Abstract
The incidence of subsequent primary cancers was assessed in relation to treatment for a cohort of 7,203 patients from the Birmingham and West Midlands Cancer Registry diagnosed between 1957 and 1976. The total of 213 cancers observed one or more years after treatment for ovarian cancer (mean follow-up = 6.5 person-years) represented a significant excess (observed (O) = 213, expected (E) = 140.07, relative risk (RR) = 1.5, 95% CI 1.3-1.7, P less than 0.001). Among patients whose treatment included chemotherapy (CT), with or without radiotherapy (RT), the risk of acute and non-lymphocytic leukaemia (A + NLL) was significantly increased (O = 5, E = 0.18, RR = 27.8, 95% CI 9.0-64.8, P less than 0.001). The relative risks of A + NLL following RT without CT (RR = 4.5) and after other treatments (RR = 2.9) were not significantly in excess of 1.0. Significant excesses of subsequent cancers were observed at several sites: breast (RR = 1.7, 95% CI 1.3-2.2), lung (RR = 2.0, 95% CI 1.3-3.4), colon and rectum (RR = 1.6, 95% CI 1.1-2.3), urinary system (RR = 1.9, 95% CI 0.9-3.7), nervous system (RR = 3.3, 95% CI 1.2-7.3) and connective tissue (RR = 6.7, 95% CI 1.8-17.1) but the relationship with type of treatment was not so clearly defined as that for leukaemia. Although the treatment groups were broad and based on routinely collected data, they can enhance the use of cohort analyses for exploratory and monitoring purposes.