Risk for cancer in a cohort of patients hospitalized for schizophrenia in Denmark, 1969–1993 Susanne Oksbjerg Dalton a, * , Lene Mellemkj&r a , Lars Thomassen a , Preben B. Mortensen b , Christoffer Johansen a a The Danish Cancer Society, The Institute of Cancer Epidemiology, Strandboulevarden 49, DK-2100 Copenhagen, Denmark b National Centre for Registerbased Research, Aarhus University, Taasingegade 1, DK-8000 Aarhus, Denmark Received 17 May 2004; received in revised form 16 November 2004; accepted 18 November 2004 Available online 13 December 2004 Abstract We investigated the cancer risk of patients hospitalized for schizophrenia in a nationwide cohort study. All 22766 adults admitted for schizophrenia, ICD-8 295, in Denmark between 1969 and 1993 were followed up for cancer through 1995. The incidence of site-specific cancers was compared with national incidence rates, adjusted for sex, age and calendar time. The risk for cancer was increased for both men and women during the first year of follow-up. When the first year of follow-up was excluded, the risk for all tobacco-associated cancers and for prostate and rectal cancers was reduced for male patients with schizophrenia. The standardized incidence ratio (SIR) of lung cancer was marginally reduced (SIR, 0.86; 95% CI: 0.65, 1.02) for male patients with schizophrenia; this was due, however, to a reduction in risk for older patients. An increased risk for breast cancer found for female patients with schizophrenia (SIR, 1.20; 95% CI: 1.05, 1.38) should be interpreted with caution, given the high proportion of nulliparous women with schizophrenia in Denmark. The data might support reduced risks for prostate and rectal cancer among male patients with schizophrenia, whereas a changing smoking pattern might explain the reduced risk for tobacco-related cancers. D 2004 Elsevier B.V. All rights reserved. Keywords: Schizophrenia; Cancer; Cohort studies A possible negative association between schizo- phrenia and cancer was discussed and investigated throughout the 20th century. In earlier studies on cancer risk, cancer mortality rates were used to evaluate the hypothesis, but these studies were often hampered by methodological problems, and the question remained unresolved (Fox and Howell, 1974; Baldwin, 1979; Fox, 1995). Six more recent studies of this association were based on incidence rates (Dupont et al., 1986; Nakane and Ohta, 1986; Mortensen, 1989, 1994; Gulbinat et al., 1992; 0920-9964/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2004.11.009 * Corresponding author. Tel.: +45 3525 7500; fax: +45 3525 7734. E-mail address: sanne@cancer.dk (S.O. Dalton). Schizophrenia Research 75 (2005) 315 – 324 www.elsevier.com/locate/schres