Comparative performance of prognostic models in predicting mortality in cirrhotic patients with spontaneous bacterial peritonitis

Abstract

BACKGROUND: Prognostic models are widely used to predict mortality and management outcomes in liver cirrhosis patients with spontaneous bacterial peritonitis (SBP). However, it remains unclear whether these prognostic models can be applied to SBP.

OBJECTIVES: To determine the predictive value of prognostic models, including Child-Turcotte-Pugh (CTP), Model for End-stage Liver Disease score- Sodium (MELD-Na), Albumin to Bilirubin (ALBI), and Neutrophil-to-Lymphocyte Ratio (NLR) in cirrhotic patients with SBP.

METHODS: Seventy-four hospitalized cirrhotic patients with SBP were selected. Diagnosis was based on clinical, biochemical, ultrasonographic, and ascitic fluid analysis. CTP, MELD-Na, ALBI, and NLR scores at admission were calculated. The area under the ROC curve (AUC) was used to measure accuracy. Sensitivity and specificity were calculated for the optimal cut-off points.

RESULTS: Our results revealed that patients who died had higher scores in NLR (MD: 7.51; 95% CI: 1.46–13.56; p: 0.0150), MELD-Na (MD: 10.09; 95% CI: 6.91–13.27; p: 0.0000), CTP (MD: 2.57; 95% CI: 1.82–3.32; p: 0.0000), and ALBI (MD: 0.47; 95% CI: 0.22–0.73; p: 0.0000) compared to survivors. Among these scores, the highest AUC in univariate logistic regression analysis were CTP, MELD-Na, NLR, and ALBI with 0.87, 0.82, 0.73, and 0.72, respectively.

CONCLUSION: The combination of CTP and MELD-Na scores was superior to ALBI and NLR prognostic models. These can be used to assess liver function and prognosis in cirrhotic patients with SBP.