BackgroundMinor (i.e., <20% prevalence) drug-resistant human immunodeficiency virus (HIV) variants may go undetected, yet be clinically important ObjectivesTo compare the prevalence of drug-resistant variants detected with standard and ultra-deep sequencing (detection down to 1% prevalence) and to determine the impact of minor resistant variants on virologic failure (VF) MethodsThe Flexible Initial Retrovirus Suppressive Therapies (FIRST) Study (N = 1397) compared 3 initial antiretroviral therapy (ART) strategies. A random subset (n = 491) had baseline testing for drug-resistance mutations performed by use of standard sequencing methods. Ultra-deep sequencing was performed on samples that had sufficient viral content (N = 264). Proportional hazards models were used to compare rates of VF for those who did and did not have mutations identified ResultsMutations were detected by standard and ultra-deep sequencing (in 14% and 28% of participants, respectively; P<.001). Among individuals who initiated treatment with an ART regimen that combined nucleoside and nonnucleoside reverse-transcriptase inhibitors (hereafter, “NNRTI strategy”), all individuals who had an NNRTI-resistance mutation identified by ultra-deep sequencing experienced VF. When these individuals were compared with individuals who initiated treatment with the NNRTI strategy but who had no NNRTI-resistance mutations, the risk of VF was higher for those who had an NNRTI-resistance mutation detected by both methods (hazard ratio HR, 12.40 95% confidence interval {CI}, 3.41–45.10) and those who had mutation(s) detected only with ultra-deep sequencing (HR, 2.50 95% CI, 1.17–5.36) ConclusionsUltra-deep sequencing identified a significantly larger proportion of HIV-infected, treatment-naive persons as harboring drug-resistant viral variants. Among participants who initiated treatment with the NNRTI strategy, the risk of VF was significantly greater for participants who had low- and high-prevalence NNRTI-resistant variants