- Email: [email protected]
A modified sugiura operation for bleeding varices in children
A Modified Sugiura Operation for Bleeding Varices in Children By Riccardo A. Superina, James L. Weber, and Barry Shandling Toronto, Ontario, Canada 9 Objections to portal systemic shunting in children with life-threatening hemorrhage from esophageal varices include the high incidence of postshunt encephalopathy with neurologic and psychiatric sequelae and the inability to provide an adequate shunt in very young children. W e have operated on eight children in the past 4 years for bleeding varices. The causes were: portal vein thrombosis (3), congenital hepatic fibrosis (2), chronic active hepatitis (2), and cystic fibrosis (1). The ages at operation w e r e between 2 and 17 years. These children underwent various modifications of an operation described by Sugiura. The operation w e have developed is done through a single thoracoabdominal incision, dividing and anastomosing the esophagus with a stapler, preserving the vagal innervation to the pylorus and antrum, and wrapping the fundus around the distal esophagus at the site of the anastomosis. The venous drainage of the lower esophagus and of the upper stomach is divided. The operation is therefore shorter and simpler, but adheres to the principles enunciated by Sugiura. Complications include one significant postoperative anastomotic leak and one symptomatic esophageal stricture. Longterm results have been gratifying with no evidence of rebleeding from esophageal varices. W e believe that our modification of the original Sugiura operation is the preferred therapy of bleeding esophageal varices when surgical intervention is indicated because it preserves the normal structure and function of the upper gastrointestinal tract as well as the portal venous drainage to the liver. INDEX WORDS: Esophageal varices; Sugiura procedure.
ISSATISFACTION with the potential for development of serious late complications following portosystemic shunting for the treatment of bleeding esophageal varices has led to the development of alternate forms of treatment. These include a procedure described in 1973 by
D
From the Departments o f Pediatric Surgery and Gastroenterology, The Hospital for Sick Children, Toronto, Ontario. Presented before the 14th Annual Meeting of the American Pediatric Surgical Association, Hilton Head Island, South Carolina, May 4-7, 1983. Address reprint requests to Dr Barry Shandling, Department o f Surgery, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G lX8. 9 1983 by Grune & Stratton, Inc. 0022/3468/83/1806-0028501.00/0
794
Sugiura and Futagawa ~ in which the proximal stomach and distal esophagus were extensively devascularized thus interrupting the blood flow to the intraluminal varices yet maintaining naturally occurring paraesophageal portosystemic shunts. This procedure was carried out through separate abdominal and thoracic incisions and included complete vagotomy, gastric drainage procedure, and esophageal transsection and anastomosis. A modification of this procedure has been developed2 and the following is our experience in children with this technique. MATERIALS AND METHODS The operation is carried out with the patient supine and the left flank slightly elevated (Figs. 1, 2). A single thoracoabdominal incision is used, sweeping transversely upwards from above the umbilicus across the left upper quadrant and along the seventh intercostal space. The diaphragm is divided circumferentially close to its costal origin. If possible, the spleen is preserved. It often is removed, however, either as an initial step or because its large size may impede satisfactory exposure needed for the rest of the operation. In addition, the spleen may need to be excised for resolution of problems due to hypersplenism. Veins are divided between ligatures or clips along the greater curvature of the stomach and lateral border of the esophagus. Great care is taken to preserve the vagal trunks. A proximal gastric vagotomy is carried out at the same time devascularizing the lesser curvature of the stomach and carrying the process upwards across the esophageal hiatus along the esophagus until the inferior pulmonary vein is reached. Through an anterior gastrostomy the EEA stapler is inserted into the esophagus. The stapler is used to transect and anastomose the esophagus 2 cm above the gastroesophageal junction. The size of the stapler used varies with the size of the child but generally the largest size possible is preferred. The circumference of resected tissue is carefully inspected to make sure that the entire anastomosis includes the full thickness of the esophagus. The fundus of the stomach is then loosely wrapped around the lower esophagus as in a Nissen fundoplication as the final step. The patient is not fed for ten days after which a contrast study of the esophagus using a water soluble agent is obtained to assess the anastomosis for leaks.
Patients Between 1978 and 1982 eight children underwent various modifications of the Sugiura procedure. Their ages ranged from 2 to 17 years. All were referred to the surgical service because of repeated episodes of life-threatening hemorrhage Journal of Pediatric Surgery, Vol. 18, No. 6 (December),1983
MODIFIED SUGIURA
795 SUGiURA PROCEDURE
Table 1. Severity of Bleeding
1 2 3 4 5 6 7 8
Left Lateral Thoracor Upper ~,~id[i~e Abdominal Inelson with Left Lateral Extension
Age
History (Years)
Admissions
Blood Loss(rnL)
2 14 15 13 7 5 17 15
1 11 2 2 1 5 3 8
3 9 2 4 3 8 7 3
2250 5800 3000 11000 11500 13000 32500 2300
Severity of bleeding illustrated by length of history, the numbers of admissions which often exceeded one per year, and the accumulated blood transfused. Fig. 1.
Diagrahm of Sugiura Procedure
(Table 1). Esophageal varices were demonstrated by radiological and endoscopic means. Diagnoses were: portal vein thrombosis (3), congenital hepatic fibrosis (2), chronic active hepatitis (2), and cirrhosis from cystic fibrosis (1) (Table 2). All patients fit into Child's class A group and had normal liver function tests except for mild elevation of alkaline phosphatase in the patient with cystic fibrosis. One patient had had two failed mesocaval shunts and one failed proximal splenorenal shunt in addition to sclerotherapy before her devascularization procedure. Another patient had sclerotherapy complicated by esophageal perforation and consequent mediastinal abscess treated by drainage and antibiotics. Her parents refused further attempts to control hemorrhage by sclerotherapy.
RESULTS
Two children had the operation as described originally by Sugiura, three had the modified operation, and three had only one stage of the operation (2 chests, 1 abdomen). Five of the procedures were elective and three were urgent. Total operative time and blood loss for each case are listed in Table 3. MODIFIED SUGIURA PROCEDURE
ication
The operative time was markedly shorter for the modified procedure compared to the original operation and was only minimally longer than the partial procedures. Similarly, blood loss was greatly reduced in the modified operation. There were neither intraoperative deaths nor postoperative deaths related to the operation. Two patients had immediate postoperative upper gastrointestinal bleeding which required blood transfusion but resolved spontaneously. The first patient had had only the thoracic part of the devascularization and an esophageal transection. Sepsis from an anastomotic leak further complicated his postoperative course requiring tube drainage of the left pleural cavity and antibiotic therapy. Since his discharge, however, there has been no further bleeding and the second part of the operation has not been necessary. The second patient had a limited bleed associated with a minor anastomotic leak which was detected radiologicatly but was of no clinical significance. In the longterm there has been no evidence of further bleeding. It is possible that both these patients bled from their anastomosis although this was not proven. Two patients have had more chronic problems. The first patient has required admission to hospital for investigation of chronic iron deficiency anemia without any gross bleeding. This child was the patient who had had a mediastinal Table 2. Diagnoses in Eight Patients With Bleeding Esophageal Varices
Fig. 2.
Diagrahm of Modified Sugiura Procedure
Portal vein thrombosis Congenital hepatic fibrosis Chronic active hepatitis Cystic fibrosis
3 2 2 1
796
SUPERINA, WEBER, AND SHANDLING
DISCUSSION
Table 3. Comparison
O.R. Time BloodLoss Days in Hospital Sugiura (2 atages)
2
Sugiura (partial)
1 abdominal 1 chest 1 chest 3
Sugiura ( 1 stage)
10 11 5 4 4 6 5 6
3000 3000 1300 55 200 1800 1500 2200
31 31 17 22 38 12 53 (died) 33
Operative time and blood loss were less for the one-stage procedure than for the two-stage procedure and only slightly more than for the partial operation.
abscess from the sclerotherapy and had only had the abdominal part of the operation. Endoscopy revealed residual esophageal varices and fresh bleeding coming from gastric erosions. She has been managed successfully with oral cimetidine and iron supplements and has not required further surgery. The second patient developed a symptomatic anastomotic stricture 3 weeks postoperatively. A high-calorie liquid diet was well-tolerated but solid food produced dysphagia. This problem was managed by esophageal dilations and she had been taught to pass a Maloney bougie dilator at home in order for her to eat a normal diet. She is now dilating herself once a week with a number 48 bougie and we plan to discontinue all dilations soon.
Two patients have died from causes not related to their operation. One child with cystic fibrosis died of Pseudomonas cepatia pneumonia causing progressive respiratory deterioration. She died 53 days after her operation. She had resumed a normal diet on the tenth postoperative day. The second death occurred 14 months after the operation in an 8-year-old girl with ulcerative colitis and sclerosing cholangitis. Autopsy showed severe biliary cirrhosis and small bowel infarction from mesenteric vein thrombosis. No patient has demonstrated any evidence of encephalopathy, neurologic disorders or behavioural disturbances. No child has any dietary restrictions. In short, five of the six longterm survivors have been completely cured of their esophageal bleeding with no adverse gastrointestinal or neuropsychiatric sequelae. One patient has a stricture managed by self dilations and one patient has anemia from gastric erosions.
Variceal hemorrhage in children is often caused by portal vein obstruction or nonprogressive hepatic fibrosis where hepatocellular function is normal and the patient is spared the devastating complications of gastrointestinal bleeding to which patients with advanced hepatic dysfunction are prone. Hemorrhage is relatively well-tolerated and the frequency and severity of bleeding often diminish as the child approaches adulthood. For this reason repeated attempts at conservative therapy aimed at controlling the acute episode are justified. However, the disruptive influence of frequent hospital stays on a child's development and the severe limitations on the activity of the child's family imposed by the ever present threat of unexpected bleeding may demand a more aggressive approach in some cases. Portosystemic venous shunting had been regarded as the preferred method of longterm control of variceal bleeding in children. 3'4 However, the small size of the vessels in younger children make the operation technically difficult and predispose to shunt thrombosis) Of greater consequence than technical considerations, however, is the possible harmful effect of diverting portal blood away from the liver, and the development of neurological and emotional disturbances. In a careful study, Voorhees and coworkers studied 16 patients with portosystemic shunts with neurologic and psychiatric testing. 6 All 16 were found to manifest neurological or psychiatric abnormalities which were related in severity and character to the postoperative interval and to the co-existence of liver disease. Clearly, then, any procedure which may lead to undesirable and far reaching neurological aberrations must be critically reevaluated and used with great caution. Resective procedures such as esophagectomies and esophagogastrectomies, once thought to hold great promise in the treatment of extra hepatic portal hypertension, have been virtually abandoned because of the severe associated morbidity. 7,8 In a longterm follow-up of his results, Sugiura reported a low operative mortality and a remarkably low rate of rebleeding. There was no evidence of encephalopathy. 9 Our results are similar.
MODIFIED SUGIURA
797
With the modified Sugiura operation there have been no perioperative deaths and the associated morbidity has been u n c o m m o n and easily managed. The longterm results, even in the patient with bleeding following portosystemic shunting, have been gratifying. N o patient has returned with gross bleeding. The pattern of repeated admissions and multiple transfusions has been eliminated in all cases. In a 1- to 5-year follow-up there has been no evidence of neuropsychiatric problems and there has been no need to impose dietary restrictions. Other added advantages of this procedure are its applicability in very young children and its flexibility. Three of the patients were below the age of 10. In three children the devascularization was only limited to one-half of the usual extent and there has been no need for any further therapy in two. The use of the E E A stapler has considerably shortened operating time. Because it was designed for use in adults, the currently available sizes may be too large to fit in a child's esophagus. In this case the stapler m a y be applied to the circumference of the stomach just below the gastroesophageal junction where the same intramural venous tributaries m a y also be effectively interrupted. Because the spleen m a y have to be removed, all patients receive preoperative immunization with polyvalent antipneumococcal vaccine and
are placed on longterm antibiotic prophylaxis against postsplenectomy sepsis if necessary. Perioperative broad spectrum antibiotics are used for prophylaxis against wound infection. More recently, there has been renewed interest in controlling hemorrhage by the use of sclerotherapy and results have been promising. ~~ So far, however, results seem to indicate that there is a high incidence of rebleeding. In pediatric surgery there is a particular caveat against doing any h a r m for the longterm. R e a r r a n g e m e n t of portal blood flow by shunt operations carries a significant risk of encephalopathy with its devastating psychosocial consequences. If there is an acceptable alternative to shunt surgery we believe it should be used. O u r experience supports our belief that the modification of the original Sugiura operation is the preferred therapy of bleeding esophageal varices in children when surgical intervention is indicated because it preserves the normal structure and function of the upper gastrointestinal tract as well as the portal venous drainage to the liver. Indications for the operation include: failure of sclerotherapy in good-risk patients particularly those with extrahepatic obstruction; in younger children where shunts are not feasible; and when sclerotherapy is not available. W e favor the modified Sugiura operation over shunting procedures for children with bleeding varices.
REFERENCES
1. Sugiura M, Futagawa S: A new technique for treating esophageal varices. J Thorac Cardiovasc Surg 66:677-685, 1973 2. Ginsberg RJ, Waters PF, Zeldin RA, et al:A Modified Suguira procedure. Ann Thorac Surg 34:258-264, 1982 3. Fonkalsrud EW: Surgical Management of Portal Hypertension in Childhood. Arch Surg 115:1042-1045, 1980 4. Mitra SK, Datta DV, Rao PN, et al: Extrahepatic portal hypertension: A review of 70 cases. J Ped Surg 13:51-54, 1978 5. Arcari FA, Lynn HB: Bleeding Esophageal varices in children. Surg Gyn Obst 112:101-105, 1961 6. Voorhees AB, Chairman E, Schneider ~S, et al: Portal systemic encephalopathy in the noncirrhotic patient--Effect of portal sytemic shunting. Arch Surg 107:659-663, 1973 7. Perry JF, Root HD, Miller FA: Total removal of intrathoracic esophagus and antethoracic jejunal esophageal replacement for treatment of esophageal varices due to extrahepatic portal block. Ann Surg 158:126-128, 1963
8. Bernstein EG, Varco RL, Wangensteen OH: Treatment of bleeding esophageal varices in portal systemic shunt failures. Arch Surg 99:171-178, 1969 9. Sugiura M, Futagawa S: Further evaluation of the Suguira procedure in the treatment of esophageal varices. Arch Surg 112:1317-1321, 1977 10. Clark AW, Westaby D, Silk DBA, et al: Prospective controlled trial of injection sclerotherapy in patients with cirrhosis and recent variceal hemorrhage. Lancet 552-554, 1980 11. Johnston GW, Rodgers HW: A review of 15 years experience in the use of sclerotherapy in the control of acute hemorrhage from esophageal varices. Brit J Surg 60:797800, 1973 12. Terblanche J, Northover JMA, Bornman P: A prospective controlled trial of sclerotherapy in the long term management of patients after esophageal variceal bleeding. Surg Gynecol Obstet 148:323-333, 1979
798
SUPERINA, WEBER, AND SHANDLING
Discussion Eric W. Fonkalsrud (Los Angeles): Although many children with bleeding varices can be controlled with repeated sclerotherapy, at least half will require shunt operations or direct control of bleeding. For children over age 10 years with intrahepatic block, the distal splenorenal shunt has worked well and obviates many complications of central shunts. The mesocaval shunt has been the next most favored. These two shunts have also been successful with long follow-up for extrahepatic block. Almost half the children with extrahepatic block have unshuntable veins, and the modified Sugiura procedure may be particularly helpful. Injury to the vagus nerves is common while devascularizing the esophagus and stomach; we have therefore performed a pyloroplasty in each of our six patients. We question the merit of dividing the esophagus since anastomotic leaks in a devascularized area may lead to considerable morbidity. A gastrostomy with oversewing of gastric varices is helpful since they are difficult to sclerose or eradicate with devascularization. Transabdominal exposure is adequate for esophagogastric devascularization; however, the crura must be closed adequately. Splenectomy for portal hypertension in children is rarely necessary since hypersplenism is rarely severe. Splenic artery ligation may reduce splenic size. Most authors have rarely observed encephalopathic changes in children with extrahepatic block following shunts. Why is a fundoplication necessary, since there is little evidence to indicate postoperative reflux unless the vagus nerves have been injured and a pyloroplasty not performed? Caution should be used in referring to a cure for bleeding in portal hypertension with less than a 5-year follow-up. John R. Wesley (Ann Arbor, Michigan): In Ann Arbor we have been using the central splenorenal shunt with excellent results with basically the same indications as those that you reported. We have not had shunt failure in eight successive patients, and I think that the splenorenal shunt has been a particularly useful shunt in those patients where hypersplenism with platelet counts below 70,000 constituted the principle
indication for operation. In our particular group, the diameter of the constructed shunts have ranged from 40 to 20 mm and seven of the eight patients underwent a postoperative angiogram demonstrating shunt patency. Follow-up ranging from 1 to 5 years has shown no cases of postoperative gastrointestinal bleeding and no evidence of hepatic encephalopathy. I am concerned with the Sugiura technique because of the risk of inducing gastroesophageal reflux and I wonder whether you look for postoperative GE reflux--even in /those patients where you perform a Nissen fundoplication. In the patient that you describe with the postoperative stricture, did you feel that this was due to reflux or the devascularization procedure itself? Are any of your pediatric surgical colleagues in Toronto performing shunt procedures for similar patients, and how do their results compare with the results of your modified Sugiura series? Thomas Boles (Columbus, Ohio): I would like to comment on three aspects of this paper. One, a direct operation is obviously important for portal hypertension and I hope that the Sugiura operation or modification of it will be it. I am concerned, however, that the follow-up on the procedure described by Dr Superina is not long enough to make that statement. We did direct operations such as the Tanner procedure 20 years ago; and they worked very well for about 5 years, following which all of those patients began to bleed again. So, longterm follow-up is clearly important. Secondly, it has been 10 years since Voorhees reported on encephalopathy with extrahepatic portal hypertension. We looked up our cases last year, and in some 37 with follow-up periods of 5 years or longer there were no patients that had neuropsychiatric abnormalities. A shunt is still a very good operation. It would appear that in most people's experience, the risk of encephalopathy is relatively minor. Al deLorimier (San Francisco): I have performed six modified Sugiura procedures in which a splenectomy was not done and it included a Nissen fundoplication. Every one of them have rebled. Let me warn you that this is not a panacea. It is a procedure that will allow time for growth, so that the mesenteric vessels will
MODIFIED SUGIURA
enlarge enough to provide a successful shunt. However, it does not seem to provide longterm control of bleeding. Lower esophageal stricture is a problem in some of the smaller children where the esophageal mucosa is transected and reanastomosed. This can be a formidable problem to manage. It may be that transgastric ligation of varices might be preferred until a shunt is possible. Scott J. Boley (Bronx): I was not going to comment, but after Dr deLorimier has reported his poor results I think I have to respond. I have performed nine of these operations and none of the patient s has rebled. A major point that bothered me during Dr deLorimier's comments was when he said that he had modified the operation. We have not done a modified Sugiura, we have done the procedure as Sugiura described it. We have not used the stapler, and we have not done a fundoplication. We have performed postoperative 24 hour pH studies, and none of the patients have shown reflux. The patients do get recurrent varices, but do not bleed. We have reported them and John Ransom from N Y U has reported them. I think Dr Boles is absolutely correct, we have to watch these patients a lot longer than we have, but until now they have not rebled. I believe the main problem causing conflicting results arises when we start modifying other people's procedures. We then have no right to compare our results with theirs. If you perform the operation as Sugiura does it, I believe the operation has real potential. I do not believe you should be left with the pessimistic outlook that Dr delorimier has presented. Riccardo Superina (closing): I think there are three issues that really will cover most of those questions. Regarding the question of the incidence of encephalopathy, let me point out that Voorhees really went to a lot of trouble to demonstrate encephalopathy in these children and I don't know whether anybody has really looked at
799
their patients as carefully as he has. I know we don't always have the opportunity of seeing the patients for the purpose of testing them neurologically or psychiatrically and it may be that if you look for more subtle problems that you will pick them up. These may be just learning disabilities, who knows, but the potential is there and it is not there with the Sugiura procedure we should just stick to the Sugiura. Next let me comment on the modifications and the use of the stapler. The single incision is faster and simpler to close than the two incisions and I don't think that it makes any difference to the results. The EAA stapler really saves a lot of time and the two leaks, Dr Fonkalsrud, were in the patients that had had anastomosis before we started using the stapler. The experience in adults would suggest that the use of staplers has decreased the incidence of leaks both in esophageal and low rectal anastomoses. So, maybe the use of the staplers will make a difference in problems associated with leaks. We have only had three patients in whom we used the stapler and we think that strictures occur because an anastomosis is made in a relatively devascularized esophagus, and we don't think they are due to reflux. However, the precaution is taken of doing a fundoplication to help prevent reflux because we have really destroyed most of the mechanisms that normally prevent reflux. During the procedure the phrenoesophageal ligament is divided and the lower esophageal sphincter is manipulated and devascularized. So the Nissen procedure is a precaution against reflux, as well as providing a patch graft in case there is a leak. Hopefully this will take care of it. I don't want to discredit shunters but we believe the Sugiura procedure provides more reliable results and is more easily done in small children, and it is an alternate form of therapy in older children where shunts are not so readily done. Thank you very much for your comments.
ISSATISFACTION with the potential for development of serious late complications following portosystemic shunting for the treatment of bleeding esophageal varices has led to the development of alternate forms of treatment. These include a procedure described in 1973 by
D
From the Departments o f Pediatric Surgery and Gastroenterology, The Hospital for Sick Children, Toronto, Ontario. Presented before the 14th Annual Meeting of the American Pediatric Surgical Association, Hilton Head Island, South Carolina, May 4-7, 1983. Address reprint requests to Dr Barry Shandling, Department o f Surgery, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G lX8. 9 1983 by Grune & Stratton, Inc. 0022/3468/83/1806-0028501.00/0
794
Sugiura and Futagawa ~ in which the proximal stomach and distal esophagus were extensively devascularized thus interrupting the blood flow to the intraluminal varices yet maintaining naturally occurring paraesophageal portosystemic shunts. This procedure was carried out through separate abdominal and thoracic incisions and included complete vagotomy, gastric drainage procedure, and esophageal transsection and anastomosis. A modification of this procedure has been developed2 and the following is our experience in children with this technique. MATERIALS AND METHODS The operation is carried out with the patient supine and the left flank slightly elevated (Figs. 1, 2). A single thoracoabdominal incision is used, sweeping transversely upwards from above the umbilicus across the left upper quadrant and along the seventh intercostal space. The diaphragm is divided circumferentially close to its costal origin. If possible, the spleen is preserved. It often is removed, however, either as an initial step or because its large size may impede satisfactory exposure needed for the rest of the operation. In addition, the spleen may need to be excised for resolution of problems due to hypersplenism. Veins are divided between ligatures or clips along the greater curvature of the stomach and lateral border of the esophagus. Great care is taken to preserve the vagal trunks. A proximal gastric vagotomy is carried out at the same time devascularizing the lesser curvature of the stomach and carrying the process upwards across the esophageal hiatus along the esophagus until the inferior pulmonary vein is reached. Through an anterior gastrostomy the EEA stapler is inserted into the esophagus. The stapler is used to transect and anastomose the esophagus 2 cm above the gastroesophageal junction. The size of the stapler used varies with the size of the child but generally the largest size possible is preferred. The circumference of resected tissue is carefully inspected to make sure that the entire anastomosis includes the full thickness of the esophagus. The fundus of the stomach is then loosely wrapped around the lower esophagus as in a Nissen fundoplication as the final step. The patient is not fed for ten days after which a contrast study of the esophagus using a water soluble agent is obtained to assess the anastomosis for leaks.
Patients Between 1978 and 1982 eight children underwent various modifications of the Sugiura procedure. Their ages ranged from 2 to 17 years. All were referred to the surgical service because of repeated episodes of life-threatening hemorrhage Journal of Pediatric Surgery, Vol. 18, No. 6 (December),1983
MODIFIED SUGIURA
795 SUGiURA PROCEDURE
Table 1. Severity of Bleeding
1 2 3 4 5 6 7 8
Left Lateral Thoracor Upper ~,~id[i~e Abdominal Inelson with Left Lateral Extension
Age
History (Years)
Admissions
Blood Loss(rnL)
2 14 15 13 7 5 17 15
1 11 2 2 1 5 3 8
3 9 2 4 3 8 7 3
2250 5800 3000 11000 11500 13000 32500 2300
Severity of bleeding illustrated by length of history, the numbers of admissions which often exceeded one per year, and the accumulated blood transfused. Fig. 1.
Diagrahm of Sugiura Procedure
(Table 1). Esophageal varices were demonstrated by radiological and endoscopic means. Diagnoses were: portal vein thrombosis (3), congenital hepatic fibrosis (2), chronic active hepatitis (2), and cirrhosis from cystic fibrosis (1) (Table 2). All patients fit into Child's class A group and had normal liver function tests except for mild elevation of alkaline phosphatase in the patient with cystic fibrosis. One patient had had two failed mesocaval shunts and one failed proximal splenorenal shunt in addition to sclerotherapy before her devascularization procedure. Another patient had sclerotherapy complicated by esophageal perforation and consequent mediastinal abscess treated by drainage and antibiotics. Her parents refused further attempts to control hemorrhage by sclerotherapy.
RESULTS
Two children had the operation as described originally by Sugiura, three had the modified operation, and three had only one stage of the operation (2 chests, 1 abdomen). Five of the procedures were elective and three were urgent. Total operative time and blood loss for each case are listed in Table 3. MODIFIED SUGIURA PROCEDURE
ication
The operative time was markedly shorter for the modified procedure compared to the original operation and was only minimally longer than the partial procedures. Similarly, blood loss was greatly reduced in the modified operation. There were neither intraoperative deaths nor postoperative deaths related to the operation. Two patients had immediate postoperative upper gastrointestinal bleeding which required blood transfusion but resolved spontaneously. The first patient had had only the thoracic part of the devascularization and an esophageal transection. Sepsis from an anastomotic leak further complicated his postoperative course requiring tube drainage of the left pleural cavity and antibiotic therapy. Since his discharge, however, there has been no further bleeding and the second part of the operation has not been necessary. The second patient had a limited bleed associated with a minor anastomotic leak which was detected radiologicatly but was of no clinical significance. In the longterm there has been no evidence of further bleeding. It is possible that both these patients bled from their anastomosis although this was not proven. Two patients have had more chronic problems. The first patient has required admission to hospital for investigation of chronic iron deficiency anemia without any gross bleeding. This child was the patient who had had a mediastinal Table 2. Diagnoses in Eight Patients With Bleeding Esophageal Varices
Fig. 2.
Diagrahm of Modified Sugiura Procedure
Portal vein thrombosis Congenital hepatic fibrosis Chronic active hepatitis Cystic fibrosis
3 2 2 1
796
SUPERINA, WEBER, AND SHANDLING
DISCUSSION
Table 3. Comparison
O.R. Time BloodLoss Days in Hospital Sugiura (2 atages)
2
Sugiura (partial)
1 abdominal 1 chest 1 chest 3
Sugiura ( 1 stage)
10 11 5 4 4 6 5 6
3000 3000 1300 55 200 1800 1500 2200
31 31 17 22 38 12 53 (died) 33
Operative time and blood loss were less for the one-stage procedure than for the two-stage procedure and only slightly more than for the partial operation.
abscess from the sclerotherapy and had only had the abdominal part of the operation. Endoscopy revealed residual esophageal varices and fresh bleeding coming from gastric erosions. She has been managed successfully with oral cimetidine and iron supplements and has not required further surgery. The second patient developed a symptomatic anastomotic stricture 3 weeks postoperatively. A high-calorie liquid diet was well-tolerated but solid food produced dysphagia. This problem was managed by esophageal dilations and she had been taught to pass a Maloney bougie dilator at home in order for her to eat a normal diet. She is now dilating herself once a week with a number 48 bougie and we plan to discontinue all dilations soon.
Two patients have died from causes not related to their operation. One child with cystic fibrosis died of Pseudomonas cepatia pneumonia causing progressive respiratory deterioration. She died 53 days after her operation. She had resumed a normal diet on the tenth postoperative day. The second death occurred 14 months after the operation in an 8-year-old girl with ulcerative colitis and sclerosing cholangitis. Autopsy showed severe biliary cirrhosis and small bowel infarction from mesenteric vein thrombosis. No patient has demonstrated any evidence of encephalopathy, neurologic disorders or behavioural disturbances. No child has any dietary restrictions. In short, five of the six longterm survivors have been completely cured of their esophageal bleeding with no adverse gastrointestinal or neuropsychiatric sequelae. One patient has a stricture managed by self dilations and one patient has anemia from gastric erosions.
Variceal hemorrhage in children is often caused by portal vein obstruction or nonprogressive hepatic fibrosis where hepatocellular function is normal and the patient is spared the devastating complications of gastrointestinal bleeding to which patients with advanced hepatic dysfunction are prone. Hemorrhage is relatively well-tolerated and the frequency and severity of bleeding often diminish as the child approaches adulthood. For this reason repeated attempts at conservative therapy aimed at controlling the acute episode are justified. However, the disruptive influence of frequent hospital stays on a child's development and the severe limitations on the activity of the child's family imposed by the ever present threat of unexpected bleeding may demand a more aggressive approach in some cases. Portosystemic venous shunting had been regarded as the preferred method of longterm control of variceal bleeding in children. 3'4 However, the small size of the vessels in younger children make the operation technically difficult and predispose to shunt thrombosis) Of greater consequence than technical considerations, however, is the possible harmful effect of diverting portal blood away from the liver, and the development of neurological and emotional disturbances. In a careful study, Voorhees and coworkers studied 16 patients with portosystemic shunts with neurologic and psychiatric testing. 6 All 16 were found to manifest neurological or psychiatric abnormalities which were related in severity and character to the postoperative interval and to the co-existence of liver disease. Clearly, then, any procedure which may lead to undesirable and far reaching neurological aberrations must be critically reevaluated and used with great caution. Resective procedures such as esophagectomies and esophagogastrectomies, once thought to hold great promise in the treatment of extra hepatic portal hypertension, have been virtually abandoned because of the severe associated morbidity. 7,8 In a longterm follow-up of his results, Sugiura reported a low operative mortality and a remarkably low rate of rebleeding. There was no evidence of encephalopathy. 9 Our results are similar.
MODIFIED SUGIURA
797
With the modified Sugiura operation there have been no perioperative deaths and the associated morbidity has been u n c o m m o n and easily managed. The longterm results, even in the patient with bleeding following portosystemic shunting, have been gratifying. N o patient has returned with gross bleeding. The pattern of repeated admissions and multiple transfusions has been eliminated in all cases. In a 1- to 5-year follow-up there has been no evidence of neuropsychiatric problems and there has been no need to impose dietary restrictions. Other added advantages of this procedure are its applicability in very young children and its flexibility. Three of the patients were below the age of 10. In three children the devascularization was only limited to one-half of the usual extent and there has been no need for any further therapy in two. The use of the E E A stapler has considerably shortened operating time. Because it was designed for use in adults, the currently available sizes may be too large to fit in a child's esophagus. In this case the stapler m a y be applied to the circumference of the stomach just below the gastroesophageal junction where the same intramural venous tributaries m a y also be effectively interrupted. Because the spleen m a y have to be removed, all patients receive preoperative immunization with polyvalent antipneumococcal vaccine and
are placed on longterm antibiotic prophylaxis against postsplenectomy sepsis if necessary. Perioperative broad spectrum antibiotics are used for prophylaxis against wound infection. More recently, there has been renewed interest in controlling hemorrhage by the use of sclerotherapy and results have been promising. ~~ So far, however, results seem to indicate that there is a high incidence of rebleeding. In pediatric surgery there is a particular caveat against doing any h a r m for the longterm. R e a r r a n g e m e n t of portal blood flow by shunt operations carries a significant risk of encephalopathy with its devastating psychosocial consequences. If there is an acceptable alternative to shunt surgery we believe it should be used. O u r experience supports our belief that the modification of the original Sugiura operation is the preferred therapy of bleeding esophageal varices in children when surgical intervention is indicated because it preserves the normal structure and function of the upper gastrointestinal tract as well as the portal venous drainage to the liver. Indications for the operation include: failure of sclerotherapy in good-risk patients particularly those with extrahepatic obstruction; in younger children where shunts are not feasible; and when sclerotherapy is not available. W e favor the modified Sugiura operation over shunting procedures for children with bleeding varices.
REFERENCES
1. Sugiura M, Futagawa S: A new technique for treating esophageal varices. J Thorac Cardiovasc Surg 66:677-685, 1973 2. Ginsberg RJ, Waters PF, Zeldin RA, et al:A Modified Suguira procedure. Ann Thorac Surg 34:258-264, 1982 3. Fonkalsrud EW: Surgical Management of Portal Hypertension in Childhood. Arch Surg 115:1042-1045, 1980 4. Mitra SK, Datta DV, Rao PN, et al: Extrahepatic portal hypertension: A review of 70 cases. J Ped Surg 13:51-54, 1978 5. Arcari FA, Lynn HB: Bleeding Esophageal varices in children. Surg Gyn Obst 112:101-105, 1961 6. Voorhees AB, Chairman E, Schneider ~S, et al: Portal systemic encephalopathy in the noncirrhotic patient--Effect of portal sytemic shunting. Arch Surg 107:659-663, 1973 7. Perry JF, Root HD, Miller FA: Total removal of intrathoracic esophagus and antethoracic jejunal esophageal replacement for treatment of esophageal varices due to extrahepatic portal block. Ann Surg 158:126-128, 1963
8. Bernstein EG, Varco RL, Wangensteen OH: Treatment of bleeding esophageal varices in portal systemic shunt failures. Arch Surg 99:171-178, 1969 9. Sugiura M, Futagawa S: Further evaluation of the Suguira procedure in the treatment of esophageal varices. Arch Surg 112:1317-1321, 1977 10. Clark AW, Westaby D, Silk DBA, et al: Prospective controlled trial of injection sclerotherapy in patients with cirrhosis and recent variceal hemorrhage. Lancet 552-554, 1980 11. Johnston GW, Rodgers HW: A review of 15 years experience in the use of sclerotherapy in the control of acute hemorrhage from esophageal varices. Brit J Surg 60:797800, 1973 12. Terblanche J, Northover JMA, Bornman P: A prospective controlled trial of sclerotherapy in the long term management of patients after esophageal variceal bleeding. Surg Gynecol Obstet 148:323-333, 1979
798
SUPERINA, WEBER, AND SHANDLING
Discussion Eric W. Fonkalsrud (Los Angeles): Although many children with bleeding varices can be controlled with repeated sclerotherapy, at least half will require shunt operations or direct control of bleeding. For children over age 10 years with intrahepatic block, the distal splenorenal shunt has worked well and obviates many complications of central shunts. The mesocaval shunt has been the next most favored. These two shunts have also been successful with long follow-up for extrahepatic block. Almost half the children with extrahepatic block have unshuntable veins, and the modified Sugiura procedure may be particularly helpful. Injury to the vagus nerves is common while devascularizing the esophagus and stomach; we have therefore performed a pyloroplasty in each of our six patients. We question the merit of dividing the esophagus since anastomotic leaks in a devascularized area may lead to considerable morbidity. A gastrostomy with oversewing of gastric varices is helpful since they are difficult to sclerose or eradicate with devascularization. Transabdominal exposure is adequate for esophagogastric devascularization; however, the crura must be closed adequately. Splenectomy for portal hypertension in children is rarely necessary since hypersplenism is rarely severe. Splenic artery ligation may reduce splenic size. Most authors have rarely observed encephalopathic changes in children with extrahepatic block following shunts. Why is a fundoplication necessary, since there is little evidence to indicate postoperative reflux unless the vagus nerves have been injured and a pyloroplasty not performed? Caution should be used in referring to a cure for bleeding in portal hypertension with less than a 5-year follow-up. John R. Wesley (Ann Arbor, Michigan): In Ann Arbor we have been using the central splenorenal shunt with excellent results with basically the same indications as those that you reported. We have not had shunt failure in eight successive patients, and I think that the splenorenal shunt has been a particularly useful shunt in those patients where hypersplenism with platelet counts below 70,000 constituted the principle
indication for operation. In our particular group, the diameter of the constructed shunts have ranged from 40 to 20 mm and seven of the eight patients underwent a postoperative angiogram demonstrating shunt patency. Follow-up ranging from 1 to 5 years has shown no cases of postoperative gastrointestinal bleeding and no evidence of hepatic encephalopathy. I am concerned with the Sugiura technique because of the risk of inducing gastroesophageal reflux and I wonder whether you look for postoperative GE reflux--even in /those patients where you perform a Nissen fundoplication. In the patient that you describe with the postoperative stricture, did you feel that this was due to reflux or the devascularization procedure itself? Are any of your pediatric surgical colleagues in Toronto performing shunt procedures for similar patients, and how do their results compare with the results of your modified Sugiura series? Thomas Boles (Columbus, Ohio): I would like to comment on three aspects of this paper. One, a direct operation is obviously important for portal hypertension and I hope that the Sugiura operation or modification of it will be it. I am concerned, however, that the follow-up on the procedure described by Dr Superina is not long enough to make that statement. We did direct operations such as the Tanner procedure 20 years ago; and they worked very well for about 5 years, following which all of those patients began to bleed again. So, longterm follow-up is clearly important. Secondly, it has been 10 years since Voorhees reported on encephalopathy with extrahepatic portal hypertension. We looked up our cases last year, and in some 37 with follow-up periods of 5 years or longer there were no patients that had neuropsychiatric abnormalities. A shunt is still a very good operation. It would appear that in most people's experience, the risk of encephalopathy is relatively minor. Al deLorimier (San Francisco): I have performed six modified Sugiura procedures in which a splenectomy was not done and it included a Nissen fundoplication. Every one of them have rebled. Let me warn you that this is not a panacea. It is a procedure that will allow time for growth, so that the mesenteric vessels will
MODIFIED SUGIURA
enlarge enough to provide a successful shunt. However, it does not seem to provide longterm control of bleeding. Lower esophageal stricture is a problem in some of the smaller children where the esophageal mucosa is transected and reanastomosed. This can be a formidable problem to manage. It may be that transgastric ligation of varices might be preferred until a shunt is possible. Scott J. Boley (Bronx): I was not going to comment, but after Dr deLorimier has reported his poor results I think I have to respond. I have performed nine of these operations and none of the patient s has rebled. A major point that bothered me during Dr deLorimier's comments was when he said that he had modified the operation. We have not done a modified Sugiura, we have done the procedure as Sugiura described it. We have not used the stapler, and we have not done a fundoplication. We have performed postoperative 24 hour pH studies, and none of the patients have shown reflux. The patients do get recurrent varices, but do not bleed. We have reported them and John Ransom from N Y U has reported them. I think Dr Boles is absolutely correct, we have to watch these patients a lot longer than we have, but until now they have not rebled. I believe the main problem causing conflicting results arises when we start modifying other people's procedures. We then have no right to compare our results with theirs. If you perform the operation as Sugiura does it, I believe the operation has real potential. I do not believe you should be left with the pessimistic outlook that Dr delorimier has presented. Riccardo Superina (closing): I think there are three issues that really will cover most of those questions. Regarding the question of the incidence of encephalopathy, let me point out that Voorhees really went to a lot of trouble to demonstrate encephalopathy in these children and I don't know whether anybody has really looked at
799
their patients as carefully as he has. I know we don't always have the opportunity of seeing the patients for the purpose of testing them neurologically or psychiatrically and it may be that if you look for more subtle problems that you will pick them up. These may be just learning disabilities, who knows, but the potential is there and it is not there with the Sugiura procedure we should just stick to the Sugiura. Next let me comment on the modifications and the use of the stapler. The single incision is faster and simpler to close than the two incisions and I don't think that it makes any difference to the results. The EAA stapler really saves a lot of time and the two leaks, Dr Fonkalsrud, were in the patients that had had anastomosis before we started using the stapler. The experience in adults would suggest that the use of staplers has decreased the incidence of leaks both in esophageal and low rectal anastomoses. So, maybe the use of the staplers will make a difference in problems associated with leaks. We have only had three patients in whom we used the stapler and we think that strictures occur because an anastomosis is made in a relatively devascularized esophagus, and we don't think they are due to reflux. However, the precaution is taken of doing a fundoplication to help prevent reflux because we have really destroyed most of the mechanisms that normally prevent reflux. During the procedure the phrenoesophageal ligament is divided and the lower esophageal sphincter is manipulated and devascularized. So the Nissen procedure is a precaution against reflux, as well as providing a patch graft in case there is a leak. Hopefully this will take care of it. I don't want to discredit shunters but we believe the Sugiura procedure provides more reliable results and is more easily done in small children, and it is an alternate form of therapy in older children where shunts are not so readily done. Thank you very much for your comments.