Background:  The POWER 1, 2, and 3 trials demonstrated high efficacy of darunavir/ritonavir (DRV/r) in treatment-experienced patients. We defined the upper and lower clinical cut-offs for DRV/r within these trials datasets by evaluating week-4 HIV RNA outcomes.

Methods:  Phenotyping of a set (n = 184) of baseline samples was by PhenoSense HIV (Monogram Biosciences). This dataset was restricted to patients in whom the on-study regimen did not include enfuvirtide (T-20). The lower clinical cut-off was defined as the fold change where the HIV RNA response was first observed to decline relative to the wild type reference population. The upper clinical cut-off was defined as the fold change above which the attributable HIV RNA change from baseline was less than –0.3 log₁₀ copies/mL. The effect of the on-study background therapy was explored by deriving PhenoSense specific continuous phenotypic susceptibility scores for the drugs in each regimen. Linear regression and local linear fitting by the function lowess were used to define the optimal correlation between the baseline DRV fold change and the week-4 change in HIV RNA (log₁₀ copies/mL).

Results:  Among the 184 analyzed subjects, 87 received functional DRV/r monotherapy, ie, their optimized background regimen was scored as phenotypic susceptibility scores = 0. Within this subset the baseline DRV fold change correlated with the week-4 HIV RNA change from baseline (R = 0.42, p <0.0001). From the lowest DRV fold change (0.42) to fold change 10 the week-4 HIV RNA reductions were constant (median –1.99 log₁₀ copies/mL). Thus, the lower clinical cut-off was defined as a fold change of 10. For DRV fold change >10 a gradual reduction in HIV RNA response was observed as fold change increased (R = 0.37, p = 0.015, median week-4 HIV RNA change –1.11 log₁₀ copies/mL). Within this distribution the upper clinical cut-off for DRV was defined as a fold change of 90. The median HIV RNA change for samples with fold change 10 to <40 and 40 to <90 was –1.34 (–2.66 to 0.15) and –0.39 (–2.93 to 0.28) log₁₀ copies/mL, respectively. Analyses at study weeks 2 and 8 were concordant with the week-4 findings. Applying these clinical cut-offs to the 184 phenotyped baseline samples, the proportions fully susceptible were: DRV/r, 53.8%; tipranavir (TPV)/r, 33.7%; amprenavir (APV)/r, 18.5%; atazanavir (ATV)/r, 14.6%; lopinavir (LPV)/r, 11.4%, and saquinavir (SQV)/r, 9.8%. By contrast the proportions above the upper clinical cut-offs were:  DRV/r, 3%; TPV/r, 31%; APV/r, 66.3%; LPV/r, 70%, and SQV/r, 71.7%.

Conclusions:  In the POWER trials, DRV/r demonstrated optimal anti-HIV activity with a fold change £10 and reduced anti-HIV activity with fold change >10. These analyses defined the phenotypic lower and upper clinical cut-offs for DRV/r at 10-fold change and 90-fold change, respectively.