Background

Right lateral minithoracotomy(RT)has become the standard approach for mitral valve surgery in many centers and is considered to be minimally invasive mitral valve surgery (MIMVS). There is still ongoing debate about the benefits of minimally invasive interventions [1, 2].

Although these benefits, criticisms have been raised as MIMVS is technically more complex, requires a distinct learning curve [3]. Furthermore, the minimally invasive access is technically more demanding and is usually limited to centers with extensive expertise [4, 5]. Patients with left ventricular dysfunction, advanced age, poorer New York Heart Association functional class, larger body mass index, carotid disease, and tricuspid valve regurgitation are more likely to undergo a median sternotomy [6, 7]. Therefore, basic characteristics of patients selected for median sternotomy and minimally invasive access are usually not comparable, and prospective randomized trials are not available. Since 2011, we started our MIMVS program and after few years, RT approach has become the standard approach for the treatment of isolated mitral valve disease. The aim of our study is to report early and mid-term outcomes of consecutive patients who had undergone mitral valve surgery using RT during a 6-year period.

Methods

Patient selection and data collection

From January 2011 to January 2017, 1120 patients underwent primary isolated mitral valve surgery. Minimally invasive surgery was performed in 283 (25.3%) patients, and conventional full sternotomy was performed in 837 (74.7%) patients. The present study was a retrospective, observational review. The institutional review board approved the study and waived patient consent. Patient selection was predicated by surgeon preference. The patient demographics, medical history, and operative and in-hospital outcomes were collected at each patient’s admission and at each consecutive follow-up visit. Mid-term mortality was documented, with a review of the data collected from routine clinic follow-up visits, in addition to a query of the Social Security Death Index. Propensity scores were then generated based on the regression model and matched pairs created using 1:1 nearest neighbor matching. There were 165 matched pairs in the analysis for a total sample size of 1120(sternotomy group, n = 165; minimally invasive group, n = 165).

Exclusion criteria

Patients with concomitant coronary artery bypass graft surgery, aortic or tricuspid valve procedures, and surgical ablation of atrial fibrillation were excluded from the study. Studies including mainly redo surgical procedures were excluded. The cause of the mitral valve disease was not taken into consideration for inclusion or exclusion of the studies.

Definitions

Hospital mortality included all deaths within 30 days of operation irrespective of where the death occurred and all deaths in hospital after 30 days among patients who had not been discharged after the index operation. The diagnosis of stroke was also confirmed by computed tomography or magnetic resonance imaging whenever possible and documented by staff neurologists. Renal complications included acute renal failure, defined as the requirement for hemodialysis or an increased creatinine level (> 200 mmol/L). A diagnosis of postoperative myocardial infarction was based on the presence of new Q waves greater than 0.04 milliseconds and/or a reduction in R waves greater than 25% in at least 2 contiguous leads on an electrocardiogram. Pulmonary complications included chest infection, ventilation failure, reintubation, and tracheostomy.

Surgical technique

Briefly, MIMVS by a way of right anterior thoracotomy was performed through a 5–7 cm skin incision placed at 4th intercostal space. After incision a soft tissue retractor is inserted and the intercostal space is gently spread with a retractor. Two trocars are inserted in the thorax to allow positioning of a ventricular vent, CO2 insufflator, camera device and pericardial stay sutures. Cardiopulmonary bypass was initiated by way of the femoral artery and vein cannulation through a small transverse incision in the groin,and direct transthoracic aortic clamping. QuickDraw Venous Cannula single stage (Edwards Lifesciences) were inserted through the femoral vein into the right atrium and the correct position was achieved with the Seldinger technique under transesophageal echocardiographic guidance. Direct ascending aorta cannulation is performed under direct vision. After vacuum-assisted cardiopulmonary bypass (− 40 to − 60 mmHg) was established, the patients were cooled to 34 C°. The lungs must be deflated before aortic cannulation. The aorta was crossclamped using the Chitwood aortic clamp (Cardiomedical GmbH, Langenhagen, Germany) directly through the thoracotomy incision, and antegrade cold crystalloid blood cardioplegia is delivered directly into the ascending aorta by a needle vent catheter. The mitral valve is approached with a traditional left paraseptal atriotomy and exposed using a specially designed atrial retractor held by a mechanical harm inserted through a right parasternal port. Mitral valve procedures were performed under a combination of direct vision and thoracoscopic assistance. All patients received an accurate intraoperative transoesophageal echocardiogram before and after weaning from cardiopulmonary bypass machine.

Full sternotomy

Standard operative technique was a median sternotomy and cardiopulmonary bypass using aortic and bicaval cannulation. Cardiac arrest was induced by the instillation of antegrade cardioplegia. The operative strategy was individualized but aimed towards curative resection.

Follow-up

Complete follow-up could be achieved in 92.1% (mean follow-up 4.4 ± 1.2 years). Information was collected from patient’s follow-up visits, telephone interviews with the patient or the referring physician, and mailed questionnaires. Postoperative complications were analyzed according to the “Guidelines for Reporting Morbidity and Mortality after Cardiac Valvular Operations,” approved by The Society of Thoracic Surgeons.

Statistical analysis

Continuous data were expressed as mean ± standard deviation or median with the interquartile range and categorical data as percentages. Cumulative survival was evaluated with the Kaplan–Meier method. All reported P values are two-sided, and P values of < 0.05 were considered to indicate statistical significance. All statistical analyses were performed with SPSS 22.0 (SPSS, Inc., Chicago, IL, USA). All statistical analyses were performed with the assistance of a departmental statistician.

A propensity score, indicating the predicted probability of receiving MIMVS treatment, was then calculated by the use of a non-parsimonious multiple logistic regression analysis from the logistic equation for each patient. Finally, we used the propensity score to match MIMVS to Sternotomy patients (1:1 match). Results are reported as percentage and odds ratios (ORs) and 95% confidence intervals. The propensity score included the following variables: age, body surface area, the ratio of rheumatic cause, preoperative ejection fraction (EF), sex, CPB time, New York Heart Association (NYHA) functional class (I, II versus III, IV), diabetes mellitus and hypertension.

Multiple variable models were constructed to determine independent factors influencing the following outcomes: postoperative blood transfusion, reoperation for early postoperative hemorrhage, length of postoperative ventilation support, and length of hospital stay after surgery.

Results

Patient characteristics

Table 1 compares patient demographics between the two surgical approaches. Although the patient ages and pulmonary artery pressures were similar, the preoperative New York Heart Association (NYHA) functional classification was worse in the group undergoing the sternotomy approach (2.6 ± 0.6 vs 2.1 ± 0.5, P < .001). After propensity score matching there were 165 matched pairs of patients (Table 1). Once matched, there were no longer significant differences among major baseline characteristics between groups, including gender, age, the ratio of rheumatic cause, cerebral infarction, pre-operative left ventricular ejection function, and so on.

Table 1 Characteristics of Unmatched and Propensity Matched Patients
Full size table

Operative and postoperative data

Several differences in operative and postoperative variables were identified on univariate analysis (Table 2). Minimally invasive patients had significantly longer cross-clamp time (65 ± 13 versus 49 ± 11 min, p<0.001) and bypass time (99 ± 25 versus 88 ± 17 min, p<0.001). No aortic dissections or injury occurred in either patient group. Patients undergoing minimally invasive surgery spent less time on ventilation support (6.2 ± 1.1 h vs 10.4 ± 2.7, p < 0.001).

Table 2 Perioperative Data of matched pairs
Full size table

Reoperation for bleeding was similar in matched groups (0.6%in MIMVR group vs 2.4%in the conventional sternotomy group, (P = 0.367; Table 2). However, Chest tube drainage was lower at 6 and 24 h after a minimally invasive approach (30 ± 5 mL) and 120 ± 20 mL than after conventional sternotomy 175 ± 50 mL and 400 ± 150 mL at these times (p < 0.001). Transfusion was less frequent after minimally invasive surgery than after conventional surgery (15.7% vs 40.6%, p<0.001). In-hospital complications are summarized in Table 2. Overall, no significant differences existed among major in-hospital complications between groups. There were 6 sternal wound infections (0.79%) among patients in the ST group. There was no significant difference in survival at 30 days between groups (p = 1.0). The hospital mortality was 1.2% for the sternotomy and 0.6% for the minimally invasive approach. Permanent neurologic perioperative events occurred in 1.2% of patients undergoing sternotomy and 1.8% of the patients undergoing the minimally invasive approach.

The multivariable regression analysis showed the full sternotomy was an independent risk factor for the propensity-adjusted likelihood of postoperative transfusion, re-exploration for bleeding, and postoperative ventilatory support (p < 0.05). Another independent predictor of length of hospital stay was ejection fraction (P = 0.006; Table 3). But the duration of cardiopulmonary bypass time was not an independent risk factor.

Table 3 Multivariate Analysis
Full size table

Follow-up

Late death occurred in 10 patients (2 cardiac-related deaths, and 8 non–cardiac related deaths).The mean New York Heart Association class at follow-up was 1.5 ± 0.6. There was 1 late re-intervention at 1.1 year for mitral prosthesis endocarditis (Table 4). Survival at 1, 3, and 5 years was 98.1% ± 0.9%, 93.5% ± 2.7%, and 91.7% ± 3.8%, respectively Among matched patients, survival at 1, 3, and 5 years was 99.3 ± 0.7%, 97.3 ± 1.5%, and 92.1 ± 2.6% after MIMVS surgery and 98.7 ± 1.0%, 96.2 ± 1.7%, and 92.5 ± 2.7% after conventional sternotomy. However, comparison of survival curves between the two groups revealed no significant difference (P = 0.203, Fig. 1). With regard to freedom from valve-related morbidity, there was no significant difference between groups (P = 0 .574, Fig. 2).

Table 4 Follow-up Results of Propensity Matched Patients
Full size table
Fig. 1
figure 1

Kaplan-Meier analysis of long-term survival.(Blue line = minimally invasive; green line = sternotomy

Full size image
Fig. 2
figure 2

Kaplan-Meier estimate of freedom from mitral valve-related morbidity of patients with either a right minithoracotomy (green line) or a median sternotomy (blue line)

Full size image

Discussion

Over the past decade, the field of MIMVS cardiothoracic surgery has seen rapid growth, and Mitral valve surgery has been particularly amenable to minimal access approaches, and the described benefits in the literature include improved patient satisfaction, improved cosmesis, decreased length of hospitalization, and decreased overall resource utilization [8,9,10]. Unfortunately, previous studies have been limited either by descriptions of MIMVS outcomes with no reference group for comparison, or a sternotomy reference group with significant differences in baseline risk [11,12,13]. To address these challenges, we used a propensity-matched comparison technique to simulate outcomes after pseudo randomization of patients to Sternotomy versus MIMVS approach for mitral valve surgery.

Bleeding was minimized, but this did not translate into less frequent surgical re-exploration or lower transfusion requirements. However, previous series [14,15,16], including our propensity- matched comparison of mitral valve procedures with a larger series of patients (n = 165) receiving isolated, less invasive valve surgery, have shown benefits compared with full sternotomy. The reasons for reduced bleeding are most likely associated with a smaller incision and less dissection of tissue and thus a smaller wound surface. Diffuse bleeding should therefore be less. And serratus anterior muscle was blunt dissected and only the intercostal muscle should be cut off, without disruption of the integrity of the thorax, so that the surgery was with little trauma, better for postoperative recovery. Using propensity adjustment, we were able to demonstrate an advantage of a MIMVS approach in diminishing postoperative blood transfusion, reoperation for hemorrhage, or postoperative length of stay.

Not surprisingly, sternal wound infections were significantly decreased with mini-MVS versus conventional MVS (0% vs 3.6%). This is a result that is consistent with expectations, as the mini-MVS approach was by definition via a thoracotomy rather than via sternotomy, whereas the conv-MVS approach was by definition only via median sternotomy in this meta-analysis [17, 18].

However, despite the potential benefits of mini-MVS and the results of our study to suggest the mid-term durability to be maintained using these techniques, mini-MVS has potential drawbacks that still need to be addressed. An increased risk of stroke, aortic dissection, and groin complications and increased crossclamp and cardiopulmonary times have all been mentioned as being of greater risk when performing mini-MVS versus conventional techniques [19, 20]. In our particular series, such complications have remained low. The current study demonstrates that despite significantly longer cross-clamp and bypass times, the early outcomes of MIMVS are similar to those of an open approach through median sternotomy.

Similar results were described by the Society of Thoracic Surgeons of the adult cardiac surgery database as well as by several meta-analyses confirming the main points of the aforementioned consensus statement [5, 6, 21]. Our results are in line with the current literature; however, despite these excellent outcomes, many criticism still remain regarding MIMVS as it is technically more complex, requires a distinct learning curve (prolonged cross-clamp and cardiopulmonary bypass times). Finally, for many surgeons, the decision to utilize MIMVS is more related to the cosmetic results than better clinical outcomes, because no large randomized trial has been performed.

Despite these differences, in both the current series and others [22, 23], early mortality was not increased in the port group. With regard to survival, there was no difference in 1-year or 3-year mortality between groups, and both MIMVS and ST patients had excellent short-term survival. Mid-term survival was also similar between groups, with both groups achieving survival rates above 92% at 5 years after surgery. Thus, in our series, a minimal access approach for mitral valve surgery does not appear to compromise morbidity or mortality when compared with matched ST controls.

Limitations

The main limitation of this study is its retrospective nature. This is also a single-institution study, which limits its generalizability. Within this single-institution experience, we acknowledge that selection bias cannot be completely reversed by propensity-based methods and in this study cannot completely overcome distinct surgeon preferences. Whether a larger series of patients with more power would have shown more benefits is unknown.

Conclusions

In this series, we demonstrate through propensity matching that a MIMVS approach for mitral valve surgery is associated with slightly increased CPB times and cross-clamp times when compared with an ST approach. Moreover, we show that a MIMVS approach is associated with equivalent rates of morbidity and mortality with the benefit of decreased postoperative blood transfusion, reoperation for hemorrhage, or postoperative length of stay, and reduced assisted ventilation duration. In addition to improved cosmetic results, MIMVS provides equally durable midterm results as the standard sternotomy approach. In conclusion, MI mitral valve surgery represents a safe and effective surgical technique that we believe should be used more routinely in the surgical management of mitral valve disease. However, widespread acceptance of this technique requires further advances in technique and proof of benefits in propensity-matched patients, because a randomized trial appears unlikely.