Summary The incidence of colorectal cancer was compared in nine towns in England and Wales, chosen to encompass a range of socio-economic conditions and spread of latitude in the country. Cases were ascertained through pathology records, supplemented by clinical notes. The pattern of variation in incidence was different for men and women. Among men incidences were highest in towns with better socio-economic conditions, whereas among women the trend was reversed. This supports the hypothesis that the dominant aetiological influences causing colorectal cancer differ in the two sexes. Mortality rates did not correlate closely with incidence. This, together with the markedly different patterns of incidence of colorectal cancer and appendicitis in the nine towns, casts doubt on the significance of a reported inverse correlation between regional mortality from colonic cancer and the consumption of pentosic fibre.Geographical variation in the incidence of colorectal cancer within Britain offers a method of exploring the role of dietary influences in the aetiology of the disease. There is a reported inverse correlation between mortality from colonic, but not rectal cancer and the consumption of pentosic dietary fibre (Bingham et al., 1979). However, this correlation was based on mortality data by region. Survival in the disease is comparatively high, around 50 per cent one year after diagnosis and 30 per cent at 5 years, and varies from one part of the country to another (H.M.S. O., 1980). Mortality is therefore a varying underestimate of incidence. The value of regional analyses is also limited by regions being large and heterogeneous geographical units. This paper describes the incidence of colorectal cancer in nine British towns, as determined from the most detailed and complete data source available, that is pathology records supplemented where necessary by information from case notes.
MethodsThe towns were selected to encompass a range of socio-economic conditions and latitude in England and Wales. The method of their selection has been described elsewhere (Barker et al., 1979). In summary the 83 largest county boroughs in England and Wales were classified into three equal groups having "better", "intermediate", and "worse" social and economic conditions. This classification was effected using a range of intercorrelated social and economic variables. The county boroughs were also divided into three groups according to latitude. December, 1980 were included. To conform with previous surveys residence was defined using the county borough boundaries in force before the 1974 boundary changes. Incidence rates were calculated using the 1981 census data adjusted to the pre-1974 boundaries. The rates were directly age-sex standardised using the total population of the nine towns as the standard: the same standard was used in the calculation of sex specific rates. Because the diagnosis of colorectal cancer is often not confirmed histologically in elderly patients, and in accord with other studies, this survey was confin...