From the Dep

The Netherlan

Dr. Debonnair

entific Corpor

Research Gra

ical Research

the Royal Aus

Electrophysio

Medical Cente

ing, St. Jude

Left Atrial Function by Two-DimensionalSpeckle-Tracking Echocardiography in Patients

with Severe Organic Mitral Regurgitation: Associationwith Guidelines-Based Surgical Indicationand Postoperative (Long-Term) Survival

PhilippeDebonnaire,MD,Darryl P. Leong,MBBS,MPH, PhD, Tomasz G.Witkowski,MD, Ibtihal Al Amri,MD,Emer Joyce, MB, BCH, BAO, Spyridon Katsanos, MD, Martin J. Schalij, MD, PhD, Jeroen J. Bax, MD, PhD,

Victoria Delgado, MD, PhD, and Nina Ajmone Marsan, MD, PhD, Leiden, The Netherlands

Background: Left atrial (LA) mechanics in patients with severe mitral regurgitation (MR) remain largely unex-plored. The aim of the present evaluation was to assess the effect of severe MR on LA function, its potentialrelation with conventional surgical indications, and long-term postoperative survival.

Methods: Two-dimensional speckle-tracking strain and volumetric indices of LA reservoir, conduit, and con-tractile function were assessed in 121 patients with severe MR and 70 controls. Patients were divided accord-ing to the presence (n = 46) or absence (n = 75) of one or more guidelines-based criteria for mitral surgery(symptoms, left ventricular ejection fraction # 60%, left ventricular end-systolic diameter $ 40 mm, atrialfibrillation, or systolic pulmonary arterial pressure >50 mm Hg).

Results: In patients with severe MR compared with controls, significant LA reservoir and contractile dysfunc-tion was observed, which was more pronounced in patients with mitral surgery indication (P < .05 for all strainand volumetric indices). Of all indices of LA function, LA reservoir strain was an independent predictor (oddsratio, 0.88; 95%confidence interval, 0.82–0.94;P < .001) and had the highest accuracy to identify patients withindications for mitral surgery (area under the receiver operating characteristic curve, 0.8; 95% confidenceinterval, 0.72–0.87). A total of 117 patients underwent mitral valve surgery. Patients with LA reservoir strain#24% showed worse survival at a median of 6.4 years (interquartile range, 4.7–8.7 years) after mitral surgery(P = .02), regardless the symptomatic status before surgery. LA reservoir strain, on top of mitral surgery indi-cations, provided incremental predictive value for postoperative survival.

Conclusions: Impaired LA reservoir strain in patients with severe organic MR relates to long-term survival aftermitral valve surgery, independently of and incremental to current guidelines-based indications for mitral sur-gery. (J Am Soc Echocardiogr 2013;26:1053-62.)

Keywords: Atrium, Mitral valve, Regurgitation, Strain, Surgery

Severe chronic mitral regurgitation (MR) causes a significant left atrial(LA) volume overload, leading to LA dilatation as a compensatorymechanism tomaintain LA pressure homeostasis and prevent pulmo-nary congestion.1 Measures of LA remodeling, such as LA diameterand LA volume index, have been shown to predict outcomes in pa-tients with severe organic MR and are helpful for risk stratification

artment of Cardiology, Leiden University Medical Center, Leiden,

ds.

e is supported by a Sadra Medical Research Grant from Boston Sci-

ation and holds a European Association of Cardiovascular Imaging

nt for 2013. Dr. Leong is supported by the National Health and Med-

Council of Australia, the National Heart Foundation of Australia, and

tralasian College of Physicians and is a recipient of the Earl Bakken

logy Fellowship. The Department of Cardiology of Leiden University

r has received grants from GE Healthcare, Lantheus Medical Imag-

Medical, Medtronic, Boston Scientific Corporation, Biotronik, and

and clinical decision making in these patients.2-4 However, chronicMR may induce significant LA ultrastructural changes, potentiallyaffecting LA myocardial contractility and relaxation before LAdilatation occurs.3,5-8 Therefore, assessment of LA function, ratherthan dimension, might be of significant value to guide therapy inpatients with severe MR. However, data on the effect of severe MR

Edwards Lifesciences. Dr. Delgado has received consulting fees from St. Jude

Medical and Medtronic.

Reprint requests: Nina Ajmone Marsan, MD, PhD, Leiden University Medical Cen-

ter, Department of Cardiology, Albinusdreef 2, 2333 ZA Leiden, The Netherlands

(E-mail: n.ajmone@lumc.nl).

0894-7317/$36.00

Copyright 2013 by the American Society of Echocardiography.

http://dx.doi.org/10.1016/j.echo.2013.05.019

1053

Abbreviations

CI = Confidence interval

LA = Left atrial

LAVI = Left atrial volume

indexed to body surface area

LV = Left ventricular

LVEF = Left ventricular

ejection fraction

MR = Mitral regurgitation

OR = Odds ratio

2D = Two-dimensional

1054 Debonnaire et al Journal of the American Society of EchocardiographySeptember 2013

on LA function are scarce, andthe potential incremental valueof LA function assessment topredict timing for surgery isunexplored.9-12

Different LA functions can beevaluated, including reservoir(storage of pulmonary venous in-flow during ventricular systole),conduit (passive emptying duringearly diastole), and contractile(active emptying at late diastole)function. These sequential LAfunctions play a crucial role incardiac performance by optimiz-ing left ventricular (LV) filling13,14

and can bemeasured using conventional two-dimensional (2D) echo-cardiography by detecting the phasic changes of LAvolume during thecardiac cycle. More recently, 2D speckle-tracking longitudinal defor-mation (strain and strain rate) imaging has also showed accurate char-acterization of all phases of LA function.14-18

Therefore, the aims of this study were (1) to characterize LA reser-voir, conduit, and contractile function using 2D speckle-tracking de-formation imaging in patients with chronic severe organic MR; (2)to determine its clinical relation to the presence of guidelines-basedconventional indications for mitral surgery; and (3) to explore its asso-ciation with long-term survival after mitral valve surgery.2,19

METHODS

Patient Population

The patient population consisted of patients presenting with chronicsevere organic MR, who were referred to our center within the pastdecade. Clinical and echocardiographic data were prospectivelycollected in the department’s cardiology information system(EPD-Vision; Leiden University Medical Center, Leiden, TheNetherlands) and retrospectively analyzed.

Clinical assessment included demographics, medications, identifi-cation of comorbidities, and symptoms according to the New YorkHeart Association functional class. Echocardiographic evaluation in-cluded conventional measurements and speckle-tracking analysisfor LA deformation analysis. Patients with prior cardiac surgery ormyocardial infarction, congenital heart disease, concomitant mitralstenosis exceeding mild severity (mean gradient $5 mm Hg), or sig-nificant aortic valve disease were excluded. In addition, to avoid con-founders for increased pulmonary pressure, patients with significantpulmonary disease were excluded.

Patients were further dichotomized on the basis of the presence ofone or more conventional criteria for mitral surgery indication, ac-cording to current guidelines: New York Heart Association class IIIor IV symptoms, LV ejection fraction (LVEF) # 60%, LV end-systolicdiameter $ 40 mm, atrial fibrillation, or systolic pulmonary arterialpressure at rest >50 mm Hg.2 Accordingly, phasic LA functionswere evaluated in both patient groups.

In addition, 70 individuals with similar ages, body surface areas,and gender distribution served as a control group. These subjects un-derwent clinically indicated echocardiography for the evaluation ofpotential cardiac symptoms, murmurs, or increased cardiovascularrisk profiles, but all showed absence of significant structural or func-tional abnormalities.

A total of 117 patients (97%) underwent mitral valve surgerya mean of 85 6 19 days after the baseline echocardiographic exam-ination. This practice is in line with current guidelines advocatingthat early mitral valve surgery may be performed at experienced valvecenters with high repair rates for patients presenting with severe or-ganic MR, despite an absence of symptoms or LV dysfunction.2

All-cause mortality was assessed in all operated patients by reviewingpatients’ medical files and by evaluation of the official Dutch NationalSurvival Registry for patients who were followed by the referring cen-ter postoperatively.

Echocardiography

Transthoracic 2D echocardiography was performed in the left lat-eral decubitus position using commercially available ultrasoundsystems (Vivid-5, Vivid-7, and E9; GE Vingmed Milwaukee, WI)equipped with a 3.5-MHz transducer. Electrocardiographically trig-gered standard 2D grayscale and color Doppler images were ac-quired in cine-loop format and transferred to a workstation foroffline analysis (EchoPAC 110.0.0; GE Medical Systems, Horten,Norway). Chamber quantification was performed conforming tocurrent recommendations.20 LA anteroposterior linear diameterwas measured in the 2D left parasternal long-axis view. LV andLA volumes were assessed using Simpson’s biplane method andwere indexed to body surface area.20 LVEF was calculated fromLV volumes as recommended. In accordance with recent guide-lines, a multiparametric integrative approach was used to assessMR severity (additional transesophageal images were used whenavailable).21 In particular, an effective regurgitant orifice area$40 mm2 or regurgitant volume $60 mL, using the proximal iso-velocity surface area method, as well as the presence of flail mitralleaflet or ruptured papillary muscle defined severe MR. If none ofthese factors were present or the proximal isovelocity surface areamethod was not feasible (because of eccentricity of regurgitant jet),presence of vena contracta width $7 mm in combination with oneor more qualitative (large central jet, severe coanda effect, largeflow convergence radius, dense mitral envelope on continuous-wave Doppler) and/or semiquantitative (pulmonary Doppler sys-tolic venous flow reversal, E-wave dominance $1.5 m/sec)variable(s) defined severe MR.21 MR etiology was categorized asflail (free mitral leaflet edge reversing into the left atrium), prolapse(mitral leaflet coaptation line behind annular plane without edgereversing into the left atrium), or degenerative (degenerative mitralvalve abnormalities without flail or prolapse).20 All Doppler mea-surements represented the average of three beats (five beats if atrialfibrillation was present). Mitral inflow was analyzed to assess earlyflow (E wave), deceleration time, and mitral valve pressure gradi-ent.22 Early diastolic peak velocity (E0) was derived from the lateralwall on the apical four-chamber view color tissue Doppler acquisi-tion. Systolic pulmonary arterial pressure at rest was calculated us-ing the maximal tricuspid regurgitant jet velocity and right atrialpressure estimation, on the basis of diameter and respiratory vari-ation of the inferior caval vein.20

LA Function: Volumetric Indices. LA volumes were assessedjust before mitral valve opening (maximal LA volume [Volmax]), atmitral valve closure (minimal LA volume [Volmin]), and at P-waveonset on electrocardiography just before atrial contraction, if in sinusrhythm (pre-A LA volume [VolP]). As previously reported, LA reser-voir function was calculated as LA expansion index (Volmax �Volmin/Volmin � 100), LA conduit function as LA passive emptyingfraction (Volmax � VolP/Volmax � 100), and LA contractile function

Figure 1 Assessment of LA phasic function by 2D speckle-tracking echocardiography in a control subject. (A) Four-chamber longi-tudinal strain, (B) two-chamber longitudinal strain, (C) four-chamber longitudinal strain rate, and (D) two-chamber longitudinal strainrate. The white dotted line represents the mean value of the tracked LA segments. Reservoir, conduit, and contractile phase arerepresented by color-coded arrows (yellow, pink, and blue respectively). Measurements are averaged over the four-chamber andtwo-chamber views. In this subject, reservoir, conduit, and contractile strain and strain rate were 38.6%, �22.1%, and �16.5%and 1.75, �1.8, and �2.0 sec�1, respectively.

Journal of the American Society of EchocardiographyVolume 26 Number 9

Debonnaire et al 1055

as LA active emptying fraction (VolP � Volmin/VolP � 100).14,17 Inpatients with atrial fibrillation, LA contractile function could notbe calculated.

LA Function: Deformation Indices. LA longitudinal strain andstrain rate was assessed using 2D speckle-tracking analysis withQRS onset as the reference point, applying a commercially availableLV strain software package to the left atrium (EchoPAC version110.0.0).23 In summary, the region of interest was adjusted to includethe LA myocardium in both the four-chamber and two-chamber api-cal views that included both the left atrium and left ventricle. Manualcorrection was performed to optimize tracking results if needed. Thevalues of LA strain and strain rate were calculated as the average valueof the four-chamber and two-chamber views, as described in Figure 1.As previously reported, LA reservoir function was calculated as peaksystolic LA strain (LA reservoir strain), while LA contractile functionwas measured as the LA strain value at P-wave onset on electrocardi-ography, if in sinus rhythm (LA contractile strain). Finally, LA conduitfunction was assessed as the difference between LA reservoir andcontractile strain (LA conduit strain).15 In addition, LA reservoir strainrate was measured as peak systolic positive value, LA conduit strainrate as the early diastolic negative peak, and LA contractile strainrate as the late diastolic negative peak (if in sinus rhythm)(Figure 1). In patients with atrial fibrillation, LA contractile strain(rate) could not be calculated.

Fifteen subjects were randomly selected to perform a blinded inter-observer and intraobserver (1 week after the first observation) agree-ment test for LA strain measurements.

For simplicity, absolute deformation values are reported as positivenumbers.

Statistical Analysis

Continuous variables were compared using independent Student’st tests and Mann-Whitney U or Kruskal-Wallis tests if not normallydistributed and are expressed as mean 6 SD. Categorical valueswere compared using c2 tests and are expressed as percentages.Multiple-group comparisons were performed using one-way analysesof variance with post hoc Bonferroni correction. Receiver operatingcharacteristic curve analysis was performed to assess the accuracyof various deformation and volumetric indices of LA function to pre-dict the presence of conventional indications for mitral surgery. Theoptimal cutoff value was obtained by maximizing the sum of sensitiv-ity and specificity. Subsequently, multiple logistic regression analysisidentified independent predictors of the presence of mitral surgery in-dications. Covariates with P values < .05 at the univariate level wereincluded in themultivariate logistic regression analysis, whichwas per-formed using a backward elimination approach. In addition, Cox re-gression analysis including conventional guidelines-based criteria formitral surgery and LA reservoir strain was performed to identify theindependent determinants of long-term mortality after mitral surgery.To maximize the possibility to detect any relationship of this clinicalmodel with postoperative mortality, covariates with univariate signif-icance of P < .10 were included in the multivariate analysis. Kaplan-Meier curves were constructed to explore differences in long-termsurvival after mitral valve surgery in different patient subgroups strat-ified according to LA reservoir strain, using the previously mentionedreceiver operating characteristic curve–based cutoff value and com-pared using the log-rank test. The incremental value of successive con-ventional guidelines-based criteria for mitral surgery and LA reservoirstrain to predict postoperative mortality was assessed using the likeli-hood ratio test: the model containing the covariate of interest was

Table 1 Baseline characteristics of controls and the overall study population, with dichotomization according to the presence orabsence of mitral surgery indication

Variable Controls (n = 70) All patients (n = 121) No surgical indication (n = 46) Surgical indication (n = 75) P*

Clinical characteristicsAge (y) 63 6 12 63 6 13 59 6 14 66 6 11 .008

Men 47 (67%) 77 (64%) 32 (70%) 45 (60%) .30

BSA (kg/m2) 1.9 6 0.21 1.9 6 0.20 1.9 6 0.21 1.9 6 0.20 .80

AF 0 (0%) 26 (21%) 0 (0%) 26 (35%) <.001

NYHA class <.001

I 70 (100%) 38 (32%) 20 (43%) 18 (24%)II 0 (0%) 49 (40%) 26 (57%) 23 (31%)

III 0 (0%) 30 (25%) 0 (0%) 30 (40%)IV 0 (0%) 4 (3%) 0 (0%) 4 (5%)

Hypertension 31 (45%) 11 (10%) 4 (10%) 7 (11%) 1.00

Hypercholesterolemia 20 (29%) 11 (10%) 4 (10%) 7 (11%) 1.00

Diabetes 9 (13%) 4 (3%) 0 (0%) 4 (5%) .30

Smoking history 25 (43%) 21 (21%) 7 (18%) 14 (23%) .20

Medication useb-blockers 8 (12%) 41 (34%) 8 (17%) 33 (44%) .003

Diuretics 9 (13%) 36 (30%) 7 (15%) 29 (39%) .006

ACE inhibitors/ARBs 22 (32%) 54 (45%) 17 (37%) 37 (49%) .20

Statins 21 (31%) 20 (17%) 5 (11%) 15 (20%) .20

MR etiology .13

Degenerative — 7 (6%) 1 (2%) 6 (8%)Flail — 40 (33%) 12 (26%) 28 (37%)

Prolapse — 74 (61%) 33 (72%) 41 (55%)Echocardiographic data

LVESD (mm) 25.2 6 4.1 34.8 6 6.4 31.9 6 4.5 36.6 6 6.7 <.001LVEDVI (mL/m2) 55 6 11 85 6 21 84 6 18 87 6 23 .70

LVESVI (mL/m2) 20 6 6 28 6 11 25 6 7 30 6 13 .20LVEF (%) 64 6 6 68 6 8 70 6 5 66 6 9 .047

LA diameter (mm) 35 6 4 49 6 8 46 6 6 50 6 8 .001LAVI (mL/m2) 31 6 8 76 6 33 64 6 23 83 6 36 .002

Vena contracta (mm) 1.7 6 1.25 8.0 6 0.92 7.6 6 0.81 8.2 6 0.91 <.001

E wave (m/sec) 0.6 6 0.2 1.3 6 0.2 1.2 6 0.2 1.4 6 0.2 <.001

E0 7.3 6 2.0 8.5 6 2.1 8.5 6 1.6 8.6 6 2.4 .9

DT (msec) 218 6 49 180 6 43 189 6 34 174 6 47 .01

Systolic PAP at rest (mm Hg) 26 6 5 43 6 16 32 6 8 49 6 17 <.001

ACE, Angiotensin-converting enzyme; AF, atrial fibrillation; ARB, angiotensin receptor blocker; BSA, body surface area; DT, deceleration time;

LVEDVI, LV end-diastolic volume indexed to body surface area; LVESD, LV end-systolic diameter; LVESVI, LV end-systolic volume indexed to

body surface area; NYHA, New York Heart Association; PAP, pulmonary artery pressure.

Data are expressed as mean 6 SD or as number (percentage).*For comparison between patient groups with and without mitral surgery indication.

1056 Debonnaire et al Journal of the American Society of EchocardiographySeptember 2013

tested against the nested model not containing the covariate. Finally,linear regression analysis was performed to explore the correlationbetween LA reservoir strain, LA diameter, and LA volume indexedto body surface area (LAVI), respectively. Statistical analysis was per-formed using SPSS version 17.0. (SPSS, Inc., Chicago, IL) and Stataversion 12.1 (StataCorp LP, College Station, TX). All tests were twosided, and P values < .05 were considered statistically significant.

RESULTS

Patient Population

A total of 160 patients with chronic severe organic MR who under-went complete clinical and echocardiographic examinations wereincluded. Thirty-nine patients (24%) with poor acoustic windows

were excluded. The mean frame rate for echocardiographic exami-nations performed within the first half of the past decade did notsignificantly differ from those performed during the most recenthalf (56.4 vs 59.4 frames/sec, respectively, P = .31). Clinical andechocardiographic characteristics of the remaining 121 patients(mean age, 63 6 13 years; 77% men) are summarized in Table 1.Overall, preserved LV systolic function was observed, with mildLV dilatation and severe LA dilatation. Severe MR was documented,with a mean mitral vena contracta width of 8 6 0.9 mm, mean Ewave of 1.3 6 0.2 m/sec, and mean proximal isovelocity surfacearea radius of 11 6 2 mm. Forty-six patients did not meet anyconventional criteria for mitral surgery, whereas 75 patients metone or more criteria. In particular, symptoms were present in 34patients (28%), LVEF # 60% in 19 (16%), LV end-systolicdiameter $ 40 mm in 27 (22%), atrial fibrillation in 26 (21%),

Table 2 Phasic LA function indices of controls and the overall study population with dichotomization according to presence orabsence of mitral surgery indication

LA function variable Controls (n = 70) All patients (n = 121)

No surgical

indication (n = 46) Surgical indication (n = 75) P*

Reservoir functionStrain (%)† 31 6 6.1 22 6 8.4† 27 6 6.6† 19 6 7.7† <.001

Strain rate (sec�1) 1.25 6 0.28 0.95 6 0.32† 1.12 6 0.27 0.85 6 0.31† <.001

Expansion index (%)† 191 6 61 111 6 64† 150 6 70† 87 6 46† <.001

Conduit functionStrain (%)† 17 6 5.1 15 6 5.0† 17 6 4.7 13 6 4.7† <.001

Strain rate (sec�1) 1.08 6 0.38 1.0 6 0.32 1.14 6 0.34 0.98 6 0.32† .04

Passive emptying fraction (%)† 35 6 9.9 34 6 9.6 37 6 9.9 33 6 9.1† <.001

Contractile function‡

Strain (%)† 14 6 3.0 9.2 6 3.5† 10 6 3.5† 8 6 3.4† .03

Strain rate (sec�1) 1.65 6 0.55 1.04 6 0.39† 1.15 6 0.4† 0.94 6 0.35† .06

Active emptying fraction (%)† 44 6 9.4 28 6 11† 32 6 12† 24 6 10† .001

All deformation indices are presented as positive absolute numbers.

*For comparison between patient groups without and with mitral surgery indication.†P < .05 for comparison with controls.‡Patients in sinus rhythm only.

Table 3 Receiver operating characteristic analysis to assesspredictive value for presence of conventional guidelines-based mitral surgical indication(s)

Parameter AUC 95% CI

Reservoir functionStrain (%) 0.80 0.72–0.87

Strain rate (sec�1) 0.75 0.66–0.84

Journal of the American Society of EchocardiographyVolume 26 Number 9

Debonnaire et al 1057

and systolic pulmonary arterial pressure at rest >50 mm Hg in 29(24%). Of note, no patients presented with LVEFs # 30%. Bydefinition, the presence of atrial fibrillation, New York HeartAssociation class III or IV symptoms, LV end-systolic diameter,LVEF, and pulmonary arterial pressure significantly differed betweenpatients with conventional criteria for mitral surgery and those with-out (Table 1). In addition, age, use of b-blockers, use of diuretics,LAVI, MR vena contracta width, E wave, and deceleration timewere significantly different between the groups (P < .05 for all).

Expansion index (%) 0.79 0.71–0.87

Conduit functionStrain (%) 0.73 0.64–0.82

Strain rate (sec�1) 0.66 0.56–0.76

Passive emptying fraction (%) 0.60 0.49–0.71

Contractile function*Strain (%) 0.65 0.54–0.76

Strain rate (sec�1) 0.66 0.55–0.77

Active emptying fraction (%) 0.70 0.59–0.81

AUC, Area under the receiver operating characteristic curve.

Positive absolute numbers were used for all deformation indices.

*Measured only in 95 of 121 patients in sinus rhythm.

LA Function in Severe MR

Patients with severe organic MR showed significantly impaired LAreservoir function compared with control subjects, assessed bothby volumetric expansion index or strain and strain rate (P < .001for all), as indicated in Table 2. LA conduit function assessed withvolumetric passive emptying fraction (P = .20) or LA strain rate(P = .57) did not differ significantly between patients with severe or-ganic MR and controls, whereas LA conduit strain was significantlyimpaired in patients with severe organic MR (P = .01). In 95 patients(79%) who were in sinus rhythm during echocardiography, LA con-tractile function was reduced when measured by volumetric activeemptying function, LA strain, and strain rate (P < .001 for all).

LA Function to Predict Surgical Indication

As summarized in Table 2, LA dysfunction was more pronounced inpatients withmitral surgery indications versus patients without, and allphases of LA function (reservoir, conduit, and contractile) were af-fected, assessed by either volumetric or deformation parameters.Patients without mitral surgery indications showed impairment ofLA reservoir and contractile function but preserved conduit function.

Subsequently, receiver operating characteristic curve analysiswas performed to determine the value of all different deforma-tion and volumetric indices of phasic LA function to predictthe presence of guidelines-based mitral surgery indications(Table 3). This analysis showed that LA reservoir strain had thehighest predictive value among all indices of LA function, evi-denced by an area under the curve of 0.80 (95% confidence in-

terval [CI], 0.72–0.87; Figure 2A). In particular, a value of LAreservoir strain #24% predicted the presence of mitral surgeryindications with sensitivity and specificity of 76% and 72%, re-spectively (Figure 2B).

LA reservoir strainwas therefore implemented in further analysis toexplore the independent predictive value of LA function for the pres-ence of mitral surgery indications in patients with severe organic MR.Univariate logistic regression analysis , as indicated in Table 4, showedthat age (odds ratio [OR], 1.04; 95% CI, 1.01–1.07; P = .01), use ofb-blockers (OR, 3.7; 95% CI, 1.5–9.1; P = .004), use of diuretics(OR, 3.5; 95% CI, 1.4–8.9; P = .008), LAVI (OR, 1.02; 95% CI,1.01–1.04; P = .003), MR vena contracta width (OR, 2.29; 95% CI,1.43–3.66; P = .001), and LA reservoir strain (OR, 0.86; 95% CI,0.81–0.92; P < .001) were significantly related to the presence ofmitral surgery indications in patients with severe MR. Multivariatelogistic regression analysis subsequently identified LA reservoir strain

Figure 2 LA reservoir strain to predict the presence ofmitral sur-gery indication in patients with severe MR. (A) Receiver operat-ing characteristic curve. (B) Sensitivity-specificity curve withvertical line indicating a reservoir strain cutoff value of #24%to predict the presence of surgical indication with sensitivity of76% and specificity of 72%.

Figure 3 Likelihood ratio test for the prediction of long-termmortality after mitral valve surgery. LA reservoir strain shows in-cremental value to predict long-term postoperative mortality insevere organic MR when assessed on top of conventionalguidelines-based indications for surgery. AF, Atrial fibrillation;LVESD, LV end-systolic diameter;NYHA, New York Heart Asso-ciation; PAP, systolic pulmonary artery pressure at rest.

1058 Debonnaire et al Journal of the American Society of EchocardiographySeptember 2013

as an independent predictor of the presence of mitral surgery indica-tion (OR, 0.88; 95% CI, 0.82–0.94; P < .001), together with MRvena contracta width (OR, 2.04; 95% CI, 1.21–3.47; P = .008) andthe use of b-blockers (OR, 3.2; 95%CI, 1.2–8.7; P= .03). Each 1%de-crease in LA reservoir strain increased the risk for having an indicationfor mitral surgery by 12%.

LA Strain and Long-Term Survival after Mitral ValveSurgery

After amedian follow-up period of 6.4 years (interquartile range, 4.7–8.7 years) after mitral valve surgery, all-cause mortality occurred in 20of the 117 operated patients (17%). Cox regression analysis, restrictedto LA function (reservoir strain) and remodeling parameters (LAdiam-eter and LAVI), together with conventional criteria for mitral surgery,indicated that LA reservoir strain, using the#24%cutoff value (hazardratio, 3.8; 95% CI, 1.10–12.93; P = .04), and symptoms (hazard ratio,2.3; 95% CI, 0.94–5.71; P = .07) were related to long-term mortalityafter mitral valve surgery. In addition, LA diameter$ 55 mm or LAVI$ 60 mL/m2 did not show a significant relation with postoperativemortality.3,4 At the multivariate level, only LA reservoir strain(hazard ratio, 3.5; 95% CI, 1.03–12.20; P = .045) was shown to be

independently related to mortality. In addition, the likelihood ratiotest indicated that the assessment of LA reservoir strain, using thecutoff value, on top of the assessment of conventional guidelines-based indications for mitral surgery provides significant incrementalvalue to predict long-term postoperative mortality (P = .03; Figure 3)

Kaplan-Meier survival analysis showed that patients with baselineLA reservoir strain values # 24% had worse long-term survivalthan those with LA reservoir strain > 24% (log-rank P = .02;Figure 3). This association was also observed among asymptomaticpatients before mitral surgery (n = 83 of 117, log-rank P = .04;Figure 3). Interestingly, a significant trend toward worse survivalwas observed in 44 of 117 patients with baseline LA reservoir strain# 24%who underwent early mitral valve surgery despite the absenceof any conventional surgical indication compared with patients withLA reservoir strain > 24% (log-rank P = .06; Figure 4).

LA Reservoir Strain versus Diameter and Volume

LA reservoir strain showed significant linear correlations with LA di-ameter (R = �0.57; R2 = 0.33; P < .001; 95% CI, �0.79 to �0.46)and LAVI (R = �0.57; R2 = 0.32; P < .001; 95% CI, �0.18 to�0.11) (Figure 5). However, 47 of 98 patients (48%) with LA diam-eters < 55 mm and 15 of 44 patients (34%) with LAVI < 60 mL/m2

had LA reservoir strain # 24% (Figure 5).

Observer Variability

Bland-Altman analysis showed intraobserver bias of 1.1% (95% CI,�4.2% to 6.4%), �0.02% (95% CI, �5.1% to 5.0%), and �1.1%(95% CI, �4.5% to 2.3%) for LA reservoir, conduit, and contractilestrain assessment, respectively. Similarly, respective interobserverbiases of �1.3% (95% CI, �6.6% to 3.9%), 1.2% (95% CI, �2.8%to 5.1%), and 0.1% (95% CI, �4.1% to 4.3%) were noted.

DISCUSSION

Themain findings of the present study can be summarized as follows:(1) patients with chronic severe organic MR are characterized byoverall impaired LA reservoir, conduit, and contractile function;

Figure 4 Kaplan-Meier survival after mitral valve surgery according to LA reservoir strain. Conventional surgical criteria includesymptoms, LVEF # 60%, LV end-systolic diameter # 40 mm, atrial fibrillation, and/or systolic arterial pulmonary pressure at rest>50 mm Hg.

Figure 5 Correlation between LA reservoir strain, diameter, and indexed volume. LA reservoir strain # 24% (associated with worsesurvival after mitral valve surgery) is present in 48% and 34% of patients despite LA diameter < 55 mm and LAVI < 60 mL/m2, respec-tively (purple squares).

Journal of the American Society of EchocardiographyVolume 26 Number 9

Debonnaire et al 1059

(2) LA reservoir strain is an accurate and independent predictor of thepresence of conventional guidelines-based mitral surgery indications;and (3) LA reservoir strain relates to long-term survival after mitralvalve surgery, independently of and incremental to current conven-tional guidelines-based mitral surgery indications.

Phasic LA Function in Severe MR

LA Reservoir Function. During LV systole, pulmonary venous in-flow distends the left atrium, which acts as a reservoir by storing en-ergy in the form of pressure.14 LA reservoir function is determinedmainly by LA compliance (stiffness) and LV systolic function (systolicapical motion of the LV base, facilitating LA filling), and to a lesser ex-tent by right ventricular systole (pulmonary venous inflow).13,14,24

Although increased LA reservoir function has been reported inmild MR, because of an initial increase in LA compliance,1,9,10

decreased LA reservoir function has been described in patients withsevere MR, as assessed with several imaging techniques.9,10,12

Similarly, the present study showed that severe MR is associatedwith a significant reduction in LA reservoir function.

Reduced LV systolic function may lead to impaired LA reservoirfunction, as previously demonstrated.24 However, in the presentstudy, LVEF was preserved in the majority of the patients, so this fac-tor per se may not explain the impairment of LA reservoir function.

Furthermore, significant ultrastructural changes of the LA myocar-dium, including the presence of interstitial fibrosis, myocyte hypertro-phy, chronic inflammatory changes, and decreased capillary density,have been reported in patients with severe MR and might causethe reduction in LA reservoir function.5-7,25 In particular, a recentstudy of 46 patients with severe MR who underwent mitral valvesurgery showed a strong negative correlation between histologic LAfibrosis and LA reservoir strain.26 Ultrastructural changes are due tothe chronic volume overload and increased LV filling pressures andsignificantly affect LA myocardial wall properties, such as relaxationand compliance (stiffness), ultimately leading to the observed reduc-tion in LA reservoir function.3

Table 4 Cox regression analysis for postoperative all-cause mortality

Variable

Univariate Multivariate

HR (95% CI) P HR (95% CI) P

Symptoms (NYHA class III or IV) 2.3 0.94–5.71 .07 2.1 0.86–5.26 .10

LVEF # 60% 1.1 0.31–3.63 .93

LVESD $ 40 mm 2.0 0.77–4.96 .16Atrial fibrillation 1.8 0.70–4.87 .21

Systolic PAP at rest >50 mm Hg 1.2 0.46–3.20 .67LA diameter $ 55 mm 2.5 0.57–10.8 .23

LAVI $ 60 mL/m2 1.8 0.58–5.35 .32LA reservoir strain # 24% 3.8 1.10–12.93 .04 3.5 1.03–12.20 .045

HR, Hazard ratio; LVESD, LV end-systolic diameter; NYHA, New York Heart Association; PAP, arterial pulmonary artery pressure.

The relation of currently recommended clinical and echocardiographic variables for decision making in severe organic MR with long-term post-operative mortality is explored.

1060 Debonnaire et al Journal of the American Society of EchocardiographySeptember 2013

LA Conduit Function. After the reservoir phase, the left atriumfunctions as a conduit during early diastole, allowing a passive transferof blood from the pulmonary veins toward the left ventricle. LA con-duit function is determined mainly by LA elasticity (recoil) and after-load (LV relaxation and early LV filling pressure).13,14 In the presentstudy, reduced LA conduit function was observed in patients withsevere MR and mitral surgery indication. Increased early diastolicLV filling pressures have been reported in subjects with MR, due toincreased atrioventricular pressure gradient, increased LV stiffness,and decreased LA recoil, and might therefore be an explanation forimpaired LA conduit function in these patients.9,25,27 In addition,ultrastructural changes of the LA wall may cause impaired LAelasticity, reducing LA recoil and affecting LA conduit function.3,12

LA Contractile Function. In patients in sinus rhythm, active LAcontraction finalizes LV filling during late diastole and depends onLA afterload (LV compliance and end-diastolic filling pressure) andLA intrinsic contractility, following Starling’s law.13,14,28 ImprovedLA contractile function has been reported in patients with mild MRand may be explained by the fact that the LA myocardial walloperates higher on the Frank-Starling curve (larger LA volume leadsto higher contractility).9,10,28 In the present study, including patientswith severe organic MR, impaired LA contractile function assessedusing speckle-tracking echocardiography was observed. This findingmay be related to the presence of decreased LA contractility, dueto exhaustion of the Frank-Starling mechanism by severe volumeoverload and to ultrastructural alterations of the atrial myocardium.Moreover, increased LA afterload may have also contributed tothe impairment of LA contractile function, because increased LVend-diastolic filling pressures have been observed in the presence ofsevere MR.28

LA Reservoir Strain and Mitral Surgical Indication

Optimal timing of surgical treatment of severe organic MR is crucial,particularly in asymptomatic patients.2 Current guidelines recom-mend mitral valve surgery on the basis of the presence of symptoms,LV systolic dysfunction, atrial fibrillation, or pulmonary hypertension,which reflect MR severity and its impact on cardiac performance andsymptomatic status.2,19 However, severe MR primarily affects the leftatrium, causing significant LA dysfunction and remodeling, whichimplies that LA characteristics may indicate the hemodynamicimplications of severe MR at an earlier stage than conventionalcriteria for surgical indication.29 The predictive value of LA reservoir

strain observed in the present evaluation might be explained by thefact that reduced LA compliance, and therefore impaired LA reser-voir function, is directly related to the duration and severity of MRand represents an important substrate for the occurrence of symp-toms, atrial fibrillation, or the development of pulmonary hyperten-sion. In fact, decreased atrial compliance due to increased LAstiffness may cause inadequate LA pressure homeostasis. Therefore,small increases in LA volume may lead to significant increases ofLA and pulmonary artery pressures, which can be associated withthe development of pulmonary hypertension and symptoms duringexercise.13 This has been demonstrated in other clinical conditions,for example, in patients with diastolic heart failure.30 Reduced LA res-ervoir strain, not LA volume, was shown to be associated with thepresence of symptoms and was directly related to atrial stiffness.30

In addition, LA reservoir strain was demonstrated to be inverselyrelated to LA wall fibrosis (affecting atrial compliance) and to beassociated with a higher burden of atrial fibrillation.14 Finally, atmore advanced stages, severe MR may cause LV systolic dysfunction,which can decrease LA reservoir function by reduced systolic apicalmotion of the LV base.24

LA Reservoir Strain and Mortality after Mitral ValveSurgery

The present evaluation showed that LA reservoir strain is related tolong-term survival in patients operated for severe organic MR, inde-pendently of conventional indications for mitral surgery, includingsymptoms, LV dysfunction, atrial fibrillation, and significant pulmo-nary hypertension. In particular, LA reservoir strain# 24% was asso-ciated with increased risk for postoperative mortality. In addition toconventional mitral surgery indications, LA diameter $ 55 mm andLAVI$ 60mL/m2 have been recently proposed as additional criteriafor indication to mitral surgery, because they yield increased mortalityunder conservative management, but not after mitral valve surgery.We found, however, that LA reservoir might be a more sensitive ap-proach than LA diameter or LAVI to detect the impact of MR on car-diac performance, as previously shown in patients with diabetes andhypertension.8 A significant proportion of patients in fact showed LAreservoir strain < 24%, implying increased mortality risk after mitralvalve surgery, despite LA diameter < 55 mm or LAVI < 60 mL/m2.

Limitations

Some limitations to the present study should be acknowledged. First,no invasive assessment of LA mechanical properties or afterload was

Journal of the American Society of EchocardiographyVolume 26 Number 9

Debonnaire et al 1061

performed to prove the determinants of reduced LA function, whichmust be the focus of further specific studies.

Second, limited impairment of LA function in the control group ofthe present study cannot be excluded, because of a fairly high propor-tion of subjects with cardiovascular risk factors such as diabetes andhypertension. This population, however, represents a real-life popula-tion without the presence of overt structural or functional disease, andmore stringent selection of patients without any risk factors mostlikely would have only amplified the significance of differences inLA function found between controls and patients with severe MR.

Third, assessment of LV global longitudinal strain was not includedin the present analysis, although it has been shown to have potentialvalue in predicting postoperative LV dysfunction in patients withsevere organic MR.31 Including a specific analysis to demonstratethe postoperative prognostic value of this additional parameter wouldincrease significantly the complexity of this study, whose aim was tofocus on the clinical value of LA function.

Last, to firmly establish the predictive value of LA reservoir strainindependent of conventional indications for mitral surgery, our find-ings need validation in larger patient cohorts.

CONCLUSIONS

Severe MR is associated with significant LA dysfunction that stronglyrelates to the presence of conventional indications for mitral surgeryand, more important, is associated with long-term postoperative sur-vival. LA dysfunction, measured as reservoir strain, may be a valuableclinical marker for follow-up and decision making in patients with se-vere organic MR, irrespective of conventional mitral surgery indica-tions and more sensitive than LA diameter and volume.

REFERENCES

1. Kihara Y, Sasayama S, Miyazaki S, Onodera T, Susawa T, Nakamura Y,et al. Role of the left atrium in adaptation of the heart to chronic mitral re-gurgitation in conscious dogs. Circ Res 1988;62:543-53.

2. Bonow RO, Carabello BA, Chatterjee K, de Leon AC Jr., Faxon DP,Freed MD, et al. 2008 focused update incorporated into the ACC/AHA 2006 guidelines for the management of patients with valvular heartdisease: a report of the American College of Cardiology/American HeartAssociation Task Force on Practice Guidelines (Writing Committee to Re-vise the 1998 Guidelines for the Management of Patients With ValvularHeart Disease): endorsed by the Society of Cardiovascular Anesthesiolo-gists, Society for Cardiovascular Angiography and Interventions, andSociety of Thoracic Surgeons. Circulation 2008;118:e523-661.

3. Le Tourneau T, Messika-Zeitoun D, Russo A, Detaint D, Topilsky Y,Mahoney DW, et al. Impact of left atrial volume on clinical outcome inorganic mitral regurgitation. J Am Coll Cardiol 2010;56:570-8.

4. Rusinaru D, Tribouilloy C, Grigioni F, Avierinos JF, Suri RM, Barbieri A,et al. Left atrial size is a potent predictor of mortality in mitral regurgitationdue to flail leaflets: results from a large international multicenter study.Circ Cardiovasc Imaging 2011;4:473-81.

5. AnneW,Willems R, Roskams T, Sergeant P, Herijgers P, Holemans P, et al.Matrix metalloproteinases and atrial remodeling in patients with mitralvalve disease and atrial fibrillation. Cardiovasc Res 2005;67:655-66.

6. Corradi D, Callegari S, Maestri R, Ferrara D, Mangieri D, Alinovi R, et al.Differential structural remodeling of the left-atrial posterior wall in patientsaffected by mitral regurgitation with or without persistent atrial fibrillation:a morphological and molecular study. J Cardiovasc Electrophysiol 2012;23:271-9.

7. Verheule S, Wilson E, Everett T, Shanbhag S, Golden C, Olgin J. Alter-ations in atrial electrophysiology and tissue structure in a canine model

of chronic atrial dilatation due to mitral regurgitation. Circulation2003;107:2615-22.

8. Mondillo S, Cameli M, Caputo ML, Lisi M, Palmerini E, Padeletti M, et al.Early detection of left atrial strain abnormalities by speckle-tracking inhypertensive and diabetic patients with normal left atrial size. J Am SocEchocardiogr 2011;24:898-908.

9. Borg AN, Pearce KA, Williams SG, Ray SG. Left atrial function and defor-mation in chronic primary mitral regurgitation. Eur J Echocardiogr 2009;10:833-40.

10. Cameli M, Lisi M, Giacomin E, Caputo M, Navarri R, Malandrino A, et al.Chronic mitral regurgitation: left atrial deformation analysis by two-dimensional speckle tracking echocardiography. Echocardiography 2011;28:327-34.

11. Cameli M, Lisi M, Righini FM, Focardi M, Alfieri O, Mondillo S. Left atrialspeckle tracking analysis in patients with mitral insufficiency and history ofparoxysmal atrial fibrillation. Int J Cardiovasc Imaging 2012;28:1663-70.

12. Moustafa SE, Alharthi M, Kansal M, Deng Y, Chandrasekaran K,Mookadam F. Global left atrial dysfunction and regional heterogeneity inprimary chronic mitral insufficiency. Eur J Echocardiogr 2011;12:384-93.

13. Pagel PS, Kehl F, Gare M, Hettrick DA, Kersten JR, Warltier DC. Mechan-ical function of the left atrium: new insights based on analysis of pressure-volume relations and Doppler echocardiography. Anesthesiology 2003;98:975-94.

14. Rosca M, Lancellotti P, Popescu BA, Pierard LA. Left atrial function: path-ophysiology, echocardiographic assessment, and clinical applications.Heart 2011;97:1982-9.

15. Kim DG, Lee KJ, Lee S, Jeong SY, Lee YS, Choi YJ, et al. Feasibility of two-dimensional global longitudinal strain and strain rate imaging for the as-sessment of left atrial function: a study in subjects with a low probabilityof cardiovascular disease and normal exercise capacity. Echocardiography2009;26:1179-87.

16. Saraiva RM, Demirkol S, Buakhamsri A, Greenberg N, Popovic ZB,Thomas JD, et al. Left atrial strain measured by two-dimensional speckletracking represents a new tool to evaluate left atrial function. J Am SocEchocardiogr 2010;23:172-80.

17. Vianna-Pinton R,MorenoCA, Baxter CM, Lee KS, Tsang TS, AppletonCP.Two-dimensional speckle-tracking echocardiography of the left atrium:feasibility and regional contraction and relaxation differences in normalsubjects. J Am Soc Echocardiogr 2009;22:299-305.

18. Motoki H, Dahiya A, Bhargava M, Wazni OM, Saliba WI, Marwick TH,et al. Assessment of left atrial mechanics in patients with atrial fibrillation:comparison between two-dimensional speckle-based strain and velocityvector imaging. J Am Soc Echocardiogr 2012;25:428-35.

19. Joint Task Force on the Management of Valvular Heart Disease of the Eu-ropean Society of Cardiology, European Association for Cardio-ThoracicSurgeryVahanian A, Alfieri O, Andreotti F, Antunes MJ, et al. Guidelineson the management of valvular heart disease (version 2012). Eur Heart J2012;33:2451-96.

20. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA,et al. Recommendations for chamber quantification: a report from theAmerican Society of Echocardiography’s Guidelines and Standards Com-mittee and the ChamberQuantificationWriting Group, developed in con-junction with the European Association of Echocardiography, a branch ofthe European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440-63.

21. Lancellotti P, Moura L, Pierard LA, Agricola E, Popescu BA, Tribouilloy C,et al. European Association of Echocardiography recommendations forthe assessment of valvular regurgitation. Part 2: mitral and tricuspid regur-gitation (native valve disease). Eur J Echocardiogr 2010;11:307-32.

22. Nagueh SF, Appleton CP, Gillebert TC, Marino PN, Oh JK, Smiseth OA,et al. Recommendations for the evaluation of left ventricular diastolicfunction by echocardiography. J Am Soc Echocardiogr 2009;22:107-33.

23. Mor-Avi V, Lang RM, Badano LP, Belohlavek M, Cardim NM,Derumeaux G, et al. Current and evolving echocardiographic techniquesfor the quantitative evaluation of cardiac mechanics: ASE/EAE consensusstatement on methodology and indications endorsed by the Japanese So-ciety of Echocardiography. J Am Soc Echocardiogr 2011;24:277-313.

1062 Debonnaire et al Journal of the American Society of EchocardiographySeptember 2013

24. Barbier P, Solomon SB, Schiller NB, Glantz SA. Left atrial relaxation andleft ventricular systolic function determine left atrial reservoir function.Circulation 1999;100:427-36.

25. Zile MR, Tomita M, Nakano K, Mirsky I, Usher B, Lindroth J, et al. Effectsof left ventricular volume overload produced bymitral regurgitation on di-astolic function. Am J Physiol 1991;261:H1471-80.

26. Cameli M, Lisi M, Righini FM, Massoni A, Natali BM, Focardi M, et al. Use-fulness of atrial deformation analysis to predict left atrial fibrosis and endo-cardial thickness in patients undergoing mitral valve operations for severemitral regurgitation secondary to mitral valve prolapse. Am J Cardiol2013;111:595-601.

27. Katayama K, Tajimi T, Guth BD,Matsuzaki M, Lee JD, Seitelberger R, et al.Early diastolic filling dynamics during experimental mitral regurgitation inthe conscious dog. Circulation 1988;78:390-400.

28. Sasayama S, Takahashi M, Osakada G, Hirose K, Hamashima H,Nishimura E, et al. Dynamic geometry of the left atrium and left ventriclein acute mitral regurgitation. Circulation 1979;60:177-86.

29. Messika-Zeitoun D, Bellamy M, Avierinos JF, Breen J, Eusemann C,Rossi A, et al. Left atrial remodelling in mitral regurgitation–methodologicapproach, physiological determinants, and outcome implications: a pro-spective quantitative Doppler-echocardiographic and electron beam-computed tomographic study. Eur Heart J 2007;28:1773-81.

30. Kurt M, Wang J, Torre-Amione G, Nagueh SF. Left atrial function in dia-stolic heart failure. Circ Cardiovasc Imaging 2009;2:10-5.

31. Witkowski TG, Thomas JD, Debonnaire P, Delgado V, Hoke U,Ewe SH, et al. Global longitudinal strain predicts left ventricular dys-function after mitral valve repair. Eur Heart J Cardiovasc Imaging2013;14:69-76.