Indocyanine Green Clearance and Estimated Hepatic Blood

Journal of Clinical InvestigationVol. 43, No. 8, 1964

Indocyanine Green Clearance and Estimated Hepatic BloodFlow during Mild to Maximal Exercise in Upright

Man *

LORING B. ROWELL, JOHN R. BLACKMON,AND ROBERTA. BRUCE(From the Department of Medicine, Division of Cardiology, University of Washington School

of Medicine, Seattle, Wash.)

Krogh in 1912 (1) postulated the necessity fora large capacity venous reservoir capable ofquickly delivering an autotransfusion of bloodinto the right heart during exercise, particularlyin upright man, where Krogh felt venous returnmay at times be inadequate. Bock and co-workers(2) in their classical work reasoned that increasedblood flow to working muscle, related in part tolocal vasodilation, must be partially effected by de-creased splanchnic blood flow. Since the splanch-nic region receives 20 to 25% of the total left ven-tricular output but extracts only 10 to 25% of theavailable oxygen (3), the splanchnic bed is ideallysuited for rapid correction for any residual in-balance between left ventricular output and periph-eral distribution of flow.

The experimental evaluation of the influence ofexercise on splanchnic blood flow, however, hasproduced conflicting results. Herrick, Grindlay,Baldes, and Mann (4), and more recently Rush-mer, Franklin, Van Citters, and Smith (5) did notfind a decrease in splanchnic blood flow in dogsundergoing an intensity of exercise far belowtheir maximal capacities. Also plethysmographicstudies in man by Sj6strand (6) led him to con-clude that intrathoracic rather than splanchnicblood depots are functionally important in exer-cise.

In contrast, a study by Bradley (7) and byWade and co-workers (8), utilizing the sulfo-bromophthalein sodium (BSP) extraction method,demonstrated that significant decreases in splanch-

* Submitted for publication January 27, 1964; acceptedApril 23, 1964.

Supported in part by grant-in-aid HE 908C12 fromthe U. S. Public Health Service National Heart In-stitute and also a grant from the Washington State HeartAssociation. Presented in part at the 17th Annual Meet-ing of the Western Society for Clinical Research, Janu-ary 1964, Carmel, Calif.

nic blood flow do occur in man during mild supineexertion. Although Lowenthal, Harpuder, andBlatt (9) did not obtain evidence for diminishedsplanchnic blood flow during exercise with theBSP extraction method, the method as applied bythese workers was too insensitive to detect any butvery large changes. Further evidence regardingsplanchnic blood flow during exercise comes frommeasurements of hepatic arteriovenous (a-v)oxygen difference in normal subjects (10) and incardiac patients (11). These studies indicate thatconsiderable decreases in splanchnic blood flowoccur during moderately heavy exercise in thesupine posture particularly in cardiac patientswith a severely impaired cardiac output responseto exercise. The purpose of the present studywas to explore the applicability of a simple and in-direct means of assessing changes in hepatic bloodflow with exertion and to assess the extremes ofthese changes during maximal or near maximalexertion in upright man. Heretofore investiga-tions of splanchnic blood flow in exercising manhave been carried out using mild to moderate exer-cise in recumbency; however, cardiovascular re-sponses during exercise in this posture are funda-mentally different from those in upright posture(12). Indocyanine green dye (ICG) was usedsince it is extracted exclusively by the liver (13-15). Changes in hepatic blood flow in response toexercise were estimated from changes over rest-ing values of plasma clearance rate of ICG aftera single injection. Evidence regarding the validityof these estimates was obtained by measuring si-multaneously ICG clearance rate and hepatic ex-traction efficiency for ICG during rest and uprightexercise.

MethodsThe plasma clearance rate of ICG was studied in 10

normal men 19 to 37 years of age with a mean age of

1677

.L. B. ROWVELL, J. R. B3LACKMON,AND R. A. BRUCE

26 years. Both hepatic extraction and plasma clearanceof ICG were determined in five additional normal sub-jects (17 to 25 years old) and in two cardiac patients,17 and 25 years old. The two patients were class I,asymptomatic cardiacs who had marked aortic regurgi-tation. Each exhibited a normal response to the exercisecapacity test of Bruce, Blackmon, Jones, and Strait (16).Hepatic circulation was considered to be normal fromclinical criteria [an assumption borne out later by nor-mal resting estimated hepatic blood flow (EHBF) andextraction of ICG]. Two additional subj ects were de-leted from this study because of problems with hepaticvenous sampling.

The subjects exercised on a motor-driven treadmillin an air conditioned room maintained at 780 ± 2° Fand a relative humidity of 55 to 65%. Several days be-fore an experiment the subjects worked at each in-tensity (and duration) to be employed in the study.Oxygen intake and heart rate were determined at thistime, thus familiarizing the subject with these details ofthe procedure. The maximal oxygen intake was deter-mined by the procedure of Taylor, Buskirk, and Henschel(17).

Plasma disappearance of ICG. All experiments werecarried out in the morning with the subjects allowed onlyfruit juice and black coffee at breakfast. Two 50-cm,Lehman no. 5, thin-walled catheters were introducedpercutaneously or by venous cut-downs as necessary atthe antecubital fossa of one arm. The tip of the in-jection catheter rested in the cephalic vein above themid-portion of the upper arm. The withdrawal catheterextended via the median cubital vein into the axilla.

The subject was seated in a chair, and 10 minutes aftera priming dose of ICG, sufficient to fill the catheter, a15-ml sample of blood to be used as a zero-absorbancyreference (blank) and for standard curves was taken.Immediately afterward 12.5 mg of ICG was injectedfrom a calibrated 5-ml syringe. Thereafter, 3-ml sam-ples of blood were taken at precisely timed intervals for12 minutes after an initial 3- to 4-minute delay.

Ten minutes later the subject was positioned on thetreadmill with the catheterized arm resting on an arm-board to minimize support. Heart rate was monitoredby electrocardiogram, and expired air was collectedthrough a low resistance Collins "triple j valve" (re-sistance, 2 cm H20 at 200 L per minute) into a balanced350 L Collins spirometer connected to the valve by wide-lumen (11 inch) pipe and rubber tubing. Even at maxi-mal ventilations [150 L per minute, standard temperature,pressure, dry (STPD) ] it was felt that resistance re-mained low enough to minimize increases in intrathoracicpressure above those normally occurring at such highventilations. Expired gas samples were analyzed by theHaldane technique or the Scholander microtechnique, orboth, with duplicates required to check within 0.05%for both oxygen and carbon dioxide.

Except at near maximal exercise, the subject walkedon the treadmill at 3.5 miles per hour (mph) at variousgrades for 44 minutes at which time the blood sample

to serve as spectrophotometric blank was obtained. Atthe fifth minute of exercise 12.5 mg of ICG was in-jected, and 3-ml blood samples were withdrawn every 2minutes thereafter for 10 minutes. After exercise thesubject rested 20 to 25 minutes, and the same procedurewas repeated at a higher level of exercise. Two menrepeated the same level. At higher levels of exercise,where subjects ran at 7 mph on appropriate grades,which were exhausting in less than 15 minutes, the sampleof blood to be used as a spectrophotometric blank wastaken after 1 to 14 minutes of exercise. Dye was thenimmediately injected, after which additional sampleswere taken every 1 or 2 minutes depending upon theanticipated duration of exercise.

Blood loss usually ranged from 90 to 120 ml during asingle experiment, and the largest single loss was 220 ml.The total volume of isotonic saline injected ranged from100 to 150 ml. Initially the studies required that 6-mlsamples be taken so that a third level of exercise was notstudied in those subjects. The men were given water adlibitum; no weight loss was observed during the ex-periment. No subject was studied again within 3 weeksafter an experiment.

After withdrawal into heparinized syringes, bloodsamples were centrifuged at 1,000 X g for 30 minutes andafterwards analyzed spectrophotometrically in a Beck-man DU spectrophotometer at 805 mg. Lipemic serumfrom two subjects who ate breakfast was cleared byultracentrifugation (1,000,000 g-minutes) to obtain clearserum below the suspended chylomicrons. A known dyeconcentration of approximately 3 mg per l was prc-pared in a samnple of plasmia, and from it successive di-lutions of 1.5 and 0.75 mg per L. This was done foreach subject, using a sample of dye from the injectionreservoir that contained approximately 75 mg of dye in30 ml of solvent. The standards were read against thesubject's plasma blank, and each blank was read againstwater.

Estimated hepatic blood flow (EHBF). Estimates ofhepatic blood flow were obtained by the single-injectionmethod described by Wiegand, Ketterer, and Rapaport(18) and Caesar and co-workers (15). A 120-cm no. 6,venous Cournand catheter was wedged into a right he-patic vein under fluoroscopic guidance. The catheter wasretracted just sufficiently to allow free withdrawal ofblood and yet remain well into the hepatic vein (3 to 5cm). Care was taken to see that the catheter tip re-mained in the hepatic vein during maximal respiratoryexcursions. When the catheter was in a wedged or nearlywedged position, a 2- to 3-ml injection of saline producedright upper quadrant discomfort in all subjects; this pro-cedure proved useful in determining the position of thecatheter tip. A Cournand needle was placed in the radialartery, and an ICG injection catheter was introduced asdescribed above. All catheters were in the same arm.

The studies were carried out exactly as those outlinedabove with two exceptions: resting measurements wereconducted with all but two of the subjects supine; theinjected dye dose was 25 mg in subjects JA and GJ and

t678

HEPATIC BLOODFLOWDURING SEVEREEXERCISE IN UPRIGHT MAN

31.3 mg in the remaining subjects. In no case was theexact quantity of dye injected known, since the weightof dye in several vials was found to vary from 25 to 29mg. However, only differences in absorbency betweenthe two sampling sites and their changes with time arerequired for use of this method. Changes in plasma vol-ume could be roughly estimated from the slope of ICGconcentration extrapolated to zero time when variations involume or concentration of injected dose were knownnot to occur.

Hepatic venous samples were half-drawn 10 to 15 sec-onds after the mid-point in volume of arterial sampling.This was done to correct for the transit time of bloodacross the hepatic sinusoids. At the end of exercise theposition of the catheter tip was again viewed by fluoros-copy as a second check on its position in the hepatic vein.

Hepatic a-v oxygen difference. Blood samples weretaken simultaneously from the radial artery and hepaticvein immediately after the final resting and exercise ICGclearance sample. The oxygen content was determinedby the manometric method of Van Slyke and Neill.

Calculations. Plasma ICG concentrations were plottedagainst time on a semilogarithmic scale from which thehalf-life (ti) of ICG in plasma and the disappearancerate constant or fractional clearance rate (K) were de-rived. The constant K was calculated from loge2/ti de-rived from the more familiar form log.C1 - log.C2/tl -t2,when tl - t2 tj and represents the fraction of the totalquantity of injected dye that disappears from the plasmain 1 minute, or alternatively, the fraction of the plasmavolume cleared in 1 minute. The percentage of the rest-ing value for ICG clearance is most simply expressed as(tj,/tj.. X 100), where ties and tor represent exerciseand resting values, respectively, for ICG half-life.

The arterial-hepatic venous plasma extraction ratio(E) was calculated from the form Ao - AObh,/Ao.,where Ao. and Aohv are arterial and hepatic venous ab-sorbancy (or concentration, since either may be used)extrapolated to zero time. Plasma clearance (millilitersper minute) was calculated from the disappearance rateconstant (K) times estimated plasma volume, or 45 ml perkg times body weight, the 45 ml per kg being selectedfrom several sources (19). The EHPF was calculatedfrom plasma clearance/E and the EHBFfrom EHPF/1-hematocrit.

The percentage of the resting value for hepatic bloodflow was also approximated from the following relation-ship: (av,/av.. X 100), where av, and av.. are thehepatic a-v oxygen difference during rest and during ex-ercise, respectively.

Results

Fractional clearance of ICG. The restingplasma clearance rate of ICG was in close agree-ment with the findings of others (14, 15, 18, 20).The mean tj was 3.0 minutes (range, 2.3 to 3.7minutes), and the fractional clearance rate (K)

0.1-o0.08- Vt2 3.0 min

. 0.06- \ 00%

0.04-

0.03-

0.02-

0.01~0 2 4 6 8 10 12 14

Minutes

FIG. 1. RELATIONSHIP OF FRACTIONAL CLEARANCERATEOF INDOCYANINE GREEN (ICG) FROMPERIPHERAL VENOUSBLOOD TO WORKINTENSITY. The change in absorbencywith time of ICG in plasma is shown for subject DHat rest and during treadmill exercise at 3.5 miles perhour (mph) on grades of 10, 15, and 174%. The half-life (tI) of each slope is specified in minutes and per-centage of the resting tf, which was taken as 100%v.ICG was injected at the 5th minute of exercise; timevalues on the abscissa are relative to the time of injec-tion of ICG (at zero time).

was 23.5% per minute (range, 30.1 to 18.7% perminute). During exercise of various intensities(requiring from 26 to 97%o of maximal oxygen in-take) peripheral venous concentration again de-creased linearly when plotted on a semilogarithmicscale against time (Figure 1), but the t1 was pro-longed over the resting values. The data fromthe clearance studies on 10 men are presented inTable I.

The distribution of fractional clearance ratesof ICG expressed as percentage of resting values(taking rest as 100%o) is plotted against oxygenintake as percentage of maximal oxygen intake(Figure 2). The values represent a total of 47observations from 10 subjects (Table I). Theregression line was fitted by the method of leastsquares from the regression equation y = - 1.16x+ 127.89, where y is the fractional clearance ofICG as percentage of the resting value and x is

1679

B. ROWELL, J. R. BLACKMON,AND R. A. BRUCE

the oxygen intake expressed as percentage ofmaximal oxygen intake. The slope of the line is- 1.16 and the intercept, 127.89. The correlationcoefficient is - 0.89 and the SE of estimate is9.9.

The prolongation of ICG clearance by the liveris inversely correlated with relative total oxygen

intake expressed as the fraction of the maximaloxygen intake of which the individual is capable.However, when the ICG clearance is plottedagainst the absolute oxygen intake (milliliters per

kilogram per minute), the correlation is dimin-ished (r = - 0.77) (Figure 3) and the SE ofestimate increased to 13.6. The regression equa-

tion is y = - 1.81 x + 113.18, with the exceptionthat x in this equation represents oxygen intakein milliliters per kilogram per minute.

In two instances, where work required 76 and97% of maximal oxygen intake, the fractional

clearance of ICG was only 19% of the restingvalue, the largest decrement in clearance observed.

The first study on JA followed by only 3 hoursa very large, high-fat breakfast, whereas the sec-

ond study adhered to the regular protocol. Nolarge differences are noted in ICG clearance rateswhen the two studies are compared, therefore sug-

gesting that the presence of food in the digestivetract does not influence hepatic clearance of ICG.

The agreement found in repeat determinationswas usually close (Table I). Duplicate tests were

carried out on JB on two separate occasions (Ta-ble I). The four tests, all at 3.5 mph on a 10%

grade, showed ICG clearances that varied from52 to 46% of the resting value with three at 46%.Subject LJ was tested on three separate occasions.Despite repeated commands to this man, he leanedheavily on the support for the catheterized arm,

thereby producing inadvertently a series of tests

BLE I

Plasma, clearance of indocyanine green during rest and mild to maximal exercise in normal upright men

02 intake, ICG%of Heart clearance, Maximal

Subject Age Wt 02 intake maximal rate t K %of resting 02 intake

yrs kg mi/'kg/mzill %7X0 beats/m'?i lMistt min'-1 %X0 nl/kg/minJ. B. 33 84 Rest 3.1 0.224 41

31.7 77 165 6.2 0.112 5032.2 78 164 6.8 0.102 46Rest* 3.1 0.22414.6 36 3.4 0.204 9132.5 79 6.7 0.103 4633.0 80 6.7 0.103 46

J. L. 19 95.7 Rest 2.5 0.277 4328.9 67 188 5.6 0.124 4231.3 73 196 8.2 0.084 35Rest* 2.8 0.24830.3 70 195 7.4 0.094 3830.7 72 198 8.2 0.084 3433.8 79 202 9.0 0.077 31Rest* 3.1 0.22415.4 36 125 4.0 0.173 7823.6 54 152 4.0 0.173 7831.1 72 190 6.1 0.114 51

J. A. 23 74 Rest 3.0 0.231 5830.8 53 126 3.8 0.182 7940.0 69 147 4.8 0.144 62Rest* 3.4 0.20414.8 26 83 3.4 0.204 10030.2 52 130 4.5 0.154 7637.4 65 143 5.6 0.124 58

M. S. 21 68.6 Rest 2.3 0.301 5529.2 53 142 3.5 0.198 66Rest* 2.8 0.24829.4 53 142 4.0 0.173 7037.2 68 168 6.2 0.112 45

D. H. 21 78 Rest 3.0 0.231 5731.0 55 148 5.0 0.139 60

1680


TABLE i-Continued

02 intake, ICG%of Heart clearance, Maximal

Subject Age Wt 02 intake maximal rate tj K %of resting 02 intake

yrs kg ml/kg/min % beags/min min min % ml/kg/min31.2 55 152 5.2 0.133 5839.5 72 175 9.8 0.071 31Rest* 3.0 0.23129.5 52 148 6.1 0.114 4940.0 70 181 11.3 0.061 2743.2 76 187 16.0 0.043 19

L. R. 33 77.6 Rest 3.0 0.231 5827.0 47 128 3.2 0.216 9334.8 60 157 5.5 0.126 5552.0 90 188 11.0 0.063 27Rest* 2.7 0.25728.9 49 140 4.1 0.169 6638.4 65 166 7.2 0.096 3857.0 97 188 14.0 0.050 19Rest* 3.3 0.21027.0 47 136 3.7 0.187 9134.7 60 162 5.3 0.131 6352.0 90 188 9.6 0.072 35

W. D. 24 77 Rest 3.7 0.187 5030.2 60 162 6.8 0.102 5438.0 76 186 9.6 0.072 39Rest* 3.0 0.23114.7 29 120 3.5 0.198 8629.3 58 164 5.1 0.136 5936.5 73 184 8.5 0.082 35

L. J. 23 78 Rest 2.8 0.248 4630.6 66 175 5.2 0.133 5437.1 81 193 11.0 0.063 25

P. W. 22 74.6 Rest 2.7 0.257 6530.1 46 115 3.6 0.192 7538.5 60 147 4.7 0.147 58

R. L. 37 86.1 Rest 3.0 0.231 4012.3 30 94 3.4 0.204 8825.6 63 146 4.6 0.150 65

* Repeat studies 3 to 4 weeks after the previous study.

that required approximately the same amount ofoxygen despite the increases in percentage ofgrade of the treadmill. With the exception of thefinal value, as listed in Table I, the clearance slopesare a fairly sensitive indicator of the severity ofexercise. Although subject LR was studied onthree occasions, the second study produced clear-ance rates that were markedly different fromthose obtained in the first and third studies (TableI). It is not understood why these discrepanciesoccurred. It is only known that this subject "feltdifferent" during the second study.

Volume distribution of ICG. The volume dis-tribution of ICG in 17 men averaged 49.5 ml perkg body weight (range, 35 to 62 ml per kg). Theaverage change in the volume distribution of ICGwith exercise on the treadmill set at 3.5 mph and

a 10% grade was 12.6% (the range for 14 menwas 0 to 27%o).

Estimated hepatic blood flow (EHBF). Esti-mates of hepatic blood flow demonstrated that thedecrease in ICG clearance (Figure 1) observedduring exercise is primarily the result of de-creased EHBF and not a decrease in extractionratio. The values from seven men are presentedin Table II. The curves and calculations for asingle subject at rest and during exercise areshown in Figure 4.

The mean resting EHBFwas 1,614 ml per min-ute, whereas during exercise of various intensi-ties, EHBF fell to values ranging from 820 to390 ml per minute. The average extraction ratioat rest was 0.77 (range, 0.47 to 0.94 for 7 ob-servations), whereas the average value during ex-

1681

L. B. ROWELL, J. R. BLACKMON,AND R. A. BRUCE

100- \

r =-0.89

~~~~80~~~~~~~~~y =- 1. I6x + 1 27.89

, 60 *

CZ

~40

~20-Azo

020 40 60 80 100Oxygen intake- percent of maximal

FIG. 2. CORRELATION BETWEEN FRACTIONAL CLEARANCERATE OF ICG AS PERCENTAGEOF THE RESTING VALUE(100%) AND OXYGENINTAKE AS PERCENTAGEOF MAXIMALOXYGENINTAKE. About the least squares regression lineare a) the 95%c confidence interval (shaded area) forthe true mean value of y (ICG clearance) for a givenvalue of r (oxygen intake) from the regression equationy = - 1.16 x + 127.89 and b) the wider 95% confidenceinterval for the true y-value of an individual having thegiven x-value, indicated by the (lashed lines.

ercise of various intensities was 0.84 (range, 0.70to 0.95 for 12 observations). In three subjectssmall decreases in the extraction ratio of ICGwere observed during exercise, and in one of thesemen (MM) a 22% fall in extraction ratio wasnoted at the highest level of exercise studied (20%grade, 3.5 mph). Due to the decrease in plasmavolume that occurs during exercise, percentagechanges in EHBF still agreed closely with thepercentage changes in ICG clearance rate in thosewhere extraction ratio did decrease. As expected,in the five subjects who showed increased extrac-tion of ICG during exercise, the estimates of thedecreases in hepatic blood flow using only thechange in the peripheral venous disappearance rateof ICG underestimated the measured decreasess inElIBF.

The average difference between a.) percentagechange in the fractional clearance rate of ICGand b) percentage change in EHBFwith exercisewas 8.2% (range, 20 to - 5o) with b beinggreater than a. There was, therefore, a tendencyfor the peripheral sampling method to slightly

underestimate the extent to which EHBF de-creased during exercise.

Hepatic a-v oxygen differencc. In the five sub-jects used for measurements of arterial andhepatic venous oxygen content, a marked wideningof the hepatic a-v oxygen difference occurred dur-ing exercise. The percentage change in hepaticblood flow estimated from avr/ave, x 100 closelyparalleled the percentage change in ICG clearancerate and EHBF. The mean difference betweenpercentage change from rest to exercise in ICGclearance slope and the percentage change in he-patic a-v oxygen difference was 8.6%o with thelatter being higher (range, 19 to - 1%). Themean difference between a) the percentage changefrom rest to exercise in EHBFand b) percentagechange in hepatic a-v oxygen difference from restto exercise was 2% (range, 10 to - 12%), thechange in a being greater than b. Thus, the changein EHBFwith exercise was also very closely par-alleled by changes in hepatic a-v oxygen differ-ence. In the case of RC, whose EHBFdecreasedto 16% of the resting value during exercise, whichwas severe for this subject, the hepatic venousoxygen content fell from 16.00 ml per 100 ml atrest to 0.99 ml per 100 ml during exercise.

T'he estimated mean splanchnic oxygen uptakeat rest (Table II) was 68 ml per minute (range,104 to 41 ml per minute), and during exercisethe mean value was 69 ml per minute (range, 87to 44 ml per minute).

The product of arterial oxygen content duringexercise and the volume of blood shunted awayfrom the hepatic bed (i.e., resting less exerciseEHBF) provides an estimate of increased oxy-gen transport to other regions. The values (Ta-ble II) ranged from 128 to 429 ml oxygen perminute.

Discussion

In an excellent discussion of the concept of he-patic clearance, IFauivert (21 ) summarizes thebasic assulmll)tiolns inherent in clearance techniques.MNfost inlm)ortant is the nature of the substance tobe cleared, since the measurement of organ clear-ance, i.e., the virtual volume of plasma completelycleared of that substance in 1 minute, is valid onlywhen that substance is cleared from the bloodexclusively by that organ. Numerous publications

1682

HIEPATIC BLOODFLOWDURING SEVEREEXERCISE IN UPRIGHT MAN

Oxygen intake ml/kg min1

FIG. 3. CORRELATION BETWEEN FRACTIONAL CLEARANCE RATE OFICG AS PERCENTAGEOF THE RESTING VALUE (100%) AND OXYGENIN-TAKE IN MILLILITERS PER KILOGRAM PER MINUTE. About the leastsquares regression line is shown the 95% confidence interval (dashedlines) for the true value for y (ICG clearance) for an individualhaving the given values of x (oxygen intake) from the regressionequation y = - 1.81 x + 113.18.

have shown that in dogs (13, 22), in humans (13-15), and in other animals (13) ICG, unlike BSP,is cleared exclusively by the liver. Indeed, theperipheral clearance of ICG has been shown tobe a sensitive and reliable index of hepatic func-tion in resting man (14, 15, 18, 20).

During exercise, however, there are possibleextrahepatic routes of removal of ICG. The dyecombines immediately with serum albumin, whichpasses rapidly with other proteins through capil-lary walls during exercise (23). However, ac-cording to De Lanne, Barnes, and Brouha (24)the filtered albumin is rapidly returned to the cir-culation because of accelerated lymph circulationduring exercise, precluding any net loss of albu-min from the blood. The result is an extrahepaticcirculation of ICG that is probably quantitativelyvery small. To offset the initial changes in albu-min concentration, in plasma volume, and inplasma optical density (25), a period of 41 min-utes after the start of exercise and before thepre-injection plasma blank was drawn was pro-vided in the experimental design. The major vol-

ume shifts occur within a 5-minute period duringmoderate to heavy exertion (26, 27). The upper-most slope in Figure 1 (17.5% grade, 3.5 mph)indicates an initial nonlinearity of the ICG disap-pearance slope if injection or sampling, or both,are started too soon after the onset of exercise.The fact that the disappearance of ICG followsfirst-order kinetics, after the initial alterations involume and concentration, argues against the pos-sibility of either significant extrahepatic circula-tion or of changing hepatic flow with time underthese experimental conditions.

Although no ICG is lost into urine during rest-ing measurements (14-16), during exercise con-siderable quantities of albumin may be lost inurine (28). In checking urine samples of twosubjects at the end of a study, no evidence of thedye was found in urine that was made up to a 5 %solution with human serum albumin to stabilizeany ICG (14, 15).

Although the large and consistent changes dur-ing exercise indicate a marked impairment of he-patic capacity to remove ICG from blood, there is

1683

1684 I. B. ROWELL, J. R. BLACKMON,AND R. A. BRUCE

.4J 0- en It 00 VcOC r:10 ~~~~~(1') OV if ) '-4 CIu -4o0 - 0oN-o

zcd Ln~~~-4Of Of) t. VI14 00r 'o1c3>.f,-W0> 't~'jI) 0- 0 j.O) 0- -- ~ 0

O02~~~~~~~~~~~~~~~~~~~~~~I -t t- 00 \ +Xo

0) Cl)~~~~~~0 -

00q Ccq - C- \ - ,~w~~~~~~~~ 0t = _ 0-0 V e sos

~~~0 4 oa rI 0=

*Ws~~~~~~C~ V '1 f)' 'f 0 c

X) OC)00r-X0.1 -!LoC _

e ~ ~~~~~~~~--' \1

m c ', 00Ic tn t-t nI t- CNIt 'I It r

0) 0

ci r ICI ciCOO0 ' -Oif C3i Ci l *t) 0m

ctl Q CID°'JU)

l

9 I

CN\0

tn

di ds0s0N0 m 0t 0-to0 m00t C- t-0 C. t- t-

--Is21 0Xo~ t cO O c dk Wuoco o~' CV10, ,C °_/c

¢ v S o o o o o o o o4 r o6o, o o6o io o o oo

4 .1

0 0; t 0

coroeo-o c-cc00 \Cic\0 00 "C-t,

.S ~ ~ ~ ~ i,~-co4- o~-0c

0~~~~~~~~~~~~~~~~~~~~~

t~~~~~~~~~~~~~~~~t It -! tnm c It t t- '1 v,z &> 4Oo o ,o_

0~~~~~~~~~~~~~~~~~~~~~

.0 4Lf) 000000-=us0 0 0 v'\ -4V) ) CI\ C

C~~ 06#I o-4 '6 r '~o6 OcOldt-~

-U)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-

> Y > an ane tente~~~~++ n Ln n e . :

0 I--V,-.0 c.V V V0

*~~~~~~~~~~~~~~~~~~~~~~~~~~~~~' t6o;eC;t.3 'uUo z

X) -a-~4 N-4-- -04~~~~~~~~~~~~~4J + - - - -40~~~0U)ci V'i~~~~~~~~~00 1)L C)i? C)0 v4--=i) n

0 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~e4-~~~~~~~.0U)~~~~~~~~~0\00 if~~~~~~~~~~~~~i?) - 00V 0~~~~~~~~~~uoe > 6 o6 o ~ r 6o. o o ..

0) L,0-ti)O10 Of) Ci81,Of) 4

-4~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U) 0 -) 4

(n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~+ bO 0i2

ci)O)4-J


Half life-t1,,2.7

Fractional cleronce 'slope' Ka log.2

K = Quffa 0.253 min.-'

Plasma clearance K x est. plasma volumes

PC = 0.253 min- x 2.48L 0.656L min:'

Minutes

FIG. 4. PERIPHERAL ARTERIAL AND HEPATIC VENOUSPLASMA CLEARANCE

OF ICG DURING REST AND EXERCISE. The change in absorbency with timeof ICG in plasma from the two sampling sites is shown for subject EFseated at rest (solid lines) and during treadmill exercise at 3.5 mph on a10% grade (dashed lines). The method of calculating estimated hepaticblood flow (EHBF) is shown. Plasma volume was taken as 45 ml timesweight in kg. During exercise EHBF and extraction efficiency for ICGchanged from 1,515 ml per minute and 71% (at rest) to 765 ml per minuteand 77%, respectively.

no way of differentiating from these data anychanges due to hemodynamic factors from thosedue to altered hepatic cellular function. By defi-nition, clearance of a substance from any organdepends upon blood flow through the organ andthe removal efficiency (extraction ratio) for thatsubstance.

There are reasons for assuming that the he-patic extraction efficiency for a substance such asICG would increase during severe upright exer-cise, since alterations such as abdominal com-pression (29, 30), orthostasis (30, 31), syncope(31), and mild supine exercise (7, 8), which de-crease hepatic blood flow, increase the extractionefficiency for BSP. This is logical, since blood

remains for a longer period in the hepatic sinus-oids. Brauer (32) has shown in the isolated ratliver that BSP extraction efficiency bears an in-verse relationship to tissue perfusion rate. Evengeneralized hepatic ischemia has little influenceon BSP extraction efficiency (29, 32). It isreasonable to assume that ICG and BSP arehandled similarly, for despite their removal bydifferent hepatic mechanisms (14) the two sub-stances are cleared and extracted similarly withrespect to time in the same individual with im-paired hepatic function (14, 15).

Experimentally, however, extraction efficiencyis difficult to measure accurately. The majordifficulty is one of obtaining hepatic venous sam-

1685

L. B. ROWELL, J. R. BLACKMON,ANI) R. A. BRUCE

Mminutes

FIG. 5. PERIPHERAL ARTERIAL AND HEPATIC VENOUS

PLASMA CLEARANCEOF ICG DURING REST AND EXERCISE.

The change in absorbency with time of ICG in plasma isshown for subject RC lying at rest (solid lines) andduring treadmill exercise at 3.5 mph on a 10% grade(dashed lines). Irregularities in hepatic venous levelsof ICG during rest and exercise probably result froman excessively deep hepatic venous wedge (see text).Numerals about the lowermost dashed line represent ex-

traction efficiency at each point during exercise.

ples. Since "hepatic venous blood" does not existas a pooled volume representing the mixed con-

tents of all lobes of the liver, but rather as sepa-

rate pools released into their own venous systems,truly representative samples cannot be obtained.Bradley, Ingelfinger, Bradley, and Curry (33)have shown that the variation in BSP extractionefficiency at different locations in the same liverwas as great as 20% in two cases. Another im-portant problem is the variation in hepatic venous

pressure and instantaneous flow introduced byrespiration. Bradley and co-workers (33) notedreflux of caval blood during heavy breathing, andthey stressed the necessity of quiet breathing dur-ing the sampling period. Brauer (32) reportsrecent cineangiographic evidence that during quietrespiration, hepatic venous outflow is arrestedduring inspiration, and is maximal just before

peak expiratory excursion of the diaphragm. In-(leed, even regurgitation of vena caval blood intohel)atic veins occurs in some animals during in-spirat ioi (32), indicating a positive pressuregradient from the cava to hepatic vein at this time.Also position of the catheter is critical. Althoughretraction of the catheter tip increases greatlythe risk of caval reflux into the catheter, advance-ment of the catheter to wedge position in a hepaticvein results in increased oxygen content of with-drawxn blood (32). In this situation deflection ofportal flow to adjacent regions results, as evi-denced lby detection of substances injected into thehepatic artery and failure to detect all but minutequantities injected into the portal route (32).

Regardless of whether the indicator is ICG oroxygen, the extraction efficiency is likely to be un-derestimated either when the catheter is too deeplywedged, or when it is too close to the caval orificeof the vein. The irregularity of hepatic venousconcentration of ICG seen in several of our ex-periments (Figure 5) may have been the resultof one or several of the problems stated above.The results from RC are a case in point. Thissubject showed an extraction efficiency for ICGwhile at rest that was very low, 47% (Figure 5).If a line parallel to the resting arterial slope weredrawn through the first and final points whereextraction at rest was greater, efficiency wouldincrease from 47 to 72%. However, the fourpoints through which the slope was drawn showthe same ICG concentration-time relationship thatwas found in arterial blood. This would make arandom dilution from caval blood seem unlikely.The variation in extraction efficiency during restand exercise might be accounted for by an exces-sively deep hepatic venous wedge, since samplingfrom this individual was frequently difficult. De-spite this variation, however, there is little ques-tion that hepatic extraction of ICG was higherduring exercise; (extraction efficiency rangedfrom 88 to 94% and averaged 92%o). A similarproblem, although less marked, occurred duringthe first exercise study with LP and was resolvedin the same manner. After slight retraction of thecatheter, a repeat exercise study with the samework intensity provided a perfect set of slopesand close agreement with the EHBF of the firsttest. Studies on two other subjects were dis-

1686


carded because of obvious dilution of hepatic ve-nous blood with caval blood.

The single injection technique has never beenemployed for the study of EHBFduring exercise,but has been employed at rest, where it was shownto provide results corresponding closely to thoseobtained by the constant infusion technique in thesame individuals (15, 20). In theory neither issuperior, since one simply provides an integral(constant infusion) and the other a differential.The single injection technique has the advantageof requiring only the fractional clearance rate ofICG and the extraction efficiency for the dye.Also no long period for equilibration of dye isneeded, and repeated determinations are possiblewith less risk of saturating the hepatic extractionmechanism for the dye. The method has the dis-advantage of requiring that plasma volume beknown. As in earlier studies (33) a figure repre-sentative of the per kilogram plasma volume wasselected, since the principle aim of this studywas to assess only changes in hepatic blood flow.Accordingly, it became necessary to add an addi-tional approximation to this study; namely, thedecrease in plasma volume resulting from exer-cise. It was possible to select from a literaturesource (34) carefully determined values of plasmavolume decreases during an intensity of exercise(8.6% grade, 3.5 mph) close to the lowest in-tensity employed in this study (10%o grade, 3.5mph). Additional investigations provided cluesas to probable decreases at higher levels of exer-cise (27, 34, 35). Despite the rather close agree-ment in this study of the mean value for plasmavolume with the findings of others (19) and, aswell, the average change with exercise (10%grade, 3.5 mph) (34), the wide variability ob-served (due principally to variation in the injecteddose of dye), discouraged the individual use ofthis estimator of plasma volume or changestherein. Others, however, have shown the validityof estimates of plasma volume using a single in-jection of ICG (15, 18).

Despite the problems inherent in the methods,particularly during severe upright exercise, theresults (i.e., changes in EHBF, hepatic extractionefficiency, and hepatic a-v oxygen difference) ofthis study were consistent during rest (14, 15, 18,20) and mild exercise with the findings of Brad-

ley (7) and Wade and associates (8). In addi-tion, derived values for splanchnic oxygen uptakeduring rest and exercise agreed quite closely withthe findings of these investigators (7, 8). Fur-thermore, they supported the conclusion of Wadeand his colleagues that splanchnic oxygen uptakeis not significantly altered during exercise. Paren-thetically, the latter finding may provide somephysiologic basis for the time-honored procedureof assuming an unaltered metabolic rate duringexercise for tissues that remain essentially at rest.This tacit assumption is inherent in the estimationof oxygen debt that is calculated from the differ-ence between total recovery oxygen uptake lessthe pre-exercise resting oxygen consumption.

Although the exclusive extraction of ICG bythe liver restricts its use to measurement of hepa-tic rather than total splanchnic blood flow, in theabsence of portacaval shunting of blood thesplanchnic flow is also estimated. Such shuntsare known to exist in normal man (36), but quan-titatively may be in the order of 10% of the totalsplanchnic blood flow (37). On the contrary,oxygen is removed by extrahepatic, portal or-gans; therefore, a change in hepatic a-v oxygendifference does not necessarily reflect a change inhepatic blood flow, particularly if portacaval shunt-ing is quantitatively significant. Since the changesin hepatic a-v oxygen difference agreed on theaverage to within 2% of changes in EHBF, porta-caval shunting was not significantly altered in pro-portion to total portal flow during exercise, andtherefore changes in EHBFmust be closely equiv-alent to changes in total splanchnic blood flow.

On the basis of the findings in this study, it isconcluded that large decreases in EHBFand totalsplanchnic blood flow do occur in normal menduring moderate to severe upright exercise. Be-cause of the difficulty in obtaining hepatic venousblood samples, particularly when ventilatoryrates are high, EHBFmeasurements were not at-tempted during exercise either at or near the levelof maximal oxygen intake. However, since thedecreases in fractional clearance rates of ICG dur-ing sub-maximal exertion are usually rather closeapproximations of the measured decreases inEHBF, and in fact frequently underestimatethem, it is reasonable to assume that very largedecreases (approximately 80%) in hepatic blood

1687

L. B. ROWELL, J. R. BLACKMON,AND R. A. BRUCE

TABLE III

The cardiac output at a fixed level of oxygen intake that represents varying proportions of maximal oxygenintake in six sedentary men before and after physical conditioning and in four endurance athletes*

and the relationship to the predicted EHBF

02iintake, Mean cardiac* EHBF, %ofSubject 02intake* %0of maximal* output resting

ml/kg/min CLo L,'ninmm e eeeSix sedentarysmens(agess18s24syrsss28.5 i 1. 3 58.5 17.5 +i- 3.2$ 60Same six men after 3 months of physical 28.5 n4± 1.3 50.1 17.2 +4- 2.0$ 70

trainingFour endurance athletes (ages, 18-21 yrs) 28.5 +4- 1.3 45.6 17.4§ 75

* Reference 39.t Predicted from Table 1.I Mean and SD of 36 measurements.§ Mean of 24 measurements.

flow occur during maximal exercise. The obser-vation of maximal decreases of 81%o in ICG clear-ance and 84%o in EHBFconcurs with the estimatemade by \Vade and Bishop (3), who extrapolatedtheir data from mild to moderate supine exerciseto estimated maximal levels. This was also con-

firmed by the observation of a decrease in hepaticblood flow of approximately 80 %o (estimated fromthe increase in hepatic a-v oxygen difference) in a

patient with rheumatic heart disease during su-

pine exercise that maximally taxed the circu-latory adaptations to exertion of this patient (11).

An interesting analogy exists between the en-

durance athlete, the sedentary man, and the cardiacpatient. Although their maximal work capacitiesare markedly different, each shows very similaralterations in the partitioning of left ventricularoutput for distribution to the hepatic portal sys-

tem, and, quite likely, to other vascular beds also(3). The levels of work or metabolism, or both,at which these changes in hepatic blood flow be-come very marked are closely related to the pro-

portion of maximal oxygen consumption of theparticular individual, rather than to the absolutelevel of oxygen intake or cardiac output, althoughthe latter has not been examined in patients withcardiac impairment.

The relationship between ICG disappearancerate (as percentage of resting) and oxygen in-take (milliliters per kilogram per minute) wouldbe considerably less evident than it appears inFigure 3 were a larger number of endurance ath-letes with high values of maximal oxygen intakeincluded. For example, MM(Table II) had a

maximal oxygen intake of 75.3 ml oxygen per kg

per minute. At an oxygen uptake of 47.5 ml perkg per minute, which exceeded the maximal valuefor four men (Table I), the decrease in EHBFfor this subject was only 51%.

Despite the marked differences between athletesand sedentary men in EHBF at a given level ofsubmaximal aerobic metabolism (expressed perunit of body weight), the cardiac output is thesame for the two types of individuals (Table III)(38-40). Thus, changes in EHBFwith exerciseare not related to differences in total blood flowin the two types of individuals, but rather to theproportional demands for oxidative metabolism.

The responsible mechanisms) for thesechanges in EHBF during exertion is unknown.There are in addition, however, other physiologicchanges that are closely correlated to the relativeoxygen intake. For example, during levels of ex-ertion requiring approximately 50% of maximaloxygen intake, the following changes occur: Bloodlactic acid concentration (41) and respiratory quo-tient (42) begin to show a marked rise, and heartrate (in sedentary men) (42) breaks from a linearto asymptotic relationship to oxygen intake. Thesealterations may be concurrent with the onset ofanaerobiosis in exercising skeletal muscles thatdevelops as work intensity is increased.

It is also concluded from this study that inter-pretation of circulatory and other changes rela-tive to a level of aerobic metabolism are meaning-ful only when this metabolic rate is expressed asthe proportion of the maximal oxygen intake ofwhich each individual is capable. In this way thecirculatory and metabolic response to a fixed taskcan be meaningfully compared when individuals

1688


of widely varying maximal oxygen intake are

studied.Summary

The fractional clearance rate of indocyaninegreen (ICG) was determined from peripheralvenous samples during rest and upright exerciserequiring from 26 to 97% of maximal oxygen in-take. An inverse linear relationship (r = - 0.89)was observed between fractional ICG clearancerate as percentage of the resting value and oxygen

intake only when the latter was expressed as per-

centage of maximal oxygen intake.To test the validity of this procedure as a means

of assessing alterations in hepatic blood flow withexertion, the fractional clearance rate and hepaticextraction efficiency for ICG and estimated hepaticblood flow (EHBF) were determined in seven

men during rest and at one or two rates of work.The percentage change in ICG clearance under-

estimated percentage changes in EHBF, on theaverage, by 8.2%. Percentage changes in EHBFagreed, on the average, to within 2% of corre-

sponding changes in hepatic arteriovenous (a-v)oxygen difference.

It was concluded that large decreases, in ex-

cess of 80%, in EHBFoccur in upright man dur-ing severe exertion. The magnitude of change inblood flow varied inversely with relative ratherthan absolute oxygen intake at all rates of workwhere relative oxygen intake is defined as theproportion of maximal oxygen intake.

Acknowledgments

The assistance of Drs. John Mazzarella and RobertLamppert during numerous catheterization proceduresand the technical assistance of Miss Julie Drinkard, Mrs.Eric Merrifield, and Mr. Philip Weiser are gratefullyacknowledged. We also thank Drs. E. J. Masoro andWade Volwiler for their critical evaluation of thismanuscript.

References1. Krogh, A. The regulation of the supply of blood to

the right heart. Skand. Arch. Physiol. 1912, 27,227.

2. Bock, A. V., C. Vancaulaert, D. B. Dill, A. F6lling,and L. M. Hurxthal. Studies in muscular ac-

tivity. III. Dynamical changes occurring in man

at work. J. Physiol. (Lond.) 1928, 66, 136.3. Wade, 0. L., and J. M. Bishop. Cardiac Output

and Regional Blood Flow. Oxford, BlackwellScientific Publications, 1962, pp. 87, 107.

4. Herrick, J. F., J. H. Grindlay, E. J. Baldes, andF. C. Mann. Effect of exercise on the blood flowin the superior mesenteric, renal and common iliacarteries. Amer. J. Physiol. 1940, 128, 338.

5. Rushmer, R. F., D. L. Franklin, R. L. Van Citters,and 0. A. Smith. Changes in peripheral blood flowdistribution in healthy dogs. Circulat. Res. 1961,9, 675.

6. Sj6strand, T. The regulation of the blood distribu-tion in man. Acta physiol. scand. 1952, 26, 312.

7. Bradley, S. E. Hepatic blood flow. Effect of pos-ture and exercise upon blood flow through the liverin Transactions of the Seventh Conference onLiver Injury. New York, Josiah Macy, Jr., Foun-dation, 1948, p. 53.

8. Wade, 0. L., B. Combes, A. W. Childs, H. 0.Wheeler, A. Cournand, and S. E. Bradley. The ef-fect of exercise on the splanchnic blood flow andsplanchnic blood volume in normal man. Clin.Sci. 1956, 15, 457.

9. Lowenthal, M., K. Harpuder, and D. Blatt. Periph-eral and visceral vascular effects of exercise andpostprandial state in supine position. J. appl.Physiol. 1952, 4, 689.

10. Bishop, J. M., K. W. Donald, S. H. Taylor, andP. N. Wormald. Changes in arterial-hepatic ve-nous oxygen content difference during and aftersupine leg exercise. J. Physiol. (Lond.) 1957,137, 309.

11. Bishop, J. M., K. W. Donald, and 0. L. Wade.Changes in the oxygen content of hepatic venousblood during exercise in patients with rheumaticheart disease. J. clin. Invest. 1955, 34, 1114.

12. Reeves, J. T., R. F. Grover, G. F. Filley, and S. G.Blount, Jr., Circulatory changes in man duringmild supine exercise. J. appl. Physiol. 1961, 16,279.

13. Hunton, D. B., J. L. Boolman, and H. N. Hoffman.Studies of hepatic function with indocyanine green.Gastroenterology 1960, 39, 713. .

14. Cherrick, G. R., S. W. Stein, C. M. Leevy, and C. S.Davidson. Indocyanine green: observations onits physical properties, plasma decay, and hepaticextraction. J. clin. Invest. 1960, 39, 592.

15. Caesar, J., S. Shaldon, L. Chiandussi, L. Guevara,and S. Sherlock. The use of indocyanine greenin the measurement of hepatic blood flow and asa test of hepatic function. Clin. Sci. 1961, 21, 43.

16. Bruce, R. A., J. R. Blackmon, J. W. Jones, andG. Strait. Exercise testing in adult normal sub-jects and cardiac patients. Pediatrics supply. )1963, 32, 742.

17. Taylor, H. L., E. Buskirk, and A. Henschel. Max-imal oxygen intake as an obj ective measure ofcardio-respiratory performance. J. appl. Physiol.1955, 8, 73.

1689

L. B. ROWELL, J. R. BLACKMON,AND R. A. BRUCF,

18. Wiegand, B. D., S. G. Ketterer, and E. Rapaport.The use of indocyanine green for the evaluation ofhepatic function and blood flow in man. Amer.J. dig. Dis. 1960, 5, 427.

19. Gregersen, M. I. Blood volume, hemorrhage, andshock in Medical Physiology, 11th ed., P. Bard,Ed. St. Louis, C. V. Mosby, 1961, p. 280.

20. Reemtsma, K., G. C. Hottinger, A. C. DeGraff, Jr.,and 0. Creech, Jr. The estimation of hepatic bloodflow using indocyanine green. Surg. Gynec. Ob-stet. 1960, 110, 353.

21. Fauvert, R. E. The concept of hepatic clearance.Gastroenterology 1959, 37, 603.

22. Ketterer, S. G., B. D. Wiegand, and E. Rapaport.Hepatic uptake and biliary excretion of indo-cyanine green and its use in estimation of hepaticblood flow in dogs. Amer. J. Physiol. 1960, 199,481.

23. Keys, A., and H. L. Taylor. The behavior of theplasma colloids in recovery from brief severework and the question as to the permeability ofthe capillaries to proteins. J. biol. Chem. 1935,109, 55.

24. De Lanne, R., J. R. Barnes, and L. Brouba. Changesin concentration of plasma protein fractions dur-ing muscular work and recovery. J. appl. Physiol.1958, 13, 97.

25. Ebert, R. V., and E. A. Stead, Jr. An error inmeasuring changes in plasma volume after exer-cise. Proc. Soc. exp. Biol. (N. Y.) 1941, 46, 139.

26. Holmgren, A. Circulatory changes during muscu-lar work in man, with special reference to ar-terial and central venous pressures in the systemiccirculation. Scand. J. clin. Lab. Invest. 1956, 8,(suppl. 24), 1.

27. Kaltreider, N. L., and G. R. Meneely. The effect ofexercise on the volume of the blood. J. clin. In-vest. 1940, 19, 627.

28. Nedbal, J., and V. Seliger. Electrophoretic analysisof exercise proteinuria. J. appl. Physiol. 1958, 13,244.

29. Selkurt, E. E. Effect of acute hepatic ischemia onsplanchnic hemodynamics and on BSP removal byliver. Proc. Soc. exp. Biol. (N. Y.) 1954, 87, 307.

30. Bradley, S. E., F. J. Ingelfinger, and G. P. Bradley.Determinants of hepatic haemodynamics in Vis-ceral Circulation, Ciba Foundation. London, J. &A. Churchill, 1952, p. 225.

31. Bradley, S. E. Liver function as studied by hepaticvein catheterization in Transactions of the FifthConference on Liver Injury. New York, JosiahMacy, Jr., Foundation, 1946, p. 38.

32. Brauer, R. W. Liver circulation and function.Physiol. Rev. 1963, 43, 115.

33. Bradley, S. E., F. J. Ingelfinger, G. P. Bradley, andJ. J. Curry. The estimation of hepatic blood flowin man. J. clin. Invest. 1945, 24, 890.

34. Cassels, D. E., and M. Morse. Blood volume anl ex-ercise. J. Pediat. 1942, 20, 352.

35. Cullumbine, H., and A. C. E. Koch. The changes inplasma and tissue fluid volume following exercise.Quart. J. exp. Physiol. 1949, 35, 39.

36. Knisely, M. H., F. Harding, and H. Debacker. He-patic sphincters, brief summary of present-dayknowledge. Science 1957, 125, 1023.

37. Caesar, J., K. M. Barber, E. Baraona, and S. Sher-lock. The estimation of portal-systemic collateralflow in man using intrasplenic injection of radio-active indicator. Clin. Sci. 1962, 23, 77.

38. Taylor, H. L., Y. Wang, L. B. Rowell, and G. Blom-qvist. The standardization and interpretation ofsubmaximal and maximal tests of working ca-pacity. Pediatrics (suppl.) 1963, 32, 703.

39. Wang, Y., J. T. Shepherd, R. J. Marshall, L.Rowell, and H. L. Taylor. Cardiac response toexercise in unconditioned young men and inathletes (abstract). Circulation 1961, 24, 1064.

40. Wang, Y., L. B. Rowell, G. Blomqvist, and H. L.Taylor. The effects of a conditioning programon the cardiovascular function in exercise ofsedentary college males. In preparation.

41. Wyndham, C. H., N. B. Strydom, C. G. Williams,and M. von Rahden. A physiological basis forthe 'optimum' level of energy expenditure. Na-ture (Lond.) 1962, 195, 1210.

42. Rowell, L. B., H. L. Taylor, and Y. Wang. Limi-tations to the prediction of maximal oxygen intake.J. appl. Physiol. 1964, in press.

1690

Documents

Indocyanine Green Clearance and Estimated Hepatic Blood