Post-operative left atrial volume index is a predictor of the occurrence of permanent atrial fibrillation after mitral valve surgery in patients who undergo mitral valve surgery

Increased left atrial (LA) size is associated with the occurrence of atrial fibrillation (AF) [1]. Therefore, AF is frequently observed in patients with chronic mitral valve (MV) disease, which invariably induces LA remodeling [2‐4]. On the other hand, reduction of LA size (reverse LA remodeling) can also occur after correction of MV pathology [3, 5]. It has been shown that the degree of reverse LA remodeling varies, particularly according to pre-operative cardiac rhythm. Moreover, pre-operative sinus rhythm is associated with larger degrees of reverse LA remodeling [3]. In addition, this structural reversal can also induce reversal of electrophysiologic abnormalities that are predisposed to the occurrence of AF. Therefore, reversal of these changes by treatment, i.e., MV surgery, could potentially have important implications for the prevention of AF [5, 6]. Nevertheless AF can occur even after correction of MV pathology in patients who have undergone MV surgery [7‐17]. However, the incidence and predictors of new onset permanent AF after MV surgery have not been clearly defined. Therefore, the aim of this study is to investigate the prevalence and predictors of the occurrence of new onset permanent AF in patients with MV diseases who have undergone MV surgery and sinus rhythm pre-operatively.

The enrolled patients were classified into the two groups according to the occurrence of permanent AF or maintained sinus rhythm after MV surgery (normal sinus rhythm [NSR] group vs. AF group). Post-operative new onset permanent AF occurred in 81 (18%) patients. Baseline characteristics of the study subjects are shown in Table 1. The mean age at the time of surgery was similar in both groups. The majority of patients (81%) had rheumatic etiology and others (19%) had non-rheumatic etiology, such as infective endocarditis, MV prolapse, or chordae rupture. The proportion of rheumatic valve disease was significantly higher in patients with new onset permanent AF (p = 0.018). The median interval from the surgery to the occurrence of AF was 9.2 years (110.6 ± 78.9 months), and the mean follow-up duration was not significantly different in both groups (9.8 ± 5.9 years in the NSR vs. 9.4 ± 6.5 years in the AF, p = 0.566). The TTE parameters before and after surgery are listed in Table 2. Pre- and post-operative LV EF were significantly lower in the AF group (63 ± 10% in the NSR vs. 59 ± 14% in the AF, p = 0.026 & 61 ± 10% in the NSR vs. 57 ± 14% in the AF, p = 0.006), although the LV EF of both groups were in normal range. Pre-operative LA AP dimension (51 ± 9 mm in the NSR vs. 53 ± 7 mm in the AF, p = 0.080) and LAVI (58 ± 25 ml/m² in the NSR vs. 63 ± 23 ml/m² in the AF, p = 0.399) were not significantly different in both groups. However, the degree of reduction of LAVI (21 ± 21 ml/m² in the NSR vs. 11 ± 23 ml/m² in the AF, p = 0.041) and percentage reduction of LAVI (28 ± 28% in the NSR vs. 10 ± 37% in the AF, p < .001) were significantly smaller in the AF group (Fig. 2). Therefore, post-operative LA size was significantly larger in the AF group, shown as AP dimension (43 ± 6 mm in the NSR vs. 51 ± 8 mm in the AF, p < .001) and LAVI (38 ± 13 ml/m² in the NSR vs. 52 ± 16 ml/m² in the AF, p < .001). When the degree of reduction of LA dimension was compared according to the direction (AP, ML or SI), the change in SI direction was most prominent (Table 2). Regarding the hemodynamic variables, the post-operative mean diastolic pressure gradient (MDPG) of the MV was significantly higher in patients with permanent AF group (3.5 ± 1.4 mmHg in the NSR vs. 3.9 ± 1.5 mmHg in the AF, p = 0.023). The grade of pre and post-operative tricuspid regurgitation (TR) were slightly higher in the AF group. Post-operative estimated pulmonary artery pressure (PAP) was also slightly higher in the AF group. In univariate analysis, rheumatic etiology (odds ratio [OR] = 2.474, 95% confidence interval [CI] = 1.143–5.355, p = 0.021), lower pre (OR = 0.971, 95% CI = 0.945–0.997, p = 0.028) and post-operative LV EF (OR = 0.970, 95% CI = 0.949–0.992, p = 0.007), higher post-operative MDPG across mitral prosthesis (OR = 1.212, 95% CI = 1.024–1.434, p = 0.025), lesser degree of reduction in LA size after surgery (OR = 0.790, 95% CI = 0.960–0.980, p < 0.001), and large post-operative LA size (OR = 1.064, 95% CI = 1.045–1.083, p < 0.001) were risk factors for the occurrence of AF. Presence of MS rather than pure MR (OR = 1.767, 95% CI = 0.981–3.182, p = 0.058) was associated with the occurrence of AF with borderline significance (Table 3). Interestingly, none of the parameters reflecting pre-operative LA size was associated with post-operative AF. Pre and post-operative TR (OR = 1.864, 95% CI = 1.328–2.617, p < 0.001 & OR = 2.641, 95% CI = 1.922–3.630, p < 0.001) and postoperative higher PAP (OR = 1.067, 95% CI = 1.025–1.110, p = 0.001) were associated with the occurrence of AF. In multivariate analysis, post-operative LAVI was an independent predictor for the occurrence of AF (Table 4). The predictive ability of the LAVI was determined by the area under the curve of the receiver operating curve and post-operative LAVI > 39 ml/m² (cut-off value) was associated with new onset permanent AF (sensitivity: 79%, AUC: 0.762, SE: 0.051, p < 0.001). The AF-free survival curves of patients with post-op LAVI < 39 ml/m2 or ≥39 ml/m2 are shown in Fig. 3 (p = 0.06).

In the present study, the prevalence and predictors of new onset permanent AF after MV surgery were investigated. The results showed that new onset permanent AF is not uncommon, occurring in about 20% of patients even after successful MV surgery and pre-operative sinus rhythm. Although several parameters, such as rheumatic etiology of MV, presence of mitral stenosis (MS), lower LV EF, higher post-operative mean diastolic pressure gradient across mitral prosthesis, and lesser degree of reverse LA remodeling after surgery, were shown to be associated with the occurrence of post-operative AF, post-operative LAVI was found to be an independent predictor for the occurrence of AF in multivariate analysis. Interestingly, none of the parameters reflecting pre-operative LA size was associated with post-operative AF. These findings underscore the importance of post-operative echocardiographic assessment before dismissal of the evaluation of LV function and LA size even in patients who have undergone successful MV surgery and pre-operative sinus rhythm.

Newly developed postoperative permanent AF is not uncommon, occurring in 18% of patients who have undergone successful MV surgery despite pre-operative sinus rhythm. Rheumatic etiology, the presence of MS, lower LV EF, lesser degree of LA reverse remodeling, and larger post-operative LAVI were associated with new onset permanent AF. Post-operative LAVI > 39 ml/m2 was an independent predictor for the occurrence of AF.