FDA AMDAC – Developing Single-Pathogen agents for P. aeruginosa and A. baumannii

Dear All (and with apologies … long note alert!):

Yesterday’s AMDAC was very productive. For reference, all the materials are here and a reasonably detailed set of notes are found below my signature. For further color, David Shlaes has also posted a summary on his Perfect Storm blog.

FDA nicely positioned the question for the day as “It will be easy to approval a new narrow-spectrum agent if clean non-inferiority or superiority data are possible, but we think it likely that this will be the exception rather than the rule. If we are presented instead with a product that seems to have all the right stuff preclinically but where the best available clinical data are unclear, how should we proceed?”

In response, my core takeaway is that the committee overall (some dissent, some nuances) recognized:

1. the pressing public health threat,
2. the potential role that narrow-spectrum agents might play for selected pathogens,
3. the challenges in predictably generating typical amounts of data, and
4. the need to cautiously permit a pathway built on lots of animal model data, lots of human PK data, a good understanding of safety, and at least some clinical efficacy data.

But, also coming through loudly and clearly was the message that this is really a pathway of last resort … every effort should be made to generate the best possible clinical data. Put in Tier A-B-C-D language (Lancet ID 13:269-75, 2013), approval could be based only Tier C-quality data if you truly have no choice … but, please, please, PLEASE get as close to the quality of data expected in Tier B as you can.

All of this makes good sense to me and I think this is a great outcome for the day – perhaps not a home run but certainly a triple if you’ll permit a baseball metaphor. I especially liked the balance around the idea that “you can rely more heavily on animal models + human PK if you really have no other options” vs. “If you bring a rock-bottom clinical package, the drug will be viewed skeptically … it is in everyone’s interest if you do more.”

Let me close by thanking our colleagues at FDA for the many hours of effort that went into bringing us to yesterday’s AMDAC. The two prior workshops have definitely moved the conversation forward and I can see how this could become a guidance document in short order. Exciting!

And finally, the topic of forward conversations brings up this list of events (in addition of course to ECCMID, etc.). I hope to see you at some or all of:

• NIAID’s PKD-PD workshop (14-15 June)
• The ASM-ESCMID conference (6-8 Sep) with its associated 5 Sep CARBX-GARDP Antibiotic Bootcamp
• FDA/CDRH’s 13 Sep workshop on Antimicrobial Susceptibility and Resistance Testing Devices

All best wishes and safe travels,

John

John H. Rex, MD | Chief Medical Officer, F2G Ltd. | Chief Strategy Officer, CARB-X | Expert-in-Residence, Wellcome Trust. Follow me on Twitter: @JohnRex_NewAbx

== Detailed notes ==
The day begain with these presentations:

1. FDA provided a thorough review of the regulatory background (Ed Cox, Sumati Nambiar) and the two prior workshops (Yuliya Yasinskaya)
   1. 18-18 July 2016: https://www.fda.gov/Drugs/NewsEvents/ucm497650.htm
   2. 1 Mar 2017: https://www.fda.gov/Drugs/NewsEvents/ucm534031.htm
2. Two case studies
   1. Entasis (Robin Isaacs) presented the strategy for their anti-Acinetobacter compound
   2. FDA (Peter Kim) presented a summary of the currrent strategy for Polyphor’s anti-Pseudomonas compound
   3. Details on these two can be found in the 1 Mar 2017 materials online – the presentations given today were largely identical to those given in March
3. IDSA (Trish Perl-DeLisle) presented the IDSA view
   1. New drugs are desperately needed; ID clinicians can handle uncertainty; We must find a way!
   2. IDSA noted there is an in press paper at JID that provides further details (disclaimer: I am a co-author on this)
4. Open Public Hearing
   1. There were 3 presentations that focused principally on critiques of non-inferiority. Their core themes were that (a) new drugs should be superior and (b) non-inferiority studies are unethical. On (a), the speakers failed to understand that new antibiotics offer superior microbiology that can’t be shown directly. On (b), the speakers failed to understand that merely having a superiority hypothesis doesn’t mean that the drug will be so. The confusion on these points is so long-standing and frustrating! As at least a partial further response, a group of us have a paper in press at CID in which we explore these issues at greater length – I’ll share it in a future note.
   2. I was the 4th OPH speaker and my comments were that the crisis that’s upon us mandates work to enable pathways for needed agents. Narrowing the discovery focus to a single species can be helpful and we simply must have paths for the two real examples before us now (Entasis, Polyphor). I closed by urging the committee to endorse the idea of approvals built on animal model data + human PK + safety data + best possible clinical data. In short, I favored sins of Commission over sins of Omission (so to speak).

There were no votes taken but rather the meeting ended with the committee providing verbal commentary on two questions (paraphrased):

1. What are you views on the ideas presented for program progression for narrow-spectrum Pae or Abau drugs (NI, superiority, classical Animal Rule, microbiology as surrogate, PK as surrogate)?
2. What should FDA do if, despite all best efforts to generate high quality data, the clinical trial results are murky due to such confounding factors as prior therapy, comorbidities, etc.?

The discussion prior to commenting on the questions was wide-ranging and the committee need to do some exploratory thinking via its questions. This was perhaps not surprising — this group is mostly new to the specific issues of antibacterial agents. When it came time to comment on the questions, these themes were dominant:

1. The proposed agents are valuable and should progress!
   1. We are facing a clear public health threat
   2. It’s not that narrow-spectrum per se is a goal, but we need a predictable path forward when that’s what you find
2. FDA has done a thorough job of surveying the possibilities for progressing narrow-spectrum agents for Pae or Abau.
   1. None of the possibilities is perfect
   2. There’s no alternative approach that is obviously missing
   3. The Animal Rule per se is not a great fit for these pathogens as (i) at least some clinical data can be generated and (ii) these pathogens lack the explosive virulence that gives such clarity with Yersinia, etc.
3. Superiority would be lovely but should not be expected
   1. Indeed, Lynn Marks hit the nail on the head when he reminded the committee that we should work hard to have drugs available so that superiority is never possible.
4. Non-inferiority is perfectly fine as a tool
   1. See comments above regarding the negative comments about NI from the first 3 OPH speakers. Committee comments often focused on correcting their misunderstandings.
5. But, sometimes it will not be possible to do a clean NI or superiority study
   1. When this is true, a combination of these elements should be used to (cautiously) support (a limited form of) approval
      
      1. exhaustive and varied animal models PLUS
         
      2. careful documentation of adequate PK in man PLUS
         
      3. adequate characterization of the safety profile PLUS
         
      4. at least some clinical data
         
   2. The committee made it clear that such an approval would be viewed with a critical eye
      
      1. It would be considered very much as second-line therapy
         
      2. There was strong encouragement to seek post-approval data by ensuring release of the drug was monitored
         
      3. There was strong encouragement for clinical trial networks to support such studies
         
      4. There was a strong wish for use of some sort of conditional approval mechanism
         
         1. As an aside and on the point of conditional approval, note that this is something available to EMA but FDA does NOT have this specific tool in its regulatory/legal toolkit. FDA’s accelerated approval tool is a different thing … it’s not a conditional approval but rather an approval grounded in a surrogate measure. All this said, FDA does (at least to my eye) can approximate the idea of a conditional mechanism via post-marketing commitments. Further materials on this nuance are available online:
            1. An FDA presentation that I think is from approximately 2006: https://www.fda.gov/downloads/AboutFDA/CentersOffices/CDER/ucm119996.pdf
            2. FDA’s Expedited program guidance: https://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm358301.pdf
            3. Van Norman 2016: http://www.sciencedirect.com/science/article/pii/S2452302X16300638
               
      5. In summary, and for avoidance of doubt, this path should only be pursued as a last resort. Every effort should be made to produce strong & clear clinical data.