Ultraviolet light, autoimmune disorders and the etiology of malignant lymphomas

Abstract: Malignant lymphomas constitute a clinically and morphologically diverse group of malignancies that may also differ etiologically. The incidence of the most common non-Hodgkin lymphomas (NHL) has increased dramatically worldwide during the past decades. Established risk factors together only explain a minority of the cases, let alone the NHL increase. Observations of a positive link between skin cancer and NHL have fostered the hypothesis that frequent sun exposure could be a risk factor for both malignancies and a contributing factor to the rise in NHL incidence. The aim of this thesis was to test the hypothesis of a positive association between sun exposure and malignant lymphomas, and to evaluate the role of autoimmune and chronic inflammatory disorders, especially rheumatoid arthritis (RA) and celiac disease, in the development of malignant lymphoma subtypes, and lymphomagenic mechanisms in this context. We performed a population-based case-control study in all of Denmark and Sweden (the Scandinavian Lymphoma Etiology, or SCALE, study) between 1999 and 2002, including 3,740 patients with malignant lymphomas and 3,187 control subjects aged 18 to 74 years. Based on structured telephone interviews, information was collected on history of ultraviolet (UV) light exposure, sun sensitivity, skin cancer history and other potential risk factors. In contrast with the a priori hypothesis, frequent sun exposure in Denmark/Sweden and abroad and sun burns at different ages were associated with a statistically significant 30-40% reduction in risk of overall NHL, with clear indications of inverse doseresponse trends (all Ptrends, < .003). There was similar but weaker evidence of inverse associations for Hodgkin lymphoma (HL). Self-reported skin cancer history was associated with a doubling in risks of NHL and HL. To test the hypothesis that the established excess risk of lymphomas in RA is due to risk factors shared by both disorders, we undertook a retrospective registry-based cohort study of Swedish patients hospitalized with RA (n=76,527) and the first-degree relatives (n=70,290) of a subset of these patients. Relative risks of malignant lymphomas were assessed by matching the respective cohorts with the population-based cancer register. The RA patients had a doubled risk of malignant lymphomas overall, although the excess risk did not persist beyond 20 years of follow-up. Firstdegree relatives were generally not at increased risk of malignant lymphomas, and thus a prominent role of shared risk factors in RA-related lymphomagenesis was not supported. To evaluate the spectrum of lymphoma subtypes associated with celiac disease, we re-classified 56 malignant lymphoma cases occurring in a large cohort of patients previously hospitalized for celiac disease (n=11,650). Our results indicated that celiac disease patients are at increased risk, not only of the well-described enteropathy-type T-cell lymphoma, but also of non-intestinal T-cell lymphomas and the common B-cell lymphomas compared to the general population. Finally, we estimated relative risks of NHL overall and by subtype in association with several autoimmune and chronic inflammatory disorders, disease phenotype and treatment, using selfreported data in SCALE. We confirmed an increased risk of NHL in RA, systemic lupus erythematosus, primary Sjögren's syndrome and celiac disease, but not in type I diabetes, inflammatory bowel disorders, sarcoidosis or psoriasis. The first four disorders were all specifically associated with diffuse large B-cell lymphoma and with a few more uncommon NHL subtypes. Data suggested a tendency towards higher risks in severe and long-standing inflammation, but there was little to support previous notions of risk associated with medical treatments in these conditions.