Bacterial infections in patients with liver cirrhosis: clinical characteristics and the role of C-reactive protein

Authors Melanie Deutsch, Spilios Manolakopoulos, Ioannis Andreadis, Markos Giannaris, George Kontos, Hariklia Kranidioti, Maria Pirounaki, John Koskinas.


Background The diagnosis of bacterial infection in cirrhotic patients may be difficult, because of the absence of classical signs such as fever and raised white blood cell count. The role of C-reactive protein (CRP) in this context has not been clearly defined.

Methods Clinical and laboratory characteristics of 210 consecutive cirrhotic patients with (n=100) or without (n=110) bacterial infection were compared with a control group of noncirrhotic patients with infection (n=106).

Results Significantly fewer patients with cirrhosis had a body temperature ≥37°C when presenting with bacterial infection (56% cirrhotic vs. 85.5% non-cirrhotic patients, P=0.01). Mean leukocyte count was 6.92 × 103/mm3 in patients with cirrhosis and infection, 5.75 × 103/mm3 (P=0.02) in cirrhotic patients without infection, and 11.28 × 103/mm3 in non- cirrhotic patients with infection (P<0.001). Multivariate analysis revealed that CRP level and model for end-stage liver disease score were significantly associated with the presence of infection in patients with cirrhosis. A cutoff level of CRP>10 mg/L indicated the presence of infection with a sensitivity of 68%, a specificity of 84.5% and an area under the receiver operating characteristic curve of 0.8197. CRP cutoff level differed according to the severity of the liver disease: Child-Pugh score (CPS) A: 21.3 mg/L, B: 17 mg/L, and C: 5.78 mg/L.

Conclusions CRP at admission could help diagnose infection in cirrhotic patients. Since the severity of liver disease seems to affect the CRP values, lower CRP levels might indicate infection. Clinical suspicion is necessary to avoid delay in diagnosis and initiate antibiotic treatment.

Keywords Bacterial infection, liver cirrhosis, C-reactive protein

Ann Gastroenterol 2018; 31 (1): 77-83

Original Articles