Download Heart Rate Variability in Time Domain after Acute Myocardial Infarction

136
Heart Rate Variability
45
Heart r a t e variability in time domain after acute myocardial
infarction
Thomas Fetsch, Lutz Reinhardt, Markku Mikijtiwi, Dirk Bijckcr, Michael
Block, Martin Borggrefe and Giinter Breithardt.
Westftilische Wilhelms-Universitiit, Medizinische Klinik und Poliklinik,
lnnere Medizin C,D-48129 Monster
Introduction
The identification of individuals at high risk for development of life
threatening ventricular arrhythmias after myocardial infarction remains a
difficult problem that requires further refinement. The screening of postinfarction patients necessitates non-invasive risk predictors that have a hi$
level of sensitivity and specificity. Conventional analysis of Holter
recordings corrclates poorly with the devclopmcnt of serious vcntricular
arrhythmias and sudden death since there is a low sensitivity. Many patients
dying suddenly do not exhibit major spontancous ventricular arrhythmias.
This does not deny the well-known increasc in risk that is associated with
the presence of frequent and complex ventricular arrhythmias on Holter
monitoring. Furthermore, the analysis of late potentials from the signal
averaged electrocardiogram (ECG) is of additional prognostic significance
which. like Holter monitoring, is limited in its usefulness by a high number
of false positive results. In recent years, it has been shown that the
autonomic nervous system plays an important role in triggering serious
ventricular arrhythmias. Therefore, a markcr for the autonomic control on
cardiac clcctrophysiologic properties such as heart rate variability might
provide information useful for a more accurate risk stratification togcthcr
with other non invasivc methods.
Pathophysiological background
Confrol of the heart by oiifononiic nervoits p f h w a y
Affcrciit signals that originate from scvcrnl sciisors located in the carotid
sinus, the aorta, the grcat veins and both atria and vcntriclcs, give input to a
variety of cardio-vascular rcflcscs controlling heart and circulation. This
includes baroreceptors measuring blood pressure as well as chemoreceptors
responding to hypoxia and acidosis. The efferent autonomic nervous
pathways of these reflexes are located at the sinus node, the atrioventricular
(AV) node, the atrial and ventricular myocardium and the small coronary
arteries and arterioles. The right sympathetic and vagal nerves affect the
sinus node more than the AV node and vice versa for the nerval input from
the left side. Sympathetic nerves distribute mostly in the superficial
subepiwdiwn while v q a l fibers are located in the subendocardium
penetrating intramurally towards the epieardium ([I], Fig. I). Thc
activation of different cardiac reflexes leads to diverse reactions depending
on the type, location and distribution of the efferent nerves involved.
Stimulation of parasympathetic fibers prolongs the cycle length of the sinus
node and the conduction time of the AV node, produces more depression of
atrial than of ventricular contractility, shortens the atrial refractory period
and lengthens the ventricular refractory period. In contrast, the stimulation
of cffcrent sympathetic fibers leads to a shortening of the sinus cycle length
and of the AV conduction time, to an increase in contractility of the
ventricles more than of the atria and to a shortening of the refractory period
of the atrial and ventricular myocardium (21. The total effect on myocardial
conduction and contractility depends on the balance of sympathetic and
parasympathctic activation.
Modulation of cardiac aufonomic innervation by ischemia
Thc autonomic responses to myocardial infarction are complex. They
dcpend on the changes of the afferent and efferent limbs of the autonomic
cardiac reflexes due to~bcalisationand size of myocardial damage.
Trnnsmural infarction that involves the subepicardium leads to both, altered
vagal and sympathetic innervation, while subendocardial infarction
interrupts the parasympathetic fiben but spares the sympathetic nerves.
Vagal dcnervation and intact sympathctic innervation may lead to
anhythmogenic conditions [3.4]. If the area of damaged tissue is small, the
effect on thc balance of the cardiac autonomic tonus is minor. Especially in
the border zone of infarction, damaged, altered and normal myocardial cells
arc closely ncighboured. This zone of heterogeneous autonomic innervation
may also be prone to developing arryhthmias [I]. An additional
phenomenon observed in case of complete dcncrvation is the ensuing
Fig. I : Schematic of sagittal view of left ventricular wall showing
pathways of vagal and sympathetic afferent and efferent nerves.
Postganglionic sympathetic ~ X O N are located superf~callyin periadventia
of coronary artcrics; postganglionic vagal axons cross the AV groove in
subepicardium but are located in subendocardium. Cx
circumflex
coronary artery; LAD - left ventricular descending coronary artery (from
Zipcs DP. Circulation 1990;82: 1096.Reprinted with permission)
-
supcrscnsitivity to catecholamines. It manifests an overshooting response to
norepinephrine in myocardium adjacent to the denervated tissue [5]. This
again can result in arrhythmogenic conditions.
i'hc aiifonomic nervous sysfem and fhe mechanisms of arrhythmia
Thcrc is increasing evidence that the sympathetic nervous system and
circulating catecholamines interact with the three major mechanisms
involved in the generation of arrhythmias: [I] Enhanced automaticity;
cntccholnmincs iiicrcase the slow inward calcium currents in phase 4 of the
action potcntial of paccmaker cells. Infarcted myocardium with abnormal
elcctrophysiological properties may respond to sympathetic stimulation by
cnhniiccd autoinnticity generating arrhythmias, even sustained ventricular
tachycardia 16. 7. 8. 9, 101. (2) Triggered automaticity; low amplitude
oscillations, triggcred by the preceding action potential, often occur at the
completion of the action potential as a result of calcium inward currents.
These soulled delayed afterpotentials mostly are subthreshold. In altered
myocardial cells, the calcium influx can be increased by catecholamines
lading to augmented afterpotentials reaching the threshold potential and
triggering a spontaneous action potential [lo, 111. (3) R a w , a narCry
circuit consists of two pathways ~ 0 ~ e c t - dproximally and distally
exhibiting different conduction velocities and refractory periods. In case of
unidircctional block, the entering impulse is conducted antegradely through
one pathway, but blocked in thc second pathway. From the distal connection
the impulse is conducted retrogradely throu&, the second limb. If the
impulse reaches the proximal connection with the first pathway at an
appropriate time after depolarisation, it may reenter the circuit. If this
process repeats itsclf, a reentrant tachycardia is generated. Stimulation of
the sympathetic nervous system may increase impulse conduction velocity.
Since the distribution of thc sympathetic innervation in the myocardium is
nonhomogencous, sympathetic stimulation leads to nonuniform effects on
the elcctrophysiological characteristics. Especially in the borderline of
infarcted myocardium, abnormal inhomogeneity of impulse conduction may
thcrcby be augmented, thus, generating conditions for reentrant
tachycardias.
Hearf rafe voriahility as a marker of aufonomic innervation
Periodic changcs in heart rate have been welt known for a long time [12.
13, 141. The major component of this modulation of sinus rhythm is the
respiratory sinusarrhythmia (10-30 peaks per minute), closely correlated to
the actual vagal tone [ 15, 161. These changes are obvious and visible in the
ECG without mathematical calculations. The second element of heart rate
variability (HRV) reflects fluctuations of the blood pressure (3-6 peaks per
minutc), mostly cffected by sympathetic innervation [IS, 161. The third
fraction. slow changcs with 1-2 peaks per minute, is influenced by the level
of circulating catecholamines as wcll as vasomotorical regulations [16].
Hcart rate variability can be analyzed in the time and frequency domain
using 24 hour ECG recordings of Holter tapes [IS, 16, 17, 18, 19, 20, 21,
221. The tcchnical.details have becn presented elsewhere at this workshop.
lo table I. the most widcly used parameters of time domain measurements
of heart rate variability arc presented. r'or better comparability of clinical
results, the different paramctcrs used for calculation of HRV and their
mutual dependency should be considered.
Heart Rate Variability
Tab. 1 : Time domain paranictcrs of heart ratc variability measurements.
Long-rerm
MEAN
mean of all normal RR intervals
(24-hour)
SDNN
standard deviation of all normal RR
intcrvals
TI
width of the triangle interpolation of the
RR histogram [23.24]
SDANN
standard deviation of the 5 minute mean
RR intcrvals
SD
mmn of the 5 min RR standard deviations
cv
mean of the 5 min variation coefficients
pNNSO
percentage of subsequent RR differences
greater than 50 ms [20]
RMSSD
root mean square of subsequent RR
diffcrcnccs
Shorr-term
(5-min)
Bear-lo- hear
HRV and mortality
In 1965, Schncidcr et al. [25] observed that patients after acutc
niyocardial infarction with a decrease in the degree of sinus arrhythmia
compared to controls wcrc more likcly to die during follow-up. In 1978,
Wolf ct al. [26] wcrc the first to show in 230 prospcctivcly studied patients
after myocardial infarction that there was a significantly increased inhospital mortality in thosc patients who had dccrcased heart rate variability
compared to thosc presenting a normal HRV. In this study, the variance of
30 consccutivc RR intcrvals from ECG rhythm strips was calculated to
cxprcss HRV. In 1987, thc first large clinical prospective trial on the
association of dccrwcd heart ratc variability and mortality after acute
myocardial infarction was published [27]. In 808 survivors of acute
myocardial infarction, Holter recordings were performed before discharge.
During the mean follow-up time of 31 months, 127 deaths of all causes
occurred. The Holter parameters calculated to express the distribution and
variability of heart rate were the averagc RR interval of normal cycles
(MEAN) and the standard deviation of the RR intervals around the average
(SDNN). A mean SDNN of 82 f 34 ms was found, with 16% of patients
showing a SDNN < 50 ms and 26% a SDNN t 100 ms. The survival curve
for all patients presenting a SDNN < 50 ms was significantly decreased
compared to all others @<0.0001, Fig. 2). The independent relative risk for
all-cause mortality in these patients was 2.7 compared to controls. These
results were confirmed by Casolo et al. [28] in 54 consecutive patients with
acutc myocardial infarction. During a mean follow-up of 12 months, an allcause mortality of 11% was obscrved. All patients who died showed a
0.5
0
1.0
2.0
30
40
TIME AFTER M I IYsorrI
Fig 2: Survival curves of 808 pts. after myocardial infarction up to 4 years
of follow-up. The SDNN of a 24 h Holter recording was divided into 3
groups. The pts. with an SDNN < 50 ms prcscnted a significantly
dccrcased survival probability (from Klcigcr ct al., Am. 1. Cardiol. 1987,
rcprintcd with permission)
SDNN < 50 ms. Uwreased HRV in post myocardial patients was also
associated with low day-night variations.
Farrell et al. [24] introduced another HRV parameter, based on the
baseline width of a triangular interpolation to the main peak of the
probability density h c t i o n of all RR intervals (TI). Holter recordings, late
potential analysis and conventional risk parameters were obtained. In this
study of 416 consecutive postmyocardial infarction patients, 47 all-causc
137
cardiac deaths and 24 arrhythmic events occurred during a mean-follow-up
period of 612 days (I to 1,112 days). Reduced HRV. expressed as TI, was
the most sensitive parameter for arrhythmic evcnts (92%) but had a low
specificity (77%). The relative risk for cardiac mortality was 6.7 using
HRV measurement and 2.2 using late potential calculation. The
combination of both non-invasive parameters lead to a relative risk of 18.5.
A stcpwise Cox regression analysis for the prediction of arrhythmic events
selected reduced HRV at step one followed by the presence of late potentials
and repetitive ventricular arryhthmias in the Holtcr recordings. All other
variables including left ventricular e j d o n fraction and exercise testing
wcre excluded from the model. Bigger et al. [29] found SDANN the
strongest predictor for arrhythmic deaths in 673 post myocardial infarction
patients, even when the calculation for relative risk was adjusted for five
other risk predictors age, New York Hcart Association functional class,
pulmonary rales in the coronary care unit, left ventricular ejection fraction
and frequency of ventricular arrhythmias in Holter recordings (Tab. 2). In
250 patients that wcre prospectively studied afte! acute myocardial
infarction, we [30] found arrhythmic events in 6% during 6 months of
follow-up. Univariate analysis of differences in all time domain paramcters
(Tab. I ) showcd significant results only in beat-to-beat parameters @NN50.
RMSSD) and slight differences in some short-term parameters (SD, CV,
Tab. 3). The long-term variables wcre not significantly different. Variables
significant in univariatc analysis werc entered as covariates into the Cox
proportional hazards model in order to identify independent prognostic
factors for cstimation of arrhythmic events. The most powcrful prognostic
factor was the presence of latc potentials followed by RMSSD. All othcr
HRV paramctcrs wcrc rcmovcd stcpwise. With the use of rccciver operator
characteristiccurves, we found the optimal cutpoint for RMSSD at 36 ms.
-
Tub. 2: Association of time domain measures of HRV with mortality in 715
pts. post myocardial infarction with a follow-up of 2-4 years. The relative
risk is calculated as the probability of dying if below/above the cutoff (in
brackets). Relative risk for arrhythmic deaths adjusted to age, NYHA
functional class, rales in thc coronary care unit, left ventricular ejection
fradion and frequency of ventricular arrhythmias (modification from
Bigger et al, Am J Cardiol 1992).
Survival curves on the basis of these findings are presented in Fig. 3.
Patients with RMSSD>36ms presented a significantly higher survival
probability compared to the rest of the population (p<0.007). For risk
modelling, thc estimated cumulative survival functions were plotted for
different cutpoints of RMSSD with all covariates set to mean or standard
deviation (Fig. 4). In principle, this risk model yields survival probabilities
for each individual patient based on his test result.
Only few data arc availablc on the response of HRV in the early phase
of myocardial infarction in mcn. Pipilis et al. [3 I] recorded Holter ECGs in
70 patients suffcring acute myocardial infarction .at a mean of 13 hours
after onset of symptoms. They calculated MEAN, SDNN and SDDRR
from two 4 hour periods during the night (1:OO am to 5:OO am) and the day
(8:OO am to 12:OO am). At night, MEAN was significantly longer and
SDNN smaller, but the beat-to-beat variability, expressed as SDDRR,
showed no differences betweenday and night. This reflects a small range of
heart rates during the night with similar beat-to-beat variations. In this
study, the subsequent development of heart failure was the major endpoint
of investigation. A strong inverse conelation of HRV,expressed as SDNN,
and the incidence of heart failure was found, while the differences were
more prominent at night-time. The sensitivity and specificity were 83% and
63% using SDNN < 50 ms as cutpoint.
Heart Rate Variability
138
Tub. 3: Time domain parameters of heart rate variability in 250 pts. after
myocardial infardion. In a follow-up of 6 months 15 arrhythmic events
occured (9 pts. dtvelopped sustained VT, 6 sudden cardiac deaths). The
beat-to-beat parameters reached the highest level of significant differenccs
between both groups of pts., foillowed by some of the short-term
parameters. The long-term parameters were not different. W=Wilcoxon test,
T=Student’s u n p a i d t-test, other abbreviations see text.
Parameters
arrh. event
(n=15)
no arrh. event
(n=235)
79Oi138
9w45
46kt.244
83Oi121
102i34
475*170
p-value
long-term
(24 h)
Mean RR [ms]
SDNN [ms]
TI fms]
n s . (W)
n.s. (T)
llrm illerlnhrcllon In dy.
Fig. 4: Calculation of the event free probability in a Cox proportional
hazards model based on 250 pts. post myocardial infarction with a followup of 180 days using several cut-offs for RMSSD. The individual risk
profile of post infarction patients can be estimated based on these findings.
short-term
(5 min.)
SDANN [ms]
SD [nis]
cv
I
n.s. (W)
p=0.02 0
108*63
38.4i15.3
0.046i0.014
heof-to- bear
2.89i2.48
26.2*7.9
pNN50 [%]
RMSSD [ms]
0
20
10
60
7.2*6.2
36.W 13.5
w
1w
Tlms In day.
1m
p=0.005 (W)
14
T I
(I
Fig. 3: Survival analysis of RMSSD in 250 patients after myocardial
infarction. In a follow-up of 6 months 15 arrhythmic events occurred
(9 pts. developped sustained VT, 6 sudden cardiac deaths). Patients with
RMSSD > 36 ms presented a significantly higher survival probability
compared to the rest of the population @<0.007).
In addition to the well selected patients after myocardial infarction of the
prcviously demonstrated trials, Alga et al. [32] studied retrospectively
6,693 consecutive patients of four Rotterdam hospitals with a variety of
underlying cardiac diseases in whom Holter monitoring had been performed
for various clinical indications. There were 245 sudden deaths (3.7%)
during two years of follow-up. A comparison of HRV variables and other
clinical panmcters in these sudden death patients as well as in 230 patients
randomly selected out of the rest of the large population showed SD and
SDANN as best risk predictors, exhibiting a risk for sudden death about
four timcs higher than in controls. After adjustment for age. evidence of
cardiac dysfunction and history of previous infarction, the relative risk
cstiniatcs for SD and SDANN reached 2.2 and 2.6, respectively.
The rclative risk for cardiac and arrhythmic deaths based on all variables
in the study by Algra et al. [32], by Bigger et al. (291, by Kleiger et al. [27]
and in ours [30] was significantly lowcr than in the study by Farrell et al.
[24]. Wcthcr the latter results represent a selection bias or are due to some
mcthcdological factors remain unresolved. Opposite to the findings in
patients affcr acute myocardial infarction, patients presenting unstable
angina showed no significant differences in HRV compared to n o d
subjects [281.
HRV and other clinical parameters
The results of scvcral studics showed HRV to be an independent
predictor for mortality after myocardial infarction [29, 24, 27, 301.
However, a strong rclationship to convcntional parameters used for risk
stratification was cspcctcd. Klcigcr ct al. [27] found that decreased HRV
increased the risk of death after myocardial infarction irrespective of
MEAN, left ventricular ejection fraction, ventricular ectopic activity, New
York Heart Association functional class or drug treatment with &blockers
and digitalis. Casolo et al. [28] found HRV to be independent but closely
associated to other clinical parameters. It was significantly decreased in Qwave compared to non-Q-wave infarction togcther with an inverse relation
to peak CK-MB level, left ventricular ejection fraction, enddiastolic left
ventricular diameter and Killip class. Several studies found no influence of
infarct localisation [28,24] or the number of diseased coronary arteries [24]
on HRV measurements. In contrast, some studies that used ECG recordings
during the early phase of myocardial infarction (6-24 hours after onset of
pain) found substantial differences in HRV with regard to infara
localisation [31,33]. In 34 patients with anterior wall infarction, the degree
of HRV reduction and heart rate increase was significantly stronger
compared to 36 patients presenting inferior wall infarction [31]. Few hours
after onset of infarction, both groups were similar in clinical evidence of
hcart failure, enzyme release and drug trcatment. These divergent results
can be cxplaincd by the diffcrcnt rccording times and their relation to the
course of recovery after myocardial infarction. Early differences in HRV
bctwecn anterior and inferior wall infarct might be due to reflex increase in
sympathetic tonc bascd on the stimulation of sympathetic afferent nerves
particularly by anterior ischemia and of vagal affcrcnt fibers especially in
inferior infarction.
Odemuyiwa et aI. [34] examined the prognostic value of reduced HRV.
espressed as TI (Table I), in post infarction patients in comparison to left
ventricular ejection fraction, ventricular premature beats > IOihour and
presence of ventricular latc potcntials. They found that at a high level of
sensitivity, the ejection fraction was the risk predictor with the lowest
specificity, whereas HRV showcd tlic highest specificity. The combination
of scvcral risk parameters incrcascd specificity if HRV was included. The
highest specificity of 95% at a sensitivity level of 75% was reached if HRV
and left vcntricular cjcction fraction wcrc combined. The addition of the
prcsence of late potentials yicldcd no further iniprovcment.
HRV and thrombolytie treatment after myocardial infarction
Early thrombolysis after myocardial infarction is known to reducc the
degree of myocardial damage and the total prognostic outcomc. Differing
grades of sympatho-vagal impairment will occur due to the size and
localisation of the infarctcd area. Casolo et al. 12x1 found that carly
thrombolytic therapy affected the degrcc of HRV reduction recorded on day
2 to 3 after myocardial infarction. In I8 patients receiving systemic
recombinant tissue-type plasminogen activator (rt-PA) within 2 hours after
admission, SDNN was significantly higher compared to those not rcceiving
thrombolysis. However, the relation to CK elevation 3s a marker of
thrombolytic success was only slight. In contrast, Farrell ct al. [24] found
no effect of early thrombolysis on HRV, expressed as TI, left ventricular
ejection fraction or the presence of ventricular late potcntials in 200 patients
suffering acute myocardial infarction. Although thc incidence of death
during follow-up (612 days, range 1 to I , I 12 days) was significantly lowcr
in patients receiving thrombolytic therapy, the incidcncc of subscquent
arrhythmic events tcnded to be lower but did not rcach statistical
significance. The major differences bctwecn both studies were the timc of
Holter recording (Casolo et al. on day 2 to 3, Farrell et al. on day 6 to 7
after myocardial infarction) and the HRV indcx uscd (Casolo ct al. SDNN, Farrell et at. -TI).
Heart Rate Variability
Time course of HRV after myocardial infarction
The variability indices of heart rate represent the degree of sympathctic and
parasympathetic control of the hcart and the level of balance between both
limbs. Time dcpcndcnt changes in HRV might be expected in rclation to
variations in autonomic control of the heart during thc recovery aftcr
myocardial infarction. Lombardi ct al. [IS] recorded Holter ECGs in 70
patients 2 wwks, 6 months and 12 months aftcr myocardial infarction. In a
subgroup of these patients. the effects of sympathetic stimulation were
studicd at 2 wccks and 1 ycar after myocardial infarction using a tilt tablc
protocol. Thcy found that the sympathetic predominance is detectable by
HRV analysis 2 wccks after acutc myocardial infarction and that a recovery
of vagal tone and a nomialisation of sympathovagal interaction occurs over
6 to 12 months post myocardial infarction, not only during resting
conditions, but also in response to a sympathetic stimulus. Casolo et al. [28]
performed 3 Holter recordings on post myocardial infarction patients on day
2 to 3, after 30 days and 60 days. He found that HRV, expressed as SDNN,
increased in all surviving patients to near normal counts in 60 days after
infarction with linear intermediate values obtained at 30 days. These results
indicate a progressive increase in HRV over time in patients with previously
decreased HRV, suggesting a normalisation of impaired parasympathetic
control of the heart directly after acute myocardial infarction.
Short-time variations in autonomic reflexes of the heart over the day were
expected to influence the outcome and stability of HRV measurements.
Malik et al. (35) examined circadian variations of HRV in 40 patients after
myocardial infarction, 20 of whom developed arrhythmic events over 6
months, 20 other patients without events during follow-up were selected to
match those with events with regard to several clinical parameters. From
each Holter recording, a total number of 5.1 13 portions of different duration
(20 minutes to 24 hours) and time of onsct (varied. by 20 minutes) were
analyzed. Thcy found that the most significant differences in HRV,
expressed as TI, could be found in recordings starting at 6 0 0 am, lasting
for approximatcly 8 'hours and starting at 8:00 pm, lasting for
approximately 6 hours. Diffcrcnces in HRV between day and night-time
wcrc also obscrved by Biggcr ct al. [36]. They found that patients
presenting low HRV, expressed as SDNN, showed significantly higher
heart rates with substantially low differences between day and night-time
compared to controls with normal HRV. Heart rate spikes, defined as
increases in heart ratc of at least 10 beatshinute that lasted from 3 to 15
minutes, occurred much less frequently in patients with a low overall
variability of hurt rate.
It is known that the level of autononiic control of the heart and the baianccd
innervation of sympathetic and pnrasympathctic effects change with age.
Ewing ct al. 1371 described an invcrsc corrclation of age and the total
number of RR intervals counts in 24 hours in subjccts without cardiac
disorder. Identical findings in HRV were prcscntcd by Casolo ct al. in post
niyocardial patients 1281.
The reproducibility of HRV after myocardial infarction
Before HRV can be used in clinical practice as a routine risk predictor, the
stability and reproducibility of results need to be known. Hohnloser et al.
[38] compared grouped and individual changes in reproducibility of HRV in
normal subjects, patients with documented coronar). artery disease and
those with m o t e myocardial infarction using linear correlation coefficients
of SD, pNN50 and RMSSD measurements as well as frcqucncy domain
parameters. Holter recordings were performed on day 1. 7 and 28 in all
subjects. Short-term reproducibility (day 1 to 7) was superior to long-tcnn
reproducibility (day 1 to 28 or day 7 to 28). The best correlation
coefficients were found in the beat-to-beat parameters pNN50 ( ~ 0 . 9 3and
)
RMSSD ( ~ 0 . 8 9 ) .However, the individual stability of HRV measurements
over time was largely varying in some patients with day-today differences
up to 23.5 i 14.6 %. Normal subjects and patients with cardiac disease
presented comparable levels of individual variations. Similar results have
been observed by others [39.40,41]. In 33 normal subjects and 22 patients
with congestive.heart %lure and coronary artery disease Hoogenhuyre et al.
[39] found strong correlationsof SD, SDANN and CV in two consecutive
Holter recordings. Individual day-today variations occurred up to 50%.
Low HRV results have been more consistent compared to high values. Since
CV as the mean of the 5 min Mn'ation cocfficients minimises the effect of
heart rate on the calculation of HRV, this parameter was most stabile over
time.
Summary
These results suggest that analysis of heart rate variability recorded even
very early after acute myocardial infarction ( I to 2 d a y after onset of pain)
139
is feasible in clinical routine and strongly related to subsequent arrhythmic
events and cardiac mortality. Decrcascd HRV is considered not only a
marker of impaired vagal activity of the heart but complete autonomic
impairment, strongly associated with the degree of myocardial damage.
HRV represents the integrated response of the cardiovascular system to a
variety of different influences: the plasma lcvcl of catecholamines, thc
baroreflex activation and the direct sympathetic and vagal activity. The
HRV profile is dynamic HRV reduction caused by myocardial damage
changes over time presenting a progressive increase up to normality over a
2-month follow-up. The. observed early differences between anterior and
inferior myocardial infarction disappear later in the healing phase.
However, group analysis results of different HRV indices are stabile over
time and highly reproducible, but presenting large individual variations.
HRV parameters which are adjusted to heart rate (e. g. CV) seemed to be
more stabile.
Thc role of HRV analysis in risk stratification of patients after myocardial
infarction is strongly related to the actual model of the genesis of ventricular
arrhythmias. Multiple experimental and clinical studies described the
development of life threatening ventricular arrhythmias as a multifactorial
cvcnt which can not be described adequately using just one risk parameter
for stratification.The arrhythmogenie substrate representing the underlying
inhomogeneity of electrical behaviour of adjacent myocardial areas might be
detectable by the analysis of ventricular late potentials or frequency
disturbances from the signal averaged ECG. The autonomic modulation of
this substrate is represented by an altered heart rate variability. Possible
trigger factors to initiate arrhythmias in a modulated arrhythmogenic
substrate like ventricular premature. beats or transient myocardial ischemia
can be detected by conventional arrhythmia and ST segment analysis from
Holter tapes. The optimized combination of these non-invasive risk
prcdicton together with well known evident clinical risk parameters, like
left ventricular ejection fraction, may lead to a valid set of screening
paramcters for individual risk estimation after myocardial infxction
-
This papcr was partly supported by the Deutsche Forschungsgemcinschaft
(DFG Project Nr. Br 59/2-l), Bonn-Bad Godesberg, Gcrmany
References
1. Zipes DP. Influence of myocardial ischemia and infarction on
autonomic innervation of heart. Circulation 1990; 8 2 1095-1105.
2. Levy MN. Sympathetic-parasppathetie interactions in the heart. Circ
Res 1971; 29: 437-445.
3. Webb SW, Adgey AAJ, Pantridge JF. Autonimic disturbance at onset
of acute myocardial infarction. Br Med J 1972; 3: 89-92.
4. Stanton MS, Tuli MM, Radtke NL. Regional sympathetic denemtion
following myocardial infarction in humans detected noninvasively using
1-123-metaiodobenzyIguanadine. J Am Coll Cardiol 1989; 14: 15191526.
5. Zipes DP. Ischemic modulation of cardiac autonomic innervation. J Am
Coll Cardiol 1991; 17: 1424-1425.
6. Lauara R, El-Sherif N, Scherlag BJ. Electrophysiological properties of
canine Purkinje cells in one day old myowdial infarction. Circ Res
1973; 33: 122-734.
7. Camcron JS, Han J. Effetcs of epinephrine on automaticity and the
incidence of arrhythmia in Purkinje fibers surviving myocardial
infarction. J Pharmacol Exp ?her 1982; 223: 573-579.
8. Marec H, Pangman KH, D a d o P, Rosen MR. An evaluation of
automaticity and triggered activity in canine heart one to four days after
myocardial infarction. Circulation 1985; 71: 1224-1236.
9. Martins JB. Autonomic control of ventricular tachycardia - sympathetic
neural influences in spontaneous tachycardia 24 hours after coronary
occlusion. Circulation-1985; 72: 933-942.
10. Kimura S, Bassett AL, Kohya T, Kozlovskis PL, Myerburg RJ.
Automaticity, triggered activity, and responses to adrenergic stimulation
in cat subendocardial Purkinje fibers after healing of myocardial
infarction. Circulation 1987; 75: 651660.
I I. Priori SG, Mantica M, Schwartz PJ. Delayed afterdepolarizations
elected in vivo by left stellate ganglion stimulation. Circulation 1988;
78: 17X-IX5.
12. Fleisch A, Beckmann R. Die raschen Schwankungen der Pulsfrequenz
rcgistricrt mit den1 Pulsschrciber. Z ges exper Med 1932; 80: 487-5 10.
13. Wilhelnison B. Die Schwankungen der Pulsfrequenz bei Belastung des
Herzens. 2 exper Med 1932; 85: 248-261.
140
Heart Rate Variability
14. Schlomka G. Untersuchungen iiber die physiologische UnregeMigkeit
der
dcs Hemchlages. 111. Mitteilung: h e r die Ab-gkeit
respiratorischcn Ruhc-Arrhythmie von der Schlagfiequm und vom
Lebensalter. Z Kreislaufforschung 1937; 2 9 510-524.
Garimoldi M, Cerutti
15. Lombardi F. Sandrone G, Pcrnpruner S, Sala
S, Baselli G. Pagani M, Malliani A. Heart Rate Variability As An
Index Of Sympathovagal Interaction After Acute Myocardial Infarction.
Am J Cardi01 1987; 60: 1239-1245.
16. Pomeranz B, Maeaulay RJB, Caudill MA, Kutz I, Adam D, Gordon D,
Kilborn KM, Bargcr AC Shannon,DC Cohen,RJ Benson,H. Assessment
of autonomic function in humans by heart rate spectral analysis. Am J
Physiol 1985; 248: H151-HI53.
17.Appel ML. Bergcr RD, Saul P, Smith JM, Cohen RJ. Beat to beat
variability in cardiovascular variables: noise or music ? I Am Coll
Cardiol 1989; 14: 1139-1 l4R.
18. Baselli G, Ccrutti S, Civardi S, Lombardi F, Malliani A, M e m M,
Pagani hl. Riuo G. Heart rate variability signal processing: a
quantitative approach as an aid to diagnosis in cardiovascular
pathologies. Int J Bio Medical Computing 1987; 20: 51-70.
19. Carter JE Jr, Childcrs RW. The importance of autonomic dysfunction
as assessed by heart rate variability. Clin Cardiol 1992; 15: 769-772.
20. Ewing DJ, Neilson JMM, Travis P. New mcthod for assessing cardiac
pansympathetic activity using 24 hour electrocardiograms. Br Heart J
1984; 52: 396402.
2 I. Hayano J, Sakakibara Y, Yamada A, Yamada M, Mukai S, Fujinami T,
Y o k o y m K. Watanabe Y, Takata K. Accuracy of assessment of
cardiac vagal tone by heart rate variability in normal subjects. Am J
Cardiol 1991; 67: 199-204.
22. Molgaard H. Evaluation of the Rcynolds Pathfinder I1 system for 24 h
heart n t c variability analysis. Eur Heart J 1991; 12: 1153-1162.
23. Cripps TR, Malik M, Farrell TG, Camm AJ. Prognostic value of
reduced heart rate variability aftcr myocardial infarction: clinical
evaluation of a new analysis mcthod. Br Heart J 1991; 65: 14-19.
24. Farrcll TG, Bashir Y, Cripps T, Malik M, Poloniecki J, Bennett D,
Ward DE. Camm AJ. Risk stratification for arrhythmic cvents in
postinfarction paticnts based on heart rate variability, ambulatory
electrocardiographic
variables
and
the
signal-averaged
electrocardiogram. J Am Coll Cardiol 1991; 18: 687-697.
25. Schneider RA, Costiloe JP. Relationship of sinus arrhythmia to age and
its prognostic significance in ischemic heart disease. Clin Res 1965; 13:
219.
26. Wolf MW, Varigos GA, Hunt D, Sloman JG. Sinus arrhythrma in acute
myocardial infarction. Med J Austral 1978; 2: 52
27. Kleiger RE, Miller JR, Bigger JT, Moss AJ and the Multicenter PostInfarction Research Group. Decreased heart rate variability and its
association with increased mortality after acute myocardial infarction.
Am J Cardiol 1987; 59: 256-262.
28. Casolo G C, Stroder P, Signorini C, Calzolari F, Zucchini M, Balli E,
Sulla A, Lazzcrini S. Heart Rate Variability During The Acute Phase
Of Myocardial Infarction. Circulation 1992; 85: 2073-2079.
29. Bigger JT, Fleiss JL, Steinman RC, Rolnitzky LM, Kleiger RE,
Rottman JN. Corrclations Among Time And Frequency DOMeasures Of Heart Period Variability Two Weeks After Myocardial
Infarction. Am J Cardiol 1992; 69: 891-898.
30. Reinhardt L, Miikijani M, Fetsch T. Martinez-Rubio A, BBcker D,
Block M. Borggrcfe M, Breithardt G. The progno~tics i g d k m c e of
heart rate variability aftcr acute myocardial infarction. Experience from
the Post-Infarction Late Potential study. Submitted for publication.
31. Pipilis A, Flather M, Ormerod 0,Sleight P. Heart rite variability in
acute myocardial infarction and its association with infarct site and
clinical course. Am I Cardiol 1991; 67: 1137-1139.
32. Algra A, Tijsscn J G P, Roelandt J R T C, Pool J, Lubsen 1. Heart rate
variability from 24-hour elcctricardiography and the 2-year risk for
suddcn death. Circulation 1993; 88: 180-185.
33. McAreavcy D, Neilson JMM, Ewing DJ, Russell DC. Cardiac
parasympathetic activity during thc cnrly hours of acute myocardial
infarction. Br Heart J 1989; 62: 165-170.
34. Odcmuyiwa 0, Malik M, Farrcll T. Bashir Y, Poloniecki I, Camm J.
Comparison of heart rate variability indcs and Icft ventricular ejection
fraction for all-causc mortality. arrhyThmic cvcnts and sudden death
aRcr acute myocardial infarction. Am J Cardiol 1991; 68: 434439.
35. Malik M, Farrell T, Camm AJ. Circadian rhythm of heart rate
variability after acute myocardial infarction and its influencc on the
prognostic value of heart rate variability. Am J Cardiol 1990; 66: 10491054.
36. Bigger TJ, Kleiger RE, Fleiss JL, Rolnitzky LM, Steinman RC, Miller
JP, Multicenter Post-Infarction Research Group. Coniponents of heart
rate variability measured during healing of acute myocardial infarction.
Am J Cardiol 1988; 61: 208-215.
37. Ewing DJ, Neilson JMM, Shapiro CM, Stewart JA, Reid W. Twenty
four hour heart rate variability: effects of posture, slecp, and time of day
in healthy controls and comparison with bedside tests of autonomic
function in diabetic patients. Br Heart J 1991; 65: 239-244.
38. Hohnloscr SH, Klingenheben T, Zabel M, Schriider F, Just H.
Intraindividual Rcproducibility Of H a r t Rate Variability. PACE 1992;
15: 2210-2214.
39. Hoogenhuyze D van, Wcinstein;N Martin GI, Wciss JS, Schaad JW,
Sahyouni N, Fintel D, Rcmme WJ, Singer DH. Rcprducibility and
rcalction to mean heart rate of h a r t nte varibility in normal subjects
and in paticnts with congestive heart failure secondary to coronary
artery disease. Am J Cardiol 1991; 68: 1668-1676.
40. Klcigcr RE, Biggcr JT, Bosncr MS, Chung MK, Cook JR, Rolnitzky
LM, Steinman R, Fleiss JL. Stability over timc of variables measuring
heart rate variablity in normal subjects. Am J Cardiol 1991; 68: 626630.
41. Bigger JT, Fleiss JL, Rolnitzky LM, Steinmann RC, with data
contributed by CAPS and ESVEM Investigators. Stability ovcr timc of
heart period variability in paticnts with prcvious myocardial infarction.
Am J Cardiol 1992; 69: 718-723.