Download Chronic heart failure patients with restrictive LV filling pattern have

International Journal of Cardiology 100 (2005) 5 – 12
www.elsevier.com/locate/ijcard
Chronic heart failure patients with restrictive LV filling pattern
have significantly less benefit from cardiac resynchronization
therapy than patients with late LV filling patternB
Tushar V. Salukhea,b,T, Darrel P. Francisa,c, Jonathan R. Clagueb, Richard Suttona,b,
Philip Poole-Wilsona,b, Michael Y. Heneina,b
a
National Heart Lung Institute, Royal Brompton Hospital, Sydney Street, London SW3 6 NP, UK
b
Royal Brompton Hospital, London, UK
c
St. Mary’s Hospital, London, UK
Received 1 October 2004; received in revised form 29 January 2005; accepted 30 January 2005
Abstract
Background: Cardiac resynchronization fails to improve symptoms in up to one third of patients meeting criteria for this treatment, for
reasons which are unclear. Indeed, the very mechanism of benefit from resynchronization is controversial. Resynchronization may work by
improving ventricular filling: we tested the hypothesis that benefit from resynchronization depends on filling pattern.
Methods and results: We assessed symptoms (NYHA class) and LV filling of 40 patients with chronic heart failure and prolonged QRS who
underwent resynchronization. Fifteen had restrictive filling pattern (E velocity z1.0 m/s, E/A ratio N1 and E wave deceleration time V140 ms) and
25 had late filling pattern (single isolated A wave or summation wave filling in late diastole). At 6 months, the patients with restrictive filling failed
to show the improvements observed in those with late filling. They failed to reduce NYHA class (DNYHA: 27% improved one class, 66%
unchanged, 7% worsened one class, P=NS; vs. 8% improved two classes, 72% improved one class and 20% unchanged, Pb0.001; difference
between groups, Pb0.001). They failed to reduce LV end-diastolic dimension (DLVEDD 0.04 cm, P=NS; vs. 0.6, Pb0.001; difference between
groups, Pb0.05) or end-systolic dimension (DLVESD 0.01 cm, P=NS; vs. 0.6, Pb0.001; difference between groups, Pb0.05). They failed to
improve cardiac cycle efficiency (Dtotal isovolumic [wasted] time 2.1 s/min, P=NS; vs. 5.4 s/min; difference between groups, Pb0.001).
Conclusion: Among patients routinely eligible for resynchronization, those with restrictive filling may show significantly less (and possibly
no) improvement in symptom class and ventricular dimensions after resynchronization. Their failure to improve cardiac cycle efficiency may
account for their attenuated clinical benefit.
D 2005 Elsevier Ireland Ltd. All rights reserved.
Keywords: Cardiac resynchronization theraphy; LV filling pattern; Total isovoluminic time
1. Introduction
A succession of trials has shown that groups of patients
with heart failure and broad QRS duration have significant
relief of symptoms from cardiac resynchronization therapy
B
DOI of linked article: 10.1016/j.ijcard.2005.02.002.
T Corresponding author. National Heart Lung Institute, Royal Brompton
Hospital, Sydney Street, London SW3 6 NP, UK. Tel.: +44 20 7352 8121;
fax: +44 20 7351 8510.
E-mail address: [email protected] (T.V. Salukhe).
0167-5273/$ - see front matter D 2005 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ijcard.2005.01.010
[1–4] and subsequently show reduction in LV cavity size
[5–8]. Exactly which patients are most likely to benefit
remains controversial. Large clinical trials may eventually
address this question, but would be best designed in the
light of a rational physiological basis for subgrouping
patients. The inclusion criteria of the trials encompass a
range of cardiac physiological states, and so permitting
more than one interpretation of the mechanism of benefit
from resynchronization.
The broad QRS, an entrance criterion for all of these
trials, can delay the onset of mechanical contraction in
6
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
some segments of the ventricle, so that they are still
shortening beyond the end of LV ejection. This delayed
segmental shortening can impede ventricular filling in
early diastole and can therefore leave ventricular filling
heavily dependent on atrial contraction, which occurs late
in diastole (blate filling patternQ) [9]. In some patients with
heart failure and broad QRS, however, the elevation of left
atrial pressures is severe enough to drive early diastolic
filling (brestrictive filling patternQ). The exact response of
these two physiologically different conditions to cardiac
resynchronization therapy is not known, and may give
insights into the mechanism of benefit of resynchronization. Whether patients with blate fillingQ and brestrictive
fillingQ have similar or disparate responses to resynchronization depends on (and may be informative about) the
mechanism of benefit of resynchronization.
One hypothesis of the mechanism of benefit from cardiac
resynchronization is that it makes systole more efficient.
Under this bglobal improvementQ hypothesis, patients would
be expected to benefit to a similar extent regardless of filling
pattern. An alternative hypothesis is that resynchronization
works primarily through enhancing time available for
filling. This bimprovement of filling timeQ hypothesis makes
a specific testable prediction: benefit would depend on the
type of filling pattern. In patients with broad QRS, the onset
of systole in certain LV segments is delayed. Subsequently,
systole continues beyond LV ejection into early diastole,
thus blunting early ventricular filling [9]. In these patients,
resynchronization therapy–by allowing filling to begin
earlier–would have the best potential to enhance total
filling. In contrast, if left atrial filling pressure is sufficiently
elevated–enough to overwhelm segmental asynchrony–filling occurs early, is prompt and rapid (restrictive).
In this study we set out to examine whether there was a
difference in response to resynchronization in symptoms,
cardiac cycle efficiency, and in left ventricular dimensions
between patients with restrictive filling and late filling.
2. Methods
2.1. Study design
In this prospective study, patients were identified
through referral from heart failure clinics and enrolled
over an 18-month period from August 2001 to February
2003. Patients were referred for resynchronization therapy
based on symptomatic heart failure due to dilated
cardiomyopathy and QRS prolongation on electrocardiography. All patients had clinical assessment, electrocardiography and echocardiography at baseline. Classification of
LV filling pattern as restrictive or late diastolic was made at
baseline.
Eligible patients were admitted for implantation of atriobiventricular pacemakers. Prior to discharge, the programmed atrio-ventricular delay was optimized guided by
Doppler echocardiography to maximize LV filling time. At 6
months, baseline observations were repeated. The patients
and physicians who implanted the devices were blinded to
classification of LV filling. The investigator reprogramming
the device was also blinded to the filling pattern; although, the
instructions given to them during echo-guided atrio-ventricular delay optimization were by a physician who could see the
LV filling pattern on the echocardiograph monitor.
2.2. Patient eligibility
Patients were eligible if they had chronic heart failure due
to either ischaemic or idiopathic dilated cardiomyopathy and
had symptoms consistent with NYHA functional class III–IV
dyspnoea. Patients who were classified as bunstable class IIQ
patients (who often described class III symptoms) were also
included. Patients had to be in sinus rhythm, have a left
ventricular end-diastolic dimension of 5.5 cm or more, a left
ventricular ejection fraction of 40% or less and a QRS
duration of 130 ms or more. All patients had received the
maximum tolerated medical therapy for heart failure which
was stable for at least one month prior to pacemaker
implantation. Medical therapy included angiotensin-converting enzyme inhibitors (or angiotensin II receptor blockers),
beta-blockers and diuretics.
Patients were excluded if they already had a pacemaker
or defibrillator prior to biventricular pacing, if they had a
cardiac or cerebral ischaemic event 3 months prior to
biventricular pacemaker implantation, or if they had a
history of atrial tachyarrhythmia.. The study was approved
by the Royal Brompton and Harefield Hospitals ethics
committee.
2.3. Echocardiographic evaluation and measurements
Resting transthoracic echocardiography was performed
using a Philips (Andover, Massachusetts, USA) Sonos 5500
echocardiograph with a multifrequency transducer. Left
ventricular transverse axis dimensions at end-diastole,
LVEDD (onset of the q wave of the ECG), and at endsystole, LVESD (onset of the second heart sound on the
phonocardiogram) were measured from the M-mode recording-according to the recommendations of the American
Society of Echocardiography. This was taken from the
parasternal long axis view with the M-mode cursor
positioned adjacent to the tips of the mitral valve leaflets,
using leading edge methodology.
Doppler flow velocities were recorded from the apical
four-chamber view, using the same transducer while in the
pulsed wave Doppler mode, with the sample volume
positioned at the tip of the mitral valve leaflets. Peak early
and late diastolic flow velocities and total left ventricular
filling time (time from the onset to the end of forward flow)
were measured. LV filling was classified as restrictive
according to a standard definition: E wave velocity greater
than 1.0 m/s, E/A ratio greater than one and E wave
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
deceleration time less than 140 ms [10,11]. We pre-specified
that the patients with a pseudonormal pattern would be
classified as part of the restrictive group. When present,
duration of mitral regurgitation was measured.
Left ventricular ejection time was measured from the
LV outflow tract velocity profile. LV filling and ejection
times were expressed as absolute values (ms) and as a
product of their duration and heart rate, expressed in
seconds per minute. The total isovolumic time (s/min) was
calculated as: 60 (total ejection time+total filling time);
these measurements are independent of heart rate [12]. The
total isovolumic time represents the time wasted within the
cardiac cycle when the LV is neither filling nor ejecting.
All recordings were made at a speed of 100 mm/s with a
superimposed electrocardiogram and phonocardiogram.
Images were stored digitally and analyzed off-line.
2.4. Electrocardiography
All patients had standard 12-lead resting electrocardiogram, recorded on a Hewlett-Packard Xli Page Writer
(calibration 0.1 mV/mm, paper speed 25 mm/s) on which
QRS duration was measured using built-in computer
software.
2.5. Device implantation
Patients were admitted 1 day prior to implantation of
atrio-biventricular pacemakers (or atrio-biventricular pacemaker-defibrillators). Right ventricular leads were positioned at the RV apex. A venogram of the coronary sinus
was performed to help optimize coronary sinus lead
position. The target site was preferably the lateral wall of
the LV, mid-way between its base and apex using the
lateral vein. Other lateral and posterior sites were also
accepted (postero-lateral or posterior vein). Lead positions
were assessed and documented with a chest X-ray.
Pacemakers were programmed to a therapeutic biventricular pacing mode (BV-DDD) and prior to discharge the
programmed atrio-ventricular delay was optimized using
Doppler echocardiography to maximize LV filling time,
while ensuring atrial sensing and ventricular pacing, even
at higher heart rates.
2.6. Statistical analysis
Primary endpoints of the study were NYHA functional
class, LVEDD and LVESD. Secondary endpoints were
total isovolumic time, LV filling time and LV ejection time.
Statistical calculations were performed using Statview
4.5 (Abacus Concepts, Berkeley, CA, USA). Continuous
data are expressed as a meanFstandard deviation. Statistical
comparisons of continuous variables within groups and
between groups were assessed using the paired and unpaired
t-test, respectively. Comparisons of non-parametric variables were assessed using the Mann–Whitney U test for
7
unpaired samples and the Wilcoxon Signed Rank test for
paired samples. A P value of b0.05 was considered
significant. Comparison of changes in variables between
groups was performed by unpaired t-test. The predictive
value of echocardiographic parameters and QRS duration
for changes in the primary endpoints was determined using
a simple linear regression model. When appropriate, a
multivariate model was then constructed by a forward
stepwise method.
3. Results
3.1. Patient population
Of 43 potentially eligible patients, 1 was excluded
because resynchronization involved upgrading a pre-existing dual-chamber pacemaker and 2 were excluded because
of they were in atrial fibrillation. Hence, 40 patients were
enrolled in this study: 15 patients with restrictive filling
(including 1 with pseudo-normalized LV filling) and 25
patients with late diastolic LV filling. The groups were
comparable at baseline (Table 1) in age, sex, QRS
duration, QRS morphology, LV dimensions, ejection
fraction, NYHA class, aetiology of heart failure, blood
pressure and medication profile. The range of systolic
blood pressures of patients enrolled was between 80 and
154 mm Hg. As expected the patients with prompt early
diastolic filling (restrictive filling pattern) had a shorter
duration of functional mitral regurgitation. None of the 40
patients had significant (more than mild) mitral regurgitation. Among the patients with restrictive filling, E wave
velocity was 1.2F0.2 m/s, E wave deceleration time was
113F19 ms and E/A ratio was 4.8F2.9.
3.2. Device implantation, coronary sinus lead position and
pacing parameters
All 40 patients were successfully implanted with an
atrio-biventricular pacemaker or a combined device (atriobiventricular pacemaker-defibrillator). Adequate right
atrial and right ventricular lead positions were achieved
in all patients. Although two patients required coronary
sinus lead repositioning due to loss of myocardial capture
of pacing, acceptable coronary sinus lead positions were
achieved in all forty patients. The type of pacing devices
used, LV lead positions achieved and activation sequence
proportions at follow-up did not differ between the two
groups (see Table 1).
Two patients died before follow-up. Average follow-up
time in the 38 remaining patients was 6.1F1.1 months. Over
the follow-up period, mean atrial-sensed percentage was
76F30% and mean (bi-)ventricular-paced percentage was
98F3% in the group as a whole. There was no significant
difference in these parameters between the two groups (see
Table 1).
8
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
Table 1
Patient characteristics
Table 2
Effect of resynchronization on the whole population
Restrictive
filling (N=15)
Baseline characteristics
Males (%)
Age (years)
Systolic BP (mm Hg)
Diastolic BP (mm Hg)
Diabetes (%)
Ischaemic
cardiomyopathy (%)
NYHA class (%)
IV
III
unstable II
Medication
ACE inhibitors (%)
Beta-blockers (%)
Spironolactone (%)
QRS duration (ms)
LBBB (%)
LVESD (cm)
LVEDD (cm)
Ejection fraction (%)
R–R interval (ms)
LV filling time (ms)
LV ejection time (ms)
Functional MR
duration (ms)
LV filling time (s/min)
LV ejection time (s/min)
Total isovolumic
time (s/min)
Pacing characteristics
Device manufacturer
Guidant (%)
Medtronic (%)
Device
Pacemaker (%)
Pacemaker–defibrillator (%)
LV lead position
Lateral (%)
Posterior–lateral (%)
Posterior (%)
Activation sequence
at follow-up
Mean %age A-sensed
Mean %age (B)V-paced
Late filling
(N=25)
P
87
62F13.8
122F20
61F14
27
60
76
66.8F7.7
121F20
64F13
28
56
0.24
0.16
0.89
0.51
0.28
0.25
13.3
80
6.7
4
84
12
93
60
53
162F21
100
6.3F1.1
7.6F1.2
35F5
784F129
357F171
232F37
423F58
88
60
56
153F16
100
6.0F1.3
7.3F1.1
36F9
820F184
331F149
248F26
501F55
0.38
0.26
0.25
0.11
1
0.54
0.53
0.79
0.44
0.67
0.13
0.003
26.2F8.7
17.9F2.3
15.9F8.5
23.3F5.6
18.6F3.6
18.0F4.7
0.22
0.47
0.32
0.21
0.23
73
27
64
36
40
60
56
44
67
20
13
56
28
16
75F29
98F3
80F28
98F3
0.16
0.21
0.6
0.76
New York Heart Association, NYHA; Left ventricular (LV) end-systolic
dimension, LVESD; LV end-diastolic dimension, LVEDD. Sex, medication
profile, coronary artery disease and diabetes prevalence were compared by
Fisher’s exact probability test. NYHA class was compared by Mann–
Whitney U test. All other comparisons were made using unpaired t-test.
3.3. Effect of resynchronization in whole population
The effects of resynchronization on the whole population
are shown in Table 2. NYHA class improved, with 62% of
patients improving by at least one functional class ( Pb0.0001).
Left ventricular cavity size fell, LVEDD by 0.34F1.6 cm
( Pb0.01) and LVESD by 0.41F0.9 cm ( Pb0.01). Ejection
NYHA class
IV (%)
III (%)
II (%)
LVESD (cm)
LVEDD (cm)
Ejection fraction (%)
R–R interval (ms)
LV filling time (ms)
LV ejection time (ms)
Functional MR duration (ms)
LV filling time (s/min)
LV ejection time (s/min)
Total isovolumic time (s/min)
Baseline
Follow-up
10
82.5
7.5
6.2F1.2
7.4F1.1
35F8
810F164
342F156
241F32
463F68
24.5F7.0
18.3F3.2
17.2F6.4
2.5
35
62.5
5.7F1.2
7.0F1.1
43F14
854F109
398F129
252F39
432F58
27.6F6.7
17.9F2.6
14.5F7.0
P
b0.0001
0.009
0.003
0.0002
0.07
0.01
0.09
0.84
0.01
0.25
0.03
New York Heart Association, NYHA; Left ventricular (LV) end-systolic
dimension, LVESD; LV end-diastolic dimension, LVEDD; Mitral regurgitation, MR; NYHA class was compared by Wilcoxon Signed Rank test.
All other comparisons were by paired t-test.
fraction increased by 8F16% ( Pb0.001). Efficiency of the
cardiac cycle also improved, in that total isovolumic time
shortened by 2.6F7.7 s/min ( Pb0.01) and LV filling time
increased by 3.0F8.7 s/min ( P=0.01). There was no
Table 3
Effect of resynchronization in restrictive vs. late filling
Restrictive
filling (N=14)
DNYHA class
%Improving
by 2 class
%Improving
by 1 class
%With no change
%Worsening
by 1 class
%Worsening
by 2 class
DLVESD (cm)
DLVEDD (cm)
DEjection
fraction (%)
DR–R interval (ms)
DLV filling
time (ms)
DLV ejection
time (ms)
DFunctional MR
duration (ms)
DLV filling
time (s/min)
DLV ejection
time (s/min)
DTotal isovolumic
time (s/min)
0.2F0.6
0
Late filling
(N=24)
0.9F0.6
8
P for difference
in effect size
between groups
b0.001
6.7
72
66.6
26.7
20
0
0
0
0.0F0.9
0.0F0.8
4F13
0.6F0.8
0.6F0.7
11F12
0.04
0.03
0.14
71F127
3F161
25F149
83F92
0.34
0.06
9F38
12F38
0.81
38F149
47F48
0.09
1.3F8.8
5.6F4.1
0.002
0.9F2.2
0.2F2.6
0.46
2.1F8.6
5.4F4.4
0.001
New York Heart Association, NYHA; Left venticular (LV) end-systolic
dimension, LVESD; LV end-diastolic dimension, LVEDD. P values for the
difference in effect size between groups were obtained using the Mann–
Whitney U test for non-parametric variables (NYHA class) and unpaired ttest for all other comparisons.
10
10
9
9
8
8
LVEDD (cm)
Restrictive LV filling
(N=14)
LVESD (cm)
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
7
P=0.98
6
5
4
6
5
3
Pre-CRT
Pre-CRT
Post-CRT
10
10
9
9
8
8
7
6
5
P<0.001
LVEDD (cm)
LVESD (cm)
P=0.85
7
4
3
Late LV filling
(N=24)
9
Post-CRT
7
P<0.001
6
5
4
4
3
3
Fig. 1. Changes in LV end-systolic dimension (LVESD) and end-diastolic dimension (LVEDD), before and after resynchronization. Observations in individual
patients with restrictive LV filling are shown on the top and those of the late diastolic LV filling group on the bottom. The heavy lines represent the changes in
mean values and standard deviation of the mean. Comparisons before and after pacing were made by paired t-test.
3.4. Symptomatic response in restrictive and late filling
patterns
The restrictive filling patients showed no significant
improvement in NYHA class after resynchronization. In
contrast, the patients with late diastolic LV filling had a
significant reduction in NYHA class. These patients had a
significantly greater reduction in NYHA class after resynchronization ( Pb0.0001, see Table 3).
3.5. Changes in ventricular dimensions in restrictive and
late filling patients
left ventricular filling and ejection times were unchanged. In
patients with late filling LV filling, time increased from
334F149 ms to 414F123 ms, Pb0.001 (or 23.3F5.6 s/min
to 28.9F5.9 s/min, Pb0.0001). LV ejection time did not
change.
Total isovolumic time did not change after resynchronization in patients with restrictive filling but was significantly reduced in patients with late filling (from 18.0F4.7 s/
min to 12.6F5.7 s/min, Pb0.0001, Fig. 2). There was a
significant difference between restrictive filling and late
filling patients in the change in total isovolumic time and LV
40
Restrictive LV filling (N=14)
35
Total isovolumic time
(seconds per minute)
significant change in heart rate or LV ejection time after
resynchronization.
On univariate analysis, none of the baseline echocardiographic parameters (LV filling time, ejection time, t-IVT,
MR duration or ejection fraction) or QRS duration were
significant predictors of improvement in NYHA functional
class, LVESD or LVEDD.
30
25
20
P = 0.34
15
10
5
0
Pre-CRT
Post-CRT
40
35
Total isovolumic time
(seconds per minute)
While the restrictive filling patients showed no significant
change in LV dimensions after resynchronization, the late
filling patients did show significant reduction in both LVESD
(from 6.1F1.3 cm to 5.5F1.2 cm, Pb0.001) and LVEDD
(from 7.3F1.1 cm to 6.7F1.0 cm, Pb0.001) as shown in Fig.
1. There was a significant difference between restrictive
filling and late filling patients in the change in ventricular
dimensions after resynchronization, with Pb0.05 for LVEDD
and for LVESD on assessment by unpaired t-test (Table 3).
Late LV filling (N=24)
30
25
20
15
P < 0.0001
10
5
0
3.6. Changes in cardiac cycle timing in restrictive and late
filling patients
Heart rate (R–R interval) did not change in either group
after resynchronization. In patients with restrictive filling,
Fig. 2. Changes in total isovolumic time. Observations of individual
patients with restrictive LV filling are shown on top and those of the late
diastolic LV filling group below. Mean changes and standard deviation of
the mean are represented by the heavy lines. Comparisons before and after
pacing were made by t-test.
10
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
filling time with resynchronization, with P=0.001 for total
isovolumic time and P=0.002 for LV filling time (Table 3).
3.7. Reproducibility
The reproducibility of measurements used in our analysis
has been published in a previous report by our department
[5,12]. This study reported a coefficient of variation of 1–2%
and 2–3% for intra- and inter-observer variability respectively for measuring LV ejection and filling times. The
coefficients of variation were 10–11% (intraobserver) and
12–14% (interobserver) for measurement of LV cavity size.
4. Discussion
In this study we found that among patients with dilated
cardiomyopathy and intraventricular conduction delay,
those with restrictive pattern of left ventricular filling had
a significantly smaller benefit from cardiac resynchronization, than patients with a late filling pattern. This smaller
benefit manifested as no significant symptomatic improvement, no significant reverse remodelling and no significant
improvement in cardiac cycle efficiency (each significantly
different from late filling patients).
These findings help narrow down the range of mechanisms of benefit from cardiac resynchronization therapy.
Under the bglobal improvementQ hypothesis, there would be
no grounds for one filling pattern to respond differently
from another. The limited peer reviewed data available in
the literature include a post hoc analysis of fractional
shortening in 19 patients, which found a greater increment
in LV fractional shortening in patients with higher transmitral E wave velocity [13]. However, E wave velocity is
neither monotonically related to severity of disease nor to
left atrial pressure, and on its own is not an unambiguous
classifier of physiological status. Moreover LV fractional
shortening does not necessarily reflect symptomatic status
and may be especially unreliable in the settings of
asynchrony and of artificial resynchronization.
Our data favor the benhancement of filling timeQ
hypothesis over the bglobal improvementQ hypothesis, since
there appears to be a specificity of benefit to the late filling
group rather than restrictive filling group (DNYHA class
0.9 vs. 0.2, DLVESD 6 vs. 0 mm, DLVEDD 6 vs. 0
mm). These data are supported by another study which
described significantly greater E wave velocities in nonresponders (those who did not show LV reverse remodeling)
to resynchronization [14].
Aside from these data, there are theoretical physiological
grounds favoring the benhancement of filling timeQ hypothesis. Left bundle branch block, regardless of ventricular
cavity size, prolongs the isovolumic contraction and
relaxation times at the expense of filling and ejection times
[15,16], thus grossly reducing the temporal efficiency of
ventricular function. LBBB regionally delays onset of LV
shortening so that parts of the ventricle are continuing to
contract beyond the end of ejection, thus impeding LV
filling during early diastole [9]. In some cases early diastolic
filling can be completely suppressed (classic blate filling
patternQ), leaving all filling to occur in late diastole
(simultaneously with the atriogenic contribution to filling)
and thereby raising left atrial pressure and rendering the
patient susceptible to breathlessness [9]. Cardiac resynchronization therapy brings about more uniform activation of
the LV and therefore a more regionally consistent onset of
relaxation. This has two benefits. First, more of the time and
energy of LV shortening is applied productively in stroke
volume ejection, rather than being wasted in early diastole.
Second, the alleviation of intra-ventricular tension during
early diastole now allows the ventricle to fill more promptly
rather than late in diastole (see Fig. 3). These effects prolong
LV ejection and filling times and reduce the total isovolumic
time, thus making the cardiac cycle more efficient. Therefore, under the benhancement of filling timeQ hypothesis,
patients with late filling would be expected to have a good
chance of benefit.
The benhancement of filling timeQ hypothesis also
explains why the restrictive patients may have a smaller
benefit. In restrictive filling, elevated left atrial pressure
forces early diastolic filling to begin promptly after then end
of LV ejection. The elevated end-diastolic ventricular
pressures then forces early cessation of passive filling
(manifest as a short E wave deceleration time). The result of
the short duration of this filling waveform is that even if
additional time were available to it in early diastole, the net
flow into the ventricle would not be increased.
The underlying processes involved in late filling and
restrictive filling can co-exist. This means that some patients
with restrictive filling and underlying intraventricular
conduction delay could still benefit from resynchronization
if, for example, addition of vasodilators were able to
unmask latent late diastolic filling [17]. Hence, the pattern
of ventricular filling represents the dominant physiological
mechanism of impaired function; late filling suggesting
delayed relaxation and restrictive filling suggesting raised
atrio-ventricular filling pressures.
Failure of reverse re-modeling in the ventricles of
patients with restrictive filling may result from failure to
improve cycle efficiency (i.e. failure to shorten total
isovolumic time). This possibility is supported by MUSTIC
trial [5] data which showed the temporal sequence of
physiological events after resynchronization. First, total
isovolumic time fell within 3 months (and fell no further
thereafter). Then, LV cavity size fell and continued to fall
for up to 12 months after resynchronization. In resynchronization, it may be the reduction of total isovolumic time
and consequent increase in time available for filling which
allows reverse re-modelling to occur.
How can these findings be reconciled with the multiplicity
of trials showing clinical benefit from resynchronization? We
believe it is noteworthy that there was a statistically
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
11
Fig. 3. Pulsed wave Doppler recording of late diastolic LV filling (above) and LV ejection (below) in a patient prior to resynchronization is shown on the left
side panel. Doppler recording of LV filling (above) and ejection (below) of the same patient after resynchronization is shown on the right. Of note the early
phase of diastole made available through resynchronization becomes occupied with a dominant E wave with and total isovolumic time (represented by the
coloured bars) has reduced by two-thirds after resynchronization with no change in cycle length.
significant benefit in our patient population as a whole. This
would accord with the benefit seen in previous larger trials.
However, none of the large multicenter trials set out to
systematically pre-classify patients according to filling
pattern. Therefore our data are consistent with the possibility
that within these trials, there existed a (large) minority of
patients who may never have had much prospect for benefit.
4.1. Limitations
It may be a limitation of the study that it included a
population with a mixed aetiology of cardiomyopathy.
However, the proportions of ischaemic and non-ischaemic
patients were comparable in the two groups. Similarly, other
confounding factors which could potentially explain results,
for example, ejection fraction, LV lead-tip position, blood
pressure, medication profile and amount of therapeutic
pacing during follow-up were comparable in both groups.
Patients were selected and grouped according to two
distinct and dichotomous LV filling patterns. It is possible
that some patients, who could be classified between these
two paradigm groups, may demonstrate co-dominance of
the mechanisms which limit ventricular function and therefore may demonstrate an intermediate functional response to
resynchronization therapy.
Data was only available for a medium-term follow-up (6
months). Previous studies have demonstrated symptomatic
improvement [3,7], reduction in total isovolumic time [12]
and LV reverse remodeling [5,8] occur as early as three
months and continue to improve up to one year after
resynchronization. We therefore believe the findings at 6
months are a fair window over which to judge the mediumterm effects of resynchronization.
4.2. Conclusions
Patients with restrictive left ventricular filling are unlike
patients with late diastolic filling, in that they fail to show
symptomatic benefit from resynchronization, fail to show
reverse re-modelling of the LV cavity. We believe that this
failure to respond is physiologically explicable on the basis
of their lower potential to increase the quantity of filling
when more time is made available for filling.
Fundamentally, the difference between the two groups
may well be the relative importance of the two competing
factors disrupting normal diastolic performance. When the
dominant factor is pure dyssychrony (which is potentially
correctable with resynchronization) filling occurs late in
diastole. When the dominant factors are elevated left atrial
and end-diastolic pressures, filling occurs in early diastole
with a restrictive pattern.
If large-scale trials substantiate this relationship between
filling pattern and clinical benefit, then there will be several
clinical implications. First, it may help provide physiological
refinement to the criteria currently available for selecting
patients for resynchronizaton. This may reduce the flow of
patients (meeting the current indications) to resynchronization. Second, the size of the expected benefit in patients subselected using the additional filling pattern criteria would
presumably be correspondingly greater, since it would be
enriched for likely responders. Third, it may broaden the
range of patients considered eligible for resynchronization to
12
T.V. Salukhe et al. / International Journal of Cardiology 100 (2005) 5–12
include patients with late filling from a variety of causes (not
limited to broad QRS).
References
[1] Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C, et
al. Effects of multisite biventricular pacing in patients with heart
failure and intraventricular conduction delay. N Engl J Med 2001;
344:873 – 80.
[2] Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronization
in chronic heart failure. N Engl J Med 2002;346:1845 – 53.
[3] Auricchio A, Stellbrink C, Butter C, et al. Clinical efficacy of cardiac
resynchronization therapy using left ventricular pacing in heart failure
patients stratified by severity of left ventricular conduction delay. J
Am Coll Cardiol 2003;42(12):2109 – 16.
[4] Young JB, Abraham WT, Smith AL, Leon AR, Lieberman R, Wilkoff
B, et al. Combined cardiac resynchronisation and implantable
cardioversion defibrillation in advanced chronic heart failure: the
MIRACLE ICD trial. JAMA 2003;289(20):2685 – 94.
[5] Duncan A, Wait D, Gibson D, Daubert JC. MUSTIC (Multisite
stimulation in cardiomyopathies trial). Left ventricular remodelling
and haemodynamic of multisite biventricular pacing in patients with
left ventricular systolic dysfunction and activation disturbances in
sinus rhythm: substudy of the MUSTIC (multisite stimulation in
cardiomyopathies) trial. Eur Heart J 2003;24(5):430 – 41.
[6] Yu C-M, Chau E, Sanderson JE, Fang K, Tang MO, Fung WH, et al.
Tissue Doppler evidence of reverse remodelling and improved
synchronicity by simultaneously delayed regional contraction after
biventricular pacing therapy in heart failure. Circulation 2002;
105(4):438 – 45.
[7] Saxon LA, De Marco T, Schafer J, Chaterjee K, Kumar UN, Foster E.
VIGOR Congestive heart failure investigators. Effects of long term
biventricular stimulation for resynchronization on echocardiographic
measures of remodelling. Circulation 2002;105(11):1304 – 10.
[8] St John Sutton MG, Plappert T, Abraham WT, Smith AL, DeLurgio
BD, Leon AR, et al. Effect of cardiac resynchronization therapy on
left ventricular cavity size and function in chronic heart failure.
Circulation 2003;107:1885 – 990.
[9] Henein M, Gibson DG. Suppression of left ventricular early diastolic
filling by long axis incoordination. Br Heart J 1995;73(2):151 – 7.
[10] Xie GY, Berk MR, Smith MD, Gurley JC, DeMaria AN. Prognostic
value of Doppler transmitral flow patterns in patients with congestive
heart failure. J Am Coll Cardiol 1994;24(1):132 – 9.
[11] Giannuzzi P, Temporelli PL, Bosimini E, Silva P, Imparato A, Corra
U, et al. Independent and incremental prognostic value of Dopplerderived mitral deceleration time of early filling in both symptomatic
and asymptomatic patients with left ventricular dysfunction. J Am
Coll Cardiol 1996;28:383 – 90.
[12] Duncan AM, O’Sullivan CA, Gibson DG, Henein MY. Electromechanical interrelation during Dobutamine stress in normal subjects
and patients with coronary disease: comparison of changes in
activation and inotropic states. Heart 2001;85:110 – 43.
[13] Thaman R, Murphy RT, Firoozi S, Hamid SM, Gimeno JR, Sachdev
B, et al. Restrictive transmitral filling patterns predict improvements in
left ventricular function after biventricular pacing. Heart 2003;
89:1087 – 8.
[14] Yu C-M, Fung W-H, Lin H, Zhang Q, Sanderson JE, Lau C-P.
Predictors of left ventricular reverse remodelling after cardiac
resynchronization therapy for heart failure secondary to idiopathic
dilated or ischaemic cardiomyopathy. Am J Cardiol 2002;91:684 – 8.
[15] Xiao HB, Lee CH, Gibson DG. Effect of left bundle branch block on
diastolic function in dilated cardiomyopathy. Br Heart J 1991;66:
443 – 7.
[16] Zhou Q, Henein MY, Coats AJS, Gibson DG. Different effects of
abnormal activation and myocardial disease on left ventricular
ejection and filling times. Heart 2000;84:272 – 6.
[17] Henein MY, O’Sullivan CA, Coats AJS, Gibson DG. Angiotensin
converting enzyme (ACE) inhibitors revert abnormal right ventricular
filling pattern in patients with restrictive left ventricular disease. J Am
Coll Cardiol 1998;32(5):1187 – 93.