You are on page 1of 6

Ann Hematol (2008) 87:557–562 DOI 10.

1007/s00277-008-0464-1

ORIGINAL ARTICLE

Effects of bacteria and yeast on WBC counting in three automated hematology counters
Hye Ryoun Kim & Bo Rae G. Park & Mi Kyung Lee

Received: 2 November 2007 / Accepted: 11 February 2008 / Published online: 27 February 2008 # Springer-Verlag 2008

Abstract Bacteria or yeast may be observed on peripheral blood smears and may lead to spuriously elevated platelet counts. They have been reported to disturb the white blood cell (WBC) differential count if they clumped together, and a large number of such microorganisms have been shown to increase WBC counts. The purpose of this study was to evaluate the spurious rise in WBC counts according to species of microorganisms and automated hematology analyzers. The species we selected were Staphylococcus aureus, Escherichia coli, Candida albicans, C. tropicalis, C. krusei, C. dubliniensis, C. glabrata, and C. parapsilosis. We investigated the effects of bacteria and yeast on peripheral blood samples by the ADVIA 120/2120 Hematology System, Sysmex XE-2100 (TOA Medical Electronics, Kobe, Japan) and Coulter LH 750 (Beckman Coulter, Miami, FL, USA). C. albicans, C. tropicalis, C. krusei, and C. dubliniensis had an overt effect on the WBC count at concentrations of up to 1–5×107 colony-forming units (CFU)/mL in three automated cell counters, and C. glabrata and C. parapsilosis, when present at concentrations of 1–5×108 CFU/mL, caused a significant increase in the WBC count obtained by the Sysmex XE-2100 but not by the ADVIA 120/2120 system and Coulter LH 750 (p< 0.05). In conclusion, yeast may influence the results of peripheral blood smears only when the yeast concentration is unusually high. The results differed among analyzers and among species of yeast. Hematologists should be aware that samples containing bacteria and yeast may give erroneously
H. R. Kim : B. R. G. Park : M. K. Lee (*) Departments of Laboratory Medicine, Chung-Ang University College of Medicine, 65-207, 3-Ka Hangang-Ro Yongsan-Ku, Seoul 140-757, South Korea e-mail: cpworld@cau.ac.kr

high WBC counts and differential leukocyte counts and should review the peripheral blood smear by microscopy. Keywords ADVIA 120/2120 . Sysmex XE-2100 . Coulter LH 750 . Candidemia

Introduction Automated cell counters have a central role in the hematology section of modern clinical laboratories. The use of automated analyzers in hematology laboratories is now the rule rather than the exception. These instruments have enhanced the precision of results and, with optimal quality control measures in the laboratory, have improved the accuracy of tests [1]. Therefore, the vast majority of results reported from most modern hematology laboratories are released directly from automated instruments, without any microscopic review of the blood smear. Most laboratories perform microscopic examination only on samples flagged by the automated cell counter as requiring a manual review. However, there are a variety of conditions where automated parameters may be fictitious, including spuriously elevated platelet counts caused by microspherocytosis or by bacteremia [2]. Branda and Kratz [3] have reported that very high concentrations of C. glabrata and C. parapsilosis significantly cause spuriously elevated platelet counts and white blood cell (WBC) counts by the ADVIA 120/2120 hematology system (Bayer HealthCare, Diagnostics Division, Tarrytown, NY, USA). Thus, we investigated the effects of bacteria and yeast on peripheral blood samples by the ADVIA 120/2120 Hematology system, Sysmex XE-2100 (TOA Medical Electronics, Kobe, Japan), and Coulter LH 750 (Beckman Coulter, Miami, FL, USA). We simulated bacteremia and candidemia by adding known

neutrophils. and two samples with pancytopenia (WBC count. 4. glabrata. In each case. Samples were coded. and S. aureus. Franklin Lakes. differential data. 25922. one sample with thrombocytopenia (WBC count. glabrata. C. resulting in their separation from other forms of WBCs. two channels are used to analyze WBCs: (1) a peroxidase channel in which a peroxidase reagent and light scatter are used to differentiate WBCs by myeloperoxidase content and size and (2) a lobularity/nuclear density channel in which differential WBC lysis is combined with light scatter analysis to determine the WBC count. Red cell lysis is performed using a reagent that selectively suppresses the degranulation of basophils. E. and the technologists were blinded to the meaning of the codes. coli. albicans. 1–5×106. and basophils. Bacteria and yeast Isolates of S. USA). Sysmex XE-2100. NJ. The cells are then categorized according to their side-scattered light and intensity of fluorescence.5×109/L. The Coulter LH 750 uses Volume. 150×109/L). and C. platelet count. Serial tenfold dilutions were prepared from the undiluted stock suspensions. A value of p<0. with four tenfold dilutions in addition to the stock suspension for each yeast or bacterium strain. 1–5×106. platelet count. and scatter (VCS) technology to count and classify leukocytes. conductivity. tropicalis. Highly turbid saline suspensions were prepared from overnight cultures by picking numerous colonies and suspending them in sterile 0. and Coulter LH 750. and values were expressed as mean±SD. Statistical tests were performed using SPSS version 13. krusei. 1–5×107. Simulated candidemia We made 500-μL aliquots with 50 μL of suspensions containing C. the peripheral blood specimens were collected using BD vacutainer K2EDTA plus plastic tubes (Becton Dickinson.9% sodium chloride by the use of an electric vortex. and information about the nuclear maturity of WBCs. platelet count. In the DIFF channel.05 was considered statistically significant. provides information about cell surface characteristics and cell granularity. C. Manassas. C.558 Ann Hematol (2008) 87:557–562 concentrations of two species of bacteria and six species of yeast to the blood and analyzed the specimens with the automated cell counters. Instruments The tested instruments were the ADVIA 120/2120 hematology system. obtained as the cells pass through the helium–neon laser beam. Cells are identified and classified by simultaneous threedimensional analysis using volume. and 22019. dubliniensis at concentrations of 1–5×107. 750. tropicalis. VA. E. krusei. <100×109/L). . monocytes. 6258. C. and C. WBCs are permeabilized to have their deoxyribonucleic acid and ribonucleic acid stain with a fluorescent dye. Materials and methods Subjects For the purpose of performing the WBC differential count and the platelet count. In the immature myeloid information channel. 25923.and sidescattered light information.5× 109/L. The cell density of each undiluted stock suspension was verified by colony counts of the final tenfold dilution on Sabouraud dextrose agar or Sheep blood agar to determine colony-forming units (CFU) per milliliter.5×109/L. C. Volume measured by direct current is used to identify the size of the cells. The Sysmex XE-2100 measures WBCs by flow cytometry using a semiconductor laser to detect forward. parapsilosis at concentrations of 1–5×108. conductivity. 1–5× 105. which were made by diluting them with their own plasma. whereas mature leukocytes are disrupted. respectively. one aliquot represented a negative control sample. 2001. aureus. parapsilosis were acquired from the American Type Culture Collection (ATCC nos. respectively.0. albicans. a special reagent acts on the lipid pattern of the cell membrane to selectively protect immature WBC against disruption. C. In leukocyte differential analysis by the ADVIA 120/ 2120. and light scatter. and C. 14053. C. 70×109/L). 5. MYA 646. containing the same amount of normal saline instead of bacteria or yeast suspension. A fivepart WBC differential is created from the WBC populations of lymphocytes. <2. USA). and 1–5×105 CFU/mL. dubliniensis. eosinophils. coli. Light scatter measurement. Cell count The analyzers were operated by the hospital technologists according to standard laboratory procedures. Statistical analysis Significant increment was defined as WBC count increment greater than 20% compared to specimens without bacteria or yeast. WBC counts for each sample are performed independently in the lobularity/nuclear density channel and the peroxidase channel. We used one sample with normal counts (WBC count. Conductivity or radio frequency measurement provides information about the internal characteristics of the cells. C. and 1–5×104 CFU/mL.

5±1. the spurious effect of WBC counts increased.993. coli (lower row) had no effect on WBC counts in the three automated cell counters . Fig. each largesized particle (greater than the size of platelet) that is not destroyed by hemolytic agents can be identified as a WBC on most hematology analyzers [9. Candida parapsilosis (middle row) caused a significant increase in WBC counts obtained by the Sysmex XE-2100 but not by the ADVIA 120/2120 system and Coulter LH 750 at a concentration of 1–5× 108 CFU/mL. The effects of yeast on WBC counting are shown in Fig 1.957. dubliniensis had an overt effect on WBC counts at concentrations of up to 1–5×107 CFU/mL in three automated cell counters. as a lymphocyte and a large unstained cell by ADVIA 120/ 2120 (concentrations. and Coulter LH 750. Many hematology analyzers that enumerate WBCs generate WBC scattergrams. C. In blood cell analysis. quick and accurate results. for the most parts.6. After enumeration. E. 1–5×107 and 1–5×106 CFU/mL). 10]. C. Yeast was misclassified mainly as a basophil by Sysmex XE-2100 (concentrations. ADVIA 120/2120 system. even at the highest concentration assayed (∼1–5×108 CFU/mL). C.Ann Hematol (2008) 87:557–562 559 Results The presence of bacteria and yeast in blood samples did not affect red blood cell (RBC) counts and platelet counts on the three automated cell counters. 1–5×107 CFU/mL) and as an eosinophil by Coulter LH 750 (concentrations. or peroxidase-staining intensity is used. Although the WBC count and the WBC differential count are possibly erroneous in many instances. when present at a concentration of 1–5×108 CFU/mL. p<0.05). caused a significant increase in WBC counts obtained by the Sysmex XE-2100 (6. C. Each bacterium had no effect on the WBC count and the differential leukocyte count. impedance with lowand high-frequency electromagnetic or direct current. Discussion The widespread use of automated hematology analyzers has led to a major improvement in cellular hematology through. krusei. regardless of the concentration of bacterium and yeast or type of species. tropicalis. albicans.0±1. 1 The presence of Cadida tropicalis (upper row) had an overt effect on WBC counts at concentrations of up to 1–5×107 CFU/mL in the Sysmex XE-2100. 2). either individually or together [11. glabrata and C. parapsilosis.2) but not by the ADVIA 120/2120 system and Coulter LH 750 (3. WBC scattergrams allow the detection of abnormalities related to spurious counts [4–8]. which are the basis for the automated WBC differential count. according to the types of particles. 1–5×107 CFU/mL. Table 1.637. 12]. and C. laser light scattering. Fig. As the number of WBC counts in the samples decreased.110.

560 Table 1 Data of complete blood cell counts in three automated hematology counters Cell counter Sysmex XE-2100 No bacteria or yeast Hemoglobin (g/dL) WBC (×109/L) Platelet (×109/L) Differential count (per μL) Segmented neutrophil Lymphocyte Monocyte Eosinophil Basophil LUC 7. C.30 112 1. tropicalis. glabrata and C. We reviewed peripheral blood smears by microscopy and found that the yeast formed clusters. electrical conductivity/cell volume. containing the same amount of normal saline instead of bacteria or yeast suspension. krusei.0×109/L).922 599 226 583 ADVIA 120/2120 No bacteria or yeast 7. The WBC differential count in the Sysmex XE2100 is based on impedance with direct current. eosinophils. 2 Scattergram generated by a Sysmex XE-2100 hematology analyzer. monocytes. which involves volumetric impedance. blue. green. dubliniensis were bigger than those of C. parapsilosis. is used. C. neutrophils and basophils. glabrata and C. and C. yeast clusters may be shown as large as eosinophils and may increase cytoplasmic granules. VCS technology.285 385 90 367 376 Ann Hematol (2008) 87:557–562 Coulter LH750 No bacteria or yeast 7. red.25 110 1. the instrument flagged a comparison error of the WBC parameter for all samples with falsely elevated WBC counts. parapsilosis affected the WBC count and the differential leukocyte count only at the highest concentration (1–5×108 CFU/mL) and usually by the Sysmex XE-2100 and sometimes by the ADVIA120/2120 (only WBC count<2. and a white cell scatter plot. Because of the discrepant effect on the two WBC channels. parapsilosis as a neutrophil.014 198 28 16 Candidemia 7. Thus. LUC large unstained cell Our results demonstrated that candidemia models of C.044 1. glabrata and C.089 961 213 90 7 61 Candidemia 7.5 2. Our results showed that only the Sysmex XE-2100 had false increments of WBC counts in candidemia of C. krusei. tropicalis.405 954 163 675 WBC White blood cell.2 2. glabrata and C. krusei and C. b A sample from Cadida tropicalis had an overt effect on WBC counts at concentrations of up to 1–5×107 CFU/mL (clusters: sky blue. lymphocytes. and C.3 3.0 5. The majority of these spurious signals were misidentified by the instruments as lymphocytes because they were small and peroxidase negative. C. pink.42 117 1. The methodology used for the differential count varies between different automated counters.2 117 1. ghost) . albicans. albicans.21 102 900 999 210 91 11 Candidemia 7. parapsilosis had a smaller effect on WBC counts than the other Candida species.344 2. dubliniensis was misclassified as a basophil and that of C. Because the sizes of C. and C. The ADVIA 120/2120 uses two WBCcounting channels: the basophil/lobularity channel and the peroxidase channel. Fig. dubliniensis had a significant effect on the WBC count measured by all of the automated cell counters in our simulated candidemia model (concentrations. A variety of sizes create differences in WBC identification. C. which reflects the cell size and radio frequency and the internal structure of cells. In Coulter LH 750. C. C.3 2. tropicalis. 1–5× 107 CFU/mL).48 119 1.6 4. albicans. a A negative control sample. C. The spurious increment of C.922 1. Eosinophils were identified on volumetric impedance (12–14 μm) and scatter plot (high-signal cells) by Coulter LH 750.

Carandente P. regardless of the concentration of bacterium and yeast or type of species. Tschopp C. Genevieve F. our study showed that the presence of microorganisms in blood samples did not affect platelet counts in the three automated cell counters. Bacteria and yeast may induce falsely elevated platelet counts because of their presence in vivo. De Caterina M (2005) Evaluation of the monocyte counting by two automated haematology analysers compared with flow cytometry. Am J Hematol 52:69 . Sysmex XE 2100. 13–15]. For this reason. Pavlovic R. Although it is a rare situation. Furthermore. Yeast may influence the results of peripheral smears only when the concentrations are unusually high. RBCs may not be destroyed by lysis reagents under certain circumstances. unlike the simulated samples presented here. including C. The Candida species may show the same size as platelets and may be observed on peripheral blood smears. Zandecki M. haemoglobin. Pellegrino M. Branda and Kratz [3] have reported that extremely high concentrations (1–5×108 CFU/mL) of C. Kratz A (2006) Effects of yeast on automated cell counting. Since the size of yeast ranged from 2 to 4 μm. However. Subsequent analysis of larger yeast. which can lead to a false increase in WBC counts. In summary. the presence of bacteria or yeast sometimes interferes with the differential leukocyte count regardless of its concentration. Dahinden CA (2005) Flowcytometric analysis of basophil counts in human blood and inaccuracy of hematology analyzers. References 1. Elghetany MT. They have recently been reported as increased platelet counts in thrombocytopenic patients infected by Candida [3. Godon A (2007) Spurious counts and spurious results on haematology analysers: a review. the size of the microorganism clumps was bigger than that of platelets. which incompletely separates the WBC population from neutrophil granulocytes and from RBC remnants. Lee KM. Pseudoeosinophilia sometimes presents with no increment of WBC counts in the Coulter LH 750. However. leaving the platelet count unaffected by bacteria and yeast. The Candida species were mainly misidentified as lymphocytes in a simulated candidemia model. which was similar to that of our study. Genevieve F. spurious results will be flagged by the analyzers for manual review. Bishop CA (1993) A parallel evaluation of four automated hematology analyzers. Chung W (2005) Pseudoeosinophilia associated with malaria infection determined in the Sysmex XE2100 hematology analyzer. All of the three Candida species significantly increased automated WBC counts and in a dose-dependent manner at concentrations of greater than or equal to 1–5×106 CFU/mL. parapsilosis. Int J Lab Hematol 29:4–20 3. some bacteria may be observed on the peripheral blood smear and are associated with positive blood cultures. Am J Clin Pathol 100:626–632 2. Bisogni R. Am J Clin Pathol 126:248–254 4. Am J Clin Pathol 118: 235–241 10. Part II: white blood cells. the inability of the analyzer to distinguish yeast from platelets is readily explained. Ducrest S. Osei ES. Thus. Gerard J. Branda JA. we demonstrated a cause and effect relationship between the presence of bacteria and yeast and the differences of automated hematology analyzers. Yoon H. albicans and C. Lee YK. Clin Lab Haematol 27:91–97 7. Hur M. glabrata and C. Romano MF. which had numerous clumps of bacteria or yeast. Gerard J. Ann Hematol 84:400–402 8.Ann Hematol (2008) 87:557–562 561 glabrata and C. No such effect was observed with C. Part I: platelets. Jung J. Allergy 60:1446– 1450 9. but a linear relationship between yeast concentrations and WBC counts was not observed. Because most microorganisms tended to clump. In addition. Scopacasa F. Hematologists should be aware that samples containing bacteria and yeast may give erroneously high WBC counts and differential leukocyte counts and should review the peripheral blood smear by microscopy. Huh J. it is conceivable that different lysis reagents used in three instruments may give different measurements. and Advia 120: a comparative analysis using the RBC/platelet ratio reference method. The results differ in analyzers and among yeast species. Bentley SA. Hudnall SD (1996) Spurious automated white cell count with Coulter STKS in the myelodysplastic syndromes suggests the presence of a red cell membrane defect. Cho HI (2004) Pseudobasophilia as an erroneous white blood cell differential count with a discrepancy between automated cell counters: report of two cases. even in septic patients. Meyerson HJ (2002) Platelet counting by the Coulter LH 750. Although the presence of yeast can influence the WBC counts generated by automated hematology analyzers. parapsilosis cause spuriously elevated platelet counts. Moreover. parapsilosis. Agrawal NN. the microorganisms were classified as WBCs. did not show the increase. probably because yeast might show atypical light-scattering patterns. which was different from the results of a previous study [3]. The scattergrams generated by the Sysmex XE-2100 hematology analyzer showed an atypical distribution (the increment of RBC debris and neutrophils) of the WBC/baso channel. red cell indices and reticulocytes. red blood cells. Int J Lab Hematol 29:21–41 6. Johnson A. albicans. Grimaldi E. WBCs tended to increase in a dose-dependent manner. another cause might be the number of platelets in mild thrombocytopenic samples (>100×109/L). Sandhaus LM. Godon A (2007) Spurious counts and spurious results on haematology analysers: a review. Kim HJ. Clin Lab Haematol 26:287–290 5. This might be attributed to the result of the moving threshold. Meier F. a linear relationship between yeast concentrations and WBC counts was noted [3]. Zandecki M. Dillman CA. True bacteremia and candidemia usually results in a small number of bacteria and yeast in the blood smear.

Bayle J (2003) Evaluation of the Sysmex Xe-2100 hematology analyzer in hospital use. Br J Haematol 104:1 . Veillon DM. Oberle A. Latif S. Bourner G. Am J Clin Pathol 120:882–885 15. Nakul-Aquaronne D. Curry S. Arnold JA. Cotelingam JD (2003) Spurious automated platelet count. side-by-side study. Kakkar N (2004) Spurious rise in the automated platelet count because of bacteria.562 11. Linscott AJ. Candida glabrata in a blood film. Sudaka-Sammarcelli I. Starck B. commercial laboratory setting: a 4-way. J Clin Lab Anal 17:113– 123 Ann Hematol (2008) 87:557–562 13. Brown D. Kaltenbach J. Enumeration of yeast forms as platelets by the cell-DYN 4000. Lab Hematol 11:285–297 12. Sumner J (2005) Performance evaluation of the latest fully automated hematology analyzers in a large. Ferrero-Vacher C. J Clin Pathol 57:1096–1097 14. Dhaliwal J. Jowzi Z (1999) Images in haematology.