To study the role of tobacco smoking and alcohol drinking in the etiology of non-Hodgkin's lymphoma (NHL), we conducted a multicenter case-control study in Spain, France, Germany, Italy, Ireland and Czech Republic between 1998 and 2004, which included 1,742 cases of NHL and 2,465 controls matched on age, sex and recruitment area. Tobacco smoking was not associated with the risk of NHL overall or with risk of specific histological subtypes. Similarly, there was no association between alcohol drinking and the risk of NHL overall or across histological subtypes. However, a protective effect of alcohol drinking was observed among men (OR 5 0.76, 95% CI 5 0.62-0.93) and in non-Mediterranean countries (OR 5 0.73, 95% CI 5 0.61-0.86). There was no evidence of interaction between alcohol drinking and tobacco smoking in NHL etiology. The results of this large-scale European study did not support an association between tobacco and NHL and suggested a protective effect of alcohol on development of NHL for men and in non-Mediterranean countries. ' 2006 Wiley-Liss, Inc.Key words: lymphoma; non-Hodgkin's; tobacco; alcohol; EuropeThe incidence of non-Hodgkin's lymphoma (NHL) has increased considerably in the last few decades in all countries with reliable information. 1 In Europe, between the early and the late 1990s, the rise in mortality was 5% for men and 8% for women. 2 There is no clear explanation of this phenomenon. To date, the established risk factors for NHL are autoimmune diseases, 3 infection with human immunodeficiency virus and human T-cell leukemia/lymphoma virus, 4 but these factors explain only a small proportion of the etiology of NHL. The role of tobacco smoking and alcohol drinking in the development of NHL remains uncertain partly due to the small sample sizes of previous epidemiologic studies. Recently, the International Lymphoma Epidemiology Consortium (Interlymph) has reported results based on pooled data of studies from Europe, 32-34 USA 22,28,39,45 and Australia. 46 In this analysis based on 6,492 cases of NHL and 8,683 controls, a slight inverse association was found between alcohol intake and the development of NHL. 47 The analysis of the effect of tobacco smoking, based on 6,594 cases of NHL and 8,892 controls, reported a slightly increased risk of NHL overall and of B-cell follicular lymphoma in particular. 48 Thus, we have attempted to confirm the results of the pooled analysis using data from an independent European multicenter case-control study (Epilymph).
Material and methodsOur investigation was conducted in 7 European countries (Czech Republic, Finland, France, Germany, Ireland, Italy, and Spain), but data from Finland were not included in this analysis because of methodological issues, in particular, controls were not matched to cases on age, sex and residence area, as done in other study centers, and the present study is based on data from 6 countries.Histologically or cytologically confirmed incident cases of lymphoma were recruited from 1998 to 2004 (Table I). Cases were classified ac...