International Journal of Laboratory Hematology
The Official journal of the International Society for Laboratory Hematology
LETTER TO THE EDITOR INTERNAT IONAL JOURNAL OF LABORATO RY HEMATO LOGY
Letter to the Editor
Comparison between automated and microscopic air-dried hematological staining of May-Grunwald). Dif-
analysis in body fluids cytology ferential counts were determined by counting at least
100–300 cells in a magnification of 10009. White blood
Sir, Body fluid cytological analysis is an important tool cell differential counts were expressed in percentage
for the diagnosis and management of several diseases. basis, whereas macrophages, mesothelial cells, and neo-
Although not conclusive, the differential leukocyte plastic cells were reported in a qualitative manner.
counts may suggest the degree of inflammatory response. Validation protocol included studies of precision, line-
For example, a neutrophil predominance is associated arity, and carryover for WBC, RBC, mononuclear cells
with the early phase of the inflammatory response as (MN), and polymorphonuclear cells (PMN). Precision
seen in several infectious and noninfectious diseases. On was obtained from the analysis of 20 replicates of two
the other hand, lymphocyte predominance is associated individual samples (low and high cell count fluids). Line-
with subacute or chronical conditions [1]. arity was performed by a serial dilution of fluid samples
Routinely, determination of cell counts in body fluids with the analyzer diluting fluid (Cellpackâ). Carryover
(BF) is performed by microscopic examination. Conven- effect for WBC and RBC was assessed by testing a high-
tional methods for the cytological analysis require man- level sample consecutively in triplicate (H1, H2, and H3)
ual chamber counting of red and white blood cells followed by a low-level sample (L1, L2, and L3). Carry-
(WBC) and leukocytes differentiation using a cytocentri- over was calculated using the following formula: carry-
fuged and stained preparation. It is a time-consuming over (%) = [(L1–L3)/(H1–L3)] 9100. In the lack of
procedure, requiring at least 60 min to be completed. specific analytical performance goals in BF, we adopted
Besides that, as it is a manual technique, it is prone to a the desirable specifications for total error, imprecision,
subjective interpretation of the cytological findings and bias, derived from intra- and interindividual biologic
resulting in significant interobserver variability. variation [5]. According to this database, the total allow-
A promising alternative could be the automated cell able inaccuracy for erythrocytes, leukocytes, lympho-
analysis of these fluids. However, the use of automated cytes, and neutrophils counts is 4.4%, 14.6%, 16.0%,
hemocytometers in this situation is not a usual and stan- and 22.4%, respectively. For linearity, our goal was a
dardized practice. In 2007, Sysmex Corporation intro- correlation >0.99.
duced a fully automated hematology analyzer with a Diagnostic performance of the automated method was
dedicated body fluid mode (XE-5000), permitting to evaluated by comparison with microscopic analysis. Sim-
overcome these limitations and opening up the possibility ple linear regression (least-square method), paired t-test,
of applying automated analysis of cells in liquid cavity and kappa agreement were used to the statistical analysis.
for diagnostic purposes [2–4]. Our results met the requirements for analytical qual-
To evaluate the performance of the body fluid mode ity regarding precision (CVs < desirable specifications for
on the Sysmex XE-5000, we studied 194 serous fluids imprecision, except for RBC at a very low level) and line-
(108 peritoneal and 86 pleural) after the approval of the arity (r > 0.99). Precision (within-run) ranged from 1.6
Faculty of Medical Sciences Ethics Committee to 15.1% depending on the considered parameter
(# 139634). All samples were sent in an anticoagulant- (Table 1). Linearity for RBC and WBC was excellent
treated tube [ethylenediaminetetraacetic acid (EDTA) or (r > 0.99; P < 0.001). The analytical sensitivity was 1
sodium citrate 3.8%] and analyzed up to 2 h after collec- cell/lL for nucleated cells and 1000/lL for RBC. Carry-
tion. The laboratory analysis included automated total over effect was minimal (<0.1%).
and differential cell counts (Sysmex XE-5000), and man- The automated WBC and RBC counts were highly
ual erythrocytes (RBC) and WBC total and differential correlated with that of the microscopic reference method
counts (Neubauer improved chamber and cytocentrifuged (r ≥ 0.95 in both cases).
e16 © 2014 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem. 2015, 37, e16–e18
LETTER TO THE EDITOR e17
cells reflect the degree of the inflammatory response,
Table 1. Precision of XE-5000
being innumerous in acute process.
LEVEL 1 (low) LEVEL 2 (high) Microscopic analysis is considered the golden standard
for BF cytology. However, this procedure presents some
Cell/lL Mean SD CV Mean SD CV difficulties as it requires skilled personnel, and it is a
time-consuming method with high imprecision and vari-
RBC 16 500 970 5.9% 119 000 2000 2.1%
ability. Besides that, it is not available 24 h/day in most
WBC 169 7 4.1% 7136 116 1.6%
laboratories which can cause delays in reporting the
MN 104 9 9.3% 7143 47 6.6%
results to the clinicians. As BF are generally sent for
PMN 545 60 11% 6340 397 6.3%
urgent analysis, an automated method seems to be the
answer to this issue.
A good correlation between both methods was also The use of automated hematology analyzer without a
observed for mononuclear (r = 0.95) and polymorphonu- dedicated BF mode to enumerate WBCs and RBCs in BF
clear cell counts (r = 0.98). Eosinophils (EOSs) are poses some difficulties. It is well known that most ana-
reported separately as a research parameter and demon- lyzers used in the hematological routine have a func-
strated a good correlation with microscopy (r = 0.94). tional detection limit not suitable to BF cytology. There
Table 2 resumes RBC, WBC, MN, PMN, and EOS counts are other restrictions in using these devices in the body
by both methods. fluid routine as the matrix of BF is quite different from
WBC and MN statistical analysis revealed a difference blood, and presence of other elements as mesothelial or
between both methods (P < 0.05) without clinical tumor cells and macrophages may interfere in the WBC
impact. The agreement between the cell type predomi- counting [2, 9].
nance by both methods was 94% (kappa = 0.86, In 2007, Sysmex Corporation introduced the XE-5000
P < 0.0001). Macrophages, mesothelial cells, or malig- analyzer which is a hybrid hematology analyzer with a
nant cells are reported by the analyzer as high fluores- dedicated BF mode. To increase precision in samples with
cence cells. The presence of these elements can be small numbers of cells, the BF mode has an extended cell
evidenced at the upper border of the DIFF scattergram. counting that allows the passage of three times more
The accurate analysis of BF is of great importance in cells by the detector than previous generation of analyz-
the diagnosis and management of several diseases [2]. ers as the Sysmex XE-2100 [10]. Erythrocytes are enu-
Spontaneous bacterial peritonitis is suspected when merated by flow impedance with hydrodynamic focusing
PMN > 250/lL are present in ascites [6, 7]. Similarly, in technology and do not interfere with the leukocyte count
continuous ambulatory peritoneal dialysis (CAPD) fluid, or differentiation once the fluorescence dye used in the
WBC > 100/lL with PMN ≥ 50% is indicative of perito- XE-5000 only stains cells containing nucleic acid. WBCs,
nitis [8]. Also in pleural fluids, the WBC differential PMNs, and MNs are counted in the DIFF channel using
count provides useful information in the diagnosis and fluorescent flow cytometry. Macrophages, mesothelial
management of patients with pleural effusion [1]. Acute cells, or malignant cells (high fluorescence cells also
inflammatory diseases (e.g., pneumonia) present neutro- called HF-BF) are also counted in the DIFF scattergram
phil predominance in pleural fluid, whereas lymphocytic and excluded from the WBC differential count. The
predominance can be seen in tuberculosis. Mesothelial change from the whole blood mode to BF mode is a sim-
Table 2. Cell counts: comparison between automated and manual methods
RBC/lL WBC/lL MN/lL PMN/lL EOS/lL
XE- XE- XE- XE- Micro- XE-
Microscopy 5000 Microscopy 5000 Microscopy 5000 Microscopy 5000 scopy 5000
Mean 27 920 30 146 690 792 258 368 413 423 14 18
SD 88 812 113 714 1424 1610 703 579 1198 1263 58 78
Median 1420 2000 207 242 95 153 42 54 0 1
Minimum 0 0 3 4 0 2 0 1 0 0
Maximum 726 400 1 187 000 11 360 14 287 8320 8363 11 360 13 144 476 711
RBC, erytrhrocytes; WBC, white blood cells; MN, mononuclear cells; PMN, polimorphonuclear cells; EOS, eosinophils.
© 2014 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem. 2015, 37, e16–e18
e18 LETTER TO THE EDITOR
ple procedure and has an automatic rinse programme to cytologist is not available [2]. The difference observed in
eliminate carryover [4]. This equipment permits the MN percentages is probably due to changes of size and
analyses of an urgent body fluid sample as it easily shape of these cells which are frequently observed in
changes from total blood mode to body fluid mode [3]. these fluids.
As observed by others [2–4], the BF mode on the XE- In summary, automated RBC, WBC, and differential
5000 proved to have the ability to accurately analyze leukocyte counts of pleural and peritoneal fluids show
pleural and peritoneal fluids received in a general labora- good correlation with the manual method. Presence of
tory. Our data showed that XE-5000 is a precise method high fluorescence cells must be investigated by micro-
that certainly exceeds that obtainable by manual cell scopic analysis. Considering that pleural fluids are gener-
counts. Linearity was also excellent, as shown by de Jon- ally sent for urgent analysis, its laboratory routine requires
ge et al. [9] and Williams et al. [4]. It is noteworthy that a skilled personal, and microscopic analysis is not available
high linearity is generally not a limiting factor in fluid 24 h/day in most laboratories, the use of this hematologi-
analysis. cal automated analyzer has the potential of reducing the
We observed good agreement between automated, time to report a preliminary result to the clinician.
and manual RBC and WBC counts in pleural and perito-
neal fluids. The WBC differential counts provided by the P. V. Bottini, D. B. Pompeo, M. I. Souza, C. R. Garlipp
XE-5000 proved to be accurate when compared to
manual differentials performed on cytospin preparations, Division of Clinical Pathology, University of Campinas,
providing a 2-part differential WBC counts (mononuclear Campinas, SP, Brazil
and polymorphonuclear cells). Despite the fact that most E-mail: paula@hc.unicamp.br
laboratories perform a complete differentiation of all cells,
the 2-part differential may support a primary clinical doi: 10.1111/ijlh.12228
decision in an urgent situation or when an experienced
4. Williams JE, Walters J, Kabb K. Gaining cytometric analysis of peritoneal fluid. J Lab
References
efficiency in the laboratory – automated Autom 2013;18:240–4.
1. Connor BD, Lee G, Branca P, Rogers JT, body fluid cell counts: evaluation of the 8. Piraino B, Bailie GR, Bernardini J, Boescho-
Rodriguez RM, Light RW. Variations in body fluid application on the Sysmex XE- ten E, Gupta A, Holmes C, Kuijper EJ, Li PK,
pleural fluid WBC count and differential 5000 hematology analyzer. Lab Med Lye WC, Mujais S, Paterson DL, Fontan MP,
counts with different sample containers 2011;42:395–401. Ramos A, Schaefer F, Uttley L; ISPD Ad Hoc
and different methods. Chest 2003;123: 5. Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Advisory Committee. Peritoneal dialysis-
1181–7. Hernandez A, Jimenez CV, Minchinela J, related infections recommendations: 2005
2. de Jonge R, Brouwer R, de Graaf MT, Lui- Perich C, Simon M. Desirable specifications update. Perit Dial Int 2005;2005:107–31.
twieler RL, Fleming C, de Frankrijker-Mer- for total error, imprecision, and bias, 9. de Jonge R, Brouwer R, van Rijn M, van
kestijn M, Sillevis Smitt PA, Boonstra JG, derived from intra- and inter-individual Acker BA, Otten HJ, Lindemans J. Auto-
Lindemans J . Evaluation of the new body biologic variation. Available at: http://www. mated analysis of pleural fluid total and dif-
fluid mode on the Sysmex XE-5000 for westgard.com/biodatabase1.htm. Accessed on: ferential leukocyte counts with the Sysmex
counting leukocytes and erythrocytes in January, 2014. XE-2100. Clin Chem Lab Med 2006;44:
cerebrospinal fluid and other body fluids. 6. Runyon BA. Management of adult patients 1367–71.
Clin Chem Lab Med 2010;48:665–75. with ascites due to cirrhosis: an update. 10. Perne A, Hainfellner JA, Womastek I,
3. Paris A, Nhan T, Cornet E, Perol JP, Malet Hepatology 2009;49:2087–107. Haushofer A, Szekeres T, Schwarzinger I.
M, Troussard X. Performance evaluation of 7. Lippi G, Cattabiani C, Benegiamo A, Performance evaluation of the Sysmex XE-
the body fluid mode on the platform Sysmex Gennari D, Pavesi F, Caleffi A, Pipitone S. 5000 hematology analyzer for white blood
XE-5000 series automated hematology ana- Evaluation of the fully automated hemato- cell analysis in cerebrospinal fluid. Arch
lyzer. Int J Lab Hematol 2010;32:539–47. logical analyzer Sysmex XE-5000 for flow Pathol Lab Med 2012;136:194–8.
© 2014 John Wiley & Sons Ltd, Int. Jnl. Lab. Hem. 2015, 37, e16–e18
