Welcome to the Schizophrenia Resource Centre

Welcome, this website is intended for healthcare professionals in EMEA with an interest in the treatment of schizophrenia. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

Suicide in first episode psychosis: A nationwide cohort study

Schizophrenia Research, Volume 157, Issues 1–3, August 2014, Pages 1–7



Relatively little is known about suicide in diagnostic subtypes of first episode psychosis (FEP). Our aim was to assess suicide rates and potential risk factors for suicide in FEP.


This is a national register-based cohort study of patients born in 1973–1978 in Sweden and who were hospitalized with a FEP between ages 15 and 30 years (n = 2819). The patients were followed from date of discharge until death, emigration, or 31st of December 2008. The suicide rates for six diagnostic subtypes of FEP were calculated. Suicide incidence rate ratios (IRRs) were calculated to evaluate the association between suicide and psychiatric, familial, social, and demographic factors.


In total 121 patients died by suicide. The overall suicide rate was 4.3 (95% confidence interval [CI] 3.5–5.0) per 1000 person-years. The highest suicide rates were found in depressive disorder with psychotic symptoms and in delusional disorder. In an adjusted model, the strongest risk factors for suicide were self-harm (IRR 2.7, CI 1.7–4.4) or a conviction for violent crime (IRR 2.0, CI 1.3–3.2). Also having a first-degree relative with a schizophrenia/bipolar diagnosis (IRR 2.1, CI 1.2–3.6) or substance use disorder (IRR 2.0, CI 1.2–3.2) were significant risk factors for suicide.


Impulsive behavior such as self-harm as well as having a family history of severe mental disorder or substance use are important risk factors for suicide in FEP.

1. Introduction

Patients with a first episode psychosis (FEP) have an elevated suicide risk and higher all-cause mortality (Allebeck, 1989, Craig et al, 2006, Dutta et al, 2010, and Bjorkenstam et al, 2013). In a cohort study of FEP patients in England and Scotland the suicide mortality was increased 12 times compared to the general population ( Dutta et al., 2010 ). Previous studies have identified depressive symptoms, suicidal ideation/intent, negative events, non-suicidal self-harm, and depressive symptoms as risk factors for suicide in FEP ( Fedyszyn et al., 2012 ). Dutta et al. (2011) also found male gender and presence of more pronounced psychiatric symptoms to be predictors of suicide in FEP. In a recent literature review, a substantial proportion of psychotic patients examined after violent suicide attempts, self-mutilation, homicide, and assault resulting in serious injury were found to be in their first episode of psychosis ( Nielssen et al., 2012 ). Moreover, a substantial proportion of first-episode patients committed an act of less serious violence or attempted suicide prior to initial treatment ( Nielssen et al., 2012 ). A history of violent offending has also been shown to be a risk factor for suicide among people diagnosed with schizophrenia ( Webb et al., 2011 ). However, to the best of our knowledge, no earlier study has investigated the role of violent offending on suicide risk in FEP.

Although a delay in the clinical diagnosing of schizophrenia is common, most studies of suicide risk and risk factors for suicide in FEP have been performed on patients with established schizophrenia (Foley et al, 2008 and Pompili et al, 2011). However, self-harm has been shown to be common during the pre-treatment phase of first-episode psychosis (Harvey et al, 2008 and Nielssen et al, 2012) and one can assume that some patients may die by suicide early in the course of mental illness, before they have developed or been formally diagnosed with schizophrenia. Moreover, most studies of suicide in FEP have not been population based and have included a relatively small number of patients and suicides (Radomsky et al, 1999 and Bertelsen et al, 2007). For these reasons, more knowledge is needed about suicide rates and risk factors for suicide in FEP and in specified subtypes of FEP. Because the suicide risk is most increased in the early phase of psychotic disorder ( Osby et al., 2000 ), this time period is particularly important to study.

We performed a nationwide register study of completed suicide across the broad spectrum of patients with a first inpatient diagnosed psychotic disorder in a cohort of young adults. Our aims were, first, to calculate suicide rates and second, to assess the significance of a range of potential risk factors for suicide across different FEP diagnoses.

2. Methods

2.1. Study population

National registers make it possible to study the entire Swedish population and to perform linkage of data between different registers on an individual level. In the present study, the unique personal identity number assigned to each permanent resident in Sweden was used to link information from ten population-based registers. The Medical Birth Register, established in 1973 and held by the Swedish National Board of Health and Welfare, includes information on almost all births in Sweden. The National Patient Register includes all individuals admitted to psychiatric or general hospitals, with nearly complete coverage for psychiatric care since 1973 and for somatic care since 1987 ( Ludvigsson et al., 2011 ). Through the Multi-Generation Register one is able to link children and parents (biological and adoptive) together. The Causes of Death Register comprises information on all deaths of Swedish residents since 1952. The Register of Court Convictions contains information on all court convictions in Sweden for persons 15 years of age or older. The Swedish Register of Children and Young Persons subjected to child welfare measures were used to obtain records on out-of-home care foster family and residential care. The Total Enumeration Income Survey contains data on the income of and governmental benefits provided to all Swedish residents. The Total Population Register, established in 1968, includes information such as age, sex, and place of residence. Finally, the National School Register, administered jointly by the Swedish National School Administration and Statistics Sweden, encompasses information on each individual's educational achievement, i.e. grades by subject as well as grade point average, for all students graduating from both compulsory school (years 1–9) and secondary school (years 10–12) since 1988. The selection of the study population is illustrated in Fig. 1 . We identified all individuals born between 1973 and 1978 in the Medical Birth Register (n = 604 072). Individuals who died (n = 7097) or emigrated (n = 17 508) before age 15 were excluded. We only included individuals with at least one biological parent born in Sweden and for whom we also could obtain the personal identity number for the biological mother (excluding another 40 087 individuals). Of the remaining 551 590 individuals, we identified as the study cohort all individuals with a first diagnosis of hospital-treated psychosis (as defined by the International Classification of Disease (ICD) ICD-9: 291–292, 295, 296A, 296C, 296D, 296E, 296W, 297B, 297C, 297W, 297D, 298; ICD-10: F10.5–F19.5, F10.7–F19.7, F20–F31, F32.3, F33.3) sometime between the age 15 to 30 years (n = 2819). Thus, the diagnoses were given between the years 1988 and 2003. Seven diagnostic groups were created: (1) schizophrenia and schizoaffective disorder (ICD-9: 295; ICD-10: F20, F25), (2) bipolar disorder (ICD-9: 296A, 296C–296E, 296W; ICD-10: F30–F31), (3) depressive disorder with psychotic symptoms (ICD-10: F32.3, F33.3), (4) persistent delusional disorder (ICD-9: 297B–297C, 297W; ICD-10: F22), (5) psychotic disorder due to substance use (ICD-9: 291, 292; ICD-10: F10.5–F19.5, F10.7–F19.7), (6) acute and transient psychotic disorder (ICD-9: 298; ICD-10: F23), and (7) other psychoses (ICD-9: 297D, 298; ICD-10: F21, F24, F28–F29).


Fig. 1 Flow chart for the study population.

2.2. Outcome

Suicide was identified in the Cause of Death Register (ICD-9: E950–E959, E980–E989; ICD-10: X60–X84, Y10–Y34) as an underlying cause of death. To reduce spatial and secular trends in detecting and classifying cases of suicide ( Linsley et al., 2001 ), our definition of suicide includes deaths with undetermined intent. The patients were followed from date of discharge after a first-time psychosis until date of death, emigration, or until the end of the follow-up period, i.e. 31st of December 2008.

2.3. Risk factors

We categorized potential risk factors for suicide occurring before the first diagnosis of FEP into five main categories: psychiatric factors, familial factors, social factors, premorbid intellectual functional level, and other risk factors.

2.4. Psychiatric factors

We created three dichotomous variables for having a history of hospitalization (i.e. before first diagnosis for psychosis) with the following diagnoses: (1) any psychiatric diagnosis (ICD-8/ICD-9: 290–319; ICD-10: F00–F99), (2) deliberate self-harm (ICD-8/ICD-9: E950–E959; ICD-10: X60–X84), and (3) depression (ICD-8: 296.00, 300.40; ICD-9: 296B, 300E, 311; ICD-10: F32–33). Substance use disorder was defined as a hospitalization with a primary or secondary diagnosis of a substance use disorder (ICD-9: 291, 292, 303–305 and ICD-10: F10–F19) or a conviction for an alcohol or drug-related offense.

2.5. Familial factors

We created dichotomous variables for familial mental disorders, i.e. psychiatric diagnoses among first-degree relatives (biological parents or siblings) and who had been hospitalized any time between 1973 and the study subjects' first psychiatric diagnosis. Through the National Patient Register we identified the following diagnoses: schizophrenia and bipolar disorder (ICD-8/ICD-9: 295, 296A, C, D, E, W; ICD-10: F20, F25, F30–F31), substance use disorder (ICD-9: 291, 292, 303–305; ICD-10: F10–F19) or other mental disorders. As a separate risk factor, we combinedParent hospitalized for any psychiatric diagnosisandConvicted for a violent crime.

2.6. Social factors

2.6.1. Conviction for violent crime

Obtained from the National Crime Register, violent crime was defined as homicide, assault, robbery, arson, sexual offense (rape, sexual coercion, child molestation, indecent exposure or sexual harassment), illegal threats or intimidation ( Fazel and Grann, 2006 ).

2.6.2. First child welfare intervention before age 12

As previous research has found experience of child welfare interventions to be a risk factor for mental health problems later in life (Vinnerljung et al, 2006 and Leslie et al, 2010), we selected out-of-home care or provision of a contact family before age 12 as a marker for social welfare interventions.

2.6.3. Social assistance recipiency

We collected information on whether the household received social assistance between ages 12 and 17. At least one parent had to receive social assistance during at least one year, where also more than 50% of the yearly income should be generated from social assistance.

2.6.4. Premorbid school performance

In an attempt to capture intellectual capacity and social adaptation pre-diagnosis we retrieved information on school grades from the National School Register, and created two variables ( Berlin et al., 2011 ):

  • No, incomplete or low (less than one standard deviation (SD) below average) grades from compulsory school (years 1–9).
  • No graduation from secondary school (years 10–12) at age 20.

The first variable captures individuals who due to reduced cognitive functions, illness or other difficulties graduate with either low grades or who lack grades in one or more subjects. In Sweden you cannot start secondary school if you haven't graduated from compulsory school with complete grades. The second variable captures individuals who at age 20 still have not graduated from secondary school. We also calculated grade point average for the different FEP groups. During the period 1988–1997 the school system used a normally distributed scale of five grades 1 to 5 (where 5 is the highest) with the presumption that school achievements are dispersed according to a normal distribution with mean 3.0 and SD 1.0. However, this information was not used as a risk factor.

2.6.5. Head injury

We used the National Patient Register to retrieve information on hospitalization for intracranial injuries, or fractures of skull and facial bones (ICD-9: 800–801.3, 802–803.3, 804.0–804.3, 873.6–873.7, 851–854; ICD-10: S061–S069, S02) until date of first diagnosis.

2.6.6. Obstetrical complication

We selected the following complications: preeclampsia, neonatal jaundice (ICD-8: 6349, 636, 6370–6371, 6379, 6379, 661, 6370, 661; ICD-9: 6424–6427, 6429, 6461–6462; ICD-10: O11–O15, P57.8, P57.9), being born small for gestational age, prematurity (< 33 weeks), microcephaly (i.e. head circumference 32 cm) and an Apgar score of 7 at 5 min post partum.

2.7. Statistical analysis

Person-years at risk during follow-up were calculated by adding the years the study individuals were alive and living in Sweden. We performed Poisson regression analyses using the SAS GENMOD procedure to calculate incidence rate ratios (IRRs) for suicide with 95% confidence intervals (CI). We calculated IRRs separately for each risk factor as a ratio of the incidence rate of suicide among patients having vs. not having a particular risk factor. This way, each IRR estimates the association between a particular risk factor and suicide. We adjusted the IRRs for sex and year of birth. We also calculated population attributable fractions (PAF) for the comorbid disorders to estimate the proportion of suicides in the FEP cohort that is associated with a specific risk factor ( Rothman, 2013 ). The following formula was used:


stripin: si1.gif

where IRR = relative risk of suicide associated with having a specific risk factor. We used relative risks (IRRs) that were adjusted for sex in the calculation of PAF. All analyses were carried out in the statistical software SAS v. 9.1 (SAS Institute Inc. Cary, NC, USA).

2.8. Ethical considerations

This study was evaluated and approved by the regional ethical review board in Stockholm, Sweden (no: 2009/2027-31/5 and 2012/657-32).

3. Results

Table 1 presents descriptive statistics of the study cohort as well as the suicide rates across the FEP diagnoses and the prevalence of the potential risk factors for suicide. The cohort included 2819 patients (42% female). The mean age at FEP diagnosis was 22.6 and 47% were between 20 and 24 years of age. In total, 121 patients died by suicide during follow-up. The highest point estimates of suicide rates were found in depressive disorder with psychotic symptoms 7.8 (95% CI 2.4–13.1) per 1000 person years and in delusional disorder 6.9 (95% CI 0.9–13.0) per 1000 person years, whereas the overall suicide rate was 4.3 (95% CI 3.5–5.0) per 1000 person years. During the follow-up 3.8% of the women and 4.7% of the men died by suicide.

Table 1 Descriptive statistics and suicide rates with 95% confidence intervals (CI) for patients with a first episode psychosis.

  Schizophrenia/schizoaffective disorder Bipolar disorder Depressive disorder with psychotic symptoms Persistent delusional disorders Psychotic disorder due to substance use Acute and transient psychotic disorders Other psychoses Total
  n Suicide n Suicide n Suicide n Suicide n Suicide n Suicide n Suicide n Suicide
n 458 18 388 13 133 8 81 5 445 20 868 39 446 18 2819 121
Rates of suicide, no./1000 person years (CI) 3.6 (2.0–5.3)   3.4 (1.5–5.2)   7.8 (2.4–13.1)   6.9 (0.9–13.0)   4.8 (2.7–6.9)   3.9 (2.7–5.1)   4.9 (2.6–7.1)   4.3 (3.5–5.0)  
Mean age at first admission, years (SD) 21.9 (3.5)   22.9 (3.7)   24.8 (2.5)   24.0 (3.5)   23.2 (3.3)   21.3 (3.6)   24.2 (2.6)   22.6 (3.6)  
Mean time of follow-up (months)   21.6   51.8   1.2   6.5   50.7   39.7   22.0   30.7
 Women 158 6 (4%) 222 (57%) 9 (4%) 79 (59%) 4 (5%) 21 (26%) 0 (0%) 107 (24%) 4 (4%) 423 (49%) 16 (4%) 177 (40%) 6 (3%) 1187 (42%) 45 (4%)
 Men 300 12 (4%) 166 (43%) 4 (2%) 54 (41%) 4 (7%) 60 (74%) 5 (8%) 338 (76%) 16 (5%) 445 (51%) 23 (5%) 269 (60%) 12 (4%) 1632 (58%) 76 (5%)
Age at first hospitalization                                
 15–19 years 146 (32%) 6 (1%) 99 (26%) 4 (1%) 4 (3%) 0 (0%) 14 (17%) 0 (0%) 81 (18%) 2 (0%) 361 (42%) 16 (2%) 25 (6%) 0 (0%) 730 (26%) 28 (1%)
 20–24 years 222 (48%) 7 (2%) 162 (42%) 5 (1%) 66 (50%) 4 (3%) 30 (37%) 0 (0%) 224 (50%) 9 (2%) 359 (41%) 16 (2%) 248 (56%) 8 (2%) 1311 (47%) 49 (2%)
 25–30 years 90 (20%) 5 (1%) 127 (33%) 4 (1%) 63 (47%) 4 (3%) 37 (46%) 5 (6%) 140 (31%) 9 (2%) 148 (17%) 7 (1%) 173 (39%) 10 (2%) 778 (28%) 44 (2%)
Risk factors                                
Psychiatric factor                                
 Prior hospitalization with a psychiatric diagnosis 164 (36%) 6 (33%) 168 (43%) 5 (38%) 50 (38%) 4 (3%) 27 (33%) 0 (0%) 199 (45%) 10 (50%) 225 (26%) 14 (36%) 135 (30%) 6 (33%) 968 (34%) 45 (37%)
 Prior hospitalization for intentional self-harm 45 (10%) 4 (22%) 35 (9%) 2 (15%) 15 (11%) 3 (2%) 7 (9%) 1 (20%) 58 (13%) 4 (20%) 50 (6%) 5 (13%) 35 (8%) 2 (11%) 245 (9%) 21 (17%)
 Prior hospitalization with a diagnosis for depression 38 (8%) 1 (6%) 61 (16%) 4 (31%) 23 (17%) 2 (2%) 3 (4%) 0 (0%) 22 (5%) 1 (5%) 32 (4%) 0 (0%) 25 (6%) 2 (11%) 204 (7%) 10 (8%)
 Alcohol- or drug use disorder (hospitalization or conviction), including secondary diagnoses 56 (12%) 3 (17%) 46 (12%) 2 (15%) 14 (11%) 3 (38%) 16 (20%) 0 (0%) 243 (55%) 13 (65%) 86 (10%) 5 (13%) 73 (16%) 2 (11%) 534 (19%) 28 (23%)
premorbid school performance                                
 No/incomplete grades from the primary school 144 (31%) 1 (6%) 69 (18%) 2 (15%) 30 (23%) 1 (1%) 29 (36%) 0 (0%) 250 (56%) 13 (65%) 242 (28%) 14 (36%) 122 (27%) 4 (22%) 886 (31%) 35 (29%)
 Grade point average from the primary school (SD) a 3.0 (0.8)   3.2 (0.8)   3.1 (0.7)   2.8 (0.8)   2.4 (0.7)   3.1 (0.8)   3.0 (0.8)   3.0 (0.8)  
 No graduation from secondary school at age 20 225 (49%) 9 (50%) 126 (32%) 7 (54%) 33 (25%) 1 (1%) 34 (42%) 1 (20%) 273 (61%) 15 (75%) 376 (43%) 18 (46%) 171 (38%) 6 (33%) 1238 (44%) 57 (47%)
Social factors                                
 Conviction for violent crime 42 (9%) 0 (0%) 14 (4%) 1 (8%) 5 (4%) 0 (0%) 12 (15%) 0 (0%) 172 (39%) 13 (65%) 68 (8%) 5 (13%) 40 (9%) 4 (22%) 353 (13%) 23 (19%)
 First child welfare intervention before age 12 37 (8%) 1 (6%) 28 (7%) 1 (8%) 8 (6%) 0 (0%) 6 (7%) 0 (0%) 72 (16%) 1 (5%) 78 (9%) 5 (13%) 38 (9%) 2 (11%) 267 (9%) 10 (8%)
 Parental social assistance recipiency 82 (18%) 2 (11%) 41 (11%) 1 (8%) 17 (13%) 3 (2%) 17 (21%) 0 (0%) 137 (31%) 8 (40%) 142 (16%) 7 (18%) 78 (17%) 3 (17%) 514 (18%) 24 (20%)
Familial factors                                
 First degree relative hospitalized for schizophrenia or bipolar disorder 31 (7%) 1 (6%) 42 (11%) 5 (38%) 2 (2%) 0 (0%) 2 (2%) 1 (20%) 10 (2%) 2 (10%) 48 (6%) 3 (8%) 25 (6%) 2 (11%) 160 (6%) 14 (12%)
 First degree relative hospitalized for substance use disorder 31 (7%) 2 (11%) 26 (7%) 1 (8%) 12 (9%) 1 (1%) 6 (7%) 0 (0%) 86 (19%) 6 (30%) 63 (7%) 7 (18%) 42 (9%) 2 (11%) 266 (9%) 19 (16%)
 First degree relative hospitalized for any psychiatric diagnosis 175 (38%) 8 (44%) 169 (44%) 9 (69%) 47 (35%) 3 (2%) 26 (32%) 1 (20%) 229 (51%) 11 (55%) 366 (42%) 26 (67%) 182 (41%) 9 (50%) 1194 (42%) 67 (55%)
 Parent hospitalized for any psychiatric diagnosis and convicted for a violent crime 27 (6%) 1 (6%) 16 (4%) 0 (0%) 6 (5%) 1 (13%) 6 (7%) 1 (20%) 55 (12%) 1 (5%) 57 (7%) 5 (13%) 33 (7%) 0 (0%) 200 (7%) 9 (7%)
 Parental suicide 41 (9%) 2 (11%) 37 (10%) 3 (23%) 24 (18%) 0 (0%) 9 (11%) 0 (0%) 52 (12%) 4 (20%) 80 (9%) 8 (21%) 36 (8%) 0 (0%) 279 (10%) 17 (14%)
Other risk factors                                
 Head injury 2 (0%) 0 (0%) 6 (2%) 0 (0%) 2 (2%) 0 (0%) 1 (1%) 0 (0%) 22 (5%) 2 (10%) 20 (2%) 1 (3%) 10 (2%) 0 (0%) 63 (2%) 3 (2%)
 Obstetrical complication 68 (15%) 4 (22%) 71 (18%) 1 (8%) 13 (10%) 0 (0%) 20 (25%) 1 (20%) 73 (16%) 3 (15%) 133 (15%) 8 (21%) 59 (13%) 1 (6%) 437 (16%) 18 (15%)

a Based on a grade scale from 1 to 5.

Table 2 displays the IRRs for suicide in relation to the various potential risk factors. The highest IRRs were observed in patients previously hospitalized because of self-harm (IRR 2.7, 95% CI 1.7–4.4). Having a history of a substance use disorder (IRR 1.7, 95% CI 1.1–2.6), conviction for a violent crime (IRR 2.0, 95% CI 1.3–3.2), as well as having a first degree relative hospitalized for schizophrenia, bipolar disorder (IRR 2.1, 95% CI 1.2–3.6) or substance use disorder (IRR 2.0, 95% CI 1.2–3.2) also constituted significant risk factors for subsequent suicide. Having complete grades from compulsory school also increased the suicide risk IRR: 1.6 (95% CI 1.4–1.9). The following risk factors did not show a statistically significant association with subsequent suicide: prior hospitalization with any mental disorder or depression, graduation from secondary school at age 20, having experienced child welfare intervention, having a parent diagnosed with any mental disorder or convicted for a violent crime or who had died by suicide, previous head injury, or being born with an obstetrical complication.

Table 2 Risk factors for suicide expressed as incidence rate ratios (IRRs) with 95% confidence intervals (CI) among patients diagnosed with a first episode psychosis.

Risk factors n IRR 95% CI a
Psychiatric factors
Prior hospitalization with a psychiatric diagnosis 45 1.3 (0.9–1.9)
Prior hospitalization for intentional self-harm 21 2.7 (1.7–4.4)
Prior hospitalization with a diagnosis for depression 10 1.5 (0.8–2.9)
Prior hospitalization with a diagnosis for substance use disorder 28 1.7 (1.1–2.6)
premorbid school performance
Complete grades from compulsory school 86 1.6 (1.4–1.9)
No graduation from secondary school at age 20 57 1.0 (0.7–1.4)
Social factors
Conviction for violent crime 23 2.0 (1.3–3.2)
First child welfare intervention before age 12 10 0.8 (0.4–1.6)
Parental social assistance recipiency 24 1.1 (0.7–1.8)
Familial factors
First degree relative with a diagnosis of schizophrenia or bipolar disorder 14 2.1 (1.2–3.6)
First degree relative with a diagnosis of substance use disorder 19 2.0 (1.2–3.2)
First degree relative with another psychiatric diagnosis 67 1.7 (1.2–2.4)
Parent hospitalized with a diagnosis of mental disorder and convicted for a violent crime 9 1.2 (0.6–2.4)
Parental suicide 17 1.5 (0.9–2.5)
Other risk factors
Head injury 3 1.2 (0.4–3.7)
Obstetrical complication 18 0.9 (0.6–1.5)

a The IRR reflects the relative risk of suicide among patients having vs. not having the respective risk factor.

Of the suicides in the cohort of patients with FEP, 11% were statistically attributable to prior hospitalization for intentional self-harm (population attributable fraction, PAF). PAF for prior hospitalization for substance use disorder or depression was estimated to be 10% and 3%, respectively.

4. Discussion

In this national population-based cohort study of consecutively admitted patients with a first-episode psychosis, we found that suicide rates were highest in depressive disorder with psychotic symptoms and in delusional disorder. The most prominent risk factors for subsequent suicide were having a history of self-harm, or having a conviction for violent crime. Further, having a first-degree relative with a history of schizophrenia, bipolar disorder or substance use disorder as well as having complete grades from compulsory school were significant risk factors for suicide.

Our finding that the suicide rate was highest in depressive disorder with psychotic symptoms is in agreement with many previous reports of depression being a highly potent risk factor for suicide ( Hawton et al., 2005 ), although studies of suicide in specific subtypes of depression are few (Black et al, 1987 and Rihmer et al, 1990). As for the risk of completed suicide in delusional disorder, for which we found the second highest point estimate, we have not found other comparable reports. However, in a meta-analysis of schizophrenia, delusions were not found to increase the suicide risk in schizophrenia ( Hawton et al., 2005 ). One explanation for our finding may be that delusions confer a significant risk increase for suicide in the early course of psychosis, whereas the relative contribution of delusions to the full risk profile among patients who have survived longer and developed schizophrenia is lower.

The overall suicide mortality rate of 4.3 per 1000 person years is comparable with earlier reports of a one year suicide mortality of 0.3%–0.4% in patients diagnosed with FEP (Nordentoft et al, 2002 and Addington et al, 2004) although other studies have found higher rates ( Fedyszyn et al., 2012 ). Our result of a total of 4.7% of male and 3.8% of female patients who died by suicide is slightly lower than what was found in Danish long term follow-up study of up to 36 years where suicide rates in schizophrenia and other schizophrenia like disorders were 6.6% and 5.9%, respectively, in men and 4.9% and 4.1%, respectively, in women ( Nordentoft et al., 2011 ). Conversely, the overall proportion (4.3%) of patients who died by suicide in our study is slightly higher than the 1.9% reported among FEP patients in the United Kingdom ( Dutta et al., 2010 ). Nevertheless, contemporary reports of suicide mortality in FEP appear to agree with recent estimations of a lifetime suicide rate in schizophrenia of about 5%, rather than previously communicated higher rates in this disorder ( Palmer et al., 2005 ).

Concurring to our results, a history of self-harm has been found to be a risk factor for subsequent suicide in many populations (Harris and Barraclough, 1997, McGirr et al, 2006, Hawton and van Heeringen, 2009, and Fedyszyn et al, 2012), including patients with FEP (Bertelsen et al, 2007, Harvey et al, 2008, and Nordentoft et al, 2002). The same is true for history of substance use disorders ( Harris and Barraclough, 1997 ). Also in agreement with our finding of a higher suicide risk in patients convicted for a violent crime, previous studies have found not only a higher risk of suicide in the general population with a history of juvenile delinquency but also that a major proportion of the risk increase was due to concomitant mental illness and substance use disorder (Fazel et al, 2009 and Bjorkenstam et al, 2011b). Nevertheless, previous research suggests that there is a higher suicide risk connected with delinquency in itself (Gallagher and Dobrin, 2006 and Thompson et al, 2006). A Swedish cohort study of patients with schizophrenia found that violent offending was a significant risk factor for suicide, but only in men ( Webb et al., 2011 ).

A potential common denominator for having a history of self-harm, substance use disorder, and criminal delinquency, may be that they all include some degree of impulsivity, which in turn is associated with an increased risk of suicide (Mann et al, 1999, Swann et al, 2005, and McGirr et al, 2008). Swann et al. (2005) found that a history of severe suicidal behavior in patients with bipolar disorder was associated with impulsivity, manifested as a tendency toward rapid, unplanned responses. The patterns of suicide risk factors found in our present study may suggest that also subsets of patients with FEP may be prone to impulsive behaviors. This is also supported by the population attributable fractions calculated in our study which indicated that a history of hospitalization for self-harm or substance abuse was associated with 11% and 10% of the suicides, respectively.

Our finding of a higher suicide risk among patients who had completed compulsory school is comparable with prior results in schizophrenia ( Reutfors et al., 2009 ) and among inpatient treated psychiatric patients in general ( Agerbo, 2007 ). These results are reverse to findings in the general population where poorer school performance has been shown to increase the suicide risk (Alaraisanen et al, 2006 and Bjorkenstam et al, 2011a). It has been speculated that patients with higher education may feel more stigmatized and shameful when developing a mental disorder ( Agerbo, 2007 ), and that this may lead to higher suicide risk.

We also identified higher suicide risk among patients with a family history of a first-degree relative hospitalized for schizophrenia or bipolar disorder, substance use disorder or other mental disorder that had required inpatient treatment. Similar associations have been reported in some studies (Gould et al, 1998 and Brent and Mann, 2005), although such an association may not be evident among individuals with schizophrenia and other severe mental illness (Qin et al, 2002, Hawton et al, 2005, and Reutfors et al, 2009). Contrasting with findings in the general population ( Fazel et al., 2014 ) we did not find head injury to be a risk factor for suicide. One reason for this may be that the substantially increased baseline suicide risk among patients with FEP may override other relatively weaker risk factors, and make them more difficult to detect in this population.

We wish to acknowledge a number of limitations of the study. The validity of all types of FEP diagnoses has not been evaluated against DSM criteria for the various FEP. However, previous validations of the clinical schizophrenia diagnoses in Swedish registers have indicated high agreement with DSM-III, DSM-IIIR, and DSM-IV ( Ludvigsson et al., 2011 ). Because we were restricted to study the variables available in population registers, unrecorded factors, such as social support, clinical symptoms, type and adherence to medication as well as other interventions were not studied. Further, we do not know the exact timing of onset of psychosis symptoms, or the duration of untreated psychosis. As our study population included only inpatients, it is likely that a number of individuals had committed suicide before they were admitted to hospital, although they may have been diagnosed with FEP as outpatients. Nevertheless, our study is likely to encompass a relatively large number of individuals who may have committed suicide before a diagnosis of schizophrenia. Finally, it should be recognized that the suicide rates in the various FEP subtypes in our study did not differ statistically significantly from each other; this may be partly due to the relatively small number of suicides in certain diagnostic groups.

To the best of our knowledge, no earlier study has reported suicide rates and risk factors for completed suicide in a national population-based cohort of patients with FEP. Compared to previous studies in this field, our study adds precision in the risk estimates of suicide risk in various subtypes of FEP. That the highest point estimate for suicide rate in our study was found for depressive disorder with psychotic symptoms highlights the importance of depression as a risk factor for suicide in psychotic disorder. Impulsivity may be a common denominator for some of the factors associated with subsequent completed suicide.

In addition to family history of mental disorder and history of self-harm, being convicted of a violent crime is a prominent risk factor for suicide in FEP. Having complete grades from the nine year compulsory school may also indicate higher risk of suicide in this population. Identification of these risk factors may increase our understanding of the suicidal process in FEP and aid in focusing interventions to those at highest risk of suicide.

Role of funding



RB and JR originated the idea. EB and CB analyzed the data in consultation with RB and JR. CB and JR wrote the manuscript draft. All authors contributed in designing the study and to the interpretation of the results and to the writing and approval of the final article.

Conflict of interest

Author JR has been a speaker for Eli Lilly, has received unrestricted grant support from Schering-Plough, and has been in research collaboration with AstraZeneca and Janssen-Cilag for which Karolinska Institutet has received grant support. All other authors declare that they have no conflicts of interest.




  • Addington et al., 2004 J. Addington, J. Williams, J. Young, D. Addington. Suicidal behaviour in early psychosis. Acta Psychiatr. Scand.. 2004;109(2):116-120 Crossref
  • Agerbo, 2007 E. Agerbo. High income, employment, postgraduate education, and marriage: a suicidal cocktail among psychiatric patients. Arch. Gen. Psychiatry. 2007;64(12):1377-1384 Crossref
  • Alaraisanen et al., 2006 A. Alaraisanen, J. Miettunen, E. Lauronen, P. Rasanen, M. Isohanni. Good school performance is a risk factor of suicide in psychoses: a 35-year follow up of the Northern Finland 1966 Birth Cohort. Acta Psychiatr. Scand.. 2006;114(5):357-362 Crossref
  • Allebeck, 1989 P. Allebeck. Schizophrenia: a life-shortening disease. Schizophr. Bull.. 1989;15(1):81-89 Crossref
  • Berlin et al., 2011 M. Berlin, B. Vinnerljung, A. Hjern. School performance in primary school and psychosocial problems in young adulthood among care leavers from long term foster care. Child Youth Serv. Rev.. 2011;33(12):2489-2497 Crossref
  • Bertelsen et al., 2007 M. Bertelsen, P. Jeppesen, L. Petersen, A. Thorup, J. Ohlenschlaeger, P. le Quach, T.O. Christensen, G. Krarup, P. Jorgensen, M. Nordentoft. Suicidal behaviour and mortality in first-episode psychosis: the OPUS trial. Br. J. Psychiatry. 2007;(Suppl. 51):s140-s146 Crossref
  • Bjorkenstam et al., 2011a C. Bjorkenstam, G.R. Weitoft, A. Hjern, P. Nordstrom, J. Hallqvist, R. Ljung. School grades, parental education and suicide—a national register-based cohort study. J. Epidemiol. Community Health. 2011;65(11):993-998 Crossref
  • Bjorkenstam et al., 2013 E. Bjorkenstam, C. Bjorkenstam, A. Hjern, J. Reutfors, R. Boden. A five year diagnostic follow-up of 1,840 patients after a first episode non-schizophrenia and non-affective psychosis. Schizophr. Res.. 2013;150(1):205-210 Crossref
  • Bjorkenstam et al., 2011b E. Bjorkenstam, C. Bjorkenstam, B. Vinnerljung, J. Hallqvist, R. Ljung. Juvenile delinquency, social background and suicide—a Swedish national cohort study of 992,881 young adults. Int. J. Epidemiol.. 2011;40(6):1585-1592
  • Black et al., 1987 D.W. Black, G. Winokur, A. Nasrallah. Suicide in subtypes of major affective disorder. A comparison with general population suicide mortality. Arch. Gen. Psychiatry. 1987;44(10):878-880 Crossref
  • Brent and Mann, 2005 D.A. Brent, J.J. Mann. Family genetic studies, suicide, and suicidal behavior. Am. J. Med. Genet. C: Semin. Med. Genet.. 2005;133C(1):13-24 Crossref
  • Craig et al., 2006 T.J. Craig, Q. Ye, E.J. Bromet. Mortality among first-admission patients with psychosis. Compr. Psychiatry. 2006;47(4):246-251 Crossref
  • Dutta et al., 2010 R. Dutta, R.M. Murray, M. Hotopf, J. Allardyce, P.B. Jones, J. Boydell. Reassessing the long-term risk of suicide after a first episode of psychosis. Arch. Gen. Psychiatry. 2010;67(12):1230-1237 Crossref
  • Dutta et al., 2011 R. Dutta, R.M. Murray, J. Allardyce, P.B. Jones, J. Boydell. Early risk factors for suicide in an epidemiological first episode psychosis cohort. Schizophr Res.. 2011;126(1–3) 11–9. (Mar)
  • Fazel and Grann, 2006 S. Fazel, M. Grann. The population impact of severe mental illness on violent crime. Am. J. Psychiatry. 2006;163(8):1397-1403 Crossref
  • Fazel et al., 2009 S. Fazel, N. Langstrom, A. Hjern, M. Grann, P. Lichtenstein. Schizophrenia, substance abuse, and violent crime. JAMA. 2009;301(19):2016-2023 Crossref
  • Fazel et al., 2014 S. Fazel, A. Wolf, D. Pillas, P. Lichtenstein, N. Langstrom. Suicide, fatal injuries, and other causes of premature mortality in patients with traumatic brain injury: a 41-year Swedish population study. JAMA Psychiatry. 2014;71(3):326-333 Crossref
  • Fedyszyn et al., 2012 I.E. Fedyszyn, J. Robinson, M.G. Harris, S.J. Paxton, S. Francey. Predictors of suicide-related behaviors during treatment following a first episode of psychosis: the contribution of baseline, past, and recent factors. Schizophr. Res.. 2012;140(1–3):17-24 Crossref
  • Foley et al., 2008 S. Foley, D. Jackson, S. McWilliams, L. Renwick, M. Sutton, N. Turner, A. Kinsella, E. O'Callaghan. Suicidality prior to presentation in first-episode psychosis. Early Interv. Psychiatry. 2008;2(4):242-246 Crossref
  • Gallagher and Dobrin, 2006 C.A. Gallagher, A. Dobrin. Deaths in juvenile justice residential facilities. J. Adolesc. Health. 2006;38(6):662-668 Crossref
  • Gould et al., 1998 M.S. Gould, R. King, S. Greenwald, P. Fisher, M. Schwab-Stone, R. Kramer, A.J. Flisher, S. Goodman, G. Canino, D. Shaffer. Psychopathology associated with suicidal ideation and attempts among children and adolescents. J. Am. Acad. Child Adolesc. Psychiatry. 1998;37(9):915-923 Crossref
  • Harris and Barraclough, 1997 E.C. Harris, B. Barraclough. Suicide as an outcome for mental disorders. A meta-analysis. Br. J. Psychiatry. 1997;170:205-228 Crossref
  • Harvey et al., 2008 S.B. Harvey, K. Dean, C. Morgan, E. Walsh, A. Demjaha, P. Dazzan, K. Morgan, T. Lloyd, P. Fearon, P.B. Jones, R.M. Murray. Self-harm in first-episode psychosis. Br. J. Psychiatry. 2008;192(3):178-184 Crossref
  • Hawton et al., 2005 K. Hawton, L. Sutton, C. Haw, J. Sinclair, J.J. Deeks. Schizophrenia and suicide: systematic review of risk factors. Br. J. Psychiatry. 2005;187:9-20 Crossref
  • Hawton and van Heeringen, 2009 K. Hawton, K. van Heeringen. Suicide. Lancet. 2009;373(9672):1372-1381 Crossref
  • Leslie et al., 2010 L.K. Leslie, S. James, A. Monn, M.C. Kauten, J. Zhang, G. Aarons. Health-risk behaviors in young adolescents in the child welfare system. J. Adolesc. Health. 2010;47(1):26-34 Crossref
  • Linsley et al., 2001 K.R. Linsley, K. Schapira, T.P. Kelly. Open verdict v. suicide — importance to research. Br. J. Psychiatry. 2001;178:465-468 Crossref
  • Ludvigsson et al., 2011 J.F. Ludvigsson, E. Andersson, A. Ekbom, M. Feychting, J.L. Kim, C. Reuterwall, M. Heurgren, P.O. Olausson. External review and validation of the Swedish national inpatient register. BMC Public Health. 2011;11:450 Crossref
  • Mann et al., 1999 J.J. Mann, C. Waternaux, G.L. Haas, K.M. Malone. Toward a clinical model of suicidal behavior in psychiatric patients. Am. J. Psychiatry. 1999;156(2):181-189
  • McGirr et al., 2008 A. McGirr, J. Renaud, A. Bureau, M. Seguin, A. Lesage, G. Turecki. Impulsive–aggressive behaviours and completed suicide across the life cycle: a predisposition for younger age of suicide. Psychol. Med.. 2008;38(3):407-417
  • McGirr et al., 2006 A. McGirr, M. Seguin, J. Renaud, C. Benkelfat, M. Alda, G. Turecki. Gender and risk factors for suicide: evidence for heterogeneity in predisposing mechanisms in a psychological autopsy study. J. Clin. Psychiatry. 2006;67(10):1612-1617 Crossref
  • Nielssen et al., 2012 O.B. Nielssen, G.S. Malhi, P.D. McGorry, M.M. Large. Overview of violence to self and others during the first episode of psychosis. J. Clin. Psychiatry. 2012;73(5):e580-e587 Crossref
  • Nordentoft et al., 2002 M. Nordentoft, P. Jeppesen, M. Abel, P. Kassow, L. Petersen, A. Thorup, G. Krarup, R. Hemmingsen, P. Jorgensen. OPUS study: suicidal behaviour, suicidal ideation and hopelessness among patients with first-episode psychosis. One-year follow-up of a randomised controlled trial. Brit. J. Psychiatry. 2002;(Suppl. 43):s98-s106
  • Nordentoft et al., 2011 M. Nordentoft, P.B. Mortensen, C.B. Pedersen. Absolute risk of suicide after first hospital contact in mental disorder. Arch. Gen. Psychiatry. 2011;68(10):1058-1064 Crossref
  • Osby et al., 2000 U. Osby, N. Correia, L. Brandt, A. Ekbom, P. Sparen. Mortality and causes of death in schizophrenia in Stockholm county, Sweden. Schizophr. Res.. 2000;45(1–2):21-28 Crossref
  • Palmer et al., 2005 B.A. Palmer, V.S. Pankratz, J.M. Bostwick. The lifetime risk of suicide in schizophrenia: a reexamination. Arch. Gen. Psychiatry. 2005;62(3):247-253 Crossref
  • Pompili et al., 2011 M. Pompili, G. Serafini, M. Innamorati, D. Lester, A. Shrivastava, P. Girardi, M. Nordentoft. Suicide risk in first episode psychosis: a selective review of the current literature. Schizophr. Res.. 2011;129(1):1-11 Crossref
  • Qin et al., 2002 P. Qin, E. Agerbo, P.B. Mortensen. Suicide risk in relation to family history of completed suicide and psychiatric disorders: a nested case–control study based on longitudinal registers. Lancet. 2002;360(9340):1126-1130 Crossref
  • Radomsky et al., 1999 E.D. Radomsky, G.L. Haas, J.J. Mann, J.A. Sweeney. Suicidal behavior in patients with schizophrenia and other psychotic disorders. Am. J. Psychiatry. 1999;156(10):1590-1595
  • Reutfors et al., 2009 J. Reutfors, L. Brandt, E.G. Jonsson, A. Ekbom, P. Sparen, U. Osby. Risk factors for suicide in schizophrenia: findings from a Swedish population-based case–control study. Schizophr. Res.. 2009;108(1–3):231-237 Crossref
  • Rihmer et al., 1990 Z. Rihmer, J. Barsi, M. Arato, E. Demeter. Suicide in subtypes of primary major depression. J. Affect. Disord.. 1990;18(3):221-225 Crossref
  • Rothman, 2013 K.J. Rothman. Modern Epidemiology. (Lippincott Williams and Wilkins, 2013)
  • Swann et al., 2005 A.C. Swann, D.M. Dougherty, P.J. Pazzaglia, M. Pham, J.L. Steinberg, F.G. Moeller. Increased impulsivity associated with severity of suicide attempt history in patients with bipolar disorder. Am. J. Psychiatry. 2005;162(9):1680-1687 Crossref
  • Thompson et al., 2006 M.P. Thompson, J.B. Kingree, C.H. Ho. Associations between delinquency and suicidal behaviors in a nationally representative sample of adolescents. Suicide Life Threat. Behav.. 2006;36(1):57-64 Crossref
  • Vinnerljung et al., 2006 B. Vinnerljung, A. Hjern, F. Lindblad. Suicide attempts and severe psychiatric morbidity among former child welfare clients — a national cohort study. J. Child Psychol. Psychiatry. 2006;47(7):723-733
  • Webb et al., 2011 R.T. Webb, N. Langstrom, B. Runeson, P. Lichtenstein, S. Fazel. Violent offending and IQ level as predictors of suicide in schizophrenia: national cohort study. Schizophr. Res.. 2011;130(1–3):143-147 Crossref


a Department of Clinical Neuroscience, Insurance Medicine, Karolinska Institutet, Stockholm, Sweden

b Department of Public Health Sciences, Karolinska Institutet, Stockholm, Sweden

c Centre for Health Equity Studies (CHESS), Stockholm University/Karolinska Institutet, Stockholm, Sweden

d Department of Medicine, Clinical Epidemiology Unit, Karolinska Institutet, Stockholm, Sweden

e Department of Neuroscience, Psychiatry, Uppsala University, Uppsala, Sweden

lowast Corresponding author at: Karolinska Institutet, Division of Insurance Medicine, SE-171 77 Stockholm, Sweden. Fax: + 46 524 862 06.