A creatinine-based model for predicting recurrent bleeding after modified percutaneous transhepatic variceal embolization in patients with cirrhosis

Background

Patients who survive initial esophagogastric variceal bleeding (EVB) are at an increased risk of recurrent bleeding and death; however, a reliable predictive model is lacking. We aimed to develop a model for rebleeding prediction in patients with EVB after modified percutaneous transhepatic variceal embolization (PTVE) with cyanoacrylate.

Methods

A total of 122 patients with EVB who underwent PTVE from January 2015 to November 2020 were enrolled. Multivariate logistic analyses were conducted to determine independent risk factors for nomogram construction. The discrimination, calibration, and clinical utility of the nomogram were compared with the Model for End-stage Liver Disease score (MELD) and the Child–Pugh model. Risk stratification was performed according to the nomogram.

Results

Rebleeding within 3 months of PTVE occurred in 32 patients (26.2%). Independent rebleeding indicators included prior history of endoscopic therapy, Child–Pugh score, partial splenic embolization, and creatinine level. The nomogram incorporating these four predictors achieved excellent calibration and discriminatory abilities, with a concordance index of 0.85, which was confirmed to be 0.83 through bootstrapping validation. The nomogram demonstrated superior discrimination and clinical applicability than the MELD and Child–Pugh models. As shown in the Kaplan–Meier curves, high-risk patients had a high probability of rebleeding (P ​< ​0.001).

Conclusions

The creatinine-based nomogram had a superior ability to predict rebleeding after PTVE in patients with EVB. Risk stratification may help identify high-risk patients and lead to the earlier implementation of aggressive treatments and formulation of intensive follow-up plans.