The analysis of body fluids plays an important role in the diagnosis of inflammatory, infectious, malignant, and other pathological conditions. In addition to the total leukocyte count, a key component of laboratory evaluation is leukocyte differentiation, which is traditionally performed by microscopic examination of cytological slide. However, this method is time-consuming and subjective. The aim of this master’s thesis is to compare the microscopic examination of cytological slide with automated flow cytometric method using the Sysmex XN-1000 analyzer with body fluid module for leukocyte differentiation, and to evaluate the use and interchangeability of both methods in clinical practice. A total of 1013 body fluid samples were included in the study (289 cerebrospinal fluid, 214 peritoneal fluid, 355 pleural fluid, and 155 synovial fluid samples). The results were statistically evaluated using Spearman’s coefficient, Passing–Bablok regression, and Bland–Altman plot. A good agreement between methods was found for both absolute and relative polymorphonuclear and mononuclear counts. On average, the analyzer measured 2,4% more polymorphonuclear leukocytes and 2,4% fewer mononuclear leukocytes, with the largest differences observed in peritoneal fluid, where the automated method measured 7,7% more polymorphonuclear leukocytes and 7,6% fewer mononuclear leukocytes. The comparability of the methods for measuring individual cell populations (neutrophils, lymphocytes, and monocytes) was generally good. The greatest differences for lymphocytes were observed in pleural fluid, where the automated method measured 10,8% more lymphocytes. The agreement between methods was poorer for monocyte differentiation, showing a negative proportional bias (Passing–Bablok regression equation for relative monocyte count across all fluids: y = 1,638 + 0,669x). Measuring eosinophils using the automated method was challenging due to their low concentrations and the large relative deviations between methods. The observed differences between the microscopic and automated method were substantially smaller than the diagnostic decision limits for key clinical conditions. Therefore, it is unlikely that these differences would lead to misdiagnosis of spontaneous bacterial peritonitis, meningitis, pleural empyema, or septic arthritis.
|