Customize your JAMA Network experience by selecting one or more topics from the list below.
Macknin ML, Piedmonte M, Calendine C, Janosky J, Wald E. Zinc Gluconate Lozenges for Treating the Common Cold in ChildrenA Randomized Controlled Trial. JAMA. 1998;279(24):1962–1967. doi:10.1001/jama.279.24.1962
Context.— The common cold is one of the most frequently occurring illnesses and
is responsible for substantial morbidity and economic loss. Biochemical evidence
suggests that zinc may be an effective treatment, and zinc gluconate glycine
(ZGG) lozenges have been shown to reduce the duration of cold symptoms in
Objective.— To determine the efficacy of ZGG treatment of colds in children and
Design.— A randomized, double-masked, placebo-controlled study.
Setting.— Two suburban school districts in Cleveland, Ohio.
Patients.— A total of 249 students in grades 1 through 12 were enrolled within
the first 24 hours of experiencing at least 2 of 9 symptoms of the common
Intervention.— Zinc lozenges, 10 mg, orally dissolved, 5 times a day (in grades 1-6)
or 6 times a day (in grades 7-12).
Main Outcome Measures.— Time to resolution of cold symptoms based on subjective daily symptom
scores for cough, headache, hoarseness, muscle ache, nasal congestion, nasal
drainage, scratchy throat, sore throat, and sneezing.
Results.— Time to resolution of all cold symptoms did not differ significantly
between students receiving zinc (n=124) and those receiving placebo (n=125)
(median, 9 days; 95% confidence interval [CI], 8-9 days; median, 9 days, 95%
CI, 7-10 days, respectively; P=.71). There were no
significant differences in the time to resolution of any of the 9 symptoms
studied. Compared with controls, more students in the zinc group reported
adverse effects (88.6% vs 79.8%; P=.06); bad taste
(60.2% vs 37.9%; P=.001); nausea (29.3% vs 16.1%; P=.01); mouth, tongue, or throat discomfort (36.6% vs 24.2%; P=.03); and diarrhea (10.6% vs 4.0%; P=.05).
Conclusions.— In this community-based, randomized controlled trial, ZGG lozenges were
not effective in treating cold symptoms in children and adolescents. Further
studies with virologic testing are needed to clarify what role, if any, zinc
may play in treating cold symptoms.
THE COMMON cold is one of the most frequently occurring illnesses in
the world. More than 200 viruses can cause common colds in adults, including
rhino viruses (the most frequent cause), coronaviruses, adenoviruses, respiratory
syncytial virus, and parainfluenza viruses. In the United States each year,
adults have an average of 2 to 4 colds and children have an average of 6 to
8 colds.1,2 The morbidity resulting
from these illnesses and the subsequent financial costs in terms of loss of
time from work are substantial.3 Previously
described treatments have provided neither consistent nor well-documented
relief of symptoms. However, even a treatment that is only partially effective
in relieving cold symptoms could markedly enhance recovery and reduce economic
losses in a large population.
Different formulations of oral zinc have been evaluated as a cold remedy.
Ten double-masked, placebo-controlled clinical trials have been reported,
but each used a different dose or formulation of zinc. Five of these studies
showed that zinc had a beneficial clinical effect and 5 found no effect.4-11
Zinc treatment had no effect on viral shedding.5,8
All studies reported to date have been performed on adults.
The objective of this study was to determine the efficacy of zinc gluconate
lozenges in reducing clinical symptom scores in children and adolescents with
the common cold.12,13
This study was similar in design to a previous study of zinc for cold
treatment in adult patients at the Cleveland Clinic Foundation.7
The current study is a prospective, randomized, double-masked, placebo-controlled
investigation designed to assess whether zinc gluconate glycine (ZGG) lozenges
would reduce the time to resolution of cold symptoms in children (Figure 1).
The study was approved by the Cleveland Clinic institutional review
board. Presentations were made to school administrators and school boards
in the communities, and their permission to conduct the study in the school
systems was obtained.
Students were recruited from the Beachwood and Mayfield school districts
in the eastern suburbs of Cleveland, Ohio, during the winter cold season,
from October 7, 1996, through March 13, 1997. Before enrollment began, students
willing to participate in the study were identified by responses to a consent
form and cover letter from school administrators and the principal investigator,
which were mailed to all parents and guardians of children in the study schools.
The principal investigator and study personnel attempted to contact the family
of every potential enrollee by telephone to answer any questions about the
study. The principal investigator spoke about scientific research in general
and the study in particular to most science classes in the participating high
schools and middle schools. School-based study personnel collected consent
forms. As an incentive, all students who returned signed consent forms were
entered into a raffle with a grand prize of a trip for 4 to a popular theme
park or the cash equivalent. Students did not have to become ill and be enrolled
in the study to be eligible to win the raffle. Only students with informed
consent forms on file at school and signed by their parents or guardians (for
students younger than 18 years) were eligible for enrollment in the study.
Study personnel enrolled patients before school, at lunchtime, and at the
end of the school day, or in the patients' homes if they were contacted on
nonschool days, during the first 24 hours of the student's cold symptoms.
Students were required to report having at least 2 of the following
9 symptoms: cough, headache, hoarseness, muscle ache, nasal congestion, nasal
drainage, scratchy throat, sore throat, or sneezing. Students were excluded
if they had an oral temperature greater than 37.7°C, had previously taken
the zinc preparation (Cold-Eeze, Quigley Corporation, Doylestown, Pa), were
pregnant, had a known adverse reaction to zinc, or had a known immune deficiency.
Other reasons for exclusion were an acute illness other than the common cold
(eg, pneumonia, gastroenteritis) or cold symptoms lasting more than 24 hours.
A computer-generated randomization code was provided to the pharmacist,
who held the code and prepared the packages of medication. The packages were
identical in appearance, except for the identifying code number, and were
distributed to the study personnel, all of whom were masked to the group assignments.
Students had 2 packages of identical medication, 1 for home and 1 for school.
Home medication was delivered by study personnel to the students' homes, while
a parent or guardian was present, on the day that the student was enrolled
in the study.
All students were asked to take 3 lozenges per day in the school study
personnel offices: before school, at lunchtime, and before school was dismissed.
If the students did not come to the office to receive their medication, the
study personnel went to their classes to distribute it. If students missed
school, they were instructed to take their regularly scheduled medications
from their home medication package. Students in grades 1 through 6 were instructed
to take 2 lozenges at home on school nights and 5 lozenges per day at home
on weekends. Students in grades 7 through 12 were instructed to take 3 lozenges
at home on school nights and 6 per day at home on weekends. Students were
instructed to let the lozenges dissolve in their mouths and not to chew them.
Patients were asked to take study lozenges until their cold symptoms had been
completely resolved for 6 hours.
Adherence was assessed by a daily diary of the medication taken and
by the number of lozenges returned at the end of the study. Whenever there
was a discrepancy between the diary and medication returned, the number of
lozenges returned was used. Based on the duration of the cold, the number
of prescribed lozenges was calculated. Adherence was defined as taking at
least 70% of the prescribed medication.
Students were given exactly enough lozenges for 3 weeks of treatment,
126 and 105 lozenges for secondary and elementary students, respectively.
However, students were followed up until their cold symptoms resolved, even
if their symptoms persisted beyond 21 days. They were asked to take no other
cold preparations, if possible, during the study. Oral digital thermometers
were given to students at the time of enrollment. All students had a brief
examination at the time of enrollment by trained study personnel, who confirmed
the presence of at least 1 sign of a cold (cough, hoarseness, nasal drainage,
nasal congestion, throat redness and exudate, enlarged tonsils, and sneezing),
and another brief examination at discharge from the study. Cold symptoms,
adverse effects from medications, and other medications taken were recorded
daily on school days by the study personnel who distributed the lozenges.
On nonschool days and missed school days, students phoned all information
into a voice mail recording. If study personnel did not receive a voice mail
message, they called students at home.
Students graded each symptom on a numerical scale of 0 to 3 each day,
but parents occasionally questioned the symptom rating assigned by students.
In cases of dispute, the parent's evaluation was used instead of the child's.
Students reported daily on the severity of 9 symptoms: cough, headache, hoarseness,
muscle ache, nasal drainage, nasal congestion, scratchy throat, sneezing,
and sore throat. Severity for each symptom was measured as either none (0),
mild (1), moderate (2), or severe (3). The overall severity score was computed
as the sum of severity scores of all 9 symptoms, yielding a number between
0 and 27. Resolution of the cold was defined as the time at which the total
severity score reached 0, indicating the absence of all symptoms. On day 2
and on the final day of the study, students were asked to guess whether the
medication they were taking was "active drug, placebo, or don't know." We
identified adverse effects by each day asking students an open-ended question
about adverse effects and by offering a list of potential adverse effects
to choose from at the conclusion of the study.
The ZGG and placebo lozenges were supplied by the Quigley Corporation.
The zinc lozenges consisted of a hard-candy base prepared with approximately
equal proportions of sucrose and corn syrup, zinc gluconate trichydrate (AKZO
Chemie, Amersfoort, the Netherlands), a molar proportion of glycine (aminoacetic
acid), and cherry flavoring oils. The mixture was formed into lozenges that
weighed 3.75 g and contained 10 mg of zinc. Placebo lozenges were prepared
from the same flavored hard-candy base and contained calcium lactate pentahydrate
instead of zinc. Placebo and active lozenges were as identical as possible
in appearance, flavoring content, and texture.
The objective of the current study was to assess whether the response
to ZGG in children and adolescents with presumed viral upper respiratory tract
infection was superior to placebo. The projected sample size of 220 ensured
a power of at least 98% to detect differences that were at least similar to
or were greater than differences found in our study in adults—a 42%
reduction in the median time of resolution of all cold symptoms with zinc
therapy (7.6 days in the placebo group vs 4.4 days in the zinc group).7 This sample size also ensured a power of at least
80% to detect differences of the same magnitude (42%) separately in elementary
school patients (grades 1-6) and in junior and senior high school patients
(grades 7-12). Power calculations assumed that accrual would occur over a
6-month period with 21 days of follow-up per patient. The design ensured that
the 2-sided, type I error rate would not exceed 0.05.
Ten students' (3 placebo, 7 active) colds did not resolve during the
period of observation. Thus, statistical methods for incomplete data were
used. Resolution rates were calculated using the method of Kaplan and Meier,14 and 95% confidence intervals (CIs) for the estimates
of median time to resolution were calculated using the method of Brookmeyer
The primary analyses of resolution time were performed using Cox proportional
hazards regression models16 and the method
devised by Efron17 to adjust for the large
number of patients with identical resolution times. The suitability of a proportional
hazards model was first assessed by modeling resolution time as a function
of treatment group and as a time-dependent factor representing an interaction
between treatment group and time. Although the observed hazards were not strictly
proportional, the extent of nonproportionality was not statistically significant
(P=.64). Two students, who were determined to be
ineligible after assignment, were included in these analyses, in keeping with
the intent-to-treat principle, and were treated as censored observations with
a cold duration of 0.001 days. Estimates of the effect of treatment group
assignment on the probability of school absence on a given day were obtained
using generalized estimating equations.18
Adverse effects, success of masking, medication use, and adherence were
compared between groups using χ2 tests, unless the expected
cell frequencies were small, in which case the Fisher exact test was used.
In all cases, P values of .05 or less were considered
to be statistically significant. All statistical analyses were performed using
software from the SAS Institute, Cary, NC, and were independently verified
by 2 biostatisticians (M. P. and J. J.).
A total of 249 students received either ZGG lozenges (n=124) or an identically
packaged placebo (n=125). Baseline characteristic were similar between the
2 groups (Table 1). Distributions
of race and sex within the students on the study (overall, 92.4% white, 4.0%
black, and 3.6% other, and 52.2% female) were similar to those of the entire
school population (88.0% white, 7.3% black, and 4.7% other, and 48.4% female).
Baseline characteristics of the students, including prevalence of allergies,
proportion who smoked, frequency of colds and other infections, and temperatures
between 37.1°C and 37.7°C were similar between the groups (Table 1). The only characteristic for which
there was an imbalance was asthma. Seventeen (14%) of 120 patients in the
placebo group and 9 (7.5%) of 120 in the zinc group reported a history of
asthma (P=.10). (Cox regression analysis revealed
that this imbalance had no significant effect on time to cold resolution.)
Medications used by the groups at enrollment were similar. However,
31 subjects (25.0%) in the zinc group and 20 subjects (16.0%) in the placebo
group were taking vitamins or mineral supplements (P=.08).
At enrollment, all students were asked to discontinue taking any zinc-containing
vitamins or mineral supplements during the course of the study.
Two students reported that they deliberately provided false information
at enrollment because they wanted to participate in the study with their friends
who had colds. In keeping with the intent-to-treat principle, these students
were included in the primary analyses of time to resolution of symptoms, but
they were excluded from secondary analyses and from the tables involving initial
symptoms. The proportion of students with each initial symptom, and the results
of comparisons between groups are shown in Table 2. The distributions of symptoms in the 2 groups were similar,
except that fewer students in the zinc group presented with cough (71.8% vs
The median score for overall severity of initial symptoms, computed
as the sum of the initial scores for each symptom, was 10 (range, 3-22; mean±SD,
10.1±3.9) for the placebo group, vs 9 (range, 2-22; mean±SD,
9.2±3.9) for the zinc group. This difference was statistically significant
(P=.03), but not clinically important because an
increase of 1 point in total symptom score represents an increase of 1 level
of severity for 1 symptom or 1 additional mild symptom. The severity score
6 hours later was available for 231 patients; the placebo group had a mean
score of 8.7±4.35 and a median of 8, and the zinc group had a mean
of 7.7±4.1 and a median of 7 (P=.09).
The median time to resolution of all cold symptoms was 9.0 days (95%
CI, 8-9 days) in the placebo group and 9.0 days (95% CI, 7-10 days) in the
zinc group (P=.71; Figure 2). In the elementary grades, 57 students who received placebo
and 56 students who received zinc experienced resolution of all symptoms in
a median of 9.0 days (95% CI, 8-11 days) and 8.0 days (95% CI, 6-11 days),
respectively (P=.44). In the junior and senior high
schools, 68 students who received placebo and 68 students who received zinc
had a median time to resolution of all symptoms of 8.5 days (95% CI, 7-9 days)
and 9.5 days (95% CI, 7-10 days), respectively (P=.83).
The lack of statistical differences between the groups remained (P=.73) after adjusting for age and initial severity of illness level.
For 8 students (5 aged 7 years, 2 aged 8 years, and 1 aged 9 years)
in whom their symptom ratings disagreed with their parents' ratings, the parents'
evaluation was used.
Separate models were also fit to assess whether the resolution time
of individual symptoms was related to treatment group. Because the presence
and severity of individual symptoms often fluctuated during the course of
the cold, for these analyses, the symptom was considered to be resolved when
the score for that symptom reached 0 for the last time, or until the last
day the patient was seen if it had not resolved. Treatment groups had no significant
effect on the time for resolution of any of the individual symptoms (Table 3).
There were a total of 85 days of school absence in the 2454 days (1260
placebo, 1194 active) the students were in the study, including 53 days of
absence (among 26 children) in the placebo group, and 32 days of absence (among
23 children) in the zinc group. Children taking zinc were therefore less likely
to be absent than children taking placebo (odds ratio, 0.60; 95% CI, 0.32-1.13),
but this difference was not statistically significant (P=.12).
Slightly more students in the zinc group (n=109) experienced at least
1 adverse effect than in the placebo group (n=99). The students who received
zinc experienced significantly more bad taste reactions; nausea; mouth, tongue,
or throat irritation; and diarrhea than those in the placebo group; there
were no significant differences in the frequency of vomiting, abdominal pain,
constipation, dizziness, headache, or dry mouth between the groups (Table 4).
The median percentage of prescribed lozenges taken was 83.3% overall
and did not differ significantly between groups (83.3% in the placebo group
and 82.5% in the zinc group, P=.45). Overall, 74.1%
(183/247) of subjects took at least 70% of the medication prescribed: 73.4%
(91/124) of the placebo group and 74.8% (92/123) of the zinc group (P=.80). Forty-six percent (57/124) of the placebo group
and 47.2% (58/123) of the zinc group reported taking more lozenges, by a median
of 6 lozenges in both groups, than verified by pill counts; no patients underreported
the number of lozenges taken. The extent of misreporting was not significantly
different between the groups (P=.29 by Wilcoxon rank
If the zinc lozenges had a beneficial effect, students in the zinc group
with the highest adherence rates would be expected to have the shortest duration
of symptoms; however, the proportional hazards regression model found no statistically
significant association (P=.36) between adherence
and duration of symptoms, and there was also no statistically significant
association (P=.33) between the dose of zinc per
body surface area per day and duration of symptoms. Excluding all nonadherent
patients from the data analysis did not change the results that the median
time to resolution of all symptoms was 9 days (95% CI, 8-10 days) in the placebo
group and 9 days (95% CI, 7-10 days) in the zinc group and was not statistically
Students were asked to indicate whether they thought they were taking
the active drug, the placebo, or whether they didn't know on day 2 and at
the conclusion of the study. Defining the guesses on day 2 and at the end
of the study as either correct or incorrect (which included the response of
"don't know"), 35% (85/242) of the patients guessed correctly on day 2. A
significantly higher proportion of students receiving zinc (46% [55/119])
guessed correctly than did controls (24% [30/123]); P=.001)
on day 2. At the end of the study the results were similar, with 56% (67/119)
of students receiving zinc guessing correctly vs 42% (51/123) of controls
(P=.02). Six patients who had previously taken Cold-Eeze
were inadvertently enrolled in the study. Of these 6, 1 of 3 taking placebo
and 1 of 3 taking zinc correctly identified their study medication. When we
performed the analysis excluding students with the "don't know" responses
(n=103 students on days 2 and 41 at the end of the study), the results similarly
showed that students who received zinc were more likely to guess their group
assignment than those receiving placebo.
Ten previous double-masked, placebo-controlled studies of zinc for treatment
of the common cold have been reported.4-11
Half of these studies reported beneficial effects of zinc4-7
and half did not.8-11
The major criticisms of the studies with negative results are that the formulations
of zinc used may inactivate zinc salts, the studies had small sample sizes,
and too low a dose of zinc was used. Studies with positive results have been
criticized for inadequate masking because of poor taste matching of placebo
and zinc medications, too many patients being excluded from data analysis,
small sample sizes, and subjective outcome measures. The controversies over
the efficacy of zinc treatment for the common cold are summarized in a recent
meta-analysis.19 The dosages and formulations
of zinc used, clinical settings, number and type of patients, and possible
shortcomings and results of these studies have varied widely.19
The mechanisms by which zinc may affect the common cold remain to be
determined, but several possibilities have been suggested. Zinc prevents the
formation of viral capsid proteins, thereby inhibiting in vitro replication
of several viruses, including rhinovirus.20-24
Zinc ions combine with the carboxyl termini (negatively charged canyons) of
rhinovirus coat proteins, which may prevent the virus from combining with
the tissue-surface protein (intracellular adhesion molecule type 1) and entering
the cell. Inhibition of entry of virus into the cell stops further reproduction.25,26 Extracellular zinc also may exert
antiviral effects by stabilizing and protecting cell membranes by an unknown
In vitro studies have suggested that zinc may induce the production of interferon.31 Zinc ions also have properties that inhibit human
prostaglandin metabolism at 0.01 to 0.1 mmol,32
which may also allow zinc to help relieve symptoms of the common cold.
Two studies with different doses of the same formulation of ZGG lozenges
in adults found a 42% decrease in the duration of symptoms with zinc treatment6,7 compared with placebo. The discrepant
results between these studies in adults and the current study in children
may be explained by the different dosages or flavoring of the formulation,
the ages of the subjects, the time of year when the studies were performed
(ie, the viruses involved may have been different), or chance differences
between the placebo and zinc groups.
The first study in adults of the same formulation of ZGG that we studied
for the common cold was conducted in college students.6
The dose of 23.7 mg was given about 8 times per day; this was more than 3
times the dose used in the current study. The second study in adults was conducted
in hospital employees.7 A dose of 13.3 mg was
given approximately 6 times per day. In the current study, a dose of 10 mg
was administered 5 or 6 times a day, so that the children's doses would be
approximately proportional in milligrams per square meter of body surface
area to the dose used in the second study7
in adults. Also, the dosage of zinc used in the first study in adults achieved
intraoral zinc concentrations well in excess of those needed to inhibit rhinovirus
in vitro.20,21 Despite what seems
to be a sufficient dosage, the dosage in our current study may have been too
low for children and adolescents. Because the mechanism(s) of action of zinc
in treating the common cold is unknown, the optimal dose of medication is
also unknown. In another study of zinc gluconate performed in adults, no beneficial
effect was found using a dose of 4.5 mg.11
The current study used cherry-flavored lozenges, whereas the adult study used
lemon lime–flavored lozenges. It is possible that the cherry flavoring
in some unknown way inactivated the zinc.
Mild subclinical zinc deficiency, present predominantly in adults, may
impair cellular immunity; accordingly, zinc supplementation might enhance
cellular immunity in adults.33 Arguing against
this explanation is the fact that 1 of the 2 previous studies using the same
formulation of zinc that we used in the current study was done in college
students6 whose ages were closer to many of
the students in the current study than to the adults in our previous study.7 Furthermore, a study of another zinc preparation reported
increased serum zinc levels without a decreased length of common cold symptoms.8
Students may have been less subject to a "placebo effect" if they were
less able than adults to guess which substance they were receiving. This explanation
seems unlikely because students guessed their group assignments more accurately
than did the adults, and subjects' guesses about which medication they were
receiving were not associated with their response to treatment in either study.
Students' ability to more accurately "break the blind" than adults would,
if anything, be expected to bias results in favor of a beneficial effect of
treatment in the current study.
Adherence is another issue that might have influenced the results. If
zinc were beneficial, but students were less adherent than adults, the beneficial
effect would not be observed. However, the students were more closely monitored
to ensure adherence in this study than in either of the 2 previous studies
performed in adults; both diaries and pill counts in this study reflected
good adherence. Furthermore, there was no statistically significant correlation
between days to symptom resolution and either dosage of zinc in milligrams
per square meter of body surface area per day or adherence.
Viruses may vary in their susceptibility to zinc. Theoretically, zinc
ions may be most effective against rhinovirus,26
which is most prominent at both ends of the respiratory season.2
Our study was performed throughout the cold season. The previous 2 studies
each enrolled patients over approximately 1 month, one near the beginning
and the other near the end of the cold season. Although we did not perform
diagnostic viral studies, rhinovirus almost certainly would not have been
the predominant virus isolated throughout this entire study. In a subgroup
analysis, students who were enrolled in October (n=51) were analyzed separately
and compared with adults who had been enrolled at about the same time of year
in the previous trial of zinc gluconate in Cleveland,7
but the effect of zinc treatment in reducing cold symptoms was nonsignificant
(P=.63). However, the small number of patients involved
(ie, those enrolled in October) may have been insufficient to demonstrate
a difference, even if one were present.
We investigated the effect of asthma on outcome because more students
receiving placebo had asthma then those students receiving zinc, and also
assessed whether the higher rate of vitamin use in the group receiving zinc
influenced the outcome. Cox regression analysis showed that neither asthma
nor vitamin use significantly influenced outcome. Considering the placebo
group had statistically significantly higher symptom scores at baseline and
more coughs (which may be related to the higher incidence of asthma in the
placebo group) at baseline than the zinc group should, if anything, bias the
results toward a beneficial outcome in the zinc group.
The current study has several limitations. First, students, particularly
in the early elementary grades, may not have reliably reported their symptoms.
However, with a few exceptions, it was our impression that the students were
reliable. Second, we did not evaluate objective measures of cold severity,
such as tissue counts or nasal mucus weight, because of the impracticality
of obtaining these measures in our student population, and because we believed
symptom scores were the most important clinical outcome measure. Although
virus cultures or serologic assays might have been desirable, we decided not
to perform these tests because of their cost and because they are seldom performed
in the course of standard care. Also, the placebo lozenges used in our study
were not exactly the same as the zinc lozenges; however, any bias resulting
from patients correctly guessing their assigned study medication would have
favored assessments showing efficacy of zinc lozenge treatment. We also studied
only 1 dose of zinc lozenge (which was lower than doses previously shown to
be effective against cold symptoms), and we studied only 1 formulation of
In conclusion, ZGG lozenges in the dosages studied were ineffective
in relieving cold symptoms in children and adolescents in this placebo-controlled,
randomized, community-based trial. Additional studies in all age groups with
different dosages and formulations of zinc lozenges and with virologic testing
are needed to define what role, if any, zinc has in the treatment of common