356

HEPAT 01104

Journal oj Hepatology, 1992; 15: 356-360 $2 1992 Else&r Science Pc!biishers B.V. All rights reserved. 0168-8278/92/$05.00

Echo-Doppler evaluation of acute flow changes in portal hypertensive patients: flow velocity as a reliable parameter

Carlo SabbB’, Giovanna Ferraioli2, Paolo Buonamico’*3, Elsa Berardi l, Gianfranco Antonica’,

Kenneth J.W. Taylor3 and Ottavio Albano’

‘ktituto di Clirku Medico 1 I Unimm’td degli Studi di Bari, Bari, Italy; ‘Dioisione di Medicina, Ospedale Cicile di Scafati. Scafati, Italy; and ‘Department ofDiagnostic Radiology, Yale University School of Medicine. New Haven, CT. U.S.A.

(Received 23 July 199 I)

1n order to evaluate the behavior of the portal vein cross-sectional area during changes in portal flow, two groups of subjects were analyzed in two blinded cross-over studies using echo-Doppler flowmetry. The first group (I) consisted of 21 patients with cirrhosis and 16 controls. They received a standardized meal which is known to increase portal flow. The second group (II) consisted of 31 patients with cirrhosis who received a dose of propranolol which is known to decrease portal flow. In Group I, 30 min after the meal, the portal vein blood velocity increased by 35 + 6% (p < 0.01) in cirrhotic patients and by 55f 5% (p<O.Ol), in normal subjects. The portal vein cross-sectional area increased significantly in normal subjects (22 + 2%, pcO.01) but not in cirrhotic patients (4+ 2%, n.s.). In Group II, 2 h after propranolol, there was a significant decrease in portal blood velocity (- 14+2%), whereas the portal vein cross- sectional area did not show any significant changes. These data demonstrate that, in portal hypersensitive patients, the portal area measured by echo-Doppler flowmetry can be assumed to be constant and hence its calculation to estimate changes in portal blood flow can be omitted. Therefore, the use of blood velocity alone is suggested to monitor acute changes in flow in portal hypertension using Doppler flowmetry. The elimination of the portal vein cross-sectional area measurement simplifies the quantitative calculation of portal hemodynamics and increases the reliability of the technique by avoiding a source of error.

Keq’ Mlords: Splanchnic vessels; Blood flow measurements; Cirrhosis: Doppler studies; Ultrasound

Despite certain limitations in quar,tifying blood flow, Doppler flowmetry is currently used to study portal hemodynamics (l-4). The most widely used formula to calculate blood flow for the portal vein (PV) is F = A x V, where F is portal blood flow, A is portal vein cross-sectional area and V is mean velocity. When using the echo-Doppler technique, measurements of both parameters are subject to potential errors (5,6). If changes m flow could be estimated from only one of these parameters, at least one of two potential sources of error could be eliminated {7).

It has been established in an experimental model that the portal vein flow of the normal rabbit increases significantly over a range of pressures from 2 to 15

mmHg, but that the slope of the volume-pressure curve decreases when the pressure approaches 15 mmHg (8).

In human portal hypertension there is an enlarged portal system which is relatively insensitive to respiratory variations (9). We hypothesized that, in portal hyperten- sive patients, the high venous pressure stretches the vein and reduces changes in area in response to changes in flow. Therefore acute changes in flow measured by echo- Doppler should be independent of portal vein area measurements.

To verify this hypothesis, we analyzed data from two previously reported studies (10,ll) with additional data (12 patients). All data were re-analyzed together.

Two different stimuli which are known to have oppo-

Corresponderlce 10: Carlo Sabb& M.D., Istituto di Clinica Medica I, Universite degli Studi di Bari, Piazza Giulio Cesare 11, Bari 70124, Italy.

EVALUATION OF FLOW CHANGES BY ECHO-DOPPLER 357

site effects on portal blood flow were used in a controlled

and blinded study of patients with cirrhosis and control subjects: a meal which increases portal flow (12- 14), and

a dose of propranolol which decreases portal flow (15-l 7). In the present report, attention is focused on

the effects of these stimuli on the portal vein cross-

sectional area and velocity, measured separately, in order to distinguish the contribution of each on flow changes.

Materials and Methods

Patients

Fifty-two patients with cirrhosis and 16 normal volun-

teers entered the study. Two additional patients were

excluded because of ultrasound (US) evidence of a

paraumbilical vein. The diagnosis of cirrhosis was based

on liver biopsy findings in 21 subjects and on clinical, laboratory, US and endoscopic findings for the remain-

ing patients. Forty patients had esophageal varices

observed with upper gastrointestinal endoscopy; the remaining patients had US signs of portal hypertension.

The subjects were distributed in two groups depending

on the stimulus received: meal/sham meal (test I) and

propranolol/placebo (Test II).

Test I This group included 21 patients with cirrhosis and 16 normal volunteers (Table 1). The etiology cf

cirrhosis was chronic hepatitis B virus infection in seven

patients, and alcoholic in 14. According to the Child-

Turcotte classification (IS), 13 patients were Class A and

eight were Class B. Test II This group included 31 cirrhotic patients

(Table 1). The etiology of cirrhosis was chronic hepati-

tis B virus infection in 17 patients, alcoholic in ten,

cryptogenic in three, and hemochromatotic in one. Nine

patients were Class A, 16 were Class E. and six were

Class C. Thir;y subjects were examined at the West Haven

Veterans Administration Medical Center, the Clinical

TABLE 1

Physical characteristics of the studied population

Controls (meal test) (n = !6)

Sex

Age Weight BSA BMI

12M 4F 16M SF 23M SF - 68 f I.5 67 &- 1.7 59+ 1.8 72 + 2.3 74 12.8 65 2 2.0

1.93 + 0.04 1.93 + 0.05 1.77 * 0.03 25.7 + 0.8 25.8 _+ 0.9 23.3 + 0.6

Cirrhotics

Meal test (n=21)

Piopranolol test (n = 31)

Mean -(- SE. BSA = body surface area; BMI = body mass index.

Campus of the Yale School of Medicine (by 6.~. and C.S.), 26 at the University of Bari, Istituto di Clinica

Medica I, Italy (by C.S. and P.B.) and the remaining 12

tit the ‘Ospedale Civile’ of Scafati,, Italy (by G.F.). The three operators performing the examinations had a

background of preliminary work together which pro-

vided a warranty against methodological differences among them.

The study was approved by the Human Investigation Committee of Yale University School of Medicine and of the VA West Haven Medical Center. Informed written

consent was obtained from each patient

Equipment

The Doppler equipment used included a Toshiba sonolayer SSAjSSD-1OOA duplex system with 2.5 and

3.75 MHz sector electronic probes, an ATL Ultramark

8 duplex system with a 3.0 MHz sector mechanical

probe, and an ATL Illtramark 9 color-duplex system with a 3.5 MHz convex electronic probe.

Methods

All subjects were studied in the morning, after an

overnight fast. In both groups of subjects, after 15 min

of rest, portal vein baseline measurements were recorded

with the subject in a supine position. The operator performing the examination was unaware of the test

substcance received by the patients. Test I Subjects were randomized to receive either a

standardized mixed liquid meal (355 kcal; 16.7% proteins, 30% fats, 53.3% carbohydrates; Ensure plus. Abbott

Lab.) or a sham meal (an equivalent volume of water)

in a cross-over design on two different days. Measure-

ments of the portal vein parameters after either the meal

or sham meal were obtained after 30 min.

Tesr II Thirty-one patients were randomly assigned

to receive propranolol(40 mg p.o.) or placebo in a cross- over design on two different days. Two hours after propranolol or placebo administration, echo-Doppler

measurements of portal vein parameters were taken.

Cakulations

Portal cein cross-sectional area (A: min’) Area was

obtained from measurement of a cross-section of the

vessel visualized in B-mode after defining the major and minor diameters (Dl, D2) of the vessel. The computer

software of the duplex system calculated the area using

the ellipse formula: A=n Dl *D2/4.

Mean velocity ( V; cm/s) The maximum velocity meas-

urements were obtained during suspended inspiration and averaged over a few seconds. They were taken on

358 C. SABBA et al.

a longitudinal section of the vessel using an incident angle of less than 60 degrees. The mean velocity was calculated as reported in the literature (V,,,,,,=

0.57 * V,,,) (19). At each time period, portal vein and mean velocity

were obtained by averaging three consecutive measure- ments for each parameter.

Statistics The baseline results for each group are given as mean

*SE of absolute values. The changes over time are

expressed as percentages over baseline values. Comp- arisons are performed using a paired Student t-test p-Values < 0.05 are considered significant.

Results

Group I (meal test in conrrols and cirrhotics) The baseline absolute values of A and Y are show in

Table 2. Area When compared to placebo, the portal vein

cross-sectional area increased significantly in normal subjects after food intake (+22 +2% vs. +1 +l%,

p< 0.01). Patients with cirrhosis showed no significant

postprandial changes (+ 4f 2% vs. + 1 f I%, n.s.) (Fig. 1). Velocity When compared to the placebo, the portal

vein velocity increased significantly 30 min after food intake in both normal subjects (+55*5% vs. O+ 1%,

~~0.01) and patients with cirrhosis (+ 35 +6% vs. -3f2%, p<O.Ol).

Group II (propranolol test in cirrhotics) The baseline absolute values of A and V are given in

Table 2.

Area The portal vein cross-sectional area failed to

show any difference induced by either propranolol or placebo (Fig. 2).

TABLE 2

Baseline values

Controls (meal test) (n = 16)

Cirrhoiics

Meal test (n=Zl)

Propranolol test ln=31~

PV area (rrrm2) 130&5 201 k 16* 189 & 5* PV velocity

(cm/s) 14& 1 11&l* 11+1*

Mean + SE Compared to normals, the two groups of patients with cirrhosis p,esented a significantly greater portal vein (PV) cross- sectional area and a significantly lower velocity. Statistical significance (cirrhotics U. controls): *p c 0.001.

% ‘L

80 - 80

1 60

4c

a 0

0

-20

60

-20 AREA UELOCITY

i?normals sham neai r3 normals neal q cirrhotics sham meal Tclrrhotics meal

Fig. 1. Test 1. The percentage changes from baseline values in portal vein cross-sectional area and velocity 30 min after a standardized meal or a sham meal in normal subjects and in patients with cirrhosis are

shown. Statistical significance: meal vs. sham meal p<O.Ol.

i 10 1

i 1

1 -104 -10

I -201 -20

AREA UELOCI-ITY

q placebo

0 propranolol

Fig. 2. Test II. The percentage changes from baseline values in portal- vein area and velocity 2 h after placebo or propranolol in patients with cirrhosis are shown. Statistical significance: propranolol vs. pla-

cebo. pcO.01.

Velocity When compared to the placebo, there was a

significant decrease in portal blood velocity (-14f2%

vs. + 1 +l%, p<O.Ol) 2 h after propranolol admin-

istration.

Discussion

The cross-sectional area of the portal vein under physiological conditions is not circular and therefore

cannot be calculated by measuring the diameter of the

vessel on one plane using the formula R D2/4. Instead,

it must be calculated by measuring the two diameters

in the cross-section using the formula for an eilipse

R Dl -D2,‘#. A given percentage error in estimating DZ

EVALUATION OF FLOW CHANGES BY ECHO-DOPPLER,

and 02 will result in an exceptional error in estimating

flow (5).

Respiratory changes, physiological stimuli, and ana-

tomical relationships with other vessels influence varia-

tions in the cross-section of the portal vein in normal

subjects over a given period of time. Therefore, in spite

of all the limitations in measuring the changes in area

in physiological states, this measurement is necessary

and must be taken in account. This was confirmed by

the results of this study. In normal subjects the portal

vein cross-sectional area increased after a standardized

meal, and the changes in cross-sectional area contributed

significantly to increasing the portal blood flow.

There are reports in the literature which attribute

portal vein flow changes in patients with cirrhosis to the

variations in both portal vein cross-sectional area and

flow velocity (20-22). However, due to an increase in

portal pressure in patients with cirrhosis, the portal

system behaves differently (23-25). In fact, once the steep

ascending part of the pressure-volume curve of a vessel

is reached, any change in flow must be achieved by a

change in velocity rather than by any further increase

in the maximally stretched wall diameter. Therefore, in

portal hypertension, recorded changes in the portal vein

cross-sectional area measured by echo-Doppler after

physiological stimuli could be due to ihe variability of

the measurement technique by the operator, and not to

real changes in vessel size.

In this study the operator was unaware of the test

substance used, thus any variability in measurement was

taken into account by the blinded nature of the study.

Only slight changes in the portal vein cross-sectional

area were observed both when portal vein flow was

increased after a meal and when it was decreased with

propranolol. Indeed, the changes in area were not statis-

tically significant. Moreover, they fell into the range of

the intra-observer variability observed by us in the

previous study (6). The results in this study were reported

by three operators who had been using this non-invasive

methodology together in the same laboratory for 2

consecutive years. They had performed a great number

of examinations together and only when the methodol-

ogy was standardized was the protocol comple:ed in

different locations. The preliminary work is the warranty

against methodological differences among them. On the

other hand, the use of different operators in this particu-

lar study was not a limit to the methodology or the test

hypothesis but rather a further confirmation. We

observed a reaction in the area and velocity changes

after acute stimuli in pai:ents with a common denomina-

359

tor, i.e. portal hypertension, independent of the operators.

Our results show that, in portal hypertension, acute

changes in portal blood flow after administration of well-known stimuli, demonstrated by echo-Doppler, are due to portal blood Bow velocity and not to ciiangcs in

the cross-sectional area of the portal vein. In portal hypertensive states, the area of the portal vein can be

reasonably assumed to be constant when acute changes in portal blood flow are induced by methods similar to those used in this study. We therefore suggest that

measurement of portal vein velocity alone may be suffi- cient to monitor acute changes in portal flow in portal

hypertension (additional studies are necessary fo evalu-

ate the possibility of measurable changes in area occur-

ring following larger reductions in portal blood flow and pressure than those induced in this one).

However, since researchers and physicians are more

familiar with data expressed in terms of blood flow

rather than of blood velocity, the cross-sectional area of

the portal vein could be measured at baseline: this value can then 5e used as a constact in order to calculate

blood flow. In conclusion, elimination of the repeated area meas-

urement has two advantages when measuring the effect of acute stimuli on portal hemodynamics: an enormous simplifica:ion of the quantitative measurements of portal

hemodynamics, and increased reliability of the echo- Doppler technique. This technical simplification of por-

tal blood flow calculation in portal hypertensive patients by echo-Doppler gives this non-invasive technique a

new promising potential for clinical trials.

Acknowledgemeoas

The authors wish to thank Prof. R.J. Groszmann for his valuable suggestions and Ms. Paulene Butts for her

assistance in the preparation of the manuscript.

References

Patriquin H, Lafortune M. Bums PN, Dauzat M. Duplex Doppler examination in portal hypertension: technique and anatomy. Am J Roentgen01 1987; 149: 71-6. Ohnishi 1, Saito M, Sato S, Sugita S, Tanaka H. Okuda K. Clinical utility ol pulsed Doppler flowmetry in patients with portal hyper- tension. Am J Gastroenterol 1986; 81: l-8. Moriyasu F, Ban Id, Nishida 0, et ai. Clinical application of an ultrasonic duplex system in the quantitative measurement of porta! blood Row. J Clin Ultrasound 1986; :4: 579-88. Moneta GL, TayJor DC, Scott Helton, W, Mullholland MW,

360

S&an&less DE. Duplex ultrasound measurement of post-prandial intestinal blood flow: effect of meal composition. Gastroenterology 1988; 95: 1294-1301.

5 Gill RW. Measurement of blood flow by ultrasound: accuracy and source of error. Ultrasound Med Bioi 1985; 11: 625-41.

6 Sabbi C, Weltin GG, Cicchetti DV, et al. Observer variability in echo-Doppler measurements of portal flow in cirrhotic patients and normal volunteers. Gastroenterology 1990; 98: 1603-l I.

7 Burns P, Taylor KJW, Biei AT. Doppler flowmetry and portal hypertension. Gastroenteroiogy 1987; 92: 824.

8 Brown BP, Heistad DD. Capacitance of the rabbit portal vein and inferior vena cava. J Physiol (Lond.) 1986; 38 I: 417-25.

9 Bolondi L. Gandoifi C, Arienti V, et al. Ultrasonography in the diagnosis of portal hypertension: diminished response of portal vessel to respiration. Radiology 1982; 142: 167-72.

IO Sabba C, Ferraioli G, Genecin P, et al. Evaluation of postprandial hyperemia in superior mesenteric artery and portal vein in healthy and cirrhotic humans: an operation-blind echo-Doppler study. Hepatoiogy 1991; 13: 714-8.

1 I Sabba C, Ferraioli G, Buonamico P, et al. Post-prandial hyperemia in cirrhosis: the effect of propranolol. Gastroenterology 1990; 98: A626.

12 Chou CC. Splanchnic and overall cardiovascular hemodynamics during eating and digestion. Fed Proc 1983; 42: 1658-61.

13 Okazaki K, Myazaki M, Onishi S, Ito K. Effects of food intake and various extrinsic hormones on portal blood flow in patients with liver cirrhosis demonstrated by pulsed Doppler with the Octoson. Stand J Gastroenteroi 1986; 21: 1029-38.

I4 Bellamy EA, Bossi MC, Cosgrove DO. Ultrasound demonstration of change in the normal portal venous system following a meal. Br J Radioi 1984; 57: 147-9.

15 Lebrec D, Poynard T, Hillon P, Benhamou JP. Propranoiol for

16

17

18

I9

20

C. SABBA et al.

prevention of recurrent gastrointestinal bleeding in patients with cirrhosis: a controlled study. N Engi J Med 1981; 305: 1371-4. Pomier-Layrargues G, Viiieneuve JP, Wiliems B, Huet PM, Marieau D. Systemic and hepatic hemodynamics after variceal hemorrage: effects of propranofol and placebo. Gastroenteroiogy 1987; 93: 1218-24. Coiombo M, De Franchis R, Tommasini M, Sangiovanni A, Dioguardi N. Beta blockade prevents recurrent gastrointestinal bleeding in well-compensat d patients with alcoholic cirrhosis: a multicenter randomized controlled trial. Hepatoiogy 198% 9: 433-8. Conn HO. A peek at the Child-Turcotte classification. Hepatology 1981; 1: 673-6. Moriyasu F, Ban N, Nishida 0, et al. Clinical application of an ultrasonic duplex system in the quantitative measurement of portal blood Sow. J Ciin Ultrasound 1986; 14: 579-88. Zoii M, Marchesini G, Brunori A, Cordiani MR, Pisi E. Portal venous flow in response to acute fi-blocker and vasodilatory treatment in patients with liver cirrhosis. Hepatology 1986; 6: 1248-51.

21 Onishi K, Saito M. Nakayama T, Hatano H, Okuda K. Effects of vasopressin on portal hemodynamics in patients with portal hyper- tension. Am J Gastroenterol 1987; 82: 135-8.

22 Goyal AK, Pokharna DS, Sharma SK. ultrasonic measurements of portal vasculature in diagnosis of portal hypertension. A contro- versial subject reviewed. J Ultrasound Med 1990; 9: 45-8.

23 Murray JF, Dawson AM, Sherlock S. Circulatory changes in chronic liver disease. Am J Med 1958; 24: 358-67.

24 Kontos HA, Shapiro W, Mauck HP, Patterson JL. General and regional circulatory alterations in cirrhosis of the liver. Am J Med 1964; 37: 526-35.

25 Sikuler E, Kravetz D, Groszmann RJ. Evolution of portal hyperten- sion and mechanisms involved in its maintenance in a rat model. Am J Physiol 1985; 248: G618-25.