Influence of Coronary Arterial Thrombus Location on Left Ventricular Function in Acute Myocardial Infarction Florence H. Sheehan, MD, Wolfgang G. Schmidt, MD, and Edward L. Bolson, MS
In 226 patients with acute myocardial infarction studied during intracoronary streptoldnase therapy, the circumferential extent of left ventricular hypsidnesis was measured by 5 methods and correlated with the tocation of the infarct-related coronary artery segment and with l-year survival. Of the 5 methods, 1 focused only on the infarct region, and 4 varied in the complexity of the noise filter. Hypohinetic segment length measurements by all 5 methods correlated significantly with the iocation of occlusion along the left anterior descending coronary artery. No method yielded measurements that correlated with occlusion location along the right coronary artery. Measurements by all methods correlated significantly with survival, but the method that focused on the infarct region performed least well. Thus, the circumferential extent of hypoidnesis in patients with acute myocardial infarction is greater for proximal than mid- or distal occlusions of the left anterior descending but not the right coronary artery. Survival is influenced by the function of periinfarct and noninfarct regions and by the function of the infarct region. Complex noise filters provide no advantage over simpler filters in measuring the extent of hypohinesis. (Am J Cardid 1990;66:16-21)
From the Cardiovascular Research and Training Center, University of Washington, Seattle, Washington, and the Department of Internal Medicine I, Rheinisch-Westfaelische Technische Hochschule, Aachen, West Germany. This study was supported in part by grant HL-19451 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, and the John L. Locke, Jr. Charitable Trust, Seattle. Manuscript received January 12, 1990, revised manuscript received and accepted February 26, 1990. Address for reprints: Florence H. Sheehan, MD, University of Washington, RG-22, Seattle, Washington 98195.
16
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
66
easurement of the severity of hypokinesis as a parameter of regional left ventricular (LV) function has proven most useful in studies of thrombolytic therapy, because of its sensitivity to the effect of treatment. Also, its independence of the infarct location and of coronary anatomy simplifies the analysis of clinical trials by allowing subgroups of patients with anterior and inferior infarcts, or with 11 coronary artery severely narrowed, to be analyzed together. In contrast, measurement of the circumferential extent of an infarct as a parameter of regional function is more variable, and therefore less sensitive, to the effect of treatment.’ One reason is that myocardial infarcts involving the left anterior descending coronary artery are more extensive, due to larger perfusion territory, than infarcts involving the other coronary arteries.2-4 With the increasing use of nonsurgical revascularization procedures, the ability to predict the hemodynamic impact of occluding a given coronary artery segment may be useful. The present study correlates the circumferential extent of LV dysfunction with the location of thrombotic occlusion of the left anterior descending and right coronary artery. Methods for measuring the extent of hypokinesis were also studied and compared.
M
METHODS Patient population: The study group comprised 238
patients with acute myocardial infarction (AMI) treated with intracoronary streptokinase therapy at the Rheinisch Westfaelische Technische Hochschule.5p6 All patients had total occlusion of either the left anterior descending or right coronary artery, and an analyzable contrast ventriculogram recorded in the 30’ right anterior oblique projection. Angiographii analysis: The coronary angiograms were assessed visually to identify the infarct-related artery segment (Figure 1). The ventriculograms were projected and traced at end-diastole and end-systole. The xy coordinates were analyzed in Seattle, Washington. LV volume was calculated using the area length method’ and wall motion was measured using the centerline method8 (Figure 2). Several approaches were used to measure the circumferential extent of hypokinesis at each of 3 thresholds: below -1 standard deviation, below -2 standard deviation or akinesis and dyskincsis. Method A measured the number of contiguous chords with motion below the threshold lying within a defined artery territory. Each territory corresponded to the span of chords likely to be affected by a large infarct due to occlusion of the
respective coronary artery. The territories were restricted for patients with multivessel disease, in order to distinguish the wall motion abnormality due to the current infarct from that due to disease in the other vessels (Table I, Figure 2). To filter occasional noise in the motion curves, noncontiguous chords with motion below the threshold were included in the -1 standard deviation hypokinetic segment if separated by 15 chords, in the -2 standard deviation segment if the gap was 110 chords, and in the akinetic segment if the gap was I1 5 chords (Figure 2). The circumferential extent of hypokinesis measured by method A has been shown to correlate closely with infarct size estimated from creatine kinase release.* Method B measured the length of the longest hypokinetic segment meeting the threshold, regardless of location on the LV contour (Figure 2, Table I). No adjustment was made for multivessel disease. The noise filter or allowable gap between noncontiguous segments was the same as for method A. Method Cl resembled method B but further refined the noise filter to consider the function of patients with hypokinetic segments separated by wide gaps. By method Cl, noncontiguous seg-
FIGURE 1. Coronary artery map used to identRy the proximal, mid- and dlstallegmntsoftho fir&t -v artery (1, 2,3, respectively) and the rkhtpoaerior-w artery (4), and the proximal, mid- and distal segmentsoftheleftantefiar dewnding artery (6, 7, 8, rospmtlvoly). AC = atrlal circumllox (branch); AM = acute marginal (branch); AV = at&mm+ bhlar (node branch); CB =cowsbrandqClRC= circumilox; Dl = 1st diagonal (branch); LCA = left wronary artery; Ml = 1st obtuse marginal (branch); PD = posterior dewending; RPD = rigM postdor descemhg; SN = sinus node (branch); V = (right) ventriculer (branch).
ments could be included in calculating the hypokinetic segment length segment if (1) the gap was 15 chords (10 chords at the -2 standard deviation threshold, 15 chords for akinesis), or (2) at least 50% of the chords in a wider gap had motion depressed below the percentage of mean normal motion that approximates the hypokinesis threshold (Figure 2). The -1 standard deviation threshold was approximated as motion depressed to 70% of normal, because in the normal reference population, motion 1 standard deviation below normal is motion depressed to 68 f 8% of normal (median 69%, range 43 to 83). Similarly, the -2 standard deviation threshold was approximated as motion depressed to 40% of normal, because an abnormality of -2 standard deviation is motion depressed to 36 f 17% of normal (median 39%, range -14 to 65). Akinesis, of course, is motion = 0% of normal. Methods C2 and C3 are similar to Cl except that the percent normal threshold was relaxed by 5 and lo%, respectively (Table I). Statistical analysis: The length of the hypokinetic segment resulting from the proximal, mid- or distal portions of the infarct artery, and the influence of single versus multivessel disease on these relations, were com-
V
SN
THE AMERICAN
JOURNAL
OF CARDIOLOGY
JULY 1, 1990
17
A
B
,
I
CHORD NUMBER
..‘.‘.. 1I
-4. -4
.
:
: ‘. ‘. ., ‘.” I iI
I
‘.
I 50
10
I
I 90
CHORD NUMBER FtGURE 2. Measurement of hypokinetic enddlastolll and end-systolii contours.
segment 6, motion
length by the csnterline method. A, a csntdine is drawn midway between ths is measured as the length of each of 100 chards comtmctsd perpediilar to in heart size, ths measured motion is normalized by the andeveniyspacedakmgthecentedine.C,toadjustfordiffemmes ~oftheend-dia~olicperimeter.Theakinetic~lengthisthepereentofthele*v~lareontowwith~e devlatlon. This motion SO cm. The paSent% motion is the se/id curve. The dotted curves indicate ths -aI mean f 1 sitar&d eas 2 a+hetic seymts. Fthod A interrog+s.only the infarct artery territory, chords 51 to 85 in this patient with ~and~~ck~lre,~oa~~~gmentlengthis~~~fnnnthe~mentrpamingchonlr79to 82.Theakineticsegmentlen%hiscaleulatedfnnnthelongerofthe?,chordD41to47,bymethodB.Sincethellkinetic segmeds are separatad by a gap >15 chords wide, the shorter one is ignored. Methods Cl, C2 and C3 will yield the same reRln~~BBbeecluoehaffoftheehordsbetweenthe2akinetiesegmentsdonotfaHbelowO%,S%orlO%ofnormal, ~~.~,a~inmotionisexprwrsedinunitsof~ndanl&~onsfrom~meanofammnalrefermce segmentlsngthiscakculatedasthenumberofcantigususcbordswithmotian pqndatm. The -1 stadard deviation hypokinetic more depmmed than -1 standard deviation. Since method A itiarrogates only ths tenitory of the itiarct artery,
the-l~devia~hypdcinesis~gmentextendsfnrmchonlsylto85only.In~sBandCl,the nmmdgwm
hypokinetic
segmeA
from
chords
21 to 49 is considered
part
of the
~~~ng~5qto1oObecausethegapbetweenthese2segmentsisnarrow(~5ehords).Thenoneontiguous ~~ow$ 1 to 4 IS separated by a wider gap, but half of the intervening the hypokinetic
18
nom~l,75% segmant according
THE AMERICAN
JOURNAL
of normal or 80% to methods Cl,
OF CARDIOLOGY
of -I.
Therefore,
C2 or C3, respedively.
VOLUME
66
-1 standard deviation hypddnetic
chords the noncxMiguws
do not have motion chords are not included
in
TABLE I Methods
for Measurement
of the Circumferential
Extent
of Hypokinesis Falter
Cn3rrh “CYl _I, Territory
Method
Subgroup
A
LAD alone
5-85
LAD + others
5-67
Rrght alone
25-85
Rrght + others
51-85
B
All
l-100
Cl
All
l-100
c2
All
l-100
c3
All
l-100
LAD = left anterior
TABLE
descendmg
II Coronary
coronary
Anatomical
artery. SD = standard
rn Artery
Right coronary (%) Proximal Mid Distal Posterior descendrng Left anterror descendrng Proxrmal Mid Distal
Only 1
>l
30 (28) 15(14)
35 (27) 20(15)
4 (4) 3 (3)
6(5)
-1 SD -2SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm
5 10 15 5 10 15 5 10 15 5 10 15 5 10 15 5 10 15 5 10 15 5 10 15
70 40 0 75 45 5 80 50 10
23 (21) 33 (30) 0
27 (21) 36 (28)
One-Year
Ill Correlation
Of the 238 patients, 115 (48%) had occlusion of the right coronary artery and 112 (47%) had l-vessel disease.The incidence of proximal, mid- and distal infarct artery occlusion is listed in Table II. anatomy:
hypokinetic segment length and artery occlusion: The location of
Segment
Threshold
Method
Uwariate
-1
A 0 Cl c2 c3 A B Cl c2 c3 A B Cl c2 c3
9.7 23.6 21.6 20.0 16.2 9.3 16.1 15.8 16.3 16.2 10.5 10.8 11.0 11.0 10.4
-2
SD
SD
(%)
4 (3)
of Hypokinetic
Length
with
Survival
2(l)
RESULTS
Relation between location of coronary
9’~ Normal Mean
TABLE
Characteristics
pared using analysis of variance. The prognostic power of the parameters of hypokinetic segment length were compared using discriminant analysis for l-year survival.
Coronary
Non contiguous
dewtfon
No. of Coronary Arteries Severely Narrowed Artery Occluded and Locatron of Occlusion
Threshold
Akinesis
SD = standard
F
p Value
In 226 patients with acute myocardial infarction studied during intracoronary streptoldnase therapy, the circumferential extent of left ventricular hypsidnesis was measured by 5 methods and correlated with the tocation of the infarct-related coronary artery segment and with l-year survival. Of the 5 methods, 1 focused only on the infarct region, and 4 varied in the complexity of the noise filter. Hypohinetic segment length measurements by all 5 methods correlated significantly with the iocation of occlusion along the left anterior descending coronary artery. No method yielded measurements that correlated with occlusion location along the right coronary artery. Measurements by all methods correlated significantly with survival, but the method that focused on the infarct region performed least well. Thus, the circumferential extent of hypoidnesis in patients with acute myocardial infarction is greater for proximal than mid- or distal occlusions of the left anterior descending but not the right coronary artery. Survival is influenced by the function of periinfarct and noninfarct regions and by the function of the infarct region. Complex noise filters provide no advantage over simpler filters in measuring the extent of hypohinesis. (Am J Cardid 1990;66:16-21)
From the Cardiovascular Research and Training Center, University of Washington, Seattle, Washington, and the Department of Internal Medicine I, Rheinisch-Westfaelische Technische Hochschule, Aachen, West Germany. This study was supported in part by grant HL-19451 from the National Heart, Lung, and Blood Institute, Bethesda, Maryland, and the John L. Locke, Jr. Charitable Trust, Seattle. Manuscript received January 12, 1990, revised manuscript received and accepted February 26, 1990. Address for reprints: Florence H. Sheehan, MD, University of Washington, RG-22, Seattle, Washington 98195.
16
THE AMERICAN
JOURNAL
OF CARDIOLOGY
VOLUME
66
easurement of the severity of hypokinesis as a parameter of regional left ventricular (LV) function has proven most useful in studies of thrombolytic therapy, because of its sensitivity to the effect of treatment. Also, its independence of the infarct location and of coronary anatomy simplifies the analysis of clinical trials by allowing subgroups of patients with anterior and inferior infarcts, or with 11 coronary artery severely narrowed, to be analyzed together. In contrast, measurement of the circumferential extent of an infarct as a parameter of regional function is more variable, and therefore less sensitive, to the effect of treatment.’ One reason is that myocardial infarcts involving the left anterior descending coronary artery are more extensive, due to larger perfusion territory, than infarcts involving the other coronary arteries.2-4 With the increasing use of nonsurgical revascularization procedures, the ability to predict the hemodynamic impact of occluding a given coronary artery segment may be useful. The present study correlates the circumferential extent of LV dysfunction with the location of thrombotic occlusion of the left anterior descending and right coronary artery. Methods for measuring the extent of hypokinesis were also studied and compared.
M
METHODS Patient population: The study group comprised 238
patients with acute myocardial infarction (AMI) treated with intracoronary streptokinase therapy at the Rheinisch Westfaelische Technische Hochschule.5p6 All patients had total occlusion of either the left anterior descending or right coronary artery, and an analyzable contrast ventriculogram recorded in the 30’ right anterior oblique projection. Angiographii analysis: The coronary angiograms were assessed visually to identify the infarct-related artery segment (Figure 1). The ventriculograms were projected and traced at end-diastole and end-systole. The xy coordinates were analyzed in Seattle, Washington. LV volume was calculated using the area length method’ and wall motion was measured using the centerline method8 (Figure 2). Several approaches were used to measure the circumferential extent of hypokinesis at each of 3 thresholds: below -1 standard deviation, below -2 standard deviation or akinesis and dyskincsis. Method A measured the number of contiguous chords with motion below the threshold lying within a defined artery territory. Each territory corresponded to the span of chords likely to be affected by a large infarct due to occlusion of the
respective coronary artery. The territories were restricted for patients with multivessel disease, in order to distinguish the wall motion abnormality due to the current infarct from that due to disease in the other vessels (Table I, Figure 2). To filter occasional noise in the motion curves, noncontiguous chords with motion below the threshold were included in the -1 standard deviation hypokinetic segment if separated by 15 chords, in the -2 standard deviation segment if the gap was 110 chords, and in the akinetic segment if the gap was I1 5 chords (Figure 2). The circumferential extent of hypokinesis measured by method A has been shown to correlate closely with infarct size estimated from creatine kinase release.* Method B measured the length of the longest hypokinetic segment meeting the threshold, regardless of location on the LV contour (Figure 2, Table I). No adjustment was made for multivessel disease. The noise filter or allowable gap between noncontiguous segments was the same as for method A. Method Cl resembled method B but further refined the noise filter to consider the function of patients with hypokinetic segments separated by wide gaps. By method Cl, noncontiguous seg-
FIGURE 1. Coronary artery map used to identRy the proximal, mid- and dlstallegmntsoftho fir&t -v artery (1, 2,3, respectively) and the rkhtpoaerior-w artery (4), and the proximal, mid- and distal segmentsoftheleftantefiar dewnding artery (6, 7, 8, rospmtlvoly). AC = atrlal circumllox (branch); AM = acute marginal (branch); AV = at&mm+ bhlar (node branch); CB =cowsbrandqClRC= circumilox; Dl = 1st diagonal (branch); LCA = left wronary artery; Ml = 1st obtuse marginal (branch); PD = posterior dewending; RPD = rigM postdor descemhg; SN = sinus node (branch); V = (right) ventriculer (branch).
ments could be included in calculating the hypokinetic segment length segment if (1) the gap was 15 chords (10 chords at the -2 standard deviation threshold, 15 chords for akinesis), or (2) at least 50% of the chords in a wider gap had motion depressed below the percentage of mean normal motion that approximates the hypokinesis threshold (Figure 2). The -1 standard deviation threshold was approximated as motion depressed to 70% of normal, because in the normal reference population, motion 1 standard deviation below normal is motion depressed to 68 f 8% of normal (median 69%, range 43 to 83). Similarly, the -2 standard deviation threshold was approximated as motion depressed to 40% of normal, because an abnormality of -2 standard deviation is motion depressed to 36 f 17% of normal (median 39%, range -14 to 65). Akinesis, of course, is motion = 0% of normal. Methods C2 and C3 are similar to Cl except that the percent normal threshold was relaxed by 5 and lo%, respectively (Table I). Statistical analysis: The length of the hypokinetic segment resulting from the proximal, mid- or distal portions of the infarct artery, and the influence of single versus multivessel disease on these relations, were com-
V
SN
THE AMERICAN
JOURNAL
OF CARDIOLOGY
JULY 1, 1990
17
A
B
,
I
CHORD NUMBER
..‘.‘.. 1I
-4. -4
.
:
: ‘. ‘. ., ‘.” I iI
I
‘.
I 50
10
I
I 90
CHORD NUMBER FtGURE 2. Measurement of hypokinetic enddlastolll and end-systolii contours.
segment 6, motion
length by the csnterline method. A, a csntdine is drawn midway between ths is measured as the length of each of 100 chards comtmctsd perpediilar to in heart size, ths measured motion is normalized by the andeveniyspacedakmgthecentedine.C,toadjustfordiffemmes ~oftheend-dia~olicperimeter.Theakinetic~lengthisthepereentofthele*v~lareontowwith~e devlatlon. This motion SO cm. The paSent% motion is the se/id curve. The dotted curves indicate ths -aI mean f 1 sitar&d eas 2 a+hetic seymts. Fthod A interrog+s.only the infarct artery territory, chords 51 to 85 in this patient with ~and~~ck~lre,~oa~~~gmentlengthis~~~fnnnthe~mentrpamingchonlr79to 82.Theakineticsegmentlen%hiscaleulatedfnnnthelongerofthe?,chordD41to47,bymethodB.Sincethellkinetic segmeds are separatad by a gap >15 chords wide, the shorter one is ignored. Methods Cl, C2 and C3 will yield the same reRln~~BBbeecluoehaffoftheehordsbetweenthe2akinetiesegmentsdonotfaHbelowO%,S%orlO%ofnormal, ~~.~,a~inmotionisexprwrsedinunitsof~ndanl&~onsfrom~meanofammnalrefermce segmentlsngthiscakculatedasthenumberofcantigususcbordswithmotian pqndatm. The -1 stadard deviation hypokinetic more depmmed than -1 standard deviation. Since method A itiarrogates only ths tenitory of the itiarct artery,
the-l~devia~hypdcinesis~gmentextendsfnrmchonlsylto85only.In~sBandCl,the nmmdgwm
hypokinetic
segmeA
from
chords
21 to 49 is considered
part
of the
~~~ng~5qto1oObecausethegapbetweenthese2segmentsisnarrow(~5ehords).Thenoneontiguous ~~ow$ 1 to 4 IS separated by a wider gap, but half of the intervening the hypokinetic
18
nom~l,75% segmant according
THE AMERICAN
JOURNAL
of normal or 80% to methods Cl,
OF CARDIOLOGY
of -I.
Therefore,
C2 or C3, respedively.
VOLUME
66
-1 standard deviation hypddnetic
chords the noncxMiguws
do not have motion chords are not included
in
TABLE I Methods
for Measurement
of the Circumferential
Extent
of Hypokinesis Falter
Cn3rrh “CYl _I, Territory
Method
Subgroup
A
LAD alone
5-85
LAD + others
5-67
Rrght alone
25-85
Rrght + others
51-85
B
All
l-100
Cl
All
l-100
c2
All
l-100
c3
All
l-100
LAD = left anterior
TABLE
descendmg
II Coronary
coronary
Anatomical
artery. SD = standard
rn Artery
Right coronary (%) Proximal Mid Distal Posterior descendrng Left anterror descendrng Proxrmal Mid Distal
Only 1
>l
30 (28) 15(14)
35 (27) 20(15)
4 (4) 3 (3)
6(5)
-1 SD -2SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm -1 SD -2 SD Ocm
5 10 15 5 10 15 5 10 15 5 10 15 5 10 15 5 10 15 5 10 15 5 10 15
70 40 0 75 45 5 80 50 10
23 (21) 33 (30) 0
27 (21) 36 (28)
One-Year
Ill Correlation
Of the 238 patients, 115 (48%) had occlusion of the right coronary artery and 112 (47%) had l-vessel disease.The incidence of proximal, mid- and distal infarct artery occlusion is listed in Table II. anatomy:
hypokinetic segment length and artery occlusion: The location of
Segment
Threshold
Method
Uwariate
-1
A 0 Cl c2 c3 A B Cl c2 c3 A B Cl c2 c3
9.7 23.6 21.6 20.0 16.2 9.3 16.1 15.8 16.3 16.2 10.5 10.8 11.0 11.0 10.4
-2
SD
SD
(%)
4 (3)
of Hypokinetic
Length
with
Survival
2(l)
RESULTS
Relation between location of coronary
9’~ Normal Mean
TABLE
Characteristics
pared using analysis of variance. The prognostic power of the parameters of hypokinetic segment length were compared using discriminant analysis for l-year survival.
Coronary
Non contiguous
dewtfon
No. of Coronary Arteries Severely Narrowed Artery Occluded and Locatron of Occlusion
Threshold
Akinesis
SD = standard
F
p Value
