Journal of Affective Disorders 156 (2014) 119–125

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research report

Risk factors for suicide among 34,671 patients with psychotic and non-psychotic severe depression Anne Katrine K. Leadholm a,d,n, Anthony J. Rothschild b, Jimmi Nielsen c, Per Bech d, Søren D. Østergaard a,e,f a

Unit for Psychiatric Research, Aalborg Psychiatric Hospital, Aalborg University Hospital, Aalborg, Denmark University of Massachusetts Medical School and UMass Memorial Health Care, Worcester, MA, USA c Center for Schizophrenia, Aalborg Psychiatric Hospital, Aalborg University Hospital, Aalborg, Denmark d Psychiatric research Unit, Psychiatric Center North Zealand, University of Copenhagen, Hillerød, Denmark e Institute of Clinical Medicine, Aarhus University Hospital, Aarhus, Denmark f Depression Clinical and Research Program, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA b

art ic l e i nf o

a b s t r a c t

Article history: Received 8 August 2013 Received in revised form 25 November 2013 Accepted 3 December 2013 Available online 18 December 2013

Background: Severe unipolar depression is associated with increased risk of suicide, but it remains unknown whether the same risk factors are present in the non-psychotic (non-PD) and psychotic (PD) subtypes respectively. Therefore, this study aimed to identify risk factors for suicide in non-PD and PD separately, and to investigate if the presence of psychotic symptoms is an independent risk factor for suicide in severe depression. Methods: This register-based, nationwide, historical prospective cohort study used logistic regression analyses to ascertain risk factors for suicide among all adults diagnosed with severe depression at Danish psychiatric hospitals between January 1, 1994 and December 31, 2010. The risk for suicide was expressed as adjusted odds ratios (AOR). Results: A total of 34,671 individuals with severe depression (non-PD: n ¼ 26,106 and PD: n ¼12,101) were included in the study. Of these, 755 completed suicide during follow up. PD was not found to be an independent risk factor for suicide in severe depression (AOR ¼0.97 [0.83–1.15]). Older age (non-PD AOR¼ 1.05 [per year], PD AOR ¼1.04 [per year]), male sex (non-PD AOR ¼1.89, PD AOR ¼1.98), and a previous incident of self-harm (non-PD AOR ¼5.02, PD AOR ¼ 5.17) were significant risk factors for both groups. Limitations: As the study population was comprised only of patients with contact to psychiatric hospitals, the results cannot be extrapolated to the primary care setting. Conclusion: The following risk factors for non-PD and PD were identified: older age, male gender and previous incidents of self-harm. In suicide prevention efforts, equal attention should be paid to non-PD and PD patients. & 2013 Elsevier B.V. All rights reserved.

Keywords: Depressive disorder Affective disorders Psychotic Suicide Suicide prevention Register study

1. Introduction Every year, approximately 1 million people complete suicide worldwide (World Health Organization, 2011). Psychological autopsy studies suggest that at least 90% of the victims fulfilled criteria for a mental disorder at the time of suicide (Cavanagh et al., 2003). There has therefore been considerable interest in identifying potential risk factors for suicide among people suffering from mental disorders (Mortensen et al., 2000; Oquendo et al., 1997). Knowing such risk factors can help identify patients n Corresponding author at: Unit for Psychiatric Research, Aalborg Psychiatric Hospital, Aalborg University Hospital, Mølleparkvej 10, DK-9000 Aalborg, Denmark. Tel.: þ 45 29 29 51 60; fax: þ45 72137235. E-mail address: [email protected] (A.K.K. Leadholm).

0165-0327/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jad.2013.12.003

at risk and elicit targeted preventive measures (Berman, 2006; Erlangsen et al., 2011; Madsen et al., 2012; Nordentoft et al., 2007; Nordentoft, 2007). Unipolar depression, particularly if severe, is associated with a significantly increased risk of suicide (Agerbo et al., 2001; Jeon et al., 2010, 2013; Qin and Nordentoft, 2005). Severe unipolar depression is subdivided into non-psychotic depression (non-PD) and psychotic depression (PD) in both the International Classification of Disease, 10th revision (ICD-10) (World Health Organization, 1993) and the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (American Psychiatric Association, 2000) (see Fig. 1). Patients with PD suffer from delusions and/or hallucinations, in addition to the symptoms of a depressive episode (Ostergaard et al., 2013b; Rothschild, 2009). A number of studies have

120

A.K.K. Leadholm et al. / Journal of Affective Disorders 156 (2014) 119–125

Fig. 1. Prevalence of depression and its severe subtypes.

identified significant differences, beyond the psychosis, between non-PD and PD (Ostergaard et al., 2012b). The differences concern heritability (Leckman et al., 1984), genetic and environmental risk factors (Domschke, 2013; Ostergaard et al., 2013a), symptomatology (Ostergaard et al., 2012a), treatment response (Coryell et al., 1996; Kessing, 2004; Leadholm et al., 2013; Loo et al., 2011), prognosis (Coryell et al., 1996) and mortality (Vythilingam et al., 2003). It has also been suggested that the risk of suicide is higher in PD than in non-PD (Park et al., 2010; Roose et al., 1983), but this has not been observed in all studies (Rothschild, 2009). Furthermore, the extensive literature on risk factors for suicide in depression has not reported whether the identified risk factors differ in quality and quantity between non-PD and PD respectively (Coryell and Young, 2005; Oquendo et al., 1997). At present, there has only been one study investigating markers associated with suicidality among patients with PD (Schaffer et al., 2008), and none examining risk factors for completed suicide. Previous studies examining the risk factors for completed suicide have focused particularly on risk factors easily observable in a clinical setting. These include: 1.1. Socio-demography In both the general population and among patients with mental disorders, suicide is more frequent among males than females (Large et al., 2011; Qin, 2011) and the risk is positively correlated with age (Qin, 2011). The relationship between educational level and suicide remains controversial. While some studies report that suicidality is not affected by years of education (Schaffer et al., 2008), others have found that both low education level (Wiktorsson et al., 2010) and completion of higher education (Wenzel et al., 2011) are associated with increased risk of suicide, depending on the population in question. Also, the nature of a person0 s affiliation to the labor market affects the risk

of suicide. Recent unemployment has been shown to increase the risk of suicide among the mentally ill (Hoyer et al., 2009). Furthermore, recipients of disability pension seem to be at increased risk of suicide (Ahs and Westerling, 2006). Solitary habitation is shown to be a risk factor for suicide, even when adjusted for age, sex and other demographic variables (Mortensen et al., 2000; Wiktorsson et al., 2010). Finally, institutionalization (long-term psychiatric and/or social care at a residential institution) has previously been shown to lower the risk of suicide among people with schizophrenia in Denmark (Uggerby et al., 2011). 1.2. Psychiatric morbidity Mental disorders co-morbid to depression have been reported to increase the risk of suicide (Bolton et al., 2010). This is the case for mental disorders due to substance abuse (Ganz and Sher, 2009), organic mental disorders (Haw et al., 2009), anxiety (Bolton et al., 2010), obsessive compulsive disorder (Torres et al., 2011), and personality disorders (Dumais et al., 2005). In studies examining mixed populations of psychiatric patients, the relation between psychotic symptoms and suicidality/suicide varies (Angst et al., 2005; Caetano et al., 2006; Grunebaum et al., 2001). Finally, a previous suicide attempt is a strong risk factor for later completed suicide (Wiktorsson et al., 2010). 1.3. Physical co-morbidity Physical illness increases the risk of depression (Goldberg and Harrow, 2010). This relationship is particularly strong for e.g. cardiovascular disease (Lippi et al., 2009), stroke (Cumming et al., 2010), diabetes (Kokoszka et al., 2009), chronic pulmonary disease (Goodwin, 2011; Putman-Casdorph and McCrone, 2009), and cancer (Spiegel, 1996). A number of studies have shown that

A.K.K. Leadholm et al. / Journal of Affective Disorders 156 (2014) 119–125

the co-morbidity between physical illnesses and depression is associated with a very high risk of suicide (Chen et al., 2010; Forsstrom et al., 2010; Larsen et al., 2010; Llorente et al., 2005). In the present study we aimed to identify risk factors for completed suicide among patients with severe depression in general, as well as in PD and non-PD. We also wished to investigate whether the psychotic subtype acts as an independent risk factor for suicide. Based on the literature, we focused on potential risk factors within the three aforementioned areas of interest, namely socio-demographics, psychiatric morbidity and physical co-morbidity.

2. Methods This study was designed as a population-based, nationwide, historical prospective cohort study and used data from a number of Danish registers, which are described below. The data for each subject in the sample was obtained by register linkage via the unique personal registration numbers, that are assigned to all Danish residents at the time of birth, or with the achievement of residency (Pedersen, 2011). Permission to use the registers was given by the Danish Data Protection Agency, The Danish National Board of Health and Statistics Denmark. Data from the Danish registers is rendered anonymous when used for research purposes. 2.1. Definition of sample In Denmark, all diagnoses assigned at psychiatric hospitals throughout the country are reported to the Danish Central Psychiatric Research Register (DPCRR), which contains complete electronic data from 1969 and onwards (Mors et al., 2011; MunkJorgensen and Ostergaard, 2011). The diagnoses are assigned at discharge by the treating psychiatrist taking into account clinical history/observation, laboratory tests, potential brain imaging, and treatment response. Hospital treatment is free of charge in Denmark (financed by taxes) and there are no private psychiatric hospitals in the country. Consequently, the DPCRR contains information regarding all patients treated for mental disorders at hospital-based settings. The diagnoses in the DPCRR were labeled according to the 8th International Classification of Diseases (ICD-8) (Health, 1971) until January 1, 1994, when the ICD-8 was replaced by the ICD-10 (the ICD-9 was never used in Denmark). Diagnoses from outpatient settings were included from January 1, 1995. In this study, the sample was defined as all Danish adults (Z 18 years of age) assigned with a diagnosis of unipolar severe depression, non-psychotic subtype (ICD-10: F32.2 or F33.2) or psychotic subtype (F32.3 or F33.3) registered in the DPCRR, between January 1, 1994 and December 31, 2010. We chose not to make these diagnoses mutually exclusive, as the same person may have received both diagnoses during the follow-up period. This ambiguity accurately reflects the patients seen in a clinical setting, where patients can have a history of both diagnoses. Patients with ICD-8/ICD-10 diagnosis of schizophrenia, schizoaffective disorder or bipolar disorder prior to the diagnosis of nonpsychotic or psychotic depression, were excluded from the analyses. The remaining population was followed from the date of onset of severe depression to either death by suicide, death by other causes than suicide, lost to follow-up, or December 31, 2010. 2.2. Definition of primary outcome The primary outcome of the study was the risk of suicide after an episode of non-PD or PD within the period between January 1, 1994 and December 31, 2010. Data on cause of death, including suicide, was obtained from the “Danish Cause of Death Register”,

121

which contains information on the date and cause (according to ICD-10 codes) of all deaths in Denmark in the study period (Juel and Helweg-Larsen, 1999). In accordance with previous studies, completed suicide was defined as a primary cause of death, labeled with the ICD-10 codes from X60 to X84 (Mortensen et al., 2000; Qin and Nordentoft, 2005). 2.3. Assessment of socio-demographics Data on age and sex was extracted from the Civil Registration System (Pedersen, 2011). Information on the highest level of education, habitation status (solitary or co-habitation), disability retirement was derived through the integrated database for Labor Market Research (IDA) (Statistics Denmark, 1991). Information regarding institutionalization was derived from the Danish Accommodation Database. Data from 2008 (latest update) or the last recording prior to death was used. 2.4. Assessment of psychiatric history A number of explanatory variables with potential relationship to suicide were defined based on the diagnostic codes recorded in the DPCRR. These included age at diagnosis of severe depression, lifetime number of psychiatric bed-days (dichotomized at medianvalue), a diagnosis of schizophrenia, schizoaffective disorder or bipolar disorder after initial episode of severe depression, as well as a lifetime history (from beginning of register to either death, lost to follow-up or December 31, 2010) of any of the following: organic mental disorder, mental disorder due to use of psychoactive substances, anxiety/obsessive–compulsive disorder, personality disorder, intentional self-harm and poisoning with analgesics. The two last variables were based on diagnoses recorded both in the DPCRR and the Danish National Patient Register (DNPR), which contains data on all medical diseases treated at somatic hospitals in Denmark from 1977 and onwards (Lynge et al., 2011). We chose to focus on analgesic poisoning, as this is a major public health concern and may be preventable by limiting over-the-counter sale (Nordentoft et al., 2007; Nordentoft, 2007; Qin et al., 2009). For the present study, DNPR diagnoses recorded until December 31, 2009 were available. 2.5. Assessment of physical illness Each subject0 s clinical history (from 1977 to 2009) regarding 17 diagnostic categories of medical diseases with known relation to depression were assessed through the DNPR: ischemic heart disease, chronic hepatitis, chronic renal disease, chronic lower pulmonary disease, CNS infection/inflammation, Cushing syndrome, cancer, dementia, diabetes, disorders of the thyroid gland, epilepsy, hypertension, intracranial injury, inflammatory bowel disease, Parkinson0 s Disease, rheumatoid arthritis and stroke (for exact definition of diagnostic categories see Ostergaard et al. 2013 (Ostergaard et al., 2013)). A dichotomous variable (1/0) covering physical illnesses was defined based on the aforementioned diagnoses. If a patient had been assigned with at least one of the diagnoses he/she was considered to be “physically ill” (a value of 1 on the variable). 2.6. Statistical analysis All statistical analyses were performed with STATA 12 at Statistics Denmark via remote access. Sample characteristics: Comparisons of mean age at diagnosis of severe depression and number of bed-days between groups were performed by Wilcoxon rank sum test. All between group comparisons of proportions were performed by chi-square tests. Assessment of risk factors for

122

A.K.K. Leadholm et al. / Journal of Affective Disorders 156 (2014) 119–125

suicide: In order to identify markers associated with suicide, univariate and multivariate logistic regression analyses considering a number of explanatory variables were carried out. Risk was expressed as Odds Ratios (OR) in the univariate analyses and as Adjusted Odds Ratios (AOR) in the multivariate analysis (mutually adjusted for all explanatory variables and for date of diagnosis of PD). Two-tailed tests with a significance level of 5% were used in all analyses. Bonferoni correction was not performed due to the exploratory nature of the study.

3. Results We identified 38,352 patients diagnosed with severe depression during the study period, of whom 2481 had a previous diagnosis of schizophrenia, schizoaffective disorder or bipolar disorder and were not included in the analyses. A further 1200 were excluded due to age under 18 at the time of diagnosis. Thus, a total of 34,671 individuals were included in the study. Of these, 12,150 had received a diagnosis of PD and 26,106 a diagnosis of non-PD (not mutually exclusive). During follow up 755 individuals completed suicide, corresponding to 2.1% of patients with non-PD, 2.3% of patients with PD, and 9.5% of all 7918 recorded deaths among subjects. The socio-demographic and clinical characteristics of both patients who completed suicide and those who did not are shown in Table 1. Table 1 also shows the results of the logistic regression analysis assessing the association of the socio-demographic and clinical characteristics with completed suicide in the total sample of patients with severe depression (both psychotic and non-psychotic). In both the univariate and multivariate analyses, male gender, non-fatal intentional self-harm, and older age were identified as

significant risk factors for suicide. Receiving disability pension, living in an institution and suffering from a physical illness was associated with a decreased risk of suicide among patients with severe depression. Notably, PD was not found to be an independent risk factor for suicide. Since complete data regarding sociodemographic variables were missing for 3662 subjects in the total sample, the multiple logistic regression analysis only included the remaining 31,009 individuals. When patients with severe depression were stratified by psychotic and non-psychotic sub-type, a few differences appeared between the two groups (Table 2). Only PD patients had a lower risk of committing suicide if suffering from an organic mental disorder. For the non-PD patients, psychiatric bed days above median and previous analgesia poisoning were significant risk factors for suicide.

4. Discussion Through a register linkage study we have examined potential risk factors for suicide in more than 30,000 Danes diagnosed with unipolar severe depression between 1994 and 2010. This study is the largest to examine the psychotic subtype as an independent risk factor for suicide in severe depression and the first to investigate risk factors for suicide in non-PD and PD separately. We found that 9.5% of the participants, who died during follow-up, died by suicide. This proportion is much higher than the 1.3% (12,337 of 982,983) of deaths attributed to suicide in the general Danish population in the same period according to the Cause of Death Register (Statistics Denmark; Juel and Helweg-Larsen, 1999). One of the main aims of this study was to determine whether the popular clinical assumption that the psychotic subtype is

Table 1 Prevalence of potential risk factors for suicide and their association with completed suicide among patients with severe depression. Variable

Controls (no suicide) (n¼ 33,916)

Cases (suicide)(n¼ 755)

OR (n¼ 34,617)

AORa (n ¼31009)

Non-fatal intentional self-harm (%) Male gender (%) Hospital contact due to poisoning with analgesics (%) Psychotic episodes other than schizophrenia (%) Psychiatric bed-days above median (%) Mood-incongruent psychosis (%) Mental disorder due to psychoactive substances (%) Age at diagnosis of severe depression (years) Schizophrenia after severe depression (%) Unipolar psychotic depression (%) Schizoaffective disorder after severe depression (%) Education level above public school (%) Personality disorder (%) Anxiety/OCD (%) Living alone (%) Bipolar disorder after severe depression (%) Organic mental disorder (%) Living in an institution (%) Suffering from physical illness (%) Receiving disability pension (%)

9.5 (9.2–9.8) 36.8 (36.3–37.4) 6.5 (6.2–6.7) 12.2 (11.9–12.6) 49.9 (49.4–50.6) 0.9 (0.8–1.0) 18.6 (18.2–19.0) 51.2 (SD 18.9) 3.7 (3.5–3.9) 35.0 (34.5–35.5) 1.5 (1.4–1.7) 58.9 (58.3–59.4) 24.7 (24.2–25.1) 40.6 (40.1–41.1) 78.4 (78.0–78.9) 7.1 (6.8–7.4) 12.3 (11.9–12.6) 10.8 (10.5–11.2) 54.8 (54.2–55.3) 51.5 (51.0–52.1)

33.4 (30.0–36.8)nnn 54.4 (50.9–58.0)nnn 10.9 (8.6–13.0)nnn 15.8 (13.2–18.4)nn 61.1 (57.6–64.6)nnn 1.2 (0.4–2.0) 24.8 (21.9–27.9)nnn 53.5 (SD 15.6)nnn 2.7 (1.5–3.8) 37.1 (33.6–40.5) 1.6 (0.7–2.5) 61.8 (58.2–65.4) 26.1 (23.0–29.2) 37.9 (34.4–41.4) 68.8 (65.5–72.1)nnn 7.0 (5.2–8.9) 7.4 (5.5–9.3)nnn 4.9 (3.3–6.3)nnn 42.5 (39.0–46.1)nnn 40.8 (37.3–44.3)nnn

4.9 (4.1–5.6)nnn 2.1 (1.8–2.4)nnn 1.8 (1.4–2.2)nnn 1.3 (1.1–1.6)nn 1.6 (1.4–1.8)nnn 1.3 (0.7–2.6) 1.4 (1.2–1.7)nnn 1.0 (1.0-1.0)nnb 0.7 (0.5–1.1) 1.1 (0.9–1.3) 1.1 (0.6–1.9) 1.1 (1.0–1.3) 1.9 (0.9–1.3) 0.9 (0.8–1.0) 0.6 (0.5–0.7)nnn 1.0 (0.8–1.3) 0.6 (0.4–0.8)nnn 0.4 (0.3–0.6)nnn 0.6 (0.5–0.7)nnn 0.7 (0.6–0.8)nnn

5.1 (4.3–6.1)nnn 1.9 (1.6–2.2)nnn 1.5 (1.1–1.9)nn 1.4 (1.1–1.7)nn 1.3 (1.1–1.6)nn 1.2 (0.6–2.5) 1.2 (1.0–1.4) 1.1 (1.0–1.1)nnnb 1.0 (0.6–1.6) 1.0 (0.8–1.2) 1.0 (0.5–1.8) 1.0 (0.8–1.2) 1.0 (0.8–1.2) 0.9 (0.8–1.1) 0.8 (0.7–1.0) 0.8 (0.6–1.1) 0.6 (0.5–0.9)nnn 0.6 (0.4–0.9)nn 0.5 (0.4–0.6)nnn 0.3 (0.2–0.3)nnn

Comparison between controls (non-psychotic and psychotic severely depressed who did not complete suicide) and cases (non-psychotic and psychotic severely depressed who completed suicide during follow-up). Numbers in brackets represent 95% confidence intervals or standard deviation (SD). Analyses used for comparison between groups: chi-square tests for dichotomous variables and the two-sample Wilcoxon rank-sum (Mann–Whitney) test for difference in age. Analysis of association between potential risk factors and completed suicide: univariate logistic regression with risk expressed as odds ratio (OR) and multivariate logistic regression with risk expressed as adjusted odds ratio (AOR). n

po 0.05. p o0.01. nnn p o0.001. a Including adjustment for date of diagnosis. Due to missing data, there were only 33,841 controls and 753 cases for the univariate analyses concerning living alone, living in an institution, receiving disability pension, and 30,308 controls and 702 cases for the univariate analysis concerning education level above primary school. The multiple logistic regression analysis was based on the 31,009 individuals, with complete data. b Odds-ratio per year. nn

A.K.K. Leadholm et al. / Journal of Affective Disorders 156 (2014) 119–125

123

Table 2 Risk factors for suicide among patients with non-psychotic depression (non-PD) and psychotic depression (PD) respectively. Variable

Non-fatal intentional self-harm Male gender Hospital contact due to poisoning with analgesics Mental disorder due to psychoactive substances Psychotic episodes other than schizophrenia Psychiatric bed-days above median Mood-incongruent psychosis Age at diagnosis of severe depression Schizophrenia after severe depression Schizoaffective disorder after severe depression Education level above public school Personality disorder Bipolar disorder after severe depression Living alone Anxiety/OCD Organic mental disorder Living in an institution Suffering from physical illness Receiving disability pension

AORa

OR Non-PD (n¼ 26,106)

PD (n ¼12,150)

Non-PD (n¼ 23,808)

PD (n¼ 10,386)

4.5 (3.8–5.4)nnn 2.0 (1.7–2.4) nnn 1.7 (1.3–2.2) nnn 1.4 (1.2–1.7)nn 1.4 (1.1–1.9)nn 1.7 (1.5–2.1)nnn 1.8 (0.6–5.6) 1.0 (1.0-1.0)nnnb 0.5 (0.3–1.1) 1.0 (0.4–2.2) 1.1 (0.9–1.4) 1.1 (0.9–1.3) 1.1 (0.8–1.5) 0.6 (0.5–0.7)nnn 0.9 (0.8–1.1) 0.7 (0.6–1.0) 0.5 (0.3–0.7)nnn 0.7 (0.6–0.8)nnn 0.7 (0.6–0.9)nn

4.7 (3.7–6.1)nnn 2.2 (1.7–2.8)nnn 1.8 (1.2–2.6)n 1.4 (1.1–1.9)n 1.3 (1.0–1.6) 1.2 (1.0–1.6) 1.2 (0.7–2.5) 1.0 (1.0-1.0)b 0.8 (0.4–1.3) 1.1 (0.6–2.1) 1.3 (1.0–1.7)n 1.0 (0.8–1.4) 1.1 (0.7–1.6) 0.6 (0.5–0.8)nn 0.8 (0.6–1.1) 0.4 (0.2–0.6)nnn 0.3 (0.2–0.5)nnn 0.5 (0.4–0.6)nnn 0.4 (0.3–0.5)nnn

4.8 (3.9–5.9)nnn 1.9 (1.6–2.3)nnn 1.4 (1.1–1.9)n 1.1 (0.9–1.4) 1.3 (0.9–1.7) 1.3 (1.1–1.6)nn 1.4 (0.4–4.7) 1.1 (1.0–1.1)nnnb 0.8 (0.4–1.6) 0.9 (0.4–2.3) 1.0 (0.9–1.3) 1.0 (0.8–1.2) 0.9 (0.6–1.2) 0.8 (0.7–1.0)n 1.0 (0.8–1.2) 0.8 (0.5–1.1) 0.7 (0.4–1.0)n 0.5 (0.4–0.6)nnn 0.3 (0.2–0.4)nnn

5.2 (3.8–7.0)nnn 2.0 (1.5–2.6)nnn 1.5 (1.0–2.4) 1.3 (1.0–1.9) 1.3 (1.0–1.8) 1.2 (0.9–1.6) 1.2 (0.6–2.5) 1.0 (1.0–1.1)nnnb 1.0 (0.6–1.9) 1.0 (0.5–2.1) 0.9 (0.7–1.2) 0.9 (0.7–1.3) 0.9 (0.6–1.4) 0.9 (0.6–1.2) 0.8 (0.6–1.1) 0.6 (0.3–0.9)n 0.5 (0.3–0.9)n 0.5 (0.4–0.6)nnn 0.2 (0.1–0.3)nnn

Analysis of association between potential risk factors and completed suicide: univariate logistic regression with risk expressed as odds ratio (OR) and multivariate logistic regression with risk expressed as adjusted odds ratio (AOR). n

po 0.05. p o0.01. nnn p o 0.001. a Including adjustment for date of diagnosis. Due to missing data, there were only 26,071 non-PD and 12,101 PD for the univariate analyses concerning living alone and receiving disability pension, 26,069 non-PD and 12,100 PD for the univariate analysis concerning living in an institution, and 23,809 non-PD and 10,386 PD for the univariate analyses concerning education level above primary school. The multiple logistic regression analysis was based on the 31,009 individuals, with complete data. b Odds-ratio per year. nn

particularly susceptible to suicide (Mork et al., 2009; Roose et al., 1983) held true when subjected to investigation based on a large dataset consisting of patients with severe depression. Strikingly, psychotic depression was not identified as an independent risk factor for suicide in this population, supporting a previous finding of no difference between the two subtypes0 suicide risk (Black et al., 1988). The reason for this finding is probably that the examined population is comprised solely of severely depressed individuals, i.e. the analysis is, to some extent, adjusted for depressive severity (Fig. 1). This is in agreement with a recent Danish study comparing the symptom profile of patients with the psychotic and non-psychotic subtypes of severe depression. This study showed that these two groups of patients were both equally depressed and equally prone to self-harm (Ostergaard et al., 2012a). Among the remaining investigated factors in the present study, there were noteworthy findings in each of the three areas: 4.1. Socio-demographic factors The study confirmed the previously well-established risk factor of male gender (Large et al., 2011; Perkins et al., 2009; Qin, 2011). This association has been found to be related to the gender0 s more violent methods of attempting suicide, their reluctance to seek help, and their use of suicide as a stress response in situations where they have lost control or mastery (Moller-Leimkuhler, 2003). In addition, older age at diagnosis was associated with a higher risk of suicide in both non-PD and PD. Other studies have found a younger age of onset to be a risk factor for suicide, possibly due to the long exposure to severe illness (Oquendo et al., 1997). Due to our cessation of follow-up in December 2010, we are unable to discern cause of death past this date. This may have created a bias, which inadvertently censors the suicides of participants receiving a diagnosis of severe depression at a young age, thereby skewing our results. In other words, age at diagnoses should be interpreted

as “age at beginning of follow up” in the present study, and the findings indicate that persons of higher age with severe depression are at high risk of committing suicide. This is also in accordance with a previous Danish register-based study, which found higher population attributable risk for suicide with increasing age among depressed individuals (Qin, 2011). This study also identified socio-demographic factors associated with not committing suicide (protective factors). Notably, previous studies have identified recipients of disability pension as being at increased risk of suicide (Ahs and Westerling, 2006), while recent unemployment has been shown to increase the risk of suicide among the mentally ill (Hoyer et al., 2009). Nonetheless, in this study we found an overall protective effect against suicide when receiving disability pension in the total population and when stratified into subgroups. A potential explanation for this relationship could be that the individuals in question have been removed from situations that might otherwise heighten the level of psychological stress, feelings of guilt/worthlessness and hence, the risk of suicide. 4.2. Psychiatric morbidity As previously well documented (Nordentoft et al., 2011; Wiktorsson et al., 2010), a history of intentional self harm proved to be a highly significant risk factor for suicide in all analyses. The same was true of hospital contact due to poisoning with analgesia, except in the PD group – probably due to the lower number of cases and not a lack of association. These findings reinforce the need for taking a thorough history of previous self-harm and suicide attempts in the clinical setting, as part of a larger assessment of suicide risk. This clinical practice is of equal importance in both psychotic and non-psychotic depression. Another identified risk factor, psychiatric bed-days above median, may reflect treatment resistance and a more severe course of illness. Finally, our results suggest institutionalization (long-term psychiatric and/or

124

A.K.K. Leadholm et al. / Journal of Affective Disorders 156 (2014) 119–125

social care at a residential institution) to be a protective factor, a trend also seen in a Danish population of patients with schizophrenia (Uggerby et al., 2011). However, a study of recently discharged patients with affective disorders associated institutionalization with increased risk of suicide (Hoyer et al., 2009). These discrepancies are likely due to different study populations, and it therefore remains unclear to what extent institutionalization plays a role in suicide outcome.

Conflict of interest A.K.K. Leadholm and P. Bech declare no conflict of interest. A.J. Rothschild has received Grant support from the National Institute of Mental Health, Cyberonics, Takeda, and St. Jude Medical and has served as a consultant to Allergan, GlaxoSmithKline, Eli Lilly, Noven Pharmaceuticals, Pfizer, Shire Pharmaceuticals, Sunovian, and Takeda. J. Nielsen has received research Grants from Lundbeck, Pfizer and Chempaq for clinical trials and speaking fees from Bristol-Myers Squibb, Astra Zeneca, Janssen-Cilag, Lundbeck and Hemocue. S.D. Østergaard has received speaking fees, consultant honoraria and travel support from Janssen-Cilag until April 2011. Furthermore, he has received travel support on one occasion in 2010 from Bristol Myers Squibb.

4.3. Physical morbidity Despite the strong association between physical illness and depression (Goldberg and Harrow, 2010; Ostergaard and Foldager, 2011), as well as a number of studies showing that the co-morbidity between physical illnesses and depression is associated with a very high risk of suicide (Chen et al., 2010; Forsstrom et al., 2010; Larsen et al., 2010; Llorente et al., 2005), our study found co-morbid physical illness to be a significant protective factor against suicide for the whole study population. The suicide risk studies mentioned above, however, investigated all depression severities, while our study only examined the severe subtype (Fig. 1). It might then be hypothesized that the increased suicide risk seen in depression with co-morbid physical illness only exists in mild and moderate depression, where the physical illness may be more likely to be part of the factors leading to (secondary) depression, than is the case in severe depression, which may be a more “genetic” (primary) disorder (Ostergaard et al., 2013). Furthermore, the physical illnesses assessed in the present study are generally relatively severe, and the negative association with suicide may simply be due to the fact that the affected individuals are more likely to die as a direct cause of the illness in question. Another contributing factor could be that those treated for physical illnesses have more frequent contacts with medical services, which may pick up on worsening in psychopathology or suicidality and report these to the patients0 psychiatrist or refer directly to the psychiatric emergency room in severe cases.

4.4. Strengths and Limitations The results discussed here must be considered in light of the study population. The subjects in the present study were all severely depressed, while previous studies have been conducted on a range of depression severities. Furthermore, the subjects were included based on contact with secondary care facilities (inpatient/ outpatient care at psychiatric hospitals), and therefore it is difficult to extrapolate these results to the primary care setting. Finally, we were not able to include treatment history as a variable as data on inpatient treatment is not available in the Danish registries. The strength of our findings lies in our study population, as it was large, nationwide, and contained a representative proportion of psychotically depressed individuals (35%) (Coryell et al., 1984; Gournellis and Lykouras, 2006; Meyers and Greenberg, 1986; Ohayon and Schatzberg, 2002). In conclusion, suicide prevention efforts should pay equal attention to patients with PD and non-PD. Of special interest in both groups are older age, male gender and previous incidents of self-harm, as they were all found to be significant risk factors for suicide in both univariate and multivariate analyses in the present study.

Role of funding source The study was partly funded by a Grant from the Lundbeck Foundation (stipend for A.K. Leadholm). The remaining authors were funded by their respective institutions as listed under affiliations.

Acknowledgments The authors are grateful to programmer Søren Skadhede, Department M, Aarhus University Hospital, Risskov, Denmark and statistician Signe Olrik Wallenstein Jensen, Unit for Psychiatric Research, Aalborg Psychiatric Hospital, Aalborg University Hospital. The Project was supported by a Grant from the Lundbeck Foundation.

References Agerbo, E., Mortensen, P.B., Eriksson, T., Qin, P., Westergaard-Nielsen, N., 2001. Risk of suicide in relation to income level in people admitted to hospital with mental illness: nested case-control study. Br. Med. J. (Clinical research ed.) 322, 334–335. Ahs, M.A., Westerling, R., 2006. Mortality in relation to employment status during different levels of unemployment. Scand. J. Public Health 34, 159–167. American Psychiatric Association, 2000. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. A.P.A., Washington, DC. (Text Revision). Angst, J., Angst, F., Gerber-Werder, R., Gamma, A., 2005. Suicide in 406 mooddisorder patients with and without long-term medication: a 40 to 44 years0 follow-up. Arch. Suicide Res. 9, 279–300. Berman, A.L., 2006. Risk management with suicidal patients. J. Clin. Psychol. 62, 171–184. Black, D.W., Winokur, G., Nasrallah, A., 1988. Effect of psychosis on suicide risk in 1593 patients with unipolar and bipolar affective disorders. Am. J. Psychiatry 145, 849–852. Bolton, J.M., Pagura, J., Enns, M.W., Grant, B., Sareen, J., 2010. A population-based longitudinal study of risk factors for suicide attempts in major depressive disorder. J. Psychiatr. Res. 44, 817–826. Caetano, S.C., Olvera, R.L., Hunter, K., Hatch, J.P., Najt, P., Bowden, C., Pliszka, S., Soares, J.C., 2006. Association of psychosis with suicidality in pediatric bipolar I, II and bipolar NOS patients. J. Affect. Disord. 91, 33–37. Cavanagh, J.T., Carson, A.J., Sharpe, M., Lawrie, S.M., 2003. Psychological autopsy studies of suicide: a systematic review. Psychol. Med. 33, 395–405. Chen, C.K., Tsai, Y.C., Hsu, H.J., Wu, I.W., Sun, C.Y., Chou, C.C., Lee, C.C., Tsai, C.R., Wu, M.S., Wang, L.J., 2010. Depression and suicide risk in hemodialysis patients with chronic renal failure. Psychosomatics 51 (528–528), e526, http://www.ncbi. nlm.nih.gov/pubmed/21051686. Coryell, W., Young, E.A., 2005. Clinical predictors of suicide in primary major depressive disorder. J. Clin. Psychiatry 66, 412–417. Coryell, W., Pfohl, B., Zimmerman, M., 1984. The clinical and neuroendocrine features of psychotic depression. J. Nerv. Ment. Dis. 172, 521–528. Coryell, W., Leon, A., Winokur, G., Endicott, J., Keller, M., Akiskal, H., Solomon, D., 1996. Importance of psychotic features to long-term course in major depressive disorder. Am. J. Psychiatry 153, 483–489. Cumming, T.B., Churilov, L, Skoogm, I., Blomstr, C., Linden, T., 2010. Little evidence for different phenomenology in poststroke depression. Acta. Psychiatr. Scand. 121, 424–430. Domschke, K., 2013. Clinical and molecular genetics of psychotic depression. Schizophr Bull 39, 766–775. Dumais, A., Lesage, A.D., Alda, M., Rouleau, G., Dumont, M., Chawky, N., Roy, M., Mann, J.J., Benkelfat, C., Turecki, G., 2005. Risk factors for suicide completion in major depression: a case-control study of impulsive and aggressive behaviors in men. Am. J. Psychiatry 162, 2116–2124. Erlangsen, A., Nordentoft, M., Conwell, Y., Waern, M., De Leo, D., Lindner, R., Oyama, H., Sakashita, T., Andersen-Ranberg, K., Quinnett, P., Draper, B., Lapierre, S., 2011. Key considerations for preventing suicide in older adults: consensus opinions of an expert panel. Crisis 32, 106–109. Forsstrom, E., Hakko, H., Nordstrom, T., Rasanen, P., Mainio, A., 2010. Suicide in patients with stroke: a population-based study of suicide victims during the years 1988–2007 in northern Finland. J. Neuropsychiatry Clin. Neurosci. 22, 182–187. Ganz, D., Sher, L., 2009. Suicidal behavior in adolescents with comorbid depression and alcohol abuse. Minerva Pediatr. 61, 333–347. Goldberg, F., Harrow, M., 2010. The detection and treatment of depression in the physically ill. World Psychiatry: Off. J. World Psychiatr. Assoc. (WPA) 9, 16–20. Goodwin, R.D., 2011. Is COPD associated with suicide behavior? J. Psychiatr. Res. 45, 1269–1271. Gournellis, R., Lykouras, L., 2006. Psychotic (delusional) major depression in the elderly: a review. Curr. Psychiatry Rev., 235–244.

A.K.K. Leadholm et al. / Journal of Affective Disorders 156 (2014) 119–125

Grunebaum, M.F., Oquendo, M.A., Harkavy-Friedman, J.M., Ellis, S.P., Li, S., Haas, G.L., Malone, K.M., Mann, J.J., 2001. Delusions and suicidality. Am. J. Psychiatry 158, 742–747. Haw, C., Harwood, D., Hawton, K., 2009. Dementia and suicidal behavior: a review of the literature. Int. Psychogeriatr. 21, 440–453. Health, D.N.B.o., 1971. Classification of Diseases: Extended Danish-Latin Version of the World Health Organization International Statistical Classification of Diseases and Related Health Problems, 8th Revision. Danish National Board of Health, Copenhagen, Denmark. Hoyer, E.H., Licht, R.W., Mortensen, P.B., 2009. Risk factors of suicide in inpatients and recently discharged patients with affective disorders. A case-control study. Eur. Psychiatry: J. Assoc. Eur. Psychiatr. 24, 317–321. Jeon, H.J., Lee, J.Y., Lee, Y.M., Hong, J.P., Won, S.H., Cho, S.J., Kim, J.Y., Chang, S.M., Lee, H.W., Cho, M.J., 2010. Unplanned versus planned suicide attempters, precipitants, methods, and an association with mental disorders in a Korea-based community sample. J. Affect. Disord. 127, 274–280. Jeon, H.J., Peng, D., Chua, H.C., Srisurapanont, M., Fava, M., Bae, J.N., Man Chang, S., Hong, J.P., 2013. Melancholic features and hostility are associated with suicidality risk in Asian patients with major depressive disorder. J. Affect. Disord. 148, 368–374. Juel, K., Helweg-Larsen, K., 1999. The Danish registers of causes of death. Dan. Med. Bull. 46, 354–357. Kessing, L.F., 2004. Severity of depressive episodes according to ICD-10: prediction of risk of relapse and suicide. Br. J. Psychiatry: J. Ment. Sci. 184, 153–156. Kokoszka, A., Pouwer, F., Jodko, A., Radzio, R., Mucko, P., Bienkowska, J., Kuligowska, E., Smoczynska, O., Sklodowska, Z., 2009. Serious diabetes-specific emotional problems in patients with type 2 diabetes who have different levels of comorbid depression: a Polish study from the European Depression in Diabetes (EDID) Research Consortium. Eur. Psychiatry: J. Assoc. Eur. Psychiatr. 24, 425–430. Large, M., Sharma, S., Cannon, E., Ryan, C., Nielssen, O., 2011. Risk factors for suicide within a year of discharge from psychiatric hospital: a systematic metaanalysis. Aust N Z J Psychiatry 45, 619–628. Larsen, K.K., Agerbo, E., Christensen, B., Sondergaard, J., Vestergaard, M., 2010. Myocardial infarction and risk of suicide: a population-based case-control study. Circulation 122, 2388–2393. Leadholm, A.K., Rothschild, A.J., Nolen, W.A., Bech, P., Munk-Jorgensen, P., Ostergaard, S.D., 2013. The treatment of psychotic depression: is there consensus among guidelines and psychiatrists? J. Affect. Disord. 145, 214–220. Leckman, J.F., Weissman, M.M., Prusoff, B.A., Caruso, K.A., Merikangas, K.R., Pauls, D. L., Kidd, K.K., 1984. Subtypes of depression. Family study perspective. Arch. Gen. Psychiatry 41, 833–838. Lippi, G., Montagnana, M., Favaloro, E.J., Franchini, M., 2009. Mental depression and cardiovascular disease: a multifaceted, bidirectional association. Semin. Thromb. Hemost. 35, 325–336. Llorente, M.D., Burke, M., Gregory, G.R., Bosworth, H.B., Grambow, S.C., Horner, R.D., Golden, A., Olsen, E.J., 2005. Prostate cancer: a significant risk factor for late-life suicide. Am. J. Geriatr. Psychiatry: Off. J. Am. Assoc. Geriatr. Psychiatry 13, 195–201. Loo, C.K., Mahon, M., Katalinic, N., Lyndon, B., Hadzi-Pavlovic, D., 2011. Predictors of response to ultrabrief right unilateral electroconvulsive therapy. J. Affect. Disord. 130, 192–197. Lynge, E., Sandegaard, J.L., Rebolj, M., 2011. The Danish National Patient Register. Scand. J. Public Health 39, 30–33. Madsen, T., Agerbo, E., Mortensen, P.B., Nordentoft, M., 2012. Predictors of psychiatric inpatient suicide: a national prospective register-based study. J. Clin. Psychiatry 73, 144–151. Meyers, B.S., Greenberg, R., 1986. Late-life delusional depression. J. Affect. Disord. 11, 133–137. Moller-Leimkuhler, A.M., 2003. The gender gap in suicide and premature death or: why are men so vulnerable? Eur. Arch. Psychiatry Clin. Neurosci. 253, 1–8. Mork, E., Mehlum, L., Walby, F., 2009. Selvmord ved psykotisk depresjon og bipolar lidelse: Forekomst og risikofaktorer. Suicidologi 14, 8–12. Mors, O., Perto, G.P., Mortensen, P.B., 2011. The Danish psychiatric central research register. Scand. J. Public Health 39, 54–57. Mortensen, P.B., Agerbo, E., Erikson, T., Qin, P., Westergaard-Nielsen, N., 2000. Psychiatric illness and risk factors for suicide in Denmark. The Lancet 355, 9–12. Munk-Jorgensen, P., Ostergaard, S.D., 2011. Register-based studies of mental disorders. Scand. J. Public Health 39, 170–174. Nordentoft, M., 2007. Prevention of suicide and attempted suicide in Denmark. Epidemiological studies of suicide and intervention studies in selected risk groups. Dan. Med. Bull. 54, 306–369. Nordentoft, M., Qin, P., Helweg-Larsen, K., Juel, K., 2007. Restrictions in means for suicide: an effective tool in preventing suicide: the Danish experience. Suicide Life-Threat. Behav. 37, 688–697.

125

Nordentoft, M., Mortensen, P.B., Pedersen, C.B., 2011. Absolute risk of suicide after first hospital contact in mental disorder. Arch. Gen. Psychiatry 68, 1058–1064. Ohayon, M.M., Schatzberg, A.F., 2002. Prevalence of depressive episodes with psychotic features in the general population. Am. J. Psychiatry 159, 1855–1861. Oquendo, M.A., Malone, K.M., Mann, J.J., 1997. Suicide: risk factors and prevention in refractory major depression. Depress. Anxiety 5, 202–211. Ostergaard, S.D., Foldager, L., 2011. The association between physical illness and major depressive episode in general practice. Acta Psychiatr. Scand. 123, 290–296. Ostergaard, S.D., Bille, J., Soltoft-Jensen, H., Lauge, N., Bech, P., 2012a. The validity of the severity-psychosis hypothesis in depression. J. Affect. Disord. 140, 48–56. Ostergaard, S.D., Rothschild, A.J., Uggerby, P., Munk-Jorgensen, P., Bech, P., Mors, O., 2012b. Considerations on the ICD-11 classification of psychotic depression. Psychother. Psychosom. 81, 135–144. Ostergaard, S.D., Petrides, G., Dinesen, P.T., Skadhede, S., Bech, P., Munk-Jorgensen, P., Nielsen, J., 2013. The association between physical morbidity and subtypes of severe depression. Psychother. Psychosom. 82, 45–52. Ostergaard, S.D., Waltoft, B.L., Mortensen, P.B., Mors, O., 2013a. Environmental and familial risk factors for psychotic and non-psychotic severe depression. J. Affect. Disord. 147, 232–240. Østergaard, S.D., Leadholm, A.K.K., Rothschild, A.J., 2013b. Persistent delusional theme over 13 episodes of psychotic depression. Acta Neuropsychiatr. 25, 370–373. Park, M.H., Kim, T.S., Yim, H.W., Jeong, S.H., Lee, C., Lee, C.U., Kim, J.M., Jung, S.W., Lee, M.S., Jun, T.Y., 2010. Clinical characteristics of depressed patients with a history of suicide attempts: results from the CRESCEND study in South Korea. J. Nerv. Ment. Dis. 198, 748–754. Pedersen, C.B., 2011. The Danish civil registration system. Scand. J. Public Health 39, 22–25. Perkins, R., Sanddal, T.L., Howell, M., Sanddal, N.D., Berman, A., 2009. Epidemiological and follow-back study of suicides in Alaska. Int. J. Circumpolar. Health 68, 212–223. Putman-Casdorph, H., McCrone, S., 2009. Chronic obstructive pulmonary disease, anxiety, and depression: state of the science. Heart Lung: J. Acute Crit. Care 38, 34–47. Qin, P., 2011. The impact of psychiatric illness on suicide: differences by diagnosis of disorders and by sex and age of subjects. J. Psychiatr. Res. Qin, P., Nordentoft, M., 2005. Suicide risk in relation to psychiatric hospitalization: evidence based on longitudinal registers. Arch. Gen. Psychiatry 62, 427–432. Qin, P., Jepsen, P., Norgard, B., Agerbo, E., Mortensen, P.B., Vilstrup, H., Sorensen, H. T., 2009. Hospital admission for non-fatal poisoning with weak analgesics and risk for subsequent suicide: a population study. Psychol. Med. 39, 1867–1873. Roose, S.P., Glassman, A.H., Walsh, B.T., Woodring, S., Vital-Herne, J., 1983. Depression, delusions, and suicide. Am. J. Psychiatry 140, 1159–1162. Rothschild, A.J., 2009. Clinical Manual for Diagnosis and Treatment of Psychotic Depression. American Psychiatric Publishing, Inc., Washington DC, USA. Schaffer, A., Flint, A.J., Smith, E., Rothschild, A.J., Mulsant, B., Szanto, K., PeasleyMiklus, C., Heo, M., Papademetriou, E., Meyers, B.S., 2008. Correlates of suicidality among patients with psychotic depression. Suicide Life-Threat. Behav. 38, 403–414. Spiegel, D., 1996. Cancer and depression. Br. J. Psychiatry (Supplement), 30, 109–116. Statistics Denmark, Deaths by Region, Sex, Age and Cause of Death (StatBank.dk/ fod507). Statistics Denmark, 1991. IDA: An integrated Database for Labor Market Research. Statistics Denmark Press, Copenhagen. Denmark. (in Danish). Torres, A.R., Ramos-Cerqueira, A.T., Ferrao, Y.A., Fontenelle, L.F., do Rosario, M.C., Miguel, E.C., 2011. Suicidality in obsessive–compulsive disorder: prevalence and relation to symptom dimensions and comorbid conditions. J. Clin. Psychiatry 72, 17–26. (quiz 119). Uggerby, P., Nielsen, R.E., Correll, C.U., Nielsen, J., 2011. Characteristics and predictors of long-term institutionalization in patients with schizophrenia. Schizophr. Res. 131, 120–126. Vythilingam, M., Chen, J., Bremner, J.D., Mazure, C.M., Maciejewski, P.K., Nelson, J.C., 2003. Psychotic depression and mortality. Am. J. Psychiatry 160, 574–576. Wenzel, A., Berchick, E.R., Tenhave, T., Halberstadt, S., Brown, G.K., Beck, A.T., 2011. Predictors of suicide relative to other deaths in patients with suicide attempts and suicide ideation: a 30-year prospective study. J. Affect. Disord. 132, 375–382. Wiktorsson, S., Runeson, B., Skoog, I., Ostling, S., Waern, M., 2010. Attempted suicide in the elderly: characteristics of suicide attempters 70 years and older and a general population comparison group. Am. J. Geriatr. Psychiatry: Off. J. Am. Assoc. Geriatr. Psychiatry 18, 57–67. World Health Organization, WHO, 1993. The ICD-10 Classification of Mental and Behavioural Disorders. Diagnostic Criteria for Research. WHO, Geneva. World Health Organization, WHO, 2011. Suicide Prevention (SUPRE).