Available via license: CC BY 4.0
Content may be subject to copyright.
RESEARCH ARTICLE
The Effect of Cold Showering on Health and
Work: A Randomized Controlled Trial
Geert A. Buijze
1
*, Inger N. Sierevelt
2
, Bas C. J. M. van der Heijden
3
, Marcel G. Dijkgraaf
4
,
Monique H. W. Frings-Dresen
5
1Department of Orthopaedic Surgery, Academic Medical Center, Amsterdam, The Netherlands,
2Department of Orthopaedic Surgery, Medical Center Slotervaart, Amsterdam, The Netherlands, 3Risk
Management, Achmea, Zeist, The Netherlands, 4Clinical Research Unit, Academic Medical Center,
Amsterdam, The Netherlands, 5Department Coronel Institute of Occupational Health, Academic Medical
Center, Amsterdam, The Netherlands
*g.a.buijze@amc.nl
Abstract
Purpose
The aim of this study was to determine the cumulative effect of a routine (hot-to-) cold
shower on sickness, quality of life and work productivity.
Methods
Between January and March 2015, 3018 participants between 18 and 65 years without severe
comorbidity and no routine experience of cold showering were randomized (1:1:1:1) to a (hot-
to-) cold shower for 30, 60, 90 seconds or a control group during30 consecutive days followed
by 60 days of showering cold at their own discretion for the intervention groups. The primary
outcome was illness days and related sickness absence from work. Secondary outcomes
were quality of life, work productivity, anxiety, thermal sensation and adverse reactions.
Results
79% of participants in the interventions groups completedthe 30 consecutive days protocol.
A negative binomial regression model showed a 29% reduction in sickness absence for
(hot-to-) cold shower regimen compared to the control group (incident rate ratio: 0.71, P =
0.003). For illness days there was no significant group effect. No related serious advents
events were reported.
Conclusion
A routine (hot-to-) cold shower resulted in a statistical reduction of self-reported sickness
absence but not illness days in adults without severe comorbidity.
Trial Registration
Netherlands National Trial Register NTR5183
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 1 / 15
a11111
OPEN ACCESS
Citation:Buijze GA, Sierevelt IN, van der Heijden
BCJM, Dijkgraaf MG, Frings-Dresen MHW (2016)
The Effect of Cold Showering on Health and Work: A
Randomized Controlled Trial. PLoS ONE 11(9):
e0161749.doi:10.1371/journal.pone.0161749
Editor: Jacobus van Wouwe, TNO, NETHERLANDS
Received:December 17, 2015
Accepted:August 11, 2016
Published:September 15, 2016
Copyright:© 2016 Buijze et al. This is an open
access article distributed under the terms of the
CreativeCommons Attribution License, which permits
unrestricted use, distribution,and reproduction in any
medium, provided the original author and source are
credited.
Data Availability Statement: All relevant data are
within the paper and its Supporting Information files.
Funding: The authors received no specific funding
for this work. One author [BCJMvdH] was employed
by a commercial company [Achmea] but did not
receive any support in either time or salary by his
employer. All work was performed in the author’s
spare time. The commercial company did not have
any role in the study design, data collection and
analysis, decision to publish,or preparation of the
manuscript. The specific roles of these authors are
articulated in the ‘author contributions’ section.
Introduction
Cold bathing is a common custom in many parts of the world. Ever since the introduction of
civilized bathing, humans have experimented with water temperature variation to expose the
body to extreme conditions. In ancient times, Roman bathing was based around the practice of
moving through a series of heated rooms culminating in a cold plunge at the end.[1] In modern
times, the traditional ritual of the frigidarium has been kept in most saunas and spas around
the world.
Cold bathing has been claimed to have multiple beneficial effects on health such as improve-
ment of the immune system, cardiovascular circulation and vitality, but any true association
remains unclear.[2] Previous investigations on the short-term effects of cold exposure have
shown increases of cortisol and norepinephrine concentrations with modulation of the physio-
logical response but showed minimal or no immune modulation.[3–7] However, the cumula-
tive clinical effect and relevance for health after adaption of cold exposure (response
conditioning) in healthy humans remain speculative as randomized controlled trials are
lacking.
The primary objective of this trial was to determine whether perceived illness could be mod-
ulated after repeated pragmatic cold exposure by taking a cold shower for at least 30 consecu-
tive days. Secondaryobjectives were to determine whether there was any effect on quality of
life, work productivity and anxiety as well as adverse reactions. A doseresponse relationship
was investigated by varying in the duration of the cold shower.
Methods
Study Design
This parallel group, unblinded, randomized controlled trial was designed following CONSORT
guidelines and took place in The Netherlands, named the Cool Challenge. Between December
7
th
and December 30
th
2014, we recruited participants through advertisements and (social)
media. Inclusion, randomization and data collection were all performed via a web based appli-
cation using surveys only. Written informed consent was obtained from all participants. The
study was designed as a pragmatic trial and compliance to the intervention could not be veri-
fied. The primary aim was to look at any effect of a routine cold shower and the secondary aim
was to look at dose-dependency effects. The study protocol was approved by our institutional
review board based on ethical considerations (September 3
rd
2014, Academic Medical Center,
Amsterdam, The Netherlands). Being exempt from formal medical ethical review as it was con-
sidered non-medical research, this non-clinical trial was not registered in a clinical trial registry
before recruitment of the first participant but on June 25th 2015 prior to data analysis (August
5
th
—September 13
th
2015) with The Netherlands National Trial Register (NTR), approved by
the WHO, number NTR5183. The authors confirm that all ongoing and related trials for this
intervention are registered.
Participants
Participants were adults aged 18–65 without routine experience of (hot-to-) cold showering
who were employed when they entered the study. As no harmful effects of cold showering have
previously been reported, the only exclusion criterion was significant comorbidity, including
cardiac, pulmonary or any other severe disease. Exclusion criteria were primarily self-assessed.
Significant comorbidity was defined by either a subject’s positive answer to the self-assessment
question: “Do you have a severe medical condition to the heart or lungs?” or at investigator’s
judgement of the subject’s self-reported medical conditions. Subjects were asked to answer the
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 2 / 15
CompetingInterests: Bas C.J.M. van der Heijden
was employed by Achmea during the course of the
study. There are no patents, productsin development
or marketed products to declare. This does not alter
our adherence to all the PLOS ONE policies on
sharing data and materials.
question: “Do you have a medical condition?” Careful screening was done for any severe car-
diac, pulmonary or other systemic comorbidity at the investigator’s judgement. After informed
consent, eligible participants were randomized to one of four groups (1:1:1:1). Randomization
was performed using computerized random numbers within a custom-made Hypertext Pre-
processor (PHP) scripted web-based application for online sur veys, without applying block
randomization or stratification methods. The function RAND with PHP programming lan-
guage was used,which assigns a random numberbetween 1 and 4 each time.[8] Allocation
concealment was ensured within the web-based application.
Procedures
Participants randomizedto the intervention groupswere instructed to shower as warm and as
long as preferredbut ending with respectively30, 60 or 90 seconds showering at the coldest
available water temperature. They were instructed to use either the timer provided through a
web link for smart phones by the research team, or a timer of their own. In case they could not
complete the full period, participants were asked to time the period using a stopwatch. The
average ground temperature at the level of running water in The Netherlands was 10°C during
the study period with average cold water temperatures of 10–12°C.[9] The intervention period
was 30 consecutive days from January 1
st
-30
th
2015. During the following 60 days January 31
st
-
March 31
st
2015 participants of all three intervention groups were instructed to shower as pre-
ferred, i.e. taking cold showers as often and as long as preferred. Participants randomized to
the control group were instructed to shower as regular (not cold) during the full 90-day study
period.
Data were primarily collected through an online web-based platform and managed in
Microsoft Excel 15.0 (Microsoft Corporation, Redmond, Washington, USA). In order to pro-
vide self-reports, participants were asked to log in three times: at baseline, between 30 and 60
days, and between 90 and 120 days. Weekly reminder emails were sent to participants who had
not yet completed follow-up. Specific missing data were collected by email.
Outcomes
All outcomes were self-reported using web based surveys. The primary outcome was illness
days and related leave from work during the 90-day study period (January to March 2015).
Sickness absence was considered to be the most objective indirect parameter indicative of ill-
ness severity. Participants were asked to rate the total number of days of absence from their
work due to sickness, if possible by verifying with their employer or their agenda. Absence fre-
quency was notmeasured. Illness days were defined as thetotal amount of days a participant
felt ill (including symptoms of cold and flu). Participants were asked to rate the number of
days that they had “symptoms of illness, cold or flu” during the study period. If participants
rated sickness absence or illness over five days, they were asked for a reason. The secondary
outcomes were timeof subjective sickness,quality of life, work productivity, thermal sensation
and anxiety.
• Qualityof life was assessed using theShort Form 36 (SF-36), a 36-item patient-reported sur-
vey of health.[10] Results were analyzed including the physical component summary (PCS)
and mental component summary (MCS), with higher scores representing better quality of
life. The Dutch 4-week recall version was used, as adapted and validated by Aaronson et al.
[11]
• Work productivitywas assessedusing the Utrecht Work Engagement Scale (UWES), a
9-item survey with a 7-point Likert scale of engagement during work activities (0 = “never”
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 3 / 15
to 6 = “every day”) and the total sum scores (range 0–54) were used for analysis, with higher
scores representing better work productivity. [12–13]
• Mediation of the mind-body neurosensory pathways were assessed as changes in daily ther-
mal sensation of the body and extremities, respectively (expressed as warmer, colder or not
different from their habitual daily sensation before the start of the trial).
• To explore anxiety, we selected the six questions of the subscale scoringanxiety from the
Brief Symptom Inventory.[14] The total anxiety score (range 0–24) of this self-report inven-
tory with a 5-point Likert scaleof distress (0 = "not at all" to 4 = "extremely") wasused for
analysis with lower scores representing less anxiety. The Dutch version was used, as adapted
and validated by De Beurs and Zitman.[15]
• At each follow-up moment, participants were asked to report any positive and negative
effects. Adverse reactions other than influenza or influenza-related symptoms that were pos-
sibly or likely related to (hot-to-)cold showering were recorded by asking participants for any
negative experiences and events as well as reasons to discontinue the intervention.
Statistical Analysis
We calculated that 575 individuals were required to achieve 80% power to detect a difference
of 0.5 days of sickness absence (SD 3.03) during the 90 days period, based on previous data.
[16] The significance level (alpha) of the test was set at P<0.05. Accounting for a 20% lost to
follow-up, we set target enrolment at 720 individuals for each intervention group and the con-
trol group. We collected participants’ characteristics and baseline values for primary and sec-
ondary outcome measures to allow comparisons between groups at baseline (Table 1). Analysis
was conducted using intention to treat principles. For the primary outcome (sickness absence
days and illness days at 90 days follow-up)a negative binomial regression modelwith log link
Table 1. Baseline characteristics, according to study group.
Charachteristics 30s Group (n = 798) 60s Group (n = 727) 90s Group (n = 775) Control Group (n = 718)
Women 473 (59) 423 (58) 466 (60) 399 (56)
Mean (SD) age (years) 39.7 (11.3) 38.9 (10.6) 39.6 (10.6) 39.2 (10.6)
Mean (SD) body mass index (kg/m2) 23.7 (3.4) 23.9 (3.7) 23.6 (3.3) 23.9 (3.4)
Good subjective health 770 (96) 694 (95) 752 (97) 684 (95)
Median (interquartile range) SF-36 physical component
score
84.2 (77.2–89.2) 84.2 (76.2–90.2) 85.2 (77.2–90.4) 84.2 (77.3–90.2)
Median (interquartile range) SF-36mental component score 81.4 (69.8–87.6) 81.3 (69.7–88.2) 81.1 (69.7–88.8) 81.9 (69.6–88.6)
Median (interquartile range) work engagement score 41 (33–46) 41 (32–45) 41 (32–46) 41 (32–46)
Median (interquartile range) anxiety score 1 (3) 1 (3) 1 (3) 1 (3)
Regular physical activity 661 (83) 600 (83) 664 (86) 614 (86)
Fulltime employee 315 (39) 283 (39) 279 (36) 269 (37)
Residence conditions
Single 207 (26) 196 (27) 171 (22) 190 (26)
Living with partner 237 (30) 206 (28) 224 (29) 209 (30)
Living with (partner and) children 354 (44) 325 (45) 380 (49) 319 (44)
This study investigated theeffect of cold showering onhealth and work: a trial randomizing a (hot-to-)cold shower for 30, 60,90 seconds or a control group
during 30 consecutive days followed by 60 days of showering cold at their own discretion for the intervention groups. Values are numbers (percentages)
unless stated otherwise
doi:10.1371/journal.pone.0161749.t001
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 4 / 15
was performed which was preferred over the Poisson model because of over-dispersion in the
count data. The followingparameters were tested by an analysis of deviance:group, age, sex,
body mass index, regular physical activity (y/n), and fulltime employee (y/n). Statistical meth-
ods used 2-sided testing. For secondary outcomes this included Kruskal Wallis or Mann-Whit-
ney U tests for non-normal distributed continuous variables and Chi-squaretests for
categorical variables. The level of significance of theprimary outcome was set at P<0.05. For
secondary outcome variables a Bonferroni adjustment was used to correct for multiple out-
come variable testing, with the level of significance set at P<0.005. Analyses were performed
with SPSS 23 (IBM SPSS, Chicago, USA) and SAS 9.4 (SAS Institute Inc., Cary, USA).
Results
Of the 4229 candidates screened for eligibility 3018 participants were enrolled (Fig 1). Loss to
follow-up was 12% after 30 days, and 19.6% after 90 days. Table 1 shows that baseline charac-
teristics as well as data on primary and secondary outcome measures were similar between the
intervention groups and the control group. Results in text and tables are reported in respective
order of the groups as 30s cold shower, 60s cold shower, 90s cold shower and control group.
For the primary outcome sickness absence the individual cold shower regimes all differed sta-
tistically significant from the hot shower regimen (for 30s, 60s, and 90s: p = 0.014, p = 0.0268,
and p = 0.0065, respectively). Analysis of deviance showed no statistically significant effect
between the three cold shower groups (p = 0.98 for sickness absence, S1 Table). For illness days
only the 60s cold shower regime differedstatistically significant from the hot shower regimen:
(for 30s, 60s, and 90s: p = 0.235, p = 0.014, and p = 0.383,respectively). Analysis of deviance
showed no statistically significant effect between the three cold shower groups (p = 0.15 analy-
sis of deviance, S2 Table). There were no trends between doses towards illness or absenteeism
benefit.
Seventy-nine percent of participants in the interventions groups completed the 30 consecu-
tive days protocol (82% vs 79% vs 79%; P = 0.14) and 64% continued the (hot-to-) cold shower
on regular basis (66% vs 63% vs 62%; P = 0.36). A negative binomial regression model showed
a 29% reduction in sicknessabsence for the (hot-to-) cold shower regimen compared to the
control group (Incidence Rate Ratio (IRR): 0.71, P = 0.003). No significant difference between
the intervention groups (P = 0.992) was observed, therefore parameter group was transformed
into a factor with two levels, all intervention groups versus control group (Table 2). The only
associated parameter of influence in the model was regular physical activity (IRR: 0.65,
P = 0.003), which reduces the sickness absence by 35%. For illness days at 90 days follow-up
there was no significant group effect, only a gender effect, with males showing a 14% reduction
compared to females (IRR: 0.86, P = 0.010).
Median quality of life MCS after 30 days was slightly higher for all intervention groups
(84.7, interquartile range 76.4–90.2 v85.1, interquartile range 76.7–90.6 v85.7, interquartile
range 78–90.8) compared to the control group (83.9, interquartile range 72.9–89.4)(Table 3).
However, after 90 days significant differences were not observed anymore (Table 4). None of
the other secondary outcomes were significantly different between groups at 30 and 90 days
follow-up (Tables 3and 4).
Twenty serious adverse events were reported, that were all considered unrelated to the inter-
vention. One participant in the 90 seconds intervention group died unexpectedly of occult
chronic pulmonary embolism at 56 days follow-up. This occult condition was not diagnosed at
the time of enrolment and her medical history included hypertension only. Critical assessment
by the team of treating (intensive care) physicians showed no possible relationship to the (hot-
to-)cold shower. There were eight participants with a mild pneumonia, two urinary tract
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 5 / 15
infections, two had middleear infections, one pneumothorax, one glaucoma, two hand
wounds, one with multiple rib contusions after a fall, one with concussionand head wound
after a fall, one bike and one ski-accident both with multiple minor contusions, distorsions and
lacerations. No related serious adverse events were reported. The most common related mild
Fig 1. Study flow diagram.
doi:10.1371/journal.pone.0161749.g001
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 6 / 15
adverse event was a persistent cold sensation after the cold shower in the body (196 participants)
as well as in the hands and/or feet (257 participants), specificallyin 3 participants with Raynaud’s
phenomenon. Other possibly related adverse events included muscle ache or cramps in eight, itch
in six, insomnia in four (related to cold shower in the evening), dizzinessin four, lumbago in two,
head ache in one, nose bleedingin one, diarrhea in one, palpitations in one and transient swelling
and erythema of three digits of one hand in one participant after the cold shower.
Discussion
In this pragmatic randomized controlled trial, routinely showering (hot-to-) cold resulted in a
29% reduction of self-reported sick leave from work but not illness days at 90 days follow-up in
adults without severe comorbidity. The contrast between the results of both primary outcome
parameters is suggestiveof the fact that the intensity rather than the durationof symptoms is
modulated by the intervention. Regular physical activity resulted in a 35% reduction of sickness
absence. The combination of routine (hot-to-) cold shower and regular physical activity
resulted in an expected 54% reduction of sickness absence compared to people who don’t do
either. The duration of the cold shower did not influence outcome as there was no significant
difference between intervention groups. The only secondary outcome that showed a slight ben-
eficial effect–on the short run–was quality of life (mental component summary) although this
was deemed too small to be clinically relevant. Even though the vast majority of participants
reported a variabledegree of discomfortduring cold exposure,the fact that 91% of participants
reported the will to continue such routine (and 64% actually did) is perhaps the most indicative
of any health or work benefit. The most commonly reported beneficial effect was an increase in
perceived energy levels (including many reported comparisons to the effect of caffeine). The
most common discomfortable related adverse reaction was persistent cold sensation in body,
hands and/or feet in up to 13% of participants. Other related harmful effects were mild and
uncommon.
Table 2. Negative binomial regression model of the primary outcome.
Outcome Median (interquartile
range) per group Range
[Min, Max] Percentage with any
sickness respectively
illness
Parameter Maximum
Likelhood Estimate
(95% CI)
Exponential
Estimate (95% CI) P
value
90 days
sickness
absence
30s Group: 0 (0–1) [0, 62] 29,4% Intercept 0.80 (0.49, 1.11) 2.23 (1.63, 3.03) <
.0001
60s Group: 0 (0–1) [0,29] 34,0% Group (inter vention
groups vs control)*
-0.35 (-0.58, -0.12) 0.71 (0.56, 0.89) 0.003
90s Group: 0 (0–1,5) [0,40] 33,1% Regular physical
activity (yes vs no)*
-0.42 (-0.70, -0.15) 0.65 (0.5, 0.86) 0.003
Control Group: 0 (0–2) [0,51] 34,8% Dispersion 4.64 (4.17, 5.15)
90 days illness 30s Group: 2 (0–7) [0,56] 65,0% Intercept 1.27 (1.14, 1.39) 3.55 (3.13, 4.02) <
.0001
60s Group: 2 (0–6) [0,60] 63,3% Group (inter vention
groups vs control)*
-0.12 (-0.26, 0.01) 0.89 (0.77, 1.01) 0.073
90s Group: 2 (0–6) [0,70] 64,5% Gender (Male vs
Female)*
-0.15 (-0.27, -0.04) 0.86 (0.76, 0.96) 0.0097
Control Group: 2 (0–7) [0,90] 69,3% Dispersion 1.53 (1.41, 1.66)
This study investigated theeffect of cold showering onhealth and work: a trial randomizing a (hot-to-)cold shower for 30, 60,90 seconds or a control group
during 30 consecutive days followed by 60 days of showering cold at their own discretion for the intervention groups.
*The exponential of the estimates are Incident Rate Ratios (IRR)
doi:10.1371/journal.pone.0161749.t002
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 7 / 15
Table 3. Secondary outcomes at 30 days. Values are numbers (percentages) unless stated otherwise.
Outcomes 30s Group
(n = 700)
60s Group
(n = 660)
90s Group
(n = 680)
Control Group
(n = 615)
Group difference P
value#
Intervention/control differenceP
value##
Median
(interquartile
range) sickness
absence (days)
0 (0) 0 (0) 0 (0) 0 (0) 0.544 0.648
Median
(interquartile
range) illness
(days)
0 (0–3) 0 (0–3) 0 (0–3) 1 (0–4) 0.232 0.047
Completed
(hot-to) cold
shower protocol
during first 30
days*
573 (82) 513 (79) 530 (79) N.A. 0.138
Will to continue
(hot-to) cold
shower after
first 30 days**
634 (93) 571 (89) 609 (91) N.A. 0.024
Median
(interquartile
range) SF-36
physical
component
score
86.2 (78.8–91.4) 87.2 (80.5–91.2) 87.2 (79.8–91.4) 85.4 (77.8–90.4) 0.017 0.006
Median
(interquartile
range) SF-36
mental
component
score
84.7 (76.4–90.2) 85.1 (76.7–90.6) 85.7 (78–90.8) 83.9 (72.9–89.4) 0.003 0.001
Median
(interquartile
range) work
engagement
score
42 (33–46) 42 (33–46) 42 (34–47) 40 (32–46) 0.108 0.020
Median
(interquartile
range) anxiety
score
1 (0–3) 1 (0–3) 1 (0–3) 1 (0–3) 0.003 0.001
Thermal body
sensation**
0.160
Warmer 262 (39) 265 (41) 269 (40) N.A.
Colder 55 (8) 72 (11) 69 (10) N.A.
No
difference
363 (53) 304 (48) 333 (50) N.A.
Thermal hands
and feet
sensation**
0.778
Warmer 179 (26) 170 (26) 180 (27) N.A.
Colder 79 (12) 88 (14) 90 (13) N.A.
No
difference
422 (62) 383 (60) 401 (60) N.A.
This study investigated theeffect of cold showering onhealth and work: a trial randomizing a (hot-to-)cold shower for 30, 60,90 seconds or a control group
during 30 consecutive days followed by 60 days of showering cold at their own discretion for the intervention groups.
N.A. Not applicable
*Missing data in 5, 8, 7 participants (respectively)
** Missing data in 20, 19, and 9 participants (respectively)
# Difference between all groups (Kruskal Wallis)
## Difference between all interventional groups versus control group (Mann-Whitney U)
doi:10.1371/journal.pone.0161749.t003
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 8 / 15
Influenza was the most common reason for participants’ absenteeism durations longer than
five days. This study was performed during the 2014/2015 influenza epidemic in The Nether-
lands, which lastinga total of 21 weeks had the longest duration since more than 40 years.[17]
The influenza-like illness incidence (consulting a general practitioner) was consistently above
10 per 10,000 inhabitants duringthe study period, with 5 per 10,000 inhabitants representing
the threshold for a mild epidemic. An epidemic is defined as an incidence above this threshold
for at least two consecutive weeks. The actual incidence of influenza cases was considerably
higher, because only a proportion of the patients with influenza-like symptoms consulted the
general practitioner. At the start of the season, influenza virus A(H3N2) dominated, while later
in the season, influenza virus B was most prevalent. A part of the circulating influenza A-
viruses appeared to mismatch with the influenza A-strain in the vaccine. Other prevalent
viruses during the study period included the respiratory syncytial virus (RSV), the enterovirus
and the rhinovirus.
Table 4. Primary and secondary outcomes at 90 days. Values are numbers (percentages) unless stated otherwise.
Outcomes 30s Group
(n = 673) 60s Group
(n = 611) 90s Group
(n = 595) Control Group
(n = 547) Group difference
P value# Intervention/control
difference P value##
Continued (hot-to) cold shower after
first 30 days*
446 (66) 378 (63) 363 (62) N.A. 0.355 N.A.
Median (interquartile range) frequency
of cold shower (times per week)*
3 (0–7) 3 (0–7) 2 (0–6) N.A. 0.727 N.A.
Median (interquartile range) duration
of cold shower (s)*
30 (10–50) 60 (40–80) 60 (10–110) N.A. <0.001 N.A.
Will to continue (hot-to) cold shower
after 90 days**
546 (88) 487 (84) 490 (85) N.A. 0.199 N.A.
Median (interquartile range) SF-36
physical component score
85.8 (78.9–
90.4)
86.4 (79.4–
92)
87.2 (79.8–
92)
86.4 (78.5–
91.4)
0.121 0.338
Median (interquartile range) SF-36
mental component score
84.8 (76.7–
89.6)
84.4 (75.7–
90.2)
85.8 (78.0–
90.6)
84.4 (74.3–90) 0.108 0.090
Median (interquartile range) work
engagement score
41 (33–46) 42 (32–46) 42 (32–46) 41 (31.3–46) 0.638 0.389
Median (interquartile range) anxiety
score
1 (0–3) 1 (0–3) 1 (0–3) 1 (0–3) 0.190 0.133
Reason of sickness absence if longer
than 5 days***
0.326
Influenza 27 (64) 17 (46) 13 (42) 20 (51)
Psychosocial (including burnout) 6 (14) 7 (19) 6 (19) 5 (13)
Musculoskeletal Injury 4 (10) 4 (11) 2 (6) 3 (8)
Bronchitis/pneumonia 3 (7) 0 (0) 3 (10) 2 (5)
Other upper respiratory tract
infection (excluding influenza)
2 (5) 0 (0) 2 (6) 2 (5)
Other infection(s) 0 (0) 5 (14) 1 (3) 3 (8)
Other comorbidity (including
operation)
0 (0) 4 (11) 4 (13) 4 (10)
This study investigated theeffect of cold showering onhealth and work: a trial randomizing a (hot-to-)cold shower for 30, 60,90 seconds or a control group
during 30 consecutive days followed by 60 days of showering cold at their own discretion for the intervention groups.
N.A. Not applicable
*Missing data in 2 participantsin 30s Group, and 2 participantsin 60s Group
** Missing data in 56, 30, and 20 participants (respectively)
*** Data collected in 42, 37, 31, and 39 participants (respectively)
# Difference between all groups (Kruskal Wallis)
## Difference between all interventional groups versus control group (Mann-Whitney U)
doi:10.1371/journal.pone.0161749.t004
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 9 / 15
We searched PubMed, Web of Science, the Cochrane Database of Systematic Reviews,the
Cochrane CentralRegister of ControlledTrials, and Database of Abstracts of Reviews on
Effects for articles published between Jan 1, 1980, and Oct 1, 2015 on the effect of any type of
cold bathing on health. We used the broad MeSH term “Cold Temperature” in combination
with the terms “bath” or “shower”. Our search resulted in no randomized controlled trials
that assessed health. One Cochrane review investigated the effect of cold-water immersion for
preventing and treating muscle soreness after exercise[18] and found some evidence that cold-
water immersion reduces delayed onset muscle soreness after exercise. The three randomized
controlled trials were limited to the subject of cold bathing on athletic performance[19,20] and
physiological response.[3] Positive habituation effects on the physiological response and slight
beneficial outcomes on athletic recovery have been reported. However, there is a lack of data
regarding any cumulative clinical effect and relevance for health.
Cold water has been used therapeutically for many centuries and continues using modern
technology. Hippocrates, the father of medicine, who addedrubbing to cold bathing, was
accustomed to use cold water in his treatment of the most serious illnesses.[21] Although most
cold exposure studies involved cold water immersion, different methods of cold water therapy
such as cold bathing and cold showering are used interchangeably and seem to have similar
effects.[19] The latest form of cold therapy (or stimulation) is called whole-body cryotherapy
and consists of exposureto very cold air that is maintained at -110°C to -140°Cin special tem-
perature-controlled cryochambers,generally for 2–3 minutes. It was initially proposed for the
treatment of rheumatic diseases[22] but is increasingly popularized among athletes for its sup-
posedly beneficial effect on recovery and performance, even though it has not been confirmed
in a recent systematic review.[18,23–24]
In The Netherlands, there has been an increasing trend for cold bathing over the past few
years. Part of this growingpopularity is owed to the scientificapproach of a health and mindset
technique hallmarked by cold-exposureas created by an individual namedWim Hof, nick-
named the Iceman for his ability to remain constant body temperature in extreme cold condi-
tions.[25] These methods involving concentration, breathing and cold-exposure have shown to
modulate the immune response.[26] These findings served as inspiration to design the present
trial and its popularity facilitated recruitment of over three thousand participants in just one
month time.
The mechanism or explanatory pathway of any therapeutic effects of cold exposure remains
unclear. In the acute phase (during shivering) increases of cortisol and norepinephrine concen-
trations have been reported but resulted in minimal or no immune modulation.[4–7] More-
over, both immune-stimulatory and immune-inhibiting effects of cold exposure during
exercise increase controversy.[27] Data obtained mainly on small mammals suggests that cold
exposure suppresses several cellular and humoral components of the immune response but
adaptation to a given cold stimulus appears to develop over the course of 2–3 weeks.[28] Beta-
endorphin increase has been reported after cold exposure in rats and cold stress-induced mod-
ulation of cell immunity has been reported during acute Toxoplasma gondii infection in mice.
[29–30] However, these findings couldnot be reproduced in one study in humans.[31] The
current study adds data on cold adaptation following repeated cold exposure. Longterm hor-
monal and cytokineeffects of such modulationare relatively small and its significance remains
unclear as only the early steps of the immune cascade appear to be affected.[6] The fact that
there was no differencebetween 30, 60 or 90 secondsof cold showeringis consistent with pre-
vious research on the habituation of the initial responses to cold water immersion.The greatest
physiological response to cold water exposure was observed during the first 30 seconds and the
rapidity suggests that it is initiated by neurogenic pathways rather than circulating hormones.
[32]
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 10 / 15
Another physiological explanatory mechanism is the improvement of fitness level when
considering the routine cold shower as frequent engaging physical activity. In the present trial,
reduction of sickness absence of a routine cold shower (29%) was comparable to the effect of
regular physical activity (35%). A previous study in The Netherlands showed that the mean
total duration of absenteeism was 15% lower in cyclists than in non-cyclists. Cycling to work
was therefore associated with less sickness absence.[16] The more often people cycled to work
and the longer the distance travelled, the less they reported sick. This is consistent with the
findings of Nieman et al. who have shown in several studies that there is an inverse relationship
between physical activity or fitness level and the rates of upper respiratory tract infection.[33]
Recently, a meta-analysis of four randomized controlled trials determined the effects of exer-
cise on prevention of the common cold. The effect of exercise on the prevention of the com-
mon cold had a relative risk reduction of 27% and there was a mean reduction of 3.5 illness
days compared to controls.[34]
In addition, there are multiple psychological explanatory mechanisms such as expectancies
which play a major role for the treatment outcome of a broad variety of immune-mediated
conditions.[35] The outcome expectations of the present study billed as testing the hypothesis
whether “cold-showers-might-decrease-illness-and-improve-health” could potentially play a
suggestive role in the actual outcomes such as the decisionto go to work when feeling ill.The
promotion material included more positive than negative general claims to be explored such as
“Habitual cold exposure has been claimed to have positive influences including improvements
of the immune system, circulation, emotional state, skin conditions, and energy. The aim of
this study is to investigate whether such claims are true.” Other than the statement of these
unsupported claims, promotional material did not suggest that cold showers might reduce ill-
ness or absenteeism.Prior to the start of the trial,participants were informed of several out-
come parameters including vitality, energy levels, work productivity and sickness absence.
They were intentionally not informed of primary and secondary outcomes. Participants were
fully aware of the four different groups. Other communication forms such as consent form and
emails were nonsuggestive. The contrast between the results of both primary endpoints could
also suggest that the intervention made participants more resilient to absenteeism with compa-
rable intensity and duration of illness symptoms.
Our data cannot determine whether the present findings were causal or associational. More-
over, participants in this trial couldnot be blinded for theintervention nor for theirown out-
come assessment, hence potentially introducing important bias. Specifically, a placebo-effect of
this intervention cannot be ruled out. However, if such effect was causative in this trial, it
should not beconsidered as an effect of an “inert substance”.[36] Placebo effects rely on com-
plex neurobiologicpathways involving neurotransmitters such as norepinephrine and activa-
tion of specific, quantifiable, and relevant areas of the brain.[37] With the recent discovery of
the central nervous system lymphatic system represent, a neurobiologicimmunostimulatory
effect should not be ruled out.[38]
The findings of this study should be interpreted while accounting for its limitations. First,
all outcomes were self-reportedbased on our online survey design. Hence, none of our parame-
ters could be objectified. Our primary outcome sickness absence was selected for its closest
proximity of an objective parameter. Second, according to the SF-36 data, the study population
is extremely healthy compared the general Dutch population. This is most likely a correct mea-
surement due to an important sampling bias: (1) all patients with severe comorbidity were
excluded; (2) the athletic / strenuous character of the study attracted a highly motivated,
healthy and physically active group with SF-36 averages much higher than the population
norm though lower than a competitive athletic population;[39] (3) 96% of participants rated
their health to be good or excellent; (4) the prevalence of participation in sports (85%) in the
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 11 / 15
study population was higher than the average norm (53% in the national population of 12
years and above);[40] and (5) average sickness absence in thecontrol group (1,55%) was almost
a third compared to the average sickness absence in the Dutch population (4,4% during the
first quartile of 2015 corresponding to the study period).[41] Third, attrition bias could be
introduced becauseof the large loss to follow-up (20%), which is likely due to the online-only
interface of the study. Non-responders could only be contacted by the provided email address.
In the intervention groups, large numbers of participants discontinued the intervention
because of its burden or a sickness making them choose for their preferred routine. In contrast,
in the control group discontinuers were much fewer as the control group instructions to
shower as regular did not cause any burden or preference to discontinue due to sickness.
Fourth, inherent to the pragmatic design, compliance to the intervention was not verifiable.
Participants were asked to record if and how long they continued the intervention. Duringthe
first 30 days of the trial the median recorded time of the interventionwas equal to the
instructed timefor each group, which issuggestive of valid measuresfor dividing the groups.
During the last 60 days of the trial the median recorded time for the intervention was different
only in the 90s cold shower group (median 60s, interquartile range 10–110). In our opinion
there was no incentiveto report false data consideringthe anonymous character of data analy-
sis. Such bias would likely have a tendency to reduce any effect on health and work because of
the limited compliance ranging from 64 to 79 percent of participants during the study period.
Fifth, the relatively short follow-up period and the very healthy character of the study group
resulted in the fact that most participants did not have any sickness absence days at all. Sixth,
there was a variationof temperature of the coldest available shower water accordingto loca-
tion. However, less cold temperatures would underestimate the effect of the intervention.
The main strengths of this trialinclude its innovativity, the large number of participants,
the randomization to four groups and the pragmatic approach in a domestic setting. It was
designed as a straightforward study looking at cumulative subjective effect after a routine
behavioral intervention in daily life and significant relevance in terms of effect size. This prag-
matic randomized controlled trial is the first study showing that a routine cold shower has a
beneficial effect on health.
Repetitive cold showering can modulate the physiological response.[3] Our findings show
that routinely showering (hot-to-) cold for at least 30 days resulted in a reduction of self-
reported sick leave from work but not illness days in adults without severe comorbidity. Fur-
ther research using objectiveparameters is necessary to determine whether thesewere causal
or associational findings. Considering the mild effect of a routine cold shower on hormonal
and cytokine modulation, these alone are unlikely to play a significant role.[4–7] Perhaps neu-
roimaging technologiessuch as functional MRI could be used to assess any potential neurobio-
logic immunostimulatory effect.
Supporting Information
S1 Checklist. CONSORT 2010 Checklist.
(DOC)
S1 Protocol. Protocol version 1 for Institutional Review Board–Dutch version.
(PDF)
S2 Protocol. Protocol version 1 for Institutional Review Board–English version.
(PDF)
S3 Protocol. Study Protocol COOL Challenge—definitiveversion.
(DOCX)
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 12 / 15
S4 Protocol. Institutional Review Board approval (including English translation).
(PDF)
S1 Table. Final model output from SAS PROC GENMOD for sickness absence.
(DOCX)
S2 Table. Final model output from SAS PROC GENMOD for illness.
(DOCX)
Acknowledgments
We would like to thank Koen de Jong, Linda Koeman, Jan Zandberg, Bram Bakker, Klaas Kroe-
zen, Anna Chojnacka,and Wim Hof for their inspiration and contributionto this study. We
thank the periodical Quest1for supporting the recruitment of participants.
Author Contributions
Conceptualization:GAB MGD MHWF.
Data curation: GAB INS BCJMH MGD.
Formal analysis: GAB INS BCJMH MGD MHWF.
Funding acquisition: GAB MHWF.
Investigation: GAB.
Methodology: GAB MGD MHWF.
Project administration: GAB MHWF.
Resources: GAB.
Software: GAB INS BCJMH MGD.
Supervision: INS BCJMH MGD MHWF.
Validation: GAB INS BCJMH MGD.
Visualization: GAB INS BCJMH.
Writing – original draft: GAB.
Writing – review & editing: GAB INS BCJMH MGD MHWF.
References
1. DeRose Evans J. A Companion to the Archaeology of the RomanRepublic. 2013, Hoboken: Wiley-
Blackwell.
2. Cold Showers. The definitive guide to cold water therapy. http://www.cold-showers.com/
3. Eglin CM, Tipton MJ. Repeated cold showers as a method of habituating humans to the initial
responses to cold water immersion. Eur J Appl Physiol 2005; 93:624–9. PMID: 15778892
4. Brenner IK, Castellani JW, Gabaree C, Young AJ, Zamecnik J, Shephard RJ, et al. Immune changes in
humans during cold exposure: effects of prior heating and exercise. J Appl Physiol 1999; 87:699–710.
PMID: 10444630
5. Castellani JW, Brenner IKM, Rhind SG. Cold exposure: human immune responses and intracellular
cytokine expression. Med Sci Sports Med 2002; 34:2013–2020.
6. Janský L, Pospísilová D, Honzová S, Ulicný B, Srámek P, Zeman V, et al. Immune system of cold-
exposed and cold-adapted humans. Eur J Appl Physiol 1996; 72:445–450.
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 13 / 15
7. Srámek P, Simecková M, Janský L, Savlíková J, Vybíral S. Human physiological responses to immer-
sion into water of different temperatures. Eur J Appl Physiol 2000; 81:436–442. PMID: 10751106
8. PHP manual. http://php.net/manual/en/function.rand.php
9. Platform Geothermie. http://www.geothermie.nl/geothermie
10. Ware JE, Sherbour neCD. The MOS 36-item shor t-formhealth sur vey (SF-36). I. Conceptual frame-
work and item selection. Med Care 1992; 30:473–83. PMID: 1593914
11. Aaronson NK, Muller M, Cohen PD, Essink-Bot ML, Fekkes M, Sanderman R, et al. Translation, valida-
tion, and norming of the Dutch language version of the SF-36 Health Survey in community and chronic
disease populations. J Clin Epidemiol 1998; 51:1055–68. PMID: 9817123
12. Schaufeli WB, Bakker AB. Bevlogenheid: Een begrip gemeten [Work engagement: The measurement
of a concept]. Gedrag & Organisatie 2004; 17:89–112.
13. Demerouti E, Bakker AB, Janssen PPM, Schaufeli WB. Burnout and engagement at work as a function
of demands and control. Scand J Work Environ Health 2001; 27:279–286. PMID: 11560342
14. Derogatis L, Melisaratos N. The Brief Symptom Inventory: An introductory repor t.Psychol Med 1983;
13:595–605. PMID: 6622612
15. Beurs E de, Zitman F. De Brief Symptom Inventory (BSI). De betrouwbaarheid en validiteit van een
handzaam alternatief voor de SCL-90. Maandblad Geestelijke volksgezondheid 2006; 61:120–41.
16. Hendriksen IJ, Simons M, Garre FG, Hildebrandt VH. The association between commuter cycling and
sickness absence. Prev Med. 2010; 51:132–5. doi: 10.1016/j.ypmed.2010.05.007 PMID: 20580736
17. Teirlinck AC, van Asten L, Brandsema PS, Dijkstra F, Donker GA, Euser SM. Surveillance of influenza
and other respiratory infections in the Netherlands: winter 2014/2015. Report of the National Institute
for Public Health and the Environment (RIVM). 2015:1–85.
18. Costello JT, Baker PR, Minett GM, Bieuzen F, Stewart IB, Bleakley C. Whole-body cryotherapy
(extreme cold air exposure) for preventing and treating muscle soreness after exercise in adults.
Cochrane Database Syst Rev. 2015; 9:CD010789. doi: 10.1002/14651858.CD010789.pub2 PMID:
26383887
19. Juliff LE, Halson SL, Bonetti DL, Versey NG, Driller MW, Peiffer JJ. Influence of contrast shower and
water immersion on recovery in elite netballers. J Strength Cond Res. 2014; 28:2353–8. doi: 10.1519/
JSC.0000000000000417 PMID: 24531433
20. Crowe MJ, O'Connor D, Rudd D. Cold water recovery reduces anaerobic performance. Int J Spor ts
Med. 2007; 28(12):994–8. PMID: 17534786
21. Claridge RT. Hydropathy or The Cold Water Cure, as practiced by Vincent Pr iessnitz, at Graefenberg,
Silesia, Austria. ( 8th ed.) 1843; London: James Madden and Co.
22. Jastrzabek R, Straburzyńska-Lupa A, Rutkowski R, Romanowski W. Effects of different local cryothera-
pies on systemic levels of TNF-α, IL-6, and clinical parameters in active rheumatoid arthritis. Rheumatol
Int 2013; 33:2053–60. doi: 10.1007/s00296-013-2692-5 PMID: 23397259
23. Banfi G, Lombardi G, Colombini A, Melegati G. Whole-body cryotherapy in athletes. Sports Med. 2010;
40:509–17. doi: 10.2165/11531940-000000000-00000 PMID: 20524715
24. Bleakley CM, Bieuzen F, Davison GW, Costello JT. Whole-body cryotherapy: empirical evidence and
theoretical perspectives. Open Access J Spor ts Med. 2014; 5:25–36. doi: 10.2147/OAJSM.S41655
PMID: 24648779
25. Kox M, Stoffels M, Smeekens SP, van Alfen N, Gomes M, Eijsvogels TM, et al. The influence of con-
centration/meditationon autonomic nervous system activity and the innate immune response: a case
study. Psychosom Med. 2012; 74:489–94. doi: 10.1097/PSY.0b013e3182583c6d PMID: 22685240
26. Kox M, van Eijk LT, Zwaag J, van den Wildenberg J, Sweep FC, van der Hoeven JG, et al. Voluntary
activation of the sympathetic ner vous system and attenuation of the innate immune response in
humans. Proc Natl Acad Sci U S A. 2014; 111:7379–84. doi: 10.1073/pnas.1322174111 PMID:
24799686
27. Gagnon DD, Gagnon SS, Rintamäki H, Törmäkangas T, Puukka K, Herzig KH, et al. The effects of cold
exposure on leukocytes, hormones and cytokines during acute exercise in humans. PLoS One. 2014;
22;9:e110774.
28. Shephard RJ, Shek PN. Cold exposure and immune function.Can J Physiol Pharmacol. 1998; 76:828–
36. PMID: 10066131
29. Banerjee SK, Aviles H, Fox MT, Monroy FP. Cold stress-induced modulation of cell immunity during
acute Toxoplasma gondii infection in mice. J Parasitol. 1999; 85:442–7. PMID: 10386435
30. Vaswani KK, Richard CW 3rd, Tejwani GA. Cold swim stress induced changes in the levels of opioid
peptides in the rat CNS and peripheral tissues. Phar macol Biochem Behav 1988; 29:163–8. PMID:
3353422
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 14 / 15
31. Leppäluoto J, Westerlund T, Huttunen P, Oksa J, Smolander J, Dugué B, et al. Effects of long-term
whole-body cold exposures on plasma concentrations of ACTH, beta-endorphin, cortisol, catechol-
amines and cytokines in healthy females. Scand J Clin Lab Invest 2008; 68:145–53. doi: 10.1080/
00365510701516350 PMID: 18382932
32. Tipton MJ, Eglin CM, Golden FS. Habituation of the initialresponses to cold water immersionin
humans: a centralor peripheral mechanism? J Physiol 1998; 512:621–8. PMID:9763650
33. Nieman DC, Henson DA, Austin MD, Sha W. Upper respiratory tract infection is reduced in physically fit
and active adults. Br J Sports Med 2011; 45:987–92. doi: 10.1136/bjsm.2010.077875 PMID: 21041243
34. Lee Hyun Kun, Hwang In Hong, Kim Soo Young, Pyo Se Young. The Effect of Exercise on Prevention
of the Common Cold: A Meta-Analysis of Randomized Controlled Trial Studies. Korean J Fam Med
2014; 35:119–126. doi: 10.4082/kjfm.2014.35.3.119 PMID: 24921030
35. van Middendorp H, Kox M, Pickkers P, Evers AW. The role of outcome expectancies for a training pro-
gram consisting of meditation, breathing exercises, and cold exposure on the response to endotoxin
administration: a proof-of-principle study. Clin Rheumatol 2016; 35:1081–5. doi: 10.1007/s10067-015-
3009-8 PMID: 26194270
36. Kaptchuk TJ, Miller FG. Placebo Effects in Medicine. N Engl J Med. 2015; 373:8–9. doi: 10.1056/
NEJMp1504023 PMID: 26132938
37. Finniss DG, Kaptchuk TJ, Miller F, Benedetti F. Biological, clinical, and ethical advances of placebo
effects. Lancet 2010; 375:686–695. doi: 10.1016/S0140-6736(09)61706-2 PMID: 20171404
38. Louveau A, Smirnov I,Keyes TJ, et al. Structural and functional features of central nervoussystem lym-
phatic vessels. Nature 2015; 523:337–41. doi: 10.1038/nature14432 PMID: 26030524
39. Huffman GR, Park J, Roser-Jones C, Sennett BJ, Yagnik G, Webner D. Normative SF-36 values in
competing NCAA intercollegiate athletes differ from values in the general population. J Bone Joint Surg
Am. 2008; 90:471–6. doi: 10.2106/JBJS.G.00325 PMID: 18310695
40. Tiessen-Raaphorst A. Rapportage Sport 2014. Report of the Netherlands Social and Cultural Planning
Bureau 2015:1–316.
41. Centraal Bureau voor de Statistiek (CBS), March 31st 2016. National statistics on sickness absence.
[Nationale Verzuim Statistiek]. Statline, Statistics Netherlands. Available at: http://statline.cbs.nl/
StatWeb/publication/?VW=T&DM=SLNL&PA=80072NED&D1=0&D2=0-6,11-14,18,20-25,27-29,34-
35&D3=4,9,(l-6)-l&HD=120703-1034&HDR=T,G2&S TB=G1
The Effect of Cold Showering
PLOS ONE | DOI:10.1371/journal.pone.0161749 September 15, 2016 15 / 15
