Background: Retrospective and prospective studies have demonstrated the prognostic value and clinical utility of HIV resistance tests. Nevertheless, the virological benefits remain modest, and short lived. We tested the hypothesis that data from sequential tests may improve the prognostic significance of resistance assays.

Methods: We studied 100 highly antiretroviral experienced patients (median no. drugs received = 6), in whom 2 genotypic resistance assays had been undertaken, at least 6 months apart, with a subsequent therapy change within 1 month of the later test, and who underwent at least 6 months of subsequent follow-up. A modified DAP score, representing the number of “active” drugs (to which the virus was fully susceptible) prescribed at this therapy change was applied, utilising genotypic information obtained from the later test alone, and by combining the mutations observed in the 2 resistance assays.

Results: Using either method, the DAP score was related to the mean maximal viral load decline within 6 months of the later test (later test alone, p<0.001; cumulative tests, p=0.026: one way ANOVA). However, in addition, the maximum viral load decreases observed for DAP scores generated from the 2 tests were consistently greater than for scores based on the later test alone. This difference reached statistical significance for a score of 2 (-2.88 log₁₀[95% CI-3.48 to -2.28] vs -1.88log₁₀ [95% CI –2.22 to –1.54], respectively). The drivers for this difference appeared to be changes in  3TC and nNRTI resistance mutations between tests.

Conclusions: These results provide evidence that virus species which emerge on therapy but subsequently become undetectable in plasma virus may still contribute to therapy failure. In highly antiretroviral experienced patients, consideration of multiple resistance assays may enhance the clinical utility of these tests.