Systolic Papillary Muscle Dyssynchrony PredictsRecurrence of Mitral Regurgitation in Patients withIschemic Cardiomyopathy (ICM) Undergoing MitralValve Repair
Leen van Garsse, M.D.,*1 Sandro Gelsomino, M.D., Ph.D.,*1,† Orlando Parise, M.Sc.,† Fabiana Lucà, M.D.,†Emile Cheriex, M.D., Ph.D.,* Roberto Lorusso, M.D., Ph.D.,‡ Enrico Vizzardi, M.D.,‡ Carmelo MassimilianoRao, M.D.,† Gian Franco Gensini, M.D.†, and Jos Maessen, M.D., Ph.D.*
*Department of Cardiothoracic Surgery, University Hospital, Maastricht, The Netherlands; †Department ofHeart and Vessels, Careggi Hospital, Florence, Italy; and ‡Cardiac Surgery, Civic Hospital, Brescia, Italy
Objective: We investigated the impact of papillary muscle dyssynchrony (DYS-PAP) in predicting recur-rent mitral regurgitation (MR) in patients with ischemic cardiomyopathy (ICM) undergoing undersizedmitral ring annuloplasty (UMRA). Methods: One hundred forty-four ICM patients (left ventricular ejec-tion fraction <35%) in sinus rhythm undergoing UMRA between January 2001 and December 2010 atthree Institutions (University Hospital, Maastricht, The Netherlands; Careggi Hospital, Florence, Italy;Civic Hospital, Brescia, Italy) were recruited. The primary endpoint was the recurrence of MR at the lat-est echocardiographic study defined as insufficiency �2+ in patients with no/trivial MR at discharge.The assessment of DYS-PAP was performed by applying two-dimensional (2D) speckle-tracking imag-ing. Results: In patients with MR recurrence, DYS-PAP significantly worsened (84.1 ± 8.8 msecvs.65.4 ± 8.8 msec at baseline, P < 0.001) whereas in patients with no MR recurrence, DYS-PAP didnot vary (22.3 ± 5.3 msec vs. 25.9 ± 7.2 msec at baseline, P = 0.8). Recurrent MR was positively corre-lated with preoperative DYS-PAP (P < 0.001), baseline anterior mitral leaflet tethering anglea (P < 0.001) and tethering symmetry index a/b before surgery (P < 0.001). There was no significantcorrelation between MR recurrence and other echocardiographic parameters. Logistic regression analy-sis revealed that baseline values of DYS-PAP (OR: 5.4 [95% CI: 3.1–7.7], P < 0.001), a (OR: 5.0 [2.6–6.7], P < 0.001), and a/b (OR: 3.9 [2.5–5.7], p < 0.001) were predictors of recurrent MR. A DYS-PAPvalue � 58 msec predicted recurrence of MR with 100% sensitivity and 83% specificity (area underthe curve [AUC]: 0.92 [0.7–1], P < 0.001). Conclusions: A DYS-PAP cutoff value of 58 msec is useful toidentify patients in whom UMRA is likely to fail. That way decision making in ischemic functional MRmight be facilitated. (Echocardiography 2012;29:1191-1200)
Functional mitral regurgitation (FMR) iscommonly observed in ischemic cardiomyopathy(ICM) and is associated with a poor prognosis.1
Undersized mitral ring annuloplasty (UMRA) hasbeen widely used in recent years for the treat-ment of ischemic FMR. Nonetheless, while somepatients do show echocardiographic evidence ofvalve competence and left ventricular (LV)reverse remodeling,2 in a large percentage of
patients recurrence of MR associated with a per-sistent or progressive remodeling pattern hasclearly been documented over time despite aninitially successful surgical repair.3
Cardiac dyssynchrony (DYS) is associated withfunctional deterioration and poor prognosis inadvanced systolic heart failure4 and is an impor-tant determinant of postoperative outcome inpatients with severe LV dysfunction.5 Further-more, DYS is a strong predictor of LV remodelingafter acute myocardial infarction.6
Two-dimensional (2D) strain imaging basedon novel speckle-tracking echocardiography(STE) is a relatively new tool to define regional
1The first two authors equally contributed to the article.Address for correspondence and reprint requests: Dr. SandroGelsomino, Experimental Surgery Unit, Careggi Hospital VialeMorgagni 85, 50134, Florence, Italy. Fax + 39-055-7947628; E-mail: [email protected]
myocardial strain and to quantify dyssynchronybased on a more robust technique and avoidingangle of incidence.7
We tested the hypothesis that systolic papil-lary muscle dyssynchrony (DYS-PAP) evaluatedby 2D-STE can predict MR recurrence in patientswith ICM undergoing UMRA.
Methods:Patients:Patients with ICM (LV ejection fraction [LVEF]<35%) and in sinus rhythm, undergoing UMRAbetween January 2001 and December 2010 atthree Institutions (University Hospital, Maastricht,The Netherlands; Careggi Hospital, Florence,Italy; Civic Hospital, Brescia, Italy) were includedin the study. Definitions, inclusion, and exclusioncriteria were as previously reported.8 Patientswere also excluded if they had residual MRdefined as a mitral insufficiency �2 at discharge,if they had QRS > 120 msec/undergoing biven-tricular pacing, or if echocardiograms were notavailable/incomplete or images were not appro-priate for 2D-STE. One hundred forty-fourpatients met the inclusion criteria and repre-sented the study population whereas 35 wereexcluded.
The primary endpoint was the recurrence ofMR (insufficiency �2+ in patients with no/trivialMR at discharge) at the latest echocardiographicstudy performed at a median of 39.3 months (in-terquartile range [IQR] 18.9–46.5).
Ethical Committee approval was waived dueto the retrospective analysis of the study accord-ing to National laws regulating observational ret-rospective studies (Italian law nr.11960, releasedon 13/7/2004; Dutch WMO law). However, allpatients gave their informed consent to accesstheir data for scientific purposes. Preoperativecharacteristics of the patients are shown inTable I. Ninety-eight (68.1%) patients had �moderate recurrent MR at follow-up whereas 46(31.9%) did not show recurrent MR at the lastechocardiographic study. No difference wasdetected in preoperative and surgical databetween the two groups.
Surgery:Surgery was performed with standard operativetechniques, including cardiopulmonary bypassand undersized annuloplasty. Downsizing by tworing sizes was performed in all patients.
All patients underwent associated coronaryartery bypass grafting (CABG). For the purposesof this study, complete revascularization wasaccomplished when at least one graft was placeddistally to an approximately 50% diameter nar-rowing in each of the three major vascular sys-tems in which arterial narrowing of this severity
was noted in a vessel �1.5 mm in diameter. Fol-lowing this definition, 100% of patientsunderwent complete revascularization. After car-diopulmonary bypass, a transesophageal echo-cardiography (TEE) was performed to assessmitral valve repair: leaflet coaptation �0.5 cm,MR � 1, and diastolic MV area >2 cm2 wereassessed as successful repair.
Standard Echocardiographic Measurements:Transthoracic echocardiography (TTE) was per-formed following a common standard protocolat baseline (within a week before surgery) at dis-charge, 6 months, and at yearly follow-up visits.Two-dimensional echoes were carried out using acommercially available ultrasound system IE 33,(Philips Medical System, Amsterdam, The Nether-lands). Images were stored in DICOM format andtransferred to a workstation for further offlineanalysis (Tomtec Imaging system, Untersch-leissheim, Germany). Measurements and calcula-tions were made separately by one of theinvestigators (F.L.). Echocardiographic measure-ments and calculations were carried out as previ-ously reported.8 The following quantitativemeasurements were simultaneously employed tograde the severity of MR: (1) pulsed Dopplerquantitative flow methods; (2) proximal isoveloc-ity surface area (PISA).9 Measurements as well asrespective thresholds for mild, moderate, andsevere MR followed the American Society ofEchocardiography (ASE) recommendations.9
Mitral valve configuration was assessed inmid-systole using the parasternal long-axis andfour-chamber views.10 The anterior mitral leaflet(AML) tethering angle a (between the annularplane and the basal part of the AML), the poster-ior mitral leaflet (PML) tethering angleb (between the angular plane and the basal partof the PML), and the bending angle c (betweenthe bending distance [from the anterior annulusto the bending point created by the tethering ofintermediate or strut chordae in the body of theanterior leaflet] and the line from the bendingpoint to the coaptation point) were directly mea-sured using specific software (Philips DICOMViewer, Philips Medical System).10 The excursionangles aex and bex were calculated as the differ-ence between AML and PML angles in systoleand diastole. The anterior/posterior tetheringangle ratio a/b was a quantitative measurementof tethering pattern: the more this ratioapproached 1 the more symmetric was the teth-ering.8
The tenting area (TA) was measured by thearea enclosed between the annular plane andmitral leaflets from the parasternal long-axis viewat mid-systole. The coaptation height (CH) wasmeasured as the perpendicular distance between
1192
van Garsse, et al.
the coaptation point of the mitral leaflets and theline connecting the annular hinge points in thelong-axis view at end-systole. The coaptationlength (CL) was measured as the length of appo-sition of the anterior and PMLs. The coaptationdistance (CD) was measured (along the annularplane) from the anterior leaflet attachment to thepoint of coaptation.
The LV volumes and LVEF were assessed bythe biapical Simpson disk method.11 Sphericity
indexes were obtained at end-diastole and end-systole (SIDia and SISys, respectively) as the vol-ume of the LV divided by the volume of a spherewith a diameter equal to the longest axis of theLV measured in the apical view.12
The displacement of papillary muscle wasquantified as distances from well-defined ana-tomic landmarks at early and end-systole. Fromthe parasternal short-axis view, the geometricchord defined by the intersection of the right
TABLE I
Preoperative Demographic Clinical and Surgical Data (n = 144)
Normally distributed variables are presented as mean ± standard deviation; discrete variables are presented as percentages.Nonnormally distributed variables were presented as median [interquartile range]. M/F = male/female; NYHA = New York HeartAssociation; COPD = chronic obstructive pulmonary disease; CPB = cardiopulmonary bypass; CCL = (aortic) cross-clamp; CABG =coronary artery bypass grafting; MR = mitral regurgitation; TA = tenting area; EROA = effective regurgitant orifice area; RF =regurgitant fraction; RV = regurgitant volume; CH = coaptation height; CL = coaptation length; CD = coaptation distance.
1193
Papillary Muscle Systolic Dyssynchrony in ICM
and left ventricles (“septal insertions”) and themid-septal perpendicular line were used as refer-ences.10 Lateral and inferior displacements ofanterior and posterior papillary muscles weremeasured as distances from these fixed refer-ences. Separation between papillary muscles wasdirectly measured.10
The lengths between the anterolateral papil-lary muscle (ALPM) and posteromedial papillarymuscle (PMPM) tips and the contralateral ante-rior mitral annulus (l1 and l2, respectively) werealso measured in mid-systole in the apical four-and two-chamber views by using the anteriormitral annulus as a reference point to estimateoutward PM displacement.13
The WMSI of the basal and mid-posterior andinferior segments for the PMPM (PMPM–WMSI)and basal and mid-lateral and anterior segmentsfor the ALPM (ALPM–WMSI) were also calcu-lated.14
Special Echocardiographic Readings:Assessment of longitudinal peak systolic strain wasperformed by applying 2D speckle-tracking imag-ing to the apical two- and four-chamber views ofthe LV. All images were obtained at a frame rate of70–90 frames/sec. The LV was divided into six seg-ments in each apical view and the global longitu-dinal strain (GLS) was obtained in each view.14
The peak global longitudinal strain (GLSpeak) wasthe mean of the peak negative values on the twoGLS curves. The assessment of DYS-PAP was per-formed by applying 2D speckle-tracking imagingto the apical four-chamber view for ALPM and tothe apical long-axis view for PMPM.15
From an end-systolic single frame, the edge ofeach papillary muscle was detected on theendocardial cavity and an automated tracking
algorithm (free-trace method, Tomtec, TomtecImaging system) followed the papillary musclefrom this single frame throughout the cardiaccycle. The value of peak systolic longitudinalstrain for each papillary muscle was then deter-mined as ALPM longitudinal strain (ALPM-LS)and PMPL longitudinal strain (PMPM-LS). Thebeginning of the QRS complex was used as thereference point and the time to peak (TTP) sys-tolic longitudinal strain was quantified for eachpapillary muscle. For the assessment of DYS-PAP,the absolute difference in TTP between anterolat-eral and posteromedial papillary muscle wascalculated (Fig. 1).
The reliability of echocardiographic measure-ments was assessed by calculating intraobserverintervals of agreement of main direct measuresused in this study in 20 subjects randomly chosenamong the study patients. The Bland–Altmanmethod showed excellent agreement betweenintraobserver measurements in both low andhigh values of echocardiographic parameters(Table II).
Statistics:Statistical analysis was performed using a statisti-cal software program (SPSS, version 12.0, SPSSInc. Chicago, IL, USA). Data are presented asmean and standard deviations or as median andIQR after being controlled for normal distributionby the Kolmogorov–Smirnov test. Data werecompared for statistical significance using at-test, Mann–Whitney rank sum test, chi-squaretest, or Fisher’s exact test where appropriate.Multiple comparisons were carried out by analy-sis of variance (ANOVA) and Kruskal–Wallis testswith Tukey and Dunn post hoc tests. Pearson’scorrelation was used to test linear relationships
Figure 1. Representative example of measuring papillary muscle dyssynchrony (DYS-PAP) as difference in time to peak (TTP)between anterolateral and posteromedial papillary muscles. AVC = aortic valve closure; ALPM = anterolater papillary muscle;PMPM = posteromedial papillary muscle.
1194
van Garsse, et al.
between variables. We assessed, on the entirepatient population, univariable regression withpostoperative regurgitant volume as a depen-dent variable and other preoperative echocardio-graphic parameters (end-diastolic diameter[EDD], end-systolic diameter [ESD], end-diastolicvolume index [EDVI], end-systolic volume index[ESVI], LVEF, SISys, SIDia, TA, CH, CL, CD, AMLtethering angle a, PML tethering angle b, a/bratio, PMPM-WMSI, ALPM-WMSI, DYS-PAP,ALPM-LS, PMPM-LS, and GLSpeak) as indepen-dent variables. Variables reaching statisticalsignificance or borderline (P � 0.1) were intro-duced in multivariable analysis. Multivariate logis-tic regression with the forward stepwise methodwas performed to identify preoperative echocar-diographic predictors of significant postrepairMR. The goodness of fit of the final logistic regres-sion models was assessed with the Hosmer–Lemeshow (H-M) statistic and predictiveaccuracy was assessed by the concordance index(c). Optimal cutoff values were determined bythe receiver operating characteristic curve (ROC)as the rounding cutoff that gives the maximumsum of sensitivity and specificity. This valueshould be the shoulder at the top left of the ROCcurve. Results were validated using the bootstrapmethod (1000 iterations).
A P-value < 0.05 was considered statisticallysignificant.
Results:Leaflet Tethering and LV Remodeling:Table III shows changes in tethering configura-tion and LV remodeling before and after surgery.
At preoperative echo, patients who developedrecurrent MR had lower a excursion angle(P < 0.001) and higher a angle (P = 0.003), a/b
At the latest study, in patients with no recur-rent MR, a (P = 0.03), a/b ratio (P < 0.001),c (P = 0.01), and aex (P = 0.03) improved whereasb (P = 0.04) and b ex (P = 0.03) significantlyworsened. In patients with recurrent MR,a (P = 0.88), c (P = 0.6), and aex (P = 0.73) didnot change whereas b and bex (both, P < 0.001)significantly worsened. In these patients, the post-operative tethering became more asymmetric(P < 0.001). All indices of global LV remodelingsignificantly improved in MR� patients whereasthey worsened in MR+ patients. Ejection fractionincreased in MR� patients (P = 0.04), while it didnot change inMR+ patients (P = 0.85).
Regarding local LV remodeling, in patientswith no recurrent MR, posterior displacement(P = 0.03), lateral displacement (P = 0.002), andapical displacement (P = 0.02) of ALPM as well asALPM–WMSI (P = 0.01) improved at late follow-up whereas indices of displacement of PMPM sig-nificantly worsened (posterior, P = 0.001; lateral,P = 0.007; apical, P = 0.04; PLPM–WMSI, P =0.007).
In contrast, in patients with recurrent MR,indices of displacement of PMPM significantlyworsened (all, P < 0.001) whereas ALPM poster-ior displacement (P = 0.86), ALPM lateral dis-placement (P = 0.80), ALPM apical displacement,and ALPM–WMSI (both P > 0.9) did not change.
Two-Dimensional Speckle-Tracking Analysis:Results of 2D speckle-tracking analysis are shownin Table IV. In patients with MR recurrencePMPM-LS (P = 0.6), ALPM-LS (P = 0.78), andGLSPeak (P = 0.8) had not changed at the last fol-low-up when compared with baseline values. Incontrast, DYS-PAP significantly worsened(P < 0.001; Fig. 2A) at the last echocardio-graphic study compared to baseline values. Incontrast, in patients without MR recurrence,PMPM-LS (P < 0.001), ALPM-LS (P < 0.001), andGLSPeak (P = 0.001) improved whereas DYS-PAPdid not vary (P = 0.8; Fig. 2B) at the last postop-erative echocardiographic study. All these indiceswere significantly higher (all, P < 0.001) inpatients with recurrent MR compared with thepatients without MR.
Echocardiographic Predictors of RecurrentMR:Recurrent MR was positively correlated with pre-operative DYS-PAP (P < 0.001), baseline AMLtethering angle a (P < 0.001), and tetheringsymmetry index a/b before surgery (P < 0.001;Table V). There was no significant correlation
TABLE II
Bland–Altman Limits of Agreement for Intraobserver Variability
BiasStandardDeviation
95% Limit ofAgreement
ALPM-LS 0.03 0.23 �0.42 to 0.48PMPM-LS 0.05 0.16 �0.27 to 0.37GLSpeak �0.06 0.27 �0.61 to 0.48DYS-PAP �0.07 0.76 �1.57 to 1.42
Intraobserver and interobserver relative differences were <5%for all parameters. The Bland–Altman method showed excel-lent agreement between intraobserver measurements in bothlow and high values of echocardiographic parameters. ALPM-LS = anterolateral papillary muscle longitudinal strain; PMPM-LS = posteromedial papillary muscle longitudinal strain;GLSpeak = peak global longitudinal strain; DYS-PAP = papillarydyssynchrony.
1195
Papillary Muscle Systolic Dyssynchrony in ICM
TABLE
III
MitralLeafl
etTe
theringan
dLV
Remod
eling
Con
trols
All(n
=14
4)MR�
(n=46
)MR+
(n=98
)
Baselin
eFo
llow-up
PBa
selin
eFo
llow-up
PBa
selin
eFo
llow-up
P
Mitral
tethering
a(°)
23.2
±3.1
43.1
±6.4*
37.0
±5.5*
0.02
37.3
±4.6*
†28
.3±5.5*
‡0.03
48.9
±6.8*
47.0
±5.5*
0.88
b(°)
34.1
±5.8
51.3
±5.4*
58.0
±6.6*
0.00
949
.6±5.7*
54.6
±7.0*
‡0.04
53.1
±5.6*
63.3
±7.9*
0.03
a/b
0.72
±0.2
0.84
±0.1*
0.63
±0.1*
<0.00
10.75
±0.1*
†0.52
±0.1n
*‡<0.00
10.92
±0.1*
0.75
±0.1*
<0.00
1c(°)
155.1±8.4
125.4±5.7*
128.0±7.1*
0.61
122.1±3.4*
133.0±8.6*
‡0.01
127.7±7.3*
123.6±5.3*
0.6
a ex(°)
46.4
±6.7
25.5
±3.0*
31.5
±5.0*
0.03
32.1
±4.4*
†40
.6±7.5*
‡0.03
20.3
±2.7*
23.9
±4.4*
0.73
b ex(°)
28.0
±5.6
18.4
±3.5*
10.3
±3.5*
0.00
118
.2±3.3*
13.3
±3.1*
‡0.03
19.6
±3.7*
6.3±3.5*
<0.00
1Globa
lLVremod
eling
EDD(m
m)
49.1
±5
69.0
±6*
66.3
±7*
0.74
67.3
±6*
†52
.1±5*
‡0.00
474
.0±8*
80.6
±6*
0.06
ESD(m
m)
31.7
±4
58.5
±7*
55.7
±6*
0.70
58.4
±6*
44.3
±4*
‡<0.00
160
.3±8*
69.1
±0*
0.04
EDVI
(mL/m
2)
55.9
±8
158.3±24
*15
5.7±22
*0.38
149.0±20
*†12
6.0±15
*‡0.00
316
6.4±29
*18
0.1±25
*0.00
1ESVI
(mL/m
2)
25.2
±7
113.2±18
*10
9.6±15
*0.55
106.2±14
*†78
.4±9*
‡<0.00
111
9.0±23
*13
8.0±20
*<0.00
1SI
sys
0.52
±0.1
0.98
±0.1*
0.92
±0.1*
0.13
0.98
±0.1*
0.82
±0.1*
‡<0.00
10.99
±0.2*
1.09
±0.1*
0.04
SIDia:
0.64
±0.1
0.84
±0.1*
0.80
±0.1*
0.42
0.82
±0.1*
0.70
±0.1*
‡<0.00
10.86
±0.1*
0.91
±0.1*
0.02
LVEF
0.58
±0.1
0.28
±0.1*
0.29
±0.1*
0.11
0.30
±0.1*
0.40
±0.1*
‡0.04
0.25
±0.1*
0.24
±0.1*
0.85
LocalLVremod
eling
Posteriordisplacemen
tALP
M(cm)
2.1±0.2
3.1±0.3*
3.0±0.3*
0.88
2.6±0.2*
†2.3±0.2*
‡0.03
3.5±0.4*
3.6±0.3*
0.86
PMPM
(cm)
1.7±0.2
2.7±0.3*
3.3±0.5*
0.00
12.6±0.3*
3.3±0.5*
‡0.00
12.8±0.3*
3.8±0.5*
<0.00
1Laterald
isplacem
ent
ALP
M(cm)
0.9±0.2
1.6±0.4*
1.6±0.4*
>0.9
1.3±0.2*
†1.1±0.4*
‡0.00
22.0±0.5*
2.1±0.6*
0.8
PMPM
(cm)
0.4±0.2
1.9±0.3*
2.5±0.5*
0.00
71.9±0.4*
2.3±0.5*
‡0.00
71.9±0.3*
2.6±0.5*
<0.00
1Apicald
isplacem
ent
ALP
M(l 1
,cm)
2.3±0.6
4.1±0.8*
3.9±0.7*
0.80
3.3±0.7*
†3.0±05
*‡0.02
4.9±0.9*
4.9±0.8*
>0.9
PMPM
(12,c
m)
1.9±0.4
3.7±0.6*
4.4±0.6*
0.00
43.7±0.8*
3.9±0.7*
‡0.04
3.9±0.8*
4.9±0.8*
<0.00
1ALP
M-W
MSI
1.0±0
1.6±0.5*
1.5±0.5*
0.62
1.3±0.3*
†1.1±0.3*
‡0.01
1.9±0.7*
1.9±0.5*
>0.9
PLPM
-WMSI
1.0±0
1.4±0.3*
2.3±0.4*
<0.00
11.5±0.3*
2.0±0.3*
‡0.00
71.4±0.3*
2.6±0.4*
0.00
1
a=an
terio
rmitralleafl
ettetheringan
gle;
b=po
steriormitral
leafl
ettetheringan
gle;
c=be
ndingan
gle;
a ex=an
terio
rmitral
leafl
etex
cursionan
gle;
b ex=po
steriormitral
leafl
etex
cursion
angle;
EDD
=en
d-diastolic
diam
eter;ESD
=en
d-systolic
diam
eter;ED
VI=en
d-diastolic
volumeinde
x;ESVI
=en
d-systolic
volumeinde
x;SI
sys=systolic
sphe
ricity
inde
x;SI
Dia=diastolic
sphe
ricity
inde
x;LV
EF=leftventric
ular
ejectio
nfractio
n;PM
PM=po
sterom
edialp
apillarymuscle;
ALP
M=an
terolateralp
apillarymuscle;
WMSI
=wallm
otionscoreinde
x.P=Sign
ificanc
eat
follo
w-upversus
baselin
ein
theen
tireco
hortan
din
thetw
ogrou
ps.
*Significanc
eof
thetw
ogrou
psversus
controls.
†Sign
ificanc
eMR�
versus
MR+
atba
selin
e.‡Sign
ificanc
eMR�
versus
MR+
atfollo
w-up.
1196
van Garsse, et al.
between MR recurrence and other echocardio-graphic parameters.
Logistic regression analysis revealed thatbaseline values of DYS-PAP (OR: 5.4, 95% CI: 3.1–7.7, P < 0.001), AML tethering angle a (OR:5.0, 95% CI: 2.6–6.7, P < 0.001), and tetheringsymmetry (OR: 3.9, 95% CI: 2.5–5.7, P < 0.001)were predictors of recurrent MR (Table V). Themodel proved to be reliable (HL, P = 0.8) andaccurate (c = 0.7).
ROC analysis was performed to assess theutility of DYS-PAP to predict moderate or mod-erate-to-severe recurrent MR (Fig. 3). A DYS-PAP value �58 msec predicted recurrence ofMR with 100% sensitivity and 83% specificity(AUC 0.92, 95% CI: 0.7–1, P < 0.001). Ante-rior tethering angle a had 95% sensitivity and80% specificity with an optimal cutoff of�39.5° (AUC 0.86, 95% CI 0.72–0.95,P < 0.001); a/b with a cutoff of �0.75 had88% sensitivity and 79% specificity (AUC 0.82,95% CI 0.70–0.93, P < 0.001).
Discussion:General Considerations:Mitral valve repair has been preferred toreplacement in patients with chronic ischemicmitral regurgitation (CIMR) and LV dysfunctionbecause of its positive effect with respect topreservation of ventricular function andimproved survival.16,17 Nonetheless, concernhas been raised about the high incidence ofrecurrent MR following restrictive annuloplastyin CIMR patients,18 which has been reported torange between 33% and 70% at 5 years.3,19
Although predictors of reappearance or deterio-ration of MR have been related to preoperativemitral valve anatomical features, or to LVgeometry and dimensions,20,21 it seemsunquestionable that, also in the presence of in-traoperative successful repair, recurrence of MRstill represents a disappointing and not unusualcomplication in CIMR patients submitted tomitral valve annuloplasty.3,8,19 For this reason,much of current research in this area has beenfocused on identifying predictive factors ofrecurrent MR to select patients who can reallybenefit from annuloplasty. Functional MR andintraventricular dyssynchrony are common find-ings in patients with heart failure and they areassociated with a poor prognosis.4,5 The dys-synchronic contractions of LV segments are theleading causes of FMR in these patients and areindependently related to severity of FMR.22 Inour study, we investigated the impact of DYS-PAP on the recurrence of MR in patients withICM undergoing UMRA using 2D-STE, whichallows differentiation of myocardial segments
TABLE
IV
MitralLeafl
etTe
theringan
dLV
Remod
eling
Con
trols
All(n
=14
4)MR�
(n=46
)MR+
(n=98
)
Baselin
Follo
w-up
PBa
selin
eFo
llow-up
PBa
selin
eFo
llow-up
P
PMPM
-LS(%
)�1
9.4±5.9
�7.4
±2.9*
�9.1
±3.9*
0.3
�9.7
±3.5*
†�1
5.0±5.6*
‡<0.00
1�5
.2±2.5*
�3.2
±2.2*
0.06
ALP
M-LS(%
)�2
2.3±7.1
�8.2
±3.2*
�11.1±4.4*
0.1
�12.4±3.2*
†�1
8.9±6.7*
‡<0.00
1�4
.6±3.3*
�3.4
±2.2*
0.09
GLS
pea
k(%
)�2
0.7±7.0
�8.3
±3.3*
�11.2±4.0*
0.1
�10.1±4.3*
†�1
6.6±5.5*
‡0.00
1�6
.5±2.3*
�5.8
±2.4*
0.3
DYS
-PAP(m
s)4.8±2.1
45.8
±7.6*
53.2
±8.1*
0.05
25.9
±7.2*
†22
.3±7.3*
‡0.8
65.4
±8.2*
84.1
±8.8*
<0.00
1
ALP
M-LS=an
terolateralp
apillarymusclelong
itudina
lstrain;
PMPM
-LS=po
sterom
edialp
apillarymusclelong
itudina
lstrain;
GLS
pea
k=pe
akglob
allong
itudina
lstrain;
DYS
-PAP=pa
pillary
dyssyn
chrony
.P=Sign
ificanc
eat
follo
wup
versus
baselin
ein
theen
tireco
hortan
din
thetw
ogrou
ps.
*Significanc
eof
thetw
ogrou
psversus
controls.
†Sign
ificanc
eMR�
versus
MR+
atba
selin
e.‡Sign
ificanc
eMR�
versus
MR+
atfollo
w-up.
1197
Papillary Muscle Systolic Dyssynchrony in ICM
with active contractions from segments whichare passively tethered.7
Key Findings:Our study demonstrated that DYS-PAP was thestrongest independent predictor of recurrentMR. Indeed, we found that DYS-PAP �58 msecpredicted recurrence of MR with 100% sensitivityand 83% specificity. Anterior tethering anglea had 95% sensitivity and 80% specificity with anoptimal cutoff of �39.5° and tethering symme-try with a cutoff of �0.75 had 88% sensitivityand 79% specificity.
Penicka et al.5 have recently demonstratedthat high-degree LV dyssynchrony hinders LVpump function recovery in patients with ischemicheart failure undergoing myocardial revasculari-zation.
A
B
Figure 2. A. Papillary muscle dyssynchrony (DYS-PAP) at follow-up in a patient showing recurrent MR. B. Postoperative DYS-PAPin a patient without MR. AVC = aortic valve closure; TTP = time to peak; ALPM = anterolater papillary muscle; PMPM = postero-medial papillary muscle.
Mitral ring annuloplasty in patients with ischemic dilatedcardiomyopathy. OR = odds ratio; CI = confidence interval.
Figure 3. Receiver operating characteristic (ROC) curvedemonstrating the predictive value of papillary muscle dys-synchrony for recurrence of mitral regurgitation after under-sized mitral ring annuloplasty in patients with ischemic dilatedcardiomyopathy. The cutoff value is marked in the plot bya dot.
1198
van Garsse, et al.
Myocardial ischemia is one of the majorcauses of LV dyssynchrony23 and one mayspeculate that correction of ischemia by CABGsurgery should resynchronize LV contractions.
In this study, patients with severe preopera-tive dyssynchrony did not show a reduction inDYS-PAP by CABG surgery, thus confirming thatextensive LV dyssynchrony cannot be reversed byrevascularization and that persistent high-degreedyssynchrony is associated with MR recurrence.Furthermore, in patients with significant MR atfollow-up, DYS-PAP significantly worsened. Inthese patients, volume overload resulting fromMR leads to further ventricular dilatation, LV dys-function, and DYS-PAP which subsequently leadsto further regurgitation. The persistent mechani-cal dyssynchrony between LV segments support-ing the papillary muscles also producesuncoordinated regional LV mechanical activationin these segments and decreases LV contractionefficiency and closing forces, thereby impairingmitral valve tenting and resulting in further geo-metric changes in mitral leaflet attachments anda worsening of mitral leaflet tethering.
We might postulate that DYS-PAP perpetuatesthe vicious circle whereby ventricular dilatation,LV dysfunction, and changes in LV geometry pro-mote MR and in turn MR promotes dilatation,dysfunction, and geometrical remodeling. Incontrast, in patients with low baseline DYS-PAP,the preserved synchronic contractions of papil-lary muscles might turn the vicious cycle into thevirtuous cycle of normalization of LV functionand reverse remodeling. Thus, preoperative DYS-PAP might represent a landmark of advancedmyocardial damage in ICM and a sign of irrevers-ible progression of the disease. Studies on thistopic are ongoing.
Clinical Implications:The findings of our study advocate routineassessment for LV dyssynchrony, in addition toother echocardiographic parameters, in patientswith ICM undergoing CABG and UMRA. In thesepatients the presence of severe baseline DYS-PAP�58 msec was associated with high MR recur-rence. Therefore, noninvasive testing to assesspapillary dyssynchrony should be performedbefore surgery to guide patient selection. Patientswith severe dyssynchrony might be consideredfor concomitant or alternative surgical tech-niques.24,25
Limitations:Our study findings should be viewed in light ofsome inherent limitations.
First, the retrospective nature of this studydoes not allow us to draw definite conclusions.Second, viability testing was not performed in
these patients. Therefore, the absence of viablemyocardium might be responsible for irreversibledyssynchrony in patients showing unfavorableresults. This issue deserves further investigation.Third, postoperative evaluation of the coronarystatus was not assessed. It would have been help-ful to differentiate between surgical failure (valverepair and CABG) and the progress of the coro-nary disease. Fourth, the PISA method of assess-ing the severity of ischemic MR is less accurate asthe PISA is less likely to be a hemisphere. Adjust-ments of the aliasing velocity were carried outsuch that a well-defined hemisphere wasobtained. This was done by shifting the baselinetoward the direction of the flow, or by loweringthe Nyquist limit, or both (the latter reduces thewall filter whereas the former does not).9 Further-more, as the PISA method overestimates MR inpatients with eccentric jets, we therefore alsoemployed pulsed Doppler quantitative flowmethods and, in case of contradictory results,PISA was chosen in the presence of central jet orcalcific mitral valve/mitral annulus, whereaspulsed Doppler quantitative flow methods werepreferred when the jet was eccentric or multiple.Finally, as the myocardial movement is basicallythree-dimensional (3D), thus including rotationalmovements, 2D speckle tracking may have inher-ent limitations, especially in patients with dilatedLV.
Further larger prospective studies with a stan-dardized assessment of severity of MR are neces-sary to confirm our findings.
Conclusions:DYS-PAP is strongly correlated with the recur-rence of MR. A DYS-PAP cutoff value of 58 msecis useful to identify patients in whom UMRA islikely to fail. That way decision making in ische-mic functional MR might be facilitated.
Acknowledgment: We gratefully acknowledge Dr. JudithWilson for the English revision of the article.
References1. Di Mauro M, Di Giammarco G, Vitolla G, et al: Impact of
no-to-moderate mitral regurgitation on late results afterisolated coronary artery bypass grafting in patients withischemic cardiomyopathy. Ann Thorac Surg 2006;81:2128–2134.
2. Bax JJ, Braun J, Somer ST, et al: Restrictive annuloplastyand coronary revascularization in ischemic mitral regurgi-tation results in reverse left ventricular remodeling. Circu-lation 2004;110(suppl 1):II103–II108.
3. Hung J, Papakostas L, Tahta SA, et al: Mechanism ofrecurrent ischemic mitral regurgitation after annuloplas-ty. Continued LV remodelling as a moving target. Circula-tion 2004;110(suppl 1):II85–II90.
4. Cho GY, Song JK, Park WJ, et al: Mechanical dyssynchro-ny assessed by tissue Doppler imaging is a powerful
1199
Papillary Muscle Systolic Dyssynchrony in ICM
predictor of morality in congestive heart failure withnormal QRS duration. J Am Coll Cardiol 2005;46:2237–2243.
5. Penicka M, Bartunek J, Lang O, et al: Severe left ventricu-lar dyssynchrony is associated with poor prognosis inpatients with moderate systolic heart failure undergoingcoronary artery bypass grafting. J Am Coll Cardiol2007;50:1315–1323.
6. Turan B, Yilmaz F, Karaahmet T, et al: Role of left ventric-ular dyssynchrony in predicting remodeling after ST ele-vation myocardial infarction. Echocardiography 2012;29:165–172.
7. Nesser HJ, Winter S: Speckle tracking in the evaluation ofleft ventricular dyssynchrony. Echocardiography 2009;26:324–336.
8. Gelsomino S, Lorusso R, Caciolli S, et al: Insights on leftventricular and valvular mechanisms of recurrent ische-mic mitral regurgitation after restrictive annuloplasty andcoronary artery bypass grafting. J Thorac Cardiovasc Surg2008;136:507–518.
9. Zoghbi WA, Enriquez-Sarano M, Foster E, et al: Recom-mendations for evaluation of the severity of native valvu-lar regurgitation with two dimensional and Dopplerechocardiography. J Am Soc Echocardiogr 2003;16:777–802.
10. Gelsomino S, van Garsse L, Lucà F, et al: Impact of preop-erative anterior leaflet tethering on the recurrence ofischemic mitral regurgitation and the lack of left ventricu-lar reverse remodeling after restrictive annuloplasty. J AmSoc Echocardiogr 2011;24:1365–1375.
11. Schiller NB, Shah PM, Crawford M, et al: Recommenda-tions for quantitation of the left ventricle by twodimensional echocardiography. American Society ofEchocardiography Committee on Standards, Subcom-mittee on Quantitation of Two-Dimensional Echocardio-grams. J Am Soc Echocardiography 1989;2:358–367.
12. Kono T, Sabbah HN, Rosman H, et al: Left ventricularshape is the primary determinant of functional mitralregurgitation in heart failure. J Am Coll Cardiol1992;20:1594–1598.
13. Zhang H, Otsuji Y, Uemura T, et al: Different mechanismsof ischemic mitral regurgitation in patients with inferiorand anterior myocardial infarction. J Echocardiogr2008;3:74–83.
14. Cerqueira MD, Weissman NJ, Dilsizian V, et al; AmericanHeart Association Writing Group on Myocardial Segmen-tation Registration for Cardiac Imaging: Standardizedmyocardial segmentation and nomenclature for tomo-
graphic imaging of the heart: A statement for healthcareprofessionals from the Cardiac Imaging Committee ofthe Council on Clinical Cardiology of the American HeartAssociation. Circulation 2002;105:539–542.
15. Tigen K, Karaahmet T, Dundar C, et al: The importanceof papillary muscle dyssynchrony in predicting the sever-ity of functional mitral regurgitation in patients with non-ischaemic dilated cardiomyopathy: A two-dimensionalspeckle-tracking echocardiographic study. Eur J Echocardi-ogr 2010;11:671–676.
16. LaPar DJ, Kron IL: Should all ischemic mitral regurgitationbe repaired? When should we replace? Curr Opin Cardiol2011;26:113–117.
17. Bouma W, van der Horst ICC, Wijdh-den Hamer IJ, et al:Chronic ischemic-based mitral regurgitation: Currenttreatment results and new mechanism-based surgicalapproaches. Eur J Cardiothorac Surg 2010;37:170–185.
18. Shiota M, Gillinov AM, Takasaki K, et al: Recurrent mitralregurgitation late after annuloplasty for ischemic mitralregurgitation. Echocardiography 2011;28:161–166.
19. ten Brinke EA, Klautz RJ, Tulner SA, et al: Clinical andfunctional effects of restrictive mitral annuloplasty at mid-term follow up in heart failure patients. Ann Thorac Surg2010;90:1913–1921.
20. Calafiore AM, Di Mauro M, Gallina S, et al: Mitral valvesurgery for chronic ischemica mitral regurgitation. AnnThorac Surg 2004;77:1989–1997.
21. De Bonis M, Lapenna E, Verzini A, et al: Recurrence ofmitral regurgitation parallels the absence of left ventricu-lar remodeing after mitral repair in advanced dilated car-diomyopathy. Ann Thorac Surg 2008;85:932–939.
22. Hung CL, Tien SL, Lo CI, et al: The incremental value ofregional dyssynchrony in determining functional mitralregurgitation beyond left ventricular geometry after nar-row QRS anterior myocardial infarction: A real timethree-dimensional echocardiography study. Echocardiog-raphy 2011;28:665–675.
23. Forrester JS, Wyatt HL, Da Luz PL, et al: Functional signifi-cance of regional ischemic contraction abnormalities. Cir-culation 1976;54:64–70.
24. Borger MA, Alam A, Murphy PM, et al: Chronic ischemicmitral regurgitation: Repair, replace or rethink? Ann Tho-rac Surg 2006;81:1153–1161.
25. Gelsomino S, Lorusso R, Capecchi I, et al: Left ventricularreverse remodeling after undersized mitral ring annulopl-asty in patients with ischemic regurgitation. Ann ThoracSurg 2008;85:1319–1331.
