Mid-term results of mitral valve repair using flexible bands versus complete rings in patients with degenerative mitral valve disease: a prospective, randomized study

Background We aimed to compare the outcomes of mitral valve repair with flexible band (FB) versus complete semirigid ring (SR) in degenerative mitral valve disease patients. Methods From September 2011 to 2014, 171 patients were randomized and underwent successful mitral valve repair using a SR (n = 85) or FB (n = 86). There were no significant between-group differences at baseline. Results There were no early mortalities. The mean follow up was 24.7 months. The 2-year survival was 96.0 ± 2.3% (95% confidence interval [CI], 88.6–98.7%) and 94.3 ± 2.8% (95% CI, 85.5–97.9%) in the SR and FB groups, respectively (p = 0.899). The left ventricle remodeling was similar between the groups. Higher transmitral peak (8.5 [3.9–17] vs. 6 [2.1–18] mmHg, p < 0.001), mean pressure gradients (3.7 [1.3–8] vs. 2.8 [0.6–6.8] mmHg, p = 0.001), and systolic pulmonary artery pressure (34.5 [20–68] vs. 29.5 [8–48] mmHg, p < 0.001) was observed in the SR group. The 2-year freedom from recurrence of significant mitral regurgitation was significantly higher in the FB group than the SR group (p = 0.002). Residual mitral regurgitation was an independent prognostic factor of recurrence of mitral regurgitation. The 3-year freedom from reoperation was significantly higher in the FB group than the SR group (p = 0.044). Conclusion Patients with degenerative mitral valve disease may benefit from valve repair with FBs. Residual mitral regurgitation before discharge is an independent risk factor of late insufficiency recurrence. Trial registration ClinicalTrials.gov NCT03278574, retrospectively registered on 06.09.2017. Electronic supplementary material The online version of this article (10.1186/s13019-017-0679-0) contains supplementary material, which is available to authorized users.


Background
In 1957, Lillehei et al. [1] proposed the mitral annuloplasty technique, a new concept in valve surgery. Remodeling annuloplasty, developed by Carpentier in 1983, increases leaflet coaptation, prevents future annular dilatation, and preserves leaflet mobility in patients with degenerative mitral valve (MV) disease [2,3]. Currently, there are many commercially available mitral annuloplasty devices in the market, including complete or partial, rigid or flexible, and flat or saddle-shaped rings. However, no single annuloplasty device has been proven to have a clinical benefit above the others [4]. Theoretically, the flexible band does remodel mitral annulus providing reduction annuloplasty only. However, the rigid ring does not have enough flexibility for physiological annular motion during the cardiac cycle. A systematic review of clinical trials [5] showed comparable clinical outcomes between rigid and flexible rings. Currently, mitral ring selection is based on a surgeon's preference rather than evidence [6]. Semirigid rings combine flexibility and stability; however, their clinical benefit has not been completely clarified. The present study aimed to compare the outcomes of MV repair with a flexible posterior annuloplasty band versus complete semirigid ring in patients with degenerative MV disease.

Study design
In this prospective, randomized study, 171 patients with degenerative MV disease who were scheduled for isolated MV repair in our Institute from September 2011 through September 2014 were enrolled (CON-SORT flow diagram, Figure 1). Participants were randomly assigned following a simple randomization procedure to complete semirigid ring (SR group) or flexible posterior annuloplasty band (FB group), according to computerized random numbers on the day before surgery. Eligible participants were adults aged 18 years or more with degenerative MV disease [7] who met the indications for MV operation according to the American College of Cardiology/American Heart Association guidelines [8]. Exclusion criteria were previous open cardiac surgery, indication for concomitant aortic valve replacement, or left ventricle (LV) impairment (ejection fraction <40%). The Local Ethics Committee approved the study design, and all the patients provided informed consent. The present study was conducted in compliance with the Declaration of Helsinki.

Patients
Mean age of the participants in each group was 57  and 54  years, respectively. There were no differences in sex, age, preoperative New York Heart Association (NYHA) functional class, and comorbidities (Table 1).

Outcome measures
The primary endpoint was freedom from moderate or severe mitral regurgitation (MR) recurrence. Secondary endpoints included survival, freedom from reoperations, and freedom from severe MR recurrence. The severity of MR was evaluated and defined in accordance with the recommendations [9]. Valve-related complications were evaluated and defined in accordance with the guidelines [10].

Surgical techniques
Real-time two-dimensional/three-dimensional transesophageal echocardiography (TEE, Phillips iE33, Philips Ultrasound Inc., PA, USA) was performed after the induction of anesthesia for MV lesion estimation. Cold crystalloid The two most experienced surgeons in MV repair in our Institute performed the operations in the study. Surgeons were evenly split between the treatment groups. The surgical technique for MV repair was chosen according to the MV lesion ( Table 2). The most common techniques were quadrangular ("sliding maneuver") or triangular resection for posterior MV prolapse, and artificial chordal replacement (separate or loop technique) for anterior leaflet pathology. The minimally invasive approach through the right fourth intercostal space with femoral-femoral cannulation for arterial and venous lines was used in 29.4% (n = 25) and 31.4% (n = 27) of patients in the SR and FB groups, respectively.
All mitral valve repairs were performed with flexible bands ("C Flex", CardiaMed, Penza, Russia) or complete rings ("D Ring", CardiaMed, Penza, Russia), which routinely use since 2005. The ring size was selected according to the size of the anterior leaflet and the intertrigonal distance. In Barlow's cases, 36-40 mm rings were used. Flexible bands were implanted along the posterior mitral semicircle with an extension to both commissures for 3-5 mm. Valve competence was assessed intraoperatively by saline test and "ink-test" (symmetrical closure line, coaptation depth ≥ 5 mm, absence of leaflet prolapse and regurgitation jet).
After bypass weaning, patients were followed-up with TEE. In cases with residual MR ≥ grade 2, cardiopulmonary bypass was re-established, and the valve was rerepaired or replaced; residual MR grade 1 (mild) was left alone. Intraoperative data and concomitant procedures are shown in Table 2.

Patient follow-up
All patients underwent TTE evaluation before discharge. In total, 170 patients were discharged and followed-up periodically by cardiologists and surgeons. After discharge, examinations were scheduled annually. When annual clinic visits were unavailable, follow-up was performed by contact with the referring cardiologist, the patient or their family. Anticoagulation therapy with an international normalized ratio (INR) target in the range of 2.5-3.0 was prescribed for all patients after surgery. The decision to stop anticoagulation therapy with Coumadin was based on echocardiography (normal left ventricle function, presence of atrial contractility) and Holter data (absence of atrial fibrillation, flutter, tachycardia) after 3 months. Echocardiograms obtained from outside physicians were re-analyzed at our Institute by the most experienced echocardiographers. Clinical follow-up was completed in 165 patients (97.1%), and five patients were lost to follow-up. Follow-up echocardiography was performed 6-12 months after the operation and every year thereafter. Seventy-nine (94.0%) of the 84 discharged patients in the SR group and 78 (90.7%) of the 86 patients in the FB group underwent follow-up TTE. The mean duration from the operation to echocardiographic follow-up was 20.2 (95% confidence interval [CI], 18.6-21.8) months. The mean clinical follow-up period was 24.7 (95% CI, 23.5-27.0) months.

Statistical analysis
Assuming 5% significance (two sided), 80% power, event rates of 20 and 5% in the complete ring and flexible band groups, respectively, and a hazard ratio for

Availability of data and materials
The data that support the findings of this study are available from Meshalkin National Medical Research Center but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publically available. Data are however available from the authors upon reasonable request and with permission of Meshalkin National Medical Research Center.

Results
Overall repair rate was 95%. The mean (median) cross clamp time did not differ between the SR and FB groups. The subgroup analysis of variances by the surgical approach revealed no differences in cross clamp time between the minimally invasive (25 and 27 patients, respectively, p = 0.171) and median sternotomy subgroups (60 and 59 patients, respectively, p = 0.230). The incidence of systolic anterior motion (SAM) syndrome was observed in 1 patient in the FB group. Conservative management was not effective and SAM was successfully treated using posterior MV leaflet folding technique. Two patients from SR group were rerepaired due to residual moderate MR revealed by intraoperative TEE.
There were no early (at 30/60/90 days) deaths. One 65 years old man with p3 prolapse and coronary artery disease who was underwent triangular posterior leaflet resection and concomitant CABG in the SR group died after 6 months of hospital stay because of severe multiple organ failure. Two patients (one from each group) with severe LV systolic dysfunction in early postoperative period required extracorporeal life support with complete recovering after 7 days. The mean intensive care unit stay was 2 days in both groups (p = 0.453). Ventilation and inotropic support time also did not differ between the groups (Table 4). There were no significant differences in terms of heart failure, prolonged ventilation or requirement of extracorporeal life support. Electrical cardioversion for atrial fibrillation paroxysm was required in 7 (8.2%) and 3 (3.5%) cases in the SR and FB groups, respectively (p = 0.186). Five patients (SR group, 4 vs. FB group, 1; p = 0.169) were reexplored for bleeding on the first postoperative day. Pacemaker implantation rates were 5.9% (due to sinus node dysfunction, 3 patients; complete AVconductance disturbances, 2 patients) and 4.7% (due to sinus node dysfunction, 3 patients; complete AVconductance disturbances, 1 patient) in the SR and FB groups, respectively (p = 0.719).
There were no significant between-group differences in major valve-related complications. There were no cases of thromboembolic events, leakage, or structural dysfunction. There were 4 cases of transient ischemic attack with reversible neurologic deficits in the SR group versus none in the FB group (p = 0.042).

Echocardiographic results
All echocardiographic parameters significantly changed from the preoperative to immediate postoperative period at discharge, except transmitral pressure gradients and LV end-systolic diameter (LVESD) ( Table 3). There were no significant between-group differences in changes in LV function and left and right atria remodeling; however, transmitral peak and mean pressure gradients were significantly lower in the FB group. The mitral orifice area did not differ between the groups. The TTE before discharge revealed that 5 (5.9%) and 4 (4.7%) patients in the SR and FB groups, respectively, had grade 2 residual MR (p = 0.719). Follow-up echocardiography data at 24 months are shown in Table 3. The repeated-measures analysis of variance in the FB group revealed significant changes in LV end-diastolic diameter (LVEDD), LV enddiastolic volume (LVEDV), and LV end-systolic volume (LVESV) at discharge; moreover, these parameters and LVESD were also decreased at 24 months of follow up. Meanwhile, the LVEF significantly decreased at discharge; however, it significantly increased during the follow-up. In the SR group, there were no significant differences in LV function changes at serial examinations. We also revealed higher transmitral pressure gradients in the SR group, including significant differences in systolic pulmonary artery pressure among patients with sinus rhythm, with favorable results in the FB group.

Clinical status
Before surgery, 65.9% and 52.3% of patients in the SR and FB groups, respectively, had an NYHA functional classification of III and IV. Most patients demonstrated significant improvement of functional capacity at the last follow-up (43.2% and 34.6% of 81 SR patients and 44.1% and 40.3% of 77 FB patients had an NYHA functional classification of I and II, respectively). A total of 57 (75%) and 61 (84.7%) patients presented with sinus rhythm in the SR and FB groups, respectively (p = 0.201).

Thromboembolism and anticoagulation-related hemorrhage
There was no major bleeding case during follow-up. Two patients form each group had transient ischemic attack. In the FB group, 2 out of 3 patients who had stroke died. Two patients had atrial fibrillation and were kept on anticoagulation therapy, and another one was in sinus rhythm at the last follow up and was not anticoagulated. Central retina occlusion occurred in 1 case in each group. The overall freedom from thromboembolic events at 3 years after MV repair in the SR and FB groups was 94.7 ± 3.6% (95% CI, 80.6-98.7%) and

Infective endocarditis
Three patients had infective endocarditis. Two of them were in the SR group, and one of them died because of pacemaker lead endocarditis and heart failure. The 2 patients who survived were treated with antibiotics alone. Freedom from infective endocarditis at 3 years was 97.3 ± 1.9% (95% CI, 89.4-99.3%) and 98.5 ± 1.5% (95% CI, 89.9-99.8%) in the SR and FB groups, respectively (log-rank, p = 0.573).

Reoperations
Reoperations were required in 3 patients in the SR group for partial ring dehiscence in all cases. The MV was rerepaired in 2 patients and replaced in 1 case. The Kaplan-Meier estimate for freedom from reoperation at 2 years of follow-up was 97.0 ± 2.1% (95% CI, 88.4-99.3%) and 100% in the SR and FB groups, respectively ( Figure 4) (log-rank, p = 0.044). The Cox regression hazard model did not identify any preoperative or intraoperative risk factor of late reoperations.

Discussion
In the late 1960's, Carpentier developed annuloplasty rings, considered the "gold standard" for the surgical treatment of MR [11]. Subsequent studies showed that the MV annulus continually changes size and shape during the cardiac cycle [12][13][14]. This led to the development of a flexible ring that could conform to the physiologic changing annular shape [15,16]. However, there is controversy regarding the optimal mechanical characteristics of annuloplasty rings and the use of rigid rings or flexible bands in degenerative MV disease [5,17].
David et al. [18] conducted one of the first randomized studies comparing the rigid ring (13 patients) and flexible ring (12 patients). The results showed significantly better LV systolic function and stroke volume/LVEDV index in the flexible group and no significant reduction in LVESV and LVESD in the rigid ring group. However, the clinical outcomes of these 2 ring types have not been reported. In the present study, there were similar results in LV changes with no significant differences between the SR and FB groups intermediately and 2 years postsurgery. However, only the FB group showed a significant reduction in LVEDD, LVESD, LVEDV, and LVESV (when comparing postoperative values and those 2 years postsurgery).
Chang et al. conducted the largest randomized trial (363 patients) comparing the Carpentier-Edwards rigid ring and flexible Duran annuloplasty ring [19]. The mean duration from the operation to follow-up echocardiography was 26.7 ± 24.1 months. LVEF, LVESD, and LVEDD parameters changed significantly at serial examinations in both groups with no significant difference between the two ring types. There were no significant differences in survival, reoperation rates, and recurrences of significant MR at a mean follow-up of 46.6 ± 32.6 months. However, their study was limited by the heterogeneous MR etiology, the long recruitment period, and the fact that the echocardiography parameters were not presented at separate time points, which may have caused bias.
Shahin et al. [20] conducted a randomized study comparing the Carpentier-Edwards rigid Classic and There were no between-groups differences in terms of mortality, morbidity, and LV function. However, an unexplained 16% difference in mortality was considered clinically important.
At least 3 randomized studies did not demonstrate the clinical benefits of flexible or rigid types of annuloplasty devices [18][19][20]. However, one of them [18] suggested better LV systolic function with flexible rings. Similar results were reported by Okada et al. [21], and concern that rigid rings could restrict LV wall motion was raised.
In a prospective echocardiographic study by Unger-Graeber et al. [22], no significant difference was shown between the Carpentier rigid ring and flexible Duran ring in terms of transmitral velocity and pressure gradient. In contrast, our study showed higher transmitral pressure gradients in the SR group.
After David et al. criticized the rigid ring for impaired LV function [18] and Kreindel et al. [23] reported the potential risk of SAM syndrome after rigid ring annuloplasty, Carpentier et al. introduced a new concept of mitral annuloplasty with the semirigid prosthetic ring, which combines remodeling and flexibility [24]. At mid-term follow-up between 6 and 18 months, 93.2% of 94 followed-up patients were free from MR recurrence with a transmitral pressure gradient at 3.55 ± 1.93 mmHg. Thereafter, excellent mid-term results with semirigid rings have been reported by other investigators [25,26]. Since then, the concept of a semirigid ring has gained popularity, and the semirigid ring (instead of the rigid ring) has been successfully adopted in mitral annuloplasty for degenerative MV disease. Flexible posterior annuloplasty bands have also shown high effectiveness in preserving mitral annulus flexibility and provide good mid-term and long-term durability [4,27,28].
The semirigid ring and flexible ring have not been sufficiently compared. An animal randomized study [29] showed that LV function was not altered with either flexible or semirigid ring annuloplasty. A Japanese, retrospective, propensity score matched study [30] evaluated intermediate echocardiography results only. The overall cohort's LVEF decreased during the first week after surgery and then recovered gradually at 6 months and 1 year; LVEDD abruptly decreased and LVESD minimally decreased at the first week postoperatively, then gradually decreased at 6 months and later stabilized. There were no significant between-group differences in LVEF, LVEDD, and LVESD. They suggested that the semirigid ring might prevent LV impairment compared with the rigid ring. However, their study was limited by the retrospective design, low rate of followed-up patients, and difference in annuloplasty devices used in the flexible group.
Previously was shown that experience in mitral valve repair is an important determinant of operative efficiency and late survival [31], in our study both surgeons are well experienced and evenly split between groups. However, our results seem different to Castillo JG report [32]. We assume that limited experience in Barlow valves disease was influenced for our results. In this view complex mitral valve cases should be consolidated and addressed to one surgeon. It might be helpful to use scoring system [33] to allow stratification of complexity for degenerative mitral valve repair for improving results and develop local expertise.  We examined the immediate and mid-term results of patients with degenerative MV disease who underwent primary MV repair with a complete semirigid ring or flexible posterior annuloplasty band. Both groups had comparable early clinical results; however, the SR group had better coaptation depth, while the FB group demonstrated significantly lower transmitral pressure gradients. At serial echocardiographic examinations, only the FB group showed significant LV remodeling. Consistent with previous studies, there were no between-group differences in overall survival, freedom from cardiacrelated death, and follow-up LV remodeling. Our study is the first to show the superiority of the flexible posterior band over the semirigid complete ring in terms of freedom from recurrences of significant and/or severe MR and risk of MV reoperation.

Conclusion
The present study is limited by the 2-year follow-up period and single-center design. Important limitation is low number of patients at late risk. Further study with a longer follow-up is warranted. In conclusion, patients with degenerative MV disease may benefit from valve repair with flexible bands. Residual MR is an independent risk factor of late insufficiency recurrence.