Luigi Bolondi, MD Giovanna Benzi, MD

Silvia Li Bassi, MD #{149} Vittorio Santi, MD

#{149} Stefano

#{149}

Liver Cirrhosis: of Hepatic Veins’ The authors compared the Doppler ultrasonographic pattern of hepatic veins (HVs) in a group of 60 patients affected by liver cirrhosis and in 65 healthy subjects comparable for sex and age to (a) detect possible differences in HV waveform in the two groups and (b) investigate the relationship of these differences with the severity of the disease (according to Child-Pugh classification) and the modifications of systemic hemodynamics. The waveform of HVs was arbitrarily classified into three patterns: HVO, a normal waveform; HV1, lower oscillations without the reversed phase; and HV2, completely flat waveform. The resistivity index of the superior mesenteric artery, reflecting the peripheral splanchnic impedance and the hyperdynamic circulation, was also measured in a subgroup of 45 cirrhotic patients. The waveform of HVs in all healthy subjects corresponded to the HVO pattern. Among cirrhotic patients, HVO was found in 30 (50%), HV1 in 19 (31.7%), and HV2 in 11 (81.3%). The severity of functional impairment was greatest in the HV2 group and least in the HVO grOUp. This was significantly correlated with the decrease of the resistivity index in the superior mesenteric artery in the subgroup of 45 patients. Changes in the normal HV waveform could be considered a useful adjunctive tool for the noninvasive evaluation of liver disease. The pathophysiology of these changes in HV blood flow is still unclear. The significant correlation with the severity of the disease and with the decrease of splanchnic resistances indicates that these changes in the HV waveform occur in the presence of marked rearrangements of liver tissue and of hyperdynamic systemic circulation.

Gaiani, Barbara,

#{149} Luigi

Changes

H

veins

(HVs)

the

inferior

toward

drain

anechoic

blood

vena

are easily recognizable (US) examination

as tubular

Gianni

#{149}

structures

are usually

easy

with

humans of the cava.

to demonstrate

in healthy subjects. The caliber of HVs increases gradually as the yessels near the diaphragm and varies with respiration cycles. Normally, at 2 cm from the inferior vena cava the

diameter

of the

mm,

1 cm can normal

and

HVs

rarely

exceeds

be considered value (1).

5 the

maximum Abnormalities of HVs can occur in different diseases and mostly involve caliber, such as in hepatic venous congestion secondary to right yentriculan failure, in which dilatation of the

inferior

vena

cava

and

HVs

is a

common finding (2,3). HVs are easily compressible, and the reduction in size of their lumen up to a partial or total lack of visualization at neal-time imaging can occur in cases of marked hepatomegaly and in several diseases (eg, severe fatty infiltration, BuddChiani syndrome), on when the increase chyma

of the makes

irregular, sis.

fibrosis their

as can

Although

ations

in liven parencourse thin and

occur

in liver

morphologic

of these

vessels

Waveform

cimrho-

alter-

terms:

959.12984 sound

are already

Radiology

venous pressure des, 50 that HV

ed toward

studies

1991; 178:513-516

From

the

entenology,

Bologna,

Institute Policlinico

Via

17;

June revision

ed September LB. c RSNA,

and

S Orsola,

University

Massarenti,

by. Received July

of Medicine

12, received

7. Address

9, 40138 1990;

revision September

reprint

the

ventricular

Gastroof

Bologna,

Ita-

due blood

heart

to cardiac cyflow is direct-

during

diastole

and

atnial shortly

and me-

pat-

registration of Doppler flowmetry and electrocardiography (Fig ia) and is well displayed by color flow mapping. The Doppler waveform of the jugular vein is similar to that of the HVs, and the recording of the jugular vein pulse (Fig ib), which is related to

changes in size of the vein that follow changes in pressure, reflects the Doppler waveform. Respiratory cydes also seem to influence the Dopplen waveform

known

of HVs.

that

during

It is well

deep

inspiration

intraabdominal pressure increases and consequently the amount of splanchnic

flowing

toward

the

consistently reduces blood

heart.

In this analysis phasic (5)

2).

of changes diagnosis

pattern of HV in in cirrhotic differences of the de-

in HV waveform of liver disease.

The

possible relationships of these changes to the severity of the disease and to the systemic and splanchnic

requested 6; accept-

requests

to Abbreviations:

1991

(ie,

versed during atnial systole. The tern is underlined by simultaneous

tection for the I

waveform

This study compares the the Doppler waveform of healthy subjects with that patients to detect possible and to assess the accuracy

Hepatic veins, US studies, cirrhosis, 761.794 #{149} Ultra-

Doppler

that the in healthy

condition, Doppler spectral shows a lowering of normal oscillations in HV waveform

#{149} Liven,

(US),

is a triphasic

asof HVs

two negative waves and one positive or the opposite, depending on the scan used) (4). This pattern is the consequence of variations in central

(Fig Index

and pathologic flowmetry

have been less studied. It has been demonstrated Doppler pattern of HVs

at ultraappear

and

MD

known, normal pects of Doppler

cava.

poorly defined walls running obliquely from the periphery liver toward the inferior vena

They

Zironi,

of Doppler

EPATIC

They sound

MD MD

tivity

index,

HV SD

standard

hepatic

vein,

RI

resis-

deviation.

513

I

I I III I f III !II

IIII

II

I

II II I I I

I I

I

uIIIIIIItIIl1tIjllI(HItIjIIIllIIIIIIIIIIlIIII1,IIIIIIIII

a. Figure

b. 1.

volume is positioned within the middle HV. On the night the normal Doppler with the electrocardiographic signal. The reversal of venous flow (pictured by the small wave above the zero line) is strictly related to the cardiac atnial and ventricular systole. When the venous flow is toward the heart (represented by the negative waves of the Doppler waveform) atnial and ventricular diastole occurs. (b) Jugular vein pulse with the phonocardiographic and electrocardiographic signals. The a wave corresponds to atnial systole: It follows the P wave in the electrocardiogram and is seen just before the first tone. On the descending branch of the a wave a small positive wave c is recorded that is contemporaneous with the first tone and is caused by the closure of the tricuspid valve and the impulse transmitted from the carotid artery. The negative wave x corresponds to ventricular systole (systolic collapse) and is due to the downward stretching of the atrioventniculan plane and the atnial relapse. The following positive wave v corresponds to the atnial filling. The apex of the v wave indicates the tricuspid valve opening following the second tone. The second negative wave y is due to the atnial emptying into the right ventricle (diastolic collapse); its descending branch corresponds to the quick ventricular filling, whereas the ascending slower branch (where in this case a small positive h wave is visible) ends with the following a pattern

(a) Right

of the

subcostal

HV

scan.

is represented

On the left the sample

together

wave.

hemodynamic

alterations

were

also

dynamics.

analyzed.

PATIENTS

AND

METHODS

In

Hz. During

according

to the

study.

sion, or Patients

both, with

were present US evidence

hepatocarcinoma the

were

The

patients

clinical side

of the

65 years,

were US

also

healthy

± SD

47.4

investigated

vol-

aged

years

as a control was Aloka

group.

performed

US equipment

34-

± 9.2)

(Esaote-

SSD-650)

by as large

positioning as at least

done

to reduce

fluence of the changes the inferior vena cava

514

Radiology

#{149}

the

a

the samone-third

possible

la);

inspiration.

was

then

of three

Doppler

(b)

phasic

HV1,

arbigroups

signal

triphasic phase

charac-

waveform of reversed

decreased

oscillation,

trace

(ie,

ampli-

measured this purpose

similar

of a portal

direction blood

high-pass

with

deep

one

in-

of flow pattern in on the HV hemo-

ting

and in the

in

the we

filter

(100

ty on

both, who

Hz).

to the

branch

mean portal

(Fig

was

which suspend

(9). For volume as

the

visualized,

superior

splanchnic

seconds

to

2.

Right

subcostal

scan

of a Doppler

US study in a healthy subject. In a Doppler pattern of the middle HV during free inspiration (top), the reversal of flow is nepnesented by the short phases above the zero line.

In

the

same

patient

during

ration (bottom), the waveform vein shows a reduction of the

patients

mesenteric

deep

inspi-

of the phasic

same oscilla-

resistivity

artery

was

evaluated

RI

Vmax

Vmax arterial

(10) with

velocity

velocity)

the disease was obtained in each with the Child-Pugh continuous

and

to assess

impedance,

been proved to be related hyperdynamic circulation Finally, an evaluation

was

index

Vmin/Vmax

-

maximum

minimum

in

artery

the

formula =

Figure

setveboci-

tions.

clearly

(where

of 4-6

lowest

mean

changes of flow velocity or respiratory cycles, or

of 45 cirrhotic

whom

Vmin

at its

into The

are also reduced in patients respiration during examina-

tion. In a subgroup

(RI) of this

kept calculated

trace

eventual to cardiac

3b).

velocity of the vein were also

cirrhotic group used a sample

We

a Doppler

average related

the

was

to avoid

large as the vessel diameter, taking account the peripheral component.

by

of the vessel diameter, within the middle HV at a distance of 3-6 cm from the outlet of the vessel into the inferior vena cava. was

any

Sixty-five

convex probe of 3.5 MHz provided by a pulsed Doppler device operating at a frequency of 2.5 MHz. Each subject was examined while in a supine position after overnight fasting to avoid any influence of meal and posture on splanchnic hemodynamics (6-8). The morphologic appearance of the HVs was evaluated by assessing caliber and course. Doppler study was

This

out

The venous

mean

then performed pie volume,

(Fig

right

33 women

450;

of

into

(a) HVO, of a short

main-

of the phasic oscillations without the phase of reversed flow (Fig 3a); and HV2, completely flat waveform with-

of the

and

EUB

tude short

by

US pattern

to the

tenistics: presence

flow)

breathing

caused

classified

of failure

the

(32 men

of duplex

Hitachi

trarily

was

each subject

stop

Doppler

Doppler

examination

means

The

had

in

heart.

modifications

(c)

of

None

of mci-

US beam

examination

to simply

study

included

evidence

unteers

in all cases. of nodules

excluded.

Doppler

angle

between 20#{176} and 50#{176}, and the analysis was always recorded for 4-6 seconds with a filter of 100

rhosis

admitted

the

tamed spectral at least

asked

was

case

of the

A group of 60 patients (32 men and 28 women aged 40-69 years, mean ± standard deviation (SD) = 52.5 years ± 8.9) affected by histologically proved liver cmcause of the disease was alcoholic in 13 cases, virus-related in 42 cases, and cryptogenetic in the remaining five cases. Endoscopic or US signs of portal hyperten-

each

dence

which

to the systemic (1 1). of the severity

the has

of

(12),

with

classes:

A (continuous and C (score

7-9),

distinction

lies on clinical tions (bilirubin els, presence

of three score 10-15).

case score

different

5-6), B (score This system re-

and laboratory and albumin of ascites, grade

observaserum 1evof encepha-

February

1991

the

HV2

group

and

least

in the

HVO

group) (P < .005) as well as between the decreasing RI and the changes in HV waveform (RI was least in the HV2 group and greatest in the HVO group) (P < .0005).

DISCUSSION In recent b. Figure

3.

Scans

of a patient

with

cirrhosis.

(a) The sample

volume

positioned

middle HV shows a reduction in size of the phasic oscillations of the the short phase of reversed flow (the Doppler pattern is always below pattern was classified as HV1. (b) The middle HV shows a completely ened spectrum of frequencies. This pattern was classified as HV2.

lopathy, nutrition)

prothrombin to allow

ance

time, state of estimate of hepatic

an

reserve.

For statistical the

analysis,

distribution

waveforms cirrhotic groups cording

between patients of cirrhotic to cause

score

HV

control

in

Doppler

subjects

and

and among the various patients divided acwere analyzed with the

x2 test. Differences dinal

differences

of various

in the Child-Pugh

were

analyzed

with

orx2

the

test

for trend. Differences in the mean portal vein velocity and in the RI of the superior mesenteric

artery

tients

divided

among

waveform of HVs means of multiple hibiting a P value sidered significant. tion coefficient relationship

was of both

Child-Pugh

score

changes

the

according

in the

cirrhotic

to the

were

pa-

different

analyzed

by

f tests. Differences less than .05 were

The

Pearson

used the

to analyze continuous

and

the

excon-

correlathe

RI with

the

HV waveform.

In all the group

(65

subjects

of the

subjects,

waveform

100%)

of HVs

showed

control the

Doppler

the

tnipha-

sic pattern (HVO). In the cirrhotic group the tniphasic pattern was observed in 30 patients (50%), whereas in 19 patients (31.7%) the Doppler

waveform HV1 and (18.3%).

HVO

of HV was in 1 1 patients In summary,

pattern

as

considering

as the

waveform,

classified as HV2

normal

abnormalities

the

Doppler in the

waveform of HVs occurred 60 cirrhotic patients (50%) none of the control group.

in 30 of but in Statistical

analysis showed a significant difference (P < .001) in the distribution of the normal Doppler pattern (HVO)

between cirrhotic),

found

the

two

groups

whereas

in the

behavior

no

(healthy difference

of the

vs was

HV

Doppler US pattern in relation to the different causes of cirrhosis. As fan as the morphologic appear-

Volume

178

Number

#{149}

2

middle

the

HV

is concerned,

in 21 of 65 cirrhotic patients US showed a thin tortuous course of the vessel, but this finding did not reveal any relationship with the chanactenistics of the Doppler signal, since an abnormal Doppler shift was recorded in 17 of 30 cases from a vein with the completely normal US appearance. Direction of portal venous blood

flow value flow

was always hepatopetal, and the of the mean velocity of blood in the portal vein was 13.7 cm!

sec ± 2.3 in patients with HVO, 12.2 cm/sec ± 2.8 in patients with HV1, and 12.9 cm/sec ± 2 in patients with HV2. The difference proved statistically significant (P < .05) when cornparing HVO with HV1 but did not reach significance when comparing HVO and HV2.

On RESULTS

of the

within

Doppler trace without the zero line). This flat flow with broad-

the

basis

of their

Child-Pugh

scores, 33 patients were allocated to class A (55%), 26 patients to class B (43.3%), and one to class C (1.7%). The difference in the severity of the functional impairment of the liver among the three classes of HV waveform analyzed with the x2 test for trend was significantly greater between HV2 and HV1 than between HV1 and HVO (z 2.37, P < .01). In the subgroup of 45 cirrhotic patients in whom the RI of the superior mesenteric artery was evaluated, a progressive reduction of this value was observed across the three different classes: HVO (24 patients) 0.85 ± 0.03, HV1 (14 patients) 0.81 ± 0.05, HV2 (seven patients) 0.77 ± 0.04. A statistically significant differ-

ence

was

found

between

HVO

and

HV1 (P < .005) and between HVO and HV2 (P < .0005). A significant correlation was found between the increasing Child-Pugh scone and the progressive changes in HV waveform (the Child-Pugh score was greatest in

years

Doppler

flowmetry

has been widely used to evaluate abdominal vessels. Although some discussion about the accuracy and reproducibility of quantitative measurements still exists (13-17), the value of qualitative findings is accepted worldwide (18). Particular attention has been devoted to Doppler investigation of liver cirrhosis (6-8,19), but until now the use of this technique has been directed mostly toward the hernodynamics of the portal vein and its tributaries and collatemals. A cameful study of HVs in this pathologic condition is still lacking. Only mecently has the Doppler US technique

been

usefully

applied

to investiga-

tion of the Budd-Chiani syndrome, in which absence of normal phasic oscillations in HVs has been described by Hosoki et al (20). These authors advised that the finding be used as a major criterion for the diagnosis of Budd-Chiani syndrome, even though

we have

shown

that

other,

more

spe-

cific Doppler US patterns, such as absent or reversed flow in the HVs or flat flow in HVs associated with meversed flow in the inferior vena cava, or both, may be detected in this condition (21). Furthermore, in the present study we have shown that a continuous turbulent flow in HVs can be found in patients with diseases other than the Budd-Chiami syndrome, thus indicating that this pattern is related to the presence of a diffuse alteration of liver pamenchyma and not only to the obstruction of

HV outflow. In our study, this pattern of continuous turbulent HV flow has been more frequently observed in cimrhotic patients with relatively severe disease, as shown by the significant conrelation with the Child-Pugh score. This observation suggests that the underlying mechanism of the change in the HV waveform could be related to the amount of liver fibrosis, which progressively reduces phasic oscillations in HVs subsequent to a lack of compliance of the tissue. On the othem hand, even the hypertmophy of the hepatic cells (22) might exert a possible compression over the HV because of the low stretching of the liver capsule.

Radiology

515

#{149}

such

as liver

steatosis

and

chronic

12.

hepatitis are concerned, definite data are not yet available. In conclusion, we believe that, on the basis of our preliminary data, Doppler US abnormalities in HV waveform could be reasonably considered sign of liver disease

what tional

an and

related to the impairment.

13.

Scan

nous

congestion

right

side

of

of a patient

with

secondary the

heart.

hepatic

yeof the

to failure Doppler

US

of

flow

during

both

systobe

and

FS.

tion

of the

sis

2.

An

anatomic

upper

14.

disease.

Berlin:

Kane

SA.

16.

Springer-

The compressive mechanism is furthem supported by the finding of a similar pattern (HV1 or HVO) in hepatic veins compressed and dislocat-

ed by liver

masses

(23).

This

Parry

WR,

Real-time

wave ing

been

pattern both

late

with

reversed

systole

and

reported

in hepatic

gestion secondary right side of the

lated

hepatic

to failure heart (24)

veins

flow

L.

L, Gandolfi

Burns KJW,

PN. Burns

Iskan17.

5.

6.

7.

1980;

151:753-763.

Ohnishi

K, Saito

the

vena

cava Gynecol

Sato

on portal cirrhosis

C, Barbara

portal

hemodynamics

20.

portal

5, et al.

in

patients by

in patients

Zoli

in healthy

with 1989;

M,

Effect of meal

chronic

liver

Marchesini

measurement

blood

flow

in control

tients.

JCU

1986;

MR.

et al.

of splanchnic and

in cirrhotic

pa-

graphie.

Rev

141. Dannault

P. Bretagne

Siproudhis

resistance.

(Paris)

1972;

of

Doppler

US

in

hemodynamics:

con-

Italy,

12 Sep-

10:353-355.

Von

by

1987;

C, Tokunga

K, Mar-

M, Kozuka

T.

Hepatic

evaluation Radiobo-

170:733-737.

Doppler ultrasound. (in press).

H, Blendis

LM,

in

humans 1981;

JM, hepatic

Di-

Gastroenter-

Crossley

of intrahepatic in the Disse space and

IR, et a).

pressure with and hepato-

in the

rat.

Castro-

80:546-556.

Eleno

A.

veins

in

Duplex patient

pattern

Doppler with

(abstr).

of

cirrhosis:

Radiology

173(P):361. Bibra

H,

Schober

K, Kenni

R, Sebening

H, Blomer

constrictive

pericarditis

Am

echo-Doppler

J Gastroenterology

echocardiography

J Cardiol

H. by of the

1989;

R, Busch

Diagnosis pulsed hepatic

of Doppveins.

63:483-488.

126:127-

JF, Raoul JL, Noel J, J. Assessment of hypertension and liver

L, Gastard

role

failure

Neurol

Gas-

96:913-919.

value

studied

pseudoportal 24.

flowstudy.

L, Gaiani S. Li Bassi S. et al. of Budd-Chiani syndrome by

Longo

len

14:429-435.

Planilo T, Pourcebot L, Pottier JM, et al. Etude de La circulation carotidienne par les m#{233}thodes ultrasoniques et Ia thermo-

1989;

Orrego

1989;

C, Cordiani

Doppler

Portal

using

outflow obstruction: pulsed Duplex sonography.

Bobondi agnosis

the

disease.

C.

electromagnetic

T, Masuike

megaly

9:815-819.

Echo

the

and

Am

enterology

23.

on

Layrargues

of hepatic

Correlation collagen

humans

Doppler

Castro-

a comparative

disease

pulsed ology

22.

AT.

measurements

82:507-511. Hosoki T, Kuroda

gy 21.

Blei

1986;

Ohnishi K, Sato S. Pugliese D, Tsunoda T, Saito M, Okuda K. Changes of splanchnic circulation with progression of chron-

venous with

Portal

JCU

conference, Bologna, 1989. J Hepatol 1990;

ukawa

Obstet

blood flow demonstrated

L.

study

du-

measure-

92:824-826.

The

flowmetry.

pulsed Doppler with the Octoson. Scand Gastroenterol 1 986; 21:1029-1038. Gaiani 5, Bolondi L, Li Bassi 5, Santi V, Zinoni

L.

ultrasonic

hypertension.

1989;

Barbara

ic liver

Sung M,

19.

New

venous hemodynamic in chronic liver disease: effects of posture change and exercise. Radiology 1985; 155:757-761. Okazaki K, Miyazaki M, Ohnishi S, Ito K. Effect of food intake and various extrinsic

con-

11.

ultrasound.

and

venous

10.

In: Taylor Clinical ap-

in

Hepatology

in dogs:

sensus tember

It-

veins.

and 9.

Padua, 339-392.

1984;

and

has

Doppler

metry

by

of error.

11:625-641. 0, et al.

flow.

KJW,

flow

pulsed

the

Diagnostic

distribution hepatic

18.

Bo-

relationship and flow

dun-

of the (Fig 4). Di-

In:

L, Lab#{244}C.

of Doppler

diastole

found in this condition, whereas in our study abnormal Doppler pattern did not reflect size alterations of HV. As far as other pathologic conditions

vessels.

York: Raven, 1988; 46-75. Rabinovici N, Navot N. The between respiration, pressure,

hormones with liver

8.

blood

Taylor

blood

flow

sources

of an

M, Pomier

vein

154:495-

quantitative

in portal

Dauzat

1985;

Biol 1985; N, Nishida

1987;

veloci-

of blood

in the

PN,

T,

Doppler venous

and

application

enterology

inferior

Hemodynamics. PN, Wells PNT.

plications

are commonly

Abdominal

ultrasound in gastroenterology. aly: Piccin-Butterworths,

mecha-

nism seems the most logical explanation for the flat HV waveform. We also noted a correlation of this pattern with the changes in arterial splanchnic impedance and portal vein velocity. However, this comrelation does not imply a causal effect, and it does not seem likely that these latter phenomena directly contribute to the observed change in HV waveform. Regarding other liver diseases that possibly affect the normal Doppler US pattern of HVs, a magnification of the phasic oscillations causing a W-

Med F, Ban

troenterology

Bolondi bondi

4.

accuracy

system

Nakayama

cineangiographic

Radiology Measurement

Burns

H,

of portal

ultrasound:

JL,

for Br J Sung

Pulsed

with

498. Gill RW.

flowmetry

1989; 43-72. CS, Kotler MN, AS,

diagno-

Koen K.

as a criterion

Ultrasound Moniyasu

Dawson

oesophagus

vanices. M,

ment of portal 14:579-588.

echoangio-

64:1018-1025. 3.

F, Okuda

plex

vena caval ultrasonography: normal and abnormal findings and its use in assessing right-heart function. Circulation 1981;

tole.

K, Saito

Nomura

measurement.

in examina-

FS. Ultrasound

of digestive

Verlag, Mintz

guide

abdomen:

In: Weill

drian

dias-

Ohnishi

Clinical

Weill graphy.

the

middle HV shows a characteristic W-wave pattern consisting of speedy forward flow with an abrupt deceleration and subsequent reversed

1.

oesophageal 60:646-649.

ty: comparison

of func-

IM,

of the

1983;

flow

References 4.

Murray-Lyon

Transection

bleeding

15.

Figure

RWH,

et ab.

adjunctive are some-

severity U

Pugh

of portal in

lowering

Gastroenterol

splanchnic

vascular

Int

1988

1(suppl

1): 1017.

516

Radiology

#{149}

February

1991