Study findings suggest that depressive symptoms were associated with an increased likelihood of injection risk behaviors among treatment-seeking opioid-dependent users, even after adjusting for a wide range of potential confounders. Furthermore, this association was independent of co-occurring substance use (cocaine and amphetamine use). These findings concur with those reported in a large community sample (N=499) of IDUs (Mandell et al., 1999
). In their study, a greater number of depressive symptoms was positively associated with needle sharing, including the sharing of needles that had and had not been cleaned with bleach (Mandell et al., 1999
). This relationship remained significant (odds ratio [OR]=1.66) after adjusting for multiple potential confounders, and the authors suggested that depressive symptoms are a predisposing factor for needle-sharing behavior. Similarly, depressive symptoms were one of several factors associated with injection risk behaviors in a sample of IDUs receiving methadone treatment (N=152) (Kang & De Leon, 1993
). Even in a setting with adequate accessibility to sterile needles, depressed injection drug users were more likely to share needles that their non-depressed counterparts (S. A. Strathdee et al., 1997
How depressive symptoms may predispose individuals to engage in injection risk behaviors? In a community sample of adolescents, hopelessness was found to mediate the association between depressive symptoms and health-related risk taking (Testa & Steinberg, 2010
). Furthermore, increased levels of depression may be associated with suicide attempts among heroin users (Chen et al., 2010
). Not caring about injection risk behaviors may reflect indifference to the possibility of harming oneself. Using potentially infected needles is a risk-taking behavior that is not associated with thrill seeking, but instead with hopelessness and despair. Although a speculative formulation, it is based on the literature pointing to an elevated risk of suicide attempts among heroin users (Chen et al., 2010
Others have focused on the antecedents of depression among injection drug users. For example, Strathdee et al. found an association between depression and needle sharing but this association was not independent of prior sexual abuse (S. A. Strathdee et al., 1997
). Thus, depression and its antecedents are likely to be associated with a heightened needle risk through either direct or indirect pathways. For example, sexual abuse in childhood may increase risk for depression, and depression may heighten risk for injection risk behaviors. Another path may include the association between early adversity and alcohol dependence (Pilowsky, Keyes, & Hasin, 2009
), and between alcoholism and HIV-related risk behaviors (Newville & Haller, 2010
Contrary to expectations, neither depressive nor anxiety-related symptoms were positively associated with an increased likelihood of risky sexual behaviors. Because depressive symptoms have been associated with sexual risk behaviors in previous studies, this null finding deserves further consideration. Several studies reported associations between depression and sexual risk behaviors in community-based samples (Brown et al., 2006
; Khan et al., 2009
; Kosunen et al., 2003
; Lehrer et al., 2006
; Rubin et al., 2009
; Shrier et al., 2001
). Their risk profile may differ from that of drug users. Furthermore, this association is not well established in other population subgroups. A meta- analysis that included studies that focused on a variety of populations, such as men having sex with men, patients from alcohol treatment centers, and untreated substance abusing samples, found little evidence that negative affect (a construct that includes depressive symptoms, anxiety, and anger) was associated with sexual risk behaviors. A recent longitudinal study, however, found an association between depression and having sex with multiple partners (Williams & Latkin, 2005
) (OR=1.55; 95% confidence interval [CI]: 1.12, 2.14), with IDUs (OR=1.57; 95% CI: 1.11, 2..2), and with crack users (OR:1.37; 95% CI: 1.02, 186). Participants were street-recruited drug users (Williams & Latkin, 2005
). In contrast, we examined treatment-seeking individuals who are often different from those not seeking treatment, e.g. the number of comorbid psychiatric disorders is often different, with a higher prevalence of comorbid disorders among those seeking treatment (Cohen & Cohen, 1984
; Hasin, Stinson, Ogburn, & Grant, 2007
). Treatment-seeking IDUs might also be less impulsive than those not seeking treatment, and therefore less likely to engage in risky sexual behaviors, even when depressed. Although speculative, this is an intriguing possibility.
The literature includes studies that do and do not support an association between depression and sexual risk behaviors (Crepaz & Marks, 2001
). According to the DSM-IV, depressive symptoms are associated with difficulties in sexual functioning in some individuals (e.g., anorgasmia in women, erectile dysfunction in men, or a decrease of sexual desire in both genders), and these difficulties may account for the decrease of risky sexual behaviors found in some studies. Additionally, depression is often associated with hopelessness and a lack of care for one’s own survival. This feature of depression may account for the reported increase of injection and sexual risk behaviors among depressed individuals. Several other features of depression may be associated with increased or decreased sexual risk behaviors, such as taking action to relieve depression through sexual excitement, or depressed individuals assigning a greater weight to potential risks, compared to their non-depressed counterparts (Crepaz & Marks, 2001
). Thus, associations between depression and risky sexual behaviors might hinge on which features of depression predominate in a given case.
We also examined the effects associated with the use of substances other than heroin/opioids. Among these substances, only the co-occurring use of non-injecting amphetamines was found in the final model to be positively associated with risky sexual behaviors. Several studies have found that amphetamines, and especially methamphetamine, are positively associated with risky sexual behaviors among IDUs and non-IDUs (Centers for Disease & Prevention, 2006
; Koblin et al., 2007
). While the proportion of amphetamine users who used methamphetamine cannot be determined from the CTN data, large-scale addiction treatment data in the United States have estimated that more than 90% of treatment-seeking amphetamine users use methamphetamine (Substance Abuse and Mental Health Services Administration, 2004).
Several limitations are noteworthy. This is a treatment-seeking sample and the findings reported here are not necessarily generalizable to opioid-dependent individuals in the community, including the many out-of-treatment drug users. Another limitation is the ascertainment of depressive symptoms using a single question from the ASI. We sought to support its validity by examining associations with suicidal ideation/attempts, with eight subscales of the SF-36, and with an SF-36 question specifically measuring depressive symptoms (“down in the dumps”). These analyses consistently showed that participants reporting depressive symptoms were clearly distinct from participants not reporting these symptoms, when these measures were used to compare them. Nevertheless, we do not know to what extent the answer to a single question reflects depression, a complex construct that includes multiple psychological and somatic indicators. Recall bias is a minor limitation because these data are based on reports of recent events. Last, associations reported here are not necessarily causal due to the cross-sectional nature of the data.
Notwithstanding the above limitations, the multisite Clinical Trials Network has noteworthy strengths. Participants were recruited from 12 major treatment programs across the US, and therefore the sample is more likely to be representative of US treatment-seeking opioid-dependent individuals than participants in single site studies. Across sites participants were assessed using the same standardized instruments (e.g., HIV Risk Behavior Scale, SF-36, ASI) with demonstrated validity and reliability in treatment samples.
Conclusions and Clinical Implications
Depressive symptoms were independently associated with an increased level of injection risk behaviors among American treatment-seeking opioid-dependent individuals. We do not know to what extent these findings are applicable to opioid-dependent individuals who are not seeking treatment and to those treated in other countries. Although the increased likelihood of injection risk associated with depressive symptoms was modest, all factors associated with an increased HIV risk are important given the severity of this condition. Furthermore, depression is a potentially modifiable factor, i.e. depression can and often is successfully treated (Balon, 2006
; Kennedy, 2005
). Thus, treatment of depression can make a contribution to decreasing injection risk among opioid-dependent IDUs, especially if combined with other risk-reduction interventions. This study also revealed that non-injecting amphetamine use was independently associated with sexual risk behaviors among opioid-dependent individuals.