Surgical mitral valve (MV) repair/replacement is currently the gold standard for treatment of symptomatic patients with severe (3+ to 4+) mitral regurgitation (MR) (1- 3). Surgical correction of MR results in improvement in symptoms and prolongation of survival in appropriately selected patients. However, there is a significant number of patients with symptomatic MR and extensive comorbidities that put them at high risk of surgical morbidity and mortality rates. If the surgical risk is judged to be prohibitive, then the patient may be relegated to medical management only. Some reports suggest that as many as one-half of patients with symptoms and severe MR may not receive surgery (4). Patients are currently denied surgical treatment for multiple reasons, including previous bypass surgery, porcelain aorta, post-radiation mediastinum, liver disease, renal disease, significant left ventricular (LV) dysfunction, advanced age, and other comorbidities (5).

The EVEREST (Endovascular Valve Edge-to-Edge Repair Study) II clinical study was designed to assess the safety and effectiveness of the MitraClip device (Abbott Vascular, Santa Clara, California) in treating patients with significant MR. The High Risk Study (HRS), an arm of the EVEREST II trial, enrolled symptomatic patients with 3+ to 4+ MR for whom surgical risk for perioperative mortality rate was estimated to be ≥12%, using either the Society of Thoracic Surgeons (STS) calculator (6- 7) or surgeon co-investigator estimated mortality risk of at least 12% based on prespecified criteria (described in the Methods section). After assessment with transthoracic echocardiography (TTE) and transesophageal echocardiography (TEE) to establish protocol-based eligibility for the MitraClip procedure (8- 11), patients were entered into a prospective, multicenter, single-arm study to determine the safety and effectiveness of treatment with the MitraClip device. A group of patients who were screened for the HRS but were not enrolled for various reasons were consented retrospectively to determine 30-day and 1-year survival in patients with similar degrees of MR, risks, and comorbidities treated by standard of care.

For inclusion, patients had to be symptomatic with grade 3+ to 4+ MR and a predicted surgical mortality risk of ≥12%, based on either the STS risk calculator or surgeon co-investigator estimated mortality risk following prespecified protocol criteria. Potentially qualifying criteria included high-risk patients with porcelain aorta, mobile ascending aorta atheroma, post-mediastinal radiation, functional MR with left ventricular ejection fraction (LVEF) <40%, age older than 75 years with LVEF <40%, previous median sternotomy with patent bypass graft(s), >2 previous chest surgeries, hepatic cirrhosis, or ≥3 of the following STS high-risk criteria: creatinine level >2.5 mg/dl, previous chest surgery, age older than 75 years, or LVEF <35%.

We used TTE and TEE screening echocardiograms to determine patient eligibility and assessed MR severity and jet origin by TTE. The primary regurgitant jet had to originate from leaflet malcoaptation at the A2/P2 region (12). TEE was used to assess MV leaflet anatomy and corroborate MR jet origin.

We excluded patients if they had evidence of an acute myocardial infarction within 2 weeks; if they had an LVEF <20% and/or a LV end-systolic dimension >60 mm; an MV area <4.0 cm²; leaflet anatomy that might preclude successful device implantation; a history of MV leaflet surgery; echocardiographic evidence of an intracardiac mass, thrombus, or vegetation; or active endocarditis. If the patient met all inclusion and no exclusion criteria, the MitraClip procedure was explained to the patient as well as the options for continued medical management and high-risk MV surgery. Patients signed informed consent and were enrolled in the study. The U.S. Food and Drug Administration approved the study protocol, and the informed consent document was reviewed and approved by the institutional review boards at the participating institutions.

After enrollment in the trial, the patient's MR was graded by a central independent echocardiographic core laboratory (ECL) (University of California, San Francisco) at baseline and all follow-up visits using the American Society of Echocardiography criteria (13- 14). Major adverse events (MAEs) through 12 months were adjudicated by a central events committee (Harvard Clinical Research Institute). Cause of death in the HRS was adjudicated by the central events committee as cardiac or noncardiac related. Cause of death in the concurrent comparator group was site reported as cardiac or noncardiac.

We retrospectively (after results of the HRS were known) identified 58 patients with MR severity of ≥3+ and a predicted surgical mortality rate of ≥12% screened for enrollment in the HRS who did not enroll or were not anatomically eligible for MitraClip device placement. Twenty-two patients were not included in the concurrent comparator group due to lack of site institutional review board approval to include patients in the comparator group (n = 11), lack of patient informed consent (n = 5), or inability to contact the patient (n = 6). Patients were requested to participate in the study by consenting to 1-year follow-up to determine survival. Baseline transthoracic echocardiograms were assessed for MR severity by the site echocardiologist. Baseline LVEF, LV end-systolic dimension, and anatomic assessment for the ability to place a MitraClip device was performed by a central team of echocardiographers who were trained by the ECL and under contract with the sponsor. Twenty-two percent (n = 8) of the comparator group met all HRS eligibility criteria but were not enrolled in the HRS because study enrollment had ended or they elected to not enroll. Nineteen percent (n = 7) of the comparator group were determined to be eligible based on transthoracic echocardiographic assessment of MR severity; however, anatomic eligibility based on a transthoracic echocardiogram was never confirmed. The other 58% of patients (n = 21) included in the comparator group met all eligibility criteria except for 1 or more specific anatomic protocol criteria related to MitraClip device placement (10). Anatomic exclusions included MV area <4.0 cm² (n = 4), jet origin other than A2-P2 (n = 4), flail width of ≥15 mm, or flail gap of ≥10 mm (n = 2), coaptation length of ≥2 mm (n = 3), leaflet calcification (n = 7), or a severely retracted posterior leaflet (n = 1). All comparator group patients were treated according to standard of care over the 12-month period, with 86% managed medically and 14% undergoing MV surgery. The date of screening for these patients was considered time zero to determine the natural history of MR treated by standard of care and mortality rate associated with significant MR. Survival at 30 days and at 1 year in the concurrent comparator group was compared with that of the HRS patients treated with the MitraClip device.

The MitraClip procedure was performed with the patient under general anesthesia using TEE and fluoroscopic guidance in the cardiac catheterization laboratory, as previously described (8- 11). After transseptal puncture, heparin was administered. The transseptal sheath was exchanged for the steerable guide catheter and dilator. The MitraClip Delivery System was introduced into the steerable guide catheter, and the MitraClip device was advanced into the left atrium. The MitraClip device was then steered and positioned until axially aligned and centered over the origin of the regurgitant jet. The MitraClip device arms were opened, and it advanced into the left ventricle below the mitral leaflets. The MitraClip device was then retracted until both leaflets were grasped and closed to coapt the mitral leaflets across the regurgitant orifice. Leaflet insertion into the MitraClip device and MR reduction were assessed using 2-dimensional and Doppler echocardiography. If necessary, the device was reopened, the leaflets released, and the MitraClip device repositioned. If the MitraClip device needed to be withdrawn into the left atrium, the arms were inverted in the ventricle, providing a profile for easy retraction into the atrium. After adequate reduction of MR was achieved after a hemodynamic challenge, the MitraClip device was deployed, and the catheters were withdrawn. A second MitraClip device was placed at operator discretion to obtain additional MR reduction. Patients were treated with aspirin 325 mg/day for 6 months and clopidogrel 75 mg for 30 days (10- 11).

TTE was performed at baseline, before hospital discharge, at 30 days, 6 months, and 12 months after MitraClip device placement according to protocol, and the transthoracic echocardiograms were sent to the independent ECL for analysis.

Data were analyzed by intention to treat unless otherwise specified. Serial matched data are shown for surviving patients only. Continuous data are expressed as mean ± SD and are compared with either a paired or independent Student t test. Categorical data are expressed as a percentage and compared using the Fisher exact test, and ordinal data are expressed as a percentage and compared using the Bowker test. For estimation of surgical risk, either the number calculated from the STS calculator was used if ≥12% or the surgeon's estimated mortality rate was used if the calculated STS score was <12%. Data using the STS calculator score only are also shown. A Clopper-Pearson exact binomial method was used to determine whether the observed 30-day mortality rate was lower than the 1-sided 95.472% upper confidence limit of the estimated 30-day mortality rate. The Kaplan-Meier method was used for survival analysis, and the log-rank test was used to compare the 2 groups. The rate of hospitalization for congestive heart failure (CHF) (12-month pre-enrollment and post-discharge) was estimated and compared using a Poisson regression model. A p value <0.05 was considered significant.

We enrolled 78 patients who met all inclusion criteria in the HRS. Thirty-six patients were included in the concurrent comparator group.

Baseline demographics are listed in (Table 1). The majority of HRS patients were male and older than 75 years. All patients had a history of CHF, and the majority had a history of coronary artery disease. Additionally, more than half the patients had previous cardiac surgery. Fifty-nine percent of the HRS patients had functional MR, and 41% had degenerative MR. Eighty-nine percent of patients were in New York Heart Association (NYHA) functional class III or IV. The predicted perioperative mortality rate based on the protocol definition was 18.2% for the HRS group. Of the 78 patients, 30 patients had an STS calculated score of <12% (mean calculated score: 7.1%) but a surgeon-estimated mortality rate of at least 12% based on the presence of prespecified baseline comorbidities. Using only calculated STS scores for all 78 patients resulted in a mean STS score of 14.2%. The baseline demographics of the concurrent comparator group were similar, with a predicted surgical mortality rate of 17.4%, based on either an STS score of ≥12%, or a surgeon-estimated mortality rate if the STS score <12%. The STS calculator score without surgeon estimates was 14.9% for the comparator group.

In the HRS patients, 1 or 2 MitraClip MV repair devices were successfully placed in 75 patients (96%). MitraClip device placement was not successful in 3 patients (4%). In 1 of these 3 patients, MR could not be successfully reduced, and in 2 patients, the procedure was aborted due to a transseptal complication and observation of an intracardiac thrombus after induction of general anesthesia before initiating the MitraClip procedure (exclusion). Sixty-two patients (79.5%) were implanted with a device and achieved at least a 1-grade reduction in MR as determined by the ECL, and 56 patients (71.8%) had successful device placement and reduction in MR grade to ≤2+ (Figure 1).

Six patients died within 30 days of the MitraClip procedure (central events committee adjudicated deaths as probably related to the procedure), none of which occurred intraprocedurally, although 3 occurred before hospital discharge. Events leading to death included: 1) myocardial infarction secondary to diffuse triple-vessel coronary artery disease (died on day 1); 2) intraprocedural hemodynamic instability with subsequent right heart failure (died on day 4); 3) transseptal complication that resulted in cardiac tamponade and ultimately acute renal failure (died on day 9); 4) intraprocedural retroperitoneal hematoma requiring vascular surgery resulting in acute renal insufficiency, hypotension, and a stroke (died on day 11); 5) complications associated with prolonged anesthesia and gastrointestinal bleeding (died on day 13); 6) CHF (no device implanted; died on day 19). Overall 30-day mortality rate in the HRS was 7.7%, significantly less than that predicted for open-heart MV surgery (18.2% per protocol, p = 0.006; 14.2% per the STS calculator, p = 0.06) in this patient cohort. In the concurrent comparator group, 30-day mortality rate was 8.3% (3 of 36), which was not different from the mortality rate in the HRS group (p = NS). The cause of death in all 3 medically managed comparator patients was cardiac.

Additional 30-day MAEs in surviving HRS patients included a non–ST-segment elevation myocardial infarction (asymptomatic post-procedure with a 3-fold increase in creatine kinase, myocardial band) in a patient with a history of myocardial infarction; a cerebrovascular accident with persistent neurologic deficits 3 days post-procedure without evidence of intracranial abnormalities on a computed tomography scan; renal failure that responded to conservative therapy; prolonged ventilation in a patient secondary to a laryngeal tear from which the patient fully recovered. The most common MAE was transfusion of ≥2 U of blood, which occurred 22 times in 14 patients (17.9% of patients) within 30 days of the procedure. The reasons for the 22 transfusion events were access site–related bleeding (n = 9), gastrointestinal bleeding (n = 6), bleeding in an unspecified location (n = 3), and chronic anemia (n = 4). Nonhierarchical MAEs through 30 days and 12 months are listed in (Table 2).

MitraClip device embolization did not occur during the study. One patient experienced MitraClip device attachment to a single leaflet during the procedure and underwent a successful second MitraClip procedure 6 weeks later for placement of a second device.

The major effectiveness endpoints for the HRS were freedom from death at 12 months, freedom from death and MR >2+ at 12 months, and clinical measures of benefit at 12 months in surviving patients, defined as NYHA functional class, LV measurements, SF-36 Health Survey quality of life, and rehospitalizations for CHF.

Effectiveness of the MitraClip procedure, defined as MR reduction to grade ≤2+, and clinical measures of benefit at 12 months are presented in (Table 3). Data for parameters with 30-day and 12-month follow-up, when available, are also presented in (Table 3). Echocardiographic MR grade improved compared with baseline to ≤2+ in 73% of surviving patients at 30 days, and 78% of surviving patients at 12 months; of these, 33% had MR grade ≤1+ at 12 months (Figure 2A). NYHA functional class improved in the vast majority of patients (Figure 2B), as did quality of life. LV end-systolic and end-diastolic volumes were reduced at 30 days, with further significant reductions observed at 12 months. Systolic and diastolic septal-lateral mitral annular dimensions were significantly reduced from baseline to 12 months.

The number of patients with CHF hospitalizations decreased significantly from 42% (33 of 78) in the 12 months before the MitraClip procedure to 16% (12 of 75) (p < 0.02) in the 12 months after discharge after the MitraClip procedure, a 45% reduction. If readmission for CHF is imputed for patients with cardiac death, then rehospitalization for CHF is no longer significantly decreased after device treatment (p = 0.09). Using only matched data in surviving patients, CHF admissions were reduced significantly (from 35 events in the year before treatment to 18 events in the year after device treatment in 59 patients, p = 0.034).

Freedom from death is illustrated by the Kaplan-Meier survival curve comparing patients in the HRS group with the concurrent comparator group (Figure 3). Importantly, no surgery was performed in the HRS group through 1 year. Thirteen HRS patients died between 30 days and 12 months; 6 were adjudicated as cardiac deaths (Table 4). Survival at 1 year was significantly higher in the HRS group compared with the comparator group after the MitraClip procedure (76.4% vs. 55.3%, p = 0.047).

Forty-six of the 78 HRS patients had normal MV leaflet anatomy but had malcoaptation of the leaflets secondary to leaflet restriction and LV dilation (FMR). The remaining 32 patients had leaflet pathology consistent with degenerative disease (DMR). By 12 months, 12 of 46 (26%) FMR patients and 7 of 32 (22%) DMR patients died. At 1 year, 34 of the FMR patients and 20 of the DMR patients had matched baseline and 12-month follow-up data available for comparison. At 1 year, 79% of the FMR and 75% of the DMR patients had sustained MR reduction with ≤2+ MR by ECL determination. Twenty-five of 34 FMR patients (74%) and 15 of 20 DMR patients (75%) had improvement to NYHA functional class I/II at 1 year. Only 1 patient was in NYHA functional class IV at 1 year compared with 16 at baseline in the entire cohort. Eighty percent of FMR patients and 89% of DMR patients improved by ≥1 NYHA functional class at 1 year. Improvement in end-diastolic volume and end-systolic volume were significant in the FMR patients (Figure 4). Diastolic and systolic septal-lateral annular dimensions also decreased significantly at 1 year in the FMR group with a trend for a decrease in the DMR group.

Clinical experience with the MitraClip procedure began in 2004 (15). Encouraging initial results were reported by Feldman et al. (8,10) in 2005 and 2009. Results of the randomized, controlled trial comparing the MitraClip device with MV surgery were reported by Feldman et al. in 2011 (11), with 78% of randomized patients remaining free of MV surgery at 2 years. However, these cohorts were limited to patients deemed acceptable surgical candidates with relatively preserved LV function.

The present study specifically selected symptomatic patients with moderate to severe or severe MR at high risk of MV surgery. Results demonstrate that these patients who were not considered to be suitable candidates for surgery can be successfully treated with the MitraClip System to reduce the degree of MR with procedural mortality rate less than that predicted for surgical treatment and 30-day mortality rate not different from that of a concurrent comparator group receiving standard care (Figure 3).

Mortality at 30 days in the HRS was lower than the predicted surgical mortality rate for this group, although there were 6 procedure-related deaths (7.7%). Although there is no high-risk, medically treated historical control group that matches the HRS cohort available for comparison, survival was similar to that of a concurrent comparator group with significant MR and comparable baseline comorbidities. These data and an acceptable rate of MAEs suggest that placement of the MitraClip device is relatively safe and feasible in these high-risk patients. Additionally, 1-year mortality rate in the HRS was reduced compared with the concurrent comparator group. The authors believe that the potential survival benefit observed in this study merits further investigation and may be clinically important (16- 17).

At 1-year follow-up of surviving patients with matched data, the MitraClip procedure resulted in an improvement in the severity of MR, a decrease in NYHA functional class, a decrease in LV end-systolic and end-diastolic volumes, and a decrease in mitral annular dimensions in both FMR and DMR patients. The number of admissions for CHF was also significantly reduced compared with the year before MitraClip therapy in surviving patients with matched data. However, the impact of these comparisons is limited because no matched data are available for patients who did not survive 12 months.

The observed reductions in LV diastolic and systolic volumes at 30 days and 12 months for high surgical risk patients in this study as well as the reduction in mitral annular dimensions suggest that MitraClip therapy results in early reverse of LV remodeling. Furthermore, paired LV echocardiographic parameters at 12 months in FMR and DMR patient subsets demonstrate that reverse LV remodeling occurs in both patient subsets with significant MR regardless of underlying pathology. These objective improvements in LV function, in addition to the subjective improvement in symptoms and quality of life reported by the patients, suggest that percutaneous reduction of MR with the MitraClip device results in significant clinical benefit.

The stabilization of mitral annular dimensions observed after leaflet repair with the MitraClip device is encouraging because concerns have been raised that MitraClip leaflet repair without concomitant annuloplasty might result in progressive annular dilation. Longer follow-up is needed to confirm whether these initially favorable findings persist.

Several limitations of the present study should be considered. First, the comparator group was recruited retrospectively, the patient number is limited, transesophageal echocardiograms were not available for review in all patients, and several of the patients included did not have appropriate anatomic criteria for MitraClip placement. Transthoracic echocardiograms determining patient eligibility for the comparator group were read by the site and contract echocardiographers; however, they were not reviewed by the ECL. These differences could potentially cause confounding of important variables when comparing the HRS patients with the comparator group. Despite these limitations, the comparator group did all have symptomatic 3 to 4+ MR by site and contract echocardiographic assessment and had comorbidities and STS scores similar to those of the HRS group. The authors believe that following this group for vital status adds important insight into the limited survival expected in such ill patients treated by the current standard of care. However, direct comparison between HRS and concurrent comparator groups should be done with caution.

A second concern is that 12-month echocardiographic and functional data were obtained and reported for surviving patients only, and no imputation for deceased patients' data was performed. Thus, the matched data reported may represent an overestimation of the true benefit provided by MitraClip placement.

The data from this HRS suggest a role for the MitraClip device in treating symptomatic patients with 3 to 4+ MR who are at high risk of mortality with MV surgery. MitraClip device placement in this selected high-risk group is feasible, effective in reducing symptoms and improving clinical status, and relatively safe in patients who otherwise have no safe option to reduce MR. Favorable LV remodeling demonstrates that the degree of reduction in MR obtained with the MitraClip device is hemodynamically important. Long-term follow-up is ongoing and needed to confirm whether the benefits observed at 12 months are sustained.