Effects of Alcohol and Substance Use On Adherence and HIV Viral Suppression

Background:
Alcohol and substance use have been variably associated with unfavorable effects on antiretroviral adherence and HIV plasma viral load (pVL). No publications have explored the extent to which effects of alcohol and substance use are mediated through adherence. We hypothesized that adherence is an important mediator of alcohol and substance use effects on pVL.

Methods:
We conducted a retrospective cross-sectional analysis of patient-reported outcomes (PRO) and the most recent subsequent HIV viral load collected as part of routine care in an urban HIV medical care cohort. The last recorded PRO and matched pVL during the study period 2014-2108 were analyzed. Eligible patients were ≥ 18 years, HIV-infected, and consented to participate. Measures included the AUDIT and NIDA ASSIST as measures of hazardous drinking (AUDIT≥8) and recent (past 3 months) substance misuse, respectively. Past 4-week adherence was self-rated as excellent-very poor using a 5-point Likert scale. Adherence was binarily modeled (excellent vs. not excellent). Those substances found in bivariate logistic regression analysis to predict subsequent detectable (≥200 copies) pVL (p < 0.05) were modeled separately for mediation effects through adherence using the Stata paramed program. In addition, the number of substances used was modeled as an independent variable (IV). IV-mediator (M) interactions were included if significant at p<0.10. We report natural direct (NDE) and indirect effects (NIE), total effect (TE), and percentage mediated.

Results:
3125 patients met eligibility requirements, including 12.3% female, 53.6% non-white, median [IQR] age 51 [41,58], HIV risk factor IDU 10.9%, 9.8% detectable pVL, 57.7% excellent self-reported adherence, median [IQR] days between PRO and pVL 24 [2, 120]. AUDIT≥8 was reported in 3%. Past 3-month substance use, by rank order: cannabis 27.1%, methamphetamines 13.1%, inhalants 11.9%, cocaine 5.3%, sedative-hypnotics 4.5%, opioids 2.9%, hallucinogens 2.3%. By number of substances used, 0 (59.8%), 1 (23.5%), 2 (10.4%), 3 (3.8%), 4-7 (2.5%). Bivariate predictors of detectable pVL were [OR (95% C.I.)]: amphetamines 2.4 (1.8-3.2), opioids 2.3 (1.3-4.0), alcohol 3.6 (2.2-5.8), and number substances used 1.2 (1.1-1.4). Mediation results are presented in the Table. Significant IV-M interactions were included for alcohol (p < 0.01).

Conclusion:
Alcohol, methamphetamine, opioid use, and the number of substances used were predictive of detectable pVL. However, less than 40% of the effects were mediated by adherence.

References: 1. Arnsten JH, Demas PA, Grant RW, Gourevitch MN, Farzadegan H, Howard AA, et al. Impact of active drug use on antiretroviral therapy adherence and viral suppression in HIV-infected drug users. J Gen Intern Med. 2002;17(5):377-81.
2. Chander G, Himelhoch S, Fleishman JA, Hellinger J, Gaist P, Moore RD, et al. HAART receipt and viral suppression among HIV-infected patients with co-occurring mental illness and illicit drug use. AIDS Care. 2009;21(5):655-63.
3. Emsley RA, Liu H, Dunn G, L. V, VanderWeele TJ. PARAMED: Causal mediation analysis using parametric models. 2013.
4. Feldman MB, Thomas JA, Alexy ER, Irvine MK. Crystal methamphetamine use and HIV medical outcomes among HIV-infected men who have sex with men accessing support services in New York. Drug Alcohol Depend. 2015;147:266-71.
5. Morojele NK, Kekwaletswe CT, Nkosi S. Associations between alcohol use, other psychosocial factors, structural factors and antiretroviral therapy (ART) adherence among South African ART recipients. AIDS Behav. 2014;18(3):519-24.
6. Parry CD, Londani M, Shuper PA, Myers B, Kekwaletswe CT, Nkosi S, et al. Characteristics and drinking behaviour of patients on antiretroviral therapy who drink and attend HIV clinics in Tshwane, South Africa: Implications for intervention. S Afr Med J. 2019;109(10):784-91.
7. Valeri L, VanderWeele TJ. Mediation analysis allowing for exposure–mediator interactions and causal interpretation: Theoretical assumptions and implementation with SAS and SPSS macros. Psychological Methods. 2013;18(2):137-50.
8. VanderWeele TJ. Mediation Analysis: A Practitioner's Guide. Annu Rev Public Health. 2016;37:17-32.
9. VanderWeele TJ. Frontiers of Power Assessment in Mediation Analysis. Am J Epidemiol. 2020.
10. Vanderweele TJ, Vansteelandt S. Odds ratios for mediation analysis for a dichotomous outcome. Am J Epidemiol. 2010;172(12):1339-48.

Learning Objective: No publications have explored the extent to which effects of alcohol and substance use are mediated through adherence. We hypothesized that adherence is an important mediator of alcohol and substance use effects on plasma viral HIV load (pVL).