Introduction: Gastrointestinal (GI) bleeding affects 30-40 per 100,000 hospitalized patients annually. Many of these patients can have severe bleeding with hemodynamic instability and rapid clinical deterioration, often requiring higher levels of care such as intermediate units (IU) or intensive care units (ICU). This study aimed to compare various risk stratification tools with patient outcomes to find which tools can best help providers triage patients most appropriately on admission.
Methods: This is a single-center retrospective cohort analysis of patients admitted to IU or ICU with GI bleeding as their primary diagnosis from March, 2015 - March, 2021. Medical records of patients above 18 years of age were reviewed for baseline characteristics, lab values, 30-day mortality, and 90-day readmission. Charlson comorbidity index (CCI), Glasgow-Blatchford Bleeding Score (GBS), AIMS 65, Assessment of Blood Consumption (ABC), quick Sequential Organ Failure Assessment (qSOFA) scores on admission were calculated. Patients were compared according to the level of care. Pearson Chi-square and Mann Whitney U were applied to compare groups.
Results: Out of 299 patients admitted with GI bleeding, 195 (65.2%) were admitted to IU and 104 (34.8%) to ICU. Baseline characteristics are illustrated in Table 1. As for treatment, antibiotics (28.8% ICU vs. 10.3% IU; p< 0.01) and PRBC transfusions (median: 3.0 ICU vs. 2.0 IU; p< 0.01) were more frequently utilized in ICU. Outcome variables assessed included need for endoscopic intervention, time to scope, and 90-day readmission rates; no significant difference was seen between groups. Patients admitted to the IU had lower 30-day mortality (p=0.02). Out of the five scores assessed, GBS, AIM-65, and qSOFA were noted to be statistically significant with score being higher in patients admitted to ICU: median (interquartile range (IQR))– GBS: 12 (9, 14.75) vs. 11.00 (8, 13); p< 0.05; AIM-65: 1 (1,2) vs. 1(0,2); p< 0.01, and qSOFA: 0(0,1) vs. 0(0,1); p< 0.01.No significant difference was noted between the median (IQR) CCI and ABC scores between ICU vs. IU: CCI: 5(3,7) vs. IU: 5 (3,7); p >0.05 and ABC: 0 (0,1) vs. IU: 0 (0,1); p >0.05.
Discussion: Our study highlights the utility of scoring tools including GBS, qSOFA, and AIM-65 to assist with triaging GI bleed patients to an appropriate level of care. The scores should be calculated at the time of admission for GI bleed patients and those with elevated scores may benefit from closer monitoring in the ICU.
