It is sad to see a pivotal trial fail, and you have to wonder why we recently had several such late-stage ‘mishaps’. Take Tetraphase’s eravacycline which failed in a Phase 3 cUTI trial. We have analyzed this drug’s failure in earlier blogs; there is reason to believe that inadequate dosing was the reason for the poor efficacy results. Now we hear that Chimerix has a similar disaster with brincidofovir/CMX001, its long-acting, kidney-sparing anti-CMV nucleoside: no difference over placebo was seen in the SUPPRESS trial which was conducted in HSCT patients at high risk of CMV reactivation in the post-transplant phase [1].
How can it be that in both instances earlier data suggested efficacy which could not be confirmed in a subsequent larger trial? Not surprisingly, explanations provided by the companies in press releases and at investor conference calls are often unrevealing.
In the case of brincidofovir/CMX001, a previous Phase 2 trial supposedly showed efficacy for the 100mg BIW regimen [2]. In light of the failed successor trial it is worth reviewing these data again. Table 1 shows the primary efficacy results from this study, a 230 patients, 5 dosing cohort trial. Note that only in the 100 mg twice weekly group a significant treatment benefit was observed but not in any other cohort (Figure 1). There is also no evidence of a dose response as the 2x greater dose of 200 mg twice weekly had no demonstrable efficacy (P=0.24).
Had we seen activity in more than 1 cohort, we would have been more convinced that brincidofovir is truly efficacious, even in the absence of a dose-response. In summary, the claim to efficacy rested on data from a single cohort, at best a weak indicator of potency as the higher dose group failed to show a similar benefit. Given this situation, Chimerix took a big risk by proceeding to a full-fledged Phase 3 trial without further corroboration of what could be a spurious result.
In addition, significant GI side effects (diarrhea) occurred already in the Phase 2 study with alarming frequency, raising concerns that this brincidofovir AE may be misinterpreted as a sign of early GVHD and trigger treatment with steroids which – in its wake – may lead to CMV reactivation and blunt any benefit that the antiviral may have.
The team realized this dilemma and developed a patient management plan recommending discontinuation of CMX001 in such cases before starting immunosuppressive anti-GVHD therapy. Still, in the Phase 3 trial, investigators were unable to clinically differentiate GI toxicity due to CMX001 from early GVHD and treated with steroids if in doubt.
Then there is the issue of timing of the primary efficacy endpoint. In the Phase 2 study this was within a week of ending prophylaxis, in the Phase 3 trial it was 10 weeks after ending prophylaxis. During this long drug-free follow-up period the prophylactic benefit of brincidofovir evaporated. Was it the FDA which insisted on this delayed efficacy timepoint? It strikes us as strange that the company would make such a dramatic change from the successful Phase 2 protocol knowing that recrudescence is likely to catch up in the CMX001 arm once suppressive antiviral prophylaxis has ended.
A strange dynamic is often created within teams at the end of Phase 2: the drive to move swiftly to the “pot of gold at the end of the rainbow” clouds sound analysis and “group-think” takes over. Call it a case of ‘irrational exuberance’, call it ‘wishful thinking’, the example of CMX001 should give pause for reflection.
CMV has already created a small graveyard for drug candidates out there: ViroPharma’s maribavir also failed in this indication (prevention of CMV recurrence post-HCT), and it so happened in a large Phase 3 trial. In that case suboptimal dosing was possibly the culprit.
References:
[1] http://files.shareholder.com/downloads/AMDA-1QNA05/1552591840x0x876376/47AE0320-F2E5-4F84-960B-3081B69DEB39/CMRX_Corporate_Update_Call_February_22_2016.pdf
[2] F Marty. CMX001 to Prevent Cytomegalovirus Disease in Hematopoietic-Cell Transplantation. N Engl J Med 2013;369:1227-36
