Detection and validation of serum C-reactive protein and procalcitonin as diagnostic markers for bacterial infections in patients with cirrhosis of liver

Background: The role of inflammatory markers like CRP and procalcitonin in predicting various outcomes in patients with cirrhosis is gaining lot of attention. There is a need for extensive studies, to be carried out in India as there is no adequate literature available on the subject. Objectives: To assess the predictive validity of c-reactive protein and procalcitonin in predicting bacterial infection and mortality in patients with cirrhosis. Materials and methods: A prospective observational study conducted in the Department of Hepatology, at Madras Medical College, Chennai. Patients admitted to the Hepatology ward from March2016 to February 2017with acute decompensation of liver cirrhosis were studied. The serum procalcitonin level was assessed by Electro Chemi Luminescence Immuno Assay (Eclia) with a measuring range of 0.02-75ng/ml. and C-Reactive Protein level was assessed by ImmunoTurbido Metric Assay, with a measuring range of 1.00-200mg/l. The utility of CRP and procalcitonin in predicting the infection and mortality was assessed by Receiver Operative curve (ROC) analysis. Results: Procalcitonin had a better predictive validity than C-reactive protein in predicting the bacterial infection in the study population as indicated by their AUC curve as 0.99 (95% CI 0.99, 1.00, p value <0.001), for Procalcitonin and 0.84 (95% CI 0.76, 0.92, p value <0.001) for C-reactive protein. In predicting the mortality, C-reactive protein had a better predictive validity when compared to Procalcitonin as indicated by their AUC curve as 0.804 (95% CI 0.68, 0.92, p value <0.001) for C-reactive protein and 0.63 (95% CI 0.48, 0.77, p value <0.001) for Procalcitonin. Conclusion: More than one third of hospitalized Cirrhosis patients had infection and mortality rate was just over 20%. PCT has shown better predictive validity as compared to CRP in predicting infection, but CRP had better predictive validity in predicting mortality.