Abstract
The mortality rate related to variceal bleeding is high in patients with liver cirrhosis. Early detection and treatment of varices can reduce the risk of hemorrhage and thus decrease the mortality rate related to variceal bleeding. The study comprised 81 cirrhotic patients in training set, who were categorized into 2 groups: the patients with esophageal varices (EVs group) and the patients without esophageal varices (non-EVs group). The disparity in Cystatin C/albumin ratio (CAR) was assessed between these 2 groups. Subsequently, a regression model was constructed by generating a receiver operating characteristic (ROC) curve to calculate the area under the curve (AUC). Then an external validation was performed in 25 patients. Among patients with cirrhosis in training set, a statistically significant difference in CAR was observed between the EVs group and non-EVs group (P < .05). At the CAR cutoff value of 2.79*10-5, the AUC for diagnosing EVs were 0.666. Further, a multivariate logistic regression model was constructed, after adjusting the model, the AUC for EVs diagnosis were 0.855. And the external validation showed that the model could not be considered as a poor fit. CAR exhibits potential as an early detection marker for EVs in liver cirrhosis, and the regression model incorporating CAR demonstrates a strong capability for early EVs diagnosis.