Total bilirubin serves as a robust predictor of 6-week mortality in patients with liver cirrhosis and acute esophagogastric variceal bleeding

Background Acute esophagogastric variceal bleeding (AEGVB) is a frequent and life-threatening complication of liver cirrhosis. This study aimed to investigate robust factors associated with 6-week mortality in cirrhotic patients with AEGVB using advanced statistical analysis. Methods We retrospectively enrolled 656 consecutive cirrhotic patients with AEGVB from April 1st, 2021 to September 30th, 2022 in Beijing You’an Hospital, Capital Medical University. A 1:4 matched analysis was performed to adjust the effect of admission time on statistical results. Cluster analysis was used to divide the whole cohort into three groups with distinct clinical characteristics. Kaplan-Meier analysis was used to estimate 6-week overall survival among the three clustered groups. Cox regression analysis were used to investigate predictors of 6-week mortality for cirrhotic patients with AEGVB. A competing risk model was used to identify robust predictors for 6-week mortality. The prognostic value of total bilirubin (TB) was assessed using restricted cubic spline (RCS) analysis. Results Cluster analysis identified the top 10 most important variables for clustering, including TB and D-dimer. After clustering with these variables, we found that cluster 0 group had the highest rates of early death, rebleeding and hemostasis failure. Kaplan-Meier analysis demonstrated that the 6-week mortality rate was significantly higher in the cluster 0 group than in the cluster 1 and 2 groups. Cox regression analysis showed that 6-week mortality was independently associated with several variables including 6-week rebleeding (HR 21.904, 95% CI 8.446 to 56.805, P  < 0.001), TB (HR 1.011, 95% CI 1.007 to 1.014, P  < 0.001), rebleeding within 72h (HR 16.767, 95% CI 6.309 to 44.556, P  < 0.001), and rebleeding within 4–5 days (HR 10.137, 95% CI 2.338 to 43.945, P  = 0.002). However, the competing risk model demonstrated that for 6-week mortality, TB was the only significant risk factor, with an HR of 1.43 and 95% CI of 1.100 to1.860 ( P  = 0.008). Furthermore, RCS analysis indicated that TB level above 127.35µmol/L was associated with a significantly increased risk of 6-week mortality. Conclusion Cluster analysis allows the identification of distinct profiles of AEGVB that form clinically relevant subsets. TB can serve as a robust biomarker for assessing short-term mortality risk in cirrhotic patients with AEGVB.