Twilight Trilogy: A Series of 3 Blood‐Letting Gi Bleeds

1Duke University Medical Center, Durham, NC
2Duke University Medical Center, Durham, NC

Meeting: Hospital Medicine 2013, May 16-19, National Harbor, Md.

Abstract number: 444

Case Presentation:

During the past 8 months, we have had 3 patients with factitious GI bleeds. This is one case. Mrs. K. was a 36‐year‐old woman with a history of gastric bypass with a complicated postsurgical course requiring total parenteral nutrition for 3 years. She was hospitalized 26 times for GI bleeds and per her report had undergone 22 abdominal surgeries at various hospitals. Her GI bleeds were believed to be a result of ulceration at the gastrojejunostomy anastomosis. She underwent a truncal vagotomy on April 14, 2012, for refractory GI bleeds. Her hospital course after that was complicated by PICC‐associated MSSA bacteremia and septic pulmonary emboli. She remained on Lovenox because of persistent upper‐extremity thrombi when she presented to our hospital for another episode of hematemesis. Her Lovenox was held, and she underwent an EGD that showed a 5‐mm nonbleeding superficial gastric ulcer at her anastomosis. Because of recurrent episodes of hematemesis, she ended up going for 2 more EGDs, which continued to show the small ulceration without active bleeding during her 13‐day hospitalization. A heparin challenge was performed, and initially no bleeding occurred for 3 days. On hospital day 9 she reported 3 more episodes of hematemesis, so the heparin was stopped. The medical team became suspicious of factitious disorder because of the appearance of the blood in her basin. The volume of blood was always the same and never contained any gastric contents. The blood was sent for pH testing on day 8 and showed a pH of 7.4. Nursing also noted that the Hickman was found to be unclamped. On hospital day 10 she was restarted on heparin, and the patient produced another basin of blood. A PTT was checked on this blood that was >300 seconds, whereas a peripheral blood PTT was 59.2 seconds. The patient was confronted, a sitter was placed in the room, and a room search performed. Two syringes with blood were found wrapped in hospital washcloths in her personal bag. Her Hickman was subsequently removed and the patient discharged the next day.


All 3 patients identified with factitious GI bleeding at our institution had long‐term venous access and frequent contact with hospitals. Our institution demonstrated an ability to learn from its experiences this year. We developed a process that has improved our approach to patients with potential factitious disorder. It starts with a multidisciplinary meeting to confront the patient followed by continuous observation of the patient with a sitter and a very thorough room search and then psychiatric consultation. This case also involves a novel approach to testing the vomited blood for the fluid that was being infused. As a result, length of stay from the first case to the last case decreased from 30 to 6 days.


Factitious GI bleeding should be high on the differential for patients with chronic venous access who have had frequent contact with the medical system and with GI bleeding that is refractory to conventional diagnostics and therapy.

To cite this abstract:

Setji N, O'brien C. Twilight Trilogy: A Series of 3 Blood‐Letting Gi Bleeds. Abstract published at Hospital Medicine 2013, May 16-19, National Harbor, Md. Abstract 444. Journal of Hospital Medicine. 2013; 8 (suppl 2). Accessed April 2, 2020.

« Back to Hospital Medicine 2013, May 16-19, National Harbor, Md.