ORIGINAL ARTICLE
Performance evaluation of Abbott CELL-DYN Ruby for
routine use
T. LEHTO, P. HEDBERG
Laboratory, Oulu University
Hospital, Department of Clinical
Chemistry, University of Oulu,
Oulu, Finland
Correspondence:
Tiina Lehto, Laboratory, Oulu
University Hospital, Oulu, Finland,
PO Box 500, FI-90029 OYS Finland.
Tel.: +358-8-3155453;
Fax: +358-8-3154409;
E-mail: tiina.lehto@ppshp.fi
doi:10.1111/j.1751-553X.2007.00971.x
Received 2 November 2006;
accepted for publication 18 June
2007
Keywords
Abbott CELL-DYN Ruby, hemato-
logy analyzer, white blood cell
differential, automated blood cell
counts, hematology
400
INTERNATIONAL JOURNAL OF LA BO RATO RY HEMATOLOGY
SUMMARY
CELL-DYN Ruby is a new automated hematology analyzer suitable
for routine use in small laboratories and as a back-up or emergency
analyzer in medium- to high-volume laboratories. The analyzer was
evaluated by comparing the results from the CELL-DYN Ruby with
the results obtained from CELL-DYN Sapphire. Precision, linearity,
and carryover between patient samples were also assessed. Precision
was good at all levels for the routine cell blood count (CBC)
parameters, CV% being 2.6, except for platelet count (PLT) at the
low level with CV% of 6.9%. The CV% for reticulocyte percentage
was highest at the low level (10.4). In a comparative study, the CELL-
DYN Ruby results showed a good correlation (R2  0.98) with CELL-
DYN Sapphire for the CBC parameters. For the absolute reticulocyte
count, R2 was 0.82. In the white blood cell (WBC) differentials, the
between-days precision was good for all parameters (CV%: 8.0),
except basophil percentage and absolute basophil count (CV%: 31.8
and 31.3, respectively). Additionally, in the commercial control sample
with the low WBC count (2.5 · 109/l), the precisions in the WBC
differentials were 16.9%. The cell types that occur in low numbers
showed higher CVs, as expected. The methodological comparison of
the WBC differential parameters of neutrophils, lymphocytes, and
eosinophils showed excellent correlations (R2  0.97), and the corre-
lation coefficient for absolute monocyte count and monocyte percent-
age were 0.91 and 0.87, respectively. For absolute basophil count and
basophil percentage the correlations were weaker (R2 = 0.46 and 0.34,
respectively). Carryover was minimal for all the parameters studied.
The linearities of WBC, red blood cell, PLTs, and hemoglobin were
acceptable within the tested ranges. In conclusion, the results of the
evaluation showed the performance of CELL-DYN Ruby to be good.
 2007 The Authors
Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
 T. LEHTO AND P. HEDBERG
INTRODUCTION
The CELL-DYN Ruby (Abbott Laboratories, Diagnostic
Division, Abbott Park, IL, USA) is a new hematology
analyzer designed for medium-sized clinical laborato-
ries. The analyzer provides 22 parameters of blood
count, including a 5-part white blood cell count
(WBC) differential. The system utilizes the Multi-
Angle Polarized Scatter Separation (MAPSS) techno-
logy and laser flow cytometry. CELL-DYN Ruby also
provides an integral reticulocyte analysis and Nuclear
Optical Count (NOC). The instrument’s maximum
throughput of the analyzer is 76 samples per hour.
The aim of this study was to evaluate the analytical
performance of the CELL-DYN Ruby hematology ana-
lyzer. The following parameters were evaluated for
accuracy, precision, carryover, and linearity: hemo-
globin (HGB), hematocrit (HCT), red blood cell count
(RBC), WBC, 5-part automated WBC differential,
reticulocyte absolute count (RETC), and platelet count
(PLT). The routine blood cell count parameters were
compared with those of CELL-DYN Sapphire (Abbott
Laboratories), which is currently used in the core lab-
oratory. These studies were undertaken according to
protocols based mainly on Finnish Labquality recom-
mendations (Rajamäki & Laitinen, 1990), which are
referred from International Committee for standardi-
zation in hematology guidelines (England et al.,
1984). As an individual user laboratory and as a pur-
chaser the evaluation was performed in simpler way
according to these guidelines.
MATERIALS AND METHODS
Analytical methods
Abbott CELL-DYN Ruby is an integrated multipara-
meter hematology analyzer with the capacity to store
analysis data plus histograms for 10 000 samples using
a Windows NT-based platform. CELL-DYN Ruby uses
flow cytometric techniques to analyze the RBC, PLT,
WBC, and NOC populations. For WBC, RBC, and
platelets, CELL-DYN Ruby uses four different laser
beam channels (0o, 7o, 90o, and 90oD). The light
source for optics consists of 10 mW helium–neon
laser, which emits at 632.8 nm laser beam. This tech-
nology is called MAPSS technology. For counting
WBCs, CELL-DYN Ruby’s sample volume is 150 ll in
 2007 The Authors
Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
EVALUATION OF RUBY HEMATOLOGY ANALYZER 401
open mode and 250 ll in sample loader mode. When
running patient samples in the routine blood count
(CBC) test selection, the operator should suspect the
presence of fragile WBCs when the FWBC flag is dis-
played or the presence of resistant RBCs when the
RRBC and NRBC flags are displayed. In the case of
samples containing fragile WBCs or resistant RBCs,
alternative test selections are used to measure the
WBC count. The results of these test selections are
referred to as NOC. The physical dimensions of the
system are: weight 105.2 kg, height 499 mm, width
864 mm, and depth 768 mm.
The Abbott CELL-DYN Sapphire Hematology Sys-
tem was used for the comparison studies. This system
uses the MAPSS and focused-flow impedance techno-
logies combined with three-color fluorescent flow
cytometry. This system provides a fully automated
reticulocyte analysis with Immature reticulocyte frac-
tion (IRF), a 5-part WBC differential, fluorescent DNA
staining of NRBC, optical and impedance platelet mea-
surement, and fully automated monoclonal antibody
testing (CD 3/4/8 and CD61 testing). A cyanide-free
method is used to measure HGB colorimetrically.
These instruments were calibrated, maintained,
and used according to the manufacturer’s recommen-
dations and the laboratory standard procedures. Data
exclusions were only made when there was clear evi-
dence of incomplete or suboptimal sample aspiration
or processing faults. Instrument flags for abnormal cell
populations or numerical disturbances were used as
exclusion criteria.
Samples
For these studies, residual fresh (<4 h) ethylene-
diaminetetraacetic acid (EDTA)-anticoagulated sam-
ples submitted for routine full blood cell counts were
used. All samples were drawn into Becton Dickinson
Vacutainer K2EDTA tubes (Becton Dickinson Vacu-
tainer, Cat. No. 388010, Becton Dickinson Systems,
Plymouth, UK) and maintained at room temperature.
Sample selection criteria were used only to ensure a
representative balance of normal and abnormal
parameter ranges. Data exclusions were only made
when there was clear evidence of incomplete or
suboptimal sample aspiration or processing faults.
Samples were not selected according to any ward. All
samples were processed anonymously.
402 T. LEHTO AND P. HEDBERG EVALUATION OF RUBY HEMATOLOGY ANALYZER
The reagents, calibrators, and quality controls rec-
ommended by the manufacturer were used in the
study. Three levels of quality control material were
used with Abbott CELL-DYN Ruby for between-day
precision studies. For the reticulocyte between-day
precision study, two quality controls with different
concentrations were used.
Imprecision
Within-run precision was evaluated by performing 10
consecutive measurements on three fresh blood sam-
ples. Between-day open-mode precision was evaluated
by running commercial controls twice a day over a
period of 10 days (replicative runs; total n = 20), after
which the coefficient of variation (CV%) for each
parameter was calculated. The total variation (paired
precision) was evaluated by analyzing 30 patient sam-
ples without analyzer flags in duplicate. They were
analyzed randomly among other samples using both
open and closed modes. In open mode specimen is ana-
lyzed manually using open mode probe and in closed
mode specimen is analyzed automatically using sample
loader rack. Total variation was calculated by a two-
tailed paired t-test using Microsoft Excel. A P-value of
<0.05 was considered statistically significant. Paired
precision was determined by linear regression analysis
and by the CV%s of the combined results.
Carryover
Carryover was assessed by analyzing three pairs of
samples. A high-concentration sample was first ana-
lyzed consecutively in triplicate (H1, H2, and H3), fol-
lowed immediately by a low-concentration sample
consecutively in triplicate (L1, L2, and L3). The mean
percentage of carryover for each parameter was calcu-
lated using the following formula: carryover
% = (L1 ) L3)/(H3 ) L3) · 100. Carryover was calcu-
lated for WBC count, PLT, RBC count, HGB, and
RETC.
Linearity
The linearity ranges for the basic blood cell count
parameters are specified by the manufacturer. We
verified the linearities for HGB, WBC, RBC, PLT, and
RETC. Linearity studies were performed by serial
Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
dilution of fresh patient samples (from 100% to
10%). The analyzer’s primary dilution fluid was used
as diluent.
Interinstrument agreement
The interinstrument comparability studies were car-
ried out by analyzing in singlicate 100 randomly
selected blood samples for CBC analysis and 50 sam-
ples for RETC from the daily workload on the labora-
tory’s Abbott CELL-DYN Sapphire.
RESULTS
Imprecision
The results of the within-run and between-day preci-
sion analyses are shown in the Tables 1–3. Precision
was good at all levels for the CBC parameters, CV%
being 2.3, except for very low WBC with CV% of
2.6 and for low-level PLT with CV% of 6.9. CV%
for reticulocyte percentage was highest at the low
level (10.4). In the WBC differentials, the between-
days precision was good for all parameters (CV%:
8.0), except basophil percentage (B%) and absolute
basophil count (CV%: 31.8 and 31.3, respectively).
Additionally, in the commercial control level I with
the low WBC count (2.5 · 109/l), the precisions in
the WBC differentials were 16.9%. The cell types
that occur in low numbers showed higher CVs, as
expected.
The open and closed mode paired precision of
CELL-DYN Ruby (Table 4) was determined during a
2-week evaluation period. Open mode precision was
excellent (R2  0.99) for WBC, RBC, HGB, and mean
corpuscular volume (MCV). For mean corpuscular
hemoglobin (MCH), PLT, and mean platelet volume
(MPV), the correlation coefficients were 0.94, 0.97,
and 0.77, respectively. The correlation coefficients of
the closed mode were almost similar to those of the
open mode (R2  0.99) for WBC, RBC, HGB, and
MCV, with the exception of the correlation coefficients
for MCH and MPV being 0.97 and 0.90, respectively.
The comparative performance characteristics for
open and closed mode routine CBC measurements
were also evaluated by comparing the mean values of
each pair. Correlations were excellent for WBC, RBC,
HGB, MCV, and PLT (R2  0.99), and the correlation
 2007 The Authors
 T. LEHTO AND P. HEDBERG EVALUATION OF RUBY HEMATOLOGY ANALYZER 403

Table 1. Within-run precision using three patient samples with different concentrations

Sample I Sample II Sample III

Ruby Mean SD CV% Mean SD CV% Mean SD CV%

White blood cells (·109/l) 1.53 0.04 2.6 7.78 0.10 1.3 19.9 0.29 1.5
Red blood cells (·1012/l) 2.26 0.04 1.8 3.31 0.03 0.8 4.44 0.07 1.7
Hemoglobin (g/l) 81.2 0.03 0.0 121 0.03 0.3 159 0.07 0.4
Hematocrit (l/l) 0.23 0.31 1.3 0.43 0.52 1.2 0.41 0.58 1.4
MCV (fl) 75.1 0.45 0.6 88.2 0.47 0.5 107 0.85 0.8
Platelets (·109/l) 57 2.39 4.2 202 3.00 1.5 486 6.44 1.3
Reticulocytes (%) 1.10 0.07 5.9 4.15 0.18 4.4 6.97 0.23 3.3
Neutrophils (%) 36.8 1.73 4.7 61.4 0.67 1.1 72.0 0.62 0.9
Lymphocytes (%) 52.2 3.32 6.4 24.9 0.71 2.9 15.1 0.55 3.7
Monocytes (%) 5.45 3.08 56.4 8.10 0.69 8.5 9.93 0.52 5.3
Eosinophils (%) 4.7 1.07 22.7 4.67 0.19 4.0 0.67 0.28 42
Basophils (%) 0.9 0.22 25.2 0.9 0.22 24.6 2.39 0.22 9.1
Neutrophils (·109/l) 0.56 0.03 4.7 4.78 0.09 1.9 14.4 0.28 2.0
Lymphocytes (·109/l) 0.80 0.06 7.0 1.93 0.05 2.7 3.00 0.11 3.6
Monocytes (·109/l) 0.08 0.05 58.5 0.63 0.05 8.7 1.98 0.09 4.4
Eosinophils (·109/l) 0.07 0.02 21.8 0.36 0.02 4.9 0.13 0.06 42.9
Basophils (·109/l) 0.01 0.00 26.7 0.07 0.02 24.1 0.48 0.05 9.9

Table 2. Between-days precision using three different commercial controls (CELL-DYN 22 Controls, Abbott
Laboratories) with different concentrations

Level I Level II Level III

Ruby (n = 20) Mean SD CV% Mean SD CV% Mean SD CV%

9
White blood cells (·10 /l) 2.50 0.06 2.3 8.03 0.13 1.6 17.5 0.14 0.8
Red blood cells (·1012/l) 2.43 0.05 2.2 4.15 0.04 0.9 5.3 0.05 0.9
Hemoglobin (g/l) 78.1 0.10 1.3 125 0.12 1.0 153 0.16 1.0
Hematocrit (l/l) 0.19 0.44 2.3 0.31 0.27 0.8 0.39 0.39 1.0
MCV (fl) 77.3 0.39 0.5 75.3 0.24 0.3 72.9 0.29 0.4
Platelets (·109/l) 70.3 4.85 6.9 262 10.5 4.0 523 13.4 2.6
Neutrophils (%) 62.6 1.43 2.3 61.8 0.90 1.5 58.9 0.65 1.1
Lymphocytes (%) 19.6 2.90 14.9 22.9 1.87 8.2 25.7 0.58 2.3
Monocytes (%) 10.0 1.45 14.4 8.2 0.65 7.9 8.68 0.48 5.5
Eosinophils (%) 3.22 0.45 13.9 2.73 0.18 6.4 2.81 0.23 8.1
Basophils (%) 4.56 1.45 31.8 4.43 1.21 27.3 3.96 0.26 6.6
Neutrophils (·109/l) 1.54 0.05 3.14 4.96 0.10 2.0 10.3 0.15 1.4
Lymphocytes (·109/l) 0.48 0.08 16.9 1.84 0.16 0.51 4.49 0.11 2.4
Monocytes (·109/l) 0.25 0.03 13.4 0.66 0.05 7.8 1.52 0.08 5.6
Eosinophils (·109/l) 0.08 0.01 16.2 0.22 0.01 6.3 0.49 0.04 8.0
Basophils (·109/l) 0.11 0.04 31.3 0.36 0.10 26.9 0.69 0.05 6.8

coefficients for MCH and MPV were 0.97 and 0.87, counts, the R2 values varied within 0.96–0.98 in both
respectively. modes. Only the correlation coefficients for basophil
The correlation coefficients for neutrophil and lym- absolute counts were poorer in the open and closed
phocyte absolute counts in the open and closed modes modes (R2  0.62 and 0.76, respectively). In the com-
were 0.99. For monocyte and eosinophil absolute parative performance study between open and closed
 2007 The Authors
Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
404 T. LEHTO AND P. HEDBERG EVALUATION OF RUBY HEMATOLOGY ANALYZER

(CV%: 5.1) and MPV (CV%: 5.1) and lower CV%

Table 3. Between-days precision using two different
commercial controls (CELL-DYN Retic Plus Controls,
Abbott Laboratories) with different concentrations
Level I Level II
with PLT (CV%: 2.9). For the WBC differential
parameters precision was <7.5% and only exceeded
10% for basophils in both modes.

Ruby (n = 40) Mean SD CV% Mean SD CV% Carryover and linearity

Reticulocytes 1.27 0.132 10.4 4.42 0.224 5.1 The carryover effect on sample analysis was minimal
(%) for all the parameters studied (WBC, PLT, RBC, HGB,
and RETC). The mean carryover percentage was
0.47% for all parameters.
The results of linearity analysis are shown in
Table 4. CELL-DYN Ruby paired precision processed
Table 5. Excellent linearity for HGB was observed
with the open and closed mode (R2 values, n = 30)
determined by linear regression and comparison of the in the range from 10.1 to 187 g/l (observed =
mean values of the open and closed modes 1.02 · expected ) 1.47), with no significant bias
between the observed and expected values. Good lin-
Open Closed Open vs. earities for WBC (observed = 1.01 · expected ) 0.11),
Ruby mode mode closed modes
RBC (observed = 1.01 · expected + 0.04), and PLT
White blood cells (·109/l) 0.99 0.99 0.99 (observed = 0.99 · expected ) 2.70), were observed,
Red blood cells (·1012/l) 0.99 0.99 0.99 with no significant bias between the observed and
Hemoglobin (g/l) 0.99 0.99 0.99 expected values in the ranges from 0.49 to 163 · 109/l,
MCV (fl) 0.99 0.99 0.99 from 0.14 to 7.54 · 1012/l, from 13.3 to 868 · 109/l,
MCH (pg) 0.94 0.97 0.97 respectively.
Platelets (·109/l) 0.97 0.99 0.99
MPV (fl) 0.77 0.90 0.87
Neutrophils (·109/l) 0.99 0.99 0.99 Interinstrument agreement
Lymphocytes (·109/l) 0.99 0.99 0.99
Monocytes (·109/l) 0.96 0.97 0.96 In the comparative study, the CELL-DYN Ruby results
Eosinophils (·109/l) 0.97 0.98 0.99 showed good correlation (R2 > 0.98) in the following
Basophils (·109/l) 0.62 0.76 0.62
parameters compared to CELL-DYN Sapphire: WBC,
MCH, mean corpuscular hemoglobin; MPV, mean plate- RBC, HGB, HCT, MCV, and PLT. For RETC, R2
let volume. was 0.81 (Figure 2a–g). The methodological compari-
son of the WBC differential parameters of absolute and
percentage values for neutrophils, lymphocytes, and
modes the correlation coefficients for all the WBC dif- eosinophils showed excellent correlations (R2  0.97),
ferential parameters varied between 0.96 and 0.99, and the correlation coefficients for M% and absolute
except for basophils, which was 0.62. monocyte count were 0.87 and 0.91, respectively
To clarify the analytical imprecision of the open (Table 6). The correlation for B% and absolute basophil
and closed modes we also determined the CV%s of count were weaker (R2 = 0.34 and 0.46, respectively).
the combined results for the CBC and for the WBC
differential by analyzing the patient specimens among
DISCUSSION
other samples (Figure 1). This reflects close to the
analytical imprecision in authentic routine use. The The results for precision, carryover, and linearity were
combined results showed that the closed mode preci- acceptable. Further, there was a good correlation
sion for most measured parameters was <2.1% and between the evaluated instruments and the reference
only exceeded 2.1% for MPV (CV%: 3.1). Better pre- analyzer, CELL-DYN Sapphire, for most of the ana-
cision values (CV%: 0.9) were obtained for RBC, lyzed parameters, except RETC and basophils, the
HGB, MCV, and MCH when the samples were ana- overall correlation being acceptable. To achieve opti-
lyzed in the open mode, with the exception of WBC mal WBC count comparison between the methods the
 2007 The Authors
Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
 T. LEHTO AND P. HEDBERG EVALUATION OF RUBY HEMATOLOGY ANALYZER 405

Open mode Open mode

Coefficient of variation, CV %
20.52

Coefficient of variation, CV %
12 5.08 5.08
(±4.66)
40 (±15.38)
(±4.66)
2.91 30
8 (±3.31) 7.44
(±10.39)
20 3.89 4.64
4 0.78 0.85 1.01 (±4.93) (±3.25)
0.65
(±0.82) (±1.04) 10 (±0.75)
(±0.54) 0.16
(±0.20)
0 0
WBC RBC HGB MCV MCH P LT MPV N# L# M# E# B#

Closed mode Closed mode

Coefficient of variation, CV %
10
Coefficient of variation, CV %

40 15.90
8 3.15 (±13.17)
(±2.68) 30
6 2.14 7.06
(1.98) 20 7.00 (±5.59)
3.70
4 1.19 (±4.64)
(±3.54)
(±0.92) 0.61 0.63 0.63 10 0.98
2 (±0.64) (±0.62) 0.19 (±0.67) (±0.76)
(±0.16)
0
0
WBC RBC HG B M CV M CH PLT M PV N# L# M# E# B#

Figure 1. CELL-DYN Ruby’s open and closed mode paired precision determined as the CVs of the combined results
for the cell blood count and the white blood cell differential. The columns represent the mean values of CV% of
each replicate run. Error bars represent the +/)SD of these CVs.

reference analyzer, CELL-DYN Sapphire, WBC count,

Table 5. Linearity of CELL-DYN Ruby parameters
Parameter
WBC
RBC
HGB
Platelets
RETC
WBC, white blood cell; HGB, hemoglobin; RBC, red
blood cell; RETC, reticulocyte absolute count.
values included should be higher than 180 · 109/l.
The difference between CELL-DYN Ruby and CELL-
DYN Sapphire in RETC counts could be explained by
the different measuring principles (scatter vs. fluores-
cence). The poor correlation of basophils was not
unexpected because of the very low basophils counts.
White blood cell differential correlation between
CELL-DYN Ruby and the manual differential count
was not calculated in this study. Müller et al. (2006)
have already published data of the comparison of our
 2007 The Authors
Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
and manual differential count.
Tested ranges Recovery (%) Buttarello (2004) reviewed the various methods for
establishing imprecision criteria, ranging from methods
0.49–163 · 109/l 87.8–106.4 based on fractions of biologic variation to state-of-the-
0.14–7.54 · 1012/l 96.2–112.1
art technologies and methods based on questionnaires
10.1–187 g/l 90.4–108.0
13.3–868 · 109/l 89.5–109.3 presented to clinicians. In this study, the imprecision
15.4–203 · 109/l 89.1–123.2 performance appeared to be acceptable for WBC count,
RBC count, HGB, MCV, and reticulocyte when com-
pared to the maximum values (CV%) based on the
within-subject biologic goals (day-to-day variability).
In hematology, especially in cell counts, it is not
enough to know the imprecision at normal concentra-
tions, as the imprecision of cell counts is nonlinear. It
is advisable to know the imprecision in the entire
range of clinical use, particularly at low values, where
CV% increases dramatically (Buttarello 2004). Our
study also included the measurements of imprecision
at different levels using commercial control materials
in the between-days study and patient specimens with
the different concentration levels in the paired preci-
sion study. Compared with the goals based on intrain-
dividual biologic variation, the precision performance
406 T. LEHTO AND P. HEDBERG EVALUATION OF RUBY HEMATOLOGY ANALYZER

(a) WBC 10e9/L (b) RBC x 10e12/L

20
y = 0.9625x + 0.1656 6
y = 0.9533x + 0.3664
R 2 = 0.9917 R 2 = 0.9900
16 5

CELL-DYN Sapphire
CELL-DYN Sapphire

4
12
3
8
2

4 1

0
0 0 1 2 3 4 5 6
0 4 8 12 16 20
CELL-DYN Ruby
CELL-DYN Ruby

HGB g/L HCT L/L

(c) 180 (d) 0.6 y = 0.9735x + 0.0175
y = 1.0045x + 1.8105
R 2 = 0.9837 R 2 = 0.9842
160 0.5
CELL-DYN Sapphire

CELL-DYN Sapphire
140 0.4

120 0.3

100 0.2

80 0.1

60 0
60 80 100 120 140 160 180 0 0.1 0.2 0.3 0.4 0.5 0.6
CELL-DYN Ruby CELL-DYN Ruby

MCV fL PLT 10e9/L

(e) 120 (f) 800
y = 1.0176x - 0.9266
R 2 = 0.9756 y = 1.0372x - 12.233
110
R 2 = 0.9802
CELL-DYN Sapphire

CELL-DYN Sapphire

600
100

90 400

80
200
70

60 0
60 70 80 90 100 110 120 0 100 200 300 400 500 600 700 800
CELL-DYN Ruby CELL-DYN Ruby

RETC x 10e9/L
(g) 250
y = 0.9468x + 5.2907
R 2 = 0.8155
200
CELL-DYN Sapphire

150

100

50

0
0 50 100 150 200
CELL-DYN Ruby

Figure 2a–g. Correlation between CELL-DYN Ruby and CELL-DYN Sapphire cell blood count and reticulocyte abso-
lute count parameters evaluated by Deming regression analysis.

 2007 The Authors

Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407
 T. LEHTO AND P. HEDBERG EVALUATION OF RUBY HEMATOLOGY ANALYZER 407

using patient specimens on WBC differentials was

Table 6. Correlation between CELL-DYN Ruby and
CELL-DYN Sapphire white blood cell differential acceptable with all of the parameters in our study,
parameters evaluated by Deming regression analysis except with basophils.
The instrument is easy to use and suitable for routine
Ruby vs. Sapphire use in small laboratories and as a back-up or emergency
(n = 100) R2 Slope Intercept
analyzer in medium- to high-volume laboratories.
Neutrophils (%) 0.98 0.98 1.05
Lymphocytes (%) 0.99 1.01 0.78 ACKNOWLEDGEMENTS
Monocytes (%) 0.87 0.92 0.18
Eosinophils (%) 0.97 0.95 0.14 We would like to thank Abbott Diagnostics for provid-
Basophils (%) 0.34 0.57 0.12 ing the Abbott CELL-DYN Ruby system and the
Neutrophils (·109/l) 0.99 0.96 0.12
reagents and for technical help.
Lymphocytes (·109/l) 0.98 1.03 0.00
Monocytes (·109/l) 0.91 0.91 0.01
Eosinophils (·109/l) 0.97 0.92 0.01
Basophils (·109/l) 0.46 0.59 0.01

REFERENCES
Buttarello M. (2004) Quality specification
in haematology: the automated blood
cell count. Clinica Chimica Acta 346,
45–54.
England J.M., Rowan R.M., van Assend-
elft O.W., Coulter W.H., Groner W.,
Jones A.R., Koepke J.A., Lewis S.M.,
Shinton N.K. & Thom R. for the Inter-
national Committee for Standardization
in Haematology (1984) Protocol for
evaluation of automated blood cell
counters. Clinical and Laboratory Hae-
matology 6, 69–84.
Müller R., Mellors I., Johannessen B.,
Aarsand A.K., Kiefer P., Hardy J., Ken-
dall R. & Scott C.S. (2006) European
multi-center evaluation of the Abbott
Cell-Dyn Sapphire hematology analyzer.
Laboratory Hematology 12, 15–31.
Rajamäki A. & Laitinen M. (1990)
Automaattisen verisoluanalysaattorin
koestusohjelma. Moodi 4, 202–209.

 2007 The Authors

Journal compilation  2007 Blackwell Publishing Ltd, Int. Jnl. Lab. Hem. 2008, 30, 400–407