PROSPECTIVE ANALYSIS OF PEDIATRIC UPPER GASTROINTESTINAL BLEEDING: A SINGLE CENTER EXPERIENCE

Introduction: Pediatric upper gastrointestinal bleeding (UGIB) is rare, but can have significant patient morbidity, and even mortality. Thomson et al. developed the Sheffield score to help stratify pediatric patients with scores ≥8 that warrant endoscopy for hemostatic intervention.
Aims: To assess the utility of the Sheffield score as a screen for endoscopic hemostatic intervention in pediatric UGIB and report the outcomes of a prospectively collected pediatric UGIB database at a single large academic pediatric center.
Methods: This was a prospective cohort study that enrolled patients admitted for signs and symptoms suggestive of UGIB, including hematemesis, melena, gastrostomy bleeding, and anemia. The study was conducted from Nov 2022 – Nov 2023. Patients were identified via the primary inpatient GI service or consult service. Data including demographics; laboratories, vitals, procedure details, and sequalae during an admission, and up to 30 days after an UGIB event were analyzed.
Results: A total of 106 patients, 65 males (61%) and 41 females (39%), median age 6.8 years, with UGIB were enrolled in our database. During the study period there were 33,681 admissions to our center, for an incidence of 0.31% for pediatric UGIB. Endoscopy was performed in 65 unique patients (61%). There were 8 (12%) cases of variceal source, and 57 (88%) non-variceal. A variety of standard endoscopic interventions for hemostasis were carried out in 21 of the 65 procedures (32%), 44 cases (68%) did not have an identifiable or treatable source. All 21 cases with hemostatic intervention had a Sheffield score of 8 or greater, whereas 36 of the 44 procedures that did not require intervention had a score of 8 or greater: sensitivity 100%, Specificity 18%, PPV 37%, and NPV 100%. The 21 cases with intervention had a mean Sheffield score of 13.6 [SD=3.2], whereas endoscopic procedures without a treatable source was 10.8 [SD=3.9]. Forty-one, of the 106 cases, did not have endoscopy, mean Sheffield score of 4.5 [SD=4.4]. There were 21 patients (20%) out of the 106 who had a rebleed event within 30 days (16 with prior endoscopy and 5 without initial endoscopy), mean Sheffield 11.5, 12 of 65 (18%) underwent repeat endoscopy to address the rebleed, and . During the study period there were 9 deaths related to all-causes out of the 106 cases, mean Sheffield 15.9 [SD=4.4].
Conclusions: This study summarizes the largest reported pediatric cohort of admitted patients prospectively followed over a year for an UGIB event during admission at a single center. In our cohort the Sheffield score had excellent sensitivity for assessing need of endoscopic hemostatic intervention- supporting the use of this screening tool in pediatric UGIB.