From the Department of Gastrointestinal Surgery, Surgical Division and Gastroenterology Center, Army Hospital (Research and Referral), Rao Tularam Marg, Delhi 110 010, India.
Reprint requests: Lt. Col. Atul K. Sharma, Gastrointestinal and Laparoscopic Surgeon, Surgical Division and Gastroenterology Center, Army Hospital (Research and Referral), Delhi Cantt 110 010, India.
E-mail: [email protected]

Manuscript received: April 5, 1999;
Initial review completed: May 5, 1999;
Revision accepted: October 7, 1999

Extra hepatic portal venous obstruction (EHPVO) is the commonest cause of non-cirrhotic portal hypertension in children and is usually found in patients from the lower socio-economic strata(1). Management of these children who mainly present with a massive upper gastrointestinal (GI) bleed, a big spleen or hypersplenism, remains controversial, as sclerotherapy or band ligation is difficult and pediatric endoscopes not readily available. Also endoscopic therapy does not deal with hyper-splenism and in the long term may lead to ectopic varices, which tend to bleed torrentially and are usually not amenable to emergency sclerotherapy or tamponade.

Fear of post-splenectomy infections on the other hand deters most surgeons from per-forming a splenectomy and there is only one large series from India which presents the results of a side to side lieno renal shunt without splenectomy for noncirrhotic portal hyper-tension in children(2,3).

We present our experience of and the outcome of splenectomy and proximal Lieno-renal shunt in children with EHPVO.

All patients referred with non cirrhotic portal hypertension to the Department of Gastro-intestinal Surgery were evaluated by liver function tests, a hematological profile, an ultrasound/Doppler scan of the splenoportal axis and an upper GI endoscopy, to confirm the diagnosis of portal hypertension and look for its etiology. The majority underwent a splenectomy and a proximal end to side lieno-renal (Linton’s) shunt provided the splenic vein was available (>4 mm in diameter) and the liver function was not deranged. In either of the two instances above, i.e., if the splenic vein was found thrombosed, or the liver function significantly deranged (Child’s B or C), a splenectomy and gastro-esophageal devascularisation procedure was performed. A wedge and needle biopsy of the liver was done only if there was any gross evidence of nodularity at operation.

All patients were followed up calling them back for review every 3 months for the first 6 months, and yearly thereafter. The records of children (below the age of 14 years) who underwent surgery from July 1992 to June 1998 have been retrospectively analyzed to assess the outcome of our treatment protocol.

During this period a total of 98 patients underwent surgery for portal hypertension of which 32 were between the age of 4 and 12 years (mean 6.2; SD 3.1). Twenty-two were boys and 10 girls. The cause of portal hyper-tension was extrahepatic portal venous obstruction in all except one 11-year-old girl who had biliary cirrhosis due to Type IV (A) choledochal cyst. This child was, therefore, excluded from the study. The indication for surgery in the remaining 31 children was a GI bleed in 25 (80%); of whom 10 (40%) had failure of sclerotherapy (recurrence of bleeding despite two sessions of sclerotherapy). Three children (10%) underwent emergent surgery as sclerotherapy failed to stop the hemorrhage in the first instance. The mean number of bleeding episodes before the children were referred for surgery was 3.2 (range 1-6; SD 2.4) and the mean number of blood transfusions given to treat these bleeding episodes was 3.7 (range 1-5; SD 3.4). The remaining underwent surgery for hypersplenism (n = 4) and/or a painful large spleen (n = 3). All 31 children underwent a splenectomy (mean weight of spleen 1220 g; range 550 - 1580 g; SD 350.8). An end-to-side LR shunt was done in 27 with a mean shunt diameter of 7.4 (range 4.5-10.5; SD 3.4) mm. In 4 the splenic vein was thrombosed, thus gastroesophageal devascularization was performed. Liver biopsy was not performed in any of these 31 children since in all, the liver was absolutely normal in appearance and palpation, at the time of surgery.

There has been no operative mortality. Twenty-five children (80%) have been followed up for a median period of 3.7 (range 0.5-5.2) years with serial upper GI endoscopy and ultrasound/Doppler scans. On followup the variceal grade was always found to be two or less despite which 3 of the 24 shunts followed up (12.5%) were reported to be blocked on Doppler scanning. One child, after a devas-cularization procedure, had a minor rebleed following administration of oral antipyretics. No child ever developed encephalopathy.

Most Indian studies have highlighted EHPVO to be the major cause of portal hypertension and the commonest cause of major upper gastrointestinal bleeding in children(4,5). The etiology of EHPVO in the majority is unknown, but the postulated causes include congenital malformation of the portal vein or acquired thrombosis following umbilical sepsis, intraluminal trauma following exchange trans-fusion and pyelephlebitis following intestinal infection. All these are commonly seen in children from poorer sections of society. Most of these children present with recurrent episodes of massive upper GI bleed, the first usually occurring before the age of 10 years. Although almost all these patients tolerate the bleeding episode well, some children, especially below the age of 5, may develop transient ascites soon after a bleed. Most have a moderate spleno-megaly, although in some the spleen may reach the umbilicus. Some children present with repeated epistaxis and on evaluation are found to have hypersplenism. Liver function and histology is invariably normal and the diagnosis easily made by a ultrasonographic examination of the splenoportal axis, which in 90% of children reveals a block at the formation of the portal vein. In the remaining, either the entire splenoportal axis is thrombosed or the block lies in the hilum of the spleen, giving rise to left sided or "segmental" portal hypertension(1).

The treatment of an acute episode of hematemesis is fairly standardized and consists of resuscitation with blood and endoscopic sclerotherapy, which controls the bleeding in 95%. These episodes are well tolerated in the majority; however mortality rates upto 31% have been reported(6). It is in the prevention of recurrence of such bleeds that controversy exists and opinion is divided between endoscopic management on one side and surgery with or without splenectomy on the other.

We have managed these children by a uniform protocol of splenectomy and proximal lieno-renal shunt in the majority, performing a gastro-esophageal devascularisation procedure only if the splenic vein was not available. After a median follow up of 3.7 (±0.5-5.2) years, none of the patients have rebled or developed encephalopathy or overwhelming post splenectomy infection (OPSI). There has been no operative mortality and the one time procedure has at the same time dealt with the problem of hypersplenism. Our results compare well with the largest series published from India from the All India Institute of Medical Sciences, which is the premier center routinely performing an end to side lieno-renal shunt for EHPVO in children(7-9).

Compared to these results, a study from Institute of Post Graduate Medical Education and Research, Calcutta, has reported the outcome of sclerotherapy in 45 children. The number of sittings of sclerotherapy required for obliteration of varices was 5.9±1.6. Retrosternal discomfort (22%), dysphagia (22%), stricture (13%), esophageal ulceration (13%) and fever (11%) were the complications reported. Although variceal grade was reduced to 0-1 in 91%, rebleeding and recurrences rate of 13% were noted on follow-up(10). A similar study from SGPGI, Lucknow, has reported a much lower rebleed rate after 8 sessions of sclerotherapy per child, but the follow up is only for 19 months and the morbidity of the procedure has not been detailed upon(11).

Endoscopic band ligation is a relatively new procedure and may require fewer sittings and have a lower incidence of complications and the attendant morbidity(12); however, the procedure requires expertise and these patients need to be followed up much longer to record the incidence of rebleed from ectopic (fundal, duodenal and ileal varices).

Although there are various kinds of shunt procedures with splenic preservation described, the only major report from India is from PGIMER, Chandigarh, where routinely a side to side lieno-renal (SSLR) shunt is performed and they have not convincingly shown a reduction in the splenic size or dealt with the problem of hypersplenism in their reports(2,3).

Protagonists of the endoscopic procedures and SSLR shunt are keen on splenic pre-servation, but hypersplenism with a bulky painful spleen in a child has its own morbidity besides the chance of accidental rupture.

It is concluded that splenectomy and LR shunt is a safe and effective one-time treatment for EHPVO, which deals with both GI bleeds and hypersplenism.

Contributors: All authors have operated on the patients (at least two in each case). AKS is the guarantor of the manuscript. HKR and RPC have helped in formatting the proforma and in data collection. They have also critically reviewed the manuscript at each stage and their suggestions have been incorporated in the final draft.

Funding: None.
Competing interests: None stated.

1. Sahni P, Pande GK, Nundy S. Extrahepatic portal vein obstruction. Br J Surg 1990; 77: 1201-1202.

2. Mitra SK, Rao KL, Narasimhan KL, Dilawari JB, Batra YK, Chawla Y, et al. Side to side lieno renal shunt without splenectomy in noncirrhotic portal hypertension in children. J Pediatr Surg 1993; 28: 398-402.

3. Sharma BC, Singh RP, Chawla YK, Narasimhan KL, Rao KL, Mitra SK, et al. Effect of shunt surgery on spleen size, portal pressure and esophageal varices in patients with non-cirrhotic portal hypertension. J Gastroenterol Hepatol 1997; 12: 582-584.

4. Anand CS, Tandon BN, Nundy S. The causes, management and outcome of upper gastro-intestinal haemorrhage in an Indian Hospital. Br J Surg 1983; 70: 209-211.

5. Pande GK, Sahni P, Nundy S. Extrahepatic obstruction causing portal hypertension. J Gastro-enerol Hepatol 1988; 3: 99-107.

6. Howard ER, Stringer MD, Mowat AP. Assessment of injection sclerotherapy in the management of 152 children with esophageal varices. Br J Surg 1988; 75: 404-408.

7. Prasad AS, Gupta S, Kohli V, Pande GK, Sahni P, Nundy S. Proximal splenorenal shunts for extrahepatic portal venous obstruction in children. Ann Surg 1994; 219: 193-196.

8. Pande GK, Reddy VM, Kar P, Sahni P, Berry M, Tandon BN, Nundy S. Operations for portal hypertension due to extrahepatic obstruction: Results and 10 year follow up. Br Med J 1987; 295: 1115-1117.

9. Mohapatra MK, Acharya SK, Sahni P, Nundy S. Encephalopathy in patients with extrahepatic obstruction after lienorenal shunts. Br J Surg 1992; 79: 1103-1105.

10. Ganguly S, Dasgupta J, Das AS, Biswas K, Mazumder DN. Study of portal hypertension in children with special reference to sclerotherapy. Trop Gastroenterol 1997; 18: 119-121.

11. Yachha SK, Sharma BC, Kumar M, Khanduri A. Endoscopic sclerotherapy for esophageal varices in children with extrahepatic portal venous obstruction: A follow-up study. J Pediatr Gastro-enterol Nutr 197; 24: 49-52.

12. Nijhawan S, Patni T, Sharma U, Rai RR, Miglani N. Endoscopic variceal ligation in children. J Pediatr Surg 1995; 30: 1455-1456.