Patterns of heroin and cocaine injection and plasma HIV-1 RNA suppression among a long-term cohort of injection drug users

Abstract

Background: Previous studies suggest that active drug use may compromise HIV treatment among HIV-positive injection drug users (IDU). However, little is known about the differential impacts of cocaine injection, heroin injection, and combined cocaine and heroin injection on plasma HIV-1 RNA suppression.

Methods: Data were derived from a longstanding open prospective cohort of HIV-positive IDU in Vancouver, Canada. Kaplan-Meier methods and Cox proportional hazards regression were used to examine the impacts of different drug use patterns on rates of plasma HIV-1 RNA suppression.

Results: Between May 1996 and April 2008, 267 antiretroviral (ART) naïve participants were seen for a median follow-up duration of 50.6 months after initiating ART. The incidence density of HIV-1 RNA suppression was 65.2 (95%CI: 57.0-74.2) per 100 person-years. In Kaplan-Meier analyses, compared to those who abstained from injecting, individuals injecting heroin, cocaine, or combined heroin/cocaine at baseline were significantly less likely to achieve viral suppression (all p<0.01). However, none of the drug use categories remained associated with a reduced rate of viral suppression when considered as time-updated variables (all p>0.05).

Conclusions: Active injecting at the time of ART initiation was associated with lower plasma HIV-1 RNA suppression rates; however, there was no difference in suppression rates when drug use patterns were examined over time. These findings imply that adherence interventions for active injectors should optimally be applied at the time of ART initiation.

Figure 1

Kaplan–Meier estimates of cumulative rates of plasma HIV-1 RNA suppression within a sample of 267 injection drug users who initiated highly active antiretroviral therapy (ART) during the study period. All pairwise comparisons between injecting groups were non-significant (log-rank: p > 0.05). All pairwise comparisons between the non-injecting group and the three drug-injecting groups were significant (log-rank p < 0.01).

Figure 2

The relationship between substance use throughout follow-up and time to viral load suppression among a prospective cohort of HIV-positive injection drug users (n = 267). Model 1: Unadjusted association between drug use and viral load suppression. Model 2: Adjusted for baseline CD4 count, baseline log₁₀(viral load). Model 3: Adjusted for baseline CD4 count, baseline log₁₀(viral load), age, current enrollment in methadone maintenance therapy, adherence (past 6 months), and ART regimen.