Ascorbic Acid and the Common Cold

Analysis of controlled studies shows ivith statistical significance that this
vitamin has protective power.

Linus Pauling

Linus Pauling is Professor of Chemistry in Stanford University, Stanford,

California 94305.

   For many years there has existed the popular belief that ascorbic
acid has value in providing protection against the common cold, and in
ameliorating the manifestations of this viral disease.  This popular belief
has, however, not been generally shared by physicians, authorities on nu-
trition, and official bodies.

    I was puzzled by the contradiction between the popular belief and
.the official opinion, and I made a study of published reports of controlled
trials of ascorbic acid in relation to the common cold.  On the basis of this

study and of some general arguments about orthomolecular medicine (l),
(the preservation of good health and the treatment of disease by varying
the concentrations in the human body of substances that are normally present

in the body and are required for health), I reached the conclusion that ascorbic
acid, taken in the proper amounts, decreases the incidence of colds and re-

lated infections, and also decreases the severity of individual colds.  These
arguments were presented in my book V-and the Common Cold, which
w2s published in December 1970 (2).

    In this book I presented a discussion of the studies that had been made,
including several carefully controlled double-blind studies carried out by
competent medical investigators.  The evidence and arguments presented in

this book apparently were not convincing to some physicians, experts in


-2-

nutrition, and health officials.  Many statements contradicting my conclu-

sions were made within a few weeks of the publication of the book.  Al-
though the analysis in my book of the published accounts of controlled
studies in this field seemed to me to be clear and straightforward, I
have decided that, because of the importance of the question, it is de-
sirable for me to publish a more detailed account of the evidence, includ-
ing a more thorough statistical analysis of the controlled trials that have

been carried out.

The Nature of the Statistical Analysis

   Most of the reports discussed in this paper describe studies of
two groups of subjects selected at random from one population.  The sub-
jects in one group are administered the active substance (L-ascorbic acid,
vitamin C) in certain amounts once or more every day, and those in the
second group are administered an apparently identical inactive material,

a placebo.  Several of the studies were double-blind, with neither the sub-

jects nor the investigators knowing which subjects received the ascorbic acid
and which received the placebo, that information being kept by some other
person until all of the information had been collected.

   The question that I attempt to answer by analyzing the published
reports is the following: To what extent, if any, does the regular adminis-
tration of ascorbic acid over a period of time beginning before the subjects
have contracted a cold, and with the subjects exposed to cold viruses under
ordinary living conditions, have an effect greater than that of a placebo in
decreasing the incidence and the severity of the common cold? A compari-
son with a placebo, with the subjects not knowing which group they are in,
is essential because of the well-known "placebo effect" of even inactive

medications.

      The statistical methods used in the analysis are the conventional
ones, for the most part the calculation of s2 and then of the probability


- 3-

P(one-tailed) that the observed difference in effect of ascorbic acid and
pla'cebo (or a larger difference.) would be obtained by chance alone in two
groups taken at random from a uniform population if the null hypothesis of
equal effectiveness of ascorbic acid and placebo were true.  I have chosen
to give P(one-tailed) rather than'P(two-tailed) because no one contends
that the placebo (usually citric acid) has a greater effect than ascorbic
acid in preventing or ameliorating the common cold; the difference of opinion
is between those people who state that ascorbic acid is no better than a placebo
and those who say that it is better.  Moreover,  in none of the studies dis-
cussed did the investigators find a greater protective effect of the placebo
than of ascorbic acid; in every study ascorbic acid is reported to provide great-
er protection. than the placebo against the common cold, and the question to be
answered is the level of confidence with which the reported results can be ac-
cepted and the null hypothesis of equal effectiveness of placebo and ascorbic
acid can be rejected.

   In the following analysis I discuss the reported effects in three aspects:
first, the incidence of colds (number of colds per person in unit time, usually
taken as the period of the study); second, the average severity of individual

colds (as measured by days of illness per cold or number of days when
symptoms were recorded); and third, the integrated morbidity (the product
of the other two).  Mention is made also of the incidence and severity of
other infectious diseases, to the extent that they were reported by the in-

vestigators.

The Work of Ritzel

An important study (3) that gave results with statistical significance

was reported in 1961 by Dr. G. Ritzel, who is a physician with the medicd
servic? of the School District of the City of BaseI, Switzerland.  The study


-4-

was carried out in a ski resort with 279 skiers during two periods of five to

seven days.  The conditions were such that the incidence of colds during
these short periods was large enough (approximately 20 percent) to permit

results with statistical significance to be obtained.  The subjects were roughly
of the same age and had similar nutrition during the period of study.  The

investigation was double-blind, with neither the participants nor the physi-

cians having any knowledge about the distribution of the ascorbic-acid tablets
and the placebo tablets.  The tablets were distributed every morning and
taken by the subjects under observation such that the possibility of inter-
change of tablets was eliminated.  The subjects were examined daily as to
symptoms of colds and other infections,  as listed in the footnote of Table 1.

The records were largely on the basis of subjective symptoms, partially
supported,by objective observations (measurement of body temperature,
inspection of the respiratory organs, auscultation of the lungs, and so on),
Persons who showed cold symptoms on the first day were excluded from the
investigation.

   After the completion of the investigation a completely independent
group of professional people was provided with the identification numbers
for the ascorbic-acid tablets and placebo tablets, and this group carried

out the statistical evaluation of the observations.

    The principal results of the investigation are given in Table 1.
The author points out that the group receiving ascorbic acid showed only
39 percent as many days of illness, per person, as the group receiving the
placebo, and that the number of individual symptoms per person was only
35 percent as great for the ascorbic-acid group as for the. placebo group,
and states that the statistical evaluation of these differences by two-by-two
tables gives a significant difference, 0. 001 cP ~0. 01.  The author also

points out that the average number of days per cold for the ascorbic acid
group was 1. 8 (more accurately 1. 82), 29 percent less than the value for
the placebo group, 2. 6 (2. 58), and that this difference is statistically sig-

nificant, with P <O. 05 on 2 t-test.

   In Table 2 of the paper by Ritzel the values of the number of patients
shqwing different symptoms (the seven classes of symptoms listed in the


Table 1

The Study by G. Ritzel

Number in group

Number of colds


Incidence of colds

Total days of illness
Total- individual symptoms **
Severity of individual colds,
  from days of illness pe; cold

from individual symptoms per cold

Integrated morbidity

from days of illness per person        0.571        0.223       <o. oi       S!.%

from individual symptoms per person    0.850       0. 302           <o. 01     64%

Placebo
group

Ascorbic-acid
grouF

P(l+iled)*

Deere ase

140         139


31          17


0. 221       0.122


80          31


119           42

<o. 02

45%

2. 58        1. 82           <o. 05       29%


3. 84        2.47           <o. 05       36%

For rejection of null hypothesis of equal effect of ascorbic acid and placebo.

Pharyngitis, laryngitis, tonsilitis, sore throat; bronchitis, coughing; fever, chills; otitis media;
rhinitis; herpes labialis; other symptoms (muscle ache, headache, abdominal pain, vomiting, .
diarrhea, general malaise).


-6-

footnote to Table 1) are given, and the number of days of illness for each

symptom.  It is interesting that for each of these seven classes of symptoms
the number of patients showing the symptom is less for the ascorbic-acid
group than for the placebo group, and that, moreover, the number of days
of illness per patient showing the symptom is also less.

   Let us discuss separately the effect of ascorbic acid on the incidence
of the common cold and its effect on the severity of individual colds.  The
number of colds was 31 for the placebo group and 17 for the ascorbic-acid

group.  (The number of colds was not given .explicitly in the paper.  However,

the number of days of illness for each of the two groups was given (80, 31),
and the average number of days of illness per cold (2. 6, 1. 8 ).  The only
integral values for the number of colds allowed by these numbers are 31
for the placebo group and 17 for the ascorbic-acid group. ) The incidence
of colds is accordingly 0.221 per person for the placebo group and 0.122 for
the ascorbic-acid groulJ, a decrease by 45 percent for the ascorbic-acid group.
The value of x2 is found to be 4. 81, with Phone-tailed) ~0.02.  This investi-
.gation accordingly shows with statistical significance that the null hypothesis
that ascorbic acid has only the same effect as the placebo is to be rejected.

   Two values may be calculated for the effect of ascorbic acid on the
severity of individual colds.  In Table 1 the number of days of illness per
cold for the placebo group is given as 2.58, and for the ascorbic- acid group
as 1. 82, 29 percent smaller.  Moreover, the average number of individual
symptoms recorded per cold (they were recorded daily) is given as 3.84 for
the placebo group and 2. 87 for the ascorbic acid group, 36 percent smaller.
Each of these differ,ences is statistically significant, the null hypothesis that
the two populations are the same with respect to the number of days of
illness per cold and the individual symptoms per cold being rejected at the
level P(one-tailed) ~0.05.

Two values are given in Table 1 for the integrated morbidity, one as

measured by-the number of days of illness per person and the other as measured


-7-

by the number of symptoms (recorded daily) per person. ,These values are 61

percent and 64 percent less, respectively, for the ascorbic-acid subjects
than for the placebo subjects, with the differences significant at the level

P co. 01.

    This investigation seems -to have been very well planned and

executed.  Dr. Ritzel was aware of the problem of obtaining reliable results
in the study of the common cold, and he discussed the problem in some detail.
His paper is provided with an English-language summary, reading as follows:

"The possibility of preventing infection by administration of vitamin C was

investigated in a moderately large test population during a period of in-
creased exposure.  The trial was conducted in such a way as. to exclude sources
of error in assessing subjective symptoms.  Statistical evaluation of the results

confirmed the efficacy of vitamin C in the prophylaxis and treatment of colds.
Problems of therapeutic trials with pluripotential preparations which have
to be judged chiefly on the basis of subjective symptoms are discussed. "

    It is interesting that in an often-quoted review of the evidence about

ascorbic acid and the common cold, which ended with the statement that
"there is no conclusive evidence that ascorbic acid has any protective effect
against, or any therapeutic effect on, the course of the common cold in
healthy people not depleted of ascorbic acid" (4), the work of Ritzel was
covered in two sentences, stating quite erroneously that he had reported "a
reduction of-`39 percent in the number of days ill from upper respiratory in-
fections and a reduction of 35 percent in the incidence of individual symptoms
in the supplemented group as compared with the placebo group;" (the correct

values are 61 percent and 64 percent, respectively).

The Work of Cowan, Diehl, and Baker

One of the best studies of ascorbic acid and the common cold was


-8-

reported by Cowan, Diehl, and Baker in 1942 (5). Dr. Diehl was at that
time Dean of Medical Sciences in the University of Minnesota.  He is now

retired.  Dr. Cowan is Chief of the Student Health Service in the University,
and Dr. Baker is Professor of Neurology there.  The principal work on
ascorbic acid was done during the winter "cold season" of 1939-1940,  The
subjects were all students in the University of Minnesota who volunteered to
participate in this study because they were particularly susceptible to colds.
Persons whose difficulties seemed to be due primarily to chronic sinusitis
or allergic rhinitis, as shown by examination of the nose and throat and

consideration of symptoms of.allergy, were excluded from the study. The
subjects were .assigned alternately and without selection to an experimental
group and a control group.  The subjects in the control group were treated
exactly like those in the experimental group, except that they received a
placebo instead of the ascorbic acid.  The subjects were instructed to re-
port to the Health Service whenever a cold developed, so that special re-
port cards could be filled in by a physician.  Dr. Cowan has informed me

that the study was a double-blind one, with neither the. subjects nor the physi-
cians knowing which group a subject was in.  Each subject was interviewed every
three months in order to check the completeness of the reports.

    The study was continued for 28 weeks. Of the 233 students initially
in the ascorbic acid group, 183 received 200 mg per day throughout the pe-
riod of 28 weeks, and 50 received 200 mg per day for two weeks, followed
by 100 mg per day except on inception of a cold, when an additional 400
mg per day for two days was administered.  This group numbered 208
subjects at the completion of the study, 25 having dropped out.  If the com-

position of the group remained unchanged, the average intake of ascorbic
acid was 180 mg per day.  The students in the control group initially num-
bered 194, of whom 155 completed the study (Table 2).

    The authors report the observed incidence of colds by giving the
average and the probable error.  The ?orresponding values of the standard

deviation, as calculated from the probable error, -are given below in paren-


-7-

Table 2

-The Study by Cowan, Diehl, 2nd Baker

                                    Placebo     Ascorbic-acid    P(one-tailed)    Decrease
                                     i!FOUP       group

   Number in group                    155         208


    Incidence of colds                    2. 2         1. 9           <o. 02        14%

   Severity (days of illness per cold)        0.73        0. 58           <o. 02        21%

Integrated morbidity (days of illness per person)   1.6         1.1            co. 01        31%


-1a -

theses.  The average number of colds per person during the period of study

was'2. 2 + 0. 08 (S. D. 0.113) for the control group, and 1. 9 + 0. 07 (S. D.
0.099) for the ascorbic acid group.  The difference between the average

number of colds in the control group and in the experimental group is given
by the authors as one-third of a cold and also as 0. 3 + 0.11 (S. D. 0.156).

   The authors state in their paper that "The actual difference between
the two groups during the year of the study amounts to one-third of a cold

per person.  Statistical analysis of the data reveals that a difference as

large as this would arise only three or four times in a hundred through

chance alone.  One may therefore consider this as probably a significant
difference, and vitamin C supplements to the diet may therefore be judged
to give a slight advantage in reducing the number of colds experienced. "

   Because the authors rounded off the numbers giving the actual
numbers of colds per person, the difference is not known exactly. Dr.
Cowan has informed me that the original records and the original cal-
culations are no longer available.  There is evidence, however, that the
actual difference between the average number of colds in the two groups
is 0. 32, with uncertainty 0. 01.  If the difference had been less than 0. 29

the authors would have said "one quarter of a cold per person", rather
than "one third of a cold per person".  Moreover, the value of Pftwo-

tailed) calculated for a difference of 0. 31 with standard deviation 0.156

is 0.042, and that calculated for difference 0. 33 is 0. 031.  The statement
by the authors that the difference would arise only three or four times in
a hundred through chance alone accordingly restricts the difference to the

range 0. 31 to 0. 33,  with 0. 32  as the likely value.
    This difference represents a decrease by 14.4 percent in the inci-

dence of colds in the ascorbic-acid group as compared with the control

group.

The value of P(one-tailed) for difference 0. 31 to 0. 33 is 0.021 to

0.016.  We can accordingly state that the observed difference is statistically


- 11 -

significant, with P(one-tailed) less than or equal to 0.02.  The null hypothe-
sis that ascorbic acid has the same effect as the placebo is accordingly
eliminated at this level.

   The average number of days lost from school per person in the
placebo group was reported as 1.6, and in the ascorbic-acid group as 1.1,
giving a decrease of 31 percent in integrated morbidity.  The average num-
ber of days lost from school per cold was 0. 73 for the placebo group and 0. 58
for the ascorbic-acid group, a decrease in severity of individual colds by 21

percent.  For both severity and integrated morbidity the null hypothesis of

equal effectiveness of ascorbic acid and placebo is decisively rejected, with
P(one-tailed) < 0.01.

The Work of Franz, Sands. and Hevl

    A double-blind study of ascorbic acid and the common cold was
carried out by Franz, Sands, and Hey1 of Dartmouth Medical School during
the three-month period from February to May 1956, with 89 volunteer medical
students and student nurses (6).  The subjects were divided, in a random way,
into four groups, three of twenty-two subjects and one of twenty-three sub-
jects.  One group received tablets containing ascorbic acid, the second ascor-
bic acid and a bioflavonoid (naringin), the third a placebo, and the fourth
naringin only.  The daily amount of ascorbic acid was 205 mg and that of the
bioflavonoid was 1000 mg.  Symptoms of colds were systematically recorded.
The results for the bioflavonoid groups, with or without ascorbic acid, were

the same as for the corresponding groups without bioflavonoid.  The authors

concluded that the administration of a bioflavonoid had effect neither on the

incidence or the cure of colds nor on the ascorbic acid level of the blood.

    The results reported by the authors are given in Table 3.

     From this table we see that the incidence of colds in the two ascorbic-
acid groups is nearly the same as in the other groups (4. 6 percent less).


Table 3

The Study by Fran& Sands, and Hey1

Group

Number
in group

Number of colds
Total       Not cured or
        Improved in 5 da

Ascorbic acid

Ascorbic acid plus bioflavonoid

Placebo

Bioflavonoid

Total incidence of colds 4. 6% less for ascorbic-acid groups than for other two

groups, not statistically significant; incidence of severe colds (not cured or im-
proved in 5 days) 87.5% less for ascorbic-acid groups than for other groups,

statistically significant at the level P(one-tailed) < 0. 01.


- 13 -

The difference is not statistically significant.  Because of the small num-

bers of subjects and colds, a decreased incidence would have to be as great
as 50 percent to be significant at the level P(one-tailed) ~0.05.

    The authors point out that the subjects receiving ascorbic acid
showed more rapid improvement'in their colds than those not receiving
it, and that this difference is statistically significant at the 0.05 level.
This statistical analysis was made by Professor J. Laurie Snell, of the
Department of Mathematics, Dartmouth College.

    The statistically significant observation reported by the authors
is that 8 of the total of 15 colds in the placebo and bioflavonoid groups
remained uncured or unimproved in five days, whereas of the 14 colds
in the two groups receiving ascorbic acid only one remained unimproved
or uncured in five days.  The authors accordingly reported a much lower
incidence of severe colds (unimproved or uncured in five days) for the two
ascorbic-acid groups than for the two other groups.  The observed inci-

dence of severe colds (not improved or cured in five days; one in 44 ascorbic-

acid subjects, eight in 45 other'subjects) leads to&=,5. 88, and is statistically
significant at the level P(one-tailed) ~0.01.  My conclusion is that the double-
blind study carried out by Fr'anz, Sands, and Hey1 has shown with statistical
significance that ascorbic acid has a greater effect than a placebo in decreas-
ing the incidence of severe colds.  A comparison with statistical infor&mation
about the duration of colds leads to the conclusion that the integrated morbidity
for the ascorbic-acid subjects was approximately 40 percent less than for the
placebo subjects.

The Work of Wilson and Low

   During the past six years Professor C. W. M. Wilson, Chairman
of the Department of Pharmacology of the University of Dublin, has, together
      .


- 14 -

with his coworkers, been carrying out clinical trials on the effect of ascorbic
acid on school children.  In this work 200-mg tablets of ascorbic and cor-
respond.ing placebo tablets were administered daily to children in boarding
schools during winter periods of six or seven months (7).  The studies

were double-blind, and the numbers of infections were large enough to give
statistically significant results.  Several reports of this work are now in
process of publication (8).  One paper, describing a study of 108 subjects in
a girls f school, has been published (7). Of these subjects, 57 received
ascorbic acid (200 mg per day) and 46 received placebo tablets.  The authors

(Wilson and Low) report that "As .a result of computer analysis it was
found that the symptoms in all the children could be separated into two
unrelated groups, consisting of sore throat, headache, feverish and out
of sorts, defined as toxic colds; and cold in the head, cough, nasal ob-

struction and nasal discharge, defined as catarrhal colds. Ascorbic acid

reduced the incidence, duration, and severity of these symptoms in com-
parison with those in children receiving dummy tablets.  The form of the

toxic and catarrhal colds was also significantly altered, so that symptom
association was reduced in the presence of ascorbic acid.  Duration of

the symptoms, cold in the head and nasal discharge, was reduced from 14 to
8 days in children receiving ascorbic acid. . . It is concluded that the pro-
phylactic administration of ascorbic acid to young adults significantly reduces
the intensity of the symptoms, and form of their association, in the common

cold. "

    The information so far published by Wilson and Low does not permit
an independent statistical analysis to be carried out.  The statement about

statistical significance made by Wilson and Low (and confirmed in a letter
`from Professor Wilson) corresponds to rejection of the null hypothesis
at the level P(two-tailed) < 3. 05, and hence to P(one-tailed) ~0.03.  The
reduction of the average period of duration of symptoms from 14 to 8 days
indicates. a decrease in integrated morbidity by about 40 percent.


- 15 -

The Work of Glazebrook and Thomson

   . In 1942 Glazebrook and Thomson, of the Department of Clinical
Medicine and Bacteriology, University of Edinburgh, reported a study
carried out with about 1500 boys 15 to 20 years old in a large training
school in Scotland (9).  The subjects received a normal diet rather low
in ascorbic acid, the daily ration being estimated to contain only 10 to 15

mg-  The principal study, carried out over a period of six months, in-
volved 1100 control subjects and 335 ascorbic-acid subjects.  The control

subjects, in seven dining groups, received the ordinary diet. The ascor-

bic- acid subjects, in two dining groups, received the ordinary diet but

with ascorbic acid administered in the milk and cocoa that was served.

The average amount of ascorbic acid administered is somewhat uncertain.
The authors state that vitamin C was added to the supplies of cocoa or milk
serving the tables for the appropriate divisions.  In their discussion of pre-

liminary experiments carried out to determine the daily urinary excretion
of ascorbic acid it is stated that initially 200 mg per day wa s given to each
boy, 100 mg being placed in the morning cocoa and 100 mg in an evening glass
of milk, the mixing being done in bulk in the kitchens.  Analysis of the cocoa
and milk showed an average of 63 mg per cup of cocoa and 98 mg per glass
of milk, suggesting that about 160 mg per day was the average intake.

    Because a number of preliminary studies had been carried out, and
the ascorbic acid was added in the kitchens, it is likely that this investigation
can be considered to have been a blind study.  The authors mention that

careful records had been kept of the incidence of all infections for 18 months
before the observations described in their paper were begun, and that in the
preceding year there had been an epidemic of tonsilitis that had affected all
the divisions uniformly, so that they could not be regarded as separate units
within the larger population.  All of the divisions had a population more or
less the. same as regards the duration of stay in the establishment.  Records

were kept of the common cold (coryza), tonsilitis (hemolytic streptococcal


- 16 -

disease of the nose and throat, covering tonsilitis, sore throat, otitis media,
pharyngitis, and cervical adenitis), and other infective conditions (conjunc-
tivitis, .boils, impetigo, etc., as well as pneumonia and acute rheumatism).
    The total numbers of cases of colds during the g-month period of

the study are given in Table 4 for the control group and the ascorbic-acid

group.  There is a decrease in incidence in all colds by 17 percent, with
P(one-tailed) < 0. 05, and in colds serious enough to require hospitalization
(sick quarters) by 23 percent, with P(one-tailed) < 0. 02.

     For other infectious diseases a decreased incidence for the ascorbic-

acid group was also reported (except for tonsilitis with inclusion of the mild
cases).  The reported decreases of 100 percent for pneumonia and acute
rheumatism are significant at the level P c 0. 02.

    Glazebrook and Thomson in their paper point out that the difference
in incidence of pneumonia and acute rheumatism in the control group and
the ascorbic-acid group is statistically significant, and also that the period
of hospitalization for tonsilitis is statistic,ally significant.  They give the
average stay in the hospital for control subjects (83) hospitalized with ton-
silitis as 16. 7 days, standard deviation 11. 86, and for the vitamin-C subjects

(18) as 10. 05, standard deviation 6. 96, and state that analysis shows that

a difference as great as or greater than that obtained would be expected once
in 50 times in a homogeneous population.

    Glazebrook and Thomson give information in their paper that permits
the severity of individual colds or other infectious diseases and the integrated
morbidity, as measured by the number of days hospitalized, to be calculated.
These values are given in Tables 5 and 6.  The values of P(one-tailed) in the

tables have been calculated by assuming a Poisson distribution in the days of
`hqspitalization per period of illness.

    The results described in Tables 4, 5, and 6 thus indicate that ascorbic
acid has the effect of decreasing the incidence and severity of tonsilitis, pneu-
monia, and acute rheumatism, as well as the common cold, for the principal
population studied by Glazebrook and Thomson.


Number in group
Cold;;

Colds, sick quarters

Tonsilitis

Tonsilitis, sick quarters

Pneumonia

Acute rheumatism

Table 4

The Principal Study by Glazebrook and Thomson

Incidence of Illnesses

Control group

Number Incidence

1100


286     0.260


253       . 230
:`94       . 086


83       . 075


17       . 016


16       . 015

Ascorbic-acid
group


Number     Incidence



335


72        0. 215


59          .176


29          . 087


18          . 053


  0         . 000


  0          . 000

PIone -tailed)

<o. 05


co.02


/JO. 5


CO. 08

< 0.02

<o. 02

Decrease

.170/o.

23%

-1%

28%

100%

100%


Table 5

The Principal Study by Glazebrook and Thornion

Severity of illness, measured by average number of

  days hospitalized per hospitalized case

Control      Ascorbic- acid
group       group

Decrease

Common Cold

1. 47         1.11           24%

Tonsilitis

All infective conditions*

5.0         2. 5           50%

* Common cold, tonsilitis, pneumonia, acute rheumatism, conjunctivitis, boils,
impetigo, etc.


Table 6

The Principal Study by Glazebrook and Thomson

Integrated morbidity, measured by average number

of days hosp'italized per subject*

Common Cold


Tonsilitis

Control      Ascorbic-acid
group       group


0.334        0.195




. 095             . 029

Decrease



41%

69%

* Values f&`all infective conditions not available because total number of
hospitalized cases not reported.


- 20 -

The Total Evidence for the Effectiveness of Ascorbic Acid

    The foregoing careful studies of ascorbic acid have given the sta-
tisti.cally significant result that the hypothesis that ascorbic acid adminis-
tered daily to subjects who have not yet caught cold and are subjected to
ordinary conditions of exposure to cold viruses show the same incidence,
severity, and integrated morbidity of colds and related infections. as
placebo subjects is to be rejected.  I now ask what the weight of the total
body of evidence is.

    In Table 7 the values are given of P(one-tailed) at which each of.
several studies rejects the null hypothesis.
    The study by Ritzel gave results for the integrated morbidity (Table
1) rejecting the null hypothesis at the level P(one-tailed) < 0.01.  The results

for the incidence of colds and the severity of individual colds are not inde-
pendent of those for the integrated morbidity, and accordingly do not change
the level at which the null hypothesis is rejected.  The value P ~0. 01 is

entered in Table 7 for the Ritzel study.

    Similarly, the value < 0. 01 of P(one=tailed) for the integrated
morbidity as found by Cowan, Diehl, and Baker (Table 2) is also entered

in Table 7, together with the value < 0. 01 for the incidence of severe colds
reported by Franz, Sands, and Heyl, the value < 0.03 for the study by

Wilson and Low, and the value < 0. 02 for the incidence of colds requiring
hospitalization reported by Glazebrook and Thomson.

     There is no doubt that the five studies listed in Table 7 are inde-

      pendent of one another.  We may accordingly combine them to obtain a



       %
*A   measure of their total significance in rejecting the null hypothesis.  Fisher's
    method leads toX2 (10 degrees of freedom) = -22 In Pi = 42.47, and hence
       to P CO. 00001.  There is some question as to whether or not the study by

        Glazebrook and Thomson can be considered to have been a double-blind

      study.  If it is omitted, application of Fisher's method to the other four,
      which were double-blind studies involving comparison of an ascorbic-acid
     group and a placebo group, lead s to X2(8 degrees of freedom) = 34.84 and

        P < 0. 0001.


Table 7

Investigators

The Weight of Evidence for Rejecting

the Null Hypothesis of Equal Effectiveness

   of Ascorbic Acid and Placebo

Value of P(one-tailed) at
which the null hypothesis
is rejected

Rit zel                      <o. 01

Cowan, Diehl, and Baker         <o. 01

Franz, Sands, and Hey1          co. 01


Wilson and Low               <o. 03


Glazebrook and Thomson         co. 02

Combined value *

<o. 00001

* X2 (10 degrees of freedom) = -2x In Pi = 42.47


- 22 -

   We conclude that there is overwhelming evidence requiring
rejection of the null hypothesis that ascorbic acid has no more value
than a,placebo in decreasing the incidence, severity, or integrated
morbidity of the common cold when it is regularly administered in
amounts 180 mg to 1000 mg per day to subjects exposed to cold viruses
in the ordinary way (contact with other persons), over a period of time
beginning before colds have been contracted. The chance that, through
a statistical fluctuation, the five investigations listed in Table 7 would
have given the results described for two samples of a single population
(in each study), with ascorbic acid having the same effect as the placebo
(the null hypothesis) is only 1 in 100000 (1 in 10000 for the first four
studies alone).

   It is interesting that there is strong evidence in these studies that
ascorbic acid in these amounts also has a protective effect against tonsilitis,
pneumonia, acute rheumatism, and other diseases.


- 23 -

Dependence on Amount of Ascorbic Acid

   An additional test of the `positive results (protective effect of ascor-
bic acid against the common cold) reported in the five investigations de-
scribed above can be made ,by checking their consistency with respect to
the amount of protective effect.  Because of the heterogeneity of the gen-
eral population, it is not unreasonable that the effect of ascorbic acid taken
daily in addition to the amount in the normal diet would for smaller daily
added intake be proportional to this intake, and that for larger intake it
would, if ascorbic acid has in fact protective effect, approach the limit
of 100 percent effectiveness.  Values of the incidence and also of the sever-

ity of individual colds reported in the five investigations are plotted in the
upper part of Figure 1, and values of the integrated morbidity in the lower

part.  (The values of incidence and severity are plotted together because
they are observed to be approximately equal. ) Smooth curves have been
drawn as indicated by the points.

It is seen that the results of the different studies are consistent with

one another.  None of the observed values differs from the corresponding
point on the curve by an amount that is statistically significant at the level
P(two-tailed) < 0. 05.

    These curves are, of course, only approximations; the protective

effect of ascorbic acid, relative to that of a placebo, can be expected to
depend on various factors, such as the average genetic nature of the popu-

lation, the food ingested,  and the nature of the cold viruses to which the

subjects are exposed, in addition to the daily intake of the vitamin.
    The observed decreases in incidence, severity, and integrated

morbidity of the common cold for the five studies are summarized in
Table. 8, together with the results of some other studies, discussed in
the following p ar agr aphs .


Table 8

Summary of Observed Decreases in

Incidence, Severity, and Integrated

Morbidity in Controlled Studies

Daily amount
of ascorbic acid

Observed decrease in
Incidence Severity    Integrated
             morbidity

Rit zel              1000 mg         45%    29,36%    61, 64%

Cowan, Diehl, and Baker    180 mg         14%       21%      31%

F'ranz, S<ands, and Hey1    205 ing          5%      37%      40%


Wilson and Low          200 mg                 43%

Glazebrook and Thomson    160 mg         17%       24%      41%


- 24 -

The Smaller Study by Glazebrook and Thomson

    A smaller study was also reported by Glazebrook and Thomson (9),
with 150 recruits who entered the institution and were studied during the
second half of the six-months period.  The results of this trial, as reported
by the authors, are given in Table 9.  A decrease in the incidence of colds

by 12 percent was noted, with, however, little statistical significance. The
incidence of tonsilitis was 79 percent less for the ascorbic-acid group than
for the control group, statistically significant at P(one-tailed) c 0. 05.  The
value 12 percent for the decreased incidence of colds does not differ signifi-
cantly from the corresponding value 18% given by the curve of Figure 1.

The Second Study by Cowan, Diehl, and Baker

    Cowan, Diehl, and Baker (5) also reported the results of a second
study, carried out with three groups of subjects (students in the University

of Minnesota) in the winter of 1940-1941. The subjects in the first group (82)
received 50 mg of ascorbic acid per day, plus other vitamins, those in the
second group (88) received 25 mg per day (plus other vitamins), and those
in the third group received a placebo.  The investigators reported no dif-
ferences among the groups in the number and severity of colds. ,The upper
curve in Figure 1 leads an expected decrease in incidence and in severity
by 7 percent for the first group and by 4 percent for the second group.  The

differences between these values and the observed values (0 percent) are
not statistically significant.

The Work of Dahlberg, Engel, and Rydin

A study by Dahlberg, Engel, and Rydin (10) has been quoted as show-


Table 9

The Smaller Study by Glazebrook and Thomson

Incidence of colds and tonsilitis

Control group,



Number   Incidence

Ascorbic-acid    P(one-tailed) Decrease
group                     in inci-
                        dence
Number  Incidence

Number in group     90                  60


Colds             29        0. 322       17


Tonsilitis           7              . 078       `1

Colds plus tonsilitis  36              . 400       18

0.283    <o. 30     12%


I 017    `CO. 05     79%


.300    co.10     25%


- 26 -

ing that ascorbic acid ha& no value in preventing the common cold or

affecting its duration.  For example, in the book The Vitamins in Medi-

cine by Bicknell and Prescott (11) there is the following statement:
"Dahlberg, Engel, and Rydin carried out a mass experiment on 2,500
Army conscripts, one-half receiving 200 mg of ascorbic acid, the other
half acting as controls.  No difference was noted in the frequency or dura-
tion of colds, fever, endurance tests, or diseases of any description in the
two groups. "

    Dahlberg, Engel, and Rydin themselves, in the summary of their
paper, state that "No difference could be found as regards frequency or
duration of colds, degrees of fever, etc. Military competitions, arranged'
to relieve the tedium, disclosed no difference between the two groups.
Thus, the soldiers who only received the diet of the Swedish Army, and who
showed a lpathological deficit' [in ascorbic acid in the blood], did not differ
in any respect from those who had been given ascorbic acid during the entire
period of investigation.  Consequently, there is no reason to assume vitamin
C to be at all instrumental in preventing colds when supplementing the degree
of vitamin deficiency existing among soldiers in the north of Sweden. "

    Examination of the paper by Dahlberg, Engel, and Rydin shows,
however, that these statements are not true. The investigators in fact
reported a decrease in the incidence of colds, a decrease in the incidence
of other infectious diseases, a decrease in the number of subjects with

fever, and a small improvement in functioning in the endurance tests.  The

statement by the authors is misleading; presumably they meant to say that

no statistically significant differences were found.
    The study was carried out with 2, 525 infantry soldiers stationed in

.an isolated region in northern Sweden, during the 90 days from 3 March to
31 May, inclusive.  It was a double-blind test, the composition of the tablets
being kept secret from both the doctors and the soldiers.  The subjects were


- 27 -

divided into two groups, the ascorbic acid group (1259) and the placebo

group (1266), in a random way, by odd and even identity numbers, re-
spectively.  The placebo tablets contained a suitable amount of citric acid
to disguise any difference in taste.  The ascorbic- acid subjects received
200 mg per day for the first 24 days, and 50 mg per day for the remaining
66 days, an average of 90 mg per day.  After 24 days and after 90 days a
statistically significant difference was found between the average ascorbic-
acid levels in the urine of the two groups, both during a fasting period and
after ingestion of 200 mg or 300 mg of ascorbic acid, in a loading test.

(The loading test results are referred to in the words "pathological deficit"
in the summary of their paper. ) The ascorbic-acid tablets and placebo tab-
lets were dispensed at the first meal of the day, and special steps were taken
to see that they were consumed at that time, and did not ga to the wrong

person.  The soldiers were told what the investigation was for, and were
requested not to eat any other food or other medicines during the time of
observation than what was provided in camp.  About half of the subjects
(in certain companies of soldiers) in each group were carefully checked,
and the average intake of 90 milligrams per day of ascorbic acid is reli-

able for them.  For the other half, in other companies, there were some
periods when some proportion of the subjects did not always take the tablets
regularly.  The authors present the results separately, but in fact they are
closely similar, and in the following discussion all ascorbic-acid subjects
are grouped together, and all placebo subjects.  The failure to check the
regular ingestion of the tablets occurred during only a part of the 50-mg-
per-day period, and it seems likely that the average ingestion of ascorbic
acid, taken as 90 mg per day, is not more than 10 percent high.

    The observations, presented in the original paper in five tables,

are summarized in Table 10.  The second row gives the number of colds
for the placebo group and the ascorbic-acid group.  These numbers cor-


- 28 -

respond to a 7.4 percent smaller incidence of colds for the ascorbic-
acid group than for the placebo group:  The next three rows give further
inform,ation about colds; namely, the numbers of subjects with colds (one
or more during the period of the study), registered as ill with colds, and
with colds and fever.  In these three categories, too, there are reported
decreases in incidence in the ascorbic-acid group, ranging from 2.5 per-
cent to 3. 7 percent.  All four values may be compared with the value 12
percent given by the upper curve of Figure 1 for the average intake of 90
mg per day.  The differences are not statistically significant, and the in-
,vestigation does not require`rejection either of the null hypothesis that
ascorbic acid has no more protective effect than a placebo nor of the hy-
pothesis that it has the amount of protective effect indicated by the curves
of Figure 1.

   It is interesting that the reported amount of protective effect for
all infectious diseases (last four lines in Table 10) is somewhat larger than
that for the common cold alone (average of four values 8.0 percent, as
compared with 4. 3 percent).

A field competition was held, participated in by 359 members of

the placebo group and 357 members of the ascorbic-acid group.  The
median ranking of the ascorbic-acid participants was 2. 0 percent higher
than that of the control participants.  Some superiority of the ascorbic-
acid group over the placebo group was accordingly reported in this test
(presu-mably the endurance test mentioned by Bicknell and Prescott),
even though the superiority is small, and not statistically significant.

    Dahlberg, Engel, and Rydin mentioned that they had recorded the
number of days each patient was on the sick list, and how many days,
if any, he had been treated in hospital.  These numbers are, however,

not given in the paper, and it is accordingly not possible to use them
in assessing the severity of individual colds.


Table 10

The Study by Dahlberg, Engel, and Rydin

Incidence of colds and of all infectious diseases

TotaI number of subjects

Total number of colds*

Subjects with common cold

Subjects registered as ill with common cold

Subjects with common cold and fever

Registered cases of disease

Diseased subjects

Subjects registered as diseased

Subjects diseased and with fever
                     .

Number,        Number,
placebo .        as corbic- acid
group          group

1266           1259


152             140


130            126


94             90


73             70


162            145


141**           131


103             95


80             73

Decreased
incidence in
ascorbic-acid
group

7.4%


2: 5%


3. 7%

3.6%

10.0%

6.6%


7. 3%


8. 2%

* From Table 1, corrected for other acute infections

**  Average of 142 in Table 1 and 140 in .Table 5


- 30 -

    The statistical significance of the results of this large-scale
study, involving 2525 subjects, is les's than that of the study of Cowan,
Diehl, and Baker, involving only 363 subjects, for two reasons. First,
the period of time was less than half as great in the former study, and
second, the incidence of colds was much less, presumably because the
soldiers were in an isolated `camp in northern Sweden, and not exposed
to many cold viruses.  The total number of colds reported by Dahlberg,
Engel, and Rydin is 292, whereas the total number reported by Cowan,
Diehl, and Baker is about 735.  Moreover, the amount of ascorbic acid

per day in the Scandinavian study was less than half as much as in the
Minnesota study, so that an effect only about half as great would be an-
ticipated.

    The study by Dahlberg, Engel, and Rydin indicates that ascorbic
acid in the average amount 90 mg per day has some protective effect, but
that the null hypothesis of no protective effect is not eliminated with sta-
tistical significance.  On the other hand, the hypothesis of the amount of
protective effect indicated in Figure 1 is also not eliminated with statistical
significance, and it is not justified to claim that this work has shown ascorbic

acid to have no value in controlling the common cold.
   So far as I am aware, there have not been published any other re-

ports of carefully controlled studies of the protective value of ascorbic
acid against the common cold when it is administered over a period of time
to subjects who had not yet contracted colds and were exposed to cold viruses
in the ordinary way, by contact with other people.

    In the effort .to make my search complete, I have written to about a
score of persons who have stated that it has been shown that ascorbic acid
has no protective value against the common cold, asking for--reference to any
controlled investigation that has given statistically significant results causing


- 31 -

rejection of the hypothesis that ascorbic acid has the protective effect
indicated by the curves in Figure 1 when administered over a period of
time to subjects exposed to cold viruses in the ordinary way.  This ef-

fort was unsuccessful; most of my correspondents replied, and all who
replied either stated that they di'd not know about any such investigation
or referred me to one or another of the publications discussed in this

article.

The Work of Walker, Bynoe, and Tyrrell

    The study by Walker, Synoe, and Tyrrell (12) is often mentioned
as having shown that ascorbic acid has no protective value against the

common cold.  Of the 91 subjects, 47 received 3g of ascorbic acid per day
for three days before inoculation with viruses (rhinoviruses, influenza
B virus, or B814 virus) and for six days after inoculation, and 44 subjects
received a placebo.  The incidence of colds was only 6 percent less for the
ascorbic-acid group (18147) than for the placebo group (18/44); this dif-
ference is not statistically significant.  The upper curve in Figure 1 sug-
gests a decreased incidence of colds by about 60 percent for an intake of
3g per day.  The reported result, Gcpercent decrease, causes the hypothesis
that the protective effect is this great to be rejected with statistical signifi-
cance (P(one-tailed) < 0. 01).  The investigators concluded that there is no

evidence that the administration of ascorbic acid has any value in the pre-
vention or treatment of colds produced by five known viruses.  In fact,
their study rejects with statistical significance (P(one-tailed) ~0.05) a
protective effect greater than 45 percent, but not a smaller effect.

I conclude that it is probable that under the conditions of the study


- 32 -

carried out by Walker, Bynoe,. and Tyrrell ascorbic acid in the amount
                                          *

3g per day does not have protective effect as large as 45 percent.  The

extrapolation of the upper curve of Figure 1 to a value of about 60 per-

cent decrease in incidence seems to me to be reasonable.  The results

of this investigation accordingly disagree with those of the five investi-
gaticns represented by the points in Figure 1.

    A possible explanation of the disagreement is that the process of
inoculation with a virus suspension introduces so many virus particles
into the nose and throat as to overcome the protective effect of the ascor-

bic acid.  This possibility could be checked by studies in which the num-

ber of virus particles used for inoculation was varied.

Other Investipations

    There is some additional evidence that the protective power of
ascorbic acid increases with increase in the magnitude of the viral in-

fection.  Dr. E. Regnier (13) has reported the results of. a blind study of

137 colds, in 22 subjects (mostly professionals), over a five'$ar period..
Some colds were treated by administration of 600 mg of ascorbic acid
every three hours, beginning at the first sign of the cold.  Of 84 incipient
colds treated in this way, only 8 developed into full-blown colds, whereas
of 53 treated with a placebo, 50 developed into full-blown colds.  The dif-

ference has high statistical significance (P <8.001). Other investigators
have reported that a similar treatment is effective in stopping a cold.
Wood (14) recommends taking 1000 mg as soon as one says to himself,
"1 think I am catching a cold, " followed by 500 mg every two hours during

.waki.ng periods for a total of 4 or 5 g per day, and Stone (15) recommends

a succession of 1. 5-g doses at l-hour intervals, beginning at the first sign


- 33 -

of a cold. All three report that the treatment is unsuccessful if it is de-
layed.  I surmise that ascorbic acid can control infection with a small
number of virus particles, such as at the beginning of a cold, but cannot
control a large number, such as when the cold has developed.

   Some studies of the effect of treatment of the common cold'with
ascorbic acid beginning after the cold has developed have given negative
results.  All of these have involved smaller amounts than recommended
by Re gnier, Wood, and Stone, and with less effort to have the treatment
begin at the first sign of a cold.  An example is the work by Tebrock,
Arminio, and Johnston (16), who studied about 1900 subjects, and found
no difference between the colds treated with ascorbic acid and those
treated with a placebo.  The treatment was begun when the subject came
to the dispensary.  He then received one 50-mg tablet of asc'orbic acid
(or placebo), and was given eleven others to be taken over a three-day pe-

riod.  It seems likely that the negative result was caused by the small
amount of ascorbic acid used and by the delay in initiating the treatment.

   It is worth while to quote two sentences from a publication by Dr.
Frederick R. Klenner, who for 27 years has used ascorbic acid for the
treatment of all virus infections: "I have several hundred patients who
have taken log or more of vitamin C daily for three to fifteen years.
Ninety percent of these never have colds; the others need additional
ascorbic acid" (17).  The reported decrease in the incidence of the com-
mon cold in this uncontrolled study, 90 percent by log per day of ascorbic
acid, is not unreasonable in comparison with the upper curve of Figure 1.


- 34 -

Conclusion

An analysis has been made of published results of controlled studies

of the effect of ascorbic acid on the incidence, severity, and integrated
morbidity of the common cold in populations receiving the ascorbic acid
regularly, beginning before colds have been incurred, and with the subjects
exposed to cold viruses in the ordinary way (contact with other people).
The observations reject with high statistical significance the null hypothesis
that under these conditions ascorbic acid has the same effect as a placebo.
Ascorbic acid in the daily amount 200 mg decreases the incidence of colds
and the severity of individual colds by about 20 percent and the integrated
morbidity by about 35 percent.  In the daily amount 1,000 milligrams it de-
creases the incidence and the severity by about 40 percent and the integrated
morbidity by about 60 percent.  No controlled study under these conditions
has given results rejecting with statistical significance the hypothesis that
this amount of protective effect occurs.

     `he effectiveness of ascorbic acid taken after a cold has been in-
curred seems to depend upon the amount of ascorbic acid taken and the amount
of delay in beginning the treatment.
    Little protective effect is reported when the colds are induced by
inoculation with a virus suspension.
    There is evidence that ascorbic acid has a protective effect also against

infections other than the common cold.


References and Notes

1. Linus Pauling, Science 160, 265 (1968).

2. Linus Pauling, Vitamin C and the Common Cold, W. H. Freeman and
Company, San Francisco, 1970; also Proc. Nat. Acad. Sci. U. S. 67, 1643 (1970).,
                                                                           -

3. G. Ritzel, Helv. Med. Acta 28, 63 (1961).
                                     *-

4. Anonymous, Nutr. Rev., 25, 228 (1967).
                                 -
5. D. W. Cowan,' H. S. Diehl, and A. B. Baker, J. Am. Med. Ass. 1%
1267 (1942).

6. W. L. Franz, G. W. Sands, and H. L. Heyl, J. Am. Med. ASS. 162, 1224
(1956).

7. C. W. M. Wilson and H. S. Low, Acta Allergolica g, 367 (1970).

8. Letter' from Professor C. W. M. Wilson to L. Pauling, 19 March 19`71.

9. A. J. Glazebrook and S. Thcmson, J. Hygiene 42, 1 (1942).
                                                     --

10. G. Dahlberg, A. Engel, and H. Rydin, Acta Med. Stand. 119, 540 (1944).

11. F. Bicknell and F. Prescott, The Vitamins in Medic.ine, Grune and St;ratten,
New York, 3rd Ed. 1953, p. 475.

12. G. H. Walker, M. L. Bynoe, and D. A. J. Tyrrell, Rrit. Med. J. 1,
603 (1967).

13. E. Regnier, Rev. Allergy 22, 835, 948 (1968).

14. H. C. Wood, Jr., Calories, Vitamins and Common Sense, Belmont
Books, New York, 1962, p. 45.

15. I. Stone, Ref. 2, p. 46.

16. H. E. Tebrock,. J. J. Arminio, and J. H. Johnston, J. Am. Med. Ass.
162, 1227 (1956).
--

17. F. R. Klenner, Med. World News, 12, No. 20, 12 (1971)
                                              -


1.00

0.50

1.00

0.50

IbiClDENCE

OF

*
*    OR

SEVER

THE

I



\,

COMP

TY
ION CO

LD

l.NTEGRATED

        MORBIDITY


Legend for Figure

Fig. 1.  Points representing observed values of the incidence of the common
cold or the severity of individual colds (above) and of the integrated morbid-
ity (beldw), plotted against the milligrams of ascorbic acid taken per day.
All values are relative to the corresponding values of incidence, severity, or
integrated morbidity for control subjects.