POSTGRADUATE PROGRAM Ventricular Assist: What, When and How William L. Holman

My talk this morning is on ventricular assist devices (VADs) and when to use them, what device to use, and how to implant them.

You may wonder why I am showing you a picture of two men from the turn of the century. Partly, to spice up the morning, and partly to give you a parallel in aviation, in the development of powered flight by man that I hope will put the development of ventricular assist devices in perspective.

This is Wilbur and Orville Wright a few years after their momentous flight in North Carolina. They did not come up with the idea of powered flight or flight by man out of the blue. Obviously, people have been thinking about powered flight and gliding since they first saw birds. Many men had attempted flight before.

This is Otto Lilienthal in one of his gliders. He was a German who gave his life to the development of aviation when one of his hang-gliders dropped about 50 feet. He broke his back and his dying words were, "sacrifices must be made."

As the century turned, other investigators in aeronautics, Octave Chanute in the United States--this is at the shore of Lake Michigan--began to understand the principles of aviation.

Several engineers attempted powered flight. This is Samuel Pierpoint Langley who received a grant from the U.S. Government to develop a powered flight machine. It was launched from a houseboat in the Potomac River and promptly went into the drink. That happened twice, and the Army terminated his funding. It remained for Wilbur and Orville Wright to work out the principles of wing warping, wing lift, and power. Their work led on December 17, 1903 to the historic flight in Kitty Hawk, North Carolina. Within a few years, as shown in this picture taken at Huffman Prairie in their home state of Ohio, they had mastered the basic techniques of aviation. As a science and as a technique, aeronautics was off and running. It wasn't more than a few decades before commercial aviation was born, and powered flight became a standard way to get oneself around the country, and ultimately the world. The point is that when Wilbur and Orville Wright first flew, many people had the preconceived notion that powered flight by man was impossible. Even when they read the reports and saw the pictures that described the Wright brother's achievements, they still did not feel that it was something that would ever be of any great benefit. When these same folks saw the barnstormers flying around the country in the early '20s, suddenly the light went on, and it didn't take too much longer for the technology to gain universal acceptance.

What do we have available in 1999 for mechanical ventricular assist? I will first discuss shorter-term assistance, which I think many people in the audience are interested in applying in their practice. For shorter-term assistance, typically for postcardiotomy support, the big three right now are centrifugal pumps, the Abiomed BVS 5000 System, and the Thoratec VAD. Of the centrifugal pumps, the Biomedicus is the best known. It was never formally FDA tested, and the Biomedicus device requires a fair amount of supervision from the operator. Specifically, the operator is usually a perfusionist or an ICU nurse. We have used these pumps at UAB and have had problems with thromboembolic events, and conversely with bleeding due to anticoagulation and coagulopathy. I don't feel safe running these pumps, even in the immediate postoperative period, without some degree of anticoagulation. One must keep a constant eye on the pump to be sure its not developing thrombus on the impeller blades. But, we have had survivors, and many other groups have had survivors. Generally, the reported survival rate, that is not only weaning the patient off the device, but also discharging the patient from the hospital, runs anywhere from the high 20% range to about 50%. This is a very sick group of patients, yet there are some remarkable successes. Moreover, when the patients do survive, they typically do fairly well. We are not generating a group of cardiac cripples.

The Abiomed BVS 5000 System is a FDA approved device for short term assistance (about 7 to14 days), either as a bridge to transplant or as a bridge to recovery. One drive console can support one or two ventricles. The Abiomed console itself, I find, is easier to run than the intra-aortic balloon pump or Biomedicus console. There is, to my thinking, a major advantage here in that you don't require a trained person to be monitoring the pump continuously. The Abiomed console is quite simple. There is one switch that turns it on and off. The pumps run in an automatic (full-to-empty) mode, and then there's another switch on the console that allows you to decrease the VAD output as the patient is being weaned from the pump. The device design is a bit interesting. There are two polyurethane valves. The upper chamber is a compliance chamber. It fills with blood almost constantly, so it works a little bit like the human atrium. When the lower chamber is filling, the blood just goes straight through. When the lower pumping chamber, which acts like a ventricle, is in systole, then this compliant area of the pump will fill with blood.

The Thoratec device is based on a design that was developed at Penn State University years ago by Pierce and Donache. The pump has a polysulfone case, and inside there is a collapsible sac. The sac is smooth, and is made of segmented polyurethane. The Thoratec pump has two valves. They are mechanical (Bjork-Shiley) design, but with Delrin plastic disks instead of the classic Bjork carbon disk. The Thoratec console is pretty large. This is our console at UAB. We've been able to get rid of the gas tanks, but it is still a formidable enterprise to take the patient for a one mile walk. Most patients aren't strong enough to push the console itself. They need to have someone push it for them. I'll comment on the Thoatec portable driver in a few minutes.

One of the advantages of the Thoratec as well as the Abiomed device is that a single console can support one or both ventricles. There are a variety of cannulation sites that can be used. You can use atrial cannulation for the right or the left VADs. Recently, a right ventricular cannulation technique was described in a report from the University of Arizona. This was published in the Annals of Thoracic Surgery. If you use the Thoratec device or any of the right heart support devices, I recommend that you read about this technique. The left-sided VADs can also receive flow from the left ventricular apex. Left ventricular apical cannulation typically gives you higher VAD flows with less chance of thromboembolism, as compared to a left atrial cannulation. Left ventricular apical cannulation is not the classic way to implant an Abiomed, but it has been reported with success.

For longer-term assistance, what do we have available? FDA approved devices are currently used only as a bridge to transplant. They are not yet approved as permanent implants, but obviously the goal of all of these studies and all of these devices is to develop permanent therapy for patients with profound heart failure. The Thoratec device has been approved for longer-term (> 2 week) use as a bridge to transplantation. We have a patient at UAB now who has been on Thoratec biventricular VADs for more than a year, and is doing just fine. But he is not the record holder. There is a German patient who was successfully transplanted after more than a year of VAD support.

The HeartMate pneumatic and electric left ventricular assist devices have both been FDA approved, and the Baxter Novacor left ventricular assist device is FDA approved.

This slide shows the Thoratec device. To reiterate, it's a paracorporeal device. It is linked to the portable version of their driver in this illustration, and you can see the portable driver also has a rolling trolley. The portable driver is not a lightweight device, but it certainly can be rolled around easily. The portable driver makes the patient much more mobile than the older driver console design.

The HeartMate devices include their pneumatic pump and their electric pump. This is the pneumatic drive console. It is relatively mobile. There is a HeartPak driver which is a smaller, more portable version of the pneumatic driver that is still under investigation. We have used the HeartPak driver. I couldn't send the patient home on the HeartPak because of study restrictions, but it will be possible in the future for patients to live out of the hospital with a portable HeartPak driver. An even more mobile VAD is the HeartMate vented electric device. The motor for the pneumatic HeartMate device is in the drive console, and air goes in and out of the pump. The motor for the vented electric device is actually built into the intracorporeal can. There is a DC brushless motor that circles once for each cycle of the pump. The controller contains the alarms and some of the control mechanisms. The controller is worn like a beeper on the belt. The pump's power comes from batteries that the patient carries. These batteries will keep the pump running for about four hours before they need to be changed. The batteries are easily changed on the fly. The pump itself is placed in the left upper quadrant of the abdomen, and can be intra-peritoneal, pre-peritoneal, or in an inter-rectus sheath pocket. The pump is attached to the left ventricle at the apex, and it pumps blood into the ascending aorta. There is a percutaneous drive line that carries electrical power to the motor and vents the can to the atmosphere.

One of the goals of developing new devices is to get rid of any percutaneous connections because of the inherent problems with wound healing and infection. The HeartMate and Novacor VADs are externally vented devices. We can talk later about the need to get away from pulsatile pumps just because of the problem of having to either vent the VAD or somehow account for the air going in and out of the can as the pump diaphragm moves. The HeartMate electric VAD is attached to the system controller which is attached to an external battery pack. There are actually two of these batteries that the patient wears. Again, this is what the pump looks like. It has a titanium shell or can. There is a porcine valve sewn into the VAD inflow conduit and another porcine valve is sewn into the VAD outflow conduit.

This is about how a patient looks. This slide shows a model wearing the battery packs, but when the patients go out of the hospital, they really have nothing more showing than that.

The Novacor left VAD was approved on the same day as the vented electric HeartMate device. The Novacor VAD is a very reliable left ventricular assist device. Again, it is attached to the apex of the left ventricle and pulls blood out, then pumps it back into the ascending aorta. The patient wears a controller on the belt, and has replaceable batteries that are also worn on the patient's belt. The pump itself in a cutaway view looks like this. There are biologic valves, I believe the valve is made from bovine pericardium, in the VAD inflow as well as in the VAD outflow. The Novacor VAD is driven on electric power. Instead of having a motor per se, such as a spinning DC brushless motor, the Novacor VAD has pusher plates that are attracted to one another by electromagnetic forces when the pump is activated. The pusher plates squeeze the pumping sac, so it is a pulsatile VAD. When the power is cut off, the pusher plates pull apart and the VAD sac will fill with blood.

This is a picture that was loaned to me by Dr. Peer Portner who is the head of Novacor development. You look at this picture, and you really have a hard time figuring out if this person or that person is on a left ventricular assist device. The VAD recipient happens to be the proprietor of this delicatessen in Germany, who was rehabilitated from profound circulatory failure, and is actually back at work.

In terms of other devices, there are some innovative designs and unapproved designs that are on the horizon or close to being approved. The Cardiowest Total Artificial Heart, known initially as the Jarvik heart, was modified by Jack Copeland and Don Olsen in a collaborative effort. The results with the Cardiowest total artificial heart as a bridge to transplant has been quite impressive, and I understand that it is about to go into the final approval process of the FDA. Soon it will be another tool available for the treatment of patients with profound heart failure.

There are some distinct advantages of total artificial hearts over ventricular assist devices in terms of flow, and certain clinical situations that are better served with a total artificial heart. The obvious downside of the total artificial heart is the large amount of prosthetic material that is in the mediastinum. However, more experienced surgeons have been able to get over most of the infection problems, at least as a bridge to transplant, with this device. The other problem is that a total artificial heart must be a totally reliable device. When a ventricular assist device fails, the native heart, unless it is fibrillating, will pump an adequate amount of blood to keep the patient alive. With a total artificial heart, there is no native heart, so if the pump fails, there is no blood pressure.

Axial flow (turbine-driven) VADs are a very interesting and exciting innovation. Axial flow devices are under development by a number of people. Dr. Jarvik is working on one that is nearing clinical trials. Thermo Cardiosystems has teamed up with the Nimbus Corporation, which was responsible for developing the Hemopump (a smaller axial flow device), to design an axial flow VAD. Axial flow devices have the advantage of being relatively small and energy efficient as compared to the larger, pusher-plate type of VADs. The small size and energy efficiency of axial flow devices will allow them to be fully implantable. In other words, the energy source can be transcutaneous energy sent across the skin from opposed wire coils. The small size of the axial pumps really is going to be an advantage. Now, small adults and older children who cannot have an implanted left ventricular assist device just because of size considerations, will be candidates for prolonged support with an implanted device.

Going back for a moment to total artificial hearts, there is the Abiomed device that is on the horizon, and probably going to enter trials in a few years, and the Sarns/3M /Penn State total artificial heart.

Let us consider when to implant these devices. One of the hardest decisions a surgeon must make is when to use a VAD in the postcardiotomy situation. Our experience locally has been that whenever you know that the patient is a high risk patient, and you talk to the family ahead of time about perhaps implanting a ventricular assist device, then it is almost a guarantee that the patient is going to do just fine. On the other hand, occasionally you take a patient to the operating room and their cardiac function coming off bypass is surprisingly poor. Timing the implant of a VAD in these "surprise patients" becomes the difficult question.

When you are beginning to think about postcardiotomy support with a ventricular assist device, start by evaluating the effectiveness of your operation. Is the patient not coming off of bypass because there is something wrong with the operation? Are the grafts patent? Is the prosthetic valve well-seated and working properly? Those are the types of problems that will not be helped with a ventricular assist device. The patient will never get over a bad operation regardless of the support. If you are convinced that your operation was effective, and that you have dealt with the underlying pathology, then you need to ask a few other questions. First of all, are prosthetic valves present? Regardless of the position of the valve in the left side of the heart, if you use a VAD it becomes an important consideration because these valves become extremely thrombogenic if they don't have adequate flow going across them. That problem is relatively easy to solve for a mitral valve prosthesis. All you have to do is use the left ventricular apex for your VAD inflow instead of the left atrium. That will maintain blood flow across the prosthetic valve and minimize thrombus formation. The aortic valve is problematic, and no one has a really good solution other than to use a total artificial heart. If a prosthetic valve is present in the aortic position, and you place a left ventricular assist device and the heart is quite weak, initially the majority of the pumping work will be done by the left ventricular assist device. The amount of blood flowing through the aortic valve will be minimal, and the chance of thrombus formation on the prosthetic valve becomes extremely high. When the ventricle recovers and begins to eject blood through the prosthetic valve, obviously that thrombus on the valve prosthesis may become an embolus. As a consequence, the patient will have a very high chance of suffering a stroke.

So, an aortic valve prosthesis remains a problem. If you have performed an aortic valve replacement and the patient is not doing well, and you want to use a ventricular assist device, about the best you can do right now without committing them to a transplant is to anticoagulate the patient aggressively, and avoid noncardiac organ injury. How can you avoid noncardiac organ injury? I think the first and most important step is to minimize the duration of cardiopulmonary bypass. The duration of bypass has come out in all of the VAD trials as a risk factor for death, if the duration of bypass is prolonged above roughly 4 hours. Remember, you need to get the device in and get the patient adequately supported without four or six hours of cardiopulmonary bypass! Make the determination to implant a VAD early and implant aggressively for the best result. Remember to present a realistic outlook for the family. This recommendation echoes Dr. Jay Zwischenberger's earlier comments today about treating patients with ARDS. This is a group of patients that is extremely sick. Their anticipated mortality without the pump is hovering around 100%. What should you tell the family? That the patient has a device and their blood pressure is good and everything is going to be just fine? NO! You should tell them that you have gotten the patient out of the operating room. That you will see how their heart recovers within the next 48 to 72 hours, and if they wake up without a neurologic deficit, then the patient has a chance. But in the best of hands, you would expect only a 40-50% chance of survival. That is to say getting the patient out of the hospital alive.

With regard to bridge to transplantation patient selection, obviously the patient has to be a transplant candidate without irreversible noncardiac organ injury. The reversibility of organ damage can be very hard to define. The patient must also fail conventional medical therapy before you resort to VAD support.

Destination therapy, which is still under investigation, is an important topic for ventricular assist devices; I think the most important. The patients who are chosen for permanent VAD placement at the present time must have Class IV failure despite an optimal medical regimen, and they must have no contraindications to a rather big operation. At present, these patients should not be transplant candidates. I think the issue of when to use transplantation and when to use an assist device will be a moving target over the next several years, but right now patients who receive permanent VADs should not be transplant candidates, yet they should have acceptable neuropsychologic function and social support. These pumps are not for everyone. The patient must be sufficiently dexterous and intelligent to use the VAD, and have enough social support at home to help care for it.

If you implant a device, how about adjunct operative measures? You should have materials at hand to preclot your conduits. Generally, I use cryoprecipitate or the patient's own blood with thrombin spray. Hemostatic agents, aprotinin is the most notable among these, should be available. Transesophageal echocardiography is very important to assess cannula placement, and to look for severity of aortic insufficiency (which will diminish the effectiveness of left-sided support). If you are using only left-sided support, be sure to look for a patent foramen ovale. Right to left shunting after placing a VAD can profoundly depress the patient's arterial pO₂. Nitric oxide and prostaglandins are useful in patients who are struggling to come off of cardiopulmonary bypass with a left-sided device only.

If the device is going to be in for a very long time, I feel that pericardial reconstruction is useful, especially if you are ever going to remove the pump.

Let's talk for a moment about important pre- and intra-operative decisions when placing a VAD. Will you put the device in with, or without cardiopulmonary bypass? Generally, only the atrial cannulas can be placed without bypass. Should you use uni- or biventricular support? The sicker patients, and this has been worked by Dr. Bob Kormos and others, require biventricular assistance. The easiest way to figure out intra-operatively whether biventricular support is needed, is to place the left-sided VAD and try and take the patient off cardiopulmonary bypass. If the patient comes off bypass nicely, obviously the patient does not need a right VAD. If the patient is struggling, or if the patient doesn't come off bypass, then put in the right-sided device.

Should you use a left atrial or ventricular cannulation? If the patient has a mitral prosthesis, it is pretty much an indication for ventricular apical cannulation. For shorter-term assist in patients without a valve prosthesis, left atrial cannulation is acceptable. If the patient will have the device in place possibly for weeks, he or she is probably better off with ventricular apical cannulation, which typically has a lower thromboembolic rate, at least in the Thoratec experience. Also, ventricular apical cannulation usually gives the patient a higher VAD flow as compared to an atrial cannulation.

Which device should you use? Well, whatever is at hand, I suppose. In terms of how to do the operation; I place the atrial cannula through a purse string reinforced with pledgettes. I place the cannula into the midportion of the atrium where it is not going to suck up against the walls of the atrium itself. This will occlude flow. I then anastomose the outflow cannula to the appropriate vessel, either the pulmonary artery or the aorta.

The apical left ventricular cannula is placed after coring out a piece of the ventricle, and then sewing in a ring and inserting the stem of the VAD. There are some tricks we can talk about later to deal with the infarcted or friable left ventricle. Again, the outflow graft is anastomosed to the ascending aorta after being appropriately sized. The graft is brought down, de-aired, and attached to the pump. The pump, as I mentioned before, sits in the left upper quadrant with a percutaneous connection. The cannulas for the Thoratec VAD actually come out percutaneously, as do the Abiomed VAD cannulas, to attach to the paracorporeal pump.

What are the obstacles to success for these devices? Infection remains an important problem, particularly if the device is in for more than a few days. Thrombus formation is a problem for every pump, more for some than for others. Anticoagulants, I feel, are necessary for all pumps, although the exact prescription varies from device to device. Wear and failure of mechanical systems is a problem, but it is not a big one. These devices are fairly reliable. The current commercially available devices have been well designed. Patient acceptance, particularly with longer term VAD placement, will become an issue. It is something to talk about with a patient who is going to be bridged to transplantation, and of course cost and reimbursement are ongoing issues.

Where are we headed with mechanical circulatory support? Well, chronic support, which is another way of saying permanent placement of VADs, will head towards smaller pumps and totally implanted systems.

To bring us back to the aviation analogy, a few milestones will illustrate the development and application of aeronautic science. In 1903 we had the first mechanically powered flight by man, and by 4/4/27, we had commercial airline service, so it didn't take very long to develop. It did take awhile for the federal government to issue pilots' licenses. It was two days after the first domestic airline service was established (4/6/27) that pilots license #1 was issued. Coast to coast domestic airline service was established in 1929. The first airline stewardess flew in 1930, she was a Registered Nurse. We had instrument-guided solo flight by 1932, Chuck Yeager broke the sound barrier in 1947, and international commercial jet service by British Airways was established in 1952. Like many things, the Europeans had new technology before we did; it took another six years to establish commercial jet service in the United States. By '64, we were landing jets with computers. In contrast, Charles Lindbergh flew over the Atlantic in kind of a daredevil maneuver using very simple instruments. The technology has advanced dramatically and it is now safe and feasible to fly around the world. We are not quite to that point with VADs, but I do believe that is where we are headed.

[1999 Annual Meeting on the Web] [1999 Annual Meeting Program] [1998 Annual Meeting on the Web]