Transcatheter Aortic Valve Implantation zur Erlangung der Venia Legendi

Degenerative aortic valve stenosis accounts for the majority of native valve diseases and the prognosis is poor for symptomatic patients if untreated. Owing to the progressive aging of the population, the incidence of symptomatic aortic stenoses will further increase. Surgical aortic valve replacement provides a valuable treatment option with a well-documented favorable long-term outcome. However, the risk of surgery increases especially in older patients with significant comorbidities. For this reason, one third of symptomatic patients is not considered suitable for surgery. Transcatheter off-pump aortic valve implantation, however, using a transfemoral or a transapical approach may be a suitable therapeutic option for these patients. The balloon-expandable Edwards-SAPIEN prosthesis and the self-expandable Medtronic Core- Valve system have received CE approval and more than 10,000 patients have been treated worldwide.Both methods share the same main principles: after crossing the stenotic aortic valve, balloon valvuloplasty is performed during a short episode of rapid ventricular pacing to predilate the native valve. After careful positioning of the valve prosthesis using either transesophageal echocardiography or fluoroscopy or both, the valve is released.30-day mortality ranges between 8% and 15% in different registries in higher-risk patients. Nonetheless, larger randomized trials are warranted in order to validate the safety and effectiveness of this new approach.


Introduction
The introduction of this cumulative habilitation thesis will provide a background of the nature and prognosis of aortic valve stenosis. Treatment options of aortic stenosis will be discussed and the development of transcatheter aortic valve implantation (TAVI) will be reviewed.

Aortic stenosis
Aortic stenosis occurs due to calcific degeneration of a tricuspid or bicuspid valve, or due to rheumatic heart disease. Calcific aortic stenosis is the most frequent cause of valvular heart disease in the Western world, and the leading indication for operative valve replacement. The prevalence of aortic stenosis increases with age and reaches almost 5% in patients ≥ 75 years (Figure 1) 1 . Aortic stenosis is a mechanical problem that requires a mechanical solution. If treated with medical therapy only, patients experience a high rate of death 2, 3 . Surgical aortic valve replacement improves both symptoms and survival in patients with severe stenosis; but with advanced age, poor left ventricular function, or comorbidities, operative risk may be high or even prohibitive 4,5 . A survey of European hospitals in 2001 found that surgery was not offered in 10% of patients with severe aortic stenosis 6 . In addition, many elderly patients refuse to undergo open heart surgery. As aconsequence, during the last decade, transcatheter aortic valve implantation (TAVI) has emerged as an alternative, less invasive treatment option in such patients 7,8 .

A brief history of TAVI
The search for less invasive treatment options for patients with valvular heart disease was pioneered by Hywel Davies in 1965. Davies developed a parachute-like valve to treat aortic regurgitation. The valve was placed in the descending aorta and offered little resistance to the systolic blood flow in antegrade direction, but passively opened during diastole to obstruct the retrograde diastolic flow 9 . However, such valves were never utilized in humans. In 1985, Alain Cribier performed the first aortic balloon valvuloplasty, which was the first catheter-based intervention specifically designed to treat non-operable calcific aortic stenosis 10 . At that time, age > 75 years per se was often a contraindication to aortic valve replacement. After rapid adoption of balloon valvuloplasty and explosive growth worldwide, the lack of survival benefit and the high recurrence rate of severe stenosis led to a marked decline in its use 11 .
In 1989, Henning Andersen developed a balloon expandable aortic valve that could be implanted percutaneously in a porcine model 12 , but he did never use this technique in humans.
The first in human transcatheter valve implantation was performed by Philip Bonhoeffer in 2000. He implanted a stented valve made of a bovine jugular vein into a pulmonary artery conduit of a young patient with congenital heart disease.  With the initial high profile catheters and sheaths, not all patients could be treated through the femoral arteries. In 2006, Sam Lichtenstein, a cardiac surgeon from Vancouver, performed the first transapical TAVI through a left-sided mini thoracotomy. During the following years, other alternative access routes were developed including the transaxillary/subclavian access, the direct aortic access, and even a transcarotid access. In Switzerland, the first TAVI was performed in 2007. It is estimated that as of 2012, about 100'000 TAVI procedures have been performed worldwide.   13,14,[16][17][18][19][20][21] . From these early transarterial case series, it became evident that vascular complications were one of the limiting factors of transarterial transfemoral TAVI. These complications included dissection or rupture of the iliac and femoral arteries, the aorta, and the aortic annulus and were reported in up to 30% of patients undergoing TAVI with the early high profile catheters requiring 22-24 F sheaths. In an attempt to reduce iliofemoral complications, a surgical cut-down to expose and control the iliofemoral artery above and below the puncture site was initially performed 8,13,14,16,[21][22][23] .
Another problem that was recognized very early were paravalvular leaks. Cribier noted in his first publication that 2/6 patients had severe paravalvular aortic regurgitation (PAR) following TAVI 15 . Subsequent publications reported moderate or severe PAR in 8-22%. Patients with acute severe PAR often developed cardiogenic shock, but the relevance of moderate or mild PAR was less clear. It was known that there were different mechanisms that caused PAR.
Undersizing of transcatheter valves was the most frequent cause, but malpositioning (too low or too high), or severe calcification were other causes. Before 2010, valves were selected based on a single measurement of the annular diameter on transthoracic or transesophageal echocardiography 27 . However, the annulus is a complex, three-dimensional, nearly uniformly oval shaped structure 28 . To measure the annulus in its true plane, a three-dimensional imaging modality appeared necessary, but it was unknown if such a strategy will truly lead to less PAR.
Furthermore, there was a concern that larger valves may increase the risk for annular rupture. Initially, the rate of cerebrovascular events during and after TAVI ranged between 2-6%, although definitions did vary 13, 16, 18-21, 29, 30 . However, it was the publication of the PARTNER A trial that made the risk for strokes one of the hot research topics in TAVI 7 . In this trial, the rate of cerebrovascular events (including major strokes, minor strokes, and transient ischemic attacks) was higher in the TAVI group than in the surgical group, although the rates of strokes (4.6% vs. 2.4% at 30 days) did not differ significantly. Periprocedural strokes may result from embolization of calcified microdebris during positioning and deployment of the valve, but also during passage through the aortic arch with the guidewire and the catheter, during balloon valvuloplasty, or re-capture of a valve . Not infrequently, strokes occurred during the first 1-2 days after the procedure. These postprocedural stroke may be due to the nonendothelialised and thrombogenic bioprosthesis itself, new-onset atrial fibrillation, late calcific embolism, or possibly due to late thrombosis or hemorrhage following earlier embolism.

Patient selection: TAVI in specific patient subgroups
Typically, patients that are currently selected for TAVI are more than 80 years old, and have comorbidities such as coronary artery disease, diabetes, a prior stroke, chronic pulmonary disease, or kidney disease. These restrictions are applied more rigidly in North America than in other part of the world. Across Europe and especially in Germany, there is a trend towards that long-term survival is mainly determined by (non-cardiac) comorbidities 32 . Therefore, TAVI should not be performed in the presence of relevant comorbidities that limit survival to less than 2-3 years. It is now generally recommended that each patient should be discussed and selected by a multidisciplinary Heart Team including cardiologists and cardiac surgeons.
However, there are certain subgroups of patients that may benefit more from TAVI than others. Two of these subgroups may be patients with concomitant mitral regurgitation and women.  Compared to patients with none, trivial or mild MR, patients with moderate/severe MR had a higher mortality rate during the first 30 days (unadjusted hazard ratio (HR) 2.04, 95% confidence interval (CI) 1.11 -3.74, p = 0.02; adjusted HR 2.10, 95% CI 1.12 -3.94, p = 0.02) but no difference after 30 days (unadjusted HR 0.94, 0.58 -1.51, p = 0.80; adjusted HR 0.82, 0.50 -1.34, p = 0.42). Therefore, moderate or severe MR was associated with an increased early, but not late mortality.

TAVI in patients with concomitant mitral regurgitation
Since the severity of MR depends primarily on regurgitant orifice area and the systolic pressure gradient between the left ventricle and the left atrium, MR is expected to improve immediately after TAVI and may improve further in the mid-and long-term should positive left ventricular remodeling occur 35,36 . Indeed, in our study, we observed that a few days after TAVI, MR improved in 61% of the patients with moderate or severe MR at baseline, and rarely worsened. At 1 year follow-up MR had improved in 55%, remained unchanged in 16%, and worsened in 1%; the remaining 28% had died. Pulmonary hypertension has been identified as a predictor of mortality and adverse outcome in previous studies. Finally, atrial fibrillation may worsen MR due to atrial and annular enlargement.
The randomized Placement of Aortic Transcatheter Valves (PARTNER) study suggested that patients with moderate or severe mitral regurgitation may derive a larger benefit from TAVI compared to both medical management and surgical aortic valve replacement 7,8 . In the PARTNER B study, subgroup analysis showed that the number needed to treat to prevent 1 death at 1 year was 3 in patients with moderate or severe MR versus 7 in patients without it.
In the PARTNER A study, 1-year mortality of patients with moderate or severe MR was 24.2% after TAVI (similar to the 27.7% in our study) and as high as 35% after surgical aortic valve replacement.

TAVI in women
In the PARTNER trial, a pre-specified subgroup analysis suggested that women benefit more from TAVI than men do 7 . Two additional publications examined sex differences in outcomes after TAVI. One found no difference in mortality at 30 days and 1 year after TAVI 37 . One study analyzing 260 consecutive patients reported better 1-year survival in women, but failed to adjust for baseline characteristics, which varied substantially between men and women 38 .
In light of limited and conflicting evidence, we aimed to answer the following questions:  Are there difference in baseline characterics between men and women?
 Does periprocedural outcome differ between men and women?
 Is there a difference in long-term survival?
Paul's Hospital, Vancouver, Canada, and at the Québec Heart and Lung Institute, Québec City, Canada 39 .
We found that women had higher mean aortic gradients, worse renal function, more often a porcelain aorta, and a better left-ventricular systolic function than men. Women were more often frail, while coronary artery disease, a prior myocardial infarction, prior percutaneous revascularisazion and chronic obstructive pulmonary disease (COPD) was more frequently present in men.
We made the following observations regarding procedural outcome:  The reason for these findings remains unclear. It is unlikely that the survival advantage can be explained by the difference in baseline characteristics such as coronary artery disease. Also, the higher post-procedural complication rate may increase short-and mid-term mortality. One possible explanation may be the fact that interstitial fibrosis is more pronounced in male hearts with severe aortic stenosis, as is collagen I, II, and matrix metalloproteinase expression 40,41 . Lower levels of fibrosis in women might lead to more rapid reversal of myocardial hypertrophy after correction of aortic stenosis 42 . Another explanation for the survival advantage might be the higher life-expectancy of women. However, the Kaplan Meier figure shows that the curves separate during the first 6 months and are parallel thereafter. The reason for this early survival advantage despite the higher periprocedural complication rate remained unclear.

Measurement of the aortic annulus and valve sizing
While during open heart surgery the aortic annulus can be directly inspected and sized, TAVI operators depend on external, indirect measurements of annular size. However, correct sizing of the annulus and choice of prosthesis is of utmost importance in order to avoid paravalvular leakage, device embolization, and annular rupture. Studies have shown that about 10-20% of patients are left with moderate or severe paravalvular regurgitation after TAVI and this has been associated with increased morbidity and mortality in several independent publications [43][44][45] .
During the early experience with TAVI, the annulus was measured with 2-dimensional echocardiography using the parasternal long-axis view durig transthoracic echocardiography (TTE) or with transesophageal echocardiography (TEE) in the ~130° view. However, it is well known that the annulus is an oval structure where the long axis is in average 5-6 mm (about 20%) longer than the short axis 46,47 . Therefore, measuring an oval-shaped annulus in only one dimension may under-or overestimate it's true size.
On the other hand, multidetector computed tomography (MDCT) offers a 3-dimensional to echocardiography and allows reconstruction of the aortic annulus in it's true plane. Utilizing MDCT for annulus measurements, we aimed to address the following questions:

MDCT based annular measurements
We analyzed a total of 109 consecutive patients who underwent MDCT before TAVI at 2 centers, St. Paul's Hospital, Vancouver, Canada and Aarhus University Hospital, Aarhus, Denmark. Annular size was measured by TEE and MDCT. Valve size was chosen based on TEE measurements. A subset of patients (n = 50) also underwent MDCT before discharge to assess post-implant geometry of the balloon-expandable valves 47 .
In our study, there was no patient with annular rupture, although this has been a concern when oversizing balloon expandable valves. The mean MDCT mean annular diameter was 23.9 ± 2.4 mm which was significantly larger than the mean TEE annular diameter (22.5 ± 1.9 mm, p < 0.01).

Figure 12. In a 79 year old patient TEE measured an annulus of 2.1 cm suggesting a 23 mm Edwards SAPIEN or a 26 mm Medtronic CoreValve (A). CT angiography measured an area of 481 mm2 (C) and a mean diameter of 24 mm (B) suggesting a 26 mm Edwards SAPIEN or a 29 mm Medtronic CoreValve valve with ~10% oversizing. A 26 mm Edwards SAPIEN valve was implanted with trace paravalvular leak (D).
Undersizing the nominal diameter or area of the valve relative to the MDCT measurements resulted in significantly more moderate or even severe paravalvular regurgitation. MDCT diameter (area under the curve (AUC) 0.81) and area (AUC 0.80) were more predictive of moderate or severe paravalvular regurgitation than TEE annulus diameter (AUC 0.70). For patients with a valve diametermean annulus diameter < 1 mm, the incidence of moderate or severe paravalvular regurgitation was 21.4% vs 2.2% when the difference was ≥ 1 mm. For patients with a nominal area < 10% greater than the annular area, the incidence of moderate or severe paravalvular regurgitation was 19.1% vs. 0% when the nominal area was > 10% above the annular area.
Our results indicate that CT annular measurements have a good predictive value of moderate or severe paravalvular regurgitation following TAVI. Two-dimensional TEE measurements did often underestimate the true annular size. Our results therefore suggest that when utilizing CT measurements for valve sizing, the valve should be oversized by 10% relative to the measured area and by 1 mm relative to the mean diameter.
Based on these results, we published sizing recommendations for balloon expandable valves with the goal to oversize the valve relative to the measured area by 10-15% 48 . Based on these data, we aimed to address the following issues with our research project: The rate of vascular complications was higher when the minimal artery diameter was smaller than the external sheath diameter (24% vs. 10%, p = 0.03), in the presence of moderate or severe calcification (29% vs. 9%, p = 0.03). The rate of vascular complications fell from 32% to 9% during the study period indicating a learning curve. We also showed that major vascular complications were associated with a 30 day mortality rate of 20% and that the duration of hospital stay was more than twice as long as without complications.

Predictors and relevance of vascular complications
Transfemoral access is considered the access of choice due to its least invasive nature and is feasible in the majority of patients undergoing TAVI. Most interventional cardiologists are very familiar with the transfemoral route due to their experience with percutaneous coronary intervention. However, iliofemoral complications are the most common vascular complications in transfemoral TAVI. These complications include iliofemoral dissection, iliofemoral rupture, access site infection, stenosis, thrombosis, or occlusion, pseudoaneurysms, and femoral bleeding. To avoid these complications, a surgical cut-down was often performed in initial studies. A planned surgical cut-down can be performed at the beginning of the procedure to allow visualization and selection of the ideal puncture site and control of the artery above and below the puncture. Alternatively, a percutaneous puncture is performed and the artery is exposed for closure only. More recently, operators utilized percutaneous closure devices. Our study showed that with a fully percutaneous procedure utilizing percutaneous closure with two ProGlides or one ProStar, low vascular complication rates can be achieved.
Careful patient selection, and screening with CT angiography or fluoroscopic angiography were used to identify patients at high risk of complications. Such complications occurred more often if the minimal artery diameter was smaller than the sheath external diameter, and in the presence of moderate or severe calcification. With the available alternative access routes (transaxillary, transapical, transaortic), such patients should not undergo transfemoral TAVI.

Long-term outcomes after transcatheter aortic valve replacement
Since its introduction, the number of TAVI procedures was growing quickly world-wide. In 2012, it was estimated that in third-world countries, 20-50% of patients with severe aortic stenosis now undergo TAVI. Despite the growing popularity of TAVI, previous reports have focused on short-and mid-term outcomes, while little was known about longer-term outcomes.
A few reports have investigated outcome up to 3 years 31,54 . From these reports, it was known that with current patient selection, median survival after TAVI was about 3 years. Accelerated in-vitro testing showed that valves may last up to 15 years. In the early series, the high mortality rate was unlikely due to valve deterioration, but more likely due to the types of patients undergoing TAVI. Another important issue was and still is the occurrence of strokes.
In the PARTNER trial, TAVI has been associated with higher rates of strokes and transient ischemic attacks than surgery, although the combined endpoint of strokes and death was nonsignificantly lower with TAVI compared to surgical AVR 7 . By analyzing long-term outcome, we also looked at the timing and type of strokes. We aimed to answer the following questions:  During the observation period, a total of 6 ischemic and 4 hemorrhagic major or fatal strokes occurred. Cumulative major ischemic stroke rate at 1-5 years was 3.6%, 5.2%, 7.3%, 7.3% and 9.7%, respectively, and 3/6 (50%) of ischemic strokes were fatal. Cumulative major hemorrhagic stroke rate at 1-5 years was 2.8%, 4.4%, 4.4%, 7.3%, and 7.3%, respectively, and 2/4 (50%) of hemorrhagic strokes were fatal.
To put these findings in perspective, it is important to remember that these patients represented the first in-human experience with transfemoral and transapical TAVI in truly inoperable patients explaining the relatively poor long-term survival. However, hemodynamics were excellent and signs of moderate prosthetic valve failure were observed in only 3.4% of patients after 5 years.
Cumulative ischemic stroke rate at 5 years was 9.7%, indicating an annual risk of ischemic strokes of about 2%. At the same time, cumulative major hemorrhagic stroke rate at 5 years was 7.3%. Of note, atrial fibrillation was present in 60% of these events indicating that atrial fibrillation may cause ischemic strokes (thromboembolism) and hemorrhagic strokes (bleeding due to antithrombotic therapy). Based on our results, we were confident that transcatheter valves will last longer than 5 years and that patients with a longer life-expectancy will have long-term benefits from a durable valve.

Perspective
Further refinement of the procedure and current developments focus on minimizing vascular complications, reducing the risk for cerebrovascular accidents, improving paravalvular sealing, and facilitating accurate prosthesis implantation. Some of these issues may be addressed by next-generation valves that currently undergo early clinical evaluation. Most of these valves are constructed of self-expanding nitinol which offers the potential for recapture, repositioning, and removal, if required 56 .  The Centera valve (Edwards Lifesciences, Irvine, USA) is a relatively short self-expanding valve that does not extend into the ascending aorta for self-alignment or anchoring. The inflow is non-tubular flared, with a narrower annular segment and a larger diameter outflow.
The idea behind this shape is to facilitate self-seating in the aortic annulus and reduce paravalvular regurgitation. The Centera valve can be deployed and if necessary recaptured using a single hand operated motorized delivery system 60 .
The self-expanding Acurate valve (Symetis, Ecublens, Switzerland) has three stabilization arms meant to stabilize the valve in the ascending aorta. The valve has an upper crown that provides tactile feedback during positioning. Similar to a 'hook concept', optimal valve position is achieved by applying slight tension on the delivery system and thereby on the upper crown 61 .
With improved equipment, and more experience, there is no doubt that rates of mortality, paravalvular regurgitation, strokes, and vascular complications will decrease. Compared to open heart surgery, TAVI is less invasive, allows earlier patient ambulation, and requires a shorter recovery and rehabilitation period. There is no doubt that the indication for TAVI will be extended and that TAVI will be applied in a broader spectrum of patients with severe aortic stenosis or regurgitation.

Acknowledgments
This work and habilitation script would not have been possible without support from the following people I am indebted to: It is important that each project has a leader, but research is not a one-man show. I had the pleasure to work with a group of highly skilled and motivated fellows. We were all excited of having the opportunity to work at St. Paul's Hospital.