Abstract
Keywords
The phenomenon of suicide has sustained interest in the general and scientific community over many decades. In most developed countries, suicide is ranked among the top 10 causes of mortality [1]. The topic provokes a myriad of emotional, legal and ethical implications making it a controversial issue in many societies.
Suicide: an overview
Despite extensive research into suicide, interpretation of results and comparison across studies remain problematic. This is due in part to the variability between and within countries in the diagnostic and certification processes, which determine what constitutes a ‘suicide’ [1]. In most developed countries, causes of death, particularly sudden and unexpected deaths are ruled on by the coroner or similar authorities. The official figures on suicide thus reflect the laws which govern this process. Because of the controversial legal and moral implications of suicide, most laws are strict about the process of delivering a verdict of suicide and require unequivocal evidence before a ruling can be made [1]. In Australia for instance, only two out of the seven states routinely use suicide verdicts while in the remainder of the states, official pronouncements of suicide is discouraged [2].
Terminology
For the purposes of this review, the definition of suicide or completed suicide will be taken to mean the voluntary and intentional taking of one's own life [3]. Euthanasia or assisted suicide is beyond the scope of this review. Other forms of suicidal behaviour that will be discussed here include attempted suicide, deliberate self-harm (DSH) and suicidal ideation. Attempted suicide refers to deliberate, self-injurious behaviour (usually involving non-lethal methods) with non-fatal outcome, but for which there was an intent on the part of the person to kill himself or herself [4]. It is important to differentiate attempted or failed suicide from deliberate selfharm (DSH) through either overdose or self-injury where for the majority of cases, there is either no intent to die, or intent is low or ambivalent [5]. Suicidal ideation refers to all thoughts that may be interpreted through behaviour to endanger or threaten one's own life [3].
Suicidal behaviour in people with HIV/AIDS
In the early eighties, AIDS was regarded not only as a terminal illness but one which produced an enormous emotional and psychological burden on the patient and family, compounded by the extra burden of the social stigmatization of the disease.
Results of studies which have investigated suicide and related behaviours in HIV-seropositive individuals, must be interpreted cautiously. In addition to the problems related to the reporting of suicide statistics in general, is the additional problem of knowledge of a suicide victim's HIV status. In some countries, the HIV status of all ‘unnatural deaths’ is determined, but this is not a universal practice.
Research into suicidal behaviour in HIV-seropositive persons includes studies determining the frequency and ‘causes’ of suicidal ideation, suicide attempts and completed suicide. A variety of perspectives have been employed to investigate the association of suicide and HIV. Prevalence studies of completed suicide have primarily involved register-based surveys and psychological autopsy examinations. Evaluation of suicidal ideation and suicide attempts in live samples include retrospective assessments based on psychiatric case notes and/or in conjunction with semi-structured interviews or self-reports via questionnaires. Some studies have collected their data based solely on self-reports and structured questionnaires.
Suicidal ideation in people with HIV/AIDS
Studies investigating the prevalence of suicidal ideation in people with HIV/AIDS are presented in Table 1.
Studies investigating suicidal ideation prevalence in HIV-seropositive individuals
There is no consensus about the association between suicidal ideation and being HIV seropositive. Some studies have suggested that suicidal ideation is directly linked to HIV serostatus [6]. Others have found that suicidal thoughts have been context specific and the pathway to suicidal intent is psychologically rather than biologically driven [7]. Psychological variables such as current stressors [e.g. having a partner with AIDS or AIDS-related complex (ARC), unemployment or bereavement], poor adaptive functioning, hopelessness, higher neuroticism and lower social support have been identified as predictors of higher levels of suicidal ideation in HIVseropositive individuals compared with HIV-seronegative controls [8], [9]. Comparisons between studies presented in Table 1 suggest there is mixed evidence that acquiring HIV/AIDS per se increases the risk of suicidal ideation. Some studies have reported little to no differences in suicidal behaviour between HIV seropositive and seronegative groups [7],[10–12]
Others have reported a higher rate of suicidal ideation in persons with HIV/AIDS compared with those of unknown serostatus or who were HIV seronegative [6], [8], [16], [19], [22]. It is also unclear what influence the progression of HIV illness has on the risk of suicidal ideation. Some authors reported no differences between groups at any stage of the illness [10], [13], [23]. Others found that asymptomatic HIV seropositive patients showed more suicidal behaviour than those with advanced illness [15], [17], [18], [20] while conversely, advanced illness was reported to be associated with greater suicidal behaviour [21], [24]. The period immediately after HIV diagnosis has been made has been identified as a high-risk period for suicidal behaviour [25], [26]. However, one study, which assessed participants pre- and post-HIV testing, found no significant differences in suicidal ideation between the group that tested seropositive and those whose result was negative [12]. Another group similarly reported that a positive HIV result did not account for an increased risk of suicide in a sample of injecting drug users [27].
Methodological problems associated with study designs, sampling strategies, and definition of outcome measures are some issues of concern. Most studies are retrospective in nature and therefore data is not necessarily corroborated. Variability across studies in the types of scales or questionnaires used makes it difficult to compare results. For instance suicidal ideation was assessed using single items on questionnaires by some researchers [13], [14], [24]. Others used case notes from psychiatric consultations which involve subjective judgement on the part of assessors [10], [11], [15], [17],[19–23]. Suicidal ideation and suicide attempts are not often separately assessed but investigated jointly as suicidal behaviour. Participants in most of these studies are often convenience groups comprising psychiatric inpatient groups and homosexual males who have been reported to experience high lifetime rates of alcohol and substance abuse [28–31] and depressive symptoms [32] regardless of HIV serostatus [33]. Of interest here, however, is the finding that suicidality remains significantly associated with homosexuality even when controlling for substance abuse and depressive symptoms [34]. It should be noted here that the association between homosexuality and higher suicidal behaviour has been mainly reported in studies conducted in developed countries and may not necessarily generate to other societies. However, dearth of data from developing societies (particularly some Asian and African countries where HIV infection is common) does not allow for unequivocal conclusions to be made on the association between same-gender sexual orientation and increased suicidality. Another common problem with many of the studies is the lack of appropriate control groups. Many studies did not include control groups, or included individuals of unknown serostatus in their control samples [10], [15], [19].
Suicidal ideation in HIV seropositive females
The literature on suicidal behaviour in HIV seropositive females is not as extensive as that on males. Brown and Rundell [35] compared males and females in their study assessing psychiatric morbidity in HIV-seropositive US Air Force personnel. Only 6.7%% (n = 1) of the 15 female personnel who participated in the study admitted to mild suicidal ideation but did not make any plans, attempts or gestures. By contrast, 21%% of the male participants reported suicidal thoughts, attempts or plans since learning of their seropositivity. The mean period of time that the male and female groups had first learnt of their seropositivity was approximately equal at 13.0 and 12.8 months, respectively. There were two known cases of completed suicide among the HIV-seropositive male personnel but none among the females. The extent of the difference in suicidal ideation between both sexes surprised the authors as previous research indicates that females are three times more likely to attempt suicide than males. They suggested that several characteristics of the female sample may have been buffers against the morbidity of HIV. Only one of the 15 had AIDS, the rest were asymptomatic. All were physically healthy and none had a known history of intravenous drug use (IVDU) or alcohol abuse. There were as many black people (47%%) as white people (53%%) and the average age of the sample was 26.5 years. Prospectively, suicidal behaviour of HIV-seropositive females does not appear to increase although there is greater likelihood of having a psychiatric diagnosis at follow up [36]. Suicidal behaviour in HIV-seropositive females appears to be more influenced by a history of drug abuse than being HIV seropositive [37]. Suicidal concerns appear to be less significant for minority women than non-minority, and suicidal behaviour in HIV-seropositive females has been found to be significantly associated with finance, housing, employment, isolation and lack of social support and relationship issues rather than concern about being HIV seropositive [38].
Suicide attempts and deliberate self-harm in persons with HIV/AIDS
Many of the studies examining the characteristics of suicide attempts and DSH in HIV-seropositive individuals have found that the presence or absence of psychiatric morbidity, substance abuse and previous suicide attempts were more predictive of the acts occurring than being HIV seropositive per se [26], [37], [39]. However, it is important to note that investigating suicide attempts among HIV-seropositive individuals with drug using habits is difficult because of the need to distinguish between accidental versus deliberate overdose.
James and colleagues [37] found that in a group of 55 pregnant HIV-seropositive females, significantly more suicide attempts were made by those with a history of drug abuse (21.1%%) than those without (2.8%%). Simoni and colleagues [40], however, did not find similar results. They surveyed 230 women with HIV/AIDS and found that while the suicide attempt rate was high in the group (26%% pre-HIV and 19%% post-HIV), a history of lifetime or current substance abuse was not significantly associated with the attempts. Sixty-six percent reported a history of heavy drug use. Instead, depression, greater number of disease symptoms, loneliness, need for support and being younger were more predictive of suicide attempts post-HIV diagnosis. More than 50%% of those who attempted suicide after becoming HIV seropositive had a history of attempts prior to HIV diagnosis. The authors suggested that the risk of suicide was compounded for this group of women derived from a poor inner-city sample.
Gala and colleagues [39] reported that HIV-seropositive individuals with a past history of psychiatric illness were 7.7 times at greater risk of DSH than those without a psychiatric history. Past DSH also increased the risk fivefold of future acts of DSH. Rundell et al.[26] compared the psychiatric data of 15 HIV-seropositive suicide attempters with 15 seropositive non-attempters matched for age, sex and race. All were Air Force personnel. Examination of the psychiatric data revealed a significantly higher prevalence of alcohol abuse and psychiatric morbidity among the attempters compared to the non-attempters. There were also significant correlations between the suicide attempts and stressful life events such as separation from significant others and change of jobs because of HIV serostatus. However, the study found that the attempted suicide rate among the HIVseropositive personnel was 16–24 times higher than the rate for the entire US Air Force. It is possible that HIVseropositive individuals who attempt suicide may already be vulnerable prior to becoming infected and that contracting the HIV illness confers an added burden [20], [21].
Completed suicide in people with HIV/AIDS
Studies of completed suicide include case reports, register studies and psychological autopsy studies.
Case reports
Copeland [41] investigated 25 AIDS-related cases of suicide from 1985 to 1989 in a Florida county. Diagnosis of AIDS or fear of AIDS played a role in all the suicides. All of the suicide victims were male, 88%% were white. The mean age was 38 years with the ages ranging from 29 to 59 years. More than 90%% of the victims were homosexual. Post-mortem results indicated the presence of illicit drugs in four of the victims. There was a wide range of methods of suicide including overdosing, shooting, hanging, jumping, carbon monoxide poisoning and cutting of wrists. Death by shooting was the most common method chosen. On review, it was found that 64%% had been depressed at the time of their death. A history of previous suicide attempts was noted in four cases.
A similar profile of six AIDS-related suicides in London was reported by Pugh and colleagues [42]. All were male and five were homosexual. A history of psychiatric illness, predominantly depression, was noted in four cases. Of the two without a psychiatric history, it was believed that the main motive for the suicides was ‘to be released from a debilitating condition’ as five of the cases had advanced symptomatic HIV disease (as defined in the Centers for Disease control HIV Classification, Stage IV (CDC-IV)) at the time of death. Two had made previous suicide attempts, which predated the diagnosis of HIV.
Suicidality can also be exacerbated by HIV dementia, which is associated with labile mood, behavioural disinhibition, impaired judgement and impulsivity. Alfonso and Cohen [43] reported two cases of HIV-seropositive individuals in their 30s who exhibited suicidal behaviour. Psychiatric evaluation revealed symptoms consistent with dementia including disorientation, concrete thinking, global memory dysfunction, affective lability and suicidal ideation. Computed tomographic (CT) head scans of both individuals indicated cortical atrophy. Both individuals had a history of substance use, which may also have increased the risk of suicidal behaviour. The authors noted the need for studies to further investigate the association between HIV dementia and suicidality.
Kirchner [44] reported a case of suicide in a man who had been diagnosed with HIV/AIDS 18 months previously with no history of psychiatric illness. He suggested that this may have been the result of an early organic psychosis (although a detailed histological examination of his brain was not conducted), and in his discussion raises the possibility of a variety of neuropsychiatric syndromes due HIV/AIDS which may contribute to suicidal behaviour. Although some of the cases described by Copeland [41] and Pugh and colleagues [42] had AIDS, neither specifically comment on the possible contribution of neuropsychiatric problems such as delirium to the suicides they describe.
Halstead and colleagues [45] noted in one of the five cases they describe of an individual with no known family or personal psychiatric history and with good premorbid personality who committed suicide after being admitted to hospital for erratic behaviour. Upon admission, he had expressed suicidal thoughts. There was no known precipitant to his action. No organic symptoms were noted and no neurological examinations were performed. His HIV seropositivity was established postmortem and it was unknown if he had been aware of his serostatus.
Case studies have the advantage of providing richer insight into the clinical and psychosocial characteristics of cases and produce material that may be of heuristic value [42]. However, they lack the methodological rigour of case controlled studies. This restricts the scope of their generalisability. Nonetheless, the overall profile of HIV-related suicides presented by these case reports is consistent with those produced by case–controlled studies. The victims are almost exclusively white, homosexual males in their 30s and there appears to be a high prevalence of psychiatric morbidity at the time of death.
Register studies
Register-based reports involve matching population figures on causes of death to AIDS register data and are generally regarded as the most reliable and valid method of data collection [46]. Table 2 presents results of register studies investigating the prevalence of HIV/AIDSrelated suicides compared with the general population.
Register studies on rate of suicide in HIV seropositive and AIDS-diagnosed persons versus general community
As Table 2 indicates, the relative risk (even when age adjusted) of persons with AIDS dying from suicide is higher than that of the general population. However, the above results must be interpreted with caution due to methodological limitations of the studies. For instance, the relative risk reported by Mancoske and colleagues [51], which was much higher than the others, was inflated because the study sample was not strictly limited to those persons whose death certificates mentioned either HIV or AIDS diagnosis. It also included persons whose death certificates documented the presence of opportunistic infections commonly suffered by people with AIDS such as cytomeglovirus (CMV), Karposi's sarcoma and Pneumocyctis carinii pneumonia (PCP) but not necessarily a diagnosis of HIV/AIDS.
Although some of the studies adjusted their relative risk rates for sex and age, other relevant factors such as IVDU, sexual orientation and relationship status and race were not always controlled for. Variability between country practices of registering deaths as suicide also adds to the difficulty of establishing the true extent of suicide in the community, as does the problem of separating out those who died from accidental versus deliberate overdose of drugs. Therefore, it is not surprising perhaps that there is great variation between the relative risk (age-corrected) rates of suicide in HIV/AIDS reported by the studies in Table 2 [46]. A compounding problem is that the rate of AIDS-related suicides is actually quite low, accounting for as little as 0–3.3%% of total suicides [53]. This observation was borne out in a study, which found that suicides of males with HIV did not have any substantial effect on suicide trends in the USA [54].
It is difficult to determine the effects of possible confounding factors in these studies because most did not provide information on psychiatric histories or psychosocial issues associated with the suicides. The one study [47], which did note that 42%% of the AIDS-related suicides had contact with a psychiatrist less than 5 days before their death. Another limitation is that results of existing studies are dominated by the USA experience and may not be generalisable to other cultures, particularly developing nations, because of different management methods, support systems and a difference in profile of individuals infected by HIV. Attempts to conduct similar studies in other countries have been hampered by poor data collection and maintenance of databases on the incidence of AIDS/HIV and suicides. However, despite the limitations imposed by methodological shortfalls on the generalisability of the results of the above studies, cumulatively, they suggest an elevated rate of suicide in men with HIV/AIDS.
Some researchers have undertaken studies in defined cohorts to better understand the incidence of suicide and associated risk factors in persons with HIV and/or AIDS. In one such study that assessed rates of suicide in the US Air Force [55], HIV-seropositive personnel did not appear to be at significantly greater risk of suicide (RR = 2.08, 95%% CI = 1.00–3.82) than HIV-seronegative persons (RR = 1.67, 95%% CI = 1.07–2.48). When adjusted for age, sex and race, the suicide rates for both groups were only marginally higher than the general population.
Psychological autopsy studies
Another method to determine HIV-related deaths by suicide has been by psychological autopsy, which involves the reconstruction of suicidal death through interviews with survivors. The method was developed at the Suicide Prevention Centre in Los Angeles during the 1950s in an attempt to improve the accuracy of coroner's verdicts. The procedure analyses death through physical, psychological and social perspectives and involves interviews with family members and other key informants to provide detailed information of a kind that cannot be obtained by other methods [56]. With regard to HIV/AIDS studies, in most cases, where possible, HIV serostatus is determined post-mortem if that information is not available at the time of death. Data is also gathered from medical records. Table 3 presents the results of such studies.
Psychological autopsy studies investigating the association between HIV and suicide
Results of the studies (Table 3) indicate that suicide (as determined by the coroner) was the cause of death in approximately 26%% of cases of HIV-seropositive persons who died in suspicious or violent circumstances. Of note, all studies do not report rates of suicide in HIVseropositive individuals. More than half of the HIVrelated suicides were considered to have been in homosexual/bisexual individuals [57], [59], [60]. A history of IVDU was noted in 38%% of cases regarded as suicides [60] while in one study [57], psychoactive drugs were detected in 70%% of the cases. In the same study, 67%% had expressed suicidal intent or showed depressive symptoms before their deaths.
Two important limitations of these studies is that first, autopsies could not be conducted on all cases for reasons ranging from refusal of existing family members, to advanced decomposition of bodies precluding the removal of bodily samples for HIV testing. Therefore, there is still a percentage of cases for whom HIV serostatus was unknown and for whom conclusions cannot be drawn about cause of death and associated factors. Second, these results must be interpreted with caution, bearing in mind that coroner's practices on delivering verdicts of suicide can vary between countries and even between States.
Risk factors associated with suicidal behaviour in HIV/AIDS
Individuals infected with HIV face disease-specific stressors but also are subject to a large number of general suicide risk factors that include high rates of psychiatric morbidity and substance abuse [61]. Psychiatric disorder is associated with a high risk of suicide. The lifetime risks of suicide have been estimated at 6%% for affective disorder, 7%% for alcohol dependence, 4%% for schizophrenia [62].
Psychosocial stressors peculiar to HIV/AIDS include multiple bereavements due to the loss of partner and friends to AIDS, the stigma associated with the illness and accompanying rejection. As with chronic and debilitating illnesses, HIV/AIDS sufferers may also face financial problems due to increased inability to work. Lifetime depressive disorders also appear to be higher amongst individuals with same-gender sexual orientation [32] and lifetime rates of suicide associated with depressive illness have been reported as ranging from 6%% [62] to 15%% [63]. There appears to be a high prevalence of substance abuse in many of the HIV-seropositive groups studied, which confers an added risk of suicidal behaviour.
Comparison of suicide in HIV/AIDS with other medically ill populations
Earlier investigations into the association between suicide and medical illness reported suicide to be the cause of patients’ deaths in 10.9–51.0%% of cases [64]. Patients with cancer [65–67] and diseases of the central nervous system (CNS) including multiple sclerosis [68–70] and Huntington's disease [71], [72] were at greater risk of committing suicide as were older patients with chronic illness. As AIDS is a terminal disease which affects all organ systems [73] including the CNS, it is not unreasonable to expect that HIV/AIDS may also be associated with a greater risk of suicidal behaviour [6].
In a recent and comprehensive review, Harris and Barraclough [74] conducted a meta-analysis of all studies from 1966 to 1992 investigating the prevalence of suicide in individuals with 63 different medical disorders including HIV/AIDS. Six studies [47], [49], [50], [57], [75], [76] (out of a possible 18) on HIV/AIDS and completed suicide met criteria for the meta-analysis. All indicated an increased risk and collectively represented almost a sevenfold increase compared with the expected rate. Other groups of medical disorders that also showed significant increase in risk of suicide included Huntington's disease (threefold risk), multiple sclerosis (twofold increase), malignant neoplasms (ranged from 1.4 to 2.5- fold risk), peptic ulcers (twofold increase), renal disease (14-fold risk), spinal cord injury (fourfold risk) and systemic lupus erythematosus (fourfold risk). A consistent finding from the review was that the mental illness and/or substance abuse were influential factors in those diseases with an increased risk of suicide. Although the risk of suicide in AIDS is not the highest for all the medical conditions, it is greater than that for the ‘traditional’ diseases. This may reflect the fact that it is a relatively new illness with a variety of physical manifestations and psychological implications, and is generally regarded as ultimately fatal. In addition, less is known about effective treatment and management than is the case for most other medical disorders, at least until recently. It may also be compounded by the social stigmatization of the illness.
There is a consistently reported strong association between mental illness, particularly depression, and suicidal behaviour in physically ill patients [77], although not all researchers have found this association [78]. One possible confounding factor in discrepancies between studies is that depression has been and often remains underdiagnosed and inadequately treated, particularly in primary care settings [79].
Implications of new treatment
When HIV/AIDS was first identified in the early 1980s, it was considered a terminal illness. The introduction of more effective antiretroviral treatments in the last two decades has offered a better prognosis for many HIV-infected individuals. The data in Table 2 tentatively suggests a decrease in age-adjusted relative risk of suicide for persons with HIV/AIDS over the years, coinciding with the introduction of these newer antiretrovirals. The nature of the long-term impact of these more effective antiretrovirals on people's lives is as yet unclear – more time will be required to determine this. However, it is conceivable that one of the effects might be the relaxation of community attitudes towards preventative measures of HIV-infection, manifested by a recent increase in unsafe sexual practices, particularly amongst men who have sex with other men [80]. In Victoria, Australia for instance, there was a 41%% increase in new HIV diagnosis in 2000 [80]. Recent studies have noted an association between highly active antiretroviral therapy and decrease in depressive symptoms in HIV-seropositive individuals [81], [82]. Given the association between depressive illness and higher risk of suicidal behaviour, it is not unlikely that these newer antiretrovirals might have an impact on suicidal behaviour. However, the psychological impact of newer and better treatments may be complex. Those who experience major improvements in physical health may also be presented with paradoxical dilemmas such as what to do with their ‘new lease on life’, particularly if they had previously re-organized their life affairs in preparation for death [83]. On the other hand, those who do not respond to the new treatment may experience devastating disappointment.
Conclusion
Early epidemiological studies indicated an elevated risk of suicide in persons with HIV/AIDS compared with the general population. However, results from studies which have investigated suicidal behaviour in defined groups of subjects with HIV/AIDS suggest that although there is an association, it is mediated by a variety of other factors. A high prevalence of well-established suicide risk factors including psychiatric morbidity (particularly depression and substance abuse [62]), same-gender sexual orientation [34], [84], [85], male gender and stressors amongst HIV-seropositive individuals are major confounds. Future research, including longitudinal, needs to address these issues to determine if HIV/AIDS confers an added risk of increased suicidality above and beyond the contributions of other high-risk factors. There is little data from developing countries yet HIV has become increasingly prevalent in Asia and Africa. Studies fail to include HIV-seropositive persons of non-English speaking background perhaps because of resource constraints including lack of culturally sensitive research tools. Yet, a recent increase in HIV has been observed in individuals who originate from countries with high rates, specifically Thailand, Cambodia and Myanmar [80]. More research needs to focus on these people.
Footnotes
Acknowledgements
Funding for the project was provided by the Mental Health Branch, Victoria