Substance use and Adherence to Antiretroviral Therapy: What is known, and what is unknown

Abstract

Purpose of review:

People who use drugs face multiple challenges to achieve optimal HIV treatment outcomes. This review discusses the current knowledge in substance use and antiretroviral therapy adherence, highlighting recent findings and potential interventions.

Recent findings:

Studies continue to demonstrate the negative impacts of substance use and related disorders on ART adherence, with the exception of cannabis. Evidence-based addiction treatment, in particular opioid agonist therapy appears to improve adherence levels. Most individual-level adherence specific interventions did not provide sustained effects, and no studies evaluating structural-level interventions were found.

Summary:

Findings suggest the urgent need to scale-up opioid agonist therapy, as well as to simultaneously address multiple structural barriers to care to optimize HIV treatment outcomes among people who use drugs.

Introduction

Antiretroviral therapy (ART) has dramatically modified the natural history of HIV infection. Timely initiation of ART is associated with reduced mortality, morbidity and transmission both at the individual and community level [1–5]. This scientific consensus had led to calls to scale-up access to HIV care and treatment worldwide as a central component of efforts control the HIV pandemic. However, challenges remain as early diagnosis, linkage to and retention in care, and long-term adherence to ART are critical for attaining these optimal clinical and public health outcomes [6, 7].

While impressive gains have made in terms of scaling-up access to ART in recent years, only 38% of the estimated 36.7 million people living with HIV (PLHIV) globally were virally suppressed in 2015. A key determinant of sustained viral suppression is consistent high levels of adherence to ART. Suboptimal adherence is a risk factor for virologic rebound, and subsequent increased likelihood of virologic failure, emergence of viral drug resistance, disease progression and onward viral transmission. Unfortunately, a substantial proportion of PLHIV do not achieve optimal levels of adherence [8, 9].

A number of reviews have examined barriers and facilitators to ART adherence across different populations and settings, and among them active substance use has been identified as one of the most common predictors of poor adherence [10–14]. Given the high prevalence of substance use and related disorders among PLHIV, as well as recent explosive outbreaks of HIV driven by parenteral transmission [15], understanding patterns of ART adherence among people who use drugs (PWUD), and ways to promote adequate and sustained levels of adherence will be critical to control the HIV pandemic.

This review focuses on studies published between 2012 and 2017, summarizing and discussing the current knowledge on the epidemiology of ART adherence among PWUD, potential interventions to improve adherence among this population, as well as current gaps and future directions for research.

Recent trends in ART adherence among PWUD

Recent reviews continue to highlight active substance use as one of the major predictors of poor ART adherence among PLHIV [11, 13]. Sub-optimal levels of ART adherence have been observed among people who use injection, non-injection drugs or alcohol in low-, middle-, and high-income settings [16–18]. For example, a meta-analysis that included 15 studies of people with a history of injection drug use from seven low- and middle-income countries found a mean weighted adherence of 72% [18]. A long-term cohort study of HIV-positive PWUD in Canada showed substantial increase over time in the proportion of individuals with ≥95% adherence; however, in 2014, still only over half achieved this threshold [19]. In an analysis of the same cohort, periods of active injection drug use were associated with transitions out of optimal adherence and as barriers to becoming optimally adherent [20]. Another study conducted among 3,343 PLHIV in care across six countries from Latin America in 2012–2013 found that between 29% and 46% of people who used alcohol or illicit drugs reported missing ART doses in the prior week, compared to only 11% of individuals who did not report substance use [17].

Specific substances and ART adherence

Alcohol and problematic alcohol use are common among PLHIV, with some studies documenting up to 42% prevalence of heavy drinking [21]. A large body of research has documented negative impacts of alcohol use on HIV outcomes. A recent review examining the impacts of alcohol on the HIV cascade of care continuum published between 2010 and 2015 found an association between alcohol use and nonadherence, usually in a dose-response fashion, in most (n=16) of the 33 included studies focusing on the ART adherence step [22]. Other recent studies confirm this negative influence of alcohol on ART adherence in a variety of contexts [23, 17, 24, 25].

Recent research conducted in North America [24, 26–29] also confirms the links between heroin and cocaine use and suboptimal ART adherence found in previous reviews [14]. For example, a large study that pooled data from over 1,500 PLHIV across different sites in the United States, found independent and strong negative associations between recent heroin or cocaine use and objective measures of ART adherence (i.e., using electronic data monitoring) [27].

Of note, in many of these same studies [24, 26, 27, 25], cannabis use was not a risk for suboptimal ART adherence. This lack of association between cannabis use and ART adherence was also documented in other recent analyses [30–32]. Interestingly, in one of these studies, individuals with cannabis dependence had significantly lower rates of adherence than both non-dependent users and not users, suggesting that patterns of use may be an important factor in shaping adherence outcomes in this population [32]. In addition, these opposing findings between cannabis and other substances may relate to the fact that many PLHIV use cannabis therapeutically to address disease symptoms and medication side-effects which in turn may enhance adherence [33]. Alternatively, cannabis may be used as a substitute for other substances (e.g., alcohol, crack cocaine, opioid analgesics) with known detrimental effects on adherence [34, 35].

The literature around methamphetamine use and ART adherence is more limited, has mostly focused on men who have sex with men, and documented high rates of nonadherence among methamphetamine users [36]. Recent research continues to support this negative association [25, 37, 38]. Indeed, in one of these studies, ART non-adherence was more likely to occur in days where methamphetamine was also used [38].

Despite the high prevalence of prescription medication use (other than ART), including non-medical use, among PLHIV, we found only one study investigating the relationship between opioid analgesics and ART adherence. Among 258 marginalized PLHIV, opioid analgesic misuse, but not use, was associated with incomplete ART adherence, highlighting again that patterns of use and misuse, and associated contextual factors, rather than specific substances may be the critical factor shaping adherence behaviours [39].

Addiction treatment

Although PWUD are a heterogeneous population, accumulating evidence supports the role of evidence-based addiction treatment in improving ART outcomes among individuals with substance use disorders. To date, the largest body of evidence is for opioid agonist treatment ([OAT] e.g., methadone, buprenorphine). A recent systematic review found that engagement in OAT was associated with a two-fold increase in the likelihood of achieving optimal ART adherence [40]. Other recent studies echo this finding of OAT as a key facilitator of ART adherence [29, 41, 26, 20, 42]. Interestingly, in one study there was a positive dose-response relationship between higher daily methadone dose and the likelihood of achieving ≥95% ART adherence, underlining the need for evidence-based dosing of OAT [42]. Regarding other addiction treatment modalities, the evidence is mixed. One study showed advantages of any treatment over no treatment [43] and other no effect [14]. These differences may relate to the fact that with the exception of opioid and alcohol use disorder for which effective pharmacotherapies exist, there are currently not such medications for other substance use disorders, and the evidence for other approaches is limited. Of note, no studies evaluating the impact of pharmacological therapies for alcohol use disorders on ART adherence were identified [22]. Collectively, these findings highlight the need to expand access to OAT to optimize ART outcomes among PLHIV with opioid use disorder, as well as for further research for evidence-based treatment for stimulant use disorder or other strategies to support this population in fully benefitting from ART.

Social-structural exposures

Despite the growing acknowledgement on the importance of social-structural factors as drivers of health outcomes, particularly among marginalized populations, comparatively less research has examined the influence of structural vulnerability on ART adherence. The few existing studies, mostly for the North American setting suggest that unstable housing (e.g., homelessness, residential eviction), incarceration and prohibited income generation activities are important structural barriers to optimal ART outcomes, and mostly mediated by lower levels of adherence [44–47].

Interventions to improve adherence among PWUD

A previous systematic review comprehensively reviewed the evidence of interventions to improve ART adherence among PWUD [48]. Among the fifteen randomized clinical trials (RCTs) included in this review, the strongest evidence was for directly administered antiretroviral therapy (DAART). RCTs evaluating contingency management and multi-component, nurse delivered interventions also demonstrated improvements in short-term ART adherence. However, these benefits waned rapidly after the interventions were discontinued [48]. Given the need for lifelong ART adherence, these results suggest the need to further explore how to sustain the effects of these interventions.

Since the publication of this review, a number of other RCTs evaluating adherence interventions for PWUD were published. Lucas et al. evaluated the efficacy of a DAART intervention embedded within an OAT program [49]. In this study, rates of ART adherence and viral suppression at 12 months were overall low, with no differences between participants in the DAART and self-administered arm, contradicting previous RCTs in this area. However, most previous studies assessed shorter interventions periods and the sustained benefits after intervention cessation were equivocal [48], raising the question about the real-world effectiveness of this intervention.

Although no recent studies evaluating contingency management to promote ART adherence among PWUD were found, we identified two studies that targeted viral suppression. The first one by Farber et al —a pilot study that used a within-subject design, found a 12% (from 57% to 69%) increase in the proportion of undetectable viral loads after the introduction of cash incentives ($100) rewarding reductions in viral load (or maintaining viral suppression) [50]. In the second study, 120 ART-naïve HIV-positive PWUD were randomized to receive or not voucher incentives for achieving specific steps along the continuum of HIV care [51]. Although the voucher incentive arm was associated with improved ART initiation and monthly medication refill rates compared to the control arm, no differences were observed in viral suppression. The opposing findings in these studies may relate to differences in study design, setting (U.S. vs. India) populations (e.g., only half of the study sample in the Farber et al. study were PWUD), different types and size of incentive or other unmeasured factors. Both studies were also limited by small sample sizes.

More recently, Project HOPE went one step further and evaluated a combination approach of patient navigation with or without financial incentives to promote engagement in HIV and addiction care [52]. In this large RCT that was conducted in 11 sites across the U.S., 801 inpatient HIV-positive PWUD with detectable viral load were randomized to one of three arms: 1) usual care, 2) six months of patient navigation, or 3) six months of patient navigation plus financial incentives targeting behaviors aimed at reducing substance use and improving engagement in HIV care. At 12 months, no differences were found in viral suppression, death, or other key secondary outcomes, including ART adherence, substance use or engagement in addiction care. Possible explanations for these results may include the short duration of the intervention, the lack of addiction care options in many settings, the high proportion of participants with stimulant use disorders (70%) for which no effective treatment exists, and competing interests that were not addressed in the context of a highly marginalized population. Collectively, findings from these three studies do not support the use of financial incentives (alone) to improve progression along the HIV continuum of care among PWUD.

Finally, two pilot studies evaluated mobile health (mHealth) interventions to improve ART adherence among PWUD. The iTAB study examined the effect of personalized ART reminder text messages (in addition to daily short message service [SMS] assessing methamphetamine use) among HIV-positive methamphetamine users in San Diego, U.S. [53]. A 30-day evaluation showed preliminary evidence of the acceptability and feasibility of this intervention among this populations, with 70% response-rate to adherence reminders, 80% self-reported adherence, and high overall satisfaction with the program. The TxText intervention involved automated bidirectional daily SMS assessing mood, substance use and ART adherence for HIV-positive ART non-adherent PWUD, and subsequent personalized intervention SMS based on participant’ responses [54]. A three-month evaluation of this intervention showed consistent engagement with the tool over the study period, with responses rates between 64% and 69% to each of the categories. While preliminary feasibility and acceptability evaluations of these mHealth interventions are promising, full results of these studies are yet not available, and thus the efficacy of mHealth technology on promoting and sustaining ART adherence among PWUD remains unknown.

Conclusions

The review of current literature continues to demonstrate the negative impacts of substance use and related disorders on ART adherence, particularly when left untreated. The use of cannabis appears to be an exemption to this detrimental association, potentially due to some of its suggested therapeutic uses, though evidence in this area is limited and inconsistent, and deserves further research. Another avenue of future research relates to what level of adherence may be sufficient to achieve optimal treatment outcomes among PWUD with newer ART formulations. Indeed, while recent research has suggested that improved ART regimens with higher genetic barrier may allow for lower levels of consistent adherence than those historically required (≥95%) to achieve viral suppression, the evidence for PWUD is scant and inconclusive [55, 56]. In addition, emerging studies also suggest that these sub-optimal levels of adherence may result in higher levels of immune activation and inflammation and increased non-AIDS related morbidity [57]. Our review also points to the lack of long-term benefits of any of the individual-level adherence interventions, and the absence of studies evaluating structural-level interventions. Although there is continued need for the development and testing novel adherence strategies tailored to PWUD, these findings also highlight the urgent need to implement and scale-up interventions with proven individual and community-level benefits, particularly opioid agonist therapy. Similarly, in the context of a marginalized population with multiple structural barriers to care, findings from this review also underscore the need to simultaneously address to achieve optimal treatment outcomes and reduced health inequities faced by PWUD.