Congenital Mitral Regurgitation Repair Based on Carpentier's Classification: Long-Term Outcomes

Background There are few reports of the outcomes of standardized surgical management addressing the etiologic and morphologic aspects of mitral valve malformation according to Carpentier's classification. This study aimed to evaluate the long-term outcomes of mitral valve repair in children according to Carpentier's classification. Methods Patients who underwent mitral valve repair at our institution between 2000 and 2021 were retrospectively reviewed. Preoperative data, surgical techniques, and outcomes were analyzed according to Carpentier's classification. The proportion of patients free of mitral valve replacement and reoperation was estimated using Kaplan-Meier analysis. Results Twenty-three patients (median operative age, four months) were followed up for 10 (range, 2-21) years. Preoperative mitral regurgitation was severe in 12 patients and moderate in 11 patients. Eight, five, seven, and three patients had Carpentier's type 1, 2, 3, and 4 lesions, respectively. Ventricular septal defect (N = 9) and double outlet of the great arteries from the right ventricle (N = 3) were the most commonly associated cardiac malformations. There were no cases of operative mortality or deaths during the follow-up. The overall five-year rate of freedom from mitral valve replacement was 91%, whereas the five-year rates of freedom from reoperation were 74%, 80%, 71%, and 67% in type 1, 2, 3, and 4 lesions, respectively. Postoperative mitral regurgitation at the last follow-up was moderate in three patients and less than mild in 20 patients. Conclusions Current surgical management of congenital mitral regurgitation is generally considered adequate; however, more complicated cases required a combination of various surgical techniques.


Introduction
The surgical treatment for congenital mitral valve regurgitation (MR) in children remains a significant challenge due to variations in patient anatomy, congenital anatomic limitations in annulus size, dysmorphic valve tissue, underdeveloped subvalvular apparatus, and maldeveloped papillary muscles. [1][2][3] In addition, patients' growth potential and desire to avoid valve replacement require good outcomes of mitral valve repair in children. 1,4,5 In 1976, Carpentier published the first pathophysiological classification of mitral valve malformation, categorizing it into four types; this definition was subsequently revised in 1998. 6,7 Carpentier type 1 mitral valve pathology involves isolated annular dilation, cleft leaflet, and leaflet agenesis; type 2 includes isolated chordae elongation, papillary muscle elongation, and chordae agenesis; type 3 comprises commissural papillary muscle fusion, short chordae, and excess valvular tissue; and type 4 consists of parachute valve, hammock valve, and papillary muscle agenesis. Only a few reports of the outcomes of standardized surgical management have addressed the etiologic and morphologic aspects of mitral valve malformation according to Carpentier's classification. Therefore, this study aimed to evaluate the mid-to long-term outcomes of mitral valve repair in children with congenital MR according to Carpentier's classification.

Ethical Considerations
Ethics approval for this study was obtained from the Ethics Committee and Institutional Review Board Committee of Osaka Women's and Children's Hospital (approval no. 1537; approval date: April 12, 2022). The requirement for written informed consent was waived because of the retrospective design of the study.

Study Design and Population
This retrospective cohort study included patients who underwent mitral valve repair for congenital MR at the Osaka Women's and Children's Hospital, Osaka, Japan, between January 2000 and December 2021. Patients with MR associated with atrioventricular septal defect, Marfan's syndrome, infective endocarditis, or single ventricle morphology were excluded.
We reviewed patients' records and extracted demographic and clinical data. This included echocardiography findings, operative findings and techniques, and surgical outcomes. We then analyzed these according to Carpentier's classification.

Diagnosis and Surgical Indication
MR was diagnosed using two-dimensional and color Doppler echocardiography in the parasternal long-axis view and was graded according to the American Society of Echocardiography Guidelines as trivial, mild, moderate, or severe based on quantitative measures; namely the vena contracta width and regurgitant fraction. 8 Mitral stenosis (MS) was graded on the basis of the mean pressure gradient (PG) across the valve as mild, <5 mm Hg; moderate, ≥5 mm Hg but <10 mm Hg; and severe, ≥10 mm Hg. 9 In addition, all patients routinely underwent preoperative cardiac catheterization to estimate the left ventricular end-diastolic volume. The decision for mitral valve intervention was made by the operating surgeon according to clinical status and MR grade (moderate-to-severe MR) with regard to anatomical features of the valve (annular size, leaflet motion, and competence of the subvalvular apparatus) assessed with echocardiography.

Operative Technique
All procedures were performed by the same surgical team based on the same surgical strategy ( Figure 1). All patients underwent median sternotomy, with cardiopulmonary bypass and mild hypothermia at 32°C core temperature. Mitral valve exposure was achieved through left atriotomy or interatrial septum incision. An injection of saline solution into the left ventricle was used to assess the morphology of mitral valve malformation.
In patients with type 1 mitral valve malformation, particularly those with annular dilation, Kay-Reed's commissural annuloplasty was applied. 10 A 6-0 polypropylene suture with synthetic pledgets was used for annuloplasty. In patients with a cleft leaflet, a cleft suture was applied. The cleft was closed with interrupted fine stitches, taking care not to cause leaflet distortion or reduced leaflet coaptation.
In patients with type 2 and 3 mitral valve malformations, commissural annuloplasty according to the Kay-Reed method was the first choice. For those with type 3 mitral valve malformation, edge-to-edge paracommisural repair was also performed if commissural annuloplasty was not sufficient to control the MR. 11 The edge-to-edge technique was performed with interrupted fine stitches.
In patients with type 4 mitral valve malformation, various techniques were applied, including cusp elongation, artificial chordae placement, papillary muscle splitting, and posterior cusp quadrangular resection repair. For cusp elongation, we used glutaraldehyde-treated autologous pericardial tissue. For artificial chordae placement, we used a 5.0 Gore-Tex suture. In cases of reoperation, we applied a partial ring with glutaraldehyde-treated autologous pericardium or a rigid ring for annuloplasty.

Postoperative Medication Regimen
Postoperatively, all patients started receiving aspirin (minimal dose, 5 mg/kg/d) and enalapril (minimal dose, 0.2 mg/kg/d). Aspirin was discontinued at 3 months postoperatively, while enalapril was discontinued when the MR was less than mild at 1 year postoperatively.

Follow-Up
Follow-ups included a clinical examination and transthoracic echocardiography. All patients were routinely followed up with echocardiography before discharge, 1 year postoperatively, and every 5 years thereafter.

Statistical Analysis
Statistical analysis was performed using SPSS Statistics for Windows, version 18.0 (SPSS Inc). Frequencies are presented as absolute numbers and percentages. Continuous data are presented as mean ± standard deviation or median values with ranges. Mean values were compared using Student t test. Kaplan-Meier analysis was used to estimate the proportion of patients free of mitral valve replacement and reoperation. Freedom from reoperation rates was compared by log-rank test. For all tests, P < .05 was considered significant.

Preoperative Characteristics
Patients' demographic characteristics are described in Table 1. Twenty-three patients with a median age of 4 months (range, 1 month to 8 years) were included, of whom 14 (61%) were less than 1 year old. Ventricular septal defect (VSD) (n = 9; 39%) and double outlet right ventricle (DORV) (n = 3; 13%) were the most commonly associated cardiac malformations. Half of the patients (n = 12; 52%) had severe MR on preoperative echocardiography. The most common pathophysiological changes of the mitral valve according to Carpentier's classification were type 1 (n = 8; 35%) and type 3 (n = 7; 30%). Out of 12 patients with VSD or DORV, the type of mitral valve malformation was type 1 in five (42%), type 2 in four (33%), and type 3 in three (25%), and the median age of mitral valve repair was 8 months (range: 1 month to 10 years). The preoperative median mean PG was 2 mm Hg (range: 0-8). The preoperative median left ventricular end-diastolic volume was 207% (range: 107-462) of normal.

Primary Mitral Valve Repair
For a primary mitral valve repair, Kay-Reed's annuloplasty without prosthetic materials was applied in 16 patients, cleft suture in six, cusp elongation with an autologous pericardial patch in two, edge-to-edge paracommisural repair in one, and artificial chord reconstruction in one patient. More than two techniques were applied in five patients (22%) ( Table 2).

Follow-Up
The mean follow-up period for the entire population was 10 years (range, 2-21). There were no cases of operative mortality or deaths during the follow-up period. All patients had a normal sinus rhythm, and none required pacemaker implantation. Among the six patients who underwent reoperation, two required mitral valve replacement ( Table 3). The overall 5-year rate of freedom from mitral valve replacement was 91%, whereas the 5-year rate of freedom from reoperation was 74%, 80%, 71%, and 67% in types 1, 2, 3, and 4, respectively ( Figure 2). The postoperative MR grade on echocardiography at the last follow-up was moderate in three patients and  less than mild in 20 patients. The postoperative MS grade on echocardiography at the last follow-up, excluding two patients who underwent valve replacement, was less than moderate with a median mean PG of 1 mm Hg (range: 0-6.8) and without congestive heart failure. Postoperatively, the left ventricular enddiastolic volume improved to 139% (range: 80-268) of normal (P = .006). However, there was no association between preoperative left ventricular end-diastolic volume of patients who needed reoperation and patients who did not need reoperation (184 ± 50 vs 207 ± 99% of normal) (P = .31).
Surgical Management and Outcomes According to Carpentier's Classification Type 1: Normal leaflet motion. Cleft suture was performed in five patients, Kay-Reed's annuloplasty in two, and cleft suture and Kay-Reed's annuloplasty in one. In patients who underwent Kay-Reed's annuloplasty, the mitral valve diameter changed from 129% ± 5% to 102% ± 3% of normal. In 6 of the 8 patients (75%), the MR grade improved to less than mild. Reoperation was performed on two patients, one of whom required mitral valve replacement ( Figure 3A).
Type 2: Leaflet prolapse. Kay-Reed's annuloplasty was performed in all five patients. The mitral valve diameter changed from 126% ± 18% to 107% ± 8% of normal. In four patients (80%), the MR grade improved to less than mild. Only one patient required reoperation ( Figure 3B).
Type 3: Restricted leaflet motion with normal papillary muscles. Kay-Reed's annuloplasty was performed in six patients, and edge-to-edge repair in one. The mitral valve diameter changed from 129% ± 9% to 96% ± 6% of the normal diameter.
In four patients (57%), the MR grade improved to less than mild ( Figure 3C). According to the MS, the preoperative median mean PG was 0.5 mm Hg (range: 0-2), and there were no patients with postoperative MS more than moderate. Two patients required reoperation.
Type 4: Restricted leaflet motion with abnormal papillary muscles. This category includes three patients (patients 1, 2, 3). At the age of 8 months, "patient 1" underwent artificial chordae and Kay annuloplasty but eventually required mitral valve replacement (MVR) ( Figure 3D). At the age of one month, "patient 2" underwent posterior cusp elongation, artificial chordae, and Kay annuloplasty. At the age of 27 months, "patient 3" underwent anterior cusp elongation and posterior cusp quadrangular resection repair. The pre/post mitral valve diameter was 120% to 92% of normal in "patient 1," 124% to 97% of normal in "patient 2," and 121% to 120% of normal in "patient 3." The preoperative mean PG was 8 mm Hg in "patient 1," 0 mm Hg in "patient 2," and 4.3 mm Hg in "patient 3." Cusp elongation performed in two of them resulted in an improvement of the MR grade to less than mild but temporarily suffered from severe MS (11 and 10 mm Hg) before discharge. Nevertheless, the last follow-up mean PGs improved to 5 and 6.8 mm Hg.

Reoperation Cases According to Carpentier's Classification
Type 1: Normal leaflet motion. Two patients (cases 1 and 2, Table 3) with type 1 mitral valve malformation required reoperation. In case 1, one month after the primary repair with Kay's annuloplasty, annular dilation was observed, and we performed partial ring annuloplasty using glutaraldehyde-treated autologous pericardial material, which resulted in the reduction of the mitral valve diameter from 120% to 75% of the normal diameter.
In case 2, we performed the Yasui procedure due to DORV with coarctation of the aorta. As there was a large cleft reaching the annulus at the anterior mitral valve leaflet, we performed a cleft suture and Reed's annuloplasty. However, reoperation was required four months postoperatively due to a defect at the mitral valve repair site. We performed cleft resuture but were unable to control the mitral regurgitation; therefore, mitral valve replacement was performed. Type 2: Leaflet prolapse. Only one of the patients with type 2 mitral valve malformation (case 3, Table 3) required reoperation. This patient had congenital mitral regurgitation with leaflet prolapse caused by a defect of the strut chordae at the A2 position, for which Kay's annuloplasty was performed, resulting in a reduction of the mitral valve diameter from 126% to 113% of the normal diameter. Reoperation was performed approximately two years later using partial annuloplasty with a Duran ring, and the mitral valve diameter was reduced from 164% to 92% of the normal diameter. Artificial chordae placement was performed as well. Eventually, at the age of 16, the Duran ring was removed to prevent MS.
Type 3: Restricted leaflet motion with normal papillary muscles. Two patients with this type of mitral valve malformation (cases 4 and 5, Table 3) required reoperation. In case 4, due to the corrected complete transposition of the great arteries and VSD, we performed a double-switch operation, with the hemi-Mustard procedure and Kay's annuloplasty for posterior mitral valve leaflet tethering. Reoperation was required due to regurgitation at the anterolateral commissure, which was treated using the edge-to-edge technique as paracommisural repair. Case 5 was a patient with a VSD and congenital mitral regurgitation at the posteromedial commissure because of restricted leaflet motion due to a thickened anterior leaflet. We performed VSD closure and edge-to-edge paracommisural repair. Reoperation was performed approximately one year later using Kay's annuloplasty and the edge-to-edge technique used as a double orifice repair.
Type 4: Restricted leaflet motion with abnormal papillary muscles. Only one of the patients with this type of mitral valve malformation (case 6, Table 3) required reoperation. This patient was diagnosed with an idiopathic rupture of chordae tendineae and underwent artificial chordae placement and Kay's annuloplasty. However, two months postoperatively, a hammock mitral valve was observed, which had not been noticed previously. We performed cusp elongation, edge-to-edge paracommisural repair, and Reed's annuloplasty, but were unable to control the mitral regurgitation; therefore, mitral valve replacement was performed.

Comment
The most challenging issue in treating pediatric patients with congenital mitral valve malformation is the range of pathologies affecting leaflets and subvalvular elements of mitral valves, which complicates systematic classification and the attempts at surgical standardization. 1 However, our results showed that surgical management of congenital MR according to Carpentier's classification produces good mid-and long-term outcomes. Some reports indicate favorable long-term results after mitral valve repair, describing survival rates of 86% to 94%. Others report rates of freedom from reoperation of 69% to 94% after 10 and 20 years, respectively. [12][13][14][15][16] In our experience, mitral valve repair proved feasible and successful in all patients, with an acceptable early and late mortality and reoperation rate.
It is reported that in patients with prevalent MR, aggressive surgical repair is necessary in order to achieve a better leaflet coaptation, while minimal MS, as a consequence of mitral valve annuloplasty, is usually well tolerated. 2 Maeda et al also underlined the obligation of an annuloplasty. 16 We performed commissural annuloplasty as reinforcement and prevention of annular dilation for all patients with sufficient mitral valve diameter, whereby the mitral annular diameter was adjusted to over 80% of normal ( Figure 1). Some patients experienced postoperative MS, but this improved at the last follow-up. Additionally, in congenital MR, when a mitral valve malformation is discovered in an infant, the major objective is to delay surgery as long as possible in order to reduce the risk of a potential valve replacement, which carries a high risk of reintervention and mortality in infants. 17 However, early intervention reportedly improves the outcomes of mitral valve repair, because dilation and hypertrophy of the left heart structures can be avoided with early mitral valve repair. [18][19][20] In our series, the median surgical age of four months indicates relatively earlier intervention, and the preoperative median left ventricular end-diastolic volume was significantly improved postoperatively. These factors could also have contributed to the good results obtained in our study.
In patients with Carpentier's type 1 mitral valve malformation, the lesions were localized, and surgical outcomes were satisfactory by minimally invasive methods, such as Kay-Reed's annuloplasty and cleft suture, which has also been previously reported. 4,21,22 However, in patients with a large cleft reaching the annular ring (such as in case 2), performing only cleft suture could influence the valve leaflet dimensions and worsen the mitral regurgitation. Therefore, patch augmentation seems to be a better choice for these patients. 2 For Carpentier's type 2 mitral valve malformation, artificial chordae placement is an effective treatment. 23-25 However, Minami et al emphasized that it is difficult to decide on the size of the artificial chordae due to considerations of the patients' growth. 23 Hence, we had opted for Kay-Reed's annuloplasty as the initial treatment in these patients. If the MR could not be controlled, particularly in cases of leaflet prolapse caused by a defect of the strut chordae, such as in case 3, artificial chordae placement seems to be effective.
For Carpentier's type 3 mitral valve malformation, we performed Kay-Reed's annuloplasty and almost all patients showed good outcomes but, in fact, some patients, such as cases 4 and 5 needed edge-to-edge paracommisural repair. This indicates that it may be necessary to add edge-to-edge paracommisural repair in the management of this type, as previously reported. 26 In patients with Carpentier's type 4 mitral valve malformation, mitral valve repair is extremely difficult and surgical outcomes tend to be unsatisfactory. 7 Type 4 lesions are usually not isolated and are often characterized by a constellation of other pathological changes of the mitral valve leaflets, annulus, and ring. Therefore, the majority of patients require one or more surgical techniques. Among the techniques performed on these patients at our institution, cusp elongation seemed to be the most effective. However, it has shown some limitations; first, in cases where cusp elongation was ineffective, we had no choice other than mitral valve replacement. This is because we generally opted for minimally invasive methods, such as artificial chordae and Kay-Reed's annuloplasty, and if these were found to be insufficient to control the MR, we performed cusp elongation. The outcome of cusp elongation for our infant patient was good. If MR cannot be controlled by minimally invasive methods, cusp elongation is required even in infant patients. Second, according to the postoperative MS, the two patients who underwent cusp elongation postoperatively experienced MS. The reason seems to be a preoperative underestimation of MS or surgical technique. The mitral valve malformations, particularly in Carpentier group 4 (ie, hammock MV; and parachute MV), are often a combined lesion of incompetence and stenosis. 27 On the other hand, Kalfa et al demonstrated that excessive enlargement of leaflets resulted in the reconstruction of posterior leaflet motion. 1 It is also reported that good effects by cusp elongation can only be obtained with sufficient mitral valve diameter. 28,29 The present study had a few limitations. First, it was a singlecenter, retrospective study that included a small number of patients. Second, the effect of improvement in surgical technique with increasing surgical experience was not measured.
In conclusion, surgical management of congenital mitral valve malformation according to Carpentier's classification provided good mid-and long-term outcomes in our patients, with a surgical mortality rate of 0% and a 5-year rate of freedom from mitral valve replacement of 91%. The surgical outcomes of Carpentier's type 1, 2, and 3 lesions were good, whereas the more complicated type 4 lesions required a combination of surgical techniques, including annuloplasty, artificial chordae placement, posterior cusp quadrangular resection repair, and cusp elongation.