evidence of impaired left ventricular systolic function by doppler myocardial imaging in patients...

Pdamasdtdwmd2cpmL

Mc

A

0d

Evidence of Impaired Left Ventricular Systolic Functionby Doppler Myocardial Imaging in Patients With Systemic

Amyloidosis and No Evidence of Cardiac Involvement by StandardTwo-Dimensional and Doppler Echocardiography

Diego Bellavia, MD, Patricia A. Pellikka, MD, Theodore P. Abraham, MD,Ghormallah B. Al-Zahrani, MBBS, Angela Dispenzieri, MD, Jae K. Oh, MD, Kent R. Bailey, PhD,

Christina M. Wood, MS, Martha Q. Lacy, MD, Chinami Miyazaki, MD, andFletcher A. Miller, Jr, MD*

We examined the potential role of Doppler myocardial imaging for early detection ofsystolic dysfunction in patients with systemic amyloidosis (AL) but without evidence ofcardiac involvement by standard echocardiography. We identified 42 patients without2-dimensional echocardiographic or Doppler evidence of cardiac involvement. These pa-tients had normal ventricular wall thickness and normal velocity of the medial mitralannulus. Myocardial images were obtained in these patients and in 32 age- and gender-matched healthy controls. Peak longitudinal systolic tissue velocity (sTVI), systolic strainrate (sSR), and systolic strain (sS) were determined for 16 left ventricular segments. Radialand circumferential sSR and sS were also measured. Compared with controls in this groupof patients with AL, peak longitudinal sSR (�1.0 � 0.2 vs �1.4 � 0.2, p <0.001) and peaklongitudinal sS (�15.6 � 3.3 vs �22.5 � 2.0 p <0.001) were significantly decreased. Inconclusion, the mean sS from all 6 basal segments, or from all 16 left ventricular segmentsdifferentiated patients with AL with normal echocardiography from controls, with similaraccuracy for the mean sSR from the 6 basal segments. This distinction was not apparentfrom peak longitudinal sTVI or from radial or circumferential sSI or sSR. © 2008
Elsevier Inc. All rights reserved. (Am J Cardiol 2008;101:1039–1045)
M

Tosaw2Tg

puR2noiesebnsmv2

rimary amyloidosis (AL) is characterized by extracellulareposition of pathologic insoluble fibrillar protein in organsnd tissues.1,2 Congestive heart failure is seen in approxi-ately 25% of patients with AL, and its development is

ssociated with an average survival of �6 months.3 Recenttudies have shown the potential utility of Doppler myocar-ial imaging including systolic tissue velocity (sTVI), sys-olic strain rate (sSR), and systolic strain (sS)4,5 for earlyiagnosis of cardiac amyloid.6 The purpose of this studyas twofold: (1) to determine the potential role of Doppleryocardial imaging for identifying left ventricular (LV)

ysfunction in patients with systemic amyloid who have no-dimensional imaging or standard Doppler evidence ofardiac involvement compared with age- and gender-com-arable controls; and (2) to establish which of the Doppleryocardial imaging modalities is best for detection of earlyV dysfunction in patients with systemic amyloid.

Cardiovascular Division, Mayo Clinic and Foundation, Rochester,innesota. Manuscript received August 1, 2007; revised manuscript re-

eived and accepted November 13, 2007.This study was supported in part by a grant from the American Heart

ssociation Heartland Affiliate Grant No. 0620073Z.*Corresponding author: Tel.: 507-271-5479; fax: 507-284-3968.

aE-mail address: [email protected] (F.A. Miller).

002-9149/08/$ – see front matter © 2008 Elsevier Inc. All rights reserved.oi:10.1016/j.amjcard.2007.11.047

ethods

his study was approved by the Institutional Review Boardf the Mayo Clinic. From 103 consecutive patients withystemic AL who underwent complete echocardiographicnd Doppler exams, we identified 42 patients with ALithout evidence of cardiac involvement by standard-dimensional echocardiography and Doppler criteria.his study population was compared with 32 age- andender-matched healthy subjects who served as controls.

Subjects with AL were prospectively selected fromatients referred for an echocardiogram undergoing eval-ation in the Division of Hematology at the Mayo Clinic,ochester Minnesota, from February 2004 through June006. The diagnosis of AL was made either by subcuta-eous fat biopsy or an involved organ biopsy that dem-nstrated typical Congo red birefringence under polar-zed light. Patients with AL selected for this study had novidence of cardiac involvement by standard 2-dimen-ional echocardiography and Doppler assessment. Thus,xclusion criteria included increased LV wall thicknessy echocardiography, defined as mean value of LV thick-ess (half of the sum of the thickness of ventriculareptum and posterior walls) �12 mm for men and �11m for women (n � 35), and abnormal tissue Doppler

elocity of the medial mitral annulus (E’ velocity) (n �6), defined as �7 cm/s, indicating abnormal LV relax-
tion (Figure 1). Other exclusion criteria were familial or
www.AJConline.org

mailto:[email protected]

sm(apphmtemD

mSsteiaaaPiaciIa

apnamsdr

dyk

cadpivc

smdstDw6sviG

TCw

Vo

AMLR

LLLLEAEEAPPPAEEAP

ESC

ai

FdAvartc

1040 The American Journal of Cardiology (www.AJConline.org)

econdary AL (n � 3), senile AL (n � 1), history ofoderate or greater systemic or pulmonary hypertension

n � 5), significant valvular heart disease (n � 0), andtrial fibrillation (n � 0). Controls were selected fromatients undergoing clinically-indicated echocardiogra-hy who had no cardiac symptoms and no history ofypertension or cardiac disease. We compared Doppleryocardial imaging findings for the patients with sys-

emic AL without 2-dimensional or standard Dopplervidence of cardiac involvement with controls to deter-ine if these patients had LV systolic dysfunction byoppler myocardial imaging.All ultrasound examinations were performed with a com-

ercially available echocardiographic instrument (Vivid 7ystem, Vingmed-General Electric, Milwaukee, Wiscon-in). The thickness of the ventricular septum and LV pos-erior wall and the end-systolic and end-diastolic LV diam-ters were determined from M-mode or 2-dimensionalmaging, and LV mass and ejection fraction were calculateds previously described.7 Left atrial volume measurementnd pulsed wave Doppler of mitral inflow, pulmonary veins,nd LV outflow was performed as previously described.7ulsed wave tissue Doppler imaging was performed, plac-

ng the sample volume on the medial mitral annulus in thepical 4-chamber view. Off-line analysis of standard echo-ardiographic variables was performed with the use of ded-cated software (ProSolv CV analyzer version 3.0, ProSolv,ndianapolis, Indiana); 3 consecutive beats were measurednd averaged for each measurement.

For Doppler myocardial imaging, apical 4-chamber, longxis, and 2-chamber views were acquired, as well asarasternal short axis view at mid-LV level. In addition,arrow sector views were acquired for each LV wall frompical views, and for the parasternal short axis view at theiddle cardiac level. We collected 2-dimensional color tis-

ue Doppler recordings with frame rate �200 frames/suring brief breath holds; 3 consecutive cardiac cycles were

igure 1. Study design and group classification. Patients with a proveniagnosis of systemic AL were divided into groups: Group 1, patients withL without evidence of cardiac amyloid (normal LV thickness, normal E’elocity) (n � 42); Group 2, patients with evidence of early cardiacmyloid by pulsed wave tissue Doppler imaging (normal LV thickness,educed E’ velocity) (n � 26); Group 3, controls, age and gender similar tohe group of patients with systemic AL (n � 32). Groups 1 and 2 combinedreate AL-normal-wall-thickness (n � 68).

ecorded as 2-dimensional cine loops and the acquired raw L

ata were saved on magnetic optical disks for off-line anal-sis (Echopac BT06; Vingmed-General Electric, Milwau-ee, Wisconsin).

Sample volumes were placed on basal, middle, and api-al LV segments of the anterolateral, inferoseptal, posterior,nteroseptal, inferior, and anterior walls to assess longitu-inal Doppler myocardial imaging. Sample volumes werelaced at the middle level of the anteroseptal, posterior,nferoseptal, and lateral walls (parasternal short axisiew) to assess radial and circumferential Doppler myo-ardial imaging.

Longitudinal peak values were determined for sTVI,SR, and sS.6 Radial and circumferential peak values wereeasured for sSR and sS. For longtudinal Doppler myocar-

ial imaging, analysis was performed considering all 16 LVegments individually and combined in clusters accordingo 2 criteria: by LV level (basal, middle, and apical)—oppler myocardial imaging values (sTVI, sSR, and sS)ere averaged for the 6 basal segments (basal mean), for themiddle segments (middle mean), and for the 4 apical

egments (apical mean)—and by LV wall—sSR and sSalues were averaged for the anterolateral, inferoseptal,nferolateral, anteroseptal, inferior, and anterior walls.lobal longitudinal sS value (mean of sS values from all 16

able 1linical and standard echo variables for patients with amyloidosis (AL)ith normal echocardiograms versus controls

ariables, Mean � SDr n (%)

Patients with ALwith Normal

Echocardiograms(n � 42)

Controls(n � 32)

pValue

ge, yrs 58 � 9 61 � 12 0.34en 16 (38) 18 (56) 0.12

V thickness, mm 10 � 1 9 � 1 0.02ight ventricular free wallthickness, mm

6 � 1 6 � 1 0.38

V end-diastolic diameter, mm 47 � 5 48 � 6 0.19V end-systolic diameter, mm 29 � 4 31 � 5 0.02V mass index, gr/m2 89 � 19 80 � 17 0.14eft atrial volume index, cc/m2 33 � 9 32 � 9 0.44, m/s 0.8 � 0.2 0.7 � 0.2 0.16, m/s 0.7 � 0.2 0.8 � 0.8 0.26/A 1.2 � 0.6 1.2 � 0.5 0.63deceleration time, ms 205 � 42 208 � 36 0.74duration, ms 117 � 24 128 � 21 0.07

ulmonary vein S wave, m/s 0.6 � 0.2 0.6 � 0.1 0.21ulmonary vein D wave, m/s 0.5 � 0.1 0.5 � 0.1 0.16ulmonary vein A reversal, m/s 0.3 � 0.1 0.3 � 0.1 0.17– A reversal difference, ms 18 � 24 23 � 26 0.63

’, cm/s 8.5 � 1.5 9.0 � 2.6 0.6/E’ 9.7 � 2.5 8.6 � 2.5 0.06’, cm/s 9.2 � 2.3 9.6 � 2.8 0.75eak systolic tissue pulsed

wave Doppler velocity, cm/s7.3 � 1.5 8.4 � 2.0 0.01

jection fraction, % 65 � 6 63 � 4 0.35troke volume, ml 83 � 24 87 � 18 0.14ardiac index, l/min/m2 3.2 � 0.6 3.1 � 0.6 0.67

Descriptions are with mean � SD or count (percentage). Comparisonsre made using the Wilcoxon rank-sum test, chi-square test, or exactnference for ordered contingency tables.

V segments) ��18.5% was considered abnormal. This

tte

aNbetsapptoiuptac

a1fi

cetffepdp

R

EiTcpw(wes(tsg

Fr

1041Cardiomyopathy/Doppler Myocardial Imaging for Early Detection of Cardiac Amyloidosis

hreshold was chosen because it represents 2 SDs less thanhe mean of the global sS distribution in the healthy subjectsnrolled as controls.

Statistical analyses were performed with a commercially-vailable software program (SAS v8.2, SAS Institute, Cary,orth Carolina). Comparisons between groups were madey Wilcoxon rank-sum test, chi-square test, or exact infer-nce for ordered contingency tables. Diagnostic accuracy ofhe Doppler myocardial imaging modalities (sTVI, sSR, andS), including global mean values, mean value for clusters,nd measurements for individual LV segments, were com-ared by the receiver-operating characteristic curves foratients with AL with normal echocardiograms versus con-rols. Formal comparisons of the areas under the receiver-perating characteristic curves were performed as well, us-ng the method of Delong and Delong8 for comparing areasnder the receiver-operating characteristic curves in 2 testserformed on the same subjects. To address the question ofhe minimum number of segments or clusters required tochieve predictive accuracy, multiple logistic models wereonstructed.

Intraobserver and interobserver variability: To ex-mine intraobserver variability (repeatability), a sample of0 echocardiographic examinations were randomly selectedor masked review by the same investigator. To examine

igure 2. Boxplots of level means by groups for each of the longitudinalate imaging. (C) Strain imaging. *p �0.05; †p �0.01; ‡p �0.001.

nterobserver variability a co-investigator blinded to the clini- b

al information and to the results of the first investigatorxamined 10 randomly-selected echocardiographic exams. In-raclass correlation coefficients for the same observer and dif-erent observers were calculated using previously-describedormulas9 for single segments and for the global mean ofach Doppler myocardial imaging modality. Data are ex-ressed as mean value � SD or count (percentage). Aifference was considered statistically significant when thevalue was �0.05.

esults

ndomyocardial biopsy of the right ventricle was performedn 4 patients, and was positive for cardiac amyloid for all 4.wenty-six patients (62%) underwent peripheral blood stemell transplantation (PBSCT) with high dose melphalan/rednisone therapy. In the remaining patients, 12 (29%)ere treated with intravenous prednisone and melphalan, 3

7%) with dexamethasone and lenalidomide, and 1 (2%)ith dexamethasone alone. Three had low voltage on the

lectrocardiogram. There were no significant differences intandard 2-dimensional imaging for patients and controlsTable 1), except for LV end-systolic diameter and LVhickness. Although there were differences in these 2 mea-urements comparing the patient group with the controlroup, both measurements were within normal limits for

r myocardial imaging modalities. (A) Tissue velocity imaging. (B) Strain
Dopple
oth groups. No patient had pleural or pericardial effusion.

BDtDs

gc(Aiav

s0

grfcdFecp1i0acp0ag

n�downap1pves

TSs(

V(

sssssssssssss

TSa

V

BBBBBBMMMMMMAAAA

TS(

V

BBBBBBMMMMMMAAAA


ased on detailed 2-dimensional echocardiography andoppler analysis, there was no difference in diastolic func-

ion in patients and controls. However, peak systolic tissueoppler velocity measured at the medial mitral annulus was

ignificantly less in patients compared with controls.In patients with AL with normal echocardiograms, lon-

itudinal tissue velocities in apical segments were signifi-antly different compared with the control group. Clustersboth by wall and by level) were also different (Figure 2).pex lateral (1.9 � 0.5 vs 2.6 � 1.5, p � 0.04), apex

nferior (2.2 � 1.2 vs 2.6 � 1.0, p � 0.03), and apexnterior (1.0 � 1.2 vs 1.7 � 1.1, p � 0.003) single segment

able 2ystolic tissue velocity (sTVI), systolic strain rate (sSR), and systolictrain (sS) longitudinal cluster measures for patients with amyloidosisAL) with normal echocardiograms versus controls

ariablesmean � SD)

Patients with ALwith Normal


Controls(n � 32)

p Value

TVI basal mean 5.5 � 1.2 5.6 � 1.0 0.49TVI middle mean 3.8 � 0.9 3.9 � 1.0 0.46TVI apical mean 1.9 � 0.8 2.3 � 0.8 0.009TVI global mean 4.0 � 0.9 4.2 � 0.9 0.33SR basal mean �1.1 � 0.4 �1.6 � 0.2 �0.001SR middle mean �1.2 � 0.3 �1.4 � 0.2 0.002SR apical mean �1.0 � 0.3 �1.2 � 0.3 0.02SR global mean �1.1 � 0.2 �1.4 � 0.2 �0.001S basal mean �15.0 � 5.0 �23.7 � 2.1 �0.001S middle mean �17.7 � 3.7 �22.2 � 2.3 �0.001S apical mean �16.2 � 4.4 �21.0 � 3.5 �0.001S global mean �16.3 � 3.2 �22.4 � 2.0 �0.001S 3-basal-segments mean �15.0 � 5.7 �23.7 � 1.9 �0.001

Comparisons are made using the Wilcoxon rank-sum test.

able 3ystolic strain rate (sSR) longitudinal measures for patients withmyloidosis (AL) with normal echocardiograms versus controls

ariables, Mean � SD Patients with ALwith Normal


Controls(n � 32)

p Value

asal lateral �1.2 � 1.3 �1.9 � 0.6 0.003asal inferoseptum �1.0 � 0.4 �1.3 � 0.4 �0.001asal posterior �1.3 � 0.8 �2.0 � 0.9 �0.001asal anteroseptum �0.8 � 0.7 �1.4 � 0.4 �0.001asal inferior �0.9 � 0.8 �1.3 � 0.3 0.10asal anterior �1.1 � 0.7 �1.5 � 0.5 0.01iddle lateral �1.1 � 0.6 �1.4 � 0.5 0.002iddle inferoseptum �1.4 � 0.4 �1.4 � 0.5 0.73iddle posterior �0.9 � 0.7 �1.4 � 0.4 �0.001iddle anteroseptum �1.3 � 0.4 �1.4 � 0.3 0.13iddle inferior �1.1 � 0.6 �1.3 � 0.4 0.006iddle anterior �1.3 � 0.6 �1.4 � 0.3 0.31pex lateral �0.9 � 0.4 �1.1 � 0.4 0.02pex septum �1.2 � 0.5 �1.3 � 0.4 0.16pex inferior �1.3 � 0.5 �1.3 � 0.4 0.87pex anterior �0.7 � 0.6 �1.1 � 0.4 �0.001


elocities were all significantly reduced, as was the mean s

ystolic velocity at apex level (1.9 � 0.8 vs 2.3 � 0.8, p �.009) in patients with AL compared with controls.

In patients with AL with normal echocardiograms, thelobal mean sSR for all 16 segments was significantlyeduced compared with controls (Table 2). In addition, sSRor most of the individual basal segments was different fromontrols. sSR was also reduced for the middle lateral, mid-le inferolateral, and apical anterior segments (Table 3 andigure 2). Similarly, sSR for patients with AL with normalchocardiograms was significantly decreased in all of thelusters, with the exception of the inferior wall. Radialosterior sSR was not different in the 2 groups (1.8 � 0.7 vs.7 � 0.6, p � 0.51), but radial anteroseptal sSR was lowern patients with AL with normal echocardiograms (0.8 �.6 vs 1.1 � 0.4, p � 0.008). Circumferential sSR assessedt the inferoseptal or lateral wall was not different in patientsompared with healthy controls (�1.8 � 0.7 vs �1.9 � 0.5,� 0.31 for circumferential sSR at inferoseptum; �1.4 �

.6 vs �1.5 � 0.5, p � 0.31 for circumferential sSR atnterior wall, patients with AL with normal echocardio-rams vs controls, respectively).

Compared with controls, 80% of patients with AL withormal echocardiograms had an impaired global mean sS�18.5%, including the 4 cases with biopsy-proven car-

iac amyloid. Considering single segments and clusters,nly the middle inferoseptal and apical inferior segmentsere not significantly decreased for patients with AL withormal echocardiograms compared with controls (Table 4nd Figure 2). Radial posterior sS was not different inatients compared with controls (37.9 � 15.7 vs 37.2 �8.9, p � 0.96), but radial anteroseptal sS was reduced inatients with AL with normal echocardiograms (13.0 � 6.6s 17.0 � 5.0, p � 0.006, patients with AL with normalchocardiograms vs controls, respectively). Circumferentialystolic sS determined at the inferoseptal wall was not

able 4ystolic strain (sS) longitudinal measures for patients with amyloidosisAL) with normal echocardiograms versus controls

ariables, Mean � SD Patients with ALwith Normal


Controls(n � 32)

p Value

asal lateral �14.4 � 13.6 �25.6 � 5.7 �0.001asal inferoseptum �17.6 � 6.3 �22.9 � 3.8 �0.001asal posterior �13.3 � 10.8 �25.2 � 3.8 �0.001asal anteroseptum �14.1 � 8.9 �23 � 4.4 �0.001asal inferior �13.6 � 8.8 �23.4 � 3.8 �0.001asal anterior �17.0 � 8.5 �22.5 � 4.1 0.001iddle lateral �14.7 � 8.4 �21.0 � 5.2 0.002iddle inferoseptum �20.8 � 5.0 �22.4 � 5.3 0.19iddle posterior �15.2 � 9.7 �21.2 � 4.1 0.002iddle anteroseptum �19.4 � 5.9 �22.8 � 4.7 0.008iddle inferior �17.6 � 6.1 �23.3 � 4.0 �0.001iddle anterior �18.8 � 7.8 �22.2 � 5.3 0.04pex lateral �12.4 � 6.7 �18.7 � 4.9 �0.001pex septum �21.2 � 6.2 �24.6 � 4.6 �0.001pex inferior �20.6 � 7.6 �23.4 � 4.6 0.10pex anterior �10.9 � 6.2 �17.4 � 6.7 �0.001


ignificantly different in the groups (�26.5 � 9.3 vs �29.7

�p9

atmdmorptedsstsaou

aorp

asigsibs6udetets

Fesrt�cci


7.8, p � 0.15), but the lateral wall was increased inatients compared with controls (�25.5 � 8.0 vs �21.2 �.4, p � 0.03).

Results of the receiver-operating characteristic curvenalysis are shown in Figure 3. Peak longitudinal sS washe most accurate in all Doppler myocardial imagingodalities, followed closely by peak longitudinal sSR for

iscriminating controls from patients with AL with nor-al echocardiograms. Neither longitudinal sTVI nor any

f the radial or circumferential sSR or sS measurementseached good diagnostic accuracy in the discrimination ofatients with AL with normal echocardiograms from con-rols (p �0.001 for inferiority of global mean sTVI toither global mean sSR or global mean sS). The 3 bestiscriminators for these modalities (global mean sTVI,ystolic radial anteroseptal sS, and circumferential lateralS) achieved areas under the receiver-operating charac-eristic curves of 0.57, 0.71, and 0.65, respectively. Theimple systolic pulsed wave tissue Doppler measurementt the medial mitral annulus (area under the receiver-perating characteristic curve � 0.65, p � 0.01 vs area

igure 3. Receiver-operating characteristic analysis for the Doppler myochocardiograms versus controls. sS global mean � mean of the sS valuegments; sTVI global mean � mean of the sTVI values for 16 LV segmenadial anteroseptal sS � radial sS for the middle anteroseptal segment; sShe receiver-operating characteristic curves for patients with AL with norm

0.85; sTVI global mean � 0.57. (B) Areas under the receiver-operationtrols: sS global mean � 0.96; sSR global mean � 0.85; sTVI global meharacteristic curves for patients with AL with normal echocardiograms vnferoseptal, anteroseptal and inferolateral segments � 0.97.

nder the receiver-operating characteristic curve � 0.5), o

lthough not as accurate as the global mean sS (p �0.001)r sSR (p � 0.04), showed slightly higher diagnostic accu-acy than mean longitudinal sTVI (p � 0.12) to discriminateatients from controls.

Logistic regression and receiver-operating characteristicnalysis demonstrated that measurement of sS for anyingle LV segment is not sufficient to accurately discrim-nate between patients with AL with normal echocardio-rams and the controls. Calculation of the mean of theegments in a level, as opposed to the mean of segmentsn walls, tended to be the clustering technique providingetter discrimination. In the global and level means of sS,SR, and sTVI, the sS global mean and the sS mean of thebasal segments were the most accurate (greatest areas

nder the receiver-operating characteristic curves) foriscrimination between patients with AL with normalchocardiograms and healthy subjects (Table 2), andheir diagnostic accuracies were not significantly differ-nt (p � 0.77). In an exploratory analysis, the question ofhe minimum number of segments necessary within theS basal level was considered. It was found that the mean

imaging modalities, comparison between patients with AL with normal6 LV segments; sSR global mean � mean of the sSR values for 16 LVpulsed wave tissue Doppler systolic velocity of the mitral septal annulus;ean � mean of the sS values for 6 LV basal segments. (A) Areas under

ocardiograms versus controls: sSI global mean � 0.96; sSR global meanacteristic curves patients with AL with normal echocardiograms versus

.57; Radial anteroseptal sS � 0.71. (C) Areas under the receiver-operatingntrols: sS global mean � 0.96; sS Basal mean � 0.97; sS mean of basal

cardiales for 1ts; S’ �basal mal ech

ng charan � 0ersus co

f 3 specific segments (inferoseptal, anteroseptal, and

pao

pi0cw(t

pi00td99

D

Tiwesfiafdvtamarec

nwtdafelmdUstsW8fv

lndpff

wwbhcdftlua

pgcaaDctop

lafeAtpn

fcalmfcsrcn


osterior) yielded similar areas under the receiver-oper-ting characteristic curves (0.97 vs 0.97, p � 0.96) to thatf the basal level mean.

The intraclass correlation coefficient for intrareader re-roducibility was greater for sTVI (0.82, 95% confidencenterval [CI] 0.42 to 0.95) and sS (0.80. 95% CI 0.40 to.95) than for sSR (0.70, 95% CI 0.16 to 0.91). The intra-lass correlation coefficient for inter-reader reproducibilityas higher for sS (0.88, 95% CI 0.58 to 0.97) and sSR

0.73, 95% CI 0.23 to 0.93) than for sTVI (0.56, 95% CI 0o 0.87).

The intraclass correlation coefficient for intra-reader re-roducibility was similar for all the Doppler myocardialmaging modalities (0.99, 95% CI 0.99 to 0.99 for sTVI;.99, 95% CI 0.99 to 0.99 for sS; and 0.98, 95% CI 0.93 to.99 for sSR). For the global mean of all 16 LV segments,he intraclass correlation coefficient for inter-reader repro-ucibility was similarly high for sTVI, sS, and sSR (0.99,5% CI 0.98 to 0.99; 0.99, 95% CI 0.99 to 0.99; and 0.99,5% CI 0.97 to 0.99, respectively).

iscussion

his is the first investigation to obtain Doppler myocardialmaging measurements of all 16 LV segments in patientsith systemic AL without echocardiographic or Doppler

vidence of cardiac involvement and to compare these mea-urements with Doppler myocardial imaging measurementsrom healthy subjects. The main findings of the currentnvestigation were as follows: (1) Doppler myocardial im-ging modalities revealed evidence of impaired LV systolicunction in patients with systemic AL, despite normal stan-ard 2-dimensional echo and normal pulsed wave Dopplerelocity (E’) of the mitral annulus; (2) longitudinal sS washe most accurate diagnostic modality; and (3) mean sS ofll 16 LV segments, mean sS of the 6 basal segments, or theean sS of 3 specific basal segments (basal inferoseptal,

nterospetal, and posterior segments) were the most accu-ate Doppler myocardial imaging measurements to differ-ntiate patients with AL with normal echocardiograms fromontrols.

Koyama et al6 showed the diagnostic accuracy of theew Doppler myocardial imaging modalities for patientsith cardiac amyloid. Using a 12-segment LV model,

hey found that peak sSR and sS demonstrated significantifferences in amyloid patients without cardiac amyloidnd patients with cardiac amyloid with and without heartailure. One of the 2 goals with our investigation was toxtend the findings of Koyama et al6 by comparing Dopp-er myocardial imaging measurements for all 16 LV seg-ents between patients with AL with normal echocar-

iograms and healthy age- and gender-matched controls.sing longitudinal sS and sSR measured in the 16 LV

egments, we documented a global decrease in LV sys-olic function for patients with systemic AL who had notandard echo-Doppler evidence of cardiac involvement.

hen the threshold of ��18.5% was used for normal sS,0% of these patients had reduced longitudinal systolicunction despite having normal ejection fraction, stroke
olume, and cardiac index. Longitudinal LV fibers are
ocated in the subendocardium, and pathologic and mag-etic resonance imaging studies have highlighted suben-ocardial infiltration in patients with AL. This may ex-lain early impairment of longitudinal LV systolicunction while radial LV systolic function is preservedor these patients.

Cardiac biopsy was not often clinically-indicated andas therefore performed in only 4 of our patients with ALith normal echocardiograms, and all 4 had myocardialiopsy positive for amyloid infiltration. This supports theypothesis that the decreased longitudinal LV functionould be due to amyloid infiltration that is too small to beetected by standard echo-Doppler methods. Alternatively,or some of these patients the longitudinal systolic dysfunc-ion may arise from toxic myocardial effect of circulatingight chains, as has been previously proposed,10,11 or fromnrecognized effects of PBSCT and/or amyloid chemother-py regimens.

In addition to providing new insight into the cardiacathophysiology of patients with AL, the detection of lon-itudinal LV dysfunction for these patients may prove to belinically beneficial. Patients with AL with cardiac diseasere at higher risk of complications with many proceduresnd treatment modalities, including PBSCT. Recognition byoppler myocardial imaging analysis of earlier, subclinical

ardiac dysfunction may lead to trials of PBSCT earlier inhe course of the disease. Furthermore, serial measurementsf longitudinal LV sS may prove useful in monitoring theseatients during and after therapy.

In this investigation, the average time needed to performongitudinal sTVI, sSR, and sS acquisition and analysis forll 16 LV segments was approximately 40 minutes. It is noteasible to incorporate this extensive analysis into routinechocardiography examinations of patients with systemicL. Therefore, our second goal was to determine the best of

hese 3 modalities for detecting LV systolic dysfunction foratients with AL with normal echocardiograms and the dataeeded to confidently identify this dysfunction.

In the Doppler myocardial imaging modalities, sTVIailed to differentiate patients with AL with normal echo-ardiograms from healthy controls. Radial anteroseptal sSnd sSR and lateral circumferential sS were significantlyower for patients compared with controls. However, theost accurate Doppler myocardial imaging modality in dif-

erentiating these 2 groups was longitudinal sS, followedlosely by longitudinal sSR. Global sS mean of all 16 LVegments and the sS mean of the 6 LV basal segmentseached similar diagnostic accuracy and were the most ac-urate measurements to differentiate patients with AL withormal echocardiograms from controls.

1. Wright JR, Calkins E. Clinical-pathologic differentiation of commonamyloid syndromes. Medicine (Baltimore) 1981;60:429–448.

2. Pellikka PA, Holmes DR Jr, Edwards WD, Nishimura RA, Tajik AJ,Kyle RA. Endomyocardial biopsy in 30 patients with primary amy-loidosis and suspected cardiac involvement. Arch Intern Med 1988;148:662–666.

3. Falk RH. Diagnosis and management of the cardiac amyloidoses.Circulation 2005;112:2047–2060.

4. Heimdal A, Stoylen A, Torp H, Skjaerpe T. Real-time strain rateimaging of the left ventricle by ultrasound. J Am Soc Echocardiogr
1998;11:1013–1019.

1

1


5. Urheim S, Edvardsen T, Torp H, Angelsen B, Smiseth OA. Myocar-dial strain by Doppler echocardiography: validation of a new methodto quantify regional myocardial function. Circulation 2000;102:1158–1164.

6. Koyama J, Ray-Sequin PA, Falk RH. Longitudinal myocardial func-tion assessed by tissue velocity, strain, and strain rate tissue Dopplerechocardiography in patients with AL (primary) cardiac amyloidosis.Circulation 2003;107:2446–2452.

7. Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pel-likka PA, Picard MH, Roman MJ, Seward J, Shanewise JS, et al.Recommendations for chamber quantification: a report from the Amer-ican Society of Echocardiography’s Guidelines and Standards Com-mittee and the Chamber Quantification Writing Group, developed in
conjunction with the European Association of Echocardiography, a
branch of the European Society of Cardiology. J Am Soc Echocardiogr2005;18:1440–1463.

8. Mehta CR, Patel NR, Tsiatis AA. Exact significance testing for or-dered categorical data. Biomatrics 1984;40:819–825.

9. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing raterreliability. Psychological Bulletin 1979;86:420–428.

0. Liao R, Jain M, Teller P, Connors LH, Ngoy S, Skinner M, Falk RH,Apstein CS. Infusion of light chains from patients with cardiac amy-loidosis causes diastolic dysfunction in isolated mouse hearts. Circu-lation 2001;104:1594–1597.

1. Brenner DA, Jain M, Pimentel DR, Wang B, Connors LH, Skinner M,Apstein CS, Liao R. Human amyloidogenic light chains directly impaircardiomyocyte function through an increase in cellular oxidant stress.
Circ Res 2004;94:1008–1010.

evidence of impaired left ventricular systolic function by doppler myocardial imaging in patients...

Documents