Leucocyte esterase reagent strips for the diagnosis of spontaneous bacterial peritonitis: a systematic review.

The reported incidence of spontaneous bacterial peritonitis (SBP) is 7-30% per annum in cirrhotic patients. Timely diagnosis and treatment is crucial to reduce mortality owing to this infection. Recently, leucocyte esterase reagent strips have been tested in the diagnosis of infection in the ascitic fluid. The objective was to evaluate the diagnostic value of leucocyte esterase reagent strips in SBP in cirrhotic patients with ascites, by systematically reviewing the evidence from prospective clinical studies. We performed a comprehensive literature search in Medline up to July 2007 for adult human prospective clinical studies. Two reviewers independently checked all identified studies for fulfillment of predefined inclusion criteria, extracted data and assessed methodological quality of included studies. We had decided a priori to pool the studies via meta-analysis, only if statistical heterogeneity was found to be nonsignificant (P>0.10). Seventeen studies were included. Statistical heterogeneity among studies was found to be highly significant (P<0.001) in all analyses, precluding pooling of data for meta-analysis. Compared with the manual polymorphonuclear count ('gold standard'), leucocyte esterase reagent strips were found to have sensitivity ranging from 45 to 100%, specificity ranging from 81 to 100%, positive predictive value ranging from 42 to 100% and negative predictive value ranging from 87 to 100%. Despite the wide variation in sensitivity and positive predictive value between studies, the consistently high negative predictive value of leucocyte esterase reagent strips in SBP diagnosis should gain it a place in the ascitic tap diagnostic algorithm.