Journal of Clinical InvestigationVol. 41, No. 6, 1962

THE EFFECT OF A SIDE-TO-SIDE PORTACAVALSHUNTONHEPATIC HEMODYNAMICSIN CIRRHOSIS *

By TELFER B. REYNOLDS,WILLIAM P. MIKKELSEN, ALLAN G. REDEKERANDHARRYS. YAMAHIROt

(From the Departments of Medicine and Surgery, University of Southern California School ofMedicine, and Los Angeles County General Hospital, Los Angeles, Calif.)

(Submitted for publication November 20, 1961; accepted February 1, 1962)

Since various types of surgical portal-systemicvenous anastomoses are employed in the treat-ment of portal hypertension, it is of interest toknow the effect of these operations on hepatichemodynamics.

There is general agreement that the major he-patic hemodynamic disturbance in cirrhosis is anincrease in vascular resistance in the liver, to-gether with an increase in portal venous pressure,and a variable, moderate reduction in hepatic ve-nous blood flow. The fact that wedged hepaticvein pressure is increased to near portal vein pres-sure level in most patients suggests that the ma-jor portion of the increased vascular resistance ison the hepatic venous side of the sinusoid. Portalhypertension, then, can be regarded as a compen-satory mechanism for the increased vascular re-sistance caused by the cirrhotic liver. The pos-sible contribution of increased splanchnic or he-patic arterial inflow to portal hypertension cannotbe ignored, although if this were a major factorwe would expect to find consistent increases inhepatic venous flow in cirrhosis. Creation of asurgical portal-systemic venous anastomosis wouldbe expected to lower portal pressure and reduceportal inflow into the liver. The effect upon he-patic venous flow would be dependent upon the de-gree of reduction in portal venous inflow that ex-isted prior to surgery, together with any adjust-ments that might take place in hepatic arterial flow.

Bradley, Smythe, Fitzpatrick and Blakemore,using the standard sulfobromophthalein (BSP)method to measure liver blood flow, found anaverage fall of 22 per cent in five patients after

* Supported by grants from the National Institutesof Health (H1718) and the Los Angeles County HeartAssociation (no. 256).

t University of Southern California Research Fellowin Medicine.

spleno-renal anastomosis (1). Splanchnic oxygenconsumption did not change.

After total exclusion of portal blood from theliver by end-to-side portacaval shunt, the sameworkers found reductions in hepatic blood flowof 21, 39, and 46 per cent in three patients (1).Redeker, Geller and Reynolds noted an averagereduction of 45 per cent in ten subjects (2). Inthe latter studies, as would be expected, wedgedhepatic venous pressure levels fell in proportion tothe fall in hepatic venous flow and the calculatedpostsinusoidal hepatic resistance did not changesignificantly. Arterial-hepatic venous oxygen dif-ference increased in proportion to the drop inblood flow and there appeared to be little, if any,effect on hepatic oxygen uptake from the opera-tion. These results suggest that, in patients withportal hypertension, portal venous inflow still rep-resents a significant fraction of the total liver bloodflow. However, the failure of hepatic oxygen up-take to drop significantly and the lack of any de-monstrable deterioration in liver function afterend-to-side portacaval shunt indicates that thereduction in hepatic blood flow is well tolerated.

The end-to-side portacaval shunt has the theo-retical disadvantage of permanent withdrawal ofportal venous flow. Should liver disease improveand the vascular resistance in the liver decrease,there is no possibility of a resumption of portalflow. A side-to-side portacaval shunt leaves theportal vein connected to the liver, but has the po-tential disadvantage that some hepatic arterialblood might be diverted from its usual course tothe hepatic vein to run backward through the por-tal vein and the side-to-side anastomosis. Thiswould reduce hepatic venous flow even more thanan end-to-side portacaval shunt and could impairthe liver's oxygen supply. The satisfactory clini-cal results in the series reported by Longmire,

1242

HEMODYNAMICCHANGESAFTER SIDE-TO-SIDE PORTACAVALSHUNT 1

Mulder, Mahoney and Mellinkoff (3) encouragedus to proceed with an analysis of hepatic hemo-dynamics before and after side-to-side portacavalshunt. Using hepatic vein catheterization, wehave made pre- and post-shunt measurements ofhepatic venous blood flow (HBF), wedged hepaticvenous pressure (WHVP), postsinusoidal vascu-lar resistance (HR), and splanchnic oxygen con-sumption (SOC) in nine patients with cirrhosis.An attempt was made to keep the diameter of theopening between portal vein and vena cava be-tween 1.2 and 1.5 cm in the hope that this wouldminimize the backflow of hepatic arterial bloodthrough the shunt.

MATERIAL AND METHODS

All of the nine patients described in this report hadalcoholic cirrhosis. They had been under medical ob-servation for varying periods of time, sufficient to es-tablish the diagnosis of cirrhosis and to ensure, throughwithdrawal of alcohol, liver function adequate to with-stand surgery. They had all had upper gastrointestinalbleeding, with no demonstrable source other than esopha-geal varices.

Preoperatively, hepatic vein catheterization was per-formed by the usual technique (4). In all instances,measurements were made in a rightsided hepatic vein.WHVPwas measured initially in as many locations aspossible, using a Statham strain gage and a photographicrecorder. Pressure in the inferior vena cava was re-corded either before entering the liver or after with-drawal of the catheter from it, and this served as a zeropoint of reference for all WHVPmeasurements. Afterpressure recordings, the catheter was withdrawn in thehepatic vein to a location 3 to 4 cm from the vena cava.From 40 to 70 mg BSP was then given as a priming doseand a continuous infusion of approximately 2 mg of BSPper minute administered through an arm vein with aconstant speed pump. After 25 minutes of BSP infu-sion, three pairs of blood samples were drawn simultane-ously from a femoral artery and from the hepatic veincatheter. These were used for the determination of he-patic blood flow as described by Bradley, Ingelfinger,Bradley and Curry (4), and for the measurement ofsplanchnic oxygen consumption. BSP concentration inthe serum was determined calorimetrically after themethod of Gaebler (5). The hepatic plasma flow wascomputed from the hepatic removal rate of BSP and theperipheral arterial-hepatic venous BSP concentration dif-ference. HBF was calculated from this value and thearterial hematocrit. The rate of BSP infusion (in mil-ligrams per minute) was corrected for the rate of changeof BSP concentration (A BSP) in the arterial plasma bythe formula of Bradley and associates (4), and the re-sultant value was taken to be the hepatic removal rate

of BSP (RBsp). The validity of the values for hepaticblood flow obtained in this study was determined by therelationship of the rate of change of the arterial BSPconcentration to the hepatic removal rate of BSP. It isfelt that the correction applied for the rate of change ofthe arterial plasma BSP levels is, at best, an approxima-tion, since the plasma volume is estimated from a nomo-gram. Therefore, as this correction factor becomeslarge in proportion to the value for hepatic BSP removal,the calculation of HBF becomes progressively less re-liable. In this study, all values for RBSP were greaterthan 1.6 mg per minute and the largest A BSP was 0.019mg per minute.

The values for hepatic resistance (HR) were deter-mined by dividing WHVPby HBF and multiplying thequotient by a factor of 100 to bring the values near unity.If WHVPequals hepatic sinusoidal pressure, then HRrepresents the postsinusoidal vascular resistance.

Blood oxygen content and capacity were measured di-rectly by the method of Van Slyke and Neill in one ofthe three pairs of samples. Blood oxygen saturation wasmeasured on all samples with an oximeter, and oxygencontent was calculated from the saturation value and theoxygen capacity. The average of the four values forarterial-hepatic venous oxygen difference was multipliedby the mean value for HBF to obtain splanchnic oxygenconsumption. After the measurement of hepatic bloodflow, the catheter was again advanced into the wedgedposition in the liver and WHVPrecorded a second time.After August, 1959, the location of the catheter in awedged position was confirmed by a roentgenogram takenimmediately after the inj ection of 3 ml of 90 per centHypaque (sodium diatrizoate) through the catheter (6).

At surgery, portal pressures were measured directlywith a saline manometer, using the portal vein itself as abaseline. In most patients, the portal vein was clamped,and the effect of this maneuver on pressures on both theintestinal and hepatic side of the clamp was noted.

Postoperative catheterization was performed when thepatient's clinical state and laboratory tests returned totheir preoperative levels, which usually required from 2to 3 months. In one case (BM) postoperative catheteri-zation was delayed for 10 months because of the persist-ence of moderate jaundice, and in another (EN) the pa-tient's failure to return for clinic appointments led to de-lay in the postoperative measurement to 8 months.

RESULTS

The pre- and postoperative hepatic venouscatheterization measurements are listed in Table I.Mean preoperative values for HBF, WHVP,HR,SOC, and arterial-hepatic vein oxygen differenceare similar to those previously reported in tencirrhotic patients before end-to-side portacavalshunt (2).

In the current study, preoperative HBF rangedfrom 560 to 2,140 ml per minute, with a mean of

1 243

REYNOLDS, MIKKELSEN, REDEKERAND YAMAHIRO

H BF

2500

2000

1500

1000

500

0

cc/m in

FIG. 1. EFFECT OF SIDE-TO-SIDE PORTACAVALSHUNT ON

HEPATIC BLOODFLOW.

1,380 ml per minute. Postoperative HBF was

lower in each patient, with a range of 335 to 890and a mean of 497 ml per minute. This is a sta-tistically significant drop (p < 0.001). When ex-

pressed as a percentage of the preoperative flow,HBF fell from 23 to 79 per cent, with a mean fallof 59 per cent (Figure 1).

Preoperatively, WHVPranged from 14 to 25mmHg above inferior vena caval pressure, with a

mean of 19.0 mmHg. There was a fall in WHVPin every instance postoperatively to levels of 6 to11 mmHg (in one patient, IC, WHVPwas notmeasured). Expressed as a per cent of the origi-nal value, the drop in WHVPaveraged 55 per

cent (Figure 2).Postsinusoidal hepatic vascular resistance

ranged from 0.76 to 3.00 (mean 1.72) arbitraryunits preoperatively. The postoperative range

was 0.79 to 2.50, with a mean of 1.90 arbitraryunits. Eight of the nine patients had both a pre-

operative and postoperative measurement, and themean value did not change significantly (Figure3).

0n

02m

EA4

0

0

4/2an0

04

02

02

04

4ico

X02

04

Ul)

0

04

co

E

0

0g

0 P P4

I..

Cd

x

(n

_ 00 o _4o C.. . .

t-ms+b r- oo 0%

.

m C 00 .0 t 0in ID

M

g:essoe e + e u20)

*w 4 m eq>C N m C e

c N -I (: ro Cb - e! IR "0

.0 r..

. .,4 .-

.C

vo e ) it)

U-) to to 11 +oN m to o

0\000000 00r oo cs - t 0% -

0. oot.C .C .C .) .%- to -4 t

-X-

C4

0o o CO*O U)t tt00 6 6 i00 000%~- ~0 c

ro o ooi,0 X0 0 0O

to-

ou) Ou)v 0s o o\0 in to 0E \0 0 (uD

*-X 0 00e)

St ~ ~ ~

C5l C5 U _ t_ 0 c; v * °

+

tw O ) U) eS05 0d Ud _

N t- "-_ es0-_ e

_ -O4 N4C

. . . e.l. . . .

0 00 00 \0

mO nO qO LOOOO O

° t-Uitcs-4

00 4 t' m q0~

> 8.O. 0. O g I000g 0.

0. r% t)0

NU-

0\b - 4 \-0 ON X4

NN - - _

00 \0 0

- O-~ R OR 'd 0£ 4 C- .I - 4 N -00. . . . . . *

O -' _o0 _ _ __.

U) w m0'I 110 C4 in 0%?1 W) '" m " "41 N o in In

1: a wu¢ :U z z W, w

1244

POST-OPPRE-OP

02.

OU).0

o'

E

co

020

CU02

0~2U'-0

bU)

2.0

0

~comI J

c o

00

oMO

0C-

0>0-

.00. O

0 0

E

S0.0 v

C-2.C

2f

0.0

CZU

11.0

CZRCOCU0

0 II

CU >0

~.0.0

ba

*0.O

HEMODYNAMICCHANGESAFTER SIDE-TO-SIDE PORTACAVALSHUNT

Splanchnic oxygen consumption ranged from 37to 69 ml per minute preoperatively, with a mean

of 52 ml per minute. Postoperatively, there was a

fall in this measurement in the eight patients inwhom it was made. Values ranged from 23 to 47ml per minute, with a mean of 32 ml per minute.When expressed as a percentage of the originalvalue, postoperative fall averaged 37 per cent. Thefall was highly significant statistically (p < 0.001)(Figure 4).

Direct portal pressure measured at surgery

ranged from 31 to 48 cm saline (Table II). Inevery patient, except WE, an appreciable fall inportal pressure occurred with the creation of thesurgical shunt. In all of the five patients in whompressure studies were done after the applicationof a clamp to the portal vein, there was a rise inpressure on the intestinal side of the clamp and a

drop in pressure on the hepatic side.

HEPATIC RESISTANCE

4.00

3.00

2.00

1.00

DISCUSSION

WNrarren and Muller found a rise in pressure on

the hepatic side of a clamp placed across the por-

tal vein in three of seven patients prior to porta-

WHVP(mm.Hg)

PRE-OP POST-OP

25

20

15

10

5

0

FIG. 2. EFFECT OF SIDE-TO-SIDE PORTACAVALSHUNT ONWEDGEDHEPATIC VENOUSPRESSURE.

0

FIG. 3. EFFECT OF SIDE-TO-SIDE PORTACAVALSHUNTON

CALCULATED HEPATIC POSTSINUSOIDAL VASCULAR RESIST-ANCE (ARBITRARY UNITS).

caval shunt (7), suggesting that the portal veinoften may act as an outflow tract in cirrhosis.Such pressure changes have been an occasionalbut infrequent finding in our experience and werenot seen in any of the five patients so tested inthis study. Eighteen of the 19 patients included inthis and in a previous report (2) showed a fall in

TABLE II

Direct pressure measurements (saline manometer) at surgeryin patients undergoing side-to-side portacaval shunt

Portal vein pressure

After clamping

Before Hep. Int. AfterPatient shunt limb limb shunt

cm saline cm saline

BD 32 24LA 34 23 40 20BM 35 23ED 44 20WE 38 32 48 35IC 38 30 56 27NW 48 29 60 25CF 33 26 41 15RN 31 17

1245

REYNOLDS, MIKKELSEN, REDEKERAND YAMAHIRO

HBF after portacaval shunt, indicating that portalvein flow is toward the liver in the great majorityof patients with cirrhosis.

A marked fall in HBF was a consistent findingafter side-to-side shunt. The fall averaged 59 percent, and the mean postoperative HBF was only497 ml per minute. This drop in HBF is greaterthan that found previously after end-to-side shunt,where mean postoperative HBF was 801 ml perminute and the average fall was 44 per cent (2).Direct statistical comparison of the values forHBF after side-to-side and end-to-side shunt waspossible since preoperative HBF was similar inboth groups of patients. Mean HBFafter side-to-side shunt was 304 ml per minute lower than thatafter end-to-side shunt; the standard error of thedifference between the means was 164. Therefore,there is a 6 per cent possibility that the observeddifference is due to chance alone. If there was, infact, a greater fall in hepatic venous flow after side-to-side shunt in our patients, it can be readily ex-plained by retrograde flow of hepatic arterial blood

02 UPTAKE

80

60

40

20

0

(cc./min.)

FIG. 4. EFFECT OF SIDE-TO-SIDE PORTACAVALSHUNTON

SPLANCHNIC OXYGENCONSUMPTION.

in the portal vein through the shunt orifice. Thishas been shown to occur in normal dogs after side-to-side shunt by Murray and Mulder (8, 9) and byLong and Lombardo (10). There are severallines of evidence indicating that it also occurs inman. At surgery, after construction of a side-to-side shunt, Longmire and co-workers recoveredisotopically labeled material in the hepatic limb ofthe portal vein after its injection into the hepaticartery (3). Warren and Muller found a rise inpressure in the portal vein near the liver after theportal vein was clamped on the hepatic side of theshunt (7). Contrast media injected through acatheter placed in the portal vein near the liverhas been noted to flow toward the vena cava (3, 7).Side-to-side and "double-barreled" portacavalshunts are thought to be more effective than end-to-side shunts in relieving ascites, due to thegreater lowering of presinusoidal pressure thatoccurs when some blood leaves the liver via theportal vein ( 11, 12).

It is possible that anastomoses between the por-tal vein and vena cava larger than those (1.2 to1.5 cm in diameter) used in our patients wouldresult in more portal vein backflow and a greaterfall in hepatic venous flow. Wedo not have databearing on this question, since we have not meas-ured pressure gradients directly across the shuntsto note the size at which the orifice begins to offersignificant resistance to shunted blood flow.

The applicability of the BSP method for themeasurement of liver blood flow must be qualifiedunder conditions where some blood leaves theliver via the portal vein. The accuracy of the cal-culation (Fick principle) of the volume of bloodleaving the liver by the usual route (hepatic ve-nous blood flow) will not be affected by blood leav-ing the liver through the portal vein, providedthere is no BSP extracted from the latter. Ifthere is some BSP extraction from this blood, thenhepatic venous flow will be overestimated ac-cordingly. The presence of retrograde portal flowcannot, then, account for an underestimation ofhepatic venous flow by the standard BSP tech-nique; it can result only in an overestimation thatwill be quantitatively dependent on the amountof BSP extracted from the portal effluent. Onthree occasions, we have obtained blood from thehepatic limb of the portal vein after the creation of

1246

HEMODYNAMICCHANGESAFTER SIDE-TO-SIDE PORTACAVALSHUNT

a side-to-side anastomosis and have found BSPconcentration to be only slightly (0, 11, and 12per cent) less than in arterial blood. This sug-gests that the postoperative values obtained forhepatic venous flow in this study are a little largerthan the true values.

After a portacaval shunt, if there is no changein postsinusoidal vascular resistance, WHVPshould fall in proportion to the fall in hepatic ve-nous flow. Although there are some individualvariations, the average fall in WHVPin the pa-tients in this study was of approximately the samemagnitude as the fall in HBF, and the averagevalue for HRdid not change significantly.

Postoperative splanchnic oxygen consumptionwas lower in the patients in this study than afterend-to-side portacaval shunt. It is to be remem-bered that the preoperative figure includes theoxygen utilization for the entire splanchnic area,while the postoperative figure, if there is no portalinflow into the liver, represents only hepatic oxy-gen consumption. On the other hand, the post-operative figure probably overestimates hepaticoxygen consumption to a small extent, since he-patic blood flow is presumably overestimated.Also, it is possible that an appreciable amount ofoxygen is extracted from any hepatic arterial bloodthat flows backward into the portal vein. Al-though the side-to-side shunt appears to cause areduction in hepatic oxygen consumption, themany errors inherent in the calculations call for acautious interpretation of the data.

Warren and Muller contend that the backflowof hepatic arterial blood into the portal vein aftera side-to-side shunt is beneficial to the liver, sinceit increases the total perfusion of the organ (7).This would be true, of course, only if the bloodflowing in the retrograde fashion has made somecontact with liver parenchymal cells. They foundoxygen saturation values of 70 to 86 per cent inblood from the hepatic limb of the portal vein,favoring their contention. We feel that furtherinvestigation is necessary to settle this importantquestion about the functional value of retrogradeportal flow, and such studies are now in progressin our own and, no doubt, in other laboratories.Blood from the hepatic limb of the portal veinand from the hepatic vein can be obtained simul-taneously by catheter for comparison of BSP and

oxygen content, and the direction of blood flow inthe portal vein can be determined by the injectionof contrast media. However, it is difficult to quan-titate the flow of blood in the portal vein, and itwill be necessary to know the amount of bloodflowing in a retrograde fashion before any con-clusions can be drawn about its contribution toliver function.

In spite of the marked reduction in hepatic ve-nous blood flow and the fall in splanchnic oxygenconsumption, the clinical results in our patientswith side-to-side shunts have not been demon-strably different from the results after end-to-sideshunt. One of the nine patients (WE) has beenlost to follow-up. The other eight survived foran average of 21 months after surgery. Three(BD, NW, and BM) are doing well with no seri-ous complications from their cirrhosis. Three(LA, RN, and IC) have returned to intermittentuse of alcohol, but are still doing well. One (ED)has developed deep jaundice and serious hepaticdecompensation following prolonged heavy drink-ing. One patient (CF) has developed chronic in-termittent encephalopathy. There has been onlyone episode of gastrointestinal bleeding (IC), pre-sumably due to alcoholic gastritis, since bothupper gastrointestinal X-ray and esophagoscopywere negative.

That marked falls in hepatic blood flow and, ap-parently, in hepatic oxygen consumption after side-to-side shunt are not accompanied by any obviousclinical deterioration is difficult to explain. Oneexplanation would be that offered by Warren andMuller (7) ; i.e., that a significant volume of bloodflowing retrograde in the portal vein contributesto the oxygen supply and function of the hepaticparenchymal cells. Another explanation has beenproposed by Shaldon and co-workers; namely,that the reduction in hepatic venous flow after aportacaval shunt is largely due to elimination ofnonfunctional blood flow through intrahepaticportal-hepatic venous shunts (13). However, thisexplanation does not account for the apparent re-duction in hepatic oxygen consumption. A thirdpossibility, which seems most likely to us, is thatthe cirrhotic liver is able to adapt fairly well tomarked changes in blood flow, and that the dvs-function caused by portacaval shunting is not greatenough to be easily recognizable.

1247

REYNOLDS, MIKKELSEN, REDEKERANDYAMAHIRO

SUMMARY

Measurements of hepatic blood flow (HBF),wedged hepatic vein pressure (WHVP), post-sinusoidal vascular resistance (HR), and splanch-nic oxygen consumption (SOC) were made byhepatic vein catheterization in nine cirrhotic pa-

tients before and after side-to-side portacavalshunt. HBFdropped in each patient (average fall59 per cent), as did WHVP(average fall 55 per

cent). HR did not change significantly. SOCfell 37 per cent on the average, although arterial-hepatic vein oxygen difference increased in allpatients.

HBF fell to a degree greater than that previ-ously noted after end-to-side portacaval shunt(59 as compared to 45 per cent). The statisticalsignificance of this difference is questionable (p <0.06). If there actually is a greater fall in HBFafter side-to-side shunt, it can be accounted for byretrograde flow of hepatic arterial blood throughthe proximal portal vein and the portal-vena cavalanastomosis.

In spite of the marked fall in HBF and the ap-

parent drop in SOCafter side-to-side shunt, theclinical results were not demonstrably differentfrom those after end-to-side shunt.

REFERENCES

1. Bradley, S. E., Smythe, C. M., Fitzpatrick, H. F.,and Blakemore, A. H. The effect of a portacavalshunt on estimated hepatic blood flow and oxygen

uptake in cirrhosis. J. clin. Invest. 1953, 32, 526.2. Redeker, A. G., Geller, H. M., and Reynolds, T. B.

Hepatic wedge pressure, blood flow, vascular re-

sistance and oxygen consumption in cirrhosis be-

fore and after end-to-side portacaval shunt. J.clin. Invest. 1958, 37, 606.

3. Longmire, W. P., Jr., Mulder, D. G., Mahoney, P. S.Mellinkoff, S. W. Side-to-side portacaval anasto-mosis for portal hypertension. Ann. Surg. 1958,147, 881.

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

5. Gaebler, 0. H. Determination of bromsulphalein innormal, turbid, hemolyzed, or icteric serums.Amer. J. clin. Path. 1945, 15, 452.

6. Reynolds, T. B., Redeker, A. G., and Geller, H. M.Technique for verification of wedging of an hepaticvenous catheter. Gastroenterology 1960, 38, 799.

7. Warren, W. D., and Muller, W. H., Jr. A clarifica-tion of some hemodynamic changes in cirrhosisand their surgical significance. Ann. Surg. 1959,150, 413.

8. Murray, J. F., and Mulder, D. G. The effects ofretrograde portal venous flow following side-to-side portacaval anastomosis: A comparison withend-to-side shunts. J. clin. Invest. 1961, 40, 1413.

9. Mulder, D. G. The role of surgery in the treatmentof portal hypertension. Amer. J. Gastroent. 1960,33, 305.

10. Long, R. T. L., and Lombardo, C. R. Hemodynamicobservations on the hepatic circulation: Modifica-tions produced by portacaval shunting (abstract).J. clin. Invest. 1959, 38, 1021.

11. McDermott, W. V., Jr. The treatment of cirrhoticascites by combined hepatic and portal decompres-sion. New Engl. J. Med. 1958, 259, 897.

12. Welch, C. S., Welch, H. F., and Carter, J. H. Thetreatment of ascites by side to side portacavalshunt. Ann. Surg. 1959, 150, 428.

13. Shaldon, S., Chiandussi, L., Guevara, L., Caesar, J.,and Sherlock, S. The estimation of hepatic bloodflow and intrahepatic shunted blood flow by col-loidal heat-denatured human serum albumin labeledwith I"m. J. clin. Invest. 1961, 40, 1346.

1248