Chapter 39. Percutaneous Treatment of Aortic Valve Disease

Over the last few years, data have accumulated that untreated aortic valve disease, particularly in elderly people, is a problem despite the success of surgery and newer methods of treatment.1–32

The history of percutaneous treatment of aortic valve disease with various types of device goes back to the work of Danish researcher, H.R. Andersen, who in the late 1980s experimented in an animal lab with a balloon-expandable stented valve.15 The technology was later acquired by PVT Company and further developed and later sold to Edwards. The early work was done by Alain Cribier21,22 and subsequently the expandable stented valve was adopted in the United States and further modified. The initial approach taken by Cribier was to insert the valve via the femoral vein and then using a transeptal technique to snake the catheter through the mitral valve and left ventricle to access the aortic annulus. This turned out to be a fairly cumbersome and difficult operation to do with a fairly high mortality rate and an 8.1% stroke rate. At the same time, animal experiments were carried out by Michael Mack, Todd Dewey, and Lars Svensson,17,25 to use a transapical approach to insert the aortic valve. During this time period, John Webb and colleagues18,19 were also developing a transapical aortic valve method, and subsequently they introduced the retrograde transfemoral artery approach. This latter technique became feasible once Edwards developed a catheter that could be flexed to get around the aortic arch to access the aortic valve.

The Revival Trial was started in the United States in 55 patients using the retrograde transarterial transfemoral approach (Transapical aortic valve insert, TFAVI; Figs. 39-1, 39-2, and 39-3).24 A separate Revival trial also included a transapical approach in 40 patients.25 In the United States, the three centers involved in these studies were Cleveland Clinic in Ohio, Columbia in New York, and Medical City in Dallas. In the first transfemoral Revival series of patients, the mortality rate was 7% and the stroke rate 9.2%. The transapical (TA-AVI) (Figs. 39-4, 39-5, and 39-6) had a higher mortality rate of 17% in the first 40 patients with no immediate strokes after successful insertion, although two patients did develop strokes, one after an open replacement for aortic valve disease in a very heavily calcified ascending aorta and the other in a patient who developed atrial fibrillation a few days after surgery. Based on the feasibility studies, the FDA then approved a prospective randomized trial to further evaluate the device technology.

At the same time as the Edwards balloon-expandable valve was being developed, Core-valve introduced a nitinol-based percutaneous valve system that was also inserted transfemorally but did not require balloon expansion, but rather expanded over time as the nitinol stretched the aortic valve and seated in the aortic valve annulus.

The Edwards balloon-expandable valve has evolved over time with two principal changes, which include the stent steel cage having a higher skirt sewn to the base of it and switching from equine to bovine biological leaflets. Furthermore, the bovine leaflets are manufactured to the same thickness as those found in regular pericardial valves. In the pulse duplicators the valve was tested to exceed the equivalent of some 10.4 years of implantation. To try and reduce the bulk of the stainless steel cage, newer modifications are going to become available that involve the use of a different metal, altering the existing mounting system for the valve on the catheter for balloon expansion. Thus, the intent is that the valve will be able to be inserted in an "18-sheath system." The long-term durability of the bovine valve leaflets appears to be excellent with few reported failure of the devices at this time on long-term follow-up of the patients with the important caveat that the valve is both full and symmetrically expanded. Clearly, endocarditis is still a risk factor over time and has occurred occasionally, but leaflet calcification or tearing, at least in the first few years, has not been a problem. It is likely that further development of the valve skirt will reduce the incidence and the degree of perivalvular leak by changes in the cuff and height of the cuff of the balloon-expandable valve.

The PARTNER Prospective Randomized Trial consisted of two arms. There was much debate as how to best conduct the study as far as methods, and whether or not randomization was needed and which groups to assign patients. After negotiations with the FDA, the final decision was to have a Group A and a Group B, with Group A patients being randomized either to regular open surgery, or transfemoral arterial or transapical valve insertion. Group B patients were randomized to either transfemoral insertion or best medical treatment, including the use of balloon valvuloplasty. The Group A patients in the treatment arm were to be compared with the open surgical patients on the basis of noninferiority, and in Group B the treated patients were to be compared with the best medical treatment control arm on the basis of superiority. A total of 1040 patients were placed in the study, with 450 in the Group B arm, which was completed in the first half of 2009. The Group A part of the study was completed on August 28, 2009, and thereafter patients will still be screened and will require approval by the panel for continued access in either Group B or Group A. Randomization will no longer be required.

Entry criteria for most patients in Group A was having an STS score above 10, a valve area of less than 0.8, and, thereafter, depending on arterial access of the groins, patients were assigned either to the transapical arm or the transfemoral arm. Patients in the open surgical arm were not allowed to undergo any coronary artery bypass surgery or any other valvular procedure but were allowed to have the ascending aorta replaced for calcification using deep hypothermia and circulatory arrest. Group B patients require a combined risk of serious morbidity or mortality exceeding at least 50% and two surgeons agreeing the patient was inoperable. Exclusions included bicuspid valve pathology, left ventricular outflow tract smaller than 1.8 cm or larger than 2.4 cm, creatinine above 2.5 or dialysis, infection within 6 months, and a PCI procedure within 30 days.

There are important caveats to bear in mind when considering the trial. In our experience at the Cleveland Clinic (Table 39-1) it is only patients greater than age 80 undergoing reoperation in whom the risk of aortic valve replacement and the risk of death exceeds 5%. Hence, most patients in the trial are elderly with severe comorbid disease and usually undergoing a reoperation. Indeed, a 90-year-old patient with no comorbid disease requiring a reoperation may not qualify for the study. Another important caveat is the influence of peripheral vascular disease on outcome. We showed in a study of our patients that peripheral vascular disease increased the risk of both death and stroke after surgery.30 Hence, because TA-AVI patients have poor access by trial entry criteria, they are also at greater risk of death or complications.

The Group B patient data will clearly be available before the Group A randomized patients and in both categories, the follow-up required will be 30 days plus 1 year. The data will then be presented to the FDA and panel evaluation. In both groups, mortality at 1 year is the most important variable, although there are also the composite end points of outcomes for analysis. (See "PARTNER Results Update" at the end of this chapter.)

While the PARTNER Trial data is being evaluated, ongoing studies are being continued in Europe in which patients have not been randomized but are being followed. The method of insertion has varied somewhat from site to site, with some sites, for example the group in Leipzig, prefer a transapical approach, whereas those in other centers prefer a transfemoral approach. In the centers in which the transapical approach has been used almost exclusively in patients, the reported results have so far been equivalent to those for the transfemoral approach. We note that the perceived advantages for the transapical approach are a quicker time for insertion, less contrast, and less risk of stroke. On the other hand, patients with severe lung disease and oxygen dependency do not tolerate this procedure well because it requires a minithoracotomy. The obvious benefits for the transfemoral approach are quicker recovery and earlier mobilization of patients, but this may be burdened with a higher risk of stroke and, clearly, there is a risk of vascular injury, which was 17% in the initial series in the Revival trial. This, however, has been reduced considerably and with the 18-French device becoming available in 2010, it is likely that both the rate of injury will be reduced and also the number of patients who need a transapical approach. The other option is to use the subclavian artery or mini aortic approach for insertion of the device.

Core valve has been inserted in 79 patients in Europe via the left subclavian artery, with a 9% risk of death, but what has been more daunting has been the 38% risk of requirement for pacemaker insertion after the procedure. This may be related to the proximal stent sitting a little bit lower in the left ventricular on-flow tract, leading to continued expansion of pressure on the bundle of His, and thus resulting in the need for pacemaker insertion.

In our most recent patients at the Cleveland Clinic, the modified approach we have used for insertion of the transfemoral approach is to introduce the femoral, arterial, and venous lines percutaneously using a closure device at the end of the procedure. Thus, a venous pacing wire is inserted into the right ventricle and then two arterial wires are inserted into the groin, one for the pigtail catheter and the other for the insertion of the device. Typically, the patient is heparinized so the ACT is between 250 and 350 seconds. The sequence of insertion is the introduction of a straight wire via the straight catheter through the aortic valve and into the left ventricle. This is then exchanged with an extra stiff or Lunderquist wire, which is placed in the left ventricle after sliding the sheath across the aortic valve. A 3- or 5-cm balloon is then threaded over the guidewire and positioned across the aortic valve using fluoroscopy and then during rapid pacing, and breath holding, expanded to crack the calcium. If the balloon does not move, then the balloon is removed and the femoral 14-French sheath exchanged for progressive-sized Cook dilators and finally the largest-sized Edwards dilators are inserted into the femoral artery and into the distal abdominal aorta. The size is dependent on the valve selection, namely, a size 23 or 26 valve. The large sheath is then inserted. The size 23-mm valve is used for patients with left ventricular outflow tracts of 1.8 to approximately 2.1 to 2.2 cm, and the larger 26-mm valve is used for left ventricular outflow tract sizes from 2.2 to 2.4 cm. Generally, patients with left ventricular outflow tracts of 2.5 cm or larger are excluded from the study until the new 29-mm valve becomes available.

As a rule, patients who need the smaller-sized valve should have a femoral artery of at least 7 mm and for the size 26 valve a 7.5- to 8-mm external iliac artery, depending on the tortuosity of the vessel and degree of calcification. If there is more than 90-degree angulation of the iliac arteries, then generally it would be difficult to insert the large sheath unless the vessel straightens very easily with a Lunderquist wire. In measuring for the 22-mm-French sheath (internal diameter) the size varies from 7.45 to 7.72 mm for the internal diameter and external measurements vary from 8.21 to 8.50 mm. The 22-French sheath is used for the size 23-mm valve. In other words, the true internal size is 23-mm French and 25.5 French external diameter using a conversion factor of 0.333. For the 26-mm valve, a size 24-French sheath is used, which has an internal size measurement between 8.20 and 8.40 mm and an external size of 8.94 to 9.90 mm. By converting this to French sizes this would be equivalent to a 25.3 internal and 27.3 external French size. The 28 dilator sheath diameter is 9.24 mm, which is used before introducing the big 24-French sheath.

Once the sheath is inserted and the valve mounted on the balloon, it is checked to be correctly loaded, and is then threaded over the stiff wire and the loader pushed into the sheath to the black mark but not beyond it. Note that with the new sheath this requirement is not as critical. The valve is then pushed and threaded across the aortic arch with the steerable catheter being flexed to get around the aortic arch and then fed down to the aortic valve. With the flex-three system, the insertion and retraction of the dilators are much easier. Essentially, the valve with the leading little cone is fed across the valve to cross the valve and then the cone is pushed further into the ventricle so that the stent valve is free. Once the balloon is inflated, it can be expanded without limitation from the cone. We consider it very important to carefully check on the positioning of the valve across the hinged point of the native valve using both echocardiography and fluoroscopy and root injections with contrast via the pigtail catheter. With the flex-three system the valve is positioned at the valve hinged points, approximately 50/50 aortic/ventricular. Once everybody agrees that the position is correct, the patient's breath is held, the patient is paced at approximately 180 beats per minute, the balloon is inflated in position and then deflated, the pacer is switched off, and the patient is put back on the ventilator. There is virtually always some perivalvular regurgitation in the region of the noncoronary–left main leaflet commissure, probably because this is an area of less flexibility and a lot of calcium. Echocardiography is then used to inspect the function of the valve. Thereafter, cone is first withdrawn from the left ventricle followed by removal of the catheter and sheath, and lastly the previous closure device is tied down at the insertion sites.

In the United States, the transapical approach is reserved for those patients who do not have access and, as such, they generally have more comorbidity than transfemoral patients based on comparative studies.

The way we prefer to do it at the Cleveland Clinic is to have the patient lie flat on his or her back on pillows rather than at a traditional left thoracotomy position, because this interferes with the positioning of the C-arms. All patients before surgery undergo an aortic root injection from which the ideal planes are calculated for positioning the valve with the nadir of each cusp being in the same plane. This is also obtained with the computer tomographic scans, which are done preoperatively to check for potential calcium obstruction of the native coronary arteries.

Once the patient is positioned, he or she is prepped in the usual manner with exposure of the groin and left chest. The femoral artery and vein are exposed on the right side and the left femoral venous pacing wires inserted. The femoral artery exposure is performed in case the patient has to be placed on a pump.

We continue to have all of our patients undergo general anesthesia but, more recently, we have not used a double-lumen tube for the transapical valves. Before opening the chest, an echo probe is placed on the chest to try to better define where the left ventricular apex is, and CAT scan images are also inspected to find the ideal place to open the left chest. Typically, the fifth intercostal space is entered and the apex of the left ventricle palpated. Once the apex of the left ventricle has been palpated, a short piece of rib is removed to expose the apex. Originally we did not do this and rather spread the ribs, but we found that some patients had severe pain as with the mid-cab experience. We therefore opted to resect a short piece of rib to obtain both a better exposure and less pain after surgery, particularly because many of these patients have severe lung disease. The pericardium is then opened and stay sutures placed and then two pursestring sutures with Teflon felt pledgets are placed around the left ventricular apex. The site is selected by tapping on the left ventricular apex and then looking at the echo to make sure that the left ventricular apex is correctly identified. This is important because the heart can be anywhere from the fourth intercostal space to the seventh or eighth intercostal space. In addition, a horizontal mattress suture is also placed in the apex. A needle is then inserted followed by a guidewire into the left ventricle; then we insert a Berman catheter into the left ventricle, through the aortic valve, and across the aortic arch down to the renal arteries. An extra stiff wire is also fed down at the same time. The Berman catheter is then removed and the 14-gauge sheath exchanged for the large-bore introduction sheath. Balloon valvuloplasty is performed through this in the same sequence with rapid pacing as for the transfemoral approach. The valve is then loaded onto the sheath and pushed into position and the pusher retracted into the sheath. With the valve situated approximately 60% aortic and with rapid pacing the balloon is then inflated at this point. The balloon is then withdrawn into the left ventricle before stopping pacing. If all appears well, the sheath and wire are then removed and the pursestring sutures are tied down.

PARTNER B showed that TFAVI significantly improved 1 year survival despite a 30-day death rate of 6.4% and stroke/TIA rate of 5.5%. PARTNER A showed transcatheter aortic valve replacement (TAVR) was noninferior to open surgery for both 30-day and 1-year mortalities; however, the rate of stroke/TIA was higher for TAVR.

Percutaneous aortic valve procedures continue to improve and, despite our early skepticism with the approach based on early animal work that showed a high embolization rate in the animal valves, the procedure has established itself as a very effective way of treating high-risk patients with severe aortic valve stenosis. Undoubtedly with new iterations the procedure will become safer and even better, likely with retrievable devices and easier positioning of the devices.