Glasgow Blatchford scoring system enables accurate risk stratification of patients with upper gastrointestinal haemorrhage

OBJECTIVE: To evaluate Glasgow Blatchford (GB) scores ability for risk stratification in patients presenting with upper gastrointestinal bleeding (UGIB). METHODS: The prospective cohort study was conducted in the inpatient department of medicine at Fauji Foundation Hospital Rawalpindi, Pakistan, from April to September 2021. One hundred and thirty patients with UGIB (hematemesis, melena, and blood in the nasogastric tube) were included by consecutive sampling technique. We excluded traumatic patients with UGIB, pregnant females, patients with chronic kidney disease, anorexia nervosa, bulimia nervosa, and chronic diarrhea. Laboratory and demographic data were collected. The GB score was calculated at the time of admission. Data was analyzed through SPSS version 23, and frequencies were deduced. Groups were compared using the chi-square test. RESULTS: Mean age of patients was 61.1±13.8 years. There were 56 (43.1%) males and 74 (56.9%) females in the study. The main reason for acute gastrointestinal bleeding was Hepatitis C-associated portal hypertension (n = 103; 79.2%), followed by non-steroidal anti-inflammatory drug-induced gastrointestinal bleeding (n=13; 10.0%). There were 90 (69.2%) patients in high-risk group (Group A) and 40 (30.8%) in low-risk group (Group B). The high-risk group had a significantly higher GB score than the low-risk group (11.61±3.2 vs 3.85±1.9, p<0.001). GB score of ≥4 has sensitivity of 97.7%, a specificity of 92.5%, and an area under curve of 0.967 with a p-value of<0.001. Conclusion: GB score has an excellent accuracy for risk stratification of patients with UGIB. With a cutoff of ≥4, GB score accurately identifies 97.7% of high risk patients.