Patients with mild or moderate degenerative MR who received isolated surgical ablations for paroxysmal or persistent AF diagnosed on a 12-lead electrocardiogram in our center from 2013 to 2022 were referred. The exclusion criteria were as follows: (1) patients with organic valvular diseases who needed invasive interventions according to the guidelines; (2) patients with mechanical or biological prosthesis; (3) patients who received any other cardiac surgeries during ablations; (4) other etiologies of mitral valve diseases like rheumatic mitral valve diseases and secondary mitral valve regurgitation; (5) permanent pacemaker implantations. Patients were grouped according to the treatment methods for LAA (LAA resection, endocardial closure or preservation)

Two-dimensional echocardiography and doppler color flow imaging (IE33; Philips Medical Systems, Andover, MA, USA) were performed on all patients. Despite the routine geometric examination of left heart (left atrium and left ventricle). MR and tricuspid regurgitation (TR) grade were assessed by a multiparametric approach [13], including qualitative, semi-quantitative, and quantitative parameters, which was graded from 0 to 4, where 0, none or trivial; 1+, mild; 2+, moderate; 3+, moderate to severe; and 4+, severe. Mitral annular diameter (MAD) was acquired from a four-chamber apical view at the end of expiration. The measurement of the mitral tethering height (TH) was done by calculating the distance from the mitral annular plane to the leaflet coaptation point orthogonally.

Ablation was carried out with bipolar Cardioablate (Medtronic, Minneapolis, MN, USA) or an Atricure clamp (Atricure, West Chester, OH, USA). After the aorta clamping and cardioplegia perfusion, incising the LA and performing the LA ablation lines, which included bilateral pulmonary vein isolations (PVI), a roof connecting line between both islands of PVIs and a line from the left PVI ablation lesion to the posterior mitral annulus. Then, LAA closure ablation was performed by making a radiofrequency lesion around the base of LAA. Patients who underwent LAA interventions experienced either LAA resection or LAA endocardial closure. The LAA was excised with surgical staplers (Covidien, Medtronic, Minneapolis, MN, USA). Whether to perform LAA interventions or preservations depended on several factors, which included the anatomical features of LAA, the risk of atrial rupture evaluated by surgeons, the risk of stroke and surgeons’ preferences. All patients experienced the division of the Marshall ligament. The right atrial ablation lesions included the cavotricuspid isthmus ablation, superior vena cava to inferior vena cava, lateral free-wall to anterior-medial tricuspid valve annulus; and medial free-wall to the anterior-medial tricuspid valve annulus. More details about the above procedures can be found in our Supplementary Material.

As we described in the prior study [14], oral anticoagulation was recommended to everyone for the first three months after procedures. For patients who had LAA preserved, oral anticoagulation was discontinued if one had no AF recurrence and a CHADS2 score two within three months following the procedure. All patients who underwent LAA treatments were recommended to discontinue oral anticoagulation 3 months following the procedure. Unless contraindicated, patients received amiodarone or sotalol within 24 h of the procedure, which was discontinued at three months (blanking period). Antiarrhythmic drugs (AAds) were maintained or restarted if patients demonstrated recurrent AF. Additional postoperative medications were shown in Supplementary Material. Multiple agents, including beta-blockers, nondihydropyridine calcium-channel blockers (CCBs), digoxin and certain AADs were used for rate control in patients with AF persistence. Excluding the above treatments, angiotensin-converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), aldosterone receptor antagonists (ARA), and loop diuretics were used to relieve symptoms and delay the further exacerbation of MR during follow-up according to the patient’s degree of symptoms, grade of MR, hemodynamic situations, and presence or absence of heart failure.

Clinical visits were arranged for all patients at 1, 3, and 12 months after procedures, and then annually thereafter. Each clinical visit was composed of medical interrogation, physical examination, X-ray, Holter examination, and echocardiographic evaluation. The primary outcome was defined as the worsening of MR (MR grade increased from mild to moderate, mild to severe or moderate to severe). Adverse clinical events including all-cause death, recurrent atrial fibrillation and stroke were recorded. Excluding the routine echocardiographic parameters, items related to the anatomy of the mitral valve including MAD and tethering height (TH) were also recorded at each visit.

Continuous variables which fit the normal distribution were presented as mean (standard deviation), while other continuous variables were presented as median (interquartile range). Categorical values were presented as percentages, and odds ratios (OR) were presented with 95% confidence intervals (CIs). Unadjusted comparisons in terms of categorical variables between different cohorts were done by the chi-square test and Fisher’s test. Unadjusted comparisons in terms of continuous variables between different cohorts were done by the two-sample t test and Kruskal-Wallis H test. Propensity scores in this model were acquired through the generalized boosted model. The absolute standardized mean difference (ASMD) was used to measure the difference between two univariate distributions of a single baseline variable [15]. Imbalance was presented when a value was $\ge$0.10. Freedom from outcomes whose results were adjusted by inverse probability of treatment weighting (IPTW) for casual effects after acquiring propensity scores. All significance tests were 2-tailed, and a p value of 0.05 was considered to be statistically significant. R version 3.6.2 (R Foundation for Statistical Computing, Vienna, Austria) was used for all statistical analysis.

A total of 456 patients were enrolled in this study. Of these, 278 underwent LAA interventions (146 cases of resection, 132 cases of endocardial closure), and the remaining 177 had LAA preserved. The average age was 62 $\pm$ 7.6 years; male patients accounted for 41.4% of the overall population. Before ablations, patients with mild to moderate MR (Grade II) accounted for 95.8% of the overall population, and the remaining patients showed moderate to severe (Grade III) MR. After IPTW adjustments, no significant differences were observed among the three groups. All ASMDs were also decreased to lower than 10% after adjustments (Table 1).

All patients in this study successfully underwent the ablation procedures, and no in-hospital mortalities were observed. After a median follow-up time of 4 years, 30 deaths were observed. The follow-up rate of this study was 98.5%. The overall survival rates at 3 years and 5 years were 94.2% and 89.1%, respectively. No differences were observed among the three groups after IPTW adjustment (All p $>$ 0.05) (Fig. 1A). During follow-up, a total of 11 cases of stroke events were observed. The differences in terms of the incidence of stroke were also not significant between LAA intervention and LAA preservation (Fig. 1B).

Fig. 1.

IPTW adjusted Kaplan-Meier Curves in terms of Mortality and Stroke. (A) IPTW adjusted Kaplan-Meier curves in terms of all-cause mortality. (B) IPTW adjusted Kaplan-Meier curves in terms of stroke. IPTW, inverse probability of treatment weighting; LAA, left atrial appendage; HR, hazard ratio; CI, confidence interval.

As shown in Table 2, medications after discharges showed no differences among the three groups. Worsening of MR occurred in 20.2% (92 cases) of participants during follow-up. Of these, 52 cases of MR aggravated from mild to moderate, 10 cases aggravated from mild to severe, and 30 cases aggravated from moderate to severe. Worsening of MR was more likely to occur among patients who underwent LAA resected or closed (LAA preservation vs. LAA endocardial closure: p = 0.004, 95% CI (0.232–0.762); LAA preservation vs. LAA resection: p 0.001, 95% CI (0.201–0.611)) (Fig. 2). Patients who had LAA preserved had a lower risk of recurrent AF than those who underwent LAA interventions (LAA preservation vs. LAA endocardial closure: p = 0.020, 95% CI (0.270–0.894); LAA preservation vs. LAA resection: p = 0.010, 95% CI (0.255-0.831)) (Fig. 3). All results without IPTW adjustments can be found in the Supplementary Material.

Fig. 2.

IPTW adjusted Kaplan-Meier curves in terms of worsening of MR. IPTW, inverse probability of treatment weighting; LAA, left atrial appendage; HR, hazard ratio; CI, confidence interval; MR, mitral regurgitation.

Fig. 3.

IPTW adjusted Kaplan-Meier curves in terms of recurrent AF. IPTW, inverse probability of treatment weighting; LAA, left atrial appendage; HR, hazard ratio; CI, confidence interval; AF, atrial fibrillation.

Before ablations, the three groups of patients showed similar left atrial diameter (LAD) and MAD (all p $>$ 0.05). By 3 years postoperatively, patients who underwent LAA resection and endocardial closure showed significantly larger LAD and MAD than those who had LAA preserved (p 0.05). At any time of follow-up (1 to 3 years, 4 to 6 years, and 7 to 9 years postoperatively), patients who underwent LAA interventions (resection or endocardial closure) showed lower TH than those who had LAA preserved (Fig. 4).

Fig. 4.

Longitudinal echocardiographic features during follow-up among different LAA treatment methods. (A) LAD among different LAA treatment methods; (B) MAD among different LAA treatment methods; (C) TH among different LAA treatment methods; *p 0.05; The errors bars are SE of the mean. No corrections for multiple testing were applied. LAA, left atrial appendage; LAD, left atrial diameter; MAD, mitral annular diameter; TH, tenting height; p-values may not be interpreted as confirmatory but rather descriptive.