Portal vein ligation selectively lowers hepatic cytochrome P450 levels in rats

In rats, surgical creation of a portacaval shunt leads to hepatic atrophy and lowered levels of cytochrome P450, the key component of liver enzymes involved with drug metabolism. These effects are largely attributable to diversion of portal blood away from the liver and not to decreased hepatic blood flow. The present study has established a simpler model of portal blood diversion in order to examine the role of portal blood constituents in the regulation of hepatic cytochrome P450. Portal vein ligation was performed on male Wistar rats in which portasystemic anastomoses had been produced by subcutaneous transposition of the spleen. Portal vein ligation resulted in portal hypertension, as evidenced by splenomegaly, and in hepatic atrophy. In liver of rats with portal vein ligation, microsomal cytochrome P450 levels were significantly less than in sham-operated control rats, but cytochrome b5, NADPH-cytochrome c reductase, and glucose-6-phosphatase were unaltered. The activities of four mixed function oxidases also were reduced significantly in the liver of rats with portal vein ligation, the changes being greatest for ethylmorphine N-demethylase, a prototype substrate for the phenobarbital-inducible isoenzyme of cytochrome P450. In contrast, the activity of microsomal heme oxygenase, the rate-limiting step in catabolism of heme to bilirubin, was enhanced after portal vein ligation. Experiments in pair-fed rats showed that the changes observed in liver from rats with portal vein ligation could not be attributed to caloric deprivation. Administration of phenobarbital increased liver mass, cytochrome P450 levels, and mixed function oxidase activities both in rats with portal vein ligation and in controls, indicating that the liver of the ligated rats retained considerable protein synthetic capacity. It appears that hepatic atrophy and lowering of cytochrome P450 levels that follow portal vein ligation are consequences of altered exposure of the liver to factors normally present in portal blood, and that the same alterations may also enhance heme oxygenase activity.     

Bromodeoxyuridine uptake by early liver metastases in rats: a comparison of the hepatic artery and portal vein infusion routes

Liver metastases generated by the intraportal inoculation of ascites hepatoma cells in Donryu rats were labeled with bromodeoxyuridine (BrdU) through the hepatic artery, or through the portal vein with or without ligation of the hepatic artery, 3, 6, or 9 days after tumor inoculation. The distribution of BrdU-labeled cells was evaluated in 174 metastases, 110-1640 microm in diameter, by immunohistochemical methods. When a dual blood supply from the portal vein and hepatic artery existed, the BrdU-labeled cells were diffusely found in the metastases regardless of their size and the route of BrdU infusion. When blood supply to metastases larger than 610 microm in diameter was from a single source, namely the portal vein, the BrdU-labeled cells were located within 90-290 microm from the margin of the metastases. These results indicate first, that drug uptake by the inner part of the early metastatic liver tumors is achieved through the hepatic artery, and second, that drug uptake by early liver metastases through the portal vein is limited to within the extent of portal diffusion regardless of the size of the metastases. Thus, we conclude that prophylactic treatment against liver metastases would be more effective when given via the hepatic artery route rather than via the portal vein route.     

Immediate increase in arterial blood flow in embolized hepatic segments after portal vein embolization: CT demonstration

OBJECTIVE: This study was conducted to determine whether an immediate change occurs in the blood flow distribution in hepatic segments after segmental portal vein embolization. CONCLUSION: We found an immediate change in the distribution of blood flow in the liver after embolization; with portal vein embolization, we found an immediate increase in the hepatic artery blood flow in the affected segments.     

The transjugular stent implantation for the treatment of malignant portal and hepatic vein obstruction in cancer patients

BACKGROUND: The increase in portal vascular resistance is a significant complication of metastatic disease to the liver or locally advanced cancer, e.g., biliary cancer. PATIENTS AND METHODS: This paper describes the successful palliative treatment of two cancer patients with portal hypertension presenting with the symptoms of tense ascites, mesenteric congestion, and severe variceal bleeding. By creating a stenttract between a hepatic vein and a main branch of the portal vein and/or by placing an extendable stent into the portal vein, the transjugular intrahepatic portosystemic stent-shunt (TIPS) technique was used to decompress the portovascular system. RESULTS: The TIPS-technique offers a new, safe and effective palliation for malignant portal hypertension. In both patients, the symptoms of the portal hypertension disappeared after the procedure. This was accompanied by a significant improvement of the patients performance status allowing an early ambulation. CONCLUSION: Our findings demonstrate the feasibility and effectiveness of the TIPS procedure as a minimal invasive treatment for portal vein decompression in selected tumor patients.     

Doppler sonographic diagnosis of severe portal vein pulsatility in constrictive pericarditis: flow normalization after pericardiectomy

This case report describes the noninvasive assessment of hepatic and portal vein hemodynamics in a patient with constrictive pericarditis before and after pericardiectomy. Doppler sonography of the hepatic veins demonstrated a typical W-shaped pattern with pronounced late diastolic flow reversal that disappeared after surgery. Preoperatively, we observed severe pulsatility of the portal vein with flow reversal in systole; after pericardiectomy, portal venous flow was normal. We concluded that the high right atrial pressure in this patient might have led to increased hepatic venous outflow resistance, with subsequent trans-sinusoidal shunting between the hepatic artery and portal vein causing severe portal vein pulsatility. After pericardiectomy and a decrease in right atrial pressure, portal vein flow normalized.     

The middle colic vein: an alternative source of portal inflow in orthotopic liver transplantation complicated by portal vein thrombosis

BACKGROUND: Portal vein thrombosis (PVT) was previously considered a contraindication to orthotopic liver transplantation (OLT) since adequate portal blood supply is mandatory for graft function and patient survival. Improvements in surgical technique, however, have meant that this problem now can be circumvented in most instances. Nevertheless portal vein thrombosis remains an obstacle in OLT and is associated with increased incidence of primary non-function and long-term liver failure. METHODS: A 55-yr-old patient underwent OLT for secondary biliary cirrhosis associated with hepatitis C infection and complicated by long standing PVT. Involvement of the portal, mesenteric, and splenic veins prevented standard portal venous reconstruction. Portal inflow was accomplished by a side-to-end anastomosis between the middle colic vein and the donor portal vein. RESULTS: Hepatic reperfusion and subsequent liver function were excellent. Portal blood flow, as measured by color-enhanced Doppler ultrasound, was normal following surgery until discharge. The post-operative course was complicated by abdominal wound dehiscence and recurrent cytomegalovirus (CMV) infection. The patient was discharged in good clinical condition, with excellent liver function and patent portal vein 89 d after OLT. CONCLUSIONS: The middle colic vein is a novel, not previously described, source of portal venous inflow for OLT complicated by extensive splanchnic venous inflow thrombosis.     

Hepatic and portal surface veins: a new anatomic variant revealed during abdominal CT

OBJECTIVE: The objective of this paper is to describe a new finding on CT of hepatic and portal vein segments located in a subcapsular location on the surface of the liver. SUBJECTS AND METHODS: From a series of more than 11,000 contrast-enhanced abdominal CT scans performed from 1993 to 1997, 14 patients were identified as having hepatic or portal vein segments or both in a subcapsular location on the surface of the liver. RESULTS: We found seven portal vein surface segments in seven patients and 14 hepatic vein surface segments in 12 patients. Of the 14 patients, five had both portal and hepatic vein surface segments. Therefore, in a cohort that exceeded 11,000 patients, the incidence of this finding was 0.1%. Four patients had cirrhosis, two had small hypervascular liver lesions, and eight had healthy livers. The surface veins were not associated with any other recognized vascular anomalies or with anastomoses to extrahepatic systemic veins. CONCLUSION: Hepatic and portal veins can course to a subcapsular location on the surface of the liver. This anatomy is believed to be a normal variant and can be found in patients with healthy livers and normal hepatic vein hemodynamics and in patients with portal hypertension.     

Doppler US of helical flow in the portal vein

In helical portal venous blood flow, the usual laminar flow in the portal vein is replaced by a spiral. This changes the color Doppler ultrasound (US) appearance to one of alternating or parallel red and blue bands. Duplex US may appear to show hepatopetal, hepatofugal, or simultaneous bidirectional flow depending on placement of the cursor within the helix. Helical portal venous flow is unusual in normal individuals (2.2% of 135 patients). Its presence should prompt further scrutiny for signs of liver disease, particularly portosystemic shunts, as in 20% of 41 patients who subsequently underwent liver transplantation. It is a normal finding immediately after liver transplantation (43% of 35 patients) and transjugular intrahepatic portosystemic shunt (TIPS) creation (28% of 36 patients). In both liver transplant and TIPS recipients, helical flow is usually transient. Its persistence long after transplantation in association with a prolonged increase in portal venous velocity is a useful sign of portal vein stenosis. Helical flow may also occur in cases of neoplastic invasion or displacement of the portal vein.     

Recanalized umbilical vein in portal hypertension

Experience with splenoportography suggests that patency of the umbilical vein occurs in about 9% of the patients with portal hypertension. A widely patent umbilical vein might serve as a decompressive portosystemic shunt. Percutaneous transhepatic portography was performed in 107 patients with cirrhosis of the liver and portal hypertension. A patent umbilical vein was found in 28 patients (26%). This finding significantly paralleled the number and size of other collateral veins, apart from gastroesophageal varices. No significant relation was found between umbilical vein patency and portal pressure, extrahepatic shunting, variceal bleeding, or ascites. It is concluded that a large patent umbilical vein does not effectively relieve portal hypertension, prevent gastroesophageal varices, or protect against variceal bleeding or ascites.     

Doppler-duplex ultrasonography in the diagnosis of cavernous portal vein

Prehepatic portal hypertension caused by cavernous transformation of the portal vein has been more and more considered as a multiorgan disease with circulatory changes in numerous organs related to systemic and splanchnic vascular network [1]. Honeycomb-like, spongy, cavernous portal vein is a rare clinical and pathoanatomical entity which usually results from portal vein thrombosis. Recanalization and neovascularization processes lead to cavernomatous transformation of the portal vein lumen and formation of periportal collateral hepatopetal venous varices (Petren's veins) [5, 6]. Recently, with Doppler ultrasonography and angiography cavernous portal vein has been identified as the cause of prehepatic portal hypertension. Usage of color Doppler and duplex Doppler ultrasonography has greatly contributed to diagnostic efficiency, while therapeutically, the disease remains a serious and controversial problem. METHODS: At the Institute of Digestive Diseases, Clinical Centre of Serbia, 8 patients with cavernous portal vein were studied in the period 1995-1997. Real-time duplex and color Doppler ultrasonography (Toshiba-SSA 100A with sector duplex probe 3.75 MHz, and 9 ATL with color Doppler convex duplex probe 3.5 MHz) were used. Indirect (arterial) portography was used for imaging of lienoportal system in the venous phase of angiography as follows: catheterization (Seldinger's technique) of the coeliac trunk or lienal artery, and catheterization of the superior mesenteric artery. Indirect portography was performed by injection of 60-80 ml of the contrast medium by an automatic pump, at 10-14 ml/sec, i.e. 8-10 ml/sec by the digital technique [7]. Peroral fiberendoscopy was performed in all patients by Olympus GIF-XQ 10 endoscope. RESULTS: In our study the conventional ultrasonographic examination failed to provide an appropriate image of the normal portal vein. In hepatoduodenal ligament multiple tubular and round structures were seen, revealing an atypical honeycomb or spongycavernous shape of the venous lumen (Figs. 1 and 2). Doppler ultrasonography of the lumen of these venous collateral structures revealed a continuous, hypokinetic flow, mid-rate 7.4 cm/sec, which was always hepatopetally directed (to the liver). Color Doppler ultrasonography detected extensive portosystemic collateralls in all patients, and varices in the gallbladder wall in 1 patient. The results of indirect portography correlated well with Doppler ultrasonographic findings. In all patients hepatopetal flow was found (Figs. 3 and 4). The aetiology was diverse: idiopathic, liver cirrhosis, haematological diseases, Crohn's disease and Marfan's syndrome. Two patients had IV degree varices in the distal third of the oesophagus, and 4 patients had II/III degree varices. Patients with posthepatic liver cirrhosis and Crohn's disease had no varices in the distal third of the oesophagus and gastric fornix. DISCUSSION: Since Pick (1909) described this malformation as the hepatopetal collateral, the haemodynamic concept of this entity has not been changed. Doppler ultrasonography and angiography confirm that the blood flow in cavernomas is hepatopetal, i.e. compensated and functional. Cavernous transformation of the portal vein is clinically manifested by bleeding from oesophagogastric varices. Haemathemesis is the most alarming complication and may be the first clinical sign. The haemorrhage is usually recurrent and profuse, but in most cases it is tolerated well owing to preserved hepatic function in patients without liver cirrhosis [19]. Portosystemic collateral circulation may take place via retroperitoneal and other spontaneous venous shunts, not involving the left gastric vein or vv. gastricae breves, when oesophagogastric varices are absent (our patient with Crohn's disease and posthepatitic B cirrhosis). Splenomegaly with hypersplenism is always present with cavernous transformation of the portal vein, and usually precedes the occurrence of gastrointestinal hae     

Anti-GBM glomerulonephritis in mice lacking nitric oxide synthase type 2

To determine the relationship between quantitative Doppler parameters of portal, hepatic, and splanchnic circulation and hepatic venous pressure gradient (HVPG), variceal size, and Child-Pugh class in patients with alcoholic cirrhosis, we studied forty patients with proved alcoholic cirrhosis who underwent Doppler ultrasonography, hepatic vein catheterization, and esophagoscopy. The following Doppler parameters were recorded: time-averaged mean blood velocity, volume flow of the main portal vein flow, and resistance index (RI) of the hepatic and of the superior mesenteric artery. Doppler findings were compared with HVPG, presence and size of esophageal varices, and Child-Pugh class. There was a significant inverse correlation between portal velocity and HVPG (r = -.69), as well as between portal vein flow and HVPG (r = -.58). No correlation was found between RI in the hepatic artery or superior mesenteric artery and HVPG. No correlation was found between portal vein measurements and presence and size of varices. Severe liver failure was associated with lower portal velocity and flow. In patients with alcoholic cirrhosis, only portal vein blood velocity and flow, but neither hepatic nor mesenteric artery RI, are correlated to the severity of portal hypertension and to the severity of liver failure.     

The relationship between hepatoma and portal vein

Hepatic cells receive dual blood supply of the arterial and portal systems, but hepatoma has been thought to be supplied completely by hepatic artery. However, transcatheter hepatic artery embolization (TAE) has not been able to damage hepatoma entirely. For the study on the relationship between hepatoma and portal vein, I gave intraportal infusions of bromodeoxyuridine. (BrdU), an analogue of thymidine, to 10 patients with hepatoma at the time of surgery, and counted BrdU-positive nuclei immunohistochemically using the anti-BrdU monoclonal antibody. The labeling index, or percentage of BrdU-labelled cells, was 2.0 +/- 1.1% (mean +/- SD) in the cases without TAE and 11.9 +/- 4.2% in the other after TAE. On the other hand, examining the specimens with Microfil injected into the portal branch showed the distribution of portal branches in the hepatoma by radiograph or microscope. It is concluded that hepatoma does not receive arterial blood supply alone but the portal vein participates in vascular distribution of hepatoma, and the participation gets larger after TAE. It is suggested that this finding could be of great importance in planing treatment of patients with hepatoma.     

Effect of respiration, exercise, and food intake on hepatic vein circulation

Since the effects of respiration, nutrition, and exercise on blood flow in the hepatic vein are not well understood, the objective of this study was to determine the hemodynamic influence of these factors on hepatic venous circulation using Doppler ultrasonographic tracings. The venous blood flow of the middle hepatic vein was determined during arrested full inspiration, midinspiration, and expiration in 25 healthy subjects. The maximum velocity and the systolic-to-diastolic ratio of the blood flow were measured. The portal vein blood flow velocity was measured in 20 volunteers before and after food intake. The portal vein blood flow and the hepatic vein flow velocity were examined in eight volunteers after exercise. During inspiration, maximum blood flow velocity of the hepatic veins decreased compared to midinspiration (P < 0.001). With expiration the maximum velocity increased (P < 0.001). After food consumption, there was no change in the velocity of the hepatic veins, but the portal vein blood flow increased (P = 0.041). After physical exercises, the maximum velocity of the hepatic venous flow increased, on average, about 148% (P = 0.01), and the portal vein blood flow decreased about 44% (P = 0.027). To achieve standard measurements of hepatic venous blood flow, the state of respiration and physical exertion should be established. The nutritional status had only a minor influence on hepatic vein measurements.     

Left trisegmentectomy for hepatocellular carcinoma with tumor thrombi extending to the third branches in the remnant liver

BACKGROUND/AIMS: A successful left trisegmentectomy and tumor thrombectomy for hepatocellular carcinoma (HCC) of the right third branches in the remnant liver is described. METHODOLOGY: In this case, the hepatocellular carcinoma originated in the internal segment of an HBs- and HBe-Ag positive cirrhotic liver, involved the portal vein of the umbilical portion across the portal trunk to the contralateral third branches, and had many metastatic nodules in the anterior segment of the liver. To remove tumor thrombi by directly visualizing the lumen of the third branches of the portal vein, the portal trunk to the anterior branch of the portal vein was completely isolated by transecting the hepatic parenchyma along the intersegmental plane through a hilar approach. Tumor thrombi were removed by incising the portal trunk toward the anterior branch of the portal vein. Portal flow to the remnant liver was restored preceding liver resection to preserve hepatic function. After division of the anterior Glissonean code, the left trisegment and caudate lobes of the liver were resected. RESULTS: The patient was discharged 10 weeks after surgery and remained well for the first six months thereafter. Recurrent tumors, however, appeared in the remnant liver 7 months after surgery. CONCLUSIONS: This procedure may be applicable in cases of HCC with portal tumor thrombi extending into the third branches in the remnant liver.     

Hepatic vein reconstruction at inferior vena cava confluence using left renal vein graft

BACKGROUND/AIMS: To preserve remnant liver function, extended left hepatectomy combined with middle hepatic vein reconstruction using a left renal vein graft was performed in resection of liver metastasis from sigmoid colon cancer, involving the confluence of the middle and left hepatic veins. METHODOLOGY: The tumor, 5 cm in size, occupied the superior part of segment 4, and involved the confluence of the middle and the left hepatic veins. An extended left hepatectomy, including the left lobe, left caudate lobe and part of segment 8, together with the middle hepatic vein trunk, was performed. The left renal vein was resected as a graft from the confluence of the inferior vena cava just distal to the branches of the gonadal vein, renal-azygos, splenorenal communications and vertebral veins. The middle hepatic vein was reconstructed using the left renal vein 3 cm in length. RESULTS: Impaired values of liver function tests were normalized by the third postoperative day. Renal function was good throughout the postoperative period. The patient was discharged two weeks after the surgery. The reconstructed middle hepatic vein was patent, which was evaluated by a color Doppler ultrasonography, computed tomography and magnetic resonance imaging 60 days after the surgery. The patient remained well in the eight months thereafter. CONCLUSIONS: Hepatic vein reconstruction using a left renal vein graft is a new and preferable addition for the selection of an optimal graft.     

Serial changes in titers of antibody to hepatitis B surface antigen after immunization of infants born to mothers with hepatitis B e antigen

The purpose of this study is to identify the existence of hepatovenous intrahepatic anastomosis in normal men. A total of thirteen livers were investigated during the early autopsies of normal men who died in accidents. Perfusion venography of branches of hepatic veins using meglucamine diatrizoate was done in six cases; this method we used had not been reported in the literature. In one case, portal venography was performed. And in the other six cases, liver substance staining was done by injecting the ink through the middle hepatic vein, and such staining of the liver was observed by light microscope. The results show, (1) there are intrahepatic anastomoses between the hepatic veins within the liver; (2) there are anastomoses between the middle hepatic vein and the accessory hepatic veins; and (3) shunts exist between portal veins and hepatic veins. The above findings provide an anatomical basis for the performance of irregular hepatectomy and the rationale for one or two hepatic veins ligation should such veins were traumatized or invaded by liver cancer.     

Sinusitis in the immunocompromised host

Using retrospective studies, we have investigated the possibility of obtaining characteristic findings of inflammatory pseudotumor of the liver by magnetic resonance (MR) imaging. We examined 8 patients (involving 8 masses) who had been histologically diagnosed as having an inflammatory pseudotumor in the liver. The histological studies were performed on an excised specimen of 1 mass, and on aspiration needle biopsy specimens and the clinical courses of the other 7 masses. T1 weighted images (T1WI) and T2 weighted images (T2WI) were obtained on MR imaging. MR imagings were analyzed for visualized patterns, patterns of internal structure and patterns of contrast enhancement of dynamic MR imaging. The 8 masses were visualized as hypointense on T1WI and hyperintense on T2WI by MR imaging. Dynamic MR imaging revealed that 1 mass was markedly enhanced peripherally while another mass was homogeneously enhanced, and that enhancement was most marked immediately after injection of contrast medium and then gradually disappeared. Vessels were observed in 4 masses (the portal vein in 2 masses, the hepatic vein in 1 mass, and portal and hepatic veins in 1 mass), and these vessels were clearly visualized on T1WI. The MR imaging findings from the early stage of an inflammatory pseudotumor showed a pattern similar to that of hepatic tumors with rich blood flow. The portal vein or hepatic vein was found in the tumor in half the patients, suggesting that this characteristic was useful for diagnosis of an inflammatory pseudotumor in the liver.     

An improved technique for bloodless hepatic resection on in situ cold perfused liver

An improved technique for bloodless hepatic resection using in situ isolation and asanguinous hypothermic perfusion was described to deal with huge liver tumors involved in the liver hilum, the main hepatic veins and retrohepatic inferior vena cava. The original Fortner's technique was modified, including the choice of incision; semi-isolated perfusion of the liver portion preserved through the single portal vein; suprahepatic outlet of the perfusate and the shortening of the period of hepatic ischemia by reperfusion of hepatic artery prior to the repair or reconstruction of the portal vein. The initial successful experience of the technique applied to 2 pediatric cases with giant liver tumors was reported, and the indications, intraoperative and early postoperative courses were discussed.     

Optic neuritis with residual tunnel vision in perchloroethylene toxicity

BACKGROUND: To identify and differentiate agenesis and severe atrophy of the right hepatic lobe on computed tomography (CT). METHODS: The CT examinations of three cases of agenesis and 11 cases of severe atrophy of the right hepatic lobe were reviewed. We evaluated visibility of the three hepatic veins, the two main portal veins (including their branches if necessary), the dilated intrahepatic ducts, enlargement of the medial and lateral segments of the left lobe and caudate lobe of the liver, presence of a retrohepatic gallbladder, hyperattenuation of the atrophic liver parenchyma, posterolateral interposition of the hepatic flexure of the colon, and upward migration of the right kidney. RESULTS: In the three cases of agenesis, no structure can be recognized as the right hepatic vein, right portal vein, or dilated right intrahepatic ducts. In the 11 cases of severe lobar atrophy, the right portal vein (or its branches) was recognized in eight cases, the right hepatic vein in four cases, and the dilated right intrahepatic ducts in 11 cases. The degree of enlargement of the lateral segment does not necessarily change inversely with the size of the medial segment and the caudate lobe. The retrohepatic gallbladder is present in eight cases (two in agenesis and six in atrophy). The phenomenon of hyperattenuation of the atrophic liver parenchyma was noted in six cases. CONCLUSION: Even though a retrohepatic gallbladder and a severely distorted hepatic morphology due to compensatory hypertrophy of the left and caudate lobes may raise a suspicion of agenesis of the right lobe of the liver, absence of visualization of all of the right hepatic vein, right portal vein and its branches, and dilated right intrahepatic ducts is a prerequisite of the diagnosis of agenesis of the right hepatic lobe on CT. In severe lobar atrophy, at least one of these structures is recognizable.     

Changes in neurotransmitters in multiple sclerosis

PURPOSE: The aim of this study was to identify the route of administration of 5-FU with the greatest pharmacological advantage in a rat model using non-invasive in vivo 19F nuclear magnetic resonance (NMR) spectroscopy. METHODS: 5-FU (50 mg/kg) was administered to anesthetized Wistar rats cannulated into the hepatic artery, portal vein or tail vein and 11 NMR spectra were acquired from the liver region to 60.5 min every 5.5 min. RESULTS: With systemic i.v. (tail vein) infusion, the 19F-NMR signal for 5-FU from the liver region peaked in the first spectrum (0-5.5 min), and then gradually decreased. The signal for the 5-FU catabolite alpha-fluoro-beta-alanine (FBAL) gradually increased to the sixth spectrum (0-33.0 min) and then plateaued. Following portal vein infusion the intensity of the first 5-FU spectrum was twice as high as that following i.v. infusion, but the intensity decreased and the FBAL signal increased gradually in the sixth spectrum as systemic i.v. infusion. In contrast, the intensity of the 5-FU signal following hepatic artery infusion was the same as that following portal vein infusion in the first spectrum, and maintained a strong intensity to the final spectrum (60.5 min). The FBAL signal was detected from the second spectrum following hepatic artery infusion, but its intensity was significantly weaker than that following i.v. or portal vein infusion. CONCLUSIONS: Hepatic arterial infusion resulted in the active form of 5-FU being present for a longer time and its degradation in the liver being suppressed compared with the results following portal vein infusion. This catabolic advantage of hepatic arterial infusion could lead to a more potent anti-tumor activity against liver metastases, but could also lead to significant host toxicity including biliary toxicity. We recommend that the dose/schedule of 5-FU administered via the hepatic artery should be adjusted carefully.