Portosystemic shunting for paediatric portal hypertension

The records of 22 patients who received portosystemic shunting for portal hypertension from 1985 to 1995 inclusive at the Royal Alexandra Hospital for Children (RAHC) were retrospectively reviewed. There were 11 girls and 11 boys. The average age at operation was 8 years, 3 months (range, 2 years, 3 months to 16 years, 7 months). The aetiology was idiopathic portal cavernomatous transformation (n = 9), billiary atresia (n = 4), cystic fibrosis (n = 3), documented neonatal portal vein thrombosis (n = 3), congenital hepatic fibrosis (n = 2), and portal vein obstruction after liver transplant (n = 1). The major presenting problem was upper gastrointestinal haemorrhage. Two patients had recurrent melaena from Roux-en-Y jejunal loop and caecal varices, respectively. Before receiving shunts, 12 patients had endoscopic sclerotherapy, 1 had gastric transection, and 2 had gastric varices oversewn. Portal pressure at preoperative splenoportogram averaged 28 mm Hg (range, 20 to 41). Urgent shunts were performed on 13 patients. Two disadvantaged patients had prophylactic shunts for severe hypersplenism. The types of shunts used were reversed splenorenal (n = 13), splenoadrenal (n = 6), inferior mesenteric renal (n = 1), portocaval (n = 1), inferior mesenteric caval (n = 1), and superior and inferior mesenteric caval (n = 1). In all, 22 patients had 23 shunts. The patency rate was 96% on 6 months to 10 years follow-up (average, 5.8 years). No spleen was lost. There were 2 late deaths. Two cystic fibrosis patients and one child with extrahepatic portal hypertension experienced post-shunt encephalopathy. Three patients rebled in the early postoperative period despite a patent shunt. Two patients subsequently received liver transplantation without any additional difficulties. Thus, portosystemic shunting using a method appropriate for the patient is a reliable option for treating children with portal hypertension in whom variceal sclerotherapy is inappropriate or has failed.