Non-Traditional Antibacterials

We had our webinar-Clinical development for non-developers Part 3: Non-traditional Approaches yesterday.  I encourage everyone interested in this topic to listen.  I moderated, but the front line participants were 

Sumathi Nambiar

Director, Division of Anti-Infective Products

Office of Antimicrobial Products, CDER, FDA

Mair Powell

Senior Clinical Assessor

Health Products Regulatory Authority (Ireland) 

Ian Friedland

Clinical consultant

Friedland Strategic Consulting, LLC.

You couldn’t have a more authoritative and experienced set of presenters. 

What are “non-traditional” antibacterials? Based on our discussion yesterday, I think its better to think about what the products will do and how they will be developed.  Any protein, antibody, peptide, bacteriophage cocktail or other that can be developed and used as a stand-alone therapy can be studied clinically in a traditional indication such as ventilator-associated pneumonia, urinary tract infection, etc., using a non-inferiority trial design. This is by far the lowest risk way to proceed to demonstrate efficacy.  For these products, CMC and safety studies that will be required are quite different than those required for small molecules. 

For those products that cannot be studied as stand-alone antibacterials, such as inhibitors of virulence, critical but non-essential bacterial functions such as RecA for example, immunomodulators that might target host functions, therapies directed at the microbiome and others are more challenging to develop. One main problem will be coming up with a rationale to justify dose selection. Will you be able to use standard animal models of infection to carry out PK/PD studies to justify dose or not?  If not, you may need to carry out dose ranging trials in the clinic before proceeding to pivotal trials. A second major decision will be to determine whether the drug will be developed as part of a fixed-dose combination with a known antibiotic or whether it will be adjunctive therapy to standard of care. If it will be developed as a fixed-dose combination, you will need a strong rationale showing a dramatic effect of the new drug on its partner’s activity to justify this approach. If it will be developed as adjunctive therapy to standard-of-care, you will ultimately need to carry out superiority trials. In the latter case, the challenge to show superiority will be a significant one and not one to be undertaken lightly. 

In thinking about this, I have come to understand that there is also another issue – one that John Rex has mentioned frequently. That is, the value proposition. Regulatory requirements to achieve market approval are one set of challenges. But a plan to show the value of the new product to patients, physicians, payers and investors is another.  I think of this as a Venn diagram.  Demonstrating the value of the product to this important group of stakeholders is as important as obtaining market approval from the regulators. While non-inferiority trials for these products (when developed as fixed dose combinations) may allow for regulatory approval, they may not be sufficient to fulfil a value proposition.  Positive data from superiority trials are more likely to fulfil a value proposition, but the chance of success in these trials as currently envisioned is low.

Everyone participating in the webinar agreed that this is new territory. We do not have any real experience in developing or using non-stand-alone products like these. From the regulatory point of view, much will have to be decided on a case-by-case basis until we get a better understanding of how to develop these products. The regulators on both sides of the Atlantic are willing and even eager to hear from sponsors delving into this area.  They are open to listening to project proposals and providing advice even at the earliest stages of scientific investigation. I would encourage those of you working in this area to take advantage of this possibility before investing significant resources in product development. 

Dr. Nambiar has provided a number of links in her presentation materials to prior FDA workshops and advisory committee meetings that have examined this and related areas.

A few bottom line items from the webinar, at least for me, are:

Non-traditional products that cannot be stand-alone therapy remain a very high-risk endeavour. 

Any company working in this area should have a portfolio of projects that spread risk across more than one approach.

Have a Target Product Profile that is developed based on medical need and on a value proposition. Use experienced clinicians and scientists here. 

Speak to regulators early on about CMC, safety and efficacy.

Regulatory Conundrums

In the absence of any real news on the pull incentive front, I want to turn my attention to the regulatory side of antibiotic development. Before we get started, I want to express my thanks to John Rex who provided useful contributions to this blog.

 As I have been saying for years, for most new antibiotics, regulatory requirements are not a limiting factor. If a new antibiotic has a relatively broad spectrum or is specifically active against a very common pathogen such as Staphylococcus aureus or Neisseria gonorrhea, there are clear and feasible regulatory pathways allowing for development and approval on both sides of the Atlantic and globally. 

The remaining regulatory difficulty comes if you are trying to develop a product that is specific for a relatively uncommon or rare pathogen such as Pseudomonas aeruginosa or Acinetobacter baumannii. I have written about this previously and there have been many FDA meetings and workshops where this issue has been discussed. The conundrum as noted by Dr. Rex, is that patients, clinicians, payers and yes, regulators all want a statistically powered trial. But for these rare infections, it remains unclear that such a trial is feasible. 

Another workshop, this time to discuss the development of products for non-tuberculous mycobacterial infections, will occur on April 8.  The issue for this entire area revolves around the design of clinical trials where the enrollable population is small.  In these instances, enrolling a large, non-inferiority trial may simply be impossible or may take many years. Both FDA and EMA have struggled to provide guidance dealing with this problem. Both continue to reject the use of external or historical controls. These controls have their problems but could significantly cut the number of patients one would have to study in an active trial. 

There are at least two pathogen-specific antibiotics currently in Phase 3 trials. Murepavadin is an anti-pseudomonal drug being developed by Polyphor.  They are carrying out trails in hospital-acquired and ventilator-associated pneumonia. About 15-20% of these infections are caused by Pseudomonas. They plan to enroll about 250 patients in each of two non-inferiority trials. The feasibility of this approach remains to be seen. Entasis is studying a combination of a B-lactamase inhibitor with sulbactam for the treatment of urinary tract infection caused by Acinetobacter. In the latter case, in a phase 2 trial, patients received ExtSul plus imipenem.  Of 80 patients, 8 had imipenem-resistant isolates.  Will they be able to recruit enough of these patients and show enough of a difference in efficacy to gain approval? 

John Rex notes that, on the one hand, we do not want to have so many resistant infections that the Entasis trial enrolls easily.  On the other hand, we want to be able to approve and market drugs such as the one under study by Entasis. This topic is full of paradoxes. 

Many of the more innovative approaches currently being pursued are non-traditional in nature. These include anti-virulence, lysins, immunomodulators, bacteriophage and other targets.  Many of the resulting products will also be specific for certain uncommon pathogens. As John Rex often points out, all these approaches, traditional or not, must show to patients, clinicians and to regulators that the product under study has value to patients with unmet medical needs. Designing acceptable clinical trials that will accomplish this, especially for those products that will provide adjunctive therapy is, to say the least, challenging. 

Further, the FDA and EMA have rather divergent approaches to this topic. EMA, for example, has just published a new addendum for the development of antibacterials. They make it clear that they expect randomized clinical trial data.  Helpfully, EMA notes that the size of the data package will be based on trial feasibility including the required length of time for enrollment.  But this may conflict with criteria that FDA is using to determine trial size and makes one question the kind of statistical information EMA is willing to accept. EMA also notes that it does not consider bacteremia outside of well-documented secondary bacteremia, a clinical indication or even a disease. FDA, however, has accepted bacteremia in recent submissions. 

These apparent conflicting approaches simply signify that neither we as developers nor the regulators have come to any consensus on how to deal with antibacterial drugs targeting infections where the patient population is small.  But in order to move forward with the innovative products that are now being researched, this is a problem we must all work harder to solve. 

I am going to suggest that, as part of any legislation providing for a significant pull incentive, we insist on the inclusion of a mandate for FDA to identify feasible development pathways for products targeting small populations. Many of us thought this was already accomplished with the LPAD guidance from FDA – but in fact the guidance specifically states that there must be “substantial evidence of effectiveness.” This seems to virtually preclude small trial designs that might be feasible.

Obviously, we all have a great deal more work to do to make the global development of antibacterial drugs specific for uncommon or rare infections feasible. 

NOTE - To clarify - both FDA and EMA allow the study of bloodstream infections for products meeting an unmet medical need.  But, the label indication will not include "bacteremia" as I currently understand how this works.