-
Br Heart J 1992;68:567-73
Relation of left ventricular isovolumic relaxationtime and
incoordination to transmitral Dopplerfilling patterns
Stephen J D Brecker, Chiang H Lee, Derek G Gibson
AbstractObjective-To investigate factors dur-
ing isovolumic relaxation that determineDoppler filling patterns
in patients withleft ventricular disease, and thus toidentify the
underlying mechanisms.Design-85 patients (50 ischaemic
heart disease, 35 left ventricular hyper-trophy due to aortic
stenosis) and 26controls were studied with Doppler andM mode
echocardiography andphonocardiography. 16 patients under-went two
studies on separate occasions,to find whether changes in
isovolumicrelaxation time were reflected by achange in the Doppler
A/E ratio.Setting-A tertiary cardiac
referralcentre.Subjects-Patients referred for assess-
ment of coronary artery disease or aorticstenosis with left
ventricular hyper-trophy.Main outcomes measures-Doppler
filling velocities during early (E wave)and late (A wave)
diastole and the A/Eratio, acceleration of the E wave, digit-ised M
mode indices of incoordinaterelaxation (change in cavity
dimensionbefore mitral valve opening and timefrom minimum dimension
to mitralvalve opening), isovolumic relaxationtime, M mode measures
of diastolicfunction after mitral valve opening(peak rate of
posterior wall thinning andpeak rate of dimension increase),
andleft ventricular end diastolic pressure.Results-AIE correlated
with age in
normal subjects (r = 0 74), to a lesserextent in left
ventricular hypertrophy(r = 0-41), but not significantly
inischaemic heart disease. In all patients,isovolumic relaxation
time was signi-ficantly and negatively correlated withthe
acceleration of the E wave, showingits fundamental relation to the
force re-sponsible for early diastolic filling(r = - 0-71 for left
ventricular hyper-trophy, and -074 for ischaemic heartdisease, p
value < 0 01). In leftventricular hypertrophy and thoseischaemic
patients without left ven-tricular dilatation A/E was
correlatedboth with isovolumic relaxation time(r = 0-68 and 0-60
respectively), andwith incoordinate relaxation (r = 065and 0-61).
In those ischaemic patientswith left ventricular dilatation,
theinfluence of incoordination was lost and
isovolumic relaxation time became thedominant influence upon
A/E(r = 082). Weak correlations of enddiastolic pressure and RR
interval withA/E, became insignificant onceisovolumic relaxation
time had beentaken into account. Isovolumic relaxa-tion time and
incoordination togetheraccounted for over 50% of the variancein the
A/E ratio in our patients.Isovolumic relaxation time and the
A/Eratio were linearly related. Patients witha short isovolumic
relaxation time hadevidence of considerable diastolic
abnor-malities, despite a normal Doppler A/Eratio. In the 16
patients who had twoechocardiographic studies, changes inthe
duration of isovolumic relaxationwere accompanied by a change in
theDoppler A/E ratio. The relation betweenthese two variables,
derived from thegroup as a whole was similar.Conclusions-The main
factors
influencing the A/E ratio in patients withleft ventricular
disease are two distinctproperties of isovolumic relaxation-namely
the duration and the extent ofincoordinate wall motion. Filling
pres-sure and RR interval are not significantindependent
determinants, but act onlythrough an effect upon isovolumic
relaxa-tion time. Age is an important influencein normal people,
but this effect is atten-uated in left ventricular hypertrophyand
lost in ischaemic ventricular disease.
(Br Heart J 1992;68:567-73)
The pattern ofdiastolic inflow into the ventriclecan be
conveniently studied with pulsedDoppler echocardiography, and the
ratio ofpeak atrial to early diastolic velocities (A/Eratio) has
been used by many as an index ofunderlying left ventricular
diastolic disease.' 2Yet the precise mechanisms responsible
forfilling are complex and incompletely under-stood. In normal
individuals, age is well recog-nised to be the primary determinant
of the A/Eratio,3-5 although in disease states it is likely
thatthis effect will be attenuated by other overridingfactors. As
blood flow velocity is a function oftransmitral pressure gradient,
relations withleft atrial or left ventricular end diastolic
pres-sure have been sought and not surprisinglyhave been found. In
different studies, however,the correlation ofthe early filling
velocity and E/A with filling pressure, has been reported to
bepositive,6` negative,9 and non-existent,'°
Cardiac Department,Royal BromptonNational Heart andLung
Hospital, SydneyStreet, LondonS J D BreckerC H LeeD G
GibsonCorrespondence toDr Derek G Gibson, CardiacDepartment,
RoyalBrompton National Heartand Lung Hospital, SydneyStreet, London
SW3 6NP.Accepted for publication22 June 1992
567
on June 6, 2021 by guest. Protected by copyright.
http://heart.bmj.com
/B
r Heart J: first published as 10.1136/hrt.68.12.567 on 1 D
ecember 1992. D
ownloaded from
http://heart.bmj.com/
-
Brecker, Lee, Gibson
although striking changes in filling velocitiesdue to acute
alteration of loading conditionshave been documented both in
animals,"' andpatients.213 Disturbances of the A/E ratio andpeak
rapid filling velocity have been ascribed toabnormal
relaxation,14-16 as well as to reducedatrial and ventricular
compliance,2 17 18 whereasM mode studies have shown that in left
ven-tricular hypertrophy, the A/E ratio may benormal in the
presence of obvious diastolicdisease.'9 Our study is based on
previous find-ings that the duration of isovolumic relaxationor the
rate of left ventricular pressure declinemay affect early diastolic
filling velocity inpatients with left ventricular hypertrophy'920or
ischaemic heart disease.2' It was our aim toassess the extent to
which filling velocitiesdepended on events occurring before
mitralvalve opening, to identify the factors duringisovolumic
relaxation that might be respon-sible, and thus to gain insight
into underlyingmechanisms.
Patients and MethodsSTUDY POPULATIONWe studied 50 patients (42
men, eight women,mean age (range) 59 (37-81) years) withischaemic
heart disease, 35 patients with leftventricular hypertrophy
secondary to aorticstenosis (20 men, 15 women, mean age 56 (14-82)
years), and a control group of 26 healthyindividuals (18 men, eight
women, mean age 49(20-82) years). All patients had been referredfor
echocardiographic evaluation of ventricularfunction. Cardiac
catheterisation andmeasurement of end diastolic pressure
werecarried out in 52 patients as part oftheir
routineassessment.
M MODE AND CROSS SECTIONALECHOCARDIOGRAPHYM mode and cross
sectional echocardiogramswith an Advanced Technical
LaboratoryImager Mk 300I with a 3-0 MHz mechanicaltransducer, or a
Toshiba SSH 160A imagerwith a 3-5 MHz transducer were taken with
thepatient in the standard left lateral position.Phonocardiograms
were recorded with aLeatham microphone with a low frequencyfilter.
M mode echocardiograms were recordedwith simultaneous
eletrocardiogram andphonocardiogram, on a Honeywell (Ecoline
22)strip chart recorder at a paper speed of 10 cm/s.Only records
showing clear continuous echoesfrom the septum and posterior wall
and clearmitral cusp separation were used. Aortic valveclosure (A2)
was taken as the start of the firsthigh frequency vibration of the
aortic compon-ent of the second heart sound recorded on
thephonocardiogram, and was checked for validitywith the aortic
echogram and the aortic closureartefact on the Doppler recordings.
Leftventricular internal cavity dimensions weremeasured at end
systole (A2) and end diastole(start of the QRS complex on the
electrocar-diogram) with leading edge method, from theparasternal
long axis view. Isovolumic relaxa-tion time was measured from A2 to
the initialseparation of the mitral cusps on the M modeechogram.
All measurements were made on
three cardiac cycles and the mean taken. Mmode echocardiograms
were digitised with aTerminal Display Systems TDS 20
digitisingtablet as previously described.22 Measures ofincoordinate
relaxation were taken as theincrease in transverse cavity dimension
duringisovolumic relaxation expressed as a percentageof the total
dimension change during the car-diac cycle, and the time interval
from minimumcavity dimension to mitral valve opening. Thepeak rate
of dimension increase (dD/dt), andthe peak rate ofposterior wall
thinning (dW/dt)were also derived from the digitised traces.Once
again, all measurements were made onthree cardiac cycles, and the
mean taken.
DOPPLER ECHOCARDIOGRAPHYWe recorded Doppler signals with a
DoptekSpectrascan and a 2-0 MHz transducer, and aToshiba SSH 160A
with a 3-5 MHz trans-ducer. Peak transmitral flow velocities
wereidentified by continuous wave from the apex,and were recorded
in pulsed mode with a 3 mmgate, and 250 MHz wall filter. The
peakvelocities of early E wave and atrial A wavetransmitral flow
were recorded and the A/Eratio was calculated. The time from onset
ofearly transmitral flow to the peak velocity wastaken as
acceleration time, and the peakvelocity divided by acceleration
time as earlydiastolic acceleration. The time interval ofaortic
closure to the onset of transmitral flowwas also recorded. Records
were taken withsimultaneous electrocardiogram and phono-cardiogram
at a paper speed of 10 cm/s. Allmeasurements were made on three
cardiaccycles and the mean taken.
CARDIAC CATHETERISATION52 patients from the two study groups
under-went cardiac catheterisation as part of theirroutine
assessment, from either the brachial orfemoral approach. Pressure
was recordedbefore left ventricular angiography with a fluidfilled
7 or 8 French pigtail catheter attached to amanifold micromanometer
transducer (MedexMedical), with zero taken at mid-chest level.The
left ventricular end diastolic pressure wasmeasured at the point on
the pressurewaveform corresponding to the onset of theQRS complex
on the electrocardiogram. Themean of 10 cardiac cycles was
recorded. Allstudies were performed within 48 hours of
theechocardiographic examination.
STATISTICAL ANALYSISAll values are given as mean (SD).
Differencesbetween mean values were compared byStudent's t test.
Simple and stepwise regres-sion, multiple regression, and linear
correlationwere performed to assess relation betweenvariables.
ResultsGROUP CHARACTERISTICSTable 1 summarises the description
and com-parison of the subjects. The patients withischaemic heart
disease were older than thenormal subjects, but no other
significant dif-
568
on June 6, 2021 by guest. Protected by copyright.
http://heart.bmj.com
/B
r Heart J: first published as 10.1136/hrt.68.12.567 on 1 D
ecember 1992. D
ownloaded from
http://heart.bmj.com/
-
Relation of left ventricular isovolumic relaxation time and
incoordination to transmitral Dopplerfilling patterns
Table I Characteristics of the groups (mean (SD))
Ischaemic heart Left ventricular Controlsdisease (n = 47)
hypertrophy (n = 35) (n = 26)
Age (y) 59 (9)t 56 (19) 49 (17)LVEDD (cm) 5 9 ( 1)*t 4-8 (1-0)
5-0 (0 6)RR interval (ms) 800 (200)t 835 (140)t 930 (155)A wave
velocity (m/s) 0-43 (0 29)*t 0-68 (0-3)t 0-54 (0-12)E wave velocity
(m/s) 0 71 (0 20)* 0-86 (0 31)t 0-69 (0-17)E wave acceleration
(M/S2) 11 7 (6-0) 9-3 (4-6) 11 4 (3-1)A/E ratio 0-72 (0 60) 0-87 (0
49) 0 79 (0 28)Isovolumic relaxation time 60 (35) 60 (25) 60
(10)(measured to mitral valve opening)(ms)
Aortic closure to start of 105 (40) 85 (30) 85 (10)transmitral
flow (ms)
Change in dimension before mitral 23 (16)*t 16 (9)t 5 (4)valve
opening
Minimum dimension to mitral valve 75 (45)t 70 (45)t 35
(20)opening (ms)
Peak rate of dimension increase 9-8 (3-8)t 9 7 (3 5)t 15 5 (4
8)(cm/s)
Peak rate of posterior wall 5-9 (2 9)t 6-7 (3 3)t 9-8 (2
5)thinning (cm/s)
LVEDP(mm Hg) 22(11) 21(8)
LVEDD, left ventricular end diastolic dimension; LVEDP, left
ventricular end diastolicpressure; *p < 0-05 v group with left
ventricular hypertrophy; tp < 0 05 v controls.
ferences existed with respect to age. Left ven-tricular end
diastolic dimension was greater inpatients with ischaemic heart
disease (5 9(1 1) cm) than both those with left
ventricularhypertrophy (4-8 (1 0)) and the controls (5 0(0 6)).
TheRR interval was significantly shorterin both patient groups
compared with controls(800 (200) ms and 835 (140) v 930 (155)).
Thesedifferences did not affect within group analysis.The mean
value for peak E wave velocity was
greater in patients with left ventricular hyper-trophy than in
both other groups (0-86 (0-31) v071 (0 20) and 0-69 (0-17) m/s), in
whom themean value did not differ significantly. Themean value for
peak A wave velocity was alsogreater in patients with left
ventricular hyper-trophy than in the controls (0-68 (0 3) v 0-54(0-
12) m/s) so that the A/E ratio did not differsignificantly (0-87 (0
49) v 0 79 (0 28)), alth-ough the SD was clearly higher, reflecting
thewide range of values for the A/E ratio seen inleft ventricular
hypertrophy. Although the Awave velocity was less in patients
withischaemic heart disease than normal (0-43(0 29) m/s), once
again the mean A/E ratio
didnotdiffersignificantlyfromnormal(0*72(0*60)),the high SD
similarly reflects the wide range ofvalues seen in ischaemic heart
disease. Themean value for E wave acceleration in the groupwith
left ventricular hypertrophy was not sig-nificantly lower than the
other two groups.
Table 2 Subgroup analysis of the group with ischaemic disease
(mean (SD))
Dilated cavity Non-dilated cavity(LVEDD > 6 cm) (LVEDD < 6
cm)
Age (y) 56 (9)* 62 (9)tLVEDD (cm) 6-8 (0 5) 5 0 (0 8)Shortening
fraction (%) 14-6 (4 5)* 231 (7-1)RR Interval (ms) 765 (200)t 840
(195)A wave velocity (m/s) 0-39 (0-31)t 0 47 (0 29)E wave velocity
(m/s) 0 74 (0-17) 0-66 (0 22)A/E ratio 0 57 (0-52) 0 86 (0
65)Isovolumic relaxation time (ms) 55 (30) 70 (35)Change in
dimension before mitral valve opening (%) 21 (14)t 25 (18)tMinimum
dimension to mitral valve opening (ms) 70 (40)t 80 (45)tPeak rate
of dimension increase (cm/s) 9 2 (3-1)t 10-4 (4-3)fPeak rate of
posterior wall thinning (cm/s) 5 5 (2 8)t 6 2 (3)tLVEDP (mm Hg) 22
(9) 20 (12)
LVEDD, left ventricular end diastolic dimension; LVEDP, left
ventricular end diastolicprcssure; *p < 0-05 v group with normal
cavity dimension; tp < 0 05 v normal control group.
The mean values of isovolumic relaxationtime seen in each group
were identical (60 (35),60 (25), 60 (10) ms), although the range
ofvalues seen in both patient groups was muchwider than in the
controls. The time fromaortic closure to the start offlow was
significan-tly longer than the isovolumic relaxation timemeasured
as aortic closure to the point ofmitralcusp separation in all
groups. This discrepancywas 25 ms in left ventricular hypertrophy
andnormal people, and 45 ms in those withischaemic heart disease.
This discrepancy didnot significantly differ between those
ischaemicpatients with and without incoordinate relaxa-tion. In
both ischaemia and hypertrophy,measures of incoordination during
isovolumicrelaxation were greater than in the controls,both as the
percentage change in dimensionbefore mitral valve opening, and the
timeinterval from minimum cavity dimensions tomitral valve opening.
These two measureswere, as expected, highly correlated.Peak rates
of dimension increase and pos-
terior wall thinning were significantly reducedin both patient
groups compared with controls.
WITHIN GROUP COMPARISONCavity sizeTo investigate the effect of
cavity dilatation, wedivided patients with ischaemic heart
diseaseinto two groups according to the end diastolicdimension.
Those with left ventricular dilata-tion were younger and with a
significantlyreduced shortening fraction, but there were noother
significant differences between the twogroups in terms of Doppler
orM mode indicesof diastolic function. Indices of diastolic
func-tion in both subgroups were significantly dif-ferent from
normal (table 2).
Isovolumic relaxationFigure 1 shows that the isovolumic
relaxationtime and the Doppler A/E ratio for all patientswith left
ventricular disease were linearlyrelated. When subdivided into two
groupsdepending on isovolumic relaxation time, thosewith shorter
values had significantly lessincoordination and a greater peak rate
ofdimension increase. Importantly however, bothsubgroups had M mode
indices of diastolicfunction that were significantly different
fromnormal (table 3).
+ Dilated ischaemic heart diseaseo Left ventricular hypertrophy.
Nnn-.riBitatr inn.h2amir hp-art riisqwa-
0 0 00 0 *
+ +
000 + oDC + +
0 20 40 60 80 100 120 140Isovolumic relaxation time (ms)
2-g
._ 2C
w 1.;
a) 1-t
00.5o
Figure 1 Relation between Doppler A/E ratio andisovolumic
relaxation timefor all three patient groups.The three regression
lines were so similar that they arerepresented by a single
line.
569
Z-
on June 6, 2021 by guest. Protected by copyright.
http://heart.bmj.com
/B
r Heart J: first published as 10.1136/hrt.68.12.567 on 1 D
ecember 1992. D
ownloaded from
http://heart.bmj.com/
-
Brecker, Lee, Gibson
Table 3 Subgroup analysis of all patients with left ventricular
disease, based uponisovolumic relaxation time (mean (SD))
Variable IVRT < 60ms IVRT > 60ms
Isovolumic relaxation time (ms) 40 (15) 90 (20)Change in
dimension before mitral valve opening (%) 15 (1 1)*t 27
(15)*Minimum dimension to mitral valve opening (ms) 55 (35)*t 90
(40)*Peak rate of dimension increase (cm/s) 10 8 (3 6)*t 8-7 (3
5)*Peak rate of posterior wall thinning (cm/s) 6 8 (3 3)* 5 7
(2.8)*LVEDP (mm Hg) 23 (9) 19 (9)
IVRT, isovolumic relaxation time; LVEDP, left ventricular end
diastolic pressure; *p < 0 05 vnormal control subjects; tp <
0 05 v group with IVRT > 60 ms.
Effect of age on filling patternA/E correlated with age in
normal individuals(r = 0 74), to a lesser extent in those withleft
ventricular hypertrophy (r = 0-41), butnot significantly in those
with ischaemic heartdisease (table 4).
SEPARATE EFFECTS OF ISOVOLUMIC RELAXATIONTIME AND INCOORDINATION
UPON FILLINGPATTERNIn left ventricular hypertrophy and
thoseischaemic patients without left ventriculardilatation, A/E was
correlated both withisovolumic relaxation time (r = 0-68 and0 60
respectively), and with incoordinate relax-ation (measured as a
percentage change beforemitral valve opening, r = 0 65 and
0-61,table 4). In those ischaemic patients with leftventricular
dilatation (end diastolic dimension>6 cm), the influence of
incoordination waslost and isovolumic relaxation time became
thedominant influence upon A/E (r = 0-82).This was despite
considerable incoordinaterelaxation being evident in this
subgroup.Stepwise regression confirmed that weakcorrelations of end
diastolic pressure and RRinterval with A/E became insignificant
onceisovolumic relaxation time had been taken intoaccount. Multiple
regression analysis for thegroup with ischaemic heart disease, and
thatwith left ventricular hypertrophy, showed thatisovolumic
relaxation time (IVRT) and incoor-dinate relaxation account for
over 50% of thevariance in the Doppler A/E ratio.
In patients with ischaemic heart diseases, theregression
equation was:A/E = -0-128 + 0 011 IVRT + 0 009%
change before mitral valve opening;R2 = 52%
In left ventricular hypertrophy, it was:A/E = -0095 + 0011 IVRT
+ 0017%
change before mitral valve opening;R= 60%
Table 4 Correlation coefficients of the AIE ratio
Left Ischaemic Ischaemicventricular heart disease heart
disease
Controls hypertrophy non-dilated dilated
Age 0-74** 0 41* 0-28 0-23Isovolumic relaxation time 0-04 0-68**
0 60** 0 82**Incoordinate relaxation:Change in dimension before
mitral 025 0 65** 0 61** 0-27
valve opening (%)Minimum dimension to mitral 0-15 0 72** 0 50**
0 22
valveopening (ms)
RR Interval 002 0 17 0 09 0-54**LVEDP -0 37* - 0-48* -0 2
LVEDP, left ventricular end diastolic pressure; *p < 0 05;
**p < 0 01.
40 60 80 100 120Isovolumic relaxation time (ms)
Figure 2 Results of changes in the Doppler AIE ratioand
isovolumic relaxation time in the seven patientsstudied on two
occasions, with the largest changes. Thearrows indicate the
direction of change and the broad linerepresents the regression
line derivedfrom the group as awholefrom fig 1. Note that changes
within individualpatients behave in a similar fashion to
differences betweenpatients.
EFFECT OF ISOVOLUMIC RELAXATION TIME UPONACCELERATIONIsovolumic
relaxation time correlated sig-nificantly with acceleration of
early diastolicflow in both ischaemic heart disease(r = -0 74, p
< 0-01), and left ven-tricular hypertrophy (r = -071,p <
001).
CHANGES WITHIN PATIENTSSixteen patients were studied more than
once,and in all cases, changes in the duration ofisovolumic
relaxation time were matched by aconcomitant change in the Doppler
A/E ratio.The time between the two studies was 6 (5)months. Figure
2 shows the findings in theseven patients with the greatest
changes. Somepatients had increased isovolumic relaxationtime and
Doppler A/E ratio, and conversely,those whose isovolumic relaxation
time gotshorter had a reduced A/E ratio. Figure 3shows the relation
between the magnitude ofchange in isovolumic relaxation time and
themagnitude of change in the A/E ratio for allpatients. These
striking changes occurredwithout any evidence that the underlying
dia-stolic properties of the left ventricle hadchanged (table
5).
0 100-
xX 80-a)
E g60-L-o a)
C.f 40-
ue 2Ca)D 20-
OF
0 0
0
* S
*0Ut
II0 0:5 1:0 15Change in Doppler A/E ratio
2-0
Figure 3 Relation between the size of change in theDoppler A/E
ratio, plotted against the size of change inisovolumic relaxation
time, in the 16 patients studied ontwo occasions.
570
on June 6, 2021 by guest. Protected by copyright.
http://heart.bmj.com
/B
r Heart J: first published as 10.1136/hrt.68.12.567 on 1 D
ecember 1992. D
ownloaded from
http://heart.bmj.com/
-
Relation of left ventricular isovolumic relaxation time and
incoordination to transmitral Dopplerfilling patterns
Table 5 Doppler A/E ratio, isovolumic relaxation time, and
indices of diastolicfunction in 16 patients studied on two
occasions: data are presentedfor the lower AIEratio and higher AIE
ratio studies (mean (SD))
Value when AIE ratio Value when AIE ratiowas low was high
A/E ratio 0-46 (0-53) 1 01 (0-53)Isovolumic relaxation time (ms)
40 (35)* 80 (40)Peak rate of dimension increase (cm/s) 10-6 (3 4) 9
7 (4 1)Peak rate of posterior wall thinning (cm/s) 8.8 (2-3) 7-6 (3
6)
ip < 0 05 v group with higher A/E ratio.
DiscussionThe pattern of blood flow across the mitralvalve
during diastole is not determined bythe presence or absence of left
ventriculardisease, but by simple newtonian mechanics.Acceleration
or deceleration ofblood is directlyproportional to the applied
force. A force ismeasured in fluid as a pressure gradient-thatis,
the rate of change of pressure with distancealong the direction
offlow. A complete descrip-tion ofthe pressure gradients requires
mappingof the three dimensional pressure field-thespatial
distribution of pressure throughout theatrium, atrioventricular
junction, and ventriclethroughout the filling period. Such
measure-ments are so complex that they have never beenundertaken
experimentally, let alone in intactpatients with heart disease.
Single deter-minants of left atrial or left ventricularpressure,
particularly with respect tophysiologically irrelevant reference
points suchas atmospheric or mid right atrial pressure,have given
rise to confusing reports thatattempt to relate filling pressures
to Dopplermeasurements.6 910 In the present study,therefore, we
have explored other interrela-tions with Doppler measurements of
the trans-mitral flow pattern, and shown the importanceof two
separate components of isovolumicrelaxation in determining filling
patterns.There are an increasing number of publica-
tions relating isovolumic relaxation with earlyfilling velocity,
peak filling rate, or the A/E ratioin hypertrophic
cardiomyopathy,20 coronaryartery disease,2' left ventricular
hypertrophy,'9and dilated cardiomyopathy.23 In the study ondilated
cardiomyopathy, as in this study, A2 tomitral cusp separation Was
used to measureisovolumic relaxation time. The discrepancybetween
this and the time interval of A2 to thestart of flow on Doppler has
been shown to be25 ms, on average, in controls and in
leftventricular hypertrophy, and 50 ms in dilatedcardiomyopathy.24
We have now shown that inischaemic heart disease as a whole,
thisdiscrepancy is 45 ms. Furthermore, we wereunable to show any
significant differences in thisdiscrepancy between those ischaemic
patientswith and without incoordination, suggestingthat the primary
effect of the disturbance is todelay mitral valve opening itself,
rather thanthe start of transmitral flow once the valve hasopened.
In the present study we were able toshow an effect not simply on
peak early diastolicflow velocity but on the peak initial
accelerationof blood during early diastole. Further, theinfluence
of isovolumic relaxation could be
resolved by stepwise regression analysis, intotwo discrete
components: isovolumic relaxa-tion time itself and the extent of
incoordinationduring isovolumic relaxation. Both wereindependently
significant in patients with leftventricular hypertrophy or
ischaemic heartdisease with normal cavity size. We havealready
shown the importance of incoordina-tion in affecting rapid filling
during early dia-stole,25 and the results of our patients
withischaemic heart disease and normal cavity sizeare entirely
compatible with this study. Inthose ischaemic patients with cavity
dilatation,however, isovolumic relaxation time was theonly
influence upon the A/E ratio. This was notdue to a lack of
incoordination in the patientswith dilated ventricles; on the
contrary, theyhad particularly incoordinate ventricles.
Thisassociation between isovolumic relaxation andfilling was very
significant in all three patientgroups with values of R2 (and thus
theproportion of overall variance of filling patternaccounted for)
consistently above 50%. Thisimplies that events before rather than
aftermitral opening, were the dominant determin-ants ofboth the
initial acceleration ofblood intothe left ventricle, and of the A/E
ratio of thesepatients with ventricular disease.By definition a
force is that which causes
acceleration. In being so closely associated withinitial
acceleration, therefore, events duringisovolumic relaxation must
also be closelyrelated to the forces causing blood flow from
theleft atrium to the ventricle-that is, to theatrioventricular
pressure gradient. Severalpotential mechanisms exist to explain
thisconnection. Isovolumic relaxation time itselfhas already been
shown to be closely related toleft ventricular filling pressure,26
which maypartly explain our findings. This simplerelation, however,
would not account for theadditional effect of incoordination
duringisovolumic relaxation. Incoordinate wallmotion is due to
asynchronous termination ofsystole in different regions of the
ventricle.27Clearly, this cannot be assessed in detail fromthe
ventricular pressure pulse, but its overalleffect is likely to slow
the rate of fall ofpressure,and thus to prolong isovolumic
relaxation timeindependently of the pressure differencebetween the
aorta and left atrium. Also, theearly diastolic pressure drop
across the mitralvalve depends on the effective pressure withinthe
left ventricle as well as the left atrium.Normally, left
ventricular pressure in earlydiastole is negative to atmospheric
pressure,reflecting the effectofrestoring forces within
themyocardium at end systole.2"' Dissipation ofthese forces by
incoordinate relaxation wouldthus increase effective ventricular
pressuretowards zero by the time the mitral valve opensand so
reduce the atrioventricular pressuregradient independently of any
change inorthodox estimates of left atrial pressure (madewith
respect to the atmosphere or mid rightatrium). The absence ofthe
effect ofincoordina-tion on early diastolic acceleration in
patientswith cavity dilatation, unlike its clear effect inthe other
two groups, can also be explained onthis basis. We have already
presented evidence
571
on June 6, 2021 by guest. Protected by copyright.
http://heart.bmj.com
/B
r Heart J: first published as 10.1136/hrt.68.12.567 on 1 D
ecember 1992. D
ownloaded from
http://heart.bmj.com/
-
Brecker, Lee, Gibson
suggesting that normal ventricular restoringforces are
maintained in patients with leftventricular hypertrophy,3' but not
in those withcavity dilatation and low ejection fraction.23
Ifrestoring forces are not generated at endsystole, then
incoordinate wall motion duringisovolumic relaxation would not be
expected toha-ve any further effect on early diastolic trans-mitral
flow. These effects of incoordination arenot easy to reproduce in
any experimentalmodel, and must thus be studied clinically.The
linear relation between the Doppler A/E
ratio and isovolumic relaxation time has prac-tical
consequences. If isovolumic relaxationtime is less than 60 ms, the
A/E ratio will benormal (
-
Relation of left ventricular isovolumic relaxation time and
incoordination to transmitral Dopplerfilling patterns
19 Lee CH, Hogan JC, Gibson DG. Diastolic disease in
leftventricular hypertrophy: comparison of M mode andDoppler
echocardiography for the assessment of rapidventricular filling. Br
Heart J 1991;65:194-200.
20 Sanderson JE, Gibson DG, Brown DJ, Goodwin JF.
Leftventricular filling in hypertrophic cardiomyopathy.
Anangiographic study. Br Heart J 1977;39:661-70.
21 Fioretti P, Brower RW, Meester GT, Serruys PW. Interac-tion
of left ventricular relaxation and filling during earlydiastole in
human subjects. Am J Cardiol 1980;46:197-203.
22 Gibson DG, Brown D. Measurement of instantaneous
leftventricular dimension and filling rate in man,
usingechocardiography. Br Heart J 1973;35:1141-9.
23 Ng KSK, Gibson DG. Relation of filling pattern to
diastolicfunction in severe left ventricular disease. Br Heart
J1990;63:209-14.
24 Lee CH, Vancheri F, Josen MS, Gibson DG. Discrepanciesin the
measurement of isovolumic relaxation time: a studycomparing M mode
and Doppler echocardiography. BrHeart J 1990;64:214-8.
25 Hui WKK, Gibson DG. Mechanisms of reduced leftventricular
filling rate in coronary artery disease. Br HeartJ
1983;50:362-71.
26 Mattheos M, Shapiro E, Oldershaw PJ, Sacchetti R, GibsonDG.
Non-invasive assessment of changes in leftventricular relaxation by
combined phono-, echo-, andmechanocardiography. Br Heart J
1982;47:253-60.
27 Gibson DG, Prewitt TA, Brown DJ. Analysis of leftventricular
wall movement during isovolumic relaxationand its relation to
coronary artery disease. Br Heart J1976;38: 1010-9.
28 Suga H, Goto Y, Igarashi Y, Yamada 0, Nozawa T,Yasumura Y.
Ventricular suction under zero sourcepressure for filling. Am J
Physiol 1986;251:47-55.
29 Yellin EL, Hori M, Yoran C, Sonnenblick EH, Gabbay S,
Frater RWM. Left ventricular relaxation in the filling
andnonfilling intact canine heart. Am J Physiol 1986;250:620-9.
30 Nikolic S, Yellin EL, Tamura K, et al. Passive properties
ofcanine left ventricle: diastolic stiffness and restoring
forces.Circ Res 1988;62:1210-22.
31 Park CH, ChowWH, Gibson DG. Phase differences betweenleft
ventricular wall motion and transmitral flow in man:evidence for
involvement of ventricular restoring forces innormal rapid filling.
Int J Cardiol 1989;24:347-54.
32 Castello R, Pearson AC, Kern MJ, Labovitz AJ, Lenzen
P.Diastolic function in patients undergoing coronary angio-plasty:
Influence of degree of revascularisation. J Am CollCardiol
1990;15:1564-9.
33 Yellin E, Nikolic S, Frater RWM. Left ventricular
fillingdynamics and diastolic function. Prog Cardiovasc
Dis1990;32:247-7 1.
34 Doran JH, Traill TA, Brown DJ, Gibson DG. Detection
ofabnormal left ventricular wall movement during iso-volumic
contraction and early relaxation. Comparison ofecho-and
angiocardiography. Br Heart J 1978;40:367-71.
35 Gibson DG, Doran JH, Traill TA, Brown DJ. Abnormal
leftventricular wall movement during early systole in patientswith
angina pectoris. Br Heart J 1978;40:758-66.
36 Brecker SJ, Stojnic BB, Xiao HB, Mbaissouroum M,Gibson DG.
Prolonged right ventricular systole impairsleft ventricular filling
in pulmonary hypertension[Abstract]. J Am Coll Cardiol 1992;19
(suppl A):394.
37 Zelinsky R, Diebold B, Raffoul H, Cohen A. Modifications
ofthe mitral flow due to the presence of a left bundle branchblock
(LBBB) [Abstract]. J Am Coll Cardiol 1988;11(suppl A):174.
38 Xiao HB, Lee CH, Gibson DG. Effect of left bundle branchblock
on diastolic function in dilated cardiomyopathy. BrHeart J
1991;66:443-7.
573
on June 6, 2021 by guest. Protected by copyright.
http://heart.bmj.com
/B
r Heart J: first published as 10.1136/hrt.68.12.567 on 1 D
ecember 1992. D
ownloaded from
http://heart.bmj.com/