International Journal of Cardiology xxx (2013) xxx–xxx
IJCA-17255; No of Pages 9
Contents lists available at ScienceDirect
International Journal of Cardiology
j ourna l homepage: www.e lsev ie r .com/ locate / i j ca rd
The shape and function of the left ventricle in Ebstein's anomaly☆
Patrick J. Goleski a,1, Florence H. Sheehan a,⁎,1,2, Sylvia S.M. Chen b, Philip J. Kilner b, Michael A. Gatzoulis b
a University of Washington, Seattle, WA, United Statesb Royal Brompton Hospital, London, United Kingdom
☆ Supported in part by grants from the NationalCardiovascular Biomedical Research Unit at the RoyalFoundation Trust and Imperial College (London, UnitedVentriPoint, Inc. (Seattle, WA).⁎ Corresponding author at: University of Washington, B
6422, United States. Tel.: +1 206 543 4535; fax: +1 206E-mail address: [email protected] (F.H. Sheehan).
1 These authors take responsibility for all aspects of the rof the data presented and their discussed interpretation.
2 Founder and Chief Scientist of VentriPoint, Inc. Eqconsultant.
Background: Left ventricular (LV) failure is common in Ebstein's anomaly, though remains poorly understood.Weinvestigated whether shape deformity impacts LV function.Methods: Three-dimensional models of the right ventricle (RV) and LV from 29 adult Ebstein's patients and ninenormal subjects were generated from cardiac magnetic resonance image tracings. LV end diastolic (ED) shape,systolic function, septal motion and ventricular interaction were analyzed.Results: LV ED volume indexwas normal in Ebstein's (75 ± 19 vs. 78 ± 11 ml/m2 in normals, p = 0.50) but theLV was basally narrowed and modestly dilated apically. LV function was reduced globally (ejection fraction (EF)41 ± 7 vs. 57 ± 5% in normals, p b 0.0001) and regionally (decreased mean segment displacement at end sys-
tole (ES) in 12/16 segments, basal Z-scores −2.1 to −1.0). Septal dyskinesis was suggested by outward meansegment displacement in at least one basal septal segment in 25 patients (86%) but refuted by septal thickeningin 14 (48%), normal septal curvature at ED and ES, and by visually evident basal LV anterior translation in 27 pa-tients (93%). LV EF correlated better with normalized tricuspid annular plane systolic excursion (r = 0.70) thanwith RV EF (r = 0.42) or RVEDVI (r = 0.18).Conclusions: Although the Ebstein's LV has preserved volume, it exhibits basal narrowing, modest apical dilationand global hypokinesis. The apparent basal septal dyskinesis observed in most patients is likely attributable toanterior cardiac translation rather than true paradoxical motion. LV EF is unaffected by RV volume, correlatingwell instead with RV longitudinal shortening.
While Ebstein's anomaly is classically considered a malformation ofthe right heart, the left ventricle (LV) is also affected [1]. As the naturalhistory with Ebstein's anomaly is most commonly heart failure that isoften left sided, an understanding of failure modes is important [1,2].A few attempts have been made to describe LV abnormalities butwere limited by the use of two-dimensional (2D) modalities [2–10].Techniques based on three-dimensional (3D) reconstructions of cardiacmagnetic resonance imaging (CMR) and 3D echocardiography (echo)permitmore comprehensive descriptions of global and regional ventric-ular shape and function but have not been employed within this popu-lation. Therefore, the present study was performed to assess LV global
Institutes of Health ResearchBrompton and Harefield NHSKingdom), and by a grant from
ox 356422, Seattle, WA 98195-685 4847.
eliability and freedom frombias
uity holder and compensated
vier Ireland Ltd.
shape and function of the lef
and regional shape and function with particular emphasis on septal be-havior, in comparison to healthy controls.
2. Materials and methods
2.1. Patient populations
The study population comprises 29 patients with known Ebstein's anomaly and ninehealthy subjectswhose datawere collected previously for investigation of the right ventri-cle (RV) in Ebstein's anomaly [11]. The authors of this manuscript have certified that theycomplywith the Principles of Ethical Publishing in the International Journal of Cardiology.
2.2. Imaging
CMR images were acquired in 8–12 consecutive short axis slices covering both ventri-cles in addition to long axis and oblique views targeted to better visualize the RV outflowtract. The images were analyzed at RV end diastole (ED) and end systole (ES) because theCMR data were acquired initially for a study on the RV [11]. Although the timing of LV andRV ED and ES may differ by a frame, LV volumes do not differ statistically over such an in-terval [12]. The ventricleswere reconstructedusing thepiecewise smooth subdivision sur-face (PSSS)method, which provides anatomically accurate representation of the 3D shapeof a heart ventricle [13,14]. All reconstructionswere generated from amesh of 1328 faces,except centeraxis and centersurface analyses, where 5312 faces were used, and in curva-ture analysis, where 21,248 faces were used. The mean signed projection distance be-tween the reconstructed surfaces and the traced borders averaged 0.001 ± 1.6 mm forthe LV and 0.1 ± 1.9 mm for the RV, indicating excellent goodness-of-fit.
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
2 P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
2.3. LV shape analysis
LV shape analysis was performed using both global and regional descriptors of the LVat ED. Sphericity was calculated as the ratio of the surface area of the LV to the surface areaof a sphere with the same volume. LV long axis length was measured as the distance be-tween the mitral valve centroid and the LV apex, defined as the point on the LV endocar-dium furthest from the mitral valve centroid.
One dimensional (1D) and 2D analyses of regional shapewere performed on 20 shortaxis slices generated between themost apical point on themitral valve annulus and the LVapex (Fig. 1A). Septal–lateral wall distancewas measured along a line defined by the slicecentroid and the midpoint of the anterior and posterior septal boundary points, whichwere drawn at the respective insertions of the RV onto the septum (Fig. 1B). The anteri-or–posterior distance was measured perpendicular to the septal–lateral wall distancethrough the slice centroid. For all short axis analyses, the most apical slice was excludeddue to its extremely small cross section. The following regional metrics reported by thisapproach: LV ED slice area, slice eccentricity, septal–lateral wall distance and anterior–posterior distance. LV slice area, long axis length and short axis distanceswere normalizedby the cube root of left ventricular end diastolic volume (LVEDV). Slice eccentricity wascalculated as 4πA
P2 , where A is the slice area and P is the slice perimeter [15]: a circularcross section yields a value of 1 while obliteration of the lumen would occur at a valueof 0 [15].
The degree of LV long axis bowing at EDwasdefined as the reciprocal of the radius of acircle fitted through the centroids of each slice using themethod of least squares (Fig. 1C)normalized by cube root of LVEDV.
AoMV
A
B
C
Fig. 1.The left ventricle of a patientwith Ebstein's anomaly subdivided into slices along thelong axiswith short axis distances shown (A). Septal boundary points (white dots) used inmeasuring septal–lateral wall distance (B). Circle fitted to short axis slice centroids whoseradiuswas used tomeasure left ventricular long axis bowing at ED (C). Ao = aortic valve,MV = mitral valve.
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
LV regional shape at EDwas also analyzed in 3D using the centeraxis method in termsof distance from each vertex on the LV surface to the LV long axis (Fig. 2A) [16]. Centeraxisdistanceswere normalized by LV length, which is normal or onlymildly elongated even inadvanced LV failure [17], aggregated according to the 16 segmentmodel of the LV (Fig. 2B,C), and expressed relative to the normal group's values as Z-scores.
2.4. LV function analysis
LV function was analyzed both globally and regionally by making comparisons of thereconstructions at ED and ES. Ejection fraction (EF) was calculated directly from the EDand ES volumes. Percent LV length change was calculated from ED and ES values as well.Regional LV function was analyzed using 1D and 2D metrics in each short axis slice interms of fractional area change and fractional shortening along both short axes. Regionalwallmotionwas analyzed in 3Dby the centersurfacemethod along vectors drawn orthog-onal to a surface constructed midway between the ED and ES surface reconstructions(Fig. 3) [18–20]. Motion was normalized by the cube root of LVEDV, averaged in each ofthe 16 segments, and expressed as Z-scores.
2.5. Septal analyses and translation
As cardiac translation can cause the appearance of septal dyskinesis, three tools wereapplied to discernbetween trueparadoxicalmotion of the septumand LV translation:wallthickening, septal curvature, and visual inspection.
Wall thickening analysis is a method of assessing regional myocardial function[18,21–23] that is independent of cardiac translation. In the presence of true dyskinesis,
Fig. 2. Centeraxis distances from left ventricular long axis to endocardial surface in anEbstein's patient (A). Centeraxis distances aggregated in each segment are displayed ona reconstruction of the left ventricle (B) and in a polar map (C). Vertices basal to thevalve orifices are excluded (blackened region) (B).
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
Fig. 3.Centersurface technique to assess regionalwallmotion between end diastole and end systole (A). Display ofmotion values after aggregation according to the 16 segmentmodel on areconstruction of the left ventricle (B) and in a polar map (C) for an Ebstein's patient. Vertices basal to the valve orifices are excluded (blackened region) (B).
3P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
negative septal myocardial wall thickening (i.e. thinning) would be expected. Septalmyo-cardial wall thickeningwas determined by tracing the endocardial and epicardial contoursin one basal slice where paradoxical septal motion was frequently observed, andmeasur-ing thickening using the centerline method [24,25] at the chord on the septal border
Fig. 4.Maximumprincipal curvature is shown colormapped to the left ventricular surfacefrom a septal wall view. Only points within the septal boundary curve (shown in black)were included for septal curvature analysis.
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
nearest the midpoint of the anterior and posterior septal boundary points. Percent thick-ening was calculated by dividing change in wall thickness between ED and ES by themean ED thickness of the chord and its 10 closest neighbors [24].
Curvature also provides information regarding regional ventricular function indepen-dent of cardiac translation [26]. Septal curvaturewas determined as follows [27]: a best-fitsmooth curve was generated from the traced septal boundary points, projected onto theLV endocardial surface and used to divide the LV into septal and free wall surfaces(Fig. 4). For each vertex, principal curvaturewas determinedbyfitting a least-squares qua-dratic patch of the form f(x,y) = ax2 + by2 + cxy + dx + ey + f to the neighborhoodof surrounding vertices. The principal curvatures were calculated from the eigenvectorsand eigenvalues of the hessian and assigned a positive value if the normal vector pointedoutward (LV is convex) and a negative value if pointing toward the LV interior (concave).For our analysis, we assumed the septal surface was either concave or convex, i.e., thatboth principal curvatures would have the same sign at a given vertex. The signed principalcurvatures with the greater magnitude at each vertex were then averaged for each trian-gle, yielding themaximumprincipal curvature for that triangle, termed k1. Themean of k1for all triangles within the septal surface was computed as the mean maximum principal
Table 1Left ventricular end diastolic global shape parameters.
Shape parameter Ebstein's anomaly (n = 29) Normal (n = 9) p
4 P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
curvature, k1―
. From this analysis, the septum was deemed convex if k1―
was positive andconcave if k1
―was negative, as viewed from outside the LV. The change in curvature be-
tween ED and ES, Δk1―
was also calculated.
Fig. 5. Left ventricular enddiastolic short axis slice area (A), septal–lateral distance (B), anterior–was not included due to its small size. *p b 0.05 Ebstein's versus normal.
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
For visual examination of LV translation the 3D reconstructions of the LV in ED and ESwere superimposed using the CMR system as the external reference, and translation wasidentified from protrusion of the ES surface outside the ED surface.
posterior distance (C) and eccentricity (D) by slice along the long axis. Themost apical slice
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
Fig. 6.Mean left ventricular shape by centeraxis analysis in patients with Ebstein's anom-aly (A) and in normal subjects (B). Mean abnormality in shape in Ebstein's anomaly plot-ted as Z-scores (C). *p b 0.05 Ebstein's versus normal.
Table 2Left ventricular global function parameters.
Function parameter Ebstein's anomaly (n = 29) Normal (n = 9) p value
LV EF = left ventricular ejection fraction.a Assessed between end diastole and end systole.
5P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
2.6. Ventricular interaction
LV EF was correlated with parameters of RV size, shape and function obtained fromour prior analysis [11]: RV ED volume index (RVEDVI), RV length normalized by body sur-face area (BSA), percent RV length change, RV EF, tricuspid annular plane systolic excur-sion normalized by RV long axis length (normalized TAPSE), and RV basal bulgenormalized by BSA.
3. Statistical methods
Two-tailed t tests were used to compare continuous variables be-tween groups. Correlation between parameters was assessed via linearregression, reported as the Pearson product–moment correlation coeffi-cient, r. Statistical significance was defined as p b 0.05.
4. Results
4.1. Patient population
The normal subjects were slightly older than the Ebstein's patients(46 ± 14 vs. 36 ± 14 years, respectively, p = NS) and more often fe-male (78 vs. 41%, respectively, p b 0.001). BSAwas similar in the normaland Ebstein's groups (1.7 ± 0.2 vs. 1.8 ± 0.2 m2, respectively,p = NS). Although eight of the 29 Ebstein's patients had undergonesurgical repairs, including closure of atrial septal defect, tricuspid or pul-monary valve repair, and/or bidirectional Glenn, therewas no statisticaldifference in 233 of 241 parameters between repaired and unrepairedpatients; therefore, all Ebstein's patients were combined. None had sig-nificant mitral prolapse or regurgitation.
4.2. LV volume
Left ventricular ED volume index (LVEDVI) was similar betweenEbstein's and normal groups (75 ± 19 vs. 78 ± 11 ml/m2, respectively,p = 0.50).
4.3. LV shape
The LV was slightly less spherical in Ebstein's compared to normals,and was more elongated (Table 1). LV long axis bowing was moreprominent in Ebstein's compared to normals. The Ebstein's group hadsmaller slice areas and septal–lateral short axis distances at the baseand midventricle (Fig. 5A, B). Anterior–posterior short axis distanceswere greater near the apex and smaller just basal to the midventricle(Fig. 5C). These data are consistent with the lower observed eccentrici-ties (Fig. 5D) and agreewith the centeraxis analysis (Fig. 6),which dem-onstrated significant narrowing in five of six basal segments (basal Z-scores −1.8 to −0.8) and modest dilation in all four apical segments(apical Z-scores 0.8 to 1.0) (Fig. 6C).
4.4. LV function
LV EF and percent LV length change were reduced in Ebstein's pa-tients compared to normal (Table 2). Fractional change in area andshortening of both short axeswere uniformly depressed in Ebstein's rel-ative to normals (Fig. 7). Centersurface analysis (Fig. 8) corroboratedthat the Ebstein's group was hypokinetic in all 16 segments, reachingstatistical significance in 12 segments, including all six basal segments(basal Z-scores−2.1 to−1.0) and four of six midventricular segments(midventricular Z-scores −1.4 to−0.8) (Fig. 8C).
4.5. Septal analysis and translation
Motion in thebasal anteroseptal and inferoseptal segmentswas neg-ative (outward) between ED and ES, suggesting dyskinesis (mean seg-ment displacement of −0.03 ± 0.06 and −0.004 ± 0.08 mm/mm,respectively) (Fig. 8A). Wall motion was outward in 25 Ebstein's
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
patients in at least one basal septal segment, but 11 of these 25 (44%)had positive basal slice thickening; a negative value would be expectedwith true dyskinesis. In addition, there was no difference in septal cur-vature between Ebstein's and normals at ED ( k1
―0.05 ± 0.01 vs.
0.05 ± 0.01 mm−1, respectively, p = 0.67) or at ES (k1―
0.07 ± 0.02vs. 0.08 ± 0.02 mm−1, respectively, p = 0.31); change in septal curva-ture was also similar in Ebstein's and in normals (Δk1
―0.02 ± 0.02 vs.
0.03 ± 0.02 mm−1, respectively, p = 0.15).Finally, inspection of the LV 3D reconstructions superimposed from
ED and ES revealed visually evident anterior translation involving theentire LV in 12 patients and limited to the base in 15 patients. None ofthe normals exhibited the gross LV translation seen in Ebstein's,
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
Fig. 7. Left ventricular end diastolic short axis slice fractional area change (A), septal–lateral wall fractional shortening (B) and anterior–posterior fractional shortening (C) are plottedagainst left ventricular long axis slice number. All analyses are between end diastole and end systole. Themost apical slice was excluded due to its small size. *p b 0.05 for Ebstein's versusnormal.
6 P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
although a small amount of basal translation was identified in eightstudies (Fig. 9).
4.6. Correlation of LV shape and function
All parameters of LV shape correlated poorly with LV EF (|r| ≤ 0.4)(Table 3).
4.7. Ventricular interaction
Parameters of RV size, shape and function obtained from our prioranalysis of this patient population were compared to LV EF (Table 4)[11]. Normalized TAPSE, normalized RV basal bulge and percent RVlength change, all metrics of RV longitudinal contraction, yielded stron-ger correlation with LV EF (r = 0.70, 0.57, 0.54, respectively) than didRV EF (r = 0.42).
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
5. Discussion
We performed an analysis of LV shape at ED and of LV systolic func-tion in Ebstein's anomaly. Our patients displayed irregularly shaped LVs,which appear related to right-sided dilation. LV function was globallydepressed with apparent paradoxical septal motion at the base. Uponinvestigation, however, this septal behavior appeared to be an artifactof anterior LV translation in the majority of patients. LV RV interactionwas evidencedby the influence of RVdysfunction on LV function, partic-ularly parameters of RV longitudinal contraction.
The present study is the first attempt to quantitatively characterizeshape and function of the LV in Ebstein's using 3D analysis to our knowl-edge. Previous attempts have relied upon global descriptors of LV shapeand function obtained from cineangiography [3,6,10], radionuclide ven-triculography [7] and 2D echo [2,4,5,8,9]. Several investigators found ir-regular patterns of contour and contractility, and/or decreased EF, but
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
Fig. 8. Mean wall motion from end diastole to end systole in Ebstein's anomaly (A) andnormal subjects (B). Mean motion abnormality in Ebstein's anomaly plotted in terms ofZ-scores (C). Negative values indicating outwardmotion are seen in the basal anteroseptaland inferoseptal segments in the Ebstein's patients (A). *p b 0.05 for Ebstein's versusnormal.
Fig. 9. Superimposed reconstructions of the left ventricle at end diastole and end systole,with the right ventricle shown at end diastole for reference, in Ebstein's patients with an-terior translation of the entire left ventricle (A) or translation limited to the base (B), andin a normal volunteer, also with translation limited to the base (C).
7P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
were limited to global and qualitative descriptions by the modalitiesemployed [3,5–10].
5.1. LV volume
The normality of LV volume in Ebstein's patients agrees generallywith findings by other investigators [6,8,10]. Though some investigatorshave suggested LV ED volume should be decreased in the RV volumeoverload state due to encroachment of the RV [28,29], we did not ob-serve an absolute decrease in LV size despite dilation of the RV tomore than twice the volume in normals [11]. This finding also arguesagainst the hypothesis that LV function is reduced secondary to reduceddiastolic filling, as has been proposed elsewhere [6,8].
5.2. LV shape
Our observation of increased LV length agrees with reports of LVelongation via cineangiography by Monibi et al. and Sharma et al.[3,10] as well as Daliento et al.'s report of an increased septal length
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
on autopsy [6]. Benson et al. reported abnormal short axis eccentricityof the LV, much as we did [5]. We found decreased sphericity, crosssectional area and septal–lateral wall distance along the LV in agree-ment with the contour irregularities Monibi et al. described fromcineangiography and the irregularly distorted lateral wall noted fromautopsy [3]. Likewise, Saxena et al. observed via 2D echo that the LVwas abnormally shaped as a result of the large atrialized RVwithflatten-ing apical to the tricuspid valve [8], and Sharma et al. found contour ir-regularities in several patients via cineangiography including extrinsicimpression, inferior notching and an irregular inferior wall that agreewith ourfindings [10]. Daliento et al. noted in particular that the septum
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
LV = left ventricle. EF = ejection fraction. Basal LV = most basal short axis slice.Mid LV = short axis slice equidistant frombase and apex. Apical LV = most apicalshort axis slice.
a Normalized by cube root of LVEDV.
8 P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
was displaced leftward most frequently at the base, and occasionallymore apically, as was supported by our short axis and centeraxis analy-ses. While our studywas not designed to address the etiology of the ob-served deformities, we agree with prior investigators who theorized aneffect from encroachment of the atrialized RV [5,6,8,9], as shape irregu-larities were greatest at the base.We also foundmore LV long axis bow-ing in Ebstein's, consistent with the “banana” shape described byDaliento et al. and the “crescent” shape reported by Shiina et al. [6,9].Additionally, encroachment by the enlarged RV may contribute to LVshape abnormalities, as has been described in other RV volumeoverloadstates [28,29].
5.3. LV function and ventricular interaction
LV function was depressed in Ebstein's by EF, long axis, short axisand centersurface analyses throughout the LV, with slightly greater dys-function at the base. Thesefindings are consistentwith those of prior in-vestigators [3–8,30]. Previously, others have attributed LV dysfunctionto noncompaction, myocardial fibrosis, disruption of free wall myofibrilcontinuity, fiber geometry or mixed cardiac defects [4–9,30–34]. In ouranalysis, LV EF correlated poorly with all parameters of LV shape, sug-gesting discordant mechanisms for shape and function abnormalities.LV EF also correlated poorly with RVEDVI, suggesting that RV encroach-ment is insufficient to adversely affect LV function. We found relativelypoor correlation between LV EF and RV EF, in agreement with otheranalyses of Ebstein's anomaly [6,7,11]. Instead, we found better correla-tion between LV EF and basal bulge/TAPSE, parameters of RV long axisshape/function, respectively. Here, we speculate that RV longitudinalfunction augments LV function preferentially, as RV circumferentialfiber continuity across the septum is disrupted by the abnormally posi-tioned tricuspid valve.
Table 4Correlation between left ventricular ejection fraction and rightventricular descriptors.
RV = right ventricle. LV EF = left ventricular ejection fraction.RVEDVI = RV end diastolic volume index. TAPSE = tricuspid annularplane systolic excursion.
a Assessed between end diastole and end systole.b Normalized by RV long axis length.c Normalized by body surface area.
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
5.4. Septal analysis and translation
Several investigators have also reported septal dyskinesis as a com-ponent of LV dysfunction in Ebstein's [3,5,6,8,9]. Saxena et al. describedbiphasic motion of the septum that was paradoxical at the level of theatrialized RV using 2D echo [8]. Daliento et al. noted diastolic septal dis-placement in most cases via cineangiography [6]. Benson et al. notedparadoxical motion via 2D echo at the level of the atrialized RV [5]. In-deed, in animal models, acute RV volume overload has been shown toinduce apparent septal dyskinesis that is actually caused by diastolicleftward displacement of the septum [35,36].
In the present study, we investigated the presence of septaldyskinesis through a variety of methods. We found negative (outward)mean segment displacement frequently present in at least one septalsegment at ES. While this finding is suggestive of septal dyskinesis,these effects were small and inconsistently present. It is also possiblethat thenegativemean segment displacementswe observedwere an ar-tifact of anterior septal translation. Anterior translation of the LV duringsystole has been described in other populations with RV volume over-load, and has been implicated as contributing to “pseudoparadoxicalseptal motion” on M mode echo [37,38]. The frequent finding of wallthickening in the basal septum supports translation as the explanationfor negative displacement; whereas myocardial thinning would havebeen expected with true paradoxical septal motion. The fact that septalcurvature behaved normally in Ebstein's also argues against severe dys-function in this region. Lastly, anterior translation involving the entire LVwas grossly evident on visual inspection of the superimposed ED and ESLV reconstructions in several patients, and translationwas seen in nearlyall patients at least at the base.Whilemanynormals also displayed someanterior translation restricted to the base, this was generally of a lessermagnitude than in Ebstein's. Taken together, our findings suggest thatseptal dyskinesis at ES is not a prominent and consistent feature inEbstein's. Instead the observed paradoxical septal motion is more likelyto be an artifact of cardiac translation in most of our patients.
6. Study limitations
The present study utilized data from a small set of healthy subjects.No effort was made to control for concurrent cardiac abnormalities be-yond the presence of mitral regurgitation. We only performed analysisat ED and ES because LV function is routinely assessed from thesetime points alone, and we wished to assist in the interpretation of per-ceived abnormalities in systolic function. However, we potentiallymissed important functional behavior during other times within thecardiac cycle. Despite these limitations, we were able to observe signif-icant anomalies in the behavior of the LV, consistent with known phys-iology in Ebstein's anomaly.
7. Conclusions
Shape analysis of the Ebstein's LV demonstrated basal narrowing,mild apical dilation, mild elongation and increased long axis bowing.Functional analysis revealed generally global hypokinesis that appearsunrelated to the LV shape irregularities or to the extent of RV dilation.Instead, LV EF correlated best with normalized TAPSE. The apparentbasal septal dyskinesis observed inmost patients is likely due to cardiactranslation.
Disclosures
Dr. Sheehan is a founder and the Chief Scientist of VentriPoint, Inc.She holds equity in the company and provides consultative serviceswith compensation.
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
9P.J. Goleski et al. / International Journal of Cardiology xxx (2013) xxx–xxx
Acknowledgments
We wish to acknowledge the committed devotion of our program-mer, Edward Bolson, who came out of retirement to write “one moreprogram” for the curvature analysis.
References
[1] Attenhofer Jost CH, Connolly HM, Edwards WD, Hayes D, Warnes CA, Danielson GK.Ebstein's anomaly — review of a multifaceted congenital cardiac condition. SwissMed Wkly 2005;135:269–81.
[2] Celermajer DS, Bull C, Till JA, et al. Ebstein's anomaly: presentation and outcomefrom fetus to adult. J Am Coll Cardiol 1994;23:170–6.
[3] Monibi AA, Neches WH, Lenox CC, Park SC, Mathews RA, Zuberbuhler JR. Leftventricular anomalies associated with Ebstein's malformation of the tricuspidvalve. Circulation 1978;57:303–6.
[4] Oechslin E, Buchholz S, Jenni R. Ebstein's anomaly in adults: Doppler-echocardiographic evaluation. Thorac Cardiovasc Surg 2000;48:209–13.
[5] Benson LN, Child JS, Schwaiger M, Perloff JK, Schelbert HR. Left ventricular geometryand function in adults with Ebstein's anomaly of the tricuspid valve. Circulation1987;75:353–9.
[6] Daliento L, Angelini A, Ho SY, et al. Angiographic andmorphologic features of the leftventricle in Ebstein's malformation. Am J Cardiol 1997;80:1051–9.
[7] Hurwitz RA. Left ventricular function in infants and children with symptomaticEbstein's anomaly. Am J Cardiol 1994;73:716–8.
[8] Saxena A, Fong LV, TristamM, Ackery DM, Keeton BR. Late noninvasive evaluation ofcardiac performance inmildly symptomatic older patients with Ebstein's anomaly oftricuspid valve: role of radionuclide imaging. J Am Coll Cardiol 1991;17:182–6.
[9] Shiina A, Seward JB, Edwards WD, Hagler DJ, Tajik AJ. Two-dimensional echocardio-graphic spectrum of Ebstein's anomaly: detailed anatomic assessment. J Am CollCardiol 1984;3:356–70.
[10] Sharma S, Rajani M, Mukhopadhyay S, Aggarwal S, Shrivastava S, Tandon R. Angio-graphic abnormalities of the morphologically left ventricle in the presence ofEbstein's malformation. Int J Cardiol 1989;22:109–13.
[11] Lee CM, Sheehan FH, Bouzas B, Chen SSM, Gatzoulis MA, Kilner PJ. The shape andfunction of the right ventricle in Ebstein's anomaly. Int J Cardiol 2012:1–7.
[12] Guo A, Sheehan FH. Accuracy of quantitative ventriculographic analysis at reducedframe rates. Catheter Cardiovasc Interv 1999;48:18–21 [discussion 2–3].
[13] Hubka M, Bolson EL, McDonald JA, Martin RW, Munt B, Sheehan FH. Three-dimensional echocardiographic measurement of left and right ventricular massand volume: in vitro validation. Int J Cardiovasc Imaging 2002;18:111–8.
[14] Legget ME, Leotta DF, Bolson EL, et al. System for quantitative three-dimensionalechocardiography of the left ventricle based on a magnetic-field position and orien-tation sensing system. IEEE Trans Biomed Eng 1998;45:494–504.
[15] Gibson DG, Brown DJ. Continuous assessment of left ventricular shape in man. BrHeart J 1975;37:904–10.
[16] Munt BI, Leotta DF, Bolson EL, et al. Left ventricular shape analysis from three-dimensional echocardiograms. J Am Soc Echocardiogr 1998;11:761–9.
[17] Douglas PS, Morrow R, Ioli A, Reichek N. Left ventricular shape, afterload and surviv-al in idiopathic dilated cardiomyopathy. J Am Coll Cardiol 1989;13:311–5.
[18] Beohar N, Flaherty JD, Davidson CJ, et al. Quantitative assessment of regional leftventricular function with cardiac MRI: three-dimensional centersurface method.Catheter Cardiovasc Interv 2007;69:721–8.
Please cite this article as: Goleski PJ, et al, The shape and function of the lef10.1016/j.ijcard.2013.12.037
[19] Bolson E, Sheehan FH. CenterSurface model for 3D analysis of regional left ventricu-lar function. IEEE Computer Society Press; 1993 735–8.
[20] Hubka M, Lipiecki J, Bolson EL, et al. Three-dimensional echocardiographic measure-ment of left ventricular wall thickness: in vitro and in vivo validation. J Am SocEchocardiogr 2002;15:129–35.
[21] Sasayama S, Franklin D, Ross J, Kemper WS, McKown D. Dynamic changes in leftventricular wall thickness and their use in analyzing cardiac function in the con-scious dog. Am J Cardiol 1976;38:870–9.
[22] FeneleyMP, Hickie JB. Validity of echocardiographic determination of left ventricularsystolic wall thickening. Circulation 1984;70:226–32.
[23] Holman ER, Vliegen HW, van der Geest RJ, et al. Quantitative analysis of regional leftventricular function after myocardial infarction in the pig assessed with cine mag-netic resonance imaging. Magn Reson Med 1995;34:161–9.
[24] Yang P, Otto CM, Sheehan FH. The effect of normalization in reducing variability inregional wall thickening. J Am Soc Echocardiogr 1997;10:197–204.
[25] Sheehan FH, Bolson EL, Dodge HT, Mathey DG, Schofer J, Woo HW. Advantages andapplications of the centerline method for characterizing regional ventricular func-tion. Circulation 1986;74:293–305.
[26] Mancini GB, DeBoe SF, Anselmo E, LeFree MT. A comparison of traditional wall mo-tion assessment and quantitative shape analysis: a new method for characterizingleft ventricular function in humans. Am Heart J 1987;114:1183–91.
[27] Razdan A, Bae MS. Curvature estimation scheme for triangle meshes using biqua-dratic Bézier patches. Comput Aided Des 2005;37:1481–91.
[28] Bonow RO, Borer JS, Rosing DR, Bacharach SL, Green MV, Kent KM. Left ventricularfunctional reserve in adult patients with atrial septal defect: pre- and postoperativestudies. Circulation 1981;63:1315–22.
[29] Kelly DT, Spotnitz HM, Beiser GD, Pierce JE, Epstein SE. Effects of chronic right ven-tricular volume and pressure loading on left ventricular performance. Circulation1971;44:403–12.
[30] Louie EK, Lin SS, Reynertson SI, Brundage BH, Levitsky S, Rich S. Pressure and volumeloading of the right ventricle have opposite effects on left ventricular ejection frac-tion. Circulation 1995;92:819–24.
[31] Frescura C, Angelini A, Daliento L, Thiene G. Morphological aspects of Ebstein'sanomaly in adults. Thorac Cardiovasc Surg 2000;48:203–8.
[32] Milani RV, Lavie CJ, Mehra MR, Ventura HO, Kurtz JD, Messerli FH. Left ventriculargeometry and survival in patients with normal left ventricular ejection fraction.Am J Cardiol 2006;97:959–63.
[33] Sciancalepore MA, Maffessanti F, Patel AR, et al. Three-dimensional analysis ofinterventricular septal curvature from cardiac magnetic resonance images forthe evaluation of patients with pulmonary hypertension. Int J Cardiovasc Imaging2011:1–13.
[34] Yeo SY, Zhong L, Su Y, Tan RS, Ghista DN. A curvature-based approach for left ven-tricular shape analysis from cardiac magnetic resonance imaging. Med Biol EngComput 2009;47:313–22.
[35] Kingma I, Tyberg JV, Smith ER. Effects of diastolic transseptal pressure gradient onventricular septal position and motion. Circulation 1983;68:1304–14.
[36] Feneley MP, Olsen CO, Glower DD, Rankin JS. Effect of acutely increased right ven-tricular afterload on work output from the left ventricle in conscious dogs. Systolicventricular interaction. Circ Res 1989;65:135–45.
[37] Feneley M, Gavaghan T. Paradoxical and pseudoparadoxical interventricular septalmotion in patients with right ventricular volume overload. Circulation1986;74:230–8.
[38] Weyman AE, Wann S, Feigenbaum H, Dillon JC. Mechanism of abnormal septal mo-tion in patients with right ventricular volume overload: a cross-sectional echocar-diographic study. Circulation 1976;54:179–86.
t ventricle in Ebstein's anomaly, Int J Cardiol (2013), http://dx.doi.org/
