Management of variceal bleeding in patients with noncirrhotic portal vein thrombosis

Revisiting the place of surgical portal decompression for adults with noncirrhotic portal hypertension due to chronic extrahepatic portal vein obstruction: a scoping review

BACKGROUND

Liver cirrhosis accounts for more than 90 % of portal hypertension cases, and the other cases are due to noncirrhotic portal hypertension (NCPH). Variceal bleeding is the most life-threatening complication of portal hypertension and its primary treatment is medical according to the Baveno VII guidelines. This review discusses the evidence on surgical portal decompression for adult patients with NCPH secondary to chronic extrahepatic portal vein obstruction (EHPVO).

METHODS

This is a scoping review of the evidence for the feasibility and effectiveness of surgical portal decompression in adults with NCPH secondary to EHPVO.

RESULTS

This scoping review yielded 17 studies, including a total of 110 patients. Patient age(s) ranged from 19 to 68 years, with the majority undergoing nonphysiological (i.e., portosystemic shunts) shunts (N = 84, 76.4 %), mostly for variceal bleeding refractory to medical and endoscopic treatments. Physiological shunts (i.e., Rex shunts) had a potential advantage over nonphysiological shunts in postoperative rebleeding (5 % vs. 10 %) and hepatic encephalopathy rates (0 % vs. 13 %). Conversely, nonphysiological shunts had a potential advantage over physiological shunts in postoperative shunt thrombosis (8 % vs. 22 %).

DISCUSSION

This scoping review reported that surgical portal decompression is feasible in adults with NCPH due to EHPVO with favorable outcomes and long-term patency.

Unconventional shunt surgery for non-cirrhotic portal hypertension in patients not suitable for proximal splenorenal shunt

Backgrounds/Aims

Proximal splenorenal shunt (PSRS) is a commonly performed procedure to decompress portal hypertension, in patients with refractory variceal bleed, especially in non-cirrhotic portal hypertension (NCPH). If conventional methods are hindered by any technical or pathological factors, alternative surgical techniques may be required. This study analyzes the effectiveness of various unconventional shunt surgeries performed for NCPH.

Methods

A retrospective analysis of NCPH patients who underwent unconventional shunt surgeries during the period July 2011 to June 2022 was conducted. All patients were followed up for a minimum of 12 months with doppler study of the shunt to assess shunt patency, and upper gastrointestinal endoscopy to evaluate the regression of varices.

Results

During the study period, 130 patients underwent shunt surgery; among these, 31 underwent unconventional shunts (splenoadrenal shunt [SAS], 12; interposition mesocaval shunt [iMCS], 8; interposition PSRS [iPSRS], 6; jejunal vein-cava shunt [JCS], 3; left gastroepiploic-renal shunt [LGERS], 2). The main indications for unconventional shunts were left renal vein aberration (SAS, 8/12), splenic vein narrowing (iMCS, 5/8), portalhypertensive vascular changes (iPSRS, 6/6), and portomesenteric thrombosis (JCS, 3/3). The median fall in portal pressure was more in SAS (12.1 mm Hg), and operative time more in JCS, 8.4 hours (range, 5-9 hours). During a median follow-up of 36 months (6-54 months), shunt thrombosis had been reported in all cases of LGERS, and less in SAS (3/12). Variceal regression rate was high in SAS, and least in LGERS. Hypersplenism had reversed in all patients, and 6/31 patients had a recurrent bleed.

Conclusions

Unconventional shunt surgery is effective in patients unsuited for other shunts, especially PSRS, and it achieves the desired effects in a significant proportion of patients.

Increased Morbidity and Mortality of Patients with Non-cirrhotic Portal Vein Thrombosis After Abdominal and Pelvic Surgeries: a Study of the National Inpatient Sample 2002 to 2015

BACKGROUND

A higher rate of postoperative morbidity and mortality in patients with portal hypertension from cirrhosis is well recognized; however, the rate of postoperative morbidity and mortality among patients with portal hypertension from non-cirrhotic portal vein thrombosis (NCPVT) is largely unknown.

METHOD

All adults undergoing abdominal and pelvic surgery were identified from the National Inpatient Sample database from 2002 to 2015. Patients were then categorized into three groups: non-cirrhotic non-portal vein thrombosis (NCNPVT), NCPVT, and cirrhotic portal vein thrombosis (CPVT). Inpatient mortality, type of disposition, transfusions, length of stay, postoperative complications, and total charges were compared. Logistic regression and ordinary least squares regression analyses were performed for factors associated with inpatient mortality, transfusions, surgery-related complications, and log length of stay.

RESULTS

Patients with NCPVT had significantly higher inpatient mortality rates, surgery-related complications, and longer length of stays compared with patients with NCNPVT (2.64% vs. 0.34%, 10.26% vs. 3.26%, 8 vs. 2 days) but less than patients with CPVT (2.64% vs. 6.31%, 10.26% vs. 17.48%, 8 vs. 11 days). In multiple logistic regression analyses, NCPVT groups remained associated with increased inpatient mortality rate, transfusions, and postoperative complications with odds ratios of 3.71 (1.88, 7.32), 3.43 (2.54, 4.62), and 3.08 (2.16, 4.39), respectively. NCPVT was also associated with 2.4 times increased length of stay.

DISCUSSION

Patients with NCPVT had significantly higher risks of postoperative morbidity and mortality than patients with NCNPVT but less than patients with CPVT. Future studies with detail regarding the characteristics of PVTs are needed to confirm the findings in this study.

Current Indications and Long-Term Outcomes of Surgical Portosystemic Shunts in Adults

BACKGROUND

Surgical portosystemic shunts are rare. We reviewed indications, operative details, and outcomes of patients undergoing surgical portosystemic shunt procedures.

METHODS

We retrospectively reviewed clinical data of consecutive patients between 1997 and 2018 from a single institution. Clinical characteristics and outcomes were compared between two groups: patients with portomesenteric venous thrombosis (PMVT) vs those with cirrhosis. Endpoints included 30-day mortality, shunt-related complications, patency, and survival.

RESULTS

There were 99 patients, 45 male and 54 female, with a mean age of 46 ± 18 years, enrolled in the study. There were 63 patients (63%) with PMVT and 36 patients (36%) with cirrhosis. Both groups had similar demographics, cardiovascular risk factors, and aneurysm extent, except for more diabetes among those with cirrhosis (p < 0.05). There were no significant differences in procedural metrics and intra-procedure complications between groups, except that patients with PMVT underwent more non-selective shunts than those with cirrhosis (63% vs. 30%, p < 0.001). There were two 30-day deaths (2%), with no difference in mortality and MAEs between groups. On univariate analysis, cholangiopathy and PMVT were associated with graft thrombosis (HR = 9.22, 95% CI 1.22-70.27) while race, smoking, cardiac comorbidity, type of operative shunt, configuration of the shunt, and use of conduit were not (p > 0.05). Patients with PMVT had significantly lower 1-, 5-, and 10-year primary (77%, 71%, and 71% vs. 97%, p = 0.009) and secondary patency (88%, 76%, and 72% vs. 96%, p = 0.027) compared with those with cirrhosis. The 1-, 5-, and 10-year survival rates were 94%, 84%, and 61% for patients with PMVT compared with 88%, 58%, and 26% for those with cirrhosis (non-adjusted HR 0.40, 95% CI 0.19-0.84, p = 0.01, age-adjusted HR 0.51, 95% CI 0.24-1.09, p = 0.08). The survival of patients with PMVT without liver disease trended higher than those with liver disease; however, when adjusted for age, the survival gap narrowed, and the difference was not statistically significant (p = 0.19), survival being lowest for those with PMVT and liver disease.

CONCLUSIONS

Surgical portosystemic shunts are safe and effective for symptom relief in selected patients with portal hypertension. The odds of graft thrombosis is 9 times higher in patients with PMVT. Overall survival is similar in patients with PMVT or cirrhosis.

A Single-Center Experience in Portal Flow Augmentation in Liver Transplantation With Prior Large Spontaneous Splenorenal Shunt

Large portosystemic shunts may cause portal steal syndrome in liver transplantation (LT). Because of the possible devastating consequences of the syndrome, the authors recommend perioperative management of these large shunts. Fourteen adult recipients who underwent portal flow augmentation, including left renal vein ligation (LRVL), renoportal anastomosis (RPA), shunt ligation (SL), and splenic vein ligation (SVL) for large spontaneous splenorenal shunt (SSRS), are included in this study, and the results were analyzed. A total of 13 patients had a large SSRS, and in 1 patient, the large shunt was placed between the superior mesenteric vein and the right renal vein. LDLT was performed in 13 patients. LRVL (n = 5), SVL (n = 6), RPA (n = 2), SL (n = 1) were performed to the patients as graft inflow augmentation. The graft-recipient weight ratios (GRWR) were less than 0.8% in 5 patients (35.7%): 2 had LRVL, and 3 had SVL. Small-for-size syndrome (SFSS) occurred only in these 2 patients with LRVL (GRWR ≤0.8%) and, splenic artery ligation was performed for graft inflow modulation. No mortality or serious complications were reported during follow-up. We consider that in patients with large SSRS and small-for-size grafts, SVL can be performed safely and with satisfactory outcomes.

Reappraisal of shunt surgery for extrahepatic portal vein obstruction in adults: Report of a single-center case series

AIM

Extrahepatic portal venous obstruction (EHPVO) is a relatively rare disease in adults. The clinical significance of shunt surgery for EHPVO in adult cases remains unclear.

METHODS

We retrospectively analyzed the patient characteristics and the results of shunt surgery in 13 adult cases of EHPVO treated between March 1995 and March 2013 at a tertiary care hospital in Switzerland. The indication for shunt surgery was recurrent bleeding after endoscopic treatment. To update the outcomes of shunt surgery in adult cases of EHPVO, we performed a systematic review of published work to examine this issue.

RESULTS

The mean age of the 13 patients in the present case series was 41.8 years (range, 20-68), and the mean follow-up duration after surgery was 4.4 years (range, 1-16). The types of shunt surgery performed were mesocaval shunt (n = 8), portacaval shunt (n = 2), splenorenal shunt (n = 1) and mesorenal shunt (n = 2). Two patients (15%) developed postoperative rebleeding, which was successfully treated by endoscopic treatment. None of the patients developed postoperative hepatic encephalopathy. No operative-related deaths occurred in this series. Six studies, including our own, were identified in the published work. The overall mortality rate was 0-3.7%, and no cases of encephalopathy were observed. The rebleeding rate ranged 2.5-23%.

CONCLUSION

Shunt surgery for EHPVO in adults after the failure of endoscopic treatment is feasible, with acceptable outcomes at specialized centers. This surgical procedure should always be taken into consideration when managing adult cases of EHPVO.

[Treatment of non-cirrhotic, non-tumoural portal vein thrombosis]

Thrombosis of the splenoportal axis not associated with liver cirrhosis or neoplasms is a rare disease whose prevalence ranges from 0.7 to 3.7 per 100,000 inhabitants. However, this entity is the second most common cause of portal hypertension. Prothrombotic factors are present as an underlying cause in up to 70% of patients and local factors in 10-50%. The coexistence of several etiological factors is frequent. Clinical presentation may be acute or chronic (portal cavernomatosis). The acute phase can present as abdominal pain, nausea, vomiting, fever, rectorrhagia, intestinal congestion, and ischemia. In this phase, early initiation of anticoagulation is essential to achieve portal vein recanalization and thus improve patient prognosis. In the chronic phase, symptoms are due to portal hypertension syndrome. In this phase, the aim of treatment is to treat or prevent the complications of portal hypertension. Anticoagulation is reserved to patients with a proven underlying thrombophilic factor.

Non-cirrhotic portal hypertension

Non-cirrhotic portal hypertension (NCPH) encompasses a wide range of disorders, primarily vascular in origin, presenting with portal hypertension (PHT), but with preserved liver synthetic functions and near normal hepatic venous pressure gradient (HVPG). Non-cirrhotic portal fibrosis/Idiopathic PHT (NCPF/IPH) and extrahepatic portal venous obstruction (EHPVO) are two prototype disorders in the category. Etiopathogenesis in both of them centers on infections and prothrombotic states. Presentation and management strategies focus on repeated well tolerated episodes of variceal bleed and moderate to massive splenomegaly and other features of PHT. While the long-term prognosis is generally good in NCPF, portal biliopathy and parenchymal extinction after prolonged PHT makes outcome somewhat less favorable in EHPVO. While hepatic schistosomiasis, congenital hepatic fibrosis and nodular regenerative hyperplasia have their distinctive features, they often present with NCPH.

Non-cirrhotic portal hypertension - diagnosis and management

NCPH is a heterogeneous group of liver disorders of vascular origin, leading to PHT with near normal HVPG. NCPF/IPH is a disorder of young adults or middle aged women, whereas EHPVO is a disorder of childhood. Early age acute or recurrent infections in an individual with thrombotic predisposition constitute the likely pathogenesis. Both disorders present with clinically significant PHT with preserved liver functions. Diagnosis is easy and can often be made clinically with support from imaging modalities. Management centers on control and prophylaxis of variceal bleeding. In EHPVO, there are additional concerns of growth faltering, portal biliopathy, MHE and parenchymal dysfunction. Surgical shunts are indicated in patients with failure of endotherapy, bleeding from sites not amenable to endotherapy, symptomatic hypersplenism or symptomatic biliopathy. Persistent growth failure, symptomatic and recurrent hepatic encephalopathy, impaired quality of life or massive splenomegaly that interferes with daily activities are other surgical indications. Rex-shunt or MLPVB is the recommended shunt for EHPVO, but needs proper pre-operative radiological assessment and surgical expertise. Both disorders have otherwise a fairly good prognosis, but need regular and careful surveillance. Hepatic schistosomiasis, CHF and NRH have similar presentation and comparable prognosis.

Management of bleeding in extrahepatic portal venous obstruction

Extrahepatic portal venous obstruction, although rare in the western world, is a common cause of major and life threatening upper gastrointestinal bleeding among the poor in developing countries. Patients have large spleens and stunted growth. The diagnosis is easily confirmed by Doppler ultrasonography. Endoscopy sclerotherapy is the best option for the control of acute variceal bleeding. For secondary prophylaxis of bleeding, the choice lies between repeated sclerotherapy and a portosystemic shunt. We believe that due consideration should be given to performing a splenectomy and a lienorenal shunt. Performed by experienced surgeons, it carries a low operative mortality of 1%, a rebleeding rate of about 10%, removes the large spleen, reverses hypersplenism, and is not followed by portosystemic encephalopathy. Most importantly, it is a onetime procedure particularly suited to those who have little access to blood transfusion and sophisticated medical facilities.

Noncirrhotic portal hypertension

Portal hypertension is characterized by an increase in portal pressure (> 10 mmHg) and could be a result of cirrhosis of the liver or of noncirrhotic diseases. When portal hypertension occurs in the absence of liver cirrhosis, noncirrhotic portal hypertension (NCPH) must be considered. The prognosis of this disease is much better than that of cirrhosis. Noncirrhotic diseases are the common cause of portal hypertension in developing countries, especially in Asia. NCPH is a heterogeneous group of diseases that is due to intrahepatic or extrahepatic etiologies. In general, the lesions in NCPH are vascular in nature and can be classified based on the site of resistance to blood flow. In most cases, these disorders can be explained by endothelial cell lesions, intimal thickening, thrombotic obliterations, or scarring of the intrahepatic portal or hepatic venous circulation. Many different conditions can determine NCPH through the association of these various lesions in various degrees. Many clinical manifestations of NCPH result from the secondary effects of portal hypertension. Patients with NCPH present with upper gastrointestinal bleeding, splenomegaly, ascites after gastrointestinal bleeding, features of hypersplenism, growth retardation, and jaundice due to portal hypertensive biliopathy. Other sequelae include hyperdynamic circulation, pulmonary complications, and other effects of portosystemic collateral circulation like portosystemic encephalopathy. At present, pharmacologic and endoscopic treatments are the treatments of choice for portal hypertension. The therapy of all disorders causing NCPH involves the reduction of portal pressure by pharmacotherapy or portosystemic shunting, apart from prevention and treatment of complications of portal hypertension.

Portal vein thrombosis

Portal vein thrombosis (PVT) can be a difficult clinical problem to assess and manage. A high index of suspicion is needed for a PVT diagnosis given the subtle presentation and potentially serious long-term complications. It should be considered a clue to the presence of one or several underlying disorders, including prothrombotic disorders, whether or not a local precipitating factor is identified. The accruing evidence shows that acute PVT can and probably should be treated with anticoagulation or thrombolytic agents in an effort to prevent extension of thrombus, mesenteric vessel occlusion, and portal hypertension. However, chronic PVT should be treated conservatively with measures to control major consequences related to portal hypertension. Anticoagulation therapy duration should be tailored to the identified predisposing factors.

[19-year-old kick-boxer with hematemesis and splenomegaly]

Traumatic portal vein thrombosis is a rare cause of nonmalignant, noncirrhotic portal hypertension. We report a case of a 19-year old patient, who presented with variceal bleeding and splenomegaly. Diagnosis was based on the history of kickboxing and an otherwise negative etiological investigation. The patient underwent endoscopic therapy and portosystemic shunt operation (Warren-shunt) due to cavernous transformation and severe hypersplenism. Thereafter the patient remained asymptomatic.

Effects of modified splenocaval shunt plus devascularization on esophagogastric variceal bleeding: a comparative study of this treatment and devascularization only in cirrhotic portal hypertension

BACKGROUND

Pericardial devascularization (PCDV) and portosystemic shunt were reported to have favorable results for the management of portal hypertension in cirrhotic patients in China and the West, respectively. This study was undertaken to investigate the effects of a modified proximal splenocaval shunt plus PCDV on variceal bleeding in patients with portal hypertension.

METHODS

From January 1997 to December 2007, 168 patients with portal hypertension of cirrhotic origin received an operation for gastroesophageal variceal bleeding. Of these, 90 patients received a splenocaval shunt plus a PCDV procedure (Combined Group) and the other 78 patients received a PCDV procedure only (PCDV Group). The procedure-related morbidity and mortality, rebleeding, encephalopathy, and survival rates were analyzed.

RESULTS

Postoperative mortality was 3.3% in the combined group and 5.1% in the PCDV group (P > 0.05). Overall morbidity was 13.3% in the combined group and 15.4% in the PCDV group (P > 0.05). The rate for rebleeding, including variceal bleeding and gastropathy, was 5.1% in the combined group, which was significantly lower than that in the PCDV group, at 16.7% (P < 0.05). The incidence of encephalopathy was 6.63% in the combined group and 6.67% in the PCDV group (P > 0.05). The 1-, 3-, 5- and 10-year survival rates were 97.4, 91.7, 80.0, and 60.0% in the combined group and 96.7, 83.3, 73.3, and 53.3% in the PCDV group (P > 0.05).

CONCLUSIONS

The modified splenocaval shunt plus PCDV is a safe and effective procedure for the long-term control of variceal bleeding; the procedure may not only maintain the portal flow to the liver, but may also protect the liver function in cirrhotic patients. The better clinical outcome means that the procedure may be one of the best choices for treating portal hypertension of cirrhotic origin.

Portal vein thrombosis, revisited

This review article aims to discuss the aetiology, pathophysiology, clinical presentation, diagnostic workup and management of portal vein thrombosis, either as a primary vascular liver disease in adults and children, or as a complication of liver cirrhosis. In addition, indications and limits of anticoagulant therapy are discussed in detail.

Review article: the modern management of portal vein thrombosis

BACKGROUND

Portal vein thrombosis (PVT) is an important cause of portal hypertension. It may occur as such with or without associated cirrhosis and hepatocellular carcinoma. Information on its management is scanty.

AIM

To provide an update on the modern management of portal vein thrombosis. Information on portal vein thrombosis in patients with and without cirrhosis and hepatocellular carcinoma is also updated.

METHODS

A pubmed search was performed to identify the literature using search items portal vein thrombosis-aetiology and treatment and portal vein thrombosis in cirrhosis and hepatocellular carcinoma.

RESULTS

Portal vein thrombosis occurs because of local inflammatory conditions in the abdomen and prothrombotic factors. Acute portal vein thrombosis is usually symptomatic when associated with cirrhosis and/or superior mesenteric vein thrombosis. Anticoagulation should be given for 3-6 months if detected early. If prothrombotic factors are identified, anticoagulation should be given lifelong. Chronic portal vein thrombosis usually presents with well tolerated upper gastrointestinal bleed. It is diagnosed by imaging, which demonstrates a portal cavernoma in place of a portal vein. Anticoagulation does not have a definite role, but bleeds can be treated with endotherapy or shunt surgery. Rarely liver transplantation may be considered.

CONCLUSION

Role of anticoagulation in chronic portal vein thrombosis needs to be further studied.

Vascular disorders of the liver

This guideline has been approved by the American Association for the Study of Liver Diseases (AASLD) and represents the position of the association.

Review article: The management of non-cirrhotic non-malignant portal vein thrombosis and concurrent portal hypertension in adults

BACKGROUND

Extrahepatic portal vein thrombosis is an important cause of non-cirrhotic portal hypertension.

AIM

To provide an update on recent advances in the aetiology and management of acute and chronic non-cirrhotic non-malignant extrahepatic portal vein thrombosis.

METHOD

A PubMed search was performed to identify relevant literature using search terms including 'portal vein thrombosis', 'variceal bleeding' and 'portal biliopathy'.

RESULTS

Myeloproliferative disease is the most common risk factor in patients with non-cirrhotic non-malignant extrahepatic portal vein thrombosis. Anticoagulation therapy for at least 3 months is indicated in patients with acute extrahepatic portal vein thrombosis. However, in patients with extrahepatic portal vein thrombosis due to a prothrombotic disorder, permanent anticoagulation therapy can be considered. The most important complication of extrahepatic portal vein thrombosis is oesophagogastric variceal bleeding. Endoscopic treatment is the first-line treatment for variceal bleeding. In several of the patients with extrahepatic portal vein thrombosis biliopathy changes on endoscopic retrograde cholangiography (ERCP) have been reported. Dependent on the persistence of the biliary obstruction, treatment can vary from ERCP to hepaticojejunostomy.

CONCLUSION

Prothrombotic disorders are the major causes of non-cirrhotic, non-malignant extrahepatic portal vein thrombosis and anticoagulation therapy is warranted in these patients. The prognosis of patients with non-cirrhotic, non-malignant extrahepatic portal vein thrombosis is good, and is not determined by portal hypertension complications but mainly by the underlying cause of thrombosis.

Transabdominal modified devascularization procedure with or without esophageal stapler transection--an operation adequate for effective control of a variceal bleed. Is esophageal stapler transection necessary?

BACKGROUND

In Japan, the original Sugiura procedure reported favorable results in non-cirrhotic patients but in the West, the modified Sugiura procedure is not widely accepted because of high rebleeding, morbidity, and mortality in cirrhotics. We retrospectively analyzed the efficacy of our modified Sugiura procedure i.e., devascularization with/without esophageal transection combined with salvage endotherapy and pharmacotherapy for control of a variceal bleed.

MATERIALS AND METHODS

Between January 1999 and December 2004, 912 patients with variceal bleeding were treated. Of these, 66 (7.2%) patients were subjected to surgery after failed endotherapy/propranolol. Among these 66 patients, 52 had transabdominal devascularization (16 emergency, 36 elective); 14 patients underwent devascularization with esophageal stapler transection (group I), and 38 patients had devascularization without esophageal stapler transection (group II). Another 14 patients underwent elective end-to-side proximal splenorenal shunt surgery.

RESULTS

Postoperative mortality was 7.1% in group I, 10.5% in group II (P>0.05). Mortality for emergency surgery was 31.2% (5/16) but there were no deaths in the elective surgery group. Overall morbidity was 57.1% in group I and 21.0% in group II (P<0.05). The rates of variceal rebleeding were 7.1% and 7.8%; residual varices were 30.7% and 32.3%; recurrent varices were 7.6% and 5.8% following the group I and group II procedures, respectively, over a mean follow-up period of 39.9 (7-2) months. Esophageal transection-related morbidity (leak, stricture, and bleeding) was 21.4% (3/14) in group I.

CONCLUSIONS

Devascularization without esophageal stapler transection is a safe and effective procedure for adequate (urgent and long-term) control of variceal bleeding with similar results and less morbidity when compared to devascularization with esophageal transection in cirrhotic patients, as well as non-cirrhotic patients.

Noncirrhotic portal hypertension

Portal hypertension is characterized by an increase in portal pressure (>10 mm Hg) and could be a result of cirrhosis of the liver or noncirrhotic diseases. Noncirrhotic portal hypertension (NCPH), as it generally is termed, is a heterogeneous group of diseases that is due to intrahepatic or extrahepatic etiologies. In general, the lesions in NCPH are vascular in nature and can be classified based on the site of resistance to blood flow. Noncirrhotic portal fibrosis and extrahepatic portal vein obstruction are two diseases that are common in developing countries; they most often present only with features of portal hypertension and not of parenchymal dysfunction. These are described in detail.

Serious intestinal bleeding from vascular ectasia secondary to portal thrombosis after living-related liver transplantation in a child

Serious intestinal bleeding from vascular ectasia secondary to extrahepatic portal thrombosis is much less frequent than variceal bleeding, and its treatment is not clearly defined. We describe a 4-year-old girl with repeated intestinal bleeding from vascular ectasia, without any varix, with late extrahepatic portal vein thrombosis (PVT) and late hepatic artery thrombosis (HAT) after living-related liver transplantation. The bleeding stopped after simple splenectomy. She has presented neither bleeding nor any serious complications related to splenectomy for 1 year to date. We think uncontrollable hemorrhage from gastrointestinal vascular ectasia secondary to extrahepatic portal thrombosis in a pediatric patient can and should be treated by simple splenectomy, because patients with this complication usually have a normally functioning liver. However, it is not clear whether this procedure is effective for variceal bleeding.

Endoscopic ligation vs. sclerotherapy in adults with extrahepatic portal venous obstruction: a prospective randomized study

BACKGROUND

Endoscopic sclerotherapy is a well-established treatment for bleeding esophageal varices, although it has a substantial complication rate. A prospective randomized trial was conducted to determine whether endoscopic variceal ligation is safer and more effective than sclerotherapy in adults with bleeding esophageal varices because of extrahepatic portal venous obstruction.

METHODS

Thirty-six patients underwent sclerotherapy and 37 had band ligation.

RESULTS

Ligation and sclerotherapy were equally effective for achieving variceal eradication (94.6% vs. 91.7%, respectively; p=0.67). However, ligation achieved eradication with fewer endoscopic sessions (3.7 [1.2] vs. 7.7 [3.3]; p <0.0001) and within a shorter time interval (50.1 [17.7] days vs. 99 [54.8] days; p <0.0001). In the ligation group, recurrent bleeding was less frequent (2.7% vs. 19.4%; p=0.028; however, Bonferroni correction for multiple testing removes this significance) and the rate of major complications was lower (2.7% vs. 22.2%; p=0.014). Total cost per patient was significantly higher in the sclerotherapy vs. the ligation group ($216.6 [71.8] vs. $182.6 [63.4]; p=0.035). During the follow-up period after variceal eradication, no significant differences were found between the sclerotherapy and the ligation groups with respect to recurrent bleeding (3% vs. 2.9%; p=1.0), esophageal variceal recurrence (9.1% vs. 11.4%; p=1.0), and formation of new gastric varices (9.1% vs. 14.3%; p=0.51).

CONCLUSIONS

Variceal band ligation is superior to sclerotherapy, because it is less costly and achieves variceal eradication more quickly, with lower relative frequencies of recurrent variceal bleeding and complications.

N/A

N/A

Current state of portosystemic shunt surgery

BACKGROUND

A switch to decompressive shunt procedures is mandatory if endoscopic therapy fails to control recurrent variceal hemorrhage. Surgical shunt procedures continue to be safe, highly effective, and durable procedures to treat variceal bleeding in patients with low operative risk and good liver function.

DISCUSSION

In cirrhotics, elective operations using portal flow preserving techniques such as a selective distal splenorenal shunt (Warren) and a partial portocaval small diameter interposition shunt (Sarfeh) should be preferred. Rarely, end-to-side portocaval shunt may serve as a salvage procedure if emergency endoscopic treatment or transjugular intrahepatic portosystemic shunt insertion fails to stop bleeding. Until definitive results from randomized trials are available patients with good prognosis (Child-Pugh A and B) should be regarded as candidates for surgical shunts. For patients with noncirrhotic portal hypertension, in particular with extrahepatic portal vein thrombosis, portosystemic shunt surgery represents the only effective therapy which leads to freedom of recurrent bleeding and repeated endoscopies for many years, and improves hypersplenism without deteriorating liver function or encephalopathy. Gastroesophageal devascularization and other direct variceal ablative procedures should be restricted to treat endoscopic therapy failures without shuntable portal tributaries.

Non-cirrhotic portal fibrosis

Non-cirrhotic portal hypertension (NCPH) comprises of diseases having an increase in portal pressure (PP) due to intraheptic or prehepatic lesions, in the absence of cirrhosis. The lesions are generally vascular, either in the portal vein, its branches or in the perisinusoidal area. Because the wedged hepatic venous pressure (WHVP) is near normal, measurement of intravariceal or intrasplenic pressure is needed to assess portal pressure. The majority of the diseases included in the category of NCPH are well characterized disease entities where portal hypertension (PHT) is a late manifestation and hence, these are not discussed. Two diseases which present only with features of PHT and are common in developing countries are NCPF and extra-hepatic portal vein obstruction (EHPVO). Non-cirrhotic portal fibrosis is a syndrome of obscure etiology, characterized by 'Obliterative portovenopathy' leading to PHT, massive splenomegaly, repeated well tolerated episodes of variceal bleeding and anemia in young adults from low socio-economic strata of life. The hepatic parenchymal functions are nearly normal. Jaundice, ascites and hepatic encephalopathy are rare. Management of variceal bleeding remains the main concern as nearly 85% of patients with NCPF present with variceal bleeding. Endoscopic variceal ligation or sclerotherapy are equally effective in about 90-95% of the patients. Gastric varices are seen in about 25% patients and a bleed from them can be managed with cyanoacrylate glue injection or surgery. Other indications for surgery include failure of endoscopic therapy to control acute bleed and symptomatic hypersplenism. The prognosis of patients with NCPF is good and 5-years survival rates in patients in whom variceal bleeding can be controlled is about > 95%.

Non-cirrhotic portal fibrosis: current concepts and management

Non-cirrhotic portal hypertension (NCPH) comprises diseases having an increase in portal pressure (PP) due to intraheptic or prehepatic lesions, in the absence of cirrhosis. The lesions are generally vascular, either in the portal vein, its branches or in the perisinusoidal area. Because the wedged hepatic venous pressure is near normal, measurement of intravariceal or intrasplenic pressure is needed to assess PP. The majority of diseases included in the category of NCPH are well-characterized disease entities where portal hypertension (PHT) is a late manifestation and, hence, these are not discussed. Two diseases that present only with features of PHT and are common in developing countries are non-cirrhotic portal fibrosis (NCPF) and extrahepatic portal vein obstruction (EHPVO). Non-cirrhotic portal fibrosis is a syndrome of obscure etiology, characterized by 'obliterative portovenopathy' leading to PHT, massive splenomegaly and well-tolerated episodes of variceal bleeding in young adults from low socioeconomic backgrounds, having near normal hepatic functions. In some parts of the world, NCPF is called idiopathic portal hypertension (IPH) or 'hepatoportal sclerosis'. Because 85-95% of patients with NCPF and EHPVO present with variceal bleeding, treatment involves management with endoscopic sclerotherapy (EST) or variceal ligation (EVL). These therapies are effective in approximately 90-95% of patients. Gastric varices are another common cause of upper gastrointestinal bleeding in these patients and these can be managed with cyanoacrylate glue injection or surgery. Other indications for surgery include failure of EST/EVL, and symptomatic hypersplenism. The prognosis of patients with NCPF is good and 5 years survival in patients in whom variceal bleeding can be controlled has been reported to be approximately 95-100%.

Extrahepatic portal vein thrombosis: aetiology and determinants of survival

BACKGROUND

Malignancy, hypercoagulability, and conditions leading to decreased portal flow have been reported to contribute to the aetiology of extrahepatic portal vein thrombosis (EPVT). Mortality of patients with EPVT may be associated with these concurrent medical conditions or with manifestations of portal hypertension, such as variceal haemorrhage.

PATIENTS AND METHODS

To determine which variables have prognostic significance with respect to survival, we performed a retrospective study of 172 adult EPVT patients who were followed over the period 1984-1997 in eight university hospitals.

RESULTS

Mean follow up was 3.9 years (range 0.1-13.1). Overall survival was 70% (95% confidence interval (CI) 62-76%) at one year, 61% (95% CI, 52-67%) at five years, and 54% (95% CI, 45-62%) at 10 years. The one, five, and 10 year survival rates in the absence of cancer, cirrhosis, and mesenteric vein thrombosis were 95% (95% CI 87-98%), 89% (95% CI 78-94%), and 81% (95% CI 67-89%), respectively (n=83). Variables at diagnosis associated with reduced survival according to multivariate analysis were advanced age, malignancy, cirrhosis, mesenteric vein thrombosis, absence of abdominal inflammation, and serum levels of aminotransferase and albumin. The presence of variceal haemorrhage and myeloproliferative disorders did not influence survival. Only four patients died due to variceal haemorrhage and one due to complications of a portosystemic shunt procedure.

CONCLUSION

We conclude that mortality among patients with EPVT is related primarily to concurrent disorders leading to EPVT and not to complications of portal hypertension.

Symptomatic spleno-mesenteric-portal venous thrombosis: recanalization and reconstruction with endovascular stents

PURPOSE

To evaluate the safety and efficacy of portal reconstruction in patients with symptomatic spleno-mesenteric-portal venous thrombosis.

MATERIALS AND METHODS

Portal reconstruction was attempted in 21 patients (seven women, 14 men; mean age, 53.6 years +/- 15.2) with chronic thrombosis of the portal vein alone (n = 8), splenic vein alone (n = 3), or portal, mesenteric, and splenic veins (n = 10). Indications for the procedure were bleeding varices (n = 15), ascites (n = 2), hypersplenism (n = 2), and enteropathy (n = 2). Sixteen procedures were started transhepatically and of these seven were converted to a transjugular intrahepatic portosystemic shunt (TIPS) after successful recanalization of the thrombosed vein. In six patients reconstructions were performed using an intrahepatic portal vein as outflow. Five procedures were performed primarily as TIPS. Wallstents dilated to 7-10 mm were used for reconstruction. The mean follow-up period was 15.2 months +/- 15.9.

RESULTS

Technical success of portal reconstruction was 85.7% (18 of 21). Thirty-day mortality was 14.3% (three of 21) but was not procedural related. The cumulative rates of survival, primary patency, and palliation at 43 months of follow-up were 61.2% +/- 13.5%, 63.5% +/- 15.3%, and 31.7% +/- 15.7%, respectively. Secondary patency was 79.1% +/- 13.8%. The only predictor of mortality was the presence of liver disease (P = .001, Cox regression).

CONCLUSION

Portal reconstruction is a safe and effective treatment option for patients with symptomatic chronic portal thrombosis. Liver disease predisposes to a higher mortality.

Direct bypassing of extrahepatic portal venous obstruction in children: a new technique for combined hepatic portal revascularization and treatment of extrahepatic portal hypertension

BACKGROUND

Decompression of extrahepatic portal hypertension by directly bypassing the thrombosed portal vein has never been reported in cases of children with idiopathic (or neonatal) portal vein obstruction and cavernoma.

METHODS

Seven children (15 years or younger) with portal vein obstruction requiring surgical decompression (urgently in two cases), and in whom preoperative Doppler had shown that the intrahepatic portal branches were hypoplastic but free of thrombus, were included in a pilot study. The cavernoma was bypassed by interposing a venous jugular autograft between the superior mesenteric vein and the distal portion of the left portal vein. Patients received follow-up using routine clinical parameters, upper gastrointestinal endoscopy, and Doppler ultrasound.

RESULTS

The mesenterico-portal bypass restored a direct (physiological) hepatopetal portal flow. The operation resulted in effective portal decompression as demonstrated by decrease of the pressure gradient, rapid regression of clinical signs of portal hypertension, and definitive control of bleeding.

CONCLUSIONS

This study shows that direct bypassing of portal cavernoma is possible and results in effective portal decompression. Restoration of the hepatic portal flow is a major advantage compared with conventional surgical shunting procedures. This new technique is potentially applicable to two thirds of children with portal vein thrombosis and should be considered when shunting procedures are indicated.

Extrahepatic portal venous obstruction in childhood: etiology, clinical and laboratory findings and prognosis of 34 patients

Extrahepatic obstruction of the portal vein is a well known cause of portal hypertension in childhood, that causes severe morbidity. We evaluated 34 children (24 boys, 10 girls, age 4.5 months to 12 years, mean 5.5 +/- 3.8 years) with this diagnosis, to define the clinical picture, laboratory changes, diagnostic tools and therapeutic modalities. Gastrointestinal bleeding was the commonest mode of presentation (64.7%), with the second being splenomegaly. The cause of the obstruction could be determined in 38.2% (13/34) of the subjects. At the beginning of the study the main diagnostic procedure was splenoportography although in more recent years pulsed duplex Doppler ultrasonography has been used. The follow up period was median of 5 years (range 1-11 years). The mean number of bleeding episodes was 4.7 +/- 5.9 (range 1-26), while nine patients never bled. There was no mortality. Ten patients underwent surgery, while sclerotherapy was performed on 10. Twenty-one patients received beta-blocker drugs. No difference was found among these therapeutic modalities. It is well established that the major risk for children with extrahepatic portal vein obstruction is gastrointestinal bleeding which is tolerated quite well. Surgery should be indicated only in children where bleeding cannot be controlled by medical means including sclerotherapy.

Noncirrhotic portal vein thrombosis causing sudden unexpected death. A rare cause of fatally ruptured esophageal varices

Noncirrhotic portal vein thrombosis (PVT) is a rare disease that usually presents with small nonfatal "herald bleeding" with low mortality. Classic findings of noncirrhotic PVT include esophageal varices, splenomegaly, a normal liver, and an organized thrombus of the portal vein. We present a case of previously unreported sudden unexpected death from noncirrhotic PVT in an asymptomatic elderly woman, review the literature of this entity, and examine the pathophysiology of the formation of fatally ruptured varices. The portal vein must be carefully examined in all cases where there is no coexisting cirrhosis.

Radical esophagogastrectomy for unshuntable extrahepatic portal hypertension with bleeding varices: report of a case

A 29-year-old woman with idiopathic portal hypertension was referred to our department for the surgical management of repetitive bleeding from esophageal and gastric varices. At the age of 16 years she had undergone a splenectomy with esophageal transection followed by endoscopic sclerotherapy which had been performed a total of 24 times. Although vericeal hemorrhage was prevented for several months, bleeding from gastric varices and portal hypertensive gastropathy was not able to be controlled readily by endoscopic sclerotherapy from when she was 26 years old. On admission, angiographic studies showed a complete obstruction of the portal vein; however, a portosystemic shunt operation was not able to be performed due to her previous splenectomy. To control her repetitive bleeding, we decided to perform a total gastrectomy and distal esophagectomy with reconstruction by a Roux-en-Y esophagojejunostomy. Her postoperative course was uneventful, and no episodes of recurrent bleeding or other complications have developed, indicating that her quality of life has dramatically improved. Thus, we conclude that distal esophagectomy and total gastrectomy constitute an effective surgical treatment for unshuntable extrahepatic portal hypertension.

Surgical management of portal hypertension

Portal hypertension is frequently complicated by upper gastrointestinal tract bleeding and ascites. Hemorrhage from esophageal varices is the most common cause of death from portal hypertension. Medical treatment, including resuscitation, vasoactive drugs, and endoscopic sclerosis, is the preferred initial therapy. Patients with refractory hemorrhage frequently are referred for immediate surgical intervention (usually emergency portacaval shunt). An additional cohort of patients with a history of at least 1 episode of variceal hemorrhage is likely to benefit from elective shunt operations. Shunt operations are classified as total, partial, or selective shunts based on their hemodynamic characteristics. Angiographically created shunts have been introduced recently as an alternative to operative shunts in certain circumstances. Devascularization of the esophagus or splenectomy is done for specific indications. Medically intractable ascites is a separate indication for surgical intervention. Liver transplantation has been advocated for patients whose portal hypertension is a consequence of end-stage liver disease. In the context of an increasingly complex set of treatment options, we present an overview of surgical therapy for complications of portal hypertension.

Non-conventional portosystemic shunts in children with extrahepatic portal vein obstruction

Portal hypertension due to extrahepatic portal vein obstruction is ideally treated by the use of a selective shunt. In a four year period between July 1987 and June 1992, 50 surgical procedures were carried out in 48 children with portal hypertension-related variceal hemorrhage. The portal, splenic, mesenteric, or coronary vein was not available for anastomosis in four children who, therefore, underwent non-conventional shunts which are described here. There was no postoperative mortality. One patient had a repeat gastrointestinal hemorrhage, probably due to stenosis of the shunt and failure to ligate the coronary vein. No rebleeding occurred on follow-up of at least 18 months. We conclude that in selected children requiring surgery for portal hypertension due to extrahepatic portal venous obstruction, a non-conventional shunt may be used with beneficial results.

Surgical management of extrahepatic portal hypertension and variceal bleeding

Variceal bleeding remains an important complication in extrahepatic portal vein thrombosis (EPVT). As for portal hypertension due to other etiologies, an elective treatment to decrease the risk of subsequent rebleeding is warranted. The results of the Sugiura-Futagawa procedure (SP) in 38 patients with variceal bleeding secondary to EPVT are reported: 20 women and 18 men, with a mean age of 28 +/- 2 years (SEM). Thirty-seven patients were classified as Child-Pugh class A, and one patient as class B. In terms of diagnosis, 45% of patients had idiopathic EPVT, and 18% had associated hypercoagulability disorders; 52% of patients had associated splenic vein thrombosis. The SP was completed in two surgical stages in 18 patients and in one surgical stage in 14; 6 patients had only the abdominal stage. One patient had mild postoperative encephalopathy, and three patients rebled at long-term follow-up study. There were two operative deaths. Actuarial survival was 70% at 64 months. It is concluded that the SP is an excellent alternative for patients with variceal bleeding secondary to EPVT.

Distal splenorenal shunting for bleeding gastric varices

In patients with bleeding gastric varices from causes other than splenic vein thrombosis, endoscopic sclerotherapy and ablative surgery have yielded poor results. Over a 3-year period starting in June 1989, a total of 30 distal splenorenal shunts were performed prospectively on 19 paediatric and 11 adult patients with bleeding gastric varices and good liver function. The mean (s.d.) age was 17(12) (range 6-50) years; there were 20 male and ten female patients of whom six had cirrhosis, four non-cirrhotic portal fibrosis and 20 portal vein thrombosis. Two patients died and two more had shunt thrombosis; all four were considered failures of treatment. Hypersplenism was present in 15 patients but reverted to normal in 13. In 26 patients the gastric varices disappeared. Concomitant oesophageal varices were present in 22 patients and showed marked regression, with no rebleeding over a mean (s.d.) follow-up of 21(10) (range 7-39) months. A distal splenorenal shunt was effective in controlling gastric variceal haemorrhage in 26 of 30 patients in whom liver function was well preserved.

Treatment of bleeding esophagogastric varices due to extrahepatic portal hypertension: results of portal-systemic shunts during 35 years

From 1958 to 1990, elective therapeutic portal-systemic shunt (PSS) procedures were performed for recurrent bleeding esophagogastric varices in 162 children and adults with extrahepatic portal hypertension (EHPH) resulting from portal vein thrombosis (PVT). The onset of EHPH was in childhood for at least 74% of patients. Of the 162 patients, 147 were eligible for and received 5 years of follow-up (100%), and 117 were eligible for and received 10 years of follow-up (100%). The longest follow-up was 35 years. The cause of PVT was unknown in 68%, neonatal omphalitis in 12%, umbilical vein catheterization in 8%, peritonitis in 6%, trauma in 4%, and thrombotic coagulopathy in 2%. The number of variceal bleeding episodes ranged from 2 to 18 (mean, 5.6). None of the patients had clinical, biochemical, or liver biopsy evidence of liver disease. Esophageal varices were demonstrated by endoscopy, and/or contrast x-rays, and/or angiography in all patients. Visceral angiography was always used to demonstrate the extent of portal obstruction and the veins available for shunting. Before referral, the following procedures had failed: endoscopic sclerotherapy (68 patients), splenectomy alone (32 patients), central splenorenal shunt with splenectomy (10 patients), transesophageal varix ligation (12 patients). Three types of PSS were used: (1) central side-to-side splenorenal without splenectomy (75 patients, 46%); (2) central end-to-side splenorenal with splenectomy (34 patients, 21%); and (3) mesocaval (end-to-side cavomesenteric) (53 patients, 33%). PSS reduced the mean corrected portal pressure from 292 to 28 mm saline. All patients survived the procedure and left the hospital (100%). The actuarial survival rate for 5 years is 99%, and for 10 years is 96%. Three of the 6 deaths were unrelated to EHPH or PSS. Shunt patency for up to 35 years was demonstrated in 98% of patients by angiography and/or ultrasonography. In four patients (2%), all of whom had end-to-side splenorenal shunts, shunt thrombosis and rebleeding developed 3, 4, 4, and 6 years (respectively) after PSS. There were the only patients who experienced rebleeding. A diligent and repeated effort was made to detect portal-systemic encephalopathy (PSE), and no instance of PSE was found during 3 to 35 years of follow-up. Liver function and morphology remained normal, and hypersplenism was corrected in all patients. Quality of life was good in 98% of patients, and 5 years after PSS 96% were gainfully employed, engaged in full-time homemaking, or attending school.(ABSTRACT TRUNCATED AT 400 WORDS)

Long-term results of radical esophagogastrectomy for bleeding varices due to unshuntable extrahepatic portal hypertension

This report describes the long-term results of one-stage total gastrectomy and distal two-thirds esophagectomy, with reconstruction by esophagojejunostomy (16 Roux-en-Y; 2 interposition), in 18 adult patients with recurrent variceal hemorrhage due to unshuntable extrahepatic portal hypertension (EHPH) from occlusion of all major tributaries of the portal venous system. The etiology of portal venous occlusion was unknown in 11 patients, abdominal trauma in 3, peritonitis in 3, and thrombotic coagulopathy in 1. Almost half of the patients had their first episode of bleeding in childhood, and 83% experienced bleeding before 40 years of age. The severity of the problem was reflected by frequent previous bleeding episodes (mean: 12.8, range: 4 to 21), a large cumulative requirement for blood transfusions (mean: 129 units, range: 28 to 247 units), repeated, costly hospital admissions (mean: 15, range: 4 to 24), and numerous previous unsuccessful operations (mean: 4.4, range: 1 to 14). Blood transfusions transmitted serum hepatitis to three patients and AIDS to one, for an incidence of 22%. Bleeding recurred after repetitive endoscopic sclerotherapy in 10 patients and after various operations in 16 (failed portal-systemic shunts in 9, splenectomy in 16, devascularization procedures in 13). All patients had large esophageal and gastric varices on endoscopy, normal liver function, and widespread portal venous occlusion on visceral angiography. Radical esophagogastrectomy was usually a long and arduous operation because of dense adhesions, extensive collateral veins, and a scarred, contracted bowel mesentery due to previous operations. All patients survived the operation and are currently alive. No patient has had recurrent bleeding during 1 to 26 years of follow-up (mean: 13.9 years, 7 or more years in 14 patients). Quality of life has been good. It is concluded that radical esophagogastrectomy is the only effective treatment of unshuntable EHPH and that the operation should be performed promptly when this disease, which is associated with high mortality, high morbidity, and high costs, is diagnosed.

A 15-year experience of injection sclerotherapy in adult patients with extrahepatic portal venous obstruction

OBJECTIVE

The authors report a 15-year experience with injection sclerotherapy in the management of adult and teenage patients with esophageal varices due to extrahepatic portal venous obstruction (EHPVO).

SUMMARY BACKGROUND DATA

Extrahepatic portal venous obstruction is an uncommon cause of esophageal varices and is associated with normal liver function. Effective control of variceal bleeding is the major factor influencing survival. The results of surgery have been unsatisfactory, and therefore, more conservative management policies have been adopted.

METHODS

Fifty-five patients with proven EHPVO underwent repeated injection sclerotherapy via either a modified rigid esophagoscope under general anaesthesia or a fiber-optic endoscope under light sedation, using ethanolamine oleate as the sclerosant.

RESULTS

Esophageal varices were eradicated in 44 patients after a median number 6 injections (range 1-17) over a mean of 12.5 months (range 1-48). The mean follow-up was 6.8 years (range 1.1-14.6 years). Eleven patients were admitted on eighteen occasions with bleeding from esophageal varices before eradication and there were seven bleeding episodes in six patients from recurrent varices after initial eradication. Complications related to sclerotherapy included injection site leak (6), stenosis (11) and mucosal ulceration (32) during 362 injection sclerotherapy episodes. Four patients died during the study period.

CONCLUSIONS

Injection scelotherapy is the treatment of choice in most patients with EHPVO.

Critical appraisal of the angiographic portacaval shunt (TIPS)

The transjugular intrahepatic portacaval shunt (TIPS) is a novel angiographic method for achieving portal decompression without operation. Fifty-nine consecutive patients underwent a total of 80 consecutive TIPS procedures. The procedure was unsuccessful in 4 patients (7%) and initially succeeded in 55 (93%). Eighteen patients (30%) underwent 2 or more TIPS procedures during the same hospitalization due to technical difficulties, early rebleeding, shunt stenosis, or thrombosis. Early TIPS occlusion occurred in seven patients (12%) and led to recurrent variceal hemorrhage in five. Forty-two percent of the cases of persisting or recurrent bleeding were nonvariceal. Procedure-related complications occurred in 10% of TIPS procedures or 14% of patients. Twenty-three patients (39%) were actively bleeding at the time of the procedure, and, in 6 of these (26%), bleeding was never controlled. In-hospital mortality (25%) was related only to the presence of bleeding at the time of TIPS (56% for emergent versus 5.5% for non-emergent, p < 0.0001). Mortality was not related to the Child-Pugh classification. Hemodynamic stabilization, vasoconstrictor therapy, balloon tamponade, and sclerotherapy were underutilized in 30% to 40% of patients prior to TIPS. Aggressive medical management should be used to stop variceal hemorrhage prior to TIPS in all patients, regardless of the Child-Pugh classification. Prospective trials comparing TIPS with sclerotherapy and surgical shunt are required to demonstrate the proper role of this procedure in the management of portal hypertension and variceal hemorrhage.