VABP is clearly the main indication to be pursued by a drug like POL7080. Here is the question: How to conduct a study demonstrating efficacy for a drug which only has a single-organism spectrum? Which is actually no spectrum at all. Existing guidelines for the development of ID drugs are indication-driven, with the notable exception of MTB and malaria agents. One can study an MRSA drug in ABSSSI, but there is no equivalent indication for P. aeruginosa.
Finding sufficient numbers of the “right” patients for efficacy determinations will be incredibly difficult as long as we don’t have a rapid bedside diagnostics to prove the presence of P. aeruginosa in a VABP patient. Enrolling VABP patients indiscriminately before culture results become available means enrolling a patient population where 70- 80% will not have the target organism, i.e., P. aeruginosa, in their respiratory secretions. As the majority of VABP patients will harbor other pathogens (like S. aureus or another gram-negative), most randomized patients will not contribute to address the central efficacy question. Dropping patients with negative cultures for P.aeruginosa from the study creates havoc with the randomization and is anathema to statisticians (and the FDA) but maybe the only feasible way to deal with the situation. The current VABP Guidelines are very strict and FDA really does not care much about feasibility, to put it mildly. Following the FDA path for VABP is an option, but not a realistic one.
There is another big problem: Almost all ICU patients are started on antibiotics right after intubation and long before the onset of pneumonia. Hence, ventilated ICU patients are not “antibiotic-naïve”. This prophylaxis often includes antipseudomonas drugs. While not condoned by current ATS/IDSA guidelines it is current practice nonetheless and happens all over the world. Hence, only pretreated ‘break-through’ cases will become available for study.
Some other feasibility issues cannot be adequately addressed here but should at least be mentioned briefly. They relate to the ‘disappearance of VAP’ as a diagnosis in ICU patients, the positive impact of the ‘VAP bundle’, and the frequent practice to co-administer aminoglycosides for synergy in possible P. aeruginosa cases. Therefore, enrolling VABP patients with P. aeruginosa becomes a sisyphean effort.
You guessed it: The cleanest way to conduct a proper VABP study for POL7080 would be with the help of a rapid diagnostic test which identifies P. aeruginosa in respiratory secretions, BAL fluid or tracheal aspirates at the time of enrollment. Roche would be the ideal company to combine a diagnostic with POL7080, select appropriate patients based on rapid identification of P. aeruginosa and randomly allocate them to standard treatment vs POL7080 plus a partner antibiotic. This is a workable approach, not easy but feasible and hopefully acceptable to regulators.
Most VABP patients will not present with pure P. aeruginosa pneumonia but with a mixed gram-negative flora. P.aeruginosa will be part of the mix of pathogens, mostly gram-negatives and S. aureus. This is probably quite helpful as initial treatment with an experimental drug for pure P. aeruginosa pneumonia, a disease with a mortality of 70%, is probably not advisable. But in the presence of mixed flora, we can follow the P. aeruginosa component under POL7080 therapy. Such a study would compare a SoC broad-spectrum antipseudomonal drug (like cefepime, pipercillin/tazobactam, ceftazidime, meropenem or doripenem) with the combination of, say, ceftriaxone plus POL7080.
Of course, any VABP study with a new antipseudomonal agent should be done without the co-administration of a second antipseudomonal drug. ATS/IDSA Guidelines for VABP therapy suggest the addition of tobramycin for synergistic P. aeruginosa coverage but such combination therapy would hopelessly confounds any efficacy analysis. Remember, such synergy has never been proven convincingly in non-immunocompromised patients, and the 2005 ATS/IDSA recommendations are reasonably flexible. (Note: Work is ongoing on an updated Guidance).
Of course, questions will be raised about the contribution of P. aeruginosa to a polymicrobial disease process, i.e., when P. aeruginosa is part of a mixed flora. Such questions are valid but also academic as in clinical practice the organism – once identified – would not be ignored but covered with an anti-pseudomonal antibiotic. No single VABP trial will clarify this issue which could not be answered in over 50 years of pneumonia studies. I would think that eradication of the pathogen by POL7080 would be a meaningful endpoint in support of the efficacy of POL7080 in this patient population and indication.
FDA wants sponsors to document efficacy along the coordinates how a patient ‘feels, functions and survives’. Clearly, equal survival rates at 28-days is important. While the ‘feely’ part is probably impossible to capture in VABP patients, duration of ICU stay and duration of ventilation support can be captured in the CRFs and substitute for a missing PRO assessment.
The FERMAT / REINHART conjecture states:
There is a way to develop POL7080 but it does not fit on the margin of this page
Despite these difficulties, a workable development path can be found which should satisfy Health Authorities demanding solid proof of efficacy. POL7080 is a unique drug which requires an innovative approach and will need the support of clinicians, academics and regulators to make it to the market; Phase 3 should not become ‘Mission Impossible’.
References:
FDA Guidance for Industry. Hospital-Acquired Bacterial Pneumonia and Ventilator-Associated Bacterial Pneumonia: Developing Drugs for Treatment. 2014
Guidelines for the Management of Adults with Hospital-acquired, Ventilator-associated, and Healthcare-associated Pneumonia. Am J Respir Crit Care Med Vol 171. pp 388–416, 2005
Abbreviations:
SoC standard-of-care
CRF case report form
PRO patient reported outcome
BAL broncho-alveolar lavage
VAP/VABP ventilator-associated (bacterial) pneumonia
