Background: Genotyping for resistance testing used to manage HIV-1 therapy is usually performed on a single patient (pt) sample. It has not been addressed whether a lab generates the same real time results on duplicate clinical samples.

Methods: The Virology Quality Assurance (VQA) Program sent out 3 panels over 18 months (mos) each containing 5 plasma samples to assess proficiency in sequence-based genotyping assays. Included were dilutions of 3 clinical samples each assayed 4 or 5 times over 3 panels at high (40,000–62,000) and low (~4,000) viral loads to evaluate sensitivity and consistency of data. Labs used the ViroSeq or TRUGENE genotyping systems or in-house protocols. A group consensus sequence was made for each panel specimen for each platform (GPCS). Data returned from 32–36 labs/panel were assessed for concordance (homology) to the GPCS for the protease (PR) and reverse transcriptase (RT) genes. Identification of mutations linked to resistance was assessed.

Results: All sequences returned had ≥ 98% concordance to their respective GPCS for the samples in all 3 panels showing all labs to be proficient in generating consistent sequence data. The 3 clinical samples respectively contained 3, 4, and 14 resistance mutations. For 2 samples, all groups identified 3/4 and 14/14 mutations in all the replicates. None of the codons for these mutations contained mixed bases. The 3^rd set of replicates yielded inconsistent inter- and intra- laboratory reporting of 2/3 mutations in the sample; both codons encoding mutations L63P in PR and M184V in RT had mixed bases. Non-consensus reporting (< 80%) did not correlate with sample viral load, genotype platform or testing laboratory. The incidence of non-consensus identification tended to be higher for groups using the ViroSeq system. The VQA began an analysis to identify factors in the performance of the assays which may influence variability in identifying mixed bases. The study suggests that several features of the assay could contribute collectively to variability in the data, including the choice of sequencer and subsequent sample load and software used, dilution of the sequencing template generated by PCR, as well as the performance of the personnel doing the assay.

Conclusions: Since much significance is placed on the presence of mixed populations in resistance testing, it is important to distinguish the influence of each step of the genotyping assay on the final result and report sent back to the physician.