This retrospective single-center study was established in accordance with the Declaration of Helsinki (1964) and approved by the Ethics Committee of the Hamburg Medical Association (IRB# WF-208/20). Due to anonymous, retrospective study design, written informed patient consent was waived.

Patients undergoing full-sternotomy due to unsuitability for an endoscopic approach (eg, severe chest deformities, prior right-sided thoracic surgery, etc.) according to local heart team decision, or concomitant coronary artery bypass grafting, as well as aortic surgery were excluded. Furthermore, isolated tricuspid valve surgery (TVS) patients were excluded from the analysis.

All procedures were performed under general anesthesia by a dedicated team of cardiac surgeons, anesthesiologists, nurses, and perfusionists according to institutional standards as described elsewhere Briefly, endoscopic MVS was performed via right anterolateral mini-incision under 3D endoscopic guidance. Surgical access was established via a limited skin incision (3 to 5 cm) entering the fourth (rarely due to anatomical reasons the fifth) intercostal space, using a nonrib spreading soft tissue retractor ( Figure 1 B). Furthermore, 3 < 1 cm incisions were made to introduce a camera for visualization in the same intercostal space as the incision, a MV retractor parasternal orthogonal to the MV, as well as a transthoracic aortic clamp. Operative visualization was provided using a 3D endoscope (Aesculap Einstein Vision, Tuttlingen, Germany). Endoscopic images were displayed in high definition on a large monitor and 3D glasses were worn. Cardiopulmonary bypass (CPB) was established via percutaneous or open cannulation of femoral or axillary arteries. Venous cannulation was performed using the femoral veins. A transthoracic aortic clamp, which was inserted 1 intercostal space above the original incision, was used in addition to antegrade Bretschneider or del Nido cardioplegia, which was administered into the aortic root, to obtain cardioplegic arrest. Exposure of the MV was performed using direct left atrial access, before standardized intraoperative examination was performed to determine the etiology of MR and define the surgical strategy. MVS and concomitant surgical procedures including atrial fibrillation (AFib) ablation (ie, mainly left-sided cryoablation including a box- and mitral isthmus lesion set), left atrial appendage (LAA) closure (using an intra-atrial suture or extra-atrial clip device) or TVS were performed according to local standard of care. During the procedure transesophageal echocardiography (TOE) was used to confirm indication for surgery and appropriate surgical results.

The aim of this study was to monitor the evolution and temporal changes of endoscopic MVS at our institution. To that end perioperative data and follow-up were retrospectively analyzed.

Baseline, perioperative, and FU variables were retrospectively collected in an anonymized, standardized database. Continuous variables are shown as medians (25th percentile, 75th percentile) and compared using the Mann–Whitney U-test. Binary variables are shown as counts (frequencies) and compared using the χ2 test. For specific variables trend tests over operation year are calculated. For binary variables Cochran–Armitage test is used, whereas continuous variables are assessed by Jonckheere–Terpstra Test for ordered differences among classes. Event rate and survival curves are calculated with Kaplan–Meier estimator. Groups are compared using the log-rank test. Univariable Cox regression was conducted to identify patient characteristics associated with the occurrence of death or reoperation within 5 years. All variables with a P -value <.25 in univariable regression were chosen for further selection methods because of their relevance in the univariable model. Afterwards, a backward subset selection was performed where the chosen method selected the model based on Akaike Information Criteria (AIC). To prepare the subset selection, variables with high amount of missing values or high correlation were excluded, in order to ensure a valid selection for multivariable Cox regression. A P -value of <.05 was considered statistically significant. All analyses were performed with R statistical software version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).