A Novel Score-Based Approach by Using Routine Laboratory Tests for Accurate Diagnosis of Spontaneous Bacterial Peritonitis (SBP) in Cirrhotic Patients

• Published in: EJIFCC Vol. 34; no. 4; pp. 297 - 304
• Main Authors: Abdo, George, Nir, Uri, Rawajdey, Rasha, Abu Dahoud, Wadie, Massalha, Jammal, Hajouj, Taleb, Assadi, Mohammad H, William, Nseir
• Format: Journal Article
• Language: English
• Published: Italy The Communications and Publications Division (CPD) of the IFCC 01.12.2023
• Subjects: Original; SBP; Cirrhosis; Laboratory; Peritonitis
• ISSN: 1650-3414; 1650-3414

Spontaneous Bacterial Peritonitis (SBP) poses a significant risk to cirrhosis patients with ascites, emphasizing the critical need for early detection and intervention. This retrospective observational study spanning a decade aimed to devise predictive models for SBP using routine laboratory tests. Additionally, it aimed to propose a novel scoring system to aid SBP diagnosis. Data analysis encompassed 229 adult cirrhotic patients hospitalized for ascites between 2012 and 2021. Exclusions eliminated cases of secondary ascites unrelated to liver cirrhosis. Patients were categorized into SBP-positive (n=110) and SBP-negative (n=119) groups. Comparative analysis of demographic details and various laboratory indicators (Neutrophil-to-Lymphocyte Ratio (NLR), Mean Platelet Volume (MPV), C-Reactive Protein (CRP), Platelet (PLT), Alanine Transaminase (ALT), Aspartate Amino Transferase (AST), Potassium (K), Sodium (Na), Total Bilirubin (TB) and International Normalized Ratio (INR) was performed between the groups. The study presented effective SBP prediction models for prompt diagnosis and treatment: a multivariate logistic regression model and a simple scoring system. The study advocates early diagnosis and rapid treatment for all cirrhotic patients with ascites, regardless of cirrhosis stage. Furthermore, it recommends initiating SBP treatment for patients scoring 2-3 in the proposed scoring system while excluding SBP findings for those scoring zero. Combining age, sex, and specific laboratory tests (MPV, NLR, CRP, TB, and INR) within random forest models and a simple scoring system enables swift and accurate SBP diagnosis.