Suicide attempts and externalizing psychopathology in a nationally representative sample Amber L. Hills T , Brian J. Cox, Lachlan A. McWilliams, Jitender Sareen Department of Psychiatry, University of Manitoba, Winnipeg, MB, Canada R3E 3N4 Abstract Suicide is most often associated with internalizing disorders such as depression; however, recent evidence suggests that externalizing psychopathology (substance dependence disorders, antisocial personality disorder) may have an independent relationship with suicidal behavior. The aim in the present study was to examine the relationship between lifetime suicide attempts and lifetime externalizing psychopathology in the US National Comorbidity Survey data set (n = 5877). First, hierarchical regression was performed to explore the associations between internalizing and externalizing disorders and suicide attempts. Externalizing psychopathology was significantly associated with lifetime suicide attempts (adjusted odds ratio = 3.47; P b .001) and significantly improved the model beyond that including only the sociodemographic variables and internalizing psychopathology (v 2 difference = 73.12; df = 1; P b .001). A second logistic regression was used to investigate the association between specific patterns of psychopathology and suicidality. Externalizing disorders were significantly associated with suicide attempts even in the absence of internalizing disorders (adjusted odds ratio = 5.98; 95% confidence interval = 3.07-11.67; P b .001). These findings add to the growing literature that suggests that externalizing psychopathology is an important psychiatric correlate of suicidal behavior. D 2005 Elsevier Inc. All rights reserved. 1. Introduction Suicide was the second leading cause of death in the United States in 2000 for people 25 to 34 years old and the third leading cause of death for people 15 to 24 years old [1,2]. Identifying psychiatric factors uniquely associated with suicidal behavior is important to improve the ability to predict suicide risk and to facilitate effective and timely mental health service delivery. Using data from the US National Comorbidity Survey (NCS), Kessler et al [3] completed one of the largest studies on the prevalence and correlates of suicidal behavior. The lifetime prevalence of suicide attempts in this nationally representative sample was 4.6%. Among the sociodemo- graphic variables assessed in the NCS, Kessler et al found that suicide attempts were strongly associated with being a woman, having been previously married, being born in a recent cohort (eg, year of birth between 1966 and 1975), and having a low education level. One of the most robust correlates of suicide attempts identified has been psycho- pathology [4 -9]. Kessler et al found the association between suicide attempts and psychopathology to be substantially higher for mood disorders than for any other disorder. For example, the odds ratio for major depression was 11.0 compared with 5.7 for antisocial personality disorder and 5.6 for panic disorder. Thus, it is not surprising that investigations into psychiatric correlates of suicide have typically focused on internalizing psychopathology such as unipolar depression [4,8,10 -13]. An association between externalizing disorders (sub- stance dependence disorders and antisocial personality disorder) and suicidal behavior has been demonstrated in adolescent and adult populations [5,14 -20]. Verona et al [20] used a recently developed framework for understanding psychopathology identified by Krueger [21]. By applying confirmatory factor analysis to the wide range of individual diagnoses assessed in the NCS, Krueger identified an internalizing factor and an externalizing factor. He also found that the internalizing factor was composed of 2 sub- factors: anxious-misery disorders and fear disorders. Verona et al [20] examined the effects of unique and comorbid internalizing and externalizing psychopathology in relation to suicide attempts. They used a large community sample that extended previous research limited to selective samples of 0010-440X/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.comppsych.2005.01.004 T Corresponding author. PZ -430 PsychHealth Centre, Winnipeg, MB, Canada, R3E 3N4. Tel.: +1 204 787 5166; fax: +1 204 787 4879. E-mail address: ahills@exchange.hsc.mb.ca (A.L. Hills). Comprehensive Psychiatry 46 (2005) 334 – 339 www.elsevier.com/locate/comppsych