HIV incidence and factors associated with HIV risk among people who inject drugs engaged with harm-reduction programmes in four provinces in South Africa: A retrospective cohort study

Background:
HIV incidence among people who inject drugs (PWID) in South Africa has never been estimated. We estimated HIV incidence and associations with risk and protective factors among PWID engaged with harm-reduction services.

Methods:
We used programmatic data collected over 01/04/2019-30/03/2022 by the Networking HIV and AIDS Community of South Africa (NACOSA), which offers harm-reduction services and HIV testing to PWID. During this three-year period, services were delivered through drop-in centres and outreach in four South African provinces: Gauteng, KwaZulu-Natal, Western and Eastern Capes. PWID who did not self-report being HIV-positive and were not tested within the previous 6 weeks were offered HIV testing. Using a retrospective cohort design, we estimated HIV incidence among initially HIV-negative PWID tested at least twice, assuming seroconversions occurred at the midpoint between the last negative and first positive tests. Using Cox regression models, we assessed associations between HIV seroconversion risk and several factors, including socio-demographic characteristics, primary drug(s) used, uptake of interventions (i.e., number of harm-reduction packs, opioid agonist treatment (OAT)) and HIV testing interval.

Findings:
Of 31,182 PWID accessing NACOSA-supported services, 20,955 (including 3,409 self-reporting being HIV-positive) were not tested for HIV. Of 10,227 PWID tested at least once, 8,152 were HIV-negative at first test and, of these, 2,402 had ≥2 tests and formed the study sample. At baseline, most were male (90.4%), Black (72.4%) and homeless or unstably housed (69.6%). Median age was 30yrs and few (5.5%) received OAT in the past year. Overall, 283 (11.8%) PWID acquired HIV over 2306.1 person-years (py). HIV incidence was higher in Gauteng (16.7/100py; 95%CI: 14.5-19.1) and KwaZulu-Natal (14.9/100py; 95%CI: 11.3-19.3), than in the Eastern (5.0/100py; 95CI% 2.3-9.6) and Western Capes (3.2/100py; 95%CI: 1.9-4.9). In multivariable Cox models, HIV acquisition risk varied by race, primary drug(s) used and interval between HIV tests. Additionally, PWID who received OAT in the past year had lower HIV risk (adjusted hazard ratio: 0.48; 95%CI: 0.22-1.03), though the CI was wide and crossed the null.
Interpretation: HIV incidence among PWID engaged with harm-reduction services and receiving repeat HIV testing in South Africa is very high.

Funding:
Wellcome Trust, CIHR, Global Fund.