Albumin‑bilirubin grade and INR for the prediction of esophagogastric variceal rebleeding after endoscopic treatment in cirrhosis

Rebleeding following endoscopic treatment in patients with cirrhosis is a serious life-threatening complication. In the present study, a novel, reliable and non-invasive score for prediction of rebleeding following endoscopic therapy for esophagogastric variceal bleeding (EGVB) was developed. The present retrospective study recruited cirrhotic patients with EGVB (n=596) who underwent endoscopic therapy. Patients hospitalized from January 2015 to January 2020 were grouped into a training (n=437) cohort to develop the new score and those hospitalized from February 2020 to February 2022 were grouped into a validation (n=159) cohort to validate the score. The international normalized ratio (INR) and albumin-bilirubin (ALBI) grade were used to develop the INR-ALBI (IALBI) score to predict risk of rebleeding. In the training cohort, the prognostic performance of the IALBI score and other ALBI-associated scores (modified ALBI, platelet-ALBI and ALBI-fibrosis-4) at 1, 3 and 12 months was assessed using receiver operating characteristic (ROC) curve and Kaplan-Meier analysis. At each time point, most areas under the ROC curve of IALBI were higher than those of other ALBI-associated scores, particularly for prediction of early rebleeding. At 1 month, the rebleeding rates of patients with IALBI grade 2 and 3 were ~10.0- and 19.5-times higher than those of patients with grade 1, respectively. The negative predictive value (NPV) of IALBI for the training and validation cohort at 1 month was 100.0 and 97.8%, respectively. For viral and non-viral patients in the training cohort, IALBI showed good predictive ability and NPV for early rebleeding. The IALBI grading system successfully assessed rebleeding, particularly early rebleeding, in cirrhotic patients with EGVB following endoscopic therapy IALBI grade 1, predicted low risk of rebleeding and may not require endoscopic treatment again in the short-term.