Manual and automated reticulocyte counts

Mackelly Simionatto1, Josiane Padilha de Paula2, Michele Ana Flores Chaves3,

Márcia Bortoloso3, Domenic Cicchetti4, Maria Suely Soares Leonart1 and
Aguinaldo José do Nascimento1
1
Programa de Pós-Graduação em Ciências Farmacêuticas da Universidade Federal do Paraná,
Curitiba, PR, Brazil
2
Departamento de Ciências Farmacêuticas, Universidade Estadual de Ponta Grossa, Campus
Uvaranas, Ponta Grossa, PR, Brazil
3
Centro de Ciências Médicas e Farmacêuticas da Universidade Estadual do Oeste do Paraná,
Cascavel, PR, Brazil
4
Department of Biometry, Yale University School of Medicine, New Haven, CT, USA

Manual reticulocyte counts were examined under light microscopy, using the property whereby
supravital stain precipitates residual ribosomal RNA versus the automated flow methods, with the
suggestion that in the latter there is greater precision and an ability to determine both mature and
immature reticulocyte fractions. Three hundred and forty-one venous blood samples of patients
were analyzed of whom 224 newborn and the rest adults; 51 males and 66 females, with ages
between 0 and 89 years, as part of the laboratory routine for hematological examinations at the
Clinical Laboratory of the Hospital Universitário do Oeste do Paraná. This work aimed to compare
manual and automated methodologies for reticulocyte countings and evaluate random and
systematic errors. The results obtained showed that the difference between the two methods was
very small, with an estimated 0.4% systematic error and 3.9% random error. Thus, it has been
confirmed that both methods, when well conducted, can reflect precisely the reticulocyte counts
for adequate clinical use.
Keywords: Hematology, reticulocytes count, statistical analysis

Introduction after supravital staining, contains a minimum of two

Manual methodology is used in small and medium
sized laboratories, due to the lack of appropriate
automated equipment and the low cost in relation to
the automated methodology.1 In medium sized
laboratories, the electronic devices are not always
used due to the cost of specific reagents supplied by
the manufacturer.
The identification of the reticulocytes by both
methods must be in accordance with the standard of
the College of American Pathologists that in 1986
defined the reticulocyte as any anucleate red cell that,
Correspondence to: A. J. Nascimento, Av. Pres. Kennedy, 811, Ap. 61, 80
220-201 Curitiba, PR, Brazil
E-mail: ajnasc@ufpr.br
particles of the stained material, corresponding to
nucleic acid precipitations.2,3
Visual counting of reticulocytes under light micro-
scopy was developed in 1940 and is based on the
property whereby the supravital stain precipitates the
remaining portions of ribosomal RNA.2,4,5 This
method is rapidly being replaced by automated flow
methods, which are considered to have much greater
precision and the ability to determine the maturation
level of the reticulocytes as expressed by the
immature reticulocyte fraction as well as by a variety
of other important factors, which include both the
volume and concentration of hemoglobin.3,6,7
Studies on reticulocyte counting in clinical labora-
tories are of great importance in quality control.

ß W. S. Maney & Son Ltd 2010

Received 30 November 2009; accepted 5 March 2010
406 DOI 10.1179/102453310X12647083621128 Hematology 2010 VOL 15 NO 6

Recently, Simionatto et al.8 analyzed the agreement
among 12 cytologists evaluating 25 slides for manual
reticulocyte counting, and reported a random error of
between 7 and 40%.
Studies on errors and standardization for the
manual methods, and alternatives for the counters
reagents, are relevant. The aim of this work was the
comparison between manual and automated meth-
odologies for reticulocyte countings, evaluation of
random and systematic errors, and a proposal of
alternative dye for Cell Dyn 3500 SL – Abbott
hemocytometer (Abbott Diagnostics, Abbott Park,
IL, USA).
Material and methods
Blood samples
Three hundred and forty-one venous blood samples
were selected of patients, of whom 224 were newborn,
51 were male, and 66 were female. The sample ranged
in age from the newborn up to 89 years. The patients
were examined, as part of standard laboratory
routine for hematological examinations, after
informed consent was obtained by patients or their
responsible others. Approval was given by the Ethical
Committee for Research Involving Humans, Setor de
Saúde da Universidade Federal do Paraná. The
criterion for the sample election was the clinical
request for reticulocyte count. No restriction was
made for sex, age or clinical history of each patient
under clinical care. All the procedures were carried
out in the Laboratory of Clinical Analyses of the
Hospital Universitário do Oeste do Paraná,
Universidade Estadual do Oeste do Paraná.
Visual count reticulocytes under light microscopy
Manual reticulocyte counting was carried out accord-
ing to Lewis et al.9 with modifications as described by
Simionato et al..8 Aliquots of 100 ml of new methylene
blue stain (Laborclin, Curitiba, PR, Brazil) were added
to 200 ml of blood when packed red blood cell volume
(PCV)>30% or 250 ml of blood with PCV(29%, and
incubated at 37uC for 20–25 minutes.
Automated counting
Reticulocyte counting under automated methodology
was made in accordance with the manufacturer’s
instructions (Cell Dyn 3500 SL, Abbott Diagnostics).
Results of the reticulocyte counts were expressed in
percentages.
Commercial and alternative dyes
Comparative reticulocyte counting under automated
methodology (Cell Dyn 3500 SL, Abbott Diagnostics)
was performed on 50 samples. Commercial reagent
Simionatto et al. Reticulocyte counts
was diluted according to the manufacturer’s instruc-
tions. Brilliant cresyl blue 1 g/dl (Sigma-Aldrich
Brasil, São Paulo, SP, Brazil) was diluted in sodium
citrate 3 g/dl, NaCl 0.85 g/dl, pH 7.0, and new
methylene blue 1 g/dl (Icn Biomedicals, Irvine, CA,
USA) in phosphate buffer 150 mmol/l, pH 6.48, after
dilutions of 1 : 10, 1 : 20, 1 : 40, 1 : 80 and 1 : 100.
Statistical analyses
Statistical analyses were accomplished by the appli-
cation of analysis of variance (ANOVA), one way or
two way models, and regression analyses, using Excel
spread sheets (Microsoft) or statistical package
Statistica 8.0 (StatSoft). Statistical significance was
considered to be P(0.05.
Results and discussion
Using the protocols for evaluation of manual and
automated blood cell counts of the National
Committee for Clinical Laboratory Standards and
the International Committee for Standardization in
Haematology, both techniques were evaluated in a
medium sized clinical laboratory at Paraná, Brazil.4
During the experiments, it was found that the
automated method demanded special precautions
such as avoiding bubbles when agitating the sample
in the reagent bottle, incubating sample-stain diluent
suspensions at room temperature, and protecting the
sample from exposure to light. It should be empha-
sized that these precautions are not to be found in the
manufacturer’s manual. For this work, the reaction
occurred after 15 minute incubation and approxi-
mately 1 minute for each measurement. The whole
process was performed over a 2 hour period. Despite
the great number of samples processed, we were able
to both meet the time limit of the standard
determination and to prevent errors, even if the
suspension presented stability, as informed by the
manufacturer.
Comparison for the results of reticulocyte counts
obtained with manual and automated methods was
analyzed by appropriate ANOVA models and
demonstrates that there were no statistically signifi-
cant differences in average performance (P50.21).
The remaining results are described in Table 1.
Figure 1 illustrates the frequency histogram of
reticulocyte counts data for manual and automated
methods. The frequency histogram tells how many
items are in each class of reticulocyte counts. It is
observed that the frequency distribution profile of the
data for both methods was very similar, and that two
different samples were observed, with averages
ranging from 2 to 6% and 10 to 14%. This
Hematology 2010 VOL 15 NO 6 407
Simionatto et al. Reticulocyte counts

Manual method according to Lewis et al. (2006); automated method in Cell Dyn 3500 SL (Abbott Diagnostics); N. of
obs., number of observations. Data for patients (341) were obtained in two independent determinations
Figure 1 Frequency histogram of reticulocyte counts (%) for manual and automated data

observation is clearly due to the population hetero- around the regression line for measurements above
geneity, composed of adult and newborn patients. 10% of the reticulocyte counts.
The samples were composed of 65.4% newborns, and Davis10 pointed out high precision for electronic
34.6% children and adults, and serve to explain the counters. Riley et al.5 showed that the comparison of
major frequency of high reticulocyte counts that has the reticulocyte counting on flow cytometry with the
resulted. According to Riley et al.,5 age and other manual methodology, resulted in an excellent corre-
physiological variations may exert a strong influence lation (R50.984). Similarly, Fernández et al.11
on reticulocyte counts. obtained R50.880 for the two methods, and they
Figure 2 shows the scatterplot of the data for
reticulocyte counts obtained with manual and auto-
mated methods. The regression data indicate a
systematic error of only 0.4% higher for the
automated method relative to the manual method.
The random error calculated as 12R2 was only 3.9%.
It can be observed that the data are slightly dispersed

Table 1 Comparison in reticulocyte counts (%) between

manual and automated methods*

Conf. lim.

Methods N Means SD SE 295% z95% Min. Max.

Totals 682 6.65 4.69 0.18 6.29 7.00 0.40 21.10

Manual 341 6.42 4.62 0.25 5.93 6.92 0.40 18.60
Automated 341 6.87 4.75 0.26 6.36 7.38 0.82 21.10 Manual method according to Lewis et al. (2006); auto-
9 mated method in Cell Dyn 3500 SL (Abbott
*Manual method according to Lewis et al.; automated method
in Cell Dyn 3500 SL (Abbott Diagnostics). Diagnostics); R50.980; R250.961; solid line fitting the
N, sample sizes; SD, standard deviation; SE, standard error of equation (auto50.3917z1.00856manual); dotted lines
means; Conf. lim., 295% confidence limits; Min., minimun value; for 95% confidence limits. Data for patients (341) were
Max, maximum value. obtained in two independent determinations
Data for patients (341) were obtained in two independent Figure 2 Scatterplot for data of reticulocyte counts (%) of
determinations. manual and automated methods

408 Hematology 2010 VOL 15 NO 6


Automated counts in Cell Dyn 3500 SL (Abbott Diagnostics);
solid line fits the regression y50.3698z0.8821x; dotted lines
295% confidence limits; N550 patients in two independent
determinations
Figure 3 Scatterplot for data of reticulocyte counts (%) of
commercial and new methyle blue 1 : 10
also showed higher coefficient of variation coefficient
in reticulopenia.
Figure 3 shows the reticulocyte counts for the use
of new methylene blue dye A diluted 1 : 10, compared
with the commercial reagent resulting in a high R of
0.90. No differences were observed between averages
for the two staining (commercial52.86¡1.40; new
methylene blue 2.89¡1.40). The systematic error
between the two dyes was only 0.4%. Better results
were obtained with new methylene blue than with
brilliant cresyl blue.
This study allow us to conclude that variation
between the two methods was very low, confirming
that, if the execution criteria are strictly followed,
identification and evaluation of reticulocytes counted
Simionatto et al. Reticulocyte counts
among a large number of cells in adequate fields by
visual count microscopy or the use of automated
method, under quality control monitoring, reflect
reticulocyte counts that are precise and trustworthy
for the diagnosis and monitoring of therapies in
patients with hematological illnesses, thereby pro-
moting quality in the clinical laboratory.
References
1 Bain BJ. Células Sanguı́neas: um guia prático, 2nd edn. Porto
Alegre: Artes Médicas, 1997.
2 Pierre RV. Their usefulness and measurement in peripheral blood.
Clin Lab Med 2002; 22: 63–79.
3 Nascimento MLP. Importância do Volume Reticulocitário Médio
para a Evidência da Macrocitose. Newslab 2004; 66: 130–145.
4 Buttarello M, Bulian P, Farina G et al. Flow cytometric
reticulocyte counting parallel evaluation of five fully automated
analyzers: an NCCLS/ICSH approach. Am J Clin Pathol 2001; 115:
100–111.
5 Riley RS, Bem-Ezra JM, Tidwell A, Romagnoli G. Reticulocyte
analysis by flow cytometry and other techniques. Hematol Oncol
Clin North Am 2002; 16: 373–340.
6 Zandecki M, Genevieve F, Gerard J, Gordon A. Spurious counts
and spurious results on haematology analysers: a review. Part II:
white blood cells, red cell indices and reticulocytes. Int J Lab
Haematol 2007; 29: 21–41.
7 Banfi G, Morelli P. Relation between values of haemoglobin,
erythrocytes and reticulocytes and body mass index in elite
athletes of different sports disciplines. Int J Lab Hematol 2008;
29: 484–485.
8 Simionatto M, Paula JP, Leonart MSS, Nascimento AJ, Cicchetti
D. Analysis of manual reticulocyte counting in Clinical
Laboratories of Ponta-Grossa and Campos Gerais, PR, Brazil.
Revista Brasileira de Hematologia e Hemoterapia 2009; 31: 315–
320.
9 Lewis BH, Bain B, Bates I. Dacie and Lewis practical haematology,
9th edn. London: Churchill Livingstone, 2001; 571.
10 Davis BH. Immature reticulocyte fraction (IRF): by any name, a
useful clinical parameter of erythropoietic activity. Lab Hematol
1996; 2: 2–8.
11 Fernández LH, Gallardo A, Garcia G. Evaluación del recuento de
reticulocitos por el sistema Coulter GEN-S. Academia Biomédica
Digital, Facultad de Medicina – Universidade Central da
Venezuela, nu 30, Enero-Marzo, 2007.
Hematology 2010 VOL 15 NO 6 409