Content uploaded by Peggy L Horn
Author content
All content in this area was uploaded by Peggy L Horn
Content may be subject to copyright.
Eur J Appl Physiol
DOI 10.1007/s00421-010-1573-9
123
ORIGINAL ARTICLE
Lower white blood cell counts in elite athletes training
for highly aerobic sports
P. L. Horn · D. B. Pyne · W. G. Hopkins · C. J. Barnes
Accepted: 5 July 2010
© Springer-Verlag 2010
Abstract White cell counts at rest might be lower in
athletes participating in selected endurance-type sports.
Here, we analysed blood tests of elite athletes collected over
a 10-year period. Reference ranges were established for 14
female and 14 male sports involving 3,679 samples from
937 females and 4,654 samples from 1,310 males. Total
white blood cell counts and counts of neutrophils, lympho-
cytes and monocytes were quantiWed. Each sport was scaled
(1–5) for its perceived metabolic stress (aerobic–anaerobic)
and mechanical stress (concentric–eccentric) by 13 sports
physiologists. Substantially lower total white cell and neu-
trophil counts were observed in aerobic sports of cycling and
triathlon (»16% of test results below the normal reference
range) compared with team or skill-based sports such as
water polo, cricket and volleyball. Mechanical stress of
sports had less eVect on the distribution of cell counts. The
lower white cell counts in athletes in aerobic sports probably
represent an adaptive response, not underlying pathology.
Keywords Elite athletes · WBC counts · Sports ·
InXammation · Neutropaenia
Introduction
Most interest in the haematology of athletes focuses on the
number, size and haemoglobin content of their red blood
cells. Given the central role in oxygen delivery to the tis-
sues and ultimately exercise performance, this interest is
understandable. However, as the body’s defenders against
infection, white blood cells (WBC) also contribute indi-
rectly to performance by keeping athletes well enough
(infection free) to maintain their training programmes.
Information about WBC numbers in elite athletes has
received little attention, even though in many cases, this
information is derived at the same time and from the same
blood sample as the red cell information. Previous studies
of WBCs in athlete populations have considered only total
WBCs (Telford and Cunningham 1991), only one of the
Wve types of WBCs (Parisotto et al. 2003), or WBCs in only
one sport (Bain et al. 2000; Lesesve et al. 2000; Watson and
Meiklejohn 2001).
For more than 10 years, a haematology database has
been maintained in the Sports Science and Sports Medicine
Centre at the Australian Institute of Sport. Some of these
haematology data have been part of medical–clinical inves-
tigations on unwell athletes, but the majority have been col-
lected on healthy, elite athletes across a range of sports for
routine monitoring or research studies. Haematological
analyses in the database are linked with additional informa-
tion such as the athletes’ age (at date of collection), sex and
their sport. Our interest is in how exercise itself, in the
absence of any underlying inXammatory or immunological
responses, aVects changes in WBC numbers. Such changes
Communicated by Susan Ward.
P. L. Horn (&) · D. B. Pyne · C. J. Barnes
Australian Sports Commission, Australian Institute of Sport,
Sport Science and Sport Medicine, Bruce, ACT, Australia
e-mail: peggy.horn@ausport.gov.au
W. G. Hopkins
Institute of Sport and Recreation Research,
AUT University, Auckland, New Zealand
D. B. Pyne
Medical School,
The Australian National University, Canberra, Australia
Eur J Appl Physiol
123
might reXect adaptation to the metabolic and mechanical
stressors evident in individual and team sports (Pyne 1994),
rather than an underlying pathological response. Our aim
was therefore to establish sport-speciWc WBC reference
ranges and compare WBC values between sports against
standard clinical reference values.
Methods
Experimental approach
We retrospectively examined blood test results of rested,
healthy elite athletes (presenting without illness) collected
over a 10-year period in a haematology database that was
linked to demographic information on age, sex and sport.
All samples were from national scholarship holders thereby
deWning their elite status at the Australian level. The mean
(§SD) age for athletes in diVerent sports ranged from
16 §1 years for female gymnasts to 26 §4 years for male
cyclists.
Haematological data
Names of all athletes were removed from the analysis to
preserve the anonymity. We then removed spurious and/or
suspicious data as well as eliminating data that we sus-
pected had been collected as part of medical–clinical inves-
tigations on unwell athletes. The elimination of these results
was based upon whether the requesting oYcer was one of
the Institutes’ medical practitioners, if the sample was
accompanied by a request for pathology, or had ‘unwell’
written into the comment section on the request form.
Although our database did not identify the ethnicity of the
athletes, the overwhelming proportion was Caucasian. This
study was approved by the Ethics Committee of the Austra-
lian Institute of Sport (approval number 2007-1011).
If more than one sample was collected on a single day
from an individual athlete, we included only the Wrst sam-
ple of the day in our analysis. This step was undertaken to
ensure that the collection was an overnight ‘at rest’ sample
and not taken post-exercise. We also removed from the
dataset, all samples from athletes identiWed as disabled, less
than 14 or over 60 years of age. Our sport-speciWc retro-
spective study meant that the athlete groups were relatively
homogeneous: we included in our analysis only the sports
in which the number of diVerent individual athlete observa-
tions was greater than 20. Data presented are from 14
female and 14 male sports, and 11 of these sports were
common to both sexes. Rugby, canoeing and boxing only
had suYcient numbers of male participants, whereas gym-
nastics, netball and cricket only had suYcient female par-
ticipants for inclusion in our analysis.
ClassiWcation of sports by metabolic and mechanical stress
To understand what aspects of diVerent sports might be
inXuencing the WBC counts, we surveyed a sample of
sports physiologists at the Australian Institute of Sport. For
each of the 14 sports (for each sex), we asked physiologists
to quantify (using a 5 point Likert scale) the relative contri-
bution of the nature of each sport (during competition and
in training) in relation to perceived major metabolic energy
system used in a given sport (from dominantly aerobic to
dominantly anaerobic). We also asked each physiologist to
quantify (using another 5 point Likert scale) the relative
contribution of the mechanical nature of those sports from
purely concentric muscle demands to dominantly eccentric
muscle demands.
Analytical equipment
During the Wrst 4 years of the 10-year data collection
period, haematology results were generated on an H3 Tech-
nicon auto-analyzer (Bayer, Tarrytown, NY). The majority
of haematology results were gathered on an ADVIA-120
Hematology analyzer (Bayer Health Care Diagnostics, Tar-
rytown, NY). Across the entire 10 years, only two technical
operators had primary responsibility for equipment calibra-
tion, maintenance, data entry and monthly QC assessments
for WBC totals and diVerential counts. Selected haematol-
ogy values were also routinely submitted to the quality
assurance programme of the Royal College of Pathology,
Australasia.
Statistical analysis
Values for total WBC, neutrophils, lymphocytes, mono-
cytes, eosinophils and basophils were log transformed
before analysis with a mixed linear model using SAS soft-
ware (Statistical Analysis System, Version 9.1, SAS Insti-
tute, Cary, NC). The Wxed eVects in the model were the
identity of the sport (nominal) and age of the athlete (qua-
dratic). The random eVects were the identity of the athlete
and the residual, which represents error between repeated
measurements on a given athlete. A diVerent residual was
speciWed for each sport. Outliers were identiWed as observa-
tions with a standardised residual >5.0; these observations,
which represented typically 0.15% of the total were deleted
and the analysis was repeated. Separate analyses were per-
formed for the total WBC and individual cell types, as well
as for females and males. Some athletes were sampled sev-
eral times but these repeated measures were accounted for
in the mixed modelling analysis.
The mean values of the cell counts for each sport are the
back-transformed means adjusted to the mean age of all
athletes. The magnitude of an eVect on mean cell count was
Eur J Appl Physiol
123
assessed by standardization: the diVerence in the count was
divided by the between-subject standard deviation, which
was derived by taking the square root of the sum of the
variances for the athlete and mean residual. The resulting
standardised eVect was compared to thresholds of 0.20,
0.60 and 1.20 for small, moderate and large (Hopkins et al.
2009). These values are also compared to published refer-
ence intervals found in the literature including the deWni-
tions of what were considered to be lowered cell counts.
For example, a threshold of <2.0 £109/L was set to deWne
neutropenia.
Alpha reliabilities of the 13 completed sport surveys
comparing relative aerobic to anaerobic and relative con-
centric to eccentric nature of the 14 sports were calculated
with the Statistical Analysis System (Version 9.1, SAS
Institute, Cary, NC). Pearson’s correlation coeYcients were
used to characterise the degree of association of cell counts
with mean aerobic–anaerobic and concentric–eccentric
scores; magnitudes were interpreted using thresholds of
0.10, 0.30 and 0.50 for small, moderate and large eVects
(Hopkins et al. 2009).
Results
Cell counts
Tables 1 and 2 present WBC Wndings (total and sub-types)
from male and female sports, respectively, along with sport
speciWc and clinical reference ranges. The highest mean
WBC counts were for the male team sports of rugby and
water polo, and cricket and water polo for female sports.
The lowest total WBC counts in both males and females
were the individual sports of cycling and triathlon.
Neutrophil counts generally mirrored total WBC counts
with the lowest neutrophil counts for both male and female
sports being cycling and triathlon. DiVerences in neutrophil
counts between these team and individual sports were gen-
erally small to moderate in magnitude. Neutropenia (deW-
ned as <2 £109/L) was seen in 5% of samples across all
our sports and in 17% of cycling and 16% of triathlon sam-
ples (Table 3). Approximately 11% of athletes had at least
one episode of neutropenia but athletes in the sports like
swimming, triathlon and cycling experienced more than
twice that percentage of episodes.
Male and female swimmers had moderately higher
lymphocyte counts than other sports (yet still within the
clinical reference range), whereas the lowest lymphocytes
counts were noted in male canoeists and the female team
sports of cricket and volleyball. Lymphopenia (deWned as
<1.0 £109/L) was seen across 2.4% of all samples from all
sports (Table 3): overall 5.1% of athletes experienced at
least one episode of lymphopenia.
The lowest mean monocyte values for both sexes were
seen in cycling and triathlon. Monocytopenia (deWned as
<0.2 £109/L) was noted in 2.4% of all samples and 6.4%
of athletes experienced at least one episode (Table 3).
Female eosinophil values were consistently lower than
Table 1 White blood cell counts in male athletes
Sport nTotal white blood
cell count, 109/L
Neutrophil count,
109/L
Lymphocyte count,
109/L
Monocyte count,
109/L
Mean 95% Reference
range
Mean 95% Reference
range
Mean 95% Reference
range
Mean 95% Reference
range
Archery 21 6.7 4.1–10.9 3.7 1.8–7.6 1.9 1.1–3.3 0.42 0.22–0.83
Athletics 113 6.3 3.9–10.2 3.5 1.7–7.1 1.9 1.2–3.1 0.39 0.22–0.70
Basketball 101 6.6 4.3–10.3 3.7 1.9–7.2 1.9 1.1–3.2 0.41 0.22–0.74
Boxing 59 7.2 4.5–11.6 4.0 2.0–8.2 2.1 1.3–3.4 0.47 0.26–0.84
Canoeing 25 6.2 3.7–10.2 3.5 1.7–7.2 1.7 1.0–3.0 0.45 0.24–0.85
Cycling 173 5.7 3.7–8.8 2.8 1.5–5.5 2.0 1.2–3.3 0.36 0.20–0.65
Rowing 195 6.1 3.7–10.1 3.4 1.6–6.9 1.9 1.1–3.2 0.39 0.22–0.71
Rugby Union, AFL 150 7.4 4.7–11.6 4.2 2.3–7.9 2.1 1.3–3.2 0.45 0.26–0.79
Swimming 127 6.7 4.3–10.3 3.4 1.8–6.5 2.2 1.4–3.7 0.44 0.24–0.79
Soccer 165 6.9 4.1–11.7 3.7 1.7–8.0 2.0 1.1–3.5 0.42 0.23–0.75
Triathlon 48 5.9 3.5–9.9 2.9 1.3–6.4 2.0 1.2–3.5 0.35 0.18–0.69
Volleyball 50 6.9 4.2–11.4 4.0 1.9–8.1 1.9 1.2–3.1 0.43 0.22–0.87
Winter sports 32 6.4 4.2–9.8 3.5 1.7–7.0 1.9 1.0–3.5 0.37 0.21–0.66
Water polo 51 7.4 4.3–12.5 4.3 1.9–9.4 2.0 1.1–3.6 0.43 0.21–0.87
All sports 1,310 6.6 3.9–11.1 3.6 1.7–7.7 2.0 1.1–3.5 0.41 0.21–0.81
Normal reference range 4.5–11.0 2.0–8.0 1.0–4.8 0.2–0.78
Eur J Appl Physiol
123
male eosinophil values and swimmers had the highest mean
values across all sports. Female swimmers and male boxers
had the highest mean basophil counts.
EVects of sport-related metabolic and mechanical stresses
The alpha reliability for the Likert ratings of sports on the
aerobic–anaerobic scale, were 0.88 and 0.90 for females
and males, respectively; on the concentric–eccentric scale,
the corresponding alphas were 0.96 and 0.95. Mean aero-
bic–anaerobic scores ranged from lows of 1.2 and 1.4 for
female and male triathlon through to highs of 4.2 for vol-
leyball (both sexes). The range for the concentric–eccentric
scores was 1.5 for cycling (both sexes) to 4.0 for volleyball
(both sexes). There was a large correlation between the two
scores for the females (0.57; 90% conWdence limits §0.34)
and the males (0.58; §0.34).
The aerobic–anaerobic score had large positive correla-
tions with neutrophil and WBC counts for females and
males (0.52–0.70), a large positive correlation for mono-
cytes in males (0.53), a moderate positive correlation for
monocytes in females (0.34), a trivial correlation for lym-
phocytes in males (0.08) and a large negative correlation
for lymphocytes in females (¡0.56). The correlations
between cell counts and the concentric–eccentric score
were generally trivial to small, with the exception of mod-
erate correlations for neutrophils and WBC in males (0.36
and 0.31) and a large negative correlation for lymphocytes
in females (¡0.56). The conWdence limits for these correla-
tions ranged from §0.26 to §0.46 for the large through to
the small correlations, respectively (Fig. 1).
Discussion
This study has expanded the available published informa-
tion of normal WBC values for elite athletes by investigat-
ing a wide range of sports and detailing counts of speciWc
WBC types. The mean total WBC count across all sports
is similar to that for elite athletes in the existing literature.
Table 2 White blood cell counts in female athletes
Sport nTotal white blood
cell count, 109/L
Neutrophil count,
109/L
Lymphocyte count,
109/L
Monocyte count,
109/L
Mean 95% Reference
range
Mean 95% Reference
range
Mean 95% Reference
range
Mean 95% Reference
range
Archery 21 7.0 4.0–12.3 4.1 1.8–9.4 2.0 1.0–3.8 0.36 0.15–0.81
Athletics 66 6.2 3.8–10.1 3.5 1.6–7.6 1.9 1.1–3.0 0.33 0.18–0.62
Basketball 99 6.5 3.9–10.7 3.6 1.7–7.4 1.9 1.2–3.2 0.37 0.20–0.69
Cricket 41 7.1 4.0–12.5 4.3 1.8–10.4 1.7 1.1–2.8 0.40 0.19–0.81
Cycling 101 5.9 3.5–9.8 2.9 1.3–6.5 2.0 1.1–3.6 0.33 0.18–0.63
Gymnastics 40 6.7 3.9–11.5 3.9 1.8–8.4 1.8 1.1–3.0 0.36 0.18–0.69
Netball 122 6.8 4.2–11.1 4.0 1.9–8.2 1.9 1.1–3.3 0.38 0.20–0.71
Rowing 108 6.4 4.0–10.3 3.6 1.8–7.1 2.0 1.2–3.3 0.36 0.20–0.66
Swimming 114 6.5 4.2–10.2 3.2 1.6–6.3 2.4 1.4–4.0 0.35 0.19–0.66
Soccer 80 6.3 3.9–10.4 3.6 1.8–7.2 1.9 1.1–3.2 0.39 0.22–0.71
Triathlon 33 5.9 3.8–9.2 2.9 1.5–5.4 2.1 1.3–3.3 0.33 0.18–0.60
Volleyball 38 6.8 4.2–10.9 4.2 2.0–8.6 1.7 1.1–2.6 0.36 0.21–0.63
Winter sports 38 6.4 4.2–9.8 3.6 2.0–6.6 1.9 1.1–3.4 0.34 0.18–0.63
Water polo 35 7.3 4.1–13.0 4.6 2.0–10.5 1.8 1.1–3.0 0.36 0.18–0.70
All sports 937 6.5 3.8–11.2 3.7 1.7–8.1 1.9 1.1–3.4 0.36 0.18–0.72
Normal reference range 4.5–11.0 2.0–8.0 1.0–4.8 0.2–0.78
Table 3 Sports with low white blood cell sub-types counts
Neutrophils <2.0 £109/L Lymphocytes <1.0 £109/L Monocytes <0.2 £109/L
Across all samples 5% 2% 2%
Two sports with most samples having
low cell counts (% with low counts)
Cycling, 17%
Triathlon, 16%
Archery, 5%
Canoeing, 5%
Gymnastics, 5%
Triathlon, 5%
Eur J Appl Physiol
123
A number of other studies have also noted lower WBC
numbers in endurance-type sports compared with team-based
sports (Parisotto et al. 2003; Telford and Cunningham
1991). Here, we have shown that more aerobically oriented
sports tend to have lower WBC and neutrophil counts. This
observation has implications for sports physicians or others
involved in haematological assessment of healthy athletes
in regular training.
A Wnding in common across studies of WBC values in
high-level athletes is the observation of low neutrophil
counts (neutropenia). Several studies have reported neutro-
penia in diVerent groups of athletes including marathon
runners, cyclists, professional footballers and cyclists
(Bain et al. 2000; Lesesve et al. 2000; Parisotto et al. 2003;
Watson and Meiklejohn 2001). Our Wnding of 5.3% neutro-
penia compares with a 35% incidence of neutropenia
observed in professional footballers (Watson and Meiklejohn
2001). Other studies of cyclists reported 16, 29 and 38%
neutropenia amongst male track, road and mountain bikers,
respectively (Parisotto et al. 2003) and an 11% incidence in
male and 20% in female cyclists (Lesesve et al. 2000). The
latter diVerential cell counts were done manually on stained
blood-Wlm preparations and the 20% Wgure they reported
among their female athletes is based on only Wve individu-
als. Neutropenia, then, seems to occur relatively commonly
amongst athletes, especially in endurance athletes com-
pared to clinical reference samples. Our Wnding is lower
than other data reported but presents a more comprehensive
estimate of WBC in athletic populations given the large
numbers of sports and athletes in our analysis.
The observation of neutropenia is of clinical interest
because, in general, neutropenic individuals have increased
susceptibility to bacterial infections. Given the presence of
multidrug-resistant bacteria in the community (including
the sporting community), athletes must be vigilant with
personal hygiene and pay attention to even seemingly
trivial skin wounds (Buss et al. 2009; Redziniak et al. 2009;
Saunders 2009). As the low total neutrophil counts reported
in our studies are from active healthy people at rest, we
consider these Wndings most likely reXect a training-
induced adaptive anti-inXammatory response operating
within broader homeostatic limits. We presume that the
exclusion of unwell individuals minimised the likelihood
that immune cells had compromised functional activity.
The reasons for the exercise-induced low neutrophil counts
are unclear, but could include decreased cell production by
bone marrow stem-cell precursors, increased cell destruc-
tion, increased transit rate into tissues, increased endothe-
lial adhesion, or a mix of these mechanisms (Barreda et al.
2004).
Neutrophils are short-lived circulating cells, so they
must be constantly replaced through actions of growth fac-
tors such as granulocyte colony-stimulating factor (G-CSF)
which stimulates bone marrow hematopoietic progeni-
tor\stem cells (HPCs). Exercise elevates levels of G-CSF
(Bonsignore et al. 2002; Yamada et al. 2002) and our elite
athletes would be routinely experiencing such pulsatile
increases of G-CSF. The observation of low circulating
neutrophil counts was therefore unexpected. If transiently
elevated levels of G-CSF do not result in increased num-
bers of neutrophils, the observed neutropenia could be asso-
ciated with low numbers of HPC, the G-CSF target
population. A study of 30 elite triathletes reported lower
numbers of HPC than in 38 sedentary controls (Philip and
Bermon 2003). In contrast, another study found levels of
HPC were higher in runners compared with controls
(Bonsignore et al. 2002). A third study of HPC numbers
showed little diVerence between trained and untrained indi-
viduals (Wardyn et al. 2008). These conXicting Wndings
reXect the challenges in deWning and enumerating HPC
counts in peripheral blood. Overall though, information on
the exercise-associated pulsing of G-CSF along with nor-
mal or elevated target HPC numbers is at odds with the
observed neutropenia in our highly trained endurance ath-
letes. Stated another way, reduced neutrophil production
does not seem to be the explanation for our observed elite
athlete neutropenia.
Could the observed elite athlete neutropenia be associ-
ated with decreased neutrophil lifespan in the blood? Circu-
lating neutrophils can undergo apoptotic cell death and
exercise-induced WBC and neutrophil apoptosis gene
expression pathways have been investigated (Radom-Aizik
et al. 2008). In this study, the Jak/STAT pathway, known to
inhibit apoptosis, was signiWcantly activated by 30 min of
aerobic exercise, but another 14 genes were altered in a
way that was likely to accelerate neutrophil apoptosis.
Additionally, exercise aVected WBC gene expression in a
dose-dependent manner and genes for stress (heat shock)
proteins were substantially altered by a treadmill running
Fig. 1 Relationship between neutrophil cell count (log scale) and the
aerobic–anaerobic rating of the sport. Each point represents means for
a sport. SD bar is the within-sport, between-athlete standard deviation
in the neutrophil counts averaged over all the sports. Regression lines
are shown
2
1234
Anaerobic
Neutrophils
(109/L)
Female
Male
3
4
5
SD
Aerobic
Eur J Appl Physiol
123
protocol (Buttner et al. 2007). Unfortunately, the investiga-
tors did not include speciWc apoptosis genes in their screen-
ing procedures. Exercise-induced neutrophil apoptosis
could thus be an explanation for the observed neutropenia
by decreasing overall neutrophil life span.
A Wnal consideration to explain the low neutrophil
counts in cyclists and triathletes is the possibility of plasma
volume expansion. Expansion of plasma volume has long
been acknowledged as an adaptation to thermal and non-
thermal aspects of endurance exercise. A comprehensive
review of 11 experimental studies of short and long-term
endurance exercise (Convertino 1991) reported that the
mean percentage increase in plasma volume was less than
10% (from pre-exercise values). In comparison, our low-
end neutrophil counts (cyclists and triathletes) were more
than 20% lower than the mean neutrophil counts across all
sports. For example, the neutrophil counts in female
cyclists were 29% lower than the mean for all females.
Given that the magnitude of cell-count diVerences in
cycling and triathlon were larger than the typical change in
plasma volume, we feel that plasma volume expansion is
unlikely to fully account for the observed diVerences.
We deWned low lymphocyte counts (lymphopenia) by a
cut oV of <1.0.109/L. The only other published study
reporting on athlete lymphopenia, deWned as 1–1.5.109/L
(Lesesve et al. 2000), observed 27% among male cyclists
and 20% among female cyclists. As with these authors’
neutropenia data, this latter Wgure is based only on Wve indi-
viduals. In our study, lymphopenia was seen most often in
archery and canoeing; these sports have not previously
been included in published WBC Wndings. We have no
clinical records to suggest that these low counts were asso-
ciated with a clinical history of increased illnesses.
Compared to neutrophils, lymphocytes are relatively
long-lived cells (some surviving for months), but factors
regulating normal circulating numbers are not well under-
stood. Two published studies have reported on the inXuence
of exercise on gene expression of peripheral blood mono-
nuclear cells (PBMCs) which would consist predominately
of lymphocytes. One study had 15 healthy men run for
30 min at »80% of their peak VO2 (Connolly et al. 2004).
PBMCs were collected, RNA isolated, cRNA prepared and
then hybridized to microarrays. Circulating lymphocyte
numbers in their participants increased threefold after exercise,
and 311 genes were diVerentially regulated. Up-regulated
genes were noted in several categories including immune,
inXammatory (more pro-inXammatory than anti-inXamma-
tory) and stress (heat shock proteins and hypoxia-inducible
factor-1, HIF). A more recent paper (Radom-Aizik et al.
2009) used 20 young female participants, and found altered
expression of 622 genes in 11 diVerent gene pathways.
SigniWcant gene pathway changes were related to inXam-
mation, stress (heat shock protein-70) and apoptosis (fas
ligand) pathways. Indeed, lymphocyte apoptosis has been
speciWcally documented post-exercise (Mars et al. 1998;
Mooren et al. 2002), albeit not in the sports (archery and
canoeing) in which we noted increased incidence of
lymphopenia. Exercise-induced lymphocyte apoptosis then
could be part of an explanation of the observed low
lymphocyte counts in athletes.
Monocytes are potent innate defence cells that produce
many (and mostly) pro-inXammatory proteins. Early stud-
ies of monocytes and exercise (Bieger et al. 1980; Rivier
et al. 1994) not only quantiWed the transient, exercise-
induced monocytosis but also the functional alterations to
these cells provoked by exercise (phagocytic activity and
cytokine production). More recently, Timmerman et al.
(2008) reported that exercise training lowered blood mono-
cyte percentages as well as the LPS-stimulated monocyte
TNF- production in 65–80-year-old subjects. As these
studies did not report circulating blood cell numbers, it is
not possible to compare them directly to our younger popu-
lation where we saw 2% low monocytes counts (deWned by
a cut oV of < 0.2.0.109/L), but both sets of results are con-
sistent with exercise having generally anti-inXammatory
inXuences, including lowering WBC cell counts (Mathur
and Pedersen 2009). In general, low monocyte counts were
seen in the same sports that had low neutrophil counts. This
is consistent with exercise having an inXuence on their
common precursor cell in the bone marrow.
Eosinophil values across all sports were within the clini-
cal reference interval with male and female swimmers
having the highest eosinophil counts. Swimming is often
recommended as the sport of choice for asthmatics and high
eosinophil counts are frequently seen with asthma. More
than a decade ago, Helenius et al. (1998) reported that
eosinophils were overrepresented in sputum of elite swim-
mers (as was heightened bronchial responsiveness), but by
limiting our survey to well athletes, we hopefully excluded
athletes experiencing acute asthmatic symptoms and so do
not account for these Wndings in swimmers with a disease
situation. Furthermore, a recent overview of the association
between asthma and blood eosinophils presents a more
complex association than was previously appreciated
(Wenzel 2009).
In an eVort to partly explain the range of normal resting
white blood cell values, we characterised the association
between the physical\mechanical and biochemical\meta-
bolic diVerences and cell counts across the range of sports.
We observed a substantial relationship between the per-
ceived aerobic content of a sport and the total WBC and
subclass cell counts: the more aerobic the sport, the lower
the total WBC, neutrophil and monocyte counts (for
males). There were few substantial relationships between
the concentric\eccentric mechanical nature of diVerent
sports and the WBC. We expected a substantial relationship
Eur J Appl Physiol
123
between them, as many researchers interested in linkages
between WBC and exercise consider that eccentric muscle
damage initiates the recruitment of WBC (and especially
neutrophils and monocytes) to facilitate the subsequent
repair and healing processes. These cells are present in
eccentrically damaged tissues (Malm et al. 2000) but appar-
ently recruitment to damaged muscle is not reXected in
diminished counts of these cells in blood. In contrast, there
was a large negative correlation (¡0.56) between female
lymphocyte counts for both aerobic\anaerobic and concen-
tric\eccentric Likert scores. This latter result is somewhat
unexpected as lymphocytes are not typically associated
with eccentric muscle damage.
Limitations to this study include the absence of detailed
training and medical histories (both short and long term) of
our athletes. Nevertheless, to achieve elite status (and be a
national level scholarship holder), generally, requires years
of rigorous training and ongoing commitment to a sport. In
practice, physicians are likely to interpret results of haema-
tological assessment in view of individual patient (athlete)
presentation including medical and training history, and a
physical examination.
In summary, WBC counts can be lower in elite athletes,
particularly those participating in aerobic endurance sports.
The lower counts probably represent a training-induced
adaptive response in healthy athletes rather than an under-
lying pathological response, and are likely the result of sim-
ilar anti-inXammatory inXuences seen across even non-elite
physically active\exercising populations. Prospective longi-
tudinal studies are needed to assess the relationships
between cell counts in diVerent sports, the eVects of train-
ing type/load, and changes in inXammatory control pro-
cesses. Sports physicians need to be aware that some
athletes in highly aerobic sports will routinely present with
lower WBC, especially neutrophil counts. Our sport-spe-
ciWc reference-range intervals for WBC values should
assist physicians in interpreting haematological test results
for athletes in diagnostic and screening settings.
Acknowledgments The authors acknowledge the cooperation of
athletes and general laboratory staV of the Australian Institute of Sport
in the collection of all the samples over 10 years. We are particularly
grateful for the expert technical contribution of Robin Parisotto and
Graeme Allbon from the Haema tology and Biochemistry Laboratory at
the Australian Institute of Sport.
References
Bain B, Phillips D, Thomson K, Richardson D, Gabriel I (2000) Inves-
tigation of the eVect of marathon running on leukocyte counts of
subjects of diVerent ethnic origins: relevance to the etiology of
ethnic neutropenia. Br J Haematol 108:483–487
Barreda DR, Hanington PC, Belosevic M (2004) Regulation of mye-
loid development and function by colony stimulating factors. Dev
Comp Immunol 28:509–554
Bieger W, Weiss M, Michel G, Weicker H (1980) Exercise-induced
monocytosis and modulation of monocyte function. Int J Sports
Med 1:30–36
Bonsignore M, Morici G, Santoro A, Pagano M, Cascio L, Bonanno B,
Abate P, Mirabella F, ProWta M, Insalaco G, Gioia M, Vignola A,
Majolino I, Testa U, Hogg J (2002) Circulating hematopoietic
progenitor cells in runners. J Appl Physiol 93:1691–1697
Buss B, Mueller S, Theis M, Keyser A, Safranek TJ (2009) Population-
based estimates of methicillin-resistant Staphylococcus areus
(MRSA) infections among high school athletes. J Sch Nurs
25:282–291
Buttner P, Mosig S, Lechtermann A, Funke H, Mooren F (2007) Exer-
cise aVects the gene expression proWles of human white blood
cells. J Appl Physiol 102:26–36
Connolly P, Caiozzo V, Zaldivar F, Nenet D, Larson J, Hung S, Heck
J, HatWeld G, Cooper D (2004) EVects of exercise on gene expres-
sion in human peripheral blood mononuclear cells. J Appl Physiol
97:1461–1469
Convertino VA (1991) Blood volume: its adaptation to endurance
training. Med Sci Sports Exerc 23:1338–1348
Helenius I, Rytila P, Metso T, Haahtela T, Venge P, Tikkanen H (1998)
Respiratory symptoms, bronchial responsiveness and cellular
characteristics of induced sputum in elite swimmers. Allergy
53:346–352
Hopkins W, Marshall S, Batterham A, Hanin J (2009) Progressive Sta-
tistics for studie s in sports medicine and exercise science. Med Sci
Sports Exerc 41:3–13
Lesesve J, Guinot M, Andolfatto S, Bene M, Dine G (2000) EVect of
elite cycling on leukocyte counts. Br J Haematol 110:1006–1014
Malm C, Engstrom M, Sjodin B, Lenkei R, Ekblom B, Lundberg I
(2000) Immunological changes in human skeletal muscle and
blood after eccentric exercise and multiple biopsies. J Physiol
529:243–262
Mars M, Govender S, Weston A, Naicker V, Chuturgoon A (1998)
High intensity exercise: a cause of lymphocyte apoptosis? Bio-
chem Biophys Res Commun 249:366–370
Mathur N, Pedersen BK (2009) Exercise as a mean to control low-grade
systemic inXammation. Mediators InXamm [Epub ahead of print]
Mooren FC, Bloming D, Lechtermann A, Lerch MM, Volker K (2002)
Lymphocyte apoptosis after exhaustive and moderate exercise.
J Appl Physiol 93:147–153
Parisotto R, Pyne D, Martin D, Gore C, Fallon K, Fricker P, Hahn A
(2003) Neutropenia in elite male cyclists. Clin J Sport Med
13:303–305
Philip P, Bermon S (2003) Intensive triathlon training induces low
peripheral CD34+ stem cells. Br J Haematol 120:914–915
Pyne DB (1994) Exercise-induced muscle damage and inXammation:
a review. Aust J Sci Med Sport 22:49–58
Radom-Aizik A, Zaldivar F, Leu S, Galassetti P, Cooper D (2008)
EVects of 30 min of aerobic exercise on gene expression in human
neutrophils. J Appl Physiol 104:236–243
Radom-Aizik S, Zaldivar F, Leu S-Y, Cooper D (2009) A brief bout of
exercise alter gene expression and distinct gene pathways in
peripheral blood mononuclear cells of early- and late-pubertal
females. J Appl Physiol 107:168–175
Redziniak D, Diduch D, Turman K, Hart J, GrindstaV T, Macknight J,
Mistry D (2009) Methicillin-resistant Staphylococcus aureus
(MRSA) in the athlete. Int J Sports Med 30:557–562
Rivier A, Pene J, Chanez P, Anselme F, Caillaud C, Prefaut C, Godard
P, Bousquet J (1994) Release of cytokines by blood monocytes
during strenuous exercise. Int J Sports Med 15:192–198
Saunders J (2009) Reducing MRSA infections in college student ath-
letes. J Community Health Nurs 26:161–172
Telford RD, Cunningham RB (1991) Sex, sport and body-size depen-
dency of hematology in highly trained athletes. Med Sci Sports
Exerc 23:788–794
Eur J Appl Physiol
123
Timmerman KL, Flynn MG, Coen PM, Markofski MM, Pence BD
(2008) Exercise training-induced lowering of inXammatory
(CD14+CD16+) monocytes: a role in the anti-inXammatory inXu-
ence of exercise? J Leukoc Biol 84:1271–1278
Wardyn G, Rennard S, Brusnahan S, McGuire T, Carlson M, Smith L,
McGranaghan S, Sharp J (2008) EVects of exercise on hematolog-
ical parameters, circulating side population cells and cytokines.
Exp Hematol 36:216–223
Watson HG, Meiklejohn DJ (2001) Leucopenia in professional foot-
ball players. Br J Haematol 112:824–827
Wenzel S (2009) Eosinophils in asthma—closing the loop or opening
the door? N Engl J Med 360:1026–1028
Yamada M, Suzuki K, Kudo S, Totsuka M, Nakaji S, Sugawarra K
(2002) Raised plasma G-CSF and IL-6 after exercise may play a
role in neutrophil mobilization into the circulation. J Appl Physiol
92:1789–1794
