Background - Studies have shown that low-frequency resistance mutations can influence treatment outcome. However, the lack of a standardized high-throughput assay has precluded their detection in clinical settings. Objective - To evaluate the performance of the Roche prototype 454 UDS HIV-1 drug resistance assay (UDS assay) in a routine diagnostic laboratory. Study design - 50 plasma samples, previously characterized by population sequencing and that had shown ≥1 resistance associated mutation (RAM), were retrospectively tested by the UDS assay, including 18 B and 32 non-B subtypes; viral loads between 114–1,806,407 cp/ml; drug-naive (n = 27) and drug-experienced (n = 23) individuals. Results - The UDS assay was successful for 37/50 (74%) samples. It detected all RAMs found by population sequencing at frequencies above 20%. In addition, 39 low-frequency RAMs were exclusively detected by the UDS assay at frequencies below 20% in both drug-naïve (19/26, 73%) and drug-experienced (9/18, 50%) individuals. UDS results would lead to changes from susceptible to resistant to efavirenz (EFV) in one drug-naive individual with suboptimal response to an EFV-containing regimen and from susceptible to resistance to lamivudine (3TC) in one drug naïve subject who subsequently failed a 3TC-containing regimen and in a treatment experienced subject who had failed a 3TC-containing regimen. Conclusions - The UDS assay performed well across a wide range of subtypes and viral loads; it showed perfect agreement with population sequencing for all RAMs analyzed. In addition, the UDS assay detected additional mutations at frequencies below 20% which correlate with patients’ treatment history and had in some cases important prognostic implications. |