Abstract
Ascitic fluid infection is a major cause of morbidity and mortality in cirrhotic patients, requiring early diagnosis and therapy. We aimed to determine predictors of ascitic fluid infection in children with chronic liver disease. The study included 45 children with chronic liver disease and ascites who underwent 66 paracentesis procedures. Full history taking and clinical examination of all patients were obtained including fever, abdominal pain and tenderness and respiratory distress. Investigations included: complete blood count, C-reactive protein, full liver function tests, ascitic fluid biochemical analysis, cell count and culture. Our results showed that patients' ages ranged between 3 months to 12 years. Prevalence of ascitic fluid infection was 33.3%. Gram-positive bacteria were identified in six cases, and Gram-negative bacteria in five. Fever and abdominal pain were significantly more associated with infected ascites (p value = 0.004, 0.006). Patients with ascitic fluid infection had statistically significant elevated absolute neutrophilic count and C-reactive protein. Logistic regression analysis showed that fever, abdominal pain, elevated absolute neutrophilic count and positive C-reactive protein are independent predictors of ascitic fluid infection. Fever, elevated absolute neutrophilic count and positive C-reactive protein raise the probability of ascitic fluid infection by 3.88, 9.15 and 4.48 times respectively. The cut-off value for C-reactive protein for ascitic fluid infection was 7.2 with sensitivity 73% and specificity of 71%. In conclusion, prevalence of ascitic fluid infection in pediatric patients with chronic liver disease and ascites was 33.3%. Fever, abdominal pain, positive C-reactive protein and elevated absolute neutrophilic count are strong predictors of ascitic fluid infection. Therefore an empirical course of first-line antibiotics should be immediately started with presence of any of these predictors after performing ascitic fluid tapping for culture and sensitivity. In absence of these infection parameters, routine ascitic fluid analysis could be spared.