Epidemiological evidence for the 'two-disease hypothesis' in Hodgkin's disease.

This study relates to 1803 newly diagnosed cases of Hodgkin's Disease (HD) registered between 1984 and 1988 by a specialist leukaemia-lymphoma registry covering approximately half of England and Wales. In addition to registration data the analyses use routine census data. Rye classification is available for the majority of cases with only 7% being unclassified. The data confirm that the young adult peak in HD occurrence is attributable to the nodular sclerosing subtype (NS), and there is some evidence that NS has a different geographical distribution by county to the other Rye subtypes. Because of possible geographical biases in the classification, subsequent spatial analyses are disaggregated by age at diagnosis with particular emphasis being placed on the two age groups 0-34 and 50-79 years. For these, trends of risk by areal socioeconomic status are in opposite directions and the effects of urban-rural status while in the same direction also differ significantly. Allowance for these does not, however, explain the significant difference between the county distributions of the two groups. Patterns of spatial clustering are quite distinct with evidence for local spatial aggregation amongst younger cases.