The Take Home

The Take Home: CRT

Cardiovascular Research Technologies (CRT) was held March 2 to 5 in Washington, D.C., and Cardiology Today’s Intervention covered it live. The meeting offered up a number of topics for endovascular and interventional cardiologists, including a Vascular Leaders Forum held to discuss mortality concerns related to paclitaxel-coated devices. Weighing in on these topics are Cardiology Today’s Intervention Editorial Board Member Gary M. Ansel, MD, FACC, from OhioHealth in Columbus; Jay Giri, MD, MPH, from Hospital of the University of Pennsylvania; Michael S. Lee, MD, FACC, FSCAI, from UCLA Medical Center; and Cardiology Today’s Intervention Editorial Board Member Christopher J. White, MD, MACC, MSCAI, FAHA, FACP, FESC, from the John Ochsner Heart and Vascular Institute.

Gary M. Ansel

Gary M. Ansel, MD, FACC

Structural heart advances appeared to make the biggest impact at the meeting. In the LAMPOON trial, there was a decrease in the number of hospitalizations for people treated with the MitraClip (Abbott) for mitral valve regurgitation. That is just fantastic. The researchers used the LAMPOON technique to split the anterior mitral valve leaflet before the procedure to prevent left ventricular outflow tract obstruction.

Researchers analyzed data from 30 patients (median age, 76 years; 73% women) with high or extreme risk for surgical mitral valve regurgitation who would be ineligible for therapy. The primary endpoint, defined as procedure success at exit from the cath lab, occurred in 73% of patients.

The growth in the vascular realm at CRT has been noticeable. The fact that there was a town hall with the FDA at the end shows that this is one of those meetings where the last day is almost going to be as important as any of the days in between that, which is not usually the case.

There are so many great sessions here. I am not on our structural heart team, so I sat in on a lot of those sessions to get an update. It was interesting to see how far the field has continued to advance on an almost quarterly basis. Everything becomes more and more percutaneous, moving towards the potential for outpatient treatment.

Being a cardiology-based peripheral vascular interventionist, I was very interested in the pre-meeting International Vascular Leaders Forum (VLF) on drug-based devices and the more limited FDA town hall here at CRT. The VLF brought together almost 100 international experts to discuss a recent summary-level data meta-analysis that suggested a potential for a delayed mortality risk with drug-based devices.

Our discussions certainly were reminiscent of 2006 when a couple of meta-analyses appeared to demonstrate mortality risks with paclitaxel and -limus technology in the coronaries, which was debunked with patient-level data. Unlike the coronaries, where stent thrombosis was recognized as the potential signal, there was no such signal or mechanism described in the PAD meta-analysis.

The physicians and the FDA are concerned about any signal whether related to paclitaxel or treatment bias that may be occurring. Recent large population, patient-level meta-analyses from CMS databases do not demonstrate any signal, so we are left with looking at the patient-level data from the trial population to see if there is really anything behind these numbers. Without a doubt, patient safety comes first, but those concerns can be due to treatment or withholding treatment as well.

Jay Giri

Jay Giri, MD, MPH

I was most interested in discussions at the Vascular Leaders Forum, where stakeholders got together to examine the controversy around drug-eluting balloons and drug-eluting stents for peripheral artery disease. That has been the hottest topic in the peripheral artery arena since a controversial meta-analysis was published in December (Katsanos K, et al. J Am Heart Assoc. 2018;doi:10.1161/JAHA.118.011245) showing a long-term elevated mortality risk for patients with PAD treated with paclitaxel-coated devices. I was struck by the fact that a lot of people got together including FDA, industry representatives and of clinical leaders in the field to hash out what the next steps are going forward.

As an objective observer of the proceedings, I tend to think that the Katsanos meta-analysis raised an important point. It is great we are now looking into it and it is not so clear what the end results are going to be. Why would that have ever happened if somebody didn’t push that button to write that meta-analysis in the first place?

There has been a fair amount of justified criticism about how much you can trust a study-level meta-analysis. However, the response to it has been very positive because all stakeholders got together and said we have to look at these data more closely.

I was struck by the fact that the preliminary data that are coming out are mixed. We have had a couple analyses come out from individual companies and real-world observational data sets that suggest that things look OK, but the FDA representative at the meeting stated their own analysis also showed mortality risk, so one cannot necessarily throw out what Katsanos had observed in his meta-analysis.

As a researcher who does a lot of meta-analyses, there is no question that if you analyze individual patient data, it is always better than if you analyze published data. But a key question to ask ourselves is: how frequently does the individual patient data meta-analysis correspond with the study-level analysis? It is 80% of the time. It is in statistical literature. It is not the craziest thing to start off with that kind of gross cut, which does not take a lot of time and resources to bring something to light, and 80% of the time, you’re going to be right.

We do not want these devices to have a safety signal because we know they help patients and they are efficacious. While there are a lot of flaws in study-level meta-analyses, you always have to keep in mind that even when you cut it deeper with patient-level data, 80% of the time the original finding is indeed correct. It is going to be interesting to see what happens.

It is going to be interesting to see what happens. As a cardiologist who has been watching this story, this is going to be very interesting and we’ll probably know by summer where it settles out.

Michael S. Lee

Michael S. Lee, MD, FACC, FSCAI

The SAFE-DCB study found that the Lutonix (BD) paclitaxel-coated balloon was safe in peripheral arteries at 1 year. It included a single-arm multicenter registry of 966 patients (mean age, 69 years; 43% women) from 74 U.S. centers in April 2015 to December 2016. It assessed the safety of the paclitaxel-coated balloon to treat femoropopliteal disease. The primary effectiveness endpoint was freedom from target lesion revascularization at 12 months. The primary composite safety endpoint was freedom from device- or procedure-related and perioperative death at 30 days, target limb amputation above the ankle and target lesion revascularization.

The primary effectiveness endpoint at 1 year occurred in 89.5% of patients (95% CI, 87.3-91.3), which is similar to other Lutonix studies. In terms of the primary safety endpoint, 98.2% (95% CI, 97.2-98.9) of patients were free from any event at 30 days. At 1 year, the mortality rate was 7.1% (95% CI, 5.6-8.8), the most common cause of death being cardiac arrest with or without respiratory failure in 11 patients, MI with respiratory failure in five patients and septic shock in nine patients. At 1 year, 86.9% (95% CI, 84.5-88.9) were free of target vessel revascularization. It does suggest that there is good efficacy in terms of mortality of 7% at 1 year seems to suggest that there is safety. Unfortunately, there was no comparison arm in this study.

The BIO-RESORT trial looked at the 3-year safety and efficacy of three drug-eluting stents in 3,514 patients who required PCI. One was the biodegradable polymer-coated Orsiro (Biotronik) DES, then the other one was the Synergy (Boston Scientific) stent, another one was the Resolute (Medtronic) stent. There was no significant difference amongst the three. This is good news for Orsiro because it had no difference from the other standard therapy that is being used right now.

The AngelMed Guardian implantable cardiac monitor from the ALERTS study failed to meet its primary endpoint by identifying future MIs. That would have been a gamechanger had it been positive, but unfortunately it did not pan out.

There was a post-hoc analysis of the ATLAS-ACS-2 TIMI-51 and the COMMANDER HF trials that found that rivaroxaban (Xarelto, Janssen) can reduce adverse events in patients with ACS or heart failure on dual therapies. Patients who were on rivaroxaban twice per day had a reduction in adverse events and all-cause mortality. This was a patient-level meta-analysis of the two trials. There was a reduction in adverse events in patients with ACS or chronic HF on dual antiplatelet therapies. That is thought-provoking, though it’s unlikely that a meta-analysis would lead to an FDA approval.

Christopher J. White

Christopher J. White, MD, MACC, MSCAI, FAHA, FACP, FESC

I am really interested in the DCB concerns in the periphery. One of the most important developments of this meeting was the Vascular Leaders Forum held to discuss a mortality signal observed in paclitaxel-coated devices used for patients with PAD. I want to make sure we speak clearly about this signal and this risk.

DCBs were a major leap forward in our technology. There have been a bunch of trials that have shown they work a lot better than anything else, and they have been a godsend for doctors and patients.

The problem is that trials were powered for the functional endpoint, patency. They were not powered for mortality because relatively few people with claudication die.

The researchers who performed a study-level meta-analysis did not have patient-level information and do not know why people died. They do not even know what device a patient who died was treated with. They do not know how many balloons were put in. Some patients you treat with three balloons, some you treat with one. There is a lot of detail about the patients themselves that a study-level meta-analysis just does not have.

The researchers did find a mortality increase and those of us who checked the math found it was basically correct, but there are a couple of important things to understand about it. No. 1 is that there was no patient-level data. No. 2 is the way they had to count deaths was not the standard way we look at mortality, which is usually in a time-to-event analysis.

There is nothing wrong with that, but it means that one cannot take that study to the bank. That is not enough to make a decision. It is our responsibility to take that signal seriously and now do the patient-level meta-analysis.

The FDA is doing that and VIVA is doing that. Several groups are using the data that had been filed with the FDA to reanalyze the patient-level meta-analysis that companies had all given approval to mix their data together. It is very important that we get to the answer.

What patients should understand about this is that there is never been a signal of danger with paclitaxel at this level. I recently heard an oncologist speak who said that paclitaxel is so safe, we give it to pregnant women; pregnant women with malignancies get paclitaxel at higher doses than found in peripheral DCBs and DES. We do not even give vitamins to pregnant women. She actually said, “Given the level of safety with this drug for so many years, we are very confident that it is a safe drug.”

Then the question is if it is such a safe drug, why the signal? There are a couple of good explanations for why the signal could happen. No. 1, there are some biases that are introduced into randomized trials. The first one is that in the experimental arm, the companies tenaciously followed tenaciously the patients who got the device. The company cannot count the patient if they cannot get the follow-up. What that means is if you are in the experimental group and you die for any reason, even if you get hit by a bus, they figure that out. If you are in the control arm, you do not get chased so hard because that control arm is not relevant to their business. They just do not spend as much money or time on the control arm. You have an ascertainment bias that says we are going to find every dead person in the experimental group, we may miss a few out of this group, when you are only talking about one, two or three out of 150, you miss one, that can be significant, so that can explain why you would add up more mortalities even though there may not be a real safety problem with paclitaxel. Study-level meta-analyses will still be subject to that bias because the patients in the device groups would have been followed more intensively. That is a bias that would need to be recognized in the patient level meta-analysis.

The second bias is that mortality in PAD is not related to your femoral artery patency, but to the burden of atherosclerosis that causes stroke and MI. The leg stent does not make you live longer. What keeps you from dying is preventing a MI or stroke, and that comes from medical therapy, eg, statins, antiplatelet therapy and BP control. The control group with poorer patency has a lot more symptoms than the experimental group. In the experimental group, people feel good, they miss their clinic appointments, they do not come back because they are walking and they feel good. The control arm that got the plain balloon has restenosis, so they have recurrent events and come back to see their doctor all the time. Every time you come to see me, I say, are you taking your statins and your BP medicine? Are you not smoking? The control arm has more interactions with the health care system. If mortality is not affected by patency in your leg, but by medicines that keep your heart working, the control group gets better care. The bias is that better medical care happens in the control group, so when you are only counting a few mortalities, you end up seeing a numerical difference that may not be due to the danger of the drug.

Those are the two best answers for why there may be a mortality signal in a drug that is not dangerous. They are both very plausible. Neither one of those biases will be extinguished by the patient-level analysis because the people are going to have lived longer. Hopefully the patient-level meta-analysis will allow us to examine the cause of death and see if that can be related to any part of the signal, but the important thing for everyone to understand is that no one is in denial about this signal. We have to dig down on this.

The real question is what do you do about it? Because we do not have a high level of confidence of danger of these devices and because they are so beneficial, we ought to be able to talk to patients about the potential risk and the potential benefits. If patients want it, we should continue to use the device with informed consent.

Two European trials have stopped enrolling patients due to an abundance of caution. There are going to be some patients that will not get benefit from these devices while they wait, and while the trials are paused. I do not know if there is a right answer, but I do not think we are exposing patients to great risk by continuing to treat if they understand the potential risks and benefits.

As one of my patients said: You tell me what you think the risks are, you tell me what I can expect benefit-wise and I can decide whether I want to do this or not. The right answer is to have good informed consent during this period of uncertainty where doctors and patients can decide what they want to do. It just takes the hysteria out of it.

Also of interest was a session on endovascular therapy for acute ischemic stroke. We now have therapies that are proven beyond a shadow of a doubt to greatly impact stroke care. Unfortunately, these are endovascular therapies, which require skill to use, and the skill is largely held in the neuroradiology specialties.

Neuroradiologists are few and far between, and that’s because their daytime work only requires one of them to work at a big hospital. Well, that neuroradiologist cannot take calls every night for stroke, so how do you provide care for stroke around the clock? It turns out that in the major medical centers in the urban areas, we have enough neuroradiologists to provide this care, but away from the urban centers, we do not have anybody to provide this care.

There are two ways to solve this problem. One is you can build a medical transport system that would fly people from wherever they had their stroke into the medical center. Now, that system does not exist today and what we learned during the STEMI trials was that it just was not feasible to transport that many patients that distance. What we did for STEMIs was we made special programs in hospitals without bypass capabilities and performed angioplasties and stenting in those hospitals for acute MIs and then, if necessary, transferred patients in a less time-sensitive manner to the medical centers. Many of us are now proposing we do that for stroke: We should simply train people who are carotid-stent competent; who can safely put a catheter in the carotid artery to treat the acute stroke.

Interventional cardiologists have been treating strokes since 1997 in my hospital because we only had one neuroradiologist and he could not take calls every night. The cardiologists that were credentialed for carotid stents helped him. Subsequent to that, more neuroradiologists came aboard, so I do not treat acute strokes anymore, but for 20 years, I did and our outcomes were as good as anyone’s.

There was a famous Nobel Laureate named Amartya Sen. He talked about why famines occur. There were two reasons: No. 1 is there is not enough food, and No. 2 is there is enough food, but you cannot distribute it and it is not available to the people who are hungry. In reading that literature, it struck me that stroke is in that situation. We have plenty of people who treat strokes, but they are just not in the right place. My neuroradiologist is not going to go live in Layfette, Louisiana because there is not enough daytime work for him there.

The only way to address this geographic disparity is by working together to put together these teams of people. They can be surgeons, they can be cardiologists, they can be neurologists. They can come from multiple specialties and work together as a team.

We also have the ability with telemedicine to have a neurologist at the center talk us through a patient. I do not know all the neurological details of stroke and I do not pretend to, but I can open an artery anywhere in the body. I need the neurology knowledge, but now with telemedicine, I do not even need a neurologist in Lafayette. I can get this specialist via telemedicine who can examine that patient and tell me which artery to open, and I can do that, and once that is done, then we can take 2 hours to transfer him. We bought all the time.

That is the big argument that is happening today. The problem is the specialists in the major medical centers, they see themselves as experts, and they do not want to allow nonexperts into their area.

A paper published a few years ago looked at California. Sixty percent of Californians did not live within 1 hour of a stroke center. They were spread out. The stroke centers are all in the major metropolitan areas, so if you live more than an hour away, you lose the golden hour when stroke intervention needs to happen, so it is a fairly simple solution. But it becomes very hard to have the logical conversation about the belief that we are going to hurt patients because the “wrong” people are taking care of them. We are running into all of these stakeholder issues that make it harder to work together. It is going a lot slower than it should. – by Darlene Dobkowski

Disclosures: Ansel reports he is on the advisory board for Boston Scientific, Medtronic, Philips, SurModics and W.L. Gore and Associates. Giri reports he is on the advisory board for AstraZeneca. Lee and White report no relevant financial disclosures.

Cardiovascular Research Technologies (CRT) was held March 2 to 5 in Washington, D.C., and Cardiology Today’s Intervention covered it live. The meeting offered up a number of topics for endovascular and interventional cardiologists, including a Vascular Leaders Forum held to discuss mortality concerns related to paclitaxel-coated devices. Weighing in on these topics are Cardiology Today’s Intervention Editorial Board Member Gary M. Ansel, MD, FACC, from OhioHealth in Columbus; Jay Giri, MD, MPH, from Hospital of the University of Pennsylvania; Michael S. Lee, MD, FACC, FSCAI, from UCLA Medical Center; and Cardiology Today’s Intervention Editorial Board Member Christopher J. White, MD, MACC, MSCAI, FAHA, FACP, FESC, from the John Ochsner Heart and Vascular Institute.

Gary M. Ansel

Gary M. Ansel, MD, FACC

Structural heart advances appeared to make the biggest impact at the meeting. In the LAMPOON trial, there was a decrease in the number of hospitalizations for people treated with the MitraClip (Abbott) for mitral valve regurgitation. That is just fantastic. The researchers used the LAMPOON technique to split the anterior mitral valve leaflet before the procedure to prevent left ventricular outflow tract obstruction.

Researchers analyzed data from 30 patients (median age, 76 years; 73% women) with high or extreme risk for surgical mitral valve regurgitation who would be ineligible for therapy. The primary endpoint, defined as procedure success at exit from the cath lab, occurred in 73% of patients.

The growth in the vascular realm at CRT has been noticeable. The fact that there was a town hall with the FDA at the end shows that this is one of those meetings where the last day is almost going to be as important as any of the days in between that, which is not usually the case.

There are so many great sessions here. I am not on our structural heart team, so I sat in on a lot of those sessions to get an update. It was interesting to see how far the field has continued to advance on an almost quarterly basis. Everything becomes more and more percutaneous, moving towards the potential for outpatient treatment.

Being a cardiology-based peripheral vascular interventionist, I was very interested in the pre-meeting International Vascular Leaders Forum (VLF) on drug-based devices and the more limited FDA town hall here at CRT. The VLF brought together almost 100 international experts to discuss a recent summary-level data meta-analysis that suggested a potential for a delayed mortality risk with drug-based devices.

Our discussions certainly were reminiscent of 2006 when a couple of meta-analyses appeared to demonstrate mortality risks with paclitaxel and -limus technology in the coronaries, which was debunked with patient-level data. Unlike the coronaries, where stent thrombosis was recognized as the potential signal, there was no such signal or mechanism described in the PAD meta-analysis.

The physicians and the FDA are concerned about any signal whether related to paclitaxel or treatment bias that may be occurring. Recent large population, patient-level meta-analyses from CMS databases do not demonstrate any signal, so we are left with looking at the patient-level data from the trial population to see if there is really anything behind these numbers. Without a doubt, patient safety comes first, but those concerns can be due to treatment or withholding treatment as well.

PAGE BREAK
Jay Giri

Jay Giri, MD, MPH

I was most interested in discussions at the Vascular Leaders Forum, where stakeholders got together to examine the controversy around drug-eluting balloons and drug-eluting stents for peripheral artery disease. That has been the hottest topic in the peripheral artery arena since a controversial meta-analysis was published in December (Katsanos K, et al. J Am Heart Assoc. 2018;doi:10.1161/JAHA.118.011245) showing a long-term elevated mortality risk for patients with PAD treated with paclitaxel-coated devices. I was struck by the fact that a lot of people got together including FDA, industry representatives and of clinical leaders in the field to hash out what the next steps are going forward.

As an objective observer of the proceedings, I tend to think that the Katsanos meta-analysis raised an important point. It is great we are now looking into it and it is not so clear what the end results are going to be. Why would that have ever happened if somebody didn’t push that button to write that meta-analysis in the first place?

There has been a fair amount of justified criticism about how much you can trust a study-level meta-analysis. However, the response to it has been very positive because all stakeholders got together and said we have to look at these data more closely.

I was struck by the fact that the preliminary data that are coming out are mixed. We have had a couple analyses come out from individual companies and real-world observational data sets that suggest that things look OK, but the FDA representative at the meeting stated their own analysis also showed mortality risk, so one cannot necessarily throw out what Katsanos had observed in his meta-analysis.

As a researcher who does a lot of meta-analyses, there is no question that if you analyze individual patient data, it is always better than if you analyze published data. But a key question to ask ourselves is: how frequently does the individual patient data meta-analysis correspond with the study-level analysis? It is 80% of the time. It is in statistical literature. It is not the craziest thing to start off with that kind of gross cut, which does not take a lot of time and resources to bring something to light, and 80% of the time, you’re going to be right.

We do not want these devices to have a safety signal because we know they help patients and they are efficacious. While there are a lot of flaws in study-level meta-analyses, you always have to keep in mind that even when you cut it deeper with patient-level data, 80% of the time the original finding is indeed correct. It is going to be interesting to see what happens.

It is going to be interesting to see what happens. As a cardiologist who has been watching this story, this is going to be very interesting and we’ll probably know by summer where it settles out.

PAGE BREAK
Michael S. Lee

Michael S. Lee, MD, FACC, FSCAI

The SAFE-DCB study found that the Lutonix (BD) paclitaxel-coated balloon was safe in peripheral arteries at 1 year. It included a single-arm multicenter registry of 966 patients (mean age, 69 years; 43% women) from 74 U.S. centers in April 2015 to December 2016. It assessed the safety of the paclitaxel-coated balloon to treat femoropopliteal disease. The primary effectiveness endpoint was freedom from target lesion revascularization at 12 months. The primary composite safety endpoint was freedom from device- or procedure-related and perioperative death at 30 days, target limb amputation above the ankle and target lesion revascularization.

The primary effectiveness endpoint at 1 year occurred in 89.5% of patients (95% CI, 87.3-91.3), which is similar to other Lutonix studies. In terms of the primary safety endpoint, 98.2% (95% CI, 97.2-98.9) of patients were free from any event at 30 days. At 1 year, the mortality rate was 7.1% (95% CI, 5.6-8.8), the most common cause of death being cardiac arrest with or without respiratory failure in 11 patients, MI with respiratory failure in five patients and septic shock in nine patients. At 1 year, 86.9% (95% CI, 84.5-88.9) were free of target vessel revascularization. It does suggest that there is good efficacy in terms of mortality of 7% at 1 year seems to suggest that there is safety. Unfortunately, there was no comparison arm in this study.

The BIO-RESORT trial looked at the 3-year safety and efficacy of three drug-eluting stents in 3,514 patients who required PCI. One was the biodegradable polymer-coated Orsiro (Biotronik) DES, then the other one was the Synergy (Boston Scientific) stent, another one was the Resolute (Medtronic) stent. There was no significant difference amongst the three. This is good news for Orsiro because it had no difference from the other standard therapy that is being used right now.

The AngelMed Guardian implantable cardiac monitor from the ALERTS study failed to meet its primary endpoint by identifying future MIs. That would have been a gamechanger had it been positive, but unfortunately it did not pan out.

There was a post-hoc analysis of the ATLAS-ACS-2 TIMI-51 and the COMMANDER HF trials that found that rivaroxaban (Xarelto, Janssen) can reduce adverse events in patients with ACS or heart failure on dual therapies. Patients who were on rivaroxaban twice per day had a reduction in adverse events and all-cause mortality. This was a patient-level meta-analysis of the two trials. There was a reduction in adverse events in patients with ACS or chronic HF on dual antiplatelet therapies. That is thought-provoking, though it’s unlikely that a meta-analysis would lead to an FDA approval.

PAGE BREAK
Christopher J. White

Christopher J. White, MD, MACC, MSCAI, FAHA, FACP, FESC

I am really interested in the DCB concerns in the periphery. One of the most important developments of this meeting was the Vascular Leaders Forum held to discuss a mortality signal observed in paclitaxel-coated devices used for patients with PAD. I want to make sure we speak clearly about this signal and this risk.

DCBs were a major leap forward in our technology. There have been a bunch of trials that have shown they work a lot better than anything else, and they have been a godsend for doctors and patients.

The problem is that trials were powered for the functional endpoint, patency. They were not powered for mortality because relatively few people with claudication die.

The researchers who performed a study-level meta-analysis did not have patient-level information and do not know why people died. They do not even know what device a patient who died was treated with. They do not know how many balloons were put in. Some patients you treat with three balloons, some you treat with one. There is a lot of detail about the patients themselves that a study-level meta-analysis just does not have.

The researchers did find a mortality increase and those of us who checked the math found it was basically correct, but there are a couple of important things to understand about it. No. 1 is that there was no patient-level data. No. 2 is the way they had to count deaths was not the standard way we look at mortality, which is usually in a time-to-event analysis.

There is nothing wrong with that, but it means that one cannot take that study to the bank. That is not enough to make a decision. It is our responsibility to take that signal seriously and now do the patient-level meta-analysis.

PAGE BREAK

The FDA is doing that and VIVA is doing that. Several groups are using the data that had been filed with the FDA to reanalyze the patient-level meta-analysis that companies had all given approval to mix their data together. It is very important that we get to the answer.

What patients should understand about this is that there is never been a signal of danger with paclitaxel at this level. I recently heard an oncologist speak who said that paclitaxel is so safe, we give it to pregnant women; pregnant women with malignancies get paclitaxel at higher doses than found in peripheral DCBs and DES. We do not even give vitamins to pregnant women. She actually said, “Given the level of safety with this drug for so many years, we are very confident that it is a safe drug.”

Then the question is if it is such a safe drug, why the signal? There are a couple of good explanations for why the signal could happen. No. 1, there are some biases that are introduced into randomized trials. The first one is that in the experimental arm, the companies tenaciously followed tenaciously the patients who got the device. The company cannot count the patient if they cannot get the follow-up. What that means is if you are in the experimental group and you die for any reason, even if you get hit by a bus, they figure that out. If you are in the control arm, you do not get chased so hard because that control arm is not relevant to their business. They just do not spend as much money or time on the control arm. You have an ascertainment bias that says we are going to find every dead person in the experimental group, we may miss a few out of this group, when you are only talking about one, two or three out of 150, you miss one, that can be significant, so that can explain why you would add up more mortalities even though there may not be a real safety problem with paclitaxel. Study-level meta-analyses will still be subject to that bias because the patients in the device groups would have been followed more intensively. That is a bias that would need to be recognized in the patient level meta-analysis.

The second bias is that mortality in PAD is not related to your femoral artery patency, but to the burden of atherosclerosis that causes stroke and MI. The leg stent does not make you live longer. What keeps you from dying is preventing a MI or stroke, and that comes from medical therapy, eg, statins, antiplatelet therapy and BP control. The control group with poorer patency has a lot more symptoms than the experimental group. In the experimental group, people feel good, they miss their clinic appointments, they do not come back because they are walking and they feel good. The control arm that got the plain balloon has restenosis, so they have recurrent events and come back to see their doctor all the time. Every time you come to see me, I say, are you taking your statins and your BP medicine? Are you not smoking? The control arm has more interactions with the health care system. If mortality is not affected by patency in your leg, but by medicines that keep your heart working, the control group gets better care. The bias is that better medical care happens in the control group, so when you are only counting a few mortalities, you end up seeing a numerical difference that may not be due to the danger of the drug.

PAGE BREAK

Those are the two best answers for why there may be a mortality signal in a drug that is not dangerous. They are both very plausible. Neither one of those biases will be extinguished by the patient-level analysis because the people are going to have lived longer. Hopefully the patient-level meta-analysis will allow us to examine the cause of death and see if that can be related to any part of the signal, but the important thing for everyone to understand is that no one is in denial about this signal. We have to dig down on this.

The real question is what do you do about it? Because we do not have a high level of confidence of danger of these devices and because they are so beneficial, we ought to be able to talk to patients about the potential risk and the potential benefits. If patients want it, we should continue to use the device with informed consent.

Two European trials have stopped enrolling patients due to an abundance of caution. There are going to be some patients that will not get benefit from these devices while they wait, and while the trials are paused. I do not know if there is a right answer, but I do not think we are exposing patients to great risk by continuing to treat if they understand the potential risks and benefits.

As one of my patients said: You tell me what you think the risks are, you tell me what I can expect benefit-wise and I can decide whether I want to do this or not. The right answer is to have good informed consent during this period of uncertainty where doctors and patients can decide what they want to do. It just takes the hysteria out of it.

Also of interest was a session on endovascular therapy for acute ischemic stroke. We now have therapies that are proven beyond a shadow of a doubt to greatly impact stroke care. Unfortunately, these are endovascular therapies, which require skill to use, and the skill is largely held in the neuroradiology specialties.

Neuroradiologists are few and far between, and that’s because their daytime work only requires one of them to work at a big hospital. Well, that neuroradiologist cannot take calls every night for stroke, so how do you provide care for stroke around the clock? It turns out that in the major medical centers in the urban areas, we have enough neuroradiologists to provide this care, but away from the urban centers, we do not have anybody to provide this care.

There are two ways to solve this problem. One is you can build a medical transport system that would fly people from wherever they had their stroke into the medical center. Now, that system does not exist today and what we learned during the STEMI trials was that it just was not feasible to transport that many patients that distance. What we did for STEMIs was we made special programs in hospitals without bypass capabilities and performed angioplasties and stenting in those hospitals for acute MIs and then, if necessary, transferred patients in a less time-sensitive manner to the medical centers. Many of us are now proposing we do that for stroke: We should simply train people who are carotid-stent competent; who can safely put a catheter in the carotid artery to treat the acute stroke.

Interventional cardiologists have been treating strokes since 1997 in my hospital because we only had one neuroradiologist and he could not take calls every night. The cardiologists that were credentialed for carotid stents helped him. Subsequent to that, more neuroradiologists came aboard, so I do not treat acute strokes anymore, but for 20 years, I did and our outcomes were as good as anyone’s.

PAGE BREAK

There was a famous Nobel Laureate named Amartya Sen. He talked about why famines occur. There were two reasons: No. 1 is there is not enough food, and No. 2 is there is enough food, but you cannot distribute it and it is not available to the people who are hungry. In reading that literature, it struck me that stroke is in that situation. We have plenty of people who treat strokes, but they are just not in the right place. My neuroradiologist is not going to go live in Layfette, Louisiana because there is not enough daytime work for him there.

The only way to address this geographic disparity is by working together to put together these teams of people. They can be surgeons, they can be cardiologists, they can be neurologists. They can come from multiple specialties and work together as a team.

We also have the ability with telemedicine to have a neurologist at the center talk us through a patient. I do not know all the neurological details of stroke and I do not pretend to, but I can open an artery anywhere in the body. I need the neurology knowledge, but now with telemedicine, I do not even need a neurologist in Lafayette. I can get this specialist via telemedicine who can examine that patient and tell me which artery to open, and I can do that, and once that is done, then we can take 2 hours to transfer him. We bought all the time.

That is the big argument that is happening today. The problem is the specialists in the major medical centers, they see themselves as experts, and they do not want to allow nonexperts into their area.

A paper published a few years ago looked at California. Sixty percent of Californians did not live within 1 hour of a stroke center. They were spread out. The stroke centers are all in the major metropolitan areas, so if you live more than an hour away, you lose the golden hour when stroke intervention needs to happen, so it is a fairly simple solution. But it becomes very hard to have the logical conversation about the belief that we are going to hurt patients because the “wrong” people are taking care of them. We are running into all of these stakeholder issues that make it harder to work together. It is going a lot slower than it should. – by Darlene Dobkowski

Disclosures: Ansel reports he is on the advisory board for Boston Scientific, Medtronic, Philips, SurModics and W.L. Gore and Associates. Giri reports he is on the advisory board for AstraZeneca. Lee and White report no relevant financial disclosures.