Explanation for failure of TMC-125 (etravirine) in TMC227 study

An oral presentation by Brian Woodfall provided an analysis of why Tibotecs TMC227 study had performed so poorly. [1] This trial was closed in November 2005 by the trial Data and Safety Monitoring Board (DSMB) because treatment-experienced patients in Thailand randomised to the new NNRTI etravirine (TMC-125) failed to show minimum -1 log virological response after 12 weeks. [2]

The study randomised patients failing their first-line NNRTI-based regimen and who were PI-naive, to either the new NNRTI (n=59) or an investigator chosen PI-based regimen (n-57; 61% lopinavir/r, 32% atazanavir/r). The study took place in Argentina, Brazil, Spain, Thailand and South Africa.

Resistance tests were used to guide treatment choices and the study protocol included using two investigator selected sensitive nucleosides in the new regimen.

Pooled data from all sites showed that an initial viral load response of -1.3 logs at week 8 in the NNRTI arm (compared to >2 log reductions in the PI group) was not sustained to week 12. In Thailand the rebound occurred before week 4, and occurred after week 8 in South Africa and after week 12 in Argentina and Spain. For patients reaching week 12, only 57% of the TMC-125 groups compared to 91% in the PI group achieved at least -1.0 log viral load decline.

The baseline resistance profiles of patients in the NNRTI arm were slightly more advanced compared to the PI group, but the significant differences in response was probably more related to the fragility or an unsupported NNRTI, compared to boosted PI regimens which even shows reasonable success rates when used as monotherapy in several small trials, especially in PI-naive patients. Although TMC-125 retains activity against NNRTI-resistant virus, it is clearly as vulnerable to rapid resistance when used as virtual monotherapy as existing drugs in this class.

At baseline, the median number of NNRTI mutations was 2 (range 0-4) and median fold change (FC) to TMC125 was 2.0. The median number of NRTI mutations was 1 (range 0-7). However 9% and 12% of subjects in TMC125 and control groups, respectively, did not receive two sensitive NRTIs. Only patients with less than 2 TAMS/M184V maintained viral suppression.

Patients with higher NNRTI and RTI-associated resistance had higher FC to TMC-125. Thai patients had more extensive RTI and NNRTI resistance, and this correlated with a higher FC phenotypic resistance to TMC-125 (5.6 vs </=2.0 in other countries). Increased resistance to RTIs tended to increase with increased NNRTI mutations.

In a multivariate analysis, increased numbers of NRTI and NNRTI mutations, use of inactive NRTIs, and higher TMC125 FC were all independently associated with virologic failure.

Comment

The failure of TMC-125 in this study seems to be convincingly explained by extensive RTI and NNRTI resistance, and recycling of RTIs. It also showed that accumulating NNRTI mutations can compromise chance of benefiting from pipeline NNRTIs.

This is particularly important in countries using NNRTI-based first line regimens where there is limited or no access to viral load monitoring, where patients are treated until clinical failure. Unfortunately, this the exactly the setting that would benefit most from a second generation NNRTI.

Other Phase III trials TMC125 in NNRTI and PI experienced patients are ongoing.