This study provides information on mid-term clinical and echocardiographic outcomes with detailed mechanisms of recurrent MR focusing on bileaflet MV repair due to degenerative disease. Whereas posterior leaflet prolapse now can be repaired in practically all cases, with very low morbidity and excellent outcomes, many published series have documented decreased repair rates in the setting of anterior and bileaflet lesions [1,2,3,4,5,6].
The long-term durability of MV repair of bileaflet lesions is variable even among experienced or high-volume centers [2, 12,13,14]. In the present study, the freedom from recurrent moderate or severe MR was 87.9 ± 4.7% at 1 year and 71.1 ± 11.0% at 5 years. Although these results were inferior to the results from other centers, the high ratio of artificial neochordae placement and exclusion of limited commissural lesions in our cohort might reflect more complex MV pathology. The present study showed that 3 out of 10 patients with recurrent MR finally died of congestive heart failure. Gillinov et al.  reported that when a valve either appears unrepairable or attempts to repair fail because of complex valve pathology, neither survival nor reoperation is adversely affected by replacement. Therefore, we should know what types of MV lesions and MV repair techniques are associated with recurrent MR in complex valve pathology.
Some authors have demonstrated that the timing of failed repair can be categorized as early for procedure-related failure and later for valve-related failure, as in the present study [4, 16]. One of the challenges in categorizing late failure after the repair pertains to distinguishing between recurrent MR caused by a technical failure during surgery versus recurrent valve leakage caused by progression of native disease . However, when recurrent MR occurs late after the repair of degenerative MV disease, new valve pathology is usually the culprit and re-repair is less common. In contrast, reoperation for procedure-related failure occurs early and is often amenable to re-repair [4, 16]. Indeed, early recurrent MR due to inappropriate artificial neochordae length, which was an etiology of early recurrent MR in the present study, was successfully re-repaired. Conversely, re-repair performed on the patient with recurrent MR occurring late after commissural repair resulted in re-recurrent MR. Suri et al. demonstrated that repair is clearly beneficial, conveying improved survival with better recovery of left ventricular function and left ventricular regression compared with valve replacement . Thus, patients should be operated on at an early phase (asymptomatic or mildly symptomatic) because there is a higher probability of repair and a greater benefit on long-term survival . We experienced 1 patient who required valve replacement due to posterior leaflet degeneration 2.6 years after MV repair. Although the patient did not have any symptoms with recurrent moderate MR, we should have performed re-repair before the MR jet caused progressive degeneration of MV leaflets.
Early failure can be caused by failure of neochordoplasty that results in increased strain on remaining chords or aggressive leaflet resection that results in late fibrosis and decreased mobility . One of the challenges in MV repair for the treatment of bileaflet lesions is that neochordoplasty and leaflet resection are required on the anterior and posterior mitral leaflets, respectively. Unlike single anterior or posterior mitral leaflet repair, it is difficult to decide the area of resection or length of artificial neochordae to achieve sufficient coaptation length because the geometry may be changed after every step of repair. Quadrangular resection may preserve less valve function and leaflet kinematics than triangular resection or a non-resection technique . Adjustment of the length of the artificial neochordae is sensitive, and inappropriate length can easily cause recurrent MR. Furthermore, Gillinov et al. reported that 70% of patients who demonstrated bileaflet prolapse on echocardiography did not have any significant anterior chordal pathology . This means that bileaflet prolapse can be repaired only by posterior leaflet repair, and there is a risk of performing unnecessary neochordoplasty by wrong evaluations of the MV lesions. Considering the experiences of failure in MV repair in the present study, we suggest the following sequential approaches, in accordance with the suggestions by Castillo et al.  Repair of the posterior leaflet must be performed first by triangular resection or neochordoplasty to avoid aggressive leaflet resection. Next, whether anterior leaflet repair is needed must be re-evaluated based on the status of the chordae to the anterior leaflet . Thereafter, if neochordoplasty is required to repair the anterior mitral leaflet lesions, re-evaluation can be performed to decide if additional width or height resection of the posterior MV leaflets is required after anterior MV repair. Lastly, fine-tuning of the artificial neochordae is required.
The number of studies pertaining to the long-term outcomes after MV repair for commissural lesions is limited. Shimizu et al. reviewed 122 patients with isolated commissural prolapse, which was repaired with leaflet resection or chordal replacement . They reported that freedom from recurrent moderate or severe MR at 15 years was 87.4%. De Bonis et al. also reported the long-term outcomes of commissural plication with mitral annuloplasty for isolated commissural prolapse . In their study, the freedom from moderate-severe or severe MR at 11 years was 96.3 ± 1.7%. However, if moderate MR was taken into consideration, 13 of 121 (10.7%) patients had recurrent MR. Therefore, more than one-tenth of patients who underwent MV repair for commissural lesions can develop recurrent MR during the long-term follow-up period. The mid-term outcomes of MV repair for commissural prolapse with plication and neochordoplasty in the present study were also unsatisfactory despite the lack of recurrent MR in the early phase. Although this implies that there may be underlying unknown failure mechanisms in commissural repair, the present study could not reveal the risk factors or specific mechanisms owing to the small number of samples. Further studies are warranted to elucidate the mechanisms of failure and explore better repair techniques for commissural lesions.
This study has the inherent limitations of any observational study. Several surgeons performed the surgical procedures and the results may have been affected by each surgeon’s skill. Despite the high proportion of patient follow-up, bias may have occurred among patients presenting for follow-up assessment and those who did not. MV function was assessed in multiple echocardiography laboratories and the interpretation of the results may not have been consistent. In addition, the mechanisms of recurrent MR were mainly evaluated on transthoracic echocardiography. Thus, further studies with evaluation on transesophageal echocardiography or surgical inspection during reoperation will be required to reveal more detailed mechanisms of recurrent MR.