Background
Mitral repair is recommended for patients with significant organic mitral regurgitation (MR). The nonresectional dynamic mitral valve repair (NVR) method involves a complete flexible ring and artificial chordal insertion but without leaflet resection or annular plication. The aim of this study was to compare changes in mitral annular structure and function after the NVR technique with those after a resectional mitral valve repair (RVR) method, which involves leaflet resection and annuloplasty with a partial flexible ring.
Methods
Patients with organic severe MR undergoing mitral valve repair with either technique underwent three-dimensional transesophageal echocardiography before and after surgery. The mitral annulus was tracked offline and measured throughout the cardiac cycle. Mitral leaflet mobility was also measured.
Results
Fifteen patients underwent repair with NVR, and 13 underwent repair with RVR (age, 56 vs 61 years, respectively). Both operations reduced mitral annular area significantly (maximum area reduction, from 18.5 ± 4.6 to 6.6 ± 1.7 cm 2 and from 20.1 ± 4.8 to 6 ± 1.5 cm 2 with the NVR and RVR techniques, respectively; P < .001). In contrast to RVR, patients who underwent NVR maintained dynamic changes in mitral annular area, circumference, and anterior-posterior diameter during the cardiac cycle. Mitral leaflet mobility was reduced with both techniques, but posterior leaflet mobility was restricted with RVR.
Conclusions
The size of the mitral annulus is reduced after repair with either surgical approach. Compared with resectional valve repair, more dynamic changes in the structure of the mitral annulus are maintained during the cardiac cycle with the NVR technique early postoperatively, along with more preserved motion of the posterior leaflet.
Organic mitral regurgitation (MR) is a major cause of morbidity and mortality and is amenable to surgical repair with the aim of restoring leaflet coaptation and eliminating regurgitation. Substantial progress in surgical techniques has led to the American College of Cardiology and American Heart Association guideline recommendation for early surgical intervention in asymptomatic patients with severe MR. Mitral valve repair with the Carpentier method involves leaflet resection and annular plication, chordal transfer or shortening, and placement of a complete rigid ring with interrupted sutures. Modifications of this technique have been instituted with limited leaflet resection and/or chordal replacement using polytetrafluoroethylene sutures with the use of a complete rigid ring. Recently, a new surgical approach has been introduced, nonresectional dynamic mitral valve repair (NVR), characterized by keeping the leaflets and annulus intact, using artificial chordae for subvalvular support, and reducing the annulus with a complete flexible ring using a continuous suture technique. The NVR technique aims to preserve mitral annular (MA) function, maintaining leaflet mobility and reducing leaflets’ stress, with the ultimate goal of improving repair durability and patient outcomes. Although the feasibility and success of this technique have been documented, there are no reports on the effect of NVR on MA function and leaflet mobility. Accordingly, the aims of the present study were to evaluate whether dynamic changes in MA motion are preserved early postoperatively with the NVR method and to compare these changes with those observed with resectional mitral valve repair (RVR) method.
Methods
Study Population
Patients with organic severe MR who were already scheduled for mitral valve repair surgery with either the NVR or the RVR technique were approached for enrollment and performance of three-dimensional (3D) transesophageal echocardiography before and immediately after surgery. Patients with arrhythmias or poor echocardiographic images were excluded. All patients provided written informed consent; the studies were approved by the respective human research review boards of the Methodist Hospital and the University of Chicago Medical Center.
Surgical Techniques
The NVR operations were performed at the Methodist Hospital in Houston, whereas the RVR operations were performed at the University of Chicago Medical Center. The NVR technique involves no leaflet resection and dynamic adjustment of chordal and annular dimensions. It has been described in detail elsewhere. The alignment of the free edges of the prolapsed leaflet segments was corrected with insertion of polytetrafluoroethylene artificial chordae. Artificial chordal length adjustment and annuloplasty ring size selection were made during left ventricular inflation with pressurized normal saline. The progressive distension of the left ventricle and aortic root simulated the late diastolic and some of the isovolumic systolic changes in left ventricular, aortic root, and mitral valve geometry. During left ventricular inflation, MA size was reduced, and the leaflets’ zone of apposition was increased. At that time, the upper edges of the desired zone of apposition were marked with dots. By traction on posteriorly placed annular stay stitches, the leaflets descended progressively into the left ventricular cavity, and the zone of apposition was further increased. This maneuver was followed by chordal length adjustment. When only the dots on the leaflets marking the upper edges of the chosen zone of apposition were visible, the circumference of the annulus was sized, and the appropriate flexible ring was chosen. The ring was attached with a continuous suture.
The RVR operation involved mitral leaflet resection of the diseased segment with subsequent leaflet plasty. The prolapsed or flail segment was excised in a limited fashion (a small triangular segment of leaflet’s free edge). This was governed by the amount of leaflet tissue that was flail or unsupported by chords. The adjacent leaflet edges were then overlapped and sutured together to recreate the leaflet edge. If chordal support was not adequate, artificial chordae made of polytetrafluoroethylene suture were inserted. The artificial chordae were anchored to the body of the papillary muscle; the chords’ length was judged on the basis of the distance from the papillary muscle to the annulus. Abnormal chords and loose, torn chords were excised. A partial flexible ring or band was secured with interrupted 2-0 TI•CRON simple sutures (Covidien, Mansfield, MA) along the posterior annulus in a way that the partial ring was implanted from trigone to trigone, completely subtending the posterior annulus and just going beyond the trigones onto the anterior annulus. The valve was then tested with a passive saline injection into the ventricle and further tested with antegrade aortic root cardioplegia.
Three-Dimensional Echocardiographic Protocol
Under general anesthesia, patients underwent intraoperative 3D transesophageal echocardiography, before surgery and immediately after surgery, after bypass. All studies were performed with the chest closed. The imaging protocol included a midesophageal, full-volume or zoomed acquisition aiming to visualize the mitral annulus and leaflets throughout the cardiac cycle. Image acquisition was adjusted to maximize frame rate. The imaging was performed using an iE33 ultrasound system (Philips Medical Systems, Andover, MA) with an X7-2t probe. The 3D images were digitally stored and analyzed at a later time.
MA Analysis
All analyses were performed at the Methodist Hospital. The mitral annulus was modeled preoperatively and postoperatively using an offline analysis station (MVQ; Philips Medical Systems). Preoperatively, the native mitral annulus was traced. Postoperatively, the surgical ring was traced. Because of incomplete MA ring after RVR, tracing of the mitral annulus involved the surgical ring when present and incorporated the native mitral annulus when the surgical ring was absent ( Figure 1 ). Accordingly, the RVR model includes some of the aortic-mitral continuity.
The following MA measurements were obtained ( Figure 2 ): (1) MA area in projection plane (two-dimensional [2D]) and in minimal 3D surface area, (2) annular anterior-to-posterior diameter and anterolateral-to-posteromedial diameter, (3) annular height, and (4) MA circumference in projection plane (2D). To quantitate MA dynamics, the above measurements were repeated at end-diastole, midsystole, end-systole, middiastole, and the last frame, which represents the time point before atrial contraction, because 3D display frequently ends just before atrial contraction (the shortest R-R interval in the sequential captured images). End-diastole was taken as the frame before mitral valve closure. End-systole was defined as the first frame depicting aortic valve closure. Midsystole and middiastole were identified as the frames in the middle of systole and diastole, respectively.
Motion of Mitral Valve Leaflets
Leaflet mobility was analyzed using TomTec software (TomTec Imaging Systems, Munich, Germany). A four-chamber plane was sliced, and the maximum angle between the mitral annulus and each leaflet (at A2 and P2, respectively) was measured ( Figure 3 ). The annulus was tagged first, and then approximately 0.5 to 1 cm of the proximal adjacent leaflets were tagged to create an angle ( Figure 3 ). This measurement was performed in systole and diastole. For each leaflet, the difference in maximal angulation between systole and diastole was calculated to express the range of leaflet mobility in degrees. Each angle measurement was repeated three times and averaged.
Statistical Analysis
Continuous variables are expressed as mean ± SD and categorical variables as percentages. Mean difference between the NVR and RVR techniques were tested using Mann-Whitney U tests for continuous variables and Fisher’s exact tests for categorical variables because of the small sample size. Wilcoxon’s signed-rank test was used to compare leaflet mobility before and after surgery. Because of repeated measurements over the cardiac cycle and the patterns of change in the means of some annular parameters cannot be characterized by first-degree and second-degree polynomials in time and cannot be well approximated by polynomials in time of any order, linear or linear spline models from the longitudinal data analysis were used to investigate the effects of the surgical procedures on the patterns of change in the means of annular parameters over time. In all models, an exchangeable correlation structure was specified for the within-subject association between repeated measurements for each annular parameter, at each cardiac cycle for each patient. In linear spline models, the time axis was divided into a series of segments, and piecewise linear models were modeled. The number of segments for each annular parameter was driven by five time points measured during the cardiac cycle. Interaction between each operative technique and each cardiac cycle time point was introduced to examine if the pattern of change over the cardiac cycle or specific cardiac cycle time points in the two operations was significantly different. All statistical analyses were performed using Stata version 11 (StataCorp LP, College Station, TX). Statistical significance was defined as a two-tailed P value < .05 for all tests.
Results
There were 15 patients in the NVR operation group and 13 in the RVR operation group. Table 1 summarizes the patients’ characteristics and hemodynamics before and after mitral repair. No annular calcification was seen in either group. All patients had moderate to severe or severe MR preoperatively, and successful repair was achieved, with no or at most mild regurgitation. During the NVR procedure, a complete flexible ring was used in all patients (ATS in 13 [ATS Medical, Inc., Minneapolis, MN], St. Jude in two [St. Jude Medical, St. Paul, MN]), with insertion of artificial chordate and no leaflet resection. In the RVR technique, an incomplete flexible ring was attached in all patients (St. Jude ring/band in 11 [St. Jude Medical], Medtronic Simplici-T band in two [Medtronic, Inc., Minneapolis, MN]), along with partial leaflet resection and insertion of chordae as needed. No significant intergroup difference was observed with respect to mitral valve hemodynamics, except for higher systolic blood pressure at baseline in the patients undergoing RVR. No differences in hemodynamics were noted immediately postoperatively. The NVR operation had shorter cardiopulmonary bypass and aortic cross clamp times.
| Variable | NVR ( n = 15) | RVR ( n = 13) |
|---|---|---|
| Age (y) | 56.2 ± 13.3 | 61.3 ± 17.5 |
| Men | 10 | 8 |
| Barlow disease | 8 | 7 |
| Valve pathology | ||
| Anterior leaflet prolapse | 10 | 7 |
| Posterior leaflet prolapse | 13 | 10 |
| Posterior leaflet flail | 7 | 9 |
| Anterior leaflet flail | 1 | 0 |
| Preoperative | ||
| Systolic blood pressure (mm Hg) | 103 ± 28 | 133 ± 15 ∗ |
| Diastolic blood pressure (mm Hg) | 64 ± 12 | 73 ± 12 |
| Heart rate (beats/min) | 67 ± 12 | 67 ± 11 |
| Left ventricular ejection fraction (%) | 64 ± 10 | 63 ± 7 |
| Intraoperative | ||
| Cardiopulmonary bypass time (min) | 106.6 ± 53.6 | 176.5 ± 42.5 ∗ |
| Aortic cross clamp time (min) | 59.2 ± 55 | 137.8 ± 34.2 ∗ |
| Ring size (mm) | 32 ± 2.8 | 29 ± 3.7 |
| Postoperative | ||
| Systolic blood pressure (mm Hg) | 111 ± 16 | 117 ± 8 |
| Diastolic blood pressure (mm Hg) | 60 ± 20 | 63 ± 8 |
| Heart rate (beats/min) | 86 ± 21 | 92 ± 12 |
| Left ventricular ejection fraction (%) | 52 ± 13 | 60 ± 10 |
| Mitral doppler parameters | ||
| Postoperative peak velocity (cm/sec) | 149 ± 22 | 167 ± 43 |
| Postoperative mean gradient (mm Hg) | 3.3 ± 1.1 | 4 ± 2 |
MA Dynamics Preoperatively
MA area was found to be enlarged preoperatively in both groups: the maximal annular areas on 2D imaging were 18.5 ± 4.6 cm 2 in the NVR group and 20.1 ± 4.8 cm 2 in the RVR group. Figures 4 to 7 demonstrate the dynamic changes of the mitral annulus at baseline for both RVR and NVR operations. The results of the longitudinal data analysis showed that significant dynamic changes during the cardiac cycle in MA area ( Figures 4 A and 4 B), height ( Figure 5 ), and circumference in 2D ( Figure 6 ) were noted in both groups. During diastole, the mitral annulus enlarged and flattened, whereas it became smaller and higher (as it contracted and folded) with atrial contraction (late diastole) and toward midsystole. Preoperatively, the RVR group demonstrated significantly more dynamic changes during diastole in MA area and circumference. With regard to MA diameters, although change during the cardiac cycle in anterior-posterior diameter was similar for both techniques preoperatively ( Figure 7 A), the anterolateral-to-posteromedial diameter change was larger in the RVR group ( Figure 7 B). After adding preoperative systolic blood pressure into the models, the dynamic changes during the cardiac cycle in all annular parameters were not different.
