Persistently high HIV incidence among men who have sex with men and people who inject drugs attending integrated care centres in India: a longitudinal assessment of clinic-based data

Introduction: Globally, there have been significant declines in HIV incidence over the past two decades, but this decline is slowing, and in some settings, declines have stalled or are growing-particularly where epidemics are concentrated in key populations (KPs). Understanding temporal changes in HIV incidence among KP is critical yet, due to logistical constraints, there are few sources of longitudinal incidence data, particularly among KP. Methods: We present HIV incidence rates from June 2014 to December 2022 among cisgender men who have sex with men (MSM) and people who inject drugs (PWID) attending community-based integrated care centres (ICCs) in 15 Indian cities. ICCs, established between 2014 and 2017, provide HIV testing and other services to MSM (eight sites) or PWID (eight sites). Client HIV testing data were included in the analysis if they had >= 2 tests and were not positive on the first test. We calculated incidence rates per 100 person-years (PY), stratified by KP, city/site and year. Poisson regression explored associations of incidence with time, age, gender (PWID only) and ICC use. Results: From June 2014 to December 2022, 13,501 clients (5722 MSM, 7779 PWID) had >= 2 HIV tests over a median of 1.8 years. There were a total of 1093 incident HIV acquisitions. Overall incidence rates for MSM and PWID were 1.9/100 PY (95% CI: 1.7-2.2) and 4.1 (3.9-4.4), respectively. Among MSM sites, incidence ranged from 0.4 to 3.5 and in PWID sites from 0.6 to 17.9. From adjusted models, incidence increased by 17% annually among MSM. Among PWID, incidence increased by 11% annually up until 2020 and then decreased by 29% after 2020; when excluding the outlier of New Delhi, incidence was stable among PWID. MSM and PWID 21-25 years old had the highest risk of HIV and among PWID, those more consistently engaged in medication for opioid use disorder were at the lowest risk. Conclusions: While there was substantial geographic variability, MSM and PWID engaged in a free community-based clinic experienced persistently high HIV incidence (>2/100 > 2/100 PY). KP in low- and middle-income countries should be a focus when considering novel strategies such as long-acting pre-exposure prophylaxis to curtail incidence.