Abstract
Antiretroviral therapy (ART) reduces HIV disease burden, increases life expectancy, and prevents HIV transmission. Previous research suggests that believing that it is harmful to take ART when using substances (i.e., interactive toxicity beliefs) leads to intentional ART nonadherence; however, these associations have not been investigated among younger adults living with HIV and have not been linked to clinical outcomes. We examined the associations among interactive toxicity beliefs, intentional nonadherence, and HIV clinical outcomes in young adults living with HIV. People living with HIV younger than the age of 36 years who tested positive for at least one substance use biomarker (N = 406) completed a 1-month pretrial run-in study that included computerized interviews, substance use biomarkers, HIV viral load, and unannounced pill counts for ART adherence. Analyses compared three HIV clinical outcome groups: (1) HIV viral unsuppressed, (2) HIV viral suppressed and ART nonadherent, and (3) HIV viral suppressed and ART adherent, on substance use, interactive toxicity beliefs, and substance use-related intentional ART nonadherence. Results showed that a majority of participants reported intentional nonadherence. Participants with unsuppressed HIV reported greater interactive toxicity beliefs and intentional nonadherence. We conclude that intentional nonadherence adds to the detrimental impacts of substance use on ART adherence and interactive toxicity beliefs that foster these behaviors may be amenable to interventions.
Keywords: HIV treatment, substance use, intentional nonadherence, young adults living with HIV
Introduction
Ending HIV requires forward progression through the HIV continuum of care: identifying people who have contracted HIV, engaging them in HIV care, and providing antiretroviral therapy (ART), which effectively suppresses HIV replication.1,2 Unfortunately, not all individuals progress along the continuum of HIV care. Among the populations least likely to engage in HIV care and sustain ART adherence are individuals younger than the age of 35 years, particularly those persons who use alcohol and other drugs.3–5 An analysis of over 33,000 patients receiving care from 8 HIV clinics between 1997 and 2015 found that after controlling for multiple confounding variables, younger age was associated with poorer treatment outcomes, specifically less likelihood of achieving undetectable HIV viral load.6
In terms of substance use, alcohol consumption is closely associated with not achieving each step along the HIV continuum of care, with greater alcohol use observed in patients with suboptimal ART adherence and unsuppressed HIV viral load.7,8 In terms of illicit drug use, methamphetamine use is a reliable predictor of unsuppressed HIV viral load.9 Across multiple countries, cannabis and other drug use is associated with poorer HIV treatment engagement, adherence, and unsuppressed HIV.10,11
Multiple factors contribute to the detrimental impact of substance use on HIV treatment outcomes, including direct social, cognitive, and behavioral effects and after effects, of intoxication.12 In addition to the direct effects of substance use, beliefs that individuals often hold about the potential harms of mixing medications with alcohol and other drugs, referred to as interactive toxicity beliefs, and their associated behaviors also impede progression along the HIV continuum of care. A recent review of 17 studies of alcohol-ART interactive toxicity beliefs found that studies reliably report close associations between interactive toxicity beliefs, intentional nonadherence to ART, and poor viral suppression.13 While some individuals who believe it is harmful to mix alcohol and other drugs with ART stop their substance use when initiating ART, others continue to use alcohol and other drugs despite believing that it is harmful.
Also, for those who do continue substance use, interactive toxicity beliefs may lead some individuals to not drink or use drugs in conjunction with taking ART, while others may not take ART when they are drinking or using drugs. For example, a study of 530 people living with HIV who were actively using nonalcohol drugs found that more than one in three (35%) chose to not take ART when using drugs, and results showed that intentional nonadherence was a function of holding interactive toxicity beliefs.14 The degree to which interactive toxicity beliefs and substance use-related intentional nonadherence to ART contribute to unsuppressed HIV viral load in younger adults, however, has not yet been reported.
The current study examined interactive toxicity beliefs and intentional ART nonadherence in young adults living with HIV. Unsuppressed HIV viral load is, by definition, the clinical marker for ART nonadherence and a negative outcome for HIV treatment. Patients with unsuppressed HIV experience advancing clinical HIV disease, greater morbidity, and remain infectious.15 In addition, individuals who are HIV suppressed but demonstrate suboptimal ART adherence are at risk for viral breakthrough.16 Thus far, studies have not examined the role of substance use-related interactive toxicity beliefs and their associated intentional nonadherence among people who have unsuppressed HIV and those at-risk for unsuppressed HIV.
We hypothesized that among substance using young adults receiving ART, interactive toxicity beliefs and intentional nonadherence would distinguish both those who have an unsuppressed HIV viral load and those at-risk for an unsuppressed HIV viral load in comparison to individuals who are HIV suppressed and ART adherent. Furthermore, we hypothesized that intentional nonadherence would differentiate participants with unsuppressed HIV from those who are at-risk for unsuppressed HIV due to ART nonadherence.
Methods
Participants and setting
Young adults living with HIV (N = 591) were recruited during the period between September 2017 and December 2019. Eligible persons were aged 18 to 36 (median = 29) and all participants showed documentation of age and HIV-positive status (e.g., antiretroviral medication prescription, HIV test result, viral load laboratory result, and HIV clinic card). The study was conducted in Atlanta, Georgia, which has an annual HIV incidence of 25.9 per 100,000 population, the seventh highest rate of HIV among US cities.17
Procedures
Individuals living with HIV were recruited to take part in a 1-month run-in study to determine eligibility for a treatment engagement and ART adherence intervention trial (registered clinicaltrials.gov NCT03665532). People living with HIV were recruited through social media websites, targeted online advertisements, and a participant-driven adaptation of snowball-sampling techniques, where participants were encouraged to refer their acquaintances living with HIV to the study and were offered $25 for their efforts. The total sample for the run-in was 538 individuals assigned male at birth and 57 individuals assigned female at birth, with 61 identifying as transgender. Following informed consent, participants completed measures of demographic and health characteristics, including computer-assisted self-interviews (CASI), provided blood samples for HIV viral load testing, and urine samples for substance use. In addition, participants completed two phone-based unannounced pill counts over 1-month to prospectively assess ART adherence.
A total of 570 (96%) participants who completed the initial CASI assessment and provided biological specimens were included in this study. Participants were reimbursed $193 for completing all measures, including the CASI, two phone assessments with pill counts, daily sexual behavior surveys (not included in this article), and providing blood and urine specimens. Participant privacy was protected through a federal certificate of confidentiality and the University of Connecticut Institutional Review Board approved all procedures.
Measures
Computer assisted self-interviews
Demographic and health characteristics
Participants were asked their self-identified age, sex, gender identity, race, and income. Participants also reported the year that they tested HIV-positive and the year they initiated ART. In addition, participants completed the Center for Epidemiological Studies Depression (CES-Depression) scale to assess symptoms of depression,18 Cronbach's α = 0.91.
Self-reported substance use
To assess global alcohol use, we administered the Alcohol Use Disorders Identification Test (AUDIT), a 10-item scale designed to measure alcohol consumption and identify risks for alcohol abuse and dependence.19 The first three items of the AUDIT represent quantity and frequency of alcohol use and the remaining seven items concern problems incurred from drinking alcohol. The AUDIT has demonstrated reliability and validity,20 Cronbach's α = 0.90.
We also administered the World Health Organization's Alcohol Substance Involvement Screening Test (ASSIST) to measure self-reported drug dependence.21 Participants were asked whether they had used cannabis, cocaine, amphetamines, methamphetamine, inhalants, sedatives, hallucinogens, and opiates in the previous 4 months. Responses to questions of any use in the past 4 months were summed to yield a score of self-reported number of drugs used. For any drugs used, participants also responded to questions regarding their desire for the drug, financial problems caused by its use, being unable to meet responsibilities because of its use, having a friend or family express concern about its use, and failed attempts to cut back on its use.
Responses were made on 5-point scales, from 0 = “Never in the past four months” to 4 = “Daily or almost daily” in the past 4 months. Scores were averaged across indicators of dependence and across substances. Participants who reported no substance use scored 0 on the drug dependence scale.
Substance use interactive toxicity beliefs
Participants completed five items to assess beliefs about drinking alcohol and using other drugs in conjunction with taking ART. The items were adapted from previous research22 and represented beliefs that it is hazardous to take ART when drinking alcohol or using other drugs. Example items were, “Alcohol and HIV medications should never be mixed,” “Alcohol breaks down HIV medications so they will not work right,” “Marijuana, cocaine, and other drugs should never be mixed with HIV medications,” and “A person should skip taking their HIV medications if they are doing drugs” responded to on 6-point ratings, 1 = “Strongly disagree,” to 6 = “Strongly agree,” averaged across items, Cronbach's α = 0.70.
Substance use-related intentional nonadherence
Participants responded to five items assessing behaviors that explicitly refer to not taking ART when using substances. These behaviors were adapted from previous research on substance use-related intentional nonadherence.23,24 Example behaviors include “I skip taking my medications because I have been or will be drinking alcohol,” and “I skip taking my medications because I am drinking alcohol and do not want to mix them.” Items focused on alcohol use as lower threshold for substance use and were responded to regarding whether participants personally performed each of the five actions, 0 = “Never,” 4 = “Often.” Intentional ART nonadherence scores represent mean responses, Cronbach's α = 0.88.
Biomarkers for drug and alcohol use
Biomarkers for current drug use were assessed by a multi-panel urine dip-test to detect common illicit drug use. This test strip uses a lateral flow chromatographic immunoassay for qualitative detection of 12 drugs/metabolites, including cannabis (tetrahydrocannabinol, THC), cocaine, opiates, amphetamine, barbiturates, and methamphetamine (Reditest-12; Redwood Toxicology Laboratories). These tests are FDA approved and are reliable and valid for detecting recent use (72–96 h). In this study, cannabis use indicated by THC was examined separately from the other 11 drugs. We also used a point-of-care urine dip test for ethyl glucuronide (EtG) to index recent alcohol use. Urine EtG tests have shown 50% sensitivity for detecting alcohol use in 24 h and 100% sensitivity for alcohol use in 12 h.25 We created an index of drug use that included all substances, and a separate index that excluded cannabis and alcohol.
ART adherence
Participants completed two unannounced phone-based pill counts that occurred over a 1-month period. Unannounced pill counts are reliable and valid in assessing medication adherence when conducted in homes26 and on the phone.27,28 Following an office-based training session that included a full accounting of all prescription medications, participants were called at two unscheduled times over 30 days. All ART medications were included in the pill counts and calculating adherence. The first pill count was used to establish the initial number of pills in possession with the subsequent pill count allowing for calculating adherence, defined as the ratio of pills counted relative to pills prescribed, taking into account the number of pills dispensed.29,30 In addition, the second assessment asked participants whether they had intentionally not taken their ART in the previous month to avoid mixing medications with alcohol and other drugs.
HIV viral load
To determine HIV RNA plasma concentrations (viral load), participants provided 80 μL of fingerstick blood for dried blood spots collected in Hemaspot HF™ devices that were frozen at −20°C before laboratory delivery. HIV-1 viral load testing was conducted using the Abbott RealTime HIV-1 assay, a reverse transcription–polymerase chain reaction assay performed on the automated Abbott m2000 platform (Abbott Molecular, Inc., Des Plaines, IL).31 The target sequence for the assay is the highly conserved pol/integrase region of the HIV-1 genome. The limit of detection of the assay is 2.92 log (830 copies/mL), and it can quantify up to 7.0 log (10,000,000) copies/mL,32 a level of detection that nearly eliminates most errors caused by viral load blips or assay variability.33 All samples required upfront processing to improve assay sensitivity before subjecting to RNA extraction.
Data analyses
We first conducted descriptive analyses comparing participants who did not test positive for any substance use (n = 164) to those who did test positive (n = 406). Groups were compared using contingency table χ2 tests for categorical variables and independent t-tests for continuous measures. For the main study that only included participants who had tested positive for any substance use, we formed three HIV clinical outcome groups: Group A, unsuppressed HIV viral loads; Group B, HIV suppressed and were ART nonadherent (<70% pills taken); and Group C, HIV suppressed and ART adherent (≥70% pills taken). Descriptive analyses compared the three groups on demographic, health, and substance use variables using contingency table χ2 tests for categorical variables and analyses of variance (ANOVA) for continuous measures.
To test the main study hypotheses that interactive toxicity beliefs and intentional nonadherence would differentiate the clinical outcome groups, we conducted four a priori hierarchically ordered multivariable multinomial logistic regression models. These models were ordered: (1) distal factors (income, years since testing HIV positive, and CES-Depression scores); (2) substance use (total number of substance use biomarkers that were positive, AUDIT alcohol scores, and self-reported substance dependence); (3) substance use interactive toxicity beliefs; and (4) substance use-related intentional nonadherence.
Overall model χ2 tests with associated Cox-Snell estimates of variance accounted for (pseudo R2) are reported for each model. Odds ratios (ORs) are reported for paired contrasts comparing the HIV-unsuppressed Group A versus the HIV-suppressed and adherent Group C, and comparing the HIV-suppressed but nonadherent Group B versus the HIV-suppressed and adherent Group C. The HIV-suppressed and ART-adherent Group C was entered as the reference in all models. All statistical tests defined significance by p < 0.05 and statistical trends defined by p < 0.10.
Results
Among the 570 participants, 537 (93%) were currently receiving ART, of which 186 (32%) were not HIV viral suppressed and 267 (46%) were less than 70% adherent over the 1-month observation. Results of the substance use screening showed that 410 (71%) participants tested positive for at least 1 drug. Table 1 shows the sample characteristics partitioned by participants who did not test positive for any drug (n = 106) and those who tested positive for at least one drug (n = 406). The most common drug detected in the biomarker screening was cannabis (n = 337, 59%), followed by amphetamine or methamphetamine (n = 129, 22%) and EtG (alcohol, n = 121, 21%). Participants who tested positive for any substance use scored higher on the CES-Depression scale, demonstrated poorer ART adherence, and were less likely to have suppressed HIV. All remaining analyses focused only on those 406 participants who tested substance use positive and were receiving ART.
Table 1.
Sample Demographic and Health Characteristics Partitioned by Results from Screening for Substance Use Biomarkers
| Negative substance use biomarkers, n = 164 |
Positive substance use biomarkers, n = 406 |
t | |||
|---|---|---|---|---|---|
| M | SD | M | SD | ||
| Age | 29.0 | 4.0 | 29.1 | 3.7 | 0.7 |
| Years of education | 13.5 | 1.3 | 13.3 | 1.4 | 1.9 |
| Years since testing HIV positive | 6.1 | 5.2 | 6.4 | 5.3 | 0.7 |
| Years since initiating ART | 3.4 | 3.6 | 3.6 | 3.7 | 0.5 |
| ART adherence | 76.5 | 24.9 | 69.3 | 27.1 | 2.8** |
| Log HIV RNA copies/mL | 0.9 | 1.6 | 1.2 | 1.8 | 2.1* |
| CES-depression | 17.2 | 12.1 | 20.0 | 12.4 | 2.5** |
| AUDIT alcohol | 3.3 | 4.1 | 5.2 | 5.4 | 4.1** |
| Self-report drugs used | 1.4 | 1.6 | 2.3 | 1.7 | 5.6** |
| Self-report drug dependence | 3.3 | 6.8 | 12.5 | 13.7 | 8.1** |
| N | % | N | % | χ2 | |
|---|---|---|---|---|---|
| Male |
146 |
89 |
372 |
92 |
0.9 |
| Female |
18 |
11 |
34 |
8 |
|
| Transgender |
13 |
8 |
48 |
12 |
1.7 |
| African American |
144 |
88 |
367 |
90 |
0.9 |
| Annual income ≤$20,000 |
86 |
52 |
253 |
62 |
12.0 |
| $21,000–$39,000 |
59 |
36 |
109 |
26 |
|
| ≥$40,000 |
19 |
12 |
44 |
12 |
|
| Has a primary care provider |
152 |
93 |
377 |
92 |
0.1 |
| ART adherence ≤70% |
59 |
36 |
203 |
50 |
9.1** |
| ART adherence >70% |
105 |
64 |
203 |
50 |
|
| Undetected HIV RNA |
122 |
74 |
266 |
65 |
3.8* |
| Detected HIV RNA | 43 | 26 | 140 | 35 |
p < 0.05, **p < 0.01.
ART, antiretroviral therapy; AUDIT, Alcohol Use Disorders Identification Test; CES, Center for Epidemiological Studies; SD, standard deviation.
Among substance using participants, 54% (n = 220) endorsed at least one of the indicators of intentional nonadherence, with the most frequent behavior being “waiting to drink alcohol until I am not taking HIV medications,” endorsed by 34% (n = 140) of the sample. Overall 54% (n = 220) of participants endorsed at least one of the intentional nonadherence behaviors. During the month long study, one in five participants (n = 80) reported forgoing their ART because they were using alcohol or other drugs. Specifically, 16% (n = 66) reported not taking ART because they were drinking alcohol and 10% (n = 41) did so because they were using other drugs. Importantly, intentional nonadherence over the previous month was not associated with the clinical outcomes groups, indicating the pervasiveness of intentional nonadherence throughout the sample.
Factors associated with HIV clinical outcomes
Comparisons of the three clinical outcome groups on demographic, health, and substance use characteristics are shown in Table 2. Results indicated that participants with unsuppressed HIV had been living longer with HIV than participants with suppressed HIV. In addition, unsuppressed HIV was associated with having tested positive for more substances, particularly amphetamine and methamphetamine, and scoring higher on the drug dependence scale. In addition, participants with unsuppressed HIV and those who were suppressed/nonadherent had lower incomes than participants who were suppressed/adherent. Finally, results showed that participants with unsuppressed HIV had higher interactive toxicity belief scores and indicated greater substance use-related intentional nonadherence than the suppressed/nonadherent group, and the suppressed/nonadherent group had greater substance use-related intentional nonadherence than the suppressed/adherent group.
Table 2.
Demographic and Health Characteristics of Substance Using Young Adults in the Three Clinical Outcome Groups
| Group A: HIV unsuppressed, n = 140 |
Group B: HIV suppressed and ART nonadherent, n = 119 |
Group C: HIV suppressed and ART adherent, n = 147 |
F | ||||
|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | ||
| Years since testing HIV positive | 7.4 | 6.1 | 5.8 | 5.0 | 6.1 | 4.5 | 6.7* |
| CES-depression | 21.0 | 12.3 | 20.6 | 13.0 | 18.7 | 11.9 | 2.8 |
| ART adherence | 66.1 | 28.0 | 45.5 | 19.6 | 90.1 | 10.0 | n/aa |
| AUDIT alcohol | 5.0 | 5.4 | 5.6 | 5.8 | 5.2 | 5.2 | 0.6 |
| Number of substance use biomarkers detected | 1.8 | 1.0 | 1.5 | 0.8 | 1.5 | 0.8 | 10.5** |
| Self-report drug dependence | 13.4 | 15.9 | 14.7 | 15.0 | 9.8 | 9.3 | 10.4** |
| Substance use—ART interactive toxicity beliefs | 3.6 | 1.3 | 3.3 | 1.2 | 3.0 | 1.1 | 11.7** |
| Substance use—related Intentional ART nonadherence | 1.2 | 1.4 | 0.7 | 1.1 | 0.4 | 0.6 | 19.8** |
| N | % | N | % | N | % | χ2 | |
|---|---|---|---|---|---|---|---|
| Annual income ≤$20,000 |
93 |
66 |
76 |
64 |
84 |
57 |
8.2* |
| $21,000–$39,000 |
37 |
26 |
37 |
31 |
35 |
24 |
|
| ≥$40,000 |
10 |
7 |
6 |
5 |
28 |
19 |
|
| Substance use biomarkers detected | |||||||
| Cocaine |
25 |
17 |
27 |
22 |
22 |
14 |
2.6 |
| Amphetamine |
29 |
20 |
13 |
11 |
13 |
9 |
9.1** |
| Methamphetamine |
32 |
23 |
17 |
14 |
13 |
9 |
10.9** |
| Any drugs above usedb |
59 |
42 |
43 |
36 |
46 |
31 |
3.6 |
| THC cannabis |
111 |
78 |
106 |
88 |
120 |
81 |
5.1 |
| EtG—alcohol |
41 |
29 |
39 |
33 |
41 |
28 |
0.8 |
| Reason missed ART in the past month | |||||||
| Did not want to mix alcohol and ART |
21 |
32 |
25 |
36 |
20 |
25 |
2.96 |
| Did not want to mix other drugs and ART | 15 | 23 | 17 | 24 | 9 | 11 | 4.97+ |
Groups formed on adherence invalidates difference testing.
Excluding THC and EtG.
p = 0.08, *p < 0.05, **p < 0.01.
ART, antiretroviral therapy; AUDIT, Alcohol Use Disorders Identification Test; CES, Center for Epidemiological Studies; EtG, ethyl glucuronide; SD, standard deviation; THC, tetrahydrocannabinol.
Multivariable multinomial models for HIV clinical outcomes
Results from the four hierarchically ordered multivariable multinomial logistic regression models are shown in Table 3. Demographic characteristics entered in Model 1 were significantly associated with clinical outcomes, χ2 = 15.06, p = 0.02, and R2 = 0.037; years since testing positive significantly contributed to the model. Model 2, which added substance use variables, remained significant, χ2 = 27.21, p = 0.007, and R2 = 0.066; years since testing HIV positive remained significant and drug dependence significantly added to the model. The bivariate models showed the HIV-unsuppressed group had tested positive for more drugs and had greater drug dependence than the suppressed/adherent group, while the suppressed/nonadherent group had greater drug dependence than the suppressed/adherent group. Model 3 added interactive toxicity belief scores and remained significant, χ2 = 36.5, p = 0.001, and R2 = 0.088; only years since testing HIV positive and interactive toxicity beliefs were significant.
Table 3.
Multinomial Model Effects and Group Contrast Odds Ratios for Three Clinical Outcome Groups Along the HIV Continuum of Care
| Predictor variable | Model 1 |
Model 2 |
Model 3 |
Model 4 |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Model effects, χ2 | Group A, OR | Group B, OR | Model effects, χ2 | Group A, OR | Group B, OR | Model effects, χ2 | Group A, OR | Group B, OR | Model effects, χ2 | Group A, OR | Group B, OR | |
| Income | 5.97+ | 0.83* | 0.84 | 3.97 | 0.86 | 0.86 | 2.67 | 0.89 | 0.87 | 2.05 | 0.90 | 0.88 |
| Years since testing HIV positive | 7.02* | 1.04+ | 0.98+ | 7.02* | 1.04 | 0.98 | 6.83* | 1.04 | 0.98 | 5.50 | 1.03 | 0.97 |
| CES-depression | 1.21 | 1.01 | 1.00 | 0.24 | 1.00 | 1.00 | 0.03 | 1.00 | 1.00 | 0.01 | 1.00 | 0.99 |
| AUDIT alcohol use | 0.92 | 0.98 | 1.01 | 0.82 | 0.99 | 1.01 | 2.43 | 0.96 | 0.99 | |||
| Total number of drug/alcohol use biomarkers detected | 5.14+ | 1.32* | 1.28 | 4.77+ | 1.31* | 1.27 | 4.27 | 1.30 | 1.27 | |||
| Self-report drug dependence | 6.05* | 1.37* | 1.40* | 5.76+ | 1.36* | 1.41** | 3.64 | 1.24 | 1.34 | |||
| Substance use—ART interactive toxicity beliefs | 9.67** | 1.38** | 1.22+ | 3.34 | 1.22 | 1.13 | ||||||
| Substance use—related Intentional ART nonadherence | 25.65** | 1.97** | 1.49** | |||||||||
Clinical outcome group: Group A = HIV unsuppressed, Group B = HIV suppressed and ART nonadherent, Group C (Ref) = HIV suppressed and ART adherent.
p < 0.10, *p < 0.05, **p < 0.01.
ART, antiretroviral therapy; AUDIT, Alcohol Use Disorders Identification Test; CES, Center for Epidemiological Studies; OR, odds ratio.
In the bivariate models, both the HIV-unsuppressed and HIV-suppressed/nonadherent groups had higher interactive toxicity belief scores than the HIV-suppressed/adherent group. The final model added substance use intentional nonadherence and remained significant, χ2 = 62.23, p < 0.001, R2 = 0.145; intentional nonadherence was the only significant variable in the model and bivariate models showed that both the unsuppressed and suppressed/nonadherent groups had greater substance use intentional nonadherence than the suppressed/adherent group.
In a separate bivariate regression model comparing the unsuppressed and suppressed/nonadherent groups on all variables, results showed that HIV-unsuppressed participants had tested positive significantly longer, OR = 1.04, p = 0.034, and 95% confidence interval (CI) 1.003–1.085 and reported greater substance use-related intentional nonadherence than the suppressed/nonadherent group, OR = 1.21, p = 0.025, and 95% CI = 1.025–1.448.
Discussion
Results from the current study confirmed our hypothesis that among substance using young adults receiving ART, interactive toxicity beliefs and intentional nonadherence would differentiate participants with unsuppressed HIV as well as those at risk for becoming viral unsuppressed due to lower adherence from those who are HIV suppressed and ART adherent. Furthermore, we found that substance use-related intentional nonadherence differentiated participants who were HIV unsuppressed from those at risk for becoming unsuppressed due to ART nonadherence. This pattern of associations indicates that individuals who hold stronger substance use interactive toxicity beliefs and practice more intentional ART nonadherence behaviors demonstrate poorer overall ART adherence and treatment failure than their substance using counterparts who held fewer interactive toxicity beliefs and practiced fewer intentional nonadherence behaviors.
It should, however, be noted that substance use-related intentional nonadherence was pervasive in this sample, with one in five participants forgoing their ART in relationship to substance use during the 1-month study period. Because all of the participants in this study were actively using substances as confirmed by biomarkers, the impacts of intentional nonadherence were observed over and above any other aspects of substance use that undermine adherence. Also, added impacts of intentional nonadherence occurred in a context of lower overall ART adherence, with an average of 70% of ART taken during the study period.
As ART has expanded with combinations of medications that are far more forgiving of modest nonadherence than earlier regimens,16,34,35 concerns about occasional forgotten doses of ART have subsided, with a greater focus on factors that undermine persistent adherence.36 Forgoing ART out of concerns about mixing medications with alcohol and other drugs constitutes a source of persistent nonadherence that can ultimately lead to poor clinical outcomes. This study demonstrates that substance use intentional ART nonadherence differentiates HIV treatment outcomes in terms of HIV suppression among substance using young adults living with HIV. In addition, intentional nonadherence further distinguished levels of ART adherence among individuals who were viral suppressed.
Specifically, individuals who were viral suppressed and below a commonly accepted minimal ART adherence threshold of 70%, and therefore at risk for viral breakthrough,16 held greater interactive toxicity beliefs and reported greater intentional nonadherence than their HIV-suppressed/adherent counterparts. Efforts to improve ART adherence among people who drink and use other drugs should directly address interactive toxicity beliefs and related intentional nonadherence.
The current study findings should be interpreted in light of their methodological limitations. The sample for this study was one of convenience and cannot be considered representative of young adults living with HIV and receiving ART. In addition, our main analyses focused on individuals who tested positive for substance use biomarkers. While it is unlikely that the sample included individuals who tested false positive for substance use, we recognize that we may have excluded active substance users whose last use did not occur in a timeframe for biomarker detection. Finally, our study included self-report measures of substance use and intentional nonadherence, both of which may have been underestimated by socially desirable responding.
Overall, young adults living with HIV are in need of ART adherence support, with those who are active substance users presenting greater urgency. Unfortunately, interventions that have aimed to modify alcohol use to improve ART adherence have demonstrated limited success,37,38 even in the context of incremental improvement in adherence with retention in care.39 We speculate that failure to adequately address intentional nonadherence has undermined the potential efficacy of these interventions. Intentional nonadherence is efficiently addressed by educating patients about the impacts of substance use on adherence, a universal component of previous interventions, and the safety of taking ART when drinking and using other drugs, which is typically not included in interventions. In the absence of co-occurring liver disease,40 there is no evidence for toxicity concerns between ART and alcohol, with more nuanced and minimal concerns regarding toxicity for other drug types.
However, many care providers themselves may not be fully informed of ART regimen toxicities and potential interactions. Educating patients about the interactive safety of taking ART when using alcohol and other drugs should avoid inadvertently suggesting that substance use does not carry other health hazards and does not otherwise impede adherence. In addition, providers should not overgeneralize the safety of taking ART when using substances to other medications, some which can be quite hazardous to mix with alcohol and other drugs. Future research is needed to develop and test effective patient and provider education interventions, as well as provider communication strategies, for disputing interactive toxicity beliefs and eliminating practices of intentionally forgoing medications.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
This research was supported by National Institute on Drug Abuse grant R01-DA033067.
References
- 1. Valdiserri RO, Holtgrave DR. Ending HIV in America: Not without the power of community. AIDS Behav 2019;23:2899–2903. [DOI] [PubMed] [Google Scholar]
- 2. Holtgrave DR, Greenwald R. A SWOT Analysis of the updated national HIV/AIDS strategy for the U.S., 2015–2020. AIDS Behav 2016;20:1–6. [DOI] [PubMed] [Google Scholar]
- 3. Quinn B, Gorbach PM, Okafor CN, et al. Investigating possible syndemic relationships between structural and drug use factors, sexual HIV transmission and viral load among men of colour who have sex with men in Los Angeles County. Drug Alcohol Rev 2020;39:116–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Shanaube K, Macleod D, Chaila MJ, et al. HIV care cascade among adolescents in a “Test and Treat” community-based intervention: HPTN 071 (PopART) for youth study. J Adolesc Health 2021;68:719–727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Amico KR, Crawford J, Ubong I, et al. Correlates of high HIV viral load and antiretroviral therapy adherence among viremic youth in the United States enrolled in an adherence improvement intervention. AIDS Patient Care STDS 2021;35:145–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Nance RM, Delaney JAC, Simoni JM, et al. HIV viral suppression trends over time among HIV-infected patients receiving care in the United States, 1997 to 2015: A cohort study. Ann Intern Med 2018;169:376–384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Williams EC, McGinnis KA, Edelman EJ, et al. Level of alcohol use associated with HIV care continuum targets in a national U.S. sample of persons living with HIV receiving healthcare. AIDS Behav 2019;23:140–151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Vagenas P, Azar MM, Copenhaver MM, et al. The impact of alcohol use and related disorders on the HIV continuum of care: A systematic review: Alcohol and the HIV continuum of care. Curr HIV/AIDS Rep 2015;12:421–436. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Aralis HJ, Shoptaw S, Brookmeyer R, et al. Psychiatric illness, substance use, and viral suppression among HIV-positive men of color who have sex with men in Los Angeles. AIDS Behav 2018;22:3117–3129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Idrisov B, Lunze K, Cheng DM, et al. Role of substance use in HIV care cascade outcomes among people who inject drugs in Russia. Addict Sci Clin Pract 2017;12:30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. De Boni RB, Peratikos MB, Shepherd BE, et al. Is substance use associated with HIV cascade outcomes in Latin America? PLoS One 2018;13:e0194228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Hendershot CS, Stoner SA, Pantalone DW, et al. Alcohol use and antiretroviral adherence: Review and meta-analysis. J Acquir Immune Defic Syndr 2009;52:180–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. El-Krab R, Kalichman SC. Alcohol-Antiretroviral Therapy Interactive Toxicity Beliefs and Intentional Medication Nonadherence: Review of Research with Implications for Interventions. AIDS Behav 2021. [Epub ahead of print]; DOI: 10.1007/s10461-021-03285-x. [DOI] [PubMed] [Google Scholar]
- 14. Kalichman SC, Kalichman MO, Cherry C, et al. Intentional medication nonadherence because of interactive toxicity beliefs among HIV-positive active drug users. J Acquir Immune Defic Syndr 2015;70:503–509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Akinosoglou K, Antonopoulou S, Katsarolis I, et al. Patient-reported outcomes in HIV clinical trials evaluating antiretroviral treatment: A systematic review. AIDS Care 2021;33:1118–1126. [DOI] [PubMed] [Google Scholar]
- 16. Byrd KK, Hou JG, Hazen R, et al. Antiretroviral adherence level necessary for HIV viral suppression using real-world data. J Acquir Immune Defic Syndr 2019;82:245–251. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. CDC. Diagnoses of HIV Infection in the United States and Dependent Areas, 2018. 2020. Available at: https://www.cdc.gov/hiv/library/reports/hiv-surveillance/vol-31/content/tables.html (Last accessed July 15, 2020).
- 18. Radloff LS. The CES-D scale: A self-report depression scale for research in the general populatio. Appl Psychol Meas 1977;1:385–401. [Google Scholar]
- 19. Saunders JB, Aasland OG, Babor TF, et al. Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption II. Addictions 1993;88:791–804. [DOI] [PubMed] [Google Scholar]
- 20. Maisto SA, Conigliaro J, McNeil M, et al. An empirical investigation of the factor structure of the AUDIT. Psychol Assess 2000;12:346–353. [DOI] [PubMed] [Google Scholar]
- 21. Humeniuk R, Ali R, Babor T, et al. A randomized controlled trial of a brief intervention for illicit drugs linked to the Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) in clients recruited from primary health-care settings in four countries. Addiction 2012;107:957–966. [DOI] [PubMed] [Google Scholar]
- 22. Kalichman SC, Amaral CM, White D, et al. Alcohol and adherence to antiretroviral medications: Interactive toxicity beliefs among people living with HIV. J Assoc Nurses AIDS Care 2012;23:511–520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Altice F, Mostashari F, Friedland G. Trust and the acceptance of and adherence to antiretroviral therapy. J Acquir Immune Defic Syndr 2001;28:47–58. [DOI] [PubMed] [Google Scholar]
- 24. Sankar A, Wunderlich T, Neufeld S, et al. Sero-positive African Americans' beliefs about alcohol and their Impact on anti-retroviral adherence. AIDS Behav 2007;11:195–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Andresen-Streichert H, Müller, A, Glahn, A, et al. Alcohol biomarkers in clinical and forensic contexts. Dtsch Arztebl Int 2018;115:309–315. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Bangsberg DR, Hecht FM, Charlebois ED, et al. Comparing objective measures of adherence to HIV antiretroviral therapy: Electronic medication monitors and unannounced pill counts. AIDS Behav 2001;5:275–281. [Google Scholar]
- 27. Kalichman SC, Amaral CM, Cherry C, et al. Monitoring antiretroviral adherence by unannounced pill counts conducted by telephone: Reliability and criterion-related validity. HIV Clin Trials 2008;9:298–308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Kalichman SC, Amaral CM, Stearns HL, et al. Adherence to antiretroviral therapy assessed by unannounced pill counts conducted by telephone. J Gen Intern Med 2007;22:1003–1006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Bangsberg DR, Deeks SG. Is average adherence to HIV antiretroviral therapy enough? J Gen Intern Med 2002;17:812–813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Parienti JJ, Barrail-Tran A, Duval X, et al. Adherence profiles and therapeutic responses of treatment-naive HIV-infected patients starting boosted atazanavir-based therapy in the ANRS 134-COPHAR 3 trial. Antimicrob Agents Chemother 2013;57:2265–2271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Molecular A. RealTime HIV-1 assay. Abbott Laboratories.. 2020. Available at: https://www.molecular.abbott/us/en/products/infectious-disease/realtime-hiv-1-viral-load (Last accessed April 23, 2020).
- 32. Tang N, Pahalawatta V, Frank A, et al. HIV-1 viral load measurement in venous blood and fingerprick blood using Abbott RealTime HIV-1 DBS assay. J Clin Virol 2017;92:56–61. [DOI] [PubMed] [Google Scholar]
- 33. Crepaz N, Tang T, Marks G, et al. Durable viral suppression and transmission risk potential among persons with diagnosed HIV infection: United States, 2012–2013. Clin Infect Dis 2016;63:936–983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Parienti JJ, Das-Douglas M, Massari V, et al. Not all missed doses are the same: Sustained NNRTI treatment interruptions predict HIV rebound at low-to-moderate adherence levels. PLoS One 2008;3:e2783. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Bangsberg D, Kroetz DL, Deeks SG. Adherence-resistance relationships to combination HIV antiretroviral therapy. Curr HIV/AIDS Rep 2007;4:65–72. [DOI] [PubMed] [Google Scholar]
- 36. Kalichman SC, Kalichman MO, Cherry C. Forget about forgetting: Structural barriers and severe non-adherence to antiretroviral therapy. AIDS Care 2017;29:418–422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Scott-Sheldon LA, Carey MP, Carey KB. Alcohol and risky sexual behavior among heavy drinking college students. AIDS Behav 2010;13:845–853. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Scott-Sheldon LA, Carey MP, Vanable PA, et al. Alcohol consumption, drug use, and condom use among STD clinic patients. J Stud Alcohol Drugs 2009;70:762–770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Satyanarayana S, Rogers BG, Bainter SA, et al. Longitudinal associations of syndemic conditions with antiretroviral therapy adherence and HIV viral suppression among HIV-infected patients in primary care. AIDS Patient Care STDS 2021;35:220–230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Braithwaite RS, McGinnis KA, Conigliaro J, et al. A temporal and dose-response association between alcohol consumption and medication adherence among veterans in care. Alcohol Clin Exp Res 2005;29:1190–1197. [DOI] [PubMed] [Google Scholar]