Diagnostic model for spontaneous bacterial peritonitis in cirrhotic patients with ascites: a multicenter cohort study.

Abstract

INTRODUCTION Spontaneous bacterial peritonitis (SBP) is a potentially life-threatening complication of cirrhotic ascites. Early diagnosis and treatment of SBP are essential to improve the survival rates and prognosis of patients. We aimed to identify the predictors of SBP to establish a new noninvasive early diagnostic tool. METHODS We screened 1618 patients who underwent paracentesis due to cirrhotic ascites between January 2017 and December 2018 in three hospitals. We evaluated the symptomatic, clinical, and laboratory parameters to identify the predictors of SBP. The primary diagnostic model was displayed as a nomogram. RESULTS The model included abdominal pain, diarrhea, white blood cell count, neutrophil percentage, procalcitonin, C-reactive protein, lactate dehydrogenase, glucose, and Model for End-stage Liver Disease score. The model's diagnostic performance was good, with an AUC of 0.84 [95% confidence interval (CI), 0.81-0.87] in the training cohort. In the validation cohort, the diagnostic ability was also good, with AUCs of 0.87 (95% CI, 0.83-0.91) and 0.90 (95% CI, 0.87-0.94) for inner and outer validation queues, respectively. Moreover, the decision curve analysis confirmed the clinical utility of the nomogram model. In addition, we developed a Microsoft Excel calculation model to allow convenient adoption of the model in clinical practice. CONCLUSION We developed good performing diagnostic models for SBP.