Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Coffee, decaffeinated coffee, tea, and pancreatic cancer risk: a pooled-analysis of two Italian case–control studies Federica Turati a,b , Carlotta Galeone a,b , Renato Talamini d , Silvia Franceschi f , Marco Manzari e , Gianfrancesco Gallino c , Jerry Polesel d , Carlo La Vecchia a,b and Alessandra Tavani a To evaluate the association between coffee, decaffeinated coffee, and tea consumption and pancreatic cancer risk in a pooled analysis of two Italian case–control studies, between 1983 and 2008, we conducted two case–control studies in Northern Italy, including a total of 688 pancreatic cancer cases and 2204 hospital controls with acute, non-neoplastic diseases. We computed multivariate odds ratios (ORs) and 95% confidence intervals (CIs) for coffee drinking (mostly espresso and mocha), adjusting for age, sex, center, year of interview, education, body mass index, tobacco smoking, alcohol drinking, and diabetes. Compared with coffee nondrinkers, the multivariate OR for coffee drinkers was 1.34 (95% CI: 1.01–1.77). However, there was no trend in risk with respect to dose and duration. The OR for an increment of one cup per day was 1.05 (95% CI: 0.98–1.11). There was no heterogeneity in strata of age, sex, and other covariates, including tobacco smoking. No association emerged for decaffeinated coffee (for drinkers the OR was 0.87, 95% CI: 0.60–1.26, compared with decaffeinated coffee nondrinkers) or tea (for tea drinkers the OR was 0.92, 95% CI: 0.75–1.14). The lack of relationship with dose and duration weighs against a causal association between coffee and pancreatic cancer, which is in agreement with most evidence on the issue. European Journal of Cancer Prevention 00:000–000  c 2011 Wolters Kluwer Health | Lippincott Williams & Wilkins. European Journal of Cancer Prevention 2011, 00:000–000 Keywords: case–control study, coffee, decaffeinated coffee, hot beverages, pancreatic cancer, risk factors, tea a Istituto di Ricerche Farmacologiche ‘Mario Negri’, b Dipartimento di Medicina del Lavoro, Universita ` degli Studi di Milano, c Fondazione IRCCS, Istituto Nazionale dei Tumori di Milano, Milan, d Unit of Epidemiology and Biostatistics, Centro di Riferimento Oncologico, IRCCS, Aviano (PN), e Dipartimento di Traumatologia, Ortopedia e Medicina del Lavoro, Universita ` degli Studi di Torino, Turin, Italy and f International Agency for Research on Cancer, Lyon, Cedex, France Correspondence to Professor Carlo La Vecchia, MD, Istituto di Ricerche Farmacologiche ‘Mario Negri’, Via G. La Masa 19, 20157 Milan, Italy Tel: + 39 2 3901 4527; fax: + 39 2 3320 0231; e-mail: carlo.lavecchia@unimi.it Received 22 December 2010 Accepted 11 January 2011 Introduction The relation of coffee consumption with pancreatic cancer risk has long been investigated since a case– control study in the early 1980s showed a strong positive association (MacMahon et al., 1981). Since then, 17 cohort studies have been published. Two of these found a significant increased risk (Harnack et al., 1997; Lin et al., 2002) and one a significant decreased risk (Isaksson et al., 2002) for the highest versus the lowest coffee drinking categories. The remaining 14 cohort studies found relative risks close to unity (Nomura et al., 1981; Whittemore et al., 1983; Snowdon and Phillips, 1984; Jacobsen et al., 1986; Hiatt et al., 1988; Mills et al., 1988; Zheng et al., 1993; Shibata et al., 1994; Stensvold and Jacobsen, 1994; Michaud et al., 2001; Stolzenberg- Solomon et al., 2002; Khan et al., 2004; Luo et al., 2007; Nilsson et al., 2010). Of the 35 case–control studies (Jick and Dinan, 1981; Kessler, 1981; Goldstein, 1982; Severson et al., 1982; Wynder et al., 1983, 1986; Kinlen and McPherson, 1984; Gold et al., 1985; Hsieh et al., 1986; Mack et al., 1986; Norell et al., 1986; Falk et al., 1988; Gorham et al., 1988; Clavel et al., 1989; Cuzick and Babiker, 1989; Olsen et al., 1989; Farrow and Davis, 1990; Ghadirian et al., 1991; Jain et al., 1991; Bueno de Mesquita et al., 1992; Lyon et al., 1992; Mizuno et al., 1992; Stefanati et al., 1992; Friedman and Van den Eeden, 1993; Kalapothaki et al., 1993; Sciallero et al., 1993; Zatonski et al., 1993; Gullo et al., 1995; Partanen et al., 1995; Kokic et al., 1996; Nishi et al., 1996; Silverman et al., 1998; Mori et al., 1999; Alguacil et al., 2000; Villeneuve et al., 2000), seven found a significant positive association between the highest category of coffee intake and pancreatic cancer risk (Hsieh et al., 1986; Mack et al., 1986; Clavel et al., 1989; Lyon et al., 1992; Gullo et al., 1995; Kokic et al., 1996; Mori et al., 1999), in the absence, however, of a dose–risk relation in most of them. Thus, it is possible that the observed positive association is not causal but due to selection bias, residual confounding with cigarette smoking (the major recognized risk factor for pancreatic cancer) (Anderson et al., 2006), or other sources of bias. The odds ratios (ORs) were below unity in approximately 12 case–control studies (Jick and Dinan, 1981; Kessler, 1981; Kinlen and McPherson, 1984; Olsen et al., 1989; Ghadirian et al., 1991; Jain et al., 1991; Bueno de Mesquita et al., 1992; Mizuno et al., 1992; Friedman and Van den Eeden, 1993; Kalapothaki et al., 1993; Sciallero et al., 1993; Zatonski et al., 1993; Partanen et al., 1995; Silverman et al., Research paper 1 0959-8278  c 2011 Wolters Kluwer Health | Lippincott Williams & Wilkins DOI: 10.1097/CEJ.0b013e32834572e7