[Medscape Mental Health 3(3), 1998. © 1998 Medscape, Inc.]
Suicide is a major cause of death among patients with schizophrenia, with young adulthood to midlife being the age range of greatest risk. A past history of suicide attempts is common among schizophrenic suicide victims, as are affective symptomatology and feelings of hopelessness and demoralization. Active paranoia may increase risk, whereas negative symptoms may decrease the risk of self-destructive acts in schizophrenic patients. Many suicides occur during hospitalization or shortly after discharge. Methodical assessment of suicidal ideation and aggressive treatment with psychological, social, and pharmacologic approaches are vital aspects of patient management.
The design of effective suicide-prevention strategies hinges on the identification of risk factors characteristic of the individual or group, the strength of the causal relationship between the risk factor and the disease, and the alterability of that causal (risk) factor.[9] It is vital that efforts to prevent suicide in schizophrenic patients are tailored to the demographic and clinical characteristics of this population. Describing the clinical cases of 20 patients with schizophrenia or schizoaffective disorder who committed suicide, this review summarizes the risk factors for suicide in schizophrenia and their implications for preventive strategies.
Between August 1989 and March 1992, we used the PA method to study 141 individuals in Monroe County, N.Y., whose deaths had been ruled as suicides by the Office of the Medical Examiner. The methods we employed for identifying, enrolling, and studying these 141 subjects, along with their diagnostic distribution, have been described in detail elsewhere.[28] Fourteen (9.9%) received the diagnosis of schizophrenia and another 6 (4.3%) had schizoaffective disorder. Table II lists characteristics of each subgroup for the 20 persons having a schizophrenic disorder.
Of these 20 subjects, 15 (75%) were men, 16 (80%) were white, 2 were black, 1 was Hispanic, and 1 was Asian. Their mean age was 36.5 (± 12.0) years. Fourteen (70%) were single, 3 were married, 2 were divorced, and 1 was widowed. Five (25%) lived alone, while the remainder lived with other family or friends. They tended to use violent methods to take their lives: 6 (30%) used a firearm, 5 hanged themselves, 4 jumped from a height, and the remainder used a variety of other means (drowning, drug ingestion, carbon monoxide poisoning, exposure, and self-immolation).
Among the 14 subjects with schizophrenia, 7 had the paranoid subtype, 6 had undifferentiated schizophrenia, and 1 had the disorganized type. For those with available data (n=12), the mean duration of illness was 12.5 years. Of the victims with schizoaffective illness (n=6), 3 had the depressed type and 3 were bipolar. Their mean duration of illness was 8.7 years. Five were in a depressed or mixed phase at the time of death.
Seventeen of the 20 who committed suicide were known to have had active psychotic symptoms at the time of their suicides. Although depressive symptoms were common, separate affective disorder diagnoses were not. Active comorbid substance use disorders were present in 7 of the 20 suicides; an additional 7 had alcohol and/or other drug use disorders, which were in full remission. Informants and records documented a past history of suicide attempts in 15 victims (75%). Eighteen (90%) were in psychiatric treatment. Of 19 subjects for whom data were available concerning the timing of their last psychiatric contacts, 9 (47%) had been seen in the last week and 15 (79%) in the last month of life. None died during inpatient care, although 5 died within a month of hospital discharge, 3 within 1 week.
In nations that maintain such statistics, rates of suicide are reported to rise with age. In contrast, the risk for suicide among schizophrenic patients appears to decrease with older age.[29,30] Hu and colleagues[31] found that 35 (83%) of 42 schizophrenics who killed themselves did so before age 30. Of 19 schizophrenic suicides studied by King,[32] the mean age of death was 33.4 years, significantly lower than that for other patient groups. Data from record-linkage studies indicate that, with increasing age, schizophrenic persons are at a decreased risk.[29,33] Young adulthood, therefore, appears to be the point in the life course of persons with schizophrenia at which intervention strategies should be most aggressively targeted.
Gender
Whereas the ratio of men to women who commit suicide in the general population is approximately 4:1,[8] the ratio among schizophrenic patients seems somewhat narrower.[3,34] Among Taiwanese suicides with schizophrenia, Hu and associates[29] found a male-to-female ratio of 3:2. Goldstein and colleagues[30] found a similar ratio, whereas others have reported only slightly higher rates among men than women with the disorder.[35] In addition, age at suicide appears to vary by gender among persons with schizophrenia. The mean age at death among schizophrenic men is approximately 10 to 12 years younger than among schizophrenic women.[34,36] This difference likely reflects the earlier mean age of onset of schizophrenia in men and thus a similar duration of illness at the time of suicide by men and women.[37]
Other Factors
Some evidence supports that being unmarried, socially isolated, and unemployed are risk factors for suicide among schizophrenics as well. These same features, however, are common among all schizophrenics, for whom the illness typically strikes in young adulthood and is associated with impaired social and occupational functioning. Few studies are available that include nonsuicidal schizophrenic controls and measures of social, interpersonal, and occupational functioning sufficient to define each factor's contribution to suicide risk. Substantially more research is needed before demographic correlates of suicide among persons with schizophrenia may be established as risk factors or used to identify individuals who are particularly vulnerable. (See Modal Characteristics of Schizophrenics Who Commit Suicide.)
Characteristics of the Illness: Course/Duration
Studies indicate that schizophrenic patients who commit suicide tend to have a relatively more chronic course of illness, with acute exacerbations.[37-39] Among 92 schizophrenic suicides identified in the Finnish National Suicide Prevention Project, the mean duration of illness was 15.5 years, and victims had an average of 7.9 lifetime psychiatric admissions.[37] Havaki-Kontaxaki and colleagues[40] found a mean duration of illness of 19.3 (± 8.8) years for suicide completers, compared with 13.5 (± 9.9) years for schizophrenic controls, a significant difference. Cheng and others[6] found that, compared with nonsuicidal schizophrenics, suicide completers had more frequent psychiatric hospitalizations. Westermeyer and Harrow[39] suggested that a gradual onset of illness over time may also place schizophrenic patients at risk for suicide.
A large proportion of suicides among schizophrenics occur during or shortly after hospitalization. Allebeck and Wistedt[35] found that 34 of 63 suicides took place during or shortly after psychiatric care. Hu and coworkers[31] found that 81.1% of Taiwanese schizophrenic individuals who committed suicide were engaged in either inpatient or outpatient treatment during their final month of life, and 31% of suicides occurred while the patient was receiving inpatient psychiatric care. In a Finnish sample, 45 (51%) of 89 schizophrenics who committed suicide had their last contact with a health care provider within 4 days of death, 70% within 2 weeks, and 82% within a month.[37] Caldwell and Gottesman[3] concluded that 88.1% of schizophrenics who committed suicide were in psychiatric care at the time of death.
Of those suicides occurring during inpatient stays, the majority are among patients who are in the first week of admission or are to be discharged shortly.[31,41,42] Lloyd[41] also observed that inpatient suicides tend to occur in clusters, suggesting that there may be a heightened risk among schizophrenics when a fellow inpatient has committed suicide.
Illness Subtype
The few data available concerning suicide risk among specific subtypes of schizophrenia indicate that paranoid patients may be particularly vulnerable.[43,44] Fenton and McGlashan[45] examined the association between suicide and diagnostic subtype among 187 patients with schizophrenia spectrum disorders (schizophrenia, schizoaffective disorder, schizophreniform disorder, and schizotypal personality disorder). During a follow-up period of 6 to 32 years, 19 patients (10.2%) committed suicide. Thirteen (12%) of 112 patients with the paranoid subtype had taken their own lives, a significantly greater proportion than of hebephrenic and undifferentiated schizophrenics. Of special interest, patients who committed suicide also had significantly lower global negative symptom severity ratings at admission than did patients who survived throughout the follow-up period; this group also had greater severity of delusions and suspiciousness.
The authors concluded that negative symptoms such as blunted affect, diminished drive, and social withdrawal may be associated with reduced risk in schizophrenic patients, while prominent suspiciousness in the absence of deficit symptoms defines a relatively high-risk group.
Previous Violent Acts
As was observed in our sample, a past history of suicide attempts is common in completed suicides with schizophrenia,[6,34,40] and may be the most prominent predictor of subsequent death.[34] Seventy-one percent of Finnish schizophrenics who committed suicide had a past history of suicide attempts.[37] Allebeck and colleagues[42,46] noted this association to be especially strong among women.
Many theorists regard suicide as a form of aggression, the outward expression of which may presage self-destruction. Cheng and associates[6] found that schizophrenics who committed suicide exhibited more violent acts immediately before their last admission than did controls, while Havaki-Kontaxaki and coworkers[40] noted a significant association between homicide and subsequent suicide in a sample of inpatient schizophrenics. Studies using more specific measures of aggression and impulsiveness in schizophrenics who commit suicide have yet to be reported.
Hopelessness
Among adult clinical populations, hopelessness has been shown to correlate better with suicidal intent[47-50] and subsequent suicide[51,52] than does the severity of depression. The same relationships do not appear to hold true in adult alcoholics,[53] and the role of hopelessness in schizophrenia requires further study. Drake and Cotton[47] found that hopelessness predicted subsequent completed suicide in patients with chronic schizophrenia, which the investigators linked to increased illness awareness.[38] Although empirical data are few, it is likely that schizophrenics with a higher premorbid level of functioning are at greater risk.[34,38,39] Having formed socially congruent expectations for their futures, following the onset of illness they become despairing about its chronicity, progression, and global impact on their lives. A hopeless state, for which suicide is a perceived answer, results.
Comorbid Illness
Affective disorder is strongly correlated with suicide in people with schizophrenia.[34,54] Secondary depression follows diminution of psychotic symptoms in 25% of schizophrenic patients, and 60% suffer a major depressive episode at some point in their illness.[55] Roy[54] reviewed the data from 9 studies to show that 160 (59.3%) of 270 suicide completers were depressed at the time of their deaths. The subjective distress experienced by these patients may go unrecognized by their clinicians,[56] perhaps ascribed to psychosis or neuroleptic treatment rather than a distinguishable affective syndrome per se. Further research is needed to clarify the role of depression as a precipitant to suicidal behavior in schizophrenia.
The incidence of alcoholism in schizophrenic patients is higher than in the general population,[57] and its presence has been shown to distinguish suicide completers from schizophrenic controls in some studies[58,59] but not in others.[56] Similarly, mixed results have been reported for the association between suicide and other drug-use disorders.[5] These discrepancies may be explained, in part, by failure to distinguish active from remitted substance abuse. The striking prevalence in our sample of alcohol and drug-use disorders in remission warrants their further study as indicators of vulnerability to suicide.
The effect of physical illness on suicide in schizophrenics is not clear. Although control studies are few, there appear to be associations in the general population between suicide and cancer, diseases of the central nervous system, peptic ulcer disease, cardiopulmonary complications, rheumatoid arthritis, and urogenital disease in men.[60-62] Since individuals with schizophrenia are at greater risk for premature death from physical illness, it would be important to know whether this vulnerability contributes to their elevated risk for suicide. We are not aware of any studies that address this question.
Stressful life events. Stressful life events such as separation or divorce from a spouse, perceived or real abandonment by parents, broken relationship with a significant other, rehospitalization or discharge, change in therapists, or loss of a job have been found to precede depression and suicide in schizophrenic patients.[34] Drake and colleagues[55] found the most frequent life events prior to schizophrenic suicide completion were leaving the hospital and loss of support from family. However, Modestin and associates[63] reported no difference between normal controls and schizophrenic suicide completers in the number of antecedent life events. The role that depression, hopelessness, and positive and negative schizophrenic symptomatology may play -- either in the precipitation of these events or as mediators of the behavioral response to them -- remains uncertain.
Neurobiologic Factors
A substantial and rapidly expanding body of literature indicates an association between suicidal behavior and alterations in a range of neurochemical parameters. The serotonergic system has been the most intensively investigated. Studies of postmortem tissue have showed lower levels of serotonin (5-hydroxytryptamine, or 5-HT) and its principal metabolite (5-hydroxyindoleacetic acid, or 5-HIAA) in brain stem,[64-67] raphe nuclei,[68] putamen,[69] and hypothalamus. Also shown were decreased binding of radio-labeled ligand at the presynaptic serotonin reuptake site in the frontal cortex,[70-74] and increased density of postsynaptic 5-HT2 receptors in the same brain areas of suicide victims compared with controls.[72,75-78]
Studies of suicide attempters have repeatedly found significantly lower cerebrospinal fluid (CSF) levels of 5-HIAA than in normal controls, a finding demonstrated not only in patients with depression[74-82] but also in individuals with schizophrenia[83,84] and other psychiatric diagnoses.[85] Alterations of central serotonin measures in arsonists[86] and other violent offenders[87] have led most researchers to conclude that abnormal indolamine function predisposes to impulsive, aggressive behavior rather than suicide per se.
Although relatively few studies have specifically examined the association of serotonergic function with suicidal behavior in schizophrenics, serotonin's putative role in the pathogenesis of schizophrenia indicates the need for further study in this area. As with other diagnostic groups, schizophrenics who made violent suicide attempts have been shown to have significantly lower CSF 5-HIAA levels than nonsuicidal patient controls in some[84] but not all[88] studies. Schizophrenic patients demonstrate abnormalities of electroencephalographic (EEG) sleep, including reduced rapid eye movement (REM) latency, increased REM and decreased slow-wave sleep, and impaired sleep maintenance and efficiency. Keshavan and colleagues[89] compared suicidal with nonsuicidal schizophrenic patients and found that those with suicide attempts had significantly greater REM activity. They postulated that sleep EEG changes may be predictive for suicide risk in this population.
Alterations in other neurochemical systems have been associated with suicidal behavior as well.[90] Dopamine, because of its probable pathogenic role in schizophrenia, is of greatest interest. There is a close association between CSF levels of 5-HIAA and the major dopamine metabolite, homovanillic acid (HVA). It is unclear whether this close correlation is due to a shared transport mechanism or to a functional connection between the parent amines.[91] Several investigators have reported low levels of CSF HVA that parallel decreases in 5-HIAA in suicidal patients.[92-94] The few postmortem studies that included indices of dopaminergic function have failed to demonstrate differences between suicide victims and controls.[65,95,96] We are aware of no studies, however, that have included sufficient numbers of schizophrenic suicides and controls to investigate the role of dopaminergic, or other neurochemical, systems more specifically.
In addition to its pathogenic role in schizophrenia, degenerative change of the central dopaminergic system has been implicated in both Huntington's and Parkinson's diseases, illnesses in which suicide risk is also markedly elevated.[97,98] Further work in this area is needed.
The close association between suicidal behavior and affective symptomatology in schizophrenic patients underscores the need for effective diagnosis and treatment of depressive illness. Clinicians must remain especially alert so as not to confuse acute-phase or postpsychotic depressive symptomatology with antipsychotic side effects, and they must ensure that appropriate antidepressant therapy is initiated.
Hopelessness may be an especially important marker for the need to treat a depressive syndrome aggressively. Although negative symptomatology and the deficit syndrome may be associated with reduced suicide risk in schizophrenic patients,[45] the presence of these conditions should not reduce a clinician's vigilance, because they too may mask a clinically significant depressive syndrome.
Twenty percent of inpatients and 40% to 70% of outpatients with schizophrenia are estimated to be medication-noncompliant, the most common cause of relapse.[99] No study of which we are aware has yet examined the association between medication noncompliance and suicide risk. However, because each of these may reflect the demoralization and frustration of having a chronic and deteriorating illness, such investigation is warranted.
Similarly, little is known about the effect of antipsychotic medications on suicidal behavior, independent of their effects on the primary symptom complex of psychosis. One recent study has addressed this issue. Meltzer and Okayli[100] assessed suicidality in a prospective follow-up of 88 neuroleptic-resistant schizophrenics who were treated with clozapine. They found that suicide attempts were dramatically and significantly reduced following initiation of this drug.
Behavioral and Psychosocial Treatment
Research on behavioral and psychosocial approaches to management of suicide risk in patients with schizophrenia is sparse. Drake and colleagues[55] underscored the value of empathic support in diminishing suicide risk. They advise that the clinician acknowledge the patient's despair, address his or her losses, and help establish new, accessible goals and tasks. Families can provide support, prevent social isolation, and maintain a stable, accepting environment as key components of psychosocial treatment. A cognitive approach may be useful in helping the patient recognize suicidal urges and acknowledge them to his or her health care providers.
As noted, discharge is an especially tenuous time through which close patient-therapist alliances must be maintained. Rehabilitative measures and supportive therapeutic contacts must be established; symptoms of depression, anxiety, and hopelessness should be closely monitored; and when crises occur, hospital readmission should be considered.
Treatment of the suicidal state, and for the schizophrenic illness more generally, requires thoughtful coordination of pharmacologic and psychosocial therapies. Drug treatment must target both the acute and chronic residual schizophrenic symptoms as well as associated or secondary syndromes, such as depression. Supportive psychotherapeutic approaches may include education for both the patient and his or her family about suicide risk and warning signs. Furthermore, psychotherapeutic interventions are critical in helping the patient define realistic expectations and optimize compliance with treatment (including social and occupational rehabilitation).
Treatment of suicidal states in schizophrenic patients is, as with every other aspect of treatment in this complex and debilitating disease, a multifaceted endeavor. Additional studies are needed to refine our knowledge of risk factors, and to enable the development of more efficiently targeted and effective suicide prevention and treatment strategies.
This work was supported in part by the NIMH (Grant #T32 MH18911, Eric D. Caine, MD, Principal Investigator). The authors wish to thank Nicholas Forbes, MD, Chief Medical Examiner of Monroe, County, N.Y., and his staff for their assistance in the study of suicide, and Marge Roberts for her help in manuscript preparation.