Background: Studies have suggested that HIV-1 may develop thymidine analogue mutations (TAMs) by one of two distinct pathways - the TAM1 pathway (including mutations 41L, 210W and 215Y) or the TAM2 pathway (including mutations 67N, 70R and 219E/Q) - under the pressure of a not fully suppressive thymidine-analogue-containing regimen.
Methods: Frozen plasma samples stored in the EuroSIDA repository were selected and sent to two central laboratories for genotypic analysis. We considered 733 patients with at least one genotypic test showing > or =1 TAMs (the first of these tests in chronological order was used). TAM1 and TAM2 genotypic profiles were defined in accordance with previous literature. Statistical modelling involved logistic regression and linear regression analysis for censored data.
Results: The observed frequencies of patterns classifiable as TAM1 or TAM2 profiles were markedly higher than the probabilities of falling into these classifications by chance alone. The chance of detecting a TAM2 profile increased by 25% per additional year of exposure to zidovudine. We found that mutations 67N and 184V were not associated with a particular TAM profile. In the presence of TAM2 profiles, the adjusted mean difference in the 6-month viral reduction was 0.96 log10 copies/ml (95% confidence interval: 0.20; 1.73) higher in patients who started stavudine-containing regimens instead of zidovudine-containing regimens.
Conclusions: This study provides evidence that the suggested TAM clustering is a real phenomenon and that it may be driven by which thymidine analogue the patients has used. In patients with TAM2-resistant viruses, stavudine appears to retain greater viral activity than zidovudine.