Content uploaded by Gøran Paulsen
Author content
All content in this area was uploaded by Gøran Paulsen on Jun 26, 2019
Content may be subject to copyright.
A COX-2 inhibitor reduces muscle soreness, but does not influence
recovery and adaptation after eccentric exercise
G. Paulsen
1
, I. M. Egner
1
, M. Drange
1
, H. Langberg
2
, H. B. Benestad
3
, J. G. Fjeld
4
, J. Halle
´n
1
, T. Raastad
1
1
Norwegian School of Sport Sciences, Oslo, Norway,
2
Institute of Sports Medicine, Bispebjerg Hospital, Copenhagen, Denmark,
3
Department of Physiology, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway,
4
Department of Nuclear Medicine,
Rikshospitalet University Hospital, Oslo, Norway
Corresponding author: Gøran Paulsen, Norwegian School of Sport Sciences, PO Box 4014 U.S., N- 0806 Oslo, Norway. Fax:
147 23 26 24 51, E-mail: goran.paulsen@nih.no
Accepted for publication 16 February 2009
The aim of this study was to investigate the effect of a
cyclooxygenase (COX)-2 inhibitor on the recovery of mus-
cle function, inflammation, regeneration after, and adapta-
tion to, unaccustomed eccentric exercise. Thirty-three
young males and females participated in a double-blind,
placebo-controlled experiment. Seventy unilateral, volun-
tary, maximal eccentric actions with the elbow flexors were
performed twice (bouts 1 and 2) with the same arm, separated
by 3 weeks. The test group participants were administered
400 mg/day of celecoxib for 9 days after bout 1. After both
bouts 1 and 2, concentric and isometric force-generating
capacity was immediately reduced (40–50%), followed by
the later appearance of muscle soreness and increased serum
creatine kinase levels. Radiolabelled autologous leukocytes
(detected by scintigraphy) and monocytes/macrophages (his-
tology) accumulated in the exercised muscles, simultaneously
with increased satellite cell activity. These responses were
reduced and recovery was faster after bout 2 than 1,
demonstrating a repeated-bouteffect. Nodifferences between
the celecoxib and placebo groups were detected, except for
muscle soreness, which was attenuated by celecoxib. In
summary, celecoxib, a COX-2 inhibitor, did not detectably
affect recovery of muscle function or markers of inflamma-
tion and regeneration after unaccustomed eccentric exercise,
nor did the drug influence the repeated-bout effect. However,
it alleviated muscle soreness.
According to some reports, non-steroidal anti-inflam-
matory drugs (NSAIDs) may have detrimental effects
on regeneration and healing processes after injuries to
the different components of the muscle–skeletal system
(Almekinders, 1999). Therefore, warnings have been
voiced against the wide use of NSAIDs in exercise
and sports medicine (Almekinders, 1999; Paoloni &
Orchard, 2005; Warden, 2005; Mehallo et al., 2006).
The original NSAIDs (e.g. ibuprofen) inhibit both
the constitutive cyclooxygenase (COX)-1 and the
more inducible COX-2 enzyme, which both catalyze
the generation of prostanoids [prostaglandins (PGE
2
and PGF
2a
), prostacyclins and thromboxanes] that
are involved in numerous physiological processes
(Miller, 2006). The more recently developed selective
COX-2 inhibitors are in theory more appropriate and
specific to dampen excessive inflammatory reactions
than the non-selective (COX-1 and COX-2) inhibi-
tors (Warden, 2005). However, it has been demon-
strated that these new drugs can also have negative
long-term effects on regeneration and healing in
animals (Warden, 2005; Buvanendran & Reuben,
2006) – especially through inhibition of satellite cell
activity in skeletal muscle (Bondesen et al., 2004; Shen
et al., 2006). But there is presently no clear evidence for
detrimental effects of COX-2 inhibitors on regeneration
and healing of injuries in humans (Reuben & Ekman,
2005, 2007; Warden, 2005; Mehallo et al., 2006;
Vuolteenaho et al., 2008). Furthermore, neither clinical
nor experimental exercise-induced muscle damage stu-
dies have to our knowledge been conducted in order to
evaluate the effect of selective COX-2 inhibitors on
recovery from skeletal muscle injuries in humans.
In humans, some reports have demonstrated that
NSAIDs, such as naproxen and ibuprofen, can reduce
eccentric exercise-induced muscle damage and hasten
the recovery process (Baldwin, 2003; Cheung et al.,
2003; Connolly et al., 2003). The improved recovery
might stem from the anti-inflammatory effect of these
drugs (Sayers et al., 2001), but this assumption has
never been properly documented. Howell et al.
(1998a, b) detected no effect of ibuprofen, but a slightly
retarded recovery of maximal isometric torque after
high-force eccentric exercise was observed in subjects
administered flurbiprofen. Other negative effects of
NSAIDs, such as blunted increase in the rate of protein
synthesis after high-force eccentric exercise (Trappe
et al., 2001, 2002), and reduced satellite cell response
to long distance running (Mackey et al., 2007), have
[This article was published online on 10
th
June 2009. Error were sub-
sequently identified. This notice is included in the print and online
versions to indicate that both have been corrected as at 18
th
January 2010]
Scand J Med Sci Sports 2010: 20: e195–e207 &2009 John Wiley & Sons A/S
doi: 10.1111/j.1600-0838.2009.00947.x
e195
been reported as well. In contrast to these studies,
ibuprofen and acetaminophen may actually enhance
muscle growth in response to resistance training in
elderly subjects (Carroll et al., 2008), and celecoxib (a
COX-2 inhibitor) can allegedly boost the increased rate
of protein synthesis after a bout of high-force resistance
exercise (Burd et al., 2008). Thus, it seems fair to
conclude that the jury is still out concerning the effects
of NSAIDs on human skeletal muscle – especially
regarding the COX-2 inhibitors (Warden, 2005).
‘‘The repeated-bout effect’’ comprises the physio-
logical adaptation that takes place after a single bout
of unaccustomed high-force exercise (Clarkson et al.,
1987). Basically, the exercised muscles become sig-
nificantly more resistant toward damage from a
repeated bout, preformed days or weeks after the
first one. However, the mechanism behind this adap-
tation is still not clear (McHugh, 2003). A local
inflammatory reaction in the exercised muscle may
be involved in the strengthening of the tissue after
unaccustomed exercise, because when perturbed, a
reduced repeated-bout effect has been reported (La-
pointe et al., 2002a). In the study by Lapointe et al.
(2002a), rats were exposed two times to a muscle-
damaging exercise protocol. An NSAID was given to
one group after the first exercise bout to reduce the
post-exercise inflammation. These rats showed
clearly a smaller repeated-bout effect than the con-
trols, with larger force deficits and larger accumula-
tion of leukocytes in the muscle tissue after the
second bout. This study design has previously not
been tested with human subjects, and it is suited to
clarify both early and late effects of NSAIDs, includ-
ing the mechanisms behind the repeated-bout effect.
Therefore, the aim of the present study was to
investigate whether administration of a COX-2 inhibi-
tor would (1) enhance the recovery after maximal
eccentric exercise, by reducing the inflammatory reac-
tion during the first days after exercise, (2) slow down
the recovery in the final stage (1–3 weeks after exercise)
and (3) reduce the repeated-bout effect.
Materials and method
Subjects
Thirty-three healthy students and employees at the Norwegian
School of Sport Sciences gave written, informed consent to
participate in the study (Table 1). The subjects were physically
active and involved in different activities, such as running and
cycling, and team sports, such as basketball and soccer. None
of the subjects were familiar with maximal eccentric exercise
with the elbow flexors. No exercise was performed for 3 days
before the experiment. Subjects had a light breakfast at home
and were offered a mixed meal shortly after exercise. Water
was available ad libitum during exercise. The subjects were
asked to continue their habitual diet and not to take any form
of medications or prescription-free supplements (such as
antioxidants) and avoid exercise, stretching and massage
therapy (etc.) during the experimental periods. The study
complied with the standards set by the Declaration of Helsinki
and was approved by the Regional Ethics Committee of
Southern Norway.
Study design
This investigation was carried out with a double-blinded
design. The subjects were randomized to a celecoxib group
and a placebo group. There were no group differences in
anthropometric values (Table 1). The celecoxib group was
administrated 400 mg [Celebra (Pfizer, Oslo, Norway); 200 mg
morning and evening] for 9 days, with the first dose approxi-
mately 45 min before the exercise (bout 1; Fig. 1). This dosage
is the highest recommended dosage in the Norwegian Phar-
maceutical Product Compendium (NPPC). Maximal blood
concentration of the drug is reached after 2–3 h and the half-
life of the drug is 8–12 h (NPPC). The placebo-group received
similar looking lactose pills. The subjects were contacted every
morning and evening to ascertain that the pills were taken.
Three weeks after bout 1, the exercise was repeated (bout 2),
but without drug administration (Fig. 1).
The exercise was performed with the same arm, randomly
chosen, on both occasions (bouts 1 and 2). The other arm
served as a non-exercised control. Before and for 9 days after
the exercise bouts, tests of muscle function (force-generation
capacity), muscle soreness (pain) and swelling, as well as resting
elbow angle were performed and blood was drawn, daily (Fig.
1). Assessments of muscle function were always performed after
blood sampling and the other measurements, except immedi-
ately after exercise when muscle function was tested first.
Radionuclide imaging was used to monitor the muscle
accumulation of leukocytes 6 and 20 h after bouts 1 and 2.
Microdialysis was performed 2, 24 and 48 h after bout 1 for
measurement of prostaglandins (PGE
2
). Biopsies from m.
biceps brachii were collected from both exercised and control
muscles 1, 48, 96 and 168 h after bout 1, and 1 and 48 h after
bout 2. The different time points after bout 1 were chosen in
order to evaluate early (1 and 48 h) and late (96 and 168 h)
inflammatory reactions and regenerating processes.
No subjects withdrew during the study. However, four
subjects did not provide all the scheduled biopsies due to
technical problems or the discomfort experienced. Six subjects
Table 1. Descriptive characteristics of the subjects
Group
N
Age (years) Height (m) Weight (kg) Celecoxib mg/kg
Males Celecoxib 8 28 5 1.81 0.03 78 7 5.1 0.5
Placebo 14 26 4 1.82 0.06 77 6
Females Celecoxib 7 28 5 1.65 0.03 59 6 6.9 0.7
Placebo 4 23 5 1.67 0.03 62 3
,1<Celecoxib 15 28 5 1.73 0.08 69 12 6.0 1.1
Placebo 18 25 4 1.79 0.08 74 8
Values are means SD.
Paulsen et al.
e196
(three in the placebo and three in the celecoxib group)
performed only the first of the two bouts of exercise, because
they were specially recruited for the microdialysis experiment.
The double-blind experimental design implied that the
subjects were unaware of which group – test or control –
they belonged to, and during all analyses the subjects’ group
affiliation was concealed for the recorders.
Medication side-effect registration
The subjects were asked to inform the test leader and write
down on a form any symptom they thought could be related to
the pills.
Unilateral arm exercise
For the eccentric exercise the subjects were positioned in a
chair (Technogym, REV 9000, Gambettola, Italy) and fas-
tened with belts over the hip, chest and shoulder, and the
upper arm was supported by a cushion. Thus, the shoulder
joint was kept in a slightly flexed position (30–351from the
vertical axis) and prevented from moving during the elbow
exercise. The subjects gripped a handle connected to the lever
arm of the dynamometer. Because the handle could be rotated
about the longitudinal axis, the subjects were instructed to
supinate their forearm (elbow joint) for maximal activation of
m. biceps brachii. The exercise protocol consisted of 14 5
repetitions of maximal voluntary, eccentric actions using the
elbow flexors, with 30–35 s rest in-between sets. The lever arm
of the dynamometer was automatically returned to the start-
ing point; no muscle force being necessary for the elbow
flexion phase. The range of motion (ROM) in the elbow joint
was 40–1751(18015full extension) and the velocity was 301/s.
The subjects were verbally motivated to resist maximally
through the whole ROM and they received real time visual
feedback on their performance on a computer screen.
The total work of each eccentric action was registered and
work per set and total work per exercise were calculated. All
values registered during exercise and tests were corrected for
gravity and passive tension during the full ROM.
Muscle function
In the same position as during the exercise, maximal force-
generating capacity was measured as peak torque during two
consecutive maximal, isokinetic, concentric elbow flexions at
601/s (ROM: 175–401) and as peak torque during isometric
actions at 901in the elbow joint (5 s actions; two attempts;
Technogym, REV 9000). All subjects participated in one or
two familiarization tests on separate days within 1 week before
they entered the study. Two pre-tests were performed on day 1
of the experiment (with 30–60 min rest in-between) and the
mean of these tests were used for further analysis. Subjects did
always warm up by 3 min arm cranking (30–50 W) and four
submaximal, concentric, isokinetic actions in the dynam-
ometer. The intra-individual coefficient of variation (CV) for
the force-generating capacity measurements were o5%.
Scintigraphic monitoring of leukocyte accumulation
The method has been described by Raastad et al. (2003). In
brief, 50 mL blood was drawn, and leukocytes (mainly neu-
trophilic granulocytes) were isolated and labelled with
99m
Technetium (
99m
Tc), before being re-infused. Accumula-
tion of
99m
Tc-leukocytes (radiolabelled leukocytes) in the
subjects’ arms was quantified scintigraphically (with a gamma
camera) on anterior view images (Fig. 4). The radioactivity in
the upper arm elbow flexors was calculated with a custom-
made software (from GE Healthcare, Oslo, Norway). The
accumulation of radioactivity within a region of interest (ROI)
in the exercised arm was related to radioactivity in the same
ROI in the control arm, corrected for background radiation.
Fig. 1. Overview of the study. The dots, x’s and # show the time points for tests and measurements. Note the drug
administration after bout 1 only.
COX-2 inhibition after eccentric exercise
e197
This procedure was performed on 22 subjects (celecoxib: n59,
placebo: n513).
Muscle biopsies
A 5 or 6 mm Pelomi-needle (Albertslund, Denmark) with
manual suction was used to obtain tissue samples (2–3 30–
100 mg) from the mid-section of m. biceps brachii. Subjects were
in supine position, and the procedure was performed under local
anesthesia (10 mg/mL 15mg/mL, Xylocain
s
adrenaline, Astra-
Zeneca, So
¨derta
¨lje, Sweden). Each needle incision was placed
approximately 1–2 cm medially and laterally to the first incision
and care was taken to avoid tissue affected by earlier biopsies.
The muscle samples were rinsed in physiological saline before
visible fat and connective tissue were removed, and subsequently
frozen in isopentane on dry ice and stored at 80 1C until
analysis.
Twenty-four subjects delivered biopsies from the m. biceps
brachii of both the exercised arm and the control arm. To
reduce stress on the subjects and to reduce risk of contamina-
tion from previous biopsies, each subject was scheduled for
biopsy at three out of the six biopsy time points: 1 h, and 2, 4
and 7 days after bout 1, and 1h and 2 days after bout 2 (Fig. 1).
Immunohistochemistry
Serial cross-sections (7 mm) were incubated with antibodies (ab)
against leukocytes: CD66b (neutrophilic granulocytes, monoclo-
nal ab, M1546, PeliCluster, Sanquin, Amsterdam, the Nether-
lands; 1:500), and CD68 (monocytes/macrophages; monoclonal
ab, M0718, DAKO, Copenhagen, Denmark; 1:300), together
with antibodies against laminin (polyclonal ab, Z0097, DAKO;
1:1000) or dystrophin (polyclonal ab, ab15277, Abcam, Cam-
bridge, UK; 1:2000). The two latter antibodies were used to
visualize the sarcolemma. In order to visualize satellite cells/
myoblasts, sections were analyzed for immunoreactivity against
CD56/NCAM (monoclonal ab, ab9018, Abcam; 1:200). Sections
with overt satellite cell/myoblast activity were double stained
with CD56 and Ki67 (a marker of proliferation; polyclonal ab,
CP249A, Biocare Medical, Concord, California, USA; 1:200).
Alexa-488 (FITC) and -594 (goat anti-rabbit or goat anti-mouse;
Invitrogen-Molecular Probes, Eugene, Oregon, USA) were used
as secondary antibodies. The sections were finally counterstained
with DAPI (for nuclear staining) and mounted under coverslips
(ProLong Gold Antifade Reagent with DAPI, P36935, Invitro-
gen-Molecular Probes).
Images of the stained cross-sections were captured using an
Axiocam camera (Zeiss, Oberkochen, Germany) mounted on a
Axioskop-2 light microscope (Zeiss). Multiple images (20,
40 and 100 objectives) were taken so that the whole muscle
biopsy cross-section was captured. To quantify the number of
cells positive for a leukocyte associated antigen, a cell that
contained both DAPI and antibody staining was considered as
positive, independent of the staining intensity. Data are pre-
sented as number of positive cells per 100 myofibers.
Microdialysis and prostagandin E
2
Tissue fluid (dialysate) from both the exercised arm and the
control arm of 12 subjects (six administrated the COX-2
inhibitor and six placebo) were collected at three time points
after bout 1. All twelve subjects underwent the microdialysis
procedure 2–6 h after exercise and then again starting at either
24 (n55) or 48 h (n55) after exercise. A CMA-60 microdialysis
probe (20kDa molecular cut-off; length 30mm; CMA/Micro-
dialysis AB, Solna, Sweden) was used for collection of tissue
fluid for later measurements of PGE
2
. Before the insertion of the
CMA-60 catheter, positioned 2 cm medial to the centerline of
the m. biceps brachii muscle, the skin was locally anesthetized
with 0.3mL of Lidocain (10mg/mL; SAD, Copenhagen, Den-
mark). The CMA-60 catheters were perfused with a Ringer
acetate solution containing 5 nM
3
H-PGE
2
(specific activity,
3.7 GBq/mmol; NEN, Boston, Massachusetts, USA) to deter-
mine the recovery of PGE
2
. The subjects rested for at least
90 min before starting the experiment to ensure that any reaction
from the insertion trauma had minimized. Dialysate samples
were collected every 30 min. The samples were immediately
frozen ( 80 1C) until analysis.
The PGE
2
concentration in the dialysate was analyzed by a
commercial competitive enzyme immunoassay (ACEt, catalog
no. 514010, Cayman Chemicals Inc., Ann Arbor, Michigan,
USA). Standards and samples were analyzed in accordance with
the protocol of the kit. Samples were analyzed in duplicates.
Optical density (OD) was measured in a plate-reader (ASYS
Hitech, Eugendorf, Austria) at 405 nm.
Muscle soreness
Muscle soreness was rated on a visual analogue scale where 0
represented ‘‘not sore at all’’ and 100 mm ‘‘extremely sore.’’ The
subjects stretched and contracted their elbow flexors (both arms)
to assess soreness in m. biceps brachii and m. brachialis. In order
to assess soreness/tenderness at the lower (distal) and upper
(proximal) part of the upper arm elbow flexors, the subjects sat
with their arms extended, but resting, on a bench (901flexed and
lateral-rotated shoulder joints). The upper arm elbow flexor
muscles were then palpated manually by the test leader. In
addition, a probe (3 cm
2
cross-sectional area), giving 10 N/cm
2
of
pressure, was applied perpendicularly to the upper arm elbow
flexor muscles at predetermined locations, i.e., a distal and a
proximal spot (marked by a water-proof pen). The middle part
of the upper arm elbow flexors was discarded because of the
location of the biopsy incisions.
Muscle swelling and resting arm angle
Swelling was assessed by measuring the arm circumference
with a tape measuring device, utilizing a spring mechanism to
apply constant force (Roche, Oslo, Norway). The circumfer-
ence was measured at the person’s thickest part of the lower
half of the m. biceps brachii (and the m. brachialis under-
neath), below the biopsy incisions. A waterproof pen was used
to mark the site for repeated measurements. Resting arm angle
was measured with a goniometer. The intra-individual CV of
these measurements was o1.8%.
Blood sampling and CK
Blood was drawn from an antecubital vein into a 10 mL serum
vacutainer tube. After coagulating for 30–45 min at room
temperature (20 1C), the blood was centrifuged at 2700 g
for 10 min at 4 1C. Serum was immediately pipetted into
Eppendorf tubes and stored at 80 1C until analysis. CK
was analyzed with an automated chemistry analyzer (Modular
P, Hitachi High- Technologies Corporation, Tokyo, Japan);
analytic CV being o5%.
Statistics
All parameters were measured repeatedly after exercise and
for each variable the area under the curve (AUC) was
calculated for each subject. The AUC values, in addition to
peak/nadir values, for the celecoxib and placebo group were
Paulsen et al.
e198
compared with the unpaired Student t-test or the Mann–
Whitney test. A multivariate analysis of variance (MANOVA)
with repeated measures and a two-way ANOVA (time
group and time arm) with Bonferroni post hoc test were
also applied. For the different variables, changes over time
within each group were followed with a one-way ANOVA or
Friedman’s test with repeated measures and Dunnett’s and
Tukey’s or Dunn’s post hoc tests. We evaluated the differences
from baseline and between selected time points for the
exercised arm and the non-exercised (control) arm separately,
as well as the difference between arms. Because of the fact that
different subjects were biopsied at each sampling time point,
the unpaired Student t-test and the Mann–Whitney test were
used to analyze group differences in histological data. The
paired Student t-test or the Wilcoxon signed rank test was
applied to evaluate differences between the exercised arm and
the control arm. Data are presented as means with standard
error of the mean (SEM), if not otherwise stated in the text.
P0.05 was considered statistically significant. Cohen’s
effect size (difference between group means divided by the
pooled SD) was calculated with bias correction for unequal
numbers of subjects in the two groups (Hedges). Tests for
normality (Gaussian distribution) dictated the choice of para-
metric or non-parametric tests. The statistics were performed
with Microsoft
s
Excel 2003 (including statistiXL 1.8) InStat
s
3.06, Prism
s
5.01 and Statematet2.0 (GraphPad Software
Inc., San Diego, California, USA) and SPSS 15.0 (SPSS Inc.,
Chicago, Illinois, USA).
Results
Exercise and muscle fatigue
The total work (sum of 14 5 repetitions) performed
by each experimental subject during bouts 1 and 2
was similar within and between groups: celecoxib:
4500 500J (bouts 1 and 2) and placebo:
4800 300J (bouts 1 and 2). Total work during
the last (14th) set was reduced by 45 4% and
47 4% of the first set values for the celecoxib and
placebo group, respectively, during bout 1 (Po0.01).
This reduction was smaller in bout 2 than bout 1 for
both groups (celecoxib: 34 3% and placebo:
35 3%; Po0.01). [Correction made to above para-
graph after initial online publication.]
Muscle function
The force-generating capacity measured as isometric
and isokinetic concentric torque was markedly re-
duced (40–50%) after both exercise bouts, but
there were no significant differences between groups
at any time point (Fig. 2). The force-generating
capacity was not recovered 9 days after bout 1, but
was in both groups not detectably different from
baseline values before bout 2, 3 weeks after bout 1.
Of note, about half of the subjects in each group were
45% (CV for the test) below their baseline values 3
weeks after bout 1. After bout 2 the acute reduction
in force-generating capacity was only slightly atte-
nuated compared with after bout 1 (calculating with
the pre-values of bout 1; Po0.05; Fig. 2). However,
the recovery of both isometric and concentric force-
generating capacity was faster after bout 2 than bout
1 (comparing AUC values; Po0.01). The force-
generating capacity of the non-exercised, control
arm did not change from baseline in either group
during the experimental periods (Fig. 2).
Muscle soreness
Muscle soreness of the upper arm, as assessed during
contractions and stretching, was less in the celecoxib
group than in the placebo group, after both bouts
(Fig. 3, AUC: bout 1: P50.04, and bout 2: P50.06;
effect size: 0.75 and 0.86 for bouts 1 and 2, respec-
tively). For both groups, less soreness was reported
after bout 2 than after bout 1 (Po0.01). The peak
soreness values after bout 1 tended to be higher in the
placebo group (6.7 0.6 mm) than in the celecoxib
group (5.1 0.7 mm; P50.08). This was less evident
after bout 2 [peak values: 2.5 0.4 mm (celecoxib) vs
–60
–50
–40
–30
–20
–10
0
10
Celecoxib
Placebo
Celecoxib control
Placebo control
Pre
0
5
24
48
72
96
168
216
Pre
0
5
24
48
72
96
168
216
3 weeks
#
#
##
##
#
##
#
#
#
#
#
#
#
#
#
#
#
##
#
#
#
#
#
#
Time (hours after exercise)
Changes in isometric torque (%)
AUC bout 1 > bout 2
–60
–50
–40
–30
–20
–10
0
10
Celecoxib
Placebo
Celecoxib control
Placebo control
Pre
0
5
24
48
72
96
168
216
Pre
0
5
24
48
72
96
168
216
3 weeks
#
#
##
##
#
##
#
#
#
#
#
#
#
#
#
#
#
###
#
#
#
#
#
Time (hours after exercise)
Changes in concentric torque (%)
AUC bout 1 > bout 2
(a)
(b)
Fig. 2. Changes in force-generating capacity of the elbow
flexors exposed to two bouts of maximal eccentric exercise
(same arm) and no exercise (control). (a) Changes in max-
imal voluntary isometric torque (901) and (b) changes in
maximal voluntary isokinetic concentric torque (601/s).
N-celecoxib 512; n-placebo 515. Error bars are SEM.
#
Difference from baseline values (before bout 1). The area
under the curve (AUC) was larger for bout 1 than 2 for both
groups (Po0.01).
COX-2 inhibition after eccentric exercise
e199
3.6 0.6 mm (placebo), P50.14]. The placebo
group tended to report more soreness in the control
arm than the celecoxib group after bout 1 (AUC;
P50.1). This soreness/pain seemed to be due to the
biopsy procedure.
For soreness as assessed with palpation, lower
peak values were reported by the celecoxib group
than the placebo group, but only in the proximal part
of elbow flexors (3.4 0.8 vs 6.5 0.7; Po0.05).
After bout 2, peak soreness values during palpation
were reduced in both groups (Po0.01).
Scintigraphy
The radioactivity, reflecting autologous radiolabelled
leukocytes, was in both groups higher in the exer-
cised arm than in the control arm 6 and 20 h after
bout 1, and the difference between arms was larger at
20 than 6 h after exercise (Po0.05; Fig. 4). After
bout 2 the radioactivity at 6 h was similar to that
found in bout 1; however, no further increase at 20 h
was detected after bout 2 (Fig. 4). There were no
significant differences between groups.
The scintigrams confirmed that m. biceps brachii
was affected by the exercise protocol. We could not
precisely differentiate between the elbow flexors (due
to low image resolution), but, in addition to m.
biceps brachii, both m. brachialis and m. brachior-
adialis appeared to be affected. There was a trend
toward higher radioactivity in the lower and middle
part than the proximal part of the upper arm (Fig. 4).
To exemplify, three subjects demonstrated 1000%
difference between middle part of the exercised
muscles and control muscles 20 h after exercise. In
five subjects, high radioactivity (4100% higher in
the exercised muscles compared with control) was
found over a localized area that appeared to corre-
spond to the proximal muscle–tendon junction.
Microdialysis (PGE
2
)
The PGE
2
concentration in the interstitial fluid (dia-
lysate) of the exercised muscle was not detectably
different from the control muscle at any time point
(2, 24 and 48 h after bout 1), and there were no group
differences. The mean values (combined over groups
and time points) were 880 134 and 900 135 pg/mL
for the exercised and control muscle, respectively.
However, the PGE
2
concentration tended to decrease
in both the exercised muscle and the control muscle
during the first days after exercise in the celecoxib
group (mean of both arms: 1000 167 vs 635 187;
P50.16), but not in the placebo group (mean of both
arms: 988 412 vs 1151 296). This suggests that
celecoxib generally lowered the PGE
2
levels in the
interstitial fluid over time.
Immunohistochemistry
Leukocytes
CD68
1
cells, primarily monocytes/macrophages,
were observed in the samples from both exercised
and control muscles, but higher numbers were
counted in the exercised muscle after both bouts
(74 21 vs 22 3 CD68
1
cells per 100 myofiber;
data combined over groups and time; Po0.01). In
the exercised muscles the number of CD68
1
cells
increased from 1 to 48 h after exercise and from early
(1 and 48 h) to later time points (4 and 7 days) after
exercise (Po0.05). No differences between groups
were observed (Table 2), however, at later time
points the five highest values registered were all
found in samples from subjects in the placebo group.
CD68
1
cells seemed preferentially related to necrotic
and regenerated myofibers (Fig. 5). Necrotic myofi-
bers (probably segments) were identified as dystro-
phin negative fibers (no staining); eight of 23 subjects
displayed 2% dystrophin negative myofibers on
samples from the exercised muscle: 15% 19 [stan-
dard deviation (SD)] of the analyzed fibers. This was
only observed in samples obtained 4 and 7 days after
exercise. Dystrophin negative fibers were only seen in
four of all samples from control. Staining for laminin
indicated that the basal membrane (with few excep-
tions) stayed intact – despite loss of dystrophin
staining, and signs of the regeneration process
occurred within the basal lamina sheath.
CD66b
1
cells (neutrophils) were found in very low
numbers and no significant differences between ex-
AUC bout 1 > bout 2
AUC placebo > celecoxib (bout 1 and 2)
Pre
0
5
24
48
72
96
168
216
0
10
20
30
40
50
60
70
80
90
100 Celecoxib
Placebo
Celecoxib control
Placebo control
3 weeks
#
#
#
#
#
#
#
#
#
##
#
#
#
Pre
0
5
24
48
72
96
168
216
Time (hours after exercise)
Muscle soreness (mm)
Fig. 3. Muscle soreness after two bouts of eccentric exercise,
evaluated by visual analogue scale; 0–100 mm. N-cele-
coxib 512; n-placebo 515. Error bars are SEM.
#
Difference
from baseline. The area under the curve (AUC) was larger
for bout 1 than 2 for both groups (Po0.01). The AUC was
smaller for the celecoxib group than for the placebo group
for both bouts (bout 1: P50.04; bout 2: P50.06).
Paulsen et al.
e200
ercised and control muscles, or between groups,
could be detected (Table 2).
Satellite cells
No differences between the groups were detected in the
number of CD56
1
cells, i.e., satellite cells (located
underneath basal lamina, but outside the plasma
membrane) and myoblasts (CD56
1
cells primarily
inside necrotic myofibres; Table 2). Myoblasts were
observed only in some subjects and preferentially
clustered together (see next paragraph). For the groups
combined, the number of satellite cells/myoblasts was
higher in exercised than in control muscles at both
early (1 and 48 h) and late time points (4 and 7 days
after exercise; Po0.05), but there was an increase from
early to late time points after bout 1 (P50.01). The
number of satellite cells/myoblasts per myofiber was
0.11 0.04 (SD) in samples from the non-exercised,
control arm and 0.19 0.18 (SD) in the exercise
samples (all combined). Expressed as the proportion
of the nuclei underneath basal lamina, 4.8% 1.5
(SD) were defined as satellite cells in control samples
and 5.7% 1.5 (SD) in the exercise samples. Hence,
these latter numbers do not include myoblasts.
Intense satellite cell activity, including increased
number and cell volume, long cytoplasmatic exten-
sions, cell division activity (Ki67 positive nuclei) and
fusion of myoblasts were seen in five subjects 1 week
after exercise (Fig. 6). CD68
1
cells occasionally were
Table 2. Quantitative immunohistochemical staining: number of neutrophilic granulocytes (CD66b), monocytes/macrophages (CD68) and satellite cells/
myoblasts (CD56) in muscle cross-sections from the exercised muscle
Positive cells per
100 myofiber
Bout # Time point Celecoxib Placebo
P
-value celecoxib
vs placebo
CD66b Bout 1 Early 0 (0–3) 0 (0–29)
P
50.8
Late 0 (0–2) 0 (0–3)
P
50.9
Bout 2 Early 0 (0–2) 0 (0–3)
P
50.6
CD68 Bout 1 Early 21 (6–81) 25 (10–133)
P
50.9
Late 35 (14–150) 45 (7–968)
P
50.7
Bout 2 Early 27 (2–94) 47 (15–104)
P
50.2
CD56 Bout 1 Early 11 (7–18) 12 (8–23)
P
50.6
Late 16 (10–50) 20 (7–103)
P
50.7
Bout 2 Early 13 (11–25) 14 (10–49)
P
50.5
Values are medians and full ranges.
N
58–14 in each group per time point. The
P
-values stem from group comparisons.
Infusion of
99mTc-leukocytes
Elbow joint;
radioactive marker
Shoulder joint
M. biceps brachii/
brachialis
Exercised arm Control arm
Blood sampling
6 hours 20 hours
0
25
50
75
100
125
150 Celecoxib; bout 1
Placebo; bout 1
Celecoxib; bout 2
Placebo; bout 2
*
*
Time
Radioactivity
(% difference from control)
Fig. 4. Scintigram illustrated. The
image shows accumulation of radi-
olabelled autologous leukocytes in
the exercised arm muscles, most
prominent in the lower/middle part
of the upper arm elbow flexors (dark
areas reflect high radioactivity). The
figure shows the relative differences
in radioactivity between the exer-
cised arm and the non-exercised,
control arm 6 and 20 h after exer-
cise. N-celecoxib 59; n-placebo 5
13. Error bars are SEM.
*
denotes
difference between the 6 and 20 h
values. [Figure has been revised after
initial online publication]
COX-2 inhibition after eccentric exercise
e201
(a) (b)
Control
1 week 3 weeks
Fig. 5. Immunohistochemical experiments illustrated. The images are from a subject in the placebo group (but similar pictures
were obtained from those in the celecoxib group). Left image (a): CD68 staining of a sample obtained 168 h (1 week) after bout
1 (CD68 5red, laminin 5green, nuclei 5blue). The arrow points to a myofiber that is heavily infiltrated by CD68
1
cells
(primarily monocytes/macrophages). Note that this myofiber was negative for dystrophin (not shown) and probably in a stage
of necrosis. The presence of rounded and apparently swollen myofibers suggests that the myofibers were highly stressed in this
sample. Right image (b): CD68 staining on a sample obtained 3 weeks after bout 1 (and 1h after bout 2). Note the small
myofibers with centrally located nuclei (arrow), indicating regenerated myofibers, surrounded by CD68
1
macrophages. These
myofibers were also positive for CD56 (see Fig. 6). Note that the CD68
1
cells were most likely remnants from the damage
reaction to bout 1 (not bout 2). Thus, the present CD68
1
cells are probably macrophages/histiocytes that served regenerating
myofibers. There was little or no specific staining in the control samples (lower right corner). Scale bars 550 mm.
Fig. 6. Immunohistochemical experiments illustrated (CD56 5green, Ki67 5red, nuclei 5blue). Left column of images (a, c):
tissue samples from a subject in the celecoxib group. Right column of images (b, d): tissue samples from a subject in the placebo
group. The top row of images (a, b) shows high satellite cell/myoblast activity and Ki67
1
cells 1 week after bout 1. The inset
image on image a shows accumulation and apparent fusion of myoblasts (CD56
1
), as well as cells that seem to go through cell
division (pink stain 5Ki67 [red]1dapi [blue]). The inset images on image b show (i) a normal looking satellite cell, i.e. with
CD56 staining around the nucleus, which could be found in samples from both control and exercised muscles, and (ii) a satellite
cell positive for Ki67 (sample from a exercised muscle). The second row of images (c, d) shows tissue samples obtained 48 h
after bout 2, i.e., 3 weeks after bout 1. Note the CD56 positive myofibers with relatively small diameter and centrally located
nuclei (arrows), which indicate that these are regenerated myofibers. On image c there is a Ki67
1
nucleus apparently inside the
myofiber (arrowhead). Hence, this seems to be part of the regenerating process after bout 1 (not bout 2). Scale bars 550 mm.
Paulsen et al.
e202
positive for Ki67 as well. Thus, in areas with large
number of cells (nuclei), differentiation of the cells
could be very difficult. No detectable differences
between the celecoxib and placebo subjects could
be found.
In biopsies obtained after bout 2, relatively high
numbers of CD56
1
myofibers with centrally located
nuclei and often small diameters were observed
(10% of the myofibers in six subjects; three from
each group; Fig. 6). A great majority of these CD56
1
myofibers were also verified to contain embryonic
myosin heavy chain (F1.652, Hybridoma Bank; data
not shown). Although this was observed after bout 2,
the regeneration processes were very likely to be
related to bout 1. This implies that complete regen-
eration after severe exercise-induced muscle damage
takes more than 3 weeks. There were no signs of failed
regeneration in biopsies after bout 2, in either group.
Muscle swelling and resting arm angle
The upper arm circumference increased equally be-
tween groups and peaked 4 days after bout 1,
4.2 0.9% vs 4.6 1.0% for the celecoxib and pla-
cebo group, respectively (Po0.01 from baseline). The
resting arm angle decreased (18015extended arm)
equally between groups by 10 21and 11 2124 h
after bout 1 for the celecoxib and placebo group,
respectively (Po0.01). Both the circumference and the
resting arm angle changed less after bout 2 than bout
1(Po0.01), and no group difference were detected.
CK
The serum CK levels were markedly elevated after
bout 1, but not after bout 2 (AUC; Po0.01; Fig. 7).
There were no differences between groups.
Of note, for about half of the subjects the increase
in serum CK levels followed a biphasic time course.
Thus, there was an early increase during the first 8 h,
followed by a more or less steady phase from 8 to
24 h, before the values increased markedly until peak
values were reached typically at 96 h (4 days) after
exercise (Fig. 7).
Gender
There were no gender differences in relative changes of
muscle function, in muscle soreness or in the histolo-
gical data, neither within groups nor across groups.
However, the serum CK pre-values (before bout 1)
were lower in the females than in the males (P50.05),
and the females had a lower exercise-response to bout
1 (AUC; Po0.05). The three highest values for
accumulation of radiolabelled leukocytes (scintigra-
phy data) were observed in female subjects, but we
found no significant differences between the sexes.
Exclusion of the females from these analyses did not
change the group comparisons appreciably.
Adverse effect of the pill administration
Three subjects mentioned slight nausea, but one of
these subjects was in the placebo group. One subject
from the celecoxib group experienced an unexpected
swelling of the forearm of the exercised arm; the excess
fluid appeared to be in the subcutaneous tissue.
Discussion
The main finding in this study was that celecoxib, a
selective COX-2 inhibitor, did not affect recovery of
muscle function after a bout of unaccustomed, mus-
cle-damaging, eccentric exercise in human subjects.
The COX-2 inhibitor did, however, reduce the de-
layed onset muscle soreness (DOMS).
A repeated-bout effect was clearly demonstrated
by the faster recovery of muscle function and mark-
edly blunted DOMS and serum CK response after
the second bout of exercise. Nine days of celecoxib
administration during the recovery period after bout
1 did not have any detectably adverse effect on the
adaptation processes and recovery after the second
bout, 3 weeks after the first. The exercise protocol
clearly initiated an inflammatory reaction with accu-
mulation of leukocytes in the exercised muscles
(markedly in some ‘‘high-responder subjects’’), and
stimulated increased satellite cell activity as well.
There were no statistically significant differences
between the celecoxib and placebo group concerning
indicators of inflammation. Nevertheless, the highest
numbers of accumulated leukocytes at late time
points (4 and 7 days after bout 1) were found in
the placebo group.
Pre 18
20
24
48
72
96
168
64
128
256
512
1024
2048
4096
8192 Celecoxib
Placebo
Pre
1
8
20
24
48
72
96
168
3 weeks
###
#
#
#
#
#
###
##
Time (hours after exercise)
CK (IU
⋅
L
–1
[log 2])
AUC bout 1 > bout 2
Fig. 7. Creatine kinase in serum. N-celecoxib 512; n-
placebo 515. Error bars are SEM.
#
Difference from baseline
values. The area under the curve (AUC) was larger for bout
1 than 2 for both groups (Po0.01).
COX-2 inhibition after eccentric exercise
e203
This study is to our knowledge, the first to in-
vestigate the effect of a COX-2 inhibitor on recovery
after exercise-induced muscle damage in humans.
Furthermore, the necrosis and regeneration pro-
cesses observed in this study are rarely reported for
humans; especially in combination with measure-
ments of muscle function.
NSAIDs, changes in muscle function and inflammation
Administration of non-selective NSAIDs in the re-
covery phase after high-force, muscle-damaging ex-
ercise in human studies has resulted in equivocal
findings (Baldwin, 2003; Cheung et al., 2003; Con-
nolly et al., 2003). Some researchers have observed
enhanced recovery of muscle function during the first
days after exercise [e.g. (Dudley et al., 1997)], while
others observed no such effect in comparison with
placebo-medication [e.g. (Howell et al., 1998a)]. The
most obvious mechanism behind any effect of
NSAIDs on muscle restitution, on which our work-
ing hypothesis was founded, is reduced inflammation
(Lapointe et al., 2002a, b; Shen et al., 2005).
Accumulation of radiolabelled leukocytes in the
exercised muscles 6 and 20 h after exercise indicates
an early inflammatory reaction, but no effect of
celecoxib was detected. The blood-borne leukocytes
(primarily neutrophils) detected by scintigraphy were
probably either infiltrating the interstitial space of
the muscle tissue or merely adhering to the luminal
side of the local micro-vessels. Neutrophils were,
however, histologically not found to have accumu-
lated in the muscle tissue 2 days after exercise.
Conversely, the numbers of monocytes and/or
macrophages increased in the muscle tissue, and the
highest values were found 4 and 7 days after exercise,
in both the celecoxib and placebo group. Again, there
were no significant differences between groups, but
there was a tendency toward higher leukocyte numbers
in ‘‘high-responder’’ subjects from the placebo group.
This could mean that celecoxib had some dampening
effect when the inflammatory reaction exceeded a
certain intensity. The effect size between groups was
0.6, but our statistical power to detect such a difference
was very low (26%).
Experimental evidence indicates that high-force,
eccentric exercise protocols can cause significant
muscle damage and inflammation in humans (Jones
et al., 1986; Newham et al., 1987; Jones et al., 1989;
Hellsten et al., 1997; Child et al., 1999). However,
scant evidence of a ‘‘classical’’ local inflammatory
reaction after exercise-induced muscle damage exists
[in contrast to findings in rats and mice; (Schneider &
Tiidus 2007)]. We found that the most severe muscle
weakness occurred well before the cell damage and
inflammation had been fully manifested. In line
with this, other workers have also found that peak
accumulation of leukocytes (primarily macrophages)
among the myofibers seemed to occur several days
after exercise (Jones et al., 1986; Hellsten et al., 1997;
Child et al., 1999). Bourgeois et al., (1999) reported
enhanced recovery of muscle function after the use of
an NSAID (naproxen), but no inflammatory cells
were seen histologically in muscle samples ob-
tained 24 h after resistance exercise. Peterson et al.,
(2003) found accumulation of monocytes/macro-
phages (CD68), but not neutrophils (CD15), 24 h
after eccentric exercise, and no effect of acetamino-
phen and ibuprofen was reported. The NSAIDs
(acetaminophen and ibuprofen) blunted the exer-
cise-induced increase of PGF
2a
, but there were no
increase of the PGE
2
levels (Trappe et al., 2001). We
could not detect any increase of PGE
2
in the ex-
ercised muscles. Therefore, the lack of an NSAID
effect – as in the present study – could be due to the
absence of a sufficiently strong and early inflamma-
tory reaction, as reflected by the histological findings,
and the lack of a significant prostaglandin response.
NSAIDs and DOMS
The celecoxib group reported less muscle soreness
and pain than the control group after both bouts of
exercise. The DOMS reducing effect of celecoxib
after bout 1 may be due to the drug’s known
analgesic effect (Ekman et al., 2002; Reuben & Ek-
man, 2005), in the muscle tissue, in the central
nervous system or both (Veiga et al., 2004).
We obtained circumstantial evidence for a drug
effect on the PGE
2
generation in the muscle, but no
evidence for an exercise effect on PGE
2
. The role of
local PGE
2
in DOMS has been questioned by others,
who like us analyzed microdialysis fluid (Tegeder et
al., 2002), and by investigators who found no
NSAID effect on DOMS [e.g. (Kuipers et al.,
1985)]. Our findings are also supported by previous
reports (Paulsen et al., 2009), indicating that accu-
mulation of leukocytes (as assessed by use of radi-
olabelled leukocytes and immunohistochemistry) is
both spatially and temporally out of phase with
development of muscle soreness. To exemplify, sub-
jects with the largest accumulation of radiolabelled
leukocytes did not report more intense soreness than
others. Furthermore, large accumulations of leuko-
cytes among the myofibers were seen 1 week after
exercise, but at that time the soreness was almost
gone. Thus, our data (PGE
2
and accumulation of
leukocytes) indicate that inflammation, in the classi-
cal sense, is not a main mechanism behind DOMS.
NSAIDs and satellite cells
A remarkable finding was myofiber necrosis (degra-
dation) and regeneration observed in as much as
Paulsen et al.
e204
1/3 of our subjects. The regeneration process was
heralded by cell proliferation leading to increased
numbers of satellite cells/myoblasts (CD56-positive
cells), as well as their migration, fusion and develop-
ment of new myotubes to replace degraded segments
of damaged myofibers. This has rarely been reported
in otherwise healthy human muscles after exercise.
Indeed, degenerative and regenerative processes have
been captured by histological examination after
strenuous exercise (Hikida et al., 1983; Jones et al.,
1986; Round et al., 1987; Child et al., 1999), but
satellite cell activity in highly damage muscle
tissue assessed by immunohistochemistry has been
poorly documented. Crameri et al. (2004) applied
similar methodology as us, but found necrosis and
myoblast fusion in only one out of eight subjects
examined.
Possible retardation of regeneration processes in
skeletal muscle tissue by NSAIDs, and especially
COX-2 inhibitors (Bondesen et al., 2004; Warden,
2005; Shen et al., 2006), has not been thoroughly
investigated in humans. However, administration of
indomethacin reduced satellite cell response after long
distance running in endurance-trained subjects
(Mackey et al., 2007). The discrepancy between our
study, where the satellite cell response in the affected
muscles was not detectably reduced by celecoxib, and
that of Mackey et al. (2007) is not easily explained.
However, the contradictory findings could be due to
the large differences in the exercise protocols and
subject populations, as well as the fact that different
drugs were administered – a non-selective COX-
inhibitor vs our selective COX-2 inhibitor. Hence,
any effect on satellite cell proliferation could be due to
inhibition of both COX-1 and COX-2, which, inter
alia, might markedly lower the tissue PGE
2
levels (in
contrast to our findings).
Our data suggest that the COX-2 inhibitor cele-
coxib (at high dosage: 6 mg/kg for 9 days) leaves
regeneration and adaptation processes unaffected
after exercise-induced muscle damage. However, the
large inter-individual differences and the few subjects
with considerable necrosis in the obtained biopsies,
made any celecoxib effects of a more subtle nature
elusive. Thus, further investigations are warranted,
and it should be kept in mind that celecoxib may also
have diverse COX-2 independent effects on a cell [e.g.
(Glebov et al., 2006)].
NSAIDs, inflammation and the repeated-bout effect
The repeated-bout effect is well documented, and as
expected, faster recovery of muscle function, as well
as blunted muscle soreness and serum CK increases,
were observed after the repeated exercise bout. The
mechanisms behind the repeated-bout effect are,
however, still ambiguous (McHugh, 2003), but the
inflammatory reaction might be causative. In an
experiment with rats, Lapointe et al. (2002a) demon-
strated that the repeated-bout effect was reduced
when the accumulation of leukocytes had been
blunted by the NSAID diclofenac in the recovery
phase after the initial bout (Lapointe et al., 2002a).
The study design of the present investigation re-
sembled Lapointe et al.’s (2002a). However, we could
not detect any clear anti-inflammatory effect of our
COX-2 inhibitor (celecoxib). Moreover, the re-
peated-bout effect was seen both in subjects with
large and low accumulation of leukocytes. Conse-
quently, we question the importance of inflammation
to the adaptation process after eccentric exercise in
humans. In that sense, our findings are in agreement
with other human studies where signs of myofibrillar
remodelling have been observed, without leukocytes
detectably present during the first days after eccentric
exercise (Feasson et al., 2002; Yu et al., 2002).
Conclusion
This is the first study to investigate the effect of a
COX-2 inhibitor on inflammation, recovery, regen-
eration and adaptation after exercised-induced mus-
cle damage in humans. We found no effect of the
COX-2 inhibitor on recovery of muscle function after
damaging elbow flexor eccentric exercise. However,
the drug reduced DOMS. High doses of celecoxib
during 9 days after a bout of eccentric exercise had
no detectable influence on the repeated-bout effect
observed 3 weeks after the first bout. An extensive
inflammatory response to unaccustomed eccentric
exercise occurred in some subjects, several days after
the exercise. The accumulation of leukocytes was
apparently unaffected by celecoxib administration
and seemed to be related to segments of necrotic
and later regenerating myofibers in the exercised
muscle. Thereby, our study provided insight into
the basal regeneration processes in healthy human
skeletal muscles.
Perspectives
Celecoxib intake (400 mg/day) over 9 days neither
enhanced nor inhibited recovery, regeneration and
adaptation processes after exercise-induced damage
in skeletal muscles. The drug alleviated muscle sore-
ness, but today it is unclear whether the drug works
primarily on peripheral factors or in the central
nervous system.
Of clinical relevance, short term use of celecoxib to
reduce exercise-induced muscle damage cannot be
advocated. However, celecoxib might be a good
choice when the intention is to reduce pain and
promote rehabilitation and recovery after injuries
COX-2 inhibition after eccentric exercise
e205
such as ankle sprains (Ekman et al., 2002; Nadarajah
et al., 2006), without detectable adverse effects on
the musculature. The degree of damage from the
exercise protocol applied in this study is not as
extensive as after strain injuries (including myofiber
rupture and intramuscular bleeding). Consequently,
our conclusions should not without caution be ex-
trapolated to other kinds of muscle injuries. It should
also be noted that celecoxib, and other first genera-
tion COX-2 inhibitors, have been associated with
increased risk of cardiovascular events, such as heart
attacks (Dajani & Islam, 2008). However, equivocal
findings have been made, and the short term use (few
weeks) of 400 mg/day for healthy people, not
suffering from cardiovascular disease, seems to be
reasonably safe (Kearney et al., 2006; Solomon et al.,
2008).
Key words: muscle damage, non-steroidal anti-inflam-
matory drugs, leukocytes, satellite cells.
Acknowledgement
The authors would like to thank Prof. Lars Morkrid and the
Department of Medical Biochemistry at Rikshospitalet
University Hospital (Oslo, Norway) for performing the serum
CK analyses. This study was partly financed by Pfizer Inc.
(Norway). However, the authors have no economical interests
in Pfizer, nor are there any contract-bond restrictions or
clauses that potentially could have influenced the study.
References
Almekinders LC. Anti-inflammatory
treatment of muscular injuries in sport.
An update of recent studies. Sports
Med 1999: 28: 383–388.
Baldwin LA. Use of nonsteroidal anti-
inflammatory drugs following exercise-
induced muscle injury. Sports Med
2003: 33: 177–185.
Bondesen BA, Mills ST, Kegley KM,
Pavlath GK. The COX-2 pathway is
essential during early stages of skeletal
muscle regeneration. Am J Physiol Cell
Physiol 2004: 287: C475–C483.
Bourgeois J, MacDougall D, MacDonald J,
Tarnopolsky M. Naproxen does not alter
indices of muscle damage in resistance-
exercise trained men. Med Sci Sports
Exerc 1999: 31: 4–9.
Burd NA, Dickinson JM, LeMoine JK,
Carroll CC, Haus JM, Hughes G,
Sanders C Jr., Trappe TA.
Consumption of a COX-2 inhibitor
stimulates muscle protein synthesis
after resistance exercise in humans. The
FASEB J 2008: 22: 958.
Buvanendran A, Reuben SS. COX-2
inhibitors in sports medicine: utility
and controversy. Br J Sports Med 2006:
40: 895–896.
Carroll CC, Dickinson JM, LeMoine JK,
Haus JM, Weinheimer EM, Hollon CJ,
Trappe TA. Ibuprofen and
acetaminophen promote muscle
hypertrophy and strength gains during
resistance exercise in the elderly. The
FASEB J 2008: 22: 753.
Cheung K, Hume P, Maxwell L. Delayed
onset muscle soreness: treatment
strategies and performance
factors. Sports Med 2003: 33:
145–164.
Child R, Brown S, Day S, Donnelly A,
Roper H, Saxton J. Changes in indices
of antioxidant status, lipid
peroxidation and inflammation in
human skeletal muscle after eccentric
muscle actions. Clin Sci (London)
1999: 96: 105–115.
Clarkson PM, Byrnes WC, Gillisson E,
Harper E. Adaptation to exercise-
induced muscle damage. Clin Sci
(London) 1987: 73: 383–386.
Connolly DA, Sayers SP, McHugh MP.
Treatment and prevention of delayed
onset muscle soreness. J Strength Cond
Res 2003: 17: 197–208.
Crameri RM, Langberg H, Magnusson P,
Jensen CH, Schroder HD, Olesen JL,
Suetta C, Teisner B, Kjaer M. Changes
in satellite cells in human skeletal
muscle after a single bout of high
intensity exercise. J Physiol 2004: 558:
333–340.
Dajani EZ, Islam K. Cardiovascular and
gastrointestinal toxicity of selective
cyclo-oxygenase-2 inhibitors in man. J
Physiol Pharmacol 2008: 59(Suppl. 2):
117–133.
Dudley GA, Czerkawski J, Meinrod A,
Gillis G, Baldwin A, Scarpone M.
Efficacy of naproxen sodium for
exercise-induced dysfunction muscle
injury and soreness. Clin J Sport Med
1997: 7: 3–10.
Ekman EF, Fiechtner JJ, Levy S, Fort
JG. Efficacy of celecoxib versus
ibuprofen in the treatment of acute
pain: a multicenter, double-blind,
randomized controlled trial in acute
ankle sprain. Am J Orthop 2002: 31:
445–451.
Feasson L, Stockholm D, Freyssenet D,
Richard I, Duguez S, Beckmann JS,
Denis C. Molecular adaptations of
neuromuscular disease-associated
proteins in response to eccentric
exercise in human skeletal muscle. J
Physiol 2002: 543: 297–306.
Glebov OK, Rodriguez LM, Lynch P,
Patterson S, Lynch H, Nakahara K,
Jenkins J, Cliatt J, Humbyrd CJ,
Denobile J, Soballe P, Gallinger S,
Buchbinder A, Gordon G, Hawk E,
Kirsch IR. Celecoxib treatment alters
the gene expression profile of normal
colonic mucosa. Cancer Epidemiol
Biomarkers Prev 2006: 15: 1382–1391.
Hellsten Y, Frandsen U, Orthenblad N,
Sjodin B, Richter EA. Xanthine
oxidase in human skeletal muscle
following eccentric exercise: a role in
inflammation. J Physiol 1997: 498(Part
1): 239–248.
Hikida RS, Staron RS, Hagerman FC,
Sherman WM, Costill DL. Muscle fiber
necrosis associated with human
marathon runners. J Neurol Sci 1983:
59: 185–203.
Howell JN, Conatser RR, Chleboun GS,
Karapondo DL, Chila AG. The effect
of nonsteroidal anti-inflammatory
drugs on recovery from exercise-
induced muscle injury 1. Flurbiprofen.
J Muscoskel Pain 1998a: 6: 59–68.
Howell JN, Conatser RR, Chleboun GS,
Karapondo DL, Chila AG. The effect
of nonsteroidal anti-inflammatory
drugs on recovery from exercise-
induced muscle injury 2. Ibuprofen. J
Muscoskel Pain 1998b: 6: 69–83.
Jones DA, Newham DJ, Round JM,
Tolfree SE. Experimental human
muscle damage: morphological
changes in relation to other indices of
damage. J Physiol 1986: 375: 435–448.
Jones DA, Newham DJ, Torgan C.
Mechanical influences on long-lasting
human muscle fatigue and delayed-
onset pain. J Physiol 1989: 412:
415–427.
Kearney PM, Baigent C, Godwin J, Halls
H, Emberson JR, Patrono C. Do
selective cyclo-oxygenase-2 inhibitors
and traditional non-steroidal anti-
inflammatory drugs increase the risk of
atherothrombosis? Meta-analysis of
randomised trials. BMJ 2006: 332:
1302–1308.
Paulsen et al.
e206
Kuipers H, Keizer HA, Verstappen FT,
Costill DL. Influence of a
prostaglandin-inhibiting drug on
muscle soreness after eccentric
work. Int J Sports Med 1985: 6:
336–339.
Lapointe BM, Fremont P, Cote CH.
Adaptation to lengthening contractions
is independent of voluntary muscle
recruitment but relies on inflammation.
Am J Physiol Regul Integr Comp
Physiol 2002a: 282: R323–R329.
Lapointe BM, Frenette J, Cote CH.
Lengthening contraction-induced
inflammation is linked to secondary
damage but devoid of neutrophil
invasion. J Appl Physiol 2002b: 92:
1995–2004.
Mackey AL, Kjaer M, Dandanell S,
Mikkelsen KH, Holm L, Dossing S,
Kadi F, Koskinen SO, Jensen CH,
Schroder HD, Langberg H. The
influence of anti-inflammatory
medication on exercise-induced
myogenic precursor cell responses in
humans. J Appl Physiol 2007: 103:
425–431.
McHugh MP. Recent advances in the
understanding of the repeated bout
effect: the protective effect against
muscle damage from a single bout of
eccentric exercise. Scand J Med Sci
Sports 2003: 13: 88–97.
Mehallo CJ, Drezner JA, Bytomski JR.
Practical management: nonsteroidal
antiinflammatory drug (NSAID) use in
athletic injuries. Clin J Sport Med
2006: 16: 170–174.
Miller SB. Prostaglandins in health and
disease: an overview. Semin Arthritis
Rheum 2006: 36: 37–49.
Nadarajah A, Abrahan L, Lau FL,
Hwang LJ, Fakir-Bolte C. Efficacy and
tolerability of celecoxib compared with
diclofenac slow release in the treatment
of acute ankle sprain in an Asian
population. Singapore Med J 2006: 47:
534–542.
Newham DJ, Jones DA, Clarkson PM.
Repeated high-force eccentric
exercise: effects on muscle pain and
damage. J Appl Physiol 1987: 63:
1381–1386.
Paoloni JA, Orchard JW. The use of
therapeutic medications for soft-tissue
injuries in sports medicine. Med J Aust
2005: 183: 384–388.
Paulsen G, Crameri R, Benestad HB,
Fjeld JG, Mørkrid L, Halle
´nJ,
Raastad T. Time course of leukocyte
accumulation in human muscle after
eccentric exercise. Med Sci Sports
Exerc 2009 (in press).
Peterson JM, Trappe TA, Mylona E,
White F, Lambert CP, Evans WJ, Pizza
FX. Ibuprofen and acetaminophen:
effect on muscle inflammation after
eccentric exercise. Med Sci Sports
Exerc 2003: 35: 892–896.
Raastad T, Risoy BA, Benestad HB,
Fjeld JG, Hallen J. Temporal relation
between leukocyte accumulation in
muscles and halted recovery 10–20 h
after strength exercise. J Appl Physiol
2003: 95: 2503–2509.
Reuben SS, Ekman EF. The effect of
cyclooxygenase-2 inhibition on
analgesia and spinal fusion. J Bone
Joint Surg Am 2005: 87: 536–542.
Reuben SS, Ekman EF. The effect of
initiating a preventive multimodal
analgesic regimen on long-term patient
outcomes for outpatient anterior
cruciate ligament reconstruction
surgery. Anesth Analg 2007: 105:
228–232.
Round JM, Jones DA, Cambridge G.
Cellular infiltrates in human skeletal
muscle: exercise induced damage as a
model for inflammatory muscle
disease? J Neurol Sci 1987: 82: 1–11.
Sayers SP, Knight CA, Clarkson PM,
Van Wegen EH, Kamen G. Effect of
ketoprofen on muscle function and
sEMG activity after eccentric exercise.
Med Sci Sports Exerc 2001: 33:
702–710.
Schneider BS, Tiidus PM. Neutrophil
infiltration in exercise-injured skeletal
muscle: how do we resolve the
controversy? Sports Med 2007: 37:
837–856.
Shen W, Li Y, Tang Y, Cummins J,
Huard J. NS-398, a cyclooxygenase-
2-specific inhibitor, delays skeletal
muscle healing by decreasing
regeneration and promoting fibrosis.
Am J Pathol 2005: 167: 1105–1117.
Shen W, Prisk V, Li Y, Foster W, Huard
J. Inhibited skeletal muscle healing in
cyclooxygenase-2 gene-deficient mice:
the role of PGE2 and PGF2alpha. J
Appl Physiol 2006: 101: 1215–1221.
Solomon SD, Wittes J, Finn PV, Fowler
R, Viner J, Bertagnolli MM, Arber N,
Levin B, Meinert CL, Martin B, Pater
JL, Goss PE, Lance P, Obara S, Chew
EY, Kim J, Arndt G, Hawk E.
Cardiovascular risk of celecoxib in 6
randomized placebo-controlled trials:
the cross trial safety analysis.
Circulation 2008: 117: 2104–2113.
Tegeder L, Zimmermann J, Meller ST,
Geisslinger G. Release of algesic
substances in human experimental
muscle pain. Inflamm Res 2002: 51:
393–402.
Trappe TA, Fluckey JD, White F,
Lambert CP, Evans WJ. Skeletal
muscle PGF(2)(alpha) and PGE(2) in
response to eccentric resistance
exercise: influence of ibuprofen
acetaminophen. J Clin Endocrinol
Metab 2001: 86: 5067–5070.
Trappe TA, White F, Lambert CP, Cesar
D, Hellerstein M, Evans WJ. Effect of
ibuprofen and acetaminophen on
postexercise muscle protein synthesis.
Am J Physiol Endocrinol Metab 2002:
282: E551–E556.
Veiga AP, Duarte ID, Avila MN, da
Motta PG, Tatsuo MA, Francischi JN.
Prevention by celecoxib of secondary
hyperalgesia induced by formalin in
rats. Life Sci 2004: 75: 2807–2817.
Vuolteenaho K, Moilanen T, Moilanen E.
Non-steroidal anti-inflammatory
drugs, cyclooxygenase-2 and the bone
healing process. Basic Clin Pharmacol
Toxicol 2008: 102: 10–14.
Warden SJ. Cyclo-oxygenase-2 inhibitors:
beneficial or detrimental for athletes
with acute musculoskeletal injuries?
Sports Med 2005: 35: 271–283.
Yu JG, Malm C, Thornell LE. Eccentric
contractions leading to DOMS do not
cause loss of desmin nor fibre necrosis
in human muscle. Histochem Cell Biol
2002: 118: 29–34.
COX-2 inhibition after eccentric exercise
e207
