2018 Volume 82 Issue 6 Pages 1705-1711
Background: Although minimally invasive mitral valve repair (MIMVR) is increasingly being performed, only a few clinical studies from Japanese institutions have been reported.
Methods and Results: From 2006 to 2017, 387 consecutive patients (135 females, mean age 56±13 years) underwent an initial isolated MIMVR through a right minithoracotomy. The mitral etiology was degenerative in 348, functional in 22, and endocarditis in 13 cases. Repair techniques included leaflet resection/plication in 280, chordal reconstruction in 109, and annuloplasty alone in 24 patients, and concomitant procedures included tricuspid valve repair and atrial fibrillation ablation in 70 (18.1%) and 78 (20.2%), respectively. Hospital mortality rate was 0.26%; 2 patients (0.5%) required intraoperative conversion to a median sternotomy. Perioperative morbidity included stroke (1.3%), reoperation for bleeding (0.8%), prolonged ventilation (0.5%), and permanent pacemaker implantation (2.1%). The transfusion rate was 14.7% and median ventilation time was 4 hours. Overall 5-year survival was 96.9%. For patients with degenerative mitral regurgitation (MR), the 5-year freedom from reoperation or severe recurrent MR, and freedom from ≥moderate MR were 94.7% and 82.2%, respectively. Repair for anterior mitral leaflet prolapse and the initial 30 cases were associated with higher occurrence of recurrent MR.
Conclusions: MIMVR can be performed safely with low levels of mortality and morbidity, and provides sufficient repair durability. A learning curve exists in terms of repair durability, especially for anterior mitral leaflet repair.
Minimally invasive cardiac surgery (MICS) through a right minithoracotomy has been shown to be a feasible alternative to conventional full-sternotomy surgery, and many studies have reported excellent clinical outcomes in terms of mortality and morbidity, as well as faster patient recovery and return to normal activities.1–6 As a result, MICS is increasingly being used as a routine approach worldwide, especially for mitral valve (MV) surgery. On the other hand, several concerns have been raised because MICS is technically more complex and has a steep learning curve, especially in relation to perioperative morbidity.7 In addition, presently available clinical evidence from MICS cases is based on studies mostly conducted in large centers in Europe and North America, and only a few studies in Japan have reported clinical findings.8,9 We began our MICS program in 2006, sooner than most other Japanese institutes, and it has become the standard procedure for MV repair. The aim of this study was to review the clinical results of consecutive patients who underwent a minimally invasive MV repair (MIMVR) procedure during the 11-year period since starting the program.
From March 2006 to October 2017, a total of 476 patients underwent minimally invasive MV surgery (MIMVS) through a right minithoracotomy at The Sakakibara Heart Institute of Okayama. Of those, 387 who underwent an initial isolated MV repair were included in the present study. The mean age was 56±13 years and 135 (34.9%) were female. During the same period, 172 patients underwent an initial isolated MV repair through a median sternotomy. The numbers of MIMVR procedures by operative year are shown in Figure 1. The etiology of mitral disease was degenerative in 89.9%, functional in 5.7%, and others in 4.3%. Details regarding the preoperative patient characteristics are presented in Table 1. The institutional ethical committee approved this study and the need for individual patient consent was waived.
Annual number of cases of minimally invasive isolated mitral valve repair and conventional isolated MV repair through a standard sternotomy. Cases of reoperation and of combination with aortic valve surgery, aortic surgery, or coronary artery bypass grafting were excluded. MICS, minimally invasive cardiac surgery.
| Age, years (range) | 56±13 (19–82) |
| Female sex, n (%) | 135 (35) |
| Body surface area, m2 (range) | 1.66±0.18 (1.20–2.20) |
| BMI, kg/m2 (range) | 22.6±3.5 (14.7–37.8) |
| NYHA functional class | 1.8±0.7 |
| Class I/II/III/IV (%) | 34/50/15/1.5 |
| Hypertension, n (%) | 144 (37) |
| Hyperlipidemia, n (%) | 65 (17) |
| Diabetes mellitus, n (%) | 22 (5.7) |
| Preoperative AF, n (%) | 56 (15) |
| History of smoking, n (%) | 108 (28) |
| Peripheral vascular disease, n (%) | 1 (0.3) |
| History of cerebrovascular disease, n (%) | 11 (2.9) |
| Emergency/urgent surgery, n (%) | 4 (1.1) |
| Ejection fraction, % | 67.5±8.6 |
| Preoperative serum creatinine, mg/dL | 0.86±0.47 |
| Etiology, n | 387 |
| Degenerative, n (%) | 348 (90) |
| Functional, n (%) | 22 (5.7) |
| Endocarditis, n (%) | 13 (3.4) |
| Other, n (%) | 4 (1.0) |
| Mitral valve prolapse, n | 361 |
| Anterior leaflet, n (%) | 79 (22) |
| Posterior leaflet, n (%) | 230 (64) |
| Bileaflet, n (%) | 31 (8.6) |
| Commissural, n (%) | 21 (5.8) |
Continuous variables are expressed as mean±standard deviation. AF, atrial fibrillation; BMI, body mass index; MIMVR, minimally invasive mitral valve repair; NYHA, New York Heart Association.
Each patient was intubated with a double-lumen endotracheal tube and placed in a left lateral decubitus position. A right minithoracotomy was performed through a 4–7 cm skin incision placed lateral to the midclavicular line. Simultaneously, a small oblique incision was made in either the right or left groin to expose the femoral vessels for cannulation. Vessels were not encircled with vascular tape to avoid vascular spasm. Utilizing a modified Seldinger technique, cannulation was performed under fluoroscopic guidance. Cardiopulmonary bypass (CPB) was then instituted, and an additional venous cannula was inserted into the superior vena cava directly or percutaneously through the right internal jugular vein. Body temperature was maintained at approximately 34℃ and vacuum-assisted venous drainage was used throughout the procedure.
When there were concerns about retrograde arterial perfusion, such as inadequate size of the femoral artery or existence of atheromatous plaques in the thoracic or abdominal aorta, alternative CPB approaches were considered. Preoperative contrast-enhanced computed tomography was routinely performed to detect peripheral vascular disease and aortic pathology. Intraoperative lower extremity perfusion monitoring was also routinely done using near infrared spectroscopy (NIRS). The addition of distal leg perfusion, or conversion to central aortic cannulation, was considered when NIRS showed >20% decrease in the tissue oxygenation index. As a consequence, 14.5% of the present patients were treated with an alternative access, such as central aortic cannulation (n=40), addition of contralateral femoral arterial cannulation (n=11), or axillary arterial cannulation (n=5).
After confirming secure systemic perfusion with CPB, the ascending aorta was cross-clamped with a Chitwood transthoracic clamp (Geister Medizintechnik GmbH, Tuttlingen, Germany) or Cygnet clamp (Vitalitec, Plymouth, MA, USA), placed through the transverse sinus. Antegrade crystalloid cardioplegia (modified St. Thomas solution) at a dose of 30 mL/kg was delivered through a standard cardioplegia catheter and repeated 60 min later.10 Aortic root pressure was continuously monitored during administration of cardioplegia to secure its delivery. Retrograde cardioplegia administration was occasionally used for patients with more than mild aortic insufficiency or for those expected to undergo a complex repair, and was placed directly by the surgeon through the right atrium under transesophageal echocardiography guidance.
The left atrium was entered through the usual atrial incision and a retractor positioned. The MV procedure was performed under both direct vision and thoracoscopic assistance; 40 patients (10.3%) underwent a total endoscopic repair. The endoscope was inserted through a 5- or 10-mm port in the 4th right intercostal space.
Concomitant procedures included tricuspid valve repair in 70 (18.1%), atrial fibrillation ablation in 78 (20.1%), and atrial septal defect repair in 6 (1.6%) patients. In 2 patients with hypertrophic obstructive cardiomyopathy (HOCM) and significant mitral regurgitation (MR) caused by systolic anterior motion of the mitral leaflet, a transmitral septal myectomy with anterior leaflet augmentation and approximation of the subvalvular apparatus was performed. Mean total procedure, CPB, and cross-clamp times were 274±62, 174±49, and 119±36 min, respectively (Table 2).
| Repair technique, n (%) | |
| Leaflet resection | 199 (51) |
| With height reduction | 41 (11) |
| Without height reduction | 158 (41) |
| Leaflet plication | 81 (21) |
| Chordal replacement | 109 (28) |
| Pericardial leaflet augmentation | 5 (1.3) |
| Annuloplasty alone | 24 (6.2) |
| Concomitant procedures, n (%) | |
| Tricuspid annuloplasty | 70 (18) |
| AF ablation | 78 (20) |
| Atrial septal defect repair | 6 (1.6) |
| Type of annuloplasty ring, n (%) | |
| Flexible | 189 (49) |
| Semirigid | 196 (51) |
| Size of annuloplasty ring, mm (range) | 30.3±2.2 (26–38) |
| CPB access, n (%) | |
| Single FA | 331 (86) |
| Bilateral FA | 11 (2.8) |
| Single FA+Axillary A | 5 (1.3) |
| Ascending aorta | 40 (10) |
| Total procedure time, min (range) | 274±62 (137–503) |
| CPB time, min (range) | 174±49 (80–419) |
| Cross-clamp time, min (range) | 119±36 (51–261) |
Continuous variables are expressed as mean±standard deviation. AF, atrial fibrillation; CPB, cardiopulmonary bypass; FA, femoral artery; MIMVR, minimally invasive mitral valve repair.
Three different surgeons were in charge of the surgical procedures in this series. The early cases (2006–12) were performed by M.K. (n=137), and the late cases (2012–17) were performed by T.S. (n=203) or T.T. (n=47).
Follow-up ProceduresFollow-up information was obtained by phone contact with the patient and/or family members, with supplementary information supplied by family physicians and referring cardiologists. Clinical follow-up examinations were completed in 364 patients (94.1%), with a mean follow-up period of 36.5±32.3 months. All patients underwent a postoperative echocardiographic evaluation prior to discharge, except for 1 who died during hospitalization. Late (>6 months) echocardiographic follow-up examinations were possible in 302 patients (78.0%), with a mean period of 35.7±31.8 months. MR grade was quantitatively determined according to the guidelines of the American College of Cardiology/American Heart Association 2006.11 When quantitative assessment of MR was not applicable, MR severity was determined in a semiquantitative manner using other parameters, such as vena contracta and MR jet area. Recurrent MR was defined as moderate or greater MR noted in any echocardiographic examination performed after surgery.
Statistical AnalysisContinuous data are expressed as the mean±standard deviation or median, with interquartile range (IQR) and categorical data shown as percentages. Cumulative survival was evaluated using the Kaplan-Meier method. All reported P-values are two-sided and if <0.05 were considered to indicate statistical significance. All data were analyzed using the Statistical Analysis Systems software package JMP 13.0 (SAS Institute Inc. Cary, NC, USA).
There was 1 (0.26%) operative death from stroke; 2 patients (0.5%) underwent conversion to a sternotomy because of bleeding from the left ventricle, possibly because of perforation by a left ventricular venting tube. Retrograde aortic dissection occurred in 1 patient. Other perioperative morbidities included reoperation for bleeding in 3 patients (0.8%), stroke in 5 (1.3%), permanent pacemaker implantation in 8 (2.1%), pulmonary complications with ventilation time exceeding 72 h in 2 (0.5%), and acute renal failure requiring dialysis, myocardial infarction (MI), and low-cardiac-output syndrome (LOS) requiring mechanical support in 1 patient each (0.3%). Exogenous transfusion was required in 14.7% of the present patients. The median mechanical ventilation time was 4 h (IQR 3.0–5.5) and 82% of the patients were extubated on the same day as the operation. Median intensive care unit (ICU) and hospital stays were 2 and 14 days, respectively. Postoperative outcomes are presented in detail in Table 3.
| In-hospital death, n (%) | 1 (0.3) |
| Stroke, n (%) | 5 (1.3) |
| Re-exploration for bleeding, n (%) | 3 (0.8) |
| Renal failure requiring dialysis, n (%) | 1 (0.3) |
| Myocardial infarction, n (%) | 1 (0.3) |
| Permanent pacemaker implantation, n (%) | 8 (2.1) |
| New onset of atrial fibrillation, n (%) | 34 (8.8) |
| Re-expansion pulmonary edema, n (%) | 4 (1.1) |
| Pneumonia, n (%) | 2 (0.5) |
| Ventilation (≥24 h), n (%) | 13 (3.4) |
| Prolonged ventilation (≥72 h), n (%) | 2 (0.5) |
| Wound infection, n (%) | 0 (0) |
| Limb ischemia, n (%) | 0 (0) |
| Time to extubation, h | 4 (3.0–5.5) |
| ICU stay, days | 2 (1.0–2.0) |
| Hospital stay, days | 14 (12.0–17.3) |
Continuous variables are expressed as median with interquartile range. ICU, intensive care unit; MIMVR, minimally invasive mitral valve repair.
A total of 7 patients (1.8%) required MV replacement after a repair attempt. Repair techniques performed included leaflet resection in 199 patients (with leaflet height reduction procedure in 41), leaflet plication in 81, chordal reconstruction in 109, and annuloplasty alone in 24; 5 patients underwent leaflet augmentation with an autologous pericardium. An annuloplasty ring was used in all cases except 2, of which 1 was a repair for HOCM and 1 was a repair for an endocardial cushion defect. The type of ring was selected according to the surgeon’s preference, with a flexible band used in 189 and a semi-rigid ring in 196 patients.
Late OutcomesDuring the mean follow-up period of 36.5±32.3 months, 8 patients died. Cardiac death occurred in 1 patient, and the other deaths were pneumonia in 2, stroke in 2, cancer in 1, and unknown cause in 2. The overall 5- and 7-year survival rates were 96.9% and 94.7%, respectively. Reoperation because of recurrent MR was required in 12 patients at 27.8±20.5 months after the initial repair, of whom 11 underwent a median sternotomy and 1 had a repeat thoracotomy. Of those who underwent reoperation, 11 had an MV replacement and 1 had a repeat repair procedure.
The 5- and 7-year rates of freedom from severe MR or reoperation in patients with primary MR were 94.7% and 93.5%, respectively (Figure 2), while those for freedom from moderate or greater MR were 82.2% and 75.4%, respectively (Figure 3). Patients with isolated anterior mitral leaflet (AML) prolapse showed a significantly higher rate of recurrent MR (P<0.0001) as compared with those with an isolated posterior mitral leaflet (PML) or who underwent bileaflet repair (Figure 4). The rate of freedom from severe MR or reoperation was significantly lower in the initial 30 cases (P=0.0014) as compared with the following 324 cases (Figure 5).
Freedom from severe mitral regurgitation (MR) or reoperation after repair of primary MR (n=354 patients).
Freedom from moderate or greater mitral regurgitation (MR) after repair of primary MR (n=354 patients).
Freedom from severe MR or reoperation stratified by mitral lesion after repair of primary MR (n=354 patients). Ant, anterior leaflet; Post, posterior leaflet; Bil, bileaflet.
Freedom from severe mitral regurgitation (MR) or reoperation for initial 30 vs. later 324 patients who underwent mitral valve repair for primary MR.
According to a recent analysis of data in the Japan Cardiovascular Database (JCVSD), MIMVR is gaining popularity in Japan.9 However, only a single clinical study performed in Japan has been published, in which Ito and colleagues reported excellent mid-term outcomes of totally endoscopic MIMVS procedures.8 The present is the second and largest report of a single institution’s experience with MIMVR in Japan, and includes analysis of the initial learning curve period as well as long-term echocardiographic follow-up results. Most procedures were performed under video-assisted direct vision, which currently is more popular than totally endoscopic MIMVR. The in-hospital mortality rate of our cohort was 0.26% and the rate of major complications, which included death, conversion to sternotomy, aortic dissection, re-exploration for bleeding, stroke, valve-related reoperation during hospitalization, new dialysis, MI, and LOS with necessity for mechanical support, was 2.8% (11/387).
Previous studies conducted by high-volume centers have raised several important issues regarding the steep learning curve for MIMVS. Murzi et al investigated their institutional learning curve in terms of perioperative morbidities, such as deaths within 30 days, conversion to sternotomy, MI, aortic dissection, stroke, and in-hospital reoperation for any cause. For 936 MIMVS procedures performed by 7 surgeons, the in-hospital mortality rate was 1.8%, while the overall complication rate (8.5%) consistently decreased from approximately 16% in the first 200 cases to <5% after 600 cases.12 The most frequent complication noted was reoperation for bleeding (4.7%), which accounted for 55% of all surgical morbidities.
Another important report was presented by the Leipzig group,13 in which a total of 3,907 MIMVS procedures performed by 17 surgeons were analyzed. Major adverse events (conversion to sternotomy, re-exploration for bleeding, valve-related reoperation during hospitalization, new dialysis, stroke, MI, LOS with necessity for mechanical support, death) were observed in >20% of patients at the beginning of the surgeon’s experience, but the rate then fell to approximately 10% after 250 operations. Re-exploration for bleeding also accounted for a large proportion of those cases, and the rate was reduced from 8.2% to 1.9% after 300 operations.
Similar results were reported in an analysis of the JCVSD, in which most of the contributing institutions performed <10 cases of MIMVR annually. The rate of re-exploration for bleeding was even higher in patients undergoing MIMVR than conventional MV repair (2.9% vs. 1.2%, respectively, P<0.05).9 From these previous studies, it can be concluded that bleeding plays a dominant role in the MIMVR learning curve.
In the present study, the overall major complication rate was lower than in the previously reported studies. Notably, re-exploration for bleeding was done in only 3 patients (0.8%) and each occurred within the first 15 cases of the surgeon’s experience. Several clinical studies have shown that the amount of perioperative bleeding in patients undergoing MIMVR is significantly less as compared with conventional MVS.1–4,9 Considering that the bleeding rate decreases with surgical experience, bleeding may be avoidable by careful inspection of the thoracic wall prior to closing the thoracotomy, as it is the most common bleeding site.
Stroke occurred in 5 of the present patients (1.3%), of whom 3 (0.8%) suffered from early (intraoperative) stroke. According to the ISMICS consensus statement published in 2010, the rate of stroke in patients undergoing MIMVS was significantly higher than with conventional MVS (2.1% vs. 1.2%, respectively).14 On the other hand, a recent meta-analysis showed a similar rate for the 2 approaches.15 Grossi et al noted that the only significant risk factor for stroke was use of retrograde perfusion in high-risk patients with a diseased aorta.16 In addition, the Columbia University group reported a very low stroke rate (0.26%) for their 573 MIMVS cases, which they attributed to their preference for cannulating the aorta in a central manner.17 Therefore, we aggressively use an alternative approach if there is any concern regarding retrograde perfusion. Among the present patients, 12% underwent central aortic cannulation or had axillary cannulation added. We believe that careful preoperative assessment of the aortic pathology and a flexible cannulation strategy are important to avoid this catastrophic complication.
Pulmonary complications occurred in 1.3% of the current patients, which included prolonged mechanical ventilation (>72 h) in 2 (0.5%) and re-expansion pulmonary edema (RPE) in 4 (1.1%). RPE especially has been gaining increased attention as a unique and critical complication of MIMVS performed through a right minithoracotomy.18,19 We previously reported that RPE occurred in 2.1% of patients who underwent MICS, which resulted in prolonged ventilation time, ICU stay and postoperative hospitalization, with preoperative use of steroid or immunosuppressant medication, and prolonged aortic cross-clamping time (≥156 min) identified as independent risk factors.20 Another study noted that 25% of patients showed radiographically evident RPE, which was independently associated with preoperative chronic obstructive pulmonary disease, pulmonary hypertension, right-ventricular dysfunction, and increased CPB time.21 Administration of dexamethasone after anesthesia induction and avoidance of complete collapse of the right lung by intermittent ventilation may be effective.19 Most importantly, the indication of this procedure for high-risk cases of RPE should be carefully considered.
Vascular injury by femoral cannulation is another concern, because it is also solely related to the procedure. Lamelas et al reported a low rate of vascular complications in their 2,400 MICS cases, with no aortic dissections, 2 cases of compartment syndrome, 2 of femoral arterial pseudoaneurysm, and 174 (6.7%) groin wound seromas.22 Although most vascular complications seen in MICS are not critical and can be managed relatively easily with few long-term complications, some have reported catastrophic cases. Jeanmart et al. noted that acute aortic complications occurred in 0.9% (9/978) of their early MIMVS cases, including 8 dissections and 1 perforation by guide-wire placement.23 Abdominal compartment syndrome requiring emergency surgical decompression has also been reported by several groups.24,25 We experienced 1 case of retrograde aortic dissection caused by guide-wire injury to the iliac artery. That surgery was aborted and the patient underwent an MV repair through a standard sternotomy 1 year later after conservative treatment. Since then, we routinely use fluoroscopy to guide both femoral arterial and venous cannulation. Monitoring the perfusion of the lower extremities using NIRS is also routinely performed at our institution to avoid the risk of leg ischemia. We think that all reasonable efforts should be taken to avoid vascular complications because they can negate the benefits of MICS.
In terms of repair durability, our overall long-term results are similar to those presented in previous studies:26–28 12 patients required reoperation for recurrent MR, of whom 10 (83.3%) underwent a neochord plasty for isolated AML prolapse, and showed a significantly higher incidence of recurrent MR than patients with PML or bileaflet prolapse (Figure 4). Previous studies have shown that an isolated AML repair is the least durable as compared with PML and bileaflet repair. David et al reported that their 12-year reoperation-free rate for AML repair was 88%, which was worse than for PML (96%) and bileaflet (94%) repair procedures.29 Tabata et al also noted that the rate of freedom from moderate or severe MR for AML repair in their cases was 63.5% after 10 years, which was also worse than for PML (81.1%) and bileaflet (83.6%) repair cases.30 Kawamoto et al presented similar results, with 10-year freedom from reoperation for AML and PML repair rates of 84.6% and 96.3%, respectively.31 Repair durability for isolated AML prolapse in the present series seemed to be somewhat inferior to that presented in those studies; therefore, we performed additional analyses and found that 5 of the 12 cases of reoperation were within the initial 30 cases, of which 12 were isolated AML repairs. There were no significant differences among the 3 surgeons for Kaplan-Meier freedom from recurrent MR, so we consider that our suboptimal initial results were related to the steep learning curve associated with the MICS approach. During the MICS procedure, poor mitral visualization is often encountered, which may have affected repair quality, especially when gaining experience with the procedure. Because an AML repair is complex and may be more vulnerable to inadequate MV exposure as compared with techniques for other lesions, using the MICS procedure for such a complex MV repair should be carefully considered, especially during the initial period of introducing MIMVR.
Study LimitationsFirst, this was a retrospective study and lacked a control group for appropriate comparisons. Also, we did not statistically compare our MIMVR outcomes with those for conventional MV repair, because the approach for selection changed during the period of the study and it was difficult to eliminate confounding factors. Despite these limitations, the strengths of this study are the quite conclusive findings, with complete clinical follow-up examinations performed for nearly 400 patients, including late echocardiographic follow-up findings. Our results reaffirm the notion that MICS is an excellent interventional technique for MV disease. It can be performed safely by any qualified institution, even during the learning curve period, provided that careful consideration is taken in terms of indications and proper strategies to avoid perioperative complications.
In conclusion, MIMVR is safe, with low rates of mortality and morbidity, and sufficient repair durability. Nevertheless, a rather steep learning curve exists in terms of repair durability, especially for the AML repair procedure.
