ORIGINAL PAPER Cancer mortality in patients with psychiatric diagnoses: a higher hazard of cancer death does not lead to a higher cumulative risk of dying from cancer Ng Chong Guan • Fabian Termorshuizen • Wijnand Laan • Hugo M. Smeets • Nor Zuraida Zainal • Rene ´ S. Kahn • Niek J. De Wit • Marco P. M. Boks Received: 27 February 2012 / Accepted: 12 October 2012 Ó Springer-Verlag Berlin Heidelberg 2012 Abstract Purpose Both increased as well as decreased cancer mortality among psychiatric patients has been reported, but competing death causes were not included in the analyses. This study aims to investigate whether observed cancer mortality in patients with psychiatric disorders might be biased by competing death causes. Method In this retrospective cohort study on data from the Psychiatric Case Register Middle Netherlands linked to the death register of Statistics Netherlands, the risk of cancer death among patients with schizophrenia (N = 4,590), bipolar disorder (N = 2,077), depression (N = 15,130) and their matched controls (N = 87,405) was analyzed using a competing risk model. Results Compared to controls, higher hazards of cancer death were found in patients with schizophrenia (HR = 1.61, 95 % CI 1.26–2.06), bipolar disorder (HR = 1.20, 95 % CI 0.81–1.79) and depression (HR = 1.26, 95 % CI 1.10–1.44). However, the HRs of death due to suicide and other death causes were more elevated. Consequently, among those who died, the 12-year cumulative risk of cancer death was sig- nificantly lower. Conclusions Our analysis shows that, compared to the general population, psychiatric patients are at higher risk of dying from cancer, provided that they survive the much more elevated risks of suicide and other death causes. Keywords Cancer Á Mortality Á Psychiatric diagnoses Á Competing risk Introduction Over a century, the relationship between cancer and psy- chiatric disorders has attracted interest among researchers. The first report dates back to 1909 when the Commis- sioners in Lunacy for England and Wales reported lower cancer mortality among prison inmates [1]. Whether this finding indicates a causal relationship between mental disorders and protection against developing cancer remains a matter of debate [2–6]. Different factors might confound the comparison between psychiatric patients and the gen- eral population, such as the risk of other death causes and differences in life style and factors related to psychiatric treatments [7]. These factors may mask the association between the presence of a psychiatric disorder and cancer risk [8]. Furthermore, it remains unclear whether altered cancer mortality among psychiatric patients is specific for schizophrenia, with mixed results found in patients with schizophrenia [2–5], bipolar disorder [5, 9, 10], and depression [11, 12]. An important methodological chal- lenge when studying the association of psychiatric disor- ders with cancer mortality is the potential bias due to other competing death causes. A conventional Kaplan–Meier survival analysis may overestimate the percentage of N. C. Guan Á N. Z. Zainal Faculty of Medicine, Department of Psychological Medicine, University of Malaya, Malaya, Malaysia N. C. Guan Á F. Termorshuizen (&) Á W. Laan Á H. M. Smeets Á N. J. De Wit Julius Center for Health Sciences and Primary Care, University Medical Center (UMC) Utrecht, PO Box 85500, 3508 GA Utrecht, The Netherlands e-mail: F.Termorshuizen@umcutrecht.nl R. S. Kahn Á M. P. M. Boks Department of Psychiatry, Rudolf Magnus Institute of Neuroscience, University Medical Center (UMC) Utrecht, PO Box 85500, 3508 GA Utrecht, The Netherlands 123 Soc Psychiatry Psychiatr Epidemiol DOI 10.1007/s00127-012-0612-8