The principal therapies for valvular heart disease are surgical. Catheter-based treatment, traditionally, has had a limited role in treating stenotic lesions of the mitral, pulmonary and aortic valves. Recently, however, percutaneous valvular intervention has drastically improved, secondary to advances in technology and innovative treatment strategies. With the proven feasibility of percutaneous pulmonary1,2 and aortic valve replacement,3 and repair of mitral regurgitation,4 the field of valvular intervention has become the newest sub-speciality of the cardiovascular interventionalist.
The predominant experience in percutaneous valvular intervention is the replacement of the pulmonary and aortic valves, though only phase I trials have been completed. Percutaneous repair of mitral regurgitation using the Evalve mitral clip, however, has moved into a randomised phase II trial. In addition, several start-up companies have developed novel devices that are just now reaching animal testing, and will be soon implanted in man. The experience with all these devices has been limited to a few hundred cases in a select number of centres worldwide. This article summarises the state of affairs of on-going phase I and phase II trials, knowing the limitations of such a review in a field that is evolving rapidly.5-8
The first percutaneous aortic valve replacement was performed by Alain Cribier in April of 2002.3 Since then, multiple companies have engineered percutaneous heart valves (PHV), primarily for aortic stenosis. Only two, however, have experience with human implantation: Edwards Lifesciences (>100 patients) and CoreValve (>70 patients). The two valves are quite different, not only in their engineering aspects but also in their development strategies.
The Edwards Lifesciences (Irvine, CA) PHV is an equine pericardial valve that is sutured to a balloon-expandable, stainless steel stent frame (see Figure 1a). The stent valve is crimped using a specially designed crimper to decrease the outer diameter from 23 or 26 mm to approximately 8 mm (22-24 French compatible, respectively. See Figure 1b). The initial feasibility study in Europe9,10 primarily used femoral venous access and an antegrade trans-septal route to implant the PHV. At present, the most common approach is retrograde from the femoral artery.11
Though the initial strategy for this stent valve was a complete percutaneous technique, patients are currently undergoing general anaesthesia with mechanical ventilation during the procedure, and surgical closure of the femoral artery after the procedure.11 The outcomes from the most recent phase I trials have been encouraging: procedural success 78%, procedural mortality 0%, 30-day mortality 11%, final valve area 1.6 m2, paravalvular leak grade 1.8±0.6.11 Patients from the European phase I trial are now reaching three years post implantation without signs of valve deterioration. In addition to a femoral approach, transapical implantation in seven patients (limited antero-lateral thoracotomy, access to the left ventricular apex and catheter implantation of a PHV without cardiopulmonary support) had a 100% procedural success and a 14% 30-day mortality.12
The current generation of CoreValve (Irvine, CA) is a porcine pericardial valve sutured within a three-level self-expanding frame (see Figure 2a). In its collapsed form, the device is 18 French compatible. Because this percutaneous valve is anchored in both the aortic annulus and ascending thoracic aorta (see Figure 2b), it can be used for the treatment of inoperable aortic stenosis as well as aortic regurgitation. The device has been used to treat degenerated surgical bioprosthetic valves as well. Since its inception, the CoreValve procedure involved general anaesthesia, mechanical intubation and surgical access/closure.13 In addition, the patients were briefly supported by cardiopulmonary bypass and, most recently, percutaneous cardiac support. The earliest phase I study14,15 with the first-generation device demonstrates a 78% procedural success (peak valve gradient decreased from 74 to 14 mmHg), 14% peri-procedural mortality, and 35% in-hospital mortality. In-hospital mortality was significantly less in patients from the later part of the feasibility trial (14% or one of 7 patients).15 The subsequent generation devices have improved feasibility and safety data, which is related to a newer design as well as increased operator experience. Both CoreValve's and Edwards' percutaneous valves hold real promise for the future, and will definitely be the benchmark against which newer devices will be compared.
The first percutaneous pulmonic valve replacement (PPVR) was performed and reported by Philipp Bonhoeffer in October 2000.2 Since then, over 100 patients have had this valve implanted (Medtronic, Minneapolis, MN) in Europe and the UK, and one patient has been implanted in the US (Edwards Lifesciences).16 Procedural success with PPVR has been exceptional, with very low mortality, but the experience has been almost exclusively single operator.
The PPV is made of a bovine, jugular venous valve sutured into a platinum-iridium balloon-expandable stent. The size of the valve is currently limited to 22 mm. PPVR has been performed in patients with degenerated pulmonic conduits (with or without bioprosthetic valves) of 18-22 mm diameter to correct the untoward effects of conduit/valve stenosis and/or regurgitation17 (see Figure 3). The vast majority of patients have had previously operated tetralogy of fallot or pulmonic atresia.2,18,19 Those patients who have had patch repair of the outflow tract present the largest challenge as progressive dilation of the repaired area, often >30 mm, prohibits safe valve implantation.20 Femoral and internal jugular venous access sites have both been used successfully.
The phase I trial in Europe has demonstrated a 100% success rate with 0% procedural mortality.19 There have been some adverse events with the procedure (4.5%) such as homograft rupture, stent-valve migration and coronary artery compression.19 Stent frame fracture (16%) and 'hammock' flow-deformity of the tissue valve (12%) have been corrected on the current generation PPV.19 Implantation of the valve has been associated with significant improvements in performance status of patients.17 Additionally, right ventricular dilation, an ominous sign of impeding irreversible right ventricular (RV) failure, has been reversed as assessed by magnetic resonance imaging (MRI) measurements.17 Compared with traditional surgical outcomes, the early morbidity of PPVR is less, with a shorter hospital stay.18 Some of the earliest patients have reached five-year follow-up; 64% of patients were free from surgery or death.19 Extensive long-term data on the durability of the PPV is pending, and multicentre trials will establish the ability to generalise the procedure among different operators in Europe, Canada and the US.
Percutaneous treatment of mitral regurgitation has been focused on mitral valve repair rather than replacement. Several strategies for percutaneous mitral valve repair (PMVR) have emerged to address the differing valvular pathologies: leaflet abnormalities, chordal abnormalities, annular dilatation, papillary muscle dysfunction, and left ventricular dysfunction. The devices developed are aimed at treating degenerative (mitral valve prolapse) and functional mitral regurgitation (MR) (primarily ischaemic MR).
The first phase I trial of PMVR4 has been completed using a mitral clip (MitraClip™, Evalve, Menlo Park, CA) that mimics the surgical Alfieri stitch.21 From a trans-septal approach, a 24 French grasper is inserted into the left atrium and directed towards the central mitral regurgitant jet (see Figure 4a). The anterior and posterior leaflets are captured, and the clip is deployed after successful reduction of MR is demonstrated. The majority of patients in the phase I trial had degenerative MR (93%), and a small number (two patients, 7%) had ischaemic MR.4 The results showed that this approach was feasible (89%, 24 of 27 patients received clips), and major adverse events at one month were 15% (three patients with partial clip detachment and one with a post-procedural stroke that resolved by 30 days, not related to clip detachment). Of the 18 patients discharged with MR Ôëñ grade 2, 13 had a lasting result at six months (two patients had partial clip detachment, previously mentioned, and three patients had recurring MR grade 3). The earliest patient to have received this repair has reached two years follow-up22 with minimal MR and favourable left ventricular remodelling. This technique, similar to the surgical Alfieri stitch, may be a lasting repair for a select group of patients with degenerative MR or functional MR without significant annular dilation.23 Edwards Lifesciences has also developed a device that uses a trans-septal approach to place a stitch rather than a clip on the mitral leaflets (see Figure 4b). A phase I trial in Europe is on-going.
The devices to treat functional MR move the posterior annulus towards the anterior annulus, thus decreasing the septal-lateral distance of the mitral apparatus and increasing leaflet coaptation. The current approaches include: devices that are placed in the coronary sinus and shorten the underlying posterior annulus; devices that tether the posterior annulus to the atrial septum; devices that tether the posterior annulus and myocardium to the anterior myocardial wall; and devices that directly reduce the posterior annulus from an intra-cardiac approach (ventricular).
In four patients with ischaemic MR, a coronary sinus device, consisting of two stent anchors connected by a 'bridge' that shortens at body temperature (Edwards Lifesciences, see Figure 5a), has been implanted successfully.24 Though the MR decreased from grade 3 to 1.6, the device 'bridge' separated within 90 days in three of the patients.
Another coronary sinus device (Cardiac Dimensions, Kirkland, WA, see Figure 5a) has been implanted in four patients with non-ischaemic and ischaemic cardiomyopathy.25 With temporary deployment, the septal-lateral annular distance was decreased from 35.5±4.7 to 32.2±4.6 mm without significant impact on MR (unable to optimally place device due to sizing issues or overlapping circumflex artery in three of the four patients). A phase I trial of the coronary sinus devices is under way in the US (Cardiac Dimensions), Canada (Edwards Lifesciences and Viacor, Wilmington, MA) and Europe (Edwards Lifesciences and Cardiac Dimensions, Kirkland, WA). With improved pre-procedure sizing and patient selection, it is likely that patients with a coronary sinus at the same level or slightly below the mitral annulus will be treated with these devices successfully.
Non-coronary sinus devices have been tested in animals and human trials are beginning. One of these percutaneous techniques (Myocor, Maple Grove, MN) is modelled after a surgical device26-28 that consists of a posterior left ventricular pad, an anterior left ventricular pad, and an intra-ventricular tether that pulls the posterior annulus and myocardial wall towards the anterior annulus (see Figure 5b). Previous surgical success (no recurrence of ischaemic MR at one year in the first 11 patients of phase I study)26 makes the percutaneous version, placed intra-pericardially, an attractive strategy. Other percutaneous annuloplasty devices include the Mitralign System (Salem, NH) and Ample Medical (Foster City, CA). Both devices require localisation of a transventricular catheter (Mitralign) or a transatrial catheter (Ample) to the posterior mitral annulus. The Mitralign system subsequently places three implants into the posterior annulus and exerts tension via suture lines. The Ample technology delivers an implant into the posterior annulus and tethers the mitral apparatus to an atrial septal closure device. In cases where mitral repair may not be feasible, minimally invasive or percutaneous mitral valve replacement may be possible in the future. Boudjemline and colleagues have shown the feasibility of implanting a balloon-expandable stent valve, off-pump, into a surgical mitral bioprosthesis of an animal via the left atrium.29 This group has also reported the percutaneous replacement of an atrio-ventricular valve (tricuspid) using a self-expanding stent-valve in animals.30