Objectives: Polymorphonuclear (PMN) cell count in ascitic fluid is the most useful test for the diagnosis of spontaneous bacterial peritonitis (SBP). We evaluated the validity of an automated blood cell counter for the PMN determination in ascitic fluid by comparing it with the traditional hematologic method with a light microscope in a manual counting chamber.
Methods: A total of 130 ascitic fluid samples were collected from 74 consecutive cirrhotics. The agreement between the two techniques was assessed according to Bland and Altman's method. The sensitivity, specificity, and positive and negative predictive values of the automated blood cell counter were calculated by considering the diagnosis of SBP as a PMN count > or = 250 cells/mm(3), determined by the manual method as the "gold standard."
Results: The mean PMN counts assessed by the manual method and the automated blood cell counter were 124 +/- 301 cells/mm(3) and 130 +/- 339 cells/mm(3), respectively (p = 0.89, ns). The mean +/- SD of the difference between manual and automated measurements was 6 +/- 61 cells/mm(3), whereas the limits of agreement were +127 cells/mm(3) (95% CI = +108 to +147) and -115 cells/mm(3) (95%CI = -96 to -135). SBP was diagnosed in 11 patients. All but one were correctly identified with the automated blood cell counter, with a sensitivity of 94% and a specificity of 100%; positive and negative predictive values were 100% and 99.1%, respectively.
Conclusions: The manual method and the automated blood cell counter have a good agreement in the PMN determination in ascitic fluid, and the automated blood cell counter is a reliable tool for rapid diagnosis of SBP.