Totally endoscopic mitral replacement: outcomes when repair is infeasible

In a retrospective study from a high-volume German cardiac center, totally endoscopic mitral valve replacement was required in only 75 of 1,037 consecutive patients undergoing endoscopic mitral valve surgery, a rate of 7.2%. Most of the 1,037 patients underwent repair, producing an overall repair rate above 90%. Among patients who received a replacement, the principal reasons were sclerotic leaflet degeneration, infective endocarditis, mitral stenosis, an unsuccessful intraoperative repair attempt, and residual regurgitation after transcatheter edge-to-edge repair. The researchers report acceptable early and five-year outcomes for selected patients undergoing chordal-sparing replacement, although the retrospective comparison with matched repair patients found more early deaths, longer hospital stays, higher transmitral gradients and more permanent pacemaker implantations after replacement.

The investigation used retrospectively collected clinical and follow-up data from the University Medical Center Hamburg-Eppendorf. Patients had undergone totally endoscopic mitral valve surgery between 2012 and 2022, using a right anterolateral or transaxillary mini-incision, three-dimensional endoscopic visualization, femoral or axillary cardiopulmonary-bypass cannulation, a transthoracic aortic clamp and antegrade cardioplegia. The 75 replacement patients were compared through 1:1 propensity-score matching with 75 of the remaining 962 patients who underwent repair. Matching used age, sex, body-mass index, arterial hypertension and atrial fibrillation, while the primary endpoints were mortality and mitral valve reintervention. Kaplan–Meier methods and log-rank tests assessed survival and reintervention, and Benjamini–Hochberg adjustment accounted for multiple comparisons.

The groups were balanced on the variables included in the matching model, but important differences remained. Patients receiving replacement had a higher Society of Thoracic Surgeons Predicted Risk of Mortality score, 1.96% versus 0.66%, and lower hemoglobin, 12.0 versus 13.6 grams per deciliter. They also had more preoperative mitral stenosis and endocarditis, as well as higher peak and mean transmitral gradients: 13.5 versus 7.6 millimeters of mercury and 5.45 versus 2.5 millimeters of mercury, respectively. Left ventricular ejection fraction was similar in the two groups, at 58.2% after replacement and 56.8% after repair, and right ventricular function was also comparable. The matching therefore reduced measured demographic differences but did not establish that the groups had identical disease severity or pathology.

Surgical complexity was reflected in the reasons surgeons could not repair the valve. Sclerotic leaflet degeneration accounted for 28.0% of replacement cases and included nine patients with rheumatic disease; moderate-to-severe mitral stenosis was present in 17.3%. Acute infective endocarditis accounted for 26.7%, while 17.3% of patients had already undergone an unsuccessful repair attempt during the same operation. Another 6.7% had residual mitral regurgitation after transcatheter edge-to-edge repair. Surgeons generally preserved the posterior mitral leaflet and its chordal attachments when anatomy allowed, while the anterior leaflet was usually resected. Preservation was not attempted with substantial annular calcification or a small, restrictive annulus, and the technique was tailored to each valve’s anatomy rather than applied uniformly.

Cross-clamp time, cardiopulmonary-bypass time and total operating time were similar after replacement and repair. However, conversion from the endoscopic approach to open sternotomy was required in 9.3% of the replacement group, compared with 2.7% of the repair group. Reasons included systemic anterior motion, impaired visualization caused by severe calcification, venous congestion, pulmonary adhesions and bleeding. Six of the seven converted replacement patients survived without further sequelae. Bioprosthetic valves were implanted in 97.3% of replacement procedures, with 29 millimeters the most common size, used in 45.3% of cases. Transesophageal echocardiography was used to confirm the surgical result, and no paravalvular leakage was observed.

Early complications were broadly similar, although several differences favored repair. Re-thoracotomy for bleeding occurred in 14.7% of replacement patients and 6.7% of repair patients. Low cardiac output, defined using extracorporeal membrane oxygenation, Impella support or prolonged catecholamine requirements, occurred at rates of 6.8% and 5.5%, respectively. Replacement patients stayed in hospital longer, averaging 12.1 days versus 8.4 days, and had lower postoperative right ventricular function, with tricuspid annular plane systolic excursion of 13.0 versus 15.7 millimeters. Permanent pacemaker implantation occurred in 13.3% of the replacement group and none of the matched repair group. The authors considered both valve complexity and the potential effect of prosthetic implantation on the atrioventricular conduction system as possible contributors.

All three in-hospital deaths occurred among replacement patients, although the difference from zero deaths after repair did not meet the adjusted statistical threshold. Thirty-day mortality was 5.3% after replacement and 0% after repair, and Kaplan–Meier estimated 30-day survival was 94.7% and 100.0%, respectively. Reported causes of the three in-hospital deaths were septic shock from endocarditis, multiorgan failure with extracorporeal-support-dependent cardiogenic shock, and hemorrhagic shock following emergency re-thoracotomy for hemothorax. Postoperative peak and mean transmitral gradients remained higher after replacement, at 13.6 and 5.5 millimeters of mercury, compared with 9.4 and 3.5 millimeters of mercury after repair. The researchers did not routinely calculate indexed effective orifice area, so they could not determine how often prosthesis-patient mismatch occurred.

Follow-up data were available for the entire matched cohort, with a median follow-up of 46 months, although repair patients were followed longer on average. Overall mortality during follow-up was 13.3% after replacement and 5.3% after repair, a difference that was not statistically significant. In Kaplan–Meier analysis, five-year survival was 86.7% for replacement and 94.7% for repair. A landmark analysis that excluded the three in-hospital deaths produced five-year survival estimates of 87.1% and 92.9%, respectively, with no significant difference, suggesting that the observed separation was concentrated in the early postoperative period. Mitral valve reintervention was uncommon in both groups, occurring in 4.0% after replacement and 6.7% after repair. Replacement-group reinterventions involved prosthetic endocarditis and structural degeneration, while all five repair-group reinterventions followed recurrent significant mitral regurgitation.

The study offers a specific picture of when endoscopic replacement may be unavoidable, rather than evidence that replacement is preferable to repair. A repair-first strategy remained central at the center, and replacement was reserved for complex or non-repairable pathology. The authors also note that surgery after failed transcatheter edge-to-edge repair may become more common as clip-related leaflet damage or fibrosis complicates subsequent repair. That subgroup was small in this series, however, and the retrospective design cannot establish whether an endoscopic approach improves outcomes in such patients. A secondary inverse-probability-weighted analysis covering 692 complete cases was limited by a severe positivity violation, with a maximum control weight of 3,284, and produced less satisfactory balance than the matched cohort. The study was also confined to one expert center, included only 75 replacement patients, lacked systematic postoperative left ventricular volume data and could not exclude residual confounding. The findings therefore support chordal-sparing endoscopic replacement as a feasible option in selected difficult cases, but not as a general substitute for repair or as proof of equivalence between the two approaches.

Subject of Research: Outcomes and indications for totally endoscopic chordal-sparing mitral valve replacement when repair is not feasible

Article Title: Totally endoscopic mitral valve replacement: reasons for non-feasibility of repair and 5-year outcomes

Article References: Sobik, F., Pausch, J., Hua, X., Beyer, M., Macius, E., Bhadra, O., Alassar, Y., Kloth, B., Conradi, L., Reichenspurner, H., Pecha, S., Girdauskas, E., & Schaefer, A. (2026). Totally endoscopic mitral valve replacement: reasons for non-feasibility of repair and 5-year outcomes. Frontiers in Cardiovascular Medicine, 13, Article 1903773. https://doi.org/10.3389/fcvm.2026.1903773

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1903773

Keywords: mitral valve replacement, mitral valve repair, endoscopic surgery, chordal preservation, mitral regurgitation, infective endocarditis, mitral stenosis, cardiac surgery, bioprosthetic valve, five-year survival, Totally, endoscopic