BACKGROUND HIV incidence among people who inject drugs in South Africa has never been estimated. We aimed to estimate HIV incidence and associations with risk and protective factors among people who inject drugs engaged with harm-reduction services. METHODS For this retrospective cohort study we used programmatic data collected from April 1, 2019, to March 30, 2022, by the Networking HIV and AIDS Community of South Africa, which offers harm-reduction services and HIV testing to people who inject drugs. During this 3-year period, services were delivered through drop-in centres and outreach in four South African provinces: Gauteng, KwaZulu-Natal, Western Cape, and Eastern Cape. Our cohort comprised people who inject drugs who did not self-report being HIV positive, were HIV negative at first testing, and had at least one follow-up test. Data were collected by outreach teams. We estimated HIV incidence, assuming seroconversions occurred at the midpoint between the last negative test and first positive test. We assessed associations between HIV seroconversion risk and several factors with Cox regression models, including sociodemographic characteristics, primary drugs used, uptake of interventions (ie, number of harm-reduction packs and opioid agonist treatment [OAT]), and HIV testing interval. FINDINGS Of 31 182 people who inject drugs accessing harm-reduction services, 20 955 (including 3409 self-reporting being HIV positive) were not tested for HIV. Of 10 227 people who tested at least once, 8152 were HIV negative at first test and of these, 2402 had at least two tests and formed the study cohort. Overall, 283 (11·8%) people who inject drugs acquired HIV over 2306·1 person-years. HIV incidence was higher in Gauteng (16·7 per 100 person-years; 95% CI 14·5-19·1) and KwaZulu-Natal (14·9 per 100 person-years; 11·3-19·3), than in the Eastern Cape (5·0 per 100 person-years; 2·3-9·6) and Western Cape (3·2 per 100 person-years; 1·9-4·9). In multivariable Cox models, HIV acquisition risk varied by race, primary drugs used, and interval between HIV tests. Additionally, people who injected drugs and received OAT in the past year had lower HIV risk (adjusted hazard ratio 0·48; 95% CI 0·22-1·03) than people who did not receive OAT, although the 95% CI was wide and crossed the null. INTERPRETATION Our study highlights a pressing need for scale-up of HIV prevention strategies, particularly opioid agonist treatment, for people who inject drugs in South Africa. Dedicated investments are needed to develop monitoring systems for HIV incidence, risk behaviours, and uptake of interventions to ensure effective and equitable programmes. FUNDING Wellcome Trust, Canadian Institutes of Health Research, and Global Fund to Fight AIDS, Tuberculosis and Malaria.

中文翻译：

---

在南非四个省份参与减少危害计划的注射吸毒者的 HIV 发病率和与 HIV 风险相关的因素：一项回顾性队列研究。


背景 在南非，注射吸毒者的 HIV 发病率从未被估计过。我们旨在估计参与减害服务注射吸毒者的 HIV 发病率以及与风险和保护因素的关联。方法 对于这项回顾性队列研究，我们使用了 2019 年 4 月 1 日至 2022 年 3 月 30 日由南非网络 HIV 和艾滋病社区收集的程序数据，该社区为注射吸毒者提供减少伤害服务和 HIV 检测。在这 3 年期间，通过南非四个省份的临时中心和外展服务提供：豪登省、夸祖鲁-纳塔尔省、西开普省和东开普省。我们的队列包括自我报告未报告 HIV 阳性、首次检测时 HIV 阴性且至少进行过一次随访测试的注射吸毒者。数据由外展团队收集。我们估计了 HIV 发病率，假设血清转化发生在最后一次阴性检测和第一次阳性检测之间的中间点。我们评估了 HIV 血清转换风险与几个因素与 Cox 回归模型之间的关联，包括社会人口学特征、使用的主要药物、干预措施的接受情况 （即减少伤害包和阿片类激动剂治疗 [OAT] 的数量）和 HIV 检测间隔。结果在 31 182 名获得减害服务的注射吸毒者中，20 955 人（包括 3409 名自我报告为 HIV 阳性的人）未接受 HIV 检测。在至少检测过一次的 10 227 人中，8152 人在第一次检测时为 HIV 阴性，其中 2402 人至少进行了两次检测并形成了研究队列。总体而言，283 名 （11·8%） 注射吸毒者在 2306·1 人年内感染了 HIV。 豪登省（16·7/100 人年;95% CI 14·5-19·1）和夸祖鲁-纳塔尔省（14·9/100 人年;11·3-19·3）的 HIV 发病率高于东开普省（5·0/100 人年;2·3-9·6）和西开普省（3·2/100 人年;1·9-4·9）。在多变量 Cox 模型中，HIV 感染风险因种族、使用的主要药物和 HIV 检测间隔而异。此外，与未接受 OAT 的人相比，过去一年注射药物并接受 OAT 的人的 HIV 风险较低（调整后的风险比 0·48;95% CI 0·22-1·03），尽管 95% CI 很宽且越过零。解释 我们的研究强调了迫切需要扩大 HIV 预防策略，特别是阿片类激动剂治疗，用于南非注射吸毒者。需要专门投资来开发 HIV 发病率、风险行为和干预措施接受度的监测系统，以确保有效和公平的计划。资助 惠康信托基金、加拿大卫生研究院和全球抗击艾滋病、结核病和疟疾基金。