The primary endpoint was all-cause mortality. Secondary endpoints included: reoperation, hemorrhagic (major bleeding requiring hospitalization, transfusion, or causing death) and neurologic events (stroke or transient ischemic attack) and recurrent mitral valve dysfunction. Recurrent mitral dysfunction was defined as: mitral regurgitation > moderate (≥2+) or mitral stenosis with mean gradient ≥ 8 mmHg.

All endpoints were adjudicated based on clinical records and echocardiographic findings.

This study was approved by the Institutional Review Board of the Shisong Cardiac Center. Given the retrospective nature of the study and the use of anonymized data, the requirement for individual informed consent was waived.

All the patients underwent two-dimensional transthoracic echocardiography at the time of their admission. The routine echocardiography evaluation included the assessment of cardiac parameters (ventricular dimensions and contractility, pulmonary hypertension, etc…) and accurate analysis of mitral valve anatomy and function to define the valve etiology and orient the surgical approach. The indications for Mitral surgery were made following a case discussion by a joint team of Cardiologists, cardiac surgeons and anesthesiologists. Considering the socio-cultural nature of our environment, the patient global compliance was assessed through inclusive discussions with patients and family before each surgical procedure.

A standard median sternotomy was performed in all the patients. Cardiopulmonary Bypass was established through a conventional ascending aorta and selective bi-caval cannulation. Cardioplegic solutions (warm blood or crystalloid) for myocardial protection were chosen based on surgeon’s preference. Following a standard left atriotomy, a careful valve analysis was performed and the decision to repair or not the valve was taken by the senior surgeon. Basically, the choice between MVr and MVR was based on a combination of anatomical, clinical, and contextual factors. MVr was preferentially attempted in patients with predominant mitral regurgitation, preserved leaflet mobility and limited calcification and favorable subvalvular apparatus. MVR was favored in cases with severe leaflet fibrosis or calcification such as mixed lesions and severe stenosis with advanced rheumatic involvement with restricted mobility. Failed repair attempts (conversion from MVr to MVR intraoperatively) were analyzed within the MVR group according to an intention-to-treat principle.

Mechanical valves models and prosthetic rings (flexible, semirigid, rigid) were chosen based mainly on availability at the time of the procedure.

The ventriculo-annular continuity was preserved in MVR by leaving the posterior mitral leaflet whenever possible or by implanting artificial PTFE chords. MVr techniques included: prosthetic ring annuloplasty, artificial PTFE chords implantation, commissurotomy, leaflet shaving, leaflet pericardial patch enlargement, papillary muscle splitting and secondary chords cutting.

Intra-operative Trans-esophageal Echocardiography was used in all the cases. Residual MR ≤ 2 was tolerated in patients who had MVr whereas detected intraoperative perivalvular leak required aggressive correction in MVR patients.

A postoperative anticoagulation regimen with acenocumarol (Sintrom ®) was recommended in all the patients. The anticoagulation was initiated whenever possible from the first postoperative day with a targeted International Normalized Ratio (INR) between 2.5 and 3.5 for MVR and 2.0 to 3.0 for MVr patients. Concomitantly, a bridging anticoagulant therapy was started with subcutaneous injections of low-molecular weight heparin (Clexane, SANOFI: 60 - 100 UI/kg twice daily) until the targeted INR was reached or as substitute therapy in cases oral anticoagulation was no immediately achievable (i.e. patients with prolonged mechanical ventilation). Oral anticoagulation was recommended for three months after discharge in MVr patients whereas lifelong anticoagulation was strictly advocated to MVR group. In the MVr group, however, anticoagulation was prolonged beyond 3 months in those with atrial fibrillation (6.4% in the MVr group at follow-up).

Transthoracic echocardiograms were performed at discharge in all the cases and during follow-up in our institution. Patients living in remotes areas were reviewed during our external screening campaigns or by collaborating physicians and cardiologists. A monthly INR control was recommended in all the patients whenever possible.

Data were retrospectively collected from institutional surgical registries, operative reports, hospital medical records, and echocardiographic databases. Follow-up data were obtained through outpatient visits, institutional echocardiographic evaluations, external outreach campaigns, and communication with referring cardiologists. The median follow-up duration was approximately 7.0 years; follow-up represented 1,585 patient-years. Follow-up completeness was 89%. Patients lost to follow-up (11%) were censored at the date of last contact. Variables with missing data were analyzed using available-case analysis without imputation, given the retrospective design. Sensitivity analyses confirmed that missingness did not materially alter outcome estimates.

Continuous variables were summarized as median [interquartile range] and compared using the Mann-Whitney U test. Categorical variables were summarized as n (%) and compared using the chi-square test or Fisher’s exact test, as appropriate. Kaplan-Meier methods were used to estimate time-to-event outcomes, with comparisons by the log-rank test. In addition to p-values, standardized mean differences (SMDs) were reported to quantify baseline imbalance between the groups, as SMDs provide a sample-size independent measure of heterogeneity and help identify clinically meaningful differences.

To account for baseline differences, multivariable Cox proportional hazards models were constructed for survival and major adverse events, adjusting for clinically relevant covariates including age, sex, NYHA class, atrial fibrillation, lesion type, and concomitant tricuspid surgery. Proportional hazards assumptions were verified using Schoenfeld residuals.

Of the 231 patients of the study, 91 were male with a sex ratio of 0.76. The mean age of the whole cohort was 29.1 years (IQR, 16.1 - 40.6). The mean ages in MVR and MVr groups were 27.1 (IQR, 16.5 - 40.0) and 31.0 (IQR, 16.0 - 40.7) years respectively (p = 0.65). Rheumatic disease was the commonest etiology in both groups (MVR: 80.6% vs MVr: 76.5%, p = 0.34). Mitral regurgitation was the predominant lesion with a higher rate in MVr group (97.5 % vs 68.0 %) (Table 1 & Table 2).

Table 1. Patient’s demographics.

Note: MVR: mitral replacement; MVr: mitral repair; BSA: body surface area; NYHA: New York Heart Association; LVEF: left ventricular ejection fraction; LVEDD: left ventricular end diastolic diameter; LVESD: left ventricular end systolic diameter; PAPs: pulmonary artery pressure in systole.

Table 2. Operative data.

Note: MVR: mitral replacement; MVr: mitral repair; CPB: cardiopulmonary bypass; X-clamp: cross-clamp. ICU: intensive care unit.

Table 3. Surgical data.

The On-X mechanical valve model was implanted in 97.8 % (n = 141/150 patients) of the patients who underwent MVR, with 98.2% all the valves’ sizes being of adult size (≥25 mm). Prosthetic valves sizes extracted from operative descriptions are reported in Table 3.

An annuloplasty technique was used in 96.2% (n = 78/81) of patients in the MVr group. Ring models were: rigid or semi-rigid (66.6%) (Carpentier-Edwards, Medtronic Future Band), flexible (28.2%) (MRS system, Vascutek) and pericardial band (5.1%).

Concomitant tricuspid procedures were identified in 109 patients (44.1%), and were more frequent in the MVr group (46/81, 56.7%) than in the MVR group (63/150, 42.0%).

The median cardiopulmonary bypass and cross-clamping duration for MVR were 112.7 (65.1 - 139.0) and 72 (32.5 - 98.5) minutes, and 132.0 (95.2 - 175.0) and 63.3 (50.1 - 132.2) minutes for MVr respectively.

The 30-day mortality for the whole cohort was 5.1 % (12/231; MVR = 8/150, 5.3%; MVr = 4/81, 4.9%). The commonest cause of death was postoperative low cardiac output syndrome occurring in seven patients (MVR = 4; MVr = 3). Other causes of death were pneumonia (n = 2), acute renal failure (n = 1), pulmonary embolism (n = 1) and sudden death (n = 1). No significant differences were found between the groups regarding the intensive care unit (2.6 vs 2.5 days, p = 0.90) and hospital stay (15.9 vs 16.3, p = 0.47).

Figure 1.Kaplan-Meier estimate of overall survival.

The survival rates at 5 and 10-year were 83.6 ± 2.8% and 74.6 ± 2.7% for MVR and 87.3 ± 4.5% vs 78.1 ± 5.2% for MVr respectively (Figure 1). No significant differences were detected at Log-rank test between groups (p = 0.29). The estimate 5 and 10 year freedom from reoperation rates were 98.8 ± 1.8% and 97.1 ± 2.4% in the MVR group and 97.2 ± 2.6% and 95.1 ± 2.9% in the MVr group (p = 0.06) (Figure 2).

Figure 2.Kaplan-Meier estimate of freedom from reoperation.

The 5 and 10-year overall freedom from hemorrhagic events was 78.7 ± 7.2 and 72.0 ± 3.0% for MVR, and 97.1 ± 7.6% (5 and 10-year) for MVr respectively (p < 0.001) (Figure 3). The 5 and 10-year freedom from stroke was 94.2 ± 6.8 and 87.8 ± 2.1% for MVR, and 100% (5 and 10-year) for MVr respectively (Figure 4).

Figure 3.Kaplan-Meier estimate of freedom from hemorrhagic events.

Figure 4. Kaplan-Meier estimate of freedom from stroke.

The overall freedom from mitral dysfunction (MR > 2+ or > moderate stenosis) in MVr patients was 97.1 ± 1.8% and 76.7 ± 3.1% at 5 and 10-year respectively (Figure 5).

Figure 5. Kaplan-Meier freedom from recurrent mitral dysfunction (regurgitation > 2+ or > mild stenosis) in MVr group.

On multivariable Cox regression analysis (Table 4), there was no significant difference in long-term survival between MVr and MVR (hazard ratio [HR] 0.88, 95% confidence interval [CI] 0.54 - 1.42; p = 0.60). MVr was associated with a significantly lower risk of hemorrhagic events (HR 0.32, 95% CI 0.18 - 0.71; p = 0.003) and neurologic events (HR 0.41, 95% CI 0.19 - 0.89; p = 0.02).

Table 4. Multivariate Cox regression analysis.