We report a case of aortoesophageal fistula (AEF) caused by a fish bone that had a successful outcome. Aortoesophageal fistula is a rare complication of foreign body ingestion from which few patients survive. Over one hundred cases of AEF secondary to foreign body ingestion have been documented but only seven, including our case, have survived over 12 months. Treatment involved stabilising the patient with a Sengstaken-Blakemore tube and insertion of a thoracic aortic endovascular stent-graft. Unfortunately the stent became infected and definitive open surgical repair involved removing the stent, replacing the aorta with a homograft and coverage with a left trapezius flap while under deep hypothermic arrest. Case presentationA 59 year old man with no prior medical history, presented to a peripheral hospital emergency department with sharp pain retrosternally after eating fish the previous day. He was able to swallow fluids and soft diet but with odynophagia. A cardiac cause was ruled out and a barium swallow was organised as an outpatient. The patient was discharged home. Day five after presentation he had frank haematemesis and some malaena. He was haemodynamically stable with a haemoglobin at 130 g/L and admitted to the general surgical ward. The next day he had another large haematemesis and proceeded to urgent upper gastrointestinal (GI) endoscopy where a fish bone was seen protruding from an ulcerated area in the oesophagus, 24 cm from the teeth. There was some active bleeding after the fish bone was removed. A SengstakenBlakemore tube (SBT) was inserted and the oesophageal and gastric balloons inflated. This controlled the bleeding and the patient was admitted to the intensive care unit (ICU) with intravenous (IV) antibiotics as well as an ongoing blood transfusion. Overnight there was no further bleeding but the patient had evidence of sepsis with a high fever and hypotension despite no evidence of blood loss. He required inotropic support.The following morning, the patient was transferred to the ICU at our centre. The SBT balloons were deflated and patient observed. There was no evidence of blood loss over the ensuing 24 hours so the SBT was removed. Five hours later, he had a massive haematemesis causing hypovolaemic shock with cardiac arrest requiring reinsertion and reinflation of the SBT, massive transfusion and cardiopulmonary resuscitation. Once stabilized, a computed tomography (CT) angiogram was performed. This dem-