HIV risk behaviors among community-recruited older people who inject drugs (PWID) prescreening for HPTN 094 in Washington, DC

Abstract

Older people who inject drugs (PWID) remain at risk for HIV, yet little is known about HIV risk behavior in this population. We investigated the prevalence and correlates of HIV risk behaviors among older PWID in Washington, DC. We conducted a cross-sectional analysis of self-reported survey data collected between June 2022 and June 2023 from adults recruited through community-based outreach at known drug use hotspots. Respondents who reported injection drug use in the past three months and were aged >50 were included. We used simple and multivariable logistic regression to determine correlates of HIV risk behaviors in the last three months. Among 260 older PWID, 42% reported at least one HIV risk behavior in the past three months. Benzodiazepine injection was associated with condomless sex (OR=5.61, 95% CI: 1.36–23.06). Reporting an overdose in the last three months (aOR=3.18, 95% CI: 1.08–9.36) and other opioid injection (aOR=4.61, 95% CI: 1.04–20.40) were associated with sharing syringes/needles. Experiencing homelessness (aOR=2.20, 95% CI: 1.17–4.13), reporting an overdose (aOR=2.57, 95% CI: 1.18–5.61), and benzodiazepine injection (aOR=6.72, 95% CI: 1.49–30.24) were associated with sharing other injection equipment. Our findings suggest that older PWID in Washington, DC engage in HIV risk behaviors, particularly those experiencing homelessness, recent overdose, and injecting certain substances. Age-specific substance use treatment and HIV prevention programs are needed to address HIV risk in this population.

Introduction

The opioid overdose epidemic continues to cause significant morbidity and mortality in the United States (US). Between 1999 and 2022, more than 700,000 people died from an opioid-related overdose (1). The rate of overdose deaths in older adults is increasing, with the highest fatality rates in older Black men (2). In addition to overdose deaths, the opioid epidemic has contributed to continued transmission of infectious diseases such as HIV (3) and has been identified as a threat to progress in HIV prevention (4–6). New HIV diagnoses among people who inject drugs (PWID) increased by 5% between 2018 and 2022, with PWID accounting for a total of 7% of new HIV diagnoses in 2022 (4). HIV transmission among PWID can occur through sharing syringes or other injection equipment (7). Fentanyl injection may increase HIV risk compared to heroin injection, as fentanyl is associated with more frequent injection and increased syringe sharing (8,9). People may also be more likely to engage in HIV sexual risk behaviors when using drugs, including condomless sex and exchanging sex for money or drugs (10).

As a result of demographic trends and public health advances, a large segment of the PWID population in the US is aging (11,12). Older adults are a unique population in the context of both the opioid and the HIV epidemics. They may experience lifelong chronic opioid use (13,14) and barriers to treatment such as stigma (15,16), provider bias (15,17), masking by geriatric or comorbid conditions (15–17), and low perceived treatment need (18). Older PWID are also often overlooked in HIV prevention programs and research despite their experience of elevated risk (19). People aged 55 and older accounted for 15% of new HIV diagnoses among PWID in 2022 (4).

Washington, DC is a hotspot for opioid overdose deaths in the US (20), and older adults are highly impacted. Of opioid-related fatalities in Washington, DC in 2023, almost half occurred among those aged 50–69 (49%) (21). According to estimates from the National HIV Behavioral Surveillance (NHBS) study, in 2022, 85% of PWID in Washington, DC were aged 50 and older (22), almost two-times the average across 20 US cities (23).

Adults aged 50 or older accounted for 19%, and PWID overall accounted for 5% of newly diagnosed HIV cases in Washington, DC from 2018 to 2022 (22). However, research on HIV risk behaviors among older PWID is limited (24). We aimed to investigate the prevalence and correlates of HIV risk behaviors among older PWID (aged >50) in a recent community-based convenience sample in Washington, DC.

Methods

Data collection

We used cross-sectional data from the George Washington University HIV Prevention Trials Network (HPTN) Clinical Research Site’s (CRS) incentivized prescreening survey collected from June 21, 2022, through June 30, 2023. This brief survey was developed to identify individuals who were potentially eligible for HPTN 094, or the INTEGRA study, in Washington, DC. HPTN 094 was a randomized controlled trial (RCT) testing the use of a mobile unit and peer navigation to provide integrated care to PWID compared to peer navigation alone and is described elsewhere (25). The prescreening survey was implemented using convenience sampling during street outreach and recruitment events at drug use “hotspot” locations that were identified using local opioid overdose data. Individuals aged 18 and over were eligible to take the survey and provided informed consent. Towards the end of recruitment, to increase screening efficiency for the parent trial, we restricted the prescreening survey to individuals who reported recent injection behavior. Study data were collected and managed using REDCap electronic data capture tools (26) hosted at the Clinical and Translational Science Institute at Children’s National. Participants completed the survey as a computer-assisted self-interview on a tablet. If a participant needed assistance completing the survey, trained staff members read the questions aloud in a private area and recorded the participant’s responses. The survey contained questions about drug use and sexual behaviors as well as experience with medications for opioid use disorder (MOUD). All data were self-reported. The survey took around five minutes to complete, and respondents received a $10 gift card for their time. All study procedures were approved by the Advarra Institutional Review Board (IRB), which served as the single IRB for this study.

Measures

HIV risk behaviors

Three HIV risk behaviors were measured in the survey: condomless sex, sharing syringes/needles, and sharing other injection equipment (cotton, cookers, etc.). Condomless sex in the last three months was measured by asking respondents, “In the last three months, did you have sex with someone without a condom?”. Response options were “yes” and “no.” Sharing syringes/needles and sharing other injection equipment were measured by asking respondents, “In the last three months, which of the following have you done? Check all that apply.” Response options included 1) shared syringes or needles, 2) shared drugs, 3) shared other injection equipment such as cotton or cookers, and 4) none of these things. Sharing syringes/needles in the last three months was dichotomized as those who checked “shared syringes or needles” and those who did not. Sharing other injection equipment in the last three months was dichotomized as those who checked “shared other injection equipment such as cotton or cookers” and those who did not.

Covariates

Primary covariates of interest included age, gender, race, HIV status, experiences in the last three months (homelessness, overdose, seeking treatment for addiction, unmet healthcare need, hospitalization, and incarceration), type of drug injected in the last three months (cocaine, benzodiazepines, fentanyl, heroin, other opioids (OxyContin, Percocet, etc.), and methamphetamine), and injecting more than one substance in the last 3 months. For this analysis, age was initially categorized as 51–60, 61–70, 71–80, and 80 years and older. Due to small cell sizes, age was dichotomized in the logistic regression analyses as 51–60 and 61+. Gender was categorized as male, female, and transgender. Race was categorized as American Indian or Alaska Native, Black/African American, White, multiracial, and other. Due to small cell sizes in the non-Black race groups, race was dichotomized in the logistic regression analyses as Black/African American versus White/multiracial/American Indian or Alaska Native/other. HIV status, experiences in the last three months including homelessness, overdose, seeking treatment for addiction, unmet healthcare need, hospitalization, and incarceration, and drugs injected in the last three months including cocaine, benzodiazepine, fentanyl, heroin, other opioids, and methamphetamine, and injecting more than one substance were dichotomized as “yes” and “no.”

Data analysis

Survey respondents who reported injection drug use in the last three months and were aged >50 were included in the analysis. We used frequencies and percentages to describe sociodemographic, substance use, and HIV risk characteristics of the sample of older PWID. We utilized simple and multivariable logistic regression to determine correlates of condomless sex, sharing syringes/needles, and sharing other injection equipment in the last three months. Characteristics with a p<0.05 in bivariate analyses were included in the multivariable models. Associations with p<0.05 were considered statistically significant. All analyses were performed using Stata 17 (College Station, TX).

Results

Of 1,491 total respondents, 260 were eligible for this analysis. Among the 260 older PWID, 90% were Black/African American, 82% were male, 51% were aged >60, and 7% reported living with HIV. In the last three months, 47% sought treatment for addiction, 44% experienced homelessness, 19% reported an unmet healthcare need, 18% were hospitalized, 14% reported an overdose, and 6% were incarcerated (Table 1). The most commonly reported injection drug used in the last three months was heroin (97%), followed by fentanyl (43%), cocaine (31%), other opioids (5%), benzodiazepines (3%), and methamphetamine (2%). More than half of respondents (57%) reported more than one substance in the last three months (Figure 1). In the last three months, 28% engaged in condomless sex, 22% shared other injection equipment, and 7% shared syringes/needles. Forty-two percent reported at least one HIV risk behavior and 11% reported more than one HIV risk behavior (Figure 2).

Table 1.

Characteristics of older (aged >50) PWID in Washington, DC (N=260)

Characteristic	N (%)
Age
51–60	127 (48.9)
61–70	121 (46.5)
71–80	11 (4.2)
80 and older	1 (0.4)
Gender
Male	212 (81.5)
Female	47 (18.1)
Transgender	1 (0.4)
Race
Black/African American	235 (90.4)
Other	18 (6.9)
Multiracial	5 (1.9)
White	1 (0.4)
American Indian or Alaska Native	1 (0.4)
Living with HIV 1	18 (6.9)
Experiences in the last three months 2
Sought treatment for addiction	123 (47.3)
Experienced homelessness	115 (44.2)
Unmet healthcare need	49 (18.9)
Hospitalized	46 (17.7)
Overdosed	37 (14.2)
Incarcerated	15 (5.8)

PWID (people who inject drugs)

¹n=259, 1 observation missing.

²Participants could select multiple responses.

Figure 1. Self-reported injection drug type in the last 3 months (N=260).

Figure 2. Self-reported HIV risk behavior in the last 3 months (N=260).

Condomless sex

Benzodiazepine injection was significantly associated with higher odds of reporting condomless sex in the last three months (OR=5.61, 95% CI: 1.36–23.06) (Table 2). No other characteristics were significantly associated with condomless sex.

Table 2.

Associations between demographic and drug use characteristics and condomless sex, sharing syringes/needles, and sharing other injection equipment in the last 3 months (N=260)

Characteristic	Condomless sex	Sharing syringes/needles		Sharing other injection equipment
	OR (95% CI)	OR (95% CI)	aOR (95% CI)2	OR (95% CI)	aOR (95% CI)2
Age (Ref= 61+)
51–60	1.70 (0.98–2.94)	1.18 (0.46–3.00)	-	1.01 (0.56–1.83)	-
Gender (Ref=Male) 1
Female	1.27 (0.64–2.53)	2.24 (0.80–6.24)	-	1.27 (0.61–2.65)	-
Race (Ref=Black/African American)
AIAN/White/other/multiracial	0.63 (0.23–1.74)	0.50 (0.06–3.93)	-	0.65 (0.22–1.99)	-
Living with HIV (Ref=Not LWH) 1	1.72 (0.64–4.63)	0.73 (0.09–5.79)	-	1.86 (0.67–5.21)	-
Experienced homelessness	0.80 (0.46–1.39)	2.30 (0.87–6.04)	-	2.45 (1.34–4.47)	2.20 (1.17–4.13)
Overdosed	1.30 (0.62–2.76)	4.10 (1.50–11.24)	3.18 (1.08–9.36)	3.39 (1.62–7.04)	2.57 (1.18–5.61)
Sought treatment for addiction	0.67 (0.39–1.17)	1.00 (0.39–2.55)	-	0.84 (0.46–1.51)	-
Unmet healthcare need	0.71 (0.34–1.48)	1.16 (0.37–3.67)	-	2.01 (1.01–4.00)	1.63 (0.78–3.43)
Hospitalized	0.91 (0.44–1.87)	1.26 (0.40–4.00)	-	1.98 (0.98–4.00)	-
Incarcerated	2.42 (0.85–6.95)	3.58 (0.92–13.98)	-	0.88 (0.24–3.25)	-
Cocaine Injection	0.80 (0.44–1.46)	2.11 (0.82–5.42)	-	1.39 (0.75–2.58)	-
Benzodiazepine Injection	5.61 (1.36–23.06)	7.34 (1.68–32.12)	3.54 (0.62–20.21)	7.84 (1.90–32.43)	6.72 (1.49–30.24)
Fentanyl Injection	1.29 (0.75–2.23)	1.54 (0.60–3.93)	-	1.67 (0.93–3.02)	-
Heroin Injection	0.96 (0.18–5.04)	0.46 (0.05–4.03)	-	1.71 (0.20–14.46)	-
Other Opioid Injection	0.51 (0.11–2.38)	7.77 (2.10–28.75)	4.61 (1.04–20.40)	2.69 (0.82–8.83)	-
Methamphetamine Injection	4.04 (0.66–24.7)	9.33 (1.46–59.69)	5.76 (0.61–54.57)	2.42 (0.40–14.87)	-
Injected more than one substance	1.06 (0.61–1.84)	1.67 (0.62–4.55)	-	1.66 (0.90–3.07)	-

AIAN (American Indian/Alaska Native); LWH (Living with HIV).

Bolded cells indicate p<0.05

¹n=259, 1 observation missing.

²Adjusted for all variables in the column.

Sharing syringes/needles

In the bivariate analyses, experiencing overdose (OR=4.10, 95% CI: 1.50–11.24), benzodiazepine injection (OR=7.34, 95% CI: 1.68–32.12), methamphetamine injection (OR=9.33, 95% CI: 1.46–59.69), and other opioid injection (OR=7.77, 95% CI: 2.10–28.75) were significantly associated with a higher odds of sharing syringes/needles in the last three months. In the multivariable model, reporting an overdose (aOR=3.18, 95% CI: 1.08–9.36) and other opioid injection (aOR=4.61, 95% CI: 1.04–20.40) remained significantly associated with sharing syringes/needles (Table 2).

Sharing other injection equipment

In the bivariate analyses, reporting homelessness (OR=2.45, 95% CI: 1.34–4.47), overdose (OR=3.39, 95% CI: 1.62–7.04), unmet healthcare need (OR=2.01, 95% CI: 1.01–4.00), and benzodiazepine injection (OR=7.84, 95% CI: 1.90–32.43) were significantly associated with higher odds of sharing other injection equipment in the last three months. In the multivariable model, reporting homelessness (aOR=2.20, 95% CI: 1.17–4.13), overdose (aOR=2.57, 95% CI: 1.18–5.61), and benzodiazepine injection (aOR=6.72, 95% CI: 1.49–30.24) remained significantly associated with sharing other injection equipment in the last three months (Table 2).

Discussion

We found that ongoing HIV risk behaviors were alarmingly high among community-recruited older PWID in Washington, DC. Nearly half (42%) of participants reported at least one and 11% reported more than one HIV risk behavior, including condomless sex, sharing syringes, and sharing other injection equipment. Experiencing homelessness, recent overdose, and injecting certain substances were correlated with HIV risk behavior. Improvements in our understanding about the HIV risks of older PWID may inform the development of novel and impactful age-tailored public health interventions.

We found a lower prevalence of condomless sex in our local sample of older PWID compared to the NHBS study conducted in 20 US cities. In 2022, NHBS found that among male PWID aged 50 or older, 59% reported condomless vaginal sex, 19% reported condomless anal sex with female partners, and 4% reported condomless anal sex with male partners in the past 12 months, and among female PWID aged 50 or older, 57% reported condomless vaginal sex and 16% reported condomless anal sex with a male partner in the past 12 months (23). The national-level prevalence of condomless vaginal sex was higher compared to our finding of 28% reporting condomless sex, which may be due to different denominators, recall time frames, and sampling methods. Older adults may have lower levels of condom use due to lower perceived risk of STI acquisition (27), particularly with respect to HIV (28). Reduced concern about unplanned pregnancy may also contribute to lower condom use in older adults (29). Additionally, we found that benzodiazepine injection was associated with increased odds of condomless sex. This is consistent with findings from a study conducted in Canada that identified depressant (benzodiazepine) use as an independent risk factor for HIV infection among PWID. The cognitive effects of benzodiazepine use may contribute to impaired judgment and engagement in risk behaviors such as condomless sex and sharing injection equipment (30).

Our preliminary findings add to the existing literature by suggesting that older PWID engage in substance use and sexual behaviors that may increase their HIV risk. In 2022, NHBS found nationally that among PWID aged 50 or older who tested negative for HIV, 17% reported sharing syringes and 37% reported sharing injection equipment in the past 12 months (23). We found more moderate levels of substance use risk behaviors in our sample of older PWID, with 7% reporting sharing syringes/needles and 22% reporting sharing other injection equipment in the past three months. The lower levels of sharing syringes/needles and other injection equipment found in our study may be explained by the shorter recall period compared to NHBS (three months vs. 12 months). Overall, 42% of our sample of older PWID self-reported engaging in at least one HIV risk behavior in the last three months, indicating a level of HIV risk behavior that warrants intervention.

Almost half of our sample had recently experienced homelessness. Although experiencing homelessness was not associated with sharing syringes, as has been shown in previous studies (31–33), it was associated with sharing other injection equipment. This is consistent with previous research (31–33). PWID experiencing homelessness may use others’ syringes and injection equipment to inject quickly when in public settings, and to avoid possessing paraphernalia during police encounters (34). People experiencing homelessness may also lack a safe place to store unused injection equipment (35).

We also found a concerning prevalence of self-reported non-fatal overdose in older PWID, which was associated with increased odds of HIV risk behavior. Syringe service programs (SSPs) have been instrumental in providing support services and reducing HIV risk and overdose among PWID (36). Expanding harm reduction efforts through SSPs is critical to reducing HIV infection and overdose among both younger and older PWID in Washington, DC.

Additionally, healthcare providers may underestimate the risk of HIV in their older adult patients and may be uncomfortable initiating conversations about sexual health with older patients (37). This can result in missed opportunities for HIV prevention counseling and testing in older populations (38,39). HIV pre-exposure prophylaxis (PrEP) is also underutilized in older PWID. Data specific to older PWID are not available, but among PWID over age 18 in Washington, DC, the NHBS study found that in 2022 only 0.5% had taken PrEP in the last 30 days. However, 74% reported they would be likely to take PrEP if it were free of cost (22). One survey of primary care physicians found a lower willingness to prescribe PrEP to PWID compared to other groups, and even lower levels when PWID had a history of methadone treatment. Providers may be less likely to recommend PrEP to PWID due to negative attitudes toward PWID and concerns about their adherence to PrEP and follow-up care (40).

Age-specific substance use treatment and HIV prevention programs are needed to address the growing aging PWID population in Washington, DC. This may also apply elsewhere in the US. Nearly half of our sample reported seeking treatment for addiction in the past three months. Despite motivation to seek treatment, older PWID may encounter a lack of openings in treatment programs or may not receive age-appropriate care (41). Research is also limited on medication for opioid use disorder (MOUD) in older adults. Older adults may respond differently to MOUD and may have different preferences related to non-pharmacological interventions compared to younger users (42). Older adults who initiated heroin use after age 30 are less likely to be prescribed MOUD compared to those who initiated at younger ages (43). Additionally, healthcare providers, caregivers, family and friends, and older adults themselves may fail to recognize substance misuse in older adults. As a result, substance use disorders are underdiagnosed and therefore undertreated in older adults (44). Older adults may be more likely to seek out addiction care from primary care providers, who may be less familiar with substance use screening and treatment (45). Special considerations are needed when screening older adults for substance use disorders, including using tailored screening instruments, differentiating between symptoms of substance misuse and cognitive decline or mental illness, and creating an affirming and age-appropriate environment (45,46). Future research should explore age group differences in substance use patterns, HIV risk behavior, and access to treatment and prevention services. Additionally, future studies should develop and test interventions that integrate geriatric care, substance use treatment, and HIV prevention services.

Limitations

There are several limitations to this study. Because this survey was cross-sectional, we cannot assess temporality or causality between respondent characteristics and HIV risk behaviors. We used a convenience sample of PWID recruited through street outreach, and results may not be generalizable to all PWID in Washington, DC. Our results are also susceptible to reporting biases, particularly social desirability bias given the sensitive subject matter. The sample sizes for benzodiazepine and other opioid injection were also small, and results should be interpreted with caution. Given the brief nature of the survey and small sample size, there is also potential for residual confounding and low precision of some estimates.

Additionally, our sample was mostly composed of older male PWID (82%). Future research is needed to investigate HIV risk behaviors in older women who inject drugs (WWID). Of those aged 55 or older who were newly diagnosed with HIV in Washington, DC from 2018–2022, two in five were women, compared to one in four who were men who have sex with men (22). WWID may have a greater risk of acquiring HIV compared to their male counterparts (8,47–49). Gender-based power imbalances may restrict WWID’s agency around harm reduction behaviors and expose them to heightened levels of stigma and violence (50).

Strengths

Our study also had several strengths. One strength of this study is that our sample of older PWID was majority Black/African American. Our results offer needed preliminary insight into HIV risk behavior in older Black/African American PWID. In addition, as a community-recruited rather than clinic-based study, our data provide preliminary insight into characteristics of older PWID who may not access healthcare or other services. These data are unique in their ability to help characterize the older PWID population in ways that can inform age-tailored intervention development.

Conclusion

Our findings suggest that older PWID in Washington, DC engage in sexual and substance use behaviors that place them at increased risk for HIV. Experiencing homelessness, recent overdose, and injecting certain substances were associated with HIV risk behaviors. Age-specific HIV prevention and substance use treatment programs are urgently needed to address the growing aging PWID population in DC. Continued HIV prevention efforts are critical to maintaining the low incidence of HIV infection among PWID in Washington, DC and beyond.