Evaluation of Flow Cytometry for Cell Count and Detection of Bacteria in Biological Fluids

Abstract

The analysis of biological fluids is crucial for the diagnosis and monitoring of diseases causing effusions and helps in the diagnosis of infectious diseases. The gold standard method for cell count in biological fluids is the manual method using counting chambers. The microbiological routine procedures consist of Direct Gram staining and culture on solid or liquid media. We evaluate the analytical performance of SYSMEX UF4000 (Sysmex, Kobe, Japan) and Sysmex XN10 (Sysmex, Kobe, Japan) in comparison with cytological and microbiological routine procedures. A total of 526 biological fluid samples were included in this study (42 ascitic, 31 pleural, 31 peritoneal, 125 cerebrospinal, 281 synovial, and 16 peritoneal dialysis fluids). All samples were analyzed by flow cytometry and subsequently processed following cytological and/or microbiological routine procedures. With regard to cell counts, UF4000 (Sysmex, Kobe, Japan) showed a performance that was at least equivalent to those of the reference methods and superior to those of XN10 (Sysmex, Kobe, Japan). Moreover, the bacterial count obtained with UF4000 (Sysmex, Kobe, Japan) was significantly higher among culture or Direct Gram stain positive samples. We established three optimal cutoff points to predict Direct Gram stain positive samples for peritoneal (465.0 bacteria/μL), synovial (1200.0 bacteria/μL), and cerebrospinal fluids (17.2 bacteria/μL) with maximum sensitivity and negative predictive values. Cell count and detection of bacteria by flow cytometry could be used upstream cytological and microbiological routine procedures to improve and accelerate the diagnosis of infection of biological fluid samples. IMPORTANCE The analysis of biological fluids is crucial for the diagnosis and monitoring of diseases causing effusions and helps in the diagnosis of infectious diseases. The possibility of carrying out cytological and microbiological analyses of biological fluid samples on the same automated machine would simplify the sample circuit (addressing the sample in a single laboratory, 24/7). It would also minimize the quantity of sample required. The performance of cytological and microbiological analysis by the flow cytometer UF 4000 (Sysmex, Kobe, Japan) has never been evaluated yet. This study has shown that bacterial count by flow cytometry using UF4000 (Sysmex, Kobe, Japan) could be used upstream of microbiological routine procedures to improve and to accelerate the diagnosis of infection of biological fluid samples.

Keywords: biological fluids; cell count; detection of bacteria; flow cytometry.

FIG 1

Bland Altman plot of differences between red blood cell counts by UF and standard method. *SD = standard deviation.

FIG 2

Bland Altman plot of differences between white blood cell counts by XN and standard method. *SD = standard deviation.

FIG 3

Bland Altman plot of differences between white blood cell counts by UF and standard method. *SD = standard deviation.

FIG 4

Regression lines between automated (UF et XN) and manual mononuclear cell counts.

FIG 5

Comparison of mononuclear cell counts obtained by XN and standard methods. The Rümke curve was drawn using the Rümke table; it represents the limit value of acceptable coefficients of variation.

FIG 6

Comparison of mononuclear cell counts obtained by UF and standard methods. The Rümke curve was drawn using the Rümke table; it represents the limit value of acceptable coefficients of variation.

FIG 7

ROC curve for bacterial counts by FCM versus DGS positivity for peritoneal fluid samples. AUC = area under the curve.

FIG 8

ROC curve for bacterial counts by FCM versus DGS positivity for synovial fluid samples. AUC = area under the curve.

FIG 9

ROC curve for bacterial counts by FCM versus DGS positivity for cerebrospinal fluid samples. AUC = area under the curve.