Background: Detection of HIV-1 drug resistance mutations is complicated by the wide variation in plasma viral load, the need to isolate and reverse transcribe HIV-1 RNA prior to analysis, and the natural genetic diversity of HIV-1 viruses. The ViroSeq^TM HIV-1 Genotyping System (ViroSeq) and other population sequencing-based genotyping methods detect mutations present in the major viral population in a test sample. These assays also detect some mutations that are present at lower levels.

Methods: We compared detection of the K103N resistance mutation in subtype A, C and D HIV-1 using ViroSeq and a sensitive and quantitative point mutation assay, LigAmp, which measures the percentage of K103N-containing variants in the viral population (% K103N).

Results:  Both assays were used to analyze 305 samples collected from Ugandan and Malawian women 6-8 weeks after administration of single dose nevirapine (NVP): 146 with subtype A, 64 with subtype C, and 95 with subtype D. ViroSeq detected K103N in 100% of samples with >20% K103N, 77.8% of samples with 10-20% K103N, 71.4% of samples with 5-10% K103N, and 16.9% of samples with 1-5% K103N. The sensitivity of ViroSeq for detection of K103N was similar for subtypes A, C and D. Detection of K103N at levels between 1-20% in 107 (35.1%) of the 305 samples was unlikely to represent false positives in the LigAmp assay, since K103N was detected in only 1 of 238 available samples (0.4%) collected from these women prior to NVP administration.

Conclusions: The ViroSeq system reliably detects the K103N mutation at levels above 20%, and frequently detects the mutation at lower levels. Further studies are needed to compare the sensitivity of different assays for detection of HIV-1 drug resistance mutations, and to determine the clinical relevance of HIV-1 minority variants.