Hematology

ISSN: (Print) 1607-8454 (Online) Journal homepage: https://www.tandfonline.com/loi/yhem20

Advances in Platelet Counting

Paul Harrison, Carol Briggs & Sam Machin

To cite this article: Paul Harrison, Carol Briggs & Sam Machin (2001) Advances in Platelet
Counting, Hematology, 5:6, 421-427, DOI: 10.1080/10245332.2001.11746538
To link to this article: https://doi.org/10.1080/10245332.2001.11746538

Published online: 13 Jul 2016.

Submit your article to this journal

Article views: 4135

View related articles

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=yhem20
Hematology, 2001, Vol. 5, pp. 421-427 © 2001 OPA (Overseas Publishers Association) N.V.
Reprints available directly from the publisher Published by license under
Photocopying permitted by license only the Harwood Academie Publishers imprint,
part of The Gordon and Breach Publishing Group.

Hemostasis & Thrombosis

Advances in Platelet Counting
PAUL HARRISON*, CAROL BRIGGS and SAM MACHIN

Department of Haematology, 98, Chenies Mews, University Co/lege Hospital, London, WCIE 6HX, UK

(Received 6 September 2000; ln final form 20 September 2000)

Accurate and reliable platelet counting is critical for method (flow cytometric immunocounting using
the clinical management of platelet disorders, espe- the PLT/RBC ratio) should therefore improve the
cially in thrombocytopenia. The platelet count is overall reliability of platelet counting especially in
used to determine if the patient requires a platelet thrombocytopenia. In this review, the history and
transfusion. As the prophylactic transfusion trigger recent advances in platelet counting methodologies
is now set anywhere between 10-20X109/L (depend- will be presented. The relative advantages and
ing upon the institution) it is therefore important disadvantages of each technology will then be
that reliable and accurate counts are obtained in discussed along with their potential impact on
severely thrombocytopenic samples. The accuracy improved accuracy of platelet counting.
and precision of automated platelet counts is totally
reliant upon optimal discrimination of platelets
from other cells and interfering particles. However, Keywords: Platelets, immunoplatelet count, optical platelet
count, impedance platelet count
clinicians often still rely upon counts that have been
generated using so called "1-dimensional" cell size
analysis systems, which not only fail to discriminate
platelets from cell fragments of similar size but
exclude large platelets from the final count. Also INTRODUCTION
the current reference method for platelet counting
(the manual phase count), upon which analysers are
usually calibrated is highly imprecise, time consum-
The assessment of the platelet count is of para-
ing and unreliable. Thus there has been a demand mount importance in the routine haematology
for improvements in platelet counting technology in laboratory to assist in the diagnosis of various
order to improve accuracy of counting in thrombo- clinical disorders and the monitoring of patients
cytopenia so that correct clinical decisions can be
made. More recent developments including the receiving cytotoxic treatments. Accurate and
introduction of "2-dimensional" optical counting precise platelet counts in severely thrombocyto-
and immunoplatelet counting within automated penic patients are critical in clinical practice to
systems are significant advances. The availability
of new technologies coup led with the recent develop- facilitate correct clinical decisions in or around
ment of a new candidate international reference prophylactic platelet transfusion thresholds.

* Corresponding au thor. Tel.: 44-(0)-207-380-9580. Fax: 44-(0)-207-380-9584. E-mail: paul.harrison@ucl.ac.uk

421
422 P. HARRISON et al.

Increasing demand for reducing the frequency specifie to a platelet cell surface antigen, an indir-
of platelet transfusions has resulted in a drop ect platelet immunocount can be obtained by
in the prophylactic decision making threshold flow cytometry by simply counting the ratio of
from 20 x 109 /1 to 10 x 109 /1, but without any Red Blood Ceil Events to fluorescent platelets
evidence of a significant risk of spontaneous (RBC:PLT ratio). This method is a significant
bleeding [1]. Other laboratories and haemato- improvement over the manual phase count in
logists have even proposed that the trigger that it resolves ail platelets from debris and red
threshold might be reduced further to 5 x 109 /1 cells, it is rapid and easy to perform and is
[2--6]. However, in arder to be confident with statistically superior as at least 1000 platelet
platelet counts in and around these thresholds, events are always counted even in thrombocyto-
haematologists and clinicians must not only under- penic samples. The development and evolution
stand the limitations associated with a particular of this method and it's potential impact on plate-
technology but be aware of alternative methods let counting is discussed in detail below.
which may have significant impact upon the ac- Although electronic impedance counters have
curacy of platelet counting in thrombocytopenia. improved the enumeration of platelets because
Four analytical procedures are now currently a higher number of ceils are counted, these
available for platelet counting; (1) Manual phase methods still have significant limitations despite
contrast microscopy; (2) Impedance analysers; their widespread use. One of the major problems
(3) Optical analysers (using light scatter or fluor- in accurate platelet counting is that "1-dimen-
escence); and (4) Immunoplatelet counting. sional" ceil size analysis cannat discriminate
The Manual phase microscopy method is time platelets from other similarly sized particles
consuming, subjective, tedious as weil as impre- such as white ceil debris or small or fragmented
cise, and at low platelet counts the imprecision red cells within the sample [13]. These may be
will be even greater due to the low number of erroneously included in the platelet count, and
ceils being counted. Nevertheless, this is still the in severely thrombocytopenic patients the num-
international reference method [7,8], and man- ber of interfering particles may often exceed the
ual confirmation of low automated counts is still number of true platelets. Large or giant platelets
often a routine practice owing to the potential may be excluded from the count on the basis of
inaccuracy of automated counts in samples con- their size as they cannat be resolved from red
taining small platelet-like particles. Given the cells. For a list of potential interferences see
problems with manual counting, a new immuno- Table 1. There is also significant variation in the
platelet counting procedure has recently been results obtained on different analysers within
advocated as a candidate reference method [9-12]. the same sample due to differences in the
By using a fluorescent monoclonal antibody method of analysis. In response to these prob-

TABLE I List of potential error sources on platelet counts

Error source Effect on count Clinical condition Frequency
Fragmented red cells False increase Hemolytic anemia, DIC, HUS Rare
Fragmented white cells False increase Leukemia, chemotherapy Occasional
Large platelets False decrease Macrothrombocytopenia, Rare
Bernard Soulier disease, ITP
Platelet aggregates False decrease Aged sample, EDTA effect Occasional
Platelet-WBC complexes False decrease EDTA effect, Platelet activation Occasional
Absent Gpllb/illa Low immunocounts Glanzmann' s thrombasthenia Very rare
Autoantibodies Low immunocounts ITP Rare
Adapted from W. Groner, Draft ICSH reference method for obtaining the ratio of the RBC count to the platelet count.
 PLATELET COUNTING 423

lems with impedance counters, despite refining the chamber are counted with a phase contrast
computer algorithms to define the platelet popu- microscope. The subjectivity and number of cells
lation, a number of manufacturers have devel- being counted limits the accuracy of this method.
oped new technologies. Recent developments At low platelet numbers because fewer cells are
include the introduction of "2-dimensional" being counted, observed CVs increase propor-
optical counting by analysing either scattered tionally. The degree of imprecision with the
light and/ or fluorescence to optimally resolve manual count is exemplified by reported inter-
the platelets from either noise/ cellular debris observer coefficient of variations (CV) in the
and red cells.These methods will also be dis- range of 10 to 25% [10]. Other disadvantages of
cussed in detail below. this technique are the possible errors in the
Given the apparent high degree of confidence manual dilution of the blood and that red cell
with platelet counting, it is therefore important debris may be often be mistaken for platelets.
not only to highlight the problems with existing Despite this, it is still common practice to use
counting methodologies, but to demonstrate that manual counting methods in the routine labora-
more reliable and accurate counts are feasible tory if the platelet count is low or if there are
with sorne of the newer technologies that have atypical platelets within the sample. This method
become available. This should lead to improved remarkably is still the international reference
accuracy of platelet counting in thrombo- method for platelet counting and the calibration
cytopenia and perhaps facilitate further studies of platelet counts on automated blood cell coun-
into proving whether current platelet trans- ters and quality control material is still routinely
fusion thresholds can be safely lowered further. performed by manufacturers using this relative!y
Given the high cast and risks associated with imprecise method. Given the problems with this
platelet transfusions this may eventually methodology a new flow cytometric immuno-
decrease the frequency of unnecessary transfu- counting method has recently been put forward
sions without increasing the risk of bleeding. as a new candidate reference procedure [9-12].
This will hopefully have significant impact upon
the transfusion budget and management of the
thrombocytopenic patient. AUTOMATED PLATELET COUNTING

There are now several methods on individual

MANUAL PLATELET COUNTING analysers for counting platelets (Table II).
Whichever automated method is used for plate-
The current international reference method for let counting it must be demonstrated that the
platelet counting is performed by the manual method is not only precise but shows minimum
procedure using phase contrast microscopy fluctuation in repeated results on the same
[7,8]. EDTA anticoagulated blood is diluted sample. It is also important that any analyser
1:20 in 1% filtered ammonium oxalate solution gives linear results over the entire analytical
and mixed for 10 minutes [7,8,14]. The ammo- range, especially at either low or high counts.
nium oxalate lyses the red cells but leaves the In samples with high counts, there is a growing
platelets and white cells intact. The dilution is probability of coincidence with two or more cells
used to fill the counting chamber (haemocyto- passing through the sensing zone at the same
meter) and it is left to settle in a moist atmo- time, as well as possible sample carryover.
sphere. The haemocytometer should be of Within thrombocytopenic samples it is import-
Neubauer design and have an inaccuracy of the ant that added counts due to background counts
specified volume of less than 1%. The platelets in (spurious signais caused by electronic noise) are
424 P. HARRISON et al.

TABLE II List of major haematological analysers currently available

Manufacturer Instrument Principle of platelet count
Abbott diagnostics CELL-DYN 4000 Impedance, optical and immunological
ABX diagnostics Pentra-120 Impedance
Bayer corporation ADVIA 120 Optical
Beckman Coulter STKS Impedance
GEN-S Impedance
Sysmex corporation SE-9500 Impedance
XE-2100 Impedance and optical

not included within the reported value. It is also fied in the distribution. A log-normal curve is
desirable that there is minimal variation be- fitted to these points. The curves have a range of
tween different analysers, results obtained with 0-70 fl and the platelet count and parameters are
different systems on the same sample should be derived from this curve. Within the Sysmex
comparable. The different types of analyser counting systems (e.g. SE 9500 and XE 2100)
available are now discussed in detail. platelets are also counted by the orifice imped-
ance method [15]. A platelet size distribution
plot is produced using three thresholds. One is
fixed at the 12 fl level and the other two are
IMPEDANCE PLATELET COUNTERS allowed to hunt the upper and lower ends of
the platelet population between certain limits.
Wallace Coulter first patented the impedance The lower platelet size threshold may move
counting method in 1956 [15]. Using this method between 2 and 6 fl, the higher between 12 and
cells in suspension are continually passed 30 fl. The purpose of these thresholds is to
through an orifice and an electric current endeavour to distinguish platelets from small
applied. Each cell as it passes through the orifice red cells or red cell fragments at the upper end
will cause an impedance change in the orifice of the platelet population and debris at the lower
determined by the size of the cell. These imped- end. Analysers using the standard impedance
ance changes can then be used to size and measurements are able (for most samples) to
count the individual cells. This method was provide an accurate platelet count dawn to
originally used for the counting of red cells and 20 x 109 /1. Below this level impedance analysers
white cells and the fust Coulter platelet counter become less accurate due to decreasing statist-
required the use of platelet rich plasma to avoid ical confidence, fewer events analysed, and the
counting red cells as platelets. It was not until increasing influence of background and plasma
the 1970s that improvements in technology, non platelet particulate matter. Interfering sub-
including hydrodynamic focusing, allowed the stances in the blood can potentially cause im-
discrimination of platelets from red cells to pedance counting problems (see Table Il). False
enable an accurate platelet count to be produced increases will occur when there are erythrocyte
from a whole blood sample. Within the presently and white cell fragments, microcytic red cells,
available Coulter analysers (e.g. STKS and immune complexes, bacteria, cell debris or carry-
GEN S) particles between 2 and 20 femto- over from a previous sample [13]. False under-
litre (fl) only are counted as platelets [14]. Pulses estimates within the count will occur when there
are obtained from three red cell/platelet orifices are large platelets present, platelet dumping,
to obtain 64-channel size distribution histograms as seen with pseudo-thrombocytopenia by ethy-
for each orifice. These histograms are smoothed lene diamine tetraacetic acid (EDTA)-dependent
and a high point and two low points are identi- agglutinins or with platelet satellism [13].
 PLATELET COUNTING 425

OPTICAL PLATELET COUNTING counting especially within thrombocytopenic

specimens [17,18].
More recently opticallight scatter methods have
been introduced. In one-dimensional analysis
platelets are analysed by laser flow cytometry OPTICAL FLUORESCENT PLATELET
based on light scatter measurement at a single COUNTING
angle (2° to 3°). In these automated systems,
(H*3, Bayer, or the CELL-DYN 3500, Abbott) a An optical fluorescent platelet count has been
series of algorithms or a smoothing or fitting introduced on the Sysmex XE 2100 analyser,
routine on the platelet volume histogram are but an impedance count is still available [19]. A
used to establish validity of each platelet count. polymethine dye is used to stain the RNA/DNA
The ADVIA 120 analyser (Bayer Corporation) of reticulated cells and the platelet membrane
uses two dimensional platelet analysis, volume and granules. This technology facilitates the
and refractive index of effectively sphered indi- simultaneous counting of the red cell reticulo-
vidual platelets are simultaneously determined cytes, erythrocytes and fluorescent platelets.
on a cell-by-cell basis by measuring two angles Within the flow cell each single cell is passed
of laser light scatter at 2-3° and at 5-15° [16]. through the light bearn of a semiconductor diode
The two scatter measurements are converted to laser, and the differences in their fluorescence
volume (platelet size) and refractive index (plate- intensities allows the separation of the platelets
let density). The platelet scatter cytogram map from the red cells and reticulocytes. The fluores-
resolves volumes between 1 and 30 fl and refract- cent staining of the platelets not only potentially
ive index values between 1.35 and 1.44. Large excludes non-platelet particles from the count
platelets, red cell fragments, red cell ghosts, but includes large or giant platelets. The optical
microcytes and cellular debris are thus potentially fluorescent count is more reliable at levels below
discriminated. Recent data suggests that the 100 x 109 /1 and so should facilitate more appro-
optical count improves the accuracy of the plate- priate clinical decisions to be made, particularly
let count in thrombocytopenic samples [10,16]. with regard to platelet transfusions [19].
The CELL-DYN 4000 instrument (Abbott
Diagnostics) also routinely reports an optical
platelet count (as well as an impedance count) IMMUNOLOGICAL PLATELET COUNTING
based on two light scatter parameters, inter-
mediate light scatter (7°) and a wide angle scat- Recently, a flow cytometric platelet counting
ter (90°) [16,17]. An algorithm is used to identify procedure has been suggested as a candidate
platelets using these two parameters to exclude reference method and has been the subject of
non-platelet particles while counting all plate- review by the International Council for Standard-
lets. This is a two dimensional analysis in which ization in Haematology {ICSH) expert panel on
platelets must fall within a region that defines cytometry [9-12]. The ICSH panel originally
the correlation between the two light scatter identified the variables associated with this
parameters (a sloping window) and between a methodology, which in tum has enabled the
lower threshold and an upper discrimina tor. The International Society of Laboratory Haemato-
three discriminator lines are set dynamically; logy (ISLH) task force panel to develop, test
the lower threshold is fixed. Comparisons of and evolve a candidate consensus reference
platelet counts generated in samples containing method [11,12]. The principle of the method
either noise or large platelets demonstrate that simply involves labelling platelets within
optical scatter improves the accuracy of platelet EDTA anticoagulated whole blood with suitable
426 P. HARRISON et al.

anti-platelet monoclonal antibodies (conjugated scatter and impedance characteristics. This

to FITC) and detecting the platelets by flow opened the possibility of performing an Immuno-
cytometry. The platelet count is simply derived platelet count within an automated setting.
by dividing the red cell count (determined by a Currently the only commercial instrument that
calibrated impedance counter) by the ratio of can perform this test is the Abbott CELL-DYN
red cells to fluorescent platelets detected 4000 [13,17,18]. Their ImmunoPLTlM method is
(RBC:PLT ratio). The main advantage of the fully automated and labels platelets within
RBC ratio is that providing the blood sample is whole blood by using anti-CD61 antibodies
mixed and that coïncident events (RBC/RBC and contained within a lyophilised pellet within
RBC/platelet) are eliminated by optimal dilution special evacuated tubes (Becton Dickinson, San
then the count obtained is totally independ- Jose, California). During analysis, the CELL
ent of potential pipetting artefacts. Alternative DYN 4000 simply aspira tes 40 Ill of blood into
immunoplatelet counting procedures have used an antibody containing tube, performs an in-
known amounts of added fluorescent calibration cubation (with mixing) and then analysis based
beads to derive the platelet count [20]. However, upon a plot of fluorescence (FLl) versus light
although these methods can provide accurate scatter. The final count includes PLT /RBC coïn-
results, they are totally dependent upon a stable cidence events and is counted within a given
head preparation (with an accurate head count) volume and not based upon a ratio. The method
in combination with very accurate/precise provides an accurate platelet count especially in
pipetting. In comparative studies although the thrombocytopenic samples and studies have
within assay CV's are superior to the manual shawn that the method could potentially reduce
phase method, the RBC ratio provides a simpler, unnecessary platelet transfusions by 11.6% and
cheaper but more accurate means for determin- 16.3% respectively when compared to optical or
ing platelet counts [10]. impedance methods [17]. It is also not surprising
Many recent studies as well as an ISLH inter- that recent studies have also demonstrated that
laboratory comparison have demonstrated that this method correlates extremely well with the
the RBC ratio satisfies the criteria for a reference immunoplatelet count derived from the RBC
method particularly as it is superior to the ratio by flow cytometry [Luc van Hove persona!
manual phase count. The new method is rapid, communication]. Bath immunological methods
simple, reproducible (with excellent CV's) and may therefore help to improve the accuracy
can be set up by any laboratory with a fluores- of platelet counting in thrombocytopenia.
cent flow cytometer [9-12]. Hopefully the method However, a fully automated immunological
will replace manual phase microscopy and help technique has obvious advantages and may
to improve the calibration of haematology ana- occasionally be extremely useful in clinical
lysers, QC materials and the accuracy of platelet situations where accurate platelet counts are
counting in thrombocytopenia. required.

IMMUNOPLATELET COUNTING WITHIN CONCLUSIONS

ANALYSERS
Many clinical haematologists and physicians
With the introduction of fluorescent flow cyto- have been guilty of taking automated platelet
metry principles into conventional analysers it counts for granted and are not aware of the
became feasible to simultaneously measure significant limitations that exist with so called
fluorescent cells in conjunction with their light "1-dimensional" analysers e.g. impedance or
 PLATELET COUNTING 427

single angle light scatter instruments. The con- Thom, R., Van Assendelft, O. W. and Verwilghen, R. L.
(1998). Recommended methods for the visual determi-
current development of a new reliable candidate nation of white celland platelet counts, WHO LAB, 88, 1.
reference method and the increasing availability [8] Brecher, G., Schneiderman, M. and Cronkite, E. P.
(1953). The reproducibility of the platelet count, Amer-
of "2-dimensional" analysis using light scatter ican Journal of Clinical Pathology, 23, 15-21.
and/ or fluorescence will improve our ability to [9] Davis, B. and Bigelow, N. C. (1999). Indirect Immuno-
give accurate platelet counts in thrombocytope- platelet counting by flow cytometry as a reference
method for platelet count calibration, Laboratory
nia. Extemal QC programmes (e.g. NEQAS) will Hematology, 5, 15-21.
now also be able to develop suitable stabilised [10] Harrison, P., Horton, A., Grant, D., Briggs, C. and
Machin, S. J. (2000). Immunoplatelet counting: A pro-
and calibrated materials to assess the accuracy of posed new reference procedure, British Journal of
counting in thrombocytopenia. For the first time Haematology, 108, 228-235.
[11] International Council for Standardization in Haematol-
haematologists may be able to report and rely ogy (ICSH) Expert panel on Cytometry and Interna-
upon law platelet counts with improved confid- tional Society of Laboratory Haematology (ISLH). Task
Force on Platelet Counting (2000). Platelet Counting by
ence. Comparative studies with different ana- the PLT/RBC ratio - a reference method, American
lysers and immunocounting will also facilitate Journal of Clinical Pathology (manuscript to be submitted).
the re-evaluation of current platelet transfusion [12] Harrison, P., Ault, K. A., Chapman, S. E., Davis, B.,
Fujimoto, K., Houwen, B., Kunicka, J., Lacombe, F.,
thresholds and may lead to substantial reduc- Machin, S., Raynor, R., Van Hove, L. and van
tions in blood bank expenditure with obvious Assendelft, O. W. (2000). An lnter-laboratory Study of
a Candidate Reference Method for Platelet Counting,
clinical benefits. American Journal of Clinical Pathology (manuscript to be
submitted).
[13] Ault, K. A. (1996). Platelet counting. Is there room for
improvement? Laboratory Haematology, 2, 139-143.
References [14] Rowan, R. M. (1991). Platelet counting and assessment
of platelet function. ln Practical Laboratory Hematology,
[1] Rebulla, P., Finazzi, G., Marangoni, F., Avvisati, G., edited by J. A. Koepke, p. 157, New York, NY. Churchill
Gugliotta, L., Barbui, T., Mandelli, F. and Sirchia, G. Livingstone.
(1997). The threshold for prophylactic platelet transfusion [15] Patterson, K. (1997). Platelet parameters generated by
in adults with acute myeloid leukaemia, New England automated blood counters, CME Bulletin Haematology,
journal of Medicine, 337, 1870-1875. 1, 13-16.
[2] Gmur, J., Burger, J., Schanz, U., Fehr, J. and Schaffner, A. [16] Kunicka, J. E., Fischer, G., Murphy, J. and Zelmanovic, D.
(1991). Safety of stringent prophylactic platelet transfu- (2000). Improved platelet counting using two-dimen-
sion policy for patients with acute leukaemia, Lancet, 338, sional laser light scatter, American Journal of Clinical
1223-1226. Pathology, 114, 283-289.
[3] Murphy, W. G. (1992). Prophylactic platelet transfusion [17] Kickler, T. S., Rothe, M., Blosser, T., Schiasano, T. and
in acute leukaemia, Lancet, 339, 120. Van Hove, L. (1998). Improving platelet transfusion
[4] Springer, W., von Ruecker, A. and Dickerhoff, R. (1998). therapy using the ImmunoPLT method on the CELL-
Difficulties in Determining Prophylactic Transfusion DYN 4000, Laboratory Heamatology, 4, 80-87.
Thresholds of Platelets in Leukemia Patients, Blood, 92, [18] Ault, K. A., Mitchell, J., Knowles, C. and Van Hove, L.
2183-2184. (1997). Implementation of the immunological platelet
[5] Ancliff, P. J. and Machin, S. J. (1998). Trigger factors count on a haematology analyzer - the Abbott CELL-
for prophylactic platelet transfusion, Blood Reviews, 12, DYN 4000, Laboratory Haematology, 3, 125-128.
234-238. [19] Briggs, C., Harrison, P., Grant, D., Staves, J. and
[6] Norfolk, D. R., Ancliff, P.J. and Contreras, M. et al. (1998). Machin, S. J. (2000). New quantitative parameters on a
Consensus Conference on Platelet Transfusion, Royal recently introduced automated blood cell counter- the
College of Physicians of Edinburgh, 27-28 November XE 2100, Clinical and Laboratory Haematology (in Press).
1997. Synopsis of background papers, British Journal of [20] Dickerhoff, R. and Von Ruecker, A. (1995). Enumeration
Hematologtj, 101, 609~17. of platelets by multiparameter flow cytometry using
[7] England, J. M., Rowan, R. M., Bins, M., Bull, B. S., platelet specifie antibodies and fluorescent platelet
Groner, W., Jones, A. R., Koepke, J. A., Lewis, S. M., particles, Clinical and Laboratory Hematology, 17, 163-172.