Background:  Transmission of drug-resistant HIV is a threat to successful ART. Prior studies have suggested that mutations acquired in primary infection often persist over time, but these studies have been limited in size and duration of observation.

Methods:  We identified patients meeting the following criteria from 2 clinical cohorts (1 in San Francisco, California and 1 in São Paulo, Brazil):  acute/recent HIV infection (<12 months), presence on an initial genotype of ≥1 drug resistance mutation, using published guidelines which exclude common polymorphisms, and at least 3 months of subsequent follow-up off ART. Genotyping was performed both at baseline and subsequently.  In persons with mutation reversion, we calculated the average time to first wild-type genotype.

Results:  We identified 61 patients (46 in San Francisco, and15 in São Paulo) with a total of 164 mutations. The mean duration of follow-up was 2.6 years (range 3 months to 10.3 years). Of the 164 mutations, 31 (18.9%) reverted to wild type. The lamivudine (3TC) -associated mutations M184V/I were lost significantly more often than were all other mutation groups:  77% mutations at this codon reverted vs 15% of TAM (p <0.001) (M41L, D67N, K70R, L210W, T215Y/F, and K219Q/E ), 0% of T215 revertants (p <0.001), 26% of NNRTI mutations (p = 0.002), and 9% of PI mutations (p <0.001). No K65R mutations were noted.

Conclusions:  Transmitted drug resistance mutations revert to wild type in many patients, although in highly variable proportions. The M184V/I mutations revert most often, potentially because they reduce viral fitness and are less likely to be fixed by compensatory mutations. They revert significantly more often than TAM, T215 revertants, NNRTI or PI mutations. NNRTI mutations were commonly transmitted, but reverted infrequently; PI mutations reverted least frequently. Several partial reversions were also observed. Thus, genotyping ARV-naive individuals with chronic HIV might only reveal a fraction of what was acquired during primary infection. Differential rates of reversion may partially explain the relative infrequency of M184V mutants among ARV-naïve persons compared to treated populations. Rates of reversion need to be considered in surveillance of transmitted drug resistance.