Author + information
- Received September 11, 2013
- Revision received December 2, 2013
- Accepted December 9, 2013
- Published online March 1, 2014.
- Liam Ring, MD∗∗ (, )
- Bushra S. Rana, MD†,
- Francis C. Wells, MD†,
- Anna C. Kydd, MD∗ and
- David P. Dutka, MD∗
- ∗Department of Medicine, University of Cambridge, Addenbrooke's Hospital, Cambridge, United Kingdom
- †Papworth Hospital NHS Foundation Trust, Papworth Everard, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. Liam Ring, Department of Medicine, University of Cambridge, Addenbrooke's Hospital, Hills Road, Cambridge CB2 0QQ, United Kingdom.
Objectives The purpose of this study was to determine the clinical utility of left atrial (LA) functional indexes in patients with mitral valve prolapse (MVP) and mitral regurgitation (MR).
Background Timing of surgery for MVP remains challenging. We hypothesized that assessment of LA function may provide diagnostic utility in these patients.
Methods We studied 192 consecutive patients in sinus rhythm with MVP, classified into 3 groups: moderate or less MR (MOD group, n = 54); severe MR without surgical indication (SEV group, n = 52); and severe MR with ≥1 surgical indication (SURG group, n = 86). Comparison was made with 50 control patients. Using 2D speckle imaging, average peak contractile, conduit, and reservoir atrial strain was recorded. Using Simpson's method we recorded maximal left atrial volume (LAVmax) and minimal left atrial volume (LAVmin), from which the total left atrial emptying fraction (TLAEF) was derived: (LAVmax−LAVmin)/LAVmax × 100%.
Results TLAEF was similar in the MOD and control groups (61% vs. 57%; p = NS), was reduced in the SEV group (55%; p < 0.001 vs. control group), and markedly lower in the SURG group (40%; p < 0.001 vs. other groups). Reservoir strain demonstrated a similar pattern. Contractile strain was similarly reduced in the MOD and SEV groups (MOD 15%; SEV 14%; p = NS; both p < 0.05 vs. control group 20%) and further reduced in the SURG group (8%; p < 0.001 vs. other groups). By multivariate analysis, TLAEF (odds ratio [OR]: 0.78; p < 0.001), reservoir strain (OR: 0.91; p = 0.028), and contractile strain (OR: 0.86; p = 0.021) were independent predictors of severe MR requiring surgery. Using receiver-operating characteristic analysis, TLAEF <50% demonstrated 91% sensitivity and 92% specificity for predicting MVP with surgical indication (area under the curve: 0.96; p < 0.001).
Conclusions We report the changes in left atrial function in humans with MVP and the relationship of LA dysfunction to clinical indications for mitral valve surgery. We propose that the findings support the utility of quantitative assessment of atrial function by echocardiography as an additional tool to guide the optimum timing of surgery for MVP.
Mitral valve prolapse (MVP) is the second most common valvular lesion requiring cardiac surgery (1). If severe mitral regurgitation (MR) is not treated, there is increased morbidity and a constrained prognosis (2,3). Optimum timing of surgical intervention is of paramount importance, but remains clinically challenging. Symptomatic deterioration is a clear indication that surgery should be undertaken (4), but this may occur late in the natural history of the condition and be secondary to irreversible left ventricular (LV) dysfunction (5). Current guidelines highlight the importance of LV size and function to identify those with latent LV dysfunction who should be considered for surgery (6). Additionally, the development of atrial fibrillation (AF) or the presence of elevated pulmonary artery (PA) pressures are associated with an adverse prognosis, and indicate the need for detailed assessment of the asymptomatic patient regarding surgical referral (6–9).
The effect of MR on left atrial (LA) behavior and function has received less attention. The LA dilates as a consequence of the regurgitant load, and an LA volume index >60 ml/m2 predicts an adverse cardiovascular outcome (10,11). Characterization of atrial function using speckle tracking technology has been proposed to offer insights into the atrial adaptation to chronic MR. In a heterogeneous group of patients with severe MR, atrial strain was impaired, which may relate to the presence of atrial fibrosis (12,13). We hypothesized that atrial function may be closely linked to adverse pathophysiological findings in MVP, and offer potential as a useful additional tool to help define the optimum timing of surgery in these patients.
We studied consecutive patients with MVP and associated MR assessed within our service between August 2009 and November 2012. Only patients with MVP or a flail leaflet were included (MVP was defined as displacement of the tip of 1 or more segments of the mitral valve by ≥2 mm relative to the hinge points of the leaflets, and flail as the systolic eversion of the leaflet tip into the LA). We excluded subjects who were in AF, patients with more than moderate coexistent aortic valve disease, and those with prior mitral valve surgery (Fig. 1). All patients underwent complete clinical assessment including comprehensive echocardiography and were included prospectively in a clinical database. Patients were classified according to MR severity and the presence of established indications for surgical intervention according to American guidelines (6), including symptoms, left ventricular internal diameter in systole (LVIDs) ≥40 mm, ejection fraction (EF) <60%, or PA pressure >50 mm Hg at rest or 60 mm Hg with exercise. Comparison was made with a group of 50 patients in sinus rhythm, with normal LV size and function, normal LA size, and normal mitral valves with less than mild MR. There was institutional approval for the study.
Studies were performed using an S5-1 transducer and an iE33 imaging platform (Philips, Andover, Massachusetts) and were analyzed offline using Xcelera (Philips). The LV and LA dimensions were determined from the parasternal long-axis window, and MR severity was defined using a multiparametric approach, including an assessment of the effective regurgitant orifice area (EROA), the regurgitant volume, and the presence of pulmonary venous flow reversal (14). The PA pressure was estimated through assessment of the tricuspid regurgitant jet and right atrial pressure. Diastolic function was described with the E/E′ ratio, where E was measured using pulsed Doppler of mitral inflow from the 4-chamber view, and E′ as the average of the septal and lateral mitral annular diastolic velocities using tissue Doppler imaging. The LV volumes and EF were calculated using Simpson's biplane method; similarly, maximal left atrial volume (LAVmax) (defined as volume immediately prior to mitral valve opening) and minimal left atrial volume (LAVmin) (defined as volume immediately after mitral closure) were determined, from which the total left atrial emptying fraction (TLAEF) was derived (15): (LAVmax – LAVmin)/LAVmax × 100%.
The LA strain was obtained from apical 4- and 2-chamber views using semiautomated software (Cardiac Motion Quantification, Qlab version 9.0, Philips). Sector width, depth, and gain settings were optimized. After manually defining the basal septal, lateral, and apex of the LA, the software produces a region of interest, which is adjusted to optimize inclusion of the LA walls and enable the time-deformation graph. Similar to a previous definition (16), 3 aspects of atrial strain were recorded: contractile, describing deformation after the p-wave; conduit, describing passive atrial filling; and reservoir, representing the sum of these elements (Fig. 2). The process was repeated from the 2-chamber view, and the values for all segments were averaged. Analysis of atrial functional indexes was performed retrospectively by subjects blinded to the clinical categorization of the patients.
Statistical analysis was performed with SPSS version 19.0 (SPSS Inc., Chicago, Illinois). Continuous parameters are expressed as mean ± SD or median (interquartile range). Comparisons of continuous parameters are made using analysis of variance with post-test Bonferroni correction or the Kruskal-Wallis test with post-test Dunn correction where appropriate. Categorical parameters are compared using the chi-square test.
Associations of atrial functional indexes were determined with linear regression analysis. Clinical and echocardiographic characteristics significant or borderline (p ≤ 0.20) on univariate analysis were entered into separate multivariate models with the atrial parameter in question as the dependent variable. Independent associations are expressed as the B co-efficient and 95% confidence intervals, with the model r2 for the explained variance.
To determine whether atrial function could predict severe MR requiring surgery, multivariate logistic regression analysis was employed. To avoid colinearity, separate models were constructed for each atrial parameter, to which we added age, LAVmax, EROA, NYHA functional class, LVIDs, EF, and PA pressure. Receiver-operating characteristic (ROC) curves were generated for independent indexes to obtain optimal cut-offs, as defined by the Youden index. Measurement reproducibility was determined by selecting 10 patients at random, and the measurements were repeated by the original and a second operator while blinded to the original results. The within-subject coefficient of variation and 95% limits of agreement were calculated using the Bland-Altman method. A p value <0.05 was considered statistically significant.
Data was acquired for 192 consecutive patients, classified into 3 groups: moderate or less MR (MOD group, n = 54); severe MR without surgical indication (SEV group, n = 52), and severe MR with ≥1 surgical indication (SURG group, n = 86). Indications for surgery were: isolated cardiovascular symptoms (n = 28); elevated PA pressures with or without symptoms (n = 34); and LV dilation or dysfunction (n = 24). Baseline characteristics are presented in Table 1. The SURG patients were older than the remaining cohorts, with advanced cardiovascular symptoms (median NYHA functional class 2.0 [interquartile range: 2.0 to 3.0]; p < 0.001 vs. other groups). As expected by design, LVIDs was increased in SURG patients, whereas the remaining groups did not differ in this regard.
Atrial size and function
LAVmax increased progressively from control patients to the MOD and SEV groups (Table 1). Despite increased LAVmax in SEV patients, TLAEF was only slightly reduced compared with control patients and did not differ from the MOD patients (SEV 55 ± 6% vs. MOD 57 ± 8%; p = 0.549; control patients 61 ± 5%; p < 0.001 vs. SEV). In the SURG group, LAVmax was similar to the SEV group, but TLAEF was markedly reduced (40 ± 7%; p < 0.001 vs. other groups) (Fig. 3).
Reservoir strain exhibited a similar pattern to that shown by TLAEF (Table 2, Fig. 4). Contractile strain was similarly reduced in the MOD and SEV groups compared with control patients (MOD 15% [12% to 18%] vs. SEV 14% [10% to 18%]; p = NS; both p < 0.05 vs. control patients 20% [16% to 23%]), and was further reduced in the SURG group (8% [5% to 11%]; p < 0.001 vs. other groups). Conduit strain was similar for the control, MOD, and SEV groups, but was reduced in SURG patients.
Linear regression results are presented in Table 2. For TLAEF (model r2 = 0.636), the strongest independent association was NYHA functional class (B = −6.01; p < 0.001), followed by age and PA pressure. Similarly, symptomatic status was the strongest association with reservoir (model r2 = 0.533; NYHA functional class B = −4.95; p < 0.001) and contractile strain (model r2 = 0.426; NYHA functional class B = −2.73; p < 0.001). Only conduit strain (model r2 = 0.379) was associated with EROA (B = 3.12; p = 0.030), although again NYHA functional class was an important association. LAVmax was an independent association with all functional indexes.
To determine whether atrial function could predict the need for surgical intervention in MVP, logistic regression was employed. All atrial functional indexes were significant cofactors in univariate analysis and were therefore entered into separate multivariate models (Table 3). After analysis, TLAEF, reservoir, and contractile strain remained independent predictors, but conduit strain did not. ROC curves were constructed to define optimal cut-offs in these indexes and for the indication for MV surgery using the guideline criteria outlined above. All indexes performed well, with an area under the curve of 0.96, 0.90, and 0.89 for TLAEF, reservoir strain, and contractile strain, respectively. The optimal cut-off value of TLAEF was <50%, which offered 91% sensitivity and 92% specificity for predicting severe MR requiring surgery, suggesting that assessment of this parameter may have clinical utility. This index performed better than established echocardiographic indexes used in MVP: ROC analysis of EF, LVIDs, PA pressure, and LAVmax produced an area under the curve of 0.54, 0.73, 0.84, and 0.85, respectively (Fig. 5).
Sequential multivariate models were used to compare the predictive value of TLAEF with established echocardiographic findings. A baseline model including age and NYHA functional class (chi-square: 188.5) was improved with the addition of TLAEF (chi-square: 236.8; p < 0.001). The model was not improved further with the addition of LVIDs, EF, PA pressure, LAVmax, and EROA (chi-square: 244.8; p = 0.158) (Fig. 6), suggesting that TLAEF provides equivalent predictive information to these echocardiographic markers combined.
Measurement variability (95% limits of agreement and coefficient of variation) for intraobserver differences were: LAVmax: 4.5 ± 12 ml; TLAEF: 4.6 ± 3.7%; reservoir: 4.9 ± 2.2%; contractile: 5.7 ± 1.0%; and conduit: 8.5 ± 2.3%. Interobserver differences were: LAVmax: 6.3 ± 17 ml; TLAEF: 7.8 ± 6.4%; reservoir: 7.7 ± 3.3%; contractile: 8.0 ± 1.6%; conduit: 8.9 ± 2.1%.
We report the changes in left atrial function in humans with MVP and the relationship of LA dysfunction to clinical indications for mitral valve surgery. We propose that the findings support the utility of quantitative assessment of atrial function by echocardiography as an additional tool to guide the optimum timing of surgery for MVP.
LA dilation in response to MR has been noted for some time and is an important factor predicting adverse outcomes in MVP (11). Assessment of atrial function has received less attention. Using volumetric methods, atrial EF has been shown to be relatively preserved in MR, although later reports indicate that this marker is reduced after surgical intervention, suggesting that pre-operative atrial dysfunction is masked by the presence of a regurgitant load (10,17). Moustafa et al. (18) described atrial function using volumetric data in 43 patients with MVP. Atrial emptying was preserved in mild MR, but was reduced in patients with severe MR. This group included symptomatic patients who, therefore, had an indication for surgical intervention; these results are consistent with our findings.
The advent of speckle tracking technology allows the analysis of tissue deformation independent of extraneous motion. Primarily, speckle tracking has been used in the assessment of LV deformation, but there have been increasing reports of its use to assess atrial function (12,16,19). Saraiva et al. (16) examined the use of atrial speckle tracking in a cohort of normal subjects; they concluded that speckle tracking is feasible in the thin-walled atrium and that it provides valuable insights into atrial function.
Atrial speckle imaging has been used in mitral valve disease. Cameli et al. (12) used this tool in a heterogeneous group of patients with asymptomatic MVP. In mild MR, strain was increased, whereas patients with severe MR had reduced indexes. Of note, the patients with severe MR had a mean EF of 57%, suggesting that a significant proportion of the subjects included had latent LV dysfunction. Moreover, values for PA pressure were not presented. In the current study, patients were classified according to MR severity and the presence of an indication for surgical intervention. We demonstrated a reduction in contractile strain in mild to moderate MR, suggesting that the active element of atrial function is impaired early in the disease process. Conversely, conduit strain, which is influenced by the regurgitant volume, is maintained (Table 2, Fig. 3). Overall, atrial strain is therefore relatively preserved in moderate to severe MR. The presence of an accepted surgical indication is associated with a profound reduction in all phases of atrial strain, suggesting that the LA myocardium has exceeded its tolerance.
Observed reductions in atrial strain may relate to the presence of atrial fibrosis. Experimental and observational studies have shown that atrial fibrosis develops in humans and animals with MR (20). More recently, atrial strain derived using speckle tracking imaging has been shown to correlate with atrial fibrosis found at the time of surgery for MVP (13). This is consistent with echocardiographic work demonstrating an inverse relationship between atrial strain and burden of atrial fibrosis defined using delayed-enhancement cardiac magnetic resonance imaging in patients with AF (21). The assessment of LA function provides an insight into histopathological changes occurring within the atrium.
The LV adapts to chronic MR through the development of eccentric hypertrophy, maintaining normal systolic wall stress (22). The transition phase, during which LV dysfunction develops, is characterized by increasing wall stress and LV end-diastolic pressure (23): in diastolic heart failure, increased LV end-diastolic pressure is associated with reduced LA functional indexes (19). Atrial strain is additionally linked to pulmonary pressures (19). The present study affirms the intimate relationship between LA function and the accepted markers of an adverse prognosis in MVP. We demonstrated that TLAEF is independently associated with PA pressure and LV systolic volumes. Moreover, NYHA functional class had a strong independent association with all atrial indexes. These observations suggest that the atrium may play a key role in protecting the pulmonary vasculature in mitral valve disease: as the atrium becomes dysfunctional, it is no longer able to buffer the regurgitant jet. Symptoms and, eventually, PA hypertension will therefore ensue. Previous work has drawn similar conclusions regarding the important association between atrial strain and cardiovascular capacity in patients with normal LV function (24).
Reduced atrial function therefore corresponds to adverse pathophysiological adaptations to MR, and, accordingly, the presence of impaired atrial function is highly predictive of the need for surgery in MVP. The assessment of TLAEF provides predictive information equivalent to established echocardiographic findings, including LV dimensions, EF, PA pressure, LAVmax, and EROA combined (Fig. 6). At times, the decision to undertake surgery for MVP is challenging, and atrial indexes may provide clinical utility in these situations. In less active subjects, or patients with comorbidities, assessment of cardiovascular symptoms may be problematic. Both American and European guidelines focus on the prevention of latent LV dysfunction, but they differ as to the value of systolic dimension that should trigger surgery; moreover, there is no accepted indexed value, making interpretation difficult at the extremes of patient size (6,25).
This work represents a single-center experience. A prospective validation of atrial functional indexes is required to confirm the usefulness in prediction of need for cardiac surgery and whether atrial function influences clinical outcomes. Patients are referred to our service for consideration of mitral surgery, and, as such, a disproportionate number of subjects had severe MR, although we feel this is unlikely to have affected the results of the analysis. Strain values may vary according to imaging platform, and therefore, these results may not be applicable to all software systems.
The LA performs a key role in chronic MR. Atrial adaptation is characterized by increasing LA volume, but with preserved total LA emptying fraction and reservoir function. These parameters are relatively maintained even in severe MR. The loss of normal atrial function corresponds to the development of an indication for surgical intervention. The quantitative assessment of LA function may be a useful additional tool in guiding clinicians in the optimal timing of surgery for MVP.
Dr. Wells has served as a consultant to St. Jude Medical. Dr. Dutka has received unrestricted research support from Sorin and Merck Sharp & Dohme; and has received grants from the National Institute for Health Research, Medical Research Council and British Heart Foundation. All other authors have reported that they have no relationships relevant to this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- ejection fraction
- effective regurgitant orifice area
- left atrial/left atrium
- left atrial volume
- left ventricular
- left ventricular internal diameter in systole
- mitral regurgitation
- mitral valve prolapse
- New York Heart Association
- pulmonary artery
- total left atrial emptying fraction
- Received September 11, 2013.
- Revision received December 2, 2013.
- Accepted December 9, 2013.
- American College of Cardiology Foundation
- Iung B.,
- Baron G.,
- Butchart E.G.,
- et al.
- Avierinos J.-F.,
- Gersh B.J.,
- Melton L.J.,
- et al.
- Rosenhek R.
- Bonow R.O.,
- Carabello B.A.,
- Chatterjee K.,
- et al.
- Grigioni F.,
- Enriquez-Sarano M.,
- Ling L.H.,
- et al.
- Grigioni F.,
- Avierinos J.-F.,
- Ling L.H.,
- et al.
- Le Tourneau T.,
- Messika-Zeitoun D.,
- Russo A.,
- et al.
- Cameli M.,
- Lisi M.,
- Righini F.M.,
- et al.
- Kurt M.,
- Wang J.,
- Torre-Amione G.,
- Nagueh S.F.
- Verheule S.,
- Wilson E.,
- Everett T.,
- Shanbhag S.,
- Golden C.,
- Olgin J.
- Kuppahally S.S.,
- Akoum N.,
- Burgon N.S.,
- et al.
- Corin W.J.,
- Monrad E.S.,
- Murakami T.,
- Nonogi H.,
- Hess O.M.,
- Krayenbuehl H.P.
- Gaasch W.H.,
- Meyer T.E.
- Kusunose K.,
- Motoki H.,
- Popovic Z.B.,
- Thomas J.D.,
- Klein A.L.,
- Marwick T.H.
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.