DIAGNOSTICS

Advanced MAPSS™ Technology of Alinity hq

Analyzer: A Multi-Platform Comparison
Author: Zainab Mukhtar

Abbott Diagnostics, Hematology, Santa Clara, CA 95054

DIAGNOSTICS
ABSTRACT laboratory strives to offer the best diagnostic and
monitoring tools to its patient population.
This study evaluated the performance of the Abbott
Alinity hq, a high-throughput hematology analyzer. The purpose of this evaluation was to compare the
performance of the Abbott Alinity hq hematology
Five hundred and fifty-one samples were selected for analyzer to the Siemens Advia 2120i and the
method comparison with the Beckman Coulter DxH Beckman Coulter DxH 800, and to assess the
800 and the Siemens Advia 2120i. In addition, these correlation and agreement of Alinity hq WBC
samples were used to perform a correlation study of differential with results obtained by manual
Alinity hq white blood cell (WBC) differentials with microscopic WBC differential (reference method).
manual microscopic WBC differentials, including
immature granulocytes (IGs) and nucleated red blood Alinity hq from Abbott is a high-throughput
cells (NRBCs). hematology analyzer. It is a fully optical system
that reports 29 parameters and offers a set of
The evaluation demonstrated a high level of Research Use Only (RUO) measurands. The system
correlation and agreement between the three reports a 6-part WBC differential that includes IG
analyzers for the majority of CBC and WBC concentration. An NRBC count is provided with
differential parameters, except for basophils. In every WBC differential without using additional
addition, a bias was observed between Alinity hq reagents or an alternate test mode. The operating
and Advia 2120i for % monocytes, as Advia 2120i principle is based on an advanced version of Abbott’s
underestimated the monocyte count compared patented Multi Angle Polarized Scatter Separation
to manual differential. Mean platelet volume (MAPSS™) Technology, employing optical and
(MPV) also showed weak correlation between the fluorescence flow cytometry methods. A total of 8
two analyzers. A bias was also observed between detectors including seven light detectors (axial light
Alinity hq and DxH 800 red cell distribution width loss, polarized side scatter, depolarized side scatter,
(RDW), likely due to differences in the technology and four intermediate angles of light scatter) in
used for red blood cell (RBC) count and volume combination with a fluorescence detector, create a
measurements. Alinity hq WBC differential showed a unique signal signature of each cell, allowing for the
high level of agreement with the manual differential differentiation of WBC subpopulations2 as well as
results, including strong correlation for %IG (r=0.88) reliable separation of platelets (PLTs) and RBCs3,4.
and NRBC (r=0.99).
The technology of Alinity hq is compared to the
Overall, Alinity hq demonstrated equivalent Advia 2120i and the DxH 800 in Table 1.
performance with both DxH 800 and Advia 2120i
hematology analyzers, and a high level of accuracy Table 1. Alinity hq, Advia 2120i and DxH 800 technology comparison
compared to manual differentials.
Alinity hq Advia 2120i DxH 800
INTRODUCTION Measurand
(Abbott) (Siemens) (Beckman Coulter)
Cyanide-free Cyanide-free Cyanide-free
Medical biology has been going through HGB absorption absorption absorption
transformative changes in the public and private photometry photometry photometry
Optical: light Optical: light Impedance
sectors of the French health care system since RBC
scatter scatter (Coulter Principle)
legislative reforms in 2010. These reforms Optical: light Optical: light Impedance
drive laboratory consolidation and new quality PLT
scatter scatter (Coulter Principle)
standards based on accreditation1. The evaluation Advanced Light Scatter VCSn: Volume
was performed at MEDILAB66, one of the 19 MAPSS™ (two angles), (Impedance),
subsidiaries of the Inovie Group. MEDILAB66 technology: differential Conductivity
WBC and
Light scatter WBC lysis, (high frequency
comprises multiple sites organized in a network Differential
(seven angles) cytochemical current), Scatter
around Perpignan, France. It has a team of 19 clinical and fluorescence (myeloperoxi- (multiple angles)
pathologists and serves almost 450,000 patients per flow cytometry dase) staining
year with three specialized departments. The

1
DIAGNOSTICS
METHODS agreement between the two instruments for WBC,
%NEU and %LYM. Moderate correlation was
Approximately 50 random samples were tested daily obtained for %EOS, and weak correlation for %BASO
for 10 days on all three instruments in closed mode. (Table 2). In addition, good correlation was obtained
The samples were from two laboratories of the for NRBC, expressed as the number of NRBC per 100
MEDILAB66 group and consisted of both outpatients WBC (NR/W) (r=0.87), with Alinity hq showing a
and inpatients. positive bias compared to Advia 2120i (Table 2).
All samples were leftover, de-identified specimens
Table 2. Alinity hq vs Advia 2120i method comparison results
stored at room temperature. Analysis was completed
within 8 hours of blood collection. A peripheral
Range tested Passing-Bablok
blood smear (PBS) was made for each sample and (Alinity hq) Pearson’s
Measurand N
stained per a standard May-Grünwald-Geimsa Inter- r
Min Max Slope cept
protocol within 2 hours after testing on all three
analyzers. PBS analysis was performed by two RBC
(x 1012/L) 470 2.13 5.89 1.02 -0.07 0.99
experienced hematologists, each performing a 100-
MCV 470 64.60 116.00 0.96 2.72 0.98
cell manual WBC differential. The average of the two (fL)
readings was used as reference. When there was a RDW 470 10.70 25.70 0.96 -0.78 0.97
disagreement for a positive smear finding (e.g., blasts (%)
present or absent), a third reader was involved to HCT 470 19.20 55.70 1.04 -1.94 0.99
(%)
resolve the discrepancy.
HGB 470 6.60 18.40 1.00 0.00 1.00
(g/dL)
A total of 551 samples were tested. Not all results
PLT 460 12.20 847.00 1.00 -1.00 0.99
were available on every sample; in addition, (x 109/L)
invalidated results were excluded on a measurand- MPV 465 5.07 15.60 0.97 -0.134 0.33
by-measurand basis. Therefore, the number of (fL)
samples in each comparison may be different. WBC 459 0.85 71.40 1.03 0.04 1.00
(x 109/L)
Quantitative data analysis was performed according
to CLSI EP09-A3, using either Passing-Bablok or %NEU* 442 5.11 95.98 1.01 -1.93 0.99
Deming regression models with estimation of slope %LYM 441 2.62 86.50 1.03 0.52 1.00
and intercept to determine the agreement between
two analyzers, or between Alinity hq results and %MONO 438 0.92 59.20 1.47 -0.40 0.90
manual WBC differential. Bland-Altman mean bias
was also calculated for the comparison of Alinity %EOS 435 0.01 21.90 1.04 -0.20 0.83
hq and manual WBC differential. Correlation was %BASO 394 0.01 5.82 1.25 -0.29 0.37
assessed by determining the Pearson’s correlation
coefficient (r). Analyse-it for Microsoft Excel NR/W** 470 0.00 84.30 1.43 -0.38 0.87
software version 5.11.3 was used for the analysis.
*Advia 2120i does not enumerate and report Immature Granulocyte (IG) param-

RESULTS
eter. Alinity %NEU and %IG were combined for comparison with Advia 2120i
%NEU
**NR/W: Deming fit

Abbott Alinity hq vs Siemens Advia 2120i Despite the strong correlation for %MONO, the
A wide range of values was tested for each CBC slope of 1.47 suggested a significant bias between
measurand, covering the majority of the analytical Alinity hq and Advia 2120i results. To investigate
measuring ranges (AMR). Strong correlation and the cause of the bias, we assessed the correlation
agreement was obtained for all RBC parameters, as and agreement between the Advia 2120i and manual
demonstrated by correlation coefficients and Passing- %MONO results, and obtained a slope of 0.59,
Bablok slope and intercept values (Table 2). implying that the Advia 2120i underestimated the
monocyte count (Figure 1). The Advia 2120i reports
There was also high level of correlation and
a category of WBC called Large Unstained Cells

2
DIAGNOSTICS
(LUCs) in addition to the standard 5-part differential. r=0.99 r=0.97

We hypothesized that some monocytes might have

been included in this category. When %MONO and
%LUC were combined, and compared to Alinity hq
%MONO, the bias significantly decreased (slope=1.12;
r=0.89) (Figure 1). These results suggest that some
monocytes were classified by the Advia 2120i as LUCs.

Figure 2: Regression graphs for PLT comparison between Alinity hq and Advia
r=0.90 r=0.89 2120i.
Blue line: Identity; Black line: Passing-Bablok fit
A) PLT regression graph for all samples (n=460)
B) PLT regression graph for samples with PLT count of < 200 x 109/L (n=160)

Abbott Alinity hq vs Beckman Coulter DxH 800

Similar to the Alinity hq vs Advia 2120i comparison,
close correlation and agreement were obtained for
all RBC parameters in the Alinity hq vs DxH 800
r=0.89
comparison, as demonstrated by the slope, intercept
and correlation coefficient values (Table 3). However,
the comparison of RDW (measured as %CV) produced
a slope of 0.83, despite the excellent overall agreement
between MCV results (Table 3 and Figure 3). Bland-
Altman analysis revealed a -1.96% mean bias between
the two methods (Figure 3).

Table 3. Alinity hq vs DxH 800 method comparison results

Figure 1: Regression graphs for %Monocytes comparison between Alinity
hq, Advia 2120i and manual WBC differential. Range tested
Blue line: Identity; Black line: Passing-Bablok fit (Alinity hq) Passing-Bablok
A) Regression graph of Alinity hq %MONO versus Advia 2120i Measurand N Pearson’s r
Min Max Slope Inter-
%MONO. cept
B) Regression graph of Advia 2120i %MONO versus manual WBC
differential %MONO. RBC 479 2.13 5.89 1.02 -0.02 0.99
(x 1012/L)
C) Regression graph of Alinity hq %MONO versus Advia 2120i
%MONO + %LUC MCV (fL) 479 64.60 116.00 0.98 1.42 0.98
RDW (%) 479 10.70 25.70 0.83 0.63 0.97
Strong correlation (r=0.99) and agreement was HCT (%) 478 19.20 53.10 1.02 -0.49 0.99
observed between Alinity hq and Advia 2120i PLT
HGB 479 6.60 18.44 1.02 -0.15 1.00
counts across the AMR, but correlation was weak for (g/dL)
MPV (r=0.33) (Table 2 and Figure 2-A). To focus on PLT 467 19.40 847.00 1.04 -3.26 0.99
(x 109/L)
the clinically important low PLT counts, a separate
comparison was performed on a subset of samples MPV (fL) 470 5.17 12.20 0.78 2.24 0.80
with PLT counts of < 200 x 109/L (n=160), and the WBC 474 0.29 79.00 1.01 0.03 1.00
(x 109/L)
results demonstrated good agreement between the
%NEU* 439 0.53 95.98 1.00 -0.54 0.99
two instruments with a slope of 1.06 and an r value of
0.97 (Figure 2-B). %LYM 439 2.62 93.50 1.00 0.84 0.99
%MONO 433 0.58 61.80 0.99 0.20 0.93
%EOS 439 0.00 21.90 1.02 -0.05 0.98
%BASO 436 0.00 5.82 0.83 -0.24 0.57
NR/W** 439 0.00 7.35 1.21 -0.12 0.81

*DxH 800 does not enumerate and report Immature Granulocyte (IG) parameter.
Alinity hq %NEU and %IG were combined for comparison with DxH 800 %NEU.
**NR/W: Deming fit
3
DIAGNOSTICS
r=0.98 r=0.97
Alinity hq vs manual WBC differential
Alinity hq results closely matched the manual %NEU

Alinity hq RDW
and %LYM results; moderate correlation was observed
for %MONO and %EOS, and weak correlation
for %BASO (Table 4 and Figure 5). Bland-Altman
statistics showed a small negative bias for %NEU, and
B DxH 800 RDW a small positive bias for %LYM and %MONO.
Table 4. Alinity hq vs manual WBC differential method comparison results

Range tested
Alinity hq Mean RDW

Passing-Bablok
Measu- (Alinity hq) Pear- Bland-Altman Bias
N
rand Inter- son’s r (95% CI)
Min Max Slope cept
%NEU* 470 0.10 93.40 1.06 -6.37 0.95 -3.23 (-3.71 to -2.75)

%LYM 470 2.62 93.50 1.07 -0.45 0.96 2.03 (1.61-2.46)

DxH 800
%MONO 462 0.58 61.80 1.00 1.40 0.83 1.57 (1.26 to 1.87)
Figure 3: MCV and RDW regression and Bland-Altman plots %EOS 470 0.00 21.90 0.98 0.21 0.78 0.20 (0.03 to 0.37)
A and B: Blue line: Identity; Black line: Passing-Bablok fit. C: Blue solid
line: Mean; Dashed lines: Limits of agreement; Black solid line; Identity %BASO 466 0.00 5.82 0.73 0.03 0.48 -0.18 (-0.24 to -0.12)
A) Regression graph of Alinity hq MCV versus DxH 800 MCV
%IG** 470 0.00 26.70 0.83 0.08 0.88 -0.54 (-0.73 to -0.35)
B) Regression graph of Alinity hq RDW versus DxH 800 RDW
C) Bland-Altman plot for Alinity hq RDW versus DxH 800 RDW. Mean NR/W** 473 0.00 84.30 1.35 0.06 0.99 0.03 (-0.02 to 0.09)
Difference -1.96 (95%CI) (-2.043 to -1.880)
Differences are plotted against the averages of the results obtained with *Alinity %NEU and %IG were combined for comparison with manual %NEU
**NR/W: Deming fit
the two analyzers.
r=0.83
Strong correlation and agreement were also obtained r=0.96

for WBC and most WBC differential parameters,

including %MONO. However, a weak (but greater
than the Alinity hq vs Advia 2120i comparison)
correlation was observed for %BASO results.
High level of correlation and agreement was also
demonstrated for NR/W results (r=0.81) (Table 3).
Assessment of PLT results yielded strong correlation r=0.78 r=0.48
and agreement between the two instruments both
for the entire sample population, and on a subset of
samples with PLT counts < 200 x 109/L, analyzed
separately (n=167) (Figure 4). MPV values also
showed a high level of agreement (Table 3).

r=0.98 r=0.99
r=0.95
Alinity hq %Neutrophil

A DxH 800 Platelet B DxH 800 Platelet

E Manual % Neutrophil

Figure 4: Regression graphs for PLT comparison between Alinity hq vs Figure 5: Regression graphs for WBC differential comparison between
DxH 800. Alinity hq and manual result
Blue line: Identity; Black line: Passing-Bablok fit Blue line: Identity; Black line: Passing-Bablok fit
A) PLT regression graph for all samples (n=467) A) Regression graph for %LYM; B) Regression graph for %MONO;
B) PLT regression graph for samples with PLT count of < 200 x 109/L (n=167) C) Regression graph for %EOS; D) Regression graph for %BASO;
E) Regression graph for %NEU
4
DIAGNOSTICS
IGs are defined as the combination of promyelocytes,
myelocytes and metamyelocytes. When the Alinity hq
DISCUSSION AND CONCLUSION
%IG was compared to the manual microscopic results, Good overall agreement was demonstrated between
good correlation (r=0.88) was observed (Table 4 and CBC and WBC differential results generated by the
Figure 6-A). Bland-Altman analysis suggested slight Abbott Alinity hq, the Siemens Advia 2120i and the
negative bias (Table 4). Beckman Coulter DxH 800 hematology analyzers.
However, several differences were observed for a few
A separate analysis was performed on a subset of
measurands.
samples with a clinically relevant 1.0% or higher %IG
concentration by manual microscopy (n=63), and In the Alinity vs Advia 2120i comparison, significant
demonstrated good results (Figure 6-B). bias was observed for %MONO results between the
two analyzers. It was shown to be related to the
inclusion of some monocytes in the LUC parameter
r=0.88 r=0.90
by the Advia 2120i, and consequent underestimation
of monocyte count compared to the manual WBC
differential.
Weak correlation was observed between Alinity hq
and Advia 2120i MPV results. Significant differences
in MPV values and reference ranges have previously
A Manual %IG B Manual %IG
been published by studies comparing different
analyzers, and have been attributed to technology
Figure 6: Regression graphs for %IG comparison between Alinity hq and
manual WBC differential.
differences7. No international standards exists for
Blue line: Identity; Black line: Passing-Bablok or Deming fit MPV, and manufacturers of hematology analyzers
A) Regression graph of Alinity hq %IG versus manual %IG on all samples use different technology principles (impedance
B) Regression graph of Alinity hq %IG versus manual %IG on samples
with > 1% IG (n=63) or optical) to obtain this measurand. The low
correlation between Alinity hq and Advia 2120i
MPV results was somewhat unexpected, because
The comparison of Alinity hq NR/W results to those both analyzers use optical technology for counting
obtained by manual microscopy revealed strong PLTs and obtaining the MPV. It is possible that the
correlation (r=0.99), with a slope of 1.35, implying method for calculating MPV from the light scatter
slight positive bias (Table 4). signal is the critical factor that is responsible for the
differences between the results.
Only twelve samples within this cohort had
a clinically significant NR/W of 1.0 or higher Alinity hq RDW exhibited a negative bias compared
by manual differential. A separate calculation to DxH 800. RDW is also a technology-dependent
was performed on this subset of samples and parameter and measurements are not standardized.
demonstrated good correlation and agreement It represents the coefficient of variation of the
(r=0.99; slope=1.00) (Figure 7). erythrocyte volume distribution around the MCV.
At least two factors contribute to inter-analyzer
r=0.99
differences: the technology used to measure
RBC volume and the method used to calculate
RDW8. DxH 800 uses impedance technology to
obtain the RBC volume, whereas Alinity hq uses
optical technology based on light scattering. MCV
correlation and agreement were excellent, despite
the different technologies. Since both analyzers
use the same formula for obtaining RDW (standard
deviation [MCV] / MCV), the likely cause of the bias
Figure 7: Regression graph for NR/W comparison between Alinity hq and is the lack of harmonization between manufacturers
manual NRBC count for samples with manual NR/W ≥ 1.0.
Blue line: Identity; Black line: Passing-Bablok fit

5
DIAGNOSTICS
in volume distribution curve measurement REFERENCES
methodologies, particularly related to truncation of
1. Barrand L, Delabranche X. Clinical pathologists and physician
extreme values of the distribution9. in France: which partnership and which future? Ann Biol Clin
(Paris). 2017 Aug 1;75(4):375-392.
Alinity hq WBC differential results were equivalent
to the manual WBC differential, although weak 2. Kim YR, Ornstein L. Isovolumetric sphering of erythrocytes
correlation was observed for basophil granulocytes. for more accurate and precise cell volume measurement by flow
cytometry.Cytometry. 1983 May;3(6):419-27.
This is not an unusual finding, as several studies have
shown a low correlation between basophil counts 3. de Grooth BG, Terstappen LW, Puppels GJ, Greve J. Light-
generated by different instruments and between scattering polarization measurements as a new parameter in flow
cytometry.Cytometry. 1987 Nov;8(6):539-44.
instruments and the reference method10.
4. Van der Beken Y, Van Dalem A, Van Moer G, Segers E, Damiaens
The IG count is increased in various infectious S, Hoffmann J, Lakos G, Jochmans K. Performance evaluation of
and inflammatory conditions as well as in certain the prototype Abbott Alinity hq hematology analyzer. Int J Lab
Hematol. 2019 Aug;41(4):448-455.
hematological malignancies11. Microscopic evaluation
of patient samples for presence of IG is time 5. Harris N, Kunicka J, Kratz A. The ADVIA 2120 hematology
consuming; therefore, automated detection and system: flow cytometry-based analysis of blood and body fluids in
the routine hematology laboratory. Lab Hematol. 2005;11(1):47-61.
enumeration of these cells as offered by the Alinity
hq provides an important advantage to clinical 6. Brown W, Keeney M, Hedley BD. Initial performance evaluation
laboratories and could potentially reduce the number of the UniCel® DxH slide maker/stainer Coulter® cellular analysis
system. Int J Lab Hematol. 2014 Apr;36(2):172-83.
of manual differentials, although this was not
evaluated in this study. 7. Beyan C, Beyan E. Were the measurements standardized
sufficiently in published studies about mean platelet volume?
The presence of NRBCs in peripheral blood is an Blood Coagul Fibrinolysis. 2017 Apr;28(3):234-236.
indicator for pathologic conditions and the number 8. Hoffmann JML. Red cell distribution width and mortality risk.
of NRBCs has been shown to be associated with Clinica Chimica Acta. 2012;413:824–825
unfavorable prognosis in preterm neonates and 9. Brugnara C, Mohandas N. Red cell indices in classification
critically ill patients12,13. This evaluation demonstrated and treatment of anemias: from M.M. Wintrobes’s original 1934
strong concordance of Alinity hq NRBC counts with classification to the third millennium. Curr Opin Hematol. 2013
manual microscopy results. May;20(3):222-30

10. Amundsen EK, Henriksson CE, Holthe MR, Urdal P. IIs the
This study has a few limitations. It did not include blood basophil count sufficiently precise, accurate, and specific?:
the evaluation of reticulocyte and related parameters. three automated hematology instruments and flow cytometry
In addition, the manual differential included 100 compared. Am J Clin Pathol. 2012 Jan;137(1):86-92.
WBCs by each reader instead of the recommended 11. Buttarello M, Plebani M. Automated blood cell counts: state of
200 cell count. the art. Am J Clin Pathol. 2008 Jul;130(1):104-16.

Overall, the Alinity hq provides accurate CBC 12. Baschat AA, Gungor S, Kush LM, Berg C, Gembruch U, Harman
CR. Nucleated red blood cell counts in the first week of life: a
and WBC differential results, including IG and critical appraisal of relationships with perinatal outcome in
NRBC concentrations, and is suitable for clinical preterm growth-restricted neonates. Am J Obstet Gynecol. 2007
laboratories serving a diverse patient population. Feb;197(3):286.e1-e8

13. Purtle SW, Horkan CM, Moromizato T, Gibbons FK,

Christopher KB. Nucleated red blood cells, critical illness survivors
and postdischarge outcomes: a cohort study. Crit Care. 2017 Jun
21;21(1):154.

www.corelaboratory.abbott/hematology
Alinty hq is a Class I laser product. For in vitro diagnostic use only.
MAPSS and Alinity are trademarks of Abbott Laboratories in various jurisdictions.
All other trademarks are property of their respective owners.
Alinity h-series is available in select countries, not including the U.S.
© 2019 Abbott Laboratories. ADD-00068103.
6