performance evaluation of the advia 2120 hematology analyzer

ABSTRACT

Automated cell counters are widely used in modernclinical laboratories to provide reliable, fast, and cost-effective complete blood counts (CBCs), white blood celldifferentials, and reticulocyte measurements. In addition,some advanced instruments provide novel parameters,such as the hemoglobin content of reticulocytes or the per-centage of hypochromic cells, and are capable of analysisof a variety of body fluids. Bayer recently introduced theADVIA 2120 system as an automation-ready cell counterfor mid- to high-volume testing in the clinical laboratory.This instrument, which builds on the established technol-ogy of the ADVIA 120 system, operates with a cyanide-free method for hemoglobin measurement, has a new userinterface, and can routinely analyze biological fluid sam-ples in addition to blood. We used 749 samples from 6worldwide trial sites to evaluate the clinical performance ofthis new device. Accuracy of the ADVIA 2120 system ver-sus its predecessor model, the ADVIA 120 system, wasexcellent for all CBC and white cell differential parametersand reticulocyte counts (all correlation coefficients exceptfor basophils >0.9). Correlation of the white cell differen-tial with the standard manual method and within-run pre-cision of the ADVIA 2120 system also was very good. Useof the novel cyanide-free method for hemoglobin measure-ment had no clinically significant impact on hemoglobinresults, even in patients with hemoglobinopathies. We

concluded that the ADVIA 2120 system has clinicallyequivalent performance to the ADVIA 120 system.Lab Hematol. 2005;11:62-70.

KEY WORDS: Cell counters · Hematology analyzers ·ADVIA 2120 system · ADVIA 120system · Automated cell analysis ·Performance evaluation

INTRODUCTION

Complete blood cells counts (CBCs) and white blood cell(WBC) differentials are among the most frequently ordered labo-ratory tests. Historically these assays were manually performed,labor intensive, and associated with large coefficients of variation.Progress in the design of laboratory instrumentation has made itpossible to replace the manual performance of the vast majorityof these tests with highly automated assays. These laboratoryinstruments have been shown to provide reliable clinical resultswith a much shorter turnaround time, lower cost, and higherreproducibility than the manual approach [1]. In addition, theseanalyzers can generate a large number of novel parameters thatare not achievable through the manual method and are increas-ingly also used for cell counting in fluids other than blood [2,3]and even innovative applications such as determination of fetallung maturity [4]. Such trends continue to dominate the field.Newer models add further automation by a direct interface witha specimen management system and/or a slide maker-stainer,more user-friendly and clinically reliable software components,and increased awareness of environmental concerns.

The ADVIA 120 Hematology System (Bayer HealthCare,Diagnostics Division, Tarrytown, NY, USA) is a well-

Performance Evaluation of the ADVIA 2120 HematologyAnalyzer: An International Multicenter Clinical Trial

NEIL HARRIS,1 JOSEP MARIA JOU,2 GIANLUIGI DEVOTO,3 J. LOTZ,4 JAMES PAPPAS,5

DAVID WRANOVICS,1 MARY WILKINSON,1 SUZANNE R. FLETCHER,6 ALEXANDER KRATZ6

1University of Florida/Shands Hospital Core Laboratory, Gainesville, Florida, USA; 2Servei d´Hemoteràpia i Hemostàsia, CoreLaboratory, Hospital Clínic, University of Barcelona, Spain;3 Laboratorio Analisi Ospedale Civile, Lavagna, Italy; 4JohannGutenberg University Central Laboratory, Mainz, Germany; 5Loma Linda University Medical Center, San Bernardino, California,USA; 6Department of Pathology, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA

Received August 20, 2004; accepted August 23, 2004

Laboratory Hematology 11:62-70© 2005 Carden Jennings Publishing Co., Ltd.doi: 10.1532/LH96.04064

62

Correspondence and reprint requests: Dr. Alexander Kratz, ClinicalHematology Laboratory, Division of Laboratory Medicine, Massachu-setts General Hospital, 55 Fruit St, GRJ 235, Boston, MA 02114,USA (e-mail: [email protected]).

ISLH

Official Publication

established modern cell counter for blood and biologicalfluid analysis that is used in clinical laboratories worldwide.The system has 2 methods of measuring hemoglobin: amodification of the manual cyanmethemoglobin methodand a cell-based hemoglobin measurement [5]. Red bloodcells and platelets are analyzed on a cell-by-cell basis with 2angles of light scatter. This innovative method allows report-ing of novel parameters, such as percentage of hypochromicred cells [6] and mean platelet component (MPC), and pro-vides increased accuracy in the enumeration of platelets [7-8].Myeloperoxidase staining, cell size, and nuclear lobulationare used to obtain a 6-part differential consisting of neu-trophils, lymphocytes, monocytes, eosinophils, basophils,and large unstained cells (LUC). Staining with oxazine 750allows enumeration of reticulocytes and, in combinationwith light-scatter analysis, reporting of the hemoglobin con-tent of reticulocytes, a US Food and Drug Administration–cleared marker of iron metabolism [9-12].

The ADVIA 2120 system was recently introduced as a newhematology analyzer designed for mid- to high-volume testingin the routine clinical laboratory. Building on technology suc-cessfully used in the ADVIA 120 model, the ADVIA 2120system operates with a cyanide-free hemoglobin method[13,14], has a new user interface, has more sensitive criteria forthe enumeration of blasts and monocytes, is automation ready,and can be used for routine analysis of biological fluid samplesin addition to peripheral blood [3]. To assess the analyticalperformance of this new instrument, we conducted an inter-national multicenter clinical trial comparing its performancewith that of the ADVIA 120 system and with results of theNational Committee for Clinical Laboratory Standards(NCCLS) reference manual method for WBC differential.

MATERIALS AND METHODS

Study SitesThe study was performed at 6 sites: Johann Gutenberg

University Central Laboratory, Mainz, Germany; Laborato-rio Analisi Ospedale Civile, Lavagna, Italy; Loma Linda Uni-versity Hospital, San Bernardino, California, USA; Massa-chusetts General Hospital, Boston, Massachusetts, USA;University of Florida/Shands Hospital Core Laboratory,Gainesville, Florida, USA; and Core Laboratory, HospitalClinic, University of Barcelona, Spain.

Instrument SoftwareThe ADVIA 2120 system was run with software version

4.05 with the sensitive flagging option selected for data col-lection. The ADVIA 120 systems were run with software ver-sion 2.2.2a, 3.0.1, or 3.0.12. None of the ADVIA 120 sys-tems had the sensitive flagging software installed.

Manual Differentials (Reference Method)Manual differentials were performed on blood smears

obtained at 5 of the study sites. A total of 593 blood smearswere evaluated. For each smear, 2 independent technolo-gists performed a 200-cell WBC differential according tothe H-20 NCCLS guidelines [15].

Study ProtocolThe study was performed over 5 days at each site.

Instruments were operated by hospital employees trainedin the use of the ADVIA 120 and ADVIA 2120 systems.Control materials were assayed, and a 20-sample precision

Performance Evaluation of the ADVIA 2120 63

TABLE 1. Accuracy of the ADVIA 2120 versus the ADVIA 120*

Parameter r Slope Intercept Syx ADVIA 120 Mean ADVIA 2120 Mean Bias

WBC 0.997 0.97 0.06 0.49 8.15 8.05 –0.10

RBC 0.996 0.98 0.03 0.08 4.03 3.98 –0.05

Hgb 0.997 0.99 0.1 0.17 12.0 12.0 0.0

MCV 0.979 0.99 1.2 1.6 90.2 90.2 0.0

RDW 0.989 0.99 –0.1 0.3 15.1 14.8 –0.3

HDW 0.986 0.94 0.13 0.09 2.90 2.85 –0.05

Plt 0.992 0.96 4 15 246 240 –6

% Neutrophils 0.996 0.98 0.8 1.5 66.2 65.8 –0.4

% Lymphocytes 0.990 0.98 0.5 1.9 23.0 22.9 0.1

% Monocytes 0.939 1.01 0.8 1.0 5.4 6.2 0.8

% Eosinophils 0.980 0.98 0.0 0.5 2.6 2.6 0.0

% Basophils 0.772 0.79 0.1 0.3 0.5 0.5 0.0

% LUC 0.945 1.07 –0.5 1.2 2.4 2.0 –0.4

% Reticulocytes 0.962 0.91 0.2 0.6 2.2 2.1 –0.1

*Syx indicates standard error of the estimate;WBC, white blood cell count; RBC, red blood cell count; Hgb, hemoglobin; MCV, mean corpuscular volume;RDW, red cell distribution width; HDW, hemoglobin distribution width; Plt, platelets; LUC, large unstained cells.

64 N. Harris et al

FIGURE 1. Correlation of complete blood count and reticulocyte results from the ADVIA 2120 system compared with results fromthe ADVIA 120 system. WBC indicates white blood cells; Syx, standard error of the estimate; RBC, red blood cells; Hgb, hemo-globin; MCV, mean corpuscular volume; RDW, red cell distribution width; HDW, hemoglobin distribution width; Plt, platelets.


FIGURE 2. Correlation of white blood cell differential results from the ADVIA 2120 system compared with white cell differential resultsfrom the ADVIA 120 system. LUC indicates large unstained cells.

66 N. Harris et al

FIGURE 3. Correlation of white blood cell differentials from the ADVIA 2120 system compared with manual differentials. The plots includethe binomial envelope surrounding the manual counts as recommended by National Committee for Clinical Laboratory Standards.

run with whole blood was performed each day at eachsite. Twenty-five randomly selected patient samples wereassayed each day with both the ADVIA 2120 and theADVIA 120 systems in CBC/Diff mode for a total of 125samples at each site, for an overall total of 750 samples.One of these samples was removed from analysis becauseit was a short sample on the ADVIA 120 system. As aresult, the sample size for the accuracy study was 749samples. Of these 749 samples, 575 were also analyzed inreticulocyte mode. A reliability log for the ADVIA 2120system was kept at each site.

Statistical AnalysisRegression analysis was performed to compare the results

between the ADVIA 2120 system, the ADVIA 120 system,and the reference method. Binomial envelopes surroundingthe manual counts were also shown for the manual counts, asrecommended by NCCLS H-20A [15].

RESULTS

Accuracy of the ADVIA 2120 System Compared with theADVIA 120 System

Agreement between the ADVIA 2120 and the ADVIA120 systems for CBC, WBC differential, and reticulocytecounts was determined on 749 samples. Regression analy-sis results are shown in Table 1. The x,y regression plotsare shown in Figures 1 and 2. Correlation was excellentfor all CBC parameters, all correlation coefficients beinghigher than 0.97. The correlation for platelet count wasexcellent even in the clinically crucial low platelet countrange (r = 0.969 for platelet counts less than 50,000).Correlation also was very good for the WBC differentialand reticulocyte results. Correlation was lowest for thecell types occurring in low numbers (eg, basophils). Therewas also a slightly lower correlation for monocytes (r = 0.939)than for other differential parameters.


TABLE 2. Accuracy of the ADVIA 2120 versus Manual WhiteBlood Cell Differential*

ADVIA

Manual 2120

Parameter r Slope Intercept Syx Mean Mean

% Neutrophils 0.948 0.90 7.3 5.1 65.4 66.1

% Lymphocytes 0.953 0.91 2.5 4.4 24.6 24.9

% Monocytes 0.693 0.47 2.7 1.9 7.1 6.1

% Eosinophils 0.877 0.82 0.6 1.2 2.3 2.5% Basophils 0.311 0.18 0.4 0.4 0.5 0.5

*Lymphocyte count of the ADVIA 2120 system includes large unstainedcells (LUCs). Syx indicates standard error of the estimate.

TABLE 3. Within-Run Precision of Complete Blood Cell CountParameters on the ADVIA 2120 System*

Observed

Site Mean SD CV, %

White blood cell count, × 103/µL1 5.75 0.14 2.42 5.02 0.09 1.83 6.76 0.15 2.24 11.96 0.20 1.75 8.32 0.13 1.66 5.73 0.13 2.3

Red blood cell count, × 106/µL1 4.81 0.03 0.62 4.76 0.03 0.63 4.47 0.04 0.94 5.10 0.03 0.65 4.41 0.03 0.76 4.45 0.04 0.9

Hemoglobin, g/dL1 14.8 0.07 0.52 13.9 0.06 0.43 13.2 0.08 0.64 15.2 0.08 0.55 13.1 0.07 0.56 13.4 0.08 0.6

Mean corpuscular volume, fL1 86.5 0.15 0.22 87.9 0.12 0.13 84.4 0.25 0.34 85.5 0.11 0.15 89.5 0.27 0.36 88.5 0.21 0.2

Cell hemoglobin concentration mean, g/dL1 35.6 0.08 0.22 34.3 0.06 0.23 35.1 0.13 0.44 35.9 0.06 0.25 33.4 0.14 0.46 34.7 0.08 0.2

Red cell distribution width, %1 12.6 0.07 0.62 13.2 0.07 0.53 13.4 0.07 0.54 12.9 0.06 0.55 13.3 0.06 0.56 15.4 0.09 0.6

Hemoglobin distribution width, g/dL1 2.54 0.01 0.42 2.61 0.01 0.43 2.74 0.02 0.74 2.66 0.01 0.45 2.29 0.01 0.46 3.04 0.02 0.7

Continued

68 N. Harris et al

Accuracy of the ADVIA 2120 System versus Manual Differential

WBC differential results from 593 samples obtained at5 sites and determined with the ADVIA 2120 system werecompared with 200-cell manual differentials performed by2 technologists. For the purpose of this comparison, theLUC counts of the ADVIA 2120 system were added to sys-tem lymphocyte count and compared with the manual lym-phocyte count. Accuracy statistics are shown in Table 2, andplots of the results are shown in Figure 3. As for correlationwith the ADVIA 120 system, correlation of the automateddifferential results from the ADVIA 2120 system with themanual counts was very good.

ADVIA 2120 System Within-Run PrecisionResults from within-run replicate analysis of blood sam-

ples on the ADVIA 2120 system are shown in Tables 3 and4. A total of 600 CBCs, white cell differentials, and reticulo-cyte counts were performed on 30 samples. Results from the 5runs at each site were pooled to give a summary for each site.Within-run precision for the ADVIA 2120 system was verygood and met the ADVIA 120 system specifications. Asexpected, cell types that occur in lower numbers in periph-eral blood had higher coefficients of variation.

Samples from Patients with HemoglobinopathiesThe ADVIA 2120 system operates with a novel cyanide-

free method for measurement of hemoglobin level [13]. Toascertain whether this method was reliable for measurementof hemoglobin in patients with congenital hemoglo-binopathies, we used the ADVIA 120 system with the stan-dard cyanide-dependent method and the ADVIA 2120 sys-tem with the cyanide-free method to analyze samples frompatients with sickle cell trait and other structural and quan-titative abnormalities of hemoglobin synthesis. No clinicallysignificant difference in hemoglobin levels was found withthe 2 methods (Table 5).

TABLE 4. Within-Run Precision of White Cell Differential Parametersand Reticulocyte Counts on the ADVIA 2120 System*

Observed

Site Mean SD CV, %

% Neutrophils1 55.9 0.73 1.32 55.9 0.81 1.43 57.0 0.65 1.14 71.8 0.52 0.75 66.0 0.59 0.96 65.2 0.79 1.2

% Lymphocytes1 32.4 0.74 2.32 31.7 0.77 2.43 30.9 0.82 2.74 19.2 0.52 2.75 24.0 0.59 2.56 23.6 0.66 2.8

% Monocytes 1 6.7 0.42 6.32 6.7 0.41 6.13 6.4 0.38 5.94 5.4 0.28 5.25 6.2 0.33 5.36 6.7 0.49 7.3

% Eosinophils 1 2.6 0.28 10.82 2.6 0.20 7.73 2.9 0.26 9.04 1.7 0.15 8.85 1.8 0.17 9.46 2.8 0.26 9.3

% Basophils 1 0.8 0.14 17.52 0.9 0.14 15.63 0.6 0.44 73.34 0.3 0.07 23.35 0.5 0.07 14.06 0.6 0.11 18.3

% Large unstained cells 1 1.6 0.26 16.32 2.3 0.30 13.03 1.5 0.23 15.34 1.5 0.22 14.75 1.6 0.23 14.46 1.1 0.22 20.0

% Reticulocytes 1 1.7 0.13 7.62 1.4 0.09 6.43 2.1 0.14 6.74 1.4 0.09 6.45 1.9 0.15 7.96 2.7 0.18 6.7

*Results from the 5 runs at each site are pooled to give a summary foreach site. 1 indicates Johann Gutenberg University Central Laboratory,Mainz, Germany; 2, Laboratorio Analisi Ospedale Civile, Lavagna, Italy; 3, LomaLinda University Hospital, San Bernardino, California, USA; 4, MassachusettsGeneral Hospital, Boston, Massachusetts, USA; 5, University of Florida/ShandsHospital Core Laboratory, Gainesville, Florida, USA; 6, Core Laboratory, Hos-pital Clinic, University of Barcelona, Spain; CV, coefficient of variation.

TABLE 3—Continued

Platelets, × 103/µL1 241 6.7 2.82 249 5.3 2.13 257 6.8 2.64 253 6.3 2.55 302 8.0 2.66 203 5.6 2.8

*Results from the 5 runs at each site were pooled to give a summary foreach site. 1 indicates Johann Gutenberg University Central Laboratory,Mainz, Germany; 2, Laboratorio Analisi Ospedale Civile, Lavagna, Italy; 3, LomaLinda University Hospital, San Bernardino, California, USA; 4, MassachusettsGeneral Hospital, Boston, Massachusetts, USA; 5, University of Florida/ShandsHospital Core Laboratory, Gainesville, Florida, USA; 6, Core Laboratory,Hospital Clinic, University of Barcelona, Spain; CV, coefficient of variation.

System ReliabilityTwo reliability incidents on the ADVIA 2120 system

were reported as part of the study. In 1 incident, 2 minutesof downtime were recorded for changing a clot filter. Severalclotted samples had been run on the ADVIA 2120 system,and the clot filter was changed as a precautionary measure.In the other incident, a leak was observed in the RBCreagent area when the instrument was set up prior to thestudy. It was found that the sheath reagent line had discon-nected from its reagent pump. No downtime was recorded,because the study had not yet started.

DISCUSSION

We evaluated the accuracy, precision, and reliability ofthe ADVIA 2120 automated cell counter in an internationalmulticenter clinical trial. Our data indicated very good cor-relation between CBC, white cell differential, and reticulo-cyte counts obtained with the ADVIA 2120 system andresults obtained with the ADVIA 120 system, an instrumentwith a well-established clinical track record and years of suc-cessful use in a large number of laboratories worldwide.Most correlation coefficients for reported parameters weregreater than 0.95. The slightly lower correlation coefficientfor monocytes (0.939) was most likely due to the use ofnew, more sensitive flagging software on the ADVIA 2120system. In this software, the threshold between monocytesand neutrophils has been changed, leading to slightly (1%)increased recovery of monocytes compared with previoussoftware versions. The lower correlation coefficient forbasophils (0.772) was most likely caused by the very lownumbers in which these cells occur in blood. An even lowercorrelation coefficient for basophils was recently reported byKaplan and co-workers for a different line of cell counters[16]. Correlation of the white cell differentials obtained onthe ADVIA 2120 system with results of the referencemethod, manual 200-cell counts by 2 technologists, wasalso very good. Within-run precision of the ADVIA 2120

system was comparable with the performance of the ADVIA120 system.

The ADVIA 2120 system operates with a novel cyanide-free method of measuring hemoglobin. Interference ofabnormal hemoglobins with results from automated cellcounters has been reported in the literature [17]. We wereconcerned that abnormal hemoglobins, such as hemoglobinS, E or C, might give different measurements on theADVIA 2120 system. Our results indicated that at least forthe abnormal hemoglobins we investigated, the cyanide-freemethod gave results clinically equivalent to those obtainedwith the standard ADVIA 120 method.

In summary, our findings indicated that the performanceof the ADVIA 2120 system is clinically equivalent to that ofthe ADVIA 120 system, its predecessor model. In addition tothe proven clinical track record of its predecessor model, thenew instrument provides a cyanide-free method of hemoglo-bin determination, a more user-friendly software interface, agating system more sensitive to the presence of blasts, theability to analyze biological fluids as a standard feature, andcapability for interface with an automated slide maker-stainerand a laboratory automation system. These features shouldmake this instrument attractive to medium- to high-volumelaboratories looking for a modern cell counter capable of pro-viding clinically reliable CBC and differential results as well asnovel parameters and fluid analysis, either as a stand-alonesystem or as part of a laboratory automation system.

ACKNOWLEDGMENTS

This study was supported by a grant from Bayer HealthCare,Diagnostics Division, Tarrytown, NY, USA.

The authors also acknowledge the support of the R&DHematology Department of Bayer HealthCare, especially Dr.Jolanta Kunicka, William Canfield, and Gail Kling.

The investigators gratefully acknowledge the invaluable effortsof the technical and administrative staff at their institutions. Thiswork would not have been possible without their contributions.


TABLE 5. Hemoglobin Determinations in Patients with Hemoglobinopathies*

Hemoglobinopathy Hgb on ADVIA 120, g/dL† Hgb on ADVIA 2120, g/dL‡

Sickle cell trait (Hgb A, 56.4%; Hgb S, 38.5%) 13.4 13.5



Hgb E trait and α-thalassemia trait (Hgb A, 73.6%; Hgb E, 22.9%) 12.1 12.2

Hgb C trait and α-thalassemia trait (Hgb A, 69.0%; Hgb C, 28.2%) 12.1 11.9

Sickle cell disease, transfused (Hgb A, 40.6%; Hgb S, 27.8%; Hgb C, 27.1%) 8.3 8.4Sickle cell disease, transfused (Hgb A, 65.1%; Hgb S, 17.6%; Hgb C, 17.3%) 11.4 11.7

*Values for hemoglobin F and A2 are omitted. Hgb indicates hemoglobin.†Cyanide-containing Hgb method.‡Cyanide-free Hgb method.

70 N. Harris et al

REFERENCES

1. Simson E, Groner W. The state of the art for the automated WBCdifferential. Lab Hematol. 1995;1:13-22.

2. Aulesa C, Mainar I, Prieto M, Cobos N, Galimany R. Use of theAdvia 120 hematology analyzer in the differential cytologic analy-sis of biological fluids (cerebrospinal, peritoneal, pleural, pericar-dial, synovial, and others). Lab Hematol. 2003;9:214-224.

3. Aune MW, Becker JL, Brugnara C, et al. Automated flow cyto-metric analysis of blood cells in cerebrospinal fluid. Am J ClinPathol. 2004;121:690-700.

4. Chapman JF, Ashwood ER, Feld R, Wu AH. Evaluation of two-dimensional cytometric lamellar body counts on the ADVIA 120hematology system for estimation of fetal lung maturation. ClinChim Acta. 2004;340:85-92.

5. Kunicka J, Malin M, Zelmanovic D, et al. Automated quantita-tion of hemoglobin-based blood substitutes in whole blood sam-ples. Am J Clin Pathol. 2001;116:913-919.

6. d’Onofrio G, Zini G, Ricerca BM, Mancini S, Mango G. Auto-mated measurement of red blood cell microcytosis and hypochro-mia in iron deficiency and beta-thalassemia trait. Arch Pathol LabMed. 1992;116:84-89.

7. Kunicka JE, Fischer G, Murphy J, Zelmanovic D. Improvedplatelet counting using two-dimensional laser light scatter. Am JClin Pathol. 2000;114:283-289.

8. Chapman ES, Sorette M, Hetherington E, et al. A rapid, auto-mated flow cytometric method to measure activated degranu-lated platelets by density determination. Thromb Haemost.2003;89:1004-1015.

9. Fishbane S, Shapiro W, Dutka P, Valenzuela OF, Faubert J. A ran-

domized trial of iron deficiency testing strategies in hemodialysispatients. Kidney Int. 2001;60:2406-2411.

10. Kotisaari S, Romppanen J, Penttila I, Punnonen K. The Advia 120red blood cells and reticulocyte indices are useful in diagnosis ofiron-deficiency anemia. Eur J Haematol. 2002;68:150-156.

11. Brugnara C. Use of reticulocyte cellular indices in the diagnosisand treatment of hematological disorders. Int J Clin Lab Res.1998;28:1-11.

12. Thomas C, Thomas L. Biochemical markers and hematologicindices in the diagnosis of functional iron deficiency. Clin Chem.2002;48:1066-1076.

13. Malin MJ, Sclafani LD, Wyatt JL. Evaluation of 24-secondcyanide-containing and cyanide-free methods for whole bloodhemoglobin on the Technicon H*1TM analyzer with normal andabnormal blood samples. Am J Clin Pathol. 1989;92:286-294.

14. Benezra J. Mechanism of automated alkaline methods for thedetermination of hemoglobin in whole blood based on the micel-lization of ligated heme in the presence and absence of cyanide.Anal Chim Acta. 1992;262:67-77.

15. National Committee for Clinical Laboratory Standards. ReferenceLeukocyte Differential Count (Proportional) and Evaluation of Instru-mental Methods: Approved NCCLS Document H20-A. Villanova,PA: National Committee for Clinical Laboratory Standards; 1992.

16. Kaplan SS, Johnson K, Wolfe N, et al. Performance characteristicsof the Coulter LH 500 hematology analyzer. Lab Hematol.2004;10:76-87.

17. Sato S, Hirayama K, Koyama A, Harano T, Nagasawa T,Ninomiya H. Pseudoreticulocytosis in a patient with hemoglobinKoln due to autofluorescent erythrocytes enumerated as reticulo-cytes by the Cell-Dyn 4000. Lab Hematol. 2004;10:65-67.

performance evaluation of the advia 2120 hematology analyzer

Documents