Substance use is a public health challenge in the United States among many different vulnerable populations. Substance use is associated with a range of comorbidities among people living with HIV (PLWH), including viral hepatitis, tuberculosis, bacterial infections, kidney disease, atherosclerosis, cancer, and mental health disorders.1
In particular, current substance use is associated with increased rates of depressive symptoms and depression among PLWH.2–7 Managing mental health disorders is challenging and, if untreated, the mental health disease will negatively impact HIV care along the HIV care continuum, including delayed linkage to care, delayed initiation of antiretroviral therapy, suboptimal medication adherence, and worse clinical outcomes.3 Achieving abstinence in PLWH with substance use disorders is challenging,1 often requiring multifaceted and interdisciplinary approaches. Addressing substance use, however, is a key intervention for improving depressive symptoms in PLWH, especially because these conditions often co-occur and impact the HIV care continuum.1–7
We hypothesized that a reduction in substance use may have benefits in terms of depressive symptoms, even if abstinence is not achieved. Substance use can be considered a process of negative reinforcement, due to a decrease in the function of normal reward-related neurocircuitry, which results in increased depressive symptoms among users.8 Previous studies, however, have predominantly focused on achieving complete abstinence, which is often the main goal of substance use treatment. Harm reduction, the reduction in drug use, remains an important goal, but it is crucial to determine whether such reductions have resultant improvements in health, including depression outcomes. Our objective was to evaluate the impact on depressive symptoms of reducing substance use, with or without achieving abstinence. We considered this question in PLWH enrolled in the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort.9
This study includes PLWH aged 18 years or older and enrolled at 1 of 6 CNICS sites (Johns Hopkins University; University of Alabama at Birmingham; University of California, San Diego; University of California, San Francisco; University of North Carolina at Chapel Hill; and University of Washington, Seattle). CNICS is a longitudinal observational study of PLWH receiving primary care at CNICS sites from January 1, 1995, to the present.9 In total, 9905 participants were eligible for inclusion in analyses.
All participants completed longitudinal assessments of substance use frequency including cocaine/crack; amphetamine-type substances (ATS) that included methamphetamines; illicit opioids; and marijuana. Participants had to have 2 or more substance use assessments during cohort follow-up to be eligible for this study. We formed subcohorts of participants who were using a specific substance to ensure measurement of reduced use or abstinence were performed among participants eligible to reduce use due to current use.
As a sensitivity analysis, to enrich the number of opioid users, we also pooled CNICS with Project STRIDE10 from the Seek, Test, Treat, and Retain cohort.10 Project STRIDE is a randomized controlled trial of PLWH with substance use disorders that includes a buprenorphine plus naloxone (BPN) intervention, is registered at www.clinicaltrials.gov (NCT 01550341), and has similar measures of depressive symptoms and substance use (making it ideal for pooling).
In the CNICS cohort, PLWH completed an approximately 10-minute clinical assessment with touchscreen tablets, with a planned frequency of every ∼6 months, as part of routine clinical care.11 The CNICS clinical assessment includes measures of substance use (modified Alcohol, Smoking, and Substance Involvement Screening Test)12 as well as a broad group of other patient-reported measures and outcomes (including alcohol use). Assessments of substance use included a Likert scale for frequency of use in the past 30 days.
We categorized frequency of use into less than weekly, 1–3 times per week, and daily or almost daily, based on participant responses to the study instruments. Changes in substance use frequency were defined as a change from one category to another. Possible changes included abstinence (change to no use), reduced use (change to a lower frequency category), and nondecreasing or increasing use (the same or higher category of use, as compared to baseline).
Depressive symptom scores were assessed using the Patient Health Questionnaire (PHQ)-9. CNICS administered the PHQ-9 to participants repeatedly over follow-up, through a Computer-Assisted Self-Interview (CASI) system. The PHQ-9 has been validated as a scale for detecting depressive symptoms both in outpatient populations13 and in diverse/international populations.14 It has high screening utility and is used in clinical care to screen for depression among PLWH.15,16
Linear mixed models with time-updated change in substance use frequency and depressive symptom scores were used to examine associations between changes in the use of individual substances and depressive symptoms. For each substance, a specific cohort was formed of participants who were users at study baseline. These models were adjusted for other substance use, with random slopes and intercepts at the participant level to handle repeated measures over follow-up of both substance use and depressive symptoms. This type of mixed-model approach to link changes in drug exposure levels to changes in continuous outcomes has been used in pharmacoepidemiology contexts17 and is a well-known approach to handle irregular data collection and participants lost to follow-up.18,19 All estimates were adjusted for age, sex, use of other substances, alcohol use, and calendar year.
We were also interested in whether findings were the same if the outcome of interest was screening positive for depression rather than change in depression symptom score as a continuous variable. Therefore, we dichotomized depression scores using a cutoff point of ≥10 (high PHQ-9) to indicate high predictive value for screening positive for depression20 and repeated analyses using a joint longitudinal and survival model21 to examine the impact of decreasing drug use and of abstinence for each drug due to known limitations with less complex models.22
We also considered the possibility of a bidirectional association. We examined change in days of substance use among participants experiencing a drop in PHQ-9 score over follow-up, using a linear mixed-model approach. All statistical analysis was performed in STATA 14 (StataCorp, 2017, Stata Statistical Software: Release 15, College Station, TX: StataCorp LLC).
CNICS enrolled patients in clinical care, resulting in a large diverse cohort. Among 9905 PLWH, the mean age was 44 years, 16% of participants were female, and 40% used a substance at baseline (Table 1). Overall, 728 participants reported cocaine/crack use, 1016 ATS use, 290 illicit opioid use, and 3277 reported marijuana use at baseline.
Changes in ATS use were associated with the greatest improvements in depressive symptoms. Abstinence from ATS was associated with a mean difference in the depressive symptom score of ∆ −2.2 PHQ-9 points [95% confidence interval (CI): −2.7 to −1.8] or a 61% lower odds of screening positive for depression (95% CI: 0.30 to 0.52) compared with those who continued their ATS use without a decrease in frequency. Decreasing ATS without abstinence was associated with a mean difference of ∆ −1.7 PHQ-9 points (95% CI: −2.3 to −1.2) or 62% lower odds of screening positive for depression (95% CI: 0.25 to 0.56) compared with those who did not decrease their use (Table 2).
Decrease in use of other substances had more modest associations with depressive symptoms. Stopping marijuana was associated with a mean decrease in the depressive symptom score of ∆ −0.5 PHQ-9 points (95% CI: −0.7 to −0.3) or 28% lower odds of screening positive for depression (95% CI: 0.58 to 0.88), and decreasing marijuana use was associated with a mean decrease of ∆ −0.4 PHQ-9 points (95% CI: −0.7 to −0.1) or 30% lower odds of screening positive for depression (95% CI: 0.59 to 0.84). Stopping cocaine/crack was associated with a mean difference in the PHQ-9 score of ∆ −0.8 points (95% CI: −1.3 to −0.4) or 24% lower odds of screening positive for depression (95% CI: 0.56 to 1.03); however, decreasing cocaine/crack use without abstinence was not associated with a significant change in PHQ-9 score or the odds of screening positive for depression [∆PHQ-9 score = −0.5 points/odds ratio (OR) = 0.97]. Finally, neither stopping or reducing use of opiates over follow-up was associated with a significant reduction of depressive symptom (∆PHQ-9 score = −0.6 points/OR = 0.79 and ∆PHQ-9 score = −0.5/OR = 1.20, respectively) (Table 2). Results for opioid use were similar to the CNICS-only results (Table 2) when we included an additional study, Project STRIDE, with a high-level of opioid use to improve precision on the illicit opioid estimates (Table 1, Supplemental Digital Content, http://links.lww.com/QAI/B196).
Looking for a bidirectional association, we considered the association between reducing PHQ-9 score over follow-up and frequency of substance use. There was an association for cocaine/crack (∆days of cocaine/crack use −0.07; 95% CI: −0.12 to −0.02), ATS (∆days of ATS use −0.21; 95% CI: −0.30 to −0.13), and marijuana (∆days of marijuana use −0.25; 95% CI: −0.44 to −0.06) but not for opiates (∆days of opiate use −0.03; 95% CI: −0.08 to 0.02) (Table 2, Supplemental Digital Content, http://links.lww.com/QAI/B196, individual logistic and survival models, by substance, are Tables 3–10, Supplemental Digital Content, http://links.lww.com/QAI/B196).
We demonstrated that both substance use reduction and abstinence are associated with improvements in depressive symptoms over longitudinal follow-up in PLWH. Relative to other substances, reducing ATS use was more strongly associated with alleviation of depressive symptoms, perhaps suggesting that it has greater detrimental clinical impact, including on depression, than other drugs in this population.
Our results showing that cessation of substance use is associated with improved depression scores are not surprising, although it was reassuring to see the level of concrete improvement in routine care environments. It is well known that depression is associated with substance use.23–25 There is also evidence that substance use may interfere with pharmacologic treatment for depression in PLWH, resulting in less benefit of depression treatment in randomized trials,26 although there may be less impact for cognitive–behavioral therapy.27,28
Treatment for depression in PLWH is important because depression is a known barrier to HIV medication adherence29,30 and a source of morbidity and mortality in its own right.31,32 Substance use interferes with improving depression and should be addressed as a part of any treatment plan. Our results should provide additional support for studying interventions that lead to reductions in substance use among PLWH even when complete abstinence is not feasible.
Bidirectional associations seem to be present, in that participants who reduced their PHQ-9 score over follow-up also had fewer days of substance use for most substances. Although in clinical care there is often an emphasis on substance use cessation before diagnosis and treatment of mental health disorders due to the confounding impact of substance use on diagnosis of mental health disorders, treatment for depression can be effective in people with substance use disorders,33 including PLWH.26 These results suggest that there may also be a role for treating depression in parallel with efforts to treat substance use disorders.
Our study had key strengths including the longitudinal nature of the data and the high-quality measures of both our exposure and outcome. The use of linear mixed models allowed us to account for the irregular nature of the data and loss to follow-up in this population.
Our study had several key limitations. The population in these studies had lower levels of opioid use than other substances; so, we may have been underpowered to detect opioid associations. The study is inherently observational and although we adjusted our estimates for obvious confounders, the possibility of residual confounding remains. The measures of substance use were self-reported and were collected using categories of frequency of use, preventing us from establishing a clear threshold for reduction in order for participants to show benefit.
The results of this study suggest that reduction in substance use can result in better psychological outcomes for PLWH dealing with depressive symptoms. These benefits are particularly striking among ATS users (mostly methamphetamine), who showed the greatest benefits from cessation or reduction of use. Although clearly cessation of substance use should remain the target of public health interventions, our findings show that reduction in level of use also confers benefit for PLWH dealing with substance use issues.
The authors thank all CNICS and Project STRIDE participants and study personnel for their essential contributions to this work.
1. Altice FL, Kamarulzaman A, Soriano VV, et al. Treatment of medical, psychiatric, and substance-use comorbidities in people infected with HIV
who use drugs. Lancet. 2010;376:367–387.
2. Berger-Greenstein JA, Cuevas CA, Brady SM, et al. Major depression in patients with HIV
/AIDS and substance abuse. AIDS Patient Care STDS. 2007;21:942–955.
3. Tegger MK, Crane HM, Tapia KA, et al. The effect of mental illness, substance use
, and treatment for depression on the initiation of highly active antiretroviral therapy among HIV
-infected individuals. AIDS Patient Care STDS. 2008;22:233–243.
4. Bengtson AM, Pence BW, Moore R, et al. Relationship between ever reporting depressive symptoms
and all-cause mortality in a cohort of HIV
-infected adults in routine care. AIDS. 2017;31:1009–1016.
5. Dalessandro M, Conti CM, Gambi F, et al. Antidepressant therapy can improve adherence to antiretroviral regimens among HIV
-infected and depressed patients. J Clin Psychopharmacol. 2007;27:58–61.
6. Yun LW, Maravi M, Kobayashi JS, et al. Antidepressant treatment improves adherence to antiretroviral therapy among depressed HIV
-infected patients. J Acquir Immune Defic Syndr. 2005;38:432–438.
7. Cook JA, Grey D, Burke J, et al. Depressive symptoms
and AIDS-related mortality among a multisite cohort of HIV
-positive women. Am J Public Health. 2004;94:1133–1140.
8. Koob GF. Negative reinforcement in drug addiction: the darkness within. Curr Opin Neurobiol. 2013;23:559–563.
9. Kitahata MM, Rodriguez B, Haubrich R, et al. Cohort profile: the Centers for AIDS Research Network of Integrated Clinical Systems. Int J Epidemiol. 2008;37:948–955.
10. Chandler R, Gordon MS, Kruszka B, et al. Cohort profile: seek, test, treat and retain United States criminal justice cohort. Subst Abuse Treat Prev Policy. 2017;12:24.
11. Crane HM, Lober W, Webster E, et al. Routine collection of patient-reported outcomes in an HIV
clinic setting: the first 100 patients. Curr HIV
12. WHO ASSIST Working Group. The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST): development, reliability and feasibility. Addiction. 2002;97:1183–1194.
13. Arroll B, Goodyear-Smith F, Crengle S, et al. Hatcher S.Validation of PHQ-2 and PHQ-9 to screen for major depression in the primary care population. Ann Fam Med. 2010;8:348–353.
14. Cholera R, Gaynes BN, Pence BW, et al. Validity of the Patient Health Questionnaire-9 to screen for depression in a high-HIV
burden primary healthcare clinic in Johannesburg, South Africa. J Affect Disord. 2014;167:160–166.
15. Crane PK, Gibbons LE, Willig JH, et al. Measuring depression levels in HIV
-infected patients as part of routine clinical care using the nine-item Patient Health Questionnaire (PHQ-9). AIDS Care. 2010;22:874–885.
17. Delaney JA, Moodie EE, Suissa S. Validating the effects of drug treatment on blood pressure in the General Practice Research Database. Pharmacoepidemiol Drug Saf. 2008;17:535–545.
18. Diggle PJ. An approach to the analysis of repeated measurements. Biometrics. 1988;44:959–971.
19. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38:963–974.
20. Manea L, Gilbody S, McMillan D. Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CMAJ. 2012;184:E191–E196.
21. Gould AL, Boye ME, Crowther MJ, et al. Joint modeling of survival and longitudinal non-survival data: current methods and issues. report of the DIA Bayesian joint modeling working group. Stat Med. 2015;34:2181–2195.
22. Sweeting MJ, Thompson SG. Joint modelling of longitudinal and time-to-event data with application to predicting abdominal aortic aneurysm growth and rupture. Biom J. 2011;53:750–763.
23. Reddon H, Pettes T, Wood E, et al. Incidence and predictors of mental health disorder diagnoses among people who inject drugs in a Canadian setting. Drug Alcohol Rev. 2017;37 (suppl 1):S285–S293.
24. Tsuyuki K, Pitpitan EV, Levi-Minzi MA, et al. Substance use
disorders, violence, mental health, and HIV
: differentiating a syndemic factor by gender and sexuality. AIDS Behav. 2017;21:2270–2282.
25. Mukerji S, Haghighat R, Misra V, et al. Longitudinal modeling of depressive trajectories among HIV
-infected men using cocaine. AIDS Behav. 2017;21:1985–1995.
26. Grelotti DJ, Hammer GP, Dilley JW, et al. Does substance use
compromise depression treatment in persons with HIV
? Findings from a randomized controlled trial. AIDS Care. 2017;29:273–279.
27. Labbe AK, O'Cleirigh CM, Stein M, et al. Depression CBT treatment gains among HIV
-infected persons with a history of injection drug use varies as a function of baseline substance use
. Psychol Health Med. 2015;20:870–877.
28. Safren SA, O'Cleirigh CM, Bullis JR, et al. Cognitive behavioral therapy for adherence and depression (CBT-AD) in HIV
-infected injection drug users: a randomized controlled trial. J Consult Clin Psychol. 2012;80:404–415.
29. Shubber Z, Mills EJ, Nachega JB, et al. Patient-reported barriers to adherence to antiretroviral therapy: a systematic review and meta-analysis. Plos Med. 2016;13:e1002183.
30. Lima VD, Geller J, Bangsberg DR, et al. The effect of adherence on the association between depressive symptoms
and mortality among HIV
-infected individuals first initiating HAART. AIDS. 2007;21:1175–1183.
31. Ickovics JR, Hamburger ME, Vlahov D, et al. HIV
Epidemiology Research Study Group. Mortality, CD4 cell count decline, and depressive symptoms
-seropositive women: longitudinal analysis from the HIV
Epidemiology Research Study. JAMA. 2001;285:1466–1474.
32. Farinpour R, Miller EN, Satz P, et al. Psychosocial risk factors of HIV
morbidity and mortality: findings from the Multicenter AIDS Cohort Study (MACS). J Clin Exp Neuropsychol. 2003;25:654–670.
33. Zhou X, Qin B, Del Giovane C, et al. Efficacy and tolerability of antidepressants in the treatment of adolescents and young adults with depression and substance use
disorders: a systematic review and meta-analysis. Addiction. 2015;110:38–48.