Myocardial Blood Flow in Man: Effects of Coronary

Collateral Circulation and Coronary Artery Bypass Surgery

SIDNEY C. SMITH, JR., RICHARDGoRLIN, MICHAELV. HERMAN,WARRENJ. TAYLOR, and JoHN J. COLLINS, JR.

From the Cardiovascular Division, Department of Medicine, and the Division ofThoracic and Cardiac Surgery, Department of Surgery, Peter Bent BrighamHospital and Harvard Medical School, Boston, Massachusetts 02115

A B S T R A C T The effects of coronary artery bypassgraft (CAB) and coronary collaterals (CC) on myo-cardial blood flow (MIBF) were studied in 24 patientsundergoing 29 CAB's. MIBF after CAB was comparedto preexisting MBFby intraoperatively injecting 'xenonvia distal CAB with proximal CAB first occluded thenopen. Pressure gradients across bypassed obstructionswere measured. The results were correlated w'ith preop-erative coronary arteriograms to determine the effectsof CC on MBFand postobstructive perfusion pressures.Mean MIBF was increased by CABfrom 32+6 (SE) ml/min per 100 g (CAB occluded) to 118±13 ml/min per100 g (CAB open). The 'Xe clearance curves withCAB open were resolved into slow (19±2 ml/min per100 g) and rapid (133±12 ml/min per 100 g) phases,suggesting that MBFremained heterogeneous after CAB.Vessels with less than 80% stenosis by angiography hadpressure gradients less than 20 mmHg across obstruc-tions, high postobstructive perfusion pressures (75+7mmHg), and normal MBF (87±6 ml/min per 100 g)even with CAB occluded. Vessels with greater than 80%stenosis or total occlusion by angiography had significantpressure gradients with marked reduction of postobstruc-tive MBF. No significant difference in postobstructiveMBFwas found when vessels with CC (21±4 ml/minper 100 g) were compared to those without CC (17+4ml/min per 100 g) (P > 0.4).

These studies demonstrate that (a) mean MBF in-creased 268% after CAB, (b) heterogeneous MBFper-sisted after CAB, (c) CC were not associated with sig-nificant increases in MUBF, and (d) vessels with lessthan 80% stenosis had less than 20 mmHg gradient withminimal effect on resting MBF.

Presented in part at the Annual Meeting of the AmericanHeart Association, 11 November 1971, Anaheim, Calif.

Received for puiblication 11 February 1972 anzd in revisedformi 30 Mlay, 1972.

INTRODUCTION

Coronary artery bypass surgery provides a unique op-portunity to study in man the hemodynamics and flowpatterns of the diseased coronary artery. Previous experi-mental investigations (1, 2) have demonstrated in caninehearts the alterations in coronary hemodynamics whichfollow acute and chronic coronary arterial constriction.These studies indicate the collateral circulation is ca-pable of restoring flow to normal distal to coronary ar-terial obstructions. Recent clinical observations in man(3-7) are in conflict with the experimental studies, how-ever, and suggest that coronary collateral flow providesan inadequate response to myocardial ischemia. Implicitin the concept of coronary bypass surgery is the capa-bility of restoring flow to a myocardium rendered ischemicby coronary arterial obstruction. To justify a surgicalapproach, the improvement in myocardial flow resultingfrom bypass graft should significantly exceed that pro-vided by collateral circulation.

The present study was designed (a) to measure therelative improvement in myocardial blood flow (MBF)1after coronary bypass graft, and (b) to correlate thepreexisting MBFwith the presence of collaterals and de-gree of coronary artery disease.

METHODSPatientt populationz. 24 patients (21 male, 3 female) rang-

ing in age from 33 to 68 yr (mean age 48 yr) underwentcoronary artery bypass procedure using saphenous veingrafts. A total of 29 coronary arterial obstructions were by-passed and evaluated in this study. 10 grafts were placed onright coronary arteries, 15 to left anterior descending coro-nary arteries, and 4 to left circumflex coronary arteries.\Vith the exception of four patients (A. M., C. B., H. C.,and J. R.), all operations involved single bypass grafts.

1'Abbreviations ised in this paper: CAB, coronary arterybypass graft; CC, coronary collaterals; MBF, myocardialblood flow.

2 556 The Journal of Clinical Investigation Volume 51 October 1972

Intraoperative determination of nzyocardial blood flow andpressuire. Aorta to coronary artery bypass operations wereperformed under morphine anesthesia using saphenous veingrafts. Total cardiopulmonary bypass with induced ventricu-lar fibrillation and hypothermia to 28-30'C was employedduring the period of graft anastomosis. Measurement of coro-nary arterial pressure and blood flow was performed intra-operatively under stable hemodynamic conditions 20-30 minafter terminating cardiopulmonary bypass, at normal tem-peratures, and in normal sinus rhythm. For the vessels stud-ied in this series, the mean aortic pressure at time of flowmeasurements ranged from 62 to 100 and averaged 79+11(SD) mmHg. A small polyethylene catheter PE No. 160 wasintroduced into the distal saphenous vein bypass graft (Fig.1) near its anastomosis to the distal coronary artery. Thepreexisting MBF determinations were made with the graftcross clamped proximally. 'Xe (0.5 ml., 1.1 mCi) was in-jected into the distal saphenous vein bypass graft via the PENo. 160 tubing and flushed into the distal coronary arterywith 3 ml normal saline solution. A shielded scintillation probewith 1-inch sodium iodide crystal was positioned approxi-mately 10 cm above the myocardium perfused by the distalcoronary artery. Counts were recorded with an analog, linearratemeter (Picker Labmeter I, Picker Corp., Cleveland,Ohio) set at a time constant of either 0.1 or 0.3 sec, connectedto a Picker 2990 rectilinear strip chart recorder. After aperiod of 40-60 sec the clamp was removed from the proxi-mal bypass graft permitting the myocardium to be perfusedby flow through the coronary artery bypass graft in additionto the preexisting flow from collaterals or antegrade acrossthe coronary arterial stenosis. Precordial counting with thegraft open was continued for 2i-3 min. In this manner, myo-cardial blood flows in the region supplied by the distal coro-nary artery were obtained with the graft open and occluded,thus permitting comparison of MBFbefore and after bypassgraft. The sequence of occlusion and patency of the graftwas randomized. Repeat injections were made with thegraft open or occluded at 5-min intervals in several patientsto document reproducibility of the method. In one patientnot reported herein who underwent subsequent cardiaccatheterization postoperatively, the 3Xe clearance curvesobtained by direct injection into the bypass graft were com-parable to the intraoperative measurements obtained withthe bypass graft open (110 vs. 113 ml/min per 100 g, re-spectively). These findings suggest the clearance rates ob-tained intraoperatively are similar to those which exist dur-ing life apart from effects of anesthesia and a previous pe-riod of extracorporeal support of the circulation. In sevenpatients direct injections (0.1 ml 3Xe in saline) were madethrough a 25 gauge needle into epicardial fat to determinethe washout pattern and ti for lipid tissue.

Myocardial blood flows were calculated from the formulaF = (K * X *w) /P in which the clearance constant K = 0.6931/half time in minutes, X = myocardium-to-blood partition co-efficient for '3Xe (0.7) (8), w= 100 g, the actual weight ofmyocardium being unknown, but by convention flow is ex-pressed per 100 g of tissue, and P = 1.05, the estimatedspecific gravity of myocardium (8-10). As discussed in theresults, the majority of washout curves were best resolvedinto two exponential components (11) on the basis of a semi-logarithmic plot.

Perfusion pressure in the distal coronary artery was deter-mined via manometer connected directly to the PE tubing(Fig. 1). Pressures were obtained (a) during clamp occlu-sion of the proximal coronary bypass graft and (b) withgraft open. The pressures obtained with graft open (effectivecentral artery pressure) were compared to radial artery mean

FIGURE 1 Schematic to demonstrate method for measuringintraoperative '3Xe flow and pressures.

pressure. The pressure gradient across a given coronaryarterial obstruction was determined by cross clamping thebypass graft proximally and simultaneously measuring pres-sure in the radial artery and distal bypass graft.

Analysis of coronary arteriograms. Selective coronary ar-teriograms were performed preoperatively on each patientwithin 3 wk of the operative procedure utilizing 9-6 inchGeneral Electric (General Electric Company, Schenectady,N. Y.) or Siemens (Siemens Corp., Iselin, N. J.) 10-5 inchbiplane image intensifier and recording on 16 mmfilm at 60or 75 frames per sec, respectively. All patients had films ofexcellent radiographic quality. The arteriograms in bothright and left anterior oblique proj ections were evaluatedindependently by three observers to determine the site anddegree of coronary obstructions.

The per cent obstruction was calculated from the ratio ofthe smallest diameter of vessel visible at the lesion in eitherprojection to the diameter of normal vessel immediately ad-jacent to the lesion in the same projection. The films werereviewed carefully to detect the presence of collateral vessels,and if present, the collaterals were classified anatomically aseither epicardial or intramyocardial in origin. The time, inmilliseconds, for the coronary artery distal to an obstructionto opacify after injection of contrast agent was measuredusing a projector-linked frame counter, taking the frame im-mediately preceding appearance of contrast agent at the ori-gin of the coronary artery as to, and the first frame demon-strating opacification distal to the obstruction as the end point.This value, termed arteriographic appearance time, was cal-culated at least three times for each obstruction to documentreproducibility of the method (within 30 msec). In thosevessels with both antegrade and collateral filling of coronaryartery beyond an obstruction, the distal coronary arterial fill-ing times were calculated for both antegrade and collateralsources of filling. The distinction between antegrade and col-lateral filling times was made from separate injections ofcontrast agent first into the affected artery and then intothe donor artery giving rise to the collateral vessel. Fig. 2demonstrates schematically the filling of contrast agent invessel distal to a subtotal obstruction where early appearance

Myocardial Blood Flow: Collaterals and Coronary Bypass Surgery 2557

E. 500mSec

FIGURE 2 Angiographic recognition of variable coronary

arterial transit times. The filling of contrast agent in a vesseldistal to a subtotal obstruction may occur directly by ante-grade flow from the proximal vessel (C) or slightly laterfrom collateral flow (D).

is associated with antegrade flow (Fig. 2C) and slightlylonger appearance time with collateral flow (Fig. 2D).

The vessels receiving bypass grafts were classified angio-graphically (Fig. 3) as follows: group A, vessels with lessthan 80% obstruction and no collaterals; group B, those withgreater than 80% obstruction and no collaterals; group C,those with greater than 80% obstruction and collaterals; andgroup D, those with total obstruction and distal vessel filledby collaterals only.

RESULTS

The results of preoperative arteriographic analyses andcorresponding intraoperative determinations of pressureand flow for the 29 vessels bypassed are summarized inTable I. The vessels are listed by arteriographic groups

based on degree of obstruction and presence or absenceof collaterals (Fig. 3). Intraoperative pressures were notmeasured in three patients (C. B., D. P., and W. B.),and flow rates were not measured with the graft clampedin two patients (J. R. and R. B.).

Effect of bypass graft on myocardial blood flow. Forthe total series, mean MBFwith bypass graft clampedwas 32+6 (SEM) ml/min per 100 g and increased to

KEY:

FIGURE 3 Arteriographic stages of atherosclerosis based on

per cent obstruction and presence or absence of collaterals.

118+13 ml/min per 100 g with the graft open (Fig. 4).This represented a 268% mean increase in MBFfor thetotal series after saphenous vein bypass graft. The levelof myocardial flow was not systematically influenced bythe artery receiving the graft or its site of attachment(Table I). Of interest were several vessels in which onlyminimal improvement in flow occurred (Fig. 4). Withthree exceptions (E. M., J. S., and H. W.), these were

all vessels from group A in which preexisting flow andpressure beyond the obstruction was near normal withthe graft clamped.

0

C,0

.C 150-

ET11t 118Or 11

50 2

321

INITIAL FLOW FINAL FLOWIN DISTAL CA IN DISTAL CA

FIGURE 4 133Xe myocardial blood flow determinations withbypass graft closed (initial flow) and open (final flow).Mean values ±+SE are shown for the group of initial and finalflows.

2558 Smith, Gorlin, Herman, Taylor, and Collins

STAGES OF CORONARYATHEROSCLEROSISA B C D

<80% >80% >80% TOTALOBST. OBST. OBST. OBST.

ANTEGRADE I

F-COLLATERALS-

TABLE IResults-of Preoperative Arteriography and Intraoperative Pressure-Flow Determinations

Arteriography

Dye appearance time Graft damped Graft open flowPressure

Patient Vessel* Antel Collat§ Flow Pressure gradient Fast Slow

msec ml/min mmHg mmHg ml/min

21716

15016

24016

350267

66283283350

333133311601173627693567

10671033

1347204020001093

7335520310018001817

20271167146716172283271715171250

per 100 g

78106

N.M.11948276

4231813

N.M.25

per 100 g

100 0 110 166 17 10670 0 8372 0 9490 0 82 150 20 76 1

4422566547

N.M.

54746855

N.M.60615347

50

24N.M.N.M.

60346343

3250101326

N.M.

7 3033 1924 3020 3410 N.M.12 3025 2126 3016 15

10213830111723

N.M.

2038

N.M.N.M.

22342133

1613

78

1313

81 14113 1527 -

287 14124 37103 17

30240238114170

35167

89128

42322319

262119

114 857

106 32106 16143 1549

190 1348

* RCA, right coronary artery; LAD,coronary artery.

left anterior descending coronary artery; LCF, left circumflex

I Ante, antegrade source of coronary opacification.§ Collat, collateral source of coronary opacification.11 N.M., not measured.

Fig. 5 summarizes the intraoperative evaluation in a

patient (H. D., Table I, group C) with severe obstructionto the left anterior descending coronary artery. Distalcoronary pressure (Fig. 5, lower left panel) rose from a

mean of 21 mmHg with graft clamped to 82 mmHgwith graft open. The calculated pressure gradient across

the coronary arterial obstruction was 61 mmHg. A large

increase in the precordial clearance of 'Xe was observedwith the graft open (Fig. 5, upper right panel). Thecalculated MBF from exponential plot (Fig. 5, lowerright panel) revealed a preexisting MBF (graftclamped) of 25 ml/min per 100 g. With augmented flowafter opening the bypass graft, the 'AXe clearance was

expressed as a double exponential with a rapid phase

Myocardial Blood Flow: Collaterals and Coronary Bypass Surgery

Group AC. B.L. B.J. R.J. S.M. C.M. R.

Group BA. M.E. Y.E. M.B. H.R. B.C. B.

Group CJ. S.J. C.M. C.R. D.D. P.H. W.H. C.C. D.A. M.

Group DN. G.R. S.C. B.WV. B.H. C.J. R.S. B.R. B.

RCALADRCALADLCFLAD

LADLADLADRCALADLCF

RCALADRCARCALADLADLADLADRCA

LADLADRCARCALCFLADRCALCF

2559

CC.ORONJARY FLCOW

OSED OPENI

B ;V

SISTAIL C)ORQtNARP PRESSURE11O

..O ifS- It

.j

_. QRQNARY--FLQW-YY

V) PE

9m.m r

flow 2 5f~-f,n:

8minf̂ low 26ccm

0 , m.!

'f'ow 167cr min - 10

FA --M-

FIGURE 5 Results of arteriogram and intraoperative pressure and 133Xe flow measure-

ments for coronary artery with severe obstruction. SVB, saphenous vein bypass.

of 167 ml/min per 100 g and a slow plhase of 26 ml/minper 100 g. In a majority of patients, the clearance of33Xe after bypass graft could not be resolved by a

monoexponential analysis, but was closely approximatedby two exponential delays which were termed "fast" and"slow" (Table I). In five patiei-.'s (J. R., E. Ml., J. S.,R. D., and N. G.) immediately after opening the bypassgraft, a flow rate 5-6 times greater than the fast flowclearance was seen which lasted for 20-25 sec.

Preoperative arteriographic analyses. The mean val-ues for arteriographic appearance time for each of thefour arteriographic groups are summarized in Fig. 6

(upper left panel) .

The mean appearance time for contrast agent to appearin vessel distal to an obstruction was significantly lower( P < 0.025) for group A vessels. Two appearance timeswere calculated for group C vessels since distal filling oc-

curred by both antegrade and collateral sources. Themean antegrade arteriographic appearance time for groupC vessels where collaterals were present was significantlylonger than that for vessels in groups A and B where no

collaterals were present. Mean appearance times for fill-

ing by collaterals were similar in groups C and D al-

though collateral filling occurred earlier in group D ves-

sels in which antegrade flow was not observed.Intraoperative pressure-flow mteasurements. The mean

values for preexisting MBF (graft clamped), preexisting

distal vessel pressure (graft clamped), and pressure gra-

dients across obstructions for each of the four arterio-graphic groups are summarized in Fig. 6. The preexist-ing blood flows with graft clamped (upper right panel)were near normal in group A vessels with less than 80%obstruction and no collaterals. In striking contrast,groups B, C, and D with greater than 80% occlusions hadvery low flows before opening the bypass graft. Similarly,the perfusion pressure in the distal coronary artery (lowerleft panel) for group A approached normal with thegraft clamped, whereas in groups B, C. and D pressures

were markedly reduced. Mean pressure gradients across

obstructions (lower right panel) were very low forgroup A vessels with less than 80% occlusion, whereasfor groups B, C, and D where obstructions were greaterthan 80%, the gradients were high. Groups C and D ves-

sels with collateral blood supply did not have signifi-cantly higher preexisting flow (upper right panel) thangroup B vessels with a similar degree of obstruction andno collaterals (P > 0.4).

Fig. 7 demonstrates typical intraoperative data from a

patient (L. B.) in group A with less than 80% obstruc-tion in the proximal left anterior descending coronary

artery. The intraoperative pressure gradient for thislesion (lower left panel) was 17 mmHg. No change inprecordial clearance of '3Xe was seen after opening thebypass graft (upper right panel). The exponential plot of

2560 Smith, Gorlin, Herman, Taylor, and Collins

iU.0

'I

2161

887

267

a00

aw-

< .4.

e- E< 3:VC °

90so70605040302010

o1

E 80

-1 70

Z 60Su0 504

26 28 0 40

X C C =~~~320

87

+11

KEY:

STAGES OF CORONARYATHEROSCLEROSISA a C D

<R0% >W0% >0% TOTALOEST. oBsT. OBST. oBst.

F- ANTEGRADE--COLLATERALS--

#iFiFl9 2

A B C D

I ~~~~~~59I ~47 [7IL 46

6__S

E V T X

A B C D

FIGURE 6 Summary of the mean arteriographic appearance times, graft clamped flows, graftclamped pressures, and pressure gradients across obstructions for each of the four stages ofatherosclerosis. Values are mean +SE. Value for both antegrade and collateral arteriographicappearance time are shown for group C.

the 'Xe clearance curve (lower right panel) showed a

rapid phase flow of 106 ml/;min per 100 g and a slowphase of 13 ml/min per 100 g.

183Xe clearance from epicardial fat. In a majority ofpatients (Table I), when MBFwas augmented by a by-pass graft, the clearance of 'tXe could not be resolved bya monoexponential analysis, but was closely approximatedby two exponential decays termed fast and slow phases.These findings suggested heterogeneity of either tissueor perfusion or possibly both, as an explanation for theslow phase of the 'Xe clearance curve. The possibilityexists that the known sequestrations of 'Xe in fat mightbe the source of tissue heterogeneity and the so-called"slow" phase of clearance. This was studied in seven pa-

tients by injecting 'Xe directly into epicardial fat. Theseclearance curves were compared with the slow componentof the curves obtained after injection into the open bypassgraft. In all seven patients (Table II), the ti for epi-cardial fat injection was significantly longer than theti for the slow phase of the MBFcurves.

The comparison of 'Xe clearance curves for myo-

cardium and epicardial fat in a patient (M. R.) are

shown in Fig. 8. The myocardial 'Xe clearance curve

(left) was best approximated by two exponential delays.The ti for the slow phase was less than one-half that ofepicardial fat (right).

DIS CUSSIONThe results of this study demonstrate the efficacy ofcoronary artery bypass surgery in restoring flow to previ-ously ischemic myocardium. The majority of vesselswith significant stenosis (groups B, C, and D) showedstriking increases in MBFwhen the bypass grafts wereopened. The values for 'Xe MBFafter bypass graft areconsistent with previous intraoperative studies usingelectromagnetic flowmeters (12-14). The clearance tech-nique measures flow per unit mass of myocardium andgives actual nutrient flow. This differs from electromag-

TABLE I I33Xe Clearance (ti) for Epicardial Fat and Slow Component

of Myocardial Blood Flow in Seven Patients

SlowEpicardial component

Patient fat t4 MBFt

msec msecM. R. 481 218D. S. 650 250J. S. 635 378W. Y. 550 155M. C. 470 227M. G. 467 349V. Y. 395 180

Myocardial Blood Flow: Collaterals and Coronary Bypass Surgery

225020001750150012501000

750500250 109

E_ X

U I-1

E-oz

ALE

_ D

a Wg

LUOo.I

7

A B C D

8070605040302010n

A B C D

2561

.ORONAP Ki

- LOSEDISVB C)PEN

D;S A ORONAR PRESSURE

o00

T IMEME

CORONARY9 W

00 .FOSEDSVB

iu-.

.. >- 10c,E 65g: b

* ;

C OSE ;.C0 vR

.)PENSV

OPEN.!ZV:B

* 3.67mc r

'N45mir

,flow i'06cc mt rr I OO9

" ME -

FIGURE 7 Results of arteriogram and intraoperative pressure and 'Xe flow measure-ments for mild coronary artery obstruction.

netic flows which measure only total flow irrespective ofits distribution. This report confirms the technical ade-quacy of bypass grafts in restoring nutrient flow to myo-cardium distal to a significant coronary arterial ob-struction.

Of considerable interest are the results for vesselswN-ith small pressure gradients (group A), in whichnegligible increases in MBFwere observed. In this groupthe preexisting pressure and flow were high. alnd the netcontribution of bypass grafts to MBF in the rest statewas minimal. Preoperative arteriography proved usefulin identifying such vessels by virtue of the small degreeof obstruction and the rapid appearance of contrast agentdistal to an obstruction after selective coronary injection.Johnson, Flemma, and Leplay (15) and others (12-14,16, 17) have observed a high incidence of graft occlusionin the presence of low graft flow rates. They attribute lowflow either to technical errors w-ith anastomosis, or tohigh resistance in distal vessels. The findings in the pres-ent study indicate that vessels with a minimal gradientacross obstructions and a high preexisting flow and post-obstructive pressure might form vet a third categorysubject to low graft flows anid high incidence of earlygraft closure. Alternatively, under conditions augmentingthe need for coronary flow, mild obstructions could as-sume functional significance, at which time flow via by-

pass graft would increase in response to a decrease inpostobstructive pressure with distal vasodilation.

Recent experimental studies by Elliot, Bloor, Jones,Mitchell, and Gregg (1) in conscious dogs revealed thatcollateral flow measure by 3Xe increased to controllevels after 4 days of progressive coronary constriction.In similar studies in anesthetized dogs, Rees and Redding(2) demonstrated '3Xe collateral flow to be 75% of con-trol rates after acute ligation of the ainterior descendingcoronary arterv, with return toward control flow ratesover a 10 day period. The results of the present study arenot in accord with these experimental observations inthe dog. Preexisting MBF was not significantly influ-enced by the presence of collaterals. The values for pre-existing MIBF in the presence of collateral circulationwere less than one-third of normal rates of perfusion.Since all patients receiving bypass grafts were sympto-matic, the vessels in the present study may have repre-sented that portion of the total population of vessels re-

ceiving inadequate collateral flow. However, it is note-

worthy that collateral flow was low in all vessels studied.Rees has alluded to variations in intercoronarv anasto-moses among species (18) and noted that they are more

frequent in the dog than in man. The low values forcollateral flow in the present study suggest that con-

clusions drawn from experimental data in animals, re-

2562 Smith, Gorlin, Herman, Taylor, and Collins

garding collateral flow, may be of limited applicability inhumans.

The low values for collateral flow documented by thepresent study may explain recent clinical observations inhumans. Sheldon (3) has commented on the inability ofcollaterals to protect against infarction, and noted thatcollaterals alone are not likely to be responsible for com-plete remission of ischemic symptoms. Miller, Zelis,Mason, and Amsterdam (4) could find no significant en-hancement of ventricular performance in association witharteriographically demonstrable collaterals in a matchedgroup of 38 patients at cardic catheterization. Helfant,Kemp, and Gorlin (5) studied 111 patients with coronarydisease and concluded the presence of collaterals did notprotect against the development of asynergy or abnormalhemodynamics. In subsequent studies on the clinical ef-fect of collaterals in 119 patients, Helfant, Vokonas, andGorlin (6) concluded that collaterals had no demon-strable beneficial effect on the resting or exercise electro-cardiogram. They found the patients with collaterals tohave an incidence of acute myocardial infarction duringa follow-up period which was nearly identical to a con-trol group without collaterals.

The results of the arteriographic analyses from thepresent study suggest that coronary arterial obstructionsbecome functionally significant as an 80% constriction ofthe lumen is approached. This observation is in accordwith studies by Tuna and Amnplatz (7), and Mason et al.(19) both of which found evidence of collateral formationwith obstructions of 75% or greater. The precision bywhich per cent stenosis can be accurately determinedfrom an arteriogram is variable. The results from thepresent study indicate that delayed appearance of con-trast agent distal to an obstruction correlates well withthe severity of the obstruction and provides a valuableadjunct in determining the significance of a given ar-teriographic obstruction. In the present study, the com-bination of 80% stenosis or greater, with prolonged ar-teriographic appearance times, was predictive of a highpressure gradient vwith marked reduction of postobstruc-tive MBFeven in the presence of collaterals.

In a majority of vessels, after MBFwas augmentedwith a bypass graft, the clearance of 'Xe could not beresolved by a monoexponential analysis but was closelyapproximated by two exponential decays. It is unlikelythat the slow component was due either to recirculationor background count since greater than 95% of 'Xe isremoved on circulation through the lungs, the backgroundcounts in the room did not vary appreciably during theprocedure and the proximity of probe to myocardiumassured minimal counts from 'Xe in the lungs andneighboring tissue. In general, the clearance curves wereunaffected by either the duration of graft occlusion orvariation in the sequence of isotope injection initially

Graf t

in0

x

IL

u

0

x

I

'C

A.

0

Lesion 70% LAD

PressureGradient 20mmHg

EpicardialFat

tv.2 515 seconds2

100 200

Time -seconds

300

M.R. #14-62-68

FIGURE 8 Comparative results for mXe intraarterial myo-cardial blood flow and 'lXe direct epicardial fat injection.LAD, left anterior descending artery.

with graft open or closed. However, as commented earlier,five patients did show a highly transient rapid clearancewith initial opening of the graft and this may have beendue to reactive hyperemia.

The persistence of a double exponential 'Xe clearancepattern after augmentation of myocardial flow by bypassgraft is theoretically consistent with a heterogeneity ofmyocardial tissue or perfusion pathways, or possibly both.Ross, Ueda, Lichtlen, and Rees (8) studied MBFusing1Xe in normal dogs and found clearance from the nor-mal myocardium to be best expressed by a monoexpo-nential curve with good correlation with rotameter flows.Johansson, Linder, and Seeman (11) studied MBFusing'krypton during acute coronary arterial occlusion indogs. Monoexponential clearance curves were obtainedbefore coronary arterial occlusion. After occlusion, clear-ance of 'Kr was best resolved by two separate expo-nential components which they related to heterogeneity offlow. They observed a rapid phase related to the elimina-tion of 'Kr from normally perfused regions of the myo-cardium, and a slow phase due to decreased perfusion ofthe ischemic myocardium. In similar studies on anasto-motic blood flow during experimental myocardial in-

Myocardial Blood Flow: Collaterals and Coronary Bypass Surgery 2563

farction in the dog using 'Xe, Rees and Redding (2)also found that clearance curves were best resolved intofast and slow components which they related to unevendistribution of flow through the infarct.

Heterogeneity of myocardial flow has also been dem-onstrated in patients with coronary artery disease. Instudying the effects of sublingual nitroglycerin on re-gional MBF, Horwitz, Gorlin, Taylor, and Kemp (20)found a double exponential decay pattern for diseasedmyocardium and suggested the slow phase was associ-ated with collateral flow. Cannon, Dell, and Dwyer (21),using a multiple-crystal scintillation camera to studypatients with coronary artery disease, found regionalvariations in '33Xe clearance constants demonstratingheterogeneity of myocardial perfusion. They observed asignificant reduction in clearance recorded by crystalsoverlying myocardium distal to arterial constrictionsgreater than 75%. Klocke et al. (22) demonstrated non-uniform myocardial clearance of hydrogen in patientswith coronary artery disease which they related to hetero-geneous flow.

Alternatively, the nonuniform myocardial clearance of133Xe may result from tissue heterogeneity. The partitioncoefficient for fat is 8-10 times that for mvocardium(23). Bassingthwaighte, Strandell, and Donald (24) ob-served a persistent tailing of 1"Xe MBF curves in nor-mal dogs which they attributed to changes in partitioncoefficient due to variations in the amount of adipose tis-sue from basal to apical regions of the heart. In thepresent study, clearance of 'Xe from epicardial fat wassignificantly slower than the slow phase from myocardialclearance curves, indicating that fat alone did not accountfor the nonuniform myocardial clearance of 1"Xe.

The most plausible explanation for persistence of non-uniform mXeclearance in the revascularized myocardiumis a combined heterogeneity of altered perfusion pathwaysand varving myocardial partition coefficients. Both ofthese are the direct consequence of coronary artery dis-ease. The revascularized myocardium probably includesa wide spectrum of pathologic changes with varying de-grees of inflammation, lipid deposition, and fibrosis. Thepartition coefficient for each of these tissues is differentand would result in altered clearance of 'Xe. The dif-ferences in perfusion may be explained by persistent ab-normalities in the distal large coronary arteries not cor-rected by revascularization and deviations in the micro-circulatory pathways due to prior destruction of tissuewith fibrosis.

ACKNOWLEDGMENTSThe authors wish to express their appreciation to DoctorsLawrence D. Horwitz, Edward S. Kirk, Stanley Baldwin,and Lawrence Smith for their advice and assistance. Thisstudy was supported by U. S. Public Health Service GrantsPO-1 HE-11306 and IT 1 HE-5679.

REFERENCES1. Elliot, E. C., C. M. Bloor, E. L. Jones, W. J. Mitchell,

and D. E. Gregg. 1971. Effect of controlled coronaryocclusion on collateral circulation in conscious dogs. Am.J. Physiol. 220: 857.

2. Rees, J. R., and V. J. Redding. 1967. Anastomotic bloodflow in experimental myocardial infarction. Cardiovasc.Res. 1: 169.

3. Sheldon, W. C. 1969. On the significance of coronarycollaterals. A)t. J. Cardiol. 24: 303.

4. Miller, R. R., R. Zelis, D. T. Mason, and E. A. Amster-dam. 1971. Relation of coronary collateral vessels to ven-tricular function in patients with equal extent of coronaryartery disease. Circuilation. 44(Suppl. II): 202. (Abstr.)

5. Helfant, R. H., H. G. Kemp, and R. Gorlin. 1970. Cor-onary atherosclerosis, coronary collaterals, and their re-lation to cardiac function. Annii. Interni. Med. 73: 189.

6. Helfant, R. H., P. S. Vokonas, and R. Gorlin. 1971.Functional importance of the human coronary collateralcirculation. N. Entgl. J. Med. 284: 1277.

7. Tuna, N., and K. Amplatz. 1970. The significance ofcoronary collateral circulation. Coronary arteriographicand electrovector-cardiographic correlations. Ami. J.Cardiol. 26: 663. (Abstr.)

8. Ross, R. S., K. Ueda, P. R. Lichtlen, and J. R. Rees.1964. Measurement of myocardial blood flow in animalsand man by selective injection of radioactive inert gasinto the coronary arteries. Circ. Res. 15: 28.

9. Kety, S. S. 1949. 'Measurement of regional circulationby the local clearance of radioactive sodium. Ami. HeartJ. 38: 321.

10. Herd, J. A., WI. Hollenberg, G. D. Thorburn, H. H.Kopald, and A. C. Barger. 1962. Myocardial blood flowdetermined with 'krypton in unanesthetized dogs. Am.J. Phlvsiol. 203: 122.

11. Johansson, B., E. Linder, and T. Seeman. 1964. Col-lateral blood flow in the myocardium of dogs measuredwith 'krypton. Acta Physiol. Scand. 62: 263.

12. Gron(din, C. NI., G. Lepage, Y. R. Castonguay, C. Meere,and P. Grondin. 1971. Aortocoronary bypass graft: initialblood flow through the graft, and early postoperativepatency. Circutlationt. 44: 815.

13. Danielson, G. K., G. T. Gau, and G. D. Davis. 1971.Early results of vein bypass grafts for coronary arterydisease. Circutlationi. 44(Suppl. II): 101. (Abstr.)

14. Walker, J. A., H. D. Friedberg, R. J. Flemma, and W.D. Johnson. 1971. Determinants of angiographic patencyof aortocoronary vein bypass grafts. Circulationt. 44(Suppl. II) :108. (Abstr.)

15. Johnson, W. D., R. J. Flemma, and D. Lepley, Jr. 1970.Direct coronary surgery utilizing multiple-vein bypassgrafts. Anni. Thorac. Suirg. 9: 436.

16. Field, P., A. S. Trimble, G. F. Trobridge, and H. E.Aldridge. 1970. Direct myocardial revascularization:saphenous vein bypass grafts to the distal coronaryartery. Ann1. Thorac. Sulrg. 10: 112.

17. Bourassa, M. G., J. Lesperance, C. M. Grondin, andL. Campeau. 1971. Factors influencing patency of aorto-coronary vein grafts. Circutlationi. 44(Suppl. II) : 107.(Abstr.)

18. Rees, J. R. 1969. The myocardial collateral circulation.Br. Heart J. 31: 1.

19. Mason, D. T., E. A. Amsterdam, R. R. Miller, J. L.Hughes III, J. A. Bonanno, A. B. Iben, E. J. Hurley,

2564 Smith, Gorlin, Herman, Taylor, and Collins

R. A. Massumi, and R. Zelis. 1971. Consideration of thetherapeutic roles of pharmacologic agents, collateral cir-culation and saphenous vein bypass in coronary arterydisease. Am. J .Cardiol. 28: 608.

20. Horwitz, L. D., R. Gorlin, W. J. Taylor, and H. G.Kemp. 1971. Effects of nitroglycerin on regional myo-cardial blood flow in coronary artery disease. J. Clin.Invest. 50: 1578.

21. Cannon, P. J., R. B. Dell, and E. M. Dwyer. 1970. Re-gional myocardial perfusion in man. J. Clin. Invest. 49:16a. (Abstr.)

22. Klocke, F. J., R. C. Koberstein, D. E. Pittman, I. L.Bunnell, D. G. Greene, and D. R. Rosing. 1968. Effectsof heterogeneous myocardial perfusion on coronary venousH2 desaturation curves and calculations of coronary flow.J. Clin. Invest. 47: 2711.

23. Conn, H. L., Jr. 1961. Equilibrium distribution of radio-xenon in tissue: Xenon-hemoglobin association curve.J. Appl. Physiol. 16: 1065.

24. Bassingthwaighte, J. B., T. Strandell, and D. E. Donald.1968. Estimation of coronary blood flow by washout ofdiffusible indicators. Circ. Res. 23: 259.

Myocardial Blood Flow: Collaterals and Coronary Bypass Surgery 2565