Article Text
PDF
Abstract
Background Left atrial (LA) volume is an independent prognosticator in various cardiac diseases. The authors assessed the changes of LA volume after successful percutaneous mitral valvuloplasty (PMV) and the impact of LA enlargement on long-term clinical outcome after PMV.
Methods and Results From a prospective PMV registry started in 1988, 303 patients (242 women, age: 39.3±10.8 years) who had undergone successful PMV were followed for 4-20 years (median 11 years). Echocardiographic examination including LA volume measurement was performed before PMV and repeated after PMV. LA volume decreased from 92±50 to 69±42 ml (p < 0.001) immediately after PMV and remained stationary until 1 year after PMV. Since then, LA volume subsequently increased exceeding the pre-PMV level by 8 years after PMV. Multivariate analysis showed that LA volume increase at 10 years after PMV was independently related to the post-PMV mitral valve area, the echo score, the presence of atrial fibrillation and post-PMV LA volume. On multiple regression analysis, pre-PMV LA volume and percentage change of LA volume immediately after PMV emerged as independent predictors of event-free survival along with age, pre-PMV tricuspid regurgitation and post-PMV mitral valve area. Ten-year survival rate was 93% in patients with smaller LA before PMV (≤72 ml/m2), whereas it was only 60% in those with larger LA (>72 ml/m2).
Conclusions Progressive increase of LA volume was observed even after successful PMV. Larger pre-PMV LA volume was associated with poor prognosis.
- Cardiac remodelling
- valvuloplasty
- mitral stenosis
Statistics from Altmetric.com
Increase of left atrial (LA) size is an important part of cardiac remodelling observed in various cardiovascular diseases. Because LA enlargement reflects the severity and chronicity of the underlying cardiac diseases, LA size is regarded as a robust predictor of cardiovascular outcome.1–3 Therefore, detecting and quantifying LA enlargement is an important part of the cardiac evaluation.
Percutaneous mitral valvuloplasty (PMV) has been established as an effective treatment option for haemodynamically significant mitral stenosis (MS) in patients with suitable valvular morphology.4–8 It is well known that long-term prognosis after PMV are dependent upon the clinical, haemodynamic and echocardiographic features at baseline.6 9 10 Larger studies have shown that the independent predictors of prognosis include age, New York Heart Association functional class, echocardiographic deformity score, post-PMV mitral regurgitation (MR) and pulmonary artery pressure.6 11 However, the impact of pre-PMV LA enlargement on the long-term outcome has not been fully elucidated. Moreover, there has been few studies examining the changes of LA volume during long-term follow-up after successful PMV and the impact of LA remodelling on clinical outcome, although it has been suggested that successful PMV decreases the intensity of spontaneous echo contrast in LA, reduces the LA size, and improves LA mechanical function.11 12 In the present study, we reviewed our prospective PMV registry data to evaluate the impact of pre-PMV LA enlargement on clinical outcome. In addition, we also evaluate the immediate and long-term LA volume changes after PMV and identify their determinants and impact on clinical outcome.
Methods
Study population
We started a prospective registry for PMV in 1988 and enrolled consecutive patients with moderate to severe MS undergoing PMV at our hospital.13 Standard case report form was obtained in all patients, which included demographics, symptomatic status, and echocardiographic and cardiac catheterisation data. From April 1988 to December 2005, a total of 438 patients who underwent PMV as a first procedure were enrolled. Among 438 patients, we included only patients with successful results of PMV defined as complication-free procedure with postprocedure mitral valve area (MVA) ≥1.5 cm2 by echocardiography and postprocedure MR grade ≤2/4 by left ventriculography according to Sellers classification. The procedure was unsuccessful in 44 patients (significant MR in 21 patients, suboptimal postprocedure MVA in 19, systemic thromboembolic events including stroke in 4). In addition, patients with significant aortic valve disease (n=17), renal replacement therapy (n=5), history of cerebrovascular accident within 6 months before PMV (n=8) or history of myocardial infarction (n=3) were excluded. Patients without available LA volume data (n=58) were also excluded. Finally, the remaining 303 patients were followed for 4–20 years with median 11 years and were the subjects of this study. The study protocol was approved by the Institutional Review Board of Seoul National University Hospital.
Echocardiographic evaluation
Comprehensive 2-dimensional and Doppler echocardiographic examinations were performed in all patients 1–2 weeks before the procedure using commercially available echocardiographic devices with a 2.5 MHz transducer. MVA was determined by the 2-dimensional planimetry and the pressure half-time method. Mitral valve morphology was evaluated by grading valve leaflet thickening, mobility, calcification and subvalvular thickening, each on a scale from 1 to 4 according to the severity of Wilkins and colleagues.14 The four separate scores were added up for the echo score. MR and tricuspid regurgitation (TR) were detected and semiquantitatively graded with colour flow imaging. Pulmonary artery systolic pressure was calculated from continuous wave Doppler of TR jet with 10 mm Hg added for the right atrial pressure. Maximal LA volume was measured at the end-systole using the prolate ellipsoid model15 as LA volume=0.523×D1×D2×D3; D1 denotes anteroposterior dimension measured from the parasternal long-axis view and D2 and D3 denote mediolateral and superoinferior dimensions measured from the apical 4-chamber view, respectively. The echocardiographic examinations were repeated on the following day after PMV and annually or biennially thereafter up to 20 years.
Percutaneous mitral valvuloplasty
PMV was performed using the stepwise Inoue balloon technique16 while monitoring the conventional haemodynamic parameters. Right-sided cardiac catheterisation was performed with a Cournand catheter (USCI, Billerica, Massachusetts). Right and left heart pressure measurements, pulmonary artery pressure measurement, cardiac output and diagnostic oxygen saturation run were performed before and after PMV. MR was assessed by a left ventriculography after the procedure and graded using Sellers classification.
Follow-up studies
Upon every visit to the echocardiographic laboratory for follow-up echocardiography, standard case report form was completed, which included echocardiographic data and symptomatic status. In addition, review of patients' hospital records and telephone interviews were performed to assess the occurrence of cardiovascular events, which included death of any cause, systemic embolism including stroke, readmission due to heart failure aggravation, mitral surgery, and redo PMV.
Statistical analysis
Continuous variables were described as mean and SD, and the paired t test was used to analyse the changes after PMV. Categorical variables were reported as proportions and were compared with the χ2 test.
Each of successively measured LA volumes after PMV was compared with the baseline values using the paired t test. In these pairwise comparisons, results with p<0.007 (=0.05/7), which was calculated according to Bonferroni correction to avoid the inflation of alpha error by multiple comparisons, were considered statistically significant.
The determinants of LA reverse remodelling immediately after PMV were evaluated with the linear regression model, and the Cox proportional hazards model was used to investigate which factors affected overall survival and event-free survival. To adjust the effect of confounding factors, we exploited multivariate linear regression and Cox regression models with stepwise selection. Age, sex and variables, which have p<0.1 in the univariate analysis, were incorporated into the multivariate analysis.
The best cut-off values of LA volume index for the prediction of cardiovascular events was calculated with the R package Maxstat (http://cran.r-project.org/web/packages/maxtat/index.html) using a maximal χ2 method of Miller and Siegmund.17 Survival curves of the dichotomised patients according to the cut-off value of LA volume index were plotted with the Kaplan–Meier method. Survival was compared between these two groups using log-rank test.
Statistical analyses were performed using Statistical Package for the Social Sciences version 17 (SPSS Inc, Chicago, Illinois) and R version 2.8.1 (R Development Core Team, http://www.R-project.org). All p values are two-sided, and results with p<0.05 were considered statistically significant except the multiple comparisons of serial LA volume as previously described.
Results
Baseline characteristics
The baseline clinical and echocardiographic characteristics of the study population were summarised in table 1. Mean (SD) age was of 39.3 (10.8) years, and 242 (80%) patients were women. Pre-PMV heart rhythm was atrial fibrillation (AF) in 93 (31%) patients. Mean echo (SD) score was 6.7 (1.6). All the patients were symptomatic and 79 (26%) were in New York Heart Association functional class III or IV. The degree of TR was equal to or more than moderate in 33 (31%) patients.
Baseline characteristics of the patients (n=303)
Haemodynamic and echocardiographic data
Table 2 shows the haemodynamic and echocardiographic findings at baseline, immediate post-PMV and last follow-up (median, 9 years (mean (SD), 9 (5) years ; range, 2–19 years)) after PMV. LA pressure, mean mitral gradient and pulmonary artery systolic pressure decreased significantly after PMV with corresponding increase in MVA from 1.0±0.3 to 2.2±0.7 cm2 (p<0.001) as calculated by Gorlin formula and from 0.9±0.2 to 1.8±0.3 cm2 (p<0.001) as measured by 2-dimensional echocardiography. The MVA at last follow-up was 1.5±0.4 cm2. Post-PMV left ventricular angiography revealed the grade 2 MR in 17 patients.
Haemodynamic and echocardiographic data at baseline, immediate post-PMV, and last follow-up
Changes of LA volume after PMV
Baseline echocardiographic results showed a mean (SD) LA volume of 92.5 (50.6) ml and mean LA volume index of 60.2 (33.1) ml/m2. Immediately after PMV, LA volume and LA volume index decreased by 24% (19%) and 25% (19%) (69.3 (42.5) ml and 44.8 (26.7) ml/m2), respectively, in comparison to baseline (p<0.001). Patients with sinus rhythm showed 25% (18%) decrease of LA volume, from 74.8 (28.2) ml to 54.9 (22.9) ml, and those with AF showed 21% (19%) reduction, from 132.3 (69.9) ml to 101.4 (56.5) ml (p<0.001, table 2, figure 1 2).
Changes in left atrial (LA) volume after percutaneous mitral valvuloplasty (PMV) according to time.
Percentage of left atrial (LA) volume change after percutaneous mitral valvuloplasty (PMV) related to cardiac rhythm. B, Before PMV; I, Immediate after PMV.
LA volume decreased significantly until 1 year after PMV, and it subsequently increased exceeding the pre-PMV level by 8 years after PMV (91.0 (45.9) ml in before PMV vs 97.0 (54.5) ml in post-PMV 8 years; 163 patients; p<0.001, figure 1). Likewise, LA volume index decreased significantly after PMV as shown in table 2. When the percentage change of LA volume was represented as ((post-PMV LA volume−pre-PMV LA volume)/pre-PMV LA volume)×100, as shown in figure 2, the increase of LA volume over time was similarly observed regardless of cardiac rhythm.
Determinants of LA volume changes
The determinants of LA reverse remodelling immediately after PMV were defined by means of linear regression analysis. Pre-PMV LA volume, cardiac rhythm, TR grade, pre-PMV mean mitral gradient and age were independent predictors of LA reverse remodelling in the multivariate analysis (table 3). In other words, patients with larger LA, sinus rhythm, lower TR grade, higher mitral gradient or younger age showed greater reduction of LA size immediately after PMV.
Factors determining LA volume reduction immediately after procedure (stepwise multiple linear regression)
In 167 patients for whom follow-up echocardiogram was available at 10 years, 10-year LA volume change was calculated in comparison to LA volume immediately after PMV (LA volume at 10 years after PMV–LA volume immediately after PMV). Multivariate analysis showed that LA volume increment at 10 years was independently related to the post-PMV MVA and the echo score and more strongly related to the presence of AF and immediate post-PMV LA volume (table 4). In contrast, age did not relate to LA volume changes during long-term follow-up. In other words, progressive LA remodelling was greater in patients with AF, larger immediate post-PMV LA volume, smaller post-PMV MVA or higher echo score.
Factors determining LA volume increment at 10 years after PMV (Stepwise multiple linear regression)
Predictors of clinical outcome
During median follow-up of 11.0 years (mean (SD), 10 (5) years; range, 4–20 years), 168 cardiovascular events occurred in 106 (35%) patients. Those events included 50 deaths, 46 readmissions due to heart failure, 12 systemic embolisms including cerebral infarction, 45 mitral valve surgeries, and 15 re-PMVs. The causes of death were surgical complication after mitral valve replacement (MVR) in four, stroke in five, acute myocardial infarction in three, heart failure in three, cancer in two and infectious disease in four. Despite every effort, we could not identify the cause of death in 29 patients. In univariate analysis, pre-PMV LA volume index (per 10 ml/m2 increase) was significantly related to total mortality and cardiovascular events with hazard ratio (HR) of 1.13 and 1.14 (both p<0.001), respectively. In addition, age (1.07 and 1.04, p<0.001 and p<0.001), AF (4.32 and 2.02, p<0.001 and p<0.001), male sex (2.21 and 1.50, p=0.015 and p=0.010), post-PMV MVA (0.28 and 0.30, p=0.02 and p=0.001), pre-PMV TR grade ≥ moderate (3.59 and 3.17, p<0.001 and p<0.001), pre-PMV left ventricular ejection fraction (0.94 and 0.96, p=0.01 and p=0.02), echo score (1.11 and 1.16, p=0.22 and p=0.01), post-PMV mitral gradient (1.09 and 1.07, p=0.05 and p=0.04), post-PMV pulmonary artery pressure (1.04 and 1.03, p=0.02 and p=0.01) and percentage change of LA volume immediately after PMV (1.48 and 3.64, p=0.57 and p=0.009) were related with total mortality and occurrence of cardiovascular events.
The independent predictors of long-term mortality and combined events were identified using multiple stepwise Cox regression analysis as shown in tables 5 and 6. After adjustment for multivariates, a 10 ml/m2 incremental increase in LA volume index was associated with strong risk of cardiovascular events with an HR of 1.11 (95% CI 1.07 to 1.14, p=0.004). The best cut-off values of LA volume index for the prediction of cardiovascular events were 72 ml/m2 for overall survival and 55 ml/m2 for event-free survival. The Kaplan–Meier curves showed significant differences between the two groups according to the cut-off value of pre-PMV LA volume index (p<0.001) (figure 3A,B). Ten-year survival rate was 93% in patients with smaller LA (≤72 ml/m2) before PMV, whereas it was only 60% in those with larger LA (>72 ml/m2). Patients with a higher LA volume index had a 3.2 times higher risk of death (HR 3.22, CI 1.56 to 6.63, p<0.001) and a 2.0 times higher risk of an adverse outcome (HR 1.97, CI 1.32 to 2.91, p<0.001) than patients with a smaller LA volume index.
Independent predictors of long-term mortality (Cox regression analysis)
Independent predictors of long-term cardiovascular events (Cox regression analysis)
Kaplan–Meier survival curves of (A) cardiac death with pre-percutaneous mitral valvuloplasty (PMV) left atrial (LA) volume index ≤72 ml/m2 and LA volume index >72 ml/m2, (B) combined cumulative incidence of clinical events (death, admission of heart failure (HF), stroke, re-PMV, MVR) with pre-PMV LA volume index ≤55 ml/m2 and LA volume index >55 ml/m2.
New onset AF occurred during follow-up in 60 patients (20%). The occurrence of AF was significantly lower in patients with smaller LA (≤55 ml/m2) compared to those with larger LA (>55 ml/m2) (27/141, 19% vs 33/69, 48%, p<0.001).
Discussion
The principal findings of the present study were (1) progressive increase of LA volume was observed even after successful PMV, (2) the presence of AF, greater immediate post-PMV LA volume, lower post-PMV MVA, and higher echo score were independent predictors of progressive LA remodelling during long-term follow-up, (3) pre-PMV LA volume index and percentage change of LA volume immediately after PMV were independent predictors for long-term prognosis after PMV along with age, pre-PMV TR and post-PMV MVA in patients with severe MS. To the best of our knowledge, this is the first study to demonstrate long-term LA volume changes after PMV and their prognostic implication in patients with MS.
We observed that LA dilation was not completely reversed immediately after PMV and LA volume progressively increased over time thereafter. We found that smaller post-PMV MVA, greater immediate post-PMV LA volume, the presence of AF, and higher echo score were associated with greater increase of LA volume during long-term follow-up. Based on these findings, we can assume the underlying mechanisms of progressive LA remodelling. First, smaller post-PMV MVA as an independent determinant suggests that residual mild pressure overload on the LA may cause progressive LA remodelling. Second, greater immediate post-PMV LA volume as an independent determinant underscores the importance of LA structural changes. Chronic pressure overload on LA induces not only LA dilation but also myocardial hypertrophy, myocardial cell loss and interstitial fibrosis,18 which are often irreversible. Thus, LA dilation was only partially reversed and progressive LA remodelling occurred even after substantial relief of the pressure overload. Third, combined AF is an important cause of progressive LA remodelling. AF is very common in MS, affecting almost 40% of all patients. Although AF may be initiated by atrial dilation at the beginning,19–21 chronic AF itself induces mechanical and electrical remodelling of LA leading to further atrial dilatation.22 23 In our study, 93 (31%) patients had AF at baseline before PMV, and 60 (20%) patients had new onset AF during follow-up. The occurrence of AF was significantly lower in patients with smaller LA (≤55 ml/m2) compared to those with larger LA (>55 ml/m2) (27/141, 19% vs 33/69, 48%, p<0.001). Therefore, PMV before significant LA dilation may be beneficial in preventing the occurrence of AF. It is interesting to note that sinus conversion did not occur in any patients with AF. According to the given findings, we believe that earlier and more complete relief of the obstruction may prevent irreversible LA structural changes and improve post-PMV reverse remodeling.
Ageing might have affected the changes in LA size. Kitzman and Edwards24 showed that the endocardium of the LA undergoes physiological cellular transformation with ageing. Few previous studies have demonstrated that atrial size naturally increases with ageing.25 26 However many observational studies have shown differing reports regarding the relationship between LA size and ageing.27–30 Thomas et al27 and Casaclang-Verzosa et al.31 demonstrated that LA size does not change with chronologic ageing alone. Rather, LA enlargement and impairment of LA function suggest pathologic conditions that accompany ageing. In this study, we found that ageing was not independently related to the LA volume change during the long-term follow-up.
It is well known that LA enlargement is a strong predictor of adverse cardiovascular outcomes in various cardiac diseases.1–3 32 33 LA dysfunction directly causes adverse cardiovascular events because it exerts an important function to maintain cardiac output. In addition, LA dilation is strongly associated with LA thrombus, an important source of systemic embolisation. Furthermore, LA dilation is a marker of the severity and chronicity of left ventricular dysfunction. In this study, we observed that LA volume index represents a strong independent predictor of cardiovascular events after PMV. The Kaplan–Meier curve showed clear difference in overall survival and event-free survival between patients with smaller LA and those with larger LA. In terms of overall survival, LA volume index of 72 ml/m2 was the best cut-off value. Patients with a higher LA volume index had a 3.2 times higher risk of death. Ten-year survival rate was 93% in patients with LA volume index <72 ml/m2, whereas it was only 60% in those with LA volume index ≥72 ml/m2. In addition, greater reduction of LA volume immediately after PMV was an independent predictor of clinical events. These findings underscore the importance of LA volume as an independent prognosticator in patients undergoing PMV. We also found that age, pre-PMV TR and post-PMV MVA were independent predictors of clinical events in these patients. Therefore, in addition to the symptomatic status and pulmonary artery pressure as recommended in the current guideline, we believe that these factors may need to be incorporated in regard to deciding the optimal timing for PMV.
Study limitations
The study was conducted in a single tertiary referral hospital. Although we prospectively enrolled consecutive patients, referral bias could not be excluded and thus, the results might be difficult to be generalised. However, considering the wide range of clinical and echocardiographic parameters in our study population, the validity of our study might not be altered.
Conclusions
Progressive increase of LA volume was observed even after successful PMV. Greater post-PMV LA volume along with the presence of AF, lower post-PMV MVA, and higher echo score was independent predictor of progressive LA remodelling. In addition, greater pre-PMV LA volume was associated with poor long-term prognosis. Our findings suggest that PMV should be considered before excessive LA enlargement occurs to prevent progressive LA remodelling and improve long-term prognosis. Future studies are warranted to confirm our results and deepen our understanding of the natural history of LA remodelling after PMV and the impact of such changes on outcomes.
Acknowledgments
This study was supported by a grant from Korea Institute of Medicine and the Korea Healthcare technology R&D Project, Ministry for Health, Welfare, and Family Affairs, Republic of Korea (A090458).
References
Footnotes
Competing interests None.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the Institutional Review Board of Seoul National University Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.