- Email: [email protected]
Predictive Value of Two-Dimensional Echocardiographic and Hemodynamic Measurements on Admission With Acute Myocardial Infarction
Predictive Value of Two-Dimensional Echocardiographic and Hemodynamic Measurements on Admission With Acute Myocardial Infarction Wybren Jaarsma, MD, Cees A. Visser, MD, Machiel J. Eenige van, PhD, Freek W.A. Verheugt, MD, Albert J. Funke Kupper, MD, and Jan P. Roos, MD, Amsterdam, The Netherlands
To identify high-risk patients with acute myocardial infarction, we compared admission values of two-dimensional echocardiography and hemodynamic monitoring. Left ventricular wall motion score (WMS), left ventricular stroke work index (LVSWI), and pulmonary capillary pressure (PCP) were obtained in 77 patients without clinical signs of heart failure. Progression into Killip grade 3 or 4 was found in 16 of 77 patients (21 o/o) within 32 ± 6 hours (mean ± 1 standard deviation) after admission. Mean WMS, LVSWI, and PCP in those patients who developed severe pump failure were significantly different from those who did not: 13.4 ± 4.9 versus 7.3 ± 4, 30 ± 4 versus 46 ± ll gm/m>, and 21 ± 8 versus 12 ± 6 mm Hg, respectively. Sensitivity ofWMS of >7 and LVSWI of <35 gm/m2 in predicting Killip grade 3 or 4 was 88% and· 94%, specificity was 57% and 87%, positive predictive value was 35% and 65%, and negative predictive value was 95% and 98o/o, Sensitivity of PCP was low (50%). Early identification of patients developing myocardial rupture or reinfarction was limited by both methods. We conclude that echocardiographic examination on admission in patients with acute myocardial infarction provides an alternative approach for early identification of low-risk patients. (J AM Soc EcHo 1988;1:187-93.)
Early identification of high-risk patients during acute myocardial infarction (AMI) with clinical and hemodynamic criteria has proved to be important for prognostic and therapeutic reasons. 1-4 Initial clinical Killip classification, 5 however, has been specific but insensitive in predicting an early complicated course and may frequently vary with cardiac performance. 1•4 Routine assessment of hemodynamic measurements for identification of those who are at high risk for cardiac complications, that is, severe pump failure, is not warranted for several reasons. Conversely, twodimensional echocardiography can easily be done at the bedside with a high success rate and has been helpful in detecting and quantifying AMI. 6-11 In addition, previous studies have shown the value of a
From the Department of Cardiology, Free University Hospital; and the Interuniversity Cardiology Institute. Reprint requests: Wybren Jaarsma, MD, Department of Cardiology, Sint Antonius Hospital, Koekoekslaan l, 3435 CM Nieuwegein, The Netherlands.
wall motion score (WMS) as an expression of global left ventricular function for early identification of patients with high risk of pump failure and death. 12- 15 The purpose of tht; present study was therefore to compare the values of two-dimensional echocardiography and simuh:aneously obtained hemodynamic measurements on admission, for early risk stratification in a large group of patients with AMI without clinical signs of heart failure.
METHODS Patients
The study group consisted of 77 of 84 consecutive patients with an AMI, all in Killip grade l or 2,5 in whom adequate two-dimensional echocardiograms (success rate 94%) and hemodynamic measurements (success rate 98 o/o) were obtained. The mean age was 60 years (range 30 to 79 years). There were 62 men 187
Journ:U of the American Society of Echocardiography
188 Jaarsma et al.
WALL MOTION
Wall motion score left ventricle 3chamber
_2 chamber
,~,
@
---~4c-
2 chamber
3chamber
4 chamber
SCORE
25 r - - - - - - - - - - - - - - , n = 61 n = 16 21ZJ 0 0 0 0
15
=
Zl m
11ZJ
5
==
==
-l • ••
• ••
0
Ill 0 0
-5 ' - - - - - - - - - - - - - - - - - - ' KILLIP 1, 2 KILLIP 3,1J 13 segments Hyperkinesia -1 Normokinesia 0 Hypokinesia 1 Score 1-13
Akinesia 2 Dyskinesia 3 Aneurysm 4
Figure 1 Diagram of three apical long-axis views (lower panel) usually applied for calculation ofleft ventricular (LV) WMS. In each of three views left ventricular myocardium was divided into five segments; apex was considered common to all three apical views. If apical views were not adequate for analysis, same segments could be evaluated from short-axis cross-sections (upper panel). 4c, FourChamber view; Ao, aorta; LA, left atrium; LV, left ventricle; RA, right atrium; R V, right ventricle.
and 15 women. The diagnosis of AMI was documented in each case by a typical history, development of new pathologic Q waves on the electrocardiogram and/ or by an increase in the serum creatine kinase value. A previous AMI was not a reason for exclusion. Two-Dimensional Echocardiography
Within 2 hours after admission to the coronary care unit, all echocardiograms were obtained with a commercially available, wide-angle sector scanner (77020 A, Hewlett-Packard, Andover, Massachusetts, or an Ultramark 8 ATL transducer, Advanced Technology Laboratories, Inc., Bellevue, Washington). With the patient in the left lateral decubitus position, all segments of the entire left ventricular wall were studied with all available acoustic windows. For purposes of analysis we used three apical longaxis views, recorded at mutual angles of 60 degrees, with the aid of a calibrated fluid-filled ring containing an air bubble and fixed around the body of the transducer. 16 These three views were the apical long-axis or three-chamber view, the four-chamber
Figure 2 Distribution of WMS and mean values ± SD in 77 patients with AMI. Closed circles, Patients whose conditions progressed into Killip grade 3 or 4 during hospitalization.
view, and the two-chamber view. In each of these views the left ventricular contour was divided into five segments. The apical area was considered one segment that was common to all three views, so that a total of 13 segments was obtained. When one or more segments could not be visualized from the apex, they were visualized from the parasternal long- and short-axis views (or in some cases from subcostal views) (Figure 1). Each segment was assigned a numerical value according to the degree of systolic wall motion abnormality it showed: 0 for normokinesis, + 1 for hypokinesis, + 2 for akinesis, and + 3 for dyskinesis. Hyperkinesis (defined as exaggerated wall motion and myocardial thickening) 14 was graded - 1. Left ventricular aneurysm (defined as a bulge in the left ventricular contour persisting in both diastole and systole and showing akinesis or dyskinesis) was graded +4. 17•18 This approach has been shown to be highly reproducible with a low interobserver variability. 15 •19•20 In addition, an AMI was considered anterior if asynergy occurred in the apical, anteroseptal, or anterolateral regions and posterior if it occurred in the posterior, posterolateral, inferoseptal, or inferior regions of the left ventricle. 21 Hemodynamic measurements. Hemodynamic data were obtained on admission in all patients, irrespective of .their initial clinical classification, simultaneously with the echocardiogram. Systemic blood pressure was measured by a brachial artery catheter. Mean pulmonary capillary pressure (PCP
Volume l Number 3 May-June 1988
Value of echocardiography and hemodynamic measurements on admission 189
LV STROKE WORK INDEX
81ZJ
n =
0 0
61ZJ
1-
61
n =
PULMONARY CAPILLARY PRESSURE 51ZJ ,------------------------,
16
=ct =
"
£
• 1 •.
31ZJ
~
£
I±IZJ
'OJ)
~
Jb
•
c£l
21ZJ
IZJ
•
:r::
I
£
E
0
21ZJ 11ZJ IZJ
KILLlP 3,1±
cffi:n
1-
n
•
=
16
• •
I ii 0
'OJ)
L---------------------~
KILLIP 1,2
61
I±IZJ
00
N
=
n
II
•
••I
8
KILLIP 1,2
KILLIP 3,1±
Figure 3 Distribution of LVSWI (grams per square meter) and mean values ± SD in 77 patients with AML Closed circles, Patients whose conditions progressed into Killip grade 3 or 4 during hospitalization.
Figure 4 Distribution of PCP (millimeters of mercury) and mean values ± SD in 77 patients with AML Closed circles, Patients whose conditions progressed into Killip grade 3 or 4 during hospitalization.
in millimeters of mercury) was determined with a Swan-Ganz catheter passed into the pulmonary artery. Cardiac index (after correction for body surface area) was measured by the thermodilution technique and by averaging at least three replicate determinations. Left ventricular stroke work index (LVSWI in grams per square meter) was calculated from stroke index (SI), mean systolic arterial pressure (MAP), and PCP [0.0136 x SI x (MAP - PCP)]. For all measurements a monitor terminal (78534, HewlettPackard) with a cardiac output acquisition module (78552) and a data management module (78554) was used. All hemodynamic data were initial baseline determinations. Clinical classification and follow-up. Patients were clinically classified on admission and during the hospital stay according to the classification of Killip and Kimball. 5 Patients considered as grade 3 or 4 on admission were not included in this study. Incidence of progression into Killip grade 3 or 4 was correlated with the echocardiographic and hemodynamic findings obtained on a'dmission.
RESULTS
Statistical Methods
Data were expressed as mean ± 1 standard deviation. Differences between groups were evaluated with the unpaired Student's t test. To relate variables linear regression analysis was performed and the 95% confidence limits indicated.
The echocardiographic infarct location was anterior in 43 patients, posterior in 17 patients, and both anterior and posterior in 17 patients. Clinical Course
Progression into Killip grade 3 or 4 was found in 16 of 77 patients (21%) 32 ± 6 hours after admission (10 patients were grade 3, and six were grade 4). In anterior or anterior and posterior AMI this complication developed in 12 of 36 patients (33%) and in posterior AMI in four of25 patients (16%). Eleven of all 77 patients ( 14%) died during the hospital period. An anterior or anterior and posterior AMI was present in nine patients, and a posterior AMI was present in two of these patients. Echocardiographic characterization of clinical subsets. Figure 2 shows the distribution of WMS in patients with and without clinical pump failure during the hospital stay (Killip 1,2 and Killip 3,4, respectively). The difference between the mean WMS of the 16 patients whose conditions progressed into Killip grade 3 or 4 and the 61 patients whose conditions did not was significant ( 13.4 ± 4. 9 versus 7.3 ± 4, p < 0.001). Hemodynamic characterization of clinical subsets. Figures 3 and 4 illustrate the distribution of LVSWI and PCP in patients with and without
190
Journal of the American Society of Echocardiography
Jaarsma et a!.
Table 1
Prognostic values of WMS, LVSWI, and PCP on admission in predicting severe heart failure after AMI Killip grade Value
%
1,2
3,4
Total
35 26 61
2 14 16
37 40 77
8 53 61
15 1 16
23 54 77
57 4 61
8 8 16
65 12 77
WMS
<7 >7
TarAL
Sensitivity Specificity Positive predictive accuracy Negative predictive accuracy LVSWl (gm/m2 )
87.5 57 35 95
<35 >35
TOTAL
Sensitivity Specificity Positive predictive accuracy Negative predictive accuracy PCP (mmHg)
94 87 65 98
<20 >20
TOTAL
Sensitivity Specificity Positive predictive accuracy Negative predictive accuracy
50 93 67 88
progression into Killip grade 3 or 4. The mean value of LVWSI in patients with Killip grade 1 or 2 was 46 ± 11 and 30 ± 4 gm/m2 (p < 0.001) in patients whose conditions progressed into Killip grade 3 or 4; the mean value of PCP was 12 ± 6 and 21 ± 8 mm Hg, respectively (p < 0.001). Comparison of echocardiographic and hemodynamic measurements in predicting clinical course. To obtain the sharpest separation with the best sensitivity a cutoff value of 7 was chosen for the WMS, 35 gm/ m2 for LVSWI, and 20 mm Hg for PCP. Table 1 shows the sensitivity and specificity of the echocardiographic and hemodynamic measurements (WMS, LVSWI, and PCP) for the recognition of patients developing Killip grade 3 or 4 during the period of hospitalization. Sensitivity was 88%, 94%, and 50%, respectively; specificity was 57%, 87%, and 93%, respectively. The predictive value of a positive test (high WMS, low LVWSI, high PCP) for identification of Killip grade 3 or 4 was 35%, 65%, and 67%, respectively. The predictive value of a negative test (low WMS, high LVSWI, low PCP) for identification of patients not developing Killip grade 3 or 4 was 95%, 98%, and 88%, respectively. Nine of the 11 patients who died during hospi-
talization had developed Killip grade 3 or 4. Six of these patients (all were Killip grade 4) died of progressive pump failure. The remaining three patients (all were Killip grade 3) died of myocardial rupture (two patients) and refractory ventricular arrhythmias (one patient). The other two patients (both were Killip grade 1) died of a fatal reinfarction and a myocardial rupture. Mean WMS and PCP were significantly higher in the 11 nonsurvivors than in the 66 survivors: 12.5 ± 4 versus 7 ± 4.5 (:p < 0.01) and 21 ± 9 versus 13 ± 6 mm Hg (p < 0.001). The difference of LVSWI in nonsurvivors and survivors was also significant: 31 ± 6 versus 44 ± 11 gm/m2 (p < 0.001). All six patients who died of progressive heart failure had a WMS > 10 (mean 15 ± 3.7) and an LVSWI <35 gm/m2 (mean 31 ± 3). PCP was normal in two patients ( 11 and 14 mm Hg) and >20 mm Hg in the remaining four patients (mean 24 ± 11). Echocardiographic and hemodynamic correlations. The relation between WMS and LVSWI and between PCP and LVSWI is shown in Figures 5 and 6. There was a poor relation between WMS and LVSWI (r = -0.52) and between PCP and LVSWI (r = -0.20).
Volume 1 Number 3 May-June 1988
Value of echocardiography and hemodynamic measurements on admission
30
"'c
5~
lJ.J
qli)
J::
• X
c
w
cc
a::
20
::J
"'"'w a::
0
u
(f)
z
s>--
,_
10
-'
-'
:E
"'
30
X X Xx0 X
a:: a:
0
-' -' a:
191
CL
a:
X
0
-
20
X
X
X
,_
a:: a:
X X
z
10
* *
0 :E
-'
-llil
20
::J
o_
30
~0
50
LV STROKE WORK INDEX
70
60 gM /
X
X
u
80
0 20
30
M2
Figure 5 Relation between WMS and LVSWI (grams per square meter) in 77 patients with AMI.
DISCUSSION
LV STROKE WORK INDEX
60 gM /
70
80
M2
Figure 6 Relation between PCP (millimeters of mercury) and LVSWI (grams per square meter) in 77 patients with AMI.
. ,~D WMS
nd7
Previous work by Heger et al. 12 has shown that twodimensional echocardiography during AMI correlates well with clinical measurements of left ventricular function (Killip classification). Gibson et al. 14 demonstrated the value of a WMS as a measure of global left ventricular function for predicting later hemodynamic deterioration, particularly cardiogenic shock, and found it even more predictive than several clinical measurements such as previous infarction, location of AMI, and peak serum creatine kinase levels. An explanation for this could be that a WMS is not a measure of irreversibly damaged myocardium but rather a functional index reflecting the total mass of hypoperfused myocardium, including previous and acute necrosis and -ischemic areas. In the present study the same semiquantitative method was used, which requires little time and can be done at the bedside with a high success rate. In previous studies in our laboratory/ 5•19•20 we demonstrated a consistent interobserver agreement with this semiquantitative assessment of wall motion. Loh et al. 22 used a similar approach to ours, showed excellent interobserver agreement, and demonstrated that this method was superior to quantitative analysis in assessing hypokinesia, which was often seen in patients with nontransmural infarction. In contrast, quantitative analysis is time consuming and requires computer facilities and an optimal endocardial definition, which is sometimes difficult to obtain with still frames. This study shows that the incidence of clinical deterioration was much higher in patients with a high WMS than in patients with a low WMS. A cutoff
X
51il
qli)
--
X
L VS WI
-r-
p
c
p
n,sq
LVSW! '!' 35 PCP (
2~
-rL_
-r-
WMS > 7
LVSW! < 35
D
KILLIP 1.2
Figure 7 Graphic representation of cutoff values applied in relation to patients developing Killip grade 3 or 4.
value of > 7 was chosen to make a separation, with an acceptable sensitivity, between low- and high-risk patients. This figure makes it possible on admission to identifY patients at low risk for developing severe pump failure with a sensitivity of87.5%, a specificity of 57%, and positive and negative predictive accuracy of 35% and 95%, respectively. Figure 7 illustrates that with a WMS of ~7, 37 patients were selected. Of those 37 patients two developed severe pump failure and were missed (one patient had a borderline WMS of 7, and the other had a WMS of 6). Thus echocardiographic examination on admission may reliably identifY patients with low risk, in whom no special high level of care may be needed. The specificity of using a wall motion score of >7, however, is low because of much crossover between patients who do well and those who do not (Figure 2). Absence of segmental hyperkinesia (especially in anterior AMI) and/ or presence ofsegmental remote asyn-
Journal of the American Society of Echocardiography
192 Jaarsma et al.
ergy, on the other hand, provides additional information in identifying patients at high risk during their hospital stay. 14•19 LVSWI on admission (Figure 7) correctly identifies patients who probably will not develop severe heart failure with a high negative predictive value of 98%. This agrees with earlier reports. 2 •4 •23 The one patient who had been missed had an initial LVSWI of 38 gm/ m2. In contrast to WMS and LVSWI, PCP (Figure 7) is less useful in predicting an uncomplicated outcome because of its relatively lower negative predictive accuracy (85%). This finding is also in line with earlier reports. 3•24 Although the mean values of echocardiographic and hemodynamic measurements differ significantly in survivors and nonsurvivors, the identification of patients dying from myocardial rupture was limited. Two of the three patients with fatal myocardial rupture of the free wall had a WMS of 5 and 7 and an LVSWI of 57 and 37 gm/m2, respectively. Because myocardial rupture, however, is usually associated with early dilation and thinning of the infarct wne (that is, infarct expansion), echocardiography may potentially identify these patients, irrespective of the WMS.2s-z7 Furthermore, patients who had reinfarction during their hospitalization were correctly identified on the basis of the WMS and LVSWI in five of eight patients and in four of eight patients, respectively. Although both WMS and LVSWI can reliably identify patients who are not developing severe pump failure, the relation between these measurements was poor (Figure 6). An explanation for this might be that in contrast to WMS the LVSWI is based on both pressure and volume measurements. The present study demonstrates that echocardiographic examination on admission may provide a reliable noninvasive method for the identification of low-risk patients. Of note is that progression into Killip grade 3 or 4 was attained 32 ± 6 hours after admission. Routine application therefore may be of importance in the management of a coronary care unit in these cost-conscious times. We are indebted to Annette Brauns, Lidwin Admiraal, and Ellen Rothuis for technical assistance and to Margreet Koelemaij for secretarial assistance.
3. 4. 5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16. 17.
REFERENCES l. Rackley C, Russell RO. Left ventricular function in acute
myocardial infarction and its clinical significance. Circulation 1972;45:231-44. 2. Scheidt S, Wilner G, Fillmore S, Shapiro M, Killip T. Ob-
18.
jective hemodynamic assessment after acute myocardial infarction. Br Heart J 1973;35:908-16. Forrester JS, Diamond GA, Swan HJC. Correlative classification of clinical and hemodynamic function after acute myocardial infarction. Am J Cardiol1977;39:137-45. Weber KT, Janicki JS, Russell RO, Rackley CE. Identification of high risk subsets of acute myocardial infarction. Am J Cardiol1978;41:197-203. Killip T, Kimball JT. Treatment of myocardial infarction in a coronary care unit. A two-year experience with 250 patients. Am J Cardia! 1967;20:457-64. Heger JJ, Weyman AE, Wann LS, Dillon JC, Feigenbaum H. Cross-sectional echocardiography in acute myocardial infarction: detection and localization of regional left ventricular asynergy. Circulation 1979;60:531-8. Horowitz RS, Morganroth J, Parotto C, Chenn CC, Soffer J, Pauletto FJ. Immediate diagnosis of acute myocardial infarction by two-dimensional echocardiography. Circulation 1982;65:323-9. Weiss JL, Bulkley BH, Grover MH, Mason SJ. Twodimensional echocardiographic recognition of myocardial injury in man; comparison with postmortem srudies. Circulation 1981;63:401-8. Nixon JV, Narahara KA, Smitherman TC. Estimation of myocardial involvement in patients with acute myocardial infarction by two-dimensional echocardiography. Circulation 1980;62: 1248-55. Visser CA, Kan G, Lie KI, Becker AE, Durrer D. Detection and quantification of acute, isolated myocardial infarction by two-dimensional echocardiography. Am J Cardiol1981; 47:1020-5. Visser CA, Kan G, Lie KI, Becker AE, Durrer D. Apex twodimensional echocardiography. Alternative approach to quantification of acute myocardial infarction. Br Heart J 1982; 47:461-7. Heger JJ, Weyman AE, Wann LS, Rogers EW, Dillon JC, Feigenbaum H. Cross-sectional echocardiographic analysis of the extent of left ventricular asynergy in acute myocardial infarction. Circulation 1980;61:1113-8. Horowitz RS, Morganroth J. Immediate detection of early high-risk patients with acute myocardial infarction using twodimensional echocardiographic evaluation of left ventricular regional wall motion abnormalities. Am Heart J 1982; 103:814-22. Gibson RS, Bishop HL, Stamm RB, Crampton RS, Beller GA, Martin RP. Value of eariy two-dimensional echocardiography in patients with acute myocardial infarction. Am J Cardiol1982;49:1ll0-9. Kan G, Visser CA, Meltzer RS, Koolen JJ, Dunning AJ. Short- and long-term predictive value of admission wall motions score in acute myocardial infarction: a two-dimensional echocardiographic srudy of 345 patients. Br Heart J 1986; 56:422-7. Visser CA, Kan G, Dunning AJ. A simple device to obtain reproducible echocardiographic apical cross-sections. Am J Cardiol1985;56:206-7. Weyman AE, Peskoe SM, Williams ES, Dillon JC, Feigenbaum H. Detection of left ventricular aneurysms by crosssectional echocardiography. Circulation 1976;54:936-44. Visser CA, Kan G, David GK, Lie KI, Durrer D. Echocardiographic-cineangiographic COHelation in detecting left ventricular aneurysm: a prospective srudy of 422 patients. Am J Cardiol 1982;50:337-41.
Volume 1 Number 3 May-June 1988
Value of echocardiography and hemodynamic measurements on admission
19. Jaarsma W, Visser CA, Eenige van MJ, et al. Prognostic implications of regional hyperkinesia and remote asynergy of noninfarcted myocardium. Am J Cardiol1986;58:394-8. 20. Visser CA, David GK, Kan G, et al. Two-dimensional echocardiography during percutaneous transluminal coronary angioplasty. Am Heart J 1986;111:1035-41. 21. Kisslo JA, Robertson D, Gilbert BW, von Ramm D, Beckar VS. A comparison of real-time two-dimensional left ventricular echocardiography and cineangiography in detecting left ventricular asynergy. Circulation 1977;55:134-40. 22. Loh IK, Charuzi Y, Beeder C, Marshall LA, Ginsbury JH. Early diagnosis of transmural infarction by two-dimensional echocardiography. Am Heart J 1982;104:963-8. 23. Russel RO, Hunt D, Rackley CE. Left ventricular hemodynamics in anterior and inferior myocardial infarction. Am J Cardiol1973;32:8-16.
193
24. Forrester JS, Diamond G, Chatterjee K, Swan HJC. Medical therapy of acute myocardial infarction by application of hemodynamic subsets (first of two parts). N Eng! J Med 1976;295: 1356-62. 25. Hutchins GM, Bulkley BH. Infarct expansion versus extension: two different complications of acute myocardial infarction. Am J Cardiol1978;41:ll27-32. 26. Schuster EH, Bulkley BH. Expansion of transmural myocardial infarction: a pathophysiologic factor in cardiac rupture. Circulation 1979;60: 1532-8. 27. Eaton LW, Weiss JL, Bulkley BH, Garrison JB, Weisfeldt ML. Regional dilatation after acute myocardial infarction. N Eng! J Med 1979;310:57-62.
To identify high-risk patients with acute myocardial infarction, we compared admission values of two-dimensional echocardiography and hemodynamic monitoring. Left ventricular wall motion score (WMS), left ventricular stroke work index (LVSWI), and pulmonary capillary pressure (PCP) were obtained in 77 patients without clinical signs of heart failure. Progression into Killip grade 3 or 4 was found in 16 of 77 patients (21 o/o) within 32 ± 6 hours (mean ± 1 standard deviation) after admission. Mean WMS, LVSWI, and PCP in those patients who developed severe pump failure were significantly different from those who did not: 13.4 ± 4.9 versus 7.3 ± 4, 30 ± 4 versus 46 ± ll gm/m>, and 21 ± 8 versus 12 ± 6 mm Hg, respectively. Sensitivity ofWMS of >7 and LVSWI of <35 gm/m2 in predicting Killip grade 3 or 4 was 88% and· 94%, specificity was 57% and 87%, positive predictive value was 35% and 65%, and negative predictive value was 95% and 98o/o, Sensitivity of PCP was low (50%). Early identification of patients developing myocardial rupture or reinfarction was limited by both methods. We conclude that echocardiographic examination on admission in patients with acute myocardial infarction provides an alternative approach for early identification of low-risk patients. (J AM Soc EcHo 1988;1:187-93.)
Early identification of high-risk patients during acute myocardial infarction (AMI) with clinical and hemodynamic criteria has proved to be important for prognostic and therapeutic reasons. 1-4 Initial clinical Killip classification, 5 however, has been specific but insensitive in predicting an early complicated course and may frequently vary with cardiac performance. 1•4 Routine assessment of hemodynamic measurements for identification of those who are at high risk for cardiac complications, that is, severe pump failure, is not warranted for several reasons. Conversely, twodimensional echocardiography can easily be done at the bedside with a high success rate and has been helpful in detecting and quantifying AMI. 6-11 In addition, previous studies have shown the value of a
From the Department of Cardiology, Free University Hospital; and the Interuniversity Cardiology Institute. Reprint requests: Wybren Jaarsma, MD, Department of Cardiology, Sint Antonius Hospital, Koekoekslaan l, 3435 CM Nieuwegein, The Netherlands.
wall motion score (WMS) as an expression of global left ventricular function for early identification of patients with high risk of pump failure and death. 12- 15 The purpose of tht; present study was therefore to compare the values of two-dimensional echocardiography and simuh:aneously obtained hemodynamic measurements on admission, for early risk stratification in a large group of patients with AMI without clinical signs of heart failure.
METHODS Patients
The study group consisted of 77 of 84 consecutive patients with an AMI, all in Killip grade l or 2,5 in whom adequate two-dimensional echocardiograms (success rate 94%) and hemodynamic measurements (success rate 98 o/o) were obtained. The mean age was 60 years (range 30 to 79 years). There were 62 men 187
Journ:U of the American Society of Echocardiography
188 Jaarsma et al.
WALL MOTION
Wall motion score left ventricle 3chamber
_2 chamber
,~,
@
---~4c-
2 chamber
3chamber
4 chamber
SCORE
25 r - - - - - - - - - - - - - - , n = 61 n = 16 21ZJ 0 0 0 0
15
=
Zl m
11ZJ
5
==
==
-l • ••
• ••
0
Ill 0 0
-5 ' - - - - - - - - - - - - - - - - - - ' KILLIP 1, 2 KILLIP 3,1J 13 segments Hyperkinesia -1 Normokinesia 0 Hypokinesia 1 Score 1-13
Akinesia 2 Dyskinesia 3 Aneurysm 4
Figure 1 Diagram of three apical long-axis views (lower panel) usually applied for calculation ofleft ventricular (LV) WMS. In each of three views left ventricular myocardium was divided into five segments; apex was considered common to all three apical views. If apical views were not adequate for analysis, same segments could be evaluated from short-axis cross-sections (upper panel). 4c, FourChamber view; Ao, aorta; LA, left atrium; LV, left ventricle; RA, right atrium; R V, right ventricle.
and 15 women. The diagnosis of AMI was documented in each case by a typical history, development of new pathologic Q waves on the electrocardiogram and/ or by an increase in the serum creatine kinase value. A previous AMI was not a reason for exclusion. Two-Dimensional Echocardiography
Within 2 hours after admission to the coronary care unit, all echocardiograms were obtained with a commercially available, wide-angle sector scanner (77020 A, Hewlett-Packard, Andover, Massachusetts, or an Ultramark 8 ATL transducer, Advanced Technology Laboratories, Inc., Bellevue, Washington). With the patient in the left lateral decubitus position, all segments of the entire left ventricular wall were studied with all available acoustic windows. For purposes of analysis we used three apical longaxis views, recorded at mutual angles of 60 degrees, with the aid of a calibrated fluid-filled ring containing an air bubble and fixed around the body of the transducer. 16 These three views were the apical long-axis or three-chamber view, the four-chamber
Figure 2 Distribution of WMS and mean values ± SD in 77 patients with AMI. Closed circles, Patients whose conditions progressed into Killip grade 3 or 4 during hospitalization.
view, and the two-chamber view. In each of these views the left ventricular contour was divided into five segments. The apical area was considered one segment that was common to all three views, so that a total of 13 segments was obtained. When one or more segments could not be visualized from the apex, they were visualized from the parasternal long- and short-axis views (or in some cases from subcostal views) (Figure 1). Each segment was assigned a numerical value according to the degree of systolic wall motion abnormality it showed: 0 for normokinesis, + 1 for hypokinesis, + 2 for akinesis, and + 3 for dyskinesis. Hyperkinesis (defined as exaggerated wall motion and myocardial thickening) 14 was graded - 1. Left ventricular aneurysm (defined as a bulge in the left ventricular contour persisting in both diastole and systole and showing akinesis or dyskinesis) was graded +4. 17•18 This approach has been shown to be highly reproducible with a low interobserver variability. 15 •19•20 In addition, an AMI was considered anterior if asynergy occurred in the apical, anteroseptal, or anterolateral regions and posterior if it occurred in the posterior, posterolateral, inferoseptal, or inferior regions of the left ventricle. 21 Hemodynamic measurements. Hemodynamic data were obtained on admission in all patients, irrespective of .their initial clinical classification, simultaneously with the echocardiogram. Systemic blood pressure was measured by a brachial artery catheter. Mean pulmonary capillary pressure (PCP
Volume l Number 3 May-June 1988
Value of echocardiography and hemodynamic measurements on admission 189
LV STROKE WORK INDEX
81ZJ
n =
0 0
61ZJ
1-
61
n =
PULMONARY CAPILLARY PRESSURE 51ZJ ,------------------------,
16
=ct =
"
£
• 1 •.
31ZJ
~
£
I±IZJ
'OJ)
~
Jb
•
c£l
21ZJ
IZJ
•
:r::
I
£
E
0
21ZJ 11ZJ IZJ
KILLlP 3,1±
cffi:n
1-
n
•
=
16
• •
I ii 0
'OJ)
L---------------------~
KILLIP 1,2
61
I±IZJ
00
N
=
n
II
•
••I
8
KILLIP 1,2
KILLIP 3,1±
Figure 3 Distribution of LVSWI (grams per square meter) and mean values ± SD in 77 patients with AML Closed circles, Patients whose conditions progressed into Killip grade 3 or 4 during hospitalization.
Figure 4 Distribution of PCP (millimeters of mercury) and mean values ± SD in 77 patients with AML Closed circles, Patients whose conditions progressed into Killip grade 3 or 4 during hospitalization.
in millimeters of mercury) was determined with a Swan-Ganz catheter passed into the pulmonary artery. Cardiac index (after correction for body surface area) was measured by the thermodilution technique and by averaging at least three replicate determinations. Left ventricular stroke work index (LVSWI in grams per square meter) was calculated from stroke index (SI), mean systolic arterial pressure (MAP), and PCP [0.0136 x SI x (MAP - PCP)]. For all measurements a monitor terminal (78534, HewlettPackard) with a cardiac output acquisition module (78552) and a data management module (78554) was used. All hemodynamic data were initial baseline determinations. Clinical classification and follow-up. Patients were clinically classified on admission and during the hospital stay according to the classification of Killip and Kimball. 5 Patients considered as grade 3 or 4 on admission were not included in this study. Incidence of progression into Killip grade 3 or 4 was correlated with the echocardiographic and hemodynamic findings obtained on a'dmission.
RESULTS
Statistical Methods
Data were expressed as mean ± 1 standard deviation. Differences between groups were evaluated with the unpaired Student's t test. To relate variables linear regression analysis was performed and the 95% confidence limits indicated.
The echocardiographic infarct location was anterior in 43 patients, posterior in 17 patients, and both anterior and posterior in 17 patients. Clinical Course
Progression into Killip grade 3 or 4 was found in 16 of 77 patients (21%) 32 ± 6 hours after admission (10 patients were grade 3, and six were grade 4). In anterior or anterior and posterior AMI this complication developed in 12 of 36 patients (33%) and in posterior AMI in four of25 patients (16%). Eleven of all 77 patients ( 14%) died during the hospital period. An anterior or anterior and posterior AMI was present in nine patients, and a posterior AMI was present in two of these patients. Echocardiographic characterization of clinical subsets. Figure 2 shows the distribution of WMS in patients with and without clinical pump failure during the hospital stay (Killip 1,2 and Killip 3,4, respectively). The difference between the mean WMS of the 16 patients whose conditions progressed into Killip grade 3 or 4 and the 61 patients whose conditions did not was significant ( 13.4 ± 4. 9 versus 7.3 ± 4, p < 0.001). Hemodynamic characterization of clinical subsets. Figures 3 and 4 illustrate the distribution of LVSWI and PCP in patients with and without
190
Journal of the American Society of Echocardiography
Jaarsma et a!.
Table 1
Prognostic values of WMS, LVSWI, and PCP on admission in predicting severe heart failure after AMI Killip grade Value
%
1,2
3,4
Total
35 26 61
2 14 16
37 40 77
8 53 61
15 1 16
23 54 77
57 4 61
8 8 16
65 12 77
WMS
<7 >7
TarAL
Sensitivity Specificity Positive predictive accuracy Negative predictive accuracy LVSWl (gm/m2 )
87.5 57 35 95
<35 >35
TOTAL
Sensitivity Specificity Positive predictive accuracy Negative predictive accuracy PCP (mmHg)
94 87 65 98
<20 >20
TOTAL
Sensitivity Specificity Positive predictive accuracy Negative predictive accuracy
50 93 67 88
progression into Killip grade 3 or 4. The mean value of LVWSI in patients with Killip grade 1 or 2 was 46 ± 11 and 30 ± 4 gm/m2 (p < 0.001) in patients whose conditions progressed into Killip grade 3 or 4; the mean value of PCP was 12 ± 6 and 21 ± 8 mm Hg, respectively (p < 0.001). Comparison of echocardiographic and hemodynamic measurements in predicting clinical course. To obtain the sharpest separation with the best sensitivity a cutoff value of 7 was chosen for the WMS, 35 gm/ m2 for LVSWI, and 20 mm Hg for PCP. Table 1 shows the sensitivity and specificity of the echocardiographic and hemodynamic measurements (WMS, LVSWI, and PCP) for the recognition of patients developing Killip grade 3 or 4 during the period of hospitalization. Sensitivity was 88%, 94%, and 50%, respectively; specificity was 57%, 87%, and 93%, respectively. The predictive value of a positive test (high WMS, low LVWSI, high PCP) for identification of Killip grade 3 or 4 was 35%, 65%, and 67%, respectively. The predictive value of a negative test (low WMS, high LVSWI, low PCP) for identification of patients not developing Killip grade 3 or 4 was 95%, 98%, and 88%, respectively. Nine of the 11 patients who died during hospi-
talization had developed Killip grade 3 or 4. Six of these patients (all were Killip grade 4) died of progressive pump failure. The remaining three patients (all were Killip grade 3) died of myocardial rupture (two patients) and refractory ventricular arrhythmias (one patient). The other two patients (both were Killip grade 1) died of a fatal reinfarction and a myocardial rupture. Mean WMS and PCP were significantly higher in the 11 nonsurvivors than in the 66 survivors: 12.5 ± 4 versus 7 ± 4.5 (:p < 0.01) and 21 ± 9 versus 13 ± 6 mm Hg (p < 0.001). The difference of LVSWI in nonsurvivors and survivors was also significant: 31 ± 6 versus 44 ± 11 gm/m2 (p < 0.001). All six patients who died of progressive heart failure had a WMS > 10 (mean 15 ± 3.7) and an LVSWI <35 gm/m2 (mean 31 ± 3). PCP was normal in two patients ( 11 and 14 mm Hg) and >20 mm Hg in the remaining four patients (mean 24 ± 11). Echocardiographic and hemodynamic correlations. The relation between WMS and LVSWI and between PCP and LVSWI is shown in Figures 5 and 6. There was a poor relation between WMS and LVSWI (r = -0.52) and between PCP and LVSWI (r = -0.20).
Volume 1 Number 3 May-June 1988
Value of echocardiography and hemodynamic measurements on admission
30
"'c
5~
lJ.J
qli)
J::
• X
c
w
cc
a::
20
::J
"'"'w a::
0
u
(f)
z
s>--
,_
10
-'
-'
:E
"'
30
X X Xx0 X
a:: a:
0
-' -' a:
191
CL
a:
X
0
-
20
X
X
X
,_
a:: a:
X X
z
10
* *
0 :E
-'
-llil
20
::J
o_
30
~0
50
LV STROKE WORK INDEX
70
60 gM /
X
X
u
80
0 20
30
M2
Figure 5 Relation between WMS and LVSWI (grams per square meter) in 77 patients with AMI.
DISCUSSION
LV STROKE WORK INDEX
60 gM /
70
80
M2
Figure 6 Relation between PCP (millimeters of mercury) and LVSWI (grams per square meter) in 77 patients with AMI.
. ,~D WMS
nd7
Previous work by Heger et al. 12 has shown that twodimensional echocardiography during AMI correlates well with clinical measurements of left ventricular function (Killip classification). Gibson et al. 14 demonstrated the value of a WMS as a measure of global left ventricular function for predicting later hemodynamic deterioration, particularly cardiogenic shock, and found it even more predictive than several clinical measurements such as previous infarction, location of AMI, and peak serum creatine kinase levels. An explanation for this could be that a WMS is not a measure of irreversibly damaged myocardium but rather a functional index reflecting the total mass of hypoperfused myocardium, including previous and acute necrosis and -ischemic areas. In the present study the same semiquantitative method was used, which requires little time and can be done at the bedside with a high success rate. In previous studies in our laboratory/ 5•19•20 we demonstrated a consistent interobserver agreement with this semiquantitative assessment of wall motion. Loh et al. 22 used a similar approach to ours, showed excellent interobserver agreement, and demonstrated that this method was superior to quantitative analysis in assessing hypokinesia, which was often seen in patients with nontransmural infarction. In contrast, quantitative analysis is time consuming and requires computer facilities and an optimal endocardial definition, which is sometimes difficult to obtain with still frames. This study shows that the incidence of clinical deterioration was much higher in patients with a high WMS than in patients with a low WMS. A cutoff
X
51il
qli)
--
X
L VS WI
-r-
p
c
p
n,sq
LVSW! '!' 35 PCP (
2~
-rL_
-r-
WMS > 7
LVSW! < 35
D
KILLIP 1.2
Figure 7 Graphic representation of cutoff values applied in relation to patients developing Killip grade 3 or 4.
value of > 7 was chosen to make a separation, with an acceptable sensitivity, between low- and high-risk patients. This figure makes it possible on admission to identifY patients at low risk for developing severe pump failure with a sensitivity of87.5%, a specificity of 57%, and positive and negative predictive accuracy of 35% and 95%, respectively. Figure 7 illustrates that with a WMS of ~7, 37 patients were selected. Of those 37 patients two developed severe pump failure and were missed (one patient had a borderline WMS of 7, and the other had a WMS of 6). Thus echocardiographic examination on admission may reliably identifY patients with low risk, in whom no special high level of care may be needed. The specificity of using a wall motion score of >7, however, is low because of much crossover between patients who do well and those who do not (Figure 2). Absence of segmental hyperkinesia (especially in anterior AMI) and/ or presence ofsegmental remote asyn-
Journal of the American Society of Echocardiography
192 Jaarsma et al.
ergy, on the other hand, provides additional information in identifying patients at high risk during their hospital stay. 14•19 LVSWI on admission (Figure 7) correctly identifies patients who probably will not develop severe heart failure with a high negative predictive value of 98%. This agrees with earlier reports. 2 •4 •23 The one patient who had been missed had an initial LVSWI of 38 gm/ m2. In contrast to WMS and LVSWI, PCP (Figure 7) is less useful in predicting an uncomplicated outcome because of its relatively lower negative predictive accuracy (85%). This finding is also in line with earlier reports. 3•24 Although the mean values of echocardiographic and hemodynamic measurements differ significantly in survivors and nonsurvivors, the identification of patients dying from myocardial rupture was limited. Two of the three patients with fatal myocardial rupture of the free wall had a WMS of 5 and 7 and an LVSWI of 57 and 37 gm/m2, respectively. Because myocardial rupture, however, is usually associated with early dilation and thinning of the infarct wne (that is, infarct expansion), echocardiography may potentially identify these patients, irrespective of the WMS.2s-z7 Furthermore, patients who had reinfarction during their hospitalization were correctly identified on the basis of the WMS and LVSWI in five of eight patients and in four of eight patients, respectively. Although both WMS and LVSWI can reliably identify patients who are not developing severe pump failure, the relation between these measurements was poor (Figure 6). An explanation for this might be that in contrast to WMS the LVSWI is based on both pressure and volume measurements. The present study demonstrates that echocardiographic examination on admission may provide a reliable noninvasive method for the identification of low-risk patients. Of note is that progression into Killip grade 3 or 4 was attained 32 ± 6 hours after admission. Routine application therefore may be of importance in the management of a coronary care unit in these cost-conscious times. We are indebted to Annette Brauns, Lidwin Admiraal, and Ellen Rothuis for technical assistance and to Margreet Koelemaij for secretarial assistance.
3. 4. 5. 6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16. 17.
REFERENCES l. Rackley C, Russell RO. Left ventricular function in acute
myocardial infarction and its clinical significance. Circulation 1972;45:231-44. 2. Scheidt S, Wilner G, Fillmore S, Shapiro M, Killip T. Ob-
18.
jective hemodynamic assessment after acute myocardial infarction. Br Heart J 1973;35:908-16. Forrester JS, Diamond GA, Swan HJC. Correlative classification of clinical and hemodynamic function after acute myocardial infarction. Am J Cardiol1977;39:137-45. Weber KT, Janicki JS, Russell RO, Rackley CE. Identification of high risk subsets of acute myocardial infarction. Am J Cardiol1978;41:197-203. Killip T, Kimball JT. Treatment of myocardial infarction in a coronary care unit. A two-year experience with 250 patients. Am J Cardia! 1967;20:457-64. Heger JJ, Weyman AE, Wann LS, Dillon JC, Feigenbaum H. Cross-sectional echocardiography in acute myocardial infarction: detection and localization of regional left ventricular asynergy. Circulation 1979;60:531-8. Horowitz RS, Morganroth J, Parotto C, Chenn CC, Soffer J, Pauletto FJ. Immediate diagnosis of acute myocardial infarction by two-dimensional echocardiography. Circulation 1982;65:323-9. Weiss JL, Bulkley BH, Grover MH, Mason SJ. Twodimensional echocardiographic recognition of myocardial injury in man; comparison with postmortem srudies. Circulation 1981;63:401-8. Nixon JV, Narahara KA, Smitherman TC. Estimation of myocardial involvement in patients with acute myocardial infarction by two-dimensional echocardiography. Circulation 1980;62: 1248-55. Visser CA, Kan G, Lie KI, Becker AE, Durrer D. Detection and quantification of acute, isolated myocardial infarction by two-dimensional echocardiography. Am J Cardiol1981; 47:1020-5. Visser CA, Kan G, Lie KI, Becker AE, Durrer D. Apex twodimensional echocardiography. Alternative approach to quantification of acute myocardial infarction. Br Heart J 1982; 47:461-7. Heger JJ, Weyman AE, Wann LS, Rogers EW, Dillon JC, Feigenbaum H. Cross-sectional echocardiographic analysis of the extent of left ventricular asynergy in acute myocardial infarction. Circulation 1980;61:1113-8. Horowitz RS, Morganroth J. Immediate detection of early high-risk patients with acute myocardial infarction using twodimensional echocardiographic evaluation of left ventricular regional wall motion abnormalities. Am Heart J 1982; 103:814-22. Gibson RS, Bishop HL, Stamm RB, Crampton RS, Beller GA, Martin RP. Value of eariy two-dimensional echocardiography in patients with acute myocardial infarction. Am J Cardiol1982;49:1ll0-9. Kan G, Visser CA, Meltzer RS, Koolen JJ, Dunning AJ. Short- and long-term predictive value of admission wall motions score in acute myocardial infarction: a two-dimensional echocardiographic srudy of 345 patients. Br Heart J 1986; 56:422-7. Visser CA, Kan G, Dunning AJ. A simple device to obtain reproducible echocardiographic apical cross-sections. Am J Cardiol1985;56:206-7. Weyman AE, Peskoe SM, Williams ES, Dillon JC, Feigenbaum H. Detection of left ventricular aneurysms by crosssectional echocardiography. Circulation 1976;54:936-44. Visser CA, Kan G, David GK, Lie KI, Durrer D. Echocardiographic-cineangiographic COHelation in detecting left ventricular aneurysm: a prospective srudy of 422 patients. Am J Cardiol 1982;50:337-41.
Volume 1 Number 3 May-June 1988
Value of echocardiography and hemodynamic measurements on admission
19. Jaarsma W, Visser CA, Eenige van MJ, et al. Prognostic implications of regional hyperkinesia and remote asynergy of noninfarcted myocardium. Am J Cardiol1986;58:394-8. 20. Visser CA, David GK, Kan G, et al. Two-dimensional echocardiography during percutaneous transluminal coronary angioplasty. Am Heart J 1986;111:1035-41. 21. Kisslo JA, Robertson D, Gilbert BW, von Ramm D, Beckar VS. A comparison of real-time two-dimensional left ventricular echocardiography and cineangiography in detecting left ventricular asynergy. Circulation 1977;55:134-40. 22. Loh IK, Charuzi Y, Beeder C, Marshall LA, Ginsbury JH. Early diagnosis of transmural infarction by two-dimensional echocardiography. Am Heart J 1982;104:963-8. 23. Russel RO, Hunt D, Rackley CE. Left ventricular hemodynamics in anterior and inferior myocardial infarction. Am J Cardiol1973;32:8-16.
193
24. Forrester JS, Diamond G, Chatterjee K, Swan HJC. Medical therapy of acute myocardial infarction by application of hemodynamic subsets (first of two parts). N Eng! J Med 1976;295: 1356-62. 25. Hutchins GM, Bulkley BH. Infarct expansion versus extension: two different complications of acute myocardial infarction. Am J Cardiol1978;41:ll27-32. 26. Schuster EH, Bulkley BH. Expansion of transmural myocardial infarction: a pathophysiologic factor in cardiac rupture. Circulation 1979;60: 1532-8. 27. Eaton LW, Weiss JL, Bulkley BH, Garrison JB, Weisfeldt ML. Regional dilatation after acute myocardial infarction. N Eng! J Med 1979;310:57-62.