(Circulation. 1997;96:202-213.)
© 1997 American Heart Association, Inc.
Articles |
Risk Stratification After Myocardial Infarction Using Signal-Averaged Electrocardiographic Criteria Adjusted for Sex, Age, and Myocardial Infarction Location
From the Institut de génie biomédical, Ecole Polytechnique, Université de Montréal, and Centre de recherche, Hôpital du Sacré-Coeur, Montréal, Québec (P.S., P.M.); Département de Médecine, Institut de Cardiologie de Montréal (J.-L.R., N.P., M.T.); Memphis (Tenn) Vascular Foundation (J.F.); Department of Medicine, Ottawa (Ontario) Heart Institute (R.F.D.); Department of Medicine, Royal Victoria Hospital, Montréal (D.J.S.); Department of Medicine, Victoria General Hospital, Nova Scotia (M.G.); Département de Médecine, Centre Hospitalier Régional de l'Amiante, Thetford Mines, Québec (R.D., C.L.); Department of Medicine, Health Science Center, St John's, Newfoundland (B.S.); Département de médecine sociale et préventive, Université de Montréal (L.P.); and Department of Medicine, Foothills Hospital, Calgary, Alberta (W.W.), Canada.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Background The objectives were to investigate the factors influencing signal-averaged ECGs (SAECGs) recorded in patients after myocardial infarction (MI) and to develop criteria for predicting arrhythmic events (AEs) that account for these factors.
Methods and Results SAECGs were recorded 5 to 15 days after MI in 2461 patients without bundle-branch block. The duration (QRSd), terminal potential (VRMS), and terminal duration (LAS) of the filtered QRS were measured. During follow-up (17±8 months), AEs (arrhythmic death; ventricular tachycardia, VT; ventricular fibrillation, VF) occurred in 80 patients (3.3%). Receiver operating characteristic curves showed that QRSd discriminated patients with all types of AEs, but VRMS and LAS discriminated only VT patients; QRSd minus LAS also discriminated AE patients. Sex, age, and MI location significantly affected the SAECG; survivors without VT or VF were divided into subgroups (2 sexx4 agex2 MI), and QRSd values exceeding the 70th percentile in each subgroup predicted AEs with a sensitivity of 65.4%. An unadjusted QRSd criterion showed the same overall sensitivity and specificity but with less uniform values for each subgroup. A Cox model was constructed by use of multiple prognostic indicators, and in rank order, QRSd, previous MI, and Killip class were predictive of AEs.
Conclusions SAECG adjustments for sex, age, and MI location did not improve sensitivity and specificity but produced a more uniform predictive performance. The proposed criteria are based only on QRSd, because late potentials (VRMS and LAS) did not discriminate patients with sudden death. Duration of high-level activity during QRS (QRSd-LAS) can predict AEs, suggesting that the arrhythmogenic substrate involves a large mass of myocardium.
Key Words: tachycardia • fibrillation • arrhythmia • risk factors • electrocardiography
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The SAECG has been identified in numerous prospective studies1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 as one of the most powerful diagnostic tools for the assessment of patients at increased risk of ventricular tachyarrhythmias after MI.16 Sex and age have a significant influence on the standard ECG of normal subjects and are taken into account for defining the limits of normality.17 However, sex and age are not included in SAECG criteria.18 The QRSd measured on the SAECG is longer in men than in women in both healthy subjects19 20 21 and MI patients,22 and significant correlations have been found between age and the SAECG in MI patients.23 Also, in MI patients, it is not possible to accurately define late potentials with respect to the SAECG of normal subjects, because the activation sequence is modified by the infarct. Activation around the damaged wall occurs later in inferior MI than in anterior MI, which results in a longer QRSd and a higher incidence of late potentials in patients with inferior MI.3 24 25 SAECG criteria should thus be defined with respect to MI location in arrhythmia-free MI survivors.
The CAMI study was undertaken in 1990 to investigate the impact of new interventions on in- and out-of-hospital events, patient profile, and risk stratification methods.26 The number of patients enrolled in this study (n=2461) exceeds those of previous prospective SAECG studies1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 (n=102 to 1158), thus enabling multivariate analysis to be performed so that all factors19 20 21 22 23 24 25 that could affect the SAECG can be investigated. This large number of patients also allows the division of the population into subgroups for the definition of SAECG criteria adjusted for factors such as sex, age, and MI location.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Study Patients
All consecutive patients in eight Canadian hospitals who had sustained acute MI and met the selection criteria were considered for entry into the prospective study. Recruitment took place between June 1, 1990, and June 30, 1992, in six centers and between June 1, 1990, and December 31, 1992, in three centers. Patients were identified by daily review of all hospital admissions to the coronary care or intensive care units and to the cardiology services. In addition, cardiologists were questioned for identification of other patients who had had an acute MI but were hospitalized elsewhere in the hospital. The study design was approved by the institutional review boards of each center.
The diagnosis of acute MI required at least two of the following
criteria: (1) characteristic chest pain lasting 
20 minutes, (2) new Q
wave with a duration of >40 ms evolving in the postadmission series of
ECGs in at least two contiguous leads, and (3) creatine kinase 1.5
times the upper limit of normal and/or creatine kinase MB fraction
>5% when simultaneous reference creatine kinase exceeded
the upper limit of normal.
Of a total of 4133 patients with acute MI who were documented,26 379 patients died in the hospital before the SAECG could be recorded; of the remainder, the following groups of patients satisfied the exclusion criteria: 232 patients who were >75 years old at enrollment time, 37 patients who had malignant neoplasm, 95 patients who were in critical or unstable condition during the SAECG recording time window, 286 patients who refused to participate or whose physician refused, 131 patients because of geographic remoteness that would impair the follow-up, and 205 patients with bundle-branch block or permanent pacemakers. In addition, the following patients were excluded even if they were eligible: 302 patients because of technical reasons (equipment failure, holidays) or patient communication problems and 5 patients with an incomplete file or lost to follow-up. The remaining 2461 patients were included in the SAECG study.
All included patients had demographic variables and cardiac history assessed. In addition, the clinical course of their MI (heart failure, ventricular arrhythmias, recurrent MI, etc) and the therapeutic interventions (thrombolytic treatment, coronary angioplasty, coronary bypass, drugs, etc) were documented. The diagnosis of recurrent MI, if nonfatal, required the same criteria as the entry criteria of the study itself and if fatal, had to meet the criteria of the mortality committee (see below). The clinical management of patients was at the discretion of individual physicians who were not aware of the SAECG results.
Signal-Averaged ECG
The SAECG was recorded 5 to 15 days after MI, because SAECGs
recorded during this period have been identified as having the most
significant relation to arrhythmic events occurring during the first
year.5 Three bipolar orthogonal leads were used. After
skin preparation, electrodes were applied over the left and right
midaxillary lines at the V5 level (+X and -X), along the
left midclavicular line over the infraclavicular area and the left
iliac fossa (+Y and -Y), and over the cardiac apex and the
corresponding position on the back (+Z and -Z). Recordings
were performed in the supine position with Predictor I systems
(Corazonix). The sampling rate was 2000 s-1
and the bandwidth was 0.05 to 300 Hz. The sampling window had a width
of 300 ms. The beats were aligned by use of an updated template to
prevent the degradation of the averaging process.27
Sufficient beats (
400) were averaged to achieve a noise level of 0.3
µV in all SAECG recordings.28 Noise was
calculated as the mean of the root-mean-square noise of the averaged X,
Y, and Z leads during the 50 ms at the end of the acquisition
window.
The processing of all SAECG recordings was performed at the
coordinating center. Bidirectional high-pass filtering was repeated for
three different high-pass cutoff frequencies: 25, 40, and 80 Hz. In the
rest of this article, SAECG results are presented for 40 Hz,
which showed the highest sensitivity for predicting arrhythmic events
(see "Results"). QRS onset and offset were determined
automatically (Predictor I, version 5.0 software) and visually
validated. In rare instances (1% to 2%) when the end points could not
be automatically detected or were not correctly detected, eg, because
the P wave was smeared into the QRS complex, the end points were set
manually. Three variables were computed (see Fig 1
):
QRSd, LAS, and VRMS. In addition, QRSd-LAS was also computed.
|
Other Variables
MI location, as assessed by 12-lead ECG, was classified into
three categories: (1) non–Q-wave, (2) inferoposterior, and (3)
anterolateral. Of the 2461 patients included, 2064 (80%) had a
technically adequate measure of the LVEF. Ambulatory ECGs were
recorded and judged satisfactory (>18 hours) in 1803 patients.
After automated full disclosure analysis followed by physician
confirmation, the average frequency of PVCs (PVCs per hour) and the
occurrence of nonsustained ventricular
tachycardia were recorded.
Follow-up
Included patients had follow-up visits after 6 and 12 months
after MI to record new SAECGs and to complete a questionnaire
documenting the evolution and treatment of their disease. At 24 months
and also if a patient could not return to the participating center for
a new SAECG measurement, the questionnaire was completed by phone. This
questionnaire documented cardiac interventions and complications (eg,
angina, MI, syncope, arrhythmias) that occurred during the
follow-up period as well as the current status of the patient (eg,
Goldman scale, smoking, exercise). Excluded patients were phoned
after 1 and 2 years to assess their vital status. All
cardiovascular events were verified by chart review or
by patient records obtained from other hospitals and/or
physicians.
The details concerning the deaths of all included and excluded patients were ascertained at the participating centers. A central mortality committee, blinded with regard to SAECG results, then reviewed the information and classified deaths according to predefined criteria. Deaths were first classified according to whether they were cardiovascular or noncardiovascular. Cardiovascular deaths were classified into (1) recurrent MI, subclassified as definite according to the same criteria as used for the qualifying MI or proven at autopsy, or probable if the patient died within the same hospital admission or within 7 days of recurrent MI; (2) pump failure, defined as death in a patient with documented NYHA class IV failure or shock with at least two of the following three criteria: evidence of poor perfusion, blood pressure <90 mm Hg, or refractory congestive symptoms; (3) arrhythmic death, subclassified as definite or presumed according to the CAST criteria29 ; (4) procedural, defined as an unexpected death directly resulting from a complication related to a cardiovascular procedure; and (5) other vascular deaths (eg, stroke, aneurysms, or peripheral or renovascular disease). When a death resulted from an acute MI, it was classified as such whether the final event was low output or arrhythmias.
Study End Points
Three specific arrhythmic events were defined a priori as study
end points: (1) arrhythmic death, as defined above29 ; (2)
sustained VT, defined as a documented tachycardia of
ventricular origin with a rate 120
min-1 and lasting 30 seconds; and (3)
documented VF requiring defibrillation. When a patient experienced more
than one event, the first was taken as the study end point.
Statistical Analysis
Continuous variables are expressed as mean±SD.
Univariate comparisons between groups were conducted with
Student's t test for continuous variables, the
Mann-Whitney U test for continuous variables when the
normality assumption was questionable, and 
2
analysis for categorical variables. The type I error was
fixed at P=.05.
An ANCOVA was used to assess the effects of MI location, sex, and thrombolytic treatment factors; of the interactions between these factors; and of covariates such as age and LVEF on the QRSd, the LAS, and the logarithm of the VRMS. A logarithmic transformation was applied to the VRMS to normalize its distribution. A contrast analysis was used to detect differences between MI locations, and the Bonferroni method was applied to adjust for multiple comparisons.
ROC curves were computed to assess the sensitivity and specificity of the QRSd, VRMS logarithm, and LAS variables for predicting arrhythmic events during follow-up. The Hanley-McNeil test was used to test and to compare the discrimination power of these variables.30
The Cox proportional-hazards regression model was applied in a stepwise manner (both forward and backward) to evaluate the power of SAECG variables, LVEF, Holter variables, and other clinical features that are known predictors of survival after MI.26 The relation between arrhythmic end points and the SAECG test was assessed with the Kaplan-Meier product-limit estimate of the survival function. The statistical tests were performed with the Systat software (version 5.01 DOS), with the exception of the Hanley-McNeil test, which was performed with a custom program.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Follow-up of Included and Excluded Patients
The mortality of included and excluded patients is shown in Table 1
. The overall mortality for patients who were
ineligible because of cancer or critical condition was significantly
higher than for the included patients (28% versus 8.8%). Mortality in
patients ineligible because of bundle-branch block or pacemaker was
also higher. There was no significant difference in mortality between
included patients and those excluded because of patient or physician
refusal and for geographical or technical reasons. Among included
patients, the mean follow-up duration in the survivors was 17±8 months
(range, 12 to 24 months). After the SAECG recording, 217
patients (8.8%) died, 143 (5.8%) of cardiovascular
causes. Eighty patients (3.3%) had a primary end point during
follow-up: arrhythmic death in 51 patients (2.1%), sustained VT in 21
(0.9%), and resuscitated VF in 8 (0.3%).
|
Clinical Characteristics
The clinical characteristics of patients with and without an
arrhythmic event during follow-up are summarized in Table 2. The two groups were similar with respect to age, sex,
MI location, and the percentages of patients who underwent
coronary artery bypass graft surgery or angioplasty during
their hospitalization. Patients with previous MI or angina, those with
a lower LVEF or heart failure (Killip class III and IV), and those who
had not received thrombolytic treatment were more
likely to have subsequent arrhythmic events.
|
Factors Affecting the SAECG
An ANCOVA was performed to assess the effects of MI location, sex,
thrombolytic treatment, age, and LVEF on the SAECG
variables. Table 3 summarizes the ANCOVA results for
the 2064 patients who had LVEF data (note, there was no significant
difference in the incidence of arrhythmic events in patients with and
without LVEF data: 3.8% and 3.1%, respectively).
|
For the QRSd, a significant difference in the adjusted group means was found for each of the three independent variables: MI location, sex, and thrombolytic treatment (P<.001). A significant correlation was also found between QRSd and the two covariates (age and LVEF). The F ratios indicate that sex and LVEF had the strongest influence on QRSd, whereas thrombolytic treatment had the weakest influence. Contrast analysis shows that there was no significant difference in the QRSd between non–Q-wave and inferoposterior MI locations, whereas there was a very significant difference between either of these MI locations and anterolateral MI.
Some antiarrhythmic drugs, specifically sodium channel blockers, are known to affect the SAECG.31 In this study, only 78 patients (3.1%) received an antiarrhythmic drug within five half-lives before the SAECG recording. The SAECGs of patients receiving amiodarone, procainamide, and quinidine showed longer QRSd and LAS than those of patients who had not received any drug, whereas no significant differences were found for patients receiving sotalol and verapamil. Thus, antiarrhythmic drugs were infrequently prescribed and could have modified the SAECG in only a few patients in this study.
Risk Stratification With Unadjusted SAECG Criteria
The presence of late potentials is defined by prolonged QRSd and
LAS and reduced VRMSs. ROC curves representing the
sensitivity versus specificity obtained when different thresholds were
applied to the SAECG variables for the prediction of all arrhythmic
events are shown in Fig 2. In addition to characterizing
the interplay between sensitivity and specificity, ROC curves can also
be used to compare different tests by measuring the area under the
curve, which reflects the discrimination power of the variable. All
three SAECG variables as well as a variable
representing the duration of high-level activity during the
QRS complex obtained by subtracting the LAS from the QRSd (QRSd-LAS)
showed a significant discrimination power for predicting all arrhythmic
events. However, the discrimination powers of QRSd and QRSd-LAS were
significantly higher than those of LAS and VRMS, and there was no
significant difference between LAS and VRMS and between QRSd and
QRSd-LAS.
|
), LAS (
), VRMS (
), and QRSd-LAS. Area under a
curve measures discrimination power of variable: for prediction of
all arrhythmic events or of VT, all four variables have a
discrimination power significantly higher than random prediction,
represented by diagonal line (**P<.001,
*P<.05); however, discrimination power of QRSd is
significantly higher than that of LAS and VRMS (P<.05), and
there is no significant difference between LAS and VRMS. For prediction
of all arrhythmic events, there is no significant difference between
QRSd and QRSd-LAS; for prediction of VT, QRSd offers a higher
discrimination power than QRSd-LAS.
To further investigate the poorer performance of LAS and VRMS
for predicting all arrhythmic events, ROC curves were plotted for the
prediction of specific arrhythmic events. For predicting sustained VT
(Fig 2
, right), all four SAECG variables showed a significant
discrimination power, but the discrimination power of QRSd was the
highest. For the prediction of arrhythmic death (Fig 3, left), only the QRSd and the QRSd-LAS showed a
significant discrimination power, with no significant difference
between QRSd and QRSd-LAS. For the prediction of VF (Fig 3, right),
only the QRSd showed a significant discrimination power.
|
QRSd Percentile Distributions of Arrhythmia-Free Survivors
SAECG criteria adjusted for sex, age, and MI location were
developed for the QRSd variable alone because it could predict all
types of arrhythmic events, whereas the LAS and VRMS could not predict
arrhythmic death or VF. To develop a reference interval, percentile
values describing the distribution of the QRSd were computed for the
2174 arrhythmia-free survivors who did not receive any
antiarrhythmic drugs before the SAECG recording, taking into
account the three factors that significantly affected the QRSd: sex,
age, and MI location. The non–Q-wave and the inferoposterior MI
locations were combined because of the lack of difference between those
two MI locations. To take age into account, patients were divided into
four age groups: <50, 50 to 59, 60 to 69, and 70 to 75 years old.
Figs 4 and 5 present the QRSd
percentile curves with respect to age group and MI location for male
and female arrhythmia-free survivors, respectively. The median
(50%) QRSd was longer in men than in women, in non–Q-wave and
inferoposterior MI than in anterolateral MI, and in older patients. A
similar relation was observed for the median LAS (not shown), whereas
the median VRMS (not shown) was smaller in men than in women, in
inferoposterior MI than in non–Q-wave and anterolateral MI, and in
older patients. Since the upper percentile values can be used to define
an abnormally prolonged QRSd, these values are also given in Table 4.
|
|
|
Risk Stratification With Group-Specific SAECG Criteria
The data presented in Table 4 were used to identify
patients exceeding the 70th percentile for a given subgroup and to
predict arrhythmic events occurring during the follow-up in the entire
population (n=2461). The 70% specificity value was selected on the ROC
curve because it represents a reasonable compromise between
sensitivity and specificity. Note that the combination of the ROC curve
(Fig 2, left) and the QRSd distributions (Table 4) gives the
information needed to assess the sensitivity and specificity for other
cutpoints.
Table 5 presents the sensitivities and specificities
observed for predicting arrhythmic events for each subgroup according
to sex and MI location. In addition, the sensitivities and
specificities were also computed for an unadjusted criterion that
corresponds to the 70th percentile of the QRSd for the group of
arrhythmia-free and drug-free survivors (QRSd >109 ms). The
group-specific QRSd criteria predicted arrhythmic events with an
overall sensitivity of 65.0% and an overall specificity of 68.4%.
These values were almost identical to the values obtained with the
simpler unadjusted criterion (65.0% and 69.2%, respectively).
However, the sensitivities and specificities observed for the different
subgroups were less uniform for the unadjusted criterion. For example,
the unadjusted criterion showed a sensitivity ranging between 36.4%
for women with non–Q-wave or inferoposterior MI and 77.8% for men
with non–Q-wave or inferoposterior MI; the corresponding range was
45.5% to 71.1% for the group-specific criteria. The specificities
ranged between 63.5% and 85.6% for the unadjusted criterion and
remained at 68% to 69% for the group-specific criteria. The
specificities for the group-specific criteria were very close but not
equal to 70%, because the percentile distributions were computed for
arrhythmia-free and drug-free survivors, whereas the SAECG
specificity was evaluated for all patients without arrhythmic
events.
|
To test the uniformity of the sensitivity of the group-specific criteria, the QRSd value of each patient with an arrhythmic event was first ranked with respect to the subgroup of arrhythmia-free and drug-free survivors to which he or she belonged. Each QRSd value was thus transformed into a percentage value. A nonparametric Mann-Whitney U test then showed that there was no significant difference between the percentage values of men versus women (74±27% versus 64±30%), and a nonparametric Kruskal-Wallis ANOVA showed that there was no significant difference between the percentage values for the non–Q-wave, inferoposterior, and anterolateral MI locations (72±26%, 76±32%, and 70±29%, respectively) and for the different age groups (76±37%, 73±35%, 75±25%, and 62±29% for groups of increasing age, respectively). In addition, group-specific QRSd criteria were computed for filter frequencies of 25 and 80 Hz and showed sensitivities that were lower (60.0% and 63.1%, respectively) than the sensitivity obtained at 40 Hz, which justifies the choice of the 40-Hz filter frequency.
Relationship Between the SAECG and Other Variables for
Risk Stratification
The Cox proportional-hazards model was applied in a stepwise
manner to evaluate the power for predicting arrhythmic events of the
following variables: continuous SAECG variables (QRSd, LAS, log
VRMS), LVEF (continuous), Holter variables (>10 PVCs per hour
dichotomized; occurrence of nonsustained VT), and clinical features
that have been associated with out-of-hospital survival during the
first year after MI in this population26 (diabetes
mellitus, Killip class, previous coronary artery bypass,
previous MI, in-hospital coronary angioplasty,
thrombolytic treatment, in-hospital coronary
artery bypass). For both forward and backward stepwise procedures,
three variables were shown to have a significant and independent
contribution for the prediction of arrhythmic events. In order of
selection, these variables were (1) QRSd, (2) history of previous
MI, and (3) Killip class. After the patients for whom the LVEF and/or
the ambulatory ECG was not available were included, the same three
variables were retained by the stepwise procedure. The LVEF, which
was a univariate predictor of arrhythmic events, did not
reach statistical significance after the inclusion of the Killip
class.
Table 6 summarizes the performance of the three
variables, alone or in combination, for the prediction of
arrhythmic events. For these variables alone, a prolonged QRSd had
the highest sensitivity (65.0%) and the lowest specificity (68.4%),
whereas a Killip class 3 to 4 had the highest specificity (93.5%) and
the lowest sensitivity (17.5%). For all three variables, the
positive-predictive accuracies are low (6.5% to 8.3%) and the
negative-predictive accuracies are very high (97.1% to 98.3%) because
of the low event rate. When considered individually, a prolonged QRSd
had the highest risk ratio (3.8) of the three variables. For the
different combinations of variables, a prolonged QRSd with a
previous MI had the highest risk ratio (4.8) but a positive-predictive
accuracy that remains low (10.3%).
|
Finally, the time occurrence of arrhythmic events after MI was
investigated by plotting Kaplan-Meier survival curves (Fig 6) showing the percentage of patients that remained
arrhythmia-free as a function of time after MI for patients
with and without a prolonged QRSd by group-specific criteria (>70%).
The difference between the two survival curves is highly significant
(P<.0001).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Factors that can affect the SAECG were first investigated in a large population of MI patients by ANCOVA. This analysis revealed some significant relationships between the SAECG and sex, age, MI location, and LVEF. These relatively weak relationships were not detected in previous prospective studies with smaller patient populations for which no significant differences in sex distribution,4 5 6 average age,4 5 6 MI location distribution,5 6 7 10 and LVEF5 6 7 10 were found between MI patients with a normal SAECG and those with an abnormal SAECG. The longer QRSd observed in men can be attributed to a larger mass of myocardium, which is in turn related to a larger body size, as suggested by the cancellation of these differences after normalization for height.20 A significant increase of QRSd with age was also reported in MI patients by Malik et al23 and could be related to degenerative processes affecting conduction.32 The shorter QRSd observed in patients with anterolateral MI3 4 24 25 could be explained by the activation around the damaged wall, which occurs earlier in anterior MI.
Prediction of Arrhythmic Events With Group-Specific SAECG
Criteria
This prospective study shows that the SAECG can predict arrhythmic
events. Even if factors such as sex, age, and MI location significantly
affect the SAECG, adjustments of SAECG criteria for these factors did
not improve the overall sensitivity and specificity. However, these
adjustments equalized the predictive performance for the
different subgroups, which was not uniform with unadjusted
criteria.3 24 25 These adjustments are important for the
clinical application of the SAECG.
SAECG Differences for VT, VF, and Arrhythmic Death
Because of the large patient population among which sufficient
numbers of VTs, VFs, and arrhythmic deaths occurred, different features
of the SAECG could be associated specifically with the prediction of VT
and not with that of VF or arrhythmic death. This was not possible in
previous SAECG prospective studies, which combined all types of
ventricular tachyarrhythmias because of
their lower numbers of events.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Thus, the
group-specific SAECG criteria that are proposed are based only on the
QRSd because of the inferior predictive performance
of the VRMS and LAS variables. The predictive performances
of VRMS and LAS were acceptable only for VT patients; they were not
significant for VF patients and were very poor for patients with
arrhythmic death. The VRMS and LAS variables, along with the QRSd,
were initially proposed by Simson33 to specifically
identify VT patients in a retrospective study of MI patients, and
differences in the SAECG measured in MI patients with either VT or VF
were reported later.34 35 36 37 38 Thus, in resuscitated survivors
of spontaneous VF, the VRMS was found to be larger34 36
and the QRSd shorter35 than in patients with recurrent
sustained VT, whereas the VRMS was similar to that of normal
subjects.34 In MI patients who underwent an
electrophysiological study, Martinez-Rubio
et al37 reported that VRMSs were larger and the QRSd
shorter in patients with inducible VF than in patients with inducible
VT. In their prospective studies, Gomes et al,3 Steinberg
et al,8 and Pedretti et al14 found that only
the QRSd was significantly associated with arrhythmic events, and
Odemuyiwa et al38 reported some differences between the
predictive characteristics of the different SAECG variables for
arrhythmic death and VT.
QRSd Prolongation
Whereas the electrophysiological basis
of late potentials in VT patients is a slow, delayed, and fragmented
activation in the vicinity of the MI scar39 40 that may
constitute a substrate for reentry circuits,41 the
electrophysiological basis of the prolonged
QRSd and its possible arrhythmogenic mechanism in patients with VF and
arrhythmic death remains to be determined. The fact that the duration
of high-level activity during the QRS complex (QRSd-LAS, excluding
late potentials) had the same discrimination power for all arrhythmic
events as the QRSd (including late potentials) suggests the involvement
of a large mass of ventricular myocardium.
Because of its high amplitude, this QRSd prolongation can be detected
on the 12-lead ECG, which can give prognostic information similar to
that of the SAECG.42 The QRSd prolongation could be
explained by factors such as intraventricular
conduction defects; ventricular dilatation, which elongates
the ventricular conduction pathways without altering
conduction velocity43 44 ; and ventricular
remodeling, which increases tissue mass45 46 and could
slow conduction velocity by altering cellular coupling. The hypothesis
of a prolonged ventricular activation in dilated or
remodeled ventricles is supported by the uniformity of the
performance of the group-specific QRSd criteria: a shorter QRSd
cutoff is necessary in female patients because they have smaller
hearts, whereas a longer QRSd cutoff is necessary in older patients or
in patients with inferoposterior MI because conduction velocity is
already slower or the activation sequence is prolonged by the MI
location in these groups of patients. Zaman et al47 also
reported a correlation between the QRSd measured after MI and the
changes in the end-diastolic volume index occurring weeks
later.
Dilated48 or remodeled49 ventricles could thus constitute a specific VF substrate. For example, a significant proportion of the deaths occurring after hospital discharge and classified as arrhythmic deaths may not be caused by reentrant VT occurring in a chronic MI setting but rather by ventricular tachyarrhythmias produced by abnormal automatism resulting from an acute recurrent MI. In our patient population,50 autopsy was performed in 22 of the arrhythmic deaths and revealed the presence of a fresh, macroscopic thrombus in 5 cases (23%). This may explain in part the lack of sensitivity of late potentials for predicting arrhythmic death, because they reflect the presence of an abnormal peri-infarction activation that is potentially arrhythmogenic and not that of an arrhythmogenic focus induced by acute ischemia. On the other hand, a prolonged QRS could still play a role in the prediction of arrhythmic death associated with acute ischemia, because it may reflect an increased susceptibility to VF, whatever the mechanism of the initiating tachyarrhythmia.
Other Prognostic Indicators
The Cox proportional-hazards regression model was applied to
evaluate the predictive value of SAECG variables and a dozen other
clinical features that are known predictors of survival after MI. A
prolonged QRSd was ranked highest among the independent predictors of
arrhythmic events. Previous prospective studies that have compared the
SAECG with other prognostic indicators have also determined that the
abnormal SAECG,3 4 5 8 9 10 13 15 51 and more
specifically the QRSd,3 8 9 14 15 is one of the best
independent predictors of arrhythmic events. Two other independent
predictors were identified: a history of previous MI and a high Killip
class. Both of these support the hypothesis of a VF substrate formed by
enlarged or remodeled ventricles: the history of previous MI could
reflect the negative impact of a more extended and structurally complex
remodeled myocardium, whereas the high Killip class
reflects ventricular dysfunction and heart failure caused
by remodeling that accounts for abnormal myocardial stiffness and
pathological hypertrophy.52
A prolonged QRSd, a previous MI, and a high Killip class thus identify a group of patients with a higher risk of arrhythmic events. The risk ratio rises from 3.8 for a prolonged QRSd alone to 4.8 when the abnormal QRSd is combined with a history of previous MI; however, the positive-predictive accuracy, which rises from 6.5% to 10.6% when these variables are combined, remains low because of the low incidence of events, whereas the negative-predictive accuracy remains very high (>97%).
Role of the SAECG in Patient Management
The value of the SAECG for predicting arrhythmic events after MI
clearly outranks that of more expensive tests such as LVEF measurements
and ambulatory electrocardiography, but how
could, or how should, the SAECG affect patient management? The answer
depends on the patient population. In the CAMI study, because of the
low incidence of events, the SAECG performed very well for identifying
low-risk patients but poorly for identifying high-risk patients. The
incidence of events being a critical factor in the predictive
performance of the SAECG, how could these results be applied to
another patient population? Fig 7 illustrates the
incidence of arrhythmic events occurring after discharge in the CAMI
study and in 15 other prospective SAECG studies carried out over the
past 15 years.1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Even if these studies have different
follow-up durations and exclusion criteria, a decreasing trend in the
incidence of arrhythmic events can be observed. The positive-predictive
accuracy of an abnormal SAECG ranged between 23% and 28% in the
earlier studies1 2 3 4 5 and decreased to 11% to 17% in the
latest studies.13 14 15 A similar decreasing trend in the
post–hospital discharge mortality of MI patients occurred during the
last decades.26 The widespread use of new therapeutic and
interventional advances such as thrombolytic treatment,
ß-blocker agents, antiplatelet agents, and
revascularization procedures or changes in
lifestyle and in the management of risk factors may have improved
postinfarction survival. This reduction of the risk challenges the
clinical utility of all risk-stratification procedures because it
lowers their positive-predictive accuracy. In the context of this
decreasing trend and considering the proarrhythmic effects of
antiarrhythmic drugs as highlighted in the CAST study,29
it is clinically relevant to accurately identify low-risk patients who
would only be adversely affected by the proarrhythmic action of these
drugs. Thus, because of its high negative-prediction accuracy, the
SAECG could play an important role for patient selection in future
interventional studies on the efficacy of antiarrhythmic agents or of
automatic implantable cardioverter-defibrillators. For patient
management in the nineties, a positive SAECG test indicates an
important increase of the relative risk of arrhythmic events, but the
role of the SAECG as a screening test is otherwise severely limited by
its poor positive-predictive accuracy.
|
Study Limitations
The QRSd percentiles reported in this study are directly
applicable to other signals recorded with the same leads and
analyzed with the same algorithms. The accuracy and
reproducibility of these QRSd percentiles should be high, considering
that the accuracy of QRS end point detection by computer algorithms
increases with the signal-to-noise ratio and that all the
recordings had a very low noise level. Finally, since the
choice of the QRSd cutpoints is derived from the same population on
which the predictive performance is also assessed, the
performance of the test is likely to degrade when it is applied
to another patient sample.
Conclusions
Even if such factors as sex, age, and MI location significantly
affect the SAECG, adjustments of SAECG criteria for these factors do
not improve the overall sensitivity and specificity for predicting
arrhythmic events after MI. However, these adjustments effectively
equalize the SAECG predictive performance for the different
subgroups, which is an important clinical advantage. In contrast with
previous SAECG studies that used criteria based on QRSd, LAS, and VRMS,
we are recommending group-specific criteria for predicting arrhythmic
events that are based only on prolonged QRSds, because low VRMS and
long LAS values, which are typical of late potentials, are associated
only with VT and not with VF or arrhythmic death. The duration of
high-level activity during the QRS complex (QRSd-LAS, which excludes
late potentials) had the same discrimination power for arrhythmic
events as the QRSd (which includes late potentials), suggesting the
involvement of a large mass of ventricular
myocardium in the arrhythmogenic substrate. When multiple
prognostic indicators are considered, a prolonged QRSd, a previous MI,
and a high Killip class identify independently a group of patients with
a higher risk of arrhythmic events after MI. The investigation of the
electrophysiological basis and the possible
arrhythmogenic mechanism of the prolonged QRSd found in patients with
VF and arrhythmic death constitutes a future direction for research in
this area.
|
Selected Abbreviations and Acronyms |
|---|
|
|
Acknowledgments |
|---|
This study was financially supported through a University-Industry grant from the Medical Research Council of Canada and Bristol-Myers Squibb of Canada. The present study is an ancillary trial of CARE. The authors are grateful to Ann Baker, Vidoslav Balnozan, Danielle Beaudoin, Bonnie Cochrane, Catherine Dunne, Colette Favreau, Sheila Griffiths, Jeannette Kenny, Eric Lavallée, Charlotte Lavoie, Etel Mikes, Doris Morissette, Francine Ouimet, Denise Pagé, Andrea Serpa, Brenda Smith, and Helen Tremayne for their devoted and expert assistance. In memoriam Lise Legendre.
|
Footnotes |
|---|
Reprint requests to Pierre Savard, Centre de recherche, Hôpital du Sacré-Coeur, 5400 Gouin ouest, Montréal, Québec, Canada H4J 1C5.
Received July 27, 1995; revision received January 13, 1997; accepted January 22, 1997.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. Breithardt G, Schwartzmaier J, Borggreffe M, Haerten K, Seipel L. The prognostic significance of late ventricular potentials after myocardial infarction. Eur Heart J. 1983;4:487-495.
2.
Denniss AR, Richards DA, Cody DV, Russel PA, Young AA,
Cooper MJ, Ross DL, Uther JB. Prognostic significance of
ventricular tachycardia and fibrillation
induced at programmed stimulation and delayed potentials detected on
the signal-averaged ECG of survivors of acute myocardial
infarction. Circulation. 1986;74:731-740.
3. Gomes JA, Winters SL, Martinson M, Machac J, Stewart D, Targonski A. The prognostic significance of quantitative signal-averaged variables relative to clinical variables, site of myocardial infarction, ejection fraction and ventricular premature beats: a prospective study. J Am Coll Cardiol. 1988;13:377-384.
4. Kuchar DL, Thornburn CW, Sammel NL. Prediction of serious arrhythmic events after myocardial infarction: signal-averaged ECG, Holter monitoring and radionuclide ventriculography. J Am Coll Cardiol. 1987;9:531-538.[Abstract]
5. El-Sherif N, Ursell S, Bekheit S, Fontaine J, Turitto G, Henkin R, Caref EB. Prognostic significance of the signal-averaged ECG depends on the time of recording in the postinfarction period. Am Heart J. 1989;118:256-264.[Medline] [Order article via Infotrieve]
6. Verzoni A, Romano S, Pozzoni L, Tarricone D, SanGiorgio S, Croce L. Prognostic significance and evolution of late ventricular potentials in the first year after myocardial infarction: prospective study. Pacing Clin Electrophysiol. 1989;12:41-51.[Medline] [Order article via Infotrieve]
7. Rodriguez LM, Krijne R, van den Dool A, Brugada P, Smeets J, Wellens HJJ. Time course and prognostic significance of serial signal-averaged electrocardiograms after a first acute myocardial infarction. Am J Cardiol. 1990;66:1199-1202.[Medline] [Order article via Infotrieve]
8. Steinberg JS, Regan A, Sciacca RR, Bigger JT, Fleiss JL, Salvatore DE, Fosina M, Rolnitzky LM. Predicting arrhythmic events after acute myocardial infarction using the signal averaged electrocardiogram. Am J Cardiol. 1992;69:13-21.[Medline] [Order article via Infotrieve]
9.
Richards DAB, Byth K, Ross DL, Uther JB. What
is the best predictor of spontaneous ventricular
tachycardia and sudden death after myocardial
infarction? Circulation. 1991;83:756-763.
10. Farrel TG, Bashir Y, Cripps T, Malik M, Poloniecki J, Bennet ED, Ward DE, Camm AJ. Risk stratification for arrhythmic events in post-infarction patients based on heart rate variability, ambulatory electrocardiographic variables and the signal-averaged electrocardiogram. J Am Coll Cardiol. 1991;18:687-697.[Abstract]
11. Malik M, Kulakowski P, Odemuyiwa O, Poloniecki J, Stauton A, Millane T, Farrell T, Camm AJ. Effect of thrombolytic therapy on the predictive value of signal-averaged electrocardiography after acute myocardial infarction. Am J Cardiol. 1992;70:21-25.[Medline] [Order article via Infotrieve]
12. Odemuyiwa O, Malik M, Poloniecki J, Farrel T, Kulakowski P, Millane T, Stauton A, Camm AJ. Frequency versus time domain analysis of signal-averaged electrocardiograms, III: stratification in postinfarction patients for arrhythmic events. J Am Coll Cardiol. 1992;20:144-150.[Abstract]
13. McClements BM, Adgey AAJ. Value of signal-averaged electrocardiography, radionuclide ventriculography, Holter monitoring and clinical variables for prediction of arrhythmic events in survivors of acute myocardial infarction in the thrombolytic era. J Am Coll Cardiol. 1993;21:1419-1427.[Abstract]
14. Pedretti R, Etro MD, Laporta A. Prediction of late arrhythmic events after acute myocardial infarction from combined use of noninvasive prognostic variables and inducibility of sustained monomorphic ventricular tachycardia. Am J Cardiol. 1993;71:1131-1141.[Medline] [Order article via Infotrieve]
15. El-Sherif N, Denes P, Katz R, Capone R, Mitchel LB, Carlson M, Reynolds-Haertle R, and the CAST Investigators. Definition of best prediction criteria of the time domain signal-averaged electrocardiogram for serious arrhythmic events in the postinfarction period. J Am Coll Cardiol. 1995;25:908-914.[Abstract]
16. Kjellgren O, Gomes JA. Current usefulness of the signal averaged electrocardiogram. Curr Probl Cardiol. 1993;18:363-418.
17. MacFarlane PW, Lawrie TDV, eds. Comprehensive Electrocardiology: Theory and Practice in Health and Disease. Volume 3, Appendix 1, Normal Limits. New York, NY: Pergamon Press; 1989:1441-1526.
18. Breithardt G, Cain ME, El-Sherif N, Flowers NC, Hombach V, Janse M, Simson MB, Steinbeck G. Standards for analysis of ventricular late potentials using high-resolution or signal-averaged electrocardiography: a statement by a task force committee of the European Society of Cardiology, the American Heart Association and the American College of Cardiology. J Am Coll Cardiol. 1991;17:999-1006.[Abstract]
19. Raineri AA, Traina M, Rotolo A, Lombardo RM. Quantitative analysis of ventricular late potentials in healthy subjects. Am J Cardiol. 1990;66:1359-1362.[Medline] [Order article via Infotrieve]
20. Timmermans C, Ector H, Haisty KW, Hammil SC, Kienzle MG, Ozawa Y, Reddy BRS, Underwood DA. Signal-averaged ECG parameters in cardiac normals using Frank lead system and Fourier transform filter and gender specific differences. Pacing Clin Electrophysiol. 1994;17:303-311.[Medline] [Order article via Infotrieve]
21. Yang TF, Kennedy J, Macfarlane PW. Sex dependence of normal limits of the signal-averaged electrocardiogram. Br Heart J. 1993;69(suppl):65. Abstract.
22. de Chillou C, Sadoul N, Bruntz JF, Simon JP, Gazakur E, Aliot E. Importance of gender in the definition of an abnormal signal averaged electrocardiogram after myocardial infarction. Eur J Card Pacing Electrophysiol. 1994;4:207. Abstract.
23. Malik M, Odemuyiwa O, Poloniecki J, Kulakowski P, Farrel T, Stauton A, Camm AJ. Age-related normal values of the signal averaged-electrocardiographic variables after acute myocardial infarction. Am J Cardiol. 1991;68:440-445.[Medline] [Order article via Infotrieve]
24. Buckingham TA, Greenwalt T, Lingle A, Volgman AS, Koper P, Janosik D, Zbilut JP. In anterior myocardial infarction, frequency domain is better than time domain analysis of the signal-averaged ECG for identifying patients at risk for sustained ventricular tachycardia. Pacing Clin Electrophysiol. 1992;15:1681-1687.[Medline] [Order article via Infotrieve]
25. Elko PP, Weaver WD, Kudenchuk P, Rowlandson I. The dilemma of sensitivity versus specificity in computer interpreted acute myocardial infarction. J Electrocardiol. 1992;24:2-7.
26. Rouleau JL, Talajic M, Sussex B, Warnica W, Davies R, Potvin L, Gardner M, Stewart D, Plante S, Dupuis R, Lauzon C, Ferguson J, Mikes E, Balnozan V, Savard P. Myocardial infarction patients in the nineties, their risk factors, stratification and survival in Canada. J Am Coll Cardiol. 1996;27:1119-1127.[Abstract]
27. Lippens F, Savard P, Shenasa M. Degradation of coherent averaging by a molding effect. IEEE Trans Biomed Eng. 1989;36:856-858.[Medline] [Order article via Infotrieve]
28. Steinberg J, Bigger JT. Importance of the endpoint of noise reduction in analysis of the signal-averaged electrocardiogram. Am J Cardiol. 1989;63:556-562.[Medline] [Order article via Infotrieve]
29. The Cardiac Arrhythmia Suppression Trial (CAST) Investigators. Preliminary report: effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Engl J Med. 1989;321:406-412.[Abstract]
30.
Hanley JA, McNeil BJ. A method for comparing the
areas under receiver operating characteristic curves derived from the
same cases. Radiology. 1982;148:839-843.
31. Greenspon AJ, Kidwell GA. The effects of antiarrhythmic drugs on the signal-averaged electrocardiogram in patients with malignant ventricular arrhythmias. Prog Cardiovasc Dis. 1993;35:399-406.[Medline] [Order article via Infotrieve]
32. Jones J, Srodulski ZM, Romisher S. The aging electrocardiogram. Am J Surg Med. 1990;8:240-245.
33.
Simson MB. Use of signals in the terminal QRS
complex to identify patients with ventricular
tachycardia after myocardial infarction.
Circulation. 1981;64:235-242.
34. Freedman RA, Gillis AM, Keren A, Soderholme-Difatte V, Mason JW. Signal averaged electrocardiographic late potentials in patients with ventricular fibrillation or ventricular tachycardia: correlation with clinical arrhythmias and electrophysiologic study. Am J Cardiol. 1985;55:1350-1353.[Medline] [Order article via Infotrieve]
35. Denniss AR, Ross DL, Richards DA, Holley LK, Cooper MJ, Johnson DC, Uther JB. Differences between patients with ventricular tachycardia and ventricular fibrillation as assessed by signal-averaged electrocardiogram, radioventriculography and cardiac mapping. J Am Coll Cardiol. 1988;11:276-283.[Abstract]
36. Dolack GL, Callahan DB, Bardy GH, Green HL. Signal-averaged electrocardiographic late potentials in resuscitated survivors of out-of-hospital ventricular fibrillation. Am J Cardiol. 1990;65:1102-1104.[Medline] [Order article via Infotrieve]
37. Martinez-Rubio A, Shenasa M, Borggreffe M, Chen X, Benning F, Breithardt G. Electrophysiologic variables characterizing the induction of ventricular tachycardia versus ventricular fibrillation after myocardial infarction: relation between ventricular late potentials and coupling intervals for the induction of sustained ventricular tachyarrhythmias. J Am Coll Cardiol. 1993;21:1624-1631.[Abstract]
38. Odemuyiwa O, Malik M, Poloniecki J, Farrell T, Millane T, Kulakowski P, Staunton A, Matthies A, Camm AJ. Differences between predictive characteristics of signal averaged electrocardiographic variables for postinfarction sudden death and ventricular tachycardia. Am J Cardiol. 1992;69:1186-1192.[Medline] [Order article via Infotrieve]
39.
Gardner PI, Ursell PC, Fenoglio JJ, Wit AL.
Electrophysiologic and anatomic basis for fractionated electrograms
recorded from healed myocardial infarction.
Circulation. 1985;72:596-611.
40.
Lacroix D, Savard P, Shenasa M, Pagé P, Joly D,
Derome D, Cardinal R, Nadeau R. Spatial domain analysis
of late ventricular potentials: intraoperative and thoracic
correlations. Circ Res. 1990;66:55-68.
41. Lacroix D, Savard P, Shenasa M, Kaltenbrunner W, Cardinal R, Pagé P, Nadeau R. Intracardiac and thoracic mapping of late ventricular potentials. In: Shenasa M, Borggrefe M, Breithardt G, eds. Cardiac Mapping. Armonck, NY: Futura Publishing Co; 1993:367-382.
42. Zareba W, Moss AJ, le Cessie S. Dispersion of ventricular repolarization and arrhythmic cardiac death in coronary artery disease. Am J Cardiol. 1994;74:550-553.[Medline] [Order article via Infotrieve]
43.
Reiter MJ, Mann DE, Williams GR. Interaction of
hypokalemia and ventricular dilatation in isolated rabbit
hearts. Am J Physiol. 1993;265:H1544-H1550.
44.
Reiter MJ, Zetelakis Z, Kirchhof JH, Boersma L, Allesie
MA. Interaction of acute ventricular dilatation
and d-sotalol during sustained reentrant
ventricular tachycardia around a fixed
obstacle. Circulation. 1994;89:423-431.
45.
Pfeffer MA, Braunwald E. Ventricular
remodeling after myocardial infarction. Circulation. 1990;81:1161-1172.
46. Jugdutt B. Effect of captopril and enalapril on left ventricular function and collagen during healing after anterior and inferior myocardial infarction in a dog model. J Am Coll Cardiol. 1995;25:1718-1725.[Abstract]
47.
Zaman AG, Morris JL, Smyllie JH, Cowan JC. Late
potentials and ventricular enlargement after myocardial
infarction: a new role for high-resolution
electrocardiography?
Circulation. 1993;88:905-914.
48.
Jalal S, Williams GR, Mann DE, Reiter MJ. Effect
of acute ventricular dilatation on fibrillation thresholds
in the isolated rabbit heart. Am J Physiol. 1992;263:H1306-H1310.
49. Li HG, Jones DL, Yee R, Klein GJ. Electrophysiologic substrate associated with pacing induced heart failure in dogs: potential value of programmed stimulation in predicting sudden death. J Am Coll Cardiol. 1992;19:444-449.[Abstract]
50. Hadjis T, Talajic M, Stewart D, Sampalis J, Sussex B, Warnica W, Davies R, Gardner M, Dupuis R. The role of thrombosis in sudden cardiac death: the SAGES trial experience. Circulation. 1994;90(suppl I):I-374. Abstract.
51. Mäkijärvi M, Fetsch T, Reinhardt L, Martinez-Rubio A, Borggrefe M, Breithardt G. A new diagnostic algorithm for noninvasive risk stratification after acute myocardial infarction. Circulation. 1994;90(suppl I):I-501. Abstract.
52. Brilla CG. The cardiac structure-function relationship and the renin-angiotensin-aldosterone system in hypertension and heart failure. Curr Opin Cardiol. 1994;9(suppl 1):S2-S10.
This article has been cited by other articles:
 |
|
 |
|
  J. J. Goldberger, M. E. Cain, S. H. Hohnloser, A. H. Kadish, B. P. Knight, M. S. Lauer, B. J. Maron, R. L. Page, R. S. Passman, D. Siscovick, et al. American Heart Association/American College of Cardiology Foundation/Heart Rhythm Society Scientific Statement on Noninvasive Risk Stratification Techniques for Identifying Patients at Risk for Sudden Cardiac Death: A Scientific Statement From the American Heart Association Council on Clinical Cardiology Committee on Electrocardiography and Arrhythmias and Council on Epidemiology and Prevention J. Am. Coll. Cardiol., September 30, 2008; 52(14): 1179 - 1199. [Full Text] [PDF]
|
|
|
|
 |
|
 J. J. Goldberger, M. E. Cain, S. H. Hohnloser, A. H. Kadish, B. P. Knight, M. S. Lauer, B. J. Maron, R. L. Page, R. S. Passman, D. Siscovick, et al. American Heart Association/American College of Cardiology Foundation/Heart Rhythm Society Scientific Statement on Noninvasive Risk Stratification Techniques for Identifying Patients at Risk for Sudden Cardiac Death: A Scientific Statement From the American Heart Association Council on Clinical Cardiology Committee on Electrocardiography and Arrhythmias and Council on Epidemiology and Prevention Circulation, September 30, 2008; 118(14): 1497 - 1518. [Full Text] [PDF]
|
|
|
|
|
|
T. J. Bunch, S. H. Hohnloser, and B. J. Gersh Mechanisms of Sudden Cardiac Death in Myocardial Infarction Survivors: Insights From the Randomized Trials of Implantable Cardioverter-Defibrillators Circulation, May 8, 2007; 115(18): 2451 - 2457. [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Williams Coronary Revascularization in Diabetic Chronic Kidney Disease/End-Stage Renal Disease: A Nephrologist's Perspective Clin. J. Am. Soc. Nephrol., March 1, 2006; 1(2): 209 - 220. [Full Text] [PDF]
|
|
|
|
 |
|
 S. D. Solomon, S. Zelenkofske, J. J.V. McMurray, P. V. Finn, E. Velazquez, G. Ertl, A. Harsanyi, J. L. Rouleau, A. Maggioni, L. Kober, et al. Sudden Death in Patients with Myocardial Infarction and Left Ventricular Dysfunction, Heart Failure, or Both N. Engl. J. Med., June 23, 2005; 352(25): 2581 - 2588. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Bauer, P. Guzik, P. Barthel, R. Schneider, K. Ulm, M. A. Watanabe, and G. Schmidt Reduced prognostic power of ventricular late potentials in post-infarction patients of the reperfusion era Eur. Heart J., April 2, 2005; 26(8): 755 - 761. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. L. Lee and C.-P. Lau The use of signal-averaged electrocardiogram in risk stratification after acute myocardial infarction in the modern era Eur. Heart J., April 2, 2005; 26(8): 747 - 748. [Full Text] [PDF]
|
|
|
|
|
|
S. H. Hohnloser and B. J. Gersh Changing Late Prognosis of Acute Myocardial Infarction: Impact on Management of Ventricular Arrhythmias in the Era of Reperfusion and the Implantable Cardioverter-Defibrillator Circulation, February 25, 2003; 107(7): 941 - 946. [Full Text] [PDF]
|
|
|
|
|
|
J. J. Bailey, A. S. Berson, H. Handelsman, and M. Hodges Utility of current risk stratification tests for predicting major arrhythmic events after myocardial infarction J. Am. Coll. Cardiol., December 1, 2001; 38(7): 1902 - 1911. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L Jordaens, R Tavernier, and MIRRACLE Investigators Determinants of sudden death after discharge from hospital for myocardial infarction in the thrombolytic era Eur. Heart J., July 2, 2001; 22(14): 1214 - 1225. [Abstract] [PDF]
|
|
|
|
 |
|
 R. Mrowka, P. B. Persson, H. Theres, and A. Patzak Blunted arterial baroreflex causes "pathological" heart rate turbulence Am J Physiol Regulatory Integrative Comp Physiol, October 1, 2000; 279(4): R1171 - R1175. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A A Armoundas, D S Rosenbaum, J N Ruskin, H Garan, and R J Cohen Prognostic significance of electrical alternans versus signal averaged electrocardiography in predicting the outcome of electrophysiological testing and arrhythmia-free survival Heart, September 1, 1998; 80(3): 251 - 256. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||