Skip to main content
NCBI home page
Journal of Clinical Laboratory Analysis logoLink to Journal of Clinical Laboratory Analysis
. 2018 May 25;32(8):e22578. doi: 10.1002/jcla.22578

Automated analysis for differentiating leukocytes in body fluids using the software “biological liquid application” on ADVIA2120/2120i hematology analyzer

Alessandra Falda 1,✉, Paolo Doretto 2
PMCID: PMC6817033  PMID: 29802648

Abstract

Introduction

We evaluated the “Biological liquid application ADVIA2120” software for differentiating the percentage of polymorphonucleated (%PMN) and mononucleated cells (%MN) in ascitic, pleural, and peritoneal dialysis (PD) fluid.

Methods

Biological fluid test results of 193 specimens obtained by automated methods (87 with and 106 without dedicated software) were compared with May‐Grünwald‐Giemsa (MGG) stained blood smears. Limit of detection (LoD) and quantitation (LoQ), repeatability, and inaccuracy were assessed.

Results

Good agreement between the automated methods with dedicated software and the manual method for %PMN and %MN was obtained for leukocyte differentiation in ascitic and pleural fluids, while correlation with the manual method for PD fluid was poor, both with and without the dedicated software.

Conclusions

We demonstrated that the automated differentiation of leukocytes with dedicated software on the ADVIA2120 analyzer for body fluids is a good alternative to the microscopic reference method for peritoneal and pleural specimens, but not for PD fluids.

Keywords: automated hematology analyzer, biological liquid application ADVIA 2120 software, body fluids, differentiation of leukocytes, microscopic method

1. INTRODUCTION

Leukocyte count and differentiation in body fluids offer important information for the diagnosis and treatment of various medical conditions. In peritoneal fluid, a neutrophil count of ≥0.250 × 109/L is a sensitive indicator of spontaneous or secondary bacterial peritonitis.1, 2, 3 The presence of leukocytes (≥0.100 × 109/L; comprised by at least 50% PMN) in the dialysate effluent in PD patients is one of the diagnostic features of peritonitis.4, 5 Most exudative pleural effusions have leukocytes >1 × 109/L; if PMN cells predominate, an acute process affecting the pleural surfaces is diagnosed.3, 6 Currently, microscopic analysis is considered the gold standard for leukocyte differentiation in body fluids, but has several disadvantages: it is imprecise, time‐consuming, requires experienced personnel, and has limited round‐the‐clock availability. Consequently, over the last years, conventional blood cell counters and automated urine analyzers have been evaluated for the quantification of cells in body fluids.7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21 It is not easy to obtain a good result with automated methods for leukocyte differentiation on body fluid samples: interference from mesothelial cells, macrophages, lymphoma cells, and metastasis may lead to incorrect results. In our study, we analyzed 193 body fluid specimens (ascitic, pleural, and PD fluids), divided into two different groups of 106 and 87 samples for the two different evaluations (without and with the “Biological Liquid Application” on ADVIA2120/2120i Hematology Analyzer, respectively), and we compared results with the manual method.

2. MATERIALS AND METHODS

2.1. Reference methods

One hundred and ninety‐three body fluids were harvested into sterile tubes with K3EDTA (ethylene‐diamine‐tetraacetic acid) for routine analysis. We utilized the Fast Read 102® count method instead of a traditional chamber to obtain the WBC absolute count. The Fast Read 102® plastic slide (Biosigma, Dominique Dutscher Company) is a plastic device composed of 10 counting chambers. Each chamber contains a grid with 10 squares. Each square has a dimension of 2 × 5 mm, a depth of 0.1 mm, and a volume of 1 μL.22 The manual method was considered the reference procedure. Repeatability for the automated methods and manual counts was assessed according to the Clinical and Laboratory Standards Institute (CLSI) document EP5‐A3,23 LoD and LoQ according to the CLSI document EP17‐A2,24 and inaccuracy according to the CLSI document H20‐A2.25 We differentiated %PMN and %MN using manual counting as the gold standard with MGG stained blood smears; MGG staining of the cytospin slides (10 minutes at 400 g) was carried out manually. Double‐blind microscopic differentiation was performed on all samples by two physicians who counted 200 leukocytes on each smear, by light microscopy under oil immersion at 40× magnification. MN cells consisted of lymphocytes, monocytes, macrophages, and mesothelial cells, while PMN cells were comprised of neutrophils, eosinophils, basophils, and mast cells.26

2.2. “Biological liquid application” software on ADVIA2120/2120i hematology analyzer

The “Biological Liquid Application” on ADVIA2120/2120i Hematology Analyzer is an in vitro diagnostic test for analyzing the total nucleated cell (TNC) and red blood cell (RBC) count in peritoneal, pleural, and PD fluid samples. The Body Fluid Application Default Run Screen displays 3 cytograms: peroxidase, basophil/lobularity, and RBC scatter cytograms. The TNC count is derived from the basophil/lobularity channel that, like the peroxidase channel cytogram, has been modified for the light scatter characteristics of body fluid cells. The basophil/lobularity cytochemical reaction consists of two steps: red blood cells and platelets are lysed, and all white blood cells except basophils are stripped of their cytoplasm using phthalic acid and a surfactant and the increased temperature in the reaction chamber. The stripped white cells can be categorized as MN or PMN cells based on the size (low‐angle light scatter, 2°‐3°) and complexity of their cell nuclei (high‐angle light scatter, 5°‐15°). As no specific controls for biological fluid samples are available, our daily quality controls were executed using blood‐based products (Siemens Healthcare Diagnostics Inc., USA). All automated cell counts were performed in the open manual (blood) mode. Samples analyzed without the dedicated software were processed with the default run screen for blood samples. All measurements on ADVIA2120 were performed according to the manufacturer’s specifications.27

2.3. Statistical analysis

Comparison studies were carried out by analyzing 193 consecutive patients’ ascites, pleural, or PD fluid samples on ADVIA2120 without pretreatment to calculate LoD, LoQ, repeatability, and inaccuracy. The comparison between manual and automated methods was made by Passing‐Bablok regression analysis. Statistical analyses were performed with MedCalc version 15.8‐©1993‐2015 (MedCalc Software bvba) and Microsoft Office Excel 2007 (Microsoft, Redmond, WA).

2.4. Samples

We analyzed material following routine analysis, and as such informed consent was not required. We analyzed 87 samples as body fluid samples, with dedicated software (39 ascitic, 32 pleural, and 16 PD fluid), and 106 fluids as blood samples, without dedicated software (49 ascitic, 16 pleural fluids, 41 PD fluid samples). We carried out the company’s rinse cycle between each of the analyzed body fluids to minimize carryover from previous analyses. Our study was performed fully respecting the individuals’ rights to confidentiality.

3. RESULTS

Regarding repeatability: using dedicated software, good results were observed for TNC and %MN (range of values: 0.186‐1.666 × 109/L), with CV ranging from 3.17% to 6.66% and 2.32% to 9.07%, respectively, and CVs from 11.40% to 25.81% for %PMN. Without dedicated software, we observed good repeatability for white blood cells (WBC) and %MN (range of values: 0.282‐2.656 × 109/L) with CV ranging from 2.67% to 5.40% and from 2.28% to 7.54%, respectively, CV for %PMN was from 1.65% to 43.19%. With the manual method, leukocytes (range of values: 0.277‐3.492 × 109/L), %MN and %PMN presented a CV ranging from 10.67% to 18.52%, from 0.70% to 4.56% and from 1.76% to 28.86%, respectively. We observed a LoD of 0.027 × 109/L for TNC and 0.020 × 109/L for WBC; and a LoQ of 0.083 × 109/L for TNC and 0.060 × 109/L for WBC. Inaccuracy between the manual method and ADVIA2120 with dedicated software was observed for TNC, MN, and PMN with a mean bias of 0.62%, −11.61%, and 30.16%, respectively; mean bias between the manual method and ADVIA2120 without dedicated software was of  × 8.42%,  × 14.01%, and 17.26% for WBC, MN, and PMN, respectively.

3.1. Peritoneal fluids

Thirty‐nine peritoneal fluids were analyzed for %PMN and %MN with the body fluid software. We considered only fluids with TNC higher than the calculated LoD. Values ranged from 0.058 to 300 × 109/L. 49 peritoneal fluids were analyzed for %PMN and %MN without dedicated software; results ranged from 0.070 to 28.640 × 109/L. Figure 1 and Table 1 show results obtained using Passing‐Bablok analysis.

Figure 1.

Figure 1

Open in a new tab

Passing‐Bablok regression analysis plots for %PMN and %MN of peritoneal fluids determined by manual vs automated method [A, dedicated software; B, blood mode]. Regression line is blue solid line, 95% C.I. is between red dotted lines

Table 1.

Method comparison of manual vs ADVIA2120 count with (*) and without dedicated software for body fluid samples (CI: confidence interval). Data in bold mean statistically significant results

Body fluids Parameter n Mean manual count Mean ADVIA2120 count Intercept 95% CI Slope 95% CI
Peritoneal %PMN 39 20.39 31.63* 8.58 5.57 to 13.25 1.01 0.88‐1.52
%MN 39 79.60 65.64* −15.57 −80.00 to −3.80 1.05 0.92‐1.76
%PMN 49 18.82 27.74 10.10 5.09 to 12.95 1.00 0.85‐1.43
%MN 49 81.17 69.31 −18.49 −67.74 to −2.18 1.06 0.88‐1.62
Pleural %PMN 32 20.96 29.40* 8.61 5.41 to 12.25 1.03 0.83‐1.49
%MN 32 79.03 69.38* −12.73 −55.14 to 3.71 1.03 0.83‐1.49
%PMN 16 31.15 35.83 13.80 10.47 to 17.81 0.67 0.46‐0.90
%MN 16 68.84 61.57 13.50 −10.60 to 23.35 0.72 0.60‐1.00
Dialytic %PMN 16 53.12 52.25* 9.88 −0.45 to 21.73 0.77 0.59‐0.96
%MN 16 46.87 46.00* 9.70 2.91 to 12.34 0.77 0.62‐0.96
%PMN 41 41.97 43.52 14.41 7.94 to 20.75 0.67 0.55‐0.81
%MN 41 58.02 49.48 6.33 −0.98 to 11.58 0.79 0.68‐0.94
Open in a new tab

3.2. Pleural fluids

We analyzed 32 pleural fluids (TNC: 0.114‐18.853 × 109/L) as body fluids and 16 pleural fluids (WBC: 0.160‐ 22.180 × 109/L) without dedicated software for %PMN and %MN; we considered only fluids with elements above the calculated LoD. Results obtained with Passing‐Bablok analysis are illustrated in Figure 2 and Table 1.

Figure 2.

Figure 2

Open in a new tab

Passing‐Bablok regression analysis plots for %PMN and %MN of pleural fluids determined by manual vs automated method [A) dedicated software; B) blood mode]

With dedicated software, a single case of pleural fluid very rich in mesothelial cells (Figure 3) was observed in a patient with significant inflammation. TNC was 2.6 × 109/L. The morphological analysis revealed 5% neutrophils, 22% lymphocytes, 59% monocytes and macrophages, and 14% mesothelial cells.

Figure 3.

Figure 3

Open in a new tab

A case of pleural fluid rich in mesothelial cells (plots of ADVIA2120 with dedicated software)

3.3. Dialysis fluids

We analyzed 16 PD fluid samples as body fluids and 41 samples as blood for calculating %PMN and %MN. As for the other types of fluids, only samples with elements above the calculated LoD were considered. TNC ranged from 0.032 to 35.200 × 109/L and WBC ranged from 0.020 to 23.720 × 109/L. Results obtained with Passing‐Bablok analysis are illustrated in Figure 4 and Table 1.

Figure 4.

Figure 4

Open in a new tab

Passing‐Bablok regression analysis plots for %PMN and %MN of PD fluids determined by manual vs automated method [A, dedicated software; B, blood mode]

4. DISCUSSION

Nowadays there is an increase in the number of body fluid samples that are sent to the laboratory for urgent analysis. The manual method is not the optimal way for processing body fluids because it is time‐consuming and requires experienced personnel. For this reason, numerous automated analyzers developed a way for studying body fluids. The ADVIA2120 analyzer has a dedicated body fluid mode that provides a leukocytes differential count consisting of PMN and MN cell absolute counts and percentages. To our knowledge, this is the first study about leukocyte differentiation using ADVIA2120 dedicated software in ascitic, pleural, and PD fluid. With an automated method, we evaluated 106 samples as blood samples. The body fluid dedicated software was installed at a later time, and with it, the following 87 specimens were analyzed. We observed good performance for TNC, WBC counts, and %MN for repeatability of automated methods, and a narrower CV range for %PMN was obtained with the dedicated software. Leukocytes’ CV obtained with the manual method was higher than with automated methods, based on the high imprecision of the manual procedure. We obtained good LoD results for automated methods, comparable to the declared manufacturer low limit of linearity. For WBC count, we observed a wider inaccuracy than for TNC count. We considered the carryover (<0.3% for TNC and ≤1% for WBC) and the range of linearity as declared by the manufacturer (0.027‐400 × 109/L for the dedicated software and 0.020‐400 × 109/L for blood mode).

In our study there was a good proportionality between both automated methods and manual method for %PMN and %MN in ascites, even considering the constant over‐ and underestimation by software of ADVIA2120 for %PMN and %MN, respectively. Furthermore, there was good proportionality between the results of pleural fluids for leukocyte differentiation obtained with automated and manual methods, but as for peritoneal fluids, there was a small but constant overestimation for %PMN. We know that mesothelial cells usually cause automated flags, and this is more frequent in pleural fluids. Pleural fluids’ results analyzed without dedicated software did not have a good agreement with the manual method’s results. We consider that the constant overestimations by software of ADVIA2120 for %PMN for peritoneal and pleural fluids are not clinically relevant (even if values of %MN are bigger than %PMN, we know that the results of %PMN are clinically more important); we noticed that bigger differences for each paired count were in the lowest absolute count of leukocytes (data not shown). For %PMN and %MN of dialytic fluids, we did not observe a significant agreement between the two automated methods and the manual method. We know that lipids, insufficiently lysed RBC, protein, and cell debris are common interfering factors that can cause mistakes in WBC and/or RBC counts on automated analyzers. It is possible that these kinds of body fluids are more enriched of these factors. This is plausible because peritoneal dialysis fluid is a physiologic synthetic liquid introduced into the peritoneal cavity for normalizing fluid, electrolytes, and solute balance in the body using the principles of ultrafiltration and diffusion, and what we analyze is the lavage fluid that is removed from the cavity.

Aulesa et al. in 2003 and Froom et al. in 2013 considered performances of ADVIA120 and ADVIA2120 for leukocytes differentiation.13, 28 Considering Aulesa et al., results of repeatability of manual methods were comparable. We observed narrower CV ranges for WBC and TNC counts in ADVIA2120 with respect to ADVIA120. CV for %MN were better for ADVIA2120 with and without dedicated software, while CV for %PMN was not improved. We observed that inaccuracy for ADVIA2120 was better than ADVIA120 for leukocytes, but not for %PMN and %MN, and this is not surprising inasmuch as this manuscript highlighted misclassification in PMN and MN counts in some kinds of body fluids, in particular, dialysis fluids. Froom et al. obtained good agreement for neutrophil differential counts for ascitic fluids, even if they analyzed only 5 samples.

A limit of our study is the small number of cases for some types of fluids analyzed; we are also aware that dedicated quality control material is unavailable for body fluids. We observed that the most significantly different %PMN and %MN cells between manual and automated methods were due to the presence of mesothelial cells, lymphoma cells, or rarely nonhematopoietic malignant cells (data not shown). It was crucial to evaluate these samples by microscope. Moreover, abnormal test results or plots were verified with microscopic revision. In conclusion, we demonstrate that the body fluid dedicated software of the ADVIA2120 shows a good agreement for leukocyte differentiation for ascitic and pleural fluids compared to the manual method (“gold standard”), but not for dialysis fluids. Furthermore, automated methods eliminate the wide interobserver variability seen in manual microscopy, it reduces the number of samples submitted for microscopy and is available 24 hours a day.

Supporting information

 

Click here for additional data file. (84.5KB, doc)

 

Click here for additional data file. (133KB, doc)

Falda A, Doretto P. Automated analysis for differentiating leukocytes in body fluids using the software “biological liquid application” on ADVIA2120/2120i hematology analyzer. J Clin Lab Anal. 2018;32:e22578 10.1002/jcla.22578

REFERENCES

  • 1. Møller S, Henriksen JH, Bendtsen F. Ascites: pathogenesis and therapeutic principles. Scand J Gastroenterol. 2009;44:902‐911. [DOI] [PubMed] [Google Scholar]
  • 2. Moore KP, Wong F, Gines P, et al. The management of ascites in cirrhosis: report on the consensus conference of the International Ascites Club. Hepatology. 2003;38:258‐266. [DOI] [PubMed] [Google Scholar]
  • 3. Kjeldsberg CL, Knight JA. 1986. Body fluids. American Society of Clinical Pathologists Press, Chicago. [Google Scholar]
  • 4. Xu R, Chen Y, Luo S, et al. Clinical characteristics and outcomes of peritoneal dialysis‐related peritonitis with different trends of change in effluent white cell count: a longitudinal study. Perit Dial Int. 2013;33:436‐444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Kam‐Tao Li P, Szeto CC, Piraino B, et al. Peritoneal dialysis‐related infections recommendations: 2010 Update. Perit Dial Int. 2010;30:393‐423. [DOI] [PubMed] [Google Scholar]
  • 6. de Jonge R, Brouwer R, van Rijn M, van Acker BAC, Otten HJAM, Lindemans J. Automated analysis of pleural fluid total and differential leukocyte counts with the Sysmex XE‐2100. Clin Chem Lab Med. 2006;44:1367‐1371. [DOI] [PubMed] [Google Scholar]
  • 7. Manoni F, Tinello A, Fornasiero L, et al. Urine particle evaluation: a comparison between the UF‐1000i and quantitative microscopy. Clin Chem Lab Med. 2010;48:1107‐1111. [DOI] [PubMed] [Google Scholar]
  • 8. Block DR, Algeciras‐Schimnich A. Body fluid analysis: clinical utility and applicability of published studies to guide interpretation of today’s laboratory testing in serous fluids*. Crit Rev Clin Lab Sci. 2010;50:107‐124. [DOI] [PubMed] [Google Scholar]
  • 9. Fleming C, Brouwer R, van Alphen A, Lindemans J, de Jonge R. UF‐ 1000i: validation of the body fluid mode for counting cells in body fluids. Clin Chem Lab Med. 2014;52:1781‐1790. [DOI] [PubMed] [Google Scholar]
  • 10. Riedl JA, Dinkelaar RB, van Gelder W. Automated morphological analysis of cells in body fluids by the digital microscopy system DM96. J Clin Pathol. 2010;63:538‐543. [DOI] [PubMed] [Google Scholar]
  • 11. Bourner G, De La Salle B, George T, et al. ICSH guidelines for the verification and performance of automated cell counters for body fluids. Int J Lab Hematol. 2014;36:598‐612. [DOI] [PubMed] [Google Scholar]
  • 12. Castellone D, Peerschke EB, Francisco N, Canfield W, Kling G. Accuracy and precision study: body fluid white blood cell (WBC) analysis (peritoneal, pleural and peritoneal dialysate) sing a light scatter technology (ADVIA®2120/2120i) versus hemocytometer manual counts. Blood. 2010;116:4730. [Google Scholar]
  • 13. Aulesa C, Mainar I, Prieto M, Cobos N, Galimany R. Use of the Advia120 hematology analyzer in the differential cytologic analysis of biological fluids (cerebrospinal, peritoneal, pleural, pericardial, synovial, and others. Lab Hematol. 2003;9:214‐224. [PubMed] [Google Scholar]
  • 14. Paris A, Nhan T, Cornet E, Perol JP, Malet M, Troussard X. Performance evaluation of the body fluid mode on the platform Sysmex XE‐5000 series automated hematology analyzer. Int J Lab Hematol. 2010;32:539‐547. [DOI] [PubMed] [Google Scholar]
  • 15. De Smet D, Van Moer G, Martens GA, et al. Use of the Cell‐Dyn Sapphire hematology analyzer for automated counting of blood cells in body fluids. Am J Clin Pathol. 2010;133:291‐299. [DOI] [PubMed] [Google Scholar]
  • 16. Lippi G, Cattabiani C, Benegiamo A, et al. Evaluation of white blood cell count in peritoneal fluid with five different hemocytometers. Clin Biochem. 2013;46:173‐176. [DOI] [PubMed] [Google Scholar]
  • 17. de Jonge R, Brouwer R, de Graaf MT, et al. Evaluation of the new body fluid mode on the Sysmex XE‐5000 for counting leukocytes and erythrocytes in cerebrospinal fluid and other body fluids. Clin Chem Lab Med. 2010;48:665‐675. [DOI] [PubMed] [Google Scholar]
  • 18. Buoro S, Gustinetti R, Dominoni P, et al. Analytical evaluation of Sysmex UF‐1000i for flow cytometric analysis of peritoneal fluid. Clin Biochem. 2012;45:1263‐1265. [DOI] [PubMed] [Google Scholar]
  • 19. Danise P, Maconi M, Rovetti A, et al. Cell counting of body fluids: comparison between three automated haematology analysers and the manual microscope method. Int J Lab Hematol. 2013;35:608‐613. [DOI] [PubMed] [Google Scholar]
  • 20. Harris N, Kunicka J, Kratz A. The ADVIA 2120 hematology system: flow cytometry based analysis of blood and body fluids in the routine hematology laboratory. Lab Hematol. 2005;11:47‐61. [PubMed] [Google Scholar]
  • 21. Shu G, Lu H, Du H, Shi J, Wu G. Evaluation of Mindray BC‐3600 hematology analyzer in a university hospital. Int J Lab Hematol. 2013;35:61‐69. [DOI] [PubMed] [Google Scholar]
  • 22. Gunetti M, Castiglia S, Rustichelli D, et al. Validation of analytical methods in GMP: the disposable Fast Read 102W device, an alternative practical approach for cell counting. J Transl Med. 2012;10:112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. McEnroe RJ, Durham AP, Goldford MD, et al. Clinical and Laboratory Standards Institute. Evaluation of Precision of Quantitative Measurement Procedures; Approved Guideline—Third Edition. CLSI document PE5‐A3. Clinical and Laboratory Standards Institute 2014; 1‐56238‐967‐X.
  • 24. Person‐Perry J, VaksBarnes JE, Durham AP, et al. Clinical and Laboratory Standards Institute. Evaluation of Detection Capability for Clinical Laboratory Measurement Procedures; Approved Guideline – Second Edition. CLSI document EP17‐A2. Clinical and Laboratory Standards Institute 2012; 1‐56238‐795‐2.
  • 25. Koepke JA, Van Assendelft OW, Brindza LJ, et al. Clinical and Laboratory Standards Institute. Reference Leukocyte (WBC) Differential Count (Proportional) and Evaluation of Instrumental Methods; Approved Standard—Second Edition. CLSI document H20‐A2. Clinical and Laboratory Standards Institute 2007; 1‐56238‐628‐X.
  • 26. Szamosi DI, Bautista JM, Cornbleet J, et al. Clinical and Laboratory Standards Institute. Body Fluid Analysis for Cellular Composition; Approved Guideline. CLSI document H56‐A. Clinical and Laboratory Standards Institute 2006; 1‐56238‐614‐X.
  • 27. Siemens . Applicazione liquidi biologici ADVIA2120/2120I, Rev.A, 2009‐01.
  • 28. Froom P, Diab A, Barak M. Automated Evaluation of Synovial and Ascitic Fluids With the Advia 2120 Hematology Analyzer. Am J Clin Pathol. 2013;140:828‐830. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

 

Click here for additional data file. (84.5KB, doc)

 

Click here for additional data file. (133KB, doc)

Articles from Journal of Clinical Laboratory Analysis are provided here courtesy of Wiley

RESOURCES