ABSTRACT
BACKGROUND: We developed clinically relevant genotypic scores for resistance to fosamprenavir/ritonavir in HIV-1 protease inhibitor (PI)-experienced patients. METHODS: PI-experienced patients with virological failure receiving fosamprenavir/ritonavir as the sole PI for at least 3 months and with detectable fosamprenavir plasma levels were included. The impact of baseline protease mutations on virological response (VR, i.e. decrease in plasma HIV-1 RNA between baseline and month 3) was analysed using the Mann-Whitney test. Mutations with prevalence >10% and P value <0.10 were retained. The Jonckheere-Terpstra test was used to select the combination of mutations most strongly associated with VR. The association between score and VR was assessed by multivariate backward regression. RESULTS: In the 73 patients included, the median baseline HIV-1 RNA was 4.6 log(10) copies/mL (range: 2.7-6.9) and the mean decrease at month 3 was -1.07 +/- 1.40 log(10) copies/mL. Ninety per cent of the patients were infected by HIV-1 subtype B variants. Two fosamprenavir/ritonavir mutation scores were constructed: score A (L10F/I/V + L33F + M36I + I54L/M/V/A/T/S + I62V + V82A/F/C/G + I84V + L90M) was based only on mutations associated with a worse VR, whereas score B (L10FIV + L33F + M36I + I54L/M/V/A/T/S + A71V - V77I - N88S + L90M) also took into account favourable mutations. Both scores were independent predictors of VR, however, co-administration of tenofovir was associated with a worse VR and the presence of the N88S protease mutation and co-administration of enfuvirtide with a better VR. CONCLUSIONS: These clinically validated mutation scores should be of interest for the clinical management of PI-experienced patients. The fosamprenavir/ritonavir score A was introduced in the 2006 ANRS algorithm along with isolated mutations I50V and V32I + I47V.