Bleeding duodenal varices

Abstract

Dear Sir, Brown and Cooper (BJCP 1993; 47: 111), in their case report ‘Isolated duodenal varices as a cause of massive gastrointestinal bleeding‘, rightly draw attention to the difficulties presented by acutely bleeding varices, and show that relatively limited surgery may be effective in stopping bleeding. However, this case raises several points regarding management of ectopic varices—particularly in light of our recent experience with two patients. A 47-year-old man presented with profuse gastrointestinal haemorrhage, and at endoscopy bleeding varices were found. There were also non-bleeding oesophageal, gastric and rectal varices. Endoscopic injection sclerotherapy with ethanolamine oleate was successfully initiated. A rebleed at 24 hours was not stopped by further sclerotherapy but full control was gained by use of the long-acting somatostatin analogue octreotide intravenously (50 μg/hour) and continued for 72 hours. Emergency angiography did not locate a bleeding point but illustrated extensive small-bowel varices and occlusion of the splenic, superior mesenteric or portal veins. Cavernous transformation of the portal vein was confirmed by CT scanning. No underlying cause for the patient's venous occlusion was identified. Two further sclerotherapy sessions were performed to obliterate the bleeding duodenal varices; the other varices were not treated. He remains well on propranolol, which has been used in an attempt to reduce his portal hypertension. A 36-year-old man was investigated for recurrent rectal bleeding associated with abdominal pain which began 1 year after a seemingly uncomplicated episode of acute alcoholic pancreatitis. A right hemicolectomy had been performed for colonic varices but bleeding had continued. Mesenteric angiography performed during a bleed showed portal and splenic venous occlusion, transverse colonic varices and a large gastroduodenal artery pseudoaneurysm, which was actively bleeding into the pancreatic duct. The pseudoaneurysm was successfully embolised with metal coils, and he has had no further bleeding. He is also being treated with propranolol. Neither patient had cirrhosis but both had significant varices secondary to extra-hepatic venous occlusion. Over one-third of duodenal varices are related to pre-hepatic venous occlusion rather than cirrhosis,1 and more than 50% of patients with proven splenic vein thrombosis following pancreatitis have gastroduodenal varices.2 The usefulness of angiography in the assessment of patients with ectopic varices or unusual sites of bleeding following pancreatitis is evident. It enables detection of venous thrombosis and pseudoaneurysms which are likely to require different definitive management from cirrhotic varices. The authors used simple surgical methods to stop the bleeding, but the rate of rebleeding is high and the risks of laparotomy in cirrhotics are significant.3 In cases in which a diagnosis of duodenal varices can be established, endoscopic sclerotherapy4 and tissue adhesive injections5 have been used to control bleeding. Octreotide is effective in oesophageal variceal haemorrhage,6 and our experience shows that it may have a wider role in bleeding duodenal varices. There are no previous reports of successful use of pharmacotherapy in stopping duodenal variceal haemorrhage. Whatever emergency measures are employed, assessment of the extrahepatic venous system should be completed in patients with unusually sited varices following pancreatitis to enable detection of splenic vein occlusion. Varices in such cases can be cured by splenectomy.