Medline ® Abstract for Reference 5
of 'Spontaneous bacterial peritonitis in adults: Clinical manifestations'
Association between model for end-stage liver disease and spontaneous bacterial peritonitis.
Obstein KL, Campbell MS, Reddy KR, Yang YX
Am J Gastroenterol. 2007;102(12):2732.
OBJECTIVE: To determine whether a greater Model for End-Stage Liver Disease (MELD) score is associated with a greater risk of spontaneous bacterial peritonitis (SBP).
METHODS: Our retrospective case-control study enrolled 271 consecutive patients with cirrhosis and ascites who underwent diagnostic paracentesis upon hospital admission (2002-2005). After excluding immunosuppressed patients, those recently exposed to antibiotics, those with a potential confounding etiology for ascites, and those with a prior history of SBP, 111 were included in the study. SBP was defined as a paracentesis yielding>or=250 neutrophils/mL ascites fluid. Multivariable logistic regression was performed to determine the odds ratio for the development of SBP associated with MELD score and grouped MELD score (<or=15, 16-24,>or=25). Potential confounders assessed included age, diabetes mellitus, gender, race, alcohol use, serum sodium, and etiology of liver disease.
RESULTS: Twenty-nine of 111 hospitalized patients with cirrhosis were found to have SBP. Patient characteristics were similar between groupswith and without SBP. The mean MELD score for patients with SBP was 24 and for those without 18 (P=0.0003). The odds ratio for developing SBP by each MELD point was 1.11 (1.05-1.19, P=0.001). Patients with MELD>or=25 had an odds ratio of 9.67 (2.35-39.82, P=0.002) for SBP, compared to subjects with MELD<or=15. None of the potential confounders substantially altered the relationship between MELD and SBP.
CONCLUSIONS: Increasing MELD score is independently associated with a greater risk of SBP. For every point increase in the MELD score, the risk of developing SBP increases by 11%. Our finding has important implications for increasing the suspicion of SBP in patients with elevated MELD scores.
Department of Medicine, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA.