Barriers to Diabetic Foot Ulcer Wound Healing Clinical Trials: Insights From Regulatory and Clinical Affairs

In this podcast, Kazem Kazempour, PhD, and Kush Dhody, MD, discuss how careful study design is required to demonstrate efficacy of new wound healing treatments.

Transcript

Jaclyn Gaydos: Welcome back to "Speaking of Wounds," a podcast by the Wound Care Learning Network.

Hello, my name is Jaclyn Gaydos. I'm the Senior Managing Editor of WOUNDS. Today, I'm here with Dr Kazem Kazempour and Dr Kush Dhody. The topic of today's interview is, "Barriers to Diabetic Foot Ulcer Wound Healing Clinical Trials: Insights from Regulatory and Clinical Affairs."

During the last 10 years, the US FDA has approved only a few new treatments for diabetic foot ulcer healing. This is partially due to the way the FDA defines success in treating these wounds, and partially due to the design of clinical trials, which may fail to capture the most meaningful data, or may produce unintentionally skewed data.

Amarex's Dr Kazem Kazempour and Dr Kush Dhody believe careful study design is required to demonstrate efficacy of new wound healing treatments. Thank you both for joining us today.

Dr Kazem Kazempour, considering the lack of approved treatments for diabetic foot ulcer healing over the last 10 years, why is it so difficult for new therapies to be approved?

Dr Kazem Kazempour: This is fairly complex issue. There are many factors in responding to your question. Really, it's multifactor issue. Part of the challenge is that the regulatory agency, FDA, has a very clear, and at the same time, very narrow definition of wound healing. That is 100% wound closure with no drainage or no need for any dressing.

That is called wound healing. Time to 100% closure. 100% closure is what I define, that no drainage, no need for any dressing. Part of the challenge is that the FDA has this definition. It is generally fairly difficult to demonstrate the efficacy using this definition of wound closure.

If all the noises and all other factors affecting the wound closure are not considered at the design stage. When the trial is over, it is going to be very hard to show that, to demonstrate the efficacy of the treatment.

Investigators and study sponsors, really, they should work very hard in designing their clinical trials in a way that, as much as they can, to reduce the noise, so that the signal can be observed and presented to regulatory agency, with proper statistical power and necessary control over the patient population. I will talk more about that one.

From that angle, due to all these factors, it is difficult to show the efficacy.

Gaydos: Dr Kazempour, you mentioned the FDA's definition of wound healing is part of the problem. How exactly is the FDA making it challenging to pass these new therapies for DFU healing?

Kazempour: You are right, some consider this definition of wound closure by FDA as part of the problem, which due to the difficulty, as I mentioned earlier, it may be one of them, because as I said, 100% wound closure with no dressing requirement, no drainage, during the pre-specified time, it is challenging, no doubt. It has been challenging and it is a challenging one.

Currently, that is the clear definition. That is the rule of game to play. All the sponsors, in order to be able to approve a new product, they have to follow the rule. There is no exception whatsoever.

From a regulatory and statistical analysis point of view speaking, if the wound reduces, as I mentioned, the size, say three centimeters square to .05 centimeter square, the sponsor cannot claim the treatment is working. Although it may be that, yes, the wound size reduced by 99.5%, but that is not enough from the approval.

The clear factor is time to wound closure. That's another main factor. One of the things it's there is that, most of the clinical trials are between 12 to 24 weeks in duration. Let's say 16 weeks for the sake of this discussion. With certain wounds, they simply are not going to heal 100% during this pre-specified time—16 weeks, for example.

All it depends on the size of wound, the depth of wound, the severity of wound, and so many factors involved in the wound closure. All these factors, as I said, will create noise. Sponsors designing their trial by taking all these factors. Which when you do that, it increases the size of the study.

They may be able to go to FDA and get approval when all these factors are controlled. You cannot really get rid of all of them, but you can control them and be able to present it to regulatory agency for approval.

Gaydos: Dr Dhody, if the FDA determines the successfulness of a new therapy at 100% wound closure, that would mean the FDA does not consider a significant reduction in wound size as successful for approving new therapies.

From your clinical view and experience with clinical trials, what's the biggest challenge in the design of the clinical trial?

Dr Kush Dhody: Jaclyn, this is a very good question. As Dr Kazempour mentioned, wound healing is a multifactorial in nature. There are so many factors that determine whether the wound will heal or not. One of the biggest challenges, in my experience, is choosing the right patient population for the study.

When it comes to selection of patient population, the most crucial aspect is finding the patients with the wound that will not heal on its own, but has a potential to heal with intervention.

At the same time, you should avoid enrolling any patients with the wound that will heal on its own without any intervention, or a patient with a wound that is so bad that it is very unlikely to heal during this proposed duration of a trial, no matter what intervention you apply.

In the diabetic foot ulcer clinical trials, you typically look to enroll patients who have hard‑to‑heal ulcers. Typically, in this DFU setting, any patients who have less than 50% reduction in wound size over a four-week's treatment period with the standard of care is considered hard to heal.

One of the ways in a clinical trial how we try to identify these population, hard to heal ulcers. We use a running period to exclude patient who doesn't fit into this category. A running period is typically two or four weeks, wherein patients who agree to participate in a trial by signing the consent undergoes a standard of care treatment.

This is happening before the study treatment is started. During this period, their wound size is measured. If a patient, for example, has more than 35% reduction in size, in say, two weeks, the patient will be considered not eligible for participation in a clinical trial.

You may ask me why this is important. It's important because if a study enrolled too many patients with a wound that will heal on its own, the study will have high percentage of what we call it as a placebo responders.

For example, you conduct a trial with 300 patients. Half of these patients receive investigational product, your new drug. Other half receives a placebo. At the end of the study, 50% of these patients who received investigational product achieved a complete wound closure, 100% closure.

Say, 75 patients out of 150 who received the study drugs achieved a complete closure. Which is a good number, by the way. At the same time, you have 40% of the patients who received placebo also achieved a complete wound closure. That's about 60 out of 150 patients.

Because of these high placebo responders, the data or the difference is only 10%, which is 50% versus 40% between the two treatment groups.

Despite showing that 50% complete wound closure rate with this new growth treatment, Dr Kazempour, with his statistical expertise, will tell you the results are not statistically significant. This high placebo responder rate is one of the major reasons DFU studies are unable to show the improvement over the standard of care alone.

Gaydos: Then, what role does standard of care play in the success of these wound-healing clinical trials?

Dhody: During the trial, patient receives a study treatment or a placebo, in combination with the standard of care. There is a lot of disparity between the standard of care, and specifically, when you conduct a global clinical trial, the trial that is done in multiple countries. The level of care often varies from medical center to medical center. This can have a real impact on your outcome.

For example, the standard of care may call for a dressing to maintain a moist wound environment. It also requires debridement offloading. Debridement is cleaning up the dead or a necrotic tissue from a wound.

The question is, although the standard of care requires debridement, but to what extent does the investigator or physician going to debride the wound? What will be the frequency of debridement? These are the challenging aspects to establish as a part of standard of care.

The another aspect, as I mentioned, is offloading. Offloading means taking the pressure off the weight-bearing part of your feet. If the wound is on the plantar surface, you need to make sure pressure is off the wound, or it will not heal. The patient compliance comes into play here. To what extent is the patient following the doctor's order to stay off his or her feet, it matters.

There are devices, like boots, that patient can wear to help with offloading. That's another variable. Do all patients have access to these devices? Ideally, you want all study participants to receive the same standard of care. As much as possible, we try to specify the standard of care that should be provided.

This is one of the biggest challenges when we talk about, in terms of the clinical trial design, for chronic wounds or the diabetic foot ulcers.

Gaydos: Dr Dhody, can you elaborate on reducing noise in clinical trials? How exactly would this be done?

Dhody: That's another good question, Jackie. The variability in each patient's individual characteristics creates variability or noise in the clinical trial data. Noise can also be related to other factors. For example, unequal distribution of patients with the specific characteristics, such as wound size, location, or the severity, between the two treatment groups.

Noise in a clinical trial data makes it difficult to detect the true difference between the treatment groups. That is, between the patients who receive study treatment, or patients who receive placebo. As I mentioned before, one of the most urgent examples is the placebo response, which varies significantly between patients, and is a major source of noise in clinical trial data in a DFU setting.

I will let Dr Kazempour talk about how to minimize variance or noise during the statistical analysis.

Gaydos: Dr Kazempour, how can study design address some of these issues?

Kazempour: That's a very good question. That's the million‑dollar question. [laughs] A lot of sponsors are trying to do it. As Dr Dhody. over and over, in different ways mentioned, and I myself at the beginning, I said, that we have to find a way to reduce the noise.

Also, I want to refer, again, Dr Dhody mentioned that when we have 300 subjects—150, 150—150 getting active, 150 getting placebo or control, at the end, the observed difference between the two—let's say 10% more patients in the active arm are healed, with that 100% heal definition that I mentioned earlier, compared to that of control—this 10% may not produce statistical significant key-value which dictates the approval.

Why do we have 10%? When we designed the trial, we thought we are going to have 15%. We powered the study for 15%. That means, we calculated how many subjects we need in this study? Need 300, in order to detect 15% difference.

When the noise were not taken off control, we were not taking into account those noises, then no longer that the true difference, maybe 15%—that means that 15%—improvement compared to that of the control—we won't be able to see that.

Your question is, how can we design in such a way to reduce this potential noise so we can see the signal better. The way that we can, and we are—Amarex has done so many of them, and well‑experienced in doing so—to try to make sure that there are balanced number of subjects in each group.

For example, if it happens that we are designing a clinical trial, let's say with wound size between—I'm throwing a number between 1 and 15-centimeter square—and it happened that large number of wound size being in the control arm, which they heal faster compared to that of the active arm, then the delta of 15% no longer...We may have 10% difference.

How we can do that? How we can reduce the noise, or design the trial better? The way that we look at it is, at the beginning of the trial design, we'll look at the earlier stage of that clinical trial that they have conducted, to see what is the characteristics of that particular treatment. Based on that, we will look at to use—I'm going to use statistical language—use stratification factors.

What does stratification factors means? It makes sure that balanced number of subjects of a particular characteristic to be in both active group and control group. For example, if one of the factors is the wound size, we'll make sure that all the wound size between one and three centimeters square, for example, are not, or majority of them, are not in one arm.

Paying attention at the design stage with those factors that, potentially, will impact the end result, that is helping to be able to achieve statistical significance, with pre-specified sample size.

Another thing that we are doing, and lately, we are using—and FDA accepts that—we are referring to something as adaptive clinical trial design. Adaptive clinical trial design is one of those ways that, it takes in account the noise during the trial and in the middle of the trial.

We may adjust for the sample size. We may adjust some pre‑specified factors, in order to control the noise that are there. As I mentioned, the rule of game is 100% closure, with the way that FDA has it. We have to really use those metrics, those noise reduction factors that are in our arsenal, in the statistical analysis systems, to reduce the noise and show the efficacy.

Gaydos: Dr Kazempour, previously Dr Dhody had talked about the disparity within standard of care, and its impact on the successfulness of a trial. Along that line of thought, would a similar issue in wound imaging and measurement need to be addressed?

Kazempour: Dr Dhody was right. The answer to your question is, absolutely yes. To measure wound closure, it used to be only focused on the wound size. Nowadays, we are able to measure the depth of the wound as well. For example, there are technologies out there now that allows us to do that.

Let me give you one example. If you look at two wound size, and both of them being, let's say, five-centimeter square, but one of them, the depth is one millimeter, and the other one, the depth is five millimeters. Although on the surface, they are equal wound size, they are similar wound size, but the one that is longer, the depth is larger, this one will not heal during that pre-specified 16 weeks' timeframe that I mentioned.

The one that is only one-millimeter depth will heal faster. These are the ones that using the technology, the imaging technology that is available, it allows us to measure the depth. This is one of those noise that for long time, we knew about it. We didn't know how to take care of it by taking these new technology.

Again, you have to pre-specify them at the designer stage. Probably you have heard me and Dr Dhody over and over talking about the designer stage. When you are randomizing them, you have to make sure these noises have been taken care of. Yes, it is possible to do that.

When we are designing the study, we have to take into account ways to minimize these noise. Yes, by those who are doing that. As I said, Amarex has been involved in the approval process. The way that we have done it, we have looked at that product at the earliest stage before designing and mentioning those factors. Then presenting ways that those noises can be reduced.

Gaydos: Now, Dr Kazempour and Dr Dhody, although we're not in an ideal world. In an ideal world, if clinical trials were uniform in their standard of care, and wound imaging or measurement included more than one endpoint, and had a balanced patient population selection, what do you think the DFU wound healing options would look like?

Dhody: Let me start by saying that the hard-to-heal wounds in a diabetic foot ulcer are a challenge for a patient, the physician, and overall healthcare system. Ultimately, it's all about how we can improve the patient outcomes.

In an ideal world, we would like to see greater variety of therapies available to the patient with fewer infections incidents and fewer amputation. Amarex has been involved in development of several products, including nitric oxide, oxygen therapy, growth factors, bioengineered technologies, cell-based graft, collagen-based products, and so on.

One of the ways we would like to see is how these variety of therapies be accessible to a patient, is ensuring the clinical trials have a successful outcome. One of the ways...I would have Dr Kazempour speak about how we can expand the use of biomarkers in the wound healing studies to give it a better chance of success.

Kazempour: Thanks, Dr Dhody. Yes, if we are in the ideal world, how the DFU trials will look like, I would say that they would look like other clinical trials that we are doing.

In some areas, we are taking advantage of other endpoints. We used to call them...During HIV trials many years ago, a term that was used, surrogate endpoint. Surrogate endpoint is a replacement for the endpoint that is accepted. 100% wound closure is the accepted one.

By the way, FDA is accept two endpoints. Both of them depends on the definition of 100%, by the way. One of them, time to 100% closure. The other one, percent of closure. Remember, I said that the 15% difference between the control and active, but also, FDA accept time to closure.

If you have a product that reduced the closure from—I'm throwing a number—from, let's say, 16 weeks to only 14 weeks, versus that of the control arm, 16 weeks—two weeks' difference—as long as it comes out to be statistically-significant, FDA will accept it.

That statistical significance really depends on the pre-specified number of subjects in the trial. Meaning, if we calculated we thought we need 300 subjects, and during the trial, due to some unknown noise, end of that, we needed 350 subjects or 302 subjects, their result may not become statistically significant. That trial, from the FDA and regulatory point of view, failed.

That's what I mentioned, that we need to be able to use if we can, obviously, adaptive trial design as I mentioned, and also come up with new endpoints that are, both the regulatory agency and the investigators, they should come up with new endpoints that are acceptable to reduce this burden.

For example, in oncology, that in cancer, we have multiple acceptable endpoints. What do I mean, endpoints? You can use it as outcome. For example, one of them is overall survival, comparing to active and control, overall survival. Another one is progression-free survival. That's another endpoint. Then, the other endpoint is something like overall response rate. That's another endpoint.

These are the endpoints that clinically meaningful and regulatory-wise, accepted, that we can use them, in order to show a product is working, compared to control.

The bottom line is, we need to find a way with the investigators and sponsors, and regulatory agencies, to come to some other acceptable endpoint, so we design our clinical trial in such a way that we can come up with successful trial.

Actually, there has been an ongoing conversation with the FDA in use of these new endpoints. Hopefully, we're going to come up with that one. While we are not there yet, we have to really pay attention to the design—how we are designing our trial, and how we are trying to reduce the noise.

Gaydos: Dr Kazempour, it was great to learn about the regulatory aspects involving clinical trials. Dr Dhody, it was fantastic learning about the clinical affairs side. Thank you both so much for joining us today.

Thank you to the listeners for tuning in. Be sure to check out the woundcarelearningnetwork.com for more podcasts, articles, and videos on various topics in wound care.

Also, be sure to subscribe to our podcast on Apple Podcasts, Spotify, Spreaker, or wherever you normally listen to your podcast. We hope you tune into our next podcast and thank you for listening to, Speaking of Wounds.