platelet counting in platelet concentrates with various automated hematology analyzers

B L O O D C O M P O N E N T S

Platelet counting in platelet concentrates with variousautomated hematology analyzers

Margriet J. Dijkstra-Tiekstra, Willeke Kuipers, Airies C. Setroikromo, and Janny de Wildt-Eggen

BACKGROUND: Hematology analyzers use imped-ance, optical, and/or immunologic techniques for count-ing platelets (PLTs). PLT counting in whole blood hasbeen validated thoroughly; however, this is not the casefor PLT counting in PLT concentrates (PCs), in whichred cells (RBCs) are absent. Therefore, this study isfocused on PLT counting in PCs to study use of ethyl-enediaminetetraacetate (EDTA), carryover, and accu-racy of the analyzers.STUDY DESIGN AND METHODS: In total six hematol-ogy analyzers (AcT 8, Beckman Coulter; ADVIA 2120,Bayer; Cell-Dyn 4000, Abbott; Onyx, Beckman Coulter;K4500, Sysmex; and XT 2000i, Sysmex) were testedfor PLT counting. PC samples with various PLT concen-trations were made (0-1700 ¥ 109/L) and measured 10times. Carryover was determined five times.RESULTS: PC samples (1000 ¥ 109 PLTs/L) in EDTAtubes showed significantly higher PLT counts thansamples in “dry” tubes for all analyzers except for theCell-Dyn 4000 with the impedance technique. Carryoverwas not more than 0.3 percent for all analyzers. TheK4500 showed the most accurate results, whereas theCell-Dyn 4000 with the impedance technique had lowaccuracy due to an overestimation of more than20 percent.CONCLUSION: Most tested analyzers seemed to besuitable for counting PLTs in PCs. All hematology ana-lyzers should be validated for counting PLTs in absenceof RBCs as is the case in PCs, in addition to validationof PLT counting in whole blood.

Hematology analyzers are designed to countcells in (patient) whole blood. In bloodbanking, however, hematology analyzers arealso used to count cells in blood products.

These blood products are derived from centrifuged wholeblood; in most European blood banks, whole blood is gen-erally separated into a red cell (RBC) unit, a unit of plasma,and a buffy coat. The buffy coat contains approximately70 percent of the white blood cells (WBCs) and approxi-mately 90 percent of the platelets (PLTs). Buffy coats canbe further processed into PLT concentrates (PCs). A finalPC contains approximately 1000 ¥ 109 PLTs per L, compar-ing to a mean of 300 ¥ 109 PLT per L in normal blood, anda very low number of RBCs (<5 ¥ 109/PC) and WBCs(<1 ¥ 106/PC). PLT counting in PCs should be validatedseparately from PLT counting in whole blood becausemost analyzers are designed to count PLTs in presence ofRBCs. Thresholds for enumeration of PLTs are based onsize discrimination between PLTs and RBCs that are mea-sured in the same channel, and algorithms have beendesigned to distinguish these cells. Therefore, PLT count-ing in absence of RBCs might result in deviating PLTcounts.1-3

Hematology analyzers count cells based on imped-ance, optical technology, immunologic technology, or acombination of these. Impedance is based on monitor-ing the voltage required to send a constant currentthrough an aperture. A cell or particle passing through

ABBREVIATIONS: PC(s) = platelet concentrate(s); PPP =platelet-poor plasma.

From the Sanquin Blood Bank Northeast Region, Groningen,

the Netherlands.

Address reprint requests to: M.J. Dijkstra-Tiekstra, Sanquin

Blood Bank Northeast Region, Hanzeplein 1, 9713 GZ Gronin-

gen, PO Box 1191, 9701 BD Groningen, the Netherlands; e-mail:

[email protected].

Received for publication December 22, 2006; revision

received February 23, 2007, and accepted February 26, 2007.

doi: 10.1111/j.1537-2995.2007.01337.x

TRANSFUSION 2007;47:1651-1657.

Volume 47, September 2007 TRANSFUSION 1651

mailto:[email protected]

this aperture impedes current flow and the voltageincreases to maintain a constant current. The size ofeach particle is determined by the amplitude of thecorresponding pulse. The particle count is determinedby the number of pulses. Optical measurement occursby various angles of laser light measured by diodesthat can detect forward and sideward scatter. Basedon diameter and/or volume and structure and/ormorphology, the algorithms determine what cell haspassed the laser. The immunologic method is based onflow cytometry for which PLTs are labeled with a PLT-specific monoclonal antibody like CD41a (againstPLT membrane glycoprotein IIb/IIIa), CD42b (againstPLT membrane glycoprotein Ib-a), or CD61 (against PLTmembrane glycoprotein IIIa).4,5 Because the immuno-logic methods do not discriminate on size or impedance,it can be anticipated that there will be less influenceof background noise, that is, non-PLT events, and ofWBC or RBC fragments. Also macro- or microcytic PLTswill be detected since these have the same PLT-specificantigens as normal-size PLTs. Therefore, the Interna-tional Council for Standardization in Haematology andthe International Society of Laboratory Hematology con-sidered these methods to be reference for PLT counting.6

Beside these flow cytometric methods, there is only onehematology analyzer with an immunologic method(CD61) for PLT counting, the Cell-Dyn 4000 (Abbott,Wiesbaden-Delkenheim, Germany), which correlateswell with the flow cytometric reference method for PLTcounting.7,8

In the Netherlands, it became mandatory to registerthe PLT count on the PLT container. This makes it nec-essary to study accuracy, precision (or coefficient ofvariation [CV]), and linearity of PLT counting in PCs.

The aim of this study was to test various hematologyanalyzers that use different counting principles for PLTcounting. The hematology analyzers were tested for accu-racy, precision, linearity, and carryover. Also, the effect ofethylenediaminetetraacetate (EDTA) on PLT count anddilution of PC samples were studied with various analyz-ers. The Cell-Dyn 4000 CD61 technique was used as refer-ence in this study.

MATERIALS AND METHODS

Hematology analyzers testedIn total, six hematology analyzers were tested (Table 1).Before starting experiments, a three-level whole-blooddaily run control was performed to check all analyzersbefore use. The following analyzers were tested: the Cell-Dyn 4000 was tested for impedance, optical, and CD61measuring principles. The ADVIA 2120 (Bayer,Leverkusen, Germany) was tested for the autoloader andthe open modus, both with the optical measuring prin-ciple. The XT 2000i (Sysmex TOA, Tokyo, Japan) was testedfor the impedance and optical measurement. The remain-ing analyzers, AcT 8 (Beckman Coulter, Miami, FL), Onyx(Beckman Coulter), and K4500 (Sysmex TOA), all useimpedance measuring principles.

Preparation of PC samplesOvernight-stored whole blood was collected in citrate-phosphate-dextrose anticoagulant; centrifuged; anddivided in a plasma unit, buffy coat, and an RBC unit. Fivebuffy coats and 1 plasma unit were pooled and centri-fuged, and the PLT-rich plasma was pressed via a leukore-duction filter into a storage container. These PCs weresampled, and a dilution series was prepared for effect ofEDTA, carryover, and dilution studies (see below). Toestablish the PLT count, a PC sample was counted in trip-licate with the CD61 method on the Cell-Dyn 4000. Wechose this method due to good results obtained before.7,8

The mean PLT count was used to dilute PLT samples indiluent (Isoton II, Beckman Coulter) to the desired con-centrations. All measurements were performed on the dayof preparation.

EDTA versus dry tubes for PC samplesTo measure the effect of EDTA, samples with a dilution ofapproximately 250 ¥ 109, 500 ¥ 109, and 1000 ¥ 109 PLTsper L were prepared and transferred into both EDTA(K3EDTA, ref 454217, Greiner Bio-One, Kremsmünster,Austria) and dry tubes (no additive, ref 454241, Greiner

TABLE 1. Cell analyzers testedCell analyzer Technique Manufacturer

1a Cell-Dyn 4000 Impedance Abbott, Wiesbaden-Delkenheim, Germany1b Cell-Dyn 4000 Optical Abbott1c Cell-Dyn 4000 CD61 Abbott2 AcT 8 Impedance Beckman Coulter, Miami, FL3a ADVIA 2120 Optical autoloader Bayer, Leverkusen, Germany3b ADVIA 2120 Optical manual mode Bayer4 Onyx Impedance Beckman Coulter5 K4500 Impedance Sysmex TOA, Tokyo, Japan6a XT 2000i Impedance Sysmex6b XT 2000i Optical Sysmex

DIJKSTRA-TIEKSTRA ET AL.

1652 TRANSFUSION Volume 47, September 2007

Bio-One). Each sample was measured 10 times on the fol-lowing analyzers: AcT 8, ADVIA 2120, Cell-Dyn 4000,Onyx, and K4500.

CarryoverCarryover was determined by measuring three times asample with a concentration of approximately 500 ¥ 109

PLTs per L in an EDTA tube (H) followed by three times ablank sample (Isoton, phosphate-buffered saline [PBS] orair; L) to measure background. This was performed fivetimes for each analyzer. Carryover was calculated with theformula

L L H L ,1 3 3 3 100 9 10−( ) −( ) × .

L1 is the first blank measurement, L3 is the third blankmeasurement, and H3 is the third sample measurement.The following analyzers were tested for carryover: AcT 8,ADVIA 2120 (open modus only), Cell-Dyn 4000 (imped-ance and optical techniques only), Onyx, and K4500.

Carryover was acceptable when less than 1 percentcarryover was observed. This requirement is arbitrarilydefined, based on experience and from manuals of themachines, because no literature for this was found.

Linearity, accuracy, and CVTo determine linearity, accuracy, and CV of the variousanalyzers a dilution series was used containing 0,200 ¥ 109, 500 ¥ 109, 800 ¥ 109, 1100 ¥ 109, 1400 ¥ 109, and1700 ¥ 109 PLTs per L. For preparation of this series, the PCwas further concentrated by centrifugation. Nine percent(vol/vol) acid citrate dextrose formula A (pH 4.7, Baxter,Maurepas, France) was added to the PC, and this was cen-trifuged at 4500 ¥ g for 10 minutes. Subsequently, thepellet was dissolved in approximately 20 mL of the super-natant and an additional 80 mL of supernatant was addedto the PC. The PC was rested for 1 hour at room tempera-ture. The PLT count was determined with the CD61method on the Cell-Dyn 4000 (1:4 diluted in Isoton IIdiluent), and the dilution series was prepared (see “Prepa-ration of PC samples”). This series was measured 10 times.The Cell-Dyn 4000 CD61 measurements were performedonly in triplicate for practical reasons. Also because ofpractical reasons, not all analyzers were tested in the samepaired experiment, but in three separately performed,paired experiments with the Cell-Dyn CD61 technique asreference. The following analyzers were tested: Experi-ment 1, Cell-Dyn 4000 and ADVIA 2120; Experiment 2,Cell-Dyn 4000, AcT 8, Onyx, and K4500; and Experiment 3,Cell-Dyn 4000 and XT 2000i. For each experiment the dilu-tion series was freshly prepared.

Linearity was adequate when r2 values were greaterthan 0.98. Accuracy was the percentage of measurements

that fell within 10 percent of the expected values (of therange between 200 ¥ 109 and 1700 ¥ 109 PLT/L). An accu-racy of greater than 80 percent was considered sufficient.CV was acceptable when less than 5 percent. Theserequirements are arbitrarily defined, based on experience,because no literature for this was found.

Dilution of PC samplesTo measure the influence of dilution of the PLT samples onPLT counting, PC samples in EDTA tubes were used undi-luted (1:1) and diluted 1:2; 1:4, and 1:10 in Isoton IIdiluent, PBS, or PLT-poor plasma (PPP). PPP was preparedby centrifuging a plasma unit with 4500 ¥ g for 10minutes, and the supernatant was used. Each sample wasmeasured in 10 paired experiments on the followinganalyzers: AcT 8, Onyx, and K4500.

Statistical analysisTo compare results of samples in EDTA or dry tubes, apaired t test was used. To compare PLT count with thevarious analyzers, analysis of variance (ANOVA) followedby Dunnett’s test was used with the Cell-Dyn 4000 immu-nology measurements as control. To compare results ofthe various diluents repeated measurements, ANOVA fol-lowed by the Dunnett’s test was used with the results ofthe undiluted samples as control. A p value of less than0.05 was considered significant.

RESULTS

EDTA versus dry tubes for PC samplesTo study the effect of samples in “dry” tubes (i.e., withoutextra anticoagulant), PC samples were prepared in threedilution levels in both EDTA and dry tubes (Fig. 1). Itappeared that at the low dilution level (250 ¥ 109 PLTs/L)for most analyzers, no difference between PLT counts inEDTA and dry tubes was observed. At the high level(1000 ¥ 109 PLTs/L), however, a significantly higher PLTcount in the EDTA tube than in the dry tube was observedfor all analyzers. The only exception was the Cell-Dyn 4000impedance where no difference was detected. Themaximum difference between EDTA and dry tubes was21.4 percent for the Onyx.

CarryoverThe percentage of carryover, as measured with the highcount at about 500 ¥ 109 PLTs per L, was in all cases notmore than 0.3 percent. The highest carryover was mea-sured with the ADVIA 2120 (0.3%). The Cell-Dyn 4000optical method resulted in 0.03 percent carryover. Allother analyzers or methods showed 0 percent carryover.

PLATELET COUNTING


The Cell-Dyn 4000 impedance method, however, does notreport results below 20 ¥ 109 PLTs per L; thus, no carryoverbelow 4 percent could be measured.

Linearity, accuracy, and CVFor measuring the linearity, accuracy, and CV, the dilutionseries was tested on various analyzers on 3 days with theCell-Dyn 4000 measurements repeated on all 3 days.Because there were significant differences due to dilutionuncertainties between the results of the Cell-Dyn 4000between measuring days, the experiments of these 3 daysare presented separately.

In all three experiments (Fig. 2), the Cell-Dyn 4000impedance showed results that repeatedly were between

20 and 30 percent higher than expected.The levels at which accuracy (percentagewithin 10% of the expected value) wasequal or greater than 80 percent areshown in Table 2. The accuracy of theADVIA 2120 autoloader was greater than80 percent only at the level of 200 ¥ 109

PLTs per L (100%), because an overesti-mation of just above 10 percent was seenfor levels equal or greater than 500 ¥ 109

PLTs per L. It is possible, however, to cali-brate the ADVIA 2120 with a constantpercentage, so that all values will resultin a more accurate count. Results of themanual mode of the ADVIA 2120 wereapproximately 3 percent lower thanthose of the automatic mode, and there-fore accuracy was greater than 80percent between 200 ¥ 109 and 800 ¥ 109

PLTs per L. All analyzers showed linearresults (r2 > 0.98).

The CV was below 5 percent for theCell-Dyn 4000 (for two of three experi-ments), AcT 8, ADVIA 2120, Onyx, K4500,and XT 2000i with the impedance tech-nique over the whole range (200 ¥ 109-1700 ¥ 109 PLTs/L; Table 2). InExperiment 3 in which the Cell-Dyn4000 was compared with the XT 2000i theCV for the samples with 200 ¥ 109 PLTsper L exceeded 5 percent for both theCell-Dyn 4000 (impedance, optical, andimmunologic) and the XT 2000i (opticalonly), but remained below 7.5 percent.

The“zero” sample (Isoton II diluent)showed a maximum of 3 ¥ 109 PLTs per Lwith a mean of not more than 1 ¥ 109

PLTs per L for all tested analyzers. Onlythe AcT 8 and K4500 showed a real zerovalue for all zero sample measurements.

Dilution of PC samplesPC samples were diluted to various concentrations and invarious diluents to see whether there is an optimal (i.e., nosignificant influence on cell count) combination of dilu-tion and diluent (Fig. 3). The largest difference comparedto the undiluted sample was observed for the 1:10 dilu-tion. The 1:10 dilution in Isoton (AcT 8, K4500, and Onyx)or PBS (K4500) gave significantly lower PLT counts, but the1:10 dilution in PPP (AcT 8 and Onyx) gave significantlyhigher PLT counts compared to the less diluted samples.The samples with 1:2 or 1:4 dilution in any diluent did notdiffer significantly from the undiluted samples, except for1:2 and 1:4 diluted samples in Isoton measured on theK4500.

0

500

1000

1500

CD imp CD opt ADVIA2120

Act 8 Onyx K4500

Dilu

tio

n le

vel (¥1

09 P

LT

s/L

)

*

*

*

*

** *

*

Fig. 1. PC samples in EDTA ( ) or dry (�) tubes in three different dilutions and mea-

sured on various hematology analyzers. Data are shown as means � SD with n = 10.

CD imp = Cell-Dyn 4000 impedance; CD opt = Cell-Dyn 4000 optical. *p < 0.05

between EDTA and dry samples of the respective dilution level.

TABLE 2. Levels of each analyzer with accuracy equal or more than80 percent and CV below 5 percent*

Experiment and analyzerAccuracy, levels with �80%

within 10% of expectedCV, levels with

CV < 5%

Experiment 1Cell-Dyn 4000 CD61 500-1700 200-1700Cell-Dyn 4000 impedance 200-1700Cell-Dyn 4000 optical 500-1700 200-1700ADVIA 2120 autoloader 200 200-1700ADVIA 2120 open modus 200-800 200-1700

Experiment 2Cell-Dyn 4000 CD61 1100-1700 200-1700Cell-Dyn 4000 impedance 200-1700Cell-Dyn 4000 optical 1100-1700 200-1700AcT 8 200-1400 200-1700Onyx 200-1400 200-1700K4500 200-1700 200-1700

Experiment 3Cell-Dyn 4000 CD61 500-1700 500-1700Cell-Dyn 4000 impedance 500-1700Cell-Dyn 4000 optical 500-1700 500-1700XT 2000i impedance 500-1100 200-1700XT 2000i optical 500 500-1700

* Levels shown are ¥ 109 PLTs per L.



DISCUSSION

The techniques of the hematology analyzers for PLTcounting are based on impedance, optical methods, orimmunology. It is important to discriminate PLTs from

non-PLT events and to discriminate large PLTs from othercells.11,12 Because PCs contain almost no RBCs, countingPLTs in PCs might depend on whether RBCs are needed fordiscrimination or not.

We observed that measurements of samples at thelevel of 1000 ¥ 109 PLTs per L in presence of EDTAresulted in a higher PLT count compared to samples indry tubes. The differences between EDTA and drydepended on the analyzer, but in all cases samples inEDTA tubes were more close to the level of 1000 ¥ 109

A

0

500

1000

1500

2000

2500

200 500 800 1100 1400 1700

Expected PLT count (x109/L)

Ob

serv

ed P

LT

co

un

t (x

109 /L

)

CD imm

CD imp

CD opt

ADVIA auto

ADVIA manual

*

**

***

*

**

*

*

**

*

**

*

B

0

500

1000

1500

2000

2500

200 500 800 1100 1400 1700


Ob

serv

ed P

LT

co

un

t (x

109 /L

)

CD imm

CD imp

CD opt

AcT 8

Onyx

K4500

*

*

***

*

****

*

*****

****

*

**

*

*

C

0

500

1000

1500

2000

2500

200 500 800 1100 1400 1700


Ob

serv

ed P

LT

co

un

t (x

109 /L

)

CD imm

CD imp

CD opt

XT imp

XT opt

*

*

*

*

**

*

**

*

**

*

**

*

Fig. 2. PLT dilution series measured on various hematology

analyzers in three paired studies (A, B, and C). Data are shown

as means � SD with n = 10, except for Cell-Dyn 4000 immu-

nologic with n = 3. CD imp = Cell-Dyn 4000 impedance;

CD opt = Cell-Dyn 4000 optical; XT imp = XT 2000i impedance;

XT opt = XT 2000i optical. *p < 0.05 with the Cell-Dyn 4000

immunologic method of the respective PLT dilution level.

AcT 8

0

500

1000

1500

1:1 1:2 1:4 1:10

Dilution

Dilu

ent

leve

l (¥1

09 PL

Ts/

L)

A

*

*

Onyx

0

500

1000

1500

1:1 1:2 1:4 1:10

Dilution

Dilu

ent

leve

l (¥1

09 PL

Ts/

L)

Dilu

ent

leve

l (¥1

09 PL

Ts/

L)

B

*

*

K4500

0

500

1000

1500

1:1 1:2 1:4 1:10

Dilution

C

****

Fig. 3. PC samples diluted in Isoton ( ), PBS (�), or PPP ( ).

Samples were measured on the Act 8 (A), Onyx (B), and K4500

(C). Data are shown as means � SD and are multiplied by the

dilution factor with n = 10. *p < 0.05 with the undiluted

sample.

PLATELET COUNTING


PLTs per L. Because the CD61 method, which can be usedas a reference method for PLT counting, was used fordetermining the dilution factor of the samples, the mea-surement in EDTA seems to be more appropriate thanmeasurements in dry tubes. Lower PLT counts in drytubes are probably caused by PLT aggregates that disin-tegrate once EDTA is present.13,14 The Cell-Dyn 4000impedance was the only tested analyzer that did notshow a significant difference between samples in EDTA ordry tubes, which is confirmed by results of Johannessenand coworkers.1 This probably can be explained by theuse of floating thresholds for discrimination of PLTs fromother cells by the Cell-Dyn 4000. With fixed thresholds,PLTs of samples that are not anticoagulated with EDTA donot fall completely between the thresholds. With floatingthresholds, these PLTs still fall within the thresholds,unless abnormality is too large in which case the analyzeruses the fixed default position of the thresholds (Johan-nessen et al.1 and personal communication with productspecialist of Abbott).

Carryover for the analyzers was minimal and will notinterfere with results. The carryover that was observed fellwithin background values that are described in themanuals.

From our study no preference for impedance oroptical technology can be given, as can be concluded fromresults of accuracy, precision (CV), and linearity. This issupported by Segal and colleagues.15 One exception is theimpedance measurement of the Cell-Dyn 4000 that gavean approximately 25 percent overestimation. This is con-firmed by Johannessen and coworkers1 and Hervig andcoworkers2 counting PC samples, who found an overesti-mation of 25 and 22 percent for their impedance measure-ments at the Cell-Dyn 4000. The reason for this is that theCell-Dyn 4000 impedance technique automatically com-pensates for coincidence of PLTs and RBCs, while forsamples in the absence of RBCs no correction is needed.For the Cell-Dyn 4000 optical technique they found amean underestimation of 13 percent,1 which was notfound in this study. They partly explain this by theincreased numbers of small PLTs (up to 5.2%) that fallbelow the lower threshold.1

The results of the manual or autoloader modus of theADVIA 2120 appeared to be 3 percent different from eachother. A difference between both modes is the tube lengthto the sample shear valve, with a longer tube length for theautoloader modus. The volume of sample that will bemeasured is collected in this shear valve. Subsequently thesample will go to the RBC and PLT reaction chamberwhere cells are diluted, become spherical, and are fixed.Afterward the cells are measured in the flow cell. Forwhole blood, no difference between results derived ofboth modes was observed. PC samples are less viscousthan whole-blood samples, which probably lead to thedifferent results between both modes (results of

validation; personal communication with product spe-cialist of Bayer Healthcare).

In the comparison of the various cell analyzers andtechniques, many significant differences were observed.At the level of 1100 ¥ 109 PLTs per L a maximum differenceof 30 percent between analyzers was observed. Weemployed no further standardization or recalibration toobtain more corresponding results. Moroff and cowork-ers3 observed a comparable maximum difference inporcine PLT counts between analyzers in a send around.They were able to reduce this difference by a factor twoafter recalibration with fixed porcine PLTs as calibrant,emphasizing the importance of having a readily availablegold standard method and/or gold standard sample.

In this study, the whole-blood daily run controls atthree levels show PLT counts within the required ranges(results not shown). We found various analyzers with devi-ating results for counting PLTs, however. Thus, it is impor-tant to realize that determination of accuracy takes placein the appropriate suspension medium, that is, in theabsence of RBCs when counting PCs and in the RBC sus-pension when counting whole blood in this case.

In this study we also investigated the effect of dilu-tion. The main effect was seen in the 1:10 dilution. Theeffect of PPP can probably be explained by the fact that itwas not completely PLT free. We have shown that mostanalyzers are linear over the whole range. This makes dilu-tion unnecessary for at least the tested analyzers.

The international accepted reference method for PLTcounting is based on the RBC-to-PLT ratio5,6 and wastherefore not suitable as reference for counting PLTs.Therefore, we used in the current study the CD61 methodof Cell-Dyn 4000 as reference for preparing the dilutionseries. It is described that the CD61 technique of Cell-Dyn4000 can be used as a good reference method for PLTcounting,7,8 although this is only tested for counting PLTsin whole-blood samples. Johannessen and coworkers1

also used the CD61 technique as default for counting PLTsin PC samples. In our study, results of the Cell-Dyn 4000CD61 were in general close to the theoretical values.

In conclusion, most analyzers and methods seemedto be suitable for counting PLTs in PCs in blood bankingenvironment. There is no preference for impedance,optical, or immunologic technique for counting PLTs inthe absence of other cells. Each technique or modus of ahematology analyzer should be validated because the onetechnique of an analyzer can differ from the other as isseen between impedance and optical methods of one ana-lyzer and between the autoloader and manual mode ofanother analyzer. In some cases, it might be necessary torecalibrate the analyzer. It is important to use samples inEDTA because PLT count can be reduced for samples inthe absence of EDTA. For regular PCs with a PLT count ofapproximately 1000 ¥ 109 PLTs per L, no dilution isneeded, at least for the tested analyzers. If dilution is



needed, a limited dilution is advisable. Whole-blood con-trols do not guarantee an accurate PLT count for PCs.

ACKNOWLEDGMENTS

We thank Dirk de Korte of Sanquin Research, Department Blood

Cell Research, in Amsterdam, the Netherlands, for his permission

to use the ADVIA 2120 and for critical reading of the manuscript.

We also are grateful to Pieter van der Meer of Sanquin Blood Bank

Northwest Region, Amsterdam, the Netherlands, for critical

reading of the manuscript.

REFERENCES

1. Johannessen B, Haugen T, Scott CS. Standardisation of

platelet counting accuracy in blood banks by reference to

an automated immunoplatelet procedure: comparative

evaluation of Cell-Dyn CD4000 impedance and optical

platelet counts. Transfus Apher Sci 2001;25:93-106.

2. Hervig T, Haugen T, Liseth K, Kjeldsen-Kragh J, Scott CS,

Johannessen B. The platelet count accuracy of platelet

concentrates obtained by using automated analysis is

influenced by instrument bias and activated platelet com-

ponents. Vox Sang 2004;87:196-203.

3. Moroff G, Sowemimo-Coker SO, Finch S, et al. The influ-

ence of various hematology analyzers on component

platelet counts. Transfus Med Rev 2005;19:155-66.

4. Dickerhoff R, Von Ruecker A. Enumeration of platelets by

multiparameter flow cytometry using platelet-specific anti-

bodies and fluorescent reference particles. Clin Laboratory

Haematol 1995;17:163-72.

5. Harrison P, Ault KA, Chapman S, et al. An interlaboratory

study of a candidate reference method for platelet count-

ing. Am J Clin Pathol 2001;115:448-59.

6. Rowan M. Platelet counting by the RBC/platelet ratio

method: a reference method. Am J Clin Pathol 2001;115:

460-4.

7. Kunz D, Kunz WS, Scott CS, Gressner AM. Automated

CD61 imunoplatelet analysis of thrombocytopenic

samples. Br J Haematol 2001;112:584-92.

8. Arroyo JL, García-Marcos MA, López A, et al. Evaluation of

a CD61 MoAb method for enumeration of platelets in

thrombocytopenic patients and its impact on the transfu-

sion decision-making process. Transfusion 2001;41:1212-6.

9. Langford K, Luchtman-Jones L, Miller R, Walck D. Perfor-

mance evaluation of the Sysmex XT-20001 automated

hematology analyzer. Lab Hematol 2003;9:29-37.

10. Nakul-Aquaronne D, Sudaka-Sammarcelli I, Ferrero-

Vacher C, Starck B, Bayle J. Evaluation of the Sysmex

Xe-2100 hematology analyzer in hospital use. J Clin Lab

Anal 2003;17:113-23.

11. Kickler TS. Clinical analyzers: advances in automated cell

counting. Anal Chem 1999;71:363R-365R.

12. Harrison P, Segal H, Briggs C, Murphy M, Machin S.

Impact of immunological platelet counting (by the

platelet/RBC ratio) on haematological practice. Cytometry

B Clin Cytom 2005;67B:1-5.

13. McShine RL, Das PC, Smit Sibinga CT, Brozovic B. Effect of

EDTA on platelet count and other platelet parameters in

blood and blood components collected with CPDA-1. Vox

Sang 1991;61:84-9.

14. Ip AH, Seghatchian MJ. Quality monitoring of haemapher-

esis platelet concentrates: sampling in EDTA helps with the

standardization and improves the consistency. Blood

Coagul Fibrinol 1991;2:329-32.

15. Segal HC, Briggs C, Kunka S, et al. Accuracy of platelet

counting haematology analysers in severe thrombocytope-

nia and potential impact on platelet transfusion. Br J Hae-

matol 2005;128:520-5.

PLATELET COUNTING


platelet counting in platelet concentrates with various automated hematology analyzers

Documents