Background: Resistance genotyping is often requested due to residual HIV viraemia or for treatment optimization, but may be unsuccessful if plasma RNA levels are too low or undetectable. Analyses of proviral HIV-DNA can provide information about the viral reservoir, because integrated DNA reflects both actively and latently infected cells.
Objectives: To determine whether proviral DNA is a potential relevant alternative to HIV-RNA for resistance genotyping in this context.
Methods: The resistance mutations harboured by the proviral DNA were compared with the cumulative data for all plasma RNA genotypes previously obtained for the patient concerned. We also investigated whether various parameters, such as CD4 count, level of viraemia or drug pressure, affected the results.
Results: We collected 134 and 141 DNA genotypes with 443 and 462 corresponding RNA sequences for the reverse transcriptase and protease genes, respectively. The mean rates of concordance between DNA and RNA genotypes were 46.7% for nucleoside reverse transcriptase inhibitors (NRTIs), 26.3% for non-NRTIs and 43.7% for protease inhibitors (PIs). Mixtures were detected for most DNA mutations. The rate of concordant PI mutations was significantly higher for patients taking PIs at the time of DNA genotyping (48% versus 26%; P=0.004). The other factors studied had no impact.
Conclusions: In the context of low or undetectable viraemia, it is difficult to reach the archived mutated DNA. Classical RNA genotyping during previous periods of virological failure remains the gold standard for documenting resistance mutations and for the monitoring of future treatments.