Predictive potential of the MELD and Child-Turcotte-Pugh II scores for SBP in patients with cirrhosis and ascites

Abstract

It is very important for patients with spontaneous bacterial peritonitis (SBP) to assess the length of survival and the risk of death, primarily because of the wide range of potential complications that can lead to multisystem organ failure and fatal outcome. The aim of this study was to determine the predictive potential of MELD and Child-Turcotte-Pugh II score for SBP in patients with cirrhosis and ascites. Material and methods: The study was designed as a prospective-analytical-observational and was conducted at the University Clinic for Gastroenterohepatology in Skopje for a period of one year. The study population included 70 hospitalized patients with established liver cirrhosis, regardless of etiology, divided into two groups, 35 patients with SBP and 35 non-SBP. Prognostic scores in patients with liver cirrhosis and ascites: MELD score, according to the formula: MELD = [(0.957 x Ln Creatinin) + (0.378 x Ln Bilirubin) + (1.12 x Ln INR) + (0.643) x 10]. The Child-Turcotte-Pugh II score includes 6 parameters: serum albumin and bilirubin, amount of ascites, degree of encephalopathy ( HE), prothrombin time (PT) and serum creatinine, and assessment of the degree of hepatic encephalopathy according to the West Haven criteria. Results: The average value of the MELD score in patients with SBP was 22.6 ± 8.27 and in non-SBP the average value was lower - 17.83±5.87. According to the Mann-Whitney U test, the difference between the mean values ​​was statistically significant for p <0.05 (z = 2.41; p = 0.015). A score of 30 to 39 was registered in 25.7% of patients with SBP, and only in 2.9% in non-SBP; the percentage difference was statistically significant for p <0.05 (Difference test, p = 0.0064 ). Patients with SBP had an average Child-Pugh score of 13.09 ± 2.48 or 100.0% C-class points. In patients with non-SBP, an average child-Pugh score of 9.63 ± 1.62 was recorded, or class B in 65.7% and class C in 34.3%. The percentage difference was statistically significant for p <0.05 (Difference test, p = 0.000000). According to the Mann-Whitney U test, the difference between the mean values ​​was statistically significant for p <0.05 (z = -5.44; p = 0.00001). ROC analysis indicated that the Child-Turcotte-Pugh II score contributed to the diagnosis of SBP - 90.7% (p = 0.000) (excellent predictor), closer to the ideal value of 1.0 and above the worst value of 0.5. ROC analysis indicated that the MELD score did not contribute to the diagnosis of SBP - 66.7% (p = 0.017) (weak predictor), closer to the worst value of 0.5. Conclusion:Our research confirmed that SBP occurs in patients with severe hepatic dysfunction calculated according to the CTP II score and MELD score. Mean value of the MELD score in patients with SBP was higher then in patients with non-SBP. On the other hand all patients with SBP had an average CTP II score, C-class points, while the largest percentage of patients with non-SBP were class B-class points. MELD score is a weak predictor of SBP. The best predictor for predicting SBP is the CTP II score (rank C).