Risk of Developing Low-Level Viral Rebound Among People With HIV Receiving 2- or 3-Drug Regimens: A Case-Control Study Nested in the ICONA Cohort

Background: The risk of developing low-level viral rebound (LLVR) after achieving human immunodeficiency virus HIV-1 (HIV) RNA suppression with 2-drug regimens (2DR) compared to 3-drug regimens (3DR) remains uncertain. Methods: We conducted a matched 1:3 case-control study nested within the ICONA cohort including persons with HIV (PWH) who achieved virological suppression (≥2 consecutive HIV-RNA ≤50 copies/mL over 6 months) after 11 November 2014 (baseline [BL]). Cases were defined as PWH experiencing a single HIV-RNA of 51–199 copies/mL after BL (index date); controls were defined as those who maintained HIV-RNA ≤50 copies/mL up to the index date and were matched for history of care gaps and number of drugs failed. The cumulative incidence of LLVR after virological suppression was estimated using Kaplan–Meier methods, and conditional logistic regression models were used to evaluate the association between the current antiretroviral therapy regimen (2DR [dolutegravir/lamivudine, dolutegravir/rilpivirine, dolutegravir/doravirine, or cabotegravir/rilpivirine] vs 3DR [dolutegravir, bictegravir, rilpivirine, doravirine, boosted darunavir, or boosted atazanavir–based with a backbone of tenofovir and lamivudine or emtricitabine]) and LLVR risk. In sensitivity analyses cases were defined as PWH experiencing 2 consecutive HIV-RNA of 51–199 copies/mL. Results: Among 1033 PWH (261 cases, 772 matched controls), 21% were female, median age was 43 (interquartile range [IQR], 34–51) years, and BL CD4 count was 601 (IQR, 379–826) cells/μL; 2DR use was 25% in cases versus 29% in controls (P = .23). Two years after viral suppression, cumulative LLVR incidence was 2.7% (95% CI, 2.3%–3.1%) when considering single LLVR events and 1.8% (95% CI, 1.4%–2.1%) when limited to 2 consecutive values. After adjusting for confounding, evidence for an association with current regimen was inconclusive (adjusted odds ratio, 0.86 [95% CI, .57–1.29]). Conclusions: Despite the wide range of plausibility, we can exclude a risk of LLVR higher than 29% when using 2DR versus 3DR regimens.