A duodenocaval fistula (DCF) is seen when a connection exists between the duodenum and the inferior vena cava. It is a rare entity that is commonly missed and presents a diagnostic challenge due to its nonspecific presenting symptoms.1,2 Patients commonly present with gastrointestinal (GI) bleeding or sepsis. Here we present a case of a 37-year-old man who presented to the hospital for a workup related to melena but went into cardiac arrest prior to an esophagogastroduodenoscopy. Unfortunately, on autopsy, the patient was found to have a DCF. We highlight the diagnostic challenge associated with DCF and how in this case the presentation was confused by a diagnosis of possible transfusion-related acute lung injury (TRALI). To the best of our knowledge, this is also the first description of a case of DCF associated with food embolism to the lungs causing respiratory failure.
Case Presentation
A 37-year-old man with a history significant for bulimia presented to the hospital with a 3-day history of melena and reports of dizziness. The patient did not report being on any prescribed medications but noted that he took 4 aspirin daily to “calm his nerves.” The rest of the patient’s history was unremarkable aside from a reported history of induced emesis 3 to 4 times per week for an extended period up until 2 weeks before admission.
On admission, his vital signs demonstrated tachycardia and orthostatic hypotension. Pertinent findings on physical examination were skin pallor, a normal lung examination, mild epigastric tenderness, and guaiac-positive stools. He was alert and oriented to person, place, and time with no focal deficits. His admission laboratory tests were notable for a hemoglobin (Hb) level of 4.6 g/dL (reference range, 14-17.9), a white blood cell count of 13.5 K/cm (reference range, 4.5-11), an international normalized ratio of 1.21, a blood urea nitrogen of 61 mg/dL (reference range, 10-20), and a creatinine of 2.3 mg/dL (reference range, 0.8-1.4). The patient was placed on 2 L of oxygen via nasal cannula for comfort rather than true hypoxia. A chest X-ray on admission was negative with no signs of infiltrate, edema, or widened mediastinum. An abdominal X-ray was significant for a dilated stomach consistent with bulimia with no abdominal free air or signs of obstruction. The case was discussed with the gastroenterology service who felt that the patient needed to be more hemodynamically stable before pursuing endoscopic evaluation.
He was admitted to the intensive care unit and give a transfusion of 4 units of fresh frozen plasma and 2 packed red blood cells (PRBCs) without any issues. During the infusion of a third PRBC, he developed chills, tachycardia, and hypertension with accompanying respiratory distress characterized by wheezing, decreased breath sounds bilaterally, and a decrease in oxygen saturation to 70% on 2 L supplemental oxygen. He responded to treatment with meperidine, methylprednisolone sodium succinate, albuterol nebulizer, and acetaminophen. A new chest X-ray was read as “development of pulmonary edema vs bilateral pneumonitis.” A transfusion reaction was reported to the blood bank and a diagnosis of TRALI was considered. That evening, he completed a dose of platelets and another PRBC without difficulty after he was premedicated with meperidine, methylprednisolone sodium succinate, and acetaminophen. During the night, the patient spiked a temperature of 40.3 °C that was successfully treated with a cooling blanket and acetaminophen.
The following morning the patient was found to be tachypneic and tachycardic with his face mask off. His symptoms were corrected by replacing his face mask. He claimed he felt anxious about getting more transfusions and that he had breathing problems like this at home in the recent past. The patient requested an aspirin to calm his nerves. Over the course of the day, his Hb level dropped from 6.6 g/dL to 5.9 g/dL, and 2 washed leukopoor PRBCs were ordered.
The first unit was infused uneventfully, but after 125 cc of the second unit, the patient developed respiratory distress, rigors, and hypotension to 70/58 mm Hg despite premedication. He again was treated successfully with increased face mask support. A few rales were noted, but his fluid balance was even. A second transfusion reaction was filed with the blood bank and based on the 2 transfusion-associated events with no other clear explanation for his symptoms, the clinical team favored the TRALI diagnosis. However, the blood bank was suspicious this might not be TRALI as the previous night the patient had 2 episodes of respiratory distress with drops in oxygen saturation unassociated with transfusions. The patient was clinically stable for the remainder of the night.
Early the following morning the patient was scheduled for an esophagogastroduodenoscopy to evaluate for a source of his bleeding. At the beginning of the procedure, a unit of washed leukoreduced PRBCs was hung for a Hb level of 6.9 g/dL. No bleeding source was noted in the stomach, but as the endoscope was passed into the duodenum, and after an infusion of only 25 cc of RBCs, the patient became cyanotic and went into cardiac arrest. Despite advanced resuscitation efforts over 90 minutes, the patient could not be successfully resuscitated and died while in the endoscopy suite. A transfusion reaction workup was initiated but was unremarkable. The transfusion medicine staff was suspicious that something other than TRALI was the cause of the patient’s respiratory distress as he had respiratory distress remote to the transfusions and the unit was prepared correctly before administration. The patient’s family agreed to an autopsy.