Case Reference No. CC-497-07

A 41-year-old woman with a history of antiphospholipid syndrome and rheumatoid arthritis presented to the interventional radiology for evaluation and possible treatment of an upper GI. Early that day, the patient had brisk bleeding, but could not be brought to the department because of difficulty in maintaining adequate oxygenation. At the time of arrival, she was no longer bleeding. Duodenitis was diagnosed by endoscopy; a large blood clot was present in the duodenal bulb.

Her medications include ofloxacin, gentamicin, metronidazole, solumedrol, ranitidine, albuterol, fentanyl, heparin, and antithrombin III. Anticoagulation therapy was stopped when the GI bleeding was first noted (approximately 36 hours prior to her arrival to the IR suite). Pertinent lab values included a hemoglobin of 10.4 g despite four units of fresh frozen plasma and packed red blood cells. Her INR was 1.9.

Visceral angiography demonstrated a prominent median arcuate ligament at the celiac axis resulting in retrograde flow through the gastroduodenal artery (GDA) via superior mesenteric artery (SMA) collaterals. There was no active bleeding or extravasation.

We were unable to catheterize the GDA with an 0.035-inch lumen catheter, although it was possible with a coaxial tracker-type catheter. We elected not to empirically embolize the GDA because of the brisk "in our face" retrograde flow. Superselective catheterization through the SMA was deemed too difficult because of vessel tortuosity. In retrospect this proved to be a wise decision because the patient had no further bleeding over the next 2 weeks.

 

Fig. 1. Selective SMA injection demonstrates brisk retrograde filling of the celiac artery branches through the pancreaticoduodenal arteries.

Fig. 2. Celiac angiogram demonstrates an inflow defect at the GDA (arrow).

Fig. 3. Brisk injection in the common hepatic artery barely refluxes into the GDA.

Questions

1) Given the retrograde flow through the GDA, should we have attempted to embolize this artery and, if so, how?