Preventing HIV Outbreaks among People who Inject Drugs in the United States: Plus ça Change, Plus ça Même Chose

Introduction

In 1997, a serious HIV outbreak among people who inject drugs (PWID) occurred in Vancouver, Canada, during which HIV incidence peaked at 18.6/100 person years (py) ^[1]. The outbreak was unusual since Vancouver had the largest volume syringe services programme (SSP) in North America and HIV prevalence among PWID had previously remained low. A contributing factor to the outbreak was an influx of powder cocaine that led PWID to inject more frequently than those who injected heroin ^[2] which outpaced the number of available syringes at the city’s single fixed-site SSP ^[3]. In response, health officials expanded mobile SSPs, HIV testing and medications for opioid use disorder (MOUD). They later were among the first in North America to implement population-level antiretroviral treatment (ART) as prevention for HIV infection^[4], and implemented the first supervised injection facility (SIF) and heroin maintenance programmes^{[3, 5]}. Consequently, Vancouver’s HIV incidence among PWID plummeted ^[3], and has remained <2 per 100 py since 2011 (personal communication, Dr Thomas Kerr, July 2020). Subsequent analyses showed that the combined effects of this integrated prevention strategy on HIV incidence were substantial ^[6], with harm reduction initiatives having the greatest impact ^[7].

The US response to the Vancouver HIV outbreak was polar opposite. The paper that first described the outbreak, which was published by the first author as an AIDS Fast Track article in 1997 ^[1], was entered into the US Congressional record with the opposite interpretation: that SSPs had failed. Data from Vancouver and Montreal were then used as a political weapon to uphold the Congressional ban on the use of US federal funds to support SSPs ^[8], which was not permanently overturned until December, 2015. While SIFs now exist in Canada, Western Europe and Australia and have consistently been shown to reduce the harms associated with injection drug use ^[9], only a few ‘underground’ SIFs exist in the US and all are unsanctioned by the federal government ^[10].

The politicization of harm reduction and the War on Drugs dogma that has prevailed in the US for more than twenty years still undermines HIV prevention among PWID to this day, with repeatedly disastrous consequences, such as HIV outbreaks. The most common concerns from policymakers and community members are that SSPs and SIFs increase drug use and crime, and encourage young people to initiate injection drug use. A large body of peer-reviewed literature examining potential intended and unintended consequences of these programs has found no negative societal effects associated with either SSPs ^[11–13] or SIFs ^{[14, 15]}.

This editorial review covers current trends in the epidemiology of HIV among PWID in the US and describes four HIV outbreaks among PWID that have occurred since 2015. All were characterized by a high proportion of women who inject drugs and underlying socioeconomic drivers such as homelessness and poverty. We also discuss gaps in the prevention and treatment cascades for HIV and MOUD among PWID and propose lessons learned that may be helpful to prevent future HIV outbreaks.

The US Opioid Crisis and the Changing Epidemiology of HIV among PWID

Over the last decade, North America has been in the throes of a major opioid epidemic, due initially in part to over-prescribing of prescription opiates, followed by increasing availability of cheap (black tar) heroin, synthetic opioids (predominantly fentanyl), and stimulants, including methamphetamine ^[16]. For the last three years, annual overdose deaths in the US surpassed annual deaths due to HIV/AIDS at its peak.

While HIV risks associated with injection of heroin and methamphetamine are well documented, fentanyl injection carries additional risks. First, fentanyl has a shorter half-life than heroin, and is typically injected more frequently, placing PWID at greater risk of syringe sharing ^[17]. Second, when available as a powder, PWID can prepare fentanyl for injection without heat, which some researchers speculate can inactivate HIV and HCV^[18,19].

Changes in drug trafficking patterns and use reflect the changing epidemiologic profile of HIV infection among PWID in the US, which historically affected communities who were older, urban and Black ^{[20, 21]}. More recently, the majority of US HIV infections among PWID have been reported among persons who are younger, rural or suburban and Caucasian ^[22]. Although the overall proportion of reported HIV cases among PWID in the US declined from 2010 to 2016, that decline has stalled, especially among Caucasians ^[22]. The US Medical Monitoring Project also reported disturbing increases in distributive needle sharing and unprotected sex among HIV-positive PWID ^[23].

Women Who Inject Drugs and HIV Risk

In 2017, women comprised 28% of newly reported HIV infections among PWID in the US ^[24]. A meta-analysis of 117 studies across 14 countries suggest a modest increase (18%) in HIV risk among women who inject drugs compared to men, but only in high prevalence (>20%) settings ^[25]. Compared to men, women who inject drugs have greater overlap between their sexual and injection social networks ^[26] and are often initiated into drug use by male partners who exert significant control over their injection and sexual practices ^{[27, 28]}. Gender inequality and power imbalances often relegate women to be ‘second to the needle,’ and undermine condom negotiation. This not only increases HIV risk, but can place women at risk for physical and sexual violence ^[29]. Men who initiate women into drug use are more likely to have been incarcerated, share used syringes, inject others, and obtain syringes from informal sources, all known risk factors for HIV ^[26]. A prospective cohort study in Baltimore found HIV incidence was more than double among women injectors who had a male sexual partner that also injected drugs ^[30]. Women who engage in sex work and use drugs also experience heightened HIV risk due to shared injection equipment, condomless sex with clients and intimate partners, exposure to sexual violence, and incarceration ^{[28, 29]}.

HIV Surveillance among PWID and Recent Outbreaks

Among 11,437 PWID surveyed in 23 US cities for the National HIV Behavioral Surveillance study in 2018, overall HIV prevalence was 6% ^[31]. However, only 55% of PWID had met CDC guidelines for annual HIV testing and nearly one-third reported using a syringe after someone else had previously used it, which was more common among young PWID ^[31]. More than half of PWID reported obtaining syringes from a SSP; but variability between cities ranged from 1% to 93% ^[31]. This disparity is largely due to uneven access to SSPs in many US communities, some of which have criminalized syringe possession without a prescription or over the counter syringe sales. Barriers to sterile syringe access for PWIDs persist despite strong evidence that SSPs can significantly reduce HIV incidence ^[32]. Not surprisingly, lack of access to sterile syringes has been consistently associated with HIV outbreaks among PWID, especially in rural or semi-urban communities, as described below.

Scott County, Indiana.

After the opioid analgesic Opana® ER (an extended release oxymorphone) was approved by the FDA in 2006, it began to be liberally prescribed by medical providers, including rural regions like Southern Indiana. When the manufacturer switched to a more tamper-resistant formulation in 2012 and policy reform began limiting its prescription, many patients who had developed an opioid use disorder switched to injecting it. Qualitative research conducted in Scott County suggested that its use was associated with multiple injections per injection episode ^[33]. With their supply drying up, many PWID, including high proportions of young women, turned to injecting black-tar heroin which was more available and less expensive ^[34]. Since both SSPs and pharmacy syringe sales without a prescription were illegal in Indiana, multi-person syringe sharing was common. A moratorium on methadone expansion in the state also meant there were few medical providers offering MOUD. Several free HIV testing locations had been closed due to moral concerns that they also offered abortions. Despite reports of escalating HCV incidence in the region ^[35], few efforts were made to scale up HIV prevention ^[36].

In 2015, an astute physician in Scott County observed a small but unusual cluster of new HIV diagnoses within a short period of time and reported it to the local health department, who then notified the CDC, which began an outbreak investigation. By that time, HIV had already spread through several PWID networks and their sexual contacts in Scott County ^[34] resulting in 203 HIV infections, and a community-level HIV prevalence of 5%. Implementation of SSPs subsequently led to a decrease in syringe sharing ^[37], but a modeling analysis concluded that if SSPs and MOUD had been implemented earlier, 90% of HIV infections in the county could have been prevented ^[37]. A recent modeling analysis showed that a more proactive response to implementing SSPs could have blunted the HIV outbreak, not just among PWID, but among their sexual partners ^[37].

Due to growing concerns that similar outbreaks could be brewing, Van Handel and colleagues used surveillance data to conduct an ecological analysis examining characteristics associated with the Scott County outbreak, which they used to identify 220 US counties in 26 states that were deemed vulnerable to future outbreaks ^[38]. Of these counties, half are located in the Appalachian region, which is predominantly rural, socioeconomically disadvantaged, and, like Scott County, lacking in robust harm reduction services.

West Virginia.

West Virginia (WV) is the only US state that is located entirely in Appalachia. Since 2013, WV ranks among the highest in terms of national rates of acute hepatitis B and C ^[39] and has seen a rise in endocarditis cases related to injection drug use ^{[40, 41]}. In 2018, WV had the highest age-adjusted drug overdose death rate in the US (51.5 per 100,000) ^[42]. Unfortunately, these well-known harbingers of HIV outbreaks went unheeded. In Huntington city and surrounding Cabell County in southwestern WV, 82 HIV new cases were identified since 2018, where the region previously averaged two new HIV diagnoses annually. Of these 82 cases, 92% were among PWID, 92% were Caucasian, 40% were women, 29% exchanged sex for money or drugs, and 88% tested HCV-positive ^[43]. Of 50 individuals with available HIV genetic sequencing data, 92% had closely related infections suggestive of rapid transmission ^[44].

Although a SSP was introduced in Cabell County in 2015, strict eligibility requirements limited its access. Effective SSPs are operating in Huntington, Morgantown, and Harpers Ferry/Martinsburg, but the SSP in Charleston (WV’s capital located in Kanawha county) that served thousands of clients was closed by the city government in 2018. Subsequently, 24 cases of HIV among PWID in Kanawha County were reported ^[45]. In 2019, a bill was introduced in the WV state legislature to abolish SSPs. Although it did not pass, ongoing opposition to SSPs in the state continues.

Massachusetts.

From 2015 to 2018, two HIV outbreaks occurred in Lawrence and Lowell, both rural communities in northeast Massachusetts ^[44]. Like the Scott County outbreak, the Lawrence outbreak was first identified by an astute physician. Similar to the aforementioned outbreaks, 43% of new HIV diagnoses in Lawrence and Lowell were among women. Sexual risk behavior included exchanging sex for money or drugs among women and men ^{[45, 46]}. However, the Massachusetts outbreaks were closely associated with frequent fentanyl injection. Genetic sequencing revealed several HIV clusters reflecting multiple introductions of HIV ^[46] as opposed to a point source. Molecular links at <0.5% genetic distance also confirmed that most infections were recent. SSPs were introduced in both counties but not until the outbreaks were well underway ^[46].

Seattle, King County, Washington.

In the Northwestern US, Seattle is a mid-size city situated in King County, which was the first in the country to adopt a SSP and a PrEP program. The region has had an impressive history of ensuring that high proportions of HIV-positive people were engaged in care and receiving ART, and was not identified as one of the counties that was vulnerable to an HIV outbreak associated with injection drug use ^[38]. Despite this, and a 46% decrease in the rate of diagnosed HIV cases among PWID in King County from 2009 to 2017 ^[47], an HIV outbreak occurred in 2018 among PWID. The first case was an adult male diagnosed in the emergency department who did not report injecting drugs, but had condomless sex with a female sex worker. An epidemiologic investigation identified another 13 new HIV cases the same year that were linked via molecular sequencing. Of these 14 new HIV diagnoses, 12 were PWID and 11 were women, of whom 9 exchanged sex for money or drugs. All lived within a three mile radius that had limited access to syringes, all were homeless, and ten reported injecting heroin and methamphetamine ^[48].

The subsequent investigation revealed that the outbreak was part of a larger cluster of 23 HIV-infected persons, some of whom had been diagnosed with HIV years earlier and yet were not virally suppressed ^[48]. In response, the public health department issued an alert, expanded HIV testing, SSP and MOUD and began offering mobile services targeted to women. Given that only 1% of PWID in the region were receiving pre-exposure prophylaxis (PrEP) in 2018 ^[47], efforts to expand uptake of PrEP were also undertaken at STD clinics and mobile SSP units.

This outbreak illustrates that even in a region where HIV is relatively well controlled and a high volume SSP exists, a small number of virally unsuppressed people can spark an HIV outbreak, particularly in locations with limited SSP coverage. It also shows the importance of monitoring changes in retail drug markets that can lead PWID to inject more frequently, and to intervene upon structural determinants such as homelessness and economic hardship that increase vulnerability to HIV infection and impede access to HIV prevention and care.

Continuum of Care Considerations: HIV Treatment, PrEP, and MOUD

As these four outbreaks demonstrate, a critical component of the public health response is to ensure that exposed contacts are rapidly tested and ART is offered immediately to anyone testing HIV-positive, which reduces morbidity and mortality and curtails onward transmission once viral suppression (VS) is achieved. The US Federal government has set a bold agenda to end the HIV epidemic by reducing new HIV infections by 75% within five years and by 90% within ten years ^[49]. To succeed, early diagnosis, uptake of ART and sustained adherence need to be vastly improved, especially for PWID, who are underserved at every point in the HIV treatment cascade ^[50]. For example, by the end of Massachusetts’ outbreaks in September 2018, only 63% of HIV-positive PWID had achieved VS and 12% had not had a viral load test in the prior year, despite availability of state-supported health insurance ^[46]. Qualitative interviews with HIV-positive PWID revealed that many experienced frequent homelessness and incarceration, which are destabilizing factors that undermine ART adherence ^[46].

In 2013, PrEP based on formulations of tenofovir and emtricitabine were shown effective in reducing HIV transmission among PWID ^[51]. The CDC recommends PrEP for PWID who are HIV-negative but have an HIV-positive or injecting partner of unknown serostatus and/or engage in sharing injection equipment ^[52]. Nationally, nearly one in five PWID are indicated for PrEP, ^[53] but PrEP knowledge, access and uptake among PWID remains abysmally low ^{[31, 54–58]}, particularly in rural areas ^[59]. Use of PrEP among PWID in the US ranges from 0% to 5% in various studies and settings^{[31, 54, 56, 58]}, despite high levels of interest and willingness ^[54–58]. Gender-based differences among PWID and PrEP knowledge and utilization require urgent study, as emerging research suggests that integrating PrEP into SSPs may facilitate PrEP uptake ^[60]. Other reported barriers include low self-perceived HIV risk, concern about side effects, competing health priorities, and HIV stigma ^{[60, 61]}.

A national survey conducted in 2013–14 among HIV care providers revealed that only 1% had prescribed PrEP to PWID ^[62], perhaps due to concerns about adherence. However, a recent study of PWID revealed high HCV treatment adherence and viral clearance despite ongoing injection drug use ^[63], indicating that adherence barriers can be overcome. In 2015, nearly three-quarters of primary care physicians reported high interest in prescribing PrEP to PWID ^[64], offering hope that provision of PrEP to PWID could be expanded. Most recently, a long-acting PrEP regimen based on injectable cabotegravir was demonstrated to reduce HIV incidence in a randomized clinical trial of cisgender men and transgender women ^[65]. Long-acting injectable PrEP could provide promote uptake and adherence to PrEP among PWIDs, provided that its benefits are not offset by the potential psychological effect of receiving regular PrEP injections that could precipitate relapse or more frequent injection drug use.

Importance of Integrated OUD/SUD treatment with HIV prevention and treatment

Accurately screening for OUD, as well as other substance use disorders (SUDs), offering PrEP for those who are HIV-negative, and ART to those who are HIV-positive is essential for reducing ongoing HIV transmission. Substance use can interfere with ART adherence ^[66], which may impede VS, and increase risk of HIV transmission. Ongoing substance use can also increase the risk of acquiring HIV through condomless sex and contaminated injection equipment and may reduce PrEP adherence. Therefore, diagnosis and treatment of OUD/SUD is essential for optimizing HIV prevention and treatment.

MOUD is recognized as the most effective OUD treatment ^[67], and includes methadone, buprenorphine and extended-release naltrexone (XR-NTX). Numerous trials demonstrate that MOUD reduces opioid use, overdose, death, HIV and HCV transmission, and improved HIV VS and psychological well-being^[68–74]. Recent research found that when MOUD was offered to PLH with OUD and alcohol use disorders upon release or prior to release from prison or jail, there was an increased likelihood of achieving and maintaining VS six months post-release^{[70, 75, 76]}. Unfortunately, like the HIV treatment cascade, there are substantial gaps in the OUD treatment cascade that begins with lack of OUD screening. Among those with an OUD diagnosis in the US, <20% are initiated on MOUD and <30% are retained on MOUD six months after initiation ^[77].

To close the gaps for MOUD and HIV treatment in the US, the National Academy of Sciences, Engineering and Medicine (NASEM) recently convened a committee to evaluate integrated opioid and prevention services ^[78]. Barriers to integration of MOUD and HIV services included restrictive buprenorphine prescribing policies; burdensome prior authorization policies; lack of a motivated workforce; stigma and lack of expansion of MOUD, PrEP and other integrated services during and post-incarceration.

Yet integrated care for OUD and HIV is possible, and when offered, can improve VS among PWID ^{[70, 74, 75]}. In particular, the HIV Prevention Trials Network 074 trial demonstrated that integrating HIV and MOUD among PWID living with HIV who were randomized to receive an integrated intervention of behavioral support for SUD and ART resulted in greater self-reported ART adherence, VS, and MOUD uptake compared to standard of care ^[79]. Future research is needed to evaluate integrated MOUD with PrEP to reduce new HIV infections.

Considerations regarding Stimulant Use

Although the US overdose crisis has been largely attributed to opioid use, recent surveillance data from twelve US cities showed a resurgence in methamphetamine and cocaine use ^[80]. A recent CDC report estimated that 1.6 million adults reported using methamphetamine in the past year, 22% injected it, and over half had a methamphetamine use disorder ^[81]. In some US cities (e.g., Atlanta and New York), urban Black populations that had primarily used heroin are increasingly shifting to methamphetamine and cocaine use in combination with heroin and fentanyl. Since methamphetamine and cocaine use are associated with high risk sexual behaviors ^{[82, 83]}, there is concern that HIV incidence could surge in cities with high background HIV prevalence.

Treatment of methamphetamine use disorder has been hampered by the lack of an established pharmacologic treatment ^[84]. However, mirtazapine, a mixed monoamine agonist-antagonist that has been used to treat depression, has been hypothesized to reduce craving and withdrawal symptoms associated with methamphetamine use. Following an earlier trial that showed promise ^[85], a recent clinical trial of MSM ^[86] found that oral mirtazapine use was associated with greater number of weeks of abstinence, as well as decreases in number of male sexual partners and condomless sex. Efforts are needed to expand its use among PWID who have methamphetamine use disorders, and to explore whether mirtazapine alone or in combination with PrEP reduces high risk sexual behaviors associated with ‘chemsex’ (i.e., sexualized drug use involving substances such as methamphetamine, ecstasy, GHB and synthetic cathionones).

While there are no current effective FDA approved medication treatments for cocaine use disorder, psychostimulants, modafinil, bupropion, topiramate and disulfiram have demonstrated some modest benefits ^[87]. At present, contingency management has been found have the most benefit for cocaine and methamphetamine use disorders ^[88]; however, the benefit markedly diminishes once contingency management is ceased. Additional research is clearly needed to expand evaluation of more effective medications for stimulant use disorders.

Lessons Learned

The US experience with ongoing HIV outbreaks among PWID offers several lessons. First, harm reduction services that offer SSPs and MOUD need to be widely implemented before outbreaks occur, which means that policymakers must learn to accept their public health imperative even when they object on moral grounds. Surveillance that relies on HIV testing alone may miss outbreaks until it is too late to intervene.

In every setting that experiences rising incidence of acute HBV, HCV and endocarditis, SSPs and over-the-counter syringe sales should be available and expanded. If patterns of dispersed hotspots become more common among rural and suburban communities, surveillance programs will need to be re-structured. Since very few studies of HIV and HCV service provision exist in US suburban and rural settings ^[89], behavioral surveillance should be expanded in these regions. Molecular epidemiology with new tools such as Nanopore MinION® can be implemented in real time to inform resource allocation during active outbreaks. Novel approaches such as monitoring drug metabolites in wastewater could also inform surveillance of drug use behaviors and corresponding gaps in MOUD ^[90].

When surveillance data uncover subgroups that are over-represented among PWID and PLH, public health actors need to spring into action, tailoring the response to local needs. For example, in the US and elsewhere, women who inject drugs experience overlapping risk factors related to sex work, incarceration, gender-based and intimate-partner violence, all of which leads to risky injection and sexual practices, and ultimately increased vulnerability to HIV. To address these unique, gendered vulnerabilities, a multi-level, multi-tiered approach that integrates biomedical, structural and social prevention efforts is needed that considers gender-based violence, pregnancy, mental health, childcare, and post-release services ^{[91, 92]}. We refer interested readers to other published papers that have assessed these issues in more depth ^{[93, 94]}.

The US experience has also uncovered ‘deaths of despair’ ^[95] that reflect syndemics with common underlying drivers: HIV, viral hepatitis, STIs, SUD and more recently, COVID-19. If the root causes of poverty, discrimination and stigma can be intervened upon, the end result could reduce the burden of HIV as well as other diseases that share similar root causes ^{[96, 97]}. Recently, validated scales that measure anticipated, enacted, and internalized stigma for people with SUD show promise for identifying intervention targets to reduce stigma related to substance use and MOUD and increase MOUD uptake ^[98].

The US experience also shows that providers, policymakers and even researchers have overlooked PWID and PrEP in the HIV treatment cascade. Based on the extremely low levels of awareness and coverage of PrEP among PWID and recent findings that long-acting cabotegravir is an efficacious PrEP modality ^[65], there is an urgent need for increased education about PrEP as an HIV prevention intervention in regions that are highly vulnerable for HIV outbreaks. However, it will be important to ensure that PrEP is offered as one component of the HIV prevention response and not in lieu of harm reduction services.

Without improving the continuum of care for MOUD, OUD and its consequences will continue to weigh heavily on society and health care systems. NASEM’s recommendations to overcome these barriers ^[78] were recently summarized by co-author Springer and colleagues in a recent JAMA viewpoint ^[99] where it was recommended that structural barriers to MOUD and harm reduction programs be removed; funding be allocated to address the needs of low-income uninsured or under-insured individuals with OUD and HIV; workforce training on integrating OUD and HIV services should be improved; and timely access to health insurance, MOUD and PrEP be offered in criminal justice settings and upon release.

The US response to the HIV epidemic among PWID has been fractured. A crucial lesson for vulnerable communities is that when evidence-based responses to HIV prevention are undermined or abandoned because of moral objections, untold humanitarian and financial costs on public health will ensue. Restructuring a path forward requires that evidence-based interventions be integrated and brought to scale while simultaneously addressing underlying structural drivers. Failing to do so will mean that HIV outbreaks among PWID and the communities they live in will continue to occur in a tragic and relentless cycle.