Myocardial electrical propagation in patients with idiopathic dilated cardiomyopathy

Myocardial electrical propagation in patients

with idiopathic dilated cardiomyopathy.

K P Anderson, … , R L Lux, S V Karwandee

J Clin Invest.

1993;

92(1)

:122-140.

https://doi.org/10.1172/JCI116540

.

Myocardial propagation may contribute to fatal arrhythmias in patients with idiopathic dilated

cardiomyopathy (IDC). We examined this property in 15 patients with IDC undergoing

cardiac transplantation and in 14 control subjects. An 8 x 8 array with electrodes 2 mm apart

was used to determine the electrical activation sequence over a small region of the left

ventricular surface. Tissue from the area beneath the electrode array was examined in the

patients with IDC. The patients with IDC could be divided into three groups. Group I (n = 7)

had activation patterns and estimates of longitudinal (theta L = 0.84 +/- 0.09 m/s) and

transverse (theta T = 0.23 +/- 0.05 m/s) conduction velocities that were no different from

controls (theta L = 0.80 +/- 0.08 m/s, theta T = 0.23 +/- 0.03 m/s). Group II (n = 4) had

fractionated electrograms and disturbed transverse conduction with normal longitudinal

activation, features characteristic of nonuniform anisotropic properties. Two of the control

patients also had this pattern. Group III (n = 4) had fractionated potentials and severely

disturbed transverse and longitudinal propagation. The amount of myocardial fibrosis

correlated with the severity of abnormal propagation. We conclude that (a) severe

contractile dysfunction is not necessarily accompanied by changes in propagation, and (b)

nonuniform anisotropic propagation is present in a large proportion of patients with […]

Research Article

Find the latest version:

Myocardial Electrical

Propagation in Patients

with Idiopathic Dilated Cardiomyopathy

Kelley P. Anderson, Richard Walker, Paul Urie, Philip R. Ershler, Robert L. Lux, and Shreekant V. Karwandee

Cardiology

Division, and DepartmentofCardiothoracicSurgery, University of Utah Medical Center and NoraEcclesHarrison

CardiovascitlarTraininganlResearchInstitute, University of Utah, Salt Lake

CitY,

Utah 84132

Abstract

Myocardial propagation may contribute to fatal arrhythmias in

patients with idiopathic dilated cardiomyopathy (IDC). We

examined this property in 15 patients with IDC undergoing

cardiac

transplantation

and in 14 control

subjects. An 8 X 8

array with electrodes 2 mm apart was used to determine the

electrical

activation sequence over a small region of the left

ventricular surface. Tissue from the area beneath the electrode

array was examined in the patients with IDC. The patients with

IDC could be divided into three groups. Group I (n

=

7) had

activation

patterns

and

estimates

of

longitudinal

(OL

=

0.84±0.09

m/s)

and transverse

(OT

=

0.23±0.05

m/s)

con-duction

velocities that were no

different

from controls

(OL

=

0.80±0.08

m/s, OT

=

0.23±0.03

m/s).

Group

II

(n

=

4) had

fractionated

electrograms and disturbed transverse conduction

with normal

longitudinal activation, features characteristic of

nonuniform

anisotropic properties. Two of the control patients

also had

this pattern.

Group

III

(n

=

4)

had

fractionated

poten-tials and severely disturbed transverse and

longitudinal

propa-gation. The

amount

of

myocardial

fibrosis correlated with the

severity of abnormal

propagation.

We conclude that

(a) severe

contractile

dysfunction

is

not

necessarily accompanied

by

changes in

propagation,

and

(b)

nonuniform

anisotropic

propa-gation

is present in a large proportion of patients with IDC and

could

underlie ventricular

arrhythmias

in this disorder. (J.

Clin.

Invest.

1993.

92:122-140.) Key

words:

arrhythmia

*

car-diomyopathy *

electrophysiology

*

myocardium

*

sudden death

Introduction

Sudden death

occurs

frequently

in

patients with

idiopathic

di-lated

cardiomyopathy

(IDC),'

accounting for about half ofthe

deaths in

this disorder ( 1, 2). There is evidence that ventricular

arrhythmias

are

a

cause

of

sudden death in IDC

(1).

Unfortu-nately,

a

major

clinical

problem

persists because treatment that

Thisreportwaspresentedin part in abstract formatthe62nd

Ameri-canHeartAssociation Scientific

Sessions,

New

Orleans,

LA 1989.

Address

reprint

requests to Dr.

Kelley

P.

Anderson,

Cardiology

Division, University of Utah Medical

Center,

Salt Lake

City,

UT 84132.

Receivedfor

publication

II June 1992andinrevised

form

30

De-cember 1992.

1.Abbreviations usedinthispaper:

IDC,

idiopathic

dilated

cardiomyop-athy;NAI, nonuniformactivation

index; PVC,

prematureventricular

contraction; RMS,root meansquare;

ROC,

receiver

_operating

charac-teristic; TAT,totalactivation

time;

VT,

ventricular

_tachycardia.

could be

directed toward large

numbers of patients, such as

antiarrhythmic drugs,

has not been shown to be uniformly

ef-fective

(2, 3), and highly specific predictors

of sudden

death,

which could be used to target therapies such as cardiac

trans-plantation or defibrillator imtrans-plantation, have not been

identi-fied

( 1, 2). Therefore, there is

a pressing need

for

better

under-standing of the factors that predispose patients

with IDC to

malignant ventricular tachyarrhythmias.

Much of what is known about ventricular arrhythmias is

based on

studies of

patients with ischemic heart disease and

animal

models of

myocardial infarction.

In

these studies

reen-try has been shown

to

be

an

important mechanism of

arrhyth-mias, and abnormalities of conduction appear to play a

princi-pal

role in

their

genesis

(4).

Much

less is

known about

arrhyth-mia

mechanisms in patients with IDC. There is some evidence

of disturbed

conduction in

patients

with IDC (2, 5-8), and

there

is considerable evidence of abnormal cellular

function

(9-1 1 ),

abnormal passive electrical properties ( 12, 13), and

abnormal

morphologic features

that

could contribute to

abnor-mal

propagation

(

14-24). However, we are not aware of any

previous

attempts to systematically examine the spread

ofelec-trical

activity

in

cardiomyopathic

states in

humans or in

ani-mal

models. Because

of

the large

number of abnormalities that

could

affect conduction

in

patients with

IDC, a model

of

myo-cardial

propagation

was

used

to

predict

the

effects

that reported

functional

and structural

changes

would have on

myocardial

propagation.

The

hypothesis

was that electrical propagation in

the

myocardium

of

patients

with IDC would

differ from

that in

control

subjects,

and

that the

changes

would

conform

to

pat-terns

predicted

by the model.

Methods

Theory

Ourpredictionsabout thebehaviorofpropagating wavefronts in dis-easedmyocardiumweredrawnfromtwo sources, empirical data and models ofpropagation. Neitheris completely satisfactory since neither canimitatediseasedtissueperfectlyandfor severalpropertiesthereis

nosuitablepublished information from eithersource. To supplement

existing

evidence,we usedananalyticmodelofpropagation derived from continuouscable theory by Keener (25):

MfL _{/___l}

CmRm \

r+q~+q

where

rS

=r.

SL

2R.

SL2Re (1)

weer=Rm

i-1,if

l

R.'m

q

VRm

and where 0 is the effective conductionvelocity (26), Mfisa mem-brane function that incorporates excitability, ionic currents, and

re-latedphenomena(25),Lis thelengthof the myocyte in the direction ofpropagation, Cmis the membranecapacitance, Rmis the membrane

resistance,Sis the membrane surfaceareaofasinglemyocyte,

_R.

is the gapjunctional resistance,

Ri

is thecytoplasmic resistance,

1V

is the volumeof asingle cardiac cell,

R,

is the resistance of the extracellular

medium,and

Ve

is the volume of the extracellular medium. It is

as-J.Clin. Invest.

©TheAmerican

Society

forClinical

Investigation,

Inc.

0021-9738/93/07/122/19

$2.00

sumed that propagationcanberepresented byaplane wave traveling either paralleltofiber orientationorperpendicular to it, that thereare only two types of currents, transmembrane and axial, and that voltage potentialsareuniformly distributed along the cell and in the extracellu-lar spaceperpendiculartothe direction ofpropagation (25, 27, 28). Moreover,it is assumed that the parameters given in Eq. 1are indepen-dentof each other. Violations of the assumptions of continuous cable theoryarewell known(25, 27); in fact, Eq. 1wasderivedprimarilyto demonstrate the inadequacies of this model in comparison with a model based on the discretenatureofmyocardial propagation(25). However, this model suitedourpurpose, whichwas tocomplement previously reported resultstodevelopworkinghypotheses about the patternsofpropagationthatmight be expected for the variety of struc-tural and electrical alterationsanticipatedinmyopathiccardiac tissue ratherthan toderive specificresults.Theadvantage ofthisformulation is that it presents propagation velocityas afunction of theimportant cable parameters whichpermitsanintuitiveappreciation of the

rela-tionbetweenelectricalpropertieswithout therequirement of complex

numericalsimulations.

Thebaseline values of the parametersofthe model were obtained from the literature or from reasonable estimates, and all parameter values were varied to assess the effect on conduction velocities: The lengthof the cell for propagation longitudinal tofiber orientation: LL

= 120

gm

(0.012 cm), the length of the cell (diameter of the cell) for propagation transverse to fiber orientation: LT = 20,Am(0.002 cm), which arewithin the normal range for mammalian ventricular myo-cytes(29).The volumeofasinglemyocytewascalculatedfrom the length anddiameter assumingacylindrical shape:

_Vi

= 3.77 x 10-8

cm3.The membranesurfaceareaofamyocyte wasobtained by multi-plying thesurfaceareaof the cylinder by 10toestimate the additional surfaceprovidedby thetransverse tubular system(27, 30): S= 7.54

X l0-4 cm2 . The volume of the extracellular space was assumed to be

approximatelyone-sixth of the volume of the myocyte to represent relatively tight packing of cells in normal myocardium: Vc=6.0X10-9 cm3(31 ).Specificmembranecapacitance: Cm = 1

gF/cm2

(30, 32).

Specificmembrane resistance: Rm = 9,000Qcm2 (30). Intracellular resistance:

_Ri

=400Qcm(30, 33, 34). Extracellular resistance:

_R,

=50 Qcm(30). The gapjunctionalresistance in the longitudinal direction

rgL

=0.5mQ; in the transverse direction rgT= 1 MQ (33, 34). With the membrane function parameterarbitrarily set Mf= 30, these values resulted ineffectivelongitudinal conduction velocity AL = 0.77 m/s, effective transverse conduction velocity AT = 0.23 m/s, and 6L/0T

=3.5,which corresponded well to the average findings in control sub-jects.

Sixgeneralpatterns ofmyocardial activation,whichmightbe ob-served basedonknownalterationsinmyopathicstates,werederived from the model andprevious experimentalresults:

PatternA.This firstpattern isofuniformanisotropic propagation

(35) with values ofOL, OT,and0L/6Tnotdifferentfrom controls. Al-though few studies have beenperformedin humanventricular

myocar-dium,studies in otherspeciesdemonstrate that patternAistypicalof activationattheleft ventricular surface(36,37).

Pattern B.The second pattern is of uniformanisotropic propaga-tion withproportionately equalreductions of OL and

0T

andunchanged OL/OT(pattern Bl ),orproportionatelygreater reduction of 6L than 0T and decreased

0L/OT

(pattern B2).Reductions inrestingmembrane

potentialorinactionpotential upstroke velocity(Vm,) correspondto lowervaluesof Mf and should result in lower

0L

and0Twith no change in

0L/OT

(pattern

Bl

)according to the model. Experimental interven-tions that alter theseproperties, however,havesometimes resulted in pattern B2(37-43)aswellas

Bl

(28, 44). The modelpredictsthat increases inCm,

Rm,

orS produce theBl pattern (Fig. 1 A). The B2 pattern isexpected with increases ofcytoplasmic(Ri) and extracellular

(Re)resistance(Fig. 1 B). We could findnoreportsof experimental manipulations that supportedorrefuted the predictions of the model for two-dimensionalspreadofactivation withrespecttochangesinCm,

Rm,

S, Ri,

or

Re.

Pattern C. Third is a patternof uniform anisotropic propagation with increases Of EL and ET. A numberofchanges known to occur in the myocardium of patients with IDC would be expected actually to in-crease 0.The model predicts increases in AL and AT with no change in

0L/OT (pattern

Cl

)

with reduced S dueto

_shrinkage

of the

transverse-tubular system (Fig. 1 A). Ithas been suggestedthat reduced numbers of myofibrils ( 10, 19, 22, 24, 45) decreases

_Ri

(30, 46). Reductions of

Ri

or

Re

increaseAL, AT,and OL/OT (pattern C2, Fig. 1 B). An increase of Ve resulting from expansion of the interstitial space increases EL and

OL/OT,

with

_only

a

_slight

increaseof

OT

(Fig.

1 C). Thepatternthat accompanies myocyte hypertrophy depends whether the change in myocyte volume (Vi) is due to lengthening or widening of the cell. Elevation ofVidue toincreased cell length (LL) increases AL andOL/OT,

while AT changes little (pattern C2). In contrast,OL/0Tdeclines (pattern C3) with larger cell diameter (Fig. 1 D). The model predicts that pro-portional increases in LL and

LT

produce aC3 pattern with a gradual fall of OL/OT as EL and OT rise (Fig. 1 E).

Pattern D. Thefourth pattern is one of uniform anisotropic propa-gation with decreasedAT,increased0L/OT,and slightdecreases or no change of AL. This is the pattern predicted withageneralized increase in gapjunctional resistance with proportionately equal changes in the

longitudinal

(rgL)

and transverse

_(rgT)

directions (Fig. 1 F), and has beenobservedexperimentally (28, 47, 48).

PatternE. Next is anonuniform anisotropicpattern: variable re-ductions in

0T,

normal0L and fractionated electrograms during trans-versepropagation. This pattern of propagation was identified by Spach and co-workers who proposed that such behavior results from reduced

"side-to-side"electricalconnections(35, 49).

Pattern F. Lastisanonuniformanisotropic pattern with variable reductions of0Land0Tand fractionated electrograms. This pattern

differsfrom pattern E in that bothlongitudinaland transverse propaga-tionaredisturbedinanirregularpattern. It hasbeen observed in the borderzoneof experimental healedmyocardialinfarction (50, 51).

Patients

The studysubjects consistedofpatients with IDC undergoing cardiac

transplantationforseverecongestiveheartfailure. Coronary artery dis-ease, valvular heartdisease, hypertension, diabetes mellitus, alcohol abuse,hypertrophiccardiomyopathy, and infiltrative cardiomyopaties were excludedin all patients by an extensive evaluation which included a history, physical examination, electrocardiogram and, when indi-cated,echocardiography,radionuclide angiography, coronary and left ventricular angiography, and myocardial biopsy. The control subjects consistedofpatientswith normal cardiac function undergoing opera-tive correction ofsupraventriculararrhythmias. None of the patients werereceiving antiarrhythmic drugs or had ever received amiodarone. Informed writtenconsent tothe protocol approved by the University of UtahInstitutionalReview Board was obtained for all patients.

Electrophysiologic

data

Electricalactivationdata were obtained from an electrode array

con-sistingof 64 silver electrodes (0.6mmindiameter, 2-mm interelec-trodedistance)inan8X8 square pattern embedded in a rigid epoxy plaque. The recordings were obtained in unipolar mode referenced to a needle in theright chest wall. The data acquisition and analysis system hasbeendescribed previously (52). Electrograms were recorded at a bandwidthof0.03-500 Hz, sampled at a 1 kHz rate, and digitized by a 12-bit analog-to-digital converter. The intraoperative collection of data wasstandardized as far as possible to produce uniform experimental conditions but for ethical and practical reasons was limited to 20 min. Electrophysiological recordings were obtained before the initiation of cardiopulmonary bypass. The closely spaced electrode plaque array wasplaced on theleft ventricular epicardial surface adjacent to the left anteriordescendingartery two-thirds of the way from the base of the hearttoward the apex and thus several centimeters away from any

-_7.54104_Cm2

---- _OL

-

_OT

OLOJOT

A I _I._{._}. . . .~..

-0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.C

S

(cm2)

I-o6.0

10-9cm3

I.I . . . .V. (cm

1.0

10-8

2.0

10-8

3.0I104

V. (CM3)

4 . I .

.

. .

I.

4.0lo 5.o 10

-3.8

108cm3

._--.- ----_-.

1.0

10-7

2.010 VI(cm3)

3.0 10 4.0

107

6 -5 4

OL

3

(T

-2 6 5 4 9L

-3

OT -2 -1 6.0

109-F6

_s

160-0 120-(cm/h) 80-

40-D

F

OL

0 12 -3_0T

_(Cm/a)

8 -2

50 Qcm

ft~~~~~~oqb-*Q, _ _, -___.

100 200 300

Re

(f cm)

3.0

rgL

(Q)

Figure1. Plots oflongitudinalconductionvelocity (OL),transverse conductionvelocity (OT),and the ratio of the velocities(OL/OT)asfunctionsof severalelectrophysiologicvariablesaccordingtoacontinuouscable model(Eq. 1).The "normal" value of each variable isindicatedbyavertical line.EL, ET,and 0L/0Tareplottedasfunctions of thegivenvariable from0.1to10timesthenormalvalue inpanelsA,B,andC.(A) Increasing

membrane surfacearea(S)results inproportionately equalreductions OfELandAT;

OL/0T

remainsunchanged (patternBl).DecreasingS in-creasesALandETwithnochangeof

OL/OT

(pattern Cl). (B) Increasing RJ causesagreaterreductionofELthanETSOOL/0Tdeclines(pattern B2). Decreasing

R&

increasesALand 0L/0TwhileATincreasesslightly (pattern C2). (C) Increasing VeresultsinariseOfELand 0L/0TwhileAT

in-creasesonlyslightly(pattern C3). (D) Myocytevolume

(Vi),

whichisafunction ofthelength (LL)anddiameter(LT )ofthe cell andisassumed

tobecylindrical(V1 = _irLLLr2)wasvariedbychanging only Lr, keepingLLconstant. AsVi increases, ATand 0L/0T increase,whileELchanges

onlyslightly. (E)Thisplotwasobtainedby varyingbothLLand_Lrbutmaintainingaconstant ratio ofLLandLT; Viremains a function of

LL

andLr.AsViincreasesbothALandETrise,while 0L/0Tdeclinesslightly. (F)

EL,O

Tand 0L/0Tas afunction oflongitudinalgapjunctional

resis-tance(rL).Thisplotwasobtainedvarying rLbetween 0.1and 10times the normal value(5X

I05),

andkeepingthe ratio oflongitudinaland

transversegapjunctionalresistance constant(rL/rUT=0.5).

arrayusing rectangular impulses 2msin duration attwice diastolic threshold current. Data wereacquired afterpacing for 10s at 150

beats/min. Insome cases recordingswereobtainedafterpremature stimulidelivered after eight driven beatsat120or 150beats/min.In

thepatients with IDCthelocation ofthe plaquewasmarkedbefore removalwithagentian violetpen.

Histologic sections

Afterthe heartwasremovedin patients withIDC, thefull thickness of

myocardium beneaththe plaque electrodearray wasexcised and fixed

in 10%formalin, embedded in paraffin,andstained withhematoxylin

andeosinandMasson's trichrome stains.The subepicardialtissuewas

examinedtodeterminetheorientationandalignmentof myocardial

fibers and thepatternoffibrosis. The proportionoffibroustissue ina

given field wasdetermined usingaCUE 2 Image Analysis System

(Olympus Corp., Cherry Hill, NJ). The color andgrayscale character-istics of the fibrous tissue stained with trichromeweredetermined and automatically highlighted by the analysis system. After the

computer-selected regionswereverified and corrected by theoperator,the frac-tion of fibrous tissuewascomputed by the analysis system.

Determination ofactivation

The minimum value (i.e., thegreatestnegative value) of the first tem-poralderivative of the extracellular potential

(+.,4)

has been shown to correspond with themostrapidphase of the action potentialupstroke

(49). Hence,thetimeofthedeflection(TD)wasdefinedasthetimeat

which4Omioccurred. Numerous possibletestsfor activationwere

com-paredtoidentify thetestwith the greatest abilitytodiscriminate local fromnonlocal electricalactivity usingmethodspreviouslydescribedin detail (53). I I I I I I I I

tI

A

200.

160-o 120-(cms)) 80-

40-

C2-C

200- 160-o 120-(cm/a) 80-

40-1.

I . 400 400 -6 -5 4 L 3 _9T

-2

1 10 iL

OT

E

2001

160-0 120-(cm/s) 80- 40-I'

I - I

I.. ..I... .I ....I. ...I.... I....I ... 0

o 1 . I

1

. . . ..r. . . . _r- *-

In

v

%F-@ w w

%qb.m'm.b

.6

Conduction velocity

Strict measurement of conduction velocity requiresknowledgeof the entire path of activation in three dimensions. However, consistent val-ues of the "effective" conduction velocity (6)canbe obtained from recordingsobtainedintwodimensions ifcareis takentoinclude only

propagationin therecording plane (37, 42, 54). This isprimarilya problem forpropagationtransverse tofiber orientation. Human left ventricularmyocardiumtendstobearranged in "onionskin" layers where thefiberorientation of each successive layer isslightlyrotated (55 ). Thus,myocardial tissue subjecttoactivation viatransverse propa-gationatthesurfacecanbe"preexcited" byawavefront that travels morerapidly inasubepicardiallayer whichcanleadto an overestima-tion of 6 (56). Thiscanbe avoidedbyexcludingareasdistant from the stimulation site which demonstrate widening of isochronecontours. Twomethodsof estimating6 were used: In method A,longitudinaland transverse6

(AL

and

6AT,

respectively)werecalculatedbydividingthe distance between twoelectrode sites bythe difference in activation timesatthose sites (35,37, 42).The sites used in the calculationwere those that determinedaline normaltothe wavefront in the longitu-dinal (fast)andtransverse(slow)directions(Fig. 2).Insomecases, however, identification ofalineperpendicular tothewavefront

re-quiredsomejudgementand insome casestherewere nosites thatwere

exactly collinear with the line. For thesereasons asecond method of estimating 6wasobtained from thevelocityofaplanewaveacrossfour

adjacentsites(57,58):

D2

B

_(a

-

b)2

+

(c-

d)2'

(2)

where Distwicedistancebetweenelectrodes (4mm)and a,b,c, andd

are the activationtimesatsitessuperior,inferior,tothe

left,

andtothe right of the given site, respectively. Method B estimates of longitudinal andtransverseconduction velocities(OBLandOBT,

respectively)

con-sisted of themeanof valuescomputedatseveralsitesalongthe longitu-dinalandtransversedirections(Fig. 2).For bothmethods,sites within 4mmof the stimulation sitewere notusedtoavoid the virtual cathode

effect (59).Areasof nonuniform activation (sudden narrowingor

wid-ening ofisochrones)werealsoavoided.

Nonuniform

activation

index (NAI)

Nonuniformityof activationwasquantified by usingtheNAI,which has beendescribed indetailpreviously (56).Inbrief,the NAIat an electrodesitex

(NAIl)

is thesumofthe second

spatial

derivatives of

activation times infour directions timesthe electrodespacing. The second centraldifferenceformulawasusedtoestimate the derivative termswithAx asthe electrodespacing:

A-X B-X C-X D-X E-X

VAx A

fAx

Ax VA

f2AX

Fx

_+~+

AX

_2AX

x

+

2

F-X G-X H-X

+ ~+ ~~+ 3

Ax

_W2Ax

Ax ' (3)

where Xis theactivationtimeatsite x, and A, B,C,D, E, F,G, and H aretheactivationstimesofthesitesadjacenttositex. ApositiveNAIX occurs with deceleration of activation, i.e., when isochrones get narrower. Anegative value indicates acceleration ofactivation,or

wid-ening of isochrones.Inthis study NAIwasthe averageoftheabsolute values the

_NAIX's

computed at everypossible electrode site during pac-ingfrom each edge of the electrode plaque.

Statistical analysis

Receiver operating characteristic(ROC)curves were used to compare thecandidatetestsfor activation using methods described previously

(53, 60). The

quality

of

sensitivity

(QSE),

specificity

(Qsp),

and

effi-ciency(QE) were calculatedasproposed by Kraemer (60). QSEwas

plottedagainst

Qsp

toobtain the"quality" ROC. Student's t test was

usedforcomparison of continuous variables.The Kolmogorov-Smir-novtest wasusedwhen thedistributionswerenotnormal.Differences

wereconsidered

significant

whenP<0.05.

1 16

0~~~~~~~~

Base LV Free Wall

SepOum Apex

Figure2.Electrical activation map basedonrecordings from the 64 electrode plaque array placedonthe free wall of the left ventricular surface near the anterior interventricular groove inapatient with IDC. Its orientation with respecttothe septum, left free wall, apex, and base of the heart is indicated in the diagram below the map. The isochrone contours are drawn by the method of linear interpolation byacomputer program basedon theactivation times. The interval between isochrones is 2ms.Theperipheral electrodes sitesare onthe border of the mapdiagram, the inner electrode sitesaredepicted by +, and the labeled electrodesites (A, B, C,etc.)areindicated by filled circles (*). The numbersadjacenttoeachelectrode sitearethe acti-vationtimesrelative to the time of the stimulus artifact. The program doesnotattempttosmoothortodraw curved contours; sites with simultaneous activationareconnected with straight linesresultingin somewhat jaggedcontours.Theprintednumbersareabsentatsites where therecording channel malfunctioned or where the stimulus artifact obscured theelectrograms. Stimulationwasperformedat a cyclelength of 400msfromanelectrodealongthe loweredge (S). The sudden wideningof isochronecontoursin the upperleft corner ofthe mapprobablyresultedfromrapidsubepicardial spreadof acti-vation thatactivated several sites in that region nearly simultaneously. Thelines with arrowheadsindicate the directions oflongitudinal(L) andtransverse(T)propagation (confirmedbyhistological

examina-tion),andwereusedtoselect theelectrodesites forestimating con-ductionvelocity. Effectivelongitudinalconductionvelocityobtained by methodA(OAL)wascalculated betweensites U andC,OAL=100 m/s. Note that this valueOf OAL is greater than would have been ob-tained if sites U andA(6 =0.83 m/s),orsites U and E(6 =0.94 m/s) had been used. Logically,OAL should be bestestimated by the mostrapidvalue normaltotheactivation front because lower values may reflectsomecomponent of transverseconduction.Longitudinal conductionvelocityestimated by methodB

(OBL)

is themeanof

val-uescomputedatsites D,I,] andP( 1.1 1, 1.1 1,0.97,and1.27 m/s,

respectively).Inthis example,6BL= 1.12 m/s is greater than6AL.This may reflect the fact that sitesU andCare notprecisely inthe longi-tudinaldirection. Effectivetransverseconductionvelocity bymethod

A(6AT) wascalculated usingsites T andG

_(OAT

= 0.25m/s).This valueisslightly lower than values that would have been obtained if Hand B(6 = 0.27 m/s)orSand F(6 =0.26 m/s)had been used. Inthe absenceofimpedimentstotransversepropagation,

6AT

should beestimatedbythelowestvaluenormal totheactivation frontand

perpendiculartofiber orientation becausemorerapidvalues may have elementsoflongitudinalconduction. Effectivetransverse

con-ductionvelocitybymethod B(OBT)wasobtainedatsites Gand N

Results

Patients.

The control group consisted of

eight

males and

six

females with a mean age (±standard deviation) of 28±10 yr

(range

18-45) who underwent operation to interrupt an

acces-sory atrioventricular connection. All had normal

cardiac

func-tion

as

assessed

by history,

physical

examination,

and

echocar-diography.

The

patients

with IDC

consisted of 10 males and 5

females with

a mean

age

of

40±14 yr and a range 21-61 yr. All

were

undergoing cardiac

transplantation for

severe

congestive

heart

failure

but all were

being managed

as

outpatients

just

before the time of

transplantation.

The

mean

ejection

fraction

was

0.14±0.05

(0.10-0.23).

Determination

of

activation.

In

the presence

of

multiple

low

amplitude

deflections,

the test for activation

provided

a

consistent method

of

assigning

local

activation.

The test

vari-able

with

the greatest

ability

to

discriminate

local and nonlocal

signals was

a

linear combination variable ,

=

03/030

+

0.0037430,

where

43

=

the second

derivative ofthe deflection

and

O30

=

the

peak-to-nadir voltage

of the deflection within

±30 ms

of

TD

(53).

The

area

under the ROC

curve

for

t

_(0.96)

was

significantly

greater

than the areas

of

the next

best

vari-ables:

+3(0.93,

P

= <

0.001 )

and

gmin

(0.88,

P

<

0.001 )

(Fig.

3

A). In

addition,

the

maximum

QE for

V

(0.82)

was

significantly

greater than those

for

43

(QE

=

0.70,

P=

0.025)

and

_.min

(QE

=

0.63,

P

=

0.005).

Plots

of QE,

QSE,

and

Qsp

as a

function

off

are

shown in

Fig.

3

B.

By

design,

the threshold value

of

A/

used

to

distinguish deflections

resulting from

local activation

was

that value

(TE) which

maximized

QE.

Thus, deflections

with

values

of

A

2 0.093

were

considered

to

be due

to

local

activa-tion and

deflections with

lesser

values

to

be due

to

nonlocal

activity.

For

this

test

QE

=

0.82, QSE

=

0.70,

Qsp

=

1.00. The

conventional

statistics

were:

sensitivity

=

0.89, specificity

=

1.00,

predictive

value

of

a

positive

test

=

1.00,

predictive

value

of

a

negative

test=

0.79.

Application of

the

test

for

acti-vation

is

illustrated

in

Fig.

4

for

an

electrogram with

two

deflec-tions which

have

nearly

equal minimum

(greatest

negative)

first

derivative

(4amin)

values

(Fig.

4

A), for

an

electrogram

with

no

deflections meeting

criteria

for

activation

(Fig.

4

B),

and

for

an

electrogram with

more

than

one

deflection

meeting

criteria

for

activation

(Fig.

4

D).

Activation

patterns

during

constant

cycle length

stimula-tion.

Spread

of

excitation

in

control

subjects

tended

to

be

ellip-tical

with

relatively concentric

isochrone

contours.

Activation

patterns

from

an

18-yr-old

woman

undergoing operation

for

the

Wolff-Parkinson-White syndrome

are

shown in

Fig.

5.

Rapid

spread

of activation

was

present where fiber orientation

is

perpendicular

to

the septum. The

directions

of

rapid

and

slow

propagation

remained

the

same as

the site

of stimulation

was

changed.

Electrograms recorded

from

the

control

patients

were usually smooth with large intrinsic deflections

and a

domi-nant

minimum in the first derivative

plot

(Fig. 5,

E

and

F).

These

features

are

consistent with

uniform

anisotropic

propaga-tion

and

were observed in

12

out

of the 14 control subjects

(86%).

In

these

12

patients OAL

=

0.80±0.08

m/s,

OAT

=

_{0.23±0.03 m/s and}

_OAL/0AT

=

_{3.5±0.6. The second method}

of estimating conduction velocity produced similar values: OBL

=

0.79±0.1

1

m/s, OBT

=

0.23±0.04 m/s and

0BL/OBT=

3.3±0.3.

Two

of the control patients had moderately disturbed

activa-tion patterns similar to those present in group II IDC patients

described

below.

The

patients

with IDC were extremely heterogeneous with

respect

to

the patterns of activation that were recorded. They

A 'l

0.8-.- 0.6-(W 0.4- 0.2-0

-B

1 0.8 -0.6 --a 0 0.4 -0.2 -O

-rP<O.001l

I'l-

0

E-o0. 2

=8

0.

I

Ir

if o

1.0 ~.7 0.6

v

₄₎₀₃

_ol

I I *

0.2 0.4 0.6

Specificity 0.8 -_ ----GE -- _asp (SE

It

1~4 I I

¾

".

%.^ As \v=0.093 ".

,.

QE=0.82

SE 0--7

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 'v

0.5

Figure3. (A)ROC curvesforthe threecandidate variables for tests

for activation withthe greatest power todiscriminatelocalfrom dis-tantelectrical activity:+=

q3/030

+

_0.0037030,

the secondderivative

(43),

and theminimum first derivative

(+m).

The areaunder the

ROC curve, a measureoftestperformance, was significantly higher

for

_{0 (inset}

bar

graph). (B)

Plotof

t

versusthe

"quality"

statistics:

thequalityofefficiency(QE), the quality of specificity

(Qsp),

and the

qualityofsensitivity (QsE).The value of s which maximized QE was

TE

=0.093

(dashed

vertical

line).

could be

divided

into three groups. Group

I

consisted

of

7

out

of

the 15

patients

with IDC (47%)

with

uniform

anisotropic

propagation, i.e.,

with smooth electrograms and

few

irregulari-ties in the

activation

patterns

(Fig.

2). Although there was a

slight tendency

toward greater

EL

and

higher

0L/0T

in these IDC

patients,

the

propagation velocities did

not

differ significantly

from

the values in control

patients:

OAL

=

0.84±0.09

m/s, OAT

=

0.23±0.05

m/s

and

kAL/OAT

=

_3.7±0.7

(OBL

=

_0.83±0.08

m/s,

OBT

=

_{0.23±0.04 m/s, OBL/OBT}

=

3.7±0.7). Therefore,

propagation

in group

I

were

consistent

with pattern A

(i.e.,

no

difference from

controls) described in Methods. While it could

be

argued

that a

tendency

for pattern C2

was

observed,

i.e.,

increased AL and

0L/0T

with unchanged

ET,

patterns

B, D, E,

and F were

clearly absent.

Group

II

comprised

four

patients

with IDC (27%) who had

moderately disturbed activation patterns. An example from a

35-yr-old man

with IDC is shown in Fig. 6 A. Regions of dense

isochrome spacing

are

present in the

right

lower and in the

left

upper corners of the map.

It

is

not

possible

to

determine

whether the areas

of

dense

isochrone contours resulted from

slow

conduction

or

from conduction block.

It

is

evident,

how-1

B

(mV )\

-10

0.51

*

(mV/msoc2) .i

-0.5

E /

iiO-

/

T

_4e

_{_0} 20 40 60 80 100

(Msoc)

Dl D2

a:0.141 0.083 TD: I1 73

Dl D2 I: 0.067 0.076

TD: 64 79

DI D2 IY: 0.200 0.077 TD: 20 71

D 0D2 D3 W,: 0.115 0.077 0.122 TD: 23 63 93

Figure 4.Polyphasicrecordingsobtained fromapatientwith IDC. The toprowoftracingsis theunipolarextracellularpotential (4'),the second rowis thefirst derivativeof X

(4'),

andthethirdrowis the second derivative of 4

(4).

Thefourthrowisa

bipolar

signalobtainedbysubtracting theunipolar signalshown in the top row, from theunipolar signalof itsneighbor.This isequivalenttothesignalobtained fromapairof elec-trodes 2mmapart. Thedeflectionsarelabeledin theplotof 4 (Dl, D2, etc.)wheretheyare mosteasily recognized.For eachdeflection,the valueof the linear combination variable(4)and thetime of thedeflection (TD)with respecttothestimulus artifact(SA)aregivenin the table below each column ofplots.Anasterisk (*) indicates thedeflectionin each

electrogram

with the greatest value of 4 which exceeds the threshold TE=0.093,adagger

(i)

indicatesotherdeflectionsthat exceed TE.(A)ThetwoprominentdeflectionsDland D2 havenearly equalvalues of the minimumfirstderivative

(4min),

butDl hasagreaterpeak-to-nadirsecondderivative value within ±3msec(4').Because thetestvariable

(4'

=

43/030

+

0.0037030)

is,in essence,¢3normalized

by

the

voltage

of the

signal,

its value forDIisgreaterthan for D2.Thus,Dl meets criteria for activation whereas D2 doesnot. Themore

prominent bipolar

deflection

corresponding

toDl supports this.

(B)

Noneof the deflections in

thissignalachieve the thresholdfor4: Local activation is considered absentatthis site.(C) Dl meetsactivation criteriadespitealow 4 and lowerGaminthan D2.Thebipolarsignalsupports this choice.(D)Thechangein 4 associatedwithD3is obscuredbythe

rising

phaseofalarge, slowpotentialcausedbydistantelectrical

activity,

butthe

sharp

4 and

relatively

largeX reveal itshigh frequencycontentandaccountfor the relatively large value of

4'.

Dlalsomeetsactivation criteriathus,this site would beconsideredtohavetwoactivation times.

Multiple

activation timesat asinglesite have beenreported

previously

(49, 50).

ever, that the areas

with activation delays

do not consist

en-tirely of nonconducting tissue.

In

the

upper

left corner of

Fig. 6 A,

for

instance, propagation through

the

sites

labeled 1, 2, and 3 appears

normal. However,

stimulation

from

the lower edgeof

the plaque resulted in delayed activation through this

same area

(Fig.

6 B). The

direction

of propagation

also

affected

the

configuration

of

the electrograms.

Fig.

6 C shows X and4 re-corded

during longitudinal

propagation at the sites labeled 1, 2, and3 in

Fig.

6 A.The

intrinsic deflections

of the electrograms aresmooth and produce derivatives with a single prominent

deflection. The amplitudes of

4 and

_G.min

are similar to those

from

control

subjects

(Fig. 5). The amplitudes of4 and

_'mnin

during

transverse

propagation

at the same sites (Fig. 6 D),

however,

are much smallerand have multiple low amplitude

deflections(fractionation) (50, 61 ). This combination of

fea-tures is consistent with nonuniform anisotropic properties (pattern E).

The

effective

longitudinal conduction

velocities in the four IDC patients with moderately disturbed activation patterns tended to be high

(OAL

= 0.90±0.09 m/s, OBL = 0.88±0.07

m/s),

but were not significantly higher than control subjects with uniform activation patterns. The markedalterations in transversepropagation precluded reliable estimation of

trans-verseconduction velocity.

The

four patients with

IDC in group III had "severely" disturbed

activation

patterns. Group III patients differed from group IIpatients in that both transverse and longitudinal

-1oJ

I

nom. ..

(msuc) 0 20 40 60 80 100

Figure 5. (A-D) Electricalactivationsequence maps based onrecordingsfromacontrol subject; sameformatas Fig. 2.S=siteof stimulation;

isochronecontours are drawn at 2-msec intervals. The labeled electrodesitescorrespond to the electrograms shown in panels E and F. (E) Extracellularpotential

(4))

andfirst temporal derivative

(4)

recorded at the electrode site marked E in panel Cduringpropagation in the

longi-tudinaldirection. 4 consists ofalarge, smoothintrinsic deflection preceded byasmall R wave. Thepeaktopeakamplitude is38mV. Thefirst

derivativeconsists ofasingle predominantminimum

(4min

= _{-8.3 mV/msec). (F)Recording obtainedwhen the electrode site was in the}

transversedirection(panel D). Thepeak-to-nadiramplitude of 4 is 33 mV,

4)min

= -3.1 mV/msec. The changes in extracellular waveforms

during propagationindifferent directionswereconsistentwithprevious descriptionsin otherspecies(35). SA, stimulus artifact.

to

have

larger

differences ofactivation times

between

neighbor-ing

recording sites. Isochrone

maps

based

on

recordings

from

a group III

patients is shown

in

Fig.

7, samples of electrograms

upon

which the

maps are

based

are

shown in

Figs.

4,

8,

and 9.

The

maps are

characterized by

regions

of closely spaced

isochrone

contours

resulting

from large

differences

in

activa-tion times

between

adjacent electrode sites (e.g., 73

ms be-tween

sites C

and D

in

Fig.

7A,

which correspond

to the

elec-trograms

in Fig.

4,

C and

D,

and

1 7 ms

between

sites

Aand B

in

Fig.

7C,

which correspond

tothe

electrograms in

Fig. 9, A

and

B). The

activation path that

accounts

for the

large

gra-dients in

activation

times

cannot

be

determined with precision.

However,

based

on

previous

work

(56), the

presence

of

near

simultaneous activation times distal

to

the band

of

dense

isochrones (i.e.,

thelower

halves

ofthe

maps

in Fig.

7, A and C) suggest

that

activation spreads

to

this

area

from

an

intramural

route. On the other hand, very slow

conduction

through

the

regions of

dense isochrones cannot be

excluded.

There are also

areaslacking evidence of local

electrical activity,

e.g.,

site

B in Fig. 7 A, which

corresponds

tothe

electrogram in Fig.

4 B.

The

test or

criteria used

to

define

activation

can

obviously

affect the pattern of activation

especially

in the presence of multiple low amplitude

deflections

(53,

61 ).

The test

for

acti-vation used

to

select activation times

was

established

before activation maps were constructed so that the

choices for

activa-tion times would not be

influenced by preconceived

notions

about activation

sequences.

However,

it

is

important

to assess theimpact of a particular test for

activation

onthe observed

activation

patterns

which

can be done

by

adjusting

the test threshold T.

Reducing

rto0.070 results inamoresensitive but

less

specific test for activation with

_Qsp

=

0.50,

QSE

=

0.78,

QE

=

_{0.61, conventional sensitivity}

=

0.95, and conventional

speci-ficity = 0.60

(compare

tostatistics forr =0.093

given

above).

The effect of

using this

more

sensitive,

less

specific

test for

c

1

0-

-10-(mV)

-3

0-

.10-D

1

2

3

s s~~~

DI l2 31 DI23 D2 D3

Af: 0.089 0.182 0.156 Af: 0.123 0.230 0.123 TD: 23 36 46 To: 22 30 36

A

0-Jl

(mv)

-120-

A/

(mV) I0

-30-A

201

0

/

DD

(mV/mswc)

5_J * TD: 48

Time

(msec) 0 20 40 60 80 100

Figure 6. Electrical activationsequencesfroma35-yr-oldmanwith IDC. The interval between isochronesis 2msec.The orientation of thearray

is shown in thediagram between thetwomaps.(A) Isochronemapduring stimulationatthe leftedge of thearray(S) shows rapidactivation paralleltothelongaxesofmyocardial fibers. Propagationtransversetofiber orientation isabnormal,withareasofdelayedactivation(dense isochrones) especiallyinthe lower leftlowercorner.Propagationacrosssites labeled 1, 2,and 3appearsnormal.(B)Isochronemapobtained during stimulationatthe loweredgeof thearray.Thedirections ofrapidand slowpropagationareunchangedfrom themapinpanel A,butthe regions with slow activation have changed.Activationacrosssites labeled1, 2,and 3nowappearsabnormallyslow.(C) Unipolar potentials (4,

toprow) and first derivative plots (c,bottomrow)recordedatsites labeled1, 2,and 3inpanelA.Thevalues of and the times of the deflections

( TD, msec)aregiven in the insets. (D) andX plots recordedatsites1, 2,and3 inpanelB. Incontrast tothesignalsrecordedduring longitudinal propagationatthesamelocations(panel C), the signals recorded duringtransversepropagationexhibitmultipledeflections with lower ampli-tudes. Thetwodeflectionsinelectrode2(DI and D2) bothmeetcriteria foractivation, i.e.,both have values of4/ greater than thetestthreshold

(TE=0.093). This_suggestsasynchronous activation in adjacent muscle bundles (50). By convention, the isochronemaps weredrawnaccording

tothe deflection with thegreatestvalueof4A. Alteringthe choice of the deflectionrepresentinglocal activationwould alter the location of the densely spaced isochrones;e.g.,ifD3ofelectrode2 inpanelDhad beenchosentorepresent localactivation,thegroupof dense isochrones would switch from between electrodes 1 and 2tobetweenelectrodes 2 and 3. This effect has been illustratedpreviously byDillonetal.(61 ), but,as theyhavediscussed, switchingthe deflectiondesignatedaslocal activation doesnotalter thefinding of disturbedtransversepropagation.

multiple activation times. However, the large transitions in ac-tivationtimes and theseverity of the conduction disturbances arenotaffected.

Effects of

premature

stimulation

on activation sequence. The effects ofprematurebeats onactivationsequenceare of interest becausespontaneousprematurebeatsmayinitiate

ar-rhythmias by altering myocardial propagation (43, 61, 62). Eight ventricular stimuliweredeliveredatacycle length of400 or500msfollowedbyuptothreeprematurestimuliwith cou-pling intervals within 10 msof the refractory period of the

preceding beat. In the four control subjects in whom this

proto-colwasperformed, slight changes in activationpatternswere sometimes observed(Fig. 10), butatmostsites the activation times did notchange by more than a fewmilliseconds, and longitudinal andtransverseconductionvelocities changed only slightly. Also, the total activation time (TAT), definedasthe latest activation time minus the earliest activation time, changed littlewith prematurestimuli. Thissuggeststhat with

the stimulusstrength used (twice diastolic threshold current)

the prematurebeats couldnotbe deliveredearly enoughinthe cell cycle to severely limit sodium channel current. The

re-sponsesof OALand TATtoprematurestimuliare plottedfor

2

3

Figure 7.Electricalactivationmapsfroma35-yr-old woman with IDC. Theformatis thesame asin Fig.2 except that the interval between contoursis 4 ms. Thestippledregionsinclude electrodelocationsatwhichnodeflectionmetcriteria forlocalactivation.(A) Activation sequence

duringstimulation fromthe upperedge(S).The presence ofactivationwasbasedonthetestthresholdTE=0.093.ElectrogramsfromsitesA, B,C, and D are shown inFig. 4, A,B, C, and D,respectively.Electrograms from sites E, F, G, and Hareshown inFig.8, E, F, G, andH,

re-spectively.(B) Map basedonthesameelectrogramsasthe map inpanelAbutwith presence of local activation defined by r =0.070. (C) Activation sequence during stimulation from therightedge(S).The presenceofactivationwasbasedonthetestthreshold TE=0.093. Electro-gramsfrom sites A, B, C, and D are shown inFig.9, A, B, C, and D,respectively.(D) Map basedonthesameelectrograms as the map in panel C butwith presence of local activation definedbyr =0.070.

individual

patients

in

Fig.

11.

There was a consistent tendency

for reduction of

OAL

in the

patients with IDC with

V2and V4,

although

V3tended to result

in

less

depression of

0AL.

These

findings

were not changed when method B was used to esti-mate

0L.

Forreasons

discussed

earlier,

OT

could not be reliably

estimated

in many

of

the

patients with

IDC. TAT could be compared in all

patients

in whom

premature

stimuli were

ad-ministered

including

two

patients

with IDC in whom

activa-tion

was too

distorted

to

determine

6L.

There

was a

clear

in-crease in TAT in

some

patients with IDC, again suggesting

heterogeneity

in

the

response

to

premature

stimulation in

pa-tients with this disorder.

Althoughthe averageeffects of premature stimulation were not dramatic in the patients with IDC as a group,

striking

changes were noted in some individuals. The maps in Fig. 12 were obtained

with

a trainof

eight

stimuli

(S1)

at acycle length of 500 ms followed by three premature stimuli (S2,

S3,

S4) at intervals of 260, 230, and 210 ms. S4

did

notcapture the ventri-cles but was followed by two ectopic beats (V4 and

15).

The clearest indication of this is the reversal of

activation

order at

sites

A,

B,

andC in

Fig.

12

(the electrograms recorded

atthese sites are shown in

Fig. 13). Although site of origin

of the ec-topic beats cannot be precisely

identified,

twofeatures suggest

ok

(mY] X

.10i

01

(mV/mspc

-0.

1

S\ti1i/

(mV/m"c2) .

TlIme._0

(R)Dl

20 40 60 80 100

D2 _D2

-0.5- ~ ~ V 0.6 .7 .7

Tiff*TO25.004 60 7310

Dl )- _D

D2

*u

D1 D2 D3

W: 0.096 0.100 0.077 TD: 24 63 75

D1 D2 D3

\V: 0.053 0.105 0.080 TD: 22 64 77

DI D2 D3

V: 0.094 0.098 0.068 TD: 65 78 88

Figure 8. Electrogramsobtained fromthe electrodesiteslabeled E,F, G,and H in Fig. 7, A and B. Sameformat as Fig. 4. (E) Activation at this

siteoccursearly (DI at 25 msec) butdeflectionsdue todistantactivityareapparent at60 and 73 msec. (F) Twodeflections(DI and D2) meet criteriafor activationwithmarginalvaluesof

#.

Thissuggests that the electrode straddles two or morepoorly coupled myocardial cell bundles. (G and H) Thedeflections meetingactivation criteria occurrelativelylate inthese leads demonstrating aprogression ofactivation times across the

region ofdenseisochronespresent inFig.7A.

array.

The

first

is that

during

the

ectopic

beats, sites

depicted

in

the

upper

right corner and

center

of the

map were

activated

early compared

to

sites in the lower half and the

left

upper corner

which

were

activated

much later.

If the

site of origin

were

distant from

the plaque array, apattern closer to that

observed

during sinus

rhythm

(Fig.

12

F) would be expected,

that

is,

apattern

with much less

dispersion

of

activation times.

In

fact during

the

sinus beat, where activation of

thetissue

beneath the

plaque presumably

spread from the endocardium,

central

sites

were

activated

later than

peripheral

sites. Second,

it has been shown that

a

short

time interval

between the onset

of

the

root mean square

(RMS) voltage

curve

and the

site

of

earliest activation (5

ms

for

V4 and

V5)

indicates

proximity of

the

origin

of

electrical activity

and the

site of

earliest recorded

activation (63).

In contrast,

earliest

recorded activation

oc-curred

44 ms

after

the onset

of the RMS voltage for

a

sinus

beat

(Fig.

12 F).

Of additional interest is

the markedconduction delay that

developed with

premature beats.

During

a

sinus

beat, the

re-gion in the left

upper

corner of

the maps in

Fig.

12 is activated

early

compared to most recorded

sites

(Fig. 12 F). During

_VI,

activation in the left

upper corner

is

relatively late,

e.g., 47 ms at

site

F

(Fig.

12

A).

With each successive premature beat the

activation time of this region increases,

e.g., at

site

F 74 ms

during

V2,

119

ms

during

V3,

166

ms

during

V4,

and

176

ms

during

V5.

Although V4 and

V5 are

ectopic beats, activation

of the centralsitesoccurs

early,

and

only

about 10ms

of

the

delay

observed

at

site

F

could be

attributed

to

the change in

activa-tion

sequence

associated with the ectopic beats.

Of

further

possible significance

is the

finding

that

large

gra-dients

of activation

time occurred

over very

short distances.

Fig.

12

E

shows, for

instance,

a

difference

in

activation

times of

164

ms

between

sites

D

and

E,

which

were

only

2.8

mmapart. The

electrograms

upon

which

these

activation times

were basedareshown in

Fig.

14. In

particular,

the

electrograms

from

site

E

(Fig.

14

E)

demonstrate

progressive activation delays

between

primary deflections (those with

the

highest

valuesof

41)

and

secondary deflections (those which

meet

activation

cri-teria

but have

lesser values

of A1). Again,

the presence

of

multi-ple deflections which

meet

activation

criteria suggests that the

recording site straddled multiple

muscle bundles. If true, these

be-SA SA

±A

0

-0.5 ~~~

~~D3

D

(MV/RMC)D2D

'a'J

D1D

f

/ / D2* D2

D12

D3

D2

-0.5

0-1

I:0290050.061 IA:0.0910.141I I4:0.0830.069I AV:0.061 0.1020.085I

Lj: 80125.. TO: 80125 TO 1 124Oj4J69

Figure 9.Electrogramsobtainedfromthe electrodesites labeledA, B, C,and D in

Fig.

7,C and D. Sameformatas

Fig.

4.(A)Thedeflection

labeledDl has a valueof 4'well above thresholdfor activationanda

prominent

bipolar signal.

Laterdeflectionsarepresentat80 and 125

msec,but do notattain activation criteriaanddonotproduceadiscernable

bipolar

deflection

indicating

that D2and D3 areprobablydue to

distantactivity. (B) This electrogramrecordedfrom siteBin

Fig.

7Cincludesadeflectionthatmeetsactivation criteria thatoccurs

remarkably

latecomparedtotheactivationofthe

neighboring

siteA.

(C)

Noneofthedeflectionsmeetcriteria foractivation in this

electrogram

withthe thresholdT=0.093. Lackof activationatthis sitesuggests poorconductionandcouldindicatean

insulating

effect forthetissue

responsible

for

the very lateactivationtimeatsite B.(D)Thedeflection

meeting

activationcriteria

(D2)

occursmuch later thanactivationatsite A,butisearlier

thanthat atsiteBdespite itsmoredistal location fromthestimulationsite. This indicates

delayed

conductionbetween sites D and B and the relatively protected statusofviable tissueatsiteB.

tween

bundles

of myocardial fibers

muscle

bundles

<1 mm

apart.

It

is

noteworthy

that

these conduction

delays

were

pro-duced

by

premature

beats with

coupling

intervals

similar

to

those

which

had minimal effects

on

conduction

velocities

and

activation

patterns

in controls and IDC

patients

with uniform

anisotropic properties. Finally,

the

association between

ectopic

beats

induced by ventricular extrastimuli

and

the

development

of

large

activation

delays

raises the

possibility

that the

induced

ectopic

beats resulted

from

reentrant

excitation.

NAL The NAI

is

measure

of

the

spatial heterogeneity

of

activation

times,

and has

been usedto

quantify

the

degree

of

nonuniformity

of activation

(56).

The

patients

with IDC had

significantly

greater NAIs

(9.1±6.1

ms/mm)

than the control

patients

(5.5±1.4 ms/mm,

P=

_0.025).

The

degree

of

distur-bance

of the activation

patterns

corresponded

with the

NAIs: group I

had

amean NAI

of 6.0±1.3 ms/mm,

group II

had

a

mean NAI

of 9.6±2.2 ms/mm,

and group III

had

a meanNAI

of 13.8±10.7 ms/mm.

Histology. Myocardial

tissue

from

beneath the

recording

plaque

was

examined in 13 of the

15

patients

with IDC.

The mean

percentage of fibrous tissue

was

13.5±15.8% which is

similar

to

that

reported

previously

(

10,

15,

18, 19, 20, 23).

Patients in

group I

had

minimal fibrosis

with

a mean

fraction

of

only

4.6±5.6%

(Fig.

15,

top),

group II

patients

had

interme-diate

amounts

of

fibrosis:

9.5±7.8%

(Fig.

15,

middle),

and

pa-tients in group III

had

very

prominent fibrosis: 28.3±21% (Fig.

15,

bottom). There

was a

significant

correlation between the

NAI

and the

fraction of

fibrosis

(r

=

0.57,

P<

0.05).

More-over,

the

pattern

of

fibrosis in the

groups

with mild

and

moder-ate

disturbances of activation

patterns was

interstitial,

that

is,

primarily linear accumulations of collagen separating

bundles of myocytes with

little alteration

in

the

alignment

of

myocytes orbundles. In contrast,

replacement

fibrosis

and

microscopic

scars were

prominent

in

group III

patients

which often

was

accompanied by

deviation

or

interruption

of the

parallel

ar-rangement

of

epicardial

muscle

fibers

and bundles

(

16).

Relation between

electrophysiologicfindings

and

spontane-ous

arrhythmias.

12

of the control

patients

underwent at least

24h

of

electrocardiographic monitoring

(mean

= 63±33

h);

no

ventricular arrhythmias

were

detected in nine of these

pa-tients.

Intwo

of

the

control

patients

rare

pairs

of

premature

con-Figure 10. Effects of prematurestimulation in a 35-yr-old male control subject. (A) An electrogram from the recording site in the upper left

cornerofthe array shows thelasttwobeatsinthe

eight

beattrain

_(VI7),

andthefirst

(V2),

second

(V3),

and third

(V4)

prematurestimuli. The cyclelength of the train of stimuliwas500 msec, and thecouplingintervals of V2, V3 and V4were250, 210, and 210 msec, respectively. (B) Map ofV,.Theformat is thesame asin Fig.2.Stimulation isperformed fromtheleft edge(S),theinterval between isochrones is 2 msec. The arrowsindicate the directions oflongitudinal(L) andtransverse (T)propagation. Theelectrodesites indicated by the heads and tails of the arrows wereused toestimatelongitudinalandtransverseconductionvelocities by methodA

(6AL,

OAT) The electrodesites indicated were used toestimate longitudinal

(.)

andtransverse(m) conduction velocitiesby methodB

(OBL,

OBT ).Inthisexample,

6AL

= OBL=0.83m/sec and OAT

=

OBT

= 0.25

m/sec.

The TAT is 49msec.

(C) Map

of

V2:

6AL= 0.90m/sec,

OBL

=0.80m/sec, OAT= OBT= 0.22m/sec,TAT =45 msec.

(D)

Mapof V3:

6AL

= _{0.90 m/sec,}

6BL

=_{0.87 m/sec, OAT}=0.27 m/sec, OBT=0.25m/sec,TAT=49 msec. (E) MapofV4:

6AL

=0.90m/sec,OBL

=0.86 m/sec,OAT=0.26 m/sec OBT=0.25 m/sec, TAT=47 msec.

trol

subject

had

frequent ventricular

arrhythmias

(265

PVCs/

h,

0.1

pairs/h).

Interestingly,

this

was one

of

twocontrol

pa-tients with moderately altered activation.

10

of the 15

patients

with

IDC underwent 24-h Holter

monitoring, five in

group

I,

four in

group

II, and

one

in

group III. Because

of

the small

numbers, groups II and III were

combined.

The PVC

fre-quency

in the combined

group was

478±678 PVCs/h

(median

=

121 PVCs/h)

compared

with

group I

patients

who had a mean

of 123±240 PVCs/h (median:

21 PVCs/h) (P=

0.1).

The

frequency of PVC pairs in

the

combined

group was 25

pairs/h (median

=2

_{pairs/h) compared}

to16±25

pairs/h

(me-dian

= 0.3

pairs/h)

(P=0.26). The

frequency of unsustained

ventricular tachycardia (VT) in the combined group was

8±11

/h (median

=

6.2/h)

compared to

only

0.8±1.7/h

(me-dian

=0) in the group I

patients

(P=0.17). In

addition,

all

of

the

patients in

the

combined

group had PVC

frequencies

> 30/ hand all but one had unsustained VT during the 24-h Holter

recording,

whereas only one of the group I patients had > 30

PVCs/h, and only

two

had

an

episode of unsustained

VT.

Thus, patients

in groups II and III tended to have greater

fre-quencies

of each category of

arrhythmia,

butnoneof the differ-ences

attained statistical

significance.

On the other

hand,

it

should be

recognized that there

were

relatively

few

patients

in

each

group,

the

variances

were

high, and there

was nodata

from

three

of

the

four

group III

patients who might have

been

expected

tohave

the

most severe

ventricular arrhythmias. Only

one

of the

seven group I

patients had

a

history

of sustained

ventricular

tachycardia (lasting

atleast I min or

requiring

car-dioversion),

which occurred

during

an

infusion

of

dopexa-mine,

a

f3-adrenergic

and

dopaminergic

receptor

agonist.

In contrast, three of the

eight patients

in groups IIor III hada

history of sustained

VT.

Discussion

Propagation in patients

with IDC.

Inseven (47%)

of

the

pa-tients

with IDC

(group I)

the pattern

of activation

and the