Methods

ParticipantsWe recruited 50 volunteers from the Detroit Medical Center (Detroit, Michigan) to participate in a randomized, placebo-controlled trial of the efficacy of zinc acetate lozenges in treating the common cold. Recruitment was started in January 1999 and ended in January 2003. Participants were medical students, house staff, and employees at Wayne State University who were >18 years of age. Participants were informed of the placebo-controlled, double-blind nature of the study, and the study protocol was approved by the Human Investigation Committee of Wayne State University. Each subject was paid $10 for participation and transportation costs. We recruited healthy volunteers who were free of any illness for various laboratory tests as control subjects

Volunteers were recruited if they had had cold symptoms for 24 h or less and had at least 2 of the following 10 symptoms: cough, headache, hoarseness, muscle ache, nasal drainage, nasal congestion, scratchy throat, sore throat, sneezing, and fever. We excluded persons who were pregnant, had any known immune deficiency disorder or chronic illness, had had symptoms of the common cold for >24 h, or had previously used zinc lozenges to treat the common cold, to assure blinding. In general, subjects were recruited during fall and winter months. Subjects with history of allergies were not excluded

We chose a sample size of 50 subjects in order to detect a 50% decrease in the mean duration of symptoms, from 8 days in the placebo group to 4 days in the zinc group with a SD of 2 days, using 2-sided P⩽.05 and an approximate power of 80% or more. In view of our earlier findings, a 50% reduction in the duration of cold symptoms in the treatment group was considered to be a reasonable end point [16]

InterventionThe lozenges were cherry oil–flavored Fast Dry [18] zinc acetate lozenges, manufactured by F & F Foods (Chicago, IL). The active lozenges contained 13.3 mg of zinc as zinc acetate in a hard candy that contained 3.8 g of sucrose and corn syrup and that was prepared using the open-pot batch method, with the active ingredient added last. One hundred percent of the zinc was available at physiologic pH 7.4 in positively charged, ionic form. The placebo lozenges were of identical composition, except that they contained 0.25 mg of sucrose octaacetate rather than the active ingredient, zinc. There were no fats, metal chelators, or other zinc ion–binding agents in either the active or placebo lozenges. The placebo and zinc lozenges were identical in weight, appearance, flavor, and texture and were supplied by George Eby

A research consultant prepared the randomization code and the packages of medication [21]. The packages were identical in appearance except for the randomization numbers. A research assistant who was blinded to treatment assignments distributed the study medication. Participants were given 50 lozenges and were asked to dissolve 1 lozenge in their mouth every 2–3 h while awake for as long as they had cold symptoms. They were instructed to take no other cold preparations during the study period

Outcome measuresOur primary end point was the average duration of cold symptoms. Secondary end points were plasma levels of (1) zinc; (2) soluble interleukin (IL)–1 receptor antagonist (sIL-1ra) and soluble tumor necrosis factor (TNF) receptor (sTNF-R) 1; and (3) the plasma adhesion molecules, soluble vascular endothelial cell adhesion molecule (sVCAM)–1 and soluble ICAM (sICAM)–1

Participants were asked to complete a daily log documenting the severity of symptoms and the medications taken throughout the duration of the cold. Every day the subjects graded each symptom as 0 for none, 1 for mild, 2 for moderate, or 3 for severe. Total symptom scores were calculated by summing the scores of the 10 symptoms for each day. Resolution of cold symptoms was defined as resolution of all symptoms (a total symptom score of 0) or resolution of all but 1 mild symptom (a total symptom score of 1). The participants were not asked to rate their overall illness or level of severity in order to make their assessments unbiased

Plasma samples were collected and assayed for levels of zinc, sIL-1ra, sTNF-R1, sVCAM-1, and sICAM-1. Zinc was assayed by using methods established in our laboratory that are based on flameless atomic absorption spectrophotometry [22, 23]. Every precaution was taken to avoid contamination during collection, preparation, and analysis. Variables were analyzed using Quantikine ELISA kits for sIL-1ra, sTNF-R1, sVCAM-1, and sICAM-1 (R&D Systems)

To assess adverse effects of the treatment, participants were given a questionnaire to fill out at the end of the trial. Participants provided yes-or-no answers to questions about nausea, abdominal pain, vomiting, diarrhea, constipation, sweet taste, sour taste, bitter taste, aftertaste, dry mouth, mouth irritation, and bad taste

Participants returned to the clinic on the fifth day for a blood sampling and again for the final visit within 1 day of resolution of cold symptoms. At this time they returned unused lozenges; lozenges were collected to check adherence and to confirm that cold symptoms had resolved

Maintenance of blindingComparability in taste between zinc and placebo was tested in the participants at the beginning and the end of the trial. The participants filled out a questionnaire in which they were asked to guess whether they had received zinc or placebo lozenges. They had 5 choices: certainly placebo, certainly zinc, do not know, probably placebo, and probably zinc. Subjects who selected certainly or probably and were correct about the type of lozenges they received were considered to be correct. We therefore categorized participants as correct, incorrect, or do not know

Statistical analysisWe compared the change in outcomes before and after intervention for the zinc and placebo groups. If the changes were normally distributed in both groups, as determined by the Shapiro-Wilk test, we used the unpaired t test to compare mean changes. If the changes were not normally distributed, the differences were examined using the nonparametric Wilcoxon&rank-sum test. Multivariate analysis of variance with repeated measures was used to determine the effect of treatment × time on severity scores. Fisher’s exact test was used to determine group differences in adverse effects. χ² analyses were performed to determine group differences in correctly identifying lozenges at baseline and after treatment. Statistical analyses were completed using JMP IN software (version 5.1.2; SAS Institute) on a MacBook Pro Computer

Results

DemographicsTable 1 shows the demographic characteristics of the study subjects

Duration and severity of cold symptomsTable 2 shows the duration and severity of cold symptoms in the zinc and placebo groups. The average duration of cold symptoms was 4.0 days in the zinc group and 7.1 days in the placebo group (P<.0001). The durations of cough, nasal discharge, and muscle ache were significantly shorter in the zinc group than in the placebo group (table 2). In 56% of the zinc-group subjects the cold was completely resolved on day 4, and no subject in the placebo group was free of cold symptoms on day 4

The mean overall severity scores for cold symptoms are shown in figure 1. Repeated-measures analysis of severity scores indicated a significant effect for treatment × time for the zinc group compared with the placebo group over 10 days (P=.0002). At baseline, the average severity scores for the zinc and placebo groups were 8.32 and 7.78, respectively; by day 4, these scores were 3.45 and 5.61, respectively

Adequacy of blindingIn the zinc group at the beginning of the study, only 1 subject identified the lozenges as certainly zinc, and 2 subjects identified them as probably zinc. Thus, 3 (12%) of 25 subjects in this group were correct. At the end of the study, 2 (8%) were correct; 1 subject identified the lozenges as certainly zinc, and another subject identified them as probably zinc

In the placebo group at the beginning of the study, 1 subject said that the lozenges were certainly placebo, and another subject identified them as probably placebo. Thus, 2 subjects (8%) in this group were correct. At the end of the study, none of the subjects identified the placebo lozenge correctly

Contingency analysis of correctness by group at the beginning of the study yielded P=1.0, by the χ² test (Fisher’s exact test). At the end of the study, the same test yielded P=.489. Thus, there was no significant difference between the 2 groups

From these data, we concluded that the blinding of the subjects was adequate. In addition, we analyzed the effect of clinical variables for the 22 completely blinded subjects in the zinc group and the 23 in the placebo group; these data are shown in table 2. For completely blinded subjects, cold symptoms lasted for a mean ± SD of 3.54 ± 0.96 days in the zinc group and 7.39 ± 0.98 days in the placebo group (P<.0001). The durations of cough and nasal discharge were also significantly shorter in the zinc group (table 2)

Adverse effectsAdverse effects of the zinc and placebo lozenges are compared in table 3. The zinc and placebo groups did not differ significantly in the incidences of any of the adverse effects, including diarrhea, constipation, sweet taste, sour taste, bitter taste, aftertaste, dry mouth, mouth irritation, or bad taste. None of the subjects complained of either abdominal pain or vomiting

Adherence to therapy was determined by lozenge count. The average number of lozenges taken daily was 6.9 in the zinc group and 6.5 in the placebo group

Table 4 shows the results of treatment on plasma levels of zinc, sIL-1ra, sTNF-R1, sVCAM-1, and sICAM-1. The plasma zinc level increased significantly in the zinc group. The plasma sIL-1ra level decreased significantly in the zinc group after treatment but increased in the placebo group. A comparison of the mean changes (before vs. after therapy) in plasma sTNF-R1 and sVCAM-1 levels in the 2 groups showed no significant differences between groups. Plasma sICAM-1 levels decreased significantly after treatment in the zinc group

Discussion

Our results showed that the mean durations of cold symptoms, cough, nasal discharge, and muscle ache were significantly decreased in the zinc group compared with the placebo group. The blinding of therapy was adequate, and our analysis of clinical variables in completely blinded subjects (22 in the zinc group and 23 in the placebo group) showed that the mean durations of cold symptoms, cough, and nasal discharge were also significantly decreased in the zinc group compared with the placebo group in this subset. Zinc treatment was also effective in significantly decreasing the severity score, compared with placebo

We observed that the mean change in plasma sIL-1ra level showed a significant decrease in the zinc group, whereas in the placebo group the mean change in plasma sIL-1ra level showed a positive increase. The between-group difference in mean change was statistically significant. IL-1ra is anti-inflammatory, functions as a specific inhibitor of IL-1α and IL-1β inflammatory cytokines [24], and is produced by monocytes and macrophages [24]. The synthesis of IL-1ra and IL-1β are differentially regulated at their own promoter sites in macrophages and monocytes [24]. Our results suggest that common cold viruses increase oxidative stress, which activates macrophages and monocytes and result in increased production of both the inflammatory cytokine IL-1β (not assayed here) and the anti-inflammatory product IL-1ra. We also assayed another anti-inflammatory cytokine, sTNF-R1, which functions as a specific inhibitor of TNF activity on target tissues [24]. Levels of sTNF-R1 also decreased in the zinc-treated group, but not significantly. Thus, the decrease in sIL-1ra and sTNF-R1 levels in the zinc group only suggests that zinc decreased the oxidative stress, resulting in decreased activation of monocytes and macrophages. Our previous studies have shown that zinc is an antioxidant [23, 25]

Many of the symptoms observed in the common cold resemble the effects of proinflammatory cytokines [24]. Fever, lack of appetite, leukocytosis, hypoferremia, and induction of acute-phase reactant proteins are known effects of IL-1 production by monocytes and macrophages [24]. We have previously shown that IL-1β production by mononuclear cells is increased in zinc-deficient subjects and is normalized by zinc supplementation, suggesting that zinc modulates the proinflammatory cytokines released by monocytes and macrophages [26]

Figure 2 summarizes our hypothesis as to how zinc may be acting as an antioxidant and anti-inflammatory agent and how it decreases ICAM-1 levels. Infection and oxidative stress activate NF-κB, which increases generation and gene expression of inflammatory cytokines (such as TNF-α, IL-1β, and IL-8) and adhesive molecules (such as VCAM-1 and ICAM-1) [25, 27]. Zinc decreases oxidative stress and induces the zinc-dependent transcription factor A20 in monocytes and macrophages, which inhibits NF-κB activation via the TNF-R–associated factor pathway [25]. Zinc decreases oxidative stress by several mechanisms [23, 25]. Zinc is an inhibitor of NADPH oxidase, an enzyme that initiates the generation of free radicals; it is essential for superoxide dismutase and for generation of metallothionein, a low-molecular-weight protein that removes OH ion effectively (figure 2)

Inflammatory cytokines, generated by activated monocytes and macrophages, are also known to generate greater amounts of reactive oxygen species [27]. Zinc supplementation in healthy human subjects aged 20–50 years reduced the concentration of the oxidative stress markers, such as the oxidative stress–related by-products malondialdehyde, 4-hydroxyalkenals, and 8-hydroxydeoxyguanine in plasma; inhibited the ex vivo induction of TNF-α and IL-1β mRNA in mononuclear cells; and provided protection against TNF-α–induced NF-κB activation in isolated peripheral blood mononuclear cells [25]. We previously provided evidence that, in the human promyelocytic leukemia cell line HL-60, which differentiates to the monocyte and macrophage phenotype in response to phorbol-12-myristate-13-acetate, zinc increases the expression of A20 and the binding of A20 transactivating factor to DNA, which inhibits NF-κB activation [25, 28]. NF-κB is involved in the gene expression of TNF-α and IL-1β in monocytes and macrophages in humans and HL-60 cells, and the effect of zinc in inhibiting the gene expression of TNF-α and IL-1β is cell specific [25, 28]. In a recent study, we showed that, in elderly persons (55–87 years old), ex vivo generation of TNF-α and plasma oxidative stress markers was significantly lower in the zinc-supplemented group than in the placebo group [23]. Given that increased activation of NF-κB in monocytes and macrophages due to viral infection increases the generation and gene expression of adhesive molecules, such as ICAM-1, our observation that only the zinc group showed a significant decrease in plasma ICAM-1 levels suggests that zinc decreased the generation and gene expression of ICAM-1. Human rhinovirus type 14 “docks” with ICAM-1 on the surface of somatic cells [4, 19]. Thus, zinc may in effect act as an antiviral agent by reducing ICAM-1 levels. Another possibility is that zinc ions may form a complex with ICAM-1, preventing the binding of rhinovirus to cells [19]

Tremacamra, an sICAM-1 drug that functions as a receptor blockade, has been used to treat experimentally induced rhinovirus infection [20]. This drug was effective in decreasing the severity of common cold symptoms in the experimental model, but it was not clear whether the duration of symptoms was also decreased [20]. In our study, zinc acetate lozenges decreased both the severity and the duration of the common cold. Zinc is less expensive than tremacamra, is relatively nontoxic if ingested in the recommended dosage, and is nonmutagenic. The toxicity and mutagenicity of tremacamra remain to be ascertained

The ICAM-1 levels at baseline were higher in the zinc group than in the placebo group. Interestingly, the zinc group had more severe cold symptoms than the placebo group at the outset, suggesting that greater severity may have been responsible for higher plasma levels of ICAM-1. More studies are needed to prove this hypothesis

We conclude that zinc acetate preparation, as used in our study, was significantly effective in decreasing the mean duration of cold symptoms. Our data also show that activation of monocytes and macrophages was decreased by zinc, most likely due to its antioxidant effect. The level of ICAM-1 also showed a significant decrease after zinc therapy, probably due to decreased NF-κB activation by zinc-induced A20 zinc finger protein. We propose that the beneficial clinical effects seen in the zinc group were due to the antioxidant and anti-inflammatory effects of zinc. We also suggest that a decrease in plasma ICAM-1 levels due to zinc therapy may have decreased the docking of the cold viruses on the surface of somatic cells