A 29-year-old female with no prior medical history was transferred to our facility for evaluation of post-prandial nausea and right upper quadrant abdominal pain. These symptoms had been present for the past three months. She reported emesis of undigested food within five minutes of oral intake, which resulted in the patient being on a liquid-only diet. She denied dysphagia or odynophagia. Since symptoms started, she had a 30 lb. unintentional weight loss.
Work up at our facility included a CT scan of the abdomen and pelvis, which was remarkable for filling defects within the duodenum, as marked here by arrows. This was thought to be due to a mixture of oral contrast and food. Her small bowel was normal in caliber, without evidence of complete obstruction. This was followed by a small bowel follow thru with barium. The filling defect was again seen at the in the second and third portions of the duodenum. In this study, it appears as a windsock-shaped intraluminal duodenal diverticulum.
An upper endoscopy was performed, using both a forward-viewing endoscope and a side-viewing duodenoscope. The side-viewing duodenoscope is seen entering thru the stomach and into the duodenum. In the second portion of the duodenum, a large membrane is seen prolapsing proximally. With insufflation, this is freely mobile. There appears to be two openings, with the larger superior one likely being the main opening into the third portion of the duodenum. A smaller, secondary opening is seen inferiorly. This membrane, or windsock diverticulum, prevents advancement of the endoscope.
The duodenum was carefully inspected and the major papilla identified being proximal to the diverticulum. Two clips were placed on the superior margin of the diverticular membrane in order to delineate the plane of the planned electrosurgical incision. A needle-knife was then used to make an incision thru the membrane, the first being from the larger primary opening towards the smaller secondary opening. Another incision was made in a similar fashion in the opposite direction, from the secondary to the primary opening. In order to complete the duodenotomy, a sphincterotome was utilized. By tensing the cutting wire of the sphincterotome, this pulled the membrane away from the true duodenal mucosa, allowing for a safe incision thru the mucosa and no burn of the duodenum. Once the duodenotomy has successfully been completed, there is no further obstruction of passage of the endoscope into the third portion of the duodenum. The anatomy has now been restored to normal.
The margin of the incision did have a small amount of oozing, which was irrigated and then successfully treated with a hemostatic clip. In order to prevent further post-procedure bleeding, clips were placed in a similar fashion along the remnants of the diverticulum. Once this was completed, the forward-viewing endoscope was used to demonstrate the fully open duodenal lumen. After the procedure, a repeat small bowel follow thru was obtained. The endoscopic clips are seen at the site of the treated diverticulum. There is no further evidence of obstruction.
In the following 24 to 48 hours after the procedure, the patient had several episodes of melena and complained of lightheadedness. Lab work was significant for a decrease in hemoglobin from 9 g/dL to 7 g/dL. A repeat EGD was performed and no active sites of bleeding were identified. There was ulceration at the margin of the treated diverticulum, and two additional hemostatic clips were placed prophylactically. The patient remained hemodynamically stable and was discharged from the hospital with no further gastrointestinal symptoms.
This case highlights the etiology and diagnosis of a windsock diverticulum and demonstrates endoscopic treatment using a needle-knife and sphincterotome thru a side-viewing duodenoscope. In addition, it shows successful treatment of post-procedure bleeding.
Conflicts of interest: None
Sharlene L. D’Souza, MD, University of Michigan
Dejan Micic, MD, University of Michigan
B. Joseph Elmunzer, MD, University of Michigan