JIM-12252; No of Pages 5

Journal of Immunological Methods xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Journal of Immunological Methods

journal homepage: www.elsevier.com/locate/jim

Technical Note

A simple flow-cytometry method to evaluate peripheral blood contamination of bone

marrow aspirates
José Antonio Delgado a, Francisco Guillén-Grima b, Cristina Moreno a, Carlos Panizo c, Carmen Pérez-Robles a,
Juan José Mata a, Laura Moreno a, Paula Arana a, Silvia Chocarro a, Juana Merino a,⁎
a
Department of Immunology, Clínica Universidad de Navarra, University of Navarra, Spain
b
Department of Preventive Medicine, Clínica Universidad de Navarra, University of Navarra, Spain
c
Department of Hematology, Clínica Universidad de Navarra, University of Navarra, Spain

a r t i c l e i n f o a b s t r a c t

Article history: Bone marrow (BM) aspirates used for flow-cytometry (FCM) studies are usually obtained from a second aspira-
Received 28 July 2014 tion, as the primary aspirate is used for morphological assessment. For this reason, the FCM samples unavoidably
Received in revised form 27 December 2016 contain some blood; although, good-quality samples contain only a small amount. It is of utmost importance to
Accepted 27 December 2016
assess the quality of samples prior to FCM analysis; yet, contamination with peripheral blood (PB) is not evalu-
Available online xxxx
ated in most laboratories, possibly because the methods available are either qualitative or too complex for
Keywords:
daily practice. Here, we propose a simple FCM method to quantitatively evaluate PB contamination in BM aspi-
PB contamination rates, by analyzing the percentage of plasma cells and CD34+ cells – two cell populations nearly absent from
BM aspirates PB – and CD10+ granulocytes, which comprise the majority of the PB granulocyte population. We analyzed
Flow cytometry these three populations in 122 BM aspirates from subjects without hematological disease, and identified samples
Quality with PB contamination by performing a hierarchical cluster analysis. A discriminant analysis yielded a function,
Samples which we named the PB contamination index (PBCI). This index value gives a quantitative indication about the
degree of hemodilution of a given sample. A threshold was identified that discriminates low-quality samples.
The method and the threshold proved to be useful in BM aspirates infiltrated with malignant cells, with the ex-
ception of cases where hematological disease altered two of the three parameters included in the index. We have
easily implemented the PBCI calculation in our daily routine, and find it very helpful for an accurate interpretation
of FCM results in a large proportion of BM specimens. Limitations of the technique are discussed.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction
Bone marrow (BM) aspirates are routinely used in flow-cytometry
(FCM) laboratories for the diagnosis and follow-up of hematological
malignancies. There is a certain degree of hemodilution with this kind
of specimen, since it is usually obtained from a second aspiration with
a larger volume, as the first-pull is used entirely for morphological as-
sessment. For this reason, the percentages of BM cell populations ob-
tained from morphologic analysis are often different from those based
Abbreviations: B-CLPD, B-cell chronic lymphoproliferative disorders; BM, bone
marrow; CONT, contaminated group; FCM, flow-cytometry; GC, good-quality group;
MGUS, Monoclonal gammopathy of undetermined significance; MM, Multiple myeloma;
MRD, minimal residual disease; PB, peripheral blood; PBCI, peripheral blood
contamination index; PC, plasma cells; SSC/FSC, side scatter/forward scatter.
⁎ Corresponding author at: Department of Immunology, Clínica Universidad de
Navarra, Avda Pío XII s/n, 31008 Pamplona, Spain.
E-mail address: jmerino@unav.es (J. Merino).
on FCM analyses (Smock et al., 2007; Rawstron et al., 2008). Although
a certain amount of peripheral blood (PB) is unavoidable, there is a
limit for the sample to be representative of the BM. Low-quality BM
aspirates should be discarded, especially if they are to be used for
some specific FCM applications, such as minimal residual disease
(MRD) quantitation (Rawstron et al., 2008; Brooimans et al., 2009;
Gupta et al., 2009).
Despite its importance, PB contamination of BM aspirates is not rou-
tinely evaluated in most laboratories. Some consensus protocols recom-
mend a morphological analysis of the sample, while other guidelines do
not mention this issue when addressing the quality of BM aspirates
(Rawstron et al., 2008; Kalina et al., 2012; Johansson et al., 2014). The
morphological approach poses two difficulties. First, many FCM labora-
tories do not have morphological techniques available. Second, it is a
qualitative approach that can identify BM aspirates with a high level
of hemodilution, but it does not provide quantitative information
about lesser degrees of PB contamination. A system to estimate the
level of PB contamination of samples would be very helpful, to be

http://dx.doi.org/10.1016/j.jim.2016.12.006
0022-1759/© 2016 Elsevier B.V. All rights reserved.

Please cite this article as: Delgado, J.A., et al., A simple flow-cytometry method to evaluate peripheral blood contamination of bone marrow
aspirates, J. Immunol. Methods (2016), http://dx.doi.org/10.1016/j.jim.2016.12.006
2 J.A. Delgado et al. / Journal of Immunological Methods xxx (2016) xxx–xxx
aware of the degree of underestimation in the quantitated BM
populations.
To the best of our knowledge, only one described method quantita-
tively estimates PB contamination in BM samples (Holdrinet et al.,
1980). Some authors have found this method reliable (Brooimans et
al., 2009); however, it is not widely used, possibly because a paired sam-
ple of PB is needed. In the present study, we propose a simple FCM
method for quantitative evaluation of the degree of PB contamination
in BM aspirates. This method does not require complex calculations or
additional samples.
2. Materials and methods
2.1. Samples
A total of 188 BM aspirates were included in this study, 122 from sub-
jects without hematological disease (BM1-122) and 66 from patients
with hematological disease showing BM infiltration: 11 with Monoclonal
Gammopathy of Undetermined Significance (MGUS) (BM123-133); 35
with Multiple Myeloma (MM) (BM134-168); and 20 with B-Cell Chronic
Lymphoproliferative Disorders (B-CLPD) (6 Follicular Lymphoma, 5
Diffuse Large B-Cell Lymphoma, 3 Chronic Lymphocytic Leukemia, 2
Marginal Zone Lymphoma, and 4 Lymphoplasmacytic Lymphoma)
(BM169-188). All samples corresponded to the material remaining after
a routine study for the diagnosis or follow-up of hematological malignan-
cies. The samples were the second aspiration of BM, since in our hospital
the first pull is used entirely to make smears for diagnostic morphologic
evaluation (May Grünwald-Giemsa). The leukocyte count of samples
was obtained in a Sysmex XN-1000™ Hematology Analyzer.
2.2. Immunophenotyping of samples
We analyzed CD34+ cells, plasma cells (PC), and neutrophils
(CD10+ granulocytes). Since CD34+ cells and PC are nearly absent in
PB, contamination of BM samples with blood results in diminished BM
percentages of both populations. On the contrary, neutrophils are the
main granulocytes in PB; thus, PB contamination of BM samples
would increase the percentage of CD10+ granulocytes (CD10+ G).
The monoclonal antibody (mAb) combination used was:
anti-CD38 FITC /anti-CD34 PE /anti-CD138PerCP-Cy5.5/anti-CD10APC/anti-
CD45APC-H7. The mAbs used were: anti-CD38FITC, clone HIT2
(Pharmingen, Becton-Dickinson (BD), San Jose, CA), anti-CD34PE,
clone 8G12 (BD), anti-CD138PerCP-Cy5.5, clone MI15 (BD), anti-CD10APC,
clone HI10a (BD), and anti-CD45APC-H7, clone 2D1 (BD). Two million
leukocytes in 100 μl of BM suspension were stained with mAbs for
15 min at room temperature in the dark. After staining, FACS lysing
solution (2 ml, BD) was added. After 10 min at room temperature, sam-
ples were centrifuged and resuspended in 500 μl of phosphate buffered
saline. The samples were immediately acquired in a FACSCanto II flow
cytometer (BD) equipped with FACSDiva software (BD) (see Supple-
mentary material for calibration details). A minimum of 3 × 105 leuko-
cytes were acquired. Events acquired during the first 10–15 s for each
tube were not recorded to avoid turbulence in the leading edge of
samples. Data were analyzed using Infinicyt™ software, version 1.6
(Cytognos, Salamanca, Spain).
Leukocytes were gated on side scatter/forward scatter (SSC/FSC)
and SSC/CD45 dot plots. The CD34+ cells were identified as CD34+
CD45+ lo cells and PC as CD38++ CD138+ cells. Both cell populations
were expressed as a percentage of leukocytes. The granulocyte pop-
ulation was gated on SSC/FSC and SSC/CD45 dot plots (CD45+ SSChi
cells). Then, CD10+ neutrophils were identified and expressed as a
percentage of the granulocyte population (CD10+ G) (Domingo et
al., 2010; Sandes et al., 2013) (see Supplementary material for dot
plots).
Please cite this article as: Delgado, J.A., et al., A simple flow-cytometry method to evaluate peripheral blood contamination of bone marrow
aspirates, J. Immunol. Methods (2016), http://dx.doi.org/10.1016/j.jim.2016.12.006
2.3. Statistical analyses
Statistical analyses were performed with SPSS Statistics software,
version 20 (IBM). Differences between groups of samples were evaluat-
ed by an unpaired two-tailed t-test. The Mann-Whitney U-test was used
to compare the ratio M/C between groups. P b 0.05 was considered sta-
tistically significant in all cases.
Once normal samples were classified into the two groups, values
of the three parameters from all samples were analyzed by discrim-
inant analysis, a multivariant technique that identifies differences in
independent variables between groups, and gives a canonical func-
tion to discriminate these differences. This discriminant function
was obtained using unstandardized coefficients without missing
data imputation, and was termed the PB contamination index
(PBCI). The PBCI values from MGUS and MM samples were calculated
without including the percentage of PC in the formula, since it is in-
creased in these diseases.
3. Results
3.1. Classification of normal BM aspirates according to the quality of
samples
We analyzed the percentage of CD10+ G, CD34+ cells, and PC in
122 BM aspirates from subjects without hematological disease
(Fig. 1). We compared the results with the reference values of our
laboratory, which are consistent with published reference ranges
(Ocqueteau et al., 1998; Sandes et al., 2013) and classified samples
in two groups. Twenty-seven samples simultaneously showed an
increased percentage of CD10+ G and a decreased percentage of PC
and CD34+ cells and were considered contaminated with PB (CONT).
The rest of the samples (n = 95) formed the good-quality (GC) group.
The differences between groups were statistically significant in the
three parameters analyzed (P b 0.0001).
As expected, CONT samples showed significantly lower leukocyte
counts than GC specimens (Fig. 2A). In the samples in which morpho-
logical data were available (n = 57), we analyzed the disparity between
the percentage of PC obtained from FCM and morphological studies. The
disparity between both techniques was significantly higher in the group
of contaminated samples (Fig. 2B).
3.2. Quantitative analysis of PB contamination in BM aspirates
To develop a formula for classification of any given sample as GC
or CONT, we performed a discriminant analysis with the three pa-
rameters analyzed, as explained in the Materials and methods. The
function obtained, named the PBCI (PB contamination index), was
as follows:



PBCI ¼ −3:052 þ 0:065  %CD10þ G −0:609  %CD34þ −2:008
 ð%PCÞ:
As shown in Fig. 3, the PBCI function performed in a linear fash-
ion when an increasing amount of PB was added to a given BM
sample.
A BM aspirate with the three parameters values falling in the average
range of reference values would show a PBCI value about − 1.00;
whereas, the value for PB is approximately 3.4 (Fig. 3). As is usual in
this kind of analysis, the mean between both values (1.2) was arbitrarily
chosen as the cutoff for excessive PB in a BM aspirate. The validity of this
cutoff was tested in the 122 samples analyzed. All samples from the
CONT group had values ≥ 1.2; while, in the GC group, all samples but
one had values b 1.2 (Fig. 4).
 J.A. Delgado et al. / Journal of Immunological Methods xxx (2016) xxx–xxx 3

Fig. 1. Percentage of CD10+ granulocytes (CD10+ G), CD34+ cells, and plasma cells (PC) in BM aspirates from subjects without hematological disease, classified in two groups according to
reference values. Twenty-seven samples simultaneously showed an increased percentage of CD10+ G and a decreased percentage of PC and CD34+ cells, and therefore were considered
contaminated with PB (CONT). The rest of the samples formed the good-quality (GC) group (n = 95). The table shows the mean (standard deviation) of both groups, as well as the
reference values of our laboratory.

3.3. Application of the PBCI to evaluate the quality of BM aspirates from
patients with hematological disease
We calculated the PBCI value for 66 BM aspirates infiltrated with
malignant cells due to hematological disease, as described in the
Materials and methods. Similar to the analysis with controls, we com-
pared the results for CD10+ G, CD34+ cells, and PC for each pathological
sample with reference values. Nineteen samples simultaneously
showed an increased percentage of CD10+ G and a decreased percent-
age of PC and CD34+ cells. Therefore, they were classified into the
CONT group. The remaining 47 samples formed the GC group. Then,
we assessed the PBCI values for both groups. With the exception of
one sample (PBCI = 1.1), the CONT samples had PBCI values ≥ 1.2. All
GC samples had values b 1.2 (Fig. 5).
4. Discussion
As we explained in the Introduction, the evaluation of PB contamina-
tion in BM samples used for FCM studies remains a concern. The most
used method for this purpose is the morphological analysis of the sam-
ple prior to FCM. However, this is mainly suitable for the identification
of samples highly contaminated with PB (Rawstron et al., 2008;

Fig. 2. A) Leukocyte counts in BM aspirates from subjects without hematological disease, classified into good-quality samples (GC) (n = 95) and those contaminated with PB (CONT) (n =
27), as described in the Materials and methods. B) Ratio between the percentage of plasma cells obtained by morphological analysis (M) and by flow cytometry (C) in GC (n = 45) and
CONT (n = 12) BM aspirates without hematological disease.

Please cite this article as: Delgado, J.A., et al., A simple flow-cytometry method to evaluate peripheral blood contamination of bone marrow
aspirates, J. Immunol. Methods (2016), http://dx.doi.org/10.1016/j.jim.2016.12.006
4 J.A. Delgado et al. / Journal of Immunological Methods xxx (2016) xxx–xxx
Fig. 3. Peripheral blood contamination index (PBCI) values of a BM aspirate from a subject
without hematological disease, serially diluted into PB from the same subject. Dilutions
ranged from 1:10 (one part of PB and 9 parts of BM) to 10:10 (pure blood). The PBCI
value of the BM sample was −0.06.
Johansson et al., 2014). It is more difficult to determine that a given sam-
ple shows some degree of hemodilution and, thus, the percentages ob-
tained for the populations studied are underestimated. In the present
work, we describe a method that evaluates the degree of PB contamina-
tion in BM aspirates, simply by analyzing the percentage of CD10+ G,
CD34+ cells, and PC. Using a plain mathematical function, the results
are used to obtain a value that we have named the PBCI (peripheral
blood contamination index). A PBCI of 1.2 was demonstrated to be a
useful threshold to distinguish between GC samples and CONT samples
in BM aspirates from patients with and without hematological disease.
We analyzed 122 normal samples in the present work; of these, all
CONT samples had values ≥ 1.2 and only one of the 95 GC samples had
a PBCI higher than this threshold (1.55) (Fig. 4). Of the 66 pathological
samples analyzed, all GC samples had PBCI values b 1.2 and only one
of the 19 CONT samples had a PBCI value lower than this threshold
(Fig. 5). Thus, this cutoff discriminates BM aspirates with an unaccept-
able degree of PB contamination with a sensitivity and specificity higher
than 95%, in normal and pathological samples.
Compared with morphological analysis, this method is useful for de-
tecting lower levels of PB contamination and has the advantage of
4
3 (Holdrinet et al., 1980), our approach is more user-friendly and does
2 based on markers frequently included in mAb panels for routine FCM di-
1 neither expensive nor time-consuming.
PBCI
0 subjects with hematological disease affecting the parameters used
-1
-2
-3
-4
GC CONT
Fig. 4. Peripheral blood contamination index (PBCI) values of BM aspirates from subjects
without hematological disease, classified into good-quality samples (GC) and those
contaminated with PB (CONT), as described in the Materials and methods. All CONT
samples (n = 27) had PBCI values ≥ 1.2. In the GC group, 94 samples had values b 1.2
and 1 sample had a value of 1.55.
Please cite this article as: Delgado, J.A., et al., A simple flow-cytometry method to evaluate peripheral blood contamination of bone marrow
aspirates, J. Immunol. Methods (2016), http://dx.doi.org/10.1016/j.jim.2016.12.006
PBCI
0
-1
-2
-3
GC CONT
Fig. 5. Peripheral blood contamination index (PBCI) values of BM aspirates infiltrated with
malignant cells due to hematological disease (MGUS, MM, and B-CLPD), classified into
good-quality (GC) and PB contaminated (CONT) samples, according to reference values.
All GC samples (n = 47) had PBCI values b 1.2. In the CONT group, 18 of 19 samples had
values ≥ 1.2 and 1 sample had a PBCI value of 1.1.
providing quantitative results. The higher the PBCI value, the higher
the degree of hemodilution of the specimen and, consequently, the de-
gree that the obtained percentages for BM cell populations are
underestimated. This information about the quality of BM samples is
very useful in a clinical FCM laboratory for the correct interpretation
of results. The calculation of PBCI has been included in our daily routine
and we find it extremely helpful, especially in studies such as MRD
quantitation. We do not support recalculating cytometric results in
hemodiluted samples, as proposed by others (Loken et al., 2009),
since the accuracy of the recalculated value cannot be guaranteed. We
prefer to inform clinicians about the PBCI value of low-quality samples.
Depending on the degree of hemodilution, the number of malignant
cells detected, and the specific type of study, they can reliably decide if
a new specimen is necessary.
The PBCI value also provides invaluable information to elucidate dis-
crepancies between techniques. The large magnitude of the difference
between the percentage of PC obtained with morphological techniques
or by FCM, extensively reported in literature (Smock et al., 2007;
Rawstron et al., 2008; Gupta et al., 2009), is significantly reduced
when samples of low-quality are excluded from the analysis (Fig. 2B).
Compared with the other quantitative method previously described
not require a paired sample of PB. It is a straightforward method
agnosis and follow-up of hematological malignancies. Therefore, it is
It is important to discuss its applicability to BM aspirates from
for the calculation of the index. If the disease induces abnormal
values for one parameter (e.g. multiple myeloma samples), the
index is calculated with the other two, and the validity of the thresh-
old is maintained (Fig. 5). However, diseases affecting two of the
three parameters of the formula must be considered a limitation of
the approach. Accordingly, the method can be applied to a large pro-
portion of diagnostic specimens, with the exception of samples from
patients with myelodysplastic syndromes due to the increase in the
percentage of CD34+ cells and concomitant decrease of neutrophils
that lower the PBCI and make it not applicable. For the same reason,
the method cannot be applied for MRD studies in post-induction
samples from acute leukemia patients. However, it is very useful
for MRD assessment in MM patients, where analyses are performed
sequentially during follow-up and the assessment of the quality of
 J.A. Delgado et al. / Journal of Immunological Methods xxx (2016) xxx–xxx
samples is essential (Mailankody et al., 2015). Three MRD + samples
from MM patients were included in this study: BM151, with 0.0002%
myelomatous PC and an PBCI value of − 0.5; BM156, with 0.005%
MRD and an PBCI value of 1.3; and BM158, with 0.01% MRD and an
PBCI value of 2.8. The information provided by PBCI allowed us to in-
terpret the MRD results from these samples as accurate, slightly
underestimated, and significantly underestimated, respectively.
In conclusion, we describe a very simple FCM method that reliably
evaluates the degree of PB contamination in a large proportion of BM as-
pirates received in a clinical FCM laboratory. This method is especially
useful for studies where accuracy is essential, such as MRD studies.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.jim.2016.12.006.
References
Brooimans, R.A., Kraan, J., van Putten, W., Cornelissen, J.J., Löwenberg, B., Gratama, J.W.,
2009. Flow cytometric differential of leukocyte populations in normal bone marrow:
influence of peripheral blood contamination. Cytometry B Clin. Cytom. 76, 18–26.
Domingo, E., Moreno, C., Sánchez-Ibarrola, A., Panizo, C., Páramo, J.A., Merino, J., 2010. En-
hanced sensitivity of flow cytometry for routine assessment of minimal residual dis-
ease. Haematologica 95, 691–692.
Gupta, R., Bhaskar, A., Kumar, L., Sharma, A., Jain, P., 2009. Flow cytometric
immunophenotyping and minimal residual disease analysis in multiple myeloma.
Am. J. Clin. Pathol. 132, 728–732.
Please cite this article as: Delgado, J.A., et al., A simple flow-cytometry method to evaluate peripheral blood contamination of bone marrow
aspirates, J. Immunol. Methods (2016), http://dx.doi.org/10.1016/j.jim.2016.12.006
5
Holdrinet, R.S., von Egmond, J., Wessels, J.M., Haanen, C., 1980. A method for quantifica-
tion of peripheral blood admixture in bone marrow aspirates. Exp. Hematol. 8,
103–107.
Johansson, U., Bloxham, D., Couzens, S., Jesson, J., Morilla, R., Erber, W., Macey, M., British
Committee for Standards in Haematology, 2014. Guidelines on the use of multicolour
flow cytometry in the diagnosis of haematological neoplasms. Br. J. Haematol. 165,
455–488.
Kalina, T., Flores-Montero, J., van der Velden, V.H., Martin-Ayuso, M., Böttcher, S., Ritgen,
M., Almeida, J., Lhermitte, L., Asnafi, V., Mendonça, A., et al., 2012. EuroFlow standard-
ization of flow cytometer instrument settings and immunophenotyping protocols.
Leukemia 26, 1986–2010.
Loken, M.R., Chu, S., Fritschle, W., Kalnoski, M., Wells, D.A., 2009. Normalization of bone
marrow aspirates for hemodilution in flow cytometric analyses. Cytometry B Clin.
Cytom. 76B, 27–36.
Mailankody, S., Korde, N., Lesokhin, A.M., Lendvai, N., Hassoun, H., Stetler-Stevenson, M.,
Landgren, O., 2015. Minimal residual disease in multiple myeloma: bringing the
bench to the bedside. Nat. Rev. Clin. Oncol. http://dx.doi.org/10.1038/nrclinonc.
2014.239.
Ocqueteau, M., Orfao, A., Almeida, J., Blade, J., Gonzalez, M., Garcia-Sanz, R., Lopez-Berges,
C., Moro, M.J., Hernandez, J., Escribano, L., et al., 1998. Immunophenotypic character-
ization of plasma cells from monoclonal gammopathy of undetermined significance
patients. Implications for the differential diagnosis between MGUS and multiple
myeloma. Am. J. Pathol. 152, 1655–1665.
Rawstron, A.C., Orfao, A., Beksac, M., Bezdickova, L., Brooimans, R.A., Bumbea, H., Dalva, K.,
Fuhler, G., Gratama, J., Hose, D., et al., 2008. Report of the European Myeloma Net-
work on multiparametric flow cytometry in multiple myeloma and related disorders.
Haematologica 93, 431–438.
Sandes, A.F., Kerbauy, D.M.B., Matarraz, S., Chauffaille, M.L.L.F., López, A., Orfao, A.,
Yamamoto, M., 2013. Combined flow cytometric assessment of CD45, HLA-DR,
CD34, and CD117 expression is a useful approach for reliable quantification of blast
cells in myelodysplastic syndromes. Cytometry B Clin. Cytom. 84B, 157–166.
Smock, K.J., Perkins, S.L., Bahler, D.W., 2007. Quantitation of plasma cells in bone marrow
aspirates by flow cytometric analysis compared with morphologic assessment. Arch.
Pathol. Lab. Med. 131, 951–955.