252   A Quality Control Trial for HIV-1 Drug Resistance Testing Using Clinical Samples.

H. Walter, B. Schmidt, and K. Korn*.
Inst. of Clinical and Molecular Virology, Natl. Reference Ctr. for Retroviruses, Erlangen, Germany.

Background:Testing for HIV-1 drug resistance is becoming increasingly important for guiding therapy in HIV- infected patients. Despite the availability of commercial kits, drug resistance testing is a complicated, error-prone procedure. Checking the performance of laboratories carrying out such tests is therefore important.

Methods:A panel of 4 samples (3 HIV-1 positive, 1 HIV negative) was established and distributed to 24 laboratories. In contrast to previous quality control (QC) trials on HIV drug resistance testing, diluted plasma samples were used instead of isolates or clones. Samples had been characterized before distribution by direct sequencing and a recombinant virus assay. Participants were asked to report resistance-associated mutations, an interpretation concerning sensitivity to 14 antiviral drugs and information about the methods used.

Results:20 laboratories produced a total of 22 sets of genotypic data. The detection rate of key mutations associated with resistance to NRTI, NNRTI and PI was generally >90%. Lower reporting rates (15—75%) were observed for recently described resistance-associated mutations (e.g. RT V118I, PR N88S) and for "atypical" mutations at positions involved in drug resistance (e.g. RT K219R, PR L63N, G73T). Furthermore, 5 laboratories detected a minority species with a key mutation for PI resistance in one of the samples, which had not been evident in the pre-characterization. This is currently investigated further by clonal analysis and mutation-specific PCR. Compared to mutation detection, more heterogeneous results were observed when comparing the interpretation of mutations. Especially for PI, the same mutational pattern led to substantially different interpretations concerning sensitivity or resistance to the 5 currently licensed PI.

Conclusions:The results of this QC trial on HIV drug resistance testing show that the technical aspect of mutation detection is satisfactorily solved, with detection of minorities being the most difficult task. Problems arise mainly by the great variety of interpretation criteria, especially for PI resistance. Therefore, a generally accepted consensus for the interpretation of genotypic data on HIV drug resistance is needed.