Background: TORO 1 and TORO 2 are randomized, open-label, controlled, multi‑center, Phase III studies of patients (pts) receiving 90 mg BID of enfuvirtide (ENF, formerly T-20) by subcutaneous injection in combination with an optimized background (OB) regimen. Here we present for the first time resistance data determined with a newly developed single cycle entry inhibitor assay.

Methods: The intent to treat population included 661 pts randomized to ENF+OB. Resistance data were generated using the novel GeneSeq and PhenoSense Entry Assays on Env amplified from pt plasma samples to assess susceptibility to ENF (pEC₅₀). Combined data for TORO 1 and TORO 2 were analyzed using summary statistics. Multiple linear regression analysis identified virological factors associated with treatment-emergent (TE) change in pEC₅₀ (DpEC₅₀) and virological response.

Results: ENF pEC₅₀ were available for viral Envs from 612 (92.6%) pts at baseline (BL). By wk 24, 301 pts met protocol-defined virological failure and 204 of these had genotype and phenotype available for paired BL and failure time points. BL pEC₅₀s were approximately log-normally distributed with a geometric mean (GM) of 0.259 mg/mL (range 0.007–7.526 mg/mL) and a standard deviation (SD) of approximately 2.7-fold. Sixteen (16) of 612 pts (2.6%) had BL pEC₅₀s greater than 1.956 µg/mL (GM + 2 SD), but did not have a diminished virological response, nor were they predisposed to virological failure (p = 0.7186 and p = 0.4579, respectively). Neither BL viral tropism nor clade was associated with virological response. The GM pEC₅₀ for pts at failure was 5.645 mg/mL and DpEC₅₀ was 21-fold (range < 1–422-fold). There was a correlation with decreased ENF susceptibility at virological failure and changes in gp41 aa 36-45 (p < 0.0001). Varying DpEC₅₀ were observed for distinct TE substitutions in gp41 aa 36-45. Differences in BL ENF GM EC₅₀ between phase II and III studies may be explained by differences in the assays employed.

Conclusions: Using GeneSeq and PhenoSense Entry Assays, we characterized susceptibility to ENF at BL and at virological failure. Pts achieved similar virological suppression across the range of BL ENF susceptibility observed. Pts taking ENF who met virological failure criteria had on average a 21-fold loss of ENF susceptibility, associated with concomitant changes in gp41 aa 36-45.