Minimally invasive beating heart technique for mitral valve surgery in patients with previous sternotomy and giant left ventricle

This study was approved by the Research Ethics Committee of the First Affiliated Hospital of Zhengzhou University. Eighty cardiac patients with previous sternotomy and giant left ventricle according to the diagnostic criteria that LVEDD was ≥70 mm, who underwent mitral valve surgery at our center from January 2006 to January 2019 were analyzed. If the patient had significant aortic regurgitation (the effective regurgitate orifice area greater than 1cm²) and (or) reoperation for immediate or early surgical failures (same hospital admission or less than 30 days) was excluded. From August 2011 to January 2019, 30 patients with giant left ventricle were reoperated through a minimally invasive approach in the right fourth intercostal space without aortic cross-clamping (minimally invasive beating heart group). From January 2006 to January 2019, 50 patients with giant left ventricle were reoperated through conventional median sternotomy with aortic cross-clamping and cardioplegic arrest (median resternotomy arrested heart group). Detailed preoperative demographics and patient characteristics are shown in Table 1.

Under general anesthesia with a double lumen endotracheal intubation, patients were positioned with the right side of the chest slightly elevated. Transesophageal echocardiography (TEE) was performed and external defibrillation pads were placed in all patients. Sites of arterial and venous cannulation are described in Table 2. We preferred to use peripheral cannulation through the right internal jugular vein, right femoral vein, and right femoral artery after systemic heparinization, under TEE guidance. Especially the right internal jugular vein was percutaneously cannulated to better improve venous drainage for patients additionally undergoing tricuspid valve surgery. If only implemented the single femoral venous cannula that was inserted well into the superior vena cava (SVC) in order to avoid dislocation of the tip out of the SVC after left atrial retraction. Cardiopulmonary bypass (CPB) was established with vacuum assisted venous drainage with 30 mmHg negative pressure followed by a right anterolateral minithoracotomy through a 4 to 6 cm incision on the anterior axillary line via the right fourth intercostal space. The incision should be avoid as far as possible the breast for female patients. It had better to make a curved incision from the lower edge of the breast to protect the breast tissue. A rib spreader (Geister, Germany) was placed after a soft tissue retractor was inserted the incision. Then made two small incisions (1 cm in length) in the second and sixth intercostal space on the midaxillary line respectively. Carbon dioxide through the incision in the second intercostal space at 5 l per minute was continuously insufflated into the chest throughout the procedure to displace intracardiac air. A left ventricular drainage tube was placed through the incision in the sixth intercostal space. Under naked eye euthyphoria sheared longitudinally pericardium along anterior 2 cm of the right phrenic nerve, fixed anterior pericardium on the edge of the incision by silk thread and pulled posterior pericardium to posterolateral chest wall through two mininal incisions by suture. The aorta was not cannulated or clamped. At the beginning of experience, the patient’s body temperature was cooled to 27 or 28 °C to induce ventricular fibrillation. Subsequently, temperature was maintained between 32 and 33 °C to allow an operation on the empty beating heart [5, 7]. An aortic vent was always under continuous suction in the ascending aorta to evacuate air. With the heart perfused and beating, the left atrium was then immediately opened through interatrial groove or a trans-septal incision after a longitudinal right atriotomy in case of tricuspid surgery. For tricuspid surgery, dissociated the SVC and inferior vena cava (IVC) and completely blocked vena cava. In particular, we used 4–0 Prolene line with gasket to do the internal purse suture at SVC and IVC to right atrium opening that completely blocked vena cava, which avoided secondary injury of free vena cava because of fatal adhesion. An atrial retractor was placed through the fifth or sixth intercostal space parasternally and the left atriotomy was retracted anteriorly as a mass retraction. A left ventricular drainage tube was inserted right pulmonary vein that was used to maintain a clear operative field. The mitral valvuloplasty (MVP) or mitral valve replacement (MVR) was performed with interrupted pledgetted 2–0 braided polyester mattress sutures under direct vision. The tricuspid valvuloplasty (TVP) was performed by using the ring annuloplasty method. For coexisting paroxysmal or persistent atrial fibrillation, we preferred to use monopolar radiofrequency ablation maze approach that was easier in the minimally invasive setting. To prevent possible left atrial air embolism, the left atrium was filled with backflow of blood keeping the prosthetic or native valve open before closing the atriotomy. TEE was used to confirm proper valve and ventricular function and to ensure complete removal of air. The patient was disconnected from CPB after appropriate reperfusion and was decannulated. Placed pericardial drainage and chest drainage tubes, closed the femoral longitudinal incision and the thoracotomy in a standard fashion.

Continuous variables are expressed as the mean ± standard deviation or median and the range of values (if non-normally distributed) and categoric variables are presented as proportions. Differences between groups were assessed by using the Pearson’s chi-square test or two-tailed p-value Fisher exact test where appropriate for categoric variables, the unpaired two-tailed t test for continuous variables (Mann-Whitney U test for non-normally distributed continuous variables). All statistical analyses were performed by using the IBM SPSS software package (version 20.0; SPSS, Chicago, IL, USA). A value of p < 0.05 was considered statistically significant.

The patient demographics and preoperative characteristics in both groups were similar and not statistically significant differences (Table 1). The patients for redo mitral valve surgery tended to have older age, most of them were male and had a history of smoking. Previous cardiac surgeries were primarily mitral valve related diseases operations. The reoperation reasons were mainly mitral valve perivalvular leakage, mitral valve prosthesis dysfunction, rheumatic mitral valve disease, infective endocarditis, mitral valve perivalvular tissue hyperplasia, mitral bioprosthetic valve decay. The patients with giant left ventricle tended to have relatively low left ventricular ejection fraction (LVEF) (0.43 ± 0.08 vs 0.45 ± 0.09), New York Heart Association (NYHA) functional class above II accounted for the most part (66.7% vs 64.0%). Namely, these patients were more likely to have significant comorbidities at the time of surgery, for example, congestive heart failure, hypertension, chronic obstructive pulmonary disease (COPD), diabetes mellitus, pulmonary hypertension, atrial fibrillation, cerebrovascular disease and moderate or greater tricuspid regurgitation (TR), therefore mean preoperative logistic EuroSCORE predicted risk of operative mortality was quite high (15.3% ± 5.4% vs 14.8% ± 5.4%).

Intraoperatively, the operation times (3.75 ± 0.45 vs 4.17 ± 0.62 h, P = 0.002) and CPB times (130.2 ± 19.3 vs 165.2 ± 27.6 min, P < 0.001) were shorter in the minimally invasive beating heart group, blood loss (515 ± 188 vs 697 ± 222 ml, P < 0.001) was obviously decreased compared with median resternotomy arrested heart group (Table 2). There were 1 and 4 patients who required an intra-aortic balloon pump (IABP) to wean from extra-corporeal circulation in either group. One patient in the minimally invasive beating heart group underwent conversion to sternotomy because of extensive adhesions on the chest wall so that unable to single lung ventilation, at which time an aortic cross-clamp was placed. There were 1 and 4 patients who needed to re-exploration for bleeding in either group, but no statistically significance (3.3% vs 8.0%, P = 0.72). Arterial cannulation and myocardial protection strategies differed between the two groups. In the minimally invasive beating heart group, femoral arterial cannulation was almost used in 29 patients (96.7%), with the rest one undergoing ascending aortic cannulation (3.3%). On the contrary, in the median resternotomy arrested heart group, ascending aortic cannulation was used in 88.0% of patients (n = 44), and 6 patients (12%) underwent femoral arterial cannulation. Myocardial protection was achieved with ventricular fibrillation or superficial hypothermic empty beating heart (32.1 ± 0.46 °C core temperature) in beating heart surgery and performed with transthoracic aortic cross-clamp and cold crystalloid cardioplegia (29.0 ± 1.47 °C core temperature) in arrested heart. The distribution of valvular procedures performed and concomitant operative procedures is presented in Table 2. The most of patients for redo mitral valve surgery were performed MVR (86.7% vs 86.0%, P = 0.93). There were 8 and 14 patients who had concomitant moderate or greater TR in either group were all performed TVP by using the ring annuloplasty method. For coexisting paroxysmal or persistent atrial fibrillation, 7 patients were used monopolar radiofrequency ablation improved maze approach in the minimally invasive setting, however, 4 patients were used monopolar and 11 patients were used bipolar radiofrequency ablation improved maze approach in the conventional median resternotomy.

The early 30-day mortality was lower in the minimally invasive beating heart group, although not significant difference (6.7% [n = 2] vs 14.0% [n = 7], P = 0.52) (Table 3). Causes of death were multiple organ dysfunction syndrome (MODS)(n = 2) and coexist low cardiac output syndrome (LCOS)(n = 1) in the minimally invasive beating heart group. The postoperative complications in the minimally invasive beating heart group were all lower than that in the median resternotomy arrested heart group, such as stroke (0% vs 6.0%), acute renal failure (6.7% vs 14.0%), liver dysfunction (6.7% vs 10.0%), pulmonary complications (6.7% vs 12.0%), MODS (6.7% vs 14.0%), LCOS (3.3% vs 10.0%), ventricular fibrillation (6.7% vs 14.0%), cardiac tamponade (3.3% vs 6.0%), cardiac arrest (0% vs 4.0%), atrioventricular block (3.3% vs 8.0%), superficial wound infection (0% vs 4.0%), while there were no statistical significances. There was no statistical difference between the two groups in postoperative atrial fibrillation after radiofrequency ablation (36.7% vs 36.0%, P = 0.95). Fortunately, minimally invasive beating heart technique compared to the conventional median resternotomy arrested heart technique formitral valve surgery in the cardiac patients with previous sternotomy and giant left ventricle had significant differences in postoperative transfusion ratio (66.7% vs 90.0%, P = 0.01), median postoperative transfusion amount (2.0 vs 5.0 U, P < 0.001), median postoperative drainage volume (450 vs 800 ml, P = 0.001), median extubation time (13 vs 17 h, P = 0.04), median intensive care unit (ICU) stay time (16.5 vs 24.5 h, P = 0.04) and median postoperative hospital stay time (6 vs 9 days, P < 0.001). With patients’ agreement, we obtained the 6 months postoperative echocardiograms, which were available in all patients through consulting related cases data and telephone follow-up. The 6 months postoperative echocardiographic parameters (LVEDD, LVEF, cardiothoracic ratio, NYHA functional class) had a marked improvement compared with the preoperative circumstances, but there were no statistical significance between the two groups.