Original Contribution

Mitral Annuloplasty Device Implantation for Non-Surgical Treatment of Mitral Regurgitation: Clinical Experience After the Approval Studies

Norbert Klein, MD1;  Dietrich Pfeiffer, MD, Prof1;  Steven Goldberg, MD2;  Maika Klein, MD1 

Norbert Klein, MD1;  Dietrich Pfeiffer, MD, Prof1;  Steven Goldberg, MD2;  Maika Klein, MD1 

Abstract: Background and Aims. One option for interventional treatment of secondary mitral regurgitation is the placement of an indirect mitral annuloplasty device (Carillon Mitral Contour System) cinching the mitral annulus to facilitate coaptation of the mitral leaflets. The aim of this study is to evaluate the implantation success and hemodynamic results. Methods and Results. Seventeen consecutive patients (11 females, 6 males) suffering from moderate to severe secondary mitral regurgitation (40% ischemic, 59% non-ischemic) received a Carillon Mitral Contour System device, which was evaluated prospectively in a single-center setting. The mean procedure time was 95.5 ± 2.1 minutes, fluoroscopy time was 13.1 ± 6.1 minutes, and the amount of contrast dye was 120.6 ± 39.3 mL. The mitral regurgitation grade was reduced from 2.8 ± 2.4 to 1.9 ± 0.8 acutely (P<.01), with an additional improvement seen after 3 months to 1.5 ± 0.75 (P<.01 to baseline). In 41.2% of patients, acute reduction of coronary artery flow was seen and managed with a stepwise approach: (1) intracoronary nitroglycerin was administered; (2) release of device tension; (3) a new device was changed and placed at a different location; and (4) the device was left in place and a stent was implanted in the coronary artery. Only 1 patient (5.8%) could not receive the device due to coronary compression. No periprocedural myocardial infarction occurred. Conclusion. Indirect mitral annuloplasty is feasible outside of controlled trials, and is associated with successful improvement of mitral regurgitation. Although coronary compromise needs to be addressed as part of the implantation procedure, this issue limits implantation of the device in only a small percentage of cases.   

J INVASIVE CARDIOL 2016;28(3):115-120. Epub 2016 January 15.

Key words: mitral regurgitation, interventional treatment of mitral regurgitation, mitral annuloplasty


Secondary mitral regurgitation (SMR) is a common problem in patients with heart failure. There are no established treatments for SMR, although some newer therapies show promise.1 The CE-mark certified percutaneous MitraClip device (Abbott Vascular) has been the most common percutaneous device used to treat SMR and is currently being evaluated for this purpose with randomized trials.2 Cardiac resynchronization therapy (CRT) is indicated for patients with reduced left ventricular (LV) function and wide QRS complex and can lead to improvement in LV function and reduce SMR, with induction of reverse remodeling.3 Reverse remodeling may lead to a reduction of LV dimensions.4 With effective CRT, the interpapillary angle may be reduced, leading to improved function of the papillary muscle and reduced mitral regurgitation.5   

The only other CE-marked percutaneous device to treat SMR is an indirect mitral annuloplasty device, the Carillon Mitral Contour System (Cardiac Dimensions, Inc). The AMADEUS6 and TITAN7 clinical trials demonstrated significant reduction in SMR with the Carillon device, leading to CE-mark approval. We were trained on the use of this device in 2012 and began implanting it in patients with SMR. This paper describes our initial experience with this percutaneous device.   


This is a prospective, single-center, non-randomized observational study of 17 consecutive patients (11 females; 65%) with moderate to severe SMR despite stable optimal medical drug treatment who received a Carillon Mitral Contour System. Patients were selected for percutaneous rather than surgical approach due to high surgical risk as assessed by EuroSCORE II8 and STS score.9 The MR grade, as well as quantitative assessments of MR (regurgitant volume and fraction), were evaluated by transesophageal echocardiography (TEE) using standard approaches.10 Only 1 of these 17 patients had a previously implanted coronary sinus (CS) lead for CRT; no other patient had an indication for CRT.  

Although the procedure can be done under conscious sedation using transthoracic echocardiographic guidance, it was our preference to use TEE guidance under general anesthesia. After application of general anesthesia, the right internal jugular vein was punctured and a 9 Fr sheath was placed and sutured into place. The CS was intubated with a 9 Fr proprietary delivery catheter (Cardiac Dimensions, Inc) over an AL2 diagnostic coronary catheter (Merit Medical) and a floppy coated .035˝ Radifocus Guidewire M (Terumo Medical Corporation). After removal of the AL2 catheter, a marker catheter was placed (Figure 1), which was used as a scaling device, allowing for measurements of the CS dimensions. The Carillon Mitral Contour System is composed of two self-expanding nitinol anchors, with a curvilinear nitinol connecting section. There are 37 different combinations of distal (7-14 mm) and proximal (16-20 mm) anchor sizes and device lengths (60, 70, and 80 mm) available. After selecting an appropriately sized device, based upon the venous dimensions at the site of each intended anchor and the length of the vein, the device was delivered via the 9 Fr delivery catheter. The distal anchor was delivered into a specified location, based upon distance from the CS ostium, venous characteristics (taper), and relationship to the circumflex coronary artery, since the circumflex artery also passes within the atrioventricular groove, along with the CS. The distal anchor was unsheathed in the desired location within the great cardiac vein, and then locked. After deployment of the distal anchor, slow pulling on the delivery catheter was performed to provide a plicating force on the posterior aspect of the mitral annulus (Figure 2). The intention was to pull approximately 4 cm of tension. Simultaneous imaging with TEE typically demonstrated cinching of the atrial site of the atrioventricular groove as tension was being applied. Coronary angiography was then performed to look for coronary artery compromise by the device. If there was no significant compression or alteration of the circumflex coronary artery and branches, the proximal anchor was released, keeping the device in place by maintaining tension on the catheter and plication on the mitral ring (Figure 3).  Right coronary arteriography was then performed, as in rare cases, the distal right coronary artery can be compromised by the proximal anchor.   

The acute impact on MR was evaluated by TEE. The patient then recovered and returned for follow-up in 3 months, at which time transthoracic echocardiography was performed and compared with the baseline images.

Statistical analysis. The study analyses were performed with SPSS (IBM, Inc). Testing was done for normal distributions and a Wilcoxon test was used to calculate statistical significances.   


Patient characteristics. The patients in this study had a high surgical risk based upon EuroSCORE II with a mean value of 15.4 ± 3.3% (logistical EuroSCORE, 58.5 ± 20.7%). The STS score was 6.75 ± 4.5%, although it should be recognized that STS score is limited in patients with SMR, as very few patients with SMR undergo surgery. Significant contributors to the high risk in this cohort included advanced age (mean age, 76.8 ± 10.4 years) and poor LV function (mean ejection fraction, 41 ± 14%). In patients with EuroSCORE ≤12 (4/17 patients; 23.5%), patient preference contributed to the decision to proceed with an interventional rather than surgical approach. Forty-one percent of patients had an ischemic etiology to their cardiomyopathy, with the rest having a non-ischemic dilated cardiomyopathy.   

The MR prior to intervention was moderately severe, with a mean MR grade of 2.8 ± 2.4. The baseline regurgitant volume was 35 ± 14 mL (Table 1), with a regurgitant fraction of 43 ± 6%.   

Procedural results. All 17 patients attempted had successful implantation and device deployment; in 1 patient, the device could not be detached due to coronary compromise leading to complete removal without any sequel. Mean procedure time was 95.5 ± 21 minutes, including anesthesia preparation time. Mean fluoroscopy time was 13.1 ± 6.1 minutes and contrast dye volume was 120.6 ± 39.3 mL. The range of values for implantation and fluoroscopy time depends upon characteristics of specific individual anatomy.  Some CSs are more challenging to negotiate with wires and catheters, and can take extra time to access adequately (Figure 4). Additionally, if a Carillon device needs to be recaptured and removed (most commonly due to coronary artery compression), this adds time to the procedure.   

Coronary artery management. Because the circumflex coronary artery also lies within the atrioventricular groove, it can have a close proximity to the CS. There is a risk of coronary artery compromise, which can be managed to avoid anything other than temporary compression. Coronary angiography was performed immediately after pulling tension on the device. Also, after deployment of the proximal anchor, but before decoupling, a coronary angiography was performed. This temporary compression was observed in 7/17 patients (41.2%) (Figure 5). This was addressed by a stepwise approach. First, if there was only a partial compression (not an occlusion) of the circumflex artery, intracoronary instillation of nitroglycerin was attempted to relax the spastic artery. This was successful in 1/17 cases (5.9%). If this was unsuccessful, reduction of the tension on the system was attempted (prior to placing the proximal anchor), with assessment of MR reduction after this lesser degree of tensioning. If this was acceptable, and the coronary compromise was minimal or absent, the proximal anchor was then deployed in this location. This was successful in 1/17 cases (5.9%). The next option was to recapture the device and change to a shorter device, to allow placement of the distal anchor more proximally, to avoid compressing the artery. This was successful in 2/17 cases (11.8%). Finally, if the device under tension led to a significant reduction in MR, but coronary compression could not be solved by previous maneuvers, a coronary stent was placed in the coronary artery to protect the artery against the external compression from the device. Stenting was successfully performed in 2/17 cases (11.8%). Only 1/17 patients (5.9%) could not be implanted due to coronary artery compression of a large marginal coronary artery branch, providing an implant success rate of 94.1%. Of interest, in 1/7 cases of coronary artery compression, the compression was in the right coronary artery, which was treated with a stent. This patient underwent a coronary angiography 5 months after initial procedure and no restenosis could be observed. The other patient with need for coronary stenting of the circumflex artery was followed up non-invasively (echocardiography and treadmill exercise test), which showed normal findings.  

Previous therapies. In 1 case, a MitraClip was previously implanted 1 year before, with no favorable impact on the MR. This patient also had a CRT defibrillator implanted 2 years previously. This has been considered a contraindication to Carillon for theoretical reasons, ie, concerns that the Carillon might damage the lead, and that it may entrap the lead (which could become a problem should a pacemaker system removal be necessary in the future). Treatment with additional MitraClips was not considered feasible due to calcifications on both leaflets. Because the large left atrium seemed to support the use of the Carillon device, it was decided to remove the CS lead and implant a Carillon device, both done at one session. After 4 weeks, a new CS lead was placed easily into the same posterolateral vein as previously, as planned.   

Impact on mitral regurgitation. An acute reduction of approximately one grade in mean MR was observed, from the initial preimplantation MR grade of 2.8 ± 2.4 to 1.9± 0.8 (P<.01) (Figure 6). Further improvement after 3 months was also observed, to a mean grade of 1.5 ± 0.75. The individual changes in MR are shown in Figure 7. Acute improvement of one grade in MR was observed in 10/16 patients (62.5%), with an improvement of two grades seen in 2/16 patients (12.5%) (Figure 8). In 4/16 patients (25%) no change in MR was observed at the time of intervention, although reductions in mitral annular diameter measurements were seen. The patient who did not receive a device was excluded from the analysis of the impact on MR.  

Quantitative measurement of MR volume was also analyzed. The mean MR volume was reduced from 35 ± 14 mL at baseline to 18 ± 10 mL at 3 months (P<.01), based upon serial TEE measurements. Similar reductions were seen with regurgitant fraction (43.6 ± 6% at baseline vs 12 ± 8% at 3 months; P<.01).  

In the patients in whom acute reduction in quantitative MR measurements were not seen, the 3-month follow-up echocardiogram did show improvement in 3/4 patients, with only 1 patient not showing any improvement. Thus, only 1 patient (1/16; 6.2%) failed to have improvement in MR by 3 months. In addition, patients who showed initial improvement showed further improvement at the 3-month follow-up exam (Figure 9). After 3 months, 9/16 patients (56.3%) had a one-grade improvement and 6/16 patients (37.5%) had a two-grade improvement. Therefore, in total, 2 patients failed to demonstrate a reduction in MR (1 patient who received the device and 1 patient who could not receive the device). Therefore, clinical success was achieved in 15/17 patients (88.2%).


Percutaneous treatment for functional MR is clinically valuable11 due to the widespread prevalence of this problem and the lack of alternative available therapies. An indirect mitral annuloplasty (or annulorrhaphy) has several theoretic advantages. By plicating the surrounding (atrial) tissue, rather than the valve itself, it may be possible to treat patients even with calcification of the leaflets. This also has the potential to address inadequate coaptation for the MitraClip device. As seen from this experience, the learning curve can be quickly surmounted, as noted by the short duration and fluoroscopy times. Because the procedure is not truly echocardiography dependent (in contrast to the MitraClip and other upcoming percutaneous options), this also adds to the simplicity of the procedure. Certainly echocardiography is very interesting and helpful,12 but the procedure is guided by fluoroscopy.   

One challenge of this technology is the potential to compromise coronary arteries — a mostly solvable phenomenon that we observed in 41.2% of our patients. In the AMADEUS trial, coronary artery crossing by the CS was observed in 84% of patients, although coronary artery compression led to the patient leaving the cath lab without receiving a device in only 14% of patients. In this “real-life” series of patients, we propose a stepwise approach: (1) intracoronary nitroglycerin; (2) release of tension; (3) change to a shorter device with the distal anchor proximal to the prior location; and (4) intracoronary stent implantation. With this strategy, we were able to see a reduced implantation failure rate of only 5.8% (vs 14% in TITAN). Of course, the placement of a stent in a coronary artery carries its own set of risks, but in these cases, the clear reduction in MR appeared to outweigh the risks of the coronary stent implantation. In the single case in which we elected not to leave a Carillon device, the transient compression of the obtuse marginal coronary artery was associated with temporary ST elevation and drop in arterial pressure. In this case, the balance appeared to support not putting the coronary artery at any risk at all, so the Carillon procedure was abandoned.  

Coronary impairment is also known as a rare complication in surgical mitral annuloplasty.13 This can occur due to a direct suture, or plication and compression of surrounding tissue. If recognized during surgery (of course typically without the benefit of immediate coronary angiography), it can be treated with an acute arterial bypass graft,14 or sent to the cath lab for percutaneous repair using balloon dilatation and stent implantation.15 Even the recognition can be challenging, since obstructing the circumflex coronary artery may be electrographically silent, and thus requires careful evaluation for new posterolateral myocardial wall-motion abnormality by the echocardiographer.   

Other conceptions of interventional mitral annuloplasty16 use a protection device, which covers the coronary branches from external compression by a tube. However, this is not currently available with this generation of Carillon device.

Implanting a CS lead for CRT is possible after a Carillon device is present;17 in fact, it facilitates placement by providing a clear path for entering the CS. As mentioned above, it has been theorized to be risky to place a Carillon device on top of a previously placed CS lead, for fear it might damage the lead and entrap it, complicating cases of pacemaker infection. Because the explantation of CS leads is not very complicated,18 requiring only simple traction for leads implanted for <4 years,19 explantation of the CS lead followed by Carillon implantation was performed in 1 of our cases. By design, the patient underwent a CS lead reimplantation 4 weeks later. This time frame was selected to allow for initial healing around the Carillon device, and to allow for any possible venous dissections that might have occurred during the Carillon implantation procedure (none of which were visible during the procedure). In patients in whom a CRT indication is present, it would make sense to consider if the Carillon implantation procedure could be the first step before implantation of a CS lead to avoid reoperation.

Although there may be overlap in whom a MitraClip or Carillon may be used, it is feasible that they can be synergistic. When the coaptation zone is too large to easily grab both leaflets simultaneously, annular reduction can bring the edges closer to each other. The placement of one device does not preclude the other, and the Carillon may facilitate the ease of MitraClip placement.   

An interesting observation is the improvement of MR over time, a phenomenon also seen in the TITAN trial. It may be that decreasing the mitral ring due to pleating the mitral annulus takes time, depending upon the rigidity of the tissue. Perhaps this explains the delayed success in 20% of our patients. In the initial studies of AMADEUS and TITAN, the device was removed if acute success was not demonstrated, but it has now been demonstrated the MR can improve over time, arguing for leaving a device in place even if acute reduction in MR is not seen. Supporting this strategy, in the 3-month follow-up in this study, with patients on stable concomitant medical therapy, we saw the non-responder rate decrease from 25% in AMADEUS to 6.3% in this study.    


1.    Nickenig G, Mohr FW, Kelm M, et al. Konsensus der Deutschen Gesellschaft für Kardiologie — Herz- und Kreislaufforschung und der Deutschen Gesellschaft für Thorax-, Herz- und Gefäßchirurgie zur Behandlung der Mitralklappeninsuffizienz. Kardiologe. 2013;7:76-90.

2.    Boekstegers P, Hausleiter J, Baldus S, et al. Interventionelle Behandlung der Mitralklappeninsuffizienz mit dem MitraClip-Verfahren. Kardiologe. 2013;7:91-104.

3.    Breithardt OA, Sinha AM, Schwammenthal E, et al. Acute effects of cardiac resynchronization therapy on functional mitral regurgitation in advanced systolic heart failure. J Am Coll Cardiol. 2003;41:765-770.

4.    Ghio S, Freemantle N, Scelsi L, et al. Long-term left ventricular reverse remodelling with cardiac resynchronization therapy: results from the CARE-HF trial. Eur J Heart Fail. 2009;11:480-488.

5.    Faber L, Lamp B. Mitral valve regurgitation and left ventricular systolic dysfunction: corrective surgery or cardiac resynchronisation therapy?; Herzschr Elektrophys. 2008;19(Suppl 1):52-59.

6.    Siminiak T, Hoppe UC, Schofer J, et al. Effectiveness and safety of percutaneous coronary sinus-based mitral valve repair in patients with dilated cardiomyopathy (from the AMADEUS trial). Am J Cardiol. 2009;104:565-570.

7.    Siminiak T, Wu JC, Haude M, et al. Treatment of functional mitral regurgitation by percutaneous annuloplasty: results of the TITAN trial. Eur J Heart Fail. 2012;14:931-938.

8.    Nashefa S, Roquesb F, Sharplesc LD, et al. EuroSCORE II. Eur J Cardiothorac Surg. 2012;41:734-744.

9.    O’Brien SM, Shahian DM, Filardo G, et al; for the Society of Thoracic Surgeons Quality Measurement Task Force. The Society of Thoracic Surgeons 2008 cardiac surgery risk models: part 2 — isolated valve surgery. Ann Thorac Surg. 2009;88(1 Suppl):S23-S42.

10.    Flachskampf FA, Badano L, Daniel WG, et al; European Association of Echocardiography; Echo Committee of the European Association of Cardiothoracic Anaesthesiologists. Recommendations for transoesophageal echocardiography: update 2010. Eur J Echocardiogr. 2010;11:557-576.

11.    Schofer J, Siminiak T, Haude M, et al. Percutaneous mitral annuloplasty for functional mitral regurgitation: results of the CARILLON Mitral Annuloplasty Device European Union Study. Circulation. 2009;120:326-333.

12.    Mahmoud HM, Al-Ghamdi MA, Ghabashi AE. Real time three-dimensional transesophageal echocardiography guided coronary sinus cannulation during CARILLON mitral annuloplasty device therapy for a patient with chronic severe mitral regurgitation. Echocardiography. 2015;32:181-183. 

13.    Grande AM, Fiore A, Massetti M, Viganò M. Iatrogenic circumflex coronary lesion in mitral valve surgery: case report and review of the literature. Tex Heart Inst J. 2008;35:179-183.

14.    Nakajima H, Ikari Y, Kigawa I, et al. Rapid diagnosis and management of intraoperative myocardial infarction during valvular surgery: using intraoperative transesophageal echocardiography followed by emergency coronary artery bypass grafting without coronary angiography. Echocardiography. 2005;22:834-838.

15.    Folkmann S, Mohr FW, Garbade J. Emergency PCI after occlusion of the circumflex artery due to minimally invasive mitral valve replacement in re-do operation. Acta Cardiol. 2014;69:451-453.

16.    Kim JH, Kocaturk O, Ozturk C, et al. Mitral cerclage annuloplasty, a novel transcatheter treatment for secondary mitral valve regurgitation: initial results in swine. J Am Coll Cardiol. 2009;54:638-651.

17.    Hoppe UC, Brandt MC, Degen H, et al. Percutaneous mitral annuloplasty device leaves free access to cardiac veins for resynchronization therapy. Catheter Cardiovasc Interv. 2009;74:506-511.

18.    Williams SE, Arujuna A, Whitaker J, et al. Percutaneous lead and system extraction in patients with cardiac resynchronization therapy (CRT) devices and coronary sinus leads. Pacing Clin Electrophysiol. 2011;34:1209-1216.

19.    Sheldon S, Friedman PA, Hayes DL, et al. Outcomes and predictors of difficulty with coronary sinus lead removal. J Interv Card Electrophysiol. 2012;35:93-100.


From the 1University Hospital Leipzig, Department of Cardiology and Angiology, Leipzig, Germany; and 2Rocky Mountain Heart and Lung Kalispell Regional Medical Center, Kalispell, Montana.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr N. Klein and Dr Pfeiffer report honoraria for lectures from Cardiac Dimensions and Abbott Vascular. Dr Goldberg is a paid consultant, including stock options, and is Chief Medical Officer for Cardiac Dimensions. Dr M. Klein reports no conflicts of interest regarding the content herein.

Manuscript submitted August 19, 2015, provisional acceptance given September 23, 2015, final version accepted October 12, 2015.

Address for correspondence: Norbert Klein, MD, University Hospital Leipzig, Department of Cardiology and Angiology, Liebigstrasse 20, 04103 Leipzig, Germany. Email: norbert.klein@medizin.uni-leipzig.de