The surgeon's role in the management of portal hypertension

Clinical factors that influence the outcome of selective devascularization in the treatment of portal hypertension

There is a high incidence of death due to variceal hemorrhage in patients with portal hypertension. Factors to consider when choosing selective devascularization in the treatment of variceal hemorrhage remain a controversy. This study aims to generate the prevalent clinical risk factors that affect the outcomes of selective devascularization procedures. Elucidating these features may guide future treatment of esophageal varices in patients with portal hypertension. We retrospectively analyzed medical records of 455 patients who underwent selective devascularization procedures in our center. Patients were subject to splenectomy, selective devascularization with or without esophageal transection. The mode of surgery recurred in comparable rates in both the group with major complications postoperatively (high-risk group which consisted of 63 patients) or the group without major postoperative complications (low-risk group, 392). Risk factors that negatively influenced outcomes of surgery include severe symptoms (89% in high risk group and 71% in low risk group), large volume of blood loss in the hemorrhage before surgery (81% in high risk group and 16% in low risk group), sever liver cirrhosis (83% in high risk group and 67% in low risk group), previous endotherapy, prolonged prothrombin time, and poor liver function. Selective devascularization is a feasible option to treat variceal hemorrhage in patients with portal hypertension.

Hand-assisted laparoscopic splenectomy is a better choice for patients with supramassive splenomegaly due to liver cirrhosis

BACKGROUND

The current laparoscopic splenectomy (LS) procedure used for cirrhotic patients still has limitations. The aim of our study was to determine a standard according to the splenic size for hand-assisted LS (HALS) in patients with splenomegaly and hypersplenism due to cirrhosis.

PATIENTS AND METHODS

We conducted a retrospectively review of cirrhotic patients who underwent splenectomy between 2008 and 2011. All patients were divided into two groups: Group A (19 patients), in which patients' operations were conducted by HALS, and Group B (20 patients), in which patients were treated with LS. Then the patients in Group A were classified on the basis of the spleen size: massive splenomegaly (Group A1) and supramassive splenomegaly (Group A2). Likewise, so were patients in Group B: massive splenomegaly (Group B1) and supramassive splenomegaly (Group B2). Perioperative outcomes of these patients were compared.

RESULTS

The comparison of HALS and LS based on spleen size demonstrated that in the massive splenomegaly groups, Group A1 and Group B1 had similar estimated blood loss and morbidity, and no transfusion was required in these patients. In the supramassive splenomegaly groups, compared with Group A2, Group B2 had longer operative time, more estimated blood loss, more patients requiring transfusion, and more complications that needed surgical intervention. However, no significant differences were observed in the requirement of analgesia, time of returning to oral intake, and length of hospitalization in these paired groups.

CONCLUSIONS

In cirrhotic patients with supramassive splenomegaly, HALS should be considered because of its safety, feasibility, and effectiveness.

Transabdominal gastro-esophageal devascularization and esophageal transection for bleeding esophageal varices after failed injection sclerotherapy: long-term follow-up report

BACKGROUND

Management of continued bleeding from esophageal varices despite adequate injection sclerotherapy remains one of the medical and surgical dilemmas. Transabdominal gastroesophageal devascularization and esophageal transection (TGDET) is considered an effective and safe procedure for such patients.

AIM

This study aimed at presenting continued evaluation of TGDET. Various problems influencing the early outcome are discussed, and long-term outcome is analyzed.

DESIGN

This was a prospective clinical descriptive study.

METHODS

Prospective data was collected on 142 consecutive patients managed by one group of surgeons over a 5 year-period and 15 years follow-up after failed injection sclerotherapy for variceal bleeding. Evaluation was made in terms of effectiveness in controlling the acute bleeding, postoperative morbidity and mortality, recurrent bleeding, encephalopathy, and long-term survival.

RESULTS

There were 133 men and 9 women. Mean age was 41.8 years. Etiology of portal hypertension was bilharziasis in 54.9% and posthepatitic in 14.8%. Child-Pugh grading on admission was A: 47.2%, B: 28.8%, and C: 14%. Hemorrhage was controlled in all cases. Clinical leak was observed in 5.6%, portal vein thrombosis in 6.3%, and staple line erosion in 2.1% of cases. No patient developed encephalopathy. In-hospital mortality was 12.7%. Complete eradication of varices was observed in 70.6% patients. Recurrent variceal bleeding was noticed in 6.9% of cases. Actuarial 15-year survival for Child-Pugh A patients was 44%, B was 22.5%, and none for C.

CONCLUSION

TGDET remains a safe and effective procedure after failure of sclerotherapy when other alternatives are either not indicated or not available.

Variceal recurrence, rebleeding, and survival after endoscopic injection sclerotherapy in 287 alcoholic cirrhotic patients with bleeding esophageal varices

OBJECTIVE

This study tested the validity of the hypothesis that eradication of esophageal varices by repeated injection sclerotherapy would reduce recurrent variceal bleeding and death from bleeding varices in a high-risk cohort of alcoholic patients with cirrhosis.

SUMMARY BACKGROUND DATA

Although banding of esophageal varices is now regarded as the most effective method of endoscopic intervention, injection sclerotherapy is still widely used to control acute esophageal variceal bleeding as well as to eradicate varices to prevent recurrent bleeding. This large single-center prospective study provides data on the natural history of alcoholic cirrhotic patients with bleeding varices who underwent injection sclerotherapy.

METHODS

Between 1984 and 2001, 287 alcoholic cirrhotic patients (225 men, 62 women; mean age, 51.9 years; range, 24-87 years; Child-Pugh grades A, 39; B, 116; C, 132) underwent a total of 2565 upper gastrointestinal endoscopic sessions, which included 353 emergency and 1015 elective variceal injection treatments. Variceal rebleeding, eradication, recurrence, and survival were recorded.

RESULTS

Before eradication of varices was achieved, 104 (36.2%) of the 287 patients had a total of 170 further bleeding episodes after the first endoscopic intervention during the index hospital admission. Rebleeding was markedly reduced after eradication of varices. In 147 (80.7%) of 182 patients who survived more than 3 months, varices were eradicated after a mean of 5 injection sessions and remained eradicated in 69 patients (mean follow-up, 34.6 months; range, 1-174 months). Varices recurred in 78 patients and rebled in 45 of these patients. Median follow-up was 32.3 months (mean, 42.1 months; range, 3-198.9 months). Cumulative overall survival by life-table analysis was 67%, 42%, and 26% at 1, 3, and 5 years, respectively. A total of 201 (70%) patients died during follow-up. Liver failure was the most common cause of death.

CONCLUSION

Repeated sclerotherapy eradicates esophageal varices in most alcoholic cirrhotic patients with a reduction in rebleeding. Despite control of variceal bleeding, survival at 5 years was only 26% because of death due to liver failure in most patients.

Extrahepatic portal hypertension in children: observations on three surgical procedures

This paper presents a comparative prospective study of three modalities of surgical treatment for extrahepatic portal hypertension in children: central splenorenal shunt after splenectomy (CSS), side-to-side lienorenal shunt (SSLR) without splenectomy, and splenectomy and gastroesophageal devascularization (SGD). In an 18-month period, 27 procedures were performed: 10 CSS, 10 SSLR, and seven SGD. The outcomes were evaluated by fall in portal pressures, hematological parameters, shunt patency, splenic regression, and disappearance of esophageal varices. All three procedures were comparable in the fall of portal pressure after surgery. The average blood loss and operating time were statistically significant in favor of SSLR compared with CSS. At 3-month follow-up, shunt patency was confirmed by duplex Doppler study in all the patients in the SSLR group and in nine out of 10 patients in the CSS group. In the CSS and SGD groups, hypersplenism resolved in all the patients. In the SSLR group, blood counts improved in only five out of eight affected children. No patient re-bled during a follow-up of 3-5 years. There were no cases of hepatic encephalopathy or overwhelming postsplenectomy sepsis. In conclusion, CSS is useful when there is a large spleen, severe hypersplenism, and a shuntable splenic vein. SSLR is suitable when there is only mild splenomegaly, mild hypersplenism, and a shuntable splenic vein. Splenectomy and devascularization is the choice when there is no shuntable splenic vein.

An attempt to decrease ammonia levels after portacaval anastomosis in dogs: hepatic periarterial neurectomy

Hepatic encephalopathy and elevated serum ammonia levels occur commonly after portacaval shunt and are hypothesized to be, in part, due to decreased hepatic blood flow. Prior work has demonstrated increased blood flow to the liver following hepatic periarterial neurectomy. In this experimental study, we investigated the functional, hemodynamic, and histopathological changes in the liver and kidney occurring after the addition of hepatic periarterial neurectomy to side-to-side portacaval shunt in dogs. It is our hypothesis that the addition of hepatic periarterial neurectomy to portacaval shunt will decrease postshunt ammonia levels. Side-to-side portacaval shunt was performed in 12 dogs (group I). Hepatic periarterial neurectomy was added to portacaval shunt in 9 dogs (group II). Serum levels of ammonia, urea, creatinine, alanine aminotransferase, aspartate aminotransferase, alkaline phosphatase, albumin, and bilirubin together with hepatic blood flow were determined in both groups preoperatively and on postoperative day 21. The pre- and postoperative histopathologic changes of the liver and kidney were evaluated. There was significantly less postoperative elevation of serum ammonia and aspartate aminotransferase when hepatic periarterial neurectomy was added to the portacaval shunt procedure. Hemodynamic studies of hepatic artery and hepatic tissue indicated better blood flow in group II. The histopathologic evaluation of group II showed expansion of sinusoids, portal vessels, and portal areas and increased portal fibrosis as compared to group I. The results of this experimental study show that adding hepatic periarterial neurectomy to the portacaval shunt procedure improves postoperative serum levels of ammonia and aspartate aminotransferase and hepatic artery and tissue blood flow.

Splenocaval shunting for alleviation of portal hypertension in a dog: a case report

OBJECTIVE

To describe the construction and use of a splenocaval shunt to prevent portal hypertension in a dog with iatrogenic rupture and subsequent complete occlusion of an intrahepatic portosystemic shunt (IPSS).

STUDY DESIGN

Case report describing a single, client-owned animal.

RESULTS

During dissection, the back wall of an IPSS was torn. Complete shunt occlusion was required to control the hemorrhage. This resulted in the development of life-threatening portal hypertension. Emergency splenocaval shunt construction reduced the portal pressure from 47 to 20 cm H2O. The dog experienced minimal postoperative complications. A second surgical procedure was performed a month later to completely ligate the splenocaval shunt.

CONCLUSIONS

A splenocaval shunt can be used to divert blood from the portal to the systemic circulation to control portal hypertension. In this dog, it resulted in a successful outcome with few complications.

CLINICAL RELEVANCE

The splenocaval shunt could be constructed before the dissection of a difficult IPSS if problems arise as occurred in the dog described in this report. Complete IPSS occlusion can be performed without development of portal hypertension.

An evaluation of splenopancreatic disconnection as a modification of the distal splenorenal shunt, studied in nonalcoholic patients by sequential angiography

To evaluate the validity and complications of modifying the distal splenorenal shunt (DSRS) by performing splenopancreatic disconnection (SPD), hemodynamic changes in the portal system were assessed by visceral angiography in 93 patients with nonalcoholic portal hypertension during early postoperative follow-up after DSRS. There were 40 patients who underwent DSRS alone and 53 who underwent DSRS plus SPD. Early follow-up angiography showed that portal vein perfusion was well maintained, and that the diameter of the portal vein had decreased significantly by the same degree in both groups. Hepatofugal collaterals for the shunt had developed to a greater extent in the DSRS group, while they were almost completely absent in the DSRS with SPD group. Nevertheless, partial portal vein thrombosis was not detected in the DSRS group, although it was seen in seven (13.2%) of the patients who underwent DSRS plus SPD, in whom the left proximal splenic vein was not visible. The proximal splenic vein was seen in significantly less of the DSRS with SPD patients (47.2%) than the DSRS group patients (85%). In conclusion, SPD more effectively prevented the early postoperative development of collateral pathways for the shunt compared with standard DSRS; however, the possible stagnation of blood flow in the left proximal splenic vein may predispose to a risk of partial portal vein thrombosis developing during the early postoperative period after DSRS with SPD.

Surgical treatment for hepatocellular carcinoma and concomitant esophagogastric varices

Preventing a rupture of esophagogastric varices (EGV) is very important in aggressively treating hepatocellular carcinoma (HCC) in cirrhotic patients. We therefore performed simultaneous partial hepatic resection and direct interruption procedure on nine patients with HCC and concomitant EGV. Patients were selected on the basis of their stages of HCC and hepatic functional reserve. Postoperative hospital courses of all patients were uneventful. Six patients had recurrence of HCC and received non-surgical anti-tumor treatments. Only one patient had upper gastrointestinal bleeding at 18 months after operation, and the other eight patients have had no episodes of upper gastrointestinal bleeding during the follow-up period. The 5-year survival rate of these patients was 48%. This operative procedure is quite effective and is one of the treatments of choice for patients with less advanced HCC and concomitant risk of EGV.

Sclerotherapy versus sclerotherapy and propranolol in the prevention of rebleeding from oesophageal varices: a randomised study

BACKGROUND

This trial was carried out to assess the value of propranolol in the prevention of recurrent variceal bleeding when combined with longterm endoscopic sclerotherapy.

PATIENTS AND METHODS

Two hundred patients (161 male, 39 female, age range 20-68 years) with portal hypertension resulting mainly from schistosomal periportal fibrosis or posthepatitic cirrhosis presenting with their first episode of haematemesis or melena, or both were included. This was confirmed endoscopically to result from ruptured oesophageal varices. After initial control of bleeding, patients were randomised into two groups: group 1 treated with endoscopic sclerotherapy alone and group 2 treated with sclerotherapy plus propranolol. They were followed up for two years.

RESULTS

Group (2) had a lower rebleeding rate (14.3% v 38.6% in group 1), lower variceal recurrence after obliteration (17% v 34% in group 1), longer period between variceal obliteration and recurrence (36 weeks v 21 weeks in group 1); but no change in mortality (12% in both groups).

CONCLUSIONS

Patients treated with sclerotherapy should be given propranolol for longterm management.

Noninvasive variceal pressure measurement may be useful for predicting effect of sclerotherapy for esophageal varices

This study assessed the relationship between variceal pressure and morphological findings and hemodynamics of esophageal varices as well as the effect of sclerotherapy. Esophageal variceal pressure was measured in 40 patients with portal hypertension, using a noninvasive method. Esophageal variceal pressures were significantly higher in moderate or large varices than they were in small varices. Variceal pressures were significantly higher in patients with red color signs on the varices than in those without. According to the percutaneous transhepatic portography, the structure of the esophageal varices was classified into two types: the bar type and the palisading type. The maximum variceal pressure was significantly higher in the bar type than in the palisading type. In patients who underwent more than five sessions of sclerotherapy, the variceal pressure was significantly higher than in those receiving less than four sessions. These results suggest that variceal pressure may well reflect the vascular pattern and be useful for predicting the effect of sclerotherapy.

Surgical management of portal hypertension

The surgical management of portal hypertension depends on the location of the obstruction. Suprahepatic obstruction is usually optimally treated by a surgical portacaval shunt. In extrahepatic obstruction the treatment should be sclerotherapy. For intrahepatic obstruction in emergency situations, sclerotherapy is the first choice, with portacaval systemic shunts or transjugular intrahepatic portal systemic stent shunt the second option. Liver transplantation in other situations should, if possible, be considered ahead of a portal diversion.

Long-term management of variceal bleeding: the place of varix injection and ligation

Injection sclerotherapy remains the most widely used long-term management for patients after an esophageal variceal bleed. Sclerotherapy treatments should be repeated weekly until the varices are eradicated. Follow-up endoscopy every 6 to 12 months is required for life. Whenever varices recur, further weekly injection treatments are administered until re-eradication is achieved. Failure of sclerotherapy must be diagnosed early and an alternative salvage procedure performed. We currently recommend the distal splenorenal shunt. Although the complications of sclerotherapy are not great, they are cumulative with time. Unlike most surgical procedures for portal hypertension, the technique of performing sclerotherapy is not standardized, making the comparison of controlled trials difficult. The current status of controlled trials comparing sclerotherapy with other treatments is evaluated. We conclude that repeated injection sclerotherapy is at present the initial treatment of choice for patients after an esophageal variceal bleed. The technique of the new procedure of esophageal variceal ligation is described. As with sclerotherapy, weekly treatment sessions are recommended until the esophageal varices are eradicated, followed by long-term endoscopic surveillance and repeat ligation treatment when varices recur. The four controlled trials that have compared variceal ligation with sclerotherapy favor ligation. Ligation eradicated esophageal varices with fewer treatment sessions and a lower complication rate. One trial demonstrated improved survival. Complications due to the overtube are being increasingly reported but were not a problem in the controlled trials. Although esophageal variceal ligation or ligation plus sclerotherapy may ultimately prove to be superior to sclerotherapy alone, more data are required before a final conclusion can be reached.

Role of liver transplantation in management of esophageal variceal hemorrhage

The management of esophageal variceal hemorrhage ranges from conservative to surgical modalities. Before introduction of liver transplantation as a potentially curative therapy of the underlying etiology, decompressive portosystemic shunt operations have been the mainstay of mostly palliative procedures. Our own experience with surgery for advanced hepatic disease and portal hypertension over 20 years includes 803 liver transplantations and 201 portosystemic shunts, emphasizing our primary objective of treatment. The results after shunt surgery were favorable in Child class A candidates when performed electively and with selective decompression. After liver replacement the clinical status of the patient, including hepatic function and extrahepatic complications, had a strong influence on postoperative outcome, with the chance of excellent long-term survival. The additional risk of previous shunt surgery for subsequent transplantation could be reduced over time. Based on this experience and reports from others there are enough reasonable arguments for shunt and transplantation. Instead of the choice being controversial, the two forms of therapy should supplement each other and be available in the same center that specializes in the treatment of patients with diseases that eventually lead to liver failure and portal hypertension. Selection of either approach must depend on etiology, stage of the disease, and proper timing. Shunt procedures may be indicated in stable patients with the risk of bleeding after sclerotherapy failure, in those with contraindications to transplantation, or as a bridge to transplantation. The role of liver transplantation has been clearly established in patients with progressive or endstage (otherwise intractable) hepatobiliary disease.

Portal hypertension management

Injection sclerotherapy is the mainstay of treatment for acute variceal bleeding and for long-term management after a variceal bleed. In those few patients in whom sclerotherapy fails to control acute bleeding, either a surgical shunt or a simple esophageal transection is recommended. A surgical shunt or a more extensive esophagogastric devascularization and transection operation is advocated for the failures of long-term sclerotherapy management. The role of pharmacological agents in acute variceal bleed management remains in question, and the use of propranolol in long-term management, either as an alternative to sclerotherapy or in combination with sclerotherapy, is controversial. The definitive roles of the newly described variceal banding and transjugular intrahepatic porto-systemic shunts (TIPS) procedures have yet to be established. All patients presenting with end-stage liver disease and esophageal variceal bleeding should be evaluated for a liver transplant, although few will qualify. A possible future transplant should be kept in mind when emergency treatment is planned. Any form of prophylactic therapy for patients with esophageal varices that have not yet bled will remain unjustified until those patients at high risk of a first variceal bleed can be identified. The gastric mucosal lesion, portal hypertensive gastropathy, has been underdiagnosed in the past. Although bleeding does occur, it is seldom a major clinical problem. When necessary, bleeding can be controlled by propranolol or a surgical shunt.

Endoscopic injection sclerotherapy for 1,000 patients with esophageal varices: a nine-year prospective study

We report here the results of endoscopic injection sclerotherapy performed in 1,000 consecutively treated Japanese patients with esophageal varices. This prospective study covered the period from 1982 to 1990. Variceal bleeding was controlled in 215 (97.7%) of 220 patients. Esophageal varices were completely eradicated in 778 patients (77.8%); the mean number of sessions was 4.2. In only 3 of the 778 patients did esophageal varices of the same size recur. Small, dilated, venous vessels that required additional sclerotherapy in follow-up endoscopy at 3-mo intervals appeared in 171 (22.2%) of 778 patients. The cumulative nonbleeding rate at 5 yr was 94.5% in patients in whom the varices had been eradicated. Deaths caused by upper gastrointestinal bleeding accounted for 2.6% of cases, whereas the rates of liver failure and hepatoma were 4.6% and 47.3%, respectively. The 5-yr cumulative survival rate was 54.1% in patients without concomitant hepatoma; it was 12.0% in patients with hepatomas. Multivariate analysis showed that hepatoma, Child classification, indication (acute, elective or prophylactic) and eradication were independent factors that significantly influenced survival time. This study clearly shows that close follow-up with endoscopy and complete eradication lead to significant reduction in bleeding from esophageal varices and reduction of mortality related to this bleeding.

Issues in gastrointestinal endoscopy: oesophageal varices: inject, band, medicate, or operate

Injection sclerotherapy is the most widely used definitive treatment of acute variceal bleeding and is increasingly performed at the time of the first emergency endoscopy. Direct endoscopic ligation of varices by banding is a new technique under evaluation for both acute bleeding varices and long-term management. Repeated injection sclerotherapy is one of the major options for long-term management after variceal bleeding. More major surgical procedures are usually reserved for the failures of sclerotherapy in the management of acute variceal bleeding, whereas portosystemic shunts, particularly the distal splenorenal shunt, or an extensive devascularization and transection operation are commonly used alternative forms of therapy in long-term management. All patients with variceal bleeding should be assessed for liver transplantation, although only a few will ultimately receive a liver transplant. Medication with propranolol is widely recommended in long-term management, but its use in this context remains controversial. The most controversial area of management is prophylactic treatment before variceal bleeding. Major surgical procedures and injection sclerotherapy are not justified at present because it is difficult to identify those patients with a high likelihood of a first variceal bleed. Although medical therapy with propranolol has proved the most successful therapy to date, a case is made for treating most patients conservatively until their first variceal bleed occurs or until better predictive indices for patients at high risk of a first bleed are identified.

Surgical treatment of severe postshunt hepatic encephalopathy

Hepatic encephalopathy is a major complication of portal-systemic shunts with an incidence ranging up to 52%. A small fraction of these patients are refractory to medical therapy. Shunt ligation and colonic procedures are the main surgical approaches. The goal of the latter is to diminish the colonic absorption of nitrogenous substances which are involved in the pathophysiology of hepatic encephalopathy. Six patients, whose average age was 55.7 +/- 2.6 years, were operated for severe postshunt encephalopathy requiring 4.3 +/- 0.9 admissions for a total duration of 76 +/- 26 days over 1-11 years. One patient had undergone a splenoral shunt and 5 had a portacaval shunt. One ligation of the shunt and 5 colon exclusions were performed. The average postoperative hospital stay was 21.5 +/- 3.9 days. The mean follow-up was 47 +/- 20 months. The patient with the shunt ligation remains free of encephalopathy 94 months after the procedure and has not bled from his esophageal varices. Among the 5 colon exclusion patients, there were 1 death and 3 complications. Three patients were completely relieved of their hepatic encephalopathy. One of those 3 died of a subarachnoid hemorrhage 28 months after the surgery. The fourth still needs medication to control a persistent, although improved, encephalopathy that required 2 further hospitalizations. Colon exclusion is a useful intervention in very selected cases. It has a lower operative mortality than total colectomy and the advantage over shunt ligation of not reestablishing hypertension in the portal system.

Selective shunt versus nonshunt surgery for management of both schistosomal and nonschistosomal variceal bleeders

This clinical study included 219 (Child A/B) consecutive variceal bleeders. Electively 123 had distal splenorenal shunt (DSRS) and 96 had splenectomy with gastroesophageal devascularization (S&GD). Liver pathology was documented in 73% of patients, with schistosomal fibrosis in 41% and nonalcoholic cirrhosis or mixed pattern (fibrosis and cirrhosis) in 59%. The surgical groups were similar before operation, with a mean follow-up of 82 +/- 13 and 78 +/- 18 months, respectively (range, 60 to 120 months). The two pathologic populations were also similar before each and both procedures. The operative mortality rates were low, with incidences of 3.3% (DSRS) and 3.1% (S&GD). Rebleeding occurred significantly (p less than 0.05) more frequently after S&GD (27%) compared to DSRS (5.7%). Sclerotherapy salvaged 65% of S&GD rebleeders. Encephalopathy developed significantly (p less than 0.05) more after DSRS (18.7%) compared to S&GD (7.3%), with no significant difference among the current survivors. The difference in overall rebleeding and encephalopathy rates between both procedures was statistically related to patients with cirrhosis and mixed lesions (p less than 0.05). Distal splenorenal shunt significantly reduced the endoscopic variceal size more than S&GD (p less than 0.05). Prograde portal perfusion was documented in 94% of patients in each group, with a variable distinct pattern of portaprival collaterals in 91% (DSRS) and 65% (S&GD). The total population cumulative survival was similar with 80% for DSRS and 79% for S&GD (plus sclerosis in 23%), with hepatic cell failure the cause of death in 46% and 50%, respectively. However, in the schistosomal patients, survival was better improved after DSRS (90%) compared to S&GD (75%), with no difference among the cirrhotic and mixed group (DSRS 73%, S&GD 72%). In conclusion (1) both DSRS and S&GD have low operative mortality rates, (2) DSRS is superior to S&GD in the schistosomal patients, and (3) S&GD backed by endosclerosis for rebleeding is a good surgical alternative to selective shunt in the nonalcoholic cirrhotic and mixed population.