451

Transitional Blood Flow Zones between Ischemic andNonischemic Myocardium in the Awake Dog

Analysis Based on Distribution of the Intramural Vasculature

Robert H. Murdock, Jr., David M. Harlan, James J. Morris, III, William W. Pryor, Jr., andFrederick R. Cobb

From the Cardiology Division of the Department of Medicine at Duke University Medical Center and the Veterans AdministrationMedical Center, Durham, North Carolina

SUMMARY. The present study evaluates the transitional or border zone of intermediate blood flowreduction between nonischemic and ischemic regions after acute coronary artery occlusion inchronically instrumented dogs, using methods that minimize an admixture of ischemic and nonis-chemic myocardium in the tissue analyzed. The regions perfused by occluded and nonoccludedvessels were identified by tracing the extra and intramural distribution of the coronary vasculaturefrom postmortem angiograms. Regional blood flow was evaluated in serial 3-mm-wide epicardialand endocardial zones from outside and inside the interface between occluded and nonoccludedvessels. The zone of intermediate reduction in blood flow between nonischemic and ischemic regionsoccurred in the first 3-mm section immediately inside the region supplied by the occludedvasculature. Mean blood flow in this region was reduced to 58 ± 6% and 61 ± 5% (±SEM) ofnonischemic region blood flow at the lateral and medial epicardial margins, respectively, and 47 ±5% and 45 ± 6% at the lateral and medial endocardial margins, respectively. In the remainingischemic zone, significant differences in blood flow existed between epicardial and endocardiallayers; these differences were highly variable between animals. The data indicate that when theanalysis of regional blood flow following acute ischemia is based on the anatomic distribution of thecoronary vasculature, the transitional or border zone of intermediate reduction in blood flow islimited to a narrow zone immediately inside the occluded vasculature. Studies performed in acutelyanesthetized dogs in which the occluded region was perfused via a two-chamber blood reservoir thatallowed maintenance of perfusion and exclusion of microspheres from the circumflex region indicatethat intermediate reductions in blood flow at the border of the ischemic zone resulted from anadmixture of normal myocardium and, thus, do not represent a border zone of intermediate ischemia.(Circ Res 52: 451 -459, 1983)

Following proximal occlusion of a major coronaryartery in the canine model, the fraction of the leftventricle that becomes ischemic, the regional distri-bution of blood flow to the ischemic region, and theamount of myocardium that eventually becomes in-farcted are highly variable (Becker et al., 1973; Rivaset al., 1976; Marcus et al., 1975; Jugdutt et al., 1979a;Swain et al., 1980). Major determinants of the extentof ischemic injury and infarction include the dimen-sions of the ischemic or risk region (Jugdutt et al.,1979b; Reimer and Jennings, 1979; Factor et al., 1981)and blood flow to the ischemic myocardium (Rivas etal., 1976; Jugdutt et al., 1979b; Cobb et al., 1982).

Considerable controversy exists concerning thetransitional or border zone between ischemic andnonischemic zones. One view is that the ischemicregion consists of a central zone of greatest ischemiathat is surrounded by significant transitional zonescharacterized by intermediate reductions in bloodflow (Becker et al., 1973; Jugdutt et al., 1979a, 1979b;Vokonas et al., 1978), intermediate myocardial injury,and incomplete necrosis (Vokonas et al., 1978; Cox etal., 1968; Hearse et al., 1977; Kjekshus and Sobel,1970; Maroko et al., 1971, 1977; Braunwald et al.,1974; Sobel and Shell, 1973). According to this con-

cept, transitional zones of intermediate ischemia re-sult from preferential perfusion via the nonoccludedvasculature that borders the ischemic region. Otherinvestigators have concluded that the transition fromnonischemic to ischemic tissue is sharp and that asignificant transitional or border zone of intermediateischemia or myocardial injury does not exist (Reimerand Jennings, 1979; Factor et al., 1978; Hirzel et al.,1977; Factor et al., 1981; Yellon et al., 1981). The latterstudies suggest that an apparent rather than a realborder zone of intermediate ischemia and myocardialinjury or necrosis may result from techniques usedfor tissue sampling and analysis and/or the anatomicarrangement of the microcirculation. A transitionalzone of ischemia and/or tissue necrosis may result ifanalyses are performed on tissue samples that containan admixture of tissue even though the interfacebetween ischemic and nonischemic tissue is sharp.

The dimension of a transitional or border zone ofischemia and ischemic injury is pertinent to the con-cept of intervention therapy. It may be reasoned thata large zone of intermediate ischemia will increase thepotential for salvaging ischemic myocardium (Marokoet al., 1971, 1977; Braunwald et al., 1974; Sobel andShell, 1973; Jugdutt et al., 1979a, 1979b, 1980). Con-

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

452 Circulation Research/Vol. 52, No. 4, April 1983

versely, a sharp interface or abrupt gradient of bloodflow between nonoccluded and occluded regions sug-gests little or no functional anastomoses betweenintramural branches of the coronary vasculature atthe microcirculation level and, consequently, littlepotential for preferential salvage of the ischemic myo-cardium at the margins of the ischemic zone (Reimerand Jennings, 1979; Factor et al., 1978, 1981; Hirzel etal., 1977; Yellon et al., 1981).

The present study evaluates the transitional or bor-der zone of intermediate blood flow reduction be-tween nonischemic and ischemic regions followingacute coronary artery occlusion in chronically instru-mented dogs using methods that minimize an admix-ture of ischemic and nonischemic myocardium in thetissue analyzed. The regions perfused by occludedand nonoccluded vessels were identified by tracingthe extra and intramural distribution of the coronaryvasculature from postmortem angiograms. Regionalblood flow was evaluated in serial 3-mm-wide epi-cardial and endocardial zones from outside and insidethe interface between occluded and nonoccluded ves-sels to provide fine resolution of the distribution ofblood flow.

MethodsStudies were performed on 13 mongrel dogs weighing

15-25 kg. The dogs were anesthetized with thialmylal so-dium (30-40 mg/kg, iv), and underwent a left thoracotomy.Heparin-filled polyvinyl chloride catheters were insertedinto the left atrial cavity and the aortic root and weretunneled to a subcutaneous pouch at the base of the neck.

A pneumatic cuff occluder was placed around the leftcircumflex coronary artery either proximal (10 dogs) orimmediately distal (3 dogs) to the first marginal branch andwas tunneled to the pouch containing the catheters.

Studies were performed 7-10 days after surgery. Thepouch was anesthetized with 50 mg lidocaine and thecatheters and snare were exteriorized. Phasic and meanaortic and left atrial pressures and an electrocardiogram oflead 2 or 3 were recorded on a Sanborn eight-channelrecorder.

Myocardial blood flow was determined by injecting car-bonized microspheres 9 ± 1 (SD) fim in diameter and labeledwith y-emitting nuclides. The microspheres were obtainedas 1 mCi of each nuclide in 10 ml of 10% dextran and 0.05%polysorbate 80 (3M Co.), and were diluted in 10% dextransuch that 1.0 ml contained approximately 3.0 X 106 micro-spheres. Before each injection, the microspheres were mixedby alternate agitation for at least 15 minutes in an ultrasonicbath (3M Co., model DA 0950) and a Vortex agitator. Avolume of 1.0 ml of the microsphere suspension was in-jected into the left atrium over a period of 10-15 seconds,and the atrial catheter was flushed with 5 ml of isotonicsaline. Beginning simultaneously with each microsphereinjection and continuing for 90 seconds, a reference bloodsample was collected in counting vials from the aorticcatheter at a constant rate using a Harvard withdrawalpump. Serial injections of the microspheres resulted in nochange in heart rate during the interval of injection and nochange in aortic or left atrial pressure measured immedi-ately before and after collection of the reference bloodsamples.

After a control blood flow measurement, the left circum-

flex coronary artery was completely occluded by inflatingthe pneumatic cuff occluder. One minute later, myocardialblood flow was measured. All dogs were then anesthetizedwith thialmylal sodium and the hearts were fibrillated withconcentrated potassium chloride. The hearts were excisedimmediately, and the left main coronary artery was cannu-lated and perfused with barium sulfate gel (Fulton, 1965;Schaper, 1971). The occluder was released during the per-fusion. The hearts then were placed in 10% buffered for-malin for 3 days to allow fixation.

The epicardial distribution of the coronary vascularbranches was determined from gross examination and fromradiographs of the entire left ventricle. Each left ventriclewas then cut into eight rings from base to apex, as illustratedin Figure 1, and radiographs were made of each section.Figure 1A is a schematic of a radiograph of one ring at thelevel of the papillary muscle. From radiographs of eachtransverse section, the intramural distribution of each vesselwas traced through the myocardium in a fashion compara-ble to the branches of a tree. Transmural lines were drawnon the radiographs equidistant between medial and lateralterminal branches of the occluded and nonoccluded vessels.The angulation of the cuts varied to best conform to thedistribution of the intramural vessels. These lines were usedas a guide for transmural cuts that separated the occludedfrom the nonoccluded region, as illustrated in Figure IB.The weights of occluded and nonoccluded regions weredetermined. The rings were then cut into serial 3-mm-widesections as illustrated in Figure IB. Each 3-mm-wide sectionwas cut into epicardial and endocardial halves. Sectionsfrom each ring were combined so that myocardial bloodflow could be determined in serial 3-mm zones, 1-3 g insize, inside and outside the occluded region. The base ring1 and apex rings 7 and 8 were not included in the multiplesections and combined analyses. Ring 1 contains fibroustissue from the valve rings. The apex rings are suppliedprimarily by the left anterior descending coronary artery.The myocardium supplied by the circumflex artery tapersat the apex to supply a relatively small and variable amountof rings 7 and 8; separation of occluded and nonoccludedvasculature was more difficult in the apex rings. The present

POSTERIORMed

Lot

Med

LEFT VENTRICLE

ANTERIOR

FIGURE 1. Schematic of the /eft ventricle with the right ventricleremoved. Part A is a schematic of a radiograph of the vasculatureinjected with barium sulfate gel from one of the eight transverserings at the level of the papillary muscle. Part B is a schematic ofthe technique for sectioning each transverse ring. C = nonischemicregion, IC = ischemic zone center, Sep = septum, O r = circumflex,LAD = left anterior descending.

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

Murdock et aJ./Blood Flow to Ischemic Regions

study assesses blood flow gradients at the lateral and medialmargins of the occluded vasculature which represents themajority of the interface between the ischemic and nonis-chemic regions.

The number of zones from the occluded region rangedfrom 2 to 5 at each border, depending on the size of theregion. Two 3-mm zones were cut from the nonoccludedregion at each border. Samples were taken from the centerof the occluded and nonoccluded regions. The radioactivityin each reference blood and tissue sample was measured ina Packard Gamma Scintillation Spectrometer, using windowsettings selected to correspond with the peak energies ofeach radioactive nuclide. Blood flow per tissue sample wasdetermined using the counts per ml per minute for theblood samples and the counts per minute for the tissuesample. Blood flow was calculated using the followingformula:

Qm = Qr • Cm/Cr

where Qm = myocardial flow (ml/min), Qr = referenceblood flow (ml/min), Cm = counts/min in myocardium,and Cr = counts/min in reference blood flow. Myocardialblood flow (ml/min) was divided by the weight of the zoneand was expressed as ml/min per g.

Blood flow measurements in serial zones outside andinside the occluded regions were analyzed in two ways.Blood flow in each 3-mm epicardial and endocardial zonewas expressed as a fraction of anterior region flow in eachdog; the mean distribution of blood flow for the group thenwas analyzed. To allow comparison of the transitional zoneof intermediate blood flow reduction between individualanimals, the maximum blood flow gradient between non-ischemic and ischemic tissue was determined in each animalby subtracting the lowest blood flow value in the ischemiczone from blood flow in the anterior nonischemic zone.Myocardial blood flow in each 3-mm zone outside andinside the ischemic region was then expressed as a percentof the maximum blood flow gradient between nonischemicand maximally ischemic regions. All data points from the13 dogs are illustrated in Figures 3 and 4.

Additional studies were performed on four mongrel dogsto determine to what extent blood flow measurements tothe occluded circumflex region resulted from an admixtureof normal myocardium. This analysis was accomplished byperfusing the circumflex coronary artery using a modifica-tion of the technique of Hirzel et al. (1977) that allowedexclusion of microspheres injected into the left atrium fromthe vascular distribution of the circumflex coronary arterywhile perfusion was maintained. Each dog was anesthetizedwith thiamylal sodium (30 to 40 mg/kg, iv) and ventillatedwith 100% oxygen. A left thoracotomy was performed. Thepericardium was incised and the heart suspended in apericardial cradle. Catheters were filled with heparinizedsaline and placed in the aortic root and left atrial appendage.

The proximal left circumflex coronary artery was isolatedproximal to the first marginal branch. Myocardial bloodflow was measured during total occlusion as previouslydescribed; the occlusion then was released. Next, a catheterwas placed in the right femoral artery and connected via aHarvard withdrawal pump to an airtight reservoir bottlethat was partially filled with arterial blood. The proximalcircumflex artery then was ligated and cannulated at thepreviously occluded site and perfused from the reservoirbottle. Blood flow was interrupted to the circumflex regionfor approximately 30 seconds. Pressure in the circumflexcatheter was adjusted to equal mean arterial pressure byregulating pressure in the blood reservoir with a pressure

453

gauge and inflation bulb and adjusting the flow rate of thewithdrawal pump. A second inlet to the reservoir wasconnected to a balloon within the reservoir so that femoralartery blood could be delivered selectively into either thereservoir or the balloon without changing pressure in thecircumflex catheter. After a 15-minute steady state period,femoral arterial inflow was switched to the balloon andmicrospheres were injected into the left atrium as previouslydescribed for analysis of regional blood flow. This arrange-ment allowed perfusion of the circumflex artery at aorticperfusion pressures from blood in the reservoir and collec-tion of microspheres in the balloon. Immediately after theblood flow measurement, the heart was fibrillated withpotassium chloride, excised, perfused with barium sulfategel, and placed in 10% buffered formalin for fixation, andsubsequently sectioned according to the risk region aspreviously described. The distribution of microspheres dur-ing the latter procedure represents the distribution of thevasculature exclusive of the circumflex artery; microspherespresent in samples taken from the circumflex region rep-resent the extent of admixture or overlap of nonischemicmyocardium in the region.

Results

Figure 2 is a graph of blood flow in ml/min per gin serial 3-mm sections immediately inside and out-side the region supplied by the occluded circumflexvessel in a representative study. The region of inter-mediate flow reduction occurred in a 3-mm-widesection immediately inside the occluded region atboth the medial and lateral margins of the ischemiczone. In the nonoccluded region, endocardial flowwas greater than epicardial flow; in the occluded zone,epicardial flow was greater than endocardial flow. Inthe total group, blood flow was reduced to 61 ± 5%and 58 ± 6%, of nonischemic region flow at the medialand lateral epicardial margins, respectively, and 45± 6% and 47 ± 5% at the medial and lateral endocar-dial margins, respectively. Blood flow to the second3-mm zone inside the occluded region and the centralischemic region were not significantly different.

—

(%

Nor

ma

O—'a

8J

CO

1.2

1.0

.8

.6

.4

.2

MEDIAL

t\11I

\

\I \

\ \

\ \

tptEndo

" * \^ — . — . — -—-r

LATERAL

j

1

J2 1 1 2 3 4 5 I C I C 5 4 3 2 1 1 2

3 mm ZONES

FIGURE 2. Epicardial and endocardial regional blood flow, ml/minper g in serial 3-mm zones outside and inside the occluded regionsand in the anterior control region. The vertical lines indicate thesection between nonoccluded and occluded regions. IC = ischemiczone center.

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

454 Circulation Research/VoJ. 52, No. 4, April 1983

I 1 2 3 4 5 IC 5SECTION

3 mm Zones

4 3 2 1 1 2

FIGURE 3. Individual and mean epicardial blood flow measurementsfrom 13 dogs expressed as a percent of the gradient betweennonischemic and maximally ischemic regions. Vertical lines indi-cate the section between nonoccluded and occluded regions. IC =ischemic zone center.

Transmural gradients in the nonischemic and isch-emic regions also were similar to those illustrated inFigure 2.

Figures 3 and 4 are plots of all data points from the13 dogs studied. Each data point is plotted as a percentof the gradient between the anterior nonischemicregion and the region of maximum blood flow reduc-tion in the occluded region. The region of maximumblood flow reduction was the center of the ischemicregion in most but not all cases. The blood flowreductions in the first 3-mm-wide zone at the medial

z

GR

AD

I

z

100

80

60

40

20

0

-20

-

MED

•I

1

1 /

I-

J \ '

ENDOCARDIUM

I AT

, , . 1 , , , ,

XI

2 1 1 2 3 4 ICSECTION

5 4 3 2 1 1 2

3 mm Zones

FIGURE 4. Individual and mean endocardia! blood flow measure-ments from 13 dogs expressed as a percent of the gradient betweennonischemic and maximally ischemic regions. The vertical linesindicate the section between nonocc/uded and occluded regions.IC = ischemic zone center.

and lateral epicardial margins of the ischemic regionsaveraged 46 ± 19% and 45 ± 24%, respectively, of themaximum gradient between nonischemic and maxi-mally ischemic regions. There was considerable vari-ation in the blood flow reductions in these regions;the reduction ranged from 11 to 100%. The bloodflow reductions in the second 3-mm-wide zone at themedial and lateral epicardial margins of the ischemicregion averaged 82 ± 13% and 82 ± 16%, respectively,of the gradient between nonischemic and maximallyischemic zones. The variation in this region was less;the blood flow reduction was less than 70% of thegradient between nonischemic and maximally isch-emic zones in only 4 of 26 instances.

Blood flow reduction in the first 3-mm-wide zoneat the medial and lateral endocardial margins of theischemic region averaged 58 ± 23% and 54 ± 17%,respectively, of the gradient between nonischemic andmaximally ischemic zones. In 11 of 26 instances, thereduction exceeded 60% of the maximum blood flowgradient in the first 3-mm-wide zone. The blood flowreduction in the second 3-mm zone at the medial andlateral endocardial margins of the ischemic regionaveraged 86 ± 16% and 91 ± 6%, respectively, of thegradient between nonischemic and the maximallyischemic zones. The blood flow reduction in thesecond 3-mm zone was less than 80% of the gradientbetween the nonischemic and maximally ischemiczones in only 2 of 26 instances. Blood flow to epicar-dial and endocardial zones immediately outside theischemic region was not significantly different fromanterior region blood flow. Thus, following acutecoronary artery occlusion, the transitional zone ofintermediate reduction in blood flow occurred almostexclusively in the first 3-mm zone immediately insidethe ischemic region in both the epicardial and endo-cardial layers.

Table 1 contains blood flow data in ml/min per gin epicardia! and endocardia! layers from the nonis-chemic anterior wall and from the center of the isch-emic zone. Blood flow to the central ischemic zone isalso listed as a percentage of anterior nonischemicblood flow. Blood flow to the epicardial layers of thecentral ischemic zone was highly variable betweenanimals; mean flow was 0.15 ml/min per g, range0.07-0.32 ml/min per g, representing an average re-duction to 17%, range 7-37% of nonischemic anteriorflow. Blood flow to the endocardial layers of thecentral ischemic zone was reduced to a greater extent;mean flow was 0.06 ml/min per g, range 0-0.16 ml/min per g representing an average reduction to 6%,range 0-17% of the nonischemic blood flow. Thesedata demonstrate significant transmural differences inblood flow between the epicardial and endocardiallayers in individual animals and considerable varia-bility between animals in the transmural distributionof blood flow. Thus, although gradients of blood flowat the margins of the ischemic zone occurred over anarrow zone of 3 mm or less, substantial differencesin the transmural distribution of blood flow werepresent between the epicardial and endocardial layersof the remaining ischemic zone.

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

Murdock et al./Blood Flow to Ischemic Regions 455

TABLE l

Blood Flow to Nonischemic Zones and the Center of the Ischemic Zone after Coronary Artery Occlusion

Dogno.

1

2

3

4

56

7

8

9

10

11

12

13

Mean ± SD

Nonischemic(ml/min per g)

0.890.781.020.911.220.480.740.870.600.891.081.151.70

0.95 ± 0.31

Epicardium

Central ischemic

(ml/min per g)

0.090.130.200.180.110.120.070.320.170.110.160.080.16

0.15 ± 0.07

zone(% non-ischemic

bloodflow)

10

17

20

20

9

25

9

37

28

12

15

7

9

17 ± 9

Nonischemic(ml/min per g)

0.990.861.141.011.320.640.920.950.760.981.311.311.63

1.06 ± 0.27

Endocardium

Central ischemic

(ml/min per g)

0.040.130.090.010.020.050.010.160.11

0

0.020.030.06

0.06 ± 0.05

zone(% non-ischemic

bloodflow)

4

15

81

2

81

1714

0

2

2

4

6 ± 6

The total ischemic or risk region in the presentstudy averaged 38.8% of the left ventricular weight,range 25.5-54.6%. The 3-mm-wide zones of interme-diate reduction in blood flow at the border of theischemic zone averaged 12.9%, range 7-19% of rings2-6.

Figures 5 and 6 illustrate the distribution of bloodflow in four dogs expressed as a percent of normalzone flow, between occluded and nonoccluded re-gions during first acute ischemia and then, duringperfusion of the previously occluded vessel, via a two-chamber reservoir bottle that allowed exclusion ofmicrospheres from the perfusate. Blood flow valuesin the occluded region during no ischemia are a resultof an overlap or admixture of normal myocardium inthe sample. The difference between blood flow meas-urements in the occluded region during ischemia andno ischemia is a measure of collateral blood flow atthis interval. Table 2 lists blood flow expressed as apercent of normal zone flow in the 3-mm zone im-mediately inside the occluded region during the twoblood flow measurements; the percent of flow duringischemia that may be explained by overlap or anadmixture of nonischemic myocardium is tabulated.During no ischemia, blood flow averaged 37% and30% of normal zone flow in the medial and lateralepicardial regions, respectively, and 33% and 36% inthe medial and lateral endocardial regions, respec-tively, in the first 3-mm region inside the occludedregion. The overlap of nonischemic myocardium inthe occluded region was limited primarily to the first3-mm section; overlap blood flow averaged 6 and 3%in the epi- and endocardial regions of the second 3-mm section; no overlap blood flow occurred in theremaining sections from the occluded region. Thesedata indicate that the intermediate zone of reducedblood flow in these studies resulted primarily froman admixture of nonischemic myocardium.

Discussion

The present study analyzes the distribution ofblood flow between nonoccluded and occluded myo-cardial regions. Postmortem angiograms were used asa guide to identifying the epicardial and transmuraldistribution of the nonoccluded and occluded vascu-lature. Serial 3-mm sections were taken from insideand outside the distribution of the occluded vascula-ture to provide fine resolution of the distribution ofblood flow. The results of these analyses demonstratethat intermediate reductions in blood flow betweennonischemic and central ischemic regions occurred ina very narrow rim at the outer margin of the occludedregion. An additional group of studies was carried outto determine to what extent the zone of intermediateblood flow reduction represented an admixture ofnonischemic and ischemic myocardium. These stud-ies utilized a two-chamber reservoir that allowed ex-clusion of microspheres from the distribution of thecircumflex vasculature while maintaining perfusionas described by Hirzel et al. (1977). The results of thelatter study indicate that most of the intermediatereduction in blood flow resulted from an admixtureof nonischemic myocardium in the sample. Thesestudies support the view that the blood flow gradientbetween nonischemic and acutely ischemic myocar-dium is abrupt, with essentially no significant zone ofintermediate ischemia. Overlap of nonischemic myo-cardium in the circumflex or occluded region waslimited primarily to the first 3-mm section; overlap inthe second 3-mm section averaged 6% and 3% in theepi- and endocardium; there was no overlap in theremaining sections of the occluded zone. Thus, thedimensions of the zone of admixture or interdigitationof normal and ischemic myocardium was relativelysmall in the sections analyzed. These data indicatethat the postmortem angiographic technique provides

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

456 Circulation Research/Vol. 52, No. 4, ApriJ 1983

1.20

100

l 8°°t .60O

aj 40

I .20CD

O

MedEPICARDIUM

o Ischemia• No Ischemia

Lot

"—»• - - „ — * /

MedENDOCARDIUM

£E

120

o 100

"8 80_o

CO 60

.40

.20

=t=r- t t r4 3 2 1 1 3 42 3 4 5 I C I C 5 4 3 2 1 1 2 3 4 4 3 2 1 1 2 3 4 5 I C I C 5 4 3 2

3 mm ZONES 3 mm ZONESFIGURE 5. Blood flow in 3-mm epicardiai and endocardial zones from inside and outside circumflex regions during circumflex ischemia (opencircles) and no ischemia (closed circles) in dogs 1 and 2. Blood flow is expressed as a fraction of anterior region flow. Microspheres wereexcluded from the circumflex vasculature during no ischemia by perfusing the heart via a two-chamber reservoir bottle that allowed exclusionof microspheres from the perfusate. Vertical lines indicate the sections between nonoccluded and occluded regions. IC = ischemic zonecenter.

a guide to identify the distribution of the nonoccludedand occluded vasculature and thus to delineation ofthe ischemic zone. The technique minimizes the de-gree of overlap of ischemic and nonischemic tissue toa narrow border at the margins of the ischemic zone.Because of the interdigitation of nonischemic andischemic myocardium in this narrow zone, the tech-nique does not completely separate myocardium per-fused by the two circulations in this narrow segment.Admixture of normal and ischemic myocardium mayresult from interdigitation in a given plane as well asfrom overlap at different depths through the section.The x-ray angiogram should thus be made frommyocardial sections that are relatively thin. In thepresent study, the left ventricle was sectioned intoeight rings; this resulted in small epi-endo samplesthat were then combined from different rings to pro-vide a sample size greater than 1 g.

The view that a zone of intermediate ischemia andinjury constitutes a significant fraction of the ischemicregion is supported by a variety of studies [histochem-ical (Vokonas et al., 1978; Cox et al., 1968), biochem-ical (Cox et al., 1968; Hearse et al., 1977; Kjekshusand Sobel, 1970), myocardial blood flow (Becker etal., 1973; Vokonas et al., 1978; Jugdutt et al., 1979a,1979b, 1980), and electrocardiographic (Maroko et al.,

1971, 1977; Braunwald et al., 1974)]. It is apparentthat intermediate reductions in blood flow or injurymay result if a tissue sample that is analyzed containsan admixture of nonischemic and ischemic myocar-dium and that the dimension of the zone will dependon the technique for sampling. Most of the aforemen-tioned studies did not section the ventricle in a fashionthat would allow fine resolution of the dimension ofthe interface between occluded and nonoccluded re-gions and did not consider the extent of admixture ofnonischemic and ischemic tissue in samples analyzed.

Certain studies that have analyzed the distributionof infarction as a function of the occluded vasculatureidentified by different dye injections (Reimer et al.,1979) or postmortem angiograms (Koyanagi et al.,1982) have observed that infarction at the endocardialsurface extended to 1-2 mm or less of the lateralmargins of the occluded vasculature at 4 days to 24hours after coronary occlusion; although these studiessupport sharp separation between the occluded andnonoccluded vasculature, especially in the endocar-dium, the distribution of blood flow was not analyzedat the interface between occluded and nonoccludedregions following acute occlusion. Studies by Kirkand associates (Factor et al., 1978, 1981; Hirzel et al.,1977) have supported the concept of a sharp or abrupt,

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

Murdock et al/ Blood Flow to Ischemic Regions 457

cO

<x>

O)c

120

100

80

60

40

20

- 3

CDCO

§- 120

CO

100_oLL 80

1 .60CO

.40

.20

Med EPICAROIUM

o Ischemia• No Ischemia

Med ENDOCARDIUM Lot

4 3 2 1 1 3 2 1 1 2 3 42 3 4 5 I C I C 5 4 3 2 1 1 2 3 4 4 3 2 1 1 2 3 4 5 I C I C 5 4

3 mm Zones 3 mm ZonesFIGURE 6. Blood flow in 3-mm epicardial and endocardial zones from inside and outside circumflex regions during circumflex ischemia (opencircles) and no ischemia (closed circles) in dogs 3 and 4. Blood flow is expressed as a fraction of anterior region flow. Microspheres wereexcluded from the circumflex vasculature during no ischemia. Vertical lines indicate the sections between nonoccluded and occluded regions.IV = ischemic zone center.

rather than gradual, transition between nonischemicand ischemic tissue 24 hours after coronary occlusion.Hirzel et al. (1977) observed no lateral border zone ofintermediate creatine kinase depletion 24 hours aftercoronary occlusion in dogs when tissue supplied bythe nonoccluded vasculature was excluded from theanalyses. In subsequent studies from the same labo-

ratory, Factor et al. (1978) reconstructed the three-dimensional geometry of the border area from serialsections 24 hours after coronary occlusion in the dog.They observed an irregular interface between normaland infarcted myocardium; apparent islands of nor-mal tissue in the ischemic zone resulted from inter-digitating peninsulas of normal tissue. Okun et al.

TABLE 2

Blood Flow to First 3-mm Zone Inside Occluded Region during Ischemia and No Ischemia

Dogno.

1

23

4Mean ± SD

1

23

4Mean ± SD

Ischemia

31

36

81

76

56 ± 26

43

30

60

53

46 ± 13

Epicardium

No ischemia

11

1657

65

37 ± 28

32

18

4527

30 + 11

% Overlap

Medial zone

35

44

70

86

59 ± 23

Lateral Zone

74

60

75

51

65 ± 12

Ischemia

5116

62

• 4844 ± 20

56

21

28

57

40 ± 19

Endocardium

No ischemia

34

9

41

47

33 ± 17

53

16

23

5336 + 20

% Overlap

67

56

66

98

72+ 18

95

76

82

93

86 ± 9

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

458

(1979) and Factor et al. (1982) injected different col-ored silicone rubber into large coronary vessels ofpostmortem dog and human hearts, respectively, andobserved that the capillaries derived from an individ-ual vessel formed discrete loops without interconnec-tions or anastomoses with capillaries derived fromanother large vessel. Using this technique, Factor etal. (1981) observed a close correlation between thedistribution of the intramural vasculature and histo-Iogical infarction at the margin of infarction 24 hoursafter occlusion in dogs. Although the preceding stud-ies indicate discrete anatomic separation of the oc-cluded and nonoccluded circulation at the microcir-culation level, the blood flow distribution in thisregion was not measured; measurement of the extentof infarction in relationship to vessel distribution wasmade 24 hours after occlusion and, thus, after theinfarction process had been completed. Consequently,the possibilities of border zones of intermediate is-chemia or myocardial injury at earlier intervals afterocclusion were not excluded. These studies did notdescribe the dimensions of the interdigitating zone ofnormal and ischemic myocardium at the margin ofthe ischemic region and do not indicate that an ad-mixture of nonischemic and ischemic tissue could notbe minimized if the regions supplied by the occludedvessel could be accurately identified.

Schaper and associates (Schaper and Pasky, 1976;Schaper, 1979) demonstrated an intermediate bloodflow zone similar to the present study in certainstudies. These investigators analyzed the regional dis-tribution of blood flow between nonischemic andischemic regions in serial circumferential samplesfrom individual left ventricular rings; the sample sizewas approximately 500 mg and the dimension of eachsample was approximately 4 mm. The rings weresectioned without knowledge of the anatomic bordersof perfusion. Intermediate reductions in blood flowbetween normal and ischemic regions occurred in 1to 3 of the 4-mm sections in illustrations reported.Schaper (1979) concludes that the ischemic region issurrounded by a border zone of intermediate bloodflow reduction that measures approximately 10 mm,that the border zone is a mixture of well and poorlyperfused muscle, and that the border of perfusion isvery steep.

Analysis of transmural blood flow within the isch-emic zone demonstrated significant differences inblood flow between the endocardial and epicardiallayers in individual animals and considerable varia-tion in a given layer among animals. Epicardial flowto the center of the ischemic region varied from 7 to37%, average 17% of blood flow to anterior nonis-chemic regions. Endocardial blood flow to the centerof the ischemic region varied from 0 to 17%, average6% of nonischemic blood flow. Since the studiesperformed in acutely anesthetized dogs indicate thatadmixture of nonischemic myocardium did not ex-tend past the second 3-mm zone inside the occludedregion, these variations in blood flow result from

Circulation Research/VoJ. 52, No. 4, April 1983

differences in collateral blood flow and thus representdifferences in the degree of ischemia.

As noted in the previous discussion, numerousmethods have been utilized to characterize the tran-sitional or border zone of blood flow and/or ischemicinjury after coronary occlusion in acute open-cheststudies. Potential advantages of the angiographic tech-nique (Fulton, 1965; Schaper, 1971) are that it is apostmortem method that does not require instrumen-tation of the intact vasculature and thus is applicablefor use in studies that are performed in chronicallyinstrumented as well as acute preparations; it shouldnot be influenced by the development of epicardialcollateral vessels following coronary artery occlusionand infarction, and it provides a logical anatomic basisfor sectioning the ventricle for analysis of blood flow.As discussed previously, the technique minimizesoverlap of normal and ischemic tissue to a narrowzone at the outer rim of the occluded region, but doesnot allow complete separation of nonischemic andischemic tissue. The injection of different coloreddyes to stain myocardial regions supplied by occludedand nonoccluded vessels is another technique that hasbeen used to identify the ischemic zone (Reimer andJennings, 1979). There are several potential sources oferror that may be encountered in using dye injectiontechnique. The dye technique requires simultaneousperfusion of the occluded and nonoccluded vessels atequal pressures to minimize passage of dye via thecollateral vessels. Differential pressures may resultfrom simultaneous perfusion of different sized vas-cular beds, kinked vessels at the catheter sites, differ-ent sized perfusion catheters, and/or underperfusionof one or more of the major branches of the left maincoronary artery that often arise from immediate tri-furcations into septal, anterior descending, and cir-cumflex branches. These potential problems aregreater in the presence of enhanced collateral vesseldevelopment. Permanent occlusion causes rapid de-velopment of epicardial collateral vessels (Gregg,1974) so that 3 days after coronary occlusion, dye, aswell as radiopaque barium gel, injected into nonoc-cluded vessels may easily pass through the epicardialcollaterals and stain or fill the ischemic as well asnonischemic myocardium and vasculature. The bar-ium gel technique used in the present study dependson filling of the entire vasculature; simultaneous per-fusion of occluded and nonoccluded vessels is notrequired, and the presence of collaterals facilitatesfilling, should a branch be underperfused.

In summary, an analysis of blood flow betweenischemic and nonischemic regions, based on the dis-tribution of the intramural coronary vasculature frompostmortem angiograms, demonstrated that interme-diate reductions in blood flow occurred in a verynarrow rim of myocardium 3 mm or less in dimensionimmediately inside the ischemic zone; significant dif-ferences in blood flow in the epicardium and theendocardium were present in the remaining ischemiczone. Studies performed in anesthetized dogs in

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

Murdock et a]./ Blood Flow to Ischemic Regions 459

which microspheres were excluded from the circum-flex regions during no ischemia indicate that theintermediate reductions in blood flow at the borderof the ischemic region result primarily from an ad-mixture of normal myocardium and thus do not rep-resent a border zone of intermediate ischemia.

We wish to thank Marjie Grubb, Eric Fields, and Kirby Cooperfor their technical assistance, The Durham VA Medical MediaProduction Service for photographical assistance, and Cathie Col-lins for secretarial assistance.

This work is supported by Research Grant HL18537 from theNational Heart, Lung, and Blood Institute and the Medical ResearchService of the Veterans Administration Medical Center.

Address for reprints: Frederick R. Cobb, M.D., Division ofCardiology (111A), Veterans Administration Medical Center, 508Fulton Street, Durham, North Carolina 27705.

Received December 14, 1981; accepted for publication February3, 1983.

ReferencesBecker LC, Ferreira R, Thomas M (1973) Mapping of left ventricular

blood flow with radioactive microspheres in experimental coro-nary artery occlusion. Cardiovasc Res 7: 391-400

Braunwald E, Maroko PR, Libby P (1974) Reduction of infarct sizefollowing coronary occlusion. Circ Res 35 (suppl 3): 192-201

Cobb FR, Murdock RH, Morris KG (1982) Relationship betweenregional myocardial blood flow and the extent of MI: A modelfor assessing intervention therapy in the conscious animal. InMyocardial Infarcts, Measurements and Intervention, edited byGL Wagner. The Hague/Boston/London, Martinus Nijhoff, pp277-294

Cox JL, McLaughlin VW, Flowers NC, Horan LG (1968) Theischemic zone surrounding acute myocardial infarction in therat. Am Heart J 76: 650-659

Factor SM, Sonnenblick EH, Kirk ES (1978) The histologic borderzone of acute myocardial infarction—islands or peninsulas? AmJ Pathol 92: 111-120

Factor SM, Okun EM, Kirk ES (1981) The histological lateralborder of acute canine myocardial infarction. A function ofmicrocirculation. Circ Res 48: 640-649

Factor SM, Okun EM, Minase T, Kirk ES (1982) The microcircu-lation of the human heart: End-capillary loops with discreteperfusion fields. Circulation 66: 1241-1248

Fulton WFM (1965) The coronary arteries arteriography, micro-anatomy, and pathogenesis of obliterative coronary artery dis-ease. Springfield, 111., Charles C. Thomas, pp 3-13

Gregg DE (1974) The natural history of coronary collateral devel-opment. Circ Res 35: 335-344

Hearse D], Opie LH, Katzeff IE, Lubbe WF, VanDer Werff TJ,Peisach M, Boulle G (1977) Characterization of the "border zone"in acute regional ischemia in the dog. Am J Cardiol 40: 716-726

Hirzell HO, Sonnenblick EH, Kirk ES (1977) Absence of a lateralborder zone of intermediate creatine phosphokinase depletionsurrounding a central infarct 24 hours after acute coronaryocclusion in the dog. Circ Res 41: 673-683

Jugdutt BI, Hutchins GM, Bulkley BH, Pitt B, Becker LC (1979a)Effect of indomethacin on collateral blood flow and infarct sizein the conscious dog. Circulation 59: 734-743

Jugdutt BI, Hutchins GM, Bulkley BH, Becker LC (1979b) Myocar-dial infarction in the conscious dog: Three dimensional mappingof infarct, collateral flow and region at risk. Circulation 60:1141-1150

Jugdutt BI, Hutchins GM, Bulkley BH, Becker LC (1980) Salvage ofischemic myocardium by ibuprofen during infarction in theconscious dog. Am J Cardiol 46: 74-80

Kjekshus JK, Sobel BE (1970) Depressed myocardial creatine phos-phokinase activity following experimental myocardial infarctionin rabbit. Circ Res 27: 403-414

Koyanagi S, Eastman CI, Harrison DG, Marcus M (1982) Increasedsize of myocardial infarction in dogs with chronic hypertensiveand left ventricular hypertrophy. Circ Res 50: 55-62

Marcus ML, Kerber RE, Ehrhardt J, Abboud FM (1975) Threedimensional geometry of acutely ischemic myocardium. Circu-lation 52: 254-263

Maroko PR, Kjekshus JK, Sobel BE, Watanabe T, Lovell JW, RossJ, Braunwald E (1971) Factors influencing infarct size followingexperimental coronary artery occlusions. Circulation 43: 67-82

Maroko PR, Hillis LD, Muller JE, Tavazzi L, Heyndrickx GR, RayM, Chiariello M, Distante A, Askenazi J, Salerno J, Carpentier J,Reshetnaya NI, Radvany P, Libby P, Raabe DS, Chazov El, BobbaP, Braunwald E (1977) Favourable effects of hyaluronidase onelectrocardiographic evidence of necrosis in patients with acutemyocardial infarction. N EngI J Med 296: 898-903

Okun EM, Factor SM, Kirk ES (1979) End-capillary loops in theheart: An explanation for discrete myocardial infarctions withoutborder zones. Science 206: 565-572

Reimer KA, Jennings RB (1979) The "wavefront phenomenon" ofmyocardial ischemic cell death. II. Transmural progression ofnecrosis within the framework of ischemic bed size (myocardiumat risk) and collateral flow. Lab Invest 40: 633-644

Rivas F, Cobb FR, Bache RJ, Greenfield JC Jr (1976) Relationshipbetween blood flow to ischemic regions and extent of myocardialinfarction: Serial measurement of blood flow to ischemic regions.Circ Res 38: 439-447

Schaper W (1971) The Collateral Circulation of the Heart. Amster-dam, North-Holland Co., pp 5-18

Schaper W (1979) The pathophysiology of myocardial perfusion.Amsterdam-New York-Oxford, Elsevier/North HollandBiomedical Press

Schaper W, Pasky S (1976) Influence of collateral blood flow onthe ischemic tolerance of the heart following acute and subacutecoronary occlusion. Circulation 53 (suppl I): 57-62

Sobel BE, Shell WE (1973) Jeopardized, blighted and necroticmyocardium. Circulation 47: 215-216

Swain JL, Cobb FR, McHale PA, Roe CR (1980) Nonlinear relation-ship between creatine kinase estimates and histologic extent ofinfarction in conscious dogs: Effects of regional myocardial bloodflow. Circulation 62: 1239-1247

Vokonas PS, Malsky PM, Paul SJ, Robbins SL, Hood WB (1978)Radioautographic studies in experimental myocardial infarction:Profiles of ischemic blood flow and quantification of infarct sizein relation to magnitude of ischemic zone. Am J Cardiol 42: 67-75

Yellon DM, Hearse DJ, Crome R, Grannell J, Wyse RKH (1981)Characterization of the lateral interface between normal andischemic tissue in the canine heart during evolving myocardialinfarction. Am J Cardiol 47: 1233-1239

INDEX TERMS: Ischemic border zone • Transitional blood flowzones • Acute myocardial ischemia • Admixture of myocardialtissue

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from

R H Murdock, Jr, D M Harlan, J J Morris, 3rd, W W Pryor, Jr and F R Cobbdog. Analysis based on distribution of the intramural vasculature.

Transitional blood flow zones between ischemic and nonischemic myocardium in the awake

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 1983 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/01.RES.52.4.4511983;52:451-459Circ Res. 

http://circres.ahajournals.org/content/52/4/451World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Research Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on May 17, 2018

http://circres.ahajournals.org/D

ownloaded from