2019 年 83 巻 8 号 p. 1668-1673
Background: In 2018, Japan became the first country to have robotic cardiac surgery covered under the national health insurance. The number of patients undergoing robotic mitral valve (MV) repair has been estimated to increase remarkably, but no reports in Japan have yet described the outcomes of robotic MV repair. This study aimed to analyze the early clinical outcomes of patients undergoing totally endoscopic robotic MV repair (TERMVR) as a landmark national study for this procedure.
Methods and Results: A total of 213 patients (152 men; mean age, 55±11 years) underwent TERMVR during May 2014 to December 2018. Preoperative demographics, operative profiles, and postoperative outcomes, including follow-up echocardiography, were analyzed. Successful TERMVR was achieved in all patients. Operation, cardiopulmonary bypass, and aortic cross-clamp times were 192±49.8, 127±23.8, and 70.1±16.2 min, respectively. Intraoperative transfusion was performed in 20 patients (10%). There were no in-hospital deaths. All patients were alive during the median follow-up period of 255 days (interquartile range, 32.5–208 days). Freedom from recurrence of MR >grade 2+ was 97.3%, 95.0%, and 90.7% at 6, 12, and 24 months, respectively.
Conclusions: TERMVR is an effective and safe procedure with acceptable early postoperative outcomes.
Innovations in surgical approaches have led to the development of minimally invasive access to the mitral valve (MV). Surgical teams across the globe have reported successful outcomes in robotic MV repair; most of the reports are from the USA1–6 and the remainder are from Europe7–9 and East Asia.10,11 Observational studies have demonstrated favorable short- and long-term clinical outcomes of conventional median sternotomy or minimally invasive anterolateral thoracotomy. Consequently, robotic MV repair has been more widely used over the past decade.
In 2014, we initiated a robotic cardiac surgery program and pioneered this cutting-edge field of cardiac surgery in Japan. At NewHeart Watanabe Institute, robotic MV repair is performed via 4 ports and is a totally endoscopic surgery. All cardiac maneuvers are performed using a robot. In 2018, the Japanese government accepted endoscopic robotic MV repair to be covered under the national health insurance. Since then, totally endoscopic robotic MV repair (TERMVR) has become the most commonly used procedure for MV repair at our institution. Consequently, the number of patients undergoing robotic MV repair in Japan is estimated to increase. We expect that our report will be a landmark national study. We present the early and midterm clinical outcomes of patients undergoing TERMVR; we also evaluated their postoperative complications and long-term outcomes.
This study was approved by the Institutional Review Board of NewHeart Watanabe Institute, Japan. Written informed consent was given by all participants before the study. The da Vinci S and X Surgical Systems (Intuitive Surgical, Inc., Sunnyvale, CA, USA) were used to perform TERMVR via 4 ports as shown in Figure 1.
Port placement in totally endoscopic robotic mitral valve surgery using da Vinci X Surgical System.
MV insufficiency was measured using transthoracic echocardiography and MV repair was suggested for symptomatic patients or those with severe mitral regurgitation (MR). Preoperative examinations were performed to indicate robotic MV repair under the criteria for minimally invasive cardiac surgery in NewHeart Watanabe Institute, which has been described previously.12 There was no difference in the exclusion criteria between robotic surgery and right mini-thoracotomy approach. Patients with aortic valve disease were excluded. Computed tomography (CT) was performed to investigate anomalies of the great vessels, severe arteriosclerosis, and stenosis in the aorta, iliac artery, femoral artery, and great veins. Preoperative autologous blood donation was performed, and the amount of blood donated was based on the patient’s body weight and preoperative serous hematocrit concentration.
Surgical ProceduresPatients were placed in hemi-left side rotation under general anesthesia. Single-lung ventilation was performed. Cardiopulmonary bypass (CPB) was performed via the femoral artery, femoral vein, and jugular vein. Criteria for femoral artery cannulation performed at NewHeart Watanabe Institute have been published previously.12 After CPB was performed, the da Vinci Surgical System was docked to the patient. Robotic manipulation was initiated with pericardiotomy. Cardioplegia needle and aortic cross-clamp were endoscopically inserted. Cardiac arrest was achieved using cold cardioplegia, and the MV was approached using a right-side left atriotomy. Myocardial protection was achieved using intermittent antegrade cold cardioplegia administered every 25–30 min. Leaflet prolapse was confirmed using the saline regurgitation test. All repair techniques were performed only with the robotic arms without manual procedures. Leaflet-plasty was performed for the prolapsed lesion using standard MV repair techniques. After confirmation of the prolapsed lesion, the actual distance from the head of the papillary muscle to the MV leaflet was measured. The length of the neochordae was determined in order to obtain sufficient leaflet coaptation. After leaflet-plasty, ring annuloplasty on the valve annulus was performed with continuous suture using a V-LocTM barbed suture (Medtronic, Minneapolis, MN, USA). The ring annuloplasty technique has been presented previously by Watanabe et al.13 Concomitant procedures, such as the maze procedure and left atrium appendage closure, were performed after robotic MV repair. Tricuspid valve repair was performed after aortic declamping. CPB discontinuation and wound closure were similar to those achieved using conventional MV repair. The patients were routinely extubated on the operative day and/or the following day. Intercostal nerve block was administered to relieve postoperative pain. The patients started a fast-track recovery from the following day. Figure 2 shows a postoperative wound image. Echocardiography was performed 1 week after the operation, after which the patients were discharged. Postoperative anticoagulation therapy was prescribed for 3–6 months.
Postoperative wounds after totally endoscopic robotic MV repair.
The primary early outcomes of interest were early death (within 30 days of surgery) and major postoperative complications (bleeding requiring reoperation, acute renal failure requiring dialysis, low cardiac output syndrome requiring mechanical circulatory support, stroke, thoracotomy wound infection, and pneumonia). The primary long-term outcomes of interest were all-cause death as well as major adverse cardiac and non-cardiac events. Another outcome of interest was MV repair durability, which was assessed using echocardiography. To evaluate the long-term MV repair durability, clinical follow-up data were routinely obtained using transthoracic echocardiography 6 months after the operation.
Statistical AnalysisData were analyzed using JMP ver. 12 (SAS Institute Inc., NC., USA). Continuous values (presented as mean±standard deviation) were analyzed using Student’s independent sample t-test. Recurrence of MR was estimated using the Kaplan-Meier method; in this survival analysis, the patients were censored at the last available follow-up echocardiography. P≤0.05 was considered statistically significant for individual tests.
From May 2014 to December 2018, 214 patients were enrolled for robotic MV repair. Their characteristics are presented in Table 1. One patient had an asthma attack after administration of general anesthesia, so did not undergo robotic MV repair. Consequently, 213 patients (152 men, 61 women; mean age, 55.4±11.4 years) underwent TERMVR. From 2014 to May 2018, we used the da Vinci S Surgical System, and from June 2018 onward, we upgraded to the da Vinci X Surgical System. The da Vinci S and X Surgical Systems were used in 127 and 86 patients, respectively. All patients had MR >grade 3+. Intraoperative and postoperative data are presented in Table 2. The number of minimally invasive surgeries performed without robotic assistance was 206 and median sternotomy was performed in 3 patients during the same period. The only difference in the indication of minimally invasive surgery between with and without robotic assistance was financial. All 3 patients undergoing median sternotomy had a problem with single-lung ventilation.
| Age (years) | 55±11 |
| Male (%) | 152 (71) |
| Body mass index (kg/m2) | 22±2.9 |
| Comorbidities | |
| Hypertension (%) | 75 (35) |
| Dyslipidemia (%) | 30 (14) |
| Diabetes mellitus (%) | 15 (7) |
| Serum creatinine level (mg/dL) | 0.86±0.18 |
| Hemodialysis (%) | 0 |
| COPD (%) | 3 (1.4) |
| AF (%) | 32 (15) |
| History of CVA (%) | 3 (1.4) |
| NYHA 3 or 4 (%) | 0 |
| Previous cardiac surgery (%) | 0 |
| Echocardiographic data | |
| LVEF (%) | 69±5.4 |
| LVDd (mm) | 55±5.3 |
| LVDs (mm) | 33±5.1 |
| LAD (mm) | 45±75 |
| TR > grade 3+ (%) | 25 (12) |
| MR grade | |
| Moderate-to-severe (%) | 24 (11) |
| Severe (%) | 189 (89) |
| Etiology of MR | |
| Degernerative (%) | 207 (97) |
| Rheumatic (%) | 0 |
| Infective endocarditis (%) | 6 (2.8) |
| Congenital (%) | 2 (0.9) |
| MV prolapse involved | |
| Anterior leaflet (%) | 31 (15) |
| Posterior leaflet (%) | 125 (59) |
| Bileaflet (%) | 57 (27) |
Data given as n (%) or mean±SD. AF, atrial fibrillation; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; LAD, left atrial diameter; LVDd, left ventricular diastolic diameter; LVDs, left ventricular systolic diameter; LVEF, left ventricular ejection fraction; MR, mitral regurgitation; NYHA, New York Heart Association; TR, tricuspid.
| Variable | |
| Operation time (min) | 192±49.8 |
| CPB time (min) | 127±23.8 |
| ACC time (min) | 70.1±16.2 |
| Second aortic cross-clamp (%) | 1 (0.4) |
| Transfusion (%) | 20 (9.4) |
| MV procedures | |
| Ring annuloplasty (%) | 212 (99.5) |
| Posterior band (%) | 149 (70) |
| Complete ring (%) | 64 (30) |
| Quadrangular/triangular resection (%) | 32 (15) |
| Neochordae formation (%) | 167 (78.4) |
| Commissural repair (%) | 17 (8.0) |
| Augmentation (%) | 3 (1.4) |
| Cleft repair (%) | 21 (10) |
| Concomitant procedures | |
| MAZE procedure (%) | 29 (13.6) |
| TAP (%) | 22 (10.3) |
| ASD/PFO closure (%) | 29 (13.6) |
| LAA (%) | 23 (10.7) |
| MIDCAB (%) | 2 (0.9) |
| Lung resection (%) | 1 (0.4) |
| Thoracoplasty (%) | 2 (0.9) |
| Medinus tumor resection (%) | 1 (0.4) |
Data given as n (%) or mean±SD. ACC, aortic cross-clamp; ASD, atrial septum defect; CPB, cardiopulmonary bypass; LAA, left atrial appendage closure; MIDACB, minimally invasive direct coronary artery bypass grafting; PFO, patent foramen ovale; TAP, tricuspid annuloplasty.
Tricuspid valve repair, atrial fibrillation fixation, and atrial septum closure are the routine concomitant procedures performed at NewHeart Watanabe Institute. In 1 patient, a right lung tumor was suspected on the basis of preoperative CT, so partial lobectomy was performed. In 2 patients, concomitant left mammary artery harvesting and anastomosis with left minithoracotomy were performed. Thoracoplasty (Nuss procedure) was performed for severe pectus excavatum.
The operation, CPB, and aortic cross-clamping times were 192±49.8, 127±23.8, and 70.1±16.2 min, respectively. One patient developed moderate-to-severe residual MR during weaning of CPB, so a second cross-clamping was performed. The patients were weaned with no regurgitation. Intraoperative transfusion was performed in 20 patients (10%).
No patient died during the follow-up. Intra-aortic balloon pump (IABP) was postoperatively inserted in 3 patients because of low cardiac output syndrome; 3 patients underwent reoperation because of bleeding from the exploration; 3 patients had a postoperative cerebral infarction.
Follow-up transthoracic echocardiography data are presented in Table 3. The number of patients with MR grade 0–1 was 205 (96.2%), grade 2 was 6 (2.8%), grade 3 was 1 (0.4%), and grade 4 was 1 (0.4%). The patient with MR grade 3 caused by SAM underwent reoperation for replacement at another institute. At present, reoperation has been planned for the patient with MR grade 4. Figure 3 shows the Kaplan-Meier curves for freedom from recurrence of MR ≥grade 2+; it was 97.3%, 95.0%, and 90.7% at 6, 12, and 24 months, respectively.
| Early outcome and major morbidity | |
| Early death (<30 days) | 0 |
| LCOS requiring MCS (%) | 3 (1.4) |
| Early CVA (%) | 3 (1.4) |
| Bleeding requiring reoperation (%) | 3 (1.4) |
| New-onset dialysis (%) | 0 |
| Thoracotomy wound infection (%) | 7 (3.3) |
| Pneumonia (%) | 5 (2.3) |
| Prolonged ventilation (>48 h) (%) | 2 (0.9) |
| Groin wound lymphocele (%) | 7 (3.3) |
| Reoperation (within 30 days) | 0 |
| Extubation (h) | 7.8±10 |
| Postoperative ICU stay (days) | 1.9±1.1 |
| Postoperative hospital stay (days) | 10.8±5.1 |
| Echocardiographic data of postoperative MV regurgitation | |
| Grade 0–1 (%) | 205 (96.2) |
| Grade 2 (%) | 6 (2.8) |
| Grade 3 (%) | 1 (0.4) |
| Grade 4 (%) | 1 (0.4) |
Data given as n (%) or mean±SD. CVA, cerebrovascular accident; ICU, intensive care unit; LCOS, low cardiac output syndrome; MCS, mechanical circulatory support; MV, mitral valve.
Kaplan-Meier curve for freedom from mitral regurgitation greater than grade 2+.
MV repair, which results in excellent long-term outcomes, has become the standard of care for most patients with degenerative MV disease. The next step facing the surgical field is to minimize surgery and achieve quick recovery of the same quality. Several studies conducted worldwide have reported excellent results using minimally invasive cardiac surgery (MICS) via right minithoracotomy.14–17 Compared with conventional full sternotomy, MICS is considered less invasive, but is associated with various problems. The first and most serious problem is the visual field to the MV. Although thoracoscopes and other surgical instruments have improved, most surgeons depend on direct vision of the MV, which resulted in different operative visions in each case. In some patients, evaluation of regurgitation and manipulation of the repair were difficult. The second problem is the occurrence of rib fractures. Although MICS minimizes the wound size, rib fractures occur in some patients and these can lead to postoperative hemothorax. Furthermore, rib fractures and intercostal nerve compression prolong postoperative pain.
To solve these problems and perform the least invasive surgery, Go Watanabe and Norihiko Ishikawa aimed to perform totally endoscopic cardiac surgery using only ports to the cardiac field.18–20 This report presents the outcomes of TERMVR performed at a single institution. The advantages of robotic MV repair include 3D vision and fine dexterity. Robotic MV repair provides surgeons with better visualization of the valvular and subvalvular apparatus and precise intraoperative assessment. Surgeons can access the same surgical field in all patients, irrespective of the individual patient’s condition. Moreover, rib fractures and intercostal nerve compression are rare with the robotic MV approach. This had led to quick postoperative recovery in our patients. Furthermore, all surgical team members can observe the operation, which is a considerable advantage for young surgeons to observe and learn how to operate the valve. Dual-console systems used in robotic MV repair allow real interaction, thereby enhancing the flow of knowledge regarding the technique and procedure from an experienced surgeon to a young surgeon. It is freedom of choice whether the surgeon selects the sternotomy approach, right thoracotomy approach, or robotic approach. However, the principle of repairing the valve is the same, and that observation can be useful for young surgeons doing their mitral repairs in the future.
Robotic MV repair is similar to other approaches to MV repair performed using right minithoracotomy. Our main procedure for TERMVR, which was performed in 78.6% of the patients, was neochordae formation using Gore-Tex suture. Compared with resection technique, neochordae formation offers advantages such as reproducibility and can minimize the need for subvalvular procedures. For each prolapsed lesion, we attached >2 neochordae to the leaflet to enhance the strength of the neochordae. Only when the posterior leaflet was redundant and SAM was suspected, were resection and suture technique performed. We have also reported the usefulness of folding-plasty for patients with prolapsed posterior leaflets.21 Ring annuloplasty was performed in most patients to deepen the coaptation of the anterior and posterior leaflets. Ring annuloplasty with continuous suture using the V-Loc barbed suture was performed on the valve annulus.13 Our first choice at present is partial ring annuloplasty. From the technical point of view of annuloplasty, we consider there are 2 major advantages for the partial ring annuloplasty. One is that it can be more quickly performed than the complete ring. Another advantage is that there is no possibility of damaging the aortic valve annulus, because there is no necessity to suture the ring to the anterior leaflet annulus.
Among the concerns for adapting robotic MV repair, safety is the most essential one. Unless robotic MV repair can provide the same level of safety as conventional MV repair, the efficacy and durability of robotic MV repair are not worth considering.10 Successful outcomes of robotic MV repair performed at NewHeart Watanabe Institute were achieved via patient selection, and not all patients are suitable for this approach. Patients with significant carotid or peripheral vascular disease were screened using CT. Coronary artery stenosis >75% in a patient was a contraindication for robotic MV repair. We have been extremely cautious regarding the occurrence of cerebral complications, and we had 3 symptomatic patients after the surgery; although this number is small, future studies should aim to reduce the occurrence of cerebral complications.
A low rate of intraoperative blood transfusion (10%) was achieved in the present results, which may be attributed to sternal preservation as well as the short operation and CPB times. Furthermore, preoperative autologous blood donation may have also contributed to the result.
Concomitant procedures were planned and performed when required. In the first series, there were 2 patients who underwent minimally invasive direct coronary artery bypass (operation time, 300 and 360 min each). Both patients had a long ventilation time, so concomitant coronary artery bypass grafting with robotic MV repair was discontinued to enable quick recovery. Tricuspid valve-plasty was performed in patients with greater than moderate regurgitation. Left appendage closure and the maze procedure were performed when permanent or paroxysmal atrial fibrillation was diagnosed.
There are certain concerns associated with introducing endoscopic surgery in the cardiac region. One concern is de-airing, which causes air obstruction in the coronary artery during weaning of CPB. The Trendelenburg position or direct manipulation of the heart cannot be achieved because the operating table must not be moved when the da Vinci Surgical System is docked. We had 2 and 1 patient with intraoperative and postoperative IABP insertion, respectively, because of low output. We minimized the frequency of performing the serum regurgitation test to avoid air obstruction in the coronary artery. Moreover, we performed ECG to detect ST-elevation during weaning of CPB.
Another concern is bleeding, particularly from the ports. Bleeding from the ports does not cause early hypovolemic shock, but postoperative hemothorax is a major complication and can lead to anemia and pneumonia. We worked to ensure that port bleeding did not occur, but 3 patients underwent reoperation because of port bleeding. Surgeons must be careful about this complication and always be aware of converting to sternotomy before the result becomes catastrophic.
Many Japanese patients have a relatively small body, small body surface area and low body mass index, compared with patients from the USA and Europe. These Asian patients have small vessels, which is a risk factor for minimally invasive surgery and may cause discontinuation of CPB or leg ischemia. For the establishment for CPB, our first choice for arterial cannulation is the right femoral artery. In cases of patients with small vessels, bilateral femoral arteries or additional cannulation of the subclavian artery is performed. However, we have no contraindications for the thoracic or cardiac procedure. All cases were performed using the same procedure, regardless of the patient’s body size. We assume that this led to the safe and acceptable outcomes of TERMVR.
We also had 1 patient who underwent reoperation for recurrence of MR, and it was difficult to identify whether the relapse was secondary to any technique. During the follow-up, no dehiscence of the ring annuloplasty was reported. Based on these results, continuous suture using the V-Loc barbed suture can be considered a useful technique for ring annuloplasty on the valve annulus.
An obvious learning curve is associated with robotic MV repair. The lack of tactile feedback can potentially limit a surgeon’s ability to assess suture depth and tension, resulting in repair failure or bleeding. We have reported a special technique to control aortic bleeding.22
Japan became the first country in which robotic cardiac surgery was covered under the national health insurance. Further, the number of patients undergoing robotic MV repair is estimated to increase. Figure 4 presents the transition in the number of TERMVR performed over the years at NewHeart Watanabe Institute. In the present study, safe and acceptable outcomes of TERMVR were achieved, and we believe that this report is a landmark national study for this procedure. In the future, we aim to perform safe, totally endoscopic robotic surgery not only for MV repair but also for valve replacement and aortic surgery with minimal complications; furthermore, we will also consider the cost benefit.
Transition of the number of cases of totally endoscopic robotic mitral valve surgery at NewHeart Watanabe Institute.
This study was based on medical data collected and arranged by the NewHeart Watanabe Institute. Therefore, this study was possible by virtue of the work of all the staff in the institute.
There are no conflicts of interest for any authors and there was no significant financial support for this work that could have influenced its outcome.
