ClinicalRadiology(1993)48, 311 315

Duplex Doppler Measurements of the Portal Vein in Portal Hypertension

D. J. LOMAS, P. D. BRITTON, C. B. S U M M E R T O N * and C. A. SEYMOUR~"

Departments of Radiology and *Medicine, Addenbrooke's Hospital and Cambridge University, Cambridge, and Department of Clinical Biochemistry, St George's Hospital Medical School, London
In order to assess the validity of quantitative duplex Doppler measurements of portal vein flow, 10 patients with proven diffuse liver disease and portal hypertension were examined serially by two independent observers over a 3 month period. Multiple measurements of the portal vein were made using a consistent technique in an attempt to minimize observer errors. One patient proved unsuitable for ultrasound examination. In the remaining nine patients the intra-observer portal vein measurements for one observer (19 paired examinations) correlated significantly for flow velocity (r = 0.80, P < 0.001) and derived bulk flow (r = 0.54, P < 0.02) but not for cross-sectional area (r--0.27, P > 0.05). The intra-oliserver measurements for the second observer (11 paired examinations) were significantly correlated for cross-sectional area (r = 0.64, P < 0.05) and derived bulk flow (r = 0.61, P < 0.05) but not for flow velocity (r = 0.5, P > 0.05). The inter-observer measurements (28 paired examinations) of the portal vein were all highly significantly correlated for cross-sectional area (r=0.63, P < 0.001), flow velocity (r = 0.79, P < 0.001) and derived bulk flow (r = 0.73, P < 0.001). These results suggest that this quantitative Doppler technique may be valid for the serial study of portal vein flow in selected groups of patients with diffuse liver disease and portal hypertension. Lomas, D.J., Britton, P.D., Summerton, C.B. & Seymour, C.A. (1993). Clinical Radiology 48, 311-315. Duplex Doppler Measurements of the Portal Vein in Portal Hypertension

Accepted for Publication 20 May 1993

Alterations of portal venous blood flow are common in diffuse liver disease such as cirrhosis. Quantifying these changes is likely to provide important information about the likelihood of ensuing complications such as the development of oesophageal varices and may be useful in monitoring the effects of therapy. In the past, obtaining this information has usually required invasive techniques such as indirect vascular catheterization, direct splenic puncture or !aparotomy combined with electromagnetic flowmetry [1,2]. All of these methods m a k e the serial monitoring of portal hypertension impractical, particularly in patients who may have a related coagulopathy. Duplex Doppler ultrasound offers an apparently ideal non-invasive alternative, providing both qualitative and quantitative information on hepatic and portal venous system blood flow. The value o f this technique in the diagnosis of vascular occlusions related to the liver and portal venous system is well established [3,4]. Useful qualitative information can be provided in patients with portal hypertension by the detection of porto-systemic collateral vessels and by the demonstration of alterations in blood flow direction [5]. In addition analysis of the Doppler spectral shift in the portal vein, with appropriate corrections allows quantitative assessment of mean blood velocity and when combined with measurement o f crosssectional area portal vein bulk flow may be estimated [6,7]. Although several workers have utilized this technique for studying variations in physiological flow [8,9] and the effects of drug therapy [ 10] the validity of the quantitative Correspondenceto: Dr David J. Lomas, UniversityDepartment of Radiology, Level5, Addenbrooke'sHospital, Cambridge CB2 2QQ.

technique has been questioned [11,12]. Comparisons with invasive techniques under controlled conditions [13] and in animals [ 14,15] have been favourable but many of the criticisms relating to the numerous operator-dependent variables and the potential for cumulative measurement errors remain. Both the assessment of flow velocity and vessel diameter may be subject to measurement errors owing to a combination o f systematic and operatorrelated factors [7]. The differing calibration, sensitivity and built in analytical methods of the various commercial machines in addition to varying patient body habitus and the influence of liver pathology on sound transmission have all been postulated as further sources of potential errors [11,16,17]. Many of these objections to the use of duplex Doppler ultrasound for the assessment of portal vein flow may be overcome by carefully performed serial examinations with a single machine using relative rather than absolute interpretation o f the results. However, operator variability remains and an assessment of inter- and intraobserver variation is required before valid conclusions can be made using this technique. Such studies have been performed in normal volunteers [18] but little work has been undertaken in patients with known portal hypertension [19]. This study investigates the inter-observer and intra-observer variation of duplex Doppler portal vein measurements in a group of patients with proven liver cirrhosis and portal hypertension. P A T I E N T S AND M E T H O D S Ten patients with known diffuse liver disease and previously documented oesophageal varices were

312

CLINICALRADIOLOGY

Fig. 1- A longitudinal image of the portal vein from an anterior approach which demonstrates the location (adjacent to the right main 9 branch of the hepatic artery) used for measuring portal vein diameter. recruited for this study, which formed the initial stage of a therapeutic trial investigating the effect of drug therapy on portal blbod flow. Although some of the patients underwent endoscopy during the study no sclerotherapy treatment was undertaken. Patients who were unsuitable because of immobility, encephalopathy, major systemic disease or a known hepatic vascular occlusion were excluded. The patients were approached initially by letter and at a subsequent outpatient visit. Ten patients were recruited of whom one was later excluded, because it was not possible to obtain a satisfactory view of the portal vein. Ethical permission was obtained from the local hospital ethical committee and each patient gave informed written consent. The patients were examined following a 12 h fast at 1 pm and 3 days later at 8 am. These paired examinations were repeated on two further occasions at 4 week intervals. An Aloka SSD 650 ultrasound machine with a 3.5 MHz curvilinear array was used for all the examinations. The examinations were performed with the patient supine and during gentle respiration, if respiratory motion prevented adequate measurement the vein was examined during suspended respiration. A transhepatic approach was used to visualize the main portal vein, and all measurements were made at the point where the right branch of the hepatic artery crossed the main portal vein (Fig. 1). Two different transducer approaches were used for velocity assessment and where possible these were at 90 ~ to each other in the transverse plane in order to maximize sampling of the full range of portal vein velocities. If this was not possible because of an unsuitable Doppler angle one set of measurements was taken using an anterior approach and another set at the furthest acceptable approach lateral to this. Recordings of the Doppler signal were made for at least 2 s. Two velocity traces were obtained using each approach (using an incident beam angle of less than 60 ~) and on each trace measurements of the peak velocity and mean peak velocity (the latter requiring the operator to mark the top of the trace with a trackerball) were obtained (Fig. 2). The majority of examinations were made with the largest available sample volume of 10 mm with the aim of achieving uniform insonation of the blood flow in the portal vein.

Fig. 2 - A measurement of portal vein flow velocitytaken at the same location as the diameter measurement (Fig. 1). The upper margin of the velocity profile has been marked giving a mean peak velocity of 16.6 cm/s. The antero-posterior diameter measurements were made using exactly the same approach that was used for the anterior velocity measurements. The lateral diameter was measured using an approach at 90 ~ lateral to the first measurement, even if this second approach was unsuitable for velocity measurement because of the beam angle. The vessel diameter was maximized in longitudinal view and the image enlarged before two diameter measurements were made from each approach. The vessel area was calculated based on an elipse formula using the mean of the measurements in the two orthogonal directions. This was combined with the mean of the four average peak velocity measurements to calculate a flow value. Intra-observer variation was assessed by comparing the examinations performed on the same patients at 3 day intervals. Inter-observer variation was assessed by comparing the observations made by the two observers taken at the same occasion. The variations were analysed using correlation coefficients and probability tables. RESULTS There were nine patients with a mean age of 52 years (range 34-70 years), two were female and seven male. All nine had histologically proven diffuse liver disease" cryptogenic cirrhosis in two patients and alcoholic cirrhosis in the remaining seven. Two patients did not complete the overall study withdrawing after the second pair of examinations. Both observers could not be present at all the examinations but 28 examinations were available for inter-observer analysis. Nineteen paired examinations were available for the analysis of intra-observer variation for the first observer and 11 paired examinations for the second observer. The intra-observer measurements (Figs 3, 4) correlated significantly for portal vein bulk flow for both observers (Observer l, P < 0.02; Observer 2, P < 0.05), portal vein flow velocity for the first observer (P<0.001), and for portal vein cross-sectional area for the second observer (P < 0.05). However, the intra-observer measurements for cross-sectional area for the first observer and flow velocity for the second observer did not achieve statistical significance. Scatter plots o f the data for the inter-observer correlations are presented in Fig. 5. Significant correlations were found between observers for portal vein crosssectional area, velocity and flow (P < 0.001).

PORTAL VEIN IN PORTAL HYPERTENSION

1.4t= m 0r o

313

(a)
o [] y = 0.63 + 0.23x r = 0.286 NS
[]

1.4
1.2 1.0

(a)

1.2

-~ u o ~L~ ~o" o~

[] o

1.0

E~
0.8

~
[]

y = 0.14 + 0 . 8 0 x r = 0.636 P < 0.05

O
[] ~ ~ ' o o

,,

~ ~ . / ~ - - n
mO
~ C

E~
0.6
0 0

[]
i

o [] [] I 0.8
I I 1.0 I I 1.2 i I 1.4 0.2

0.4
i

0.4 0.2

I 0.4

I 0.6

First m e a s u r e m e n t o f portal vein cross-sectional area (square cm)

Ol 0.4

I 0.6

I 0.8

I 1.0

(a)
r

First m e a s u r e m e n t o f p o r t a l vein c r o s s - s e c t i o n a l area ( s q u a r e cm)

20

(a)

,b,
r-

16 > 15

>
u 00

(b)

18
16
[]

o o.

[]

[]

I: o o. '~

[]
14

O / / / / / / ' / 0

E.~ ~'5 ~o

14

12

E.~- 13 P'6 9 o ~

E
tO 10
~ ~ I
8

~
/
/ 10 u [, i 12

_ ~
o []

y=-1.64+~.16x
-

12

P < 0.001 ], i 14 , , 16 i 18

E
"O

[] []
[] I I i I i I i

y = 9.11 + 0 . 3 0 x r = 0.500 NS

c o r ~o

11

First m e a s u r e m e n t o f portal v e i n v e l o c i t y (cm/s)

10

12

14

16

18

20

(b) 800
o .E > 700 600
o

First m e a s u r e m e n t o f p o r t a l vein v e l o c i t y (cm/s)

(b)
0

(c)
y = 204.18 + 0 . 4 8 x r = 0.539 P < 0.02

=o
C

500 -

(c)

[]

>
0~

400

[]

[]

3O0
[] 400
o0

3
m o~ o

-o c o
Q.)

200

300 200 100

~ @ o o
i I

[]

.~I

c o (o o0 100

y = 145.23 + 0 . 5 2 x r = 0.607 P < 0.05 n

I

200

300

400

500

600

700

i0 200

300

I 400

I 500

600

First c a l c u l a t i o n o f p o r t a l v e i n f l o w ( m l / m i n )

(c)
Fig. 3 - Scatter plots o f the data and intra-observer correlations f o r the first observer. (a) Cross-sectional area; (b) flow velocity; (c) bulk flow.

First c a l c u l a t i o n o f p o r t a l v e i n f l o w ( m l / m i n )

(c)
Fig. 4 - Scatter plots of the data a n d intra-observer correlations for the second observer. (a) Cross-sectional area; (b) flow velocity; (c) bulk flow.

DISCUSSION

Duplex Doppler ultrasound offers a convenient noninvasive method of quantitative assessment of portal venous blood flow. The initial enthusiasm for this technique has been replaced by cautious acceptance following recognition of the numerous potential sources of error. Widespread clinical application is unlikely to occur until the reliability and reproducibility of this technique have been proven. In our study the duplex Doppler technique could not be

performed on one patient as it was impossible to visualize the portal vein adequately via a transhepatic approach, owing to his large size and attenuation of the ultrasound beam. Other workers have encountered similar problems and it is possible that more widespread use of this technique would be restricted to a relatively selected group of patients which may vary depending on the indigenous population [16]. Improvements in machine technology may overcome this potential limitation. Although not explicitly stated in many of the published studies an intrahepatic, anterior abdominal approach to

314
"o

CLINICAL RADIOLOGY The site of measurement in the portal vein has varied between workers. The confluence of the splenic and superior mesenteric vein, the mid-point of the portal vein and the point at which the right branch of the hepatic artery crosses the portal vein have all been described as suitable sites for assessment of portal vein size[ 12,20,21]. The majority of Doppler studies have used the 'midpoint' of the main portal vein. In our experience the most reliable site for visualizing the portal vein at an acceptable Doppler beam angle has been the point at which the right hepatic artery crosses the portal vein. In order to ensure that the diameter measurement was taken at the same site as the Doppler measurement we have used a longitudinal angled view of the portal vein for both the antero-posterior and lateral measurements. This is at variance with other workers who have rotated the transducer and taken cross-sectional views of the portal vein for the diameter measurements. This proved an impractical technique in our study using an intercostal view and curvilinear transducer, combined with the frequent portal vein tortuosity. By using two approaches, where possible at 90 ~ to each other, recording multiple Doppler measurements, and appropriately large sample volumes, we hoped to achieve 'uniform ins 9 of the portal vein and minimize velocity measurement errors. This was compromised in some of the patients where tortuosity of the portal vein meant that suitable Doppler angles could only be obtained at sites which were less than 90 ~ to each other. This aim was further limited by the l0 mm maximum sample volume available for Doppler measurement. Clearly in patients whose portal vein diameters were in excess of l0 mm 'uniform ins 9 would be impossible even without correcting for the further compromising effect of the Doppler angle. This limitation of the technique may have been partially overcome by taking the measurements during gentle respiration. The resulting slight movement of the portal vein during the recording may have improved the uniformity of coverage of the portal vein flow velocity profile. When respiratory movement prevented adequate sampling of portal vein velocity signal the patients were examined during suspended respiration. We have assumed that this has no significant effect on the measurements provided they were taken immediately after suspending respiration. A longer sampling time might reduce any respiratory effect and also any transmitted right heart variations, but in practice this would have been difficult to achieve in these patients. The calculation of mean peak velocity on our ultrasound machine required tracing manually along the peak of the Doppler trace using a trackerball device. This introduced a further potential source of error that might be overcome on other ultrasound machines where this measurement has been automated. The analysis of the inter-observer results for crosssectional area, velocity and derived bulk flow were all highly significantly correlated ( P < 0.001). Although the overall number of observations (28 paired examinations) is relatively low for such analysis, we believe this is an acceptable statistical result. The analysis of the intraobserver results is impaired by smaller numbers, but demonstrated significant correlations between all the measurements except for the diameter measurements of the first observer and flow velocity measurements for the second observer. Intra-observer variation is the most

==

==

1.4 f (a) 1.2 1.0 D []n O0 D~ []

EE
.90

0.8

0.6L- ttl

o
/ / o
t I ~

e-o

[]

0.4 ~ "
o

[] []
01 i
I t

y = 0,25 + 0.62;< r~- 0:632 P < 0.001

I I [
i

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Portal vein cross-sectional area measured by observer one (square cm)

(a) ;>
W

22(b) 20 13 18 16 14 12 10
8 i

o >,.
"O

,=

E~ ,Z. o 9 j~ n
o
> r >

[] O / [] []
I i I

I~""-0

rm

13

0 ~ = 0.13 + 1.01x r = 0.787 P < 0.001

i [ i I i I

t~ O rt

10

12

14

16

18

Portal vein velocity measured by observer one (cm/s)

(b) 'E 800 (c) o

.E

600
~

[]

.~ 400

5o
[] [] y = 42.29 + 0.84x r = 0.731 P < 0.001
I I I t I t I

"~ o,,+o > t~ o Q,.

200

100

200

300

400

500

600

"700

800

Portal vein flow calculated from measurements of observer one (ml/min)

(c)

Fig. 5 - Scatter plots o f the d a t a a n d inter-observer correlations. (a) Cross-sectional area; (b) flow velocity; (c) b u l k flow.

the portal vein appears to have been used. This avoids the problem of sound attenuation by hepatic steatosis or fibrosis but is prone to interruption by bowel gas and it may be impossible to achieve a satisfactory angle for Doppler measurement. We have used a transhepatic approach in all the patients, except one in whom the liver was so small that this was not feasible.

PORTAL VEIN IN PORTAL HYPERTENSION likely c a u s e f o r t h e lack o f c o r r e l a t i o n f o r these m e a s u r e m e n t s ; h o w e v e r , the h i g h l y significant c o r r e l a t i o n o f t h e d e r i v e d b u l k f l o w v a l u e s a n d o f all the i n t e r - o b s e r v e r m e a s u r e m e n t s raises t h e p o s s i b i l i t y o f a n o t h e r c a u s e . A s the two measurements for the intra-observer assessments w e r e c a r r i e d o u t 3 d a y s a p a r t a n d at slightly d i f f e r e n t t i m e s o f d a y (9 a m a n d 1 p m ) t h e p o s s i b i l i t y o f p h y s i o l o g i cal v a r i a t i o n in c r o s s - s e c t i o n a l a r e a a n d flow v e l o c i t y m u s t be c o n s i d e r e d . A g a i n s t this e x p l a n a t i o n is t h e observation that a reduction of physiological variability in p o r t a l v e i n size o c c u r s in p o r t a l h y p e r t e n s i o n [22]. O u r s t u d y s u p p o r t s the c o n c l u s i o n o f o t h e r s t h a t f o r l a r g e g r o u p s o f p a t i e n t s w i t h k n o w n diffuse liver disease, d u p l e x D o p p l e r u l t r a s o u n d is a v a l i d t e c h n i q u e f o r t h e serial a s s e s s m e n t o f p o r t a l v e i n f l o w f o r p h y s i o l o g i c a l a n d t h e r a p e u t i c studies [10,12,17,19]. H o w e v e r , it is i m p o r t a n t to e m p h a s i z e t h a t t h e m e t h o d m a y n o t be p o s s i b l e in all p a t i e n t s a n d t h a t t h e r e is n e e d f o r a m e t i c u l o u s t e c h n i q u e at a r e p e a t a b l e site in o r d e r to o b t a i n a c c e p t a b l e results.
REFERENCES

315

9 I0 11 12 13 14 15

16

17

1 Moreno AF, Burchell AR, Rousselot LM, Panke WF, Slafksy SF, Burke JH. Portal blood flow in cirrhosis of the liver. Journal of Clinical Investigation 1967;46:436-445. 2 Formon J J, Warren WD. Hemodynamic studies in portal hypertension. Annual Review of Medicine 1969;20:277-290. 3 Becker CD, Cooperberg PL. Sonography of the hepatic vascular system. American Journal of Roentgenology 1988;150:999 1005. 4 Lomas D J, Britton PD, Farman P, Coulden R, Egan A, Jamieson GN et al. Duplex Doppler ultrasound for the detection of vascular occlusion following liver transplantation in children. Clinical Radiology 1992;46:38-42. 5 Patriquin H, Lafortune M, Burns PN, Dauzat M. Duplex Doppler examination in portal hypertension: technique and anatomy. American Journal of Roentgenology 1987;149:71-76. 6 Gill RW. Pulsed Doppler with B-mode imaging for quantitative blood flow measurement. Ultrasound in Medicine and Biology 1979;5:223-2,35. 7 Gill RW. Measurement of blood flow by ultrasound: accuracy and sources of error. Ultrasound in Medicine and Biology 1985; 11:625641. 8 Brown HS, Halliwell M, Qamar M, Read AE, Evans JM, Wells

18

19

20 21

22

PNT. Measurement of normal portal venous blood flow by Doppler ultrasound. Gut 1989;30:503-509. Tsunoda T, Ohnishi K, Tanaka H. Portal hemodynamic responses after oral intake of glucose in patients with cirrhosis. American Journal of Gastroenterology 1988;83:398 403. Gaiani S, Bolondi L, Fenyses D, Zironi G, Rigamonti A, Barbara L. Effect of propranolol on portosystemic collateral circulation in patients with cirrhosis. Hepatology 1991;14:824-829. Burns P, Taylor K, Blei AT. Doppler flowmetry and portal hypertension. Gastroenterology 1987;92:824-826. Bolondi L, Gaiani S, Luigi B. Accuracy and reproducibility of portal vein flow measurement by Doppler ultrasound. Journal of Hepatology 1991;13:269-273. Walter JP, McGahan JP, Lantz BM. Absolute flow measurements using pulsed Doppler US. Radiology 1986;159:545-548. Dauzat M, Layrargues GP. Portal vein blood flow measurements using pulsed doppler and electromagnetic ftowmetry in dog: a comparative study. Gastroenterology 1989;96:913 919. Moriyasu F, Ban N, Nishida O, Nakamura T, Miyake T, Uchino H, Kanematsu Y, Koizumi S. Clinical application of an ultrasonic duplex system in the quantitative measurement of portal blood flow. Journal of Clinical Ultrasound 1986;14:579 588. Sabba C, Ferraioli G, Sarin S, Groszmann RJ, Taylor KJW. Feasibility for Doppler flow measurements in splanchnic vessels in non-selected patient populations and cirrhotics. Gastroenterology 1989;96:432. Zoli M, Marchesini G, Cordiani MR, Pisi P, Brunon A, Trono A et al. Echo Doppler measurements of splanchnic blood flow in control and cirrhotic subjects. Journal o f Clinical Ultrasound 1986; 14:429435. de Vries PJ, van Hattum J, Hoekstra JBL, de Hooge P. Duplex Doppler measurements of portal venous flow in normal subjects: inter- and intra-observer variability. Journal of Hepatology 1991;t3:358 363. Sabba C, Weltin GC, Cicchetti DV, Ferraioli G, Taylor KJW, Nakamura T, Moriyasu F, Groszmann RJ. Observer variability in echo-Doppler measurements of portal flow in cirrhotic patients and normal volunteers. Gastroenterology 1990;98:1603-1611. Webb LJ, Berger LA, Sherlock S. Greysclae ultrasound of the portal vein. Lancet 1977;2:675-677. Bolondi L, Gandolfi L, Arienti V, Caleni GC, Corcioni E, Gasbarrini G, Labo G. Ultrasonography in the diagnosis of portal hypertension: diminished response of portal vessels to respiration. Radiology 1982;142:167-172. Goyal AK, Pokhama DS, Sharma SK. Effects of a meal on the normal and hypertensive portal venous system: a quantitative ultrasonographic assessment. Gastrointestinal Radiology 1989; 14:t64-166.