Impact of Left Atrial Volume on Clinical Outcome in Organic Mitral Regurgitation

Study eligibility

Between 1990 and 2000, we prospectively enrolled patients: 1) with at least mild organic MR (intrinsic mitral valve lesions); 2) in sinus rhythm at baseline; and 3) in whom triple quantitation of LA volume, LV volumes, and MR severity was available. Patients with papillary muscle rupture, functional MR, associated mitral valve stenosis of any degree, associated aortic or congenital heart disease, or previous valvular surgery were excluded. Patients with symptoms or decreased EF in the context of organic MR were not excluded. The study was approved by the Mayo Clinic Institutional Review Board.

Doppler-echocardiographic measurements

Comprehensive Doppler and 2-dimensional echocardiography provided data prospectively recorded and used as measured at diagnosis without alteration. Patients underwent triple quantitation of LV characteristics (volumes and mass), LA volume, and MR. Left ventricular volumes (biplane modified Simpson rule) and EF were measured from apical views. Left atrial volume was calculated using the apical biplane area-length method (14,22) and indexed to body surface area (LA index). Left atrium is enlarged with LA index ≥40 ml/m², a threshold previously defined in a normal population (14). Accordingly, patients were stratified in 3 subgroups, as LA index within normal range (<40 ml/m², n = 158), mildly enlarged LA index (40 to 59 ml/m², n = 160), and severely enlarged LA index (≥60 ml/m², n = 174).

Mitral regurgitation was quantified by at least 2 of 3 validated methods, averaged to calculate regurgitant volume (RVol) and effective regurgitant orifice area (ERO). Methods used were the proximal-isovelocity-surface-area method (23) in 391 patients, quantitative Doppler method (using mitral and aortic stroke volumes) in 485 and quantitative 2-dimensional method (using LV volumes and aortic stroke volume) in all patients (22). Because ERO was not measurable in 32 patients, RVol, which has similar prognostic value as ERO (8) and was consistently measurable in all patients, was used as measure of MR severity (mild MR: <30 ml/beat, moderate MR: 30 to 59 ml/beat, and severe MR: ≥60 ml/beat) (24). Mitral flow profile recorded, from apical views at leaflets' tips, diastolic E and A velocities, E-wave deceleration time, and E/A ratio. Tricuspid regurgitant velocity (continuous-wave Doppler) allowed calculation of pulmonary artery systolic pressure.

Clinical assessment and follow-up

Baseline clinical characteristics were recorded at diagnosis. Symptomatic patients were those with frank dyspnea with exercise or at rest. Coronary disease diagnosis was based on a history of myocardial infarction or on angiography during the episode of care. Integrated Charlson comorbidity score was calculated. Clinical outcome was monitored by return visits, mailed questionnaires, review of medical records, and by telephone calls to patients, referring cardiologists, and primary care physicians. The patient's personal physician using all information available conducted, independently of investigators, clinical management and decided follow-up and surgical timing, guided by available recommendations. Cause of death was classified as cardiovascular or noncardiovascular according to death certificates, next-of-kin interview, and hospital and autopsy records. Congestive heart failure was diagnosed during follow-up using a combination of major and minor criteria recommended by the Framingham Heart Study (25). Electrocardiogram confirmation of AF was required.

Statistical analysis

Continuous variables are summarized as mean ± SD and compared between groups using 1-way analysis of variance or Student ttests. Post hoc comparisons used Bonferroni tests. Non-normally distributed variables (Charlson index) comparisons used nonparametric Kruskal-Wallis test. Categorical data were compared using chi-square tests. End points with conservative management were overall survival and combined cardiac events end point (cardiac death/AF/congestive heart failure), with follow-up censored at mitral surgery if performed. Surgery's effect on outcome was analyzed as time-dependent covariate using the entire follow-up, and survival curves were constructed accordingly. Event rates, estimated using Kaplan-Meier method, were compared using log-rank test. Observed mortality was compared to that expected in the U.S. population using log-rank test. Univariate and multivariate Cox proportional hazards analyzed LA index prediction of mortality and cardiac events with calculation of hazard ratios (HRs). Multivariate analyses adjusted for known predictive variables (age, sex, symptoms, EF, regurgitant volume). A value of p < 0.05 was considered statistically significant.

Baseline characteristics

Baseline characteristics (Table 1)demonstrate usual predominance of degenerative MR (prolapse in 81%, rheumatic in 5%, and miscellaneous in remainder of patients), in older men with rare symptoms at presentation. Mitral regurgitation grade was mild in 110 (22%), moderate in 110 (22%), and severe in 272 (56%) patients by quantitative criteria (24). Regurgitant volume averaged 68 ± 42 ml, ERO 42 ± 29 mm², LV diameter index 31 ± 4 mm/m²(end diastole) and 19±4 mm/m²(end systole), LA volume 103 ± 49 ml, and LA index 55 ± 26 ml/m². Left atrial index was 67 ± 26 ml/m²in severe MR compared with 47 ± 16 ml/m²(p < 0.0001) in moderate MR. Medical therapy included angiotensin-converting enzyme inhibitors in 225 (46%) patients, angiotensin-receptor blocker in 110 (22%) patients, calcium-channel blocker in 135 (27%) patients, diuretic in 188 (38%) patients, and beta-blocker in 177 (36%) patients.

Baseline characteristics stratified by LA index <40, 40 to 59, and ≥60 ml/m²are shown Table 1. With higher LA index, patients were older, were more often men and symptomatic, had more severe MR mostly related to prolapse, and had larger LV remodeling and hypertrophy. Despite LA enlargement, higher LA index was associated with higher pulmonary artery systolic pressure and early filling due to MR. Of note, patients with larger LA had no excess coronary disease, comorbidity (Charlson index mean and 25% to 75% quartiles: 0.75 [0 to 1], 0.82 [0 to 1], 0.78 [0 to 1] for LA index <40, 40 to 59, and ≥60, p = 0.68) or hypertension.

LA index and survival

During follow-up, 54 patients died under conservative management (5-year survival: 80 ± 2.9%) and 35 of cardiovascular cause (5-year cardiovascular mortality: 14 ± 2.6%). In univariate analysis (Table 2),LA enlargement strongly predicted mortality under conservative management (HR: 1.3 [95% confidence interval (CI): 1.2 to 1.5] per 10 ml/m²LA index increment, p < 0.0001), specifically with LA index ≥60 ml/m²but not significantly with LA index 40 to 59 ml/m². Patients with LA index ≥60 ml/m²incurred lower 5-year survival (53 ± 8.6%) compared with LA index 40 to 59 ml/m²(84 ± 4.8%) and <40 ml/m²(90 ± 3%, p < 0.0001) (Fig. 1).Difference in survival between LA index 40 to 59 and <40 ml/m²did not reach significance (p = 0.26).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Survival After Diagnosis According to LA Volume

Survival after diagnosis under medical management in patients with organic mitral regurgitation in sinus rhythm at baseline according to stratification by left atrial volume indexed to body surface area (left atrial index <40, 40 to 59, or ≥60 ml/m²). Note that patients with left atrial index ≥60 ml/m²incur lower survival than patients with smaller left atrium and also incur excess mortality compared with that expected in the normal population (5-year observed survival: 53% vs. 76% expected, p = 0.017). LA index = left atrial volume indexed to body surface area.

Comparison to expected survival in the population showed no difference with LA index <40 ml/m²(expected 88%, p = 0.81) or 40 to 59 ml/m²(expected 86%, p = 0.90). However, survival was significantly lower than expected in patients with LA index ≥60 ml/m²(expected 76%, p = 0.017).

In multivariate analysis, adjusting for age, sex, symptoms, EF, and regurgitant volume, LA index independently predicted survival (HR: 1.3 [95% CI: 1.1 to 1.5] per 10 ml/m²increment and 2.8 [95% CI: 1.2 to 6.5] for LA index ≥60 ml/m²) (Table 2). After further adjustment for Charlson comorbidity index, LA index predictive value remained unchanged (HR: 2.9 [95% CI: 1.3 to 6.7], p = 0.009).

LA index and cardiac events

During follow-up with medical management, 32 patients experienced new onset of AF (5-year: 10.9 ± 2.1%) and 49 patients experienced heart failure (5-year: 16.9 ± 2.6%). Thus, 86 patients experienced a cardiac event (5-year: 28 ± 3%) during follow-up under conservative management.

In univariate analysis (Table 2), LA enlargement strongly predicted cardiac events under conservative management. Patients with LA index ≥60 ml/m²had frequent cardiac events (5-year: 63 ± 8%) versus 40 to 59 ml/m²(31 ± 6%) and <40 ml/m²(9.7 ± 3%, p < 0.0001) (Fig. 2).Patients with LA index 40 to 59 ml/m²had cardiac events more frequent than with LA index <40 ml/m²(p = 0.002) but lower than with LA index ≥60 ml/m²(p < 0.0001).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Cardiac Events After Diagnosis According to LA Volume

Rates of cardiac events (cardiac death, atrial fibrillation, or congestive heart failure) after diagnosis under medical management in patients with organic mitral regurgitation in sinus rhythm at baseline according to stratification by left atrial volume indexed to body surface area (left atrial index <40, 40 to 59, or ≥60 ml/m²). Note that all event rates are significantly different from each other (all p < 0.01). Abbreviation as in Figure 1.

In multivariate analysis, adjusting for age, sex, symptoms, EF, and regurgitant volume, LA index independently predicted cardiac events under conservative management (Table 2). The LA index link to cardiac events remained unchanged after further adjustment for Charlson comorbidity index (HR: 5.4 [95% CI: 2.7 to 11.0] for LA index ≥60 ml/m²), coronary artery disease, LV mass, or pulmonary pressure (all p < 0.0001).

Subgroup analysis

Outcome analysis was stratified by RVol and LA index. Among 110 patients with RVol 30 to 59 ml, 18 had LA index ≥60 and 92 LA index <60 ml/m², and among 272 patients with RVol ≥60 ml/beat, 151 had LA index ≥60 and 121 LA index <60 ml/m².

Higher mortality under conservative management associated with LA index ≥60 ml/m²was considerable in patients with RVol ≥60 ml/beat (5-year: 54 ± 10% vs. 10 ± 6%, p = 0.0009) but did not reach significance with RVol 30 to 59 ml/beat (5-year: 38 ± 18% vs. 17 ± 6%, p = 0.39) (Fig. 3).In patients with RVol ≥60 ml/beat, observed long-term mortality was higher than expected with LA index ≥60 ml/m²(expected 27%, p = 0.041) but was not different with LA index <60 ml/m²(expected 16%, p = 0.11). Similarly, with RVol 30 to 59, observed long-term mortality was higher than expected with LA index ≥60 ml/m²(expected 19%, p = 0.0023) but was not different with LA index <60 ml/m²(expected 14%, p = 0.72).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Mortality Associated With LA Enlargement, Stratified by Regurgitant Volume

Overall mortality under medical management according to left atrial index (<60 or ≥60 ml/m²) in patients with moderate (A)or severe (B)mitral regurgitation at baseline. Note that mortality is particularly in patients with both severe mitral regurgitation and severe left atrial enlargement. Abbreviation as in Figure 1.

Cardiac events were more frequent with LA index ≥60 ml/m²whether they had RVol ≥60 ml/beat (5-year: 62 ± 9% vs. 31 ± 9%, p = 0.0014) or 30 to 59 ml/beat (5-year: 79 ± 13% vs. 17 ± 5%, p < 0.0001) (Fig. 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Cardiac Events Associated With LA Enlargement, Stratified by Regurgitant Volume

Rates of cardiac events (cardiac death, atrial fibrillation, or congestive heart failure) under medical management according to left atrial index (<60 or ≥60 ml/m²) in patients with moderate (A)or severe (B)mitral regurgitation at baseline. Note that in moderate and in severe mitral regurgitation, marked left atrial enlargement is associated with high event rates. Abbreviation as in Figure 1.

Patients with severe MR (RVol ≥60 ml/beat) and symptoms or EF <60% (n = 52) were followed medically for 0.5 ± 1.16 years and surgery was delayed because patients were initially felt to be well under medical treatment. Excluding these patients, and analyzing only patients without class I surgical indication by guidelines, results were unchanged. Indeed, LA index predicted mortality (HR: 1.25 [95% CI: 1.07 to 1.44] per 10 ml/m², p < 0.01 or 2.9 [95% CI: 1.24 to 6.5] for LA index ≥60 ml/m², p = 0.015) and cardiac events (HR: 1.29 [95% CI: 1.15 to 1.44] per 10 ml/m², p < 0.0001 or 5.4 [95% CI: 2.6 to 11.4] for LA index ≥60 ml/m², p < 0.0001 and 2.8 [95% CI: 1.5 to 5.4] for LA index 40 to 59 ml/m², p < 0.001).

Mitral valve surgery and outcome

Mitral surgery was ultimately performed in 265 patients, involving valve repair in 240 patients (90.6%) and valve replacement in 25 patients (mechanical: n = 18, bioprosthesis: n = 7). Forty patients (15.1%) underwent concomitant coronary bypass surgery and 7 had a Maze procedure. As expected, patients who underwent surgery after diagnosis versus those who remained medically managed, had larger RVol (89 ± 37 ml vs. 43 ± 32 ml, p < 0.001) and LA index (63 ± 25 ml/m²vs. 46 ± 23 ml/m², p < 0.001). During total follow-up (including pre- and post-operative period), 80 patients died (survival: 85 ± 1.8% 5 years after diagnosis), and 26 patients died after mitral surgery (survival: 91 ± 2.0% 5 years after surgery). New onset AF occurred in 26 patients and heart failure in 15 patients. There was no difference in post-operative outcome after stratification by pre-operative LA index ≥60 or <60 ml/m²(5-year post-operative mortality: 9.1 ± 2.0% vs. 8.7 ± 2.8%, p = 0.98; and cardiovascular events: 20 ± 3.9% vs. 16.8 ± 3.9%, p = 0.34).

Cox proportional hazards analysis with mitral surgery as time-dependent covariate demonstrated that surgery was associated with improved survival univariately (HR: 0.46 [95% CI: 0.29 to 0.75]; p = 0.0016). Accounting for the younger age and more frequent male sex of operated patients, in multivariate analysis adjusted for age, sex, symptoms, regurgitant volume, EF, and LA index, mitral surgery remained associated with improved survival (HR: 0.46 [95% CI: 0.26 to 0.84]; p = 0.01). Likewise, surgery was associated with decreased cardiac events univariately (HR: 0.56 [95% CI: 0.37 to 0.83]; p = 0.0045) and after adjustment (0.38 [95% CI: 0.23 to 0.62]; p = 0.0001). In patients with LA index ≥60 ml/m², mitral surgery (as time-dependent variable) was particularly beneficial compared with medical management, with decreased mortality (9 ± 3% vs. 47 ± 9%; HR: 0.19 [95% CI: 0.09 to 0.38]; p = 0.001) (Fig. 5A)and cardiac events (22 ± 5% vs. 63 ± 8%; HR: 0.22 [95% CI: 0.12 to 0.39]; p = 0.001) (Fig. 5B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Outcome in Patients With Markedly Enlarged LA Compared Between Surgical and Medical Management

Overall mortality (A)and cardiac events (B)over the entire follow-up compared between surgical management performed during follow-up (surgery is a time-dependent variable) and medical management in patients with LA index ≥60 ml/m²at baseline. Note that for both mortality and cardiac events, there is a marked outcome benefit to the surgical approach in these patients with markedly dilated LA. Abbreviation as in Figure 1.

Influence of MR on LA remodeling

Left atrial enlargement is frequent in chronic organic MR (2,14,20), is progressive (12,26), and is understood as compensatory response to volume overload (11,12,14), resulting in increased atrial compliance and contributing to persistently normal atrial and pulmonary pressures (11,12). Hence, progressive LA enlargement due to organic MR is considered as a mechanism delaying occurrence of symptoms and heart failure.

In the present study, patients with higher LA enlargement had higher RVol and LV volumes, confirming its general link to the volume overload due to MR, which elicits LA (2,14,20,27) and LV remodeling (28,29) in organic MR. However, there is considerable individual variability in LA enlargement, which is potentially linked to LV diastolic impairment (30) and end-diastolic pressure elevation (29) that affect LA pressure and further remodeling (31). Whereas normal LA is compliant with low pressure, it stretches and stiffens with chronic stress (32). Organic MR induces LA remodeling with atrial interstitial fibrosis experimentally (33) and in humans seen at a late stage (34) impairing LA elastic properties and compliance with elevated LA pressure (11,29,35,36). Indeed, in organic MR, higher LA enlargement is associated with paradoxically higher pulmonary pressure (14) and with marked hormonal (B-natriuretic peptide) activation (37) and reduced functional capacity (38). Hence, marked LA enlargement is not just a benign compensatory mechanism but likely reflects severe consequences of volume overload in organic MR with important implications for outcome.

LA enlargement and outcome in organic MR

The size of LA is considered a predictor of AF occurrence (15), which is logical but is potentially tainted by the fact that LA enlargement may be the consequence of previous undetected paroxysmal AF (14). However, it is more consequential that LA enlargement predicts poor outcomes in various clinical (18,19) or epidemiological (16,17,39) circumstances. In organic MR, LA enlargement predicts future AF (14,20,21) and limited observations suggest that it is associated with subsequent occurrence of heart failure (2). The present study confirms and most importantly extends these observations by demonstrating for the first time the strong value of LA index measured at diagnosis, in predicting death or cardiac events under conservative management, independently of known outcome predictors in patients with organic MR in sinus rhythm at diagnosis.

Patients with LA index ≥60 ml/m²incur excess mortality under medical management compared with the U.S. population and compared with patients with MR and lower LA enlargement, even after adjustment and stratification by RVol. Furthermore, LA index ≥60 ml/m²also strongly and independently predicts cardiac events with medical management, even after adjustment and stratification by RVol. Less severe LA enlargement (LA index 40 to 59 ml/m²) had less notable consequences on survival after diagnosis. Severe LA enlargement, complicating severe or even moderate regurgitation, is thus a harbinger of poor outcome in patients with organic MR, even in those without symptoms and with preserved LVEF. Therefore, LA volume should be part of comprehensive echocardiographic examination and should be used in considering surgical treatment in patients with severe organic MR.

Effect of surgery on prognosis

Mitral valve surgery improves symptoms and outcome in symptomatic patients with severe chronic MR, particularly with valve repair (3,4). Class I indications for surgery are based on symptoms or LV dysfunction in current guidelines (3,4). Nevertheless, optimal timing of surgery in organic MR remains debated (3,4,7–9). Indeed, patients operated on after they become symptomatic or after developing LV dysfunction incur excess mortality after surgery (5,6) so that such rescue surgery leaves patients with less than optimal result. Thus, ongoing research aims at defining markers of outcome that are associated with notable risk under medical management but do not affect post-operative outcome and allow life expectancy restoration after surgery. Recent data demonstrated improvement of outcome after surgery, particularly valve repair, resulting in restoration of life expectancy in patients with severe MR (2,8,9), even in patients asymptomatic before surgery (8). The present study confirms beneficial effects of mitral surgery on mortality and cardiac events, emphasizing patients with marked LA enlargement and importantly, shows that LA enlargement does not imply untoward post-operative outcome. Thus, in contrast to symptoms or low EF, LA index ≥60 ml/m²is a high-risk marker allowing restorative surgery without excess post-operative risk.

In clinical practice, our data suggest that LA index ≥60 ml/m²should be included among novel markers of risk in organic MR in conjunction with MR severity quantitation (8) as part of comprehensive risk assessment. In view of the risk attached to marked LA enlargement and of outcome improvement provided by surgery, we believe that patients with severe MR and LA index ≥60 ml/m², even those asymptomatic, should be considered for mitral surgery, particularly if valve repair is feasible. Conversely, patients with moderate MR incur initially lower risks (8) and LA index ≥60 ml/m²should lead to close follow-up to detect MR progression.

Study limitations

Our study strength relies on triple quantitative assessment of LV, LA, and MR with prospective enrollment of patients with organic MR in sinus rhythm. Atrial fibrillation is an important determinant of LA remodeling per se, which we excluded and the role of LA enlargement in patients with MR and AF will require future studies.

Clinical management was determined independently by patient's personal physicians without investigators' interference. Regular follow-up, touted as key to good outcome (7), was obtained in all patients, at Mayo Clinic or with their home physicians, and is representative of routine medical care.

Left atrial enlargement is caused by MR and may be considered a surrogate for MR severity. However, LA enlargement integrates not only MR degree but also other measures of the valve disease severity (14). The crucial point is that LA index, irrespective of this supposed “surrogate” nature, predicted survival and cardiac events, adjusting for known predictors of outcome in organic MR, including Charlson comorbidity index and quantitation of MR severity. In that respect adjustment by ERO did not affect results (HR for LA index ≥60: 2.48, p = 0.036 for mortality and 3.98, p < 0.0001 for cardiac events). Overfitting may cause detection of spurious predictors of outcome, but the prospective and hypothesis-driven natures of our study that used few adjusting variables concur to suggest overfitting risk as minimal. Thus, we believe that LA enlargement is indeed an integrator of MR consequences rather than a surrogate, which should be used for risk stratification and clinical decisions in routine practice.