In this day and age, few truly broad spectrum antibiotics are in development, although this does not hold true for narrow-spectrums. Just think of all the MRSA drugs (quinolones, FabI inhibitors), the pseudomonas drugs (NCEs and monoclonals) or some narrow spectrum drugs that will need to be tested in MDR pathogen infections.
Almost gone are the days of indication-based antibiotic development, although FDA Guidances still make you believe the indication-driven approach is the main one leading to approval. Instead, pathogen-driven development is now the norm, and it creates some very difficult problems:
- Where to find patients with infection causes by a specific pathogen
- How to enrich study populations for these pathogens if rapid diagnostic tests are not available
- How to establish outcome criteria when different sites are involved with a particular pathogen
- What is a suitable comparator drug when it is unlikely that the pathogen under study is susceptible to both
- How to blind against SoC, i.e., a cocktail of drugs that is likely required against MDR pathogens
There are no easy answers, and it takes much thought and planning to address these problems and to find workable solutions. Worse, every drug seems to require its own fairly unique testing scenario which has to be discussed and agreed-to by Regulators. Despite the great need for new drugs, showing their efficacy in patients with highly resistant populations is not straight-forward.
EXAMPLE
A new drug with specific activity against P. aeruginosa is much needed, hence the attractiveness of LpxC inhibitors and POL7080. But how to prove efficacy when therapy for P. aeruginosa is typically initiated as a combination regimen? Admittedly, the benefit of combination over monotherapy has never been proven beyond doubt but it is established clinical practice and supported by practice guidelines.
Are we to evaluate a new anti-pseudomonal in combination with an aminoglycoside? This would fly in the face of the often-made argument that (1) concomitant antibiotics with overlapping activity invalidate a case for efficacy determination; (2) even short courses of non-study antibiotics can profoundly affect the outcome; and (3) such therapy would lead to regression towards the mean and slant outcomes towards non-inferiority, resulting in a Type II error.
One often hears from KOLs that finding patients with a specific MDR profile is doable. Take CRE as an example: these organisms are no longer uncommon in India, Greece, Italy, Spain and some other locales where antibiotic stewardship is successfully avoided. According to the literature, these pathogens are popping up everywhere, including the US[1]. Only problem: there are not enough cases yet in the USA to even consider a traditional randomized trial of a new antibiotic against SoC[2]. If you do, you do it at your own risk.
Remember what happened to Advanced Life Sciences (ALS). They tried to conduct a study showing that cethromycin (Restanza®) has clinical activity against erythromycin-R respiratory pathogens. The FDA had asked for “a superiority trial design comparing Restanza to a marketed macrolide antibiotic” in not one but two trials[3]. Did KOLs convince ALS that such a trial was feasible? Well, the new Phase 3 study did not make it very far because ALS ran out of money in the futile attempt to pull off ‘mission impossible’. Only a few months after the proud announcement that ‘agreement with FDA on a Phase 3 study design’ had been reached in the form of a Special Protocol Assessment (SPA), they threw in the towel[4].
Some may still remember the difficulties Ortho / J&J had in the 90ies trying to enroll patients with penicillin-resistant (PCN-R) S. pneumoniae. The goal was to show that ofloxacin had good activity against these strains; based on MICs, this would seem a foregone conclusion. However, it took many years to collect such cases despite the fact that PCN-R S. pneumoniae was all over the news. Turns out that most PCN-R strains were found in pediatric populations, not in adults. And because we cannot expose children to fluoroquinolones, this study did not go anywhere fast.
Now Achaogen has set out to lift a similarly heavy load: they embarked on a Phase 3 study comparing plazomicin to colistin in CRE-infected patients[5]. Their lofty goal (360 patients) is certainly ambitious, to the point of appearing unrealistic. But hey, who is to say they won’t succeed where others failed? Nonetheless, I smell the good but self-serving advice of KOLs and the purist demands of regulators emanating from this endeavor.
Follow Artie’s Law[6] that says: Avoid bad advice but certainly don’t follow ‘Mission Impossible’ advice.
Also, don’t forget Lasagna’s Law[7] which states that the “incidence of patient availability sharply decreases when a study begins and returns to its original level as soon as a study is completed“.
And then there is also Muench’s Third law …
References:
[1] Gupta CID 53: 60, 2011
[2] Temkin Ann N Y Acad Sci. 2014; 1323: 22
[3] http://www.ukmi.nhs.uk/applications/ndo/record_view_open.asp?newDrugID=4445
[4] http://www.fiercebiotech.com/story/cash-shortage-puts-advanced-life-sciences-ice/2011-05-06
[5] https://clinicaltrials.gov/ct2/show/NCT01970371?term=Achaogen&rank=1
[6] Don’t look it up, I made it up
[7] http://www.appliedclinicaltrialsonline.com/improving-clinical-trial-enrollment-forecasts-using-sorm
