A middle-aged woman with recurrent haemetemesis

L C T Hung, Y T Lee, S C H Yu, W K Leung, J J Y Sung

CASE REPORT

A 48-year-old woman presented to the emergency department in September, 2001, complaining of haemetemesis of about 300ml fresh blood. On admission, her blood pressure was 99/50 mm Hg and her pulse rate was 102/min. We treated her for hypovolaemic shock. She had been in good health except for having been stabbed in the abdomen during a robbery, 20 years previously. This stab wound had necessitated a right nephrectomy and partial right hepatectomy, but she had had no complications of either in the intervening decades. On examination, we found no signs of chronic liver disease. Blood tests showed haemoglobin 84 g/L, with normal renal and liver function tests. Urgent oesophagogastroduodenoscopy (OGD) showed a small mucosal break at oesophagogastric junction and small amount of fresh blood in the stomach. We diagnosed Mallory-Weiss Syndrome, and discharged her after adequate fluid replacement and blood transfusion. 4 weeks later, she returned with another episode of haemetemesis. Her blood pressure and pulse were stable, but her haemoglobin level was 80 g/L. We repeated the OGD and found 4 columns of small oesophageal varices which we ligated with endoscopic bands.

We tested for serum markers of viral, autoimmune, and metabolic causes of liver cirrhosis; all were negative. The patient then developed ankle oedema and abdominal distension. The renal function and liver function tests remained normal but serum albumin level dropped to 25g/L. Ultrasound examination showed normal liver echotexture, mild splenomegaly and massive ascites. The right portal vein was found to be markedly dilated, forming an aneurysm. Doppler ultrasound showed that the hepatic artery had low resistance but high diastolic flow, and it was joined to the right portal vein. We diagnosed a hepatic arterioportal fistula, which was probably caused by the previous stab wound to the liver,  although the partial hepatectomy may have also contributed to its formation. We delineated the anatomy of the fistula with hepatic angiography and embolised it with a metallic coil (figure, A) Follow-up hepatic angiogram showed a completely occluded right hepatic artery (figure, B). The patient’s ascites and peripheral oedema resolved completely in 4 weeks time and when last seen, in July, 2002, she had no further complaints.

Hepatic arterioportal fistula (HAPF) was originally described in 1892,1 and has been frequently reported since.

Abstract: http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(02)09900-2/fulltext