Background: Population-based studies suggest that drug resistant HIV-1 is less likely to be transmitted than drug-susceptible viruses. We aimed to evaluate the likelihood of transmission of HIV-1 drug resistance between confirmed partner-pairs.
Methodology: Persons with primary HIV-1 infection have enrolled in an observational cohort at the University of Washington Primary Infection Clinic (PIC) since 1992. We attempted to identify sex partners of PIC enrollees. Partial gp120 env sequences were generated from transmitting and recipient partners, and relationships were confirmed by phylogenetic linkage and genetic distance analysis. Drug resistance mutations in pol encoding Pro and RT regions were assessed by pyrosequencing of >100 viral templates from plasma RNA and PBMC DNA from partner-pairs enrolled after 1995. We considered a mutation to be present if found in >1% of pyrosequences.
Results: Since 1992, 36 (72%) of 50 partner-pairs have had genetically linked infections. Plasma RNA and PBMC DNA specimens for pyrosequencing were obtained from 31 male recipient partners a median of 23 (IQR 18-40) and 29 (IQR 21-52) days after infection, respectively. Plasma and PBMC specimens from the 30 male and one female transmitter were obtained a median of 22 (IQR 2-66) and 29 (IQR 2-94) days after the estimated date of HIV-infection of the recipient. 13 (42%) transmitters had 1-4 drug resistance mutations (total=25) detected at a median frequency of 6.0% (IQR 1.5- 98.6%, range 1.0-99.5%). 11 of these 25 mutations were present in the HIV-1 population of the transmitter at a frequency >95%; 100% of these majority variants were detected in the recipient. 14 mutations were present as minority variants in the transmitter, ranging from 1.0-11.8% of the viral population; 2 (14%, 95% CI 1.8-42.8%) were detected in PBMCs of the recipient at levels above 1%. If drug resistance does not impact “transmission fitness” and if drug resistant variants are transmitted independently, then the probability of transmission would be dependent on the frequency of the variant in the viral population of the transmitter. In this case, the likelihood that two or more of the minority variants would have been identified in recipient partners is 5.3%, which is less than our point estimate but falls within the 95% CI. Requiring further investigation, 10 recipients had 12 mutations identified in PBMCs at a median of 2.0% (IQR 1.3-4.0, range 1.1-99.6%) that were not detected in plasma or PBMCs of the transmitter.
Conclusions: Drug-resistant HIV-1 variants are more likely to be transmitted when present as majority variants compared to those present as minority variants. This study does not support crude population-level data that HIV-1 drug resistance reduces “transmission fitness” compared to drug-susceptible viruses.