Gastrointestinal Bleeding Due to a Dieulafoy Lesion in the Afferent Limb of a Billroth II Reconstruction

Dieulafoy lesions, a rare cause of gastrointestinal hemorrhage, are caliber-persistent submucosal arteries without surrounding ulcers or mucosal lesions. Dieulafoy lesions rarely bleed profusely and may be missed during diagnostic endoscopy. Most of these lesions are located in the stomach; they have been reported only rarely in other parts of the gastrointestinal tract, including the small bowel. We report a case of gastrointestinal hemorrhage secondary to a Dieulafoy lesion in the afferent limb of a Billroth II gastric bypass.

Case Report

A 63-year-old African American woman with a history of hypertension, asthma, and osteoarthritis presented to the emergency room with a 3-day history of melena associated with upper abdominal discomfort. The patient denied experiencing nausea, vomiting, chest pain, dyspnea, fever, or chills. She did not have a history of smoking or substance or alcohol abuse. Her medications included diclofenac sodium/misoprostol (Arthrotec, Pfizer), lisinopril, and a fluticasone/salmeterol inhaler. Three years ago, she had presented to another institution with an episode of gastrointestinal bleeding, at which time she underwent a Billroth II gastrojejunostomy to control the bleeding.

On examination, the patient was not in acute distress. Her vital signs were stable, her oral mucosa was dry, and her conjunctiva appeared pale. Cardiovascular and pulmonary examinations were unremarkable. The patient had a large midline abdominal scar. Bowel sounds were present in all 4 quadrants, and her abdomen was soft. Mild epigastric tenderness was found without guarding or rigidity, but no organomegaly was detected. A rectal examination revealed a normal sphincter tone and black, loose stool that tested positive for occult blood. The remainder of the examination was unremarkable.

On admission, the patient's hemoglobin level measured 10.3 g/dL, with a hematocrit of 34% and a mean corpuscular volume of 103.2 fL. Other laboratory tests revealed severe nonanion gap acidosis, with a bicarbonate level of 6 mEq/L, a blood urea nitrogen level of 85 mg/dL, and a creatinine level of 1.9 mg/dL. A computed tomography scan of the abdomen showed evidence of prior bowel surgery; however, there was no evidence of bowel obstruction, free gas or fluid, or discrete mass lesions.

Several hours after admission, the patient became hypotensive and was treated with fluid resuscitation. She was admitted to the intensive care unit and was started on a continuous infusion of pantoprazole and normal saline. Eight hours after admission, a repeat blood test found a hemoglobin level of 6.9 g/dL. The patient underwent a transfusion of 2 units of packed red blood cells (PRBC). An urgent esophagogastroduodenoscopy (EGD) revealed a Billroth II anatomy, with blood clots in the gastric pouch and the afferent loop of the patient's prior gastrojejunostomy (Figure 1). Active bleeding and ulcers were not found. After the EGD, the patient experienced multiple episodes of melena and 1 episode of coffee-ground emesis. Her hemoglobin level dropped to 5.8 g/dL, and she received a transfusion of 4 more units of PRBC. A repeat EGD revealed blood clots and oozing from the afferent limb of the gastrojejunostomy, with no identifiable source of bleeding. We were unable to examine the entire length of the afferent limb. Although the patient was hemodynamically stable, she required additional transfusions due to her low hemoglobin level. At this point, a selective angiography was performed, but it did not reveal an active source of bleeding. The patient continued to have melena. By this time, she had received 11 units of PRBC, and she was taken to the operating room for intervention.

Figure 1.

An esophagogastroduodenoscopy revealed a Billroth II anatomy (A). Blood was noted in the afferent limb, but the source of the bleeding was not visible (B).

Prior to surgical intervention, an EGD with push enteroscopy was performed using a pediatric colonoscope. During this procedure, we were able to examine most of the afferent loop of the gastrojejunostomy, which showed persistent oozing of blood. The efferent loop appeared to be clean. After thorough irrigation, an oozing site that was suspicious for a Dieulafoy lesion was identified in the afferent loop of the jejunostomy (Figure 2). A total of 6 mL of epinephrine solution (1:10,000) was injected submucosally, and 3 endoclips were placed to achieve complete cessation of the bleeding. The area was tattooed with India ink to facilitate any future surgical interventions. Due to the successful endoscopic treatment, surgical intervention was not needed. The patient remained stable upon extubation. No additional episodes of bleeding were reported, and the patient's hemoglobin and hematocrit levels stabilized. Her diet was advanced, and she was discharged home in good condition.

Figure 2.

After thorough irrigation, an oozing site was identified in the afferent limb (A). Endoclips were placed after epinephrine injections (B).

Discussion

Although Dieulafoy lesions are rare, they are an important cause of gastrointestinal bleeding. These lesions were first described by Gallard in 1884.¹ When reporting on 3 patients with these lesions in 1898, the French surgeon Dieulafoy used the term “exulceratio simplex” because of the lesion's small size and large artery with normal histology.² A Dieulafoy lesion consists of a caliber-persistent arteriole that protrudes through a tiny mucosal defect, usually within 6 cm of the gastroesophageal junction on the lesser curve of the stomach.^3–5 Similar lesions have also been described in the distal esophagus, duodenal bulb, colon, rectum, and jejunum.^6–26 In approximately 4–9% of massive upper gastrointestinal bleeding episodes, a demonstrable cause cannot be found. A Dieulafoy lesion is thought to cause acute and chronic upper gastrointestinal bleeding in approximately 1–2% of these cases.²⁷ Dieulafoy lesions are more common in males, occur at a median age of 54 years, and are not related to alcohol or nonsteroidal anti-inflammatory drug use.²⁸ The etiology of Dieulafoy lesions is unknown, although they may be related to congenital or acquired vascular malformation. It has been suggested that mechanical pressure from an abnormal vessel may progressively erode the thin mucosa overlying a pulsating artery, causing it to bleed.^29,30 This artery, which protrudes through a solitary, tiny mucosal defect (2–5 mm) usually located in the upper part of the stomach, may spontaneously rupture and cause massive bleeding. These bleeding episodes can present as recurrent, massive hematemesis associated with melena (51%), hematemesis alone (28%), or melena alone (18%). On average, initial resuscitation requires a transfusion of up to 8 PRBC units.³¹

The majority of Dieulafoy lesions (75–95%) are located within 6 cm of the gastroesophageal junction. These lesions can easily be missed during endoscopy due to the intermittent nature of the bleeding. Approximately 49% of these lesions are identified during an initial endoscopic examination; however, 33% of lesions require more than 1 endoscopy for identification.^31,32 If an EGD is performed within 2 hours of admission, there is a greater likelihood of identifying the lesion. The remaining cases are identified via angiography during active bleeding or via laparotomy. An endoscopic ultrasound (EUS) may also confirm the diagnosis by revealing a tortuous submucosal vessel near the mucosal defect. EUS can also help in the surveillance and assessment of endoscopic treatment.³³

Therapeutic endoscopy is the initial treatment of choice. Different endoscopic modalities have been used, including bipolar electrocoagulation, injection sclerotherapy, heater probe, laser photocoagulation, epinephrine injection, hemoclipping, and banding.²⁹ The first endoscopic intervention is successful in 85% of patients, while 10% of patients need a repeat endoscopy, and 5% of patients may require surgical intervention.¹⁴ Angiography may have a therapeutic role, although this technique is usually reserved for patients who are not amenable to endoscopic therapy and who are poor surgical candidates.³⁴ For duodenal and proximal jejunal lesions, surgical exploration can be performed via intraoperative endoscopy to avoid unnecessary bowel resection. This exploration enables endoscopic visualization with open surgical suture ligation.³⁵ Following endoscopic management, 1 series found no recurrence of bleeding from Dieulafoy lesions over a mean follow-up period of 28 months; another series found no recurrence of bleeding over a mean follow-up period of 36 months.^7,31

Conclusion

To the best of our knowledge, this case study is the first report of a Dieulafoy lesion in the afferent limb of a gastrojejunostomy. This case study demonstrates the diagnostic problems inherent to these lesions (due to the intermittent nature of the bleeding); these problems include failure to detect a lesion during the first 2 endoscopies and a negative angiogram. This case study also illustrates successful endoscopic management that obviated the need for surgery.

Although Dieulafoy lesions are rare, they are an important and potentially fatal cause of gastrointestinal bleeding. Advanced endoscopic techniques, including capsule endoscopy, single- or double-balloon enteroscopy, and spiral enteroscopy, will increase the number of diagnosed cases and lead to better management strategies by providing access to longer segments of the small bowel. Careful examination of the mucosa during endoscopy remains the most important factor for detecting these lesions, providing appropriate therapeutic modalities, and decreasing morbidity and mortality.