Early and long-term results of minimally invasive mitral valve surgery through a right mini-thoracotomy approach: a retrospective propensity-score matched analysis

Study population

During the period from 2000 to 2016, a total of 669 isolated mitral valve procedures for isolated mitral valve regurgitation were recorded, among them 317 (47.4%) with a minimally invasive approach through right mini-thoracotomy. Degenerative mitral valve disease was the cause of mitral insufficiency in our matched cohort. The patients were divided into two groups: those who underwent mini-MVS and those who underwent conv-MVS via standard sternotomy; 227 pairs were generated.

The matched groups were similar with regard to selected preoperative co-morbidities, and the haemodynamic and demographic categories. All patients undergoing concomitant procedures were excluded. Furthermore, we considered the causes of mitral valve regurgitation; consequently, there was no statistical difference between the cohorts. Postoperative outcomes and major complications such as death, stroke, wound infection, reoperation for bleeding, re-intubation, blood product use, renal failure, atrial fibrillation, hospital and ICU lengths of stay, and survival were analyzed. This study was approved by the local ethics committee of the University of Heidelberg, medical faculty (approval no.: S-516/2016). Furthermore, this study follows the ICMJE Privacy and confidentiality guidelines. The participant consent and follow-up data were obtained by letter correspondence and clinical visits.

Definitions

Early mortality was defined as death within 30 days after surgery or before discharge from the hospital. Postoperative respiratory failure was defined as re-intubation or tracheotomy. Postoperative stroke was defined as a new and permanent neurological disability or deficit. Postoperative temporary renal failure was defined as new necessity for dialysis during the hospital stay and as permanent renal failure if dialysis was necessary at the time of discharge. Gastrointestinal complications included gastrointestinal bleeding and visceral ischemia. Cardiac decompensation included New York Heart Classification classes III and IV.

Statistical analysis

For comparison, the propensity score-matched groups were generated based on variables including demographics, co-morbidities and haemodynamic values. The patients were matched using a 1:1 nearest-neighbour algorithm. The statistical analyses were done using IBM SPSS, version 24 (IBM Corp., Armonk, NY, USA). Categorical data are described as percentages and continuous variables are presented as mean ± standard deviation or the median with the 25th and 75th quartiles. Statistical significance was defined as p ≤ 0.05. The mean values were compared using t-Student test for normally distributed variables and the Mann–Whitney U-test for variables that were not normally distributed. Cumulative survival was described using the Kaplan–Meier method. Categorical variables were compared using chi-square test or Fisher’s exact test.

Surgical technique

Our surgical approaches have been previously described in detail (Atluri et al., 2016; Glauber et al., 2015; Miceli et al., 2015). For the conventional MVS, mostly the standard median sternotomy with central bicaval cannulation was used. Briefly, the minimally invasive MVS was performed through a right anterolateral thoracotomy, mostly in the fourth intercostal space and femoral cannulation for cardio-pulmonary bypass was used. Various mitral valve repair techniques were performed. For example, we used ring annuloplasty, leaflet resection (quadrangular or triangular resection), neochordae and commissuroplasty. For the mitral valve replacement, we used the biological prosthesis Hancock II, Perimount Magna prostheses or mechanical St-Jude medical prostheses. In all cases transesophageal echocardiography was performed to evaluate the surgical results.

Baseline characteristic

The median age of the patients in the matched-propensity analysis was 63 and 62 years in the mini-MVS and the conv-MVS group, respectively (p = 0.765). The male gender was predominant in both groups. The haemodynamic profiles of the matched patients were comparable in both groups. The demographic characteristics and mitral valve pathology grouped by Carpentier’s classification of the propensity-matched cohorts are outlined in detail in Tables 1 and 2.

Clinical results

The in-hospital outcomes and complications are summarized in Table 3. In terms of in-hospital mortality, there was a significant difference between the investigated groups (mini-MVS: n = 1; 0.4% vs. n = 7; 3.1% in conv-MVS; p = 0.032). In the mini-MVS group, strokes occurred in 1.8% (n = 4) vs 0.9% (n = 2) in the conv-MVS group (p = 0.264). Furthermore, symptomatic transitory psychotic syndrome was observed in 28 patients (12.3%) in the mini-MVS group, which was similar to the incidence in the conv-MVS group (n = 23; 10.1%; p = 0.457). Moreover, there were two cases (0.9%) of myocardial infarctions after mini-MVS postoperatively, although these were without statistical significance in comparison with conv-MVS (p = 0.562). Exploration due to mediastinal bleeding did not differ significantly between the matched groups (p = 0.411). The transfusion requirements were significantly lower in the minimally invasive matched cohorts (p < 0.001). The incidence of postoperative respiratory failure requiring re-intubation (p = 0.647) or tracheotomy (p = 0.760) were comparable and did not differ significantly. Mini-MVS was at an advantage regarding acute renal failure with a necessity for dialysis (n = 6; 2.6% for mini-MVS vs. n = 15, 6.6% for conv-MVS; p = 0.044). The incidence of sepsis and pneumonia was comparable between the matched groups. The rate of pacemaker implantation did not differ between the analyzed groups (p = 0.177). There were one thoracic wound infection without necessity of surgical revision after the anterolateral thoracotomy, whereas one patient had severe wound infection requiring surgical revision in the conventional group (p = 1.000). The intensive care duration and the length of stay in hospital were comparable between the groups (p = 0.779 and p = 0.516), respectively. Postoperatively, both groups presented comparable median pain intensity rated on the visual analogue scale from 0 to 10: mini-MVS had a score 3.4 vs 3.5 for con-MVS (p = 0.862). The incidence of phrenic nerve palsy in both groups was comparable (p = 0.154).

Surgical results

In the mini-MVS group the median operation time was 282 min and the CPB time was 178 vs. 200 min and 112 min in the conv-MVS cohort. These differences were statistically significant (p < 0.001). The cross-clamp times were significantly longer in the mini-MVS than in the sternotomy cohort (p < 0.001). Patients in the mini-MVS group received mitral valve repair more frequently (81.1% vs. 46.3%; p < 0.001) despite the same distribution of mitral valve apparatus defects (Table 2). No conversion to sternotomy was necessary and there was no case of aortic dissection in our collective. There were 218 patients (96.0%) who underwent mini-MVS and 12 patients (5.3%) who underwent conventional sternotomy using femoral arterial cannulation (p < 0.001); there was aortic cannulation in 18 cases (7.9%) for mini-MVS and in 214 cases (94.7%) for conventional sternotomy (p < 0.001).

Crystalloid cardioplegia was used in all patients in the mini-MVS cohort and in 224 in the conventional cohort. Blood cardioplegia was used only in three cases of the conventional cohort. Five patients (2.2%) in the mini-MVS cohort vs. 13 patients (5.7%) in the conv-MVS group received intra-aortic balloon pump (p = 0.018), and only one patient (0.4%) received extracorporeal membrane oxygenation in the conv-MVS group (p = 0.317). The surgical data are described in Table 4.

Follow-up results

The survival data are illustrated in Fig. 1. The median follow-up time was 4.4 years. One-, five-, and 10-year survival rates were 96%, 90%, and 84% after the minimally invasive approach, and 89%, 85%, and 70% after the conventional sternotomy, respectively (log rank p = 0.004, Breslow p = 0.004). In both the mini-MVS group and in the conv-MVS group there were two valve-related re-operations in each group, (p = 0.881). In the mini-MVS cohort the patients were operated on after 30 and 72 months and in the cohort of the conventional approach after 24 and 38 months. At the final follow-up interview, 51 patients (50.5%) were presented at New York Heart Association (NYHA) Classes I and II after minimally invasive surgery which is correspondingly comparable with the conv-MVS group (50.6%, n = 44; p = 0.727). Furthermore, both groups did not differ with regard to exercise capacity (p = 0.153). During the follow-up time, five patients had strokes in the mini-MVS cohort, while no stroke occurred in the conv-MVS cohort, (p = 0.036). However, no patient suffered from myocardial infarction in the mini-MVS group, although two patients in the conv-MVS group had myocardial infarction in the follow-up period, (p = 0.124). The patients’ scar satisfaction regarding the cosmetic results was comparable in the matched groups (p = 0.771). The follow-up results are summarized in Table 5.