Just to refresh everyone’s memory: First, we had the Points-To-Consider document of 1993, then the 1998 Guidance, followed by the 2012 Guidance and now the 2015 Guidance for the conduct of registration studies in cUTI. They all have served us well and I cannot recall a situation when a drug was approved based on these guidelines but later found to be ineffective. Which begs the question: why is there this drive – almost obsession – to fine-tune a regulatory path that works just fine?
Case in point: the failed cUTI trial which recently ended the prospects of AN-3365 / GSK-052, a leucyl-transfer RNA synthetase inhibitor. The drug, rather surprisingly, was not active because resistance developed rather fast. In that case, inferiority to standard comparator was obvious and not a subtle finding; the drug would not have met the Points-to-Consider criteria for efficacy either.
In what respect is the new cUTI Guidance different from earlier versions? Where is the impact of ‘new science’ that triggered a ‘rethinking’ of the regulatory approach to cUTI? I can come up with several potential answers but they don’t seem to amount to anything substantial. The changes in the 2015 Guidance are not unreasonable from a clinical point of view but are often of a “C-III level of evidence” which amounts to ‘expert opinion’ not hard facts. Many of the changes might get changed again in the next version(s).
Does it really matter if efficacy now requires not just microbiologic cure but also clinical resolution of all core symptoms? Aren’t these two things linked? And does it really matter if urgency persist on Day 12 (an incomplete response now considered a Failure) but then resolves a few days later (making treatment outcome a Success)? Is a PRO tool really improving the evaluation of drug efficacy in infectious diseases trials that already give us the best of all possible surrogate markers of efficacy, i.e. bacterial response?
I think the 2015 Guidance is a defensive document, incorporating some concepts which are important to regulators and statisticians, less so to clinicans (or patients):
- An evidence-based justification for the 10% NI margin:
Our assessment: MUCH ADO ABOUT NOTHING
As it turns out, only a single trial (from 1918) could be found that had a sizable population of cUTI patients from pre-antibiotic days. Despite its shortcomings, it became the basis of the HESDE justification, the ‘historical evidence of sensitivity to drug effect’.
This trial, like the Snodgrass publications from 1936/1937 in cSSSI, is certainly of interest to historians but making it the basis of a statistical construct to derive M1 is really stretching the imagination. It is not often that regulators are credited with imaginative rulings but this is surely a prime example of ‘data dredging’. Or for the creative writing (A-I evidence)
Our assessment: MUCH ADO ABOUT NOTHING - Incorporating the ‘feel-function-survival’ gospel into the cUTI Guidance:
Our assessment: A GOOD THING TO IMPLEMENT
This ‘holy trinity’ of outcome criteria was conceived by FDA statisticians in an effort to make outcome determinations less biased, more objective (mortality!) and more patient-centric. While other disciplines had long worked with PROs for lack of better efficacy parameters, infectious diseases studies had always relied on microbiology results and investigator assessments.
Clearly, cUTI patients can relate their symptoms and provide unfiltered feedback. There is no reason to believe that patients could not provide information about their complaints and how they improved with treatment.
Our assessment: A GOOD THING TO IMPLEMENT - Beating a path back to reality regarding prior antibiotic use:
Enrolling patients into cUTI trials should really not be that difficult: the condition is not rare, the symptoms are tell-tale, and there are rapid diagnostics in the form of Gram-stain and diptick tests. What makes these studies hard to execute in the real world is the inelligibility of patients who received systemic antibiotics prior to enrollment.
In the past, treatment <24 hours with non-study antibiotics was acceptable; given a total treatment duration of 7-14 days this 1 day of “contamination” would not seem very consequential. However, the FDA is correct in their thinking that such early antibiotic use can have a substantial impact on outcome. At least, it is a significant confounder in the context of NI studies as it tends to blunt any true difference between study meds and controls.
In addition, some of these non-study antibiotics have long half-lives, or prolonged urinary excretion times, or accumulate in the urine, thereby resulting in treatment effects past 24 hrs.
So, the scientifically correct approach is to demand that study patients be free of such exposure. However, here is the rub: in the real world, patients will have started antibiotic therapy before coming to an ER with cUTI to be admitted. A good FMD who has a patient with known structural abnormalities or indwelling catheter will not hesitate to start therapy as he sends the patient on his way to the hospital.
FDA wants our clinical studies to reflect reality, otherwise we would not have an ITT approach but a per-protocol approach in our studies.
In an effort to find middle ground: FDA now allows up to 24 hrs of prior antibiotic use in up to 25% of patients. They encourage sponsors to reduce the amount to a single-dose of prior antibiotics, if possible. But then they still demand that >75% of all cUTI study patients should not have any prior antibiotics.
Our assessment: IT SHOULD HAVE READ: “UP TO 75% OF PATIENTS CAN HAVE A SINGLE DOSE OF NON-STUDY DRUG PRIOR TO RANDOMIZATION”.
Reference:
Complicated Urinary Tract Infections: Developing Drugs for Treatment. Guidance for Industry. http://www.fda.gov/downloads/Drugs/Guidances/ucm070981.pdf
