Journol of Gasrroenterologyand Hepatology (1991) 6 , 368-373
LIVER AND BILIARY Diagnostic utility of ultrasonography in hepatic venous outflow tract obstruction in a tropical country A. ARORA, M. P. SHARMA, S. K. ACHARYA, S. K. PANDAANDM. BERRY Department of Gastroenterology, Pathology and Radiology, All India Institute of Medical Sciences, Ansari Nagar, New Delhi, India Abstract The present study was undertaken to define the role of ultrasonography (US) in screening and diagnosis of hepatic venous outflow tract obstruction. Forty-five consecutive patients clinically suspected to have hepatic venous outflow tract obstruction were included in the study for screening by US and for assessment of patency or block in the hepatic vein (HV) and/or inferior vena cava (IVC). Four patients were excluded from the study. Eleven patients had a diagnosis other than hepatic venous outflow tract obstruction and all these patients were found to have patent HV and IVC. Thirty patients were finally diagnosed to have hepatic venous outflow tract obstruction. Using US, as a screening test 27 (90%) out of 30 such cases were correctly identified as cases of hepatic venous outflow tract obstruction and in these cases the site of block in hepatic venous outflow tract (major HV andor IVC) was correctly diagnosed in 90% of the cases. Our results indicate that US is a sensitive and accurate test and should be the initial investigation for screening and identifying the site of obstruction in patients with hepatic venous outflow tract obstruction.
Key words: hepatic vein block, hepatic venous outflow track obstruction, inferior vena cava block, ultrasonography.
INTRODUCTION Hepatic venous outflow tract obstruction is a disease caused by obstruction to the venous outflow tract of the liver which includes obstruction of the sublobular hepatic and central veins (veno-occlusive disease, VOD),obstruction to major hepatic veins (HV), or to the intra- and suprahepatic portion of the inferior vena cava (IVC). The obstruction may be due to intrinsic abnormalities of the outflow tract,'-' systemic diseases predisposing to thrombosisY6-'effect of toxins: or trauma." Once the obstruction sets in, the degree of liver damage and clinical manifestations are variable. l1 Unrelieved obstruction results in development of cirrhosis" and its sequelae. Thus, early diagnosis of the disease may help in prompt and specific therapeutic intervention which may be valuable in preventing and reversing the fibrotic changes in the liver.'3314 Traditionally, the diagnosis is made by liver biopsy and a n g i ~ g r a p h y . ' ~ ~Recently ' ~ ~ ' ~ ultrasonography (US),'7-24 computerized tomography (CT) scan,18i21324-26 and magnetic resonance imaging (MRI)" have also been used for the diagnosis of hepatic venous outflow tract obstruction, but all the reports have included only a few patients and no report has specifically assessed the role of US in screening and diagnosis of hepatic venous outflow tract obstruction. We
conducted a prospective study to evaluate this, and to assess its role in delineating the exact site of block in the outflow tract of liver.
MATERIALS A N D METHODS All the patients admitted to the gastroenterology ward of our hospital between January 1988 and January 1990 with any one of the following group of symptoms were clinically suspected to have hepatic venous outflow tract obstruction and hence were included in the study: (i) the rapid development of tender hepatomegaly, pain in the right hypochondrium andor ascites in the absence of any cardiac disease or past history of liver disease; (ii) the presence of venous collaterals on the back or abdominal wall with pedal oedema, stasis pigmentation or ulceration; or (iii) abdominal discomfort of weeks to months in duration followed by the appearance of hepatomegaly and ascites with evidence of portal hypertension. Complete work-up included a full history with special reference to oral contraceptives and herbal drugs, a thorough physical examination, full blood count, chest X-ray, urine analysis and electrocardiogram. An upper gastrointestinal endoscopy was performed for diagnosis of gastro-oesophageal varices.
Correspondence: Dr M. P. Sharma (Associate Professor), Department of Gastroenterology, All India Institute of Medical Sciences Ansari Nagar, New Delhi 110029, India. Accepted for publication 21 November 1990.
Ultrasonography and hepatic venous outflow obstruction
Ultrasonographic procedure Patients All patients underwent ultrasonographic screening before liver biopsy and angiography. The US examination was performed by the same individual (MPS) without prior knowledge of the diagnosis, and the findings were compared with subsequent venographic findings for the site of block in the hepatic venous outflow tract. The ultrasound assessment included the patency and calibre of HV and IVC and the pattern of their drainage into the IVC and right atrium. US signs of portal hypertension (PHT) (splenomegaly, dilated portal and splenic vein and the presence of collaterals) were also noted.
Controls Thirty patients, 10 with cirrhosis of the liver (six alcoholic and four post-necrotic), 10 with other liver diseases (eight viral hepatitis cases and two amoebic liver abscess cases), and 10 normal healthy volunteers were studied as controls to determine the normal patterns of visualization and dimensions of HV and IVC. The HV size was measured at a level just before its entry into the inferior vena cava, with the patient holding the breath at the end of gentle inspiration, by placing the US probe parallel to the right subcostal margin. IVC measurements were done in a similar mannei at the level of entry of the hepatic veins into the IVC, with the US probe kept in the right parasagittal position belou the costal margin.
I . Ultrasound critm'a for diagnosis of site of block A diagnosis of hepatic venous outflow tract obstruction was made when HV, or IVC, or both were blocked, as follows: (1) HV block alone: (a) visualization of at least two major hepatic veins of abnormal calibre and patency or nonvisualization of at least two hepatic veins; (b) dilated and/or tortuous hepatic veins, with the absence of demonstration of their confluence close to their opening into the IVC; these features of HV block were associated with the presence of communicating veins between the major hepatic veins (when they were visualized); in the absence of demonstrable hepatic veins, thin tortuous veins without any proper direction were also seen; these abnormal intrahepatic interconnecting veins were denoted intrahepatic venous collaterals. . (2) IVC block alone: (a) segmental narrowing of the IVC in the intrahepatic or suprahepatic portion with proximal dilatation (Fig. la); (b) dilated and patent hepatic veins draining into the IVC with intrahepatic collaterals (Fig. lb). (3) Combined IVC and HV block: see above, items 1 (a), or (b), and 2 (a) (shown in Fig. 2). I I . Non-ultrasonographic criteria for hepatic venous outflow tract obstruction and the site of block (1) Diagnosis of hepatic venous outflow tract obstruction: This Was made by finding typical histoPathological changes on liver biopsy. l6 (2) Diagnosis of site of block: The Patients with a histological diagnosis of hepatic venous outflow tract obstruction underwent inferior vena cavography to assess the patency of the IVC, by a percutaneous femoral vein puncture and
~i~~ 1 (a) ~l~~~~~~~~~~~showing marked inferior vena cava (IVC) narrowing with proximal dilatation (arrows) prior to its entry into right atrium (RA). (b) Inferior vena cavogram of the same patient showing markedly dilated inferior vena cava, dilated hepatic veins and collaterals bypassing the obstruction. A final diagnosis of isolated inferior vena cava block was made.
A. Aroru et al.
catheterization. If hepatic vein catheterization was not SUCcessful during inferior venography,a functional hepatogram was performed to assess the patency of the hepatic veins.” The diagnosis of the site of block was established on demonstration of block in the HV and/or the IVC, with evidence of collaterals (Fig. lb) proximal to the obstruction. Diagnosis of VOD was made in the presence of histological evidence of hepatic venous outflow tract obstruction with US and venographic demonstration of a patent major HV and IVC (Fig. 3a and 3b).
RESULTS Between January 1988 and January 1990, 45 patients clinically suspected of having hepatic venous outflow tract obstruction attended gastroenterology wards of our hospital (with the total number of admissions during the said period being approximately 2300). Four patients were excluded from the study (in two patients, tense ascites precluded a US evaluation and two patients refused angiography and liver biopsy). Out of the remaining 41 patients, liver histological evidence of hepatic venous outflow tract obstruction was obtained in 30. In 11 patients without histologic evidence of hepatic venous outflow tract obstruction, further investigations confirmed the correct diagnosis (Table 1). Of 30 patients finally diagnosed as suffering from hepatic venous outflow tract obstruction, fifteen were male and fifteen female. Their mean age was 30.5 k 11.68 years (1 150 years). Table 1 Analysis of 45 patients included in the study 1 Total no. patients included in the study 2 Complete work-up not possible (a) Tense ascites precluded US evaluation (b) Refusal to undergo liver biopsy and angiography 3 Total no. patients in whom diagnosis was established (A) hepatic venous outflow tract obstruction (B) others (a) non-cirrhotic portal fibrosis (b) tubercular peritonitis (c) lymphoma (d) membranous glomerulonephritis (e) constrictive pericarditis
Figure 2 Ultrasonogram showing narrowing of inferior vena cava (IVC) prior to its entry into right atrium (RA). In addition normal hepatic veins are not visualized and are instead replaced by multiple collaterals (COL). A diagnosis of combined hepatic vein and I v c block was made and subsequently confirmed by angiography.
45 4 2 2 41
30 11 5 1 1 1 3
US evaluation of HV and IVC in conttols The results of us evaluation in 30 Patients of the control group showed that there was no difference in the sizes of the IVC and the HV in these three groups of subjects (Table 2).
rqure I) (a) ultrasonogram snowmg all three patent major hepatic veins and inferior vena cava. (b) Inferior vena cavogram on digital
subtraction angiography, showing patent inferior vena cava (upper half) and patent hepatic vein (lower half). Based on histology and signs of portal hypertension, a final diagnosis of veno-occlusive disease was made.
Ultrasmgraphy and hepatic venous outflow obstruction
Based on these findings and other a HV greater than (or equal to) one cm in diameter and an IVC greater than (or equal to) 1.5 cm in diameter were considered to be abnormally dilated. Two or more hepatic veins were seen in 29 (96.6%) out of 30 patients. Similar results have been reported by Cosgrove et aL3'
The accuracy of US in delineating the patency of IVC and HV Of the 41 patients with a definite diagnosis (hepatic venous outflow tract obstruction: 30, other diseases: ll), US showed a patent IVC and HV in 14 and obstruction of either the HV, or the IVC, or both in the remaining 27 patients. In the former group of 14 patients, patency of the hepatic outflow tract was confirmed by angiographic demonstration of patent major HV and IVC in three patients with VOD, and a lack of any liver histological evidence of hepatic venous outflow tract obstruction in the remaining 11 patients with confirmed diagnoses other than hepatic venous outflow tract obstruction (Table 1). In the latter group of 27 patients, angiography, like US, revealed blockage of the IVC and/or the HV. Thus US correctly predicted the patency or blockage of the hepatic venous outflow channels in all patients, providing a 100% specificity and sensitivity.
US diagnosis of combined IVC and HV block, angiography revealed isolated obstruction of IVC in one and isolated HV block in the other two patients. The sensitivity and specificity of US in diagnosing the individual site of obstruction in the hepatic venous outflow tract is shown in Table 3.
US diagnosis of hepatic venous outflow tract obstruction irrespective of the site of block (screening of the hepatic venous outflow tract) The individual ultrasonographic features of each patient with hepatic venous outflow tract obstruction, irrespective of the site of obstruction (Table 4), were analysed. All except those with VOD (n = 3) had ultrasonographic features of intrahepatic vein-to-vein collaterals. In contrast, none of the healthy or diseased controls had intrahepatic collaterals. The presence of intrahepatic venous collaterals along with US evidence of portal hypertension, irrespective of the clarity or precision of the site of block, appeared to be diagnostic of hepatic venous outflow tract obstruction. Combining these features with abnormalities in the HV and/or IVC resulted in a positive and negative predictive value of US in the diagnosis of hepatic venous outflow tract obstruction of 100% and 78.%'0, respectively.
DISCUSSION The accuracy of US in delineating the site of block in hepatic venous outflow tract obstruction Three patients (all VOD) had patent HV and IVC on both US and angiography. US correctly diagnosed the exact site of block in 24 (90%) out of the remaining 27 patients with a block in the HV and/or IVC. In three patients with incorrect Table 2 US evaluation of hepatic vein and inferior vena cava sizes in controls
Patient group Cirrhotics Other liver diseases Normal healthy controls
HV diameter (mm)*
6.13 k 1.50 5.73 _+ 1.91 5.7 f 1.88
12.9 5 1.20 12.0 5 1.10 13.1 k 0.99
*Mean k s.d.
In our centre, patients presenting with features suggestive of hepatic venolis outflow tract obstruction undergo liver biopsy for the diagnosis, followed by angiography to determine the site of obstruction. Obviously both these procedures require hospitalization and are not without risk.31 They are also time consuming, expensive and require expert radiological facilities which are not widely available in this country, where there is a high incidence of this d i s e a ~ e . ~ ~ ~ ~ * US predicted the disease correctly in 27 (90%)of 30 patients with a diagnosis of hepatic venous outflow tract obstruction, and carried a 100% positive predictive value. The low negative predictive value (78.5%) in the present study was because of an absence of intrahepatic collaterals and abnormalities of the major HV or IVC in patients of VOD. Thus, US will diagnose hepatic venous outflow tract obstruction in all patients when it is due to IVC and/or HV block, but in cases of VOD it will only reveal signs of portal hypertension.
Table 3 Diagnostic accuracy of US and angiography for the site of block in patients of hepatic venous outflow tract obstruction S. no. 1
2 3 4 5
Site of block HV block alone IVC block alone IVC and HV combined block Correct diagnosis of site of block in HV and/or IVC Patent HV and IVC (VOD)
Specificity (%) 100 100 88.4
*Sensitivity was calculated for the individual site ot block by using the formula: true positive x loo. sensitivity = true positive + false negative
**Specificity was calculated for the individual site of block by using the formula: specificity = true negative x 100. true negative + false positive
A . Arora et al.
Table 4 US findings in patients with hepatic venous outflow tract obstruction (n = 30)
S. no. 1
Site of block VOD(n=3)
HV and IVC combined block (n = 7)
No. patients (Yo) 3 3
Patent HV & IVC Evidence of PHT Suprahepatic narrowing of IVC with proximal dilation Patent and dilated but not tortuous HV Confluence of HV draining into IVC seen Intrahepatic collaterals present
10 10 10 10
Marked dilation and tortuosity of HV Lack of visualization of at least two HV Lack of visualization of confluence of HV at the entry into IV Multiple intrahepatic collaterals Patent IVC Evidence of PHT Caudate lobe enlargement
4 6 10 10
Lack of visualization of at least two HV and their confluence of entry into IVC Narrowing of IVC (intra- & suprahepatic) Multiple intrahepatic collaterals Evidence of PHT Caudate lobe enlargement
However, the combination of US with histology will diagnose all patients with hepatic venous outflow tract obstruction, including those with VOD. US was accurate in predicting the patency or blockage of the IVC andor HV in all the patients. Correct delineation of the site of obstruction was made by US in 24 of the 27 cases of hepatic venous outflow tract obstruction with block in HV and/or IVC. In three patients the site of block on US was wrongly diagnosed as a combined IVC and HV block, whereas one had only an IVC block, and in whom the HV could not be visualized properly because of the presence of tense ascites; and in the other two cases with only isolated HV block, functionally an enlarged caudate lobe resulted in extrinsic compression of the IVC which was interpreted as a combined block, on US. However, it can be predicted that the accuracy of the delineation of the site of the block would be further improved if US is performed after a large volume paracentesis in patients with tense ascites and if, with experience, caudate lobe enlargement could be identified more accurately. Moreover the recent advent of Doppler studies during US have further established its efficacy in patients with hepatic venous outflow tract obstruction, by the demonstration of high amplitude echoes at the site of obstruction, and by delineating the reversal of blood flow proximal to the ob~tructed-segment.~~ But in a developing country like ours, facilities for’colour Doppler echo are not easily available. None of‘ the recent US st~dies’’-~~‘has evaluated US in systematically diagnosing the site of the block in the hepatic venous outflow tract. All have attempted to document the characteristic changes in the intrahepatic venous collatezals, major hepatic veins, caudate lobe and splenoportal venom, axis in patients with hepatic venous outflow tract obstruc-
10 10 10
7 7 6 6
tion. It must be accepted that despite these excellent results, US does have limitations. The technical problems encountered due to the presence of tense ascites and an enlarged caudate lobe have already been mentioned, and abdominal wall fat and bowel gas can hinder proper ultrasonographic evaluation in a small proportion of patients.33 Moreover the diagnosis of VOD cannot be made by US alone. But US has other advantages, such as simultaneous diagnosis of portal hypertension and intrahepatic space-occupyinglesions causing secondary obstruction of the hepatic venous outflow tract. We conclude that its noninvasiveness, along with a high rate of accuracy, make US the first choice modality of investigation in patients with hepatic venous outflow tract obstruction. Moreover on the basis of our study we presume that US, even in the absence of Doppler studies, is of comparable efficacy to liver biopsy and angiography, and the need for the latter can be obviated in patients with hepatic venous outflow tract obstruction unless VOD is suspected.
REFERENCES 1. SOLOMON V., JAYANTHI V. & MADANGOPALAN N . Coarctation of the inferior vena cava. Trop. Gastroenterol. 1987; 8: 127-42. 2. SIMSONI. W. Membranous obstruction of the inferior vcna cava and hepatocellular carcinoma in South Africa. Gastit:enrerologv 1982; 82: 171-8. 3. KIMURAC., MATSUDA S., KOIEH. & HIROOKA M. Mem-
branous obstruction of the hepatic portion of the inferior vena cava: Clinical study of nine cases. Surgery 1972; 72: 551-9. 4. OKUDAK. Membranous obstruction of the inferior vena cava: Etiology and relation to hepatocellular carcinoma (Editorial). Gastroenterology 1982; 82: 376-9.
Ultrasonography and hepatic venous outfow obstruction
5. GREGORY P. Clinical conference: membranous type of BuddChiari syndrome. 3.Clin. Gastroenrerol. 1984; 6: 81-8. 6. NOBLEJ. A. Hepatic venin thrombosis complicating polycythemia Vera: Successful treatment with a portocaval shunt. Arch. Intent. Med. 1967; 120: 105-8. R., RHODESR. S. & HARTMANN R. C. Throm7. PEYTERMANN bosis in paroxysmal nocturnal hemoglobinuria with particular reference to progressive diffuse hepatic vein thrombosis. Ser. Hematol. 1972; 7 : 709-18. 8. TERABAYASHI H., OKUDAK., N O M ~F., A OHNISHIK. & WONGP. Transformation of inferior vena caval thrombosis to membranous obstruction in a patient with the lupus anticoagulant. Gastroenterology 1986; 91: 219-24. K., AUDRETSCH 9. BROOKS. E. H., MILLERC. G., MCKENZIE J. J. & GRASG. Acute veno-occlusive disease of the liver. Arch. Pathol. 1979; 89: 507-20. 10. CHAMBERLAIN D. W. & WALTERJ. B. The relationship of Budd-Chiari syndrome to oral contraceptives and trauma. Can. Med. Assoc. J . 1969; 101: 97. 11. MITCHELL M. C., BOITNOTT J. K., KAUFMAN S., CAMERON J. R. & MADDREY W. C. The Budd-Chiari syndrome: Aetiology, diagnosis and management. Medicine 1982; 63: 199-218. R. G. F. Occlusion of hepatic veins in man. Medicine 12. PARKER (Baltimore) 1959; 38: 369-402. 13. GUERINM. J. & MEYERP. Need for early thrombolysis in Budd-Chiari syndrome: Correspondence. Gastroenterology 1988; 94: 1109. P. W. & BELLW. R. Thrombolytic treatment for 14. SHOLAR IVC obstruction in paroxysmal nocturnal hemoglobinuria. Ann. Intern. Med. 1985; 103: 539-41. 15. TAVIL A. S., WOODE. J., KREELJ., JONESE. A., GREGORY M. & SHERLOCK S. The Budd-Chiari syndrome: Correlation between hepatic scintigraphy and clinical, radiological and pathological findings in nineteen cases of hepatic venous outflow tract obstruction. Gasrroenrerology 1975; 68: 509-18. 16. CLAIND., FRESTON J., KREELL. & SHERLOCK S. Clinical diagnosis of Budd-Chiari syndrome: A report of 6 cases. A m . 3. Med. 1967; 43: 544-54. R. A., PICKER R. H., FULTON A. J. & LUNZER M. R. 17. COPPER The ultrasonic appearance of the liver in hepatic venous outflow tract obstruction (Budd-Chiari syndrome). 3. Clin. Ultrasound 1982; 10: 35-6. 18. LAWRENCE H. P., BARRY H. G., HILAIRE J. S., FILLY R. A. & GOLDBERG H. J. CT and sonography appearance of hepatic vein obstruction. Amer. J. Roentgen. 1982; 139: 176-8. 19. POWELLJACKSON P. R., MELIAW., CANALESE J., PICKFORD B., PORTMAN B. & WILLIAMS R. Budd-Chiari syndrome:
Clinical patterns and therapy. Quart. 3. Med. (New Series) 1982; 201: 79-88. 20. BAERTA. L., FEVERY J., MARCHALG., GODDEERIS P., WILMSG., PONETTEE. & GROOTEJ. D. Early diagnosis of Budd-Chiari syndrome by computed tomography and ultrasonography: Report of 5 cases. Gasmenurology 1983; 84: 587-95. 21. MAKAUUCHI M., HASEGAWA H., YAMAZAKIS., MORIYAMA N., TAKAYASU K. & OKAZAKIM. Primary Budd-Chiari syndrome: Ultrasonic demonstration. Radiology 1984; 152: 775-9. 22. YUES M., JEAN-MARC D. A., DOMINIQUE-VALLA L., BELIGHITI J. & NAHUM H. Budd-Chiari syndrome: US evaluation. Radiology 1985; 157: 761-4. P., KARANI J., EDE R., MERIREH. 81 23. POWEL-JACKSON WILLIAMS R. Diagnosis of Budd-Chiari syndrome by ultrasound and colloid scintigraphy (Abstr.). Gut 1985; 26: A568. 24. G U ~ TS., A BARTERS., PHILIPSG. W. L., GIBSONR. N. & HOGDSON H. J. F. Comparison of ultrasonogaphy, computed tomography and [99m]TC liver scan in diagnosis of Budd-Chiari syndrome. Gur 1987; 28: 242-7. 25. SPOSITOIN R. P., SIMONOTHI G., SPOSATO S. & GUSUMANO G. CT diagnosis of Budd-Chiari syndrome. J. Cumpt. Assist. Tomography 1981; 5: 366-9. 26. YANG P. J., CLAZERG. M. & BOWERM. Budd-Chiari syndrome: Computed tomography and ultrasound findings. 3. Compt. Assist. T m g r a p h y 1983; 7: 148-50. 27. ARNOLDF. C., RAMCHANDANI P., BLACKM., CAROLINE R. D., RADECKI P. D. & MEEGERP. Magnetic resonance imaging diagnosis of the Budd-Chiari syndrome. Gasrroenterology 1986; 91: 289-95. Y. K. & DILAWARI J. B. Hepatic venogra28. KAURU., CHAWLA phy. Trop. Gasrroenrerol. 1988; 3: 145-9. 29. HAGEM. & WEISSA. Blood vessels. In: US Atlas of RealTime Ultrasound Imaging in Internal Medicine. VCH Verlag, Weinhein. 1986; pp. 246,254. 30. COSCROVE D. O., ARCERP. H. & COLEMAN B. G. Ultrasonic anatomy of hepatic veins.3. Clin. Ultrasound 1987; 15: 231-5. 31. MADDREY W. Hepatic vein thrombosis: 'Budd-Chiari syndrome'. Hepatol. 1984; 4: 445-65. 32. KHUROOM. S. & DUTTAD. V. Budd-Chiari syndrome following pregnancy: Report of 16 cases with roentgenologic, hemodynamic and histologic studies of the hepatic outflow tract obstruction. Amer. J. Med. 1980; 68: 113-21. 33. MUWHYB., FREDERICK H. V., STERNBERG G., THOMAS SHIRES111, MARTINL. G., BERNARDINO M. E. The BuddChiari syndrome: A review. Amer. J. Roentgen. 1986; 147: 915.