Simple algorithm derived from a geno-/phenotypic database to predict HIV-1protease inhibitor resistance

Background: Resistance against protease inhibitors (PI) can either be analy sed genotypically or phenotypically. However, the interpretation of genotyp ic data is difficult, particularly for PI, because of the unknown contribut ions of several mutations to resistance and cross-resistance. Objective: Development of an algorithm to predict PI phenotype from genotyp ic data. Methods: Recombinant viruses containing patient-derived protease genes were analysed for sensitivity to indinavir, saquinavir, ritonavir and nelfinavi r. Drug resistance-associated mutations were determined by direct sequencin g, geno- and phenotypic data were compared for 119 samples from 97 HIV-1 in fected patients. Results: Samples with one or two mutations in the gene for the protease wer e phenotypically sensitive in 74.3%, whereas 83.6% of samples with five or more mutations were resistant against all PI tested. Some mutations (361, 6 3P, 71V/T, 771) were frequent both in sensitive and resistant samples, wher eas others (241, 30N, 461/L, 48V, 54V, 82A/F/T/S, 84V, 90M) were predominan tly present in resistant samples. Therefore, the presence or absence of a s ingle drug resistance-associated mutation predicted phenotypic PI resistanc e with high sensitivity (96.5-100%) but low specificity (13.3-57.4%). A mor e specific algorithm was obtained by taking into account the total number o f drug resistance-associated mutations in the gene for the protease and res tricting these to certain key positions for the PI. The algorithm was subse quently validated by analysis of 72 independent samples. Conclusion: With an optimized algorithm, phenotypic PI resistance can be pr edicted by viral genotype with good sensitivity (89.1-93.0%) and specificit y (82.6-93.3%). The reliability and relevance of this algorithm should be f urther evaluated in clinical practice. (C) 2000 Lippincott Williams & Wilki ns.