Summary The risks of second primary cancer were analysed in 2846 patients with Hodgkin's disease treated within the British National Lymphoma Investigation during 1970-87. The relative risk (RR) of leukaemia was significantly greater in women (RR = 30.1; 95% confidence limits (CL) 13.0-59.5) than in men (RR = 10.9; 95% CL 4.7-21.5), and showed a significant trend of greater risk with younger age at first treatment (P<0.001). The relative risk of solid cancers was similar between the sexes, but again significantly greater at young than at older ages of first treatment (P<0.01). Non-Hodgkin's lyrrphoma relative risks, although not related to sex or age, were significantly related to histology of the o;:iginal Hodgkin's disease, and were greatest after lymphocyte predominant Hodgkin's disease (RR = 55.6; 95% CL 18.0 -129.7). The relative risk of second cancers did not vary significantly according to whether or not splenectomy had been performed. Leukaemia risk was non-significantly greater after splenectomy than with no splenectomy, which accorded with previous evidence of a modest increased risk associated with this operation.If the greater relative risk of solid second cancers after treatment at young than at older ages persists with longer follow-up, the incidence rates of these second primaries in patients treated young for Hodgkin's disease will become very substantial as they age. This emphasises the need to maintain long-term follow-up surveillance of young Hodgkin's disease patients apparently cured of their disease, and to continue to develop new less carcinogenic treatment regimes.An increased risk of second cancers after modern intensive treatment for Hodgkin's disease is well established (Kaldor et al., 1987). After radiotherapy there are raised risks of several solid tumours (Boivin & O'Brien, 1988), and probably also of leukaemia (Boivin et al., 1984; Tucker et al., 1988;Van Leeuwen et al., 1989). After chemotherapy, a very high risk of acute or non-lymphocytic leukaemia has been shown (Pedersen-Bjergaard et al., 1987; Tucker et al., 1988), and in a recent study we found solid cancer risk also to be increased (Swerdlow et al., 1992). The risks of these second primary malignancies in relation to host factors, and subtype and stage of Hodgkin's disease, are unclear. Investigation of this needs data sets with large numbers of patients, and the results which have been published have been inconsistent (Tucker et al., 1988; Van der Velden et al., 1988;Kaldor et al., 1990;Henry-Amar et al., 1990). We present here analyses of these factors in data from the British National Lymphoma Investigation (BNLI), a long-established, large clinical collaborative investigation whose files have detailed treatment data and virtually complete follow-up information.
Materials and methodsThe BNLI is a collaborative group of over 60 participating centres in the UK, which has collected detailed data on the diagnosis, therapy and follow-up of lymphoma patients treated at the centres since 1970. Most of the patients were in rand...