Mo1899 C-Reactive Protein Predicts Early Mortality After Acute Esophageal Variceal Bleeding in Patients With Cirrhosis

Abstract

Background and Aims: Besides infections, C-reactive protein (CRP) levels may be elevated in patients with cancer or systemic inflammation. Although synthesized by the liver, elevated CRP levels can be observed in the event of advanced liver failure. The aim of this study was to investigate the prognostic value of CRP after acute variceal bleeding in cirrhotic patients. Methods: We evaluated 144 consecutive cirrhotic patients (Child-Pugh A and B: C = 80: 64) admitted with esophageal variceal bleeding between January 2009 and May 2014. 6-week mortality assessments according to risk factors were performed. We assessed the optimal CRP cutoff by receiver operating characteristic (ROC) curve and tested its impact on 6-week mortality by univariate analysis and multivariate logistic regression analysis. We next investigated the 6-week mortality of patients with elevated CRP associated with infection and patients with elevated CRP non-associated with infection. Results: The overall 6-week mortality rate was 22.3%. Endoscopic band ligation was feasible in 91.8% of patients. ChildPugh C patients showed a significant high mortality compared to Child-Pugh A or B patients (41.0% vs. 6.3%, P<0.0001). We defined a CRP level<0.53/^0.53 mg/dL as optimal cutoff for further 6-week mortality assessments in Child-Pugh C patients. In Child-Pugh C patients, CRP^0.53 mg/dL (elevated CRP) at admission was associated with 6-week mortality univariately (elevated CRP vs. CRP<0.53 mg/dL (normal CRP), 69.0% vs. 15.6%; P<0.0001). In Child-Pugh A and B patients, elevated CRP was not associated with higher 6-week mortality (elevated CRP vs. normal CRP, 12.0% vs. 3.7%, P=0.159). Elevated CRP non-associated with infection in Child-Pugh C patients was also associated with an increase in 6-week mortality (elevated CRP non-associated with infection vs. normal CRP, 63.6% vs. 18.2%, P=0.006). Child-Pugh C patients with elevated CRP associated with infection did not show significantly higher 6-week mortality, compared to those with non-associated with infection (70.6% vs. 63.6%, P=0.7005). By multivariate analysis, we determined that elevated CRP was the independent predictor for 6-week mortality (OR 10.37, 95%CI 1.86-92.00, P=