Sensitivity of a bedside reagent strip for the detection of spontaneous bacterial peritonitis in ED patients with ascites.

STUDY

OBJECTIVE

To determine the sensitivity of a highly sensitive bedside leukocyte esterase reagent strip (RS) for detection of spontaneous bacterial peritonitis (SBP) in emergency department (ED) ascites patients undergoing paracentesis.

METHODS

We conducted a prospective, observational cohort study of ED ascites patients undergoing paracentesis at two academic facilities. Two practitioners, blinded to each other's results, did a bedside RS analysis of the peritoneal fluid in each patient and documented the RS reading at 3-min according to manufacturer-specified colorimetric strip reading as either "negative", "trace", "small", or "large". The primary outcome measure was sensitivity of the RS strip for SBP (absolute neutrophil count ≥ 250 cells/mm3) at the "trace" threshold (positive equals trace or greater).

RESULTS

There were 330 cases enrolled, with 635 fluid analyses performed. Of these, 40 fluid samples had SBP (6%). Bedside RS had a sensitivity, specificity, positive predictive value, and negative predictive value of 95% (95% CI 82%-99%), 48% (95% CI 44%-52%), 11% (95% CI 10%-11%), and 99% (95% CI 97%-99%) respectively at the "trace" threshold for the detection of SBP.

CONCLUSION

Bedside use of the RS in ED ascites patients demonstrated high sensitivity for SBP. Given the wide confidence intervals, we cannot currently recommend it as a stand-alone test. We recommend further study with a larger number of SBP patients, potentially combining a negative RS result with low clinical suspicion to effectively rule out SBP without formal laboratory analysis.