Longitudinal use of phenotypic resistance testing to HIV-1 protease inhibitors in patients developing HAART failure.

An "in-house" recombinant virus protease inhibitor susceptibility assay was carried out (median of 3 per patient) retrospectively in 26 patients failing HIV protease inhibitor based therapy at regular intervals to the initiation of the first protease inhibitor. Patients were treated with either indinavir (N=6), ritonavir (N=10), or saquinavir (N=10) and two nucleoside analogues. Second line therapy was based on single or dual protease inhibitor regimens occasionally containing nelfinavir. Clinically relevant resistance cut-offs associated with a poorer virological outcome from 6 months on and the clinical outcome from 3 months on were determined tentatively as 4- to 8-fold resistance for indinavir and ritonavir and 2.5- to 8-fold to saquinavir. In addition, the degree of cross-resistance at the time of the change of protease inhibitor was associated with the response in viral load at 6 months to the second line therapy (P=0.018). Cross-resistance (>=8-fold) between ritonavir and indinavir was common (78 and 100%). Cross-resistance between indinavir or ritonavir and saquinavir was less frequent (75 and 60% respectively) than the opposite (100%, P=0.004). Cross-resistance to nelfinavir was encountered more frequently (> 70%) than to amprenavir (9%). The magnitudes of resistance were correlated between each other. In summary, the protease inhibitor susceptibility carried out longitudinally appears to be an earlier prognostic marker than viral load in a context of cross-resistance. The magnitude of resistance, as a marker of cross-resistance, should be useful to guide second line therapy. (C) 2002 Wiley-Liss, Inc.