Article Text

This article has a correction. Please see:

Download PDFPDF

Valvular heart disease
Left atrial remodelling early after mitral valve repair for degenerative mitral regurgitation
  1. F Antonini-Canterin1,
  2. C C Beladan2,
  3. B A Popescu2,
  4. C Ginghina2,
  5. A C Popescu3,
  6. R Piazza1,
  7. E Leiballi1,
  8. B Zingone4,
  9. G L Nicolosi1
  1. 1
    Cardiologia ARC, Azienda Ospedaliera “S. Maria degli Angeli”, Pordenone, Italy
  2. 2
    “Prof. Dr. C. C. Iliescu” Institute of Cardiovascular Diseases, Bucharest, Romania
  3. 3
    Cardiology Department, Elias Hospital, Bucharest, Romania
  4. 4
    Cardiochirurgia, Ospedali Riuniti, Trieste, Italy
  1. Dr Francesco Antonini-Canterin, Cardiologia, ARC, Azienda Ospedaliera S. Maria degli Angeli, Via Montereale, 24, 33170 Pordenone, Italy; cardiologia{at}


Objective: Left atrial (LA) size is an important predictor of outcome after mitral valve replacement in patients with symptomatic chronic mitral regurgitation (MR). Data on LA remodelling after mitral valve repair (MVr) for chronic non-ischaemic MR are scarce. The aim of this study was to assess changes in LA size early after MVr for chronic severe degenerative MR and to identify clinical and echocardiographic correlates of those changes.

Methods: The study analysed 225 consecutive patients who underwent MVr and were echocardiographically evaluated in our hospital within 1 month before and 1–6 months after surgery. Patients with MR aetiology other than degenerative, associated aortic valve replacement, or congenital heart disease were excluded. The remaining 79 patients (aged 60 (SD 12) years, 55 men) with isolated chronic severe degenerative MR formed the study group. LA reverse remodelling was defined as a decrease in LA volume index (LAVi) ⩾15%.

Results: LA dimensions significantly decreased after MVr (p<0.001). Mean LAVi reduction was 29% (SD 18%). LA reverse remodelling was observed in 63 patients (80%). Correlates of LAVi reduction were preoperative LAVi (p = 0.008), systolic and diastolic blood pressure (p = 0.032, p = 0.009), postoperative transmitral mean pressure gradient (p = 0.001) and residual MR (p = 0.043). LAVi reduction was lower in patients >45 years (p = 0.008) and in hypertensive patients (p = 0.031).

Conclusion: LA reverse remodelling is common early after MVr for chronic severe degenerative MR. Preoperative LAVi, blood pressure, postoperative transmitral mean pressure gradient, residual MR and age >45 are related to LAVi reduction. The prognostic value of LA reduction in this setting needs further study.

  • cardiac remodelling
  • left atrium
  • mitral regurgitation
  • mitral valve repair
View Full Text

Statistics from

According to a recent European survey mitral regurgitation (MR) accounts for 32% of single native left-sided valve diseases and the aetiology of MR is mostly degenerative (valve prolapse or flail leaflets).1 Mitral valve repair (MVr) is currently recommended in the majority of patients with severe chronic MR who require surgery when the valve is suitable for repair and appropriate surgical skill and expertise are available.2

Left atrial (LA) enlargement, through its association with left ventricular (LV) failure, arrhythmias and stroke, emerged lately as a strong marker for adverse clinical outcome in different populations including patients undergoing mitral valve replacement for chronic, symptomatic MR.35 The prognostic value of preoperative LA size and/or extent of postoperative LA reverse remodelling in patients undergoing MVr for chronic non-ischaemic MR is unknown and data on LA reverse remodelling in this setting are rather scarce.6 7 However, as the relationship between LA size and the risk for atrial fibrillation and stroke has been previously demonstrated, we would expect that a good postoperative LA reverse remodelling could translate into clinical benefit postoperatively. Therefore, the purpose of our study was to assess postoperative changes in LA size early (1–6 months) after MVr for chronic symptomatic severe isolated degenerative MR and to identify correlates of the extent of postoperative LA size reduction.


Study population

The present study is based on a retrospective analysis of 225 consecutive patients who underwent MVr. Patients with MR of aetiologies other than degenerative (ischaemic (41 patients), rheumatic (six patients), infective endocarditis (three patients), dilated cardiomyopathy (10 patients), associated aortic valve replacement (49 patients), congenital heart disease (10 patients) or incomplete echocardiographic recordings for quantitative measurements before surgery (27 patients)) were excluded. The remaining 79 patients with MVr for chronic symptomatic isolated severe degenerative MR formed the final study group. Preoperative and postoperative echocardiographic studies were available for all of them. In four patients (5%), having as main surgical indication chronic severe MR by mitral degenerative disease, concomitant coronary artery bypass grafting was performed at the time of MVr. Resection of the prolapsing portion of the leaflet, chordal shortening, chordal transfer, neochordae, edge-to-edge technique or leaflet sliding plasty were performed as needed, isolated or in association. A mitral annuloplasty with or without a prosthetic ring was also performed in 75 patients of the study group (95%). In 67 patients (84%) a mitral prosthetic ring was implanted. Surgical reduction of the LA was not performed as additional procedure to MVr in any of the 79 patients studied. The study was approved by the institutional committee on human research.

Clinical variables

Preoperative and postoperative clinical data were collected retrospectively through review of case records. Systemic hypertension was defined by documentation of the clinical diagnosis or by evidence of elevated systolic blood pressure of >140 mm Hg and/or diastolic blood pressure of >90 mm Hg on at least two occasions before surgery in the absence of acute medical illness. Ischaemic heart disease was defined as history of angina, myocardial infarction, coronary revascularisation or angiographic documentation of coronary artery disease (stenosis ⩾50% in any coronary artery by angiography). Surgical data were collected through review of surgical reports.


Two echocardiographic studies were considered for each patient: within 1 month before surgery and within 1–6 months after discharge (mean of 3 (SD 2) months). All measurements were performed offline by a single investigator, blinded to clinical data.

Left atrium linear dimensions were measured in three planes: anteroposterior, lateral and superoinferior at left ventricular end-systole (maximal LA dimensions). LA anteroposterior diameter was measured by 2D-guided M-mode echocardiography from the parasternal short-axis view at the base of the heart, from the trailing edge of the posterior aortic wall to the leading edge of the posterior LA wall.8 The lateral and superoinferior dimensions were measured by 2D-echocardiography from the apical four-chamber view using an inner-edge-to inner-edge measurement. LA area was measured by planimetry of the inner contour from the apical four-chamber view, excluding the confluences of the pulmonary veins and LA appendage from the LA tracing. LA volume was determined using the modified Simpson’s rule from the apical four-chamber view at end-systole, from the frame preceding mitral valve opening. The plane of mitral annulus was taken as the inferior LA border. LA volume and area were indexed to body surface area as recommended.8 A sphericity index was calculated as the ratio between lateral and superoinferior diameters.

LV end-systolic and end-diastolic diameters, and septal and posterior wall thickness at end-diastole, were measured in the parasternal short axis view, using 2-D guided M-mode echocardiography according to recommendations for chamber quantification.8 LV end-systolic and end-diastolic volumes and ejection fraction were determined from the apical four-chamber view according to the modified Simpson’s rule. LV volumes were indexed to body surface area. LV mass calculation was based on Devereux’s formula according to recent guidelines.9

Colour Doppler evaluation of MR was used for semiquantitative estimation of MR severity10 according to previously validated criteria, taking into account the width and depth of regurgitant jets from different views.11 Regurgitation was graded as mild (grade 1) if jet width was judged visually to be one-third or less of LA width; moderate (grade 2) if jet width was visually judged to be between one-third and one-half of LA width; and severe (grade 3) if jet width was at least one-half of LA width. The relation between jet size and LA size was assessed by visual integration of the information obtained from all the different transducer positions, to ensure exploration of the entire receiving chamber and to allow adequate evaluation of the 3-dimensional characteristics of the jet. The interobserver concordance for semiquantitative assessment of MR in our laboratory has been previously reported to be 98%.12

The postoperative peak and mean diastolic pressure gradients across the repaired mitral valve were calculated from the apical four-chamber view using CW Doppler echocardiography, according to the simplified Bernoulli equation. A careful and thorough search for the best image was performed (optimal beam alignment and sample location, complete transmitral jet envelope, proper gain settings).

The intraobserver variability in LA volume measurement was evaluated in 10 randomly selected patients by blindly repeating LA measurements 1 month apart and it was found to be 3.1% (SD 0.8%).

The postoperative reduction in LA volume indexed to body surface area (LAVi) was analysed as per cent reduction from the preoperative LAVi. Based on previously reported data on LA reverse remodelling in patients after MVr,6 LA reverse remodelling was defined as a per cent decrease in LA volume index (LAVi) ⩾15%.

Statistical analysis

Data are presented as mean values (standard deviation (SD)) for continuous variables and as numbers (percentages, %) for categorical variables. Group comparisons were performed using a standard t test or χ2 test, as appropriate. Simple associations between postoperative reduction in LAVi and clinical and echocardiographic variables were estimated and tested by Pearson’s correlation coefficient. A p value <0.05 for a 2-tailed test was considered statistically significant. Linear regression was used to assess the influence of selected continuous variables on postoperative LAVi reduction. Variables included in the multivariate analysis were selected based on their significance in univariate correlations (at a significance level of p<0.10). Statistical analysis was performed with SPSS software (version 13.0) (SPSS Inc., Chicago, Illinois).



Clinical characteristics of the study population are presented in table 1. Prolapsing leaflets, surgical details and associated surgical procedures performed at the time of MVr are shown in table 2. Postoperative residual MR was absent/trivial in 32 patients (41%), mild in 38 patients (48%), moderate in eight patients (10%) and severe in one patient (1%). Postoperative transmitral peak and mean pressure gradients were 6.7 (SD 3.2) mm Hg and 3.2 (SD 1.3) mm Hg respectively. The discharge medication consisted of: ACE inhibitors in 20 patients (25%), β-blockers in 19 patients (24%), and diuretics in 28 patients (35%).

Table 1 Clinical characteristics of the study population
Table 2 Mitral valve lesion, type of repair and associated surgical procedures (n = 79)

Changes in left atrial size after MVr

A severely increased LAVi according to the recent recommendations for chamber quantification (LAVi ⩾40 ml/m2)8 was found preoperatively in 73 patients (92%) and postoperatively in 52 patients (66%) of the study group. LA diameters, LA area and LAVi significantly decreased after MVr (p<0.001). The mean postoperative decrease in LAVi was 29% (SD 18%) with a median value of 27% and a normal distribution in the study population (fig 1). LA reverse remodelling (as previously defined) was found in 63 patients (80%). LA shape, expressed by the sphericity index, did not change after surgery (0.89 (SD 0.08) vs 0.90 (SD 0.07); p = 0.37).

Figure 1 Distribution of postoperative reduction in left atrial volume index (LAVi) in the study group.

Likewise, LV volumes significantly decreased postoperatively (p<0.001) (table 3). A significant decrease was found postoperatively in LV stroke volume indexed to BSA (58 (SD 17) ml/m2 vs 38 (SD 10) ml/m2, p<0.001). The postoperative decreases found in LV stroke volume index and LAVi respectively were similar (20 (SD 16) ml/m2 vs 21 (SD 15) ml/m2) and the two parameters were significantly related (r = 0.26, p = 0.019). However, no correlation was found between LV stroke volume decrease and postoperative LAVi reduction as we defined it.

Table 3 Overall changes of echocardiographic indices postoperatively

Correlates of LAVi reduction

Several clinical and echocardiographic variables were tested for their possible influence on the extent of early postoperative LAVi reduction (table 4).

Table 4 Correlates of postoperative LAVi reduction

Significant correlations were found between postoperative LAVi reduction and: preoperative LAVi (r = 0.30; p = 0.008) (fig 2), systolic blood pressure (r = −0.24; p = 0.03), diastolic blood pressure (r = −0.29; p = 0.009), postoperative transmitral mean pressure gradient (r = −0.38; p = 0.001) (fig 3) and residual MR (r = −0.22; p = 0.04). The reduction in LAVi was significantly higher in normotensive patients (32% (SD 18%) vs 24% (SD 16%), p = 0.031) (fig 4).

Figure 2 Linear correlation between preoperative left atrial volume index and postoperative left atrial volume index reduction (LAVi),
Figure 3 Linear correlation between postoperative transmitral mean pressure gradient and postoperative reduction in left atrial volume index (LAVi).
Figure 4 Postoperative reduction in left atrial volume index (LAVi) was significantly higher in normotensive than in hypertensive patients (p = 0.031). Boxes, interquartile range around the median line; whiskers, range of distribution; circles, outlying values. HTN, systemic hypertension.

Age showed only a trend to a significant relationship with the extent of LAVi reduction in the study group (p = 0.06); however, using age as a dichotomous variable and analysing LAVi reduction in different age subgroups we found a significantly higher reduction of preoperative LAVi in patients ⩽45 years (10 patients) than in those ⩾45 years (63 patients) (43.4% (SD 22%) vs 27.3% (SD 16%); p = 0.008). Figure 5 shows an important LAVi reduction (62%) in a 27-year-old patient.

Figure 5 (A) 27-year-old man with severe, degenerative mitral regurgitation and important preoperative left atrial (LA) enlargement (LA volume  = 72 ml; LA volume index  = 37.11 ml/m2). (B) After mitral valve repair – important LA reverse remodelling corresponding to 62% reduction in LA volume index (LA volume = 27 ml; LA volume index  = 13.91 ml/m2). The two echocardiographic images are performed on the same echocardiographic machine and with the same depth setting (15 cm).

In patients with edge–to-edge MVr the decrease in LAVi was slightly lower (22.8% (SD 18%) vs 31.2% (SD 18%); p = 0.07) and the postoperative transmitral mean pressure gradient was higher (3.6 (SD 1.7) mm Hg vs 3.0 (SD 1.1) mm Hg; p = 0.07) than in patients with other types of repair. However, no significant correlation was found between edge-to-edge technique and postoperative LAVi reduction. No correlation was found between ring size and LAVi reduction or between ring size and postoperative transmitral pressure gradients. There were no differences in postoperative LAVi reduction between patients according to discharge medication.

By entering in multivariate linear regression analysis correlates of LAVi reduction found by univariate analysis (at a significance level of p<0.10) we found that preoperative LAVi (p<0.001), postoperative transmitral mean pressure gradient (p = 0.03) and the degree of residual MR (p = 0.02) were independently related to postoperative LAVi reduction. Age was close to statistical significance, without reaching it (p = 0.07).

The extent of postoperative reduction in LAVi was not related to the time interval between surgery and the moment of the postoperative echocardiographic control (p = 0.85). Dividing the study group into two subgroups according to the time period between MVr and the postoperative control (1–3 months and 4–6 months respectively), we found no statistically significant difference in LAVi reduction between groups : 30% (SD 18%) (20 (SD 13) ml/m2) in group 1 (55 patients) vs 27% (SD 17%) (22 (SD 16) ml/m2) in group 2 (24 patients)(p = 0.50).

Excluding from the analysis the four patients with associated CABG at the time of MVr does not significantly change any of the previous results.


Preoperative LA size was found to be an important predictor of cardiac related mortality after mitral valve replacement for symptomatic chronic MR.5 Long-term predictive value of echocardiographically determined preoperative LA size and/or of its postoperative reduction in patients undergoing MVr for chronic symptomatic severe isolated degenerative MR is unknown. However, a good postoperative LA reverse remodelling is theoretically important, considering the relationship between LA enlargement and the risk of developing atrial fibrillation or stroke.3

Data on postoperative LA remodelling in patients undergoing surgery for chronic MR are scarce. Moreover, the antero-posterior diameter (obtained by M-mode or 2-D echocardiography) was used to assess LA size preoperatively or postoperatively in most of these studies. Recently, LA volume measured by 2-dimensional (2D) or 3-dimensional echocardiography (3D) emerged as a more accurate measure of LA size when compared with reference standards such as MRI or cine-computerised tomography.1315 LA volume is also a better tool to assess changes in LA size in the research field8 16 and has a stronger association than LA diameter with cardiovascular outcomes.17 18 A volumetric assessment of LA reverse remodelling after MVr was performed in two recent studies on small groups of patients (20 patients) with degenerative MR and dilated cardiomyopathy, respectively.6 7 However, to our knowledge, a specific analysis of factors that could interfere with postoperative LAVi reduction after MVr for chronic symptomatic isolated degenerative MR has not yet been performed.

Changes in LA size early after MVr

Our findings support previous data showing a significant reduction in LA size early after MVr for MR.6 7 Using the same cut-off of 15% reduction in preoperative LAVi to define LA reverse remodelling, Westenberg et al6 reported similar results in patients after annuloplasty for MR in dilated cardiomyopathy, as assessed by MRI. Even if study populations are different (dilated cardiomyopathy was an exclusion criterion in our study), both the percentage of patients with early postoperative reverse remodelling (72%) and the extent of LAV reduction in absolute values reported by Westenberg et al are similar to our findings (preoperative LAV 126 (SD 41) ml, postoperative LAV 87 (SD 30) ml in our study group vs preoperative LAV 165 (SD 48) ml, postoperative LAV 109 (SD 23) ml in the above mentioned study).

As calculation of preoperative mitral regurgitant volume was not feasible in all patients, the difference found between preoperative and postoperative LV stroke volumes was used as a surrogate index for preoperative volume overload induced by MR. The postoperative decreases in LV stroke volume index and LAVi respectively were quite similar (20 (SD 16) ml/m2 vs 21 (SD 15) ml/m2) and the two parameters were significantly related. This finding emphasises the relationship between volume overload and LA size and supports the role of postoperative LA volume unloading as primum movens of LA reverse remodelling.

Correlates of LA volume reduction

The main correlates of postoperative LAVi reduction in our study were: preoperative LAVi, systolic and diastolic blood pressure, postoperative transmitral mean pressure gradient and severity of residual MR.

The direct correlation between preoperative LA size and extent of LAVi reduction could be explained by previous histological findings showing myocardial fibre hypertrophy without important changes in the percentage of fibrosis as well as normal myocyte relaxation in enlarged, volume-overloaded atria of dogs with chronic MR.19

A smaller LAVi reduction in patients with hypertension is probably a consequence of impaired LV filling, due to increased LV stiffness in the presence of hypertensive heart disease,20 or to increased arterial stiffness. Indeed, a strong correlation was reported between central arterial stiffness indexes and LA size in hypertensive patients.21 22 LV mass was not significantly related to LAVi reduction in our study population.

The relationship between pressure overload and LA size was previously demonstrated in patients with pure mitral stenosis23 or with LV diastolic dysfunction.24 As the mean values of postoperative transmitral peak and mean pressure gradients are no higher in the present study than those reported by other authors6 25 and generally accepted after MVr, our study demonstrates that even mildly increased transmitral pressure gradients influence the postoperative LAVi reduction and cannot be considered innocent bystanders in patients undergoing MVr.

It would be difficult to assess the role of each surgical technique on postoperative LAVi reduction, having different types of repair and different surgeons.

As a complete subgroup analysis according to the type of MVr was not feasible, postoperative LAVi reduction was only assessed in the subgroup of patients with edge-to-edge MVr (18 patients, 22%). Several studies showed that mean and peak transmitral valve gradients are slightly higher during exercise in patients with degenerative MR who have undergone edge-to-edge technique with or without annuloplasty ring.26 27 A slightly lower postoperative reduction in LAVi and a slightly higher postoperative transmitral mean pressure gradient were observed in our study in these patients than in patients with other types of repair. As there were no other significant differences between the subgroup with edge-to-edge repair and the rest of the study population, the slight difference in LA remodelling could be at least partially explained by an impaired postoperative diastolic filling associated with higher mean transmitral pressure gradients.

No correlation was found between mitral prosthetic ring size and postoperative transmitral pressure gradients or postoperative decrease in LAVi in the study population.

The relatively small number of patients does not allow a definitive conclusion or an extrapolation of these observations to the whole population undergoing surgical repair for chronic, degenerative MR.

An increased severity of residual MR was significantly related in our study to a smaller postoperative LAVi reduction. The relationship between mitral regurgitant volume and LA volume was previously demonstrated.28 In a recent paper Geidel et al correlated the phenomenon of continuous LA size reduction after downsizing of the mitral valve and coronary revascularisation in ischaemic MR with a reliable mitral valve competence after surgery.29 As residual MR is known as an important prognostic factor after surgical correction of MR, the observed correlation between the severity of residual MR and postoperative reduction in LAVi supports the hypothesis of a long-term prognostic role for LAVi reduction after MVr.

As postoperative transmitral mean pressure gradient and the severity of residual MR were independently related to LAVi reduction, and both parameters are related to the surgical technique used for MVr, our study suggests that both parameters should be taken into account when the results of a specific surgical technique are evaluated. That means that better LA reverse remodelling is expected with lower postoperative transmitral mean gradients and with better MV competence. Further studies are needed, on subgroups of patients undergoing the same surgical procedure, performed by the same surgeon, in order to assess the impact of each type of repair on LA reverse remodelling.

Age-induced histological changes in LA wall (hypertrophy and fibrosis) leading to decreased elastic recoil in response to volume unloading could explain a smaller LAVi reduction in patients >45 years. An advanced age at operation also suggests a longer history of MR with less reversible morphological changes of LA wall. A better LAVi reduction in patients <45 years supports the benefits of early surgery for severe chronic degenerative MR.

We acknowledge several limitations of our study. The study was based on retrospective data. The influence of persistent atrial fibrillation on early postoperative LA reverse remodelling could not be analysed because of the small number of patients with preoperative persistent atrial fibrillation (eight patients). However, the low frequency of preoperative persistent atrial fibrillation in our study group (10%) is close to that (15%) reported recently by Berbarie and Roberts in a larger population (246 patients) who had mitral valve repair or replacement for MR secondary to mitral valve prolapse.30

The severity of MR (preoperatively) and the mitral annuloplasty (postoperatively) did not allow a reliable assessment of LV diastolic function using transmitral peak diastolic velocities and E wave deceleration time. Annular diastolic velocities by tissue Doppler imaging, less prone to the confounding effect of haemodynamic variables and changes in loading conditions, were not available in all patients.

The clinical relevance of LAVi reduction early after surgery in patients after MVr for degenerative MR needs to be demonstrated.


CCB, BAP and ACP were supported by Research Fellowships granted by the Association for Research in Cardiology, ARC, Pordenone, Italy.


View Abstract


  • Competing interests: None.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Correction
    BMJ Publishing Group Ltd and British Cardiovascular Society