Video Capsule Endoscopy in a Pediatric Patient With Hematemesis While on Ventricular Assist Device Support

**Background:** Ventricular assist devices (VADs) are used to bridge pediatric patients with heart failure to transplantation or as destination therapy. Gastrointestinal (GI) bleeding is a significant complication, occurring in approximately 13% of pediatric patients on VAD support. Standard investigations include upper and lower endoscopy, but these carry increased risks in pediatric VAD patients, including bleeding due to anticoagulation, perforation, and the need for general anesthetic. Video capsule endoscopy (VCE) is a noninvasive tool that involves swallowing a pill camera that transmits multiple images per second from within the stomach and small intestine. VCE has minimal complications, with capsule retention in 1.4% of cases and no reported deaths. In adult patients with VADs and other cardiac devices such as pacemakers, VCE has not been associated with device-related interference. The role of VCE in pediatric VAD patients is not defined due to a lack of data.

**Methods:** This is a single case report of a 16-year-old female with Shone's complex, pulmonary hypertension, and complete heart block requiring a pacemaker who had a HeartWare Ventricular Assist Device (HVAD). She presented to the emergency department with hematemesis. No hematochezia or epistaxis was reported. Physical exam was unremarkable, and bloodwork revealed a hemoglobin drop to 76 g/L from 93 g/L. Past medical history included a 5-week episode of overt GI bleeding 2 years prior, with esophagogastroduodenoscopy and an upper airway scope unable to identify a bleeding source. Upon admission, the patient was started on intravenous pantoprazole and received a blood transfusion. Endoscopy was felt to be high risk due to pulmonary hypertension, so VCE was chosen after consideration regarding potential interference between the video capsule and the VAD and pacemaker. A dissolvable patency capsule was administered first to assess for safe transit, which passed without retention. On admission day 5, a PillCam 3 capsule was swallowed by the patient.

**Key Results:** The capsule remained in the stomach for the entire recording and capsule passage occurred within 24 hours with no device-related complications. The VCE images suggested the gastric antrum was the source of the bleeding, showing active bleeding as the capsule approached the antrum, with erosion and ulcerated mucosa, and early clot formation with some fresh blood. The patient remained stable, and conservative management with intravenous pantoprazole was continued. At discharge, the patient was stable with a hemoglobin of 95 g/L and was prescribed 8 weeks of pantoprazole and 6 weeks of sucralfate.

**Clinical Implications:** This case demonstrates that VCE can be safely applied in a pediatric patient with a VAD, with a diagnostic yield that guided clinical management. The source of bleeding could have been identified with gastroscopy; however, VCE yielded the same diagnostic outcome while eliminating the need for anesthetic and potential bleeding and perforation associated with gastroscopy. In adult VAD patients, VCE is safe and effective, with diagnostic yields up to 80%. A meta-analysis comparing VCE to push enteroscopy found VCE has a diagnostic yield of 63% compared with 26% for push enteroscopy. The case also highlights limitations of VCE, including that the capsule may remain in the stomach for the recording period (likely related to reduced gastric motility in this patient), inability to perform therapeutic interventions, and the fact that VCE was performed 5 days after presentation due to lack of familiarity and time for decision-making. Optimal timing of VCE within 24 hours of presentation yields greater diagnostic accuracy and translates to a higher therapeutic intervention rate. Further studies on safety, efficacy, and complications in pediatric VAD patients are necessary to establish its role in the care of this patient population.