Within-day Physiologic Variation of Leukocyte Types
in Healthy Subjects as Assayed by Two Automated
Leukocyte Differential Analyzers
PER WINKEL, M.D., DOC. MED. SCI., BERNARD E. STATLAND, M.D., PH.D., ALEX M. SAUNDERS, M.D.,
HUGH OSBORN, AND HERBERT KUPPERMAN, M.D., PH.D.
Winkel, Per, Statland, Bernard E., Saunders, Alex M.,
Osborn, Hugh, and Kupperman, Herbert: Within-day physiologic variation of leukocyte types in healthy subjects as assayed
by two automated leukocyte differential analyzers. Am J Clin
Pathol 75: 693-700, 1981. The physiologic within-day (7002200 h) variation of leukocyte-type concentrations in blood as
determined for 21 healthy young adults is reported. All blood
specimens were obtained in duplicate such that the withinbatch analytic variation, as well as the pertinent biologic
sources of variation, was able to be determined. All specimens
were analyzed on each of two automated leukocyte differential systems: Hemalog-D® Differential System and the Hematrak-240® Analyzer. On the basis of a comparison of the performances of the two analyzers, it was decided to report the
neutrophil and the lymphocyte values as measured on both
systems, the monocyte values as measured on the Hematrak
only, and the eosinophil and basophil values as measured on the
Hemalog-D only. The intrasubject within-day physiologic
variations for the cell types on the Hemalog-D and Hematrak,
respectively, in terms of coefficient of variation were as follows: neutrophils, 19.4% and 19.6%; lymphocytes, 13.8% and
17.5%. For monocytes, as measured on the Hematrak, it was
13.4%. For eosinophils and basophils, as measured on the
Hemalog-D, it was 27.2% and 8.3%, respectively. There was
a consistent group-specific diurnal variance that amounted to
more than 40% of the total within-day variance, both for
lymphocytes and for eosinophils. The within-day physiologic
variation of the cell type concentrations for eight of the volunteers was compared with that of the plasma Cortisol concentrations, as determined on specimens derived from the same venipuncture sessions. The decrease in plasma Cortisol values was in
most instances associated with decreases in eosinophils and
increases in neutrophils. For total leukocytes and neutrophils,
the mean concentration for smokers was significantly higher
than that for nonsmokers. (Key words: Automation; Cortisol;
Cytochemistry; Diurnal; Leukocyte; Physiologic variation;
Quality control; Leukocyte; Within-day variation.)
Department of Clinical Chemistry, Finseninstitutet,
Copenhagen, Denmark; Department of Pathology,
University of California, Davis, Medical Center,
Sacramento, California; Geometric Data Corporation,
Wayne, Pennsylvania; Roche Clinical Laboratories,
Raritan, New Jersey
TYPES of automated analyzers for the differcount of leukocyte types have been introduced
clinical laboratory. For the first type of system
the Hematrak*), a blood smear is presented to
the analyzer. The system, using a pattern-recognition
technic implemented on a computer, classifies the
various leukocyte types.
For the second type of system, of which there is
only one example commercially available (the Hemalog-Dt), 10,000 leukocytes are classified in each of
three channels on the basis of size (as measured by
light scatter) and the intensity of a biochemical reaction taking place within the cell. In the peroxidase
channel, the cells are classified according to their
peroxidase activities into lymphocytes, large unstained
(peroxidase-negative) cells, neutrophils, eosinophils,
and the sum of monocytes and basophils. The fraction numbers^ of the latter two cell categories are obtained in the remaining two channels. In the monocyte channel, leukocytes are classified according to
size and esterase activity to obtain the fraction number
of monocytes (esterase-positive cells). In the basophil
channel, the basophils are separated from the rest of the
cells on the basis of a colorimetric reaction with the
heparin of the basophils. 5
Recently, we examined the intraindividual biologic
and analytic variations of the concentrations of leukocyte types determined for healthy subjects with the
Hemalog-D. 9 In that study, we examined the hour-tohour variations occurring from 800 h to 1400 h in each
of 20 healthy volunteers. The purpose of the present
paper is to present a more extensive investigation of
the within-day variation of leukocyte types in blood ob-
Received June 13, 1980; received revised manuscript and accepted
for publication November 11, 1980.
Address reprint requests to Dr. Statland: Department of Pathology, University of California, Davis, Medical Center, 2315 Stockton
Boulevard, Sacramento, California 95817.
* Hematrak-240 Analyzer, Geometric Data Corp., Wayne,
Pennsylvania.
t Hemalog-D Differential System, Technicon Instruments Corp.
Tarrytown, New York.
t A fraction number is here defined as the ratio of the number of
elements in an individual class of a number distribution to the total
number of elements, where the denominator is assumed to be 100.
This value is essentially the same as the "differential."
TWO
ential
to the
(e.g.,
0002-9173/81/0500/0693 $00.90 © American Society of Clinical Pathologists
693
WINKEL ET AL.
694
Table I. Demographic Characteristics of the 21
Subjects Participating in the Study
bject
mber
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Age
Height
Weight
Smoking
Sex
yrs
m
kg
cig/day
M
M
M
M
F
M
F
F
M
F
F
F
F
F
M
M
M
M
F
M
F
27
40
30
46
26
48
55
62
34
25
28
24
41
51
41
51
34
25
41
32
24
1.85
1.71
1.74
1.65
1.75
1.79
1.65
1.65
1.78
1.78
1.65
1.59
1.70
1.52
1.93
1.65
1.74
1.80
1.51
1.83
1.65
84
73
64
68
64
73
79
57
82
80
54
58
54
63
98
75
77
77
57
77
57
None
None
None
None
20
None
None
None
None
10
25
15
None
30
6
None
None
None
20
None
None
tained from 700 h to 2200 h on each of three days from
each of 21 healthy volunteer subjects. Moreover, we
assayed the leukocyte types on each of the two automated differential systems discussed above: the Hematrak and the Hemalog-D.
For eight of the 21 subjects, we also determined
plasma Cortisol concentrations so as to relate the
changes in leukocyte types with the fluctuations in
serum Cortisol concentration.
Materials and Methods
Subjects
Blood specimens were obtained from each of 21
healthy subjects. The demographic characteristics of
the 21 subjects are given in Table 1. There were seven
cigarette smokers in the population studied.
Experimental
Protocol
Eighteen venipuncture sessions were performed
with each subject. The dates and hours of the day
during which these sessions took place are given in
Figure 1. On each of the three experimental days, the
subjects reported to the blood-drawing area after an
overnight fast. The subjects sat quietly for at least
5 min, after which replicate blood specimens were
drawn by the phlebotomist into each of three EDTAcontaining Vacutainer® tubes. The first two tubes were
used for the leukocyte analyses; the third tube was
used (for eight of the subjects) for the Cortisol analysis. The procedure for obtaining blood was repeated
A.J.C.P. • May 1981
six times during the day: at 700 h, 1000 h, 1300 h,
1600 h, 1900 h, and 2200 h. The subjects ate three
meals during each experimental day: the first between
the 700 h and 1000 h phlebotomy sessions, the second
between the 1300 h and 1600 h sessions, and the third
between the 1900 h and 2200 h sessions. Because the
subjects were all in the same area, it was possible to
provide them with essentially the same type of meal.
The meals for each of the three experimental days were
very similar. During the experimental day, subjects who
were cigarette smokers were permitted to smoke cigarettes up to 15 min before any venipuncture session.
In addition, the subjects engaged in their normal work
activities during each of the three experimental days.
Processing and Analyzing the
Specimens
Each of the two duplicate specimens was processed
in the same manner. The tube of EDTA-containing
blood was immediately inverted after the draw. A
wedge smear was made from each tube of blood. The
smear was stained and then counted on a Hematrak
system. The remainder of the blood from each tube
was then assayed on a Hemalog-D system. Each tube
of blood was uniquely labeled and randomly ordered
before analysis. In that there were 21 subjects and
two specimens per venipuncture, 42 specimens were
assayed per batch.
The smears were counted on the Hematrak in the
following manner. The automated mode was used such
that 200 cells were identified per smear. The cells were
identified and classified into one of the following
groups: band neutrophil, segmented neutrophil, lymphocyte, atypical lymphocyte, monocyte, eosinophil, and
basophil. When the number fraction times 100 was not
a whole number, the value was rounded off to the next
integer, e.g., 0.5 rounded off to 1.0, and 17.5 rounded
off to 18.0. The Hematrak we used was in the Research
and Developmental Center of the Geometric Data
Corporation. In the analysis of the cell types, total
neutrophils was defined as the sum of band neutrophils
and segmented neutrophils. For the analysis of lympho-
§ APRL
I 4r
([^-{zXfMlMiH!)
£ APRIL 18 -
(j)-^}--@--®—®-~©
g MAY 5 L
@—@>~@-(!6M57)—@
°
•
7
i
10
i
13
I
16
HOUR OF DAY
FIG. 1. Protocol of the study.
I
19
I
22
Vol. 75 • No. 5
WITHIN-DAY VARIATION OF LEUKOCYTES
cytes, atypical lymphocytes were not added to the
lymphocyte values.
The analyses of specimens on the Hemalog-D
were done in the following manner. The Hemalog-D
we used for this study was in a nearby private clinical
laboratory. The total leukocyte count and the counts
of the cell types (neutrophil, lymphocyte, monocyte,
eosinophil, and basophil) were assayed in the usual
mode of analysis as part of the normal workload in
the private laboratory. The results for the cell types
were presented in absolute numbers, i.e., the total
concentration of leukocytes was multiplied by the
number fraction for each particular cell type as determined on each of the two instruments. In each
case, the total leukocyte count was based upon the
results determined on the Hemalog-D. The operators
of the Hematrak and the Hemalog-D were totally
unaware of the relationship of the replicates and the
number sequence used in the study. This was accomplished by using fictitious names on each tube (or
slide) of blood.
For all venipunctures performed on eight of the subjects (Subjects 3, 6, 10, 12, 13, 16, 18, and 19), the third
blood specimen was used for the determination of the
plasma Cortisol concentration. The analysis of Cortisol
was based on a radioimmunoassay procedure. All
plasma specimens from these eight subjects were
frozen at - 8 0 C and randomly ordered before being
analyzed.
Statistical Analysis
Since both analyzers have obvious strengths but
also weaknesses, we decided that the most appropriate
results would be obtained if we eliminated results
that were less reliable as measured on one analyzer
when compared with the corresponding results obtained by the other analyzer. Therefore, before analyzing the data we compared the results obtained by
the analyzers with regard to random analytic variation as well as systematic differences between the
results obtained by the two methods.
The comparison of the results obtained by the two
analyzers showed that for the monocytes there were
differences between the results obtained by the two
methods that depended on the subject examined and
the time of day the specimens were obtained. The
latter effect may be due to the specifics of the daily
analytic procedures at the private laboratory where
the Hemalog-D was situated. In all likelihood, the
monocyte channel was not functioning optimally in
this instance. This supposition is supported by the fact
that the average monocyte concentration measured by
the Hemalog-D was far below that measured by the
Hematrak. Therefore, we eliminated the monocyte
695
results measured by the Hemalog-D and relied only on
the results obtained with the Hematrak. When we
compared the random analytic variations of the two
analyzers, it was obvious that the Hemalog-D was
performing much better {i.e., having a much lower
analytic variation) than the Hematrak with regard to
basophils and eosinophils. For these reasons we
decided to report only the basophil and eosinophil
values obtained by the Hemalog-D. Thus, for the
neutrophils and the lymphocytes we report the results
obtained by both systems; for the monocytes we
report the results obtained by the Hematrak only; for
the eosinophils and the basophils we report the results
obtained by the Hemalog-D only.
Biologic
Variation
Components of Variance. Two sets of data were
analyzed separately. They are referred to as data set
A and data set B. Data set A consists of all the results
of measurements by the Hematrak and/or the HemalogD for the 21 volunteers for whom six venipunctures
were performed on each of three days. Data set B consists of these values and also the corresponding plasma
Cortisol concentrations for eight of the subjects.
In the statistical analyses of data set A, a threefactor analysis of variance (ANOVA) was performed
for the total leukocyte count and for each leukocyte
type. The Cortisol data of data set B were subjected
to a similar three-factor ANOVA. Three sources of
variation were included in the statistical model, namely,
that due to differences among subjects, that due to
differences among days, and that due to differences
among the times of day. The latter effect was considered fixed, while the former two were considered
random. The details of the statistical analysis have
been described elsewhere."
For our experimental design described above, we
briefly explain the biologic meanings of the various
terms included in the ANOVA model. For a given
quantity, a significant main effect of subject indicates
that the subjects differ from one another in regard to
the average concentration of that quantity. A significant main effect of clay implies that the mean values
computed for the various days differ. A significant
main effect of hour indicates that the mean values for
the six hours of the day as computed for all subjects
and for all days differ. The implication of a significant
subject-hour interaction is that the subjects show
hourly variations that vary from subject to subject but
are consistent from day to day for a given subject. A
significant subject-day interaction indicates that the
individual's daily levels of the quantity vary from day
to day. A significant day-hour interaction implies
that the values of the quantity vary as a function of the
WINKEL£r>tZ..
696
A.J.C.P. • May 1981
Table 2. Physiologic and Analytic Components of Variation* Determined on Two
Automated Leukocyte Differential Systems
Mean Value
Within-subject Physiologic Variation
Total
Among-subject
Variation of
Mean Values
Within-run
Analytic
Variation
7.6
14.6
19.6
4.6
19.4
19.6
12.6
12.9
23.1
23.5
26.5
26.9
5.9
7.4
2,382
2,262
13.8
17.5
9.5
10.4
16.7
20.4
28.9
32.0
7.8
11.9
Monocytes
Hematrak
534
13.4
13.4
19.0
19.5
24.8
Eosinophils
Hemalog-D
150
27.2
24.4
36.6
45.0
15.0
Basophils
Hemalog-D
55
8.3
8.3
11.8
29.4
32.2
Quantity
Measured by
cells//nl
Within-day
Day-to-day
Total leukocytes
Hemalog-D®
7,600
12.6
Neutrophils
Hemalog-D
Hematrak*
4,564
4,615
Lymphocytes
Hemalog-D
Hematrak
' Variation presented as coefficient of variation times 100.
hour of the day oh which the corresponding specimens
were obtained. In the case of the leukocyte types,
such effects are probably due to batch-to-batch analytic
variation, since all specimens obtained from the subjects at the same time on the same day were assayed
in one batch. A significant subject-day-hour interaction indicates a variation of a quantity that cannot
be explained on the basis of a general group variation or on the basis of a consistent individual subject's variation (subject-hour interaction) occurring
over the six-hour period of sampling. Such threefactor interactions are considered random biologic
fluctuations. Unbiased estimates of all components of
variance were computed on the basis of the expected
values of the mean squares of main effects and interactions. The within-batch analytic variation was
estimated from the duplicate measurements.
Analytic
Variation
Correcting Individual Patient Data for Batch-tobatch Analytic Variation. Since significant day-hour
interactions (indicating batch-to-batch analytic variation) occurred for the majority of components, the
data were corrected for this variation before graphically
displaying the data. For a given quantity according to
the ANOVA model used, we have:
Mjkv
= fi + Sj + dj + tk + sd u
+ stlk + dtjk + sdtUk + ev
where xiJkv is the quantity value measured in specimen
# v (v = 1,2) obtained from subject # i (i = 1, 21) on
day #j (j = 1,3) at hour #k(k = 1, 6); /x is the grand
mean; the parameters s(, dj, and tk represent the main
effects of subject, day, and hour of the day, respectively; the parameters sd ik , st jk , dt jt , and sdtijk represent
the corresponding two- and three-factor interaction; ev
is the error term. Each of the interaction terms dtjk
(j = 1, 3; k = 1, 6) were estimated from the data. The
estimate of dtjk was then subtracted from all observations on day #j at hour # k .
Results
Table 2 shows the mean absolute value of each cell
type as determined by each of the two automated leukocyte systems, the total within-subject physiologic
variation as well as the within-day and day-to-day
physiologic variations, the among-subject variation in
mean values, and the within-run (within-batch) analytic
variation for the Hemalog-D and the Hematrak. In all
cases the variations are presented as the coefficient
of variation (CV) times 100.
From the results presented in Table 2, we see that
the values determined by the Hematrak and the Hemalog-D (i.e., the neutrophil and lymphocyte values)
compare very favorably.
The grand means for the 21 subjects' leukocyte
types determined with the Hematrak and HemalogD compared favorably for neutrophils, lymphocytes,
and eosinophils. The means for neutrophils were 4,615/
JU.1 (Hematrak) and 4,564//xl (Hemalog-D); for lymphocytes, 2,262//xl (Hematrak) and 2,382//xl (HemalogD); for eosinophils, 166//xl (Hematrak) and 150//xl
(Hemalog-D). The monocyte values obtained by the
Hemalog-D (mean, 312//U.1) were consistently lower
than those obtained by the Hematrak (534//xl). The
WITHIN-DAY VARIATION OF LEUKOCYTES
Vol. 75 • No. 5
basophil values obtained by the Hemalog-D (55/yu.l)
were consistently somewhat higher than those obtained
by the Hematrak (46//U.1).
In Table 3, the total within-day variance for each
leukocyte type and for plasma Cortisol has been partitioned into its three components: the consistent groupdiurnal variation, the consistent subject-specific diurnal variation, and the random diurnal variation.
Each component of variance is presented as the fraction of the total within-day variance. For the neutrophils and lymphocytes, the results obtained by the two
instruments confer essentially similar relative components of variance. It is noteworthy that the consistent
group-diurnal variance for lymphocytes and eosinophils
accounts for approximately 50% of the total withinday physiologic variance, and in the case of plasma
Cortisol variances, the consistent group-diurnal variance amounts to approximately 60% of the total withinday physiologic variance.
Table 4 presents the mean values at each hour divided
by the mean of values obtained at 700 h for total leukocytes, neutrophils, lymphocytes, and eosinophils. The
eosinophils and lymphocytes showed the most dramatic, consistent group-diurnal variation.
The grand mean of plasma Cortisol determined for
eight of the subjects at six hours of the day and on three
separate days was 96.6 /u.g/1. The total intraindividual
(within-subject) CV for plasma Cortisol was 56.2%,
with the within-day CV being 53.9% and the day-to-day
CV being 15.8%. The group-specific diurnal variations
for plasma Cortisol, blood eosinophils, and blood
lymphocytes are illustrated in Figure 2.
The plasma Cortisol concentration goes down during
the day, the eosinophil concentration drops to lowest
values at 1000 h to 1600 h and then rises again, and the
concentration of blood lymphocytes gradually increases during the day (see also Table 4). The changes
697
Table 3. Components of Within-day Physiologic
Variation of Blood Concentrations of Leukocyte
Types and Plasma Cortisol Concentrations
of Healthy Subjects*
Consistentgroup
Diurnal
Variation
Consistentsubject
Specific
Diurnal
Variation
Random
Diurnal
Variation
Total leukocytes
Hemalog-D®
0.299
0.285
0.486
Neutrophils
Hemalog-D5
Hematrak"
0.109
0.060
0.326
0.356
0.565
0.584
Lymphocytes
Hemalog-D
Hematrak
0.412
0.467
0.208
0.180
0.380
0.353
Monocytes
Hematrak
0.089
0.015
0.896
Eosinophils
Hemalog-D
0.439
0.205
0.356
Basophils
Hemalog-D
0.143
0.476
0.381
Plasma cortisolt
Radioimmunoassay
0.579
0.228
0.193
Quantity
Measured by
* Variance of each component divided by the total within-day physiologic variance.
t Variances computed on the basis of results obtained from eight of the 21 subjects.
in the concentration of neutrophils are not always
correlated with the changes in the concentration of
eosinophils and also seem to be disparate with certain
of the changes in the lymphocyte concentration.
Table 5 compares the means and the standard deviations of the leukocyte types for the seven smokers and
the 14 nonsmokers. There was a statistically significant
difference in mean values, with the smokers having
higher values in leukocyte and neutrophil concentrations.
Table 4. Average Concentrations of Selected Leukocyte Types at Various Hours of the Day Based on
Observations Obtained from 21 Healthy Subjects over Three Days
Quantity
Measured By
Total leukocytes
Hemalog-D"*
Neutrophils
Hemalog-D
Hematrak"8
Lymphocytes
Hemalog-D
Hematrak
Eosinophils
Hemalog-D
Mean of Values Obtained at Stated Hours of the Day Divided
by That Obtained at 700 h Times 100
Mean of 700 h
Numerical
Values
700 h
1000 h
1300 h
1600 h
1900 h
2200 h
7,080
100
102.3
101.7
106.4
113.6
102.3
4,092
4,151
100
100
108.7
110.2
106.0
107.6
113.0
112.6
116.5
114.9
125.0
121.9
2,285
2,177
100
100
94.8
91.8
98.6
95.7
98.8
97.1
115.4
114.9
117.9
124.6
182
100
65.9
67.0
74.7
89.6
98.4
698
150 r
125
^-—-:..
UJ
_l
2 100
o
u.
A.J.C.P. • May 1981
WINKELE7-AZ..
o
75
o
t2
UJ
o 50 rr
UJ
Q.
25 -
O
a
D
O EOSINOPHILS
A LYMPHOCYTES
D NEUTROPHILS
0110
13
16
19
HOUR OF THE DAY
22
FIG. 2. The grand mean of each hour of the day divided by the 700 h
mean for plasma Cortisol, blood eosinophil, blood lymphocyte, and
blood neutrophil concentrations measured in eight healthy subjects
on six hours of the day on three days.
Discussion
Because of the unique characteristics of the two systems (i.e., the Hemalog-D classifies cells on the basis of
size and cytochemical characteristics, whereas the
Hematrak classifies cells on the basis of the attributes
observed on a microscope slide), we were interested in
evaluating the biologic sources of variation when using
the two analyzers.
Each of the automated systems had a unique contribution to give to this study. An opportunity to use
both, therefore, was very attractive. For example, it
was expected that the Hemalog-D, with its improved
precision, would be able to give more information concerning the within-day physiologic variation. As indicated in Tables 4 and 5, both instruments are virtually
identical in providing information concerning biologic
variation in total neutrophils and lymphocytes. However, for cell types that occur in rather low concentrations (eosinophils and basophils), the better precision
of the Hemalog-D allows one to penetrate the details
of the pattern of biologic variation. Thus, the individual
shifts in the daily mean levels (subject-hour interactions) and the consistent, but individual, diurnal patterns (subject-hour interactions) could not be discerned
by the Hematrak, because the corresponding mean
squares of the analysis of variance model were not
significantly different from those caused by random
biologic fluctuations^
Another example of the different contributions of the
two instruments is the interpretation of the monocyte
count. Here the methods are clearly different: the
§ The assumption that the measurements from blood samples
taken on the same day from a particular subject are necessarily independent may not be fulfilled, owing to the relatively short interval
between blood samplings. However, this lack of independence may
be taken
into account by reducing the degrees of freedom of the Ftests.2 When the latter approach—the most conservative one—was
used, the conclusions based on a 0.05 significance level were unchanged.
Table 5. Comparison of Numerical Values of Leukocyte Cell Type Concentrations of
21 Healthy Subjects: Smokers Versus Nonsmokers
Nonsmokers
Smokers
Quantity
Measured By
Total leukocytes
Hemalog-D*
Neutrophils
Hematrak0*
Hemalog-D
Lymphocytes
Hematrak
Hemalog-D
Monocytes
Hematrak
Eosinophils
Hemalog-D
Basophils
Hemalog-D
* NS, not significant.
Mean
SD
Mean
SD
Probability
of No
Difference*
9,215
876
6,793
1,055
P< 0.01
5,900
5,849
823
894
3,973
3,922
956
853
P< 0.01
P< 0.01
2,536
2,630
610
642
2,125
2,258
778
718
NS
NS
589
98
507
115
NS
178
76.0
136
66.6
NS
16.0
53.4
17.3
NS
58.3
Vol. 75 • No. 5
WITHIN-DAY VARIATION OF LEUKOCYTES
Hematrak assesses monocytes by pattern, and the
Hemlog-D assesses monocytes by enzymatic activity.
The argument may be made that without enzymatic
activity, the cell should not be called a monocyte. In
principle, we prefer the histochemical technic of the
Hemalog-D. However, the monocyte channel probably
requires very careful attention to function properly.
One of the reasons may very well be that good qualitycontrol material is not yet available for this channel on
the Hemalog-D.
Another example of differences between the two instruments is the assessment of the basophil count.
Here, the precision of the Hemalog-D makes it clearly
superior. Moreover, Hematrak also suffered from a
design deficiency in only reporting a total of 101 unique
number fraction values, i.e., all integers from " 0 " to
"100" such that in evaluating a total of 200 cells and
when counting an odd number of basophils, i.e., (2n
+ 1) for a particular cell type, the reported number is
increased to (2n + 2); i.e., a basophil value of 0.5 is
reported as 1.0.
An obvious omission of the manual-eye count results
in this study deserves some comment. We considered
that the manual-eye count procedure was not necessary, since numerous previous reports have made clear
the high degree of agreement between automated and
manual-eye count reference differential obtained from
smears from healthy subjects. 7 In addition, the increased throughput for the automated systems, the
decreased random analytic variation, the elimination of
any technician-to-technician difference, and the possible within-technician variability in reading and identifying cell types would all be avoided when using the
automated systems.
Table 6 compares values obtained from this study and
a previous study 9 for the total within-day physiologic
variance and the contribution made by the consistent
group-specific diurnal variance to the total physiologic
variance. In general, the total within-day physiologic
variance is approximately the same as earlier. The increases in the present study, as compared with the
previous study, in the within-day physiologic variations
for the total leukocyte count, the blood neutrophils,
and the eosinophils can be attributed to the fact that we
are monitoring these values over a longer period of the
day (15 hours in the present study versus six hours in
the previous study). This explanation is supported by
the fact that the systematic diurnal variation played a
much more important role in the case of the number of
the leukocyte types in the present study than it did
in the previous study (see Table 6). The magnitudes
of the within-day physiologic variation compare very
favorably with those reported in the literature for
eosinophils 11 " and for basophils".
699
Table 6. Systematic Diurnal Variation and Total
Within-day Physiologic Variation:
A Comparison of Two Studies*
Total Within-day
Physiologic Variation
in Terms of CV, %
Systematic Diurnal
Variation Divided
by Total Within-day
Physiologic Variance
Quantity
Study I
Study 11
Study I
Study II
Total leukocytes
Neutrophils
Lymphocytes
Monocytes
Eosinophils
Basophils
9.4
12.9
10.3
18.6
19.9
7.4
12.6
19.4
13.8
13.4
27.2
8.3
0.000
0.000
0.000
0.000
0.280
0.184
0.229
0.109
0.412
0.089
0.439
0.143
* Study I9: 20 volunteers—10 males. 10 females: 7 smokers and 13 nonsmokers.
Venipunctures obtained at 800. 1100. and 1400 h. Study II (present study): 21 volunteers—
11 males, 10 females: 7 smokers and 14 nonsmokers. Venipunctures obtained at 700,
1000, 1300, 1600, 1900, and 2200 h.
The changes seen in the plasma Cortisol concentration are consistent with those reported previously/ 1 The
lack of complete parallelism between the variation of
the eosinophil concentration and that of the plasma
Cortisol concentration may be due to a delayed effect of
the Cortisol on the eosinophils. Thus, there might be a
lag phase before the effect of a decline in the Cortisol
concentration can be observed. Whether the rise in the
lymphocyte concentration that is paralleled by a fall in
the plasma Cortisol values reflects a causal relationship
between those two quantities cannot be decided on the
basis of the study.
The differences in this study between smokers and
nonsmokers in terms of leukocyte types (Table 5) are
consistent with information reported from very large
epidemiologic investigations. 4 " In a subsequent paper
in this series, we investigate the question of whether the
differences between smokers and nonsmokers in this
study are due to acute or chronic effects of smoking
cigarettes. 1 - There are other demographic characteristics that must also be taken into account when evaluating reference values for leukocyte types: these include
sex and age of the subjects. 4 There is a higher proportion of women among the smokers than among the nonsmokers; however, when we compared the female
smokers with the female nonsmokers, we still found the
same type of differences between smokers and nonsmokers.
The magnitudes of the within-day physiologic variation of leukocyte types may be very dependent on the
degree of control in the experimental setting. The factors of posture of the volunteers before blood is obtained, constraints concerning physical activity, length
of time the smokers abstain from cigarettes prior to
venipuncture, and time since the last meal could all
affect the physiologic within-day variation. When
700
WINKEL £7" AL.
evaluating other studies concerned with leukocyte
counts, the critical reader must examine the experimental protocol for these factors.
Furthermore, minor differences between results
should be expected, depending on the method used.
Thus, in this present study we found that for some of
the cell types there was a small bias between the two
methods, depending on the subject examined. However, considering that the two methods are based on
quite different principles, this is not surprising.
What are the implications of this investigation concerning the timing of venipuncture in the planned clinical setting? The practical implications of the study may
be derived by comparing the systematic within-day
variation with the total within-day variation. A large
group-specific systematic variation implies that the
within-day variation may be predicted and that reference intervals may be expressed as a function of time of
day. A large subject-specific systematic variation implies that the variation to be expected when a subject is
followed-up by using his own values as control can be
diminished considerably if we always obtain the specimens at the same time of the day. When using groupspecific reference intervals, it is important to define the
time of day for venipuncture in the case of blood lymphocytes and blood eosinophils (Table 3). When monitoring a person whose own values are used as baseline
data, we diminish the within-day physiologic variation
dramatically when we define the time of day for venipuncture in all cases except for the blood monocytes.
The knowledge of the magnitudes of systematic and
nonsystematic variations of the concentration of leuko-
A.J.C.P. • May 1981
cyte types in healthy subjects should assist us in distinguishing pathologic and nonpathologic changes in the
concentrations of such quantities.
References
1. Acland JD, Gould AH: Normal variation in the count of circulating eosinophils in man. J Physiol 133:456-466, 1956
2. Greenhouse SW, Geisser S: On methods in the analysis of profile
data. Psychometrika 24: 95, 1959
3. Halberg F: Rhythmic interaction of steroidal and neural functions, Humoral steroids. Edited by L Martini et al.
Amsterdam, Excerpta Med Found, 1967, pp 966-979
4. Helman N, Rubenstein LS: The effects of age, sex and smoking
on erythrocytes and leukocytes. Am J Clin Pathol 63:35-44,
1975
5. Mansberg HP, Saunders AM, Groner W: The Hemalog-D white
cell differential system. J Histochem Cytochem 22:711-724,
1974
6. Noble RC, Penny BB: Comparison of leukocyte count and function in smoking and nonsmoking young men. Infect Immun
12:550-555, 1975
7. Pierre PV, O'Sullivan MB: Evaluation of the Hemalog-D™ Automated differential leukocyte counter. Mayo Clin Proc 49:870887, 1974
8. Rorsman H: Normal variation in the count of circulating basophil
leukocytes in man. Acta Allergol 18:49-65, 1962
9. Statland BE, Winkel P, Harris SC, et al: Evaluation of biologic
sources of variation of leukocyte counts and other hematologic quantities using very precise automated analyzers. Am J
Clin Pathol 69:48-54, 1978
10. Uhrbrand H: The number of circulating eosinophils. Acta Med
Scand 160:99-112, 1958
11. Winkel P, Statland BE, Bokelund H. Effects of time of venipuncture on variation of serum constituents. Consideration of within-day and day-to-day changes in a group of healthy young
men. Am J Clin Pathol 64:433-447, 1975
12. Winkel P, Statland BE: The acute effect of cigarette smoking on
the concentrations of blood leukocyte types in healthy young
women. Am J Clin Pathol (In press)