A. Clinical Significance

Reduction of functional mitral regurgitation (FMR) by percutaneous techniques has demonstrated significant clinical benefits, and has provided keen scientific insights into this regurgitant lesion. Exercise improvement ranging from 100 – 150 meters as demonstrated by a six minute walk test show the functional benefits of reducing FMR. Reverse left ventricular remodeling in a treated cohort provides insight into the hemodynamic consequences of FMR. Functional and hemodynamic benefit from reducing mild and moderate FMR challenge conventional wisdom regarding timing of intervention. The ability of mitral innovations to advance our clinical and scientific understanding of mitral regurgitation (MR) has been extraordinary. Years of focused and deliberate innovation created the foundation for trials that provided these insights. Awareness of the key elements of that deliberate innovative process is the catalyst to future successful innovations.

B. Introduction

Innovation for mitral valve disease has proven to be a long and complicated road, in part due to the complexity of the mitral apparatus and poor differentiation of mitral pathology in the literature. The absence of predecessor technologies upon which to conceptualize new therapeutic devices challenged us to innovate from a blank slate. Because of these challenges, the development of devices for mitral valve disease offers invaluable insights into the fundamental principles required to innovate successfully. Perseverance through the process has generated unique insights into the benefits of treating MR. Innovative efforts have now generated sufficient data to suggest that, rather than being an innocent bystander, FMR contributes significantly to patient symptoms and ventricular remodeling. Treating FMR thru percutaneous means results in dramatic symptom improvement and reverse remodeling.

Two therapies that received CE Mark approval took complementary paths through clinical trials to commercialization, one focusing initially on primary MR, the other focusing on FMR. To highlight key fundamental principles of innovation and device development, we will discuss, in depth, the experience of one of the authors (DGR) in bringing such a device from initial inspiration, through prototype development, clinical trials, and into commercial use.

C. Mitral Regurgitation: Primary vs. Functional

MR is the most common valve disease in the United States affecting more than 2.5 million people, a number expected to double by 2030. Regurgitant flow through the insufficient mitral valve during systole creates a volume load on the left ventricle, the severity of which depends on the degree of insufficiency. Volume loading is thought to initiate the signaling pathways that lead to left ventricular remodeling and related sympathetic neurohormonal activation, which may be adaptive in augmenting stroke volume in the short term, leading to pulmonary hypertension, clinical heart failure and related morbidities and death, if severe and untreated.

MR is broadly classified as either primary or functional based on a system described by Carpentier in 1983. Primary MR results from an intrinsic abnormality of one or more components of the mitral apparatus, while FMR results from left ventricular and annular dilation in the setting of heart failure. In primary MR, correcting the mechanism of regurgitation, and the volume load on the left ventricle, will presumably lead to reversal of remodeling and prevention of morbidity and mortality. This presumption is supported by studies showing that, while medical therapies are largely ineffective at improving outcomes, surgical repair of the mitral valve leads to improved outcomes.  Repair of the existing mitral valve structure is generally preferable to mechanical valve replacement, which destroys important contractile elements of the mitral valve apparatus, carries greater morbidity/mortality, and requires lifelong anticoagulation therapy. Up to 49 percent of patients with severe symptomatic MR may be deemed inoperable based on age, poor left ventricular function, or other comorbidities. Thus, the design of less invasive percutaneous techniques for mitral valve repair has become an active area of research and development.

FMR affects up to 90% of heart failure patients., The compensatory dilation of the left ventricle in a patient with ventricular dysfunction distorts the sub-valvular mitral apparatus. This leads to displacement of the papillary muscles. The vector placed on the mitral leaflets by the chordae tendineae compromises leaflet coaptation, and creates a geometry vulnerable to develop FMR. The volume overload created by this regurgitant lesion increases the hemodynamic stress on the failing left ventricle, contributes to progressive dilation, and further exacerbates systolic dysfunction. Despite mitral leaflets that are free from significant pathology, altered ventricular mechanics contribute to the development of FMR, contributing to patient symptoms, ventricular remodeling, and increased mortality.



While surgical therapy to treat primary MR has generated compelling data,, isolated mitral annuloplasty is rarely offered to heart failure patients.



D. The Mitral Apparatus

An understanding of the anatomic features of the mitral valve and its surrounding structures is critical to develop novel devices designed to treat mitral valve disease.

The mitral apparatus is comprised of five components: the annulus, the leaflets, the commissures, the chordae tendineae, and the papillary muscles (Fig. 1). When working normally, these components ensure unrestricted diastolic forward flow, prevent systolic regurgitation, and provide important structural support and contractile force TO maintain optimal left ventricular geometry and function.

Figure 1. The mitral apparatus. Reproduced from Carpentier’s Reconstructive Valve Surgery (Carpentier et al., 2010) with permission from Elsevier.

The mitral annulus is one of four contiguous fibrous valve rings, which form the connective tissue scaffolding of the heart. This scaffolding supports the tissue of all four valves and provides electrical insulation between the atria and the ventricles, allowing the conduction system to function properly. The mitral annulus is saddle-shaped and elliptical, a shape that optimizes leaflet support and coaptation. The anteromedial portion of the mitral annulus, which is continuous with the aortic annulus, is relatively stiff and not typically targeted by invasive strategies attempting to support the annulus of an insufficient valve. The remainder of the valve annulus has less reinforcement. This portion of the annulus tends to be more frequently implicated in the pathologic development of mitral regurgitation and also serves as the target of some repair techniques including partial annuloplasty rings.

The valve leaflets are made up of one continuous piece of thin fibroelastic tissue, attached circumferentially to the annulus, with two notches, or commissures, along the free edge. The commissures mark the division between the anteromedial and posterolateral leaflets. The anteromedial leaflet is triangular, with its base attached to the stiff anteromedial third of the annulus; the posterolateral leaflet is quadrangular, with its base attached to the remaining two-thirds of the annulus. The posteromedial leaflet has three scallops, termed P1-P3 from lateral to medial, that correspond to the A1-3 components of the anteromedial leaflet. The anteromedial leaflet is longer and more mobile than the posterolateral leaflet, and the area of both leaflets combined is typically more than double the mitral orifice area, which ensures a large area of coaptation during systole.

Apically, the valve leaflets are tethered along their free edges by a latticework of fibrous tissue, the chordae tendineae. The chordae converge upon the papillary muscles, which arise from the left ventricular walls in the anterolateral and posteromedial positions below the commissures. The chordae of the posterolateral leaflet are shorter than those of the anteromedial leaflet, and each papillary muscle is attached to chordae from both leaflets. Innervation from the left bundle branch causes the papillary muscles to contract just prior to ventricular systole, which optimizes tethering of the valve leaflets in the face of the high pressures generated during ventricular systole and prevents regurgitation. Additionally, the papillary muscles make up a significant proportion of the left ventricular mass and are important structural and contractile elements in determining ventricular function.

The mitral apparatus is in close approximation to several anatomical structures. The aortic annulus is in continuity with the mitral annulus, and the left and non-coronary cusps of the aortic valve are prone to injury when the mitral valve, particularly the anteromedial leaflet, is manipulated, Similarly, conducting fibers of the atrioventricular node penetrate near the right trigone, adjacent to the anteromedial mitral leaflet, and are prone to injury during invasive procedures, which may lead to transient or permanent conduction disturbances. Another potential source of conduction disturbance is injury to the AV nodal artery, which can arise from either the left or right (most frequent) coronary system. The circumflex artery courses adjacent to the posterolateral aspect of the annulus.

The great cardiac vein (GCV) and coronary sinus (CS) are of particular importance in mitral valve interventions, as several novel percutaneous annuloplasty devices have been conceived to manipulate these structures. These venous structures wrap around the superior-posterior aspect of the mitral annulus from commissure to commissure, encompassing the entire base of the posterolateral leaflet. The CS is a dynamic and contractile structure that can dilate to various degrees in patients with heart failure. Coronary arterial structures, particularly the circumflex artery and the obtuse marginal arteries, are often found in close approximation to the CS and must be carefully considered when designing devices that manipulate the GCV and CS.

E. Innovation for Functional Mitral Regurgitation

The ability to use a right-sided structure such as the CS to treat a left-sided problem, FMR, inspired many early efforts to leverage this fortuitous anatomical relationship. The initial insight to use the CS was inspired by electrophysiology procedures that developed bi-ventricular pacing leads for the CS and its tributaries to treat heart failure patients with left ventricular dyssynchrony,, and defibrillation leads positioned deep in the CS/GCV to treat atrial fibrillation with internal cardioversion. If the goal of a new percutaneous therapy was to be safer and simpler than existing surgical alternatives, the CS held great promise.

The fact that only two of the numerous CS approaches achieved CE Mark approval, and only one moved on to commercialization, highlights the challenge of the process.1,  While the schematic provided by the FDA detailing the steps of device innovation provides an accurate overview of the process (Fig. 2), success lies in attention to detail, and starts with a thorough analysis of the problem. The value of a comprehensive characterization of the anatomy and pathophysiology is that it provides insights into key design features required for a novel device to be safe and efficacious. We share highlights of the early stages of innovation to emphasize the importance of building a solid foundation to support the development process.

Figure 2. FDA’s Schematic of Device Development Process.

First Step in Innovation: Characterize Relevant Anatomy and Pathophysiology

Coronary Sinus Anatomy: While it is well known the CS runs in the atrioventricular groove, the variable superior position of the vein above the mitral annulus was important to characterize in the early stages of our innovation effort. The position of the CS in relation to the mitral valve annulus was initially quantified by Yamanouchi. Analysis of 50 human cadaver hearts demonstrated that the CS was superior to the annulus by 10.9 +/- 3.3 mm, with only modest variability in distance from P1 to P3. To complement the cadaver study of Yamanouchi, Maselli used multi-detector computed tomography to show that the distance from the CS to the mitral annulus ranged from 3.2 to 19.7 mm. While our team initially had reservations that this superior and variable position of the coronary sinus relative to the annulus may be problematic, more sophisticated insights eventually revealed clear benefits.

The saddle shape geometry of the mitral annulus was thoughtfully detailed in a study be Sheehan, documenting how the apex of the saddle at A2 and P2 facilitates leaflet coaptation. Sheehan further demonstrated that in patients with heart failure, the annular dilation contributes to loss of saddle shape geometry, thus reducing the degree of leaflet coaptation and setting the stage for greater degrees of FMR. Our preliminary investigations in the animal lab suggested that inward pressure on the left atrium from a device placed in the CS would place inward and upward force vectors on the annulus, which would reduce its septal-lateral diameter and partially restore its three-dimensional saddle shape anatomy, thus improving leaflet coaptation. Furthermore, optimal leaflet coaptation is facilitated by the dynamic three-dimensional conformational changes of the mitral valve annulus., This dynamic motion is reduced by rigid annuloplasty rings. Since devices in the CS are able to modify the annular position without directly restricting it, such a device was thought to be advantageous by maintaining the potential for dynamic motion of the annulus. We had reason to believe that a device in the CS might enhance leaflet coaptation by partially restoring the dynamic saddle shape configuration.

Coronary Artery Anatomy: The exact relationship between the coronary arteries and the coronary sinus was poorly characterized in the literature prior to innovation efforts to develop a percutaneous mitral annuloplasty device. Characterizing the anatomic variability was important to create a product requirement document. Studies at the University of Washington’s cardiac anatomy lab revealed that left dominant patients often have multiple obtuse marginal branches from the circumflex positioned in-between the CS and the mitral annulus along the length of the CS. In contrast, the circumflex artery in right dominant patients has a variable relationship with the CS in the region of P1, but the proximal portion of the CS near P2 and P3 is typically free of arterial cross-points. In addition, we noted the work of Pejkovic who studied 150 human cadaver hearts, and reported that the GCV is deep to the circumflex artery in 27% of cases, and is deep to the left anterior descending artery in 10% of cases. How these anatomical observations would translate into a device concept was not immediately clear, but the importance of the anatomical observations was unequivocal. Would we need to exclude left dominant patients? Would we need to image patients to select those in whom the artery went over (not under) the vein? Or would pressure from the device on the left atrium be sufficiently modest that compromise of arterial flow was not observed? Knowing that even a small incidence of myocardial infarction following percutaneous mitral annuloplasty would be unacceptable, and with the goal of erring on the side of caution to put patient safety first, we required that the device be recapturable in order to provide a further measure of safety.

His-Purkinje System: Just as the position of a percutaneous aortic valve is important to avoid disruption of the electrical conduction within the heart, the position of percutaneous mitral devices also needs to respect the anatomy of the His – Purkinje system. Our review of the literature revealed that the exact position of the AV Node is variable, but as a general rule it is adjacent to the ostium to the coronary sinus. Knowing that simple catheter manipulation around the AV node can cause transient heart block, we recognized the necessity of not allowing a CS device to extend into the right atrium.

CS Valves and Procedure Related Trauma: The ostium to the CS is protected by the Thebesian valve, and the junction between the CS and the GCV is the normal position of the Vieussens valve. The Vieussens valve can be either monocuspid or bicuspid, and without the proper tools or technique, can represent a challenge to deep venous access. A study of 37 cadaver specimens showed that the Thebesian valve was present in 86% of people, while the Vieussens valve was present in 57% of cases. Closer examination of the Thebesian valve confirms its anatomical variability, ranging from a crescent shape (typical) to a fenestrated barrier (rare). While study of the CS valves did not contribute to device design, it did provide insight on how variable morphology would likely affect procedure success rate, and perhaps procedure safety. Early trials of cardiac resynchronization therapy report a 2% rate of CS perforation,  and the development of specific tools to cannulate the CS have improved both their implant rate and their adverse event rate. Leveraging the tools and techniques developed by the electrophysiology community has been invaluable for CS mitral annuloplasty, especially for clinical cases with challenging venous anatomy.

Erosion, migration, and perforation are theoretical consequences of any device placed in the heart. The rare erosions caused by atrial septal defect closure devices teach that friction between a device and the myocardium is the mechanism leading to erosion. A design where the device is in intimate contact with the tissue was projected to facilitate endothelialization and minimize the risk of erosion and migration.

Dynamic Nature of FMR: While device conceptualization benefits from exhaustive characterization of the anatomy and pathophysiology, clinical trial design also benefits from the same process.

In a typical clinical setting, FMR is evaluated at rest. However, the labile nature of FMR has recently been well documented, and its implications on patient selection and reverse remodeling are only now becoming better appreciated. Lebrun and Keren both quantified the degree of FMR before and after exercise testing, and noted that the regurgitant volume of patients who underwent exercise testing almost doubled, while forward stroke volume substantially decreased., Lachman pointed out that elevated ventricular afterload may cause the increase in FMR noted by isometric exercise. If FMR is dynamic, and if the volume load on the LV is known to contribute to LV remodeling, then what patients are ideal candidates for valve repair? Surgery is only considered in select patients with severe FMR. Should the same recommendations be followed with regard to percutaneous devices, or might the risk:benefit equation shift, allowing us to factor in the dynamic nature of this regurgitant lesion? Interestingly, Tada showed that even mild FMR had detrimental effect on exercise tolerance, and Liang recently showed that patients in whom CRT therapy reduced even mild FMR demonstrated reverse ventricular remodeling.2 The contribution of FMR to morbidity 33, and mortality, , has been previously reported. While these insights may not contribute to device design parameters, they can be incorporated into a clinical trial strategy by requiring quantification of FMR, including patients with lesser degrees of FMR, and quantifying changes in LV dimensions over time.

Second Step of Innovation: Develop a Product Requirements Document

Before we could develop prototypes that would evolve into a mitral valve therapy, we needed a document that clearly and concisely defined whether the prototypes met the fundamental requirements that would enable clinical success. This document, the product requirement document, is derived from the comprehensive anatomical and physiological analysis just described. Simplicity in composition allowed the document to be a functional tool referenced real-time to evaluate the strengths and weaknesses of prototype iterations. For a device intended for the coronary sinus, the safety and efficacy product requirement details are listed in Table 1.

The above product requirement document was generated after months of careful and meticulous analysis of the coronary sinus and related structures, the mitral valve, and the hemodynamics of FMR. The document was comprehensive in scope but straightforward in format to enable functional use through the next step of the innovation process: prototype iteration.

Third Step of Innovation: Prototype Iteration

One need not set the bar too high to initiate the prototype development process. Indeed, any starting point will work, because careful analysis of the results of a failed experiment guides the next iteration. The most obvious starting point for a percutaneous device intended to simulate a mitral annuloplasty ring was a structure of similar shape. Undersized surgical rings have been shown to reduce FMR, thus we reasoned a percutaneous version may do the same, and served as our first prototype.

Analysis of a prototype where a C-shaped ring was shape-set at the end of a long piece of nitinol wire provided invaluable insights. The stiff nature of the ring was difficult to advance down the catheter, suggesting that a flexible structure might be beneficial. The device did not orient because it flipped to a point of least resistance, suggesting the need to better control orientation. Continuous inward pressure along the mitral annulus could not be achieved because there was no anchoring mechanism analogous to sutures used to stabilize a surgical annuloplasty ring. The variable position of the CS above the annulus could not be accounted for with the fixed-shaped ring because the ‘one-size-fits-all’ approach was inconsistent with the studied anatomy. With one simple prototype, an annuloplasty ring on a wire, the results were primarily a failure, but comparison of the results to the product requirement document easily pointed towards what the second iteration might look like.

The second prototype iteration tested the merits of a rigid straight bar. The shape was easier to advance down the catheter since it was less angled than the first prototype, and it oriented consistently. However, rather than applying continuous inward pressure along the annulus, it distorted the annulus near the commissures in order to push inward on the central P2 scallop of the mitral valve. Intuition suggested that consistent and optimized efficacy would not be achieved with this prototype. Analysis of the concept also suggested that an anchoring mechanism for the device would enable a continuous inward force along the peri-annular tissue.

At the end of an 18 month prototype iteration phase, we created a device concept completely unrelated to an annuloplasty ring, and one that the product requirement document confirmed should be clinically successful. The device was a fixed-length double-anchor concept where the anchors were hooped shaped helices on either end of the device, and the ribbon connecting the two anchors was arc-shaped in order to ensure proper orientation of the device as it was advanced through the curved catheter (Fig. 3). Safety and efficacy are generated by the technique, not by the device.

Figure 3. Fixed-Length Double-Anchor Mitral Annuloplasty Device.

The procedure (Fig. 4) involves advancing a 9F catheter deep into the CS / GCV. An appropriately sized device is advanced through the 9F catheter, and the first hoop-shaped nitinol anchor is unsheathed. The anchor passively expands due to the shape-set characteristics of the nitinol. The catheter is then used to open the anchor to its full height, and in so doing stretch the vein to achieve circumferential pressure and stable anchoring. With the proximal anchor still collapsed in the delivery catheter, traction is applied to the system, pulling the proximal anchor closer to the CS ostium. The amount of traction correlates with greater tissue plication, and greater reduction in the septal-lateral dimensions. This feature of ‘adjustability’ was conceived to address the anatomic variability that was so clearly noted in the first stage of the innovation process. Real-time echo is used during the procedure to optimize the amount of traction placed on the system. In parallel, real-time angiography is performed to ensure the adequacy of coronary artery flow. In the event that a flow disturbance is observed, the degree of traction can simply be reduced, and assessment of flow performed again. In the rare cases (primarily left dominant patients) in which flow disturbance can not be resolved with simply altering the amount of tissue plication, the delivery catheter can be advanced forward to collapse the deployed anchor, and recapture the device in a benign manner. If safety and efficacy are confirmed, then the proximal anchor is deployed near the CS ostium, and the device is released from the delivery system.

Figure 4. Key steps for percutaneous mitral annuloplasty procedure: 4a) Venogram defines length and dimensions of the CS; 4b) Distal anchor deployed deep in CS/GCV (Proximal anchor still collapsed in delivery system); 4c) Traction applied to delivery system to pull proximal anchor towards CS ostium, plicating tissue, reducing the septal lateral dimensions, and reducing FMR; 4d) Device permanently implanted with both distal and proximal anchors deployed to secure position.

While the device shape and features do not emulate other implantable cardiac devices, the iterative nature of the prototype analysis process helped the concept evolve. The analysis was contingent upon the initial anatomic and physiologic study that yielded the product requirements document. The unique features of any given device will of course depend on the disease process being treated; however, the process of prototype iteration driven by a product requirement document is invaluable and critical to optimize safety and efficacy of a novel therapy.

Animal Models to Assess Prototypes: Animal models were invaluable to assess serial prototypes. Crude prototypes were assessed with cadaver hearts, but animal models quickly became important to more accurately grade the merits of different prototypes. The specific animal model needs to be selected carefully in order to maximize input, and avoid false reassurance if positive results are achieved in an animal model that does not accurately represent the human anatomy.

Ovine, swine, and canine animal models were all required during the development of our percutaneous mitral annuloplasty therapy. Efficacy was assessed using an ovine tachycardiac induced cardiomyopathy model because the heart size mirrored the dimensions seen in clinical heart failure. In addition, the thinned ventricle mirrored the tissue turgor of dilated human hearts. As such, the consistency and degree to which early prototypes reduced FMR in this ovine heart failure model could be accurately analyzed.



A swine model was required to assess coronary artery flow given The circumflex artery in a sheep does not run along the atrioventricular groove; thus, providing opportunity for a false negative result. The position of the circumflex and marginal branches in a swine model simulates human anatomy, including the preponderance of right dominant anatomical features. Working in close collaboration with the group at Brigham and Women’s Hospital in Boston, we implemented IVUS, fractional flow reserve, and angiography to carefully quantify the effect of our device on coronary artery flow to rigorously assess safety, both acutely and chronically. While the majority of animals demonstrated minimal affect of the device on coronary flow, a subset of animals showed complete occlusion of arterial flow. The observations validated the device recapturability requirement. Furthermore, projecting out to clinical trials, the observations highlighted the need for careful monitoring of coronary flow during implant procedures to avoid adverse events.

A canine animal model was required to assess device anchoring. The CS anatomy in sheep and swine is distorted and dilated because the Vein of Marshall (the equivalent of a left-sided superior vena cava which drains into the mid-portion of the CS / GCV). We specifically developed a canine model of heart failure to simulate the CS dilation and altered venous compliance that is common in heart failure patients. Anticipating that anchoring would be more difficult if the compliant nature of the GCV was compromised, the canine model most closely represented the projected human anatomy.



Animal models can never fully replicate the variability of human anatomy and pathology. Our experience was that complementary animal models provided key input for different elements of device design. Extrapolation of animal data to projected human conditions is important, and challenging.

Fourth Step of Innovation: Design Freeze

It is difficult to overstate the importance and complexity of design freeze in the innovation pathway. Design freeze implies that iteration and improvement to a device and delivery system are fixed, no subsequent changes are allowed, and the entire process is going to transition from creativity and improvement to regulated and controlled testing of a fixed concept. Complicating the dynamic is the cross-functional hand-off. The mindset of an innovator (creative and analytical) is typically different than the mindset of a tester (methodical). Both play a critical role. Different teams of people are typically hired to assist with the testing phase of the process. Furthermore, the business imperative weighs in. At some point, one needs to stop developing, and start executing. Premature execution can result in sub-optimal clinical experience. Delaying execution can result in burning excessive cash and thus preventing the achievement of a critical clinical milestone that might enable future fundraising. The combination of science, team dynamics, and cross-functional coordination all play out through the process.

An example of a conundrum we faced around the time of design freeze related to the reliability of device anchoring. The canine heart failure model was developed specifically to assess device anchoring because of its anatomical and physiological similarity to human tissue. In our anchoring study, the device was clearly more temperamental than in the other animal models. Did the dog model accurately simulate people? Was there artifact created by an external pacing lead placed near the AV groove creating scar tissue? If we were to iterate further to improve anchoring, how good is good enough?

Identifying a patient population with advanced disease such that the risk:benefit of these first-in-human experiences is ethically appropriate is clearly important. Furthermore, the understanding of the clinical community that first generation devices are almost invariably receptive to further device improvement is likewise important. Our first in human studies confirmed the findings from our canine animal model that the device did not anchor reliably, thus requiring a subtle redesign. The regulatory system in Europe enabled a rapid return to our clinical trial to enroll our first 30 patients in a feasibility study. This ability to maintain clinical traction was important not only from a scientific and innovation standpoint, but also from a market perception standpoint.

Fifth Step of Innovation: Bench and Verification Testing

Verification testing is a process whereby repeated testing is performed to confirm that a device performs to the standards specified in the engineering documents. The tests include assessments of dimensional integrity, sterility testing, packaging integrity, chronic animal studies, and fatigue testing to ensure the structural integrity of the device. While part of this process can be performed by engineers in verification testing, other parts of the process require close collaboration between the individual who guided the innovation through the development process and the team doing the testing.

Fatigue testing is perhaps the most important example of where collective insights are the key to produce the most accurate test. A device that is going to be permanently implanted in the heart requires 400 – 600 million cycles of fatigue testing to ensure structural integrity matching 10 – 15 years of patient life. Setting the boundary conditions for this test is critical. Boundary conditions that are too lenient may fail to detect a flaw in the engineering design. Boundary conditions that are too rigorous and not clinically realistic may lead to further unnecessary device iteration, a delay in timelines, and a delay in generating clinical data. Getting the balance right is important.

Insights from the first and third step of the innovation process, characterizing the problem and performing device iterations in different animal models, were of paramount importance. In our canine model of heart failure, approximately 4 centimeters of proximal anchor displacement was required to achieve consistent FMR reduction. This data point helped establish the boundary conditions for a fatigue test. However, when the verification test was performed to convert 4 cms of displacement into a force, a swine animal model was used instead of a canine model. Since swine have an excessively dilated CS (due to the insertion of the Vein of Marshall), 4 cms of proximal anchor displacement primarily shifted the device from the outer curvature to the inner curvature of the vein, and a very modest force was generated by the test. This created a falsely low force value, which was used to establish the boundary condition for the fatigue test. The fact that the result was reproducible only highlights the difference between accuracy and precision. The distinction may appear subtle, but it highlights the importance of attention to detail. The device passed the fatigue test, but clinical experience revealed a 25% fatigue fracture rate in the TITAN trial. The engineering failure was clinically silent, because the device was specifically designed to be encapsulated in the tissue. Nonetheless, a third redesign was required to address the nitinol fatigue fracture, and delays in generating clinical momentum ensued. This example emphasizes how attention to detail, and cross-functional communication, are critical to optimize the innovation process.

The importance of thoughtful and conservative extrapolation of animal data to create a first pass estimate of appropriate boundary conditions in device development is important. However, the variability in clinical pathology will always be greater than the variability that one can simulate in animal studies. Prior to first in man studies, the conundrum will always exist: “How accurately do the data from animal studies represent the human disease being treated?” “How much time and money should be spent to iterate a device to optimize outcomes in an animal model, knowing that clinical data will always trump animal data.” “With design freeze and verification testing, how good is good enough?” The position one takes on these questions is different depending on whether your perspective is clinical, engineering, regulatory, or business. The responsibility of an innovation team is clearly to optimize communication to ensure that the most thoughtful multi-disciplinary decisions are made.

Sixth Step of Innovation: Clinical Trials

The purpose of early feasibility trials is to simply establish the safety of the device and procedure. When regulatory systems allow for first-in-human studies to confirm the accuracy of boundary conditions used in the development process, then the inclusion criteria should be designed to match the risk:benefit profile appropriately. When boundary condition accuracy and safety are confirmed, expanding the clinical trial inclusion criteria is appropriate.

Clinical Standard of Care: The first step to develop a Phase I trial is to identify what the standard of care is for the target patient population. For mitral innovations, this simple task has generated enormous confusion and controversy. The confusion stems in large part from the inadequate lexicon employed in the early literature. Differentiating between primary (degenerative) and secondary (functional) mitral regurgitation started about the same time that numerous mitral innovations began. Historical surgical data reporting survival curves for patients who underwent mitral valve repair failed to specify that the majority of the repair was for primary MR, not functional MR. When surgery was performed for FMR, it was typically in the context of surgical revascularization. The efficacy data were therefore confounded, and the surgical mortality was significantly higher in patients who underwent concomitant revascularization and mitral annuloplasty. While a few single arm studies suggested that surgical repair for FMR was feasible, no surgical study established the risk:benefit of this approach in the heart failure population, and as such the indication for surgery remains a class IIb indication.

Given the confounded clinical data, and insufficient differentiation between FMR and degenerative MR, it is understandable that the first mitral innovation (MitraClip) identified surgery as the standard of care, but then failed to differentiate between FMR and primary MR in its patient selection criteria, thus confounding the data analysis (2). When surgery was shown to be more efficacious but slightly less safe than the MitraClip, the regulatory approval process got complicated. Adding to the confusion from the trial was a high-risk registry arm, which included FMR patients for whom surgery is not considered the standard of care.

To define the control population, future mitral innovations need to clearly differentiate patients with primary MR from those with FMR. Since the guidelines for FMR clearly state that medical therapy is the standard of care, the study design becomes straightforward. The true value of a percutaneous intervention for FMR may be in the ability to intervene earlier in the disease course. The literature is beginning to clearly show that even mild FMR at rest contributes to reverse remodeling (likely due to its dynamic nature) (34). Therefore, for percutaneous mitral therapies that are able to clearly demonstrate the right risk:benefit profile, early intervention may be ideal for HF patients to slow the deterioration brought on by FMR. For those patients with more advanced disease, more aggressive interventions, including mitral valve replacement, may become a viable alternative.

Efficacy Endpoints: For innovations targeting heart failure patients with FMR, the input of the heart failure community is invaluable to ensure that efficacy endpoints are thoughtfully selected. Recognition that patient morbidity is derived not only from exercise intolerance but also from frequent heart failure hospitalizations suggests that exercise testing and heart failure hospitalizations might be appropriate efficacy endpoints. If a mortality endpoint is desired, then risk stratifying patients to identify those with a sufficiently high event rate is important to power a trial. Surrogate endpoints like reduction in FMR or LV dimension are theoretical options, but as a general principle surrogate endpoints are inferior to direct clinical measures.

Matching the patient population and the trial design with the efficacy endpoints is also important. If a hard endpoint like mortality is considered, risk-stratifying patients a priori with a tool like the Seattle Heart Failure Model in an effort to ensure that the anticipated event rate is sufficiently high to warrant a manageable trial size is an option. The data is still inconclusive as to if this strategy is successful. The trial duration should be sufficiently long to capture the anticipated event rate.

Patients with lesser degrees of FMR may experience significant morbidity without excessive annual mortality. In this group, one could conceive of an exercise test, like the six minute walk test, to measure changes in exercise performance. A large and durable effect size may negate the need to blind a control group, but study blinding would add to the rigor of the results if methodologically achievable.

For patients with primary MR, quantifying the change in MR may be more appropriate as an efficacy endpoint to the extent that surgery is indicated for asymptomatic patients with severe MR. Leveraging the echocardiographic tools to quantify the changes in FMR allows for a more rigorous comparison between patient cohorts.

Seventh Step of Innovation: Commercialization

The key strategy in designing a pivotal trial involves refining the patient population with inclusion / exclusion criteria in such a way that the control population has a consistent standard of care, and the primary safety / efficacy endpoints have a high likelihood of meeting statistical significance based on insights from single-arm feasibility trials. In the commercial setting, the restrictions on patient selection are lifted, and the training program is expanded. As such, the ability of a novel therapy to show comparable clinical benefits as demonstrated through clinical trials is the final innovation hurdle. Regulatory bodies thoughtfully use trial registries to track this transition from trial to commercial use. Just as the clinical community has a responsibility to understand the delicate nature of first in human studies, we likewise have a responsibility to understand the delicate nature of commercialization. The business imperative may pull in new sites such that the learning curve of new users creates data inferior to clinical trials. The heterogeneity of patients implanted may create safety and efficacy data not seen in clinical trials. The training program may require modification to efficiently transfer the critical information to a broad flux of new users eager to access a technology that might uniquely benefit their patients. If iteration and good communication is a critical component of the prototype development process, those key ingredients are likewise important through the commercialization phase of a technology so that subtle refinements can continue to be implemented to best serve the needs of all patients.

For our CS mitral annuloplasty therapy, the first three clinical trials allowed for subtle iterations to the device design to enable consistent device performance. While the process was deliberate, and progress intermittent, the final product undoubtedly benefited from the iterative process. If approximately 80 patients were enrolled during regulated clinical trials over the course of approximately 8 years, the award of the CE Mark in 2012 enabled commercialization and the treatment of approximately 100 patients in a single year. The commercialization phase easily benefits from the tremendous learning that occurred during device development and iteration phases of innovation. The ability of CS annuloplasty to achieve long-term stability is now less a function of innovation, and more a function of economics for investors or corporates who choose to invest in a percutaneous approach for the treatment of heart failure.

F. Alternative Innovations for Mitral Regurgitation

The incidence and prevalence of MR, and the heterogeneous pathology, has generated numerous alternative innovations to treat MR. The majority of these alternative devices intervene directly upon the mitral leaflets and the mitral apparatus, either with a repair technology (e.g., MitraClip (Fig. 5)) or with a replacement technology. Some of these technologies are attempting to address primary MR, and some are attempting to address severe degrees of FMR. The development process for each technology must go through the same key steps as described for CS based mitral annuloplasty.

Figure 5. MitraClip Image: (A) The MitraClip device in the open position prior to leaflet plication. (B) Edge-to-edge plication in an animal pathology specimen, creating a double-orifice valve after deployment of the device (Reproduced from European Heart Journal (2011) 32(3), 249–257) with permission from Oxford University Press.

That process starts with simply understanding the key physiology that will affect any device concept. For example, flow through the mitral orifice is low velocity and largely passive, conditions which increase the risk of thrombosis relative to a high flow, high pressure, environment like the aortic outlet, for example. Addressing the risk of thrombus formation is critical in the development of a mitral valve device positioned in the left ventricle. Indeed, mechanical valve replacements in the mitral position generally have higher and stricter anticoagulation requirements that those in the aortic position for this very reason.

Secondly, the left heart chambers, the components of the mitral apparatus, and the important surrounding anatomic structures mentioned above may be highly dynamic throughout the cardiac cycle and may vary in morphology from patient to patient, as previously discussed. Mitral valve devices must be adaptable to these conditions to be efficacious and to ensure against embolization. They must also be able to withstand the high closing pressures generated by left ventricular systole, especially in hypertensive older patients. Additionally, mitral valve devices ought to have reliable recapture mechanisms. Impingement on or damage to important related structures, including the conduction system, the aortic valve cusps, and the coronary arteries has the potential to cause serious harm, and must be easily reversible by the operator.

For those devices addressing primary MR, a second order analysis must include characterization of the relevant pathology. The most common etiology of primary MR in industrialized countries is mitral valve prolapse due to myxomatous, or degenerative, valve disease. Myxomatous degeneration is characterized by thickening and redundancy of the leaflets and can involve interchordal hooding, chordal elongation, and annular dilation. Chordal rupture can occur with myxomatous degeneration as well, resulting in flail leaflet. Mitral valve prolapse can also be due to a number of connective tissue diseases and is associated with an increased risk of infective endocarditis and sudden cardiac death in addition to the development of clinically significant MR. Leaflet damage secondary to rheumatic disease, infective endocarditis, and trauma can lead to significant MR, as can papillary muscle damage secondary to myocardial ischemia, but these etiologies are less common than myxomatous degeneration.

Similar to the product requirements document created to direct prototype iteration for CS annuloplasty, similar guidance documents must be created for these alternative strategies. An excellent example of the key challenges for a mitral replacement technology was summarized by Neovasc in their efforts to create a mitral valve replacement system. Their list of key challenges is outlined in Table 2.

The first in human studies are commencing with valve replacement technologies. CardiAQ was the first to perform a mitral replacement procedure (Fig. 6), and the insights and observations directed their focus back to the pre-clinical lab. Similarly, Neovasc (Fig. 7), reported first in human experience in 2014. The clinical data will no doubt inform the continuous iterative process that all mitral innovations undergo.

Figure 6. The CardiAQ percutaneous prosthetic mitral valve is delivered using a trans-septal approach (left) and locks onto the inferior (middle) and superior (right) surfaces of the mitral annulus (from: Chaim PT & Ruiz CE, Percutaneous transcatheter mitral valve repair: a classification of the technology. Reproduced from JACC Cardiovasc Interv 2011 Jan;4:1-13 with permission from Elsevier.

Figure 7. The Tiara transcatheter mitral valve, deployed ex vivo (A) and in vivo trans-apically (B).

Techniques for endovascular neochordal implantation for potential use in patients with ruptured or elongated chordae resulting in mitral valve prolapse have been described as well. The only currently competitive trascatheter chordal elongation technology is the NeoChord device (NeoChord, Inc, Minnetonka, MN, USA). The NeoChord device uses a transapical approach to stitch a synthetic chord through the affected leaflet. The stitch is then secured outside the ventricular wall thus providing chordal support to the leaflet while allowing the annulus and the ventricle to maintain their beneficial geometry. The first implantation of this device in a human was reported in 2010.



The most experience with percutaneous mitral repair has been generated by the MitraClip therapy. The MitraClip (Abbott, Inc, Abbott Park, IL) is a leaflet grasping device that has been evaluated for treatment of mitral valve prolapse. The device approximates and joins together the free edges of the anterior and posterior leaflets creating a double orifice while limiting the systolic motion of the prolapsing leaflet and correcting MR in a manner similar to that described surgically by Alfieri and colleagues. The device utilizes a transeptal approach to cross the mitral valve into the left ventricle. The MitraClip is a multiaxial transeptal system that approximates and joins the mitral leaflets with one or more retrievable metallic clips.

The EVEREST II trial compared the MitraClip to conventional open mitral valve surgery in patients with moderate to severe (3+ or 4+) MR using a primary composite end point of freedom from death, from surgery for mitral-valve dysfunction, and freedom from grade 3+ or 4+ mitral regurgitation at 12 months post-procedure. While the MitraClip system did not perform as well as surgery on the composite endpoint, the MitraClip group suffered fewer major adverse events and showed improvements from baseline in left ventricular size, New York Heart Association functional class, and quality-of-life measures. The MitraClip is approved in Europe, and its use has been steadily increasing.

G. Conclusions

Innovation for the mitral valve has been driven by the incidence and prevalence of MR, and the morbidity and mortality that is in direct proportion to the severity of the lesion. The greatest unmet need is for FMR since heart failure patients are poor surgical candidates, yet even mild degrees of FMR contribute to their disease progression.

Independent of the technology or specific device conceptualized, the fundamental process for mitral innovation remains the same. Accurate and comprehensive characterization of the anatomy and pathophysiology guides the development of a product requirement document. The product requirement document guides the critique of serial prototype iterations. Prototype iteration requires close collaboration between the engineering and clinical team to ensure that the key requirements are met prior to design freeze. Verification testing leverages the insights and boundary conditions established during the prototype testing phase, but in many cases require further input from first in human trials. Phase I clinical trials not only establish the safety profile of a new therapy, but also generate efficacy data to help power a pivotal trial. And lastly, commercialization enables expansion of a therapy to a broader patient population and a broader user population.

The process is complicated, multi-disciplinary, expensive, iterative, and long, but the clinical legacy generated by conceptualizing and developing a novel therapy based on rigorous scientific principles enables improved care for patients who depend on our commitment and perseverance to journey down the innovation pathway.

References

1. Siminiak T, Wu JC, Haude M, Hoppe UC, Sadkowski J, Lipiecki J, Fajadet J, Shah AM, Feldman T, Kaye DM, Goldberg SL, Levy WC, Solomon SD, Reuter DG. Treatment of functional mitral regurgitation by percutaneous annuloplasty: Results of the TITAN Trial. Eur J Heart Failure 2012; 14(8):931-938.
2. Liang YJ, Zhang Q, Fung JWF. Impact of reduction in early and late systolic functional mitral regurgitation on reverse remodelling after cardiac resynchronization therapy. Eur Heart J 2010;31:2359-2368.
3. Nkomo VT, Gardin JM, Skelton TN et al. Burden of valvular heart diseases: a population-based study. Lancet 2006; 368:1005-11.
4. Carpentier A. Cardiac valve surgery: the “French connection.” J Thorac Cardiovasc Surg. 1983; 86: 323–337.
5. American College of Cardiology/American Heart Association Task Force on Practice Guidelines, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society of Thoracic Surgeons , Bonow R.O., Carabello B.A., et al: ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing committee to revise the 1998 Guidelines for the Management of Patients With Valvular Heart Disease): developed in collaboration with the Society of Cardiovascular Anesthesiologists: endorsed by the Society for Cardiovascular Angiography and Interventions and the Society of Thoracic Surgeons. Circulation 2006; 114: e84-e231
6. Detaint D, Sundt TM, Nkomo VT, et al. Surgical correction of mitral regurgitation in the elderly: outcomes and recent improvements. Circulation 2006; 114: 265–72.
7. Enriquez-Sarano M, Schaff HV, Orszulak TA, Tajik AJ, Bailey KR, Frye RL. Valve repair improves the outcome of surgery for mitral regurgitation. Circulation 1995; 91: 1022–28
8. Mirabel M, et al. What are the characteristics of patients with severe, symptomatic, mitral regurgitation who are denied surgery. Eur Heart J. 2007. 28(11):1358-65.
9. Patel JB, Borgeson DD, Barnes ME, et al. Mitral regurgitation in patients with advanced systolic heart failure. J of Cardiac Failure 2004; 10(4):285-291.
10. Ennezat PV, Marechaux S, Asseman P, Lejemtal TH, Van Belle E, Bauters C, De Groote P. Functional mitral regurgitation and chronic heart failure. Minerva Cardioangiol 2006; 54(6) 725-733.
11. Junker A, Thayssen P, Nielsen B, Andersen PE. The hemodynamic and prognostic significance of echo-doppler-proven mitral regurgitation in patients with dilated cardiomyopathy. Cardiology 1993; 83:14-20
12. Trichon BH, Felker GM, Shaw et al. Relation of frequency and severity of mitral regurgitation to survival among patients with left ventricular systolic dysfunction and heart failure. Am J Cardiol 2003; 91:538-543
13. Braunberger E, Deloche A, Berrebi A, Carpentier A, et al. Very long-term results (more than 20 years) of valve repair with Carpentier’s techniques in nonrheumatic mitral valve insufficiency. Circulation 2001; 104(suppl I):I-8-11
14. Flameng W, Herijgers P, Bogaerts K. Recurrence of mitral valve regurgitation after mitral valve repair in degenerative valve disease. Circulation 2003; 107: 1609-1613.
15. Acker MA, Bolling S, Shemin R, Kirklin J, Oh JK, Mann DL, Fessup M, Sabbah HN, Starling RC, Kubo S. Mitral valve surgery in heart failure: Insights form the Acorn Clinical Trial. J Thorac and Cardiovasc Surg 2006; 132(3):568-577.
16. Abraham WT, Fisher WG, Smith AL, et al. Cardiac Resynchronization in Chronic Heart Failure. N Engl J Med 2002; 346:1845-53.
17. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med 2004; 350-2140-50.
18. Timmermans C, Rodriguez LM, Reuter D, Management of atrial arrhythmias secondary to severe congenital heart disease with the Atrioverter. Pacing Clin Electrophysiol. 2000 Jul;23(7):1181-3.
19. Harnek J, Webb JG, Kuck KH, Tschope C, Vahanian A, Buller CE, James SK, Tiefenbacher CP, Stone GW. Transcatheter implantation of the MONARC coronary sinus device for mitral regurgitation. J Am Coll Cardiol Intv 2011;4:115-22.
20. Yamanouchi Y, Egawa O, Hisatome I. Activation mapping from the coronary sinus may be limited by anatomic variations. PACE 1998; 21:2522-2526.
21. Maselli D, Guarracino F, Chiaramonti F, Mangia F, Borelli F, Minzioni G. Percutaneous Mitral Annuloplasty: An anatomic study of human coronary sinus and its relation with mitral valve annulus and coronary arteries. Circulation. 2006;114:377-380.
22. Kaplan SR, Bashein G, Sheehan F, et al. Three-dimensional echocardiographic assessment of annular shape changes in the normal and regurgitant mitral valve. Am Heart J 2000;139(3)378-387.
23. Aybek T, Risteski P, Miskovic A, Simon A, Dogan S, Raman U, Moritz A. Seven years experience with suture annuloplasty for mitral valve repair. J Thorac Cardiovasc Surg 2006; 131;99-106.
24. Herregods MC, Tau A, Vandeplas A, et al. Values for mitral valve annulus dimensions in normals and patients with mitral regurgitation. Echocardiography 1997; 14:529-533.
25. Pejkovic B, Bogdanovic D. The great cardiac vein. Surg Radiol Anat 1992; 14:23-28.
26. Ortale JR, Gabriel EA, Iost C, and Marquez CQ. The anatomy of the coronary sinus and its tributaries. Surg Radiol Anat 2001; 23:15-21
27. Silver MA, Rowley NE. The functional anatomy of the human coronary sinus. Am Heart J 1988; 115:1080-1084.
28. Abraham WT, Fisher WG, Smith AL, et al. Cardiac Resynchronization in Chronic Heart Failure. N Engl J Med 2002; 346:1845-53.
29. Bristow MR, Saxon LA, Boehmer J, et al. Cardiac-Resynchronization Therapy with or without an Implantable Defibrillator in Advanced Chronic Heart Failure. N Engl J Med 2004; 350-2140-50.
30. Lebrun F, Lancellotti P, Peirard LA. Quantitation of functional mitral regurgitation during bicycle exercise in patients with heart failure. J Am Coll Cardiol 2001; 38:1685-92.
31. Keren G, Katz S, Strom J, et al. Dynamic mitral regurgitation: An important determinant of hemodynamic response to load alterations and inotropic therapy in severe heart failure. Circulation 1989 Aug; 80(2):306-13.
32. Lachmann J, Shirani J, Prestis KA, et al. Mitral ring annuloplasty: An incomplete correction of functional mitral regurgitation associated with left ventricular remodeling. Curr Cardiol Rep 2001; 3:241-246.
33. Tada H, Tamai J, Takaki H, Ohnishi E, Okano Y, Yoshioka T. Mild mitral regurgitation reduces exercise capacity in patients with idiopathic dilated cardiomyopathy. International J Cardiol 1997;58:41-45.
34. Bach DS, Bolling SF. Improvement following correction of secondary mitral regurgitation in end-stage cardiomyopathy with mitral annuloplasty. Am J of Cardiology 1996. 78:966-969.
35. Blondheim DS, Jacobs LE, Kotler MN, et al. Dilated cardiomyopathy with mitral regurgitation: Decreased survival despite a low frequency of left ventricular thrombus. Am Heart J 1991; 122:763.
36. Grigioni F, Enriquez-Sarano M, Zehr KJ, et al. Ischemic mitral regurgitation: Long term outcome and prognostic implications with quantitative doppler assessment. Circulation 2001:103:1759-1764.
37. Lam BK, Gillinov AM, Blackstone EH, et al. Importance of moderate ischemic mitral regurgitation. Ann Thorac Surg 2005; 79:462-470.
38. Kaye DM, Byrne M, Alferness C, et al. Feasibility and Short-Term Efficacy of Percutaneous Mitral annular reduction for the therapy of heart failure-induced mitral regurgitation. Circulation 2003; 108:1795-1797
39. Maniu CV, Patel JB. Reuter DG. et al. Acute and chronic reduction of functional mitral regurgitation in experimental heart failure by percutaneous mitral annuloplasty. J Am Coll Cardiol 2004; 44:1652-61.
40. Schofer J, Siminiak T, Haude M, Herrman JP, Vainer J, Wu JC, Levy WC, Mauri L, Feldman T, Kwong RY, Kaye DM, Duffy SJ, Tubler T, Degen H, Brandt M, Van Bibber R, Goldberg S, Reuter DG, Hoppe UC. Percutaneous mitral annuloplasty for functional mitral regurgitation: Results of the CARILLON mitral annuloplasty device European Union study. Circulation. 2009;120:326-333.
41. Harris KM, Sundt T, Aeppli D, Sharma R, Barzilai B. Can late survival of patients with moderate ischemic mitral regurgitation be impacted by intervention on the valve? Ann Thorac Surg 2002; 74:1468-1475.
42. Bach DS, Bolling SF. Improvement following correction of secondary mitral regurgitation in end-stage cardiomyopathy with mitral annuloplasty. Am J of Cardiology 1996. 78:966-969.
43. Levy WC Mazaffarian D, Linker DT, Sutradhar SC, Anker SD, Cropp AB, Anand I, Maggioni A, Burton P, Sullivan MD, Pitt B, Poole-Wilson PA, Mann DL, Packer M. The Seattle Heart Failure Model: Prediction of survival in heart failure. Circulation. 2006;113:1424-33.
44. Sondergaard L. CardiAQ program update: Featuring the world’s. First successful transcatheter mitral valve implant [abstract]. TCT conference. Miami, 2012 [quoted Dec 2012]. Available at: http://www.tctmd.com/show.aspx?id=114961
45. Banai S, et al. Transapical mitral implantation of the tiara bioprosthesis: pre-clinical results. JACC Cardiovasc Interv. 2014 Feb;7(2):154-62.
46. Laing G, Dupont PE, Beating-heart mitral valve chordal replacement. Conf Proc IEEE Eng Med Biol Soc. 2011;2011:2476-9
47. Seeburger J, Borger MA, Tschernich H, Leontjev S, Holzhey D, Noack T, Ender J, Mohr FW. Transapical beating heart mitral valve repair. Circ Cardiovasc Interv 2010;3:611-612
48. Maisano F, Michev I, Rowe S, Addis A, Campagnol M, Guidotti A, Colombo A, Alfieri O., Transapical endovascular implantation of neochordae using a suction and suture device. Eur J Cardiothorac Surg. 2009 Jul;36(1):118-22.