates among men with prostate cancer registered during 1971–1990 were examined.
Results The annual number of new cases of prostate cancer registered in England and Wales increased by 179% between 1971 and 1993, from 6174 to 17 210. Directly age-standardized incidence rates increased by 104% between 1971 and 1993, from 29 to 59 per 100 000. The number of deaths from prostate cancer increased by 113% between 1971 and 1998, from 4027 to 8570. Directly age-standardized death rates increased by 49% between 1971 and 1995 and then decreased by 8% between 1995 and 1998, an overall increase of 38% (20 to 27 per 100 000) between 1971 and 1998. The relative survival rate for prostate cancer among men diagnosed during 1986–1990 was 77% at 1 year and 42% at 5 years, compared with 67% and 33%, respectively, for cases diagnosed during 1971–1975. The increase in survival rates was confined to men diagnosed with prostate cancer up to 1985 and no increase was seen for cases diagnosed after 1985.
Conclusions Prostate cancer is becoming a growing burden on the health service. The explanation for the large increase in prostate cancer incidence and mortality is unclear and needs further investigation. The lack of any improvement in survival rates in cases diagnosed after 1985 is of concern, and suggests that the current management of prostate cancer in both primary and secondary care may need to be reviewed.
Prostate cancer incidence and mortality have increased in many developed countries in recent years [1–3]. In England and Wales, mortality from prostate cancer increased steadily during the 1980s and by 1990, prostate cancer had become the second commonest cause of death from cancer in men . Hence, prostate cancer is becoming a more important public health and health-service issue both within the UK and elsewhere leading, for example, to a debate on whether to implement screening for prostate cancer [5,6]. Decisions about the provision of services for the diagnosis and management of prostate cancer should be based on good information about incidence, mortality and survival [7,8].
An enhanced and validated national cancer database has now been developed which allows analysis of cancer incidence rates up to 1993 and death rates up to 1998. The database has been used to examine survival rates for patients diagnosed with cancer between 1971 and 1990 . In this report, we used the new national database to examine trends in prostate cancer incidence, mortality and survival in England and Wales during 1971–1998.
The National Cancer Registry (NCR) at the Office for National Statistics collates data held by regional cancer registries, to produce national data on cancer incidence, prevalence and survival, and has access to the national mortality database. The NCR identifies duplicate registrations and converts the data supplied by each regional registry into a person-based database. Since 1990, the NCR has instituted several quality-control measures and logical checks which have resulted in a substantial increase in the accuracy of the data it holds. The NCR now produces accurate population-based cancer incidence rates for years from 1971 onwards. The annual numbers of prostate cancer registrations were examined by age group in England and Wales for 1971–1993, and the annual number of deaths from prostate cancer by age group in England and Wales for 1971–1998.
The data on registrations and deaths were used in combination with age-specific population estimates for England and Wales to calculate annual incidence and death rates for prostate cancer by age group. Directly age-standardized incidence and death rates were also calculated, using the European Standard Population. Direct standardization reduces the effect of changes in population structure on registration and death rates. This is important because prostate cancer is most common in older men. Because there are substantially more older men in the population since 1971, crude incidence and death rates would also have increased since 1971, even if age-specific rates had remained constant.
Between 1984 and 1992, the Office of Population Censuses and Surveys, which administered the registration of deaths in England and Wales until 1996, introduced a revised interpretation of the WHO Rule 3, which governs how information in the two parts of the death certificate is used to determine the underlying cause of death. Consequently, deaths from causes such as pneumonia declined steeply in 1984, whereas deaths from causes often mentioned in part II of the certificate increased . This change resulted in an increase in the death rate from prostate cancer in 1984, which was most marked in the elderly. All mortality data in the present report have been adjusted to correct for these procedural changes. Consequently, the mortality rates presented for 1984–1992 will be lower than those reported in previous publications.
The NCR Database is linked to the NHS Central Register and receives notifications of the death of all registered patients with cancer. This linkage allows the calculation of cancer survival and prevalence rates. The NCR Database was used to examine trends in survival among patients with prostate cancer registered during 1971–1990 and followed up to 31 December 1995 . Patients in whom survival status was unknown or recorded as zero (i.e. died on the day of diagnosis) were excluded from the survival analysis. Both crude and relative survival rates are presented, with relative survival rates calculated using an algorithm developed by Estève et al.. The crude survival rate is the percentage of patients still alive at a given time after diagnosis. The relative survival rate is crude survival rate adjusted for the background death rate in the general population. Consequently, relative survival rates are higher than crude survival rates. Cases diagnosed in 1992 are the latest for which survival data are currently available. The estimated prevalence of prostate cancer on 1 January 1993 was calculated as the number of men diagnosed with prostate cancer during 1983–1992 who were still alive on this date.
The annual number of new cases of prostate cancer registered in England and Wales increased by 179% between 1971 and 1993, from 6174 to 17 210. Prostate cancer is overwhelmingly a disease of older men: 96% of cases in 1993 occurred in men aged 60 years. Across the age groups, the largest increase in the number of registrations (over 200%) occurred in men aged 80 years (Table 1). The all-ages directly age-standardized incidence rate increased by 104% between 1971 and 1993, from 29 to 59 per 100 000 (Fig. 1). The largest increase in incidence rates (over 100%) occurred in men in their sixth decade.
Table 1. Prostate cancer in England and Wales: number of new cases, age-specific and age-standardized incidence and death rates, and percentage increases by age group in 1971–1993 and 1971–1998
Number of new cases
Incidence rate per 100 000
Number of deaths
Death rate per 100 000
The annual number of deaths from prostate cancer increased by 113% between 1971 and 1998, from 4027 to 8570; 97% of the deaths from prostate cancer in 1998 occurred in men aged 60 years. Across the age groups, the largest increase in the number of deaths (188%) occurred in men aged 80 years (Table 1). The directly age-standardized death rate in this age group increased by 47% during the same period, from 400 to 587 per 100 000 (Table 1). There were also large increases in death rates in men aged 60–69 and 70–79 years but the increases were not as great as those in incidence rates. There was little change in death rates during the 1970s (directly age-standardized rate 20 per 100 000 in 1971 and 21 per 100 000 in 1979) and death rates from prostate cancer only began to increase significantly from 1980 onwards (Fig. 1). Both the number of deaths (8836) and the age-standardized death rate (30 per 100 000) peaked in 1995. Death rates from prostate cancer declined in the subsequent 3 years and by 1998, the age-standardized death rate had fallen by 8% to 27 per 100,000, an overall increase of 49% between 1971 and 1995, and 38% between 1971 and 1998.
The age-standardized relative survival rate for prostate cancer among men diagnosed during 1986–1990 was 77% at 1 year and 42% at 5 years, compared with 67% and 33%, respectively, for cases diagnosed during 1971–1975 (Table 2). The increase in survival rates was confined to men diagnosed with prostate cancer up to 1985. No increase in either 1- or 5-year survival was seen for cases diagnosed after this period.
Table 2. Percentage crude and relative survival in England and Wales at 1, 5 and 10 years since diagnosis of prostate cancer for cases diagnosed during 1971–1990
Crude/relative survival rate after diagnosis at
Year of diagnosis
Among the 118 682 cases of prostate cancer diagnosed during 1983–1992, 40 909 were still alive on 1 January 1993, giving an estimated prevalence of prostate cancer on this date of 1.6 per 1000 men. Hence, a typical GP with a list of 2000 patients would be expected to have one to two patients with prostate cancer at any one time.
Both prostate cancer incidence and death rates increased substantially during the periods examined. Age-standardized incidence rates more than doubled between 1971 and 1993, and death rates increased by 38% between 1971 and 1998. However, whereas incidence rates showed an almost unbroken increase, death rates peaked in 1995 and fell in the subsequent 3 years, the first sustained decrease in annual death rates for prostate cancer for 30 years. Survival rates after a diagnosis of prostate cancer were greater in cases diagnosed during 1981–1985 than in those diagnosed during 1971–1975. However, there was no improvement in survival for men diagnosed with prostate cancer during 1986–1990 compared with men diagnosed during 1981–1985.
The main strengths of the present study are that it covered the whole of England and Wales, and used a substantially improved national database of cancer registrations. This is the first occasion that the new database has been used to examine long-terms trends in prostate cancer incidence and mortality. Hence, the results presented here are the most up-to-date and accurate national incidence and death rates published thus far. They confirm that prostate cancer is common, increasing in incidence and is imposing a growing burden on the health service.
Registration of cancer is voluntary and it is therefore possible that not all cases of prostate cancer that occurred during the study period were registered. However, cancer is registered using data from several sources, including hospital specialists, pathologists, GPs and death certificates. The proportion of cases of prostate cancer registered solely on the basis of information from death certificates is ≈ 8%. Using a formula developed by the International Agency for Research on Cancer, this implies that > 96% of all cases of prostate cancer were notified to regional cancer registries .
In contrast to registration, completion of a death certificate is compulsory and hence mortality data are complete. Although there are errors in death certification by doctors  improvements in diagnostic techniques in recent years, e.g. the development of TRUS-guided biopsy and testing for PSA, should have led to an improvement in the diagnosis of prostate cancer during the period presently examined . Increases in the rates of TURP for BPH will also have led to more cancers being diagnosed through histological examination of the excised prostatic tissue [15, 16].
Because latent carcinoma of the prostate is a common finding among elderly men [17,18] these diagnostic advances and changes in surgical practice may be responsible for some of the increased incidence of prostate cancer that has occurred in many developed countries. For example, an increased use of PSA testing was thought to be largely responsible for the 82% increase in the incidence of prostate cancer in the USA between 1986 and 1991 [19,20]. However, the incidence of prostate cancer in England and Wales has been rising steadily since 1971 and hence the increase in incidence long pre-dates the widespread use of PSA testing, which only become available in many parts of England and Wales within the last 10 years. There was an apparent acceleration in the increase in incidence of prostate cancer during 1990–1993, which may be a consequence of increased testing for PSA. Other explanations for the increase in incidence include a possible association with vasectomy , a decline in dietary selenium , and other lifestyle and environmental factors .
Disappointingly, there was no improvement in survival for men diagnosed with prostate cancer during 1986–1990 compared with 1981–1985. It is unclear why this should be so, but it may be linked to UK having relatively fewer urologists and oncologists than in most other developed countries . There is also uncertainty about how to treat prostate cancer, particularly when localized, leading to local, regional and international differences in management [25,26]. Death rates from prostate cancer decreased between 1995 and 1998, a period during which unpublished provisional data suggest that incidence rates were still increasing. Hence, survival from prostate cancer in England and Wales may be starting to improve again.
Many questions remain unanswered about prostate cancer, amongst which is an explanation for the rapid increase in its incidence and for the relatively poorer survival of men with prostate cancer in England and Wales than in many other countries. The enhanced national cancer database provides a valuable tool for research on these issues and will be used to monitor trends in prostate cancer incidence, mortality, prevalence and survival. The data held on the database can also be linked to small-area statistics thus facilitating ecological studies which can further help improve understanding of the aetiology of prostate cancer.