Acute upper gastrointestinal bleeding(AUGIB) is a common emergency. The NaviCam (ANKON) is a miniaturised wireless endoscope in a single use capsule. It can be remotely controlled with the patient in a magnetic console.. In patients with AUGIB, NaviCam has been compared to conventional esophago-gastro-duodenoscopy (EGD) in their diagnostic yields. NaviCam has been shown to detect more lesions including those in the small bowel. There are therefore several theoretical advantages to the use of NaviCam in the management of patients with AUGIB.An initial NaviCam examination allows triaging of patients. Those with low risk lesions can be discharged without EGD and hospital admission. These represent substantial reduction in resource utilisation. In the diagnosis of small bowel lesions, the yield from a video capsule examination is higher closer to the time of index bleed.

The primary objective of the study is to determine the diagnostic yield of NaviCam in patients who present with overt signs of AUGIB. In addition, the investigators aim to determine if NaviCam examinations can reduce hospital resource utilisation and compare the use of NaviCam as a triage tool to the use of risk scores such as the Glasgow Blatchford Score (GBS). The investigators hypothesize that early NaviCam examination can allow safe discharge of more patients when compared to GBS.

INTRODUCTION Background/ rationale Acute upper gastrointestinal bleeding (AUGIB) is a common emergency. The incidence varies from 50 to 110 per 100,000 population. In the majority of patients with AUGIB, bleeding is self-limiting and often has stopped spontaneously on presentation. Traditionally patients undergo early endoscopy (within 24 hours) for the purpose of finding the source of bleeding and if actively bleeding is seen from a lesion, endoscopy hemostatic treatment is then applied.

The NaviCam (ANKON) is a miniaturised wireless endoscope in a single use capsule. It can be remotely controlled with the patient in a magnetic console. The operator to the magnetic-driven NaviCam sits in a computer workstation and manipulates the capsule suspended in a waterfilled stomach. After careful examination of the stomach, the capsule is then allowed to passed through the pylorus into duodenum and small bowel. The patient wears harness with small bowel imaging capabilities. In patients with AUGIB, NaviCam has been compared to conventional esophago-gastro-duodenoscopy (EGD) in their diagnostic yields. NaviCam has been shown to detect more lesions including those in the small bowel. There are therefore several theoretical advantages to the use of NaviCam in the management of patients with AUGIB. First, the NaviCam procedure is well tolerated and in fact preferred by many patients. It is convenient to administer. Second, NaviCam can in addition examine small bowel. About 20% of patients with overt signs of AUGIB bleed from small bowel or colon. An initial NaviCam examination allows triaging of patients. Those with low risk lesions can be discharged without EGD and hospital admission. These represent substantial reduction in resource utilisation. In the diagnosis of small bowel lesions, the yield from a video capsule examination is higher closer to the time of index bleed.

Objectives The primary objective of the study is to determine the diagnostic yield of NaviCam in patients who present with overt signs of AUGIB. In addition, the investigators aim to determine if NaviCam examinations can reduce hospital resource utilisation and compare the use of NaviCam as a triage tool to the use of risk scores such as the Glasgow Blatchford Score (GBS). The investigators hypothesize that early NaviCam examination can allow safe discharge of more patients when compared to GBS.

METHODS Study design The investigators plan to conduct a prospective cohort study at the Prince of Wales Hospital, Shatin. The investigators enrol patients who present to the Accident and Emergency Department with overt signs of AUGIB (hematemesis or melena) and are hemodynamically stable (systolic blood pressure > 100 mmHg and pulse less than 100 per minute). The investigators intend to keep a log of all patients who present with AUGIB and are being referred for EGD. The investigators record their admission Glasgow Blatchford Score. The investigators exclude patients who are below the age of 18 and above 80 years, those with permanent pacemakers and other metallic implants that preclude magnetic resonance examinations, those with pharyngeal dysfunction that render them difficult to swallow, those with poor conscious state with the concern of aspiration and patients known to have risks of capsule retention e.g. previous gastric surgery, inflammatory bowel disease, patients who are unable to provide an informed consent and those in pregnancy.

Patients are required to fast for 6 hours. Before the examination, the patient ingests 400 mg simethicone dissolved in 100 ml of water to distend the stomach. He then drinks an addition of one liter of water 10 minutes before the examination. An activated NaviCam (ANKON) is then swallowed with the patient in a supine position in the MR console. This allows examination of the esophagus and the squamo-columnar junction. An operator then drives the magnetic robot and actively controls the capsule.

The stomach is examined according to a standardized protocol. (Figure 3) When the capsule reaches the stomach, the capsule is lifted away from the posterior wall, rotated, and advanced to the fundus (A), long shots (B), and close-ups (C) of the cardiac regions, and then the posterior wall (D), the lesser and greater curvature (E) and anterior wall (F) of the gastric body, followed by the angulus (G) and antrum (H) and finally the pylorus (I). Full examination of the stomach is possible with patients lying in different positions.

To enable transpyloric passage of the capsule, the capsule is oriented to obtain an en face view of the pylorus. The capsule is directed towards the pylorus. With active peristalsis, the capsule traverse across the pylorus. If this proves difficult, the patient is stood up and asked to walk freely. The capsule then drops into the duodenum. The patient is instructed to wear a jacket or harness with a data recorder. Real time examination of the duodenum is possible. The small bowel examination is completed usually after the patient is discharged or hospitalized.

The patient's comfort level is recorded over a 10 cm visual analogue scale at the end of the examination.

Decision to clinical admissions The decision to discharge the patient is left to the discretion of the on-duty GI specialist after review of the video sequences. Criteria to discharge include, Forrest IIc and III gastroduodenal ulcers, GI erosions together with minimal coffee materials or a clean stomach, admission hemoglobin of 8 g/dl and haematocrit > 0.24, good family support. Patients are discharged on appropriate medications e.g. a proton pump inhibitor. Those discharged will be asked to return the data recorder the next day. Patients are then reviewed by a GI specialist. Patients are also instructed to contact the on duty GI specialist at the Prince of Wales Hospital should they have any query. In general, a patient with a capsule examination that shows fresh blood in the GI tract, lesions with major stigmata of bleeding, evidence of esophago-gastric or upper GI varices are admitted for further management. Patients on anticoagulation therapy require to be admitted for titration of these drugs. Patients with significant comorbid illnesses that would require in hospital stabilisation.

Variables The investigators obtain patients' demographic data, medical history including concurrent illnesses and current and past drug use, biochemical and hemodynamic data.

The investigators adopt a standardised pathology reporting scheme to minimize interobserver reporting variability. Focal lesions were defined as peptic lesions (esophageal, gastric and duodenal erosions, and ulcers), vascular lesions (angioectasia, esophageal varices of all grades, gastric, and duodenal varices) and fresh/altered blood (without an obvious source). Significant lesions were those considered to be the cause of bleeding and included peptic ulcers (esophageal, gastric, and duodenal), esophageal varices of at least grade 2, and gastric varices. In fact, the computer station has an Artificial intelligence function that enables real time diagnosis of observed lesions and in addition the thoroughness and percentage of stomach examined.

All patients are followed up till day 30 after inclusion into the current study.

The investigators intend to enrol 100 consecutive patients. No formal sample size is calculated as this is considered a pilot study.

Inclusion Criteria:

• Patients who present with overt signs of acute upper gastroinestinal bleeding (hematemesis or melena) and are hemodynamically stable (systolic blood pressure > 100 mmHg and pulse less than 100 per minute).

Exclusion Criteria:

• patients who are below the age of 18 and above 80 years,
• patients with permanent pacemakers and other metallic implants that preclude magnetic resonance examinations
• patients who are difficult to swallow capsule (e.g pharyngeal dysfunction, poor conscious state
• patients who are at risk of capsule retention (e.g. previous gastric surgery, inflammatory bowel disease)
• patients who are unable to provide an informed consent
• patients who are in pregnancy