David's New Book

Monday, November 24, 2014

Will FDA Strike Again?

On December 4 and 5, the FDA is gong to hold back-to-back advisory committee meetings.  On Dec 4, the meeting will discuss clinical development approaches for antibiotics designed for patients with unmet medical need – in other words, those with serious infections caused by highly resistant bacterial pathogens.  On Dec 5, the committee will consider the submission by Cerexa/Forest/Actavis for ceftazidime-avibactam in the treatment of intra-abdominal, urinary tract and other infections where resistance is problematic.  These meetings could be of historical importance.  I won’t have the space to deal with the Dec 5 meeting preview here –the FDA has not yet released the background materials for that meeting – more later.

The FDA has already released background material for the December 4 meeting on the development of antibiotics for unmet needs. I have read this document and find myself, once again, disappointed. About half of the document is a historical review justifying the FDA approach to non-inferiority trials over the last several decades. They now say they are very satisfied with the outcome of their labors and are ready to move on to the next thing.  Unfortunately, while they may be satisfied with themselves, I think it is too early to pop the champagne. I can only conclude that this part of the backgrounder is more for FDA’s political watchdogs in congress than for anyone truly interested in bringing new antibiotics to the patients and physicians who need them. While I agree that the FDA has made significant progress in at least allowing the conduct of clinical trials that are actually feasible in the real world, the endpoints they have chosen for these trials still leave much to be desired. I won’t even begin to talk about the requirement for placebo-controlled trials for sinusitis and bronchitis – that NO ONE has undertaken since the edict came out around 2003. For skin infections, the endpoint remains a decrease in the size of the skin lesion at day 2-3 – not cure. For community-acquired pneumonia it remains symptomatic relief on day 4 or so – not cure. For hospital-acquired pneumonia it remains mortality even though about half of the mortality in this disease is completely unrelated to the infection being treated.  It should be cure of the infection. The situation is still much better and, in my view, more rational in Europe where cure of the infection remains the relevant endpoint for all these infections.

But the most worrisome section of the backgrounder is that dealing with the development of new agents targeting resistant pathogens – the unmet needs population. A few highlights (lowlights) that caught my attention are noted below.

1.     A superiority trial nested within a non-inferiority trial. Actually – this may be a highlight – I’m not sure.  We’ll have to hear the discussion.  But the last time I was involved (indirectly) with the agency on just such a design, we had designated a subgroup of the overall population, those with resistant infections, as an analysis population. Here we were going to look at failure rates for the control antibiotic and the new drug.  We carefully provided for patient safety by excluding patients with the most serious infections from study entry and by allowing a rapid switch either to the alternate regime (blinded) or exit from the study to receive another standard of care antibiotic. The FDA balked at the design because they didn’t like the subgroup analysis portion.  But the backgrounder now indicates that they would accept such an analysis.  Which FDA is which?
2.     For superiority trials, they still seem to be laboring under the illusion that head to head trials are feasible.  They’re not.  See my blog on superiority trials.
3.     As I noted in that blog on superiority designs, externally controlled trials would be the most feasible. But the FDA seems at a loss as to how to actually design such trials. They are harkening back to mortality rates as a preferred endpoint in externally controlled superiority trials – again – for the most part – completely infeasible especially in the circumstance they are discussing – multiple body site infection trials. The FDA seems to recognize the infeasibility of this approach, but they go on to flounder around about other possible approaches.  I’m worried that they don’t know what the endpoints should be at the different body sites.  (Hint – Cure)!!
a.     They suggest the possibility of an externally controlled non-inferiority design where you could actually show superiority.  This at the very least will increase trial numbers and expense at a time when we want to do the opposite and at worst is simply infeasible.
b.     They note that there may be disparate outcomes in different body sites and they give the examples of tigecycline and doripenem in ventilator-associated pneumonia as examples. Clearly they are correct – but good human PK/PD work prior to an efficacy trial should be able to deal with this.  Also, as the FDA suggests, some stratification may be necessary – but this would have to be done in a way that trial size remains small.
4.     They also discuss single pathogen trials. Here they seem to rely on rapid diagnostic approaches that (1) do not really exist and (2) when they are “available” are not sufficiently rapid. (See my blog on rapid diagnostics). Clearly, the FDA is still lost here.
5.     Finally, I will note that in several areas they mention the fatal words “further work.” The last time they said that, related to trials in skin infections, the antibiotic pipeline was set back at least 5 years.

I look forward to listening in on this meeting.  I can only hope that the AIDAC will see some of the same issues that I see and that they will help sort this out – but my expectations are low.

Now, it just remains for me to wish all of you in the US a most happy Thanksgiving holiday.  I’ll probably come back to you after the AIDAC meetings with summaries.