The Author (2005). Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.

org
doi:10.1093/fampra/cmi072

Acute cooling of the feet and the onset of

common cold symptoms
Claire Johnson and Ronald Eccles

Johnson C and Eccles R. Acute cooling of the feet and the onset of common cold symptoms.
Family Practice 2005; 22: 608613.
Background. There is a common folklore that chilling of the body surface causes the develop-
ment of common cold symptoms, but previous clinical research has failed to demonstrate any
effect of cold exposure on susceptibility to infection with common cold viruses.
Objective. This study will test the hypothesis that acute cooling of the feet causes the onset of
common cold symptoms.
Methods. 180 healthy subjects were randomized to receive either a foot chill or control pro-
cedure. All subjects were asked to score common cold symptoms, before and immediately after

Downloaded from http://fampra.oxfordjournals.org/ by guest on April 8, 2014

the procedures, and twice a day for 4/5 days.
Results. 13/90 subjects who were chilled reported they were suffering from a cold in the 4/5
days after the procedure compared to 5/90 control subjects (P = 0.047). There was no evidence
that chilling caused any acute change in symptom scores (P = 0.62). Mean total symptom score
for days 14 following chilling was 5.16 (5.63 s.d. n = 87) compared to a score of 2.89 (3.39 s.d.
n = 88) in the control group (P = 0.013). The subjects who reported that they developed a cold
(n = 18) reported that they suffered from significantly more colds each year (P = 0.007) compared
to those subjects who did not develop a cold (n = 162).
Conclusion. Acute chilling of the feet causes the onset of common cold symptoms in around
10% of subjects who are chilled. Further studies are needed to determine the relationship of
symptom generation to any respiratory infection.
Keywords. Cold exposure, common cold, infection, nose.

Introduction However, studies involving inoculation of cold

The common cold is a mild self-limiting illness usually
confined to the upper respiratory tract.1 The disease is
self-diagnosed from a range of symptoms such as nasal
stuffiness, sneezing, throat irritation and mild fever.2
There is a common folklore that associates the develop-
ment of symptoms of common cold with exposure to a
cold environment, and that the onset of a cold is a direct
result of wet clothes, feet and hair.3 Throughout the
clinical literature of the last three hundred years
there have been many reports that acute cooling of
the body surface causes the onset of symptoms of com-
mon cold, and historically it has been generally accep-
ted that acute exposure to cold is a direct cause of these
symptoms.4,5
Received 8 November 2004; Accepted 31 May 2005.
Common Cold Centre, Cardiff School of Biosciences, Cardiff
University, Cardiff CF10 3US, UK. Correspondence to
Professor Ronald Eccles, Common Cold Centre, Cardiff
School of Biosciences, Cardiff University, Cardiff CF10
3US, UK; Email: eccles@cardiff.ac.uk
viruses into the nose and periods of cold exposure
have failed to demonstrate any effect of cold exposure
on susceptibility to infection with common cold
viruses.68 Although modern textbooks of virology dis-
miss any cause-and-effect relationship between cold
exposure and common cold as erroneous folklore,9 the
belief is so widespread and longstanding it is difficult
to completely dismiss this idea as having no validity.
In 1919 Mudd and Grant studied the reactions of the
nasal mucosa in response to chilling the body surface
and showed that cooling the body surface causes a
reflex vasoconstriction of blood vessels in the nose
and a decrease in temperature of the mucous mem-
brane.10 They speculated that this reflex vasoconstric-
tion of the airway epithelium could decrease resistance
to infection and allow bacterial infection of the ton-
sils.10 Some years later Sir Christopher Andrewes sug-
gested that exposure to a cold environment may trigger
the development of a cold but only in people who are
carrying the latent cold virus.6 Eccles developed
these early observations by proposing a hypothesis
that acute cooling of the body surface causes a reflex

608
 Acute cooling of the feet and the onset of common cold symptoms
vasoconstriction in the nose and upper airways, and this
vasoconstrictor response may inhibit respiratory
defence and cause the onset of common cold symptoms
by converting an asymptomatic viral infection (sub-
clinical infection) into a symptomatic viral infection
(clinical infection).11 The novel idea in this hypothesis
was that when common cold viruses are circulating in
the community a proportion of those infected will have
sub-clinical infections, and that when any of this sub-
group are exposed to chilling of the body surface this
could aid conversion of a sub-clinical infection to a clin-
ical infection. This study was aimed at testing this hypo-
thesis, by studying the onset of common cold symptoms
after acute chilling of healthy asymptomatic subjects,
during the winter, when common cold viruses are cir-
culating in the community.
The aims of the study were to determine if acute chil-
ling caused: acute onset of common cold symptoms
within minutes of chilling; delayed onset of common
cold symptoms over a 4/5 days period after chilling;
the perception that the subjects were suffering from
a common cold over a 4/5 days period after chilling.
The study also aimed to investigate any relationship
between the history of colds incidence in the previous
year and the onset of common cold symptoms.
Methods
Subject population
180 healthy subjects were recruited from the student
population of Cardiff University. All subjects attended
the Common Cold Centre, Cardiff. All procedures were
carried out under standard conditions at a room tem-
perature of eighteen to twenty-five degrees centigrade.
Subjects were not permitted to smoke or consume food
or drink during the study period. All subjects were
given a patient information leaflet to read and were
asked to sign the consent form. After signing the con-
sent form the subjects completed a questionnaire about
their medical history and their suitability for inclusion
into the study was checked. Subjects were deemed suit-
able for inclusion in the study if the subject was over
eighteen years old and healthy as determined by med-
ical history. Subjects were not enrolled in the study if
the subject had suffered with acute upper respiratory
tract infection in the previous two weeks, or if the sub-
ject had a history of seasonal or perennial rhinitis.
Experimental procedures
Once enrolled into the study subjects were randomized
to receive chilling or control procedures. A computer
generated randomization list was used to assign subjects
to either the chill or control procedure with subjects
stratified according to the number of common colds
reported by the subject in the previous year. Subjects
with 03 colds in the previous year were allocated to the
609
next available procedure at the start of the randomisa-
tion list and subjects with 4 or more colds were assigned
to the next available procedure at the end of the list.
Ninety subjects were allocated to receive the chill pro-
cedure and ninety subjects to receive the control pro-
cedure. If allocated to the chilling procedure, the
subject was asked to remove their shoes and socks
and place their feet in a bowl containing 910 litres
of water at a temperature of 10 C for twenty minutes.
The temperature of the cold bath was monitored
(Pen shape digital multi-stem thermometer, Scientific
Laboratory Supplies Ltd, Wilford Industrial Estate,
Nottingham, UK) and ice was added if necessary to
maintain the water temperature at 10 C. If allocated
to the control procedure the subject was asked to
keep their shoes and socks on and place their feet in
an empty bowl for twenty minutes. Warm water was not
used as a control as it was believed that this stimulus
Downloaded from http://fampra.oxfordjournals.org/ by guest on April 8, 2014
could have influenced nasal blood flow.
Symptom scores
All subjects were asked if they were suffering with a
cold and to score symptoms of runny nose, blocked
nose, sore throat, sneezing and cough on a scale of
03 with 0 = not present, 1 = mild, 2 = moderate, 3 =
severe before and immediately after the procedure. The
same common cold question and symptom scores were
also used in a daily diary. The method of symptom
scores has been widely used in previous studies on com-
mon cold.12,13 All subjects were provided with a diary,
in which they were instructed to score symptoms and at
the same time to indicate if they believed they were
suffering from a common cold (day 1 PM, days 2 and
3 AM/PM, day 4 AM, and on visit two which occurred
on day 4 or 5).
Nasal airflow was measured as a Nasal Partitioning
Ratio (NPR) as described by Cuddihy and Eccles14
before the procedures and on day 4/5 using the GM
NV1 spirometer (GM Instruments Ltd, Unit 6
Ashgrove, Ashgrove Rd, Kilwinning, Scotland, UK).
NPR was believed to be useful as an objective measure
to confirm the presence of acute rhinitis. However the
measurements of NPR proved to be too variable to pro-
vide any meaningful data and these results are not
presented in the present paper.
Previous history of colds
As part of the clinical history subjects were asked how
many colds they had suffered from in the previous year.
Statistics
This was a pilot study and it was therefore not possible
to perform a power calculation, but the ratio of sub-
clinical to clinical infection was considered in order
to determine the sample size required for the study.
It was predicted that 29 subjects in the chilled group
would develop colds and 9 subjects in the control
610 Family Practicean international journal

group giving a maximum difference between the groups
of 20 and a minimum of 10 depending on the distribu-
tion of spontaneous colds. Statistical comparisons were
made between the two experimental groups of subjects;
chilled and control. The hypotheses were tested at a
0.05 level of significance. The Mann-Whitney test
was used to test for differences in symptom scores
and history of colds incidence. The immediate effects
of chilling were studied by comparing the differences
from baseline to immediately after the test procedures
in total symptom scores, between chilled and control
groups. The delayed effects of chilling were studied
by comparing the differences in total symptom scores
between chilled and control groups over the 4/5 days
period after the test procedures. Mean total symptom
scores have been used to describe the symptom score
data in the text as this descriptive shows a change in the
symptom score, whereas the median does not due to the
test was used to test for differences in the number of
colds reported by the two test groups in their diaries. A
subject was deemed to have experienced a cold if they
reported they were suffering from a cold on any occa-
sion after the test procedures on days 1 + 2 + 3 + 4/5.
Results
Subject demographics
180 subjects were enrolled in the study between
October 2003 and March 2004, 90 were randomized
to the chill procedure, and 90 to the control procedure.
The flow diagram in Figure 1 shows the flow of parti-
cipants through each stage of the study. The demo-
graphics of the two test groups are provided in Table 1
that demonstrates that the test groups were balanced
and there was no significant difference in any of the

Downloaded from http://fampra.oxfordjournals.org/ by guest on April 8, 2014

large number of zero scores. The total symptom scores
(days 1 + 2 + 3 + 4/5, maximum score 120) were also ana-
lysed as dichotomous data using the Chi-squared test,
with total scores of 08 indicating absence of a cold and
9120 indicating presence of a cold. The Chi-squared
baseline characteristics.
Acute effects of chilling
The test procedures did not cause any significant
changes in symptom scores, and all the mean scores

Assessed for eligibility

(n=188)

Excluded (n=8)
Not meeting inclusion
criteria (n=4)
Refused to participate
Randomised (n=4)
(n=180)

Allocated to chill Allocated to control

procedure (n=90) procedure (n=90)
Received chill Received control
procedure (n=90) procedure (n=90)

Lost to follow-up: Lost to follow-up:

3 subjects returned 2 subjects returned to
to the centre later the centre later than
than day 5 day 5

Analysed (n=90) Analysed (n=90)

Excluded from analysis Excluded from analysis
where comparisons we re where comparisons we re
made on results drawn made on results drawn
from day 4/5 (n=3) from day 4/5 (n=2)

FIGURE 1. Flow of subjects through each stage of the study

 Acute cooling of the feet and the onset of common cold symptoms 611
TABLE 1 Demographics of test groups 90

Control n = 90 Chill n = 90 Significance

60
Median age 20.0 (1843) 20.0 (1839) P = 0.598 (MW)

subjects
(range)
Male 29 25 P = 0.515
(Chi-squared) 30
Female 61 65
Median colds 2.0 (110) 2.0 (18) P = 0.859 (MW)
per year (range)
0
Control Chilled
MW = Mann Whitney test.
FIGURE 2. Numbers of subjects that reported they were
suffering from a common cold in their diaries during the 4/5
TABLE 2 Immediate effects of chilling days period following control or chill procedures. The
shaded area represents those subjects reporting colds
Baseline Immediately Difference Statistics
after
symptom scores for days 14/5 following the control

Downloaded from http://fampra.oxfordjournals.org/ by guest on April 8, 2014

procedure
procedure (2.89 3.39 s.d.) (P = 0.013). When the total
Control 0.02 0.13 0.11 symptom scores for the 4/5 days were analysed as dicho-
(n = 90) (0.15) (0.37) (0.35) Comparison tomous data, 26/90 (28.8%) of the chilled subjects and
of differences 8/90 (8.8%) of the control subjects were deemed to be
Chill 0.07 0.21 0.14 P = 0.62 (MW)
(n = 90) (0.29) (0.51) (0.41) suffering from a cold (total symptom score 9120), and
this difference was significant (P = 0.001).
Figures are mean (standard deviation) of total symptom scores, before The total number of subjects that reported they were
and immediately after control and chill procedures. Differences are suffering from a common cold in their diaries during the
differences from baseline.MW = Mann Whitney test. 4/5 days following the chill or control procedures is
shown in Figure 2 and this illustrates that significantly
more subjects believed they were suffering from a cold
TABLE 3 Delayed effects of chilling in the chilled group (13/90, 14.4%) compared to the
control group (5/90, 5.6%, P = 0.047). There was no
Day 1 Day 2 Day 3 Day 4/5 Total Statistics sex difference in the development of colds with 9.3%
of males and 10.3% of females developing colds (P =
Control 0.32 0.73 0.48 1.36 2.89 0.828, Chi-squared). Of those in the chilled group that
(n = 88) (0.70) (1.11) (0.77) (1.95) (3.39) Comparison
mean score of total developed colds 4/13 were male (31%) and 9/13 female
symptom (69%), but this sex difference merely reflects the pro-
Chill 0.57 1.38 1.28 1.93 5.16 scores
(n = 87) (1.12) (1.84) (1.48) (2.83) (5.63) P = 0.013
portions of males (28%) and females (72%) exposed to
mean score the chill procedure and is not significant (P = 0.749,
Fisher Exact).
Figures are mean (standard deviation) of daily symptom scores and
total scores for days 14/5 following each procedure. Colds history in previous year
There was no difference in colds incidence between the
two test groups at baseline as illustrated in Table 1.
were close to zero indicating few or no symptoms were However, when looking at both test groups combined,
present before or immediately after the procedures as those subjects who believed there were suffering from a
illustrated in Table 2. The small difference in symptom cold had a history of more colds each year (median 2.00,
scores between the two groups prior to the procedures range 110) compared to those who did not develop a
was not significant (P = 0.245). The difference in total cold (median 3.00, range 28, P = 0.007).
symptom score pre and post chill procedure was not
significantly higher than the difference in total symp-
tom score pre and post control procedure (P = 0.62). Discussion
Delayed effects of chilling Acute effects of chilling
Table 3 shows the mean daily scores and total score for The present study provides no evidence for an acute
days 1 to 4/5 following each procedure. Total symptom effect of chilling on the development of common
scores for days 14/5 following the chill procedure cold symptoms. Symptom scores were close to zero
(5.16 5.63 s.d.) were significantly higher than the total in both the control and chilled groups.
612 Family Practicean international journal
Delayed effects of chilling
A delayed effect of chilling on the incidence of colds
and symptoms was observed in the 4/5 days following
the chill procedure. Significantly more chilled subjects
than control subjects reported they were suffering from
colds in the 4/5 days following the test procedures. The
difference in the incidence of colds between the two test
groups was also supported by a significant difference in
total symptom scores over the 4/5 days following the
test procedures. Analysis of the symptom scores as
dichotomous data also demonstrated a significantly
greater symptom score (more colds) in the chilled
group. The increased incidence of reports of colds
and higher symptom scores in the chilled subjects
compared to the control subjects may be due to several
factors.
Belief in the folklore that acute chilling of the body
surface, in some way precipitates a common cold could
have caused some bias in the reporting of colds and
symptoms. The subjects were not questioned about
their beliefs but the idea was introduced and then dis-
missed in the informed consent information in the fol-
lowing way:
This study is designed to investigate the effects of
acute chilling on the development of common cold
symptoms. It is a popular belief that the develop-
ment of an upper respiratory tract infection such as
the common cold is a result of a chill. However,
previous studies have failed to demonstrate that
exposure to a cold environment increases the incid-
ence of the common cold. . . . Common cold
symptoms are very common during the winter per-
iod and it is expected that some subjects will
develop symptoms because they have been previ-
ously exposed to infected persons. Therefore the
development of any common cold symptoms
may be unrelated to any experimental procedures
in this study.
If the common cold symptoms reported after chilling
were solely a result of subject bias caused by belief
in the effects of chilling then one would have expected
an acute effect of chilling on the scoring of common
cold symptoms rather than a delayed effect.
The differences between the chilled and control
groups could have occurred as a chance finding, as it
was expected that some subjects would develop cold
symptoms due to natural exposure to common cold
viruses. The probability value for the different report-
ing of colds was just below P = 0.05 (P = 0.047) but the
P-value for the difference in total symptom scores was
more convincing with P = 0.013, and for the dichotom-
ous analysis was P = 0.001. With two different measures
of the incidence of common cold providing significant
differences between the two test groups it is unlikely
that the results are solely due to chance.
Chilling of the feet in cold water (12 C 1 C) has
been previously reported to cause an intense vasocon-
striction of both the cutaneous and upper airway blood
vessels15 and the vasoconstriction of the upper airways
has been proposed as a mechanism that reduces respir-
atory defence against infection.10,11 When common cold
viruses are circulating in the community a proportion of
subjects will have sub-clinical infections, and chilling of
these subjects may cause vasoconstriction in the upper
airway epithelium and conversion of a sub-clinical to a
clinical infection. In these cases the subject links the
causality of the common cold symptoms to the chill
and does not realise that they were already infected
before they caught a cold. Laboratory studies using
viral challenge and cold exposure do not provide any
evidence that chilling increases susceptibility to the
development of common cold symptoms7,8 but these
studies do not mimic the natural exposure to common
Downloaded from http://fampra.oxfordjournals.org/ by guest on April 8, 2014
cold viruses and they can be criticised for the small
numbers of subjects used to power the studies.
An interesting finding in the present study was that
the subjects who reported they developed a cold after
the chill or control procedures also reported that they
suffered from significantly more colds each year, than
the subjects who did not report a cold after the proced-
ures. This finding may indicate that there is a sub popu-
lation in the general population who are more
susceptible to developing common cold symptoms
each year and that they may have a common cold con-
stitution.16
The results of the present study demonstrate that
chilling is associated with the onset of common cold
symptoms but the study does not provide any objective
evidence, such as virology, that the subjects were infec-
ted with a common cold virus. Because of the great
variety of viruses causing the common cold syndrome
it is difficult to identify the causative agent responsible
for common cold symptoms in any subject when viruses
are circulating in the community. For this reason it was
decided to first study the relationship between chilling
and symptoms, and then to consider the use of virology
in a subsequent study.
In summary the results of the present study support
the folklore that exposure to chilling may cause the
onset of common cold symptoms, perhaps by some
change in respiratory defence caused by reflex vasocon-
striction of the blood vessels of the upper airways. Fur-
ther studies in this area are needed to determine if the
development of common cold symptoms following cold
exposure are associated with infection.
Declaration
Funding: the study was funded by Cardiff University.
The study sponsor had no involvement in the study
design, the collection, analysis and interpretation of
 Acute cooling of the feet and the onset of common cold symptoms
data, in the writing of the report or in the decision to
submit for publication. The corresponding author had
full access to all the data in the study and had final
responsibility for the decision to submit for publication.
Ethical approval: the study was approved by the South
East Wales Local Research Ethics Committee.
Conflicts of interest: none.
References
1
Heikkinen T, Jarvinen A. The common cold. Lancet 2003;
361(9351): 5159.
2
Johnston S, Holgate S. Epidemiology of viral respiratory infec-
tions. In Myint S, Taylor-Robinson D (eds). Viral and other
infections of the human respiratory tract. London: Chapman
& Hall; 1996, 138.
3
Helman CG. Feed a cold, starve a fever Folk models of infection
in an English suburban community, and their relation to med-
ical treatment. Cul Med Psychiatry 1978; 2: 107137.
4
Lower R. De Catarrhis 1672. London: Dawsons of Pall Mall: 1963.
5
Mackenzie M. Diseases of the throat and nose. London: J&A
Churchill; 1884.
6 15
Andrewes C. The Common Cold. New York: Norton; 1965.
7
Dowling HF, Jackson GG, Spiesman IG, Inouye T. Transmission
of the common cold to volunters under controlled conditions.
II. The effect of chilling of the subject upon susceptibility. Am
J Hygiene 1958; 66: 5965.
613
8
Douglas RGJ, Lindgren KM, Couch RB. Exposure to cold envir-
onment and rhinovirus common cold. Failure to demonstrate
effect. New Engl Med J 1968; 279: 742747.
9
White DO, Brown L. Respiratory viruses. In Gronoff A, Webser R
(eds). Encyclopedia of virology. San Diego: Academic Press;
1999.
10
Mudd S, Grant SB. Reactions to chilling of the body surface.
Experimental study of a possible mechanism for the excitation
of infections of the pharynx and tonsils. J Med Research 1919;
40: 53101.
11
Eccles R. Acute cooling of the body surface and the common cold.
12
Rhinol 2002; 40(3): 109114.
Jackson G, Dowling H, Spiesman I, Boand A. Transmission of the
common cold to volunteers under controlled conditions. 1 The
common cold as a clinical entity. Arch Intern Med 1958; 101:
267278.
13
Macknin ML, Piedmonte M, Calendine C, Janosky J, Wald E. Zinc
gluconate lozenges for treating the common cold in children
A randomized controlled trial. J Am Med Assoc 1998; 279(24):
19621967.
14
Cuddihy PJ, Eccles R. The use of nasal spirometry for the
assessment of unilateral nasal obstruction associated with
changes in posture in healthy subjects and subjects with
Downloaded from http://fampra.oxfordjournals.org/ by guest on April 8, 2014
upper respiratory tract infection. Clin Otolaryngol 2003;
28(2): 108111.
Drettner B. Vascular reactions of the human nasal mucosa on
exposure to cold. Acta Otolaryngologica (Stockholm) 1961;
Supplementum 166: 1109.
16
Ball TM, Holberg CJ, Martinez FD, Wright AL. Is there a common
cold constitution? Ambul Pediatr 2002; 2(4): 2617.