People with HIV (PWH) with substance use disorders (SUD) have worse health outcomes than PWH without SUD. Our objective was to characterize substance use patterns and their impact on longitudinal HIV RNA trajectories among those enrolled in an observational study of PWH in care in Washington, DC. Substance use by type (alcohol, cannabis, opioid, stimulant, hallucinogen, inhalant, sedative) was used to identify shared patterns of substance use using Latent Class Analysis (LCA). A multinomial logistic regression model evaluated the association between the resulting substance use classes and the membership probability in longitudinal HIV RNA trajectory groups. There were 30.1% of participants with at least one substance reported. LCA resulted in a three-class model: (1) Low-Level Substance Use, (2) Opioid Use, and (3) Polysubstance. The Opioid and Polysubstance Use classes were more likely to have a mental health diagnosis (45.4% and 53.5%; p < 0.0001). Members in the Opioid Use class were older (median age of 54.9 years (IQR 50.3-59.2) than both the Polysubstance and Low-Level Substance Use Classes (p < 0.0001). There were 3 HIV RNA trajectory groups: (1) Undetectable, (2) Suppressed, and (3) Unsuppressed HIV RNA over 18 months of follow-up. The probability of being in the unsuppressed HIV RNA group trajectory when a member of the Opioid Use or Polysubstance Use classes was 2.5 times and 1.5 times greater than the Low-Level Substance Use class, respectively. The Opioid Use and Polysubstance Use classes, with higher-risk drug use, should be approached with more targeted HIV-related care to improve outcomes.
Keywords: Group-based trajectory modeling; HIV; Latent class analysis; Substance use; Viral load.
© 2024. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.