Plasma levels of HIV protease inhibitors (PI) are often close to IC50 values of wild-type viruses when administered without ritonavir boosting. The impact of drug levels, resistance mutations, and the genotypic inhibitory quotient (GIQ) were examined on the response to saquinavir/ritonavir (SQV/r)-based salvage therapy. Patients with prior exposure to PI other than SQV and currently failing virologically were recruited prospectively in a multicenter trial. All patients began SQV/r 1000/100 mg bid plus another two antiretrovirals. A total of 139 patients was enrolled. At month 12, virological response (VR), defined as plasma HIV-RNA decline >1 log, was recorded in 68.2% of patients on treatment (41.7% in the intent-to-treat analysis). The median baseline number of protease resistance mutations was three. The VR was significantly lower in patients with >5 protease resistance mutations and in those with plasma SQV Cmin<0.1 microg/ml. At week 48, the VR was seen in 77.1% of patients with a GIQ>0.04 compared to 18.2% of those with a lower GIQ (P=0.001). In the multivariate analysis, <or=5 protease resistance mutations and SQV Cmin>0.1 microg/ml were independently associated with VR. Interestingly, drug levels had their highest predictive value of the VR at week 12, since low SQV plasma levels often permitted ruling out poorly adherent patients. In contrast, the number of protease resistance mutations had the highest impact on the VR at week 24, suggesting that for those taking the drugs, the VR is highly dependent of the presence of resistance mutations. At any time, nevertheless, the GIQ provided the most accurate prediction of the VR.
Copyright (c) 2005 Wiley-Liss, inc.