Suicide is a major public health concern in China, where it is the fifth leading cause of death.1 The overall annual suicide rate in China (19/100,000) is now above the range reported for the United States (11/100,000) and the European Union.2 A recent community prevalence survey in Beijing and Shanghai reported lifetime rates of suicidal ideation (3.1%), plans (0.9%), and attempts (1%),3 comparable with contemporaneous rates in the United States.4 Consistent with findings in North America and Europe, in China individuals with psychiatric disorders, particularly major depression and schizophrenia, have elevated past and current rates of suicidal ideation and behaviors.5 Important “at-risk” populations for suicide identified in other countries; however, seem to have been overlooked in the China context, including HIV-infected (HIV+) individuals, injection drug users (IDUs), and individuals with both conditions.6–8
In China, as elsewhere, injection drug use and HIV infection are intersecting epidemics.9 Heroin addiction has grown dramatically alongside social and economic development, with the number of registered drug users (those identified and enrolled in government treatment programs) increasing from 70,000 in the 1990s to more than 1.6 million by 2005.9 The number of “unregistered” drug users is thought to be much higher.10 More than 80% of former heroin users in treatment present with depressive symptoms.11–13 The HIV epidemic in China was first recognized in the mid-1980s among IDUs in the border region adjacent to Myanmar (Burma), Thailand, and Laos, a region known as the "Golden Triangle." Sentinel surveillance and epidemiological studies indicated that HIV prevalence among IDUs in the southern provinces exceeds 5%, and in some areas it is much higher (30%–80%).9 Injection heroin users (IDUs) account for the largest proportion of HIV+ individuals in China (more than 40%; 3). Although data from Western countries suggest that substance abusers and HIV+ individuals are at increased risk for both comorbid psychiatric disorders and suicide,6–8 to our knowledge, no prior study has examined HIV infection and suicidality in injection heroin users in China.
In 2006, the interdisciplinary NeuroAIDS research teams at the HIV Neurobehavioral Research Center at the University of California at San Diego, the China Centers for Disease Control, and the Mental Health Institute at Peking University began collaborating on a neuropsychiatric study in IDUs in treatment in the Chinese province of Yunnan, an epicenter of the heroin and HIV epidemics. This article reports the frequency of suicide-related thoughts and behaviors and characteristics associated with the risks of suicidality in HIV+ and HIV− IDUs in treatment, compared with demographically similar Chinese non–drug-using controls in the same region. We hypothesized that the rates of suicidal ideation would be elevated in HIV+ IDUs and HIV− IDUs compared with non-IDU controls and that the prevalence of suicidality would be highest among HIV+ IDUs in treatment. We explored whether other factors would increase risks for suicidal ideas, including lifetime histories of major depression and alcohol use disorders, perceived social support, and (in HIV+) AIDS status and HIV-related stress.
The parent research program was a 4-year longitudinal project to examine the neurobehavioral effects of HIV infection in injection heroin users in Yunnan, China. The study was approved by the Institutional Review Boards of the China Centers for Disease Control/National Center for AIDS, the Peking University, Beijing, China, and the University of California, San Diego, CA. All participants provided written informed consent to enroll in the research. Participants completed a baseline evaluation lasting 3.5–4 hours, and those who were IDUs in treatment were asked to return for follow-up assessments every 12 months thereafter. Compensation for each assessment was US $10. This article presents data from the baseline evaluation.
Individuals in treatment for IDU were recruited from 2 outpatient methadone clinics at Kunming and GeJio City in the province of Yunnan by flyers, announcements, and approach by the treating staff. In China, methadone clinics are sponsored by the government, and all registered injection heroin users receive free methadone, substance abuse–related counseling, and screening for HIV serostatus (HIV quick test, OraSure Technologies, Inc, Bethlehem PA). We enrolled 204 HIV-positive and 202 HIV-negative injection heroin users in treatment (more than 90% of those screened for eligibility). Participants were men and women, aged 18–55 years, with either HIV-positive serostatus or an HIV-negative screening. Participants had to be ambulatory, literate with at least 3 years of formal education, and able to comprehend study procedures and provide informed consent. Because the primary aim of the study was evaluation of HIV-associated neurocognitive impairment, we attempted to exclude conditions that might confound neuropsychological evaluation. Exclusions were as follows: (1) evidence of non–HIV-related or non–heroin-related neurocognitive disorders (eg, seizure disorder); (2) evidence of schizophrenia or bipolar disorder; and (3) individuals with “heavy substance use within the past 30 days” defined as more than 3 drinks of alcohol daily or the use of illegal psychoactive substances 3 or more times weekly. To establish norms for neurocognitive assessments, we also advertised the research generally to the community for a control sample. We enrolled 201 noninjection drug–using controls who were comparable to the 2 IDU groups in treatment with respect to age, gender, education, and ethnicity (99% Han ethnicity) and who met the same inclusion and exclusion criteria as the treated IDU groups. At least 90% of those screened for eligibility were enrolled. All controls had a negative HIV screen and were assessed with the same study protocol.
Participants underwent a structured psychiatric interview, comprehensive neuromedical and neuropsychological evaluations, and assessments of mood; perceived social support, HIV-relevant stress, and daily function as described below. All instruments were administered in the Chinese language by Chinese psychiatrists and psychiatric nurses who had been trained by the US team in the study instruments. In HIV+ participants, disease duration and staging, using US Centers for Disease Control criteria,14 and current immune status (CD4 cell count) were also determined as was treatment status (on or off antiretrovirals). The neuropsychological findings will be reported elsewhere.
World Mental Health Composite International Diagnostic Interview (Version 3.0)
We administered the Chinese version of the Composite International Diagnostic Interview, developed by the World Health Organization, as part of the World Mental Health Survey initiative.15 It is a fully structured lay-administered psychiatric interview as follows: lifetime (past) and current (within the last 30 days) rates of major depression and alcohol and opioid use disorders (abuse and dependence) were assessed using the Depression and Substance Use modules of the Composite International Diagnostic Interview.
Evaluation of Suicidality
Suicidality was based on responses to item 9 of the Beck Depression Inventor-II,16 consisting of statements ranging from 0 = “I have no thoughts of killing myself”; 1 = “I have suicidal thoughts but would not carry them out”; 2 = “I would like to kill myself”; or 3 = “I would kill myself if I had a chance.” Endorsements of ≥1 on item 9 were coded positive for suicidality. The reporting period was the past 2 weeks.
Social Support, Self-Perceived Stressors, and Everyday Function
Perceived social support, stressors due to HIV infection, and lack of independence in everyday function were examined because isolation, stressors, and inability to function were predicted to impact on suicidal ideation. Social support was measured using the Social Support Rating Scale, developed in China for psychiatric research.17 Social support is rated from 1 to 4 in response to 10 questions. Ratings are summed to yield a total score ranging from 10 to 40; a higher total score indicates higher levels of perceived social support. For example, 1 item prompt asks for a description of relationships between the respondent and work or social colleagues, with choices being 1 = “We never talk about personal concerns with each other”; 2 = “My colleagues would be somewhat concerned if I had difficulties”; 3 = “A few colleagues would be concerned about me a lot”; or 4 = “Most colleagues would be concerned about me a lot.”
Perceived stressors among HIV+ participants were assessed using the Chinese Modified HIV Stressor Scale.18 This measure was developed by translating an established English language questionnaire (the HIV Stressor Scale19) and modified by adding queries about 2 additional stressors as follows: being rejected by family members, and being discriminated against in general. The scale administered consisted of 26 HIV-relevant negative life events that participants might have experienced in the past 4 weeks. Survey items cover 5 following areas: relationships (eg, revealing HIV status to lover), medical care (eg, changing physicians), grief (eg, death of friend), financial (eg, loss of income), and work (losing health insurance). Items are rated on a scale ranging from 0 = “Never stressful” to 4 = “Extremely stressful.” Ratings for each item were summed to yield a total score ranging from 0 to 104; a higher total score indicates higher levels of self-perceived HIV-relevant stress.
To measure independence in daily functioning, we used the Chinese version of the Activities of Daily Living (ADL) Scale, a 14-item questionnaire with demonstrated validity in studies of medically ill and dementia populations in China.20–22 The14 items address both physical self-maintenance activities (eg, dressing, walking, and bathing) and instrumental ADL (eg, preparing meals, housework). Each item is rated on a 4-point scale as follows: 1 = “No difficulty at all”; 2 = “Have some difficulty”; 3 = “Need some assistance”; 4 = “Can not do at all.” Scores for each item are summed; a total score >15 indicates difficulty or need of assistance in at least 1 area of daily function and resulted in classification as “dependent” in ADL.
Descriptive analyses and group comparisons examined the demographic and clinical characteristics of non–drug-using controls and treated IDU groups with and without HIV infection.
Group comparisons were tested with 1-way analysis of variance for continuous variables and χ2 tests for categorical variables. Univariate correlates of suicidal ideas were assessed using Pearson correlation tests for continuous variables and χ2 tests for analyses of categorical variables.
Two multivariable logistic regressions examined predictors of suicidal ideation separately in HIV+ and in HIV− treated IDUs. The regression limited to HIV+ IDU participants included demographics (ie, age, years of education, employment status, gender, and marital status), level of social support, dependency in ADL, and prior (excluding current) histories of major depression or an alcohol use disorder (abuse or dependence) as potential predictors, and specific HIV-relevant characteristics [ie, level of HIV stress, years since first HIV+ test, history of an AIDS diagnosis, and current (on/off) antiretroviral status] and substance use information (ie, years of injection drug use) as potential predictors.
In the second regression with the HIV− IDU cohort, we excluded HIV-related variables as potential predictors but considered the same demographic, social support, daily living, and psychiatric and substance use variables as above. Stepwise selection based on the Akaike Information Criterion (AIC) was used to identify the optimal model.23 We chose this approach rather than standard stepwise methods because AIC is thought to generate models more likely to be applicable to populations outside the dataset from which the predictors were identified (ie, AIC avoids overfitting of individual covariates).
The critical alpha level was set at 0.05 for statistical significance. All comparisons were 2-tailed. No adjustments were made for multiple comparisons in part because some of the analyses were exploratory and because the final 2 multivariate analyses did not consider the results of the univariate analyses. Analyses were performed using the Statistical Package for the Social Sciences program (SPSS-PC, Windows V10.0) and the R statistical software package.24
Frequency of Major Depressive Disorder, Alcohol Use Disorder, and Suicidality
The 201 non–drug-using controls and 406 IDUs in treatment (202 HIV− and 204 HIV+) who completed the baseline evaluation are included in the analyses. In terms of demographic and clinical characteristics (Table 1), the treated IDU groups and matched non-IDU controls consisted of mostly men in their mid-30s, with a middle school education and above average household income for the area. There was a slight but statistically significant group difference in age, but age was not related to suicidality in this study. High proportions of individuals in the HIV− and HIV+ IDU groups in treatment were neither married nor employed, which was significantly different from non–drug-using controls. About 1 in 4 of the HIV+ individuals had AIDS. In the HIV-positive group, overall there was evidence of satisfactory immune function as evidenced by current CD4 count. With reference to psychopathology, the lifetime frequency of major depressive disorder was significantly (over 4-fold) higher in treated IDUs compared with controls (P = 0.005), but the rates did not differ meaningfully between the HIV+ and HIV− IDUs in treatment. Rates of current depressive disorder were low, with none of the controls and only 1% (n = 4) of treated IDUs having current major depression.
Lifetime frequency of any alcohol use disorder, which ranged from approximately 10%–20%, did not differ significantly between IDU groups and controls or between HIV+ and HIV− IDUs in treatment. Despite our effort to exclude current substance abusers, about 2% of the entire cohort (n = 12) reported a current (within the last 30 days) alcohol use disorder. Rates of current alcohol use disorder did not differ across groups. Nearly the entire sample (99%) of treated IDUs reported a lifetime history of heroin abuse or dependence, which is consistent with the fact that IDUs were recruited from methadone clinics. Again, despite our screening efforts, approximately 2% (n = 8) of treated IDUs met criteria for a current heroin use disorder, with HIV+ and HIV− IDUs having similar frequencies.
As also seen in Table 1, the 2 IDU groups were significantly more likely to express current suicidal ideas (as measured by item 9 of the BDI-II) than controls (P < 0.0001), with over 1 in 3 IDUs in treatment reporting suicidality compared with fewer than 1 in 10 controls. Further, about 4% of patients endorsed the most severe category (eg, “Would kill myself if I had a chance”), whereas none of the controls did so. Overall rates were slightly higher in HIV+ IDUs than HIV− IDUs in treatment, but this difference was not statistically significant. Due to the small number of cases, formal analysis of a “gradient” of severity of suicidal ideation across these latter groups was not performed.
We assessed the relationship of current suicidal ideas to history of major depression and current mood as follows: of those in our cohort who expressed suicidal ideas, a minority had a lifetime history of major depression (17% in HIV+ IDUs, 13% in HIV− IDUs, 0% in controls). Most participants who reported current suicidal ideation had total depressive symptoms in the clinically relevant range (ie, BDI-II total scores ≥ 19; 88% in HIV+ IDUs, 90% in HIV− IDUs, 41% in controls). A minority of our cohort who expressed suicidal ideas had a lifetime history of an alcohol use disorder (17% in HIV+ IDUs, 32% in HIV− IDUs, 6% in controls).
Correlates of Suicidal Ideas in IDUs and Controls
Among the entire cohort (ie, in treatment IDUs and controls; n = 607), factors significantly correlated with current suicidal thinking included being unmarried or unemployed, having a lifetime (past) history of major depression, or lifetime alcohol use disorder or heroin use disorder (all P < 0.0001). Higher BDI-II total score (excluding item 9, P < 0.0001), poor social support (P = 0.0003), and the need for assistance in ADL (P = 0.0013) were also significantly associated with current suicidality in treated IDUs and controls. Among treated IDUs (n = 407), past history of major depression (P < 0.0001), or past alcohol use disorder (P = 0.0052), and higher BDI-II total score (P < 0.0001) were associated with current suicidality as was longer duration of injection drug use (P = 0.0297) and poor social support (P = 0.0283). Among HIV+ IDUs (n = 204), experiencing HIV-relevant stress was significantly associated with current suicidality (P = 0.0041). The correlation between social support and HIV stress was small but significant (r = −0.2311, P = 0.0009).
Multivariable Predictors of Suicidal Ideas in HIV+ and HIV− IDUs
Multivariable predictions of current suicidality in the HIV+ (n = 195 with complete data) and HIV− (n = 202) IDU samples were examined separately, as explained above. The final AIC-based regression model for the HIV+ IDUs was highly significant (P < 0.0001) and included HIV-related stress, social support, and prior history of major depression. Prior history of major depression (P < 0.01) uniquely predicted current suicidal ideas in treated HIV+ IDUs, but high HIV-relevant stress (P = 0.06) and low social support (P = 0.14) did not. Past major depression was associated with a 6-fold increase in risk of current suicidal ideation [odds ratio (OR): 6.12; confidence interval (CI): 1.68 to 22.33]. In HIV− treated IDUs, the best regression model selected via AIC included prior history of major depressive disorder and prior history of an alcohol use disorder. This overall regression model also was highly significant (P < 0.001). Both prior history of major depression (P < 0.01) and prior alcohol use disorder (P < 0.001) were uniquely significant. Prior major depression increased risk by 6-fold (OR: 6.10; CI: 1.58 to 23.61), and past alcohol use disorder increased risk by 3-fold (OR: 3.44; CI: 1.58 to 23.61).
The high frequency of current suicidality in our IDU sample in treatment is consistent with previous work on treated heroin users in China, 33% of who reported suicidal thoughts in the month after detoxification,25 and is consistent with Western settings.25,26 This exceeds the 12-month prevalence rate of 3% observed in population-based Chinese community surveys.3,27 Our finding of lifetime history of major depression as a significant risk for suicidal ideas in HIV+ and HIV− IDUs in treatment agrees with risks identified in general population epidemiologic studies in China.5 Our finding of alcohol use disorder as a risk in HIV− IDUs differs from results of epidemiological surveys in the general metropolitan Chinese population.3 Studies in the United States of both the general population28 and heroin abusers report a strong association of alcohol use disorders with suicidality.29
Although the 2 treated IDU groups differed from non-IDU controls in lower employment rates and increased likelihood of ADL dependence, these indicators of poor life functioning did not predict suicidal ideation. Within the HIV+ IDU group, social factors (stress and poor support) were associated with suicidal ideation, whereas AIDS and duration of HIV disease and alcohol use disorder were not. Our prevalence rate of recent suicidal ideas is consistent with rates in US studies of HIV+ samples not selected for methadone treatment of heroin abuse,30 but in that work neither alcohol nor heroin abuse predicted suicidality.
Over the past 20 years, epidemiological research in China has observed a relatively low lifetime prevalence of major depression (eg, 3%–4%) compared with 4-fold to 5-fold higher rates in Western societies.31 Our in-treatment heroin users were more likely than the general population to have a lifetime major depressive disorder (eg, 6%–8%). Nevertheless, the high rate of current suicidal ideation may not be entirely explained by depressive syndromes. Some research argues that adverse life events and lack of social support are more related to suicidality than is depressive disorder in China.32 Our finding that HIV-related stress and perception of low social support were among the predictors of suicidal ideation in HIV+ persons is consistent with this view. Suicidal ideation in patients in heroin treatment in China also may be related to social issues (eg, loss of income), to the stigma of being a drug abuser, or to treatment-related factors.32 There are civil and criminal penalties for drug use in China, and individuals may volunteer for drug treatment or be sentenced to enforced treatment. Our study participants were in a voluntary methadone program; we are not sure whether drug treatment would increase suicidal ideation. One China study suggested that severity of depressive symptoms was equivalent between heroin users in enforced drug treatment and those who sought treatment voluntarily.13
Our study has several limitations. First, our cross-sectional design is unable to address causality in the relationship between suicidality and clinical and social risks. Second, our findings regarding suicidality in these IDU groups in treatment may underestimate the true rates in that we cannot rule out possible reluctance to reveal suicidal thoughts due to fear of stigma, or other reasons. Finally, our analysis of correlates and predictors of suicidality was based on a convenience sample that was not recruited using formal epidemiologic methods designed to ensure representativeness of a particular population.
Despite these limitations, our study provides evidence that suicidality is frequent in Chinese injection heroin users in treatment, and that comorbid mood and alcohol use disorder and HIV-relevant stressors and social isolation may contribute to risk in this vulnerable population. This suggests that assessments and interventions for suicide might be incorporated into a comprehensive care plan.
1. Phillips MR, Li X, Zhang Y. Suicide rates in China, 1995-99. Lancet. 2002;359:835–840.
2. Zhang J, Xiao S, Zhou L. Mental disorders and suicide among young rural Chinese: a case-control psychological autopsy study. Am J Psychiatry. 2010;167:773–781.
3. Lee S, Fung SC, Tsang A, et al.. Lifetime prevalence of suicide ideation, plan, and attempt in metropolitan China. Acta Psychiatr Scand. 2007;116:429–437.
4. Kessler RC, Berglund P, Borges G, et al.. Trends in suicide ideation, plans, gestures, and attempts in the United States, 1990–1991 to 2001–2003. JAMA. 2005;293:2487–2495.
5. Ma X, Xiang YT, Cai ZJ, et al.. Lifetime prevalence of suicidal ideation, suicide plans and attempts in rural and urban regions of Beijing, China. Aust N Z J Psychiatry. 2009;43:158–166.
6. Weissman MM, Bland RC, Canino GJ, et al.. Prevalence of suicide ideation and suicide attempts in nine countries. Psychol Med. 1999;29:9–17.
7. Komiti A, Judd F, Grech P, et al.. Suicidal behaviour in people with HIV/AIDS: a review. Aust N Z J Psychiatry. 2001;35:747–757.
8. Keiser O, Spoerri A, Brinkhof MW, et al.. Suicide in HIV-infected individuals and the general population in Switzerland, 1988–2008. Am J Psychiatry. 2010;167:143–150.
9. Wu Z, Sullivan SG, Wang Y, et al.. Evolution of China's response to HIV/AIDS. Lancet. 2007;369:679–690.
10. Wu Z, Luo W, Sullivan S, et al.. Evaluation of a needle social marketing strategy to control HIV among injecting drug users in China. AIDS. 2007;21(suppl 8):S115–S122.
11. Ynag H, Xin Y. Analysis of 110 Heroin dependent patients on depression. Chin J Drug Depend. 2005;14:130–132.
12. Liu JB, Dilixat Y, Li F. A study of relationship among anxiety, depression and hig-risk behavior of AIDS in heroin addicts. Clin J Behav Med Sci. 2006;15:1001–1002.
13. Li XD, Li WP, Lian Z, et al.. Analysis of depression in Heroin dependent patients on voluntary detoxication and compulsory detoxification. Clin J Drug Depend. 2009;18:500–503.
14. CDC. 1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep. 1993;17(4):802–810.
15. Kessler RC, Ustun TB. The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Int J Methods Psychiatr Res. 2004;13:93–121.
16. Chinese Behavioral Sciences Society. The Chinese Version of the Beck Depression Inventory. 2nd ed. Licensed Chinese translation. The Psychological Corporation. New York, NY: Harcourt Brace; 2000.
17. Xiao SY. The social support rate scale. Chin J Psychol. 1993:7(suppl):42–46.
18. Au A, Chan I, Li P, et al.. Stress and health related quality of life among HIV-infected persons in Hong Kong. AIDS Behav. 2004;8:119–129.
19. Thompson SC, Nanni C, Levine A. The stressors and stress of being HIV positive. AIDS Care. 1996;8:5–14.
20. Zhang MY, Katzman R, Salmon D, et al.. The prevalence of dementia and Alzheimer's disease in Shanghai, China: impact of age, gender, and education. Ann Neurol. 1990;27:428–437.
21. Zhang M. Chinese version of ADL. Shanghai Arch Psychiatry. 1989;2:68–71.
22. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist. 1969;9:179–186.
23. Akaike H. Information theory and an extension of the maximum likelihood principle. In: Petrov BN, Csaki F, eds. International Symposium on Information Theory. Budapest, Hungary: Akademia Kiado; 1973:267–281.
24. The R Foundation for Statistical Computing R Development Core Team. R: A Language and Environment for Statistical Computing. R version 2.10.1. Vienna, Austria: R Foundation for Statistical Computing; 2009.
25. Zhao XL, Liu LZ, Chen GX. Suicidal ideation in enforced detoxication subjects and nursing countermeasure. J Nurs Sci. 2003;18:684–685.
26. Borges G, Walters EE, Kessler RC. Associations of substance use, abuse, and dependence with subsequent suicidal behavior. Am J Epidemiol. 2000;151:781–789.
27. Zhang Y, Law CK, Yip PSF. Psychological factors associated with the incidence and persistence of suicidal ideation. J Affect Disord. 2011;133:584–590.
28. Kessler RC, Borges G, Walters EE. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch Gen Psychiatry. 1999;56:617–626.
29. Darke S, Ross J. Suicide among heroin users: rates, risk factors, and methods. Addiction. 2002;97:1383–1394.
30. Carrico AW, Johnson MO, Morin SF, et al.. Correlates of suicidal ideation among HIV-positive persons. AIDS. 2007;21:1199–1203.
31. Lee S, Tsang A, Zhang MY, et al.. Lifetime prevalence and inter-cohort variation in DSM-IV disorders in metropolitan China. Psychol Med. 2007;37:61–72.
32. Chen YY, Yip PSF. Rethinking suicide prevention in Asian countries. Lancet. 2008;372:1629–1630.