Gaya DR et al. (2007) Bedside leucocyte esterase reagent strips with spectrophotometric analysis to rapidly exclude spontaneous bacterial peritonitis: a pilot study. Eur J Gastroenterol Hepatol 19: 289–295

The use of leukocyte esterase reagent strips offers a rapid, cheap and simple alternative to manual counting of polymorphonuclear leukocytes (PMNs) in ascitic fluid for the diagnosis of spontaneous bacterial peritonitis (SBP). Reagent strips have not been widely adopted, however, because of concerns about the subjectivity of reading strip results.

To see whether this problem could be overcome, Gaya et al. screened 173 ascitic fluid samples from 105 cirrhotic patients with reagent strips, and read them using an automated spectrophotometer. The results were compared with PMN counts. The spectrophotometrically read reagent strip results confirmed all 14 diagnoses of SBP made by PMN counts, and also showed that 3 of these 14 patients still had SBP 48 h after initiation of antibiotic therapy.

Compared with manual PMN counts, the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of the spectrophotometrically read reagent strip results were 100%, 91%, 50%, 100% and 92%, respectively.

Since the positive predictive value was relatively poor, the authors suggest that a positive strip result should be confirmed by a manual PMN count. A negative strip result, however, reliably excludes a diagnosis of SBP and eliminates the need for a manual PMN count, which could save time, money and resources. They also highlight the need for their data to be replicated in a larger cohort with a higher prevalence of SBP before widespread application of spectrophotometric analysis of reagent strips.