C-reactive Protein Can Predict Patients with Cirrhosis at a High Risk of Early Mortality after Acute Esophageal Variceal Bleeding

Abstract

Objective The aim of this study was to identify patients with a high risk of early mortality after acute esophageal variceal bleeding by measuring the C-reactive protein (CRP) level. Methods We retrospectively evaluated 154 consecutive cirrhotic patients admitted with acute esophageal variceal bleeding. Differences between categorical variables were assessed by the chi-square test. Continuous variables were compared using the Mann-Whitney U-test. Multivariate logistic regression analyses consisting of clinical laboratory parameters were performed to identify risk factors associated with the 6-week mortality. The discriminative ability and the best cut-off value were assessed by a receiver-operating characteristic (ROC) curve analysis. Results Child-Pugh C patients showed a significantly higher 6-week mortality than Child-Pugh A or B patients (38% vs. 6%, p<0.0001). The 6-week mortality in Child-Pugh C patients was associated with the age (p<0.0001), etiology of cirrhosis (p=0.003), hepatocellular carcinoma (p=0.0003), portal vein thrombosis (p=0.005), baseline creatinine (p=0.0001), albumin (p=0.001), white blood cell count (p=0.038), baseline CRP [p=0.0004; area under the ROC (AUROC)=0.765; optimum cut-off value at 1.30 mg/dL] and bacterial infection (p=0.019). We determined that CRP ≥1.30 mg/dL was an independent predictor for 6-week mortality in Child-Pugh C patients [odds ratio (OR)=8.789; 95% confidence interval (CI): 2.080-47.496; p=0.003], along with a creatinine level of 0.71 mg/dL (OR=17.628; 95% CI: 2.349-384.426; p=0.004) (73% mortality if CRP ≥1.30 mg/dL vs. 19% if CRP<1.30 mg/dL, p<0.0001). Conclusion In Child-Pugh C patients with esophageal variceal bleeding, a baseline CRP ≥1.30 mg/dL can help identify patients with an increased risk of mortality.

Keywords: C-reactive protein; acute esophageal variceal bleeding; cirrhosis; early mortality.

Figure 1.

Flowchart of patients admitted with variceal bleeding to our hospital and the selection of the study population.

Figure 2.

ROC curves of the model for predicting the six-week mortality in acute esophageal variceal bleeding. Optimum cut-off values of the baseline CRP level and serum creatinine level for distinguishing patients with or without 6-week mortality were defined. The AUROCs of CRP and serum creatinine were not significantly different (p=0.801).