490 CLINICAL REPORTS Postgrad Med J: first published as 10.1136/pgmj.69.812.490 on 1 June 1993. Downloaded from Postgrad Med J (1993) 69, 490-493 © The Fellowship of Postgraduate Medicine, 1993 Intraperitoneal haemorrhage from anterior abdominal wall varices J.B. Hunt, M. Appleyard, M. Thursz, P.D. Carey', P.J. Guillou' and H.C. Thomas Departments ofMedicine and 1Surgery, St Mary's Hospital Medical School, Imperial College, London W2 INY, UK Summary: Patients with oesophageal varices frequently present with gastrointestinal haemorrhage but bleeding from varices at other sites is rare. We present a patient with hepatitis C-induced cirrhosis and partial portal vein occlusion who developed spontaneous haemorrhage from anterior abdominal wall varices into the rectus abdominus muscle and peritoneal cavity. Introduction Portal hypertension is most often seen in patients abdominal pain of sudden onset. Over the pre- with chronic liver disease but may also occur in ceding 3 months he had noticed abdominal and those with portal vein occlusion. Thrombosis ofthe ankle Four years earlier chronic active swelling. by copyright. portal vein is recognized in both cirrhotic patients,' hepatitis had been diagnosed in Egypt and treated those with previous abdominal surgery, sepsis, with prednisolone and azathioprine. neoplasia, myeloproliferative disorders,2 protein Examination revealed a well nourished, jaun- C3 or protein S deficiency.4 diced man with stigmata of chronic liver disease Oesophageal varices develop when the portal who was anaemic and shocked with a pulse of pressure is maintained above 12 mmHg.5 Patients 100mm and blood pressure 60/20 mmHg. The with oesophageal varices often present with severe abdomen was distended, diffusely tender and there haematemesis. The mortality associated with the was splenomegaly. Rectal examination was nor- first haemorrhage is approximately 50% but varies mal. with the severity of any underlying liver disease.6 Investigation revealed a haemoglobin of 4.5 g/ http://pmj.bmj.com/ For this reason the prognosis is better in patients dl, platelets 129 x 109/1, prothrombin time 39 sec- with prehepatic portal hypertension. onds (control <16.5 seconds). Serological tests Extraoesophageal varices are recognized in intra- were positive for anti-hepatitis B core antigen and hepatic portal hypertension but they are rare and hepatitis C(p22). Autoantibody tests were negative seldom bleed.7 In contrast in prehepatic portal as were the schistosomal serology and hepatitis B hypertension they are both more common and surface antigen. bleed more This in be due to Abdominal demonstrated as- frequently.8'9 may part ultrasonography on September 28, 2021 by guest. Protected extraoesophageal varices being of a greater size, as cites, a small irregular liver, splenomegaly and rupture is more likely when the wall tension is massively dilated umbilical and anterior abdom- high.'° We describe a patient with both cirrhosis inal wall veins. An ultrasound-guided diagnostic and partial portal vein occlusion in whom bleeding paracentesis was performed at a site distant from occurred from large anterior abdominal wall the vessels. This revealed uniformly blood-stained varices. fluid. Contrast enhanced abdominal computed tomographic (CT) scan confirmed the ultrasound Case report findings, demonstrated oesophageal varices and showed the left rectus abdominus muscle to be A 48 year old Egyptian male professional presented distended (Figure 1). Calcification was noted in the with collapse after experiencing severe central region of the portal vein. At angiography the superior mesenteric vein was narrowed near the Correspondence: J.B. Hunt, M.D., Department of head of the pancreas where calcification was again Medicine, Queen Elizabeth The Queen Mother Wing, St seen. The portal vein filled poorly and a large varix Mary's Hospital Medical School, London W2 1PG, UK. drained into the umbilical vein which was grossly Accepted: 30 November 1992 dilated. The appearances suggested haemorrhage CLINICAL REPORTS 491 Postgrad Med J: first published as 10.1136/pgmj.69.812.490 on 1 June 1993. Downloaded from Discussion The patient's presentation was due to spontaneous rupture of abdominal wall varices. The success of the operation, confirmed at postmortem, will have increased the pressure and wall tension inside the oesophageal varices. The association between variceal wall tension and risk of haemorrhage is recognizedl° and thus in ret- rospect ligation of the anterior abdominal vessels 'I- l may have precipitated oesophageal variceal haem- orrhage, which led to the patient's death. An alternative treatment would have been to decompress the portal circulation either surgically or by a transjugular intrahepatic portosystemic stent-shunt (TIPSS)."'"2 Surgical shunts are either selective, nonselective or partial.'3 A selective shunt Figure 1 Abdominal CT scan showing thickening ofleft such as the distal splenorenal anastomosis main- rectus abdominus muscle. tains the pressure in the mesenteric and portal system whilst decompressing the oesophagogastric bed.'4 This is successful therapy for oesophageal variceal bleeding but would not have been appro- priate in this case. Nonselective shunts direct the portal circulation into the inferior vena cava via end-to-side, side-to-side or prosthetic interposition into the left rectus abdominus muscle with subse- anastomoses. Although effective in preventing quent bleeding into the peritoneal cavity as a bleeding, because the anastomoses are wide and by copyright. consequence of portal hypertension in a patient allow large volume flow, they are associated with a with cirrhosis and partial portal vein occlusion. high incidence of chronic hepatic encephalo- Ten days after admission the patient suffered a pathy.l5'16 Our patient was a professional man who further episode of shock with abdominal disten- would have been incapacitated by this complica- sion. A repeat CT scan showed an increase in size of tion. He had developed severe encephalopathy the left rectus abdominus muscle. After resuscita- after both bleeding episodes and we believed the tion a laparotomy was performed. The peritoneal likelihood of encephalopathy was too great to cavity contained 5-6 litres of altered blood, the perform a nonselective shunt. spleen was massively enlarged and the liver Partial shunts are designed to avoid chronic http://pmj.bmj.com/ shrunken; there was no evidence ofsplenic trauma. encephalopathy. They have smaller anastomotic Numerous varices were seen in both the intra- windows, do not fully reduce the porto-systemic peritoneal and extraperitoneal spaces associated gradient'718 and may in part preserve portal flow.'9 with the ligamentum teres. The spontaneous The anastomosis may be made directly between the haemorrhage appeared to have occurred from one portal vein and the inferior vena cava or via a of these but the exact site could not be localized. prosthetic insert of fixed diameter.'1920 The exper- The abdominal paracentesis had been performed at ience with shunts is limited. Neither the partial on September 28, 2021 by guest. Protected a point 7-8 cm caudal to any varices. The anterior adequacy ofpressure reduction nor the duration of abdominal wall varices were under run and a wedge patency are yet well established. Thrombotic occ- liver biopsy performed. This showed cirrhosis with lusion of nonselective shunts is a well-recognized areas of inflammation and piecemeal necrosis, problem'5 and the same may be true of partial compatible with the consequences of hepatitis C shunts. Because of these uncertainties we rejected infection. this surgical option. The patient made a good initial postoperative A TIPSS is introduced by radiologically guiding recovery but on day 4 developed hypotension and a wire and subsequently a stent between a branch of hepatic failure. Attempts at resuscitation were the hepatic vein, through the parenchyma of the unsuccessful. Postmortem examination revealed liver to a radicle of the portal vein. Presumably ruptured oesophageal varices with blood in the because the decompression is not complete, ence- stomach and small bowel. There was no evidence of phalopathy has not been a major complication in recent intra-abdominal haemorrhage and the an- the limited number of reports." 12 The technique terior abdominal wall varices remained collapsed. does not need surgical intervention and repeat The portal vein was partially occluded with balloon dilatation of stents can be performed ifthe calcified thrombus. portal pressure is inadequately reduced. TIPSS 492 CLINICAL REPORTS Postgrad Med J: first published as 10.1136/pgmj.69.812.490 on 1 June 1993. Downloaded from may become the treatment of choice for portal of the corpus luteum may cause intra-abdominal hypertension when the portal vein is patent. Portal haemorrhage37. This is more easily recognized ifthe vein occlusion does not preclude the technique but patient is receiving peritoneal dialysis.38 does not increase the difficulty of deployment. As Non-gynaecological intraperitoneal bleedingmay the imaging in our patient suggested portal vein occur from primary39-42 or secondary hepatic calcification and partial portal vein occlusion we tumours43-46 or from other intra-abdominal tum- were deterred from attempting the procedure. ours.47-49 Pancreatic pseudocysts may bleed into The development of varices at points of portal- the peritoneal cavity50'51 and haemoperitoneum systemic anastomosis is well recognized. Whilst may also be seen as a complication