©1991 S. Karger AG. Basel 0008-6312/91/0791 -0064$2.75/0
Cardiology 1991;79:64-72
Relative Prognostic Value of Clinical Heart Failure and Early Echocardiographie Parameters in Acute Myocardial Infarction1 Jens Berning, Frank Steensgaarcl-Hansen. Merete Appleyard Medical Department C, Cardiac Division, CCU C 4 0 , Glostrup Hospital, and Medical Department B, Rigshospital, University of Copenhagen, Denmark
Key Words. Heart failure • Left ventricular function • Echocardiography • Acute myocardial infarction • Prognosis
Introduction Proper management of patients with acute myocardial infarction (AMI) requires knowledge about the natural history of the disease. Clinical signs of heart failure and laboratory evidence of depressed left ven tricular systolic performance are considered the strongest predictors of long-term mortal ity post-AMI [1-3]. Importantly, echocardi 1 Supported by grants from the Danish Heart Foun dation.
ography is the most readily available labora tory method for estimation of left ventricu lar performance in this setting [4, 5], Early échocardiographie examination has an es tablished prognostic value in AMI [6-8], However, it is not quite clear to which extent early determination of left ventricular per formance by echocardiography provides prognoslic information over and beyond that of accumulated observations of clinical manifestations of heart failure during the entire period of hospitalization. Similarly, the comparative prognostic value of various
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
Abstract. The relative prognostic value of clinical heart failure and early M-mode and 2-dimensional échocardiographie indexes of left ventricular performance was compared in a study of 205 consecutive patients with acute myocardial infarction (AMI). Statistical analy sis showed that an early wall motion score was a stronger predictor of 1-year mortality than the occurrence of clinical heart failure early, late or at any time during the hospital course of AMI. The finding of clinical heart failure had an independent prognostic value of interme diate strength. M-mode échocardiographie parameters only had a weak independent prog nostic value, possibly related to their content of information on left ventricular end-systolic dimension.
Prognostic Value of Heart Failure
Patients A consecutive series of 205 patients with creatine kinasc-MB-documented AMI was prospectively ex amined in the Glostrup Hospital, a nonreferral hospi tal with complete regional uptake (300.000 inhabit ants), during a 10-month period from November 1985 to September 1986 [9], Patients with associated diseases were excluded: 2 cases with valvular aortic stenosis, 1 case of cardiac amyloid and 1 case of hypertrophic myopathy. M-mode échocardiographie data were incomplete in 26 patients (15%), who were excluded from statistical analysis, leaving a study group of 175 patients. There were 125 males (age range: 38-89 years, median age: 60 years) and 50 females (age range: 35-85. median age: 65 years). No patient received thrombolytic treatment or routine treatment with [5-blockers. No patient underwent per cutaneous transluminal coronary angioplasty or coro nary bypass surgery during admission and 1-ycar fol low-up. Follow-up. All 175 patients were followed up after 1 year. Inhospital mortality was 9% (15/175). Cumu lated l-year mortality was 23% (40/175).
Methods Method o f Observation All patients with AMI were continuously ECGmonitored from the time of admission to the time of discharge. They remained under observation of the
same cardiologists in the 24-bed coronary care unit during their entire hospital stay and were discharged directly from the unit [10], Definition o f Heart Failure Clinical heart failure. Bilateral rales at ausculta tion extending at least 5 cm over the lung bases and persisting following coughing and deep inspiration, orthopnea, i.e. breathlessness in the recumbent posi tion relieved in the sitting or standing positions, the presence of an S3 gallop at auscultation, displacement of the apex beat outside the midclavicular line, neck vein distention, hepatomegaly, edema or radiographic signs of pulmonary congestion were consid ered signs of clinical heart failure. Only signs that prompted the clinician to change (i.e. institute or reinforce) treatment were accepted as an expression of heart failure. The criteria of cardiac failure were those applied in daily clinical routine of a large coro nary care unit and were not different from those used in previously published work from the department [ 10].
Early heart failure was defined as clinical signs of heart failure during hospitalization recorded until the time of the échocardiographie examination (median: 46 h. range: 12-72 h). Late heart failure was defined as clinical signs of heart failure recorded in the time interval from the échocardiographie examination un til discharge (median: 11 days, range: 3-36 days). Heart failure at any lime during admission was de fined as signs of heart failure from admission to dis charge. Echocardiography A Toshiba SSH 60A echocardiograph equipped with a 3.5-MHz. transducer was used. Patients were examined in the supine and left lateral decubitus posi tions. Conventional échocardiographie variables were selected because of their rapid bedside practicability. Time consuming 2-dimensional échocardiographie methods utilizing computerized still-frame planime try and complex algorithms for calculation of left ven tricular ejection fraction were not examined. All stud ies were performed by two highly experienced exam iners. M-Mode Echocardiography M-mode scans through the long axis of the left ventricle were obtained guided by 2-dimensional echocardiography and routinely registered on hard
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
early échocardiographie measures of global left ventricular function has not been for mally analyzed. Therefore, in a large consecutive series of patients with AMI. M-mode and 2-dimensional échocardiographie examination was performed shortly after admission and the value of traditionally used échocardiograph ie variables to predict I-year mortality was compared with that of the occurrence of clin ical heart failure early, late or at any time during admission, using a multivariate sta tistical approach.
65
Berning/Steensgaard-Hansen/Appleyard
66
copies for quantitation. The measurements repre sented an average of 3-5 consecutive beats and were performed according to the recommendations of the American Society of Echocardiography, following the ‘leading-edge’ principle [11]. The following M-mode parameters were deter mined. End-diastolic dimension (mm. 'standarddiam eter'): Distance between the endocardium of the pos terior wall and the interventricular septum measured at the level of the chordae tendineae at the beginning of the QRS complex. End-systolic dimension (mm. 'standarddiameter): Shortest vertical distance during a cardiac cycle between the endocardium of the pos terior wall and the interventricular septum measured at the level of the chordae tendineae. Eractional short ening (%): Fractional shortening = (end-diastolic di mension - end-systolic dimension)/end-diastolic di mension. E-point septal distance (mm): Minimal ver tical distance between the E point of the mitral valve and the septal endocardium [12].
face area was not performed since information on height and weight of the patients was usually not available in the setting of early AMI. Statistical Methods Descriptive statistics were given as median and range of measured echocardiographic parameters when normal distribution could not be assumed. The Kruskall-Wallis and Mann-Whitney nonparametric tests were used for the detection of statistically signif icant differences between groups of patients with the various echocardiographic measurements. Multiple, logistic regression statistics (BMDP pro gram and log likelihood method) were used to analyze the relative value of early echocardiographic esti mates of left ventricular function and signs of clinical heart failure to predict 1-year mortality.
Normal Echocardiographic Values The following values were defined as normal considering the mean age of the patients: end-diastolic dimension: < 5.5 cm. end-systolic dimension: < 3 .6 cm, FS%: 28-44, E-point septal distance: < 8 mm, WMS: > 1.6 [14]. Correction for body sur
Frequency Distributions o f Echocardiographie Measures o f Left Ventricular Function in Early AMI Distributions of all examined M-mode and 2-dimensional echocardiographic vari ables are given in figure 1. Percentage of abnormal values for each parameter is shown in table 1. Analysis of the distribution of echocar diographic variables shows that the propor tion of patients that have a normal M-mode end diastolic dimension is very high in an unselected population in early AMI (fig. 1, table 1). Furthermore, the median of this pa rameter shows little difference in the defined subsets of patients with clinical heart failure (table 2). The other M-mode echocardio graphic parameters show no differences in the groups with early and late heart failure: however, WMS is significantly different in all groups. Smaller, but statistally significant differences are present between all parame ters in the 82 patients with heart failure at
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
Results 2-Dimensional Echocardiography The heart was visualized through multiple acous tic windows with visualization of parasternal longitu dinal and cross-sectional, apical four- and two-cham ber, subxiphoid four-chamber and cross-sectional images. All examinations were recorded on video tape. The following 2-dimensional echocardiographic parameter was determined: wall motion score (WMS). The left ventricle was divided into 9 segments as orig inally described by Hegcr et al. [13] and the motion of individual segments scored as follows: +3 hyperkine sia, + 2 normokinesia, +1 hypokinesia, 0 akinesia, -1 dyskinesia. When a segment was visualized in more than one projection the average score for that segment was used, i.e. when all segments could be visualized in all projections, the motion of 35 seg ments was semiquantitated and the scores subse quently reduced to 9 scores, describing the motion in the apical, in the 4 midventricular and the 4 basal segments. The average value of the scores in these 9 segments constituted WMS, which was considered a measure of global left ventricular function.
67
Prognostic Value of Heart Failure
END DIASTOLIC DIMENSION
FRACTIONAL SHORTENING
Fig. I. Histogram showing distribution of échocardiographie variables in early AMI. Normal, abnormal and supranormal values of each variable are separated by vertical, dashed lines (n = 175).
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
WALL MOTION SCORE
Berning/Steensgaard-Hansen/Appleyard
68
any time during admission versus the 93 patients without heart failure (table 2). Clinical Data and Echocardiographie Measures o f Left Ventricular Function in Early AMI In table 2 the manifestations of clinical heart failure early, late and during the entire admission is related to échocardiographie variables. The values (median, range) of échocardiographie variables are listed for each subset of clinical heart failure and levels of statistical significance are given (KruskallWallis).
Relative Prognostic Value o f Heart Failure and Echocardiographie Variables Heart Failure at Any Time during Admis sion. Clinical heart failure observed at any time during admission had a strong predic tive value of 1-year mortality when early WMS was not included in the statistical model (table 3). However, when this param eter was introduced in the model it appeared that left ventricular global systolic perfor mance as measured by a 2-dimensional échocardiographie WMS was a stronger pre dictor of 1-year mortality post-AMI than clinical heart failure observed at any time
Table 1. Percentage of abnormal findings in the study population EDD. cm Normal values Study population
ESD, cm
< 5 .5 5.2 3.1-8.9 34
median range
Abnormal, %
EPSD. mm
FS. % 28-44 23 4-51 69
< 3 .6 4.0 1.9-6.8 71
WMS > 1.6 1.3 0.3-2.4 70
Cardiology 1991;79:64-72
Relative Prognostic Value of Clinical Heart Failure and Early Echocardiographie Parameters in Acute Myocardial Infarction1 Jens Berning, Frank Steensgaarcl-Hansen. Merete Appleyard Medical Department C, Cardiac Division, CCU C 4 0 , Glostrup Hospital, and Medical Department B, Rigshospital, University of Copenhagen, Denmark
Key Words. Heart failure • Left ventricular function • Echocardiography • Acute myocardial infarction • Prognosis
Introduction Proper management of patients with acute myocardial infarction (AMI) requires knowledge about the natural history of the disease. Clinical signs of heart failure and laboratory evidence of depressed left ven tricular systolic performance are considered the strongest predictors of long-term mortal ity post-AMI [1-3]. Importantly, echocardi 1 Supported by grants from the Danish Heart Foun dation.
ography is the most readily available labora tory method for estimation of left ventricu lar performance in this setting [4, 5], Early échocardiographie examination has an es tablished prognostic value in AMI [6-8], However, it is not quite clear to which extent early determination of left ventricular per formance by echocardiography provides prognoslic information over and beyond that of accumulated observations of clinical manifestations of heart failure during the entire period of hospitalization. Similarly, the comparative prognostic value of various
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
Abstract. The relative prognostic value of clinical heart failure and early M-mode and 2-dimensional échocardiographie indexes of left ventricular performance was compared in a study of 205 consecutive patients with acute myocardial infarction (AMI). Statistical analy sis showed that an early wall motion score was a stronger predictor of 1-year mortality than the occurrence of clinical heart failure early, late or at any time during the hospital course of AMI. The finding of clinical heart failure had an independent prognostic value of interme diate strength. M-mode échocardiographie parameters only had a weak independent prog nostic value, possibly related to their content of information on left ventricular end-systolic dimension.
Prognostic Value of Heart Failure
Patients A consecutive series of 205 patients with creatine kinasc-MB-documented AMI was prospectively ex amined in the Glostrup Hospital, a nonreferral hospi tal with complete regional uptake (300.000 inhabit ants), during a 10-month period from November 1985 to September 1986 [9], Patients with associated diseases were excluded: 2 cases with valvular aortic stenosis, 1 case of cardiac amyloid and 1 case of hypertrophic myopathy. M-mode échocardiographie data were incomplete in 26 patients (15%), who were excluded from statistical analysis, leaving a study group of 175 patients. There were 125 males (age range: 38-89 years, median age: 60 years) and 50 females (age range: 35-85. median age: 65 years). No patient received thrombolytic treatment or routine treatment with [5-blockers. No patient underwent per cutaneous transluminal coronary angioplasty or coro nary bypass surgery during admission and 1-ycar fol low-up. Follow-up. All 175 patients were followed up after 1 year. Inhospital mortality was 9% (15/175). Cumu lated l-year mortality was 23% (40/175).
Methods Method o f Observation All patients with AMI were continuously ECGmonitored from the time of admission to the time of discharge. They remained under observation of the
same cardiologists in the 24-bed coronary care unit during their entire hospital stay and were discharged directly from the unit [10], Definition o f Heart Failure Clinical heart failure. Bilateral rales at ausculta tion extending at least 5 cm over the lung bases and persisting following coughing and deep inspiration, orthopnea, i.e. breathlessness in the recumbent posi tion relieved in the sitting or standing positions, the presence of an S3 gallop at auscultation, displacement of the apex beat outside the midclavicular line, neck vein distention, hepatomegaly, edema or radiographic signs of pulmonary congestion were consid ered signs of clinical heart failure. Only signs that prompted the clinician to change (i.e. institute or reinforce) treatment were accepted as an expression of heart failure. The criteria of cardiac failure were those applied in daily clinical routine of a large coro nary care unit and were not different from those used in previously published work from the department [ 10].
Early heart failure was defined as clinical signs of heart failure during hospitalization recorded until the time of the échocardiographie examination (median: 46 h. range: 12-72 h). Late heart failure was defined as clinical signs of heart failure recorded in the time interval from the échocardiographie examination un til discharge (median: 11 days, range: 3-36 days). Heart failure at any lime during admission was de fined as signs of heart failure from admission to dis charge. Echocardiography A Toshiba SSH 60A echocardiograph equipped with a 3.5-MHz. transducer was used. Patients were examined in the supine and left lateral decubitus posi tions. Conventional échocardiographie variables were selected because of their rapid bedside practicability. Time consuming 2-dimensional échocardiographie methods utilizing computerized still-frame planime try and complex algorithms for calculation of left ven tricular ejection fraction were not examined. All stud ies were performed by two highly experienced exam iners. M-Mode Echocardiography M-mode scans through the long axis of the left ventricle were obtained guided by 2-dimensional echocardiography and routinely registered on hard
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
early échocardiographie measures of global left ventricular function has not been for mally analyzed. Therefore, in a large consecutive series of patients with AMI. M-mode and 2-dimensional échocardiographie examination was performed shortly after admission and the value of traditionally used échocardiograph ie variables to predict I-year mortality was compared with that of the occurrence of clin ical heart failure early, late or at any time during admission, using a multivariate sta tistical approach.
65
Berning/Steensgaard-Hansen/Appleyard
66
copies for quantitation. The measurements repre sented an average of 3-5 consecutive beats and were performed according to the recommendations of the American Society of Echocardiography, following the ‘leading-edge’ principle [11]. The following M-mode parameters were deter mined. End-diastolic dimension (mm. 'standarddiam eter'): Distance between the endocardium of the pos terior wall and the interventricular septum measured at the level of the chordae tendineae at the beginning of the QRS complex. End-systolic dimension (mm. 'standarddiameter): Shortest vertical distance during a cardiac cycle between the endocardium of the pos terior wall and the interventricular septum measured at the level of the chordae tendineae. Eractional short ening (%): Fractional shortening = (end-diastolic di mension - end-systolic dimension)/end-diastolic di mension. E-point septal distance (mm): Minimal ver tical distance between the E point of the mitral valve and the septal endocardium [12].
face area was not performed since information on height and weight of the patients was usually not available in the setting of early AMI. Statistical Methods Descriptive statistics were given as median and range of measured echocardiographic parameters when normal distribution could not be assumed. The Kruskall-Wallis and Mann-Whitney nonparametric tests were used for the detection of statistically signif icant differences between groups of patients with the various echocardiographic measurements. Multiple, logistic regression statistics (BMDP pro gram and log likelihood method) were used to analyze the relative value of early echocardiographic esti mates of left ventricular function and signs of clinical heart failure to predict 1-year mortality.
Normal Echocardiographic Values The following values were defined as normal considering the mean age of the patients: end-diastolic dimension: < 5.5 cm. end-systolic dimension: < 3 .6 cm, FS%: 28-44, E-point septal distance: < 8 mm, WMS: > 1.6 [14]. Correction for body sur
Frequency Distributions o f Echocardiographie Measures o f Left Ventricular Function in Early AMI Distributions of all examined M-mode and 2-dimensional echocardiographic vari ables are given in figure 1. Percentage of abnormal values for each parameter is shown in table 1. Analysis of the distribution of echocar diographic variables shows that the propor tion of patients that have a normal M-mode end diastolic dimension is very high in an unselected population in early AMI (fig. 1, table 1). Furthermore, the median of this pa rameter shows little difference in the defined subsets of patients with clinical heart failure (table 2). The other M-mode echocardio graphic parameters show no differences in the groups with early and late heart failure: however, WMS is significantly different in all groups. Smaller, but statistally significant differences are present between all parame ters in the 82 patients with heart failure at
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
Results 2-Dimensional Echocardiography The heart was visualized through multiple acous tic windows with visualization of parasternal longitu dinal and cross-sectional, apical four- and two-cham ber, subxiphoid four-chamber and cross-sectional images. All examinations were recorded on video tape. The following 2-dimensional echocardiographic parameter was determined: wall motion score (WMS). The left ventricle was divided into 9 segments as orig inally described by Hegcr et al. [13] and the motion of individual segments scored as follows: +3 hyperkine sia, + 2 normokinesia, +1 hypokinesia, 0 akinesia, -1 dyskinesia. When a segment was visualized in more than one projection the average score for that segment was used, i.e. when all segments could be visualized in all projections, the motion of 35 seg ments was semiquantitated and the scores subse quently reduced to 9 scores, describing the motion in the apical, in the 4 midventricular and the 4 basal segments. The average value of the scores in these 9 segments constituted WMS, which was considered a measure of global left ventricular function.
67
Prognostic Value of Heart Failure
END DIASTOLIC DIMENSION
FRACTIONAL SHORTENING
Fig. I. Histogram showing distribution of échocardiographie variables in early AMI. Normal, abnormal and supranormal values of each variable are separated by vertical, dashed lines (n = 175).
Downloaded by: Access provided by the University of Michigan Library 141.216.78.40 - 1/11/2019 8:08:11 AM
WALL MOTION SCORE
Berning/Steensgaard-Hansen/Appleyard
68
any time during admission versus the 93 patients without heart failure (table 2). Clinical Data and Echocardiographie Measures o f Left Ventricular Function in Early AMI In table 2 the manifestations of clinical heart failure early, late and during the entire admission is related to échocardiographie variables. The values (median, range) of échocardiographie variables are listed for each subset of clinical heart failure and levels of statistical significance are given (KruskallWallis).
Relative Prognostic Value o f Heart Failure and Echocardiographie Variables Heart Failure at Any Time during Admis sion. Clinical heart failure observed at any time during admission had a strong predic tive value of 1-year mortality when early WMS was not included in the statistical model (table 3). However, when this param eter was introduced in the model it appeared that left ventricular global systolic perfor mance as measured by a 2-dimensional échocardiographie WMS was a stronger pre dictor of 1-year mortality post-AMI than clinical heart failure observed at any time
Table 1. Percentage of abnormal findings in the study population EDD. cm Normal values Study population
ESD, cm
< 5 .5 5.2 3.1-8.9 34
median range
Abnormal, %
EPSD. mm
FS. % 28-44 23 4-51 69
< 3 .6 4.0 1.9-6.8 71
WMS > 1.6 1.3 0.3-2.4 70
