This study aims to evaluate the safety and diagnostic performance of the PillSense™ optical blood-detection capsule as a triage tool for hospitalized patients with anemia or positive FOBT but no overt gastrointestinal bleeding. The primary objective was to determine the negative predictive value of a negative PillSense result for excluding active GI bleeding. We hypothesized that a negative PillSense reading would reliably identify patients who could be safely discharged without the need for urgent inpatient endoscopy.

This was a prospective, single-center, blinded study at a tertiary U.S. hospital. Hospitalized adults with anemia and/or positive FOBT, without overt GI bleeding and without capsule contraindications, were enrolled. After informed consent, patients ingested the PillSense capsule, and results (“Blood Detected” vs “No Blood Detected”) were collected by a research team member. Treating endoscopists were blinded to capsule results. Endoscopic evaluation (EGD, enteroscopy, and/or colonoscopy) was performed per standard clinical care. Primary outcome was safety and ability of a negative PillSense result to predict the safe discharge of hospitalized patients with positive hemoccult tests and/or anaemia.

Ten patients were enrolled, with a mean age of 59.1 years and 70% female. PillSense generated a negative signal within the first 10 minutes in 9 of 10 patients (90%). Among these nine patients, extended monitoring was performed in eight between 4 and 6 hours. Five of these eight patients (62.5%) later developed a positive signal, while three patients (37.5%) remained negative throughout prolonged monitoring. In all three cases with persistently negative signals, subsequent endoscopic evaluation confirmed the absence of any identifiable bleeding source, supporting the reliability of the negative result.

Among the five patients who transitioned from an early negative to a delayed positive signal, endoscopic evaluation demonstrated non-actively bleeding lesions. These included a gastric ulcer in the upper gastrointestinal tract, a small-bowel arteriovenous malformation, and a hemorrhoidal source in the lower gastrointestinal tract. In one additional patient who was initially negative at 10 minutes, no further PillSense data could be obtained; endoscopy revealed a non-bleeding gastric polyp and diverticulosis.

Importantly, across all nine patients with a negative PillSense result at 10 minutes, no cases of active upper gastrointestinal bleeding were identified endoscopically, yielding a negative predictive value of 100% for active UGI bleeding in this cohort. No patient required an endoscopic hemostatic intervention, and no capsule-related adverse events occurred. One patient required an unrelated readmission.

In hospitalized patients with anemia or positive FOBT but no overt gastrointestinal bleeding, a negative PillSense result reliably excluded active upper GI bleeding, with an observed negative predictive value of 100% in this cohort. These findings support the safety of using PillSense as a rapid triage tool to identify patients who may not require urgent inpatient EGD and who can instead be deferred to outpatient evaluation. Because several patients in this study remained admitted solely while awaiting endoscopy, the ability to confidently rule out active upper GI bleeding has the potential to meaningfully reduce unnecessary hospital days, decrease costs, and improve bed utilization in resource-constrained settings.

PillSense also accurately detected the single case of active bleeding and demonstrated that a delayed positive signal, occurring after 4-6 hours, corresponded to clinically relevant lesions in the upper GI tract, small bowel, and even the lower GI tract that had ceased active bleeding. Together, these observations highlight the promise of PillSense as a point-of-care technology capable of both safely ruling out active bleeding early and identifying occult lesions over time, ultimately streamlining care pathways and optimizing resource allocation.