Background:
Preexisting resistance can affect antiretroviral (ARV) efficacy. Circulating HIV-1 variants with drug resistance-associated mutations (RAMs) can be archived in viral reservoirs, where they can persist and re-emerge. Given the dynamic properties of the latent reservoir, detection of these RAMs over time has not been well defined.
Methods:
Participants from bictegravir/emtricitabine/tenofovir alafenamide (B/F/TAF) switch studies (4449, 4030, 4580; virologic suppression [≥3-6 months]; no prior virologic failure) with HIV-1 RNA or proviral HIV-1 DNA genotyping (population reporting) from ≥2 pre-switch timepoints were included. For 2 timepoints only, RAMs in protease, reverse transcriptase and integrase, and other substitutions (accessory, polymorphisms) were categorized as detected 100% (2 out of 2 tests) or 50% (1 out of 2 tests; subcategorized as detection lost or gained) of the time. For >2 timepoints, RAMs were categorized as lost, gained, persistent, or fluctuating. Time between first and last tests was reported in median years (y). Nonparametric statistics were used.
Results:
In all, 235 participants had evaluable data (longitudinal tracking of ≥1 RAM or other substitution). For 206 participants with 2 timepoints, 103 of 256 RAMs reported (40.2%) had 100% detection, and 153 (59.8%) had 50% detection, with 95 (62.1%) gained and 58 (37.9%) lost (Figure). 100% detection was lower for RAMs versus other substitutions (n=3329/4339 [76.7%]; p<0.0001). Time between tests was not different between RAMs with 100% and 50% detection: 7.7 y (quartile 1-3 [IQR] 3.2-11.1) vs 5.9 y (IQR 3.2-9.7), respectively (p=0.06). For 29 participants with >2 timepoints (median 3 reports, 9.3 y [IQR 5.7-13.9] between first and last tests), 66 RAMs were categorized as fluctuating (48.5%), gained (25.8%), persistent (15.2%), or, least common, lost (10.6%). K103N (n=10) and M184V/I (n=16) were predominantly fluctuating (60.0% and 43.8%, respectively) and only lost in 10.0% and 6.3% of cases, respectively.
Conclusions:
Some RAMs were consistently reported, but the majority were newly detected or fluctuated and did not disappear significantly over time, which most likely reflects both the ongoing decay and proliferation of the latent reservoir and, ARV pressure. RAMs were not always detected, and this lack of longitudinal stability enforces the need to consider an individual’s treatment history and all past genotyping test results (and the detection sensitivities of reports) for treatment management.
