Esophageal transection versus injection sclerotherapy in the management of bleeding esophageal varices in patients at high risk

In a prospective randomized trial of 76 patients at high risk with bleeding esophageal varices, transection of the esophagus with the EEA stapling apparatus was compared with injection sclerotherapy in the management of patients with Child's class B and C liver status. Thirty-nine patients underwent transection and 37 patients, sclerotherapy with a total of 92 injection procedures (2.4 per patient). The perioperative mortality (less than 30 days) was 28.9 per cent overall; 33.3 per cent for esophageal transection and 24.3 per cent for injection sclerotherapy (chi 2 = 0.375, p greater than 0.05). Gross ascites, severe encephalopathy and emergency operations were associated with a high mortality in the transection group, but other risk factors such as age and hypersplenism did not influence the outcome in either group. Only patients in Child's class C died after transection, but patients who died in the sclerotherapy group (mainly from recurrent bleeding) included patients from both Child's class B and C. Early recurrence of nonfatal bleeding affected one of 39 patients (2.5 per cent) after transection but was evident in 18 of 37 patients (48.6 per cent) after sclerotherapy (chi 2 = 19.12, p greater than 0.0005) and six patients died. Hemorrhage did not recur after transection during a follow-up period of two years, but a further 22 episodes of bleeding were recorded in 13 patients receiving sclerotherapy with five deaths. Postoperative complications and long term morbidity were similar in the two groups. Including readmissions for bleeding and repeat procedures, the mean hospital stay per patient was shorter for transection (14.5 versus 19.1 days) and the requirements for blood were less (1.9 units per patient versus 3.6 units per patient) than for sclerotherapy. It is concluded that esophageal transection effectively protects against short term recurrence of bleeding. Preoperative control of gross ascites will further reduce the mortality and comatose patients should be excluded from operation. Sclerotherapy provides little if any protection against recurrent bleeding and its use in the management of variceal hemorrhage in patients with advanced liver disease remains questionable. It is recommended as a temporary measure in patients at high risk until such time that more effective surgical treatment can be performed.     

Portal hypertension

The management of both acute and recurrent variceal bleeding continues to be a significant challenge to the clinician. The cause and pathogenesis of portal hypertension has been described. Alcoholic cirrhosis is the most common cause of intrahepatic sinusoidal and postsinusoidal obstruction in the United States. Long term survival depends on rapid institution of an established protocol of surgical management for variceal hemorrhage. A patient who presents with variceal bleeding must be rapidly stabilized with fluid resuscitation, and specific measures, such as the use of vasopressin and balloon tamponade, must be instituted to control hemorrhage so that endoscopy can be used to establish the diagnosis. Sclerotherapy achieves a high rate of success in the acute situation, but if hemorrhage cannot be controlled, percutaneous transhepatic embolization or emergent shunting must be performed, depending on the condition of the patient. Angiography, prior to surgical treatment, is necessary to define venous anatomy and determine portal hemodynamics, both of which provide information vital in choosing the type of shunt. If bleeding is massive and the patient is unstable, H-grafts are most appropriate, for they are technically easier and give excellent short term results. In a stable Child's A or B patient with minor ascites as well as suitable anatomy and hepatopedal flow, DSRS is the procedure of choice because it produces the smallest degree of HE postoperatively and increases the survival rate for nonalcoholics. If this is not feasible or if the surgeon lacks the technical expertise to perform DSRS, PCS is the logical alternative. In view of the data from the series observed in the United States, ablative procedures cannot be recommended at the present for the treatment of variceal bleeding. In the Child's C poor-risk patient, the operative mortality rate is prohibitive, and only nonsurgical means should be used to establish control of bleeding. In the elective situation, the surgical options change. The efficacy of ES as a definitive procedure to control recurrent variceal bleeding is unproved, and rebleeding can be significant; therefore, it cannot be recommended. H-grafts have a prohibitively high rate of long term thrombosis and are also not recommended, and the Linton or proximal splenorenal shunt offers no advantages over conventional portacaval shunting. Moreover, arterialization of the hepatic stumps of the portal vein does not prevent hepatic encephalopathy or alter the survival rate. Both PCS and DSRS prevent rebleeding, yet neither alters the survival rate for alcoholic patients.(ABSTRACT TRUNCATED AT 400 WORDS)     

Long term results after portal disconnection of the esophagus using an anastomotic button for bleeding esophageal varices in cirrhosis

From 1968 to 1984, 250 patients with cirrhosis and bleeding esophageal varices underwent portal disconnection of the esophagus using either Murphy's button (before 1974) or an esophageal device developed by one of the authors (after 1974). One hundred and thirty-four patients underwent operation on an elective basis and 116 underwent emergency procedures. With the use of Child's classification, 62 patients were class A, 125 were class B and 63, class C. The over-all operative mortality rate was 24.4 per cent but this varied with the hepatic functional status and whether or not the operation was done on an elective or emergency basis. The long term survival rates were 53 per cent at one year, 36 per cent at three years, 24 per cent at five years and 8 per cent at ten years. Ninety-six per cent of the patients were without proved recurrent esophageal bleeding at one year, 88 per cent at three years, 79 per cent at five years and 66 per cent at ten years. Portal disconnection of the esophagus using an anastomotic button is a simple and effective procedure which can benefit many patients with cirrhosis who undergo an operation for bleeding varices on an elective or emergency basis. It constitutes an efficacious prophylactic means for preventing recurrent bleeding from esophageal varices.     

Portacaval shunt as emergency procedure in unselected patients with alcoholic cirrhosis.

A prospective evaluation of emergency portacaval shunt has been conducted during a 12 year period in 138 unselected, consecutive patients with alcoholic cirrhosis and bleeding esophageal varies. An extensive diagnostic evaluation was completed within seven hours of hospital admission, and the shunt operation was undertaken within a mean of 8.5 hours. Follow-up study was conducted in a special clinic, and the current status of 97.1 per cent of the patients had jaundice, ascites or encephalopathy alone or in combination on admission. Systemic intravenous administration of posterior pituitary extract temporarily controlled the hemorrhage in 94 per cent of the patients, and the emergency portacaval shunt promptly and permanently controlled the varix bleeding in 96 per cent of the patients. Contrary to recent proposals, patients with the highest portal perfusion pressure and, presumably, the largest hepatopetal portal flow had the highest survival rate and those who were presumed from pressure measurements to sustain the smallest portal flow diversion from the shunt had the lowest survival rate. The operative survival rate was 51 per cent, the predicted seven year survival rate for those operated upon seven or more years ago was 42.5 per cent. Encephalopathy requiring dietary protein restriction developed at some time in 17 per cent of the survivors. Sixty per cent of the survivors abstained from alcohol, and 53 per cent resumed gainful employment or full time housekeeping. Preoperative factors that adversely influenced survial rate were ingestion of alcohol within one month of bleeding, ascites, severe muscle-wasting and a small liver. Postoperatively, the single most important factor that compromised long term survival was resumption of alcoholism. In comparisons with our previous prospective studies, emergency portacaval shunt resulted in a significantly greater long term survival rate than did either emergency medical therapy or emergency varix ligation, followed by elective shunt. It is concluded that emergency portacaval shunt is the most effective treatment of bleeding esophageal varices in patients with alcoholic cirrhosis. Criteria for exclusion of those patients who are unlikely to derive long term benefits from portacaval shunt remain to be defined by further studies.     

Results in 100 consecutive patients with stapled esophageal transection for varices

Esophageal stapled transection has a role as an elective procedure in patients with varices who are unsuitable for shunt procedures, namely, the elderly, patients with Child's grade C, diabetics, patients with schistosomiasis and those with previous encephalopathy. It is effective in reducing variceal bleeding in the short term, although increasing rebleeding rates with longer follow-up periods are envisaged. When the alternative of injection sclerotherapy is available, there is probably little place for emergency transection in view of the high hospital mortality. A prospective randomized trial comparing chronic injection with esophageal transection is required.     

Left sided segmental portal hypertension following mesocaval interposition shunt.

Following a mesocaval interposition shunt in three patients with cirrhosis of the liver, bleeding esophageal varices recurred in two and left sided portal hypertension partially persisted in one patient. Angiographic and pressure studies of the portal system demonstrated effective decompression of the greater splanchnic venous system but continued lesser splanchnic venous hypertension. Recurrent variceal hemorrhage ceased following splenectomy done as an emergency. In contrast to a standard portacaval shunt, it is suggested that after an interposition mesocaval shunt, altered jet streaming of mesenteric blood flow may divert gastrosplenic venous drainage away from the interposition shunt with persistence of lesser splanchnic venous hypertension. Recognition of this entity and of the need for splenectomy is advocated.     

Esophageal transection by the EEA stapler for bleeding esophageal varices in schistosomal hepatic fibrosis

Thirty patients with esophageal varices due to schistosomal hepatic fibrosis underwent surgical treatment. They were divided into two groups. The elective group consisted of 20 patients who underwent splenectomy, gastroesophageal devascularization with esophageal transection using the EEA (U.S. Surgical Supply) stapler. The emergency group included ten patients, one of whom had the same procedure as those in the elective group and nine patients who had only esophageal transection with the EEA stapler. The use of the stapler in performing splenectomy and the devascularization operation did not decrease the recurrence of bleeding. Esophageal stapling in the emergency group controlled the bleeding in 90 per cent of the patients. The reappearance of esophageal varices and gastric varices in three patients in the emergency group was managed by distal splenorenal shunt operation.     

Emergency surgical treatment for nonvariceal bleeding of the upper part of the gastrointestinal tract

Endoscopic and biochemical data were collected prospectively from 1,530 patients admitted with nonvariceal bleeding of the upper part of the gastrointestinal tract between September 1985 and June 1989. Therapeutic endoscopy was done for 93 patients who underwent emergency surgical treatment for bleeding, subsequently required in 29 patients with seven postoperative fatalities. In contrast, 31 (15.7 per cent) of 198 patients (mortality rate of 9.6 per cent at 30 days) died in the hospital who had undergone emergency operation in whom therapeutic endoscopy had not been performed; data for this latter group is now presented. At admission, a greater likelihood of emergency operation was associated with a systolic blood pressure of 100 millimeters of mercury and endoscopic stigmatas of recent hemorrhage (ESRH) (p less than 0.001). Rebleeding rates for the presence of fresh blood, active spurting and oozing hemorrhage or visible vessel in an ulcer base were 26.5, 28.9 and 35.9 per cent, respectively. Endoscopic stigmatas were thus associated with an increased risk of bleeding (p less than 0.0001) and rebleeding led to a sixfold increase in the mortality rate. Congestive cardiac failure, chronic obstructive airway disease, chronic renal failure and a history of previous malignant disease were each associated with postoperative mortality rates of more than 50 per cent. An increased risk of mortality after emergency operation was related to age (p less than 0.0001), preoperative (p less than 0.002) and total (p less than 0.0001) blood transfusion requirement. Immediate operation after resuscitation and endoscopy was required in 87 patients; 11 deaths (hospital mortality rate of 12.7 per cent and 9.2 per cent at 30 days) occurred in this group compared with 20 fatalities (18.0 per cent) documented in 111 patients (9.9 per cent at 30 days) who underwent surgical treatment for rebleeding. We conclude that age, concomitant medical illness and preoperative and total transfusion requirements are each related to outcome after emergency operations. Such urgent intervention is best avoided if at all possible in patients with severe concomitant medical illness.     

Portosystemic shunt versus orthotopic liver transplantation for the Budd-Chiari syndrome.

We have analyzed the indications and results of shunt operation versus orthotopic liver transplantation (OLT) in 22 patients with Budd-Chiari syndrome (BCS). The underlying cause of the syndrome was similar between the two groups and was related to myeloproliferative disorders or the use of birth control pills in 18 of 22 patients. The results of biopsies of the liver showed centrilobular congestion and necrosis in all candidates who underwent shunting and the presence of fibrosis and cirrhosis in the OLT candidates. The indications for shunts included symptoms related to portal hypertension only and well-preserved synthetic hepatic function. Ten patients were treated with 12 shunt procedures, including mesoatrial (eight patients) and side to side portacaval shunt (four patients). Significant complications after shunt procedure included fulminant (one of ten patients) and progressive (one of ten patients) hepatic failure requiring urgent OLT; one death occurred because of pulmonary sepsis. Indications for OLT were signs of end stage liver expressed by severe portal hypertension and variceal bleeding (four of 14 patients), progressive encephalopathy (seven of 14 patients) and poor synthetic function (bilirubin greater than 3 milligrams per deciliter in eight of 14 patients and albumin less than 3.0 grams per liter, or both, in ten of 14 patients). Fourteen patients were treated with 16 OLT, three patients had retransplantation for primary nonfunction graft (two of 14 patients) or chronic rejection (one of 14 patients). There were two early deaths in the group. With a follow-up period between two months to five years, 12 of 14 patients undergoing OLT are alive, fully functional and have normal liver function tests. Seven of ten patients who had shunts are alive, six are able to maintain normal activity and one has progressive end stage hepatic disease and is not a candidate for OLT. However, the hepatic function continues progressively to be abnormal. Various options are available for the treatment of the syndrome. Portosystemic decompression is effective and should be considered at the early stage of the disease, prior to the development of significant hepatic failure. However, few of the patients will continue to have slow, but progressive hepatic failure and may require OLT. The only effective treatment for end stage hepatic disease secondary to the BCS is OLT.     

Injection thrombosclerosis of esophageal varices

It is now evident that injection sclerotherapy is a fast, effective method for controlling the acute variceal bleed. The results of sclerosis for control of acute variceal hemorrhage have been reported by many, including ourselves, and control rates vary from 78.0 to 95.0 per cent. Herein, we report a control rate of 94.7 per cent and a rebleeding rate of 22.0 per cent using a technical variation of sclerosis therapy. Thrombosclerosis is used soon after admission to the emergency room as a first line treatment to control variceal bleeding, often in lieu of the Sengstaken tube. Follow-up injections are scheduled at increasing intervals to achieve definitive control of the varices and are an essential feature of sclerotherapy as a management option.     

Portosystemic shunts for extrahepatic portal hypertension in children

Twenty-three children with prehepatic portal hypertension and hemorrhage due to ruptured esophagogastric varices had portosystemic shunts. Their ages ranged from two years and seven months to 15 years. Eleven were less than eight years of age. Twenty patients had portal vein cavernomatosis and three patients had double portal veins. In 21 patients, a mesocaval type of shunt was done. A splenorenal shunt was performed in two. There was no surgical mortality. Two shunts occluded, both in rather young infants--two years and seven months and three years of age. In all the others, there was no further bleeding, and the shunts remained patent, as shown by abdominal angiograms. Neuropsychiatric disorders, probably due to hepatic encephalopathy, occurred in only one patient. On the basis of this favorable experience, we believe that an elective portosystemic shunt should, in general, be performed upon children with prehepatic portal hypertension after one major variceal hemorrhage. We favor a mesocaval type of shunt in these children because of the larger diameter of the vessels involved in the anastomosis and because it preserves the spleen, maintaining defense against subsequent infection.     

Prospective comparison of partial versus total portal decompression for bleeding esophageal varices.

Eighty-six patients underwent portacaval shunt (PCS) to treat bleeding esophagogastric varices during a period of four years. Twenty-eight patients (group 1) underwent emergency total portal decompression, while 58 patients (group 2) underwent elective partial PCS. Age, gender, preshunt and postshunt alcohol consumption and modified Child-Pugh classification at the time of operation, and at latest follow-up evaluation, did not differ significantly between the two groups. Early mortality was higher after emergency shunts than after elective operation (p < 0.01). However, partial portal decompression, when compared with total shunt, resulted in a significantly lower likelihood of late mortality (13 versus 39 percent) (p < 0.05), as well as portasystemic encephalopathy (8 versus 56 percent) (p < 0.0005). All shunts remained patent postoperatively and no patient had variceal rebleeding during follow-up evaluation averaging 2.2 years. Duplex sonography demonstrated hepatofugal portal flow in all patients in both groups. The results of the current study suggest that partial portal decompression is technically feasible, prevents further variceal hemorrhage and confers significant protection against late mortality and the development of postshunt neuropsychologic dysfunction.     

Hepatic hemangioendotheliomas in infancy

Multicentric hepatic hemangioendotheliomas are vascular lesions of the liver that usually present in the infant with hepatomegaly, high output congestive heart failure and cutaneous hemangiomas. The diagnosis, pathologic and physiologic conditions and treatment were discussed. Two of the patients we studied and 117 from the literature were reviewed. A total of 38 patients survived and 57 died with or without medical treatment. Thirteen patients survived and five died after ligation of the hepatic artery. Five patients survived and three died after embolization of the hepatic artery. (formula; see text) Of 119 patients with MHH, 81 (68 per cent) had congestive heart failure; 40 survived and 41 died. Of 56 patients with localized hepatic angiomas, 19 had congestive heart failure. Only two of the patients died of heart failure. Fifty localized hepatic angiomas in infancy that were treated with local resection or lobectomy were reviewed. Forty-six patients survived and four died of hemorrhage. Four patients survived and two died with or without medical treatment. Plans for management of multiple hepatic hemangioendotheliomas and localized hepatic angiomas in infancy were proposed.     

Long term results of treatment of Budd-Chiari syndrome by side to side portacaval shunt

This is a report of a long term prospective study of 13 seriously ill patients with Budd-Chiari syndrome as a result of occlusion of the hepatic veins who were treated by side to side portacaval shunt from four to 78 weeks after the onset of symptoms and who were under observation for three to 16 years. The patient population was young, ranging in age from 19 to 45 years; seven were men and six were women. The presumed cause was the use of oral contraceptives in three, polycythemia rubra vera in two, Behcet disease in one patient and unknown in seven patients. All of the 13 patients had abdominal pain, marked ascites, hepatosplenomegaly, wasting and disturbed liver function. Diagnosis was based on the symptoms and signs: angiographic demonstration of hepatic vein occlusion and a patent inferior vena cava; pressure measurements that showed an inferior vena caval pressure that was normal or within the usual range for patients with massive ascites and an elevated wedged hepatic vein pressure that was much higher than the inferior vena caval pressure, and the results of biopsy of the liver showing centrilobular congestion and necrosis. Side to side portacaval shunt was very effective in decompressing the liver, reducing the mean corrected portal pressure from 240 millimeters of saline solution before to 7 millimeters of saline solution after the shunt. Operative survival rate was 92 per cent, and the long term survival rate for three to 16 years is 85 per cent. All of the survivors are free of ascites without requiring diuretic therapy.(ABSTRACT TRUNCATED AT 250 WORDS)     

Distal splenorenal shunt in treatment of bleeding esophageal varices in renal transplant recipients

We conclude from this study that bleeding esophageal varices may occur as a late complication of liver disease associated with chronic renal failure and renal transplantation. In two of the three patients reported upon, the liver disease was probably determined on the basis of cirrhosis, secondary to chronic, active hepatitis from non-A, non-B hepatitis, while the third patient had hepatic fibrosis. Such bleeding is best controlled by selective variceal decompression with a DSRS. Finally, it is technically feasible to perform a DSRS upon some patients following a left nephrectomy, and the renal vein is of adequate caliber even in the presence of nonfunctioning kidneys.     

Unresectable hepatic metastases from carcinoma of the colon and rectum

To alter the dismal prognosis of multiple unresectable metastases to the liver from carcinoma of the colon and rectum, 30 patients underwent hepatic dearterialization (ligation of the hepatic artery, transection of hepatic ligaments and cholecystectomy) and distal hepatic artery cannulation with prolonged infusion chemotherapy by a portable infusion pump followed by systemic intravenous chemotherapy. Involvement of the liver by carcinoma was less than 50 per cent in 16 and more than 50 per cent in 14 patients. The results of follow-up examinations, LFT, CEA and CT scan studies showed more than 50 per cent regression of the tumor and a decrease in alkaline phosphatase values and CEA in 29 patients (97 per cent); six had complete regression of tumor. The duration of response ranged from five to 39 months with the median of 17 or more months. The results of sequential LFT showed immediate increase in liver enzymes with return to normal in seven to 14 days. The mean CEA value decreased by 69 per cent within the first week and further decreased by 88 per cent in two months at the end of infusion chemotherapy. The over-all and adjusted survival rates from diagnosis were 79 and 91 per cent at 12 months; 56 and 81 per cent at 18 months, and 40 and 65 per cent at 24 months. The over-all and adjusted median survival rate after the treatment was 17 and 23 months, respectively. Of the 14 patients who failed this program, 11 had recurrences at sites other than the liver, with hepatic disease in remission in the majority. Of the 17 patients who died, six died of causes unrelated to the recurrence of disease. Thus, hepatic dearterialization and infusion regional chemotherapy can "effectively" control the hepatic disease and increase over-all survival time from three to six months to 23 months. However, recurrences of extrahepatic carcinoma and other causes are responsible for death and the over-all guarded prognosis.     

A comparison of hospital costs and morbidity between octogenarians and other patients undergoing general surgical operations

The hospital costs and clinical results of 304 patients who were more than 80 years old and who underwent general surgical procedures were evaluated. The over-all mortality rate was 14 per cent; 19.9 per cent occurred in patients admitted under emergency conditions as compared with 8.9 per cent that occurred in patients undergoing elective procedures (p less than 0.001). Seventy-nine per cent of the patients were discharged and 7 per cent required care in a skilled nursing facility. Survival rates were as good or better than standard life table survival rates for 80 year old patients. Costs were higher in those who were admitted under emergent conditions or who died in the hospital. Deaths were a result of complications of the primary disease rather than associated disease in most groups. Neither costs nor length of stay could accurately predict survival of individual patients. We concluded that health resources should be directed at treating problems, such as cholelithiasis, hernia or carcinoma, early before complications develop.     

Evaluation and treatment of early hemorrhage of the alimentary tract after selective shunt procedures

The cause and treatment of early variceal bleeding in 15 patients who had undergone distal splenorenal shunt were reviewed. Eight of these patients were taken from a group of 91 who underwent selective shunts from July 1983 through June 1985 and had extensive preoperative and postoperative evaluation of shunt patency and pressure gradient. Seven patients operated upon before July 1983 were reviewed because they illustrate the cause, diagnosis, successful and unsuccessful management of bleeding after selective shunt. Urgent selective arteriography combined with shunt catheterization is the key diagnostic and therapeutic maneuver. Thrombosis of the shunt can be successfully managed by revising the anastomosis. Stenosis of the shunt can be successfully treated with balloon dilation or operative revision of the anastomosis. When renal vein hypertension (RVH) occurs, there might be inadequate decompression of the varices. A gradient of 10 millimeters of mercury or greater from left renal vein to vena cava is diagnostic. Measurements of 30 patients who had no bleeding and one patient with documented RVH show the gradient decreases over time. Treatment should be supportive until this adaptation occurs. Hemorrhage can also occur in patients with a patient shunt but without a significant pressure gradient. Inadequate decompression of the varices through the short gastric veins leading to the spleen has been proposed as one cause. Termed short gastric hypertension, this syndrome could be expected to parallel RVH because the venous collaterals will enlarge and eventually decompress the varices. Treatment should be aimed toward supporting the patient until this adaptation occurs. A small number of patients continue to bleed despite these therapeutic interventions but can sometimes be salvaged with a total shunt.     

Acute renal failure following operation for aortic aneurysm.

Thirty-eight patients with acute renal failure following operation for aortic aneurysm were analyzed retrospectively in search of predictors of survival. Of 14 potential predictor variablesonsidered, none taken singly were significant; however, the combination of age, operative interruption of renal blood flow and prior renal dysfunction served as a significant predictor, p less than 0.05. Low survival rates occurred if the patient was over 70 years of age, if renal blood flow required interruption or if preoperative renal impairment was present. High survival rates occured in patients less than 70 years of age who had no interruption of renal blood flow and who had normal preoperative renal function. Although the total mortality rate of these patients with acute renal failure was 61 per cent, only one of 12 patients with favorable prognostic indicators died. Acute renal failure following aortic aneurysm repair has no worse prognosis that that stated in the literature for acute renal failure following other surgical procedures. A virorous therapeutic approach should be maintained despite the advent of complications.     

Effect of portasystemic shunts on subsequent transplantation of the liver

Portasystemic (PS) shunts have been regarded as a relative contraindication to hepatic transplantation (HT) because of the potential for increased technical difficulties during the transplant operation. We compared operative blood loss, morbidity and mortality in 27 patients with PS shunts and 147 patients with no shunts (NS) who underwent HT. The PS shunt group included 12 portocaval (PC), eight mesocaval, four central splenorenal and four distal splenorenal shunts. The PS shunt and NS groups were similar with respect to age, preoperative medical status and ABO blood group matching between donors and recipients. There were no significant differences in the mean (plus or minus S.D.) intraoperative blood transfusion (9.1 +/- 7.6 versus 9.2 +/- 11.0 units), mean (plus or minus S.D.) duration of anesthesia (8.1 +/- 1.4 versus 7.8 +/- 1.5 hours) and operative mortality (7 versus 11 per cent) between the PS shunt and NS groups, respectively. Complications of the biliary tract were significantly higher in the PS shunt group (22.0 versus 5.4 per cent, p less than 0.01) but they did not increase the mortality rate. We conclude that a prior PS shunt should not influence the decision to accept patients for HT. PS shunts remain a reasonable surgical option for patients with cirrhosis and variceal hemorrhage (refractory to sclerotherapy) who, by virtue of good hepatic function, do not merit immediate HT.