Repeated HIV-1 resistance genotyping external quality assessments improve virology laboratory performance

The performance of French virology laboratories belonging to the ANRS network has been assessed annually for 3 years. The performance of these laboratories was compared between the years 2002 and 2003. Ten and 7 coded samples were sent to 38 virology laboratories in 2002 and 45 virology laboratories in 2003, respectively. Each panel of coded samples included at least one HIV-negative control, a pair of duplicate specimens, samples with a wide range of viral loads, and samples with a large number of resistance mutations. The laboratories used their standard sequencing procedures and were asked to report the amino acids at codons associated with resistance mutations, based on the IAS-USA expert panel list. The reference amino acid sequences were defined as those most frequently reported by the participants. The specificity of detection of RT mutations was significantly better in 2003 (99.9%) than in 2002 (99.7%) (P = 0.05). There was no difference between 2002 and 2003 in the specificity of detection of protease mutations (99.6% and 99.8%) or the sensitivity of detection of RT mutations (98.8% and 98.2%). The sensitivity of detection of protease mutations improved significantly between 2002 and 2003 (97.6% and 99.0%, respectively; P = 0.037). The proportion of laboratories reporting fully accurate results, in terms of amplification, specificity, sensitivity, and reproducibility, tended to increase between 2002 and 2003 (P = 0.077). No errors were made by 19% of laboratories in 2002, compared to 42% in 2003. These results show the value of repeated external quality assessments.