The effect of various cold‑water immersion protocols

1 3
Eur J Appl Physiol
DOI 10.1007/s00421-014-2954-2
Original Article
The effect of various cold‑water immersion protocols
on exercise‑induced inflammatory response and functional
recovery from high‑intensity sprint exercise
Gillian E. White · Shawn G. Rhind · Greg D. Wells
Received: 30 January 2014 / Accepted: 10 July 2014
© Springer-Verlag Berlin Heidelberg 2014
and CWI-30 min × 10 °C conditions, while further
increases were observed for IL-8 and MPO in the CWI-
30 min × 20 °C and CWI-30 min × 10 °C conditions.
Squat jump and drop jump height were significantly lower
in all conditions immediately post-exercise and at 2 h. Drop
jump remained below baseline at 24 and 48 h in the CON
and CWI-10 min × 20 °C conditions only, while squat
jump height returned to baseline in all conditions.
Conclusions Cold-water immersion appears to facilitate
restoration of muscle performance in a stretch–shortening
cycle, but not concentric power. These changes do not
appear to be related to inflammatory modulation. CWI
protocols of excessive duration may actually exacerbate
the concentration of cytokines in circulation post-exercise;
however, the origin of the circulating cytokines is not nec-
essarily skeletal muscle.
Keywords Exercise-induced inflammation · Recovery ·
Cryotherapy · Cytokines · Skeletal muscle stress
Abbreviations
1 h 1 hour
2 h 2 hour
24 h 24 hour
48 h 48 hour
CWI Cold-water immersion
GM-CSF Granulocyte macrophage colony-stimulating
factor
IFNγ Interferon gamma
IL-1β Interleukin 1 beta
IL-6 Interleukin 6
IL-8 Interleukin 8
IL-10 Interleukin 10
IL-12 p70 Interleukin 12 p70
MPO Myeloperoxidase
Abstract
Purpose The purpose of this study was to investigate the
effects of different cold-water immersion (CWI) protocols
on the inflammatory response to and functional recovery
from high-intensity exercise.
Methods Eight healthy recreationally active males com-
pleted five trials of a high-intensity intermittent sprint
protocol followed by a randomly assigned recovery con-
dition: 1 of 4 CWI protocols (CWI-10 min × 20 °C,
CWI-30 min × 20 °C, CWI-10 min × 10 °C, or CWI-
30 min × 10 °C) versus passive rest. Circulating mediators
of the inflammatory response were measured from EDTA
plasma taken pre-exercise (baseline), immediately post-
exercise, and at 2, 24, and 48 h post-exercise. Ratings of
perceived soreness and impairment were noted on a 10-pt
Likert scale, and squat jump and drop jump were per-
formed at these time points.
Results IL-6, IL-8, and MPO increased significantly
from baseline immediately post-exercise in all con-
ditions. IL-6 remained elevated from baseline at 2 h
in the CWI-30 min × 20 °C, CWI-10 min × 10 °C,
Communicated by Fabio Fischetti.
G. E. White (*)
Graduate Department of Exercise Sciences, University
of Toronto, BN 60, 55 Harbord St., Toronto, ON M5S 2W6,
Canada
e-mail: gillian.white@mail.utoronto.ca
S. G. Rhind
Defence Research and Development Canada; Toronto Research
Centre, Toronto, Canada
G. D. Wells
Faculty of Kinesiology and Physical Education, University
of Toronto, Toronto, Canada

Eur J Appl Physiol
1 3
Pre Pre-exercise
Post Post-exercise
SSC Stretch-shortening cycle
TNFα Tumor necrosis factor alpha
Introduction
High-intensity exercise including eccentric activity, unac-
customed activity, and exercise of long duration and/or
high intensity has been shown to induce an acute inflam-
matory response (Tee et al. 2007; Bleakley and Davison
2010; Leeder et al. 2012). In response to stress and/or mus-
cle damage induced by exercise, muscle fibers and associ-
ated cells release signaling molecules (cytokines) into the
circulation (Tidball and Villalta 2010; Paulsen et al. 2012;
Suzuki et al. 2002; Peake et al. 2005b; Nieman et al. 2012),
which subsequently influence the trafficking of inflam-
matory cells (Adams et al. 2011; Stacey 2010; Butterfield
et al. 2006). This process initiates a positive feedback
mechanism characteristic of the inflammatory response,
resulting in further up-regulation of signaling molecules
and activation/infiltration of inflammatory cells into muscle
fibers that have been stressed or damaged during the exer-
cise bout (Pedersen 2011). The inflammatory response is
necessary for the resolution of any structural damage that
may have occurred, and may be important for the adaptive
response of skeletal muscle to exercise (Tidball and Villalta
2010; Pizza et al. 2002). However, if prolonged or exces-
sive, the inflammatory response can cause secondary dam-
age to surrounding cells and contribute to oxidative stress
of the muscle fibers (Tidball and Villalta 2010; Carvalho
et al. 2010; Puntel et al. 2011), thereby compounding the
stress/damage the muscle fibers experience. For this reason,
the inflammatory response has been implicated in delayed
onset muscle soreness (DOMS), edema, reduced perfor-
mance capacity, and fatigue commonly experienced follow-
ing various forms of strenuous exercise (Smith 1991).
Cold-water immersion (CWI) is a popular form of
cryotherapy that has been identified as a potentially effec-
tive anti-inflammatory intervention (Pournot et al. 2010;
Clarke et al. 1958; Swenson et al. 1996; White and Wells
2013). CWI combines the compressive hydrostatic forces
of water with low temperatures to reduce local metabolism
and induce vasoconstriction (Thorsson et al. 1985; Wilcock
et al. 2006). Following the stress of exercise, cold-induced
hypometabolism reduces muscle fibers’ O2 usage (Yanagi-
sawa et al. 2007; Yanagisawa and Fukubayashi 2010),
thereby reducing metabolic stress and the release of signal-
ing molecules from the tissue (Swenson et al. 1996). This
in turn blunts the positive feedback, whereby these mole-
cules attract and activate immune cells, which release more
inflammatory mediators. As cold has been shown to induce
vasoconstriction (Gregson et al. 2011; Yanagisawa et al.
2010; Thorsson et al. 1985), it may attenuate fluid accumu-
lation at the site of stressed muscle fibers and inflammatory
cell trafficking and signaling.
CWI is the most commonly used recovery modality
(Barnett 2006) and is used by a wide variety of athletes
to enhance recovery post-exercise; however, the protocols
used in the field are largely based on the anecdotal experi-
ence not evidence-based research. Although many studies
have investigated the cytokine response to exercise, few
have validated the anti-inflammatory effects of CWI as
the mechanism through which it may affect recovery from
intense exercise in normothermic conditions (Wilcock et al.
2006). Further, while CWI is widely used by athletes, there
continues to be ongoing debate in the literature regarding
its efficacy as a post-exercise strategy to facilitate recovery
from intense exercise. This is largely due to a high variabil-
ity in methodology used to study CWI as a recovery modal-
ity. Therefore, the objectives of this study were threefold:
(1) to measure the influence of CWI on exercise-induced
inflammation and muscle performance; (2) to investigate
inflammation as a mechanism for any recovery effect of
CWI; and (3) to identify the most effective CWI protocol
for restoring muscle function and/or attenuating exercise-
induced inflammation. We hypothesize that CWI will result
in lower circulating cytokine concentrations, which will
be related to improved muscle performance and that these
effects will be dose-dependent, such that the coldest condi-
tion (10 °C water for 30 min) will be associated with the
largest effects.
Materials and methods
Eight recreationally active males (defined as ‘participating
in structured exercise 3–6 h per week’, 23.6 ± 3.7 years,
180.8 ± 8.1 cm, 76.1 ± 8.6 kg, skin fold thickness
11.6 ± 3.5 mm, VO2peak 50.7 ± 5.1 ml·kg−1
min−1
) vol-
unteered to participate in the study. A power calculation
conducted using IL-6 (average within subject minimum
detectable difference of 1.72 ± 0.68 pg/ml) (Stacey 2010;
Peake et al. 2005b; Halson et al. 2008), the most com-
monly studied exercise-induced cytokine, resulted in a
minimum n = 6 to attain statistical power of 0.8 (G-Power
Statistical Analysis, Germany). Participants were not tak-
ing any medications, did not report history of anti-inflam-
matory/antioxidant use, were free of cardiopulmonary and/
or inflammatory conditions, and had no history of adverse
reactions to cold, including chilblains, frostbite, Rey-
naud’s phenomenon, or cold-induced urticaria. The study
was approved by the University of Toronto Research Eth-
ics Board and each participant provided written informed
consent.

Eur J Appl Physiol
1 3
Experimental overview
In a randomized crossover design, participants completed a
high-intensity intermittent sprint protocol followed imme-
diately by one of five recovery conditions assigned in a
random order over the 5 weeks of testing. Trial days were
separated by 1 week. Moderate to high-intensity exercise
(12 on the Borg 20 point scale), alcohol, and caffeine were
restricted 24 h prior to any blood collection. Nonsteroidal
anti-inflammatory drugs, steroidal anti-inflammatory drugs,
and antioxidant nutritional supplements were prohibited for
the duration of the study. Drop jump and squat jump height,
and ratings of perceived soreness and impairment were taken
pre-exercise (‘pre’), immediately post-exercise (‘post’), and
at 1, 2, 24, and 48 h post-exercise, while inflammatory mark-
ers were taken at these same time points excluding the 1 h
time point. Anthropometric measures, VO2peak, and familiari-
zation with the exercise protocol, immersion in 10 °C water,
and functional testing procedures was conducted 1 week
prior to commencing the experimental protocol. All test pro-
cedures were completed at the University of Toronto Athletic
Center. No food was consumed during the experimental tri-
als. Water was allowed ad libitum (Fig. 1).
Exercise protocol
The exercise protocol was performed on a 200-m indoor
track in the University of Toronto Athletic Center between
the hours of 9 a.m. and 5 p.m. Participants attended sessions
at the same time of day for each of the five trial weeks.
120 m was marked, and included one bend in the track, and
two straight aways. Following pre-exercise blood sampling,
participants completed a standardized dynamic warm-up
consisting of 800-m of jogging (four laps of the track), a
series of active stretches conducted on a 20-m sideline
of the track and a 5-min free stretch. The sprint-protocol
required the participants to complete 12 maximal sprints
of 120 m, performed every 3 min. This exercise protocol
was chosen to engage a large mass of muscle in high-force,
repeated stretch–shortening contractions over a period of
20–25 s (sufficient to deplete phosphocreatine stores). A
rest period of ~2.5 min was given to allow full replenish-
ment of phosphocreatine stores to support maximal effort
on subsequent bouts. This protocol was chosen to impose
both mechanical and metabolic stress, as opposed to pre-
dominantly mechanical and damaging (eccentric overload)
which would be expected to elicit a repeated-bout effect,
and therefore not appropriate for a crossover design.
A research assistant stood at the starting point to collect
pre-sprint heart rate (RS800CX, Polar, Finland). A second
research assistant stood at the finish point to collect post-
sprint heart rate acquired approximately 5 s following the
completion of the sprint. Both research assistants recorded
the time taken to complete the sprint and gave standardized
verbal encouragement. The participant then walked back
to the starting point and continued to walk along the side
Fig. 1 Schematic representation of relative timing of experimen-
tal procedures. Muscle performance and ratings of perceived sore-
ness and impairment were conducted pre-, immediately post-, and at
1, 2, 24, and 48 h post-exercise. Blood samples were taken at each
time point excluding 1 h post-exercise. The exercise protocol, includ-
ing warm-up and cool-down, was roughly 1 h in duration. Follow-
ing a 15 min lapse in which participants returned to the laboratory,
completed post-exercise testing, and changed into swim shorts, the
recovery period commenced, lasting 45 min for each experimental
condition (immersion time plus passive rest on a chair up to 45 min),
such that the next time point was 1 h post-exercise (15 min lapse plus
45 min recovery period)

Eur J Appl Physiol
1 3
of the track until 3 min had elapsed since the beginning
of the previous sprint. Upon completion of the 12 sprints
repetitions, the participant completed a 400-m cool-down
at a self-selected pace and returned to the laboratory for
post-exercise performance and blood sampling, followed
by the commencement of their randomly assigned recovery
condition.
Recovery protocol
Following post-exercise blood sampling, ratings of per-
ceived soreness and impairment, and changing into swim
shorts (15 min following cool-down on the track), par-
ticipants began a 45 min recovery intervention includ-
ing the specified time immersed and the remainder of
the 45 min spent in passive seated rest in a standard
upright chair. Immersion protocols were 10 °C water
for 10 min (CWI-10 min × 10 °C), 10 °C water for
30 min (CWI-30 min × 10 °C), 20 °C water for 10 min
(CWI-10 min × 20 °C), 20 °C water for 30 min (CWI-
30 min × 20 °C), or passive rest seated on a chair in the
laboratory for 45 min (CON). Recovery protocols were
assigned randomly by drawing a slip of paper with a letter
corresponding to one of the five conditions from an enve-
lope prior to each trial week. Participants were blinded to
their condition by covering the settings on the automated
regulator and withholding time information. The principal
researchers were blinded to the conditions by having a third
party code the conditions and set the temperature on the
regulator. Immersion took place in the laboratory in a port-
able ice bath (iCool Sport, Australia), in which participants
were immersed to the iliac crest and water was continu-
ally circulated by the ice bath’s automated regulator. When
the prescribed duration in the bath was completed in the
immersion conditions, the remainder of time until the 2-h
blood sampling point was spent quietly sitting on a chair in
the laboratory.
Muscle function
Muscle function was assessed using a squat jump and drop
jump performed on a force plate (Kistler, Switzerland) with
interfacing software (Quattro Jump 1.08). Jump tests were
conducted pre-exercise (baseline), immediately post-exer-
cise, and at 1, 2, 24, and 48 h post-exercise. Force genera-
tion is a well-established indirect indicator of muscle dam-
age (Paulsen et al. 2012; Warren et al. 1999). Squat jump
was used as a measure of maximal voluntary concentric
power, as it does not include pre-stretch activation (Byrne
and Eston 2002). As such, it gives an indirect measure of
muscle fiber activation, excitability, and functionality of
contractile elements. Participants started in a standing posi-
tion on the force plate with their hands on their hips. When
the computer indicated initiation of recording by an audi-
tory signal, the participant was instructed to squat down to
90°, pause for a second, and jump straight up maximally,
landing on the force plate and returning to a standing posi-
tion, where they remained until a second auditory signal
from the computer indicated the completion of recording.
Participants completed three squat jumps at each time point
and the mean of the best two from each time point was
used for analysis to eliminate the erroneous effect of a sub-
maximal effort.
Drop jump was used as a measure of maximal voluntary
power in a stretch–shortening cycle (SSC) with the eccen-
tric phase emphasized. As this jump includes a SSC, the
muscles derive force from pre-stretch activation including
elastic forces of passive tissues associated with the neuro-
muscular system, stretch–reflex activation, and/or altered
sarcomere length (Byrne and Eston 2002). Participants
started in a standing position on the force plate with their
hands on their hips. When the computer gave an auditory
signal indicating initiation of recording, the participant was
instructed to step on to the 40 cm box, turn, and step-off
the box, landing on the force plate, and immediately per-
forming a maximal vertical jump. Three drop jumps were
performed at each time point and the mean of the best two
jumps was used for analysis to eliminate the erroneous
effect of a submaximal effort.
Participants were familiarized with these two jumps
prior to commencing experimental trials to account for any
learning that may affect jump height, and to ensure correct
technique for the jumps. Scripted instructions were given
by a research assistant to the participants for each jump,
and participants were reminded to jump maximally for
each jump.
Blood sampling
Two, 4.0-ml vacutainer tubes treated with 7.2 mg of
EDTA (BD™ Vacutainer, USA) were collected from the
antecubital vein while the participant lay supine on an
examination table pre-, post-, 2, 24, and 48 h post-exer-
cise. Baseline samples were taken 10 min prior to starting
the exercise protocol (allowing time for functional testing
and ratings of perceived soreness and impairment follow-
ing blood sampling). Post-exercise samples were taken
immediately following the cool-down protocol on the
track, and later samples were taken 2, 24, and 48 h from
the baseline sampling time. The tubes were spun in a cen-
trifuge (Universal 320 R, Hettich, Germany) for 15 min
at 4 °C at 1,560g, and the plasma was aliquotted into 3,
0.5 ml Eppendorf tubes and frozen at −80 °C until the
time of assaying. The blood sample for a given time point
was taken prior to the bout of muscle performance testing
for that time point.

Eur J Appl Physiol
1 3
Inflammatory assays
Samples were analyzed for total plasma concentration
using electrochemiluminescence-based solid phase sand-
wich immunoassays (Meso-Scale Discovery, Gaithersbury,
MD). A 9-plex multiarray assay (Human, Ultrasensitive
Pro-inflammatory 9-plex, MSD, USA) for interleukin (IL)-
1, IL-2, IL-6, IL-8, IL-10, IL-12 p70, interferon gamma
(IFNer), tumor necrosis factor alpha (TNFαN), and granu-
locyte macrophage colony-stimulating factor (GM-CSF)
was used. A single-plex assay for myeloperoxidase (MPO)
(Human, Ultrasensitive Myeloperoxidase, MSD, USA) was
conducted. Samples and provided standards were analyzed
in duplicate according to manufacturers’ instructions. Inter-
assay coefficients of variance (CV) ranged from 0.6 to
3.4 %. Minimal detectable concentrations for the analytes
were 0.35 pg/ml for IL-2, 0.090 pg/ml for IL-8, 1.4 pg/ml
for IL-12 p70, 0.36 pg/ml for IL-1β, 0.20 pg/ml for GM-
CSF, 9.53 pg/ml for IFNγ, 0.27 pg/ml for IL-6, 0.21 pg/ml
for IL-10, 0.50 pg/ml for TNFα, and for 33 pg/ml MPO.
Ratings of perceived soreness and perceived impairment
were recorded pre-, post-, 1, 2, 24, and 48 h post-exercise.
Rating of perceived impairment was recorded on a 10-point
Likert numerical rating scale where “1” was “Fully Func-
tional” and “10” was “Debilitated”. Rating of perceived
soreness was recorded on a similar 10-point Likert scale
where “1” was “No soreness” and “10” was “Agonizing”.
Perceived soreness was rated during an eccentric con-
traction (sitting into a chair) and a concentric contraction
(standing out of the chair), with anatomical location of
soreness indicated on a diagram.
Statistical analyses
Data in text and figures represent the mean ± SD with a
95 % confidence interval, expressed as change from base-
line to account for inter-individual variation. A two-way
(five sample times × five recovery condition) repeated-
measures analysis of variance (ANOVA) was used to
determine main and interaction effects for sample time
and recovery condition for mean plasma cytokine concen-
tration change from baseline (pg/ml) using SPSS™ ver-
sion 21 (IBM, USA). A two-way (six sample times × five
recovery conditions) repeated measures ANOVA was
used to determine main and interaction effects for sample
time and recovery condition for mean jump height change
from baseline (cm) and mean rating on a 10-point Lik-
ert scale change from baseline using SPSS™ version 21
(IBM, USA). The assumption of sphericity was tested by
Mauchley’s W. In the case that sphericity was violated, the
Greenhouse–Geisser correction was used. Tukey’s least
significant difference was used to post hoc significant main
and interaction effects using GraphPad Prism software.
Relationships between average plasma cytokine concentra-
tion changes from baseline, jump height, and perceptions
of recovery were analyzed using Pearson’s product correla-
tions on SPSS™ version 21 (IBM, USA).
The standardized difference in means or effect size of
changes between each experimental group and control for
each post-intervention time point (1, 2, 24, and 48 h) was
calculated using pooled standard deviation. Cohen’s effect
size statistics were used as threshold values for effects
of a given experimental condition when compared with
control at each of the above time points, such that a ben-
eficial effect (greater than the smallest practically relevant
change), unclear, or harmful effect was assessed qualita-
tively according to the magnitude-based inference model
outlined by Hopkins et al. (2009). The inference was gen-
erated from the confidence limits derived from the p value
associated with the effect and the magnitude of the dif-
ference in means such that a 0 % probability that the true
value falls within the confidence limits corresponds with
a qualitative inference of a “most unlikely” effect; 0.5 %
“very unlikely”; 5 % “unlikely”; 25 % possibly; 75 %
likely; 95 % likely; 99.5 % very likely; 100 % most likely.
The effect was considered “unclear” if the confidence inter-
val of the effect was found to be both beneficial and harm-
ful (i.e. substantially positive/beneficial and negative/harm-
ful effect). Experimental mean ± SD, control mean ± SD,
difference in means ± 90 % CL and the associated qualita-
tive inference are reported.
Results
All participants completed the intermittent sprint protocol
on all five occasions. No significant differences between
conditions were observed on measures of exertion during
the exercise protocol or plasma cytokine concentrations at
the pre-exercise time point. A significant main effect for
time was observed in plasma IL-6, IL-8, and MPO concen-
tration change from baseline. No other inflammatory mark-
ers showed significant changes from baseline at any time
point.
Plasma IL‑6 concentration change from baseline
A significant main effect for time was observed for plasma
IL-6 (F(1.67,6.65) = 9.434, p 0.05). Plasma IL-6 was signif-
icantly greater than baseline post-exercise (p 0.05) in all
experimental conditions. At 2 h post-exercise plasma IL-6
concentration remained significantly elevated from baseline
in the CWI-30 min × 20 °C (1.118 ± 0.247 pg/ml), the

Eur J Appl Physiol
1 3
CWI-10 min × 10 °C (1.091 ± 0.532 pg/ml), and the CWI-
30 min × 10 °C conditions (1.191 ± 0.638 pg/ml). Plasma
IL-6 concentrations returned to baseline in all conditions at
24 and 48 h post-exercise (Fig. 2).
Plasma IL‑8 concentration change from baseline
A significant main effect for time (F(4,16) = 13.635,
p  0.01) and interaction effect for time and recovery
condition (F(16,64) = 2.079, p  0.05) were observed for
plasma IL-8 concentration change from baseline. Plasma
IL-8 was significantly greater than baseline post-exercise
(p  0.01) in all experimental conditions, except CWI-
10 min × 20 °C. At 2 h post-exercise, plasma IL-8 con-
centration remained significantly elevated from baseline
following CWI-30 min × 20 °C (1.350 ± 0.701 pg/ml)
and CWI-30 min × 10 °C (2.569 ± 0.841 pg/ml). Further
CWI-10 min × 10 °C was associated with significantly
lower plasma IL-8 (0.127 ± 0.340 pg/ml) compared with
CWI 30 min × 20 °C (p  0.05) and CWI 30 min × 10 °C
(p  0.05) at the 2 h time point. Plasma IL-8 concentrations
returned to baseline in all conditions at 24 and 48 h post-
exercise (Fig. 3).
Plasma MPO concentration change from baseline
A significant main effect for time (F(4,16) = 11.744,
p  0.01) and interaction effect for time and recovery
condition (F(16,64) = 1.915, p  0.05) were observed for
plasma MPO concentration change from baseline. Plasma
MPO was significantly greater than baseline post-exercise
(p  0.05) in all conditions and continued to increase, peak-
ing at 2 h post-exercise following CWI-30 min × 20 °C
(1.492 ± 0.619 pg/ml, p  0.01) and CWI-30 min × 10 °C
(2.028 ± 0.709 pg/ml, p  0.01). CWI-30 min × 10 °C was
associated with significantly higher plasma MPO when
compared with CWI-10 min × 10 °C (0.395 ± 0.121 pg/
ml, p  0.05) and CON (0.557 ± 0.235 pg/ml, p  0.05)
at 2 h post-exercise. Plasma MPO concentrations returned
to baseline in all conditions at 24 and 48 h post-exercise
(Fig. 4).
Squat jump
A significant main effect for time was observed
(F(5,20) = 49.434, p  0.01) for squat jump height change
from baseline. Squat jump height was significantly lower
than pre-exercise at all time points in all conditions, with
the greatest decrement in jump height observed at 1 h post-
exercise in all conditions. Although squat jump height was
below pre-exercise values in all of the conditions and 24 and
48 h post-exercise, it was significantly below pre-exercise
values at 48 h in CWI-10 min × 20 °C (−3.49 ± 2.3 cm)
and CWI-30 min × 10 °C (−2.10 ± 2.4 cm). No CWI con-
dition was more effective than passive rest for recovery of
squat jump performance at any time point (Fig. 5).
Magnitude-based inferences were used to make quali-
tative deductions about the effect of the respective CWI
protocols on recovery of squat jump height when com-
pared with control according to Hopkins et al. (2009).
A likely to very likely effect was found for the CWI pro-
tocols to induce a negative effect on squat jump height
at 1 h post-exercise, with the most substantially nega-
tive effects observed for CWI-10 min × 10 °C and CWI-
30 min × 10 °C. CWI-30 min × 10 °C was associated with
substantially negative effects on squat jump height at 2 h
post-exercise, while the other conditions were associated
with unclear effects. No clear inferences could be made on
Fig. 2 Mean plasma IL-6
change from baseline (pre)
following sprint exercise (post)
and 1 of 4 different CWI proto-
cols or passive rest (CON) at 2,
24, and 48 h post-exercise. Sig-
nificant change in mean plasma
concentration from baseline at a
given time point is indicated by
the letter corresponding to CON
or 1 of 4 CWI conditions listed
in the legend
-1
-0.5
0
0.5
1
1.5
2
2.5
3
Pre Post 2h 24 h 48 h
CON (*)
CWI-10minx20°C (A)
CWI-30minx20°C (B)
CWI-10minx10°C (C)
CWI-30minx10°C (D)
*
A
B
C
D
B
C
D
ChangeinplasmaIL-6concentration(pg.ml-1)
Time-point

Eur J Appl Physiol
1 3
the effects of any CWI condition at 24 and 48 h post-exer-
cise. Inferences of effects for the respective CWI condi-
tions when compared with CON at each post-cooling time
point are reported in Table 1.
Drop jump
A significant main effect for time was observed
(F(5,20) = 29.366, p  0.01) for drop jump height change
from baseline. Drop jump height was significantly
lower than pre-exercise immediately post-exercise, and
at 1 and 2 h post-exercise in all conditions. Drop jump
height remained below pre-exercise values at 24 h in
CON (−4.91 ± 3.0 cm, p  0.05), CWI-10 min × 20 °C
(−2.78 ± 2.9 cm, p  0.05), and CWI-30 min × 20 °C
(−2.47 ± 0.9 cm, p  0.05). At 48 h post-exercise, drop
jump height remained below pre-exercise values in all
conditions except CWI-10 min × 10 °C (1.96 ± 3.3 cm,
p  0.05), which was associated with significantly greater
jump height than CON (−4.63 ± 1.6 cm, p  0.01). CWI-
10 min × 10 °C was the only condition associated with
mean jump height values greater than baseline at any time
point, with positive change in mean jump height observed
at 24 and 48 h post-exercise (Fig. 6).
Magnitude-based inferences were used to make quali-
tative deductions about the effect of the respective CWI
protocols on recovery of drop jump height when compared
with control according to Hopkins et al. (2009). A likely
beneficial effect was found for CWI-10 min × 10 °C, a pos-
sible beneficial effect was found or CWI-10 min × 20 °C,
and unclear effects were found for the 30 min CWI
protocols, although no harmful effect is expected for
Fig. 3 Mean plasma IL-8
in the legend
-0.5
0
0.5
1
1.5
2
2.5
3
ChangeinplasmaIL-8concentration
(pg.ml-1)
Time-point
CON (*)
CWI-10minx20°C (A)
CWI-30minx20°C (B)
CWI-10minx10°C (C)
CWI-30minx10°C (D)
Pre Post 2h 24 h 48 h
*
B
C
D
B
D
Fig. 4 Mean plasma MPO
in the legend
-0.5
0
0.5
1
1.5
2
2.5
ChangeinplasmaMPOconcentration
(pg.ml-1)
Time-point
CON (*)
CWI-10minx20°C (A)
CWI-30minx20°C (B)
CWI-10minx10°C (C)
CWI-30minx10°C (D)
Pre Post 2 h 24 h 48 h
*
A
B
C
D
A
B
D

Eur J Appl Physiol
1 3
CWI-30 min × 10 °C. CWI-30 min × 10 °C was associated
with substantially negative effects on squat jump height at
2 h post-exercise, while the other conditions were associ-
ated with unclear effects. Beneficial effects were associ-
ated with all CWI conditions at 24 h post-exercise with
very likely beneficial effects associated with both 10 °C
conditions. Unclear effects of CWI were observed at 48 h
post-exercise, except for CWI-10 min × 10 °C, which was
associated with a very likely beneficial effect. Inferences
of effects for the respective CWI conditions compared
Fig. 5 Mean squat jump height
and 1 of 4 different CWI pro-
tocols or passive rest (CON) at
1, 2, 24, and 48 h post-exercise.
Significant change in jump
height from baseline at a given
time point is indicated by the
letter corresponding to CON or
1 of 4 CWI conditions listed in
the legend
-14
-12
-10
-8
-6
-4
-2
0
Changeinsquatjumpheight(cm)
Pre Post 1 h 2 h 24 h 48 h
Rest (*)
10min x 20°C (A)
30min x 20°C (B)
10min x 10°C (C)
30min x 10°C (D)
A
D
*
A
B
C
D
Time-point
*
A
B
C
D
*
A
B
C
D
Table 1 Mean change in squat jump height change from baseline for each CWI condition and control with qualitative inferences for the effect
of each condition on squat jump recovery following sprint-based exercise
90 %CL, add/subtract this value to the mean effect for the 90 % confidence limits of the population effect positive mean difference values reflect
a smaller negative effect on jump height in the control condition negative mean difference values reflect a smaller negative effect on jump height
in the experimental condition
Time point Control mean (cm)
± SD
Experimental condition Experimental mean (cm)
± SD
Mean differ-
ence (cm)
±90 % CL Qualitative inference
1 h CON
−6.2 ± 1.95
10 min × 20 °C −6.2 ± 4.8 2.1 ±3.6 Unclear (73 % likely to
harm)
30 min × 20 °C −7.4 ± 2.9 2.0 ±3.6 Unclear (76 % likely to
harm)
10 min × 10 °C −8.9 ± 3.3 3.5 ±3.6 Likely harmful
30 min × 10 °C −10.7 ± 2.8 4.5 ±3.6 Very likely harmful
2 h CON
−4.71 ± 1.58
10 min × 20 °C −5.2 ± 3.2 2.6 ±3.2 Unclear (likely harmful,
unlikely beneficial)
30 min × 20 °C −4.7 ± 3.4 0.8 ±3.2 Unclear
10 min × 10 °C −5.7 ± 3.9 1.9 ±3.2 Unclear
30 min × 10 °C −7.4 ± 1.7 3.4 ±3.2 Likely harmful, very
unlikely beneficial
24 h CON
−2.36 ± 0.37
10 min × 20 °C −1.0 ± 2.1 −0.1 ±2.4 Unclear
30 min × 20 °C −0.1 ± 2.8 −1.8 ±2.4 Unclear
10 min × 10 °C −0.7 ± 1.8 −1.5 ±2.4 Unclear
30 min × 10 °C −1.7 ± 1.8 −1.1 ±2.4 Unclear
48 h CON
−1.37 ± 2.59
10 min × 20 °C −2.3 ± 4.3 2.1 ±3.7 Unclear
30 min × 20 °C −0.2 ± 2.7 −1.1 ±3.7 Unclear
10 min × 10 °C −0.2 ± 5.6 −0.8 ±3.7 Unclear
30 min × 10 °C −2.3 ± 3.3 1.0 ±3.7 Unclear

Eur J Appl Physiol
1 3
with CON at each post-cooling time point are reported in
Tables 2, 3.
Muscle soreness was rated lowering into a chair (eccen-
tric) and standing (concentric) out of a chair. A significant
main effect for time was observed for rating of concentric
soreness (F(4,16) = 11.860, p  0.01), eccentric soreness
(F(4,16) = 13.709, p  0.01), and perceived impairment
(F(4,10) = 13.536, p  0.01).
Concentric soreness
Rating of perceived concentric soreness peaked post-exer-
cise (2.4 ± 0.4 points, p  0.01) in all recovery conditions
Fig. 6 Mean change from pre-
exercise (baseline) performance
for drop jump height. Signifi-
cant change in jump height from
baseline at a given time point is
indicated by the letter corre-
sponding to CON or 1 of 4 CWI
conditions listed in the legend
-10
-8
-6
-4
-2
0
2
4
Changeindropjumpheight(cm)
Pre Post 1 h 2 h 24 h 48 h
CON (*)
CWI-10minx20°C (A)
CWI-30minx20°C (B)
CWI-10minx10°C (C)
CWI-30minx10°C (D)
*
A
B
D
*
A
B
*
A
B
C
D
Time-point
*
A
B
C
D
*
A
B
C
D
Table 2 Mean change in drop jump height change from baseline for each CWI condition and control with qualitative inferences for the effect of
each condition on drop jump recovery following sprint-based exercise
90 % CL, add/subtract this value to the mean effect for the 90 % confidence limits of the population effect positive mean difference values reflect
a smaller negative effect on jump height in the control condition, negative mean difference values reflect a smaller negative effect on jump height
in the experimental condition
Time point Control mean (cm)
± SD
Experimental condition Mean (cm)
± SD
Mean differ-
ence (cm)
±90 % CL Inference
1 h CON
−6.28 ± 6.81
10 min × 20 °C −6.4 ± 3.6 0.2 ±4.6 Unclear
30 min × 20 °C −7.4 ± 5.9 0.2 ±4.6 Unclear
10 min × 10 °C −6.6 ± 7.1 0.4 ±4.6 Unclear
30 min × 10 °C −8.8 ± 2.4 1.8 ±4.6 Unclear (possible harmful)
2 h CON
−7.38 ± 3.43
10 min × 20 °C −4.6 ± 3.4 −2.1 ±4.4 Unclear (possibly beneficial)
30 min × 20 °C −7.5 ± 7.5 −0.9 ±4.4 Unclear
10 min × 10 °C −2.5 ± 5.2 −3.9 ±4.4 Likely beneficial, unlikely harmful
30 min × 10 °C −5.9 ± 3.0 −1.7 ±4.4 Unclear (unlikely harmful)
24 h CON
−3.77 ± 1.45
10 min × 20 °C −1.8 ± 2.8 −8.8 ±8.9 Likely beneficial, very unlikely harmful
30 min × 20 °C −3.6 ± 7.6 −8.8 ±8.9 Likely beneficial, very unlikely harmful
10 min × 10 °C +2.3 ± 5.9 −13.2 ±8.9 Very likely beneficial, very unlikely
harmful
30 min × 10 °C −0.6 ± 2.8 −10.8 ±8.9 Very likely beneficial, very unlikely
harmful
48 h CON
−4.39 ± 2.26
10 min × 20 °C −2.3 ± 1.4 −1.8 ±4.6 Unclear
30 min × 20 °C −2.1 ± 4.8 −2.3 ±4.6 Unclear
10 min × 10 °C +2.0 ± 6.4 −6.7 ±4.6 Very likely beneficial, most unlikely
harmful
30 min × 10 °C −2.2 ± 3.9 −2.7 ±4.6 Unclear (80 % likely to be beneficial)

Eur J Appl Physiol
1 3
and remained above baseline at 24 h (1.6 ± 0.6 points,
p  0.05). At 48 h rating of concentric soreness returned to
baseline (0.8 ± 0.4 points). No differences were observed
between conditions.
Eccentric soreness
Rating of perceived eccentric soreness peaked post-exer-
cise (1.9 ± 0.3 points, p  0.05) in all recovery conditions
and remained above baseline at 24 h (1.2 ± 0.4 points,
p  0.05). At 48 h rating of eccentric soreness returned to
baseline (0.9 ± 0.4 points). No differences were observed
between conditions.
Perceived impairment
Rating of perceived impairment peaked post-exercise
(4.4 ± 0.3 points, p  0.05) in all recovery conditions and
remained above baseline at 24 (2.8 ± 0.6 points, p  0.05)
and 48 h (1.4 ± 0.9 points). No differences were observed
between conditions.
Discussion
CWI is used widely by athletes of all types, however, the
protocols used are generally anecdotal and vary widely
with regards to duration and temperature. The present
study is the first to systematically compare different pro-
tocols of CWI for their purported anti-inflammatory and
subsequent muscle performance recovery effects following
intense exercise (Peterson and Pizza 2008). The main find-
ings of this study were that CWI following high-intensity
intermittent sprint exercise does not significantly reduce
plasma markers of inflammation or perceptions of sore-
ness and impairment. Further, CWI in both cold (10 °C)
and cool (20 °C) temperatures for 30 min can exacerbate
the response of inflammatory markers in the blood follow-
ing exercise. However, CWI-10 min × 10 °C appears to be
the most effective for restoring force production in a SSC.
The inflammatory response to intermittent sprint exercise
returned to baseline by 24 h, while perceptions of soreness
and impairment returned to baseline by 48 h and muscle
performance remained below baseline at 48 h.
The effect of CWI as a recovery modality remains con-
troversial, and although it is thought to function by blunt-
ing the exercise-induced inflammatory response, this
mechanism is not well supported in the literature (Smith,
1991). CWI has been shown to reduce intramuscular tem-
perature (Peiffer et al. 2009; Costello et al. 2012; Rupp
2012), subsequently reducing muscle perfusion (Gregson
et al. 2011; Yanagisawa et al. 2007) and O2 usage (Yanagi-
sawa et al. 2007). Reduced blood flow to the muscle is
thought to attenuate the potential for muscle fiber edema,
Table 3 Change from baseline of mean ratings of perceived soreness and perceived impairment following sprint exercise and 1 of 4 CWI proto-
cols or passive seated rest (CON)
Values are mean change from baseline on a 10-point Likert scale ± SD
Recovery condition
CON CWI-10 min × 20 °C CWI-30 min × 20 °C CWI-10 min × 10 °C CWI-30 min × 10 °C
Concentric soreness
Baseline 0 0 0 0 0
Post-ex 2.1 ± 1.0 3.1 ± 1.1 3.0 ± 1.0 1.9 ± 0.5 2.4 ± 0.6
2 h 1.4 ± 0.8 2.1 ± 0.6 1.6 ± 0.5 1.6 ± 0.5 1.1 ± 0.6
24 h 1.4 ± 0.4 2.1 ± 0.6 1.9 ± 0.6 1.4 ± 0.5 1.6 ± 0.7
48 h 0.7 ± 0.3 2.7 ± 0.9 0.6 ± 0.4 0.9 ± 0.4 0.9 ± 0.4
Eccentric soreness
Baseline 0 0 0 0 0
Post-ex 1.4 ± 0.7 2.2 ± 0.9 2.0 ± 0.5 1.7 ± 0.5 1.7 ± 0.6
2 h 1.2 ± 0.8 1.4 ± 0.5 0.8 ± 0.4 1.2 ± 0.5 1.0 ± 0.4
24 h 0.9 ± 0.3 1.4 ± 0.5 1.4 ± 0.4 1.2 ± 0.3 1.3 ± 0.6
48 h 1.2 ± 0.8 3.0 ± 2.7 0.8 ± 1.5 1.2 ± 1.3 1.0 ± 1.6
Perceived impairment
Baseline 0 0 0 0 0
Post-ex 3.3 ± 0.9 2.8 ± 1.1 3.6 ± 1.0 2.9 ± 0.8 3.4 ± 0.9
2 h 1.7 ± 0.6 2.5 ± 0.8 1.3 ± 0.6 2.1 ± 0.6 2.1 ± 0.9
24 h 1.4 ± 0.4 2.2 ± 0.8 1.5 ± 0.4 1.6 ± 0.6 2.2 ± 0.7
48 h 0.9 ± 0.5 2.6 ± 1.0 0.9 ± 0.5 1.3 ± 0.5 0.9 ± 0.5

Eur J Appl Physiol
1 3
while reducing the metabolic rate of stressed muscle fib-
ers is thought to reduce stress signaling and the ensuing
inflammatory response to intense exercise (Pournot et al.
2010). Decrements in muscle force generating capacity and
increased soreness experienced following intense exercise
are commonly attributed to edema and inflammation induc-
ing secondary damage to muscle fibers and other struc-
tures of the musculoskeletal system (Wilcock et al. 2006).
Hence, by reducing muscle perfusion and oxygen demand
of muscle following exercise, CWI is thought to attenuate
decrements in muscle performance and increased soreness
by reducing edema and inflammatory signaling.
Recovery of muscle performance
CWI is used by athletes exposed to varying physical
demands. The efficacy of CWI is equivocal, likely due to
high variability in the exercise stress and CWI protocols
studied. Our findings suggest that CWI may be effective
for restoring muscle performance in a stretch–shortening
movement in the days following exercise, but does not
significantly effect the restoration of muscle performance
in a purely concentric movement, nor is it appropriate for
short-term (2 h) restoration of force generation. Squat
jump performance was negatively affected by CWI at 1
and 2 h post-exercise when compared with CON. This is
likely due to reduced intramuscular temperature follow-
ing the four CWI protocols, slowing motor and sensory
neural conductance velocity (Herrera et al. 2010) and
contraction kinetics (Bergh and Ekblom 1979). Although
intramuscular temperature was not measured in the cur-
rent study previous studies have shown significant reduc-
tions in intramuscular and subcutaneous temperature
with similar temperatures and durations as the current
study (Yanagisawa et al. 2007). No CWI was more effec-
tive than CON for squat jump performance at 24 or 48 h.
Pointon et al. (2011) found voluntary concentric force
to be impaired 11 % at 24 h following two repetitions of
CWI for 9 min at 9 °C after intermittent sprint exercise
in the heat (Pointon et al. 2011). Similarly, Bailey et al.
(2007) found no effect of CWI for 10 min at 10 °C for
squat jump performance recovery 24 and 48 h follow-
ing shuttle running, despite improvements in maximum
voluntary contraction force when compared with passive
rest (Bailey et al. 2007). Thus, although squat jump is a
valid measure of voluntary muscle activation, it may not
be sensitive enough to detect subtle changes in specific
force generation of muscle. This may also raise the issue
of internal versus external validity of muscle performance
tests as those with greater internal validity (i.e., MVC)
may not translate to more applied muscle performance
tests (i.e., squat jump). Magnitude-based inferences indi-
cate that CWI is likely to harm squat jump performance at
1 and 2 h post-exercise, and should thus be used with cau-
tion for “pre-cooling” or between bout techniques prior to
activities requiring concentric power. Recommendations
for the use of CWI for recovery on concentric power 24–
48 h post-exercise are unclear.
In contrast to the absence of effect of CWI on squat
jump performance, drop jump height change from base-
line was significantly smaller at 2, 24, and 48 h follow-
ing CWI-10 min × 10 °C when compared with CON, as
well as following CWI-30 min × 10 °C when compared
with CON at 24 h. CWI-10 min × 10 °C was the only
condition associated with positive change from baseline
at 24 and 48 h. The major difference in force production
between squat jump and drop jump is the derivation of
force from pre-stretch activation in drop jump, as it con-
tains a SSC. Pre-stretch activation contributes force to a
movement by a combination of utilizing elastic energy of
structures of the musculature, activating proprioceptive
reflexes, increasing the time to produce force and/or by
potentiating sarcomere length (Ettema 2001; Horita et al.
1996). The discrepancy observed between squat jump and
drop jump recovery suggest that following sprint exercise,
performance reduction may be a result of neuromuscular
impairment rather than structural disruption due to the
inflammatory process.
It has been previously shown that concentric and
stretch–shortening movements follow different recovery
time courses owing to impairment in stretch reflex acti-
vation (Nicol et al. 1996). Activation of types III and IV
muscle afferents may reduce the facilitation of the stretch
reflex and its subsequent contribution to force production.
Because CWI has been shown to alter neural conductance
velocity (Herrera et al. 2010), it is possible that it may be
preserving stretch–reflex sensitivity in the days following
high-intensity sprint exercise. Previous studies have also
shown CWI to be effective at facilitating force production
in a stretch–shortening movement (i.e. counter-movement
jump) (Pournot et al. 2010), which may be of particu-
lar importance for athletes requiring recovery of muscle’s
ability to produce force in change of direction movements.
Our results suggest that CWI for 10 min at 10 °C was the
most effective for promoting recovery of force generation
in this type of movement. Magnitude-based inferences
indicate that CWI-10 min × 10 °C is likely beneficial for
recovery of power production in a SSC 2 h post-exercise.
The effects of all CWI protocols are likely to be beneficial
on recovery of drop jump performance 24 h post-exer-
cise compared with control. 10 °C conditions for 10 or
30 min are very likely to have a beneficial effect on drop
jump performance and are very unlikely to have a harm-
ful effect. CWI-10 min × 10 °C is very likely to also have
a beneficial effect at 48 h post-exercise when compared
with passive rest and, is thus, recommended for the use

Eur J Appl Physiol
1 3
for recovery of stretch–shortening force production 24 and
48 h post-exercise.
Effect on post‑exercise inflammation
IL-6 is classically up-regulated in skeletal muscle during
exercise (Peake et al. 2005a; Nieman et al. 2012; Bruun-
sgaard et al. 1997; Suzuki et al. 2002), and may also be
produced and released by other cells, including leukocytes,
fibroblasts, and endothelial cells (Tidball and Villalta 2010;
Pedersen et al. 1998). Although several studies have inves-
tigated post-exercise IL-6 responses, few have studied the
effects of CWI following exercise in normothermic con-
ditions. Nemet et al. (2009) found an increase in plasma
IL-6 following 250 m intermittent sprints, and no effect of
local ice application to the hamstring muscles, despite see-
ing reduced circulation of other cytokines and hormones
(Nemet et al. 2009). In the present study, plasma IL-6 con-
centration similarly increased following intermittent sprint
exercise. In contrast to our hypothesis, no CWI condition
was associated with reduced IL-6 concentration when
compared with CON. Indeed, plasma IL-6 concentration
was significantly elevated in both 30-min conditions when
compared with CON and CWI-10 min × 10 °C at 2 h post-
exercise. The exacerbation of the cytokine response sug-
gests that 30 min immersed in cold or cool water imposes
a compound stress on the body. A similar increase in IL-6
has been observed following CWI for recovery from exer-
cise (Lee et al. 2012a; Vaile et al. 2007). Although the dura-
tion of time was not held constant between conditions and
no control condition was used, Lee et al. (2012) observed
an increased IL-6 response after immersion in 11.7 °C
water compared with a decrease in IL-6 after immersion
in 23.5 °C water (Lee et al. 2012b). This response is fur-
ther supported by human studies of exercise fatigue dur-
ing short- or long-term cold exposure, indicating immuno-
stimulatory effects of cold (Castellani et al. 2002).
IL-6 is classically up-regulated and released from skel-
etal muscle during exercise of high intensity (Nieman et al.
2012) and is also dependent on the mass of muscle engaged
(Peake et al. 2005b). Because it is released from inflamma-
tory cells and endothelium during injury or infection, the
IL-6 response to exercise is thought to exert pro-inflamma-
tory effects and has been associated with exercise-induced
inflammation, muscle damage (Bruunsgaard et al. 1997),
and delayed onset muscle soreness (Tomiya 2004). More
recently, the effects of IL-6 released from skeletal muscle
as a ‘myokine’ have suggested that it is released from the
muscle to act in a hormone-like manner (Pedersen et al.
2003). Interestingly, IL-6 released from the muscle is also
shown to inhibit production of TNFα and enhance IL-1ra
and IL-10 anti-inflammatory cytokine production (Ped-
ersen 2007), suggesting a net anti-inflammatory effect of
IL-6 when released from the muscle. Although the elevated
IL-6 observed in these CWI conditions and the potential
role for muscle-derived IL-6 to act as an anti-inflamma-
tory cytokine may suggest an anti-inflammatory activity of
CWI, this is not consistent with the findings for IL-8 and
MPO, which are pro-inflammatory and not pleiotropic.
An alternative explanation may be impaired clearance of
IL-6 due to cold-induced vasoconstriction. Arterio-venous
difference studies have shown the plasma IL-6 changes
during exercise are almost entirely due to release from
contracting skeletal muscle (Steensberg et al. 2000). Thus,
the sustained elevation observed in plasma IL-6 concentra-
tion observed following both 30 min conditions and both
10 °C conditions may be a result of reduced blood flow to
the muscle during the recovery period. It has been previ-
ously shown that CWI induces significant reductions in
intramuscular blood flow (Gregson et al. 2011). It may be
that the reduction in blood flow to the muscle impedes the
clearance of IL-6 produced in the muscle, resulting in sus-
tained circulation in the blood during the acute recovery
phase (0–2 h post-exercise). However, without tissue spe-
cific measurement of IL-6 to directly measure production
and clearance at skeletal muscle, this is merely speculation.
Owing to previous reports of cold-induced upregulation of
IL-6, it is likely that IL-6 measured in the plasma at 2 h in
the present study represents IL-6 production from multiple
cell types.
IL-8 is a known neutrophil chemoattractant (Baggiolini
and Clark-Lewis 1992) that is up-regulated in skeletal mus-
cle during intense exercise (Nieman et al. 2012) and MPO
is an enzyme implicit in neutrophil cytotoxicity, thus is
indicative of neutrophil activation (Klebanoff 2005). IL-8 is
commonly up-regulated and released from skeletal muscle
during intense exercise (Nieman et al. 2012). As neutrophil
trafficking, thus presence in circulation, can be enhanced in
the absence of increased activity, MPO was used to indi-
rectly indicate the change in neutrophil activity associated
with exercise and subsequent CWI. This was chosen in
lieu of more general tissue oxidation detection to connect
the specific chemoattraction induced by IL-8 changes for
neutrophil trafficking and subsequent activity of neutro-
phils, rather than oxidative agents in general. Significant
increases from baseline in plasma IL-8 and MPO concen-
tration were observed following high-intensity intermittent
sprint exercise in all experimental conditions, except IL-8
in the CWI-10 min × 20 °C condition. Furthermore, plasma
IL-8 and MPO concentrations remained elevated or contin-
ued to increase beyond post-exercise concentrations fol-
lowing CWI-30 min × 20 °C and CWI-30 min × 10 °C at
2 h post-exercise, suggesting an exacerbation of the inflam-
matory response to exercise when prolonged cold exposure
is applied. Although plasma cytokines were not measured,
Stacey (2010) observed increased trafficking of neutrophils

Eur J Appl Physiol
1 3
following CWI-10 min × 20 °C following intermittent all
out exercise compared with control (Stacey 2010). As IL-8
is a neutrophil chemoattractant and was observed to be up-
regulated post-exercise, it is possible that neutrophil traf-
ficking was also altered; however, neutrophil counts were
not measured in the current study. While IL-8 returned to
near baseline in CON, CWI-10 min × 20 °C, and CWI-
10 min × 10 °C, it continued to increase following CWI-
30 min × 20 °C and CWI-30 min × 10 °C, suggesting that
while stress signals abated in the 10 min CWI conditions
and control, the stress response persisted in the 30 min
CWI conditions. This finding is supported by studies show-
ing cold changes in neutrophil activation and adhesion to
the vascular endothelium in association with increases in
IL-8 production (Bøkenes et al. 2004). Given the similar
pattern observed in plasma MPO in the 30-min conditions,
neutrophil functional activity appears to have been up-reg-
ulated, as well as chemotactic signaling. Although it was
not significant, both of the 30-min CWI conditions were
associated with a latent increase in ratings of soreness and
impairment at 24 h post-exercise. It may be that the acute
inflammatory response observed at 2 h may have contrib-
uted to secondary muscle damage resulting in enhanced
perceptions of soreness and impairment. It seems likely
that prolonged exposure to CWI may enhance the acute
(0–2 h post-exercise) inflammatory response following a
bout of high-intensity intermittent exercise. An unexpected
finding was the non-significant increase in IL-8 observed
post-exercise in the CWI-10 min × 20 °C. This may reflect
the variability in the exercise-induced cytokine response,
as well as the complexity of the response of chemokines
released from multiple cell types in response to a common
stressor. IL-8 mRNA has been shown to increase signifi-
cantly in muscle, but is not released to the same extent as
IL-6 from skeletal muscle (Pedersen et al. 2007), suggest-
ing that the increase in the plasma may be more attributable
to inflammatory or endothelial perturbation. Nevertheless,
the pattern remains consistent of elevated IL-8 at 2 h when
compared with post-exercise and sustained or enhanced
elevation of MPO at 2 h at this time point, supporting our
suggestion that these observations are related to neutrophil
trafficking and activation.
As blood samples were drawn from the anticubital veins,
we cannot infer where the additional cytokine response
originated. It may be that blood flow to the muscles was
reduced for a long enough period of time that upon exit-
ing the bath, the reperfusion of the muscles might have
resulted in an increase in inflammatory marker production
and release. Alternatively, it may be a similar ischemia/
reperfusion response from the cutaneous tissue exposed to
cooling. Muscle biopsy samples could help resolve these
uncertainties in future studies. In addition, intramuscular
temperatures were not measured, therefore, the specific
relationship between muscle temperature per se and post-
exercise inflammatory response remains unexplored.
Although participants and researchers were largely
blinded to the specifics of the study protocol, the nature of
the experimental conditions and control conditions were
such that it is very difficult to truly double-blind this type
of study.
Conclusion
CWI for 10 min at 10 °C appears very likely to be more
effective than passive recovery at restoring force generating
capacity of muscle in a SSC, but no CWI protocol used in
the current study was effective at restoring performance in
a purely concentric movement. CWI does not attenuate the
inflammatory response to an acute bout of normothermic
high-intensity intermittent sprint exercise when compared
with passive recovery. 30-min immersions to the iliac crest
in both cool (20 °C) and cold (10 °C) water appear to exac-
erbate specific aspects of the exercise-induced inflammatory
response. Performance effects CWI used following normo-
thermic sprint exercise are not likely a result of attenuation
of the inflammatory response to this type of exercise.
Acknowledgments We would like to acknowledge Research Pro-
grams in Applied Sport Sciences for funding this study, iCool Sport
Australia and the Canadian Sport Institute of Ontario for providing
equipment, and Defence Research and Development Canada for pro-
viding means for immunological analyses.
Conflict of interest The authors report no potential conflicts of
interest.
References
Adams GR, Zaldivar FP, Nance DM, Kodesh E, Radom-Aizik S,
Cooper DM (2011) Exercise and leukocyte interchange among
central circulation, lung, spleen, and muscle. Brain Behav Immun
25:658–666
Baggiolini M, Clark-Lewis I (1992) Interleukin-8, a chemotactic and
inflammatory cytokine. FEBS Lett 307:97–101
Bailey DM, Erith SJ, Griffin PJ, Dowson A, Brewer DS, Gant N, Wil-
liams C (2007) Influence of cold-water immersion on indices of
muscle damage following prolonged intermittent shuttle running.
J Sports Sci 25:1163–1170
Barnett A (2006) Using recovery modalities between training sessions
in elite athletes: does it help? Sports Med 36:781–796
Bergh U, Ekblom B (1979) Influence of muscle temperature on maxi-
mal muscle strength and power output in human skeletal muscles.
Acta Physiol Scand 107:33–37
Bleakley CM, Davison GW (2010) What is the biochemical and
physiological rationale for using cold-water immersion in sports
recovery? A systematic review. Br J Sports Med 44:179–187
Bøkenes L, Alexandersen TE, Tveita T, Osterud B, Mercer JB (2004)
Physiological and hematological responses to cold exposure in
young subjects. Int J Circumpolar Health 63:115–128

Eur J Appl Physiol
1 3
Bruunsgaard H, Galbo H, Halkjaer-Kristensen J, Johansen TL,
MacLean DA, Pedersen BK (1997) Exercise-induced increase
in serum interleukin-6 in humans is related tomuscle damage. J
Phys 499:833–841
Butterfield TA, Best TM, Merrick MA (2006) The dual roles of neu-
trophils and macrophages in inflammation: a critical balance
between tissue damage and repair. J Athl Train 41:457–465
Byrne C, Eston R (2002) The effect of exercise-induced muscle dam-
age on isometric and dynamic knee extensor strength and vertical
jump performance. J Sports Sci 20:417–425
Carvalho N, Puntel G, Correa P, Gubert P, Amaral G, Morais J, Royes
L, da Rocha J, Soares F (2010) Protective effects of therapeutic
cold and heat against the oxidative damage induced by a muscle
strain injury in rats. J Sports Sci 28:923–935
Castellani JW, M. Brenner IK, Rhind SG (2002) Cold exposure:
human immune responses and intracellular cytokine expression.
Med Sci Sports Exerc 34:2013–2020
Clarke RS, Hellon RF, Lind AR (1958) Vascular reactions of the
human forearm to cold. Clin Sci 17:165–179
Costello JT, Culligan K, Selfe J, Donnelly AE (2012) Muscle, skin
and core temperature after −110 °C cold air and 8 °C water treat-
ment. PLoS One 7:e48190
Ettema GJC (2001) Muscle efficiency: the controversial role of elas-
ticity and mechanical energy conversion in stretch-shortening
cycles. Eur J Appl Physiol 85:457–465
Gregson W, Black MA, Jones H, Milson J, Morton J, Dawson B,
Atkinson G, Green DJ (2011) Influence of cold-water immer-
sion on limb and cutaneous blood flow at rest. Am J Sports Med
39:1316–1323
Halson SL, Quod MJ, Martin DT, Gardner AS, Ebert TR, Laursen PB
(2008) Physiological responses to cold water immersion follow-
ing cycling in the heat. Int J Sports Physiol Perform 3:331–346
Herrera E, Sandoval MC, Camargo DM, Salvini TF (2010) Motor and
sensory nerve conduction are affected differently by ice pack, ice
massage, and cold water immersion. Physical Ther 90:581–591
Hopkins WG, Marshall SW, Batterham AM, Hanin J (2009) Progres-
sive statistics for studies in sports medicine and exercise science.
Med Sci Sports Exerc 41:3–13
Horita T, Komi PV, Nicol C, Kyröläinen H (1996) Stretch shortening
cycle fatigue: interactions among joint stiffness, reflex, and mus-
cle mechanical performance in the drop jump. Eur J Appl Physiol
Occup Physiol 73:393–403
Klebanoff SJ (2005) Myeloperoxidase: friend and foe. J Leuko Biol
77:598–625
Lee EC, Watson G, Casa D, Armstrong LE, Kraemer W, Vingren JL,
Spiering BA, Maresh CM (2012) Interleukin-6 responses to water
immersion therapy after acute exercise heat stress: a pilot investi-
gation. J Athl Train 47:655–663
Leeder J, Gissane C, van Someren K, Gregson W, Howatson G (2012)
Cold water immersion and recovery from strenuous exercise: a
meta-analysis. Br J Sports Med 46:233–240
Nemet D, Meckel Y, Bar-Sela S, Zaldivar F, Cooper DM, Eliakim A
(2009) Effect of local cold-pack application on systemic anabolic
and inflammatory response to sprint-interval training: a prospec-
tive comparative trial. Eur J Appl Physiol 107:411–417
Nicol C, Komi PV, Horita T, Kyröläinen H, Takala TE (1996) Reduced
stretch–reflex sensitivity after exhausting stretch-shortening cycle
exercise. Eur J Appl Physiol Occup Physiol 72:401–409
Nieman DC, Konrad M, Henson DA, Kennerly K, Shanely RA,
Wallner-Liebmann SJ (2012) Variance in the acute inflammatory
response to prolonged cycling is linked to exercise intensity. J
Interferon Cytokine Res 32:12–17
Paulsen G, Mikkelsen UR, Raastad T, Peake JM (2012) Leucocytes,
cytokines and satellite cells: what role do they play in muscle
damage and regeneration following eccentric exercise? Exerc
Immunol Rev 18:42–97
Peake JJ, Nosaka KK, Suzuki KK (2005a) Characterization of inflam-
matory responses to eccentric exercise in humans. Exerc Immu-
nol Rev 11:64–85
Peake JM, Suzuki K, Hordern M, Wilson G, Nosaka K, Coombes JS
(2005b) Plasma cytokine changes in relation to exercise intensity
and muscle damage. Eur J Appl Physiol 95:514–521
Pedersen BK (2007) IL-6 signalling in exercise and disease. Biochem
Soc Trans 35:1295–1297
Pedersen BK (2011) Muscles and their myokines. J Exp Biol
214:337–346
Pedersen BK, Ostrowski K, Rohde T, Bruunsgaard H (1998) The
cytokine response to strenuous exercise. Can J Physiol Pharmacol
76:505–511
Pedersen BK, Steensberg A, Keller P, Keller C, Fischer C, Hiscock N,
van Hall G, Plomgaard P, Febbraio MA (2003) Muscle-derived
interleukin-6: lipolytic, anti-inflammatory and immune regula-
tory effects. Eur J Physiol 446:9–16
Pedersen BK, Akerstrom TCA, Nielsen AR, Fischer CP (2007)
Role of myokines in exercise and metabolism. J Appl Phys
103:1093–1098
Peiffer JJ, Abbiss CR, Watson G, Nosaka K, Laursen PB (2009)
Effect of cold-water immersion duration on body temperature and
muscle function. J Sports Sci 27:987–993
Peterson JM, Pizza FX (2008) Cytokines derived from cultured skele-
tal muscle cells after mechanical strain promote neutrophil chem-
otaxis in vitro. J Appl Phys 106:130–137
Pizza FX, Koh TJ, McGregor SJ, Brooks SV (2002) Muscle inflam-
matory cells after passive stretches, isometric contractions, and
lengthening contractions. J Appl Physiol 92:1873–1878
Pointon M, Duffield R, Cannon J, Marino FE (2011) Cold water
immersion recovery following intermittent-sprint exercise in the
heat. Eur J Appl Physiol 112:2483–2494
Pournot H, Bieuzen F, Duffield R, Leprêtre PM, Cozzolino C, Hauss-
wirth C (2010) Short term effects of various water immersions
on recovery from exhaustive intermittent exercise. Eur J Appl
Physiol 111:1287–1295
Puntel GO, Carvalho NR, Amaral GP, Lobato LD, Silveira SO, Dauber-
mann MF, Barbosa NV, Rocha JBT, Soares FAA (2011) Therapeu-
tic cold: an effective kind to modulate the oxidative damage result-
ing of a skeletal muscle contusion. Free Radic Res 45:133–146
Rupp KA (2012) Intramuscular temperature changes during and after
2 different cryotherapy interventions in healthy individuals. J
Orthop Sports Phys Ther 42:731–737
Smith LL (1991) Acute inflammation: the underlying mecha-
nism in delayed onset muscle soreness? Med Sci Sports Exerc
23:542–551
Stacey DL (2010) Effects of recovery method on performance,
immune changes, and psychological outcomes. J Orthop Sports
Phys Ther 40:656–665
Steensberg A, van Hall G, Osada T, Sacchetti M, Saltin B, Pedersen
BK (2000) Production of interleukin-6 in contracting human
skeletal muscles can account for the exercise-induced increase in
plasma interleukin-6. J Physiol 529(Pt 1):237–242
Suzuki K, Nakaji S, Yamada M, Totsuka M, Sato K, Sugawara K
(2002) Systemic inflammatory response to exhaustive exercise.
Cytokine kinetics. Exerc Immunol Rev 8:6–48
Swenson C, Swärd L, Karlsson J (1996) Cryotherapy in sports medi-
cine. Scand J Med Sci Sports 6:193–200
Tee JC, Bosch AN, Lambert MI (2007) Metabolic consequences of
exercise-induced muscle damage. Sports Med 37:827–836
Thorsson OO, Lilja BB, Ahlgren LL, Hemdal BB, Westlin NN (1985)
The effect of local cold application on intramuscular blood flow
at rest and after running. Med Sci Sports Exerc 17:710–713
Tidball JG, Villalta SA (2010) Regulatory interactions between mus-
cle and the immune system during muscle regeneration. Am J
Physiol Regul Integr Comp Physiol 298:R1173–R1187

Eur J Appl Physiol
1 3
Tomiya A (2004) Myofibers express il-6 after eccentric exercise. Am J
Sports Med 32:503–508
Vaile J, Halson S, Gill N, Dawson B (2007) Effect of hydrotherapy on
the signs and symptoms of delayed onset muscle soreness. Eur J
Appl Physiol 102:447–455
Warren GL, Lowe DA, Armstrong RB (1999) Measurement tools
used in the study of eccentric contraction-induced injury. Sports
Med 27:43–59
White GE, Wells GD (2013) Cold-water immersion and other forms
of cryotherapy: physiological changes potentially affecting recov-
ery from high-intensity exercise. Extrem Physiol Med 2:26
Wilcock IM, Cronin JB, Hing WA (2006) Physiological response
to water immersion: a method for sport recovery? Sports Med
36:747–765
Yanagisawa O, Fukubayashi T (2010) Diffusion-weighted magnetic
resonance imaging reveals the effects of different cooling temper-
atures on the diffusion of water molecules and perfusion within
human skeletal muscle. Clin Radiol 65:874–880
Yanagisawa O, Homma T, Okuwaki T, Shimao D, Takahashi H (2007)
Effects of cooling on human skin and skeletal muscle. Eur J Appl
Physiol 100:737–745
Yanagisawa O, Takahashi H, Fukubayashi T (2010) Effects of differ-
ent cooling treatments on water diffusion, microcirculation, and
water content within exercised muscles: evaluation by magnetic
resonance T2-weighted and diffusion-weighted imaging. J Sports
Sci 28:1157–1163

The effect of various cold‑water immersion protocols

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to The effect of various cold‑water immersion protocols (20)

More from Fernando Farias (20)

Recently uploaded (20)

The effect of various cold‑water immersion protocols