Approximately 9200 cancers of the urinary bladder are registered in England and Wales each year, of which 6500 in men (8% of all malignancies) and 2700 in women (3%) (Office for National Statistics, 2003). Bladder cancer ranks as the fourth most common in men and ninth in women, and it causes approximately 4300 deaths (2800 in men) a year in England and Wales.
Incidence increases with age, most steeply above the age of 60 years. The two-fold or greater excess in men is world-wide (Ferlay et al, 2004), and the excess is three-fold above the age of 60 years. Bladder cancer is uncommon below the age of 50 years, except in countries where it is linked with chronic schistosomal infestation of the bladder. Both incidence and mortality show deprivation gradients, with higher rates in more deprived areas, but the differences are small (Quinn et al, 2001).
Incidence rose steadily in both sexes during the 1970s and 1980s, reaching a peak of approximately 31 cases per 100000 per year in men and 9 per 100000 in women by 1990. Incidence has fallen by approximately 10% since then in both sexes. Annual death rates in men have fallen by a third since the early 1990s (from 12 to 8 per 100000), but the death rate in women (3 per 100000) has not changed (Quinn et al, 2001).
Most bladder cancers (90%) are urothelial (transitional cell) carcinomas, often arising on the surface of small papillary tumours. The remaining 10% comprise mainly squamous cell carcinomas and adenocarcinomas. European recommendations for coding bladder tumours changed in 1995 to exclude some urothelial papillary tumours of the bladder that would previously have been classified as invasive (Pheby et al, 1995). Similar recommendations were implemented by UK cancer registries, but only for tumours registered from 2000 (UK Association of Cancer Registries, 2004).
Tobacco smoking confers a two- to three-fold risk. Occupational exposure to aromatic amines and polycyclic aromatic hydrocarbons in the chemical, rubber, transport and dye industries can also cause bladder cancer. Chemotherapeutic drugs, such as cyclophosphamide, and therapeutic irradiation of the pelvic region also increase the risk (Silverman et al, 2006).
Haematuria or pain often leads to the diagnosis. Approximately 75% of tumours are superficial (confined to the bladder mucosa or submucosal layer without muscle invasion) and these can be treated by transurethral resection. Tumours that have invaded muscle may require open surgery and a combination of intravesical or systemic chemotherapy, immunotherapy and radiotherapy.
Approximately 12% (21000) of all the bladder tumours registered in England and Wales during the period 1986–1999 were coded as in situ (behaviour code 2) or uncertain if benign or malignant (behaviour code 1). The proportion of tumours in these categories has risen very sharply, from virtually zero during 1986–1990 (Coleman et al, 1999) to approximately 23% during the 1990s (data not shown). Large regional variations were also seen. These tumours were considered ab initio as ineligible for inclusion in the survival analyses.
Of the 160614 eligible patients with a tumour explicitly coded as a malignant, invasive (behaviour code 3), primary cancer of the bladder, some 5% were excluded because their recorded duration of survival was zero (date of diagnosis same as date of death). Some of these patients will have been diagnosed on the day of death, but in many cases the cancer registration was based solely on the death certificate; hence, the date of diagnosis was unknown. It was not possible to distinguish these cases reliably in the available data; hence, all patients with ‘zero survival' were excluded. The proportion of patients excluded for this reason was the same for all deprivation groups (data not shown). Only patients for whom the bladder cancer was the first primary malignancy were retained in the analyses; 5% of patients known to have had a previous primary cancer at some time since 1971 were excluded. Finally, 1.6% of patients were excluded because their vital status was unknown on 5 November 2002, when the data were extracted for analysis. In all, more than 141500 patients were included in the analyses (88% of those eligible).