Board Review Infectious Disease

Clinical Trial Design in Onychomycosis

Warren Joseph, DPM

Warren S Joseph, DPM, FIDSA reviews current FDA shortfalls for conducting onychomycosis studies. Dr Joseph also reviews current acceptable endpoints and discusses how lack of trial design has lead to difficulty comparing drugs and treatment options.

CPME (Credits: 0.75)

  • CME Progress
  • Pre-Test

  • View Lecture ( CPME Credits: )

    Lecture Transcript

  • Post-Test

    Requires: Pre-Test, View Lecture
  • Survey

    Requires: Pre-Test, View Lecture, PostTest
  • Certificate

    Requires: All Content Above
Method of Participation

Complete the 4 steps to earn your CE/CME credit:

  1. Complete the Pre-Test
  2. View the Lecture
  3. Complete the Quiz (Min. 70% Passing Score)
  4. Complete the program Survey
Goals and Objectives
  1. Identify what the FDA requires to submit a pivotal trial
  2. Define the currently accepted entrance criteria and endpoints when comparing different clinical trials
  3. Identify the scientific importance of different trial types
  • Accreditation and Designation of Credits
  • CPME (Credits: 0.75)

    PRESENT eLearning Systems, LLC is approved by the Council on Podiatric Medical Education as a provider of continuing education in podiatric medicine.

    PRESENT eLearning Systems, LLC has approved this activity for a maximum of 0.75 continuing education contact hours.

    Release Date: 03/16/2018 Expiration Date: 12/31/2018

  • Author
  • Warren Joseph, DPM

    Roxborough Memorial Hospital
    Philadelphia, PA
    Editor - Journal of the APMA

  • System Requirements
  • To view Lectures online, the following specs are required:

    • PC Pentium-III class or better processor
    • 256MB minimum of RAM
    • Cable or DSL broadband Internet
    • Browsers must have javascript enabled. Most browsers have this feature enabled by default.
    • Adobe Acrobat Reader (Free) to print certificates
    • Supported Browsers:
      • Chrome
      • Firefox
      • IE 10+
      • Safari
      Supported Mobile OS:
      • Apple iOS 4.3+
      • Android 2.3+
      • Honeycomb 3.1+
      • Blackberry Playbook
  • Disclosure Information
  • It is the policy of PRESENT e-Learning Systems and it's accreditors to insure balance, independence, objectivity and scientific rigor in all its individually sponsored or jointly sponsored educational programs. All faculty participating in any PRESENT e-Learning Systems sponsored programs are expected to disclose to the program audience any real or apparent conflict(s) of interest that may have a direct bearing on the subject matter of the continuing education program. This pertains to relationships with pharmaceutical companies, biomedical device manufacturers, or other corporations whose products or services are related to the subject matter of the presentation topic. The intent of this policy is not to prevent a speaker with a potential conflict of interest from making a presentation. It is merely intended that any potential conflict should be identified openly so that the listeners may form their own judgments about the presentation with the full disclosure of the facts.


    Warren Joseph Warren S Joseph, DPM, FIDSA has disclosed that he receives Honorarium/Expenses, is a Consultant to and serves on the Speaker's Bureau for Pfizer and Merck.

  • Lecture Transcript
  • Warren Joseph: I’m Warren Joseph from Philadelphia, Pennsylvania. I’m really excited to be here today to give you a totally new PRESENT e-Learning lecture which is a brand new topic – never seen before. Never seen before on PRESENT, never seen before in any lecture you’ve attended. And that is a lecture specifically on clinical trial designed on onychomycosis. Now, I know that might sound a little bit dull, a little bit boring to you but I promise you there’s some really interesting little features. I want to give you a little bit of an insight in how our Food and Drug Administration works in looking at drug studies, not only in the antifungal world, also talk about the antibiotic world a little bit and the device world a little bit to show you how the FDA addresses the questions of clinical trial design. Well, what we’re going to look at is first, answer the question, what does the FDA want? You know, we don’t really know that anyone can answer that and the FDA has been notoriously accusing the past of changing the ground rules of what they want on the fly. But I’m going to give you my idea of what they want when it comes to onychomycosis and even some antibiotic studies. We have to look at what type of patients are enrolled, what is the demographics of the patient, what’s the extent of the disease that the patient has and are those two the same, can you compare trials? What are the currently accepted endpoints? I think this becomes a major issue especially in the world of onychomycosis because I think the endpoint is unnecessarily severe. And that answers the question or that at least gets the question, are these endpoints reasonable and I’m going to show you why I don’t think they are. And, can you compare the studies? There may be a desire of different people everywhere from you and the audience to drug reps, to want to actually compare studies, compare drugs. Say look, this is how we did our study, this is how they – they were better than they are. Well, you can’t do that. The FDA frankly does not allow that unless they are head-to-head trials, what are called non-inferiority studies. And in many cases, at least in the fungus world and no cases that I can think of, has that ever been done for onychomycosis. So let’s look at the FDA on onychomycosis studies. The FDA puts out something called a guidance for industry document which basically lays out the rules that the pharma company has to follow, to do a clinical trial in any given area. And by the way, you can find all of this information, you can find all of these different guidance for industries by going to FDA.gov and just searching for guidance documents. Well interestingly enough, there are currently no guidance for industry documents that exist on clinical trials for onychomycosis. I looked, I contacted my people that I know in the industry and could not find anybody that could tell me about an FDA guidance for industry for onychomycosis studies. Well, what that means in English is it means each and every trial has to be negotiated with the agency individually. Now just for those of you in the no, when we say the agency, that’s how some of us refer to the FDA. It’s just a little of a shorthand there. It also – and because that there are no guidelines and each of these trials are negotiated individually, this leads to inconsistent trial design between agents, making it almost impossible to really compare studies head to head. Interestingly, even though there is no guidance document for drug therapies for onychomycosis, there is in fact a guidance document for device therapy in onychomycosis. And here it is. This is the phase sheet from it. What you’ll find, it’s called medical device in clinical trial designed for the treatment or improvement in the appearance of fungally infected nails, notice that this is a draft guide. It’s relatively new. It just came out this year, January 2015, but I think they let you know in more than one place that this is a draft. It is a draft, it’s not for implementation. It contains nonbinding recommendations. Interestingly enough, frequently, these documents never get beyond the draft stage. There are lots of guidance documents that are still kept in the draft stage even though they’ve had multiple rewrites and multiple changes to them. There are some however for instance the skin and skin structure document I’m going to talk about, it is finalized. But this is the draft document for device studies of treatment of onychomycosis or I should say improvement in appearance of onychomycosis. So what is the indication? And I think this is worth reading to you. Yes I know, you’re capable of reading it yourself. But the indication that the companies making devices for onychomycosis actually have to go for is not treatment of onychomycosis.


    It is for temporary increase of clear nail. And what the document says is historically, devices have been cleared with the indication for use or IFU as they call it but for a temporary increase of clear nail in patients with onychomycosis. This indication has been used to highlight the outcomes for which the device has been – has provided support are limited to visual improvement anesthetic endpoint and not for mycologic cure. It’s anesthetic endpoint and not for mycologic cure, that is their underlying, that is not my underlying, okay? Because a mycologic cure would constitute a medical endpoint and they are not looking at that. In fact, they say this IFU is not intended to mean that the treatment eliminates fungal infection. It does however convey that the clinical studies performed in support of the indication included nails with confirmed fungal infections. Again, this is just for further down in the document. This device has not demonstrated effectiveness in the treatment of fungal infections. It is cleared only for improvement in nail appearance and that is a statement that has to go with any of these devices. So when you hear about devices being marketed for onychomycosis, please keep this in mind because they were looked at under drafts of this document. None of the devices for onychomycosis are approved to treat onychomycosis. I can’t say that enough. None of them are approved to treat onychomycosis. They are cleared, and there’s a difference between using the words approved and cleared, they are cleared for the indication of temporarily improving the appearance of the clear nail. Now that’s not bad. It’s not necessarily evil or wrong but they don’t have the data to support that they actually treat the fungal infection. So what does the FDA look for when they’re looking at studies for onychomycosis considering that there really are no written guidelines? Well, all drugs have to have two identical, what are called pivotal phase III trials comparing the active drug to either the vehicle or placebo to show superiority. Now that’s FDA-speak, that’s statistical speak. But what it means is they have to run the drug not against another drug but against the vehicle or placebo. Now please note, vehicle and placebo are two different things. Vehicle for instance for a topical can have some active agents. For instance, some of these new topical antifungals that we have, the fenticonazole and tavaborole have alcohol-based vehicles. Well, maybe the alcohol has some antifungal activity and in fact with both of those studies, you do see some effectiveness of the vehicle. Placebo, you frequently see in oral studies where the patients being given an oral drug and there they can be given a placebo, a true placebo drug that has no effect at all and the drug is looked at to show – it has to show statistical superiority over the vehicle or placebo to get its FDA approval. Head-to-head studies between two drugs have not been done nor does the FDA require them to be done. These are what are known as non-inferiority studies. When you run two drugs against each other, we see a lot of this in the antibiotic world. New drugs can get approved on non-inferiority, so drug A is tested against drug B and found to be statistically similar. We say that this drug A is non-inferior to drug B. Well, here’s the problem with non-inferiority studies. The FDA is concerned with something called non-inferiority creep. What does that mean? Well, drug A study against drug B found to be non-inferior. Drug B studied against drug C, non-inferior. Drug C is studied against drug D, non-inferior. Theoretically by doing these non-inferiority studies, that should mean that drug A is non-inferior to drug D. We can’t say that, we really can’t say that. So, there is this concern about this non-inferiority creep. You can get a drug approved through the FDA, doing non-inferiority study with an antibiotic currently with antifungals. They are strictly looking at the vehicle-controlled superiority studies to get the indication. So, recent top of the drug studies actually require 48 weeks of therapy once daily for 48 weeks with a four-week followup off therapy. So 48 weeks on, four weeks off. Likewise, some of the tinea pedis studies require two weeks on, four weeks off before the primary endpoint is evaluated. Now this is in the negotiated endpoint. What if somebody comes up with a new antifungal and says you only need to apply us for six months instead of for 48 weeks.


    You only need to apply us for three months. Well then they would have to go to the agency and say we want to run a study showing that only three months is necessarily or only six months is necessary. But for the current drugs that have been studied, they have all looked at 48 weeks followed by four weeks off therapy for a 52-week primary endpoint. So the primary time or the total time is 52. That’s an incredible amount of time when you really think about it. Think about doing a study on onychomycosis and having to follow a patient for a full year. Think about patient dropouts, think about the time involved with that, and what about followup after that? I mean people always talk about reinfection and recurrence. Well these companies have already paid for a 52-week long study. How much longer are they going to study this patient? How much further out can they go? It starts becoming a very costly ordeal. You can only imagine what the cost is to follow 15, 1600 patients for a full year. And why do they make it a full year? Because the thinking is, it could take at least that for a hallux nail to grow out. So it’s possible that using an antifungal for only a couple of months might kill the fungus and the fungus – the new nail will continue to grow with dead fungus and may come out looking clear in 48 to 52 weeks, even though the drug has only been used for a couple of weeks or a couple of months. But that’s not the way it’s currently negotiated. They have to treat for 48 weeks. So these other duration and dosings can be negotiated with the agency. What about the primary endpoint? Currently, the FDA requires for all onychomycosis studies, a primary endpoint that is called complete cure. Now complete cure is defined as, at week 52, the nail is mycologically cleared which means negative KOH and negative culture. And the nail is 100% perfect-looking. Well that’s kind of a ridiculous endpoint. Mycologic cure, negative KOH and negative culture, let’s think about that. The fungus can be dead but the KOH can still be positive because KOH doesn’t differentiate dead from live fungus. Well if the fungus is dead, what’s it matter whether it’s there or not? Complete nail care, how often do you get a patient with a nail that basically looks like your thumb nail after your treatment with an antifungal. There can still be other issues with the nail. How do you define a completely cleared nail? Has the nail been clipped back a certain way, trimmed back a certain way? All of these are variables you have to be concerned about in the study. I think this is really an unreasonable endpoint. Because this is such an unreasonable endpoint, the companies come up with secondary endpoints which the agency allows, but these include vague terms such as almost clear, clinical success, overall success, quality of life. Probably the firmest of these secondary endpoints is mycologic cure. Mycologic cure basically needs at the end of the study the nail is KOH and culture-negative. But again, it doesn’t matter if a fungus is dead, the KOH can still be positive, not a particularly good endpoint. Well, how does this very strict endpoint of negative mycology in a perfectly clear nail at 52 weeks compared to, let’s say antibiotic endpoints? Well this is the current guidance for industry that’s available for the relatively new indication known as acute bacterial skin and skin structure infections? So, if a company wants to do and study on a new antibiotic for what they define as these ABSSSI or acute bacterial skin and skin structure indication, what are the endpoints? Well, there’re a lot more basic, there’re a lot easier to achieve than you have with the antifungals. The primary efficacy endpoint for any new antibiotic approval is that at 48 to 72 hours, this cellulitis has resolved by 20%. Think about that. At 48 to 72 hours, the cellulitis is not gone, it’s only gotten better by 20%. And I will tell you when they first came out with this guidance, they didn’t have the 20% in there. The original ABSSSI guidance said that at 48 to 72 hours, the cellulitis stopped expanding. It didn’t get any worse. This I think is a kind of an easy endpoint. It’s a reasonable endpoint. It might actually be in my opinion a little bit too nonstringent.


    But this is the guidance for antibiotics versus what we have to deal with when talking about onychomycosis. So, is the FDA behind the times? Again, they want KOH and culture. Well, we know that KOH and culture are not particularly good tests. I’ve talked about the issues with KOH, what about culture? Cultures can be falsely negative over half the time. It – They’re very highly labor intensive thing to do. You think of bacteria, basically you send a swab to a lab. A lab grows a couple of colonies from that swab, takes those colonies, sticks it in a machine and that machine in couple of hours reads out for you what the organism is and the susceptibility against all antibiotics. You can’t do that with fungal culture. Fungal cultures take two to four weeks between the organ – before the organism even grows, then the growth of the organism has to be teased out by skilled mycologists who then stains the organism and has to look under the microscope for hyphae and chlamydia to identify what that organism is. There’s no mechanization here. This is all done by hand and it is very difficult to find a skilled mycologist. So, these are not particularly good or up-to-date tests to determine the presence of fungus. There are new or better test. We have PAS staining, we have molecular diagnostic techniques such as PCR and 16S ribosomal sequencing, deep sequencing. Yet, the FDA still requires positive microscopy to enroll the patient in studies or positive KOH to enroll the patient and negative KOH to be considered – and negative fungal culture to be considered a cure. And this has been around forever. I mean, this appears to be an informal FDA requirement which has been in place for these 15 years and companies have tried to shake them off but no one has been able to. So, is a completely cured nail a reasonable endpoint? We talked now already about the mycologic cure but that’s not reasonable because the testing is not particularly good. Well what about that perfect-looking nail, is that really a reasonable endpoint? Well, back in 2007, I was honored to sit on a consensus panel that looked at the definition of a cure of onychomycosis for clinical trials. And we published this in the journal, the American Academy of Dermatology, the problem is that one of the authors was at that time employed by one of the pharmaceutical companies, in fact Novartis, and for that reason, the powers that be in the FDA and other organizations felt that this was somehow biased because he worked for a drug company and never adopted this definition of cure. But I think it’s still a reasonable endpoint which the FDA and industry should consider as possible definitions for onychomycosis. So what do we say is the criteria for cure? We say that 100% absence of clinical signs of onychomycosis. Now, that doesn’t differ too much from complete cure, a perfectly look – normal-looking nail but where we differ is the mycology is not required. So in other words, as long as the nail looks good, we don’t need to do mycology which can be falsely positive, falsely negative to show that is a cure. So that’s number one. Or, we say that you can have negative mycology result with one or more of the following signs. So negative mycology, but you have distal subungual hyperkeratosis or onycholysis even less than 10% of the nail plate affected. Today, that 10% is used as a definition for almost clear or almost completely cured by some of the different companies. We actually proposed that to mean it’s pretty much what the definition is for a cure. Because as I’ve shown you and I will show you, those so-called almost clears, those 10% involvements look pretty darn good. The other instance, we can have – does not improve with treatment because of comorbid conditions. So the patients may have psoriasis, may have lichen planus, other issues, posttraumatic nail syndrome, other issues that caused the nail to be particularly thick. But this was our proposed definition. They would be a lot more reasonable, a lot easier to obtain than the FDA's current definition. So there again you see, we do not require mycology as the definition for cure and clinically, you could have up to 10% involvement. Well, this is what this looks like in clinical practice. This is actually a picture from the early trials that were for Penlac, ciclopirox lacquer, and you can see this is a complete cure. The nail was involved at baseline, at week 48 the nails were no longer involved. That is complete cure.


    I think everybody would agree that nail looks very, very good. Well what about this nail? This was from the tavaborole study, Boni E. Elewski, just published this in fact in the Journal of the American Academy of Dermatology very recently. Here you can see that, yeah, the patient has involvement beforehand and it looks pretty darn good at week 52 and in fact the KOH and culture were negative but the investigator felt there was up to 10% involvement there, maybe they just didn’t trim the nail that far enough. For whatever reason, they felt there was 10% involvement there. That was an almost clear nail. That according to the FDA is a failure, according to our definition of the cure published in 2007, that would have been a cure. Another one, also from the tavaborole study, you can see the patients got about 30% involvement. And at the end of the study, well maybe the nail is not perfect, but I would say there’s less than 10% involvement there. According to current guide and according to the current regulations, I almost said guidance but remember they don’t have guidance. According to the current regulations, that’s a failure. According to our renewed definition, that would have been a success. And finally, we’re going to talk about comparing study designs. Again, there’s a lot of tendency for people to want to say, well, we’ve got this drug and that drug, you know, how do they work? Which one is better? I get these phone calls all the time. Warren, which one of these new drugs is better than the other? You can’t say, none of them have ever been tested against each other. These are all done as vehicle-controlled studies under different conditions. Let me show you what I mean by different conditions. Patient demographics. In the ciclopirox study, the demographic was 18 to 70. In efinaconazole, 18 to 70. In the tavaborole 18 with no upper age limit of normal. The oldest patient was in fact 88 years of age. Does that make a difference in the outcomes? We don’t know. All we’re saying is, you can’t compare head to head. What about the extent involvement and how was it determined? Well, for ciclopirox, you add out between 20 and 65% involvement and that had to be done using a technique called computer planimetry where the extent of involvement of the nail was drawn with a marker, the nail was photographed and a computer used to determine how much nail was involved. Well, efinaconazole looked at 20 to 50% involvement. Tavaborole looked at 20 to 60% involvement. And they were determined visually. Basically, the investigator looked at the nail and said, yeah, it looks like there’s about 40% involvement there. Yeah, it looks like there’s 50% involvement there. It’s extremely subjective. So, how do you compare these trials with different endpoints and different modes of determining extent of involvement? Let me show you some examples. This is the – from the Penlac study. Here you can see the use of planimetry. Notice how specific they could get there. At baseline, they found 26.7% nail involvement, that’s about the specifics. Again, no one can eyeball that. And at the endpoint, that looks pretty darn good considering how it started, but because that was drawn the way it was drawn, and the computer determined it to be 17.1% involvement, that was considered a failure. I think that looks pretty good. I think your patients would be happy with that. I certainly would be happy with that. These are from the efinaconazole studies. You see here in the upper left-hand corner, 40% involvement on that patient. Maybe, I would say that sounds about right. Twenty percent involvement at week 24 but 3% at week 52. I don’t see how that’s 3%. To me, that looks almost more like 10%. How is that determined? I mean, this on the right was 17.1%, how is that 3%? You see the subjective nature of determining this. In the center column, you’ve got 50% involvement initially and by the end, I left it blank, think about it. If you were the investigator, what would you call that? How much involvement would you say that is? Well, on the left you’ve got 3% so what do we call that? Looks a little more than 3%. In fact, to me that looks like about 10, 15% if I had to guess it. Well, the investigator call that 5%. Well, 5% is still a failure. All these are still failures according to FDA. According to our 2007 guideline, that would have been okay. But if I look at it and said that’s 15%, that would have been a failure. It doesn’t matter. Look at the beginning. Look how bad that looks at the beginning. Look how good that looks at the end. But that would still be considered a failure. And finally, this last column, here you see how much involvement?


    I think most of you would probably look at this and say 80% involvement. When I show this around the country, they will say, oh yeah, 80% involvement. Well, actually the investigator called that 45% involvement. And look at the bottom, 5%. Well, if the one on the right next to it in the center is 5%, how was that also a 5%? I don’t care because the bottom line is, it’s a really nice result but the point is you could see how subjective all of these endpoints can be. And you could see how subjective the investigator's opinion can be about the extent of nail involvement. Well, let’s continue comparing studies. What about application? Ciclopirox you applied it once daily to the nail plate and the skin beneath the nail and it was removed weekly with alcohol. In fact, we were always cautioned not to apply to the nail grooves or the peri-nail tissue. Efinaconazole, we apply to the nail plate, the cuticle, the folds of the skin around the nail and under the toenail. And tavaborole was applied to the entire nail surface and under the tip, so the drugs were applied differently in different studies. Debridement and trimming. Ciclopirox, had to remove the loose nail before starting then monthly debridement by a healthcare practitioner. This action is in the package inserts. Ciclopirox, you needed to do monthly debridement, okay. Neither the efinaconazole nor the tavaborole had any debridement in the study at all. It doesn’t mean you can’t do debridement in real life, it just means that in the study and in fact the debridement may help, it just means that in the study, debridement was not done. Efinaconazole, no debridement but they did allow trimming of the nail. Tavaborole had no debridement and they did allow trimming, but they limited the trimming to within 1 millimeter of the distal nailfold. So again, trials done differently. So can you ever really compare trials? There is a technique for possibly comparing trials and this is brand new. It’s called a network meta-analysis. The first paper that I saw come out on this was by Aditya Gupta of Canada and it’s been published in the Journal of the American Podiatric Medical Association. And I actually as the editor of that journal, had to handle the review process. So I had to actually go to some world authorities in the area of network meta-analysis to get reviews of the paper because it’s a new technique. Whereas, a traditional meta-analysis conducts pairwise comparison of two drugs. A network meta-analysis compares a group of drugs to each other to direct and indirect evidence using the pairwise comparisons they have in common. Basically they try to control as much as they can and then compare all the different drugs. So this is what it looks like and this slide is a bit of a mess but it’s actually very easy to understand. This was Gupta’s network analysis of onychomycosis randomized control trials and just like – all it means, all those numbers mean are the number of trials that have been done. So between placebo and fluconazole, there were two studies. Between placebo and TNS, terbinafine nail solution, there were three studies. Terbinafine and placebo, three studies. Terbinafine and itraconazole, two studies. So you get the idea of how that’s done. Well, upon this analysis, what Gupta showed was that the most effective drug, and this is not going to surprise anyone, was terbinafine followed by itraconazole first the paused and then the continuous followed by oral fluconazole which never did get its FDA approval even though it was studied, followed by efinaconazole and ciclopirox and then some drugs that aren’t available. Notice that the tavaborole is not on this guideline and not on this particular document. The reason for that is the tavaborole data was not published at the time Gupta did this review and he did not feel it appropriate to do a network analysis based on just some press releases and posters. So, I think rightly he did not include the tavaborole in this network outcome study. So in conclusion, the FDA does not have current guidance for industry on how to do an onychomycosis study. Each trial is designed and negotiated totally differently and individually per drug and per company and this leads to difference in trial design. If there are no head to head or what we call non-inferiority trials comparing two drugs, then conclusion about the relatively – relative efficacy cannot be stated. In fact, if you happen to hear a sales rep come in and say, we’re better than drug X or we’re better than drug Y, let me know about it because frankly it’s a federal offense for them to do that. It’s not in their package label and there haven’t been any studies to compare. Thank you very much.