A Prospective Study to Evaluate the Diagnostic Performance of a Wireless Optical Sensor Capsule in Detection of Upper Gastrointestinal Bleeding

Upper gastrointestinal bleeding (UGIB) is a common and potentially lethal medical emergency, which requires hospitalization and healthcare resources. Despite the changing epidemiology and advancement of endoscopic therapy in the recent decade, the all-cause mortality of UGIB remains high (>2%). An accurate risk stratification is necessary to triage patients into low- or high-risk groups, in order to inform clinicians for the necessity and timing of urgent endoscopy. However, existing pre-endoscopy risk scores, such as the Glasgow-Blatchford score (GBS), Rockall score and AIMS65 score, are suboptimal in predicting relevant clinical outcomes. An alternative strategy for risk stratification is urgently warranted in patients with UGIB to guide the next step of management.

In a pre-clinical study, the sensitivity and specificity of HemoPill® prototype were 95% and 87.5% respectively, when the sensors were positioned close to the bleeding point. Furthermore, several human clinical studies have proven the feasibility and accuracy of HemoPill® in healthy volunteers and patients with suspected UGIB. Capsule ingestion was well tolerated with no device-related adverse event or capsule retention. All patients with negative HI were found to have no active endoscopic bleeding (true negative, 100%, 17/17). In patients with bleeding >20ml, true positive HI signals were detected (100%, 2/2) In a retrospective multi-center study, 61 patients with suspected UGIB were recruited to use HemoPill®. Among the capsule-positive cases, subsequent endoscopy confirmed active bleeding in 57% (20/35) of them. None of the capsule-negative patients rebled (0%, 0/25), which prevented unnecessary emergent endoscopy in 72% of them (18/25).