NOBLADS - External Validation of a Risk Scoring System for Severe Acute Lower Gastrointestinal Bleeding

BACKGROUND: Acute lower gastrointestinal bleeding is a common cause of hospital admission. NOBLADS is a lower gastrointestinal bleeding clinical risk score. OBJECTIVE: This study aimed to externally validate NOBLADS in predicting severe acute lower gastrointestinal bleeding and clinical outcome. DESIGN: Observational retrospective study. SETTING: This study was performed in a single large tertiary hospital. PATIENTS: Patients who were admitted with acute lower gastrointestinal bleeding during a 15-month period and underwent endoscopic evaluation were included. Patients with chronic lower gastrointestinal bleeding or acute lower gastrointestinal bleeding who were admitted for other conditions were excluded. MAIN OUTCOME MEASURES: Primary outcome was acute severe lower gastrointestinal bleeding (classified as severe if requires transfusion of >2 units of packed red blood cells and/or produces a >20% hematocrit fall). Secondary outcomes: red blood cells requirement, therapeutic intervention and hospital stay duration. NOBLADS score was applied to all patients. Its accuracy to predict acute severe bleeding and secondary outcomes were studied using receiver operating characteristic analysis. RESULTS: A total of 173 patients (50.3% males, age 69 ± 17 years) were included. The most common lower gastrointestinal bleeding etiologies were diverticular bleeding (18.5%) and ischemic colitis (15.6%). Fifty patients (28.9%) presented criteria for severe bleeding and 33 patients required instrumental intervention: endoscopic (n = 28), surgical (n = 4), and radiologic (n = 1) therapy. NOBLADS score was significantly different according with acute lower gastrointestinal bleeding causes (p < 0.001) and accurately predicted severe bleeding (area under the receiver operating characteristic curve 0.923 ± 0.018 (p < 0.001)). NOBLADS ≥2 detected acute severe lower gastrointestinal bleeding with 100% sensitivity/62.4% specificity, identifying the need for higher red blood cells requirement (3.6 vs 0.08, p < 0.001), therapeutic intervention (38% vs 13%, p < 0.001), and longer hospital stay (12.8 vs 3 days, p < 0.001). LIMITATIONS: Unicenter retrospective study; number of unprepared sigmoidoscopies/left side colonoscopies; comparison between NOBLADS and other scoring systems was not performed. CONCLUSION: NOBLADS is a practical, highly accurate tool and predicts the need of in-hospital management, early colonoscopy, red blood cells transfusion, and longer hospital stay in patients admitted with acute lower gastrointestinal bleeding.