average coronary blood flow per unit weight left...

Average Coronary Blood Flow Per Unit Weightof Left Ventricle in Patients With

and Without Coronary Artery DiseaseBy FRANCIS J. KLOCKE, M.D., IVAN L. BUNNELL, M. D., DAVIDG GREENE, M.D.,

STEPHEN M. WITTENBERG, M.D., AND JOHN P. VISCO, M.D.

SUMMARYAverage left ventricular flow per unit mass (F/W) has been evaluated at rest in 20 normal individuals and

26 patients with arteriographically proven, advanced coronary artery disease, using inert gas techniquesmodified to take into account methodological problems presented by heterogeneous perfusion within theventricle. Preliminary studies in a canine model in which corornary flow could be measured directly beforeand after coronary occlusion indicated (1) that inert gas techniques utilizing venous sampling are suitablefor abnormal situations when appropriately long periods of saturation and desaturation, and careful resolu-tion of prolonged venous-arterial differences, are employed; and (2) that traditional inert gas methods are

not adequate when F/W is abnormally heterogeneous. Using helium as a tracer, coronary disease patientsshowed a systematic reduction, in comparison to normal individuals, in average left ventricular F/W at rest

(54 ± 11 vs 70 ± 13 ml/min/100 g, P < 0.01). Simultaneous studies with traditional nitrous oxide and kryp-ton techniques did not show this difference because of the methodological limitations of these techniques inthe presence of abnormally heterogeneous F/W. We conclude: (1) that useful, accurate measurements ofF/W can be obtained in coronary disease if appropriate methodological precautions are taken; and (2) thatappreciable portions of the left ventricle have an abnormally low F/W, even at rest, in patients with ad-vanced coronary disease.

Additional Indexing Words:Coronary blood flowGas chromatographyHeterogeneous myocardial perfusion

ALTHOUGH THE CLINICAL manifestations ofcoronary atherosclerosis are agreed to be re-

lated to an imbalance between myocardial oxygen de-mand and supply, measurements of coronary bloodflow have been of limited value in clarifying specificclinical situations and general pathophysiologicalprinciples. Most existing measurements have been ob-tained using modifications of the inert gas technique

From the Department of Medicine, State University of New Yorkat Buffalo and the Buffalo General Hospital and the E. J. MeyerMemorial Hospital, Buffalo, New York.

Presented in part at the 1970 National Meeting of the AmericanFederation for Clinical Research, Atlantic City, New Jersey; the VIWorld Congress of Cardiology, London, England; and a Sym-posium on Coronary Blood Flow sponsored by Laboratorio diFisiologia Clinica C.N. R. and the Council on Clinical Science of theInternational Society of Cardiology, Pisa, Italy.

Supported by Grants HL-09587 and HL-15194 and contract PH43-69-28 from the National Heart and Lung Institute, Bethesda,Maryland.

Address for reprints: Francis J. Klocke, M.D., Room C-169,SUNYAB Clinical Center, E. J. Meyer Memorial Hospital, 462Grider Street, Buffalo, New York 14215.

Received March 25, 1974; revision accepted for publication May20, 1974.

Circutlation, Volutme 50, September 1974

Tissue-blood partition coefficientsInert gas flow measurements

Validation of flow-measurement techniques

developed by Kety and Schmidt for the cerebral cir-culation1 and subsequently modified by Eckenhoff etal.2 for the coronary circulation. There has beengeneral agreement that average flow per unit weight(F/W) for the left ventricle does not show a systematicreduction in patients with coronary artery diseasestudied at rest in the cardiac catheterizationlaboratory.3 However, in view of the knownheterogeneity of flow in coronary disease, it has notbeen clear that the reported measurements have beenmethodologically adequate to ensure inclusion oflocalized areas within the heart having a below-average perfusion. One difficulty in resolving thisquestion has been the lack of experimental datavalidating specific inert gas techniques in situations inwhich myocardial perfusion is abnormallyheterogeneous. The present investigation was in-tended (1) to examine the validity of measurementsemploying three different inert gas techniques in ananimal model in which coronary flow could bemeasured directly before and after coronary occlu-sion; and (2) to re-examine resting levels of averageF/W in conscious patients with and without coronarydisease. During the latter studies, measurements

.547

by guest on June 2, 2018http://circ.ahajournals.org/

Dow

nloaded from

http://circ.ahajournals.org/

KLOCKE ET AL.

employing recent modifications of inert gas tech-niques were compared with simultaneousmeasurements employing more traditional ap-

proaches. The findings indicate that traditionalmethodology is not adequate in the presence of cor-

onary disease and that there is indeed a systematicreduction in average F/W in patients witharteriographically advanced coronary disease studiedat rest in the cardiac catheterization laboratory.

Methods

Validation Studies

A right heart bypass preparation was utilized so that totalcoronary outflow could be measured directly by timedcollection in a graduated cylinder during inert gasmeasurements of average F/W (fig. 1). Seventeen mongreldogs were anesthetized with sodium pentobarbital, in-tubated and ventilated with a constant volume pump.Atelectasis was minimized by 2-4 cm of continuous ex-

piratory positive pressure. The chest was opened through a

median sternotomy, appropriate structures were isolatedand the heart suspended in a pericardial cradle. Sodiumheparin, 10,000 units, was administered intravenouslv andsupplemented with additional 5,000 unit doses at 30 min in-

tervals. Systemic venous return drained through superiorand inferior vena caval cannulae into an external reservoirprimed with blood from a donor animal. An occlusive rollerpump passed blood at a constant rate from the reservoirthrough a heat exchanger and a Bardic cannula inserted into

PUMP EJ

Figure 1

Canine right heart bypass preparation for validation studies.(Ao = aorta; PA = pulmonary artery; LA = left atrium; LV = leftventricle; SVC = superior vena cava; IVC = inferior vena cava;RA = right atrium; RV= right ventricle; CS = coronary sinus;LAD = left anterior descending coronary artery; RCA lig. = rightcoronary artery ligature; CS samp. cath. = coronary sinus samplingcatheter; Res. = reservoir; Heat exc. = heat exchanger).

the main pulmonary artery through a right ventricular stabwound. Retrograde leakage of blood into the right ventriclewas avoided by doubly ligating the main pulmonary arteryaround the cannula. Total coronarv outflow drained to a sec-ond reservoir through another cannula, having multiple sideholes, which was inserted into the right ventricle throughthe azygos vein. The contents of this reservoir were returnedto the systemic venous reservoir at 2-4 minute intervals.WXhen coronary flow was to be measured directly, right heartoutflow was diverted to a graduated cylinder for timed 30-60sec collections. Flow to the nonworking right ventricle wasminimized by ligating the right coronary artery high in theA-V groove and its superficial branches to the conus as theyreached the anterior portion of the-ventricle. This procedureproduced obvious cyanosis of the right ventricular free wallexcept for the few millimeters nearest the septum (suppliedby tributaries of the left coronary arterv). Surface gastransfer, which is known to occur in open-chestpreparations,4 was minimized by enclosing the ventricles ina closely fitting latex covering. Arterial pressure wasmonitored through a catheter in the aortic arch, which wasalso used to sample systemic arterial blood during inert gasmeasurements. A left ventricular catheter was available forinjection of dissolved gas tracer. Coronarv venous blood wassampled through a 7 or 8 F Sones catheter introduced intothe coronary sinus from the right atrial appendage and ad-vanced 2-3 cm beyond the coronary sinus ostium. Systemicarterial temperature was monitored with a thermistor in-serted into the descending aorta and the heat exchanger ad-justed to maintain aortic temperature constant within ±10at 35-38°C. Prior to coronary ligation, systolic and diastolicarterial pressures ranged from 80 to 175 mm Hg (average121 mm Hg) and 60 to 110 mm Hg (average 74 mm Hg),respectively. Corresponding values after coronary ligationwere 85 to 160 mm Hg (average 117 mm Hg) and 60 to 100mm Hg (average 73 mm Hg). Heart rate varied from 75 to165 beats/min (average 122 beats/min) before coronaryligation, and from 80 to 145 beats/min (average 111beats/min) after ligation.

Different inert gases were studied simultaneously toevaluate the effects of varying periods of saturation anddesaturation on the measurement of average F/W.Measurements employing helium (He) were particularlydesigned to include areas of lOW flow5 while those employingnitrous oxide (N2O) and krvpton (Kr) were intended toduplicate techniques commonly used in other laboratories.For He measurements, the inspirate was switched fromroom air to a mixture containing 75Cc He, 4C N2, and 21%02 twentv minutes prior to the onset of desaturation. ForN20 measurements, 4% N20 was substituted for the 4% N2during the last 10 minutes preceding desaturation. For Krmeasurements, 5-10 ml of isotonic saline saturated with dis-solved Kr was injected into the left ventricle at the onset ofdesaturation. Fifteen to twentv pairs of 2.0 ml arterial andcoronary venous blood samples were drawn during the lastfew minutes prior to desaturation and the 20-25 minutes im-mediately thereafter and analyzed for inert gas concen-trations with a gas chromatograph. Helium analyses wereperformed on alternate pairs using a unit describedpreviously.6 N20 and Kr analyses were performed on theremaining pairs, using a similar unit. The carrier gas washelium and the test gas sample passed successively throughan 8-10 foot column of Poropak Q (for isolation of N20 fromthe remainder of the sample), a thermistor detector (formeasurement of N20), an 8-10 foot column of activatedcharcoal (for isolation of Kr) and a second thermistor detec-

Circulatiorn Volumne 50, September 1974

5-48


Dow

nloaded from


AVERAGE LV F/W WITH AND WITHOUT CAD

tor (for measurement of Kr). Tracer concentrations werequantitated from the heights of the appropriatechromatographic peaks and expressed in arbitrary units(preliminary experiments having verified linearrelationships between tracer concentration and peakheight). Helium values for average F/W were obtained us-ing the Kety-Schmidt formula5 and desaturation curvescovering the entire 20-25 minute period of observation. N20values for F/W were obtained using the same formula andcurves covering only the initial 10 minutes of desaturation.Kr values were calculated using a monoexponentialclearance constant derived from the initial 2-4 minutes ofdesaturation.7 Tissue-blood partition coefficients werecalculated as described subsequently.

Measurements of average F/W were obtained beforeand/or approximately one hour after production of alocalized myocardial infarction. The latter was ac-complished by ligating several adjacent diagonal branchesof the left anterior descending coronary artery supplying theanterior free wall of the left ventricle (fig. 1). Total coronaryoutflow was measured repeatedly prior to the onset of traceradministration to verify the existence of a stable state.Timed outflow collections were obtained at 1-2 min intervalsthroughout desaturation and average values for directlymeasured outflow converted from total flow (ml/min) toF/W (ml/min/100 g) by dividing measured outflow by thepostmortem weight of the left ventricle, septum and smallperfused segment of adjacent right ventricular free wall (thelatter averaging 7.5 ± 0.9 [sD]% of the total weight).

Clinical StudiesForty-six patients were studied in conjunction with

diagnostic cardiac catheterization after having given in-formed consent. Twenty were classified as having normalcoronary circulations and are subsequently referred to as thenormal patient group. Thirteen of the 20 were studiedbecause of chest pain not thought clinically to representangina pectoris. All had normal coronary arteriograms. Theremaining seven individuals included five with functionalmurmurs and no cardiac complaints, one with mild mitralstenosis and a normal left ventricle, and one with mildmitral regurgitation but a normal left ventricle (related tothe systolic click-late systolic murmur syndrome). Coronaryarteriograms were not performed in view of the absence ofevidence of coronary disease. Systolic and diastolic arterialpressures for the entire group ranged from 95 to 135 mm Hg(average 116 mm Hg) and 60 to 80 mm Hg (average 73 mmHg), respectively. Heart rates varied from 58 to 90 andaveraged 73 beats/min.

Twenty-six patients had typical clinical features of anginapectoris and underwent coronary arteriography. All showedgreater than 50% occlusion of at least one of the three majorcoronary arteries and most had abnormalities of two or threeof these vessels. Systolic and diastolic arterial pressuresranged from 85 to 168 mm Hg (average 121 mm Hg) and 50to 95 mm Hg (average 71 mm Hg), respectively. Heart ratesvaried from 55 to 100 and averaged 74 beats/min.

Measurements of coronary flow utilized the principlesoutlined in the validation studies. An arterial needle or smallarterial catheter was inserted percutaneously into a femoralartery and a Sones catheter was advanced into the coronarysinus from a venotomy in the right antecubital fossa. The tipof the coronary sinus catheter was advanced at least 2 cmbeyond the coronary sinus ostium (as seen in the frontalprojection) to avoid right atrial admixture.5 The nostrilswere occluded and the patient breathed through aCirculation, Volume .50, September 197a1

mouthpiece and low resistance unidirectional valve. Allpatients breathed helium (in concentrations ranging from50-79%) throughout a 20-minute saturation period. Twenty-five patients had 2-5% N20 added to the inspirate for thefinal ten minutes of test gas breathing. In thirteen patientsin whom a left ventricular catheter was available, 10-20 mlof isotonic saline saturated with dissolved Kr was injectedinto the left ventricle at the onset of desaturation. Arterialand coronary sinus blood samples were drawn during thefinal few minutes of test gas administration and for 20minutes thereafter.

Partition CoefficientsPartition coefficients were measured directly in additional

in vitro experiments. Twenty to sixty gram specimens of leftventricular myocardium were obtained from human heartsat the time of autopsy and from dog hearts immediatelyfollowing pentobarbital anesthesia. The human tissues werefrom individuals in whom the primary cause of death wasnot heart disease; a few specimens showed patchy myocar-dial fibrosis due to coronary atherosclerosis but results werenot demonstrably different from other specimens. Allspecimens were initially stripped of epicardium and en-docardium and cut into small slices, omitting visible bloodvessels and connective tissue. The slices were divided intotwo or three portions, weighed, diluted on a 1:1, 1:2 or 1:3basis with isotonic saline and homogenized with a smallWaring blender and an ultrasonic homogenizer. Eachhomogenate was equilibrated in a Farhi tonometer8 at 37°Cwith a gas mixture containing He, N20 or Kr, or a combina-tion of these tracers. Isotonic saline was tonometered withthe same gas mixture and 2.0 ml aliquots of eachequilibrated homogenate and saline were analyzedchromatographically as outlined above. A tissue-saline parti-tion coefficient expressing the amount of tracer contained in1 g of tissue as a fraction of that contained in 1 ml of salinewas calculated for each gas from the fraction of myocardiumin the homogenate and the chromatographic peak heightsfor the test gas in the homogenate and saline (assumingmyocardial specific gravity to be 1.05). Coefficients for the 2-3 homogenates prepared from each tissue specimen wereaveraged (the individual values always agreeing within±10% of one another).

Since the solubilities of He, N,0 and Kr differsignificantly in plasma and erythrocytes, data were alsocollected relating the relative solubility of each gas in bloodand saline to hematocrit. Blood specimens were collectedfrom six healthy laboratory workers as well as seven dogsanesthetized with pentobarbital. A portion of each specimenwas centrifuged to obtain plasma and a high-hematocritconcentrate. Isotonic saline, plasma, the original specimenand the high hematocrit concentrate were equilibrated in aFarhi tonometer with a gas mixture containing He, N20 andKr and analyzed as described above. The chromatographicpeak height of each plasma or blood sample was expressedas a fraction of the peak height of the corresponding salinesample, and human and canine data were each pooled to ob-tain least-squares regression lines relating blood-saline parti-tion coefficient (the amount of tracer contained in 1 ml ofblood expressed as a fraction of that contained in 1 ml ofsaline) to hematocrit.

Tissue-blood partition coefficients for individualmeasurements of F/W were obtained by dividing theaverage human or canine tissue-saline partition coefficientbv the blood-saline partition coefficient appropriate to thehematocrit.

549


Dow

nloaded from


KLOCKE ET AL.

Table 1

Partition Coefficient Determinations

Myocardium-Saline (370C):

HeKrN20

HumanN mean - SD11 0.85 - 0.047 1.09 - 0.07

11 0.95 - 0.06

N13107

Canmean

0.85 -

1.08 =0.96 -

Blood-Saline (370C):Regression equation

HeHuman: P.C. = 0.96 - 0.17 (Hct)Canine: P.C. = 0.95 - 0.17 (Het)

KrHuman: P.C. = 1.06 + 0.50 (Hct)Canine: P.C. = 1.02 + 0.78 (Hct)

N20Human: P.C. = 0.95 + 0.16 (Hct)Canine: P.C. = 0.97 + 0.19 (Hct)

z

line G

SD W

t0.05 cv

1-t 0.04 zt 0.06 U0

0

SEE cv)

00.018 0

00.021 °m

0.0360.032

0.0460.032

Abbreviations: N = number of left ventricles tested; SD =

standard deviation; P.C. = partition coefficient; Hct =

hematocrit (ml/100 ml); SEE = standard error of estimate.

Results

Partition Coefficients

Data for tissue/saline partition coefficients andregression lines relating blood/saline partitioncoefficients to hematocrit are shown in table 1.Tissue/blood partition coefficients calculated for ahematocrit of 40 in man were 0.95 for He, 0.94 forN20 and 0.87 for Kr.*

Validation Studies

Representative sets of desaturation curves beforeand after left anterior descending ligation are il-lustrated in figures 2 and 3. Data from all 17 animalstudies are given in table 2 and figure 4. Prior to leftanterior descending ligation, values for F/W obtainedfrom He desaturation curves averaged 99 ± 10 (SD)%of directly measured F/W and the mean of thedifferences between He and metered flow was 7% ofmetered flow. Corresponding averages for N20 and Krwere 104 ± 11% and 98 + 11%. Following leftanterior descending ligation, small prolonged venous-arterial differences were apparent in the He desatura-tion curves, while differences from pre-ligation curvescould not be identified for N20 and Kr. He flowsaveraged 96 ± 9% of directly measured flow and the

*Tissue-blood partition coefficients in this paper relate theequilibrium concentrations of tracer in 1 g of myocardium and 1 mlof blood. In some other studies, these coefficients have been ex-pressed in terms of the relative concentrations of tracer in 1 ml ofmyocardium and 1 ml of blood, with a value for myocardial specificgravity being added subsequently in the calculation of F/W. Thepresent values would be 5% larger if expressed in this form.

------- = PRE - LIGATION

- = POST - LIGATION

8-

4 -

0-10 15 20 25

TIME (MIN)

Figure 2

Arterial and coronary sinus desaturation curves for He before andafter coronary ligation in a canine right heart bypass study (animal7, table 2).

mean of the differences between He and metered flowwas 9% of metered flow. N20 and Kr flows exceededdirectly measured flow (N20 = 124 + 22% of directlymeasured flow, P < 0.02 [paired t using absolutevalues for F/W] and Kr = 122 ± 14% of directlymeasured flow, P < 0.01 [also paired t using absolutevalues for F/W]).

Clinical Studies

Data from all patients, including arteriographic"scores" as described by Rowe et al.,3 are given in

z

cv

z0

z

110

0

z

00-C)

--o-°-.-- = PRE - LIGATION-- = POST - LIGATION

2 4 6 8

TIME (MIN)

R(I)

z

cr

z0uJ

z0

000£0

ii

10-

3-

1- i0 2 4 610

TIME (MIN)Figure 3

Arterial and coronary sinus desaturation curves for N20 andsemilogarithmically plotted coronary sinus desaturation curves forKr in the same canine study illustrated in figure 2.

Circulation, Volume 50, September 1974

550


Dow

nloaded from



I~ I~; c Xz m IIn

-bc: Id o rvC:X b D -

t~~~~~~~I_|q - Cqq m_N 1C-4 mS m48t>><C

q~~ ~ ~~~~~~c cc, L

,! -_L-

-Nil-H-Ht- 0 m

00m m C

- CMz 10

H -H-H00 t- 0m0mC4t

P- 0m m mOC~ - MOL

H -f m m-e:

t-ds

m LsO O- °Om t- .4 0 m0 0 0 X-4-- --

II I- 00 XS oo oo Icq t ooI bI t" " m 00 h d m = t- t_ L _l

I II _L:It-OCI - ZIsL: I_s 00 Cqt I" M,,lCl- LXU: 0

= tI-C C5 L- cC

11 11C CM mD

- r_ - =ot- LC UZ <

*0tt- m CQ L& C =O 00-H -HH I-H IH 'I,00 00 cl t- _ x:o -- 00 00 t- m0 _-

_-4 "-_ -_-_

-. C m cz0 0 0 O-000 cc C)O° 0" crm -4 c cq c,] = m cr D to C) Cnm"_- -- _---- ----- C, - -- c' -

---- - - -- ~i"..Icq m 11C C: t- X m: 0 P-4 q mX d" LC = t_ (1 * -_I4 _4 ~4 _ ~ _4 -_ Is

UD0

--0zo

551

0

a

0

0

0I0

0 -i

0.

00: R

eq

0 to0

-z--0

0

cz

a1000

0

CO

-4

-.

-4O

0e-4

G0

0

w00a'0 i,

Q4 L4

m to


0z

0.. " "t- ii-i " M Cq X 0 ---4 t.- " 004, 1-

C. L0 " c t- t- t- C. ~

In. cn N X 0 (ZO LO Intl X m X N m m LO t-= m " " m X .zo " m t- t- 1- LO ~1


Dow

nloaded from


KLOCKE ET AL.

o = CONTROL*- AFTER LAD

LIGATION

WS(N((NSf

0

He N20 Kr

Figure 4

Comparisons of inert gas and directly measured flows in caninevalidation studies. Each data point represents an inert gas flow ex-

pressed as a percentage of the simultaneous directly measured flow.Absolute values of flow are not shown. Lines connecting open andclosed circles indicate observations before and after coronary liga-tion in individual animals. Solid horizontal bars represent mean

values for each set of data; P values were calculated using paired t-tests and absolute values of inert gas and directly measured F/W.(NS = not significant).

tables 3 and 4 and figure 5. Representative studiesfrom a normal patient and a patient with coronary

artery disease are illustrated in figures 6 and 7. Valuesfor average F/W using He averaged 70 ± 13ml/min/100 g in the normal patient group and54 ± 11 ml/min/100 g in the coronary disease group

(P < 0.01). Measurements employing N20 averaged70 ± 9 ml/min/100 g in 12 normal patients(108 ± 16% of simultaneous measurements employ-ing He) and 69 + 11 ml/min/100 g in 13 coronary

disease patients (138 ± 18% of simultaneousmeasurements employing He, P < 0.01, paired t, fig.8). Measurements employing dissolved Kr averaged83 ± 19 ml/min/100 g in 12 coronary disease patients(154 ± 29% of simultaneous measurements employ-ing He, P < 0.01, paired t, fig. 8). Only one such com-

parison was available in a normal patient (figs. 6 and8).

Discussion

Methodological Considerations and PreviousValidation of Inert Gas Measurements of F/W

These studies indicate a systematic difference inaverage F/W at rest in patients with and without cor-

onary artery disease. Since this finding is at variancewith previous reports, it is pertinent to considermethodological differences between this and otherstudies and to review animal studies in which inert gas

Table 3

Data in Normal Patient Group

No.

1

2

34a

678

9101112131415

161718

1920

Age Sex

45 M

39 mI18 AvI43 mI43 M

46 m

37 m

38 m

17 F29 F22 F41 F48 F36 F28 F41 F4,5 F54 F46 F34 F

Inert gas flowsArteriographic ml/min/100 g

score (3) He N20 Kr

275 78300 52

71300 67300 60300 54300 a6300 77

71300 5)2

66300 82

861059070

300 62300 65268 67300 69Mean: 70

SD: 13

75 76.52

6767

63907180

78

666964

709

Table 4

Data in Coronary Artery Disease Patients

Inert gas flowsArteriographic ml/min/100 g

No. Age Sex score (3) He N20 Kr

1 47 M 15.5 58 S.5 842 42 m 1735 58 853 54 m 115 43 794 48 m 163 62 - 695 59 m 125 40 776 42 m 210 45 627 53 m 95 95 1228 63 AM 188 49 559 45 M 115 58 88 11810 56 m 2.5 49 6211 57 m 217 .58 7412 37 m 8.5 6413 33 m 105a 53 6214 53 M 125 47 6015. 53 m 210 52 80 8416 61 m 200 42 5117 59 M 98 6918 56 m 225 5419 42 m 193 6220 42 m 125 .i.21 53 m 210 44 J56 7522 44 M 200 57 8523 44 F 150 56 6524 37 F 238 52 7525 62 F 200 38 6826 48 F 173 46 70

Mean: .54 69 83SD: 11 11 19

Circutlation, 'oluime 50, September 1974

170-

150-

% OF130-

DIRECTLY

MEASURED

FLOW 110-

90-

70

552


Dow

nloaded from



110-

90-

AVERAGEF/W

(ml/min/lOOg)70-

50-

30-

of myocardium being evaluated. In addition, the rateof exchange of tracer between tissue and capillariesmust not be limited by the physical process of diffu-sion.

0

0

0

8of

v00

0

00

NORMALFigure 5

0

TTC000

CAD

Individual values of average LV F/W in normal individuals andcoronary disease patients. Solid horizontal bars represent mean

values for the two groups. Note reasonably good separation ofgroups ebove 70 and below 50 ml/min/100 g, but large overlapbetween 50 and 70 ml/min/100 g. As discussed in the text, thegroup means differ significantly (P < 0.01).

measurements have been compared with directmeasurements of coronary flow. As discussed in more

detail previously,9 three requirements are especiallyimportant for commonly used inert gas techniques:

(1) Concentration of tracer must be uniformthroughout the myocardium at both the onset andcompletion of the flow measurement. If F/W is trulyhomogeneous, this requirement can be achieved withany duration of saturation or desaturation. When flowis not homogeneous, venous outflow originatespredominantly from areas having a higher-than-average F/W and changes in average tissue gas con-

centration (the numerator of the Kety-Schmidt equa-

tion) which are inferred from changes in venous (orarterial) gas concentration are incorrectly high until a

uniform tissue gas concentration has been achieved.(2) Analytical methods for measuring changes in

tissue or blood gas concentration must be sufficientlysensitive to quantitate small changes in tracer concen-

tration reflecting areas of below-average F/W. Thesechanges become evident only after the initial portionof the measurement and appear as small butprolonged venous-arterial (or tissue-arterial) test gasdifferences.

(3) For calculation of F/W from a monoexponentialrate constant of desaturation following bolus injectionof tracer, F/W must be uniform throughout the area

Circttlation, volurme .50, September 1974

(A) Techniques Employing Sampling ofCoronary Venous Outflow

Previous comparisons of inert gas measurements ofF/W and direct measurements of F/W have beenlimited in number and have not included cir-cumstances in which heterogeneity of F/W has beenincreased to an abnormal degree. Direct validation oftechniques employing coronary sinus sampling restsheavily on the careful work of Gregg and co-workerswith N2O.4 After defining problems related to surfacegas loss and other factors, N20 measurements during10-15 minute saturation periods were compared tovalues of F/W obtained from measurements of leftcoronary inflow and postmortem left coronary injec-tion. In nine animals, N20 flows averaged 101 ± 14%of directly measured flows (range 78-121%). The in-itial work of Eckenhoff et al.2 also attempted this typeof comparison but, for reasons outlined by Gregg, adefinitive conclusion was not possible. More recently,Aukland et al.10 reported 19 comparisons of directlymeasured venous outflow and F/W calculated frommonoexponential clearance constants followingseveral minutes of dissolved H2 infusion in six isolated,nonworking hearts. H2 values for F/W averaged97 ± 7% (range 72-105%) of directly measured F/W,the latter being expressed as cc/min/100 g of totalheart. The successful use of the monoexponentialclearance constant implied that flow was relativelyhomogeneous in the particular experimental prepara-tion. This conclusion was supported by the similarityof additional curves following bolus injection of H2 tothose following the more prolonged H2 infusion.The present observations in the right heart bypass

preparation prior to coronary ligation support thevalidity of measurements employing conventionalN20 techniques and bolus injection of tracer with cor-onary sinus sampling in the working heart whenregional variations in F/W are small. However, inview of the findings after coronary ligation, conven-tional techniques have important limitations whencoronary perfusion is abnormally heterogeneous. Theoverestimates of average F/W with Kr seem readilyunderstandable. When employing bolus injection, dis-tribution of tracer corresponds to distribution of totalflow and it is impossible to achieve the requisite uni-formity of tissue tracer concentration when regionalF/W varies. The use of a monoexponential rate con-stant reduces the likelihood of even partial representa-tion of areas of below-average F/W in the value foraverage F/W and, when employed in conjunction

55.3


Dow

nloaded from


KLOCKE ET AL.

100-(f) o-*~~~~... = ARTERIAL (JU)H ----=CORONARY I_SINUS HF

z zzDDCD80 i 80-

cr CE~~~~~~~~~~~~~~~~~30-8-

z z z0~~~~~~~~~~~~~~0600

H60 H

4- ~ ~ 1 lo

z LizLiJ 40 0 0 wl40 :z0

0 o0 00 ~~~~~~~~~~~~~~121620 240020 N 3

20- ~~~Z20-

0 00. 0

0 4 8 12 16 20 24 0 5 10 0 4

TIME(MINUTES)

Figure 6

Arterial and coronary sinus desaturation curves for He and N,0 and semilogarithmically p.lotted desaturation curve for Krin a normal individual (case 1, table 3).

ARTERIALCORONARY

SINUS

4- ... ............

0 ,12 16 20 24

.......,. _

CEE

z0F-

zlU

z0

0c\z

.. , .v

0 4 8 12 16 20 24 0 5 10

TIME(MINUTES)

Figure 7

Arterial and coronary sinus desaturation curves for He and N20 and semilogarithmically plotted desaturation curve for Krin a coronary disease patient (case 9, table 4).


^ 100-Uf)

z

C 80-CE

z0F60-

zLii0 40z0

I 20-

00-Jin

100-

z

:D -

< 30-z0 -

H 0-io

z -

IL -0z

0 3-

00

mcJ0 4

554

_ 11. 1

0


Dow

nloaded from



220

180

% OFSIMULTANEOUS

He FLOW 140-

100-

60

o = NORMAL* = CAD

A

(p< 0.01)a

a

(NS) -

ooo..cm-

N20 Kr

Figure 8

Comparisons of N20 and Kr values of F/W with simultaneous Hevalues of F/Win human studies. Each data point represents an N20

or Kr flow expressed as a percentage of the simultaneous He flow.Absolute values ofF/W are not shown. Solid horizontal bars repre-

sent mean values for each set of data; P values were calculated usingpaired t-tests and absolute values of F/W.

with bolus injection, assumes that F/W is uniformthroughout the heart. Overestimates of flow with N20after coronary ligation were more variable than thosewith Kr, though statistically significant as a group.

While the reasons for the greater variability are notclear, it is noted that N20 F/W constituted a largerpercentage of directly measured F/W after coronary

ligation than before in each of the seven animals inwhich N20 and directly measured F/W were com-

pared in both circumstances (fig. 4). This was not thecase for measurements employing He, where agree-ment of inert gas F/W with directly measured F/Wwas as good after ligation as before. The He data in-dicate that inert gas techniques utilizing venous out-flow sampling are suitable for abnormal situationswhen appropriately long periods of saturation anddesaturation and resolution of prolonged venous-

arterial differences are employed.

(B) Coronary Sinus Drainage as an Index of LV Perfusion

As pointed out by Gregg and Fisher,"` the use ofcoronary sinus blood samples as an index of F/W forthe left ventricle and septum is less well established inman than in dog. The postmortem injection studies ofHood'2 are encouraging in demonstrating (1) drainageinto the coronary sinus of 96% of veins >1.0 mm indiameter originating in the left ventricular free walland septum; and (2) only 17% of veins entering thecoronary sinus originating from other structures. Inthe present studies, the coronary sinus catheter was in-serted at least 2 cm beyond the coronary sinus ostium(as seen in the frontal projection) to avoid right atrial


admixture.5 This no doubt minimized the contributionto our measurements of the postero-medial left ven-tricular free wall and the posterior portion of the in-terventricular septum and values for "average" F/Wshould be interpreted with this in mind. It happenedthat our coronary patient group included only one in-dividual with isolated right coronary artery disease.The position of our coronary sinus catheters alsomakes it unlikely that the observed decreases in F/Win coronary artery disease were due to sampling alarger than usual portion of right ventricular drainage.

(C) Techniques Employing Precordial Counting

Although measurements employing coronary sinussampling differ in important respects from thoseemploying precordial counting,9 our reductions inaverage F/W in coronary patients are also at variancewith findings of several groups using bolus injection ofXe'33 into the left coronary artery and a monoex-ponential rate constant derived from a conventionalscintillation crystal. While an area of below-averageF/W has a potentially larger representation in aprecordial desaturation curve than a coronary sinusdesaturation curve,9 the basic requirements of evensaturation and resolution of prolonged tissue-arterialdifferences again apply. As with Kr, bolus injectionresults in distribution of tracer according to flowrather than volume and the monoexponential calcula-tion assumes homogeneity of F/W. Special problemsare also presented by the slower rate of Xe clearancenoted after the first minute or two of desaturation.This appears related to the movement of the highlylipid-soluble tracer into adipose tissue and/or delayedelimination of tracer from alveolar gas spaces includedwithin the field " seen" by the external crystal.Calculations of F/W which include this latter portionof a desaturation curve give values which are patentlyunreasonable and have no doubt influenced the deci-sion of several groups to resort to monoexponentialcalculation. The difficulties in fitting monoexponentialclearance constants to Xe data have been emphasizedby, among others, Zierler'3 and Bassingthwaighte."Sampling of adipose tissue during Xe washout andsimultaneous studies with antipyrine and Xe'4 bothsuggest that even "early" rate constants are affectedby movement of Xe into adipose tissue, and delayedclearance therefrom. These various potentialproblems were pointed out and considered carefullyby Ross and his colleagues in their original descriptionof Xe'33 technique.'5 In addition to a number ofstudies relating rate constants of clearance tovariations of flow in relative terms, they demonstratedagreement between directly measured flow and 14values obtained with Kr85 and Xe'33 in three animals.The validity of measurements employing Kr85 and

555


Dow

nloaded from


KLOCKE ET AL.

precordial counting had also been considered by Herdand his colleagues in 1962.16 A close correlation wasfound between metered left coronary inflow and Kr8'F/W in each of four dogs, although measured inflowwas not converted into F/W because of uncertaintyabout the volume of tissue perfused. Additional com-parisons of Xe"' F/W and directly measured flowhave been reported by Bassingthwaighte et al.'4 andShaw et al.17 However, validation of the most com-monly used Xe methodology is not available in cir-cumstances in which heterogeneity of perfusion hasbeen increased to an abnormal degree. The pitfallsdemonstrated by Shaw et al. using autoradiographictechniques after canine coronary occlusion17emphasize the importance of further validation if theXe technique is to continue to be used in coronary dis-ease. While some of the difficulties related toheterogeneity of F/W can be obviated by the in-genious use of detector systems to examine localizedareas within the heart,'8' 19 the general limitations ofbolus injection and monoexponential extrapolationapply to the localized area. In view of the increasingevidence for transmural variation of flow in abnormalhemodynamic states,20 21 caution seems advisable intranslating Xe rate constants into specific values forF/W even in topographically localized areas. Direc-tionally incorrect conclusions about changes inaverage F/W related to interventions seem at leasttheoretically possible.

(D) Partition Coefficients

Direct measurements of partition coefficients havepreviously been reported for N20 by Kozam et al.22The findings were similar to those reported in table 1and there was verification that N20 solubility did notdiffer significantly in hearts with and without coronaryartery disease. Since we were unable to measure thesolubility of helium in scarred myocardium directly,we have assumed that partition coefficients for He alsodo not differ systematically in patients with andwithout coronary disease. The slope of our N20regression line relating relative blood and salinesolubilities to hematocrit is positive but small and theroutine use of a tissue-blood partition coefficient of0.94 ml/g would seem suitable. The slope of the cor-responding regression line for helium is negative(because He is less soluble in erythrocytes thanplasma), and 0.95 ml/g might be a reasonable valuefor routine use as tissue-blood partition coefficient. Inthe case of Kr, myocardial solubility is not known tohave been measured directly previously. The usualprocedure has been to assume a tissue-blood partitioncoefficient of 1.0, pointing out that this value has beenmeasured directly for brain tissue. The present datasuggest that the solubility of Kr in myocardium is

systematically less than in blood and that tissue-bloodpartition coefficients for Kr vary significantly withhematocrit.

(E) Other ConsiderationsFrom the practical point of view, the greatest

limitation in obtaining a methodologically satisfactorymeasurement of F/W is the time involved (40 minutesafter preliminary positioning of catheters, etc.). Thestandard deviations of directly measured F/Ws intable 2 indicate the degree to which the requirementof "' stationarity"23 was met in the present animalstudies. Comparable informati-on is not available inthe patient studies, although there seems no reason tosuspect a selectively unsteady state in the coronarydisease group. Also of potential concern is the con-tribution to the measured exchange of extracardiacstructures. While the amount of superficial adiposetissue is greater in the human heart than the canineheart, measurements of oil-water solubility ratios inour laboratory indicate that this parameter is 50%greater for N20 than He. This is against the possibilitythat the observed differences between He and N20F/Ws in coronary disease were due to an increasedrepresentation of adipose tissue in the He curves.Since He is inherently more diffusible than N20, thepossibility of a measurable contribution to He datafrom non-adipose structures near the heart cannot beexcluded, although this circumstance would notprovide an explanation for the observed differences inHe F/Ws in patients with and without coronary dis-ease. In any event, the volume of tissue represented inthe present measurements is the volume of distribu-tion of the sampled tracer.

Comparisons of He Values for F/W with N20 andKr Values for F/W

Even though direct measurements of F/W are notpossible in man, it is of interest to compare He, N20and Kr values of F/W in the present human studieswith each other and with previous measurements fromother laboratories.* The absolute values of F/W ob-tained with He and N20 in normal patients agreereasonably with those summarized by Rowe,24 par-ticularly when differences in assumed values formyocardial specific gravity and tissue-blood partitioncoefficients are taken into account. In addition, when

*Average values for F/W reported in this paper differ slightlyfrom those given in preliminary publications. The differences arethe result of (1) the use of a value for myocardial specific gravity of1.05 (instead of 1.00); (2) the calculation of individual tissue-bloodpartition coefficients from the data in table 1; (3) the inclusion of thesmall segment of perfused right ventricular free wall adjacent to theseptum in the calculation of directly measured F/W in the rightheart bypass studies; and (4) the inclusion of a few more cases in thehuman studies.

Circuilation, X oluinie 50, September 1974

5056


Dow

nloaded from



present N20 flows in normal individuals are expressedas percentages of simultaneous He flows (fig. 8), thetwo measurements do not differ significantly. This iscompatible with the thesis that both techniques aresatisfactory in man when flow is not abnormallyheterogeneous. In the presence of coronary artery dis-ease, absolute values for F/W with He averaged 23%lower than values obtained with He in normalpatients, a reduction approximately similar inmagnitude to that recently reported using the sametechnique in acute myocardial infarction by Hodgeset al.25 and in depressed inotropic states by Henry etal.26 Our simultaneous N20 and He flows differedsignificantly in coronary disease, N20 flows exceedingHe flows by an average of 38% (fig. 8). The absoluteN20 levels for F/W are similar to those obtained incoronary disease by others but presumably reflect aninability to include localized areas of reduced F/W inthe measurement of average F/W. Comparisons of Krand He flows were not performed systematically inpatients with arteriographically normal coronaryarteries but, in coronary patients, Kr flows showedeven larger discrepancies from He flows, averaging154 ± 29% of He flows (fig. 8). The reasons for thisdiscrepancy are no doubt the same as described abovefor the animal studies after coronary ligation. Themagnitude of the discrepancy would have been evengreater had the tissue-blood partition coefficient forKr been assumed to be 1.00.

F/W vs "Total" Flow in Man

When initial measurements of F/W did not reflect asystematic reduction in coronary artery disease, it wassuggested that measurements of total flow (ml/min)might be more useful. Measurements of total flow forthe entire heart (or, more specifically, that portion ofthe heart "seen" by four-inch external scintillationcrystals) have been performed by several groups usingRb84 and coincidence counting. Results have con-flicted, the data of some groups showing no differenceat rest between individuals with and without coronarydisease27 and the data of other groups indicating a 25-35% reduction in coronary disease.28 30 Measurementsof total left ventricular flow using a continuous ther-modilution measurement of coronary sinus outflow3'have not shown a difference between individuals withand without coronary disease. The present data forF/W do not resolve the differences among thesevarious measurements of " total" flow. However, asdiscussed below, a reduction in F/W and a normal"'total" flow in coronary disease would be compatibleif left ventricular size were increased systematically.Implications of Present Measurements inCoronary Artery Disease

Since the differences in He F/Ws between patientsCircuilation, Voltume 50, September 1974

with and without coronary disease appear not to be amethodological artifact, we conclude that average leftventricular F/W was decreased at rest in our coronarygroup. These individuals all had histories of anginapectoris and/or myocardial infarction and theirarteriographic abnormalities were rather advanced. Itseems reasonable that groups of individuals with lesssevere disease may have an abnormality of averageF/W only during stress. Attempts were made to cor-relate the present values for F/W with the degree ofarteriographic abnormality, as indicated by the Rowe"arteriographic score."3 A statistically significant cor-relation coefficient (r = 0.39, P < 0.01) could be ob-tained only if the seven patients in the normal groupnot subjected to arteriography were assumed to havenormal coronary vessels. Thus, while there probably isa "weak" correlation between average left ventricularF/W in the cardiac catheterization laboratory and thedegree of arteriographic abnormality, the correlationhas limited usefulness in attempts to derive informa-tion about average F/W from individual arteriograms.More useful correlations might well result if theregional reductions in F/W which cause the diminu-tion in average F/W could be related more specificallyto localized arteriographic abnormalities.

In addition to considering group differences inaverage LV F/W at rest, it is also pertinent to considerthe usefulness, and the limitations, of measurementsof this parameter in individual patients. As illustratedin figure 5, nine of 10 cases in which F/W was 70ml/min/100 g or more were normal individuals, whileall 10 cases in which F/W was below 50 ml/min/100 gwere coronary patients. However, the two groupsshowed considerable "overlap" in the range between50 and 69 ml/min/100 g and it is clear that themeasurement of average LV F/W at rest has limitedsensitivity and specificity in defining the status of thecoronary circulation in individual patients. A betterseparation between "normal" and "abnormal" perfu-sion might well be accomplished by additionalmeasurements during a stressful intervention, or bymeasurements of regional F/W.

Increased heterogeneity of F/W has been il-lustrated in a small series of coronary disease patientsin a previous report from our laboratory usingsemilogarithmic plots of venous-arterial test gasdifferences against time.5 Figure 9 shows a summaryplot for all individuals in the present study. Whilethere is significant overlap among individual curves innormal and coronary patients, venous-arterialdifferences for He are systematically more prolongedin the coronary group, thus again confirming the in-creased heterogeneity of F/W in this condition.Although individual curves are not shown in figure 9,there was a reasonably good separation between nor-

15.57


Dow

nloaded from


KLOCKE ET AL.

00.5 1. .

4 8 12TIME (MIN)

Figure 9

Semilogarithmic plot of average venous-arterial Hetime in normal individuals and coronary patients. Simconstructed for each study in tables 3 and 4 and indtabulated from the smoothed curves at two-minutepoints shown here are the resulting mean values (+1group.

ing myocardial infarction, Naeim et al.32 noted heartweights to average 20-25% more in infarction patientsJOMAL without rupture than in a control group of noninfarc-tion patients. In addition, Scott and Briggs33 reportedcardiac weights exceeding 450 grams in males and 400grams in females in 83 of 183 patients dying suddenlyfrom. coronary atherosclerosis in Albany County, NewYork. A similar incidence of cardiomegaly wasreported by Titus et al.34 in a series of 86 autopsiedcases of sudden unexpected death secondary toischemic heart disease in Rochester, Minnesota.

Finally, the finding of a reduction in average F/Wat rest in arteriographically advanced coronary diseasein no way diminishes the importance of evaluating the

16 ' 20 coronary circulation in localized regions of the leftventricle, and during stress as well as at rest. The ma-jor challenge in measurements during atrial pacing

differences vs and muscular exercise, and in efforts to quantitateilar plots were topographical and transmural variations in flow in!ividual values man, is still the development of methodologicallyintrvals T"he adequate, widely applicable technology.sEmL Jor eacn

mal and coronary patients in the later stages ofdesaturation. At 20 minutes, for example, venous-arterial differences were less than 1% of initial He con-centrations in all normal patients, but greater than 1%in 20 of 26 coronary patients. Attempts to constructmore specific equivalent models of coronary perfusionusing the information in figure 9 are complex and in-volve additional assumptions about diffusionequilibrium, constancy of arterial inert gas concentra-tion, etc., and are beyond the scope of this paper.

Although definitive information concerning thenature of the areas of reduced flow in coronary diseasepatients is not available, the volume of these areasmust be appreciable. Scar tissue related to previousischemic episodes no doubt constitutes at least a por-tion of the "low-flow'" tissue. We cannot exclude thepossibility that the reduction in F/W in coronarypatients is due entirely to scar tissue but were unableto document previous myocardial infarction byclinical, electrocardiographic, enzyme or angio-graphic data in four of the 10 coronary patients inwhom F/W was below 50 ml/min/100 g. Areas ofviable but "underperfused" muscle may also deserveconsideration. Reports of normal values for total flow(ml/min) in coronary patients would be compatiblewith the present data for F/W (ml/min/100 g) if leftventricular size were increased systematically inpatients with coronary disease. Interesting data in thisregard are provided by recent pathological obser-vations. While reviewing cases of cardiac rupture dur-

Acknowledgment

The authors are indebted to Drs. D. T. Arani, A. Drakonakis, H.L. Falsetti, P. Gandel, R. E. Mates, R. A. Oliveros and R. Tandonfor their participation in various phases of these studies, and to thestaffs of the Cardiac Catheterization and Angiology Laboratories ofthe Buffalo General and E. J. Meyer Memorial Hospitals for theirsuperb assistance during the studies in man.

References1. KETY SS, SCHMIDT CF: The determination of cerebral blood

flow in man by the use of nitrous oxide in low concentrations.Am J Physiol 143: 53, 1945

2. ECKENHOFF JE, HAFKENSCHIEL JH, HARMEL MH, GOODALEWT, LLBIN M, BIN-G RJ, KETY SS: Measurement of coronaryblood flow by the nitrous oxide method. Am J Physiol 152:356, 1948

3. RoW'E GG, THOMISEN JH, STENLLND RR, MCKENNA DH, SIALERS, CORLISS RJ: A study of hemodynamics and coronary bloodflow in man with coronary artery disease. Circulation 39:139, 1969

4. GREGG DE, LoNGINo FH, GREEN PA, CZERWONKA LJ: Acomparison of coronary flow determined by the nitrous oxidemethod and by a direct method using the rotameter. Cir-culation 3: 89, 1951

5. KLOCKE FJ, KOBERSTEIN RC, PITTMAN DE, BUNNELL IL,GREENE DG, RoSING DR: Effects of heterogeneous myocar-dial perfusion on coronary venous H2 desaturation curvesand calculations of coronary flow. J Clin Invest 47: 2711,1968

6. WITTENBERG SM, KLOCKE FJ, GREENTE DG, BUNNELL IL,FALSETTI HL, Zizzi JA: Measurement of cardiac output inman with a non-recirculating indicator. J Clin Invest 50:1466, 1971

7. COHEN- LS, ELLIOTT WC, GORLIN R: Measurement ofmyocardial blood flow using Kr85. Am J Physiol 206: 997,1964

8. FARHI LE: Continuous duty tonometer system. J Appl Physiol20: 1098, 1965

Uz 8L cr

LL irLL F-c:] z

Iz0

-ow1

W >z DIzD >

Z

>-)

30-

10-

5.0-

3.0-

1.0

Circutlatiotn Voluime 50, September 1974

558

O = h* = C


Dow

nloaded from



9. KLOCKE FJ, WITTENBERC SM: Heterogeneity of coronary bloodflow in human coronary artery disease and experimentalmyocardial infarction. Am J Cardiol 24: 782, 1969

10. AUKLAND K, BOWER BF, BERLINER RW: Measurement of localblood flow with hydrogen gas. Circ Res 14: 164, 1964

11. GREGG DE, FISHER LC: Blood supply to the heart. In:Handbook of Physiology, Section II: Circulation. Baltimore,Waverly Press, 1963

12. HOOD WB JR: Regional venous drainage of the human heart. BrHeart J 30: 105, 1968

13. ZIERLER KL: Equations for measuring blood flow by externalmonitoring of radioisotopes. Circ Res 16: 309, 1965

14. BASSINGTHWAICHTE JB, STRANDELL T, DONALD DE: Estimationof coronary blood flow by washout of diffusible indicators.Circ Res 23: 259, 1968

15. Ross RS, UEDA K, LICHTLEN PR, REES JR: Measurement ofmyocardial blood flow in animals and man by selective injec-tion of radioactive inert gas into the coronary arteries. CircRes 15: 28, 1964

16. HERD JA, HOLLENBERG M, THORBUR.N GD, KOPALD HH,BARGER AC: Myocardial blood flow determined with krypton85 in unanesthetized dogs. Am J Physiol 203: 122, 1962

17. SHAv DJ, PITT A, FRIESINGER GC: Autoradiographic study ofthe '13Xe disappearance method for measurement of myocar-dial blood flow. Cardiovasc Res 6: 268, 1971

18. CANNON PJ, DELL RB, DWvYER EM JR: Measurement of regionalmyocardial perfusion in man with l1UXe and a scintillationcamera. J Clin Invest 51: 964, 1972

19. CANNON PJ, DELL RB, DWVYER EM JR: Regional myocardialperfusion rates in patients with coronary artery disease. JClin Invest 51: 978, 1972

20. BUCKBERG GD, FIXLER DE, ARCHIE JP, HOFFMIAN JIE:Experimental subendocardial ischemia in dogs with normalcoronary arteries. Circ Res 30: 67, 1972

21. MOIR TW: Subendocardial distribution of coronary blood flowand the effects of antianginal drugs. Circ Res 30: 621, 1972

22. KOZANI RL, LANDAL SM, COBINA JM, LLKAS DS: Solubility ofnitrous oxide in biologic fluid and myocardium. J ApplPhysiol 29: 593, 1970

23. ZIERLER KL: Circulation times and the theory of indicator-dilution methods for determining blood flow and volume. InHandbook of Physiology, Section II: Circulation. Baltimore,Waverly Press, 1963

24. ROSVE GG, CASTILLO CA, AFONSO S, CRLMPTON CW: Coronaryflow measured by the nitrous oxide method. Am Heart J 67:457, 1964

25. HODGES M, AKIYANIA T, BIDDLE TL, ROBERTS DL, WYNAAR B:Coronary blood flow in acute myocardial infarction. Am JCardiol 31: 138, 1973

26. HENIRY PD, ECKBERG D, GAULT JH, Ross J JR: Depressedinotropic state and reduced myocardial oxygen consumptionin the human heart. Am J Cardiol 31: 300, 1973

27. KN'OEBEL SB, McHENRY PL, PHILLIPS JF, PAULETTO FJ:Coronary collateral circulation and myocardial blood flowreserve. Circulation 46: 84, 1972

28. DONATO L, BARTOLONIEI G, FEDrERIGHI G, TORREGGIANI G:Measurement of coronary blood flow by external countingwith radioactive rubidium: critical appraisal and validationof the method. Circulation 33: 708, 1966

29. COwVAN C, DURAN PVM, CORSINI G, GOLDSCHLAGER N, BING RJ:The effects of nitroglycerin on myocardial blood flow in man.Am J Cardiol 24: 154, 1969

30. MYNIIN D, SHARNIA GP: Total and effective coronary blood flowin coronary and noncoronary heart disease. J Clin Invest 53:363, 1974

31. GANZ W, TAML-RA K, MARCuS HS, DONoso R, YOSHIDA S, SWANHJC: Measurement of coronary sinus blood flow by con-tinuous thermodilution in man. Circulation 44: 181, 1971

32. NAEIM F, DE LA MAZA LM, ROBBINS SL: Cardiac rupture duringmyocardial infarction. Circulation 45: 1231, 1972

33. SCOTT RF, BRIGGS TS: Pathological findings in pre-hospitaldeaths due to coronary atherosclerosis. Am J Cardiol 29: 782,1972

34. TITUS JL, OXMIAN HA, NOBREGA FT, CONNOLLY DC: Suddenunexpected deaths as the initial manifestation of ischemicheart disease. Clinical and pathologic observations. Am JCardiol 26: 662, 1970


559


Dow

nloaded from


WITTENBERG and JOHN P. VISCOFRANCIS J. KLOCKE, IVAN L. BUNNELL, DAVID GREENE, STEPHEN M.

Without Coronary Artery DiseaseAverage Coronary Blood Flow Per Unit Weight of Left Ventricle in Patients With and

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 1974 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/01.CIR.50.3.547

1974;50:547-559Circulation.

http://circ.ahajournals.org/content/50/3/547the World Wide Web at:

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

http://circ.ahajournals.org//subscriptions/

is online at: Circulation 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 Further information about this process is available in therequested is located, click Request Permissions in the middle column of the Web page under Services.the Editorial Office. Once the online version of the published article for which permission is being

can be obtained via RightsLink, a service of the Copyright Clearance Center, notCirculationpublished in Requests for permissions to reproduce figures, tables, or portions of articles originallyPermissions:


Dow

nloaded from

http://circ.ahajournals.org/content/50/3/547

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://circ.ahajournals.org//subscriptions/


average coronary blood flow per unit weight left...

Documents