Hirui

et al.

Amervan

11. Kennedy JW, Treholme SE, Kasser IS. I,eft ventricular \ (8~ ume and mass from single plane angioyram: a ~~r~mparison (/I anteroposterior and right anterior oblique methods. Ahr H~;,.II; I .I 1970;80::344-52. 12. Fujica M, Sasayama 5, Kawai C, Eiho S. Kuwahara %I. .4urclmatic processing of rineventriculograms for analysis of re gional myocardial infarction. Circulation 1981;6:3:1065-71. ItI. Schwartz H, Leiboff RH, Bren GB. Wasserman A(+, Katz f&J. Varghese PJ, Solil AB, Ross AM. Temporal evolution of the human coronary collateral circulation after myocardial infarction. ,J Am Coil Cardiol 1984;4:1088-9X

Heart

July 1992 Journai

i I Schwartz ti. I,eiboff KL. Itarz Rd. \t~asserman A(;. Bren (;li Varghese I’.]. Ross AM. Arteriographic predictorh of sp”nt;l neous imprljvement in left ventricular function after *nyr,c’;lr dial infarct ion. rirculation 198.7:71 :#$ii-72. Ii. ‘~‘usuf S. I’eto R, Lewis J. (‘ollina R. sleight I’. Beta-blockatl~~ during and after myocardial infarction: an clcerview of r,hc, randomizt,rl trials. Prog Cardiovasc I)is 1985;27:355-71. 16. Braunwa1l.l E. Muller cJE, Iill,ner RA. Marokc PR. Role OI Ixra-adrenergic blockade in the therapy 01’ patients with my.’ in t,he t.wo settings. We postulated that if the apparent discrepancy between experimental findings and those observed clinically was due solely t,o a more protracted time course for recovery of function in humans, then a longer period of observation after rnyocardial infarction would be necessary to fully delineate this recovery process. Therefore by means of cIuantit,ative echocardiography we followed patients for I year after myocardial infarction to charact,erize the long-term pattern of recovery of postinfarction ventricular structure and function and to determine whether long-term recovery was influenced by infarct type or location. METHODS Study group. The data for this study are part of a large prospective echocardiographic study of the natural history of myocartlial infarction conducted at Massachusetts General Hospi; al. Two-dimensional echocardiography was performed in patients with suspected first myocardial infarction on admission to the hospital. Myocardial infarction was subsequently defined by the combination of a compatible clinical course including chest pain lasting at least 30 minutes, evolution of ECG changes consistent with a Q wave or non---Q wave infarction, and a significant rise in the cteatine knase level and cardiac isozyme fraction during the tirst 36 hours of hospitalization. The ECG between 24 and 72 hours after admission wit,h the largest Q wave distribution was used to assess the presence and location of pathologic Q waves. New York Heart Association criteria nere used to assign the anatomic site of Q waves’s P.atients with documented myocardial infarction who were excluded from this study included those with previous myocardial infarction, primary valvular heart disease, or cardiomyopathy, onset of symptoms more than 24 hours before screening, death before completion of 1 year of observation md those treated with interventions such as acute throm bolytic therapy, coronary angioplasty, or coronary surgery. Patients with echocardiograms of inadequare quality for quantitative analysis were also excluded. Patients selected for this analysis included those who underwent serial echocardiography on entry into the hospital,

Late improcements

in LV sizlfunction

after AMI

25

L&S- 11.5 C\IV19.7 CFhl= 22.8 CAP= 17.5 E.SL- 12.6 ESA= 219.62 AWhl AREA = 90.2 I %AWhl= 4l.l

Fig. 1. Map of left ventricular endocardial surface area with superimposed area of abnormal wall motion (black) reconstructed from echocardiographic data. Ap3C, Axis of apical four-chamber view; ApPC, axis of apical two-chamber view; LAX, long-axis length (cm); CMV, endocardial circumference of left ventricle at mitral valve level (cm); CPM, endocardial circumference of left ventricle at midpapillary muscle level (cm); CAP, endocardial circumference of left ventricle at apical level (cm); ESL, endocardial segment length (cm); ESA, endocardial surface area (cm”); .4WM area, area of abnormal wall motion (cm’); PC A WM, percentage of total ESA involved by AWM.

at 3 months after myocardial infarction, and 1 year after myocardial infarction. Seventy-two patients met the eligibility criteria and 52 completed the study. Of the 20 eligible patients who did not complete the study, nine refused follow-up, eight lived outside the geographic area and were unavailable for follow-up, and three were lost to follow-up. Those patients with evidence of recurrent infarction represented by a second episode of chest pain accompanied by a separate increase in cardiac enzyme levels or the appearance of new Q waves on the ECG at any time during the follow-up period were excluded from the analysis Cn = 1). Two-dimensional echocardiography. Two-dimensional echocardiographic studies were performed in all patients with either an Advanced Technology Laboratories, Inc., (Bothell, Wash.) Mk 600 mechanical sector scanner or a Hewlett-Packard 77020A phased-array scanner (HewlettPackard Co., Medical Products Group (Andover, Mass.) and images were recorded on one-half inch VHS videotape. All studies were performed by the same sonographer (P. A. R.). Images of the left ventricle were obtained from five standard imaging planes: parasternal short-axis views at the level of the mitral valve, the midportion of the papillary muscles, and the apex along with the apical fourchamber and two-chamber views. End-diastolic frames of the images from the five re-

26

Table

Picard et al.

I. Patient

American

Heart

July 1992 Journal

characteristics

Ch aracteristics Age

(yr)

63

Male/Female Creatine Kinase Total creatine kinase (IU) ccMB Q wave location Anterior Anterolateral Inferior Posterior None AWM location Anterior Inferior Posterolateral None Ant-exp,

Table

Anterior,

expansion

group;

Ant-nu

II. Echocardiographic

exp,

* 1

.Y5 i :,

7 :!

317

1,o.ir, t :wt 15 ’ I -t 3

1>-“89 - :19:! I”’- / + ”

599 1 I l(l I”‘, f I

5 0 0 0 Ii

11 1 0 (I II

,> II 17 1 0

*i 0 0 0

I:! II 0 0

0 I ‘1 4 (I

anterior.

no expansinn:

data Ant-exp (N =

57

ii1

inf.

inferior

infarction:

22/i

NV.

non-Q

maw

infarction.

Ant-no PXI) (N = 12)

5)

-

ESA index (cm2/m’) Entry 3 mo 1 yr

AWM (cm’) Entry 3 mo 1 yr

Anl-ewp, Anterior docardial

surface

infarction,

area;

AWM,

84.0 + 6.4 96.3 t 8.6+

59.5

(i2.1

I

.59.7 k “.o*

62.2

+ 1.9*

81.5

+ 4.2$

54.7

+ ‘.O*F

59.8

62.5 k 10.3

79.1 _+ 18.6

“8.6 “7.5

i- 5.4 ’ t 6.“*

“7.1 i :4.:;* 2l.H i xa*n

59.8

16.5

t -l.9*gj

17.1

expansion ahnormal

*

14.9$ group: wall

Ant-no motion

exp. anterior

~2 2.1.

infarction.

no expansion:

irrj, inferior

i

1.7*

ci4.n

1.9*

t “.S*t# infarction:

t ‘2.7 k

c-l.1 -t :1.,5i

NY, non-Q

57.9

4 -l.4*gll

IX.9 12.1 .i.l

k 4.5” +. r,.vq 1 2.4*#

wave

infarction:

ES:\.

FW

area

-p < 0.005 c/w ant-exp. tp < 0.001

c/w

entry.

ip < (1.05 c/w 3 months. sp < 0.01 c/w entry. lip < 0.01 c/w 3 months.

I[p < 0.05 c/w entry. #p < 0.005 c/w entry.

corded planes were analyzed off line to obtain the cardiac dimensional measurements and wall motion data required to reconstruct the endocardial surface of the left ventricle. This was performed with the assistance of a computer algorithm previously reported and validated.“, lFi. 15’.x By this method the left ventricular surface is represented graphically as a planar map composed of four quadrants (Fig. 1). The endocardial surface area (ESA) of the left ventricle, a measure of left ventricular size, was calculated as the sum of the surface area of the four map quadrants. Assessment of wall motion was performed without knowledge of the patients’ identity, ECG findings, or clinical course. In each view the region of dyssynergy was identified by repeated viewing to define the margins of normal and abnormal segments. Dyssynergy was defined as any reduc-

tion in systolic endocardial excursion and thickening. The endocardial length of the abnormal segment was measured along with its distance from predetermined internal landmarks. The area and location of this AWM was then trans ferred to the map of the total endocardial surface and rep resented the area of infarct.-related dysfunction (Fig. I i. The area of AWM was calculated as the sum of the surt’ace area of AWM in the four map quadrants. Inasmuch as body surface area has been identified as H primary determinant of normal left ventricular dimensions,“’ ” the ESA was corrected for variations in patient size by dividing it by the body surface area. The corrected ESA or ESA index was used in this analysis for comparisons between patients. Regions of AWM were present, on the initial echocardiogram in all patients in t.he study. Tht

Volume Number

124 1

Late improvements

in LV size/fu,dion

after AMI

27

ESAi (cm*/m*)

65

J

nonQ inf

14 55 -

* i*

ant, no exp

l

45

entry

pco.05

1 year

3 month

Fig. 2. Graph of endocardial surface area index(ESAiJ over l-year observation period for four subgroups. ant, exp, Anterior infarction with early expansion; ant, no exp, anterior infarction without early expansion; inf. inferior infarction; nonQ, non-Q wave infarction.

locations of AWM were classified on the basis of the predominant wall involved. 33 As previously described the intraobserver and interobserver measurement variability and the temporal variation in ESA were all less than j “;,

11, 13

Control group. Thirty normal subjects served as a control group. The group was composed of 18 men and 12 women ranging in age from 23 to 65 years (mean age 35 years). The normal range for the ESA index was 53 to 78.9 cm2,im” with a mean ESA index of 64.9 cm’/m?. Study groups. Because in previous studies we demonstrated a significant difference in early changes in AWM and ESA in Q wave infarctions based on the location of the AWM,13 patients with Q wave infarctions were separated into two subgroups on the basis of the location of the initial wall motion abnormality: anterior and inferior. Those with non-Q wave infarction by ECG were analyzed as an additional subgroup. The group with anterior infarctions was further separated on the basis of the presence of early infarct expansion. As described previously,13 infarct expansion was defined as present when the initial ESA index exceeded 77.9 cm”/m”, which was two standard deviations above the mean of the normal control population. Statistical analysis. Data are presented as mean values + standard error of the mean. Comparisons of frequency of events were performed with either Fisher’s test or chi-square analysis depending on the frequency of events. The differences and interactions between multiple subgroups were examined by multivariate repeated-measures analysis of variance. Linear regression analysis was used to assess the relationship between the change in the ESA and the initial infarct size.

RESULTS Study group. Fifty-one of the eligible patients completed the study. This study population consisted of

12 women and 39 men. The mean age of the population was 56 f 2 years. The patient demographics, clinical, and ECG data for the infarct groups are displayed in Table I. There were 40 patients with Q wave myocardial infarctions-17 with anterior and 23 with inferior

regions

of AWM.

The

subgroup

with

non-Q

wave infarctions was composed of 11 patients. Five of the patients with Q wave infarction and anterior wall motion abnormalities met the criteria for infarct expansion. No patients with non-Q wave infarction had an initial ESA index outside of the normal range. One inferior infarction had an enlarged ESA at entry because of an extensive region of AWM extending from the base of the left ventricle to and including a port,ion of the apex. Change in endocardial surface area over 1 year by infarct location. As would be expected from the defini-

tion of the subgroups, the initial mean ESA index for the expansion group was significantly larger than that of the other three groups (Table II, Fig. 2). The mean ESA of the remaining three groups at entry (anterior-nonexpander, inferior, and non-Q wave) were not significantly different from each other or from normal. In all five patients in the expansion group, the ESA index increased during the first 3 months after infarction (84.0 i 6.4 cm2/m2 to

28

Picard et al.

American

60

ESA change

Heart

July 1992 Journal

(cfn2)

r

+

+

x

anterior

+

inferior

-60 1.11---A10

10

20

30

entry

40

50

60

70

80

AWM (cm21

3. Plot of changein ESA (ESAt,,-ESA:j,,,) with entry AWM for Q wave infarctions, stratified by infarct location: anterior (asterisk) or inferior (dagger). Mean (solid line) is alsodisplayed. Whether examined by infarct location or for all infarctions, there wasno significant univariate linear correlation between these two variables.

Fig.

96.3 + 8.6 cm”/m”, p < 0.001). However, between the S-month and l-year time points the mean ESA index decreased to 81.5 k 4.2 cmZ/mZ (p < 0.02). For the anterior infarctions not demonstrating infarct expansion, the mean ESA index was not, significantly different at 3 months compared to values at entry (59.5 +- 2.1 cm2/m” vs 59.7 + 2.0 cm”/m”). However, by 1 year a decrease to 54.7 k 2.0 cm’lm” was observed (p < 0.006). A nearly identical trend in the ESA index was noted for the group with non-Q wave infarctions. The mean ESA index for this group did not change significantly from entry to 3 months (64.5 k 2.7 cm”/m” vs 64.1 + 3.1 cm”/m”). By 1 year a decrease to 57.9 k 4.4 cm”/m’ was observed (p < 0.008). Although a trend toward a decrease in the ESA index was noted for the group with inferior infarction at 1 year, this change did not reach statistical significance. Late change in endocardial surface area by initial infarct size. Fig. 3 displays the change in ESA between

3 months and 1 year for all Q wave infarctions as a function of initial infarct size. Despite differences in ESA at 3 months, the absolute change in ESA from 3 months to 1 year when examined by linear regression analysis was not significantly influenced by initial infarct size. For these 40 patients the mean decrease in ESA was 8.9 i 2.5 cm3. Similarly when this relationship was examined separately for anterior

and inferior infarctions, the initial infarct, size did not significantly influence the degree of change observed in ESA bet,ween 3 months and 1 year. Thus the rate of change in ESA between 3 months and 1 year did not vary significantly among infarctions of differing locations or size. Change in abnormal wall motion over 1 year by infarct location. At entry the mean AWM area of the group

with anterior expansion was larger than that of the other groups (p < 0.0005) (Table II, Fig. 4). This difference persisted throughout the l-year follow-up period. Significant decreases in AWM were noted in all groups during the year. The timing of these decreases, however, differed among the groups. Whereas significant, decreases in mean AWM were noted between entry and 3 months in both the inferior and non-Q wave infarction groups, significant decreases in AWM in both anterior groups were noted only after 3 months. DISCUSSION

Results of this study illustrate that if examined over a long enough period, left ventricular structure. asmeasured by the ESA, and left ventricular regional function, as measured by the extent of AWM, demonstrate variable but continued improvement, even in the absence of interventions such as thrombolytic therapy, coronary angioplasty, and bypass surgery.

Volume

124

Number

1

Late improvements

100

in LV size/function

after AMI

29

AWM (cm21 l

p to.05

80

80

**

nonQ entry

3 month

1 year

Fig. 4. Graph of area of abnormal wall motion (AWM) over l-year observation period for four subgroups. ant, exp, Anterior infarction with early expansion;ant, no exp, anterior infarction without, early expansion: inf, inferior infarction; nonQ, non-Q wave infarction.

Changes

in ventricular

structure

during

long-term

fol-

low-up. These results confirm previous observations that early (up to 3 months) increases in ventricular size after acute myocardial infarction relate primarily to the size and location of the infarctionZJs ‘sl 24-“6 Specifically infarct expansion occurs in up to 35 “;. of anterior infarctions with apical involvement “3“3 “’ and can be identified within the first 24 hours of the acute event by regional enlargement of the endocardial surface. I3 In this subset of patients, progressive dilatation and expansion of the infarct zone continues during the first 3 months after infarction in the absence of intervention. However, our results suggest that late (3 months to 1 year) decreases in both left ventricular size and regional function are independent of infarct location. Specifically, decreasesin the ES4 of a similar magnitude were observed in anterior and inferior infarctions independent of the presence of early infarct expansion. Despite these late improvements, the ventricles with early infarct expansion remain abnormally enlarged. Late improvements in ventricular size have not been noted in previous studies for three possible reasons: (1) most natural history studies examining this process have not been extended to a long enough observation periods, 24,“i,58; (2) high mortality rates in the studies with l-year follow-up of infarct expansion have excluded an adequate sample size for meaningful analysis of ventricular sizeZg;or (3) intermediate

studies were not performed and thus expansion with later contraction was simply interpreted as absence of change or expansion only?” A previous angiographic study has reported that large anterior infarctions (more than a 30’ c region of akinesis/dyskinesis) demonstrate enlargement over 1 year.’‘n In our study the rate of progression of enlargement noted from admission t,o 3 months was att,enuated by 1 year. Inasmuch as the extent of akinesis and dyskinesis at 1 year tended to be less than 3O“Oin our subgroup with infarct expansion, this apparent difference between studies may illustrate the primary importance of infarct size in determining left ventricular size. Three mechanisms could explain the recovery of ESA observed in this study: effects of pharmacotherapy, recovery of reversibly ischemic (“hibernating”) myocardium”“-“” from the lateral border zones of the infarction, or scar contraction.‘5* “‘-“a In this study population no changes in drug therapy were noted in the late post-myocardial infarction period, and thus it is unlikely that the late decrease in ESA was a pharmacologic effect. The second potential mechanism would assume that there is viable yet dysfunctioning myocardium at the border zones, and t.he third would assume that scar formation has replaced necrotic myocytes. Although the morphologic pattern of recovery of hibernating myocardium is unknown, one might hypothesize that the ESA recovery

30

Picard et al.

rate would vary markedly and depend on many factors, particularly patency of the infarct-related artery or the development of collateral vessels. Thus the timing of recovery of hibernating myocardium would not necessarily be constant for all infarctions but would relate t,o the point at which adequat,e flow was restored to this region. In our group of pat,ients this mechanism could only account for the observed changes if spontaneous reperfusion or establishment of collateral flow occurred late and relatively consistently in all of these patients. A shrinkage of the infarct zone as a function of scar contraction would lead t,o a decrease in the total ESA in addition to a decrease in the region of AWM. Since this process would be solely a mechanical one and independent of the size or location of the infarction, an identical rate of change in the ESA could be observed across the patient population. Alt,hough the purpose of t,his study was not to prove the actual mechanism of recovery, our findings of a relatively identical change in t,he ESA regardless of infarct size and location support scar contraction between 3 months and 1 year after infarction as a major factor in the decrease of the ESA. These results serve as a functional correlate of a previous histologic analysis of infarctions in humans that demonst,rated scar contraction by 3 months after the acute event.“;’ Changes in regional function during long-term followup. In most patients studied, regional dysfunction, as

measured by the extent, of AWM on the echocardiogram, demonstrated improvement over time regardless of infarct location (anterior vs inferior) or infarct t,ype (Q wave vs non-Q wave). The time course OSthe improvement, however. differed as a function of infarct location and type. The natural history of wall motion abnormalities up to 1 year aft.er infarction has not previously been systematically quantitated by echocardiography. Studies examining changes in wall motion abnormalities in t,he early postinfarction period have found improvement in regional function to be related to the size,4” type,41, and location of the infarctionl” and to t.he patency of the infarct-relat,ed artery.‘“,“” Our data also suggest that infarct location, size. and type are predictors of AWM recovery. However, although prior studies were limited to the assessment of recovery of regional function for less than 3 months after infarction, our results demonstrate that all infarctions show continued funct,ional recovery from 3 months to 1 year, whereas most of the improvement in AWM for large anterior infarctions occurs later than 3 months. These data are consistent with results of previous long-term studies of infarction in dogs. where improvement in ischemic segment length and recover!

Ameucan

July 1992 Heart Journal

of regional dysfunction has been observed within G weeks after coronary occlusion.‘4-‘” The diff’erences in the rate of recovery of function observed in canine and human studies may be related to the extensive collateral coronary circulation found in the dog. Ah observed in our clinical study, t,he canine studies also) demonstrated earlier recovery of the region of AWM aft,er circumflex occlusion than after occlusion oft he left anterior descending artery. By quantitating the condensation of microspheres in the infarct zones in the canine experiments, the decrease in the extent (rt AWM has been shown to be due to scar contrac,tion.‘,‘, “” a mechanism also supported by our data. Our study required survival to 1 year after myocardial infarction wit,hout recurrent ischemia or (*or-onary intervention. Thus these results may be represent,ative of only a port,ion of infarct sllrvivors and cannot be used to infer prognosis because of an inherent selection bias. In addition, many factors that may potentially aff’ect ventricular remodeling. such as arterial and ventricular pressure, patency ot coronary arteries and medications, were not controlled in this study. However, as noted previously. no consistent change in medications was introduced during the :i-month to 1-year follow-up period and yet the pattern of recovery was similar in all groups. We conclude that global left ventricular size decreases and regional wall motion improves in both Q wave and non-Q wave infarct ions during the year af’ ter the acute event. Recovery of’ lef’t ventricular structure occurs at a relatively constant rate from 13 months to 1 year regardless of the size or location of t,he infarction. In contrast, the time course and rate of recovery of function are related to the location and size of t,he region of dysfunction. Recnvery of regiona I function begins in both inferior Q wave and norl--Q wave infarctions within t,he first 3 months. whereas changes in the extent of AWM in anterior Q wave irlfarctions is not observed until after this time point. This variability of recovery must he taken into account when interpreting the results of’ interveIl-tions that att,empt to limit infarct size or alter vt’~ltricular remodeling. We thank ,Jt)hn B. Newell, BS, t’or his expert advice on htatlstl cal analyses. and Charles Dennis, MD, and -James I). Thomas, MD. for their helptul comments in the preparation of the manuscripr

REFERENCES

I. Heger J. 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. 2. Heger .J. Weyman AE, Wann LS, Rogers EW, Dillon .I(‘. Feigenhaum H. Cross sectional echocardiographic analysis of the extent ol’left ventricular asynergy in acute myocardial infarction. Circulalion 1980:6l:lllX-X.

Volume

124

Number

1

3. Eaton L. Weiss JL, Bulkley BH, Garrison JB, Weisfeldt ML. Regional cardiac dilatation after acute myocardial infarction. N Engl .I Med 1979;300:57-62. KA. Smitherman TC. Estimation of my4. Nixon JV. Narahara ocardial involvement in patients with acute myocardial infarction by two dimensional echocardiography. Circulation 1980;6”: 1218-55. Ft. Parisi AF. Moynihan PF, Folland ED, Feldman CL. Quantitative dttection of regional left ventricular contraction abnormalities by two-dimensional echocardiography. II. Accuracyi.1 coronary artery disease. Circulation 1981;63:761-7. R, Durrer D. Detection and fi. \:iaser CA. Lie KI, Kan G, Meltzer a uantifil*ation of acute, isolated myocardial infarction by twocimensional echocardiography. Am J Cardiol 1981;47:1020-5. 7. Gibson RS. Bishop HL, Stamm RB, Crampton RS, Beller GA. ?ilartin RP. Value of early two-dimensional echocardiography in patie,rts with accute myocardial infarction. Am .J Cardiol 1982:4S:1110-19. .JA, Weiss JL, Eaton LW. Kallman C, Weisfeldt sI. Erlebacher LIL. Bulklev BH. Late effects of acute infarct dilation on heart size: a two dimensional echocardiographic study. Am J Cardiol 198’;49: 11 “O-6. 9. \Veyman AE, Peskoe SM, Williams ES, Dillon JC, Feigenbaum H. Detection of left ventricular aneurysms by cross sectional ec,hocardiography. Circulation 1976;54:936-44. ].(I. Stamm RB, Gibson RS, Bishop HL, Carabello BA, Beller GA, Martin RP. Echocardiographic detection of infarct-localized asynerg:: and remote asynergy during acute myocardial infarction: correlation with extent of angiographic coronary disease. Circulation 1983;67:233-44. 1 I. IVilkins GT, Southern dF, Choong CY, et al. Correlation between etthocardiographic endocardial surface mapping of abnormal wall motion and pathologic infarct size in autopsied hearts. I:irculation 1988:77:978-87. AJ, Parisi AF. Postmyocar12. 1sascsol.n .JL, Earle MG, Kemper dial infi-rction pain and infarct extension in the coronary care unit: role of two-dimensional echocardiography. J Am Co11 Cardiol 1988;11:236-51. PI. I’icard MH, Wilkins GT, Ray PA, Weyman AE. Natural hist orp of et’t ventricular size and function after acute myocartiial infarction: assessment and prediction by echocardiographic endocardial surface mapping. Circulation 1990;82:48494. 14. I ;ibbon; f’,F. Hogan RD. Franklin TD, Nolting M, Weyman .4E. The natural history of regional dysfunction in a canine -neparation of chronic infarction. Circulation 1985;71:394m. EF, Hogan RD. et al. Relationship of 13 I:‘hoong CY. Gibblms mnctional recovery of scar contraction after myocardial inf