# Prevalence of Illicit Drug Detection in 5 US Cities Among Out-of-Treatment People Who Inject Drugs

## TL;DR

This study finds high rates of fentanyl and xylazine use among drug users in 5 US cities, stressing the need for better drug surveillance and harm reduction.

## Contribution

The study provides new insights into drug use patterns among marginalized populations and highlights geographic and demographic variations in illicit drug detection.

## Key findings

- Fentanyl was detected in 93% of participants, with high prevalence across all cities.
- Xylazine detection increased over time in New York City and remained high in Philadelphia.
- Unhoused individuals had higher cocaine detection, and recently incarcerated individuals had higher stimulant detection.

## Abstract

This cross-sectional study examines prevalence of and trends in illicit drugs detected over 2 years in 5 US cities among individuals who inject drugs and are not engaged in medical care.

What is the prevalence of illicit drug detection among people who inject drugs who are not engaged in medical care across 5 US cities?

In this cross-sectional study of 444 people who inject drugs, fentanyl was detected among 93%, xylazine among 53%, polysubstance drugs among 95%, and amphetamine among 67%; unhoused and recently incarcerated individuals had higher prevalence of cocaine and stimulant detection, respectively. Amphetamine prevalence increased over time in Washington, DC, and Philadelphia had the highest prevalence of xylazine detection.

The findings suggest localized, real-time drug surveillance for out-of-treatment individuals who inject drugs, often missed by traditional public health systems, is needed.

In the US, the overdose crisis continues to be driven by fentanyl, xylazine, and stimulant-involved polysubstance use among people who inject drugs, especially those who are not engaged in medical care.

To estimate the overall prevalence of illicit drugs detected among people who inject drugs in 5 US cities by city and sociodemographic characteristics and assess trends in drug detection over a 2-year period.

This cross-sectional study included data from individuals enrolled in HIV Prevention Trials Network (HPTN) 094, a randomized clinical trial to evaluate an integrated mobile unit engaging adults aged 18 years or older who inject drugs in HIV services between June 2021 and September 2023 in New York City; Houston, Texas; Los Angeles, California; Philadelphia, Pennsylvania; and Washington, DC. All analyses were completed between August 2021 and August 2025.

Baseline prevalence of toxicologic detection was assessed using liquid chromatography–high-resolution mass spectrometry. Differences in toxicologic detection by sociodemographic characteristics (age, race and ethnicity, housing status, and incarceration history), study site, and illicit drug type over time were analyzed using generalized linear models.

Across 444 participants, 303 (68.2%) were male, 267 (60.1%) were aged 30 to 49 years, 203 of 440 (46.1%) were unhoused, and 91 of 442 (20.6%) had a recent incarceration history. In all, 414 participants (93.2%) tested positive for fentanyl, 328 (73.9%) for cocaine, 299 (67.3%) for amphetamine-type stimulants, and 234 (52.7%) for xylazine. Nearly all participants (421 [94.8%]) tested positive for polysubstance drugs (fentanyl or opioids with stimulants, benzodiazepines, cocaine, and/or xylazine). Fentanyl detection was high across all sites; xylazine was most common in New York City (68 of 94 [72.3%]), Philadelphia (111 of 112 [99.1%]), and Washington, DC (31 of 41 [75.6%]). Every 6 months, xylazine detection increased in New York City by 10.3% (95% CI, 4.0%-16.5%; P = .001), and its prevalence stayed high in Philadelphia. Amphetamine-type stimulant detection increased in Washington, DC, by 15.0% (95% CI, 2.9%-27.1%) every 6 months over the enrollment period (P = .02). Across all sites, cocaine prevalence was higher among unhoused than housed participants (difference, 11.4%; 95% CI, 3.6%-19.2%; P = .004), and stimulant detection was elevated among those recently incarcerated vs not (difference, 9.9%; 95% CI, 1.4%-18.5%; P = .02).

This cross-sectional study found widespread fentanyl and polysubstance detection, with rising xylazine and stimulant detection that varied by sociodemographic and structural vulnerabilities and may be due to adulteration within the unregulated drug supply. These findings highlight urgent public health needs for real-time drug supply surveillance, targeted harm-reduction services, and integrated treatment approaches to reduce overdose risk and address social and structural vulnerabilities.