Detection of HIV-1 drug resistance in RNA and proviral DNA genotyping is variable both longitudinally and during repeat testing

Background

HIV-1 variants harboring resistance-associated mutations (RAMs) can be archived in viral reservoirs and then re-emerge. Given the dynamic properties of the latent reservoir and detection limits of genotypic assays, RAM persistence over time is incompletely understood.

Objective

This retrospective analysis investigated RAM detection in adults with HIV.

Study design

Genotyping of protease, reverse transcriptase, and integrase from 3 bictegravir/emtricitabine/tenofovir alafenamide switch studies was included. Longitudinal analyses were performed for participants with combinations of RNA and proviral DNA genotype reports from ≥2 pre-switch timepoints. Reported RAMs assessed at 2 timepoints were categorized as 100 % or 50 % detection, and those assessed at ≥3 timepoints as persistent, lost, gained, or inconsistent detection. From each whole blood sample, reproducibility (%) of proviral RAM reporting was the number of times the mutation was detected per number of assays run (2–4 replicates).

Results

In 223 participants, of 262 RAMs tracked longitudinally over 2 timepoints, 39 % (103/262) had 100 % detection and 61 % (159/262) had 50 % detection, with 64 % (101/159) detected at timepoint 2. In 25 participants with ≥3 timepoints, detection of 57 RAMs was categorized as persistent (19 %; 11/57), lost (12 %; 7/57), gained (26 %; 15/57), or inconsistent (42 %; 24/57). Mean (standard deviation) reproducibility of proviral RAM detection at 1 timepoint was 80.9 % (27.2 %) (336 RAMs from 70 participants).

Conclusions

No consistent pattern of longitudinal RAM detection was observed. Reproducibility of proviral genotyping was high but variable. Since RAMs were not always consistently detected, an individual’s cumulative resistance history should be considered for optimal treatment management.