The common cold is one of the most frequently occurring human diseases. On average, each person spends approximately 1 year of his or her life with this disease. 1 Although this disease is generally benign, it is the cause of approximately 30% of the absenteeism from work in the United States. 2
Among the many drugs that have been used to prevent common cold episodes or to reduce their duration and intensity of symptoms, vitamin C and zinc are the most common. No fewer than 30 controlled trials involving vitamin C and seven involving zinc can be found in the literature. 3,4 The attempt to find a protective and curative effect of vitamin C and zinc has a strong biological basis, as both nutrients are believed to have effects on the immune system. 5,6 Vitamin C has been reported to increase the proliferative response of T-lymphocytes both in vitro and in human subjects and to increase the production of interferon in cell culture and in mice. 7 Zinc may prevent the replication of rhinoviruses by interfering with capsid protein formation, according to in vitro assays. 8 Despite these promising findings, regularly updated meta-analyses of vitamin C and zinc trials, as well as other reviews, have failed to demonstrate any protective or curative effect on the common cold. 3,4 Some of these inconclusive reviews have been severely criticized, mainly because some of the vitamin C trials included in the review used a placebo, the taste of which could easily be distinguished from that of vitamin C. 9 Part of the controversy regarding the effect of zinc is due to the fact that some of the trials showing no benefit had used excipients with a chelating effect that may have limited the bioavailability of the zinc ion. 10
We carried out a follow-up study to address the question of whether a high intake of vitamin C and zinc in the regular diet has any protective effect on the risk of developing a common cold episode. A cohort study of vitamin C and zinc has several advantages over a randomized trial; it may examine effects of nonpharmacologic doses of those micronutrients, and it may include subjects with other diseases and taking other pharmaceutical drugs. Furthermore, the bioavailability of vitamin C and zinc contained in food may be different from that of the same micronutrients when administered as pharmacologic compounds in clinical trials. Another important advantage of a cohort study is that the data analysis is not restricted to fully compliant subjects, as is often the case in randomized trials.
Subjects and Methods
Study Population and Data Collection
The methods used in this study have been partially described elsewhere. 11
We initiated the study in October 1998 with 4,287 faculty and administrative staff members from five Spanish universities in two regions (Galicia and Canary Islands) who responded to a food frequency questionnaire designed to measure vitamin C and zinc intake in the regular diet. This number represents about 60% of the total university population invited to participate in the study. Nonrespondents or refusals included persons for whom we had the wrong addresses, those who had left the university, and those who declined to participate. A second questionnaire regarding drinking habits, tobacco consumption, stress, other lifestyle variables, and medical history was administered at the same time. We excluded from follow-up those individuals who initially reported a current common cold episode, history of asthma, or history of chronic obstructive pulmonary disease. Participants recorded information on occurrence of the common cold for 1 year using a standardized questionnaire.
To increase the response rate, all questionnaires were anonymous. To link an initial questionnaire with its corresponding follow-up, we used an eight-digit code formed by the date of birth, gender, and job category (faculty or staff) of each participant. Fifteen duplicated codes were excluded from follow-up. The study was approved by the institutional review boards of the five participating universities.
We provided participants with a calendar and instructed them to record, prospectively and on a daily basis during the duration of what they perceived as a common cold episode, the presence of the following eight symptoms of common cold: runny nose, sneezing, nasal congestion, headache, chills, sore throat, cough, and malaise. We asked them to rate the intensity of these symptoms on a four-level scale from no symptom (0) to very intense (3). We summed up the marks given to each of the eight symptoms and obtained for each subject a general severity score that ranged from 0 (none of the eight symptoms was present) to 24 (all symptoms were very intense). Every 10 weeks, we mailed a short questionnaire to the participants and asked them to transcribe the information in their calendars so that we could update our information on common cold symptoms.
We required three criteria to define a common cold episode: presence of rhinorrhea for at least 3 consecutive days, subjective sensation of having experienced a common cold episode, and a minimum symptom score of 12 out of 24 on the peak day of the episode. We selected the cutoff point of 12 because it maximized the sensitivity and specificity of the common cold diagnosis in a concurrent validation substudy (see below). However, our main study results did not substantially change when we used other cutoff points.
We determined the usual intake of vitamin C and zinc by using a 30-item semiquantitative food frequency questionnaire. Participants were asked to report the average frequency with which a standard portion of each food had been consumed during the previous year. Each food frequency was reported with an eight-grade scale ranging from “never” to “more than four times a day.” The food frequency questionnaire was designed specifically for this study with the objective of determining the usual intake of vitamin C and zinc only. Other micronutrients were not considered. To be included in our questionnaire, each food item had to be consumed frequently in our population and had to have a substantial content of vitamin C or zinc, and its use had to vary from person to person . 12 The food list is available upon request to the corresponding author. We used European food composition databases to calculate the content of vitamin C and zinc for each food item on our questionnaire. 13,14 We also asked whether the participant regularly consumed any supplement of vitamin C or zinc, or any multivitamin medication. The amount of daily vitamin C and zinc from supplements was added to that from the diet.
For each subject, the total intake of alcohol per week was calculated as the product of the content in ethanol of a drink of average size, as given by the Spanish Health Survey. 15 Cigarette consumption was assessed using the standard World Health Organization questionnaire. 16 Smoking status categories were: never-smoker, ex-smoker, occasional smoker, and current smoker of 1–19, 20–40, or more than 40 cigarettes a day. We measured four dimensions of psychological stress: negative affect, positive affect, stressful events, and perceived stress. 11
Concurrently with the main study, a sample of 69 participants was randomly drawn from the population of the study to document the validity and reproducibility of common cold diagnosis, vitamin C and zinc intake, stress measures, and alcohol consumption.
We compared the diagnosis of common cold obtained with our questionnaire with that obtained by means of the complete diagnosis method of Beare and Reed. 17 This method uses a 20-item checklist of common cold symptoms and physical signs and thus requires examination by a physician. The severity of each sign and symptom is rated from 0 to 3. Although in the original checklist all items have the same weight, we assigned them unequal weights on the basis of the relative frequency of symptoms and signs during common cold episodes in a large case series. 18 Hence, nasal manifestations contributed 60% to the total score, compared with 25% for lower-airway manifestations and 15% for general manifestations. A patient was considered as experiencing a common cold episode if his or her score was equal to or greater than 150 points (out of a total of 300 points).
For validity assessment of our measure of vitamin C and zinc intake, we compared our food frequency questionnaire with a 12-day diet record. To avoid seasonal variation, these 12 days were distributed as 3 days for each season: 2 nonconsecutive weekdays and 1 weekend day. In addition, for reproducibility assessment, approximately 1 year later we administered the same food frequency questionnaire used at the beginning of the study to the 69 participants included in the validation substudy.
Validation and reproducibility of the stress variables, used as potential confounders or effect modifiers in our analysis, have been explained elsewhere. 11
Each participant contributed person-time from the return of the initial questionnaire to the onset of the first subsequent common cold episode, the termination of the study, or the loss to follow-up, whichever came first. Secondary analyses in which we did not interrupt the follow-up after the first common cold episode but included subsequent episodes yielded similar results and are not presented here. To take into account the clustering of cases resulting from common cold episodes being contagious events, we used negative binomial regression for the determination of adjusted incidence rate ratios (IRRs) and their corresponding 95% confidence intervals (95% CIs). 19 Negative binomial regression was used to estimate models of the occurrence of an event when there was more variation than would be expected under a Poisson distribution. The use of the negative binomial distribution produces broader confidence intervals than the Poisson distribution, as it introduces an additional variance component.
Possible risk factors for common cold that were considered in the analyses as candidates for potential confounders included age, sex, faculty/staff status, university, region, smoking status, contact with children, psychological stress, and alcohol intake. Quartiles of intake of vitamin C and zinc were calculated separately for each gender.
To explore the shape of the curve relating vitamin C and zinc to the occurrence of common cold episodes and to avoid the need for prior specification of the risk function, we fitted a model with smoothing splines 20,21 and adjusted for potential confounders. We set three knots at the boundaries of the exposure categories.
We detected 1,667 first episodes of common cold among 4,272 subjects with a total of 79,240 person-weeks of follow-up between October 1, 1998, and September 30, 1999. The overall incidence of common cold was 1.4 per person-year (1.1 among men and 1.7 among women). Participants were 21–65 years of age, 46% were females, and 67% faculty members.
For zinc, only 22% of the women and 4% of the men of our study had daily intakes that reached the recommended dietary allowance of 8 and 11 mg per day, respectively. 5 In contrast, vitamin C intake in our study population is high; approximately 84% of women and 62% of men had intakes that were equal to or higher than the recommended dietary allowance of 70 and 90 mg per day, respectively. 6 As in the general Spanish population, 22 only a small proportion (less than 8%) of the participants consumed any supplement of vitamin C or zinc.
Table 1 displays the distribution of age, professional status, and other factors by daily intake of vitamin C and zinc.
Vitamin C intake was not related to occurrence of the common cold (Table 2). Relative risk point estimates ranged from 1.0 to 1.1, with narrow 95% confidence intervals; the upper-to-lower confidence limit ratio 23 varied from 1.8 to 1.9 across quartiles of distribution. Except for a weak association in the third quartile (IRR = 1.5; 95% CI = 1.0–1.9), zinc intake was also unrelated to the risk of common cold in our study. The precision of the zinc estimate was very similar to that of vitamin C.
These results did not materially change after restriction to ever-smokers (30% of the participants) or current smokers, or after further adjustment for alcohol, contact with children, psychological stress, university, and geographic region (data not shown).
Because some studies have found that vitamin C may have a protective effect in subjects under heavy stress, 24 we subsequently restricted our analysis to those participants who were on the highest quartile of the distribution of the stress variables measured in our questionnaire. 11 We did not observe any substantial change in the results.
The exposure-effect curves display the relation between daily intakes of vitamin C and zinc and the incidence rate ratio of common cold (Figure 1). For both genders, the vitamin C curves are flat with a slight and unstable increase of the risk for intakes higher than 700 mg per day. Among women, the zinc curve is similar to that of vitamin C with more unstable estimates and incidence rate ratios constantly close to 1. Among men, the incidence rate ratio estimates of zinc intake are also close to 1. For intakes higher than 20 mg per day, these estimates are too unstable to be reliable. The curves were nearly identical after adjustment for each of the four stress variables that proved to be independent risk factors for the common cold 11 and also after adjustment for alcohol intake and smoking habits.
Both history of allergic rhinitis and history of any other upper respiratory tract disease were associated with common cold. The IRRs were 2.7 (95% CI = 2.3–3.2) for history of allergic rhinitis and 3.3 (95% CI = 2.7- 4.0) for other upper respiratory tract diseases. To reduce the possibility of mistaken reports of nasal allergy episodes as common colds, we also restricted the analyses to participants with no history of allergic rhinitis or any other upper respiratory tract diseases. The results were similar to those reported above.
Twenty-two per cent of the participants abandoned follow-up before the end of the study, but the distributions of vitamin C and zinc intakes as well as other factors (sex, age, and faculty/staff status) were similar between participants with incomplete follow-up and those who completed the study (data not shown). We also recalculated the IRRs in two extreme situations. First, we assumed that all subjects lost to follow-up developed a cold within the week after their dropping out of the study. Second, we assumed that none of these subjects ever developed a cold. Again, the results were not substantially altered.
The sensitivity of our diagnosis of common cold was 94%, and the specificity of diagnosis was 84%.
The unadjusted Spearman coefficients of correlation between our food frequency questionnaire and the 12-day diet record were 0.67 for vitamin C and 0.40 for zinc intake. The crude intraclass correlation coefficients for reproducibility were 0.74 for vitamin C and 0.48 for zinc. These results indicate that both vitamin C and zinc are measured with some error. However, this error is probably nondifferential with regard to disease status because information on exposure was collected before the occurrence of the common cold episode. Although we cannot rule out the possibility of bias toward the null value, our results do not provide any evidence of an effect of these micronutrients on the incidence of common cold.
The findings from this prospective study do not provide evidence for a protective effect of a diet rich in vitamin C and zinc against the risk of common cold. However, in this analysis we did not address the effect of these micronutrients on other aspects of this syndrome such as severity or duration of the episodes. Furthermore, plasma ascorbic acid and zinc, which are presumably the true etiologic agents, were not measured in this study, which focused only on the intake of these nutrients.
The validation substudy showed that the quality of the measurement of exposure is similar to other assessments of diet carried out in the Spanish general population that used food frequency questionnaires. 25 The correlation figures for validity, especially that of zinc, would have probably been enhanced had we adjusted for total energy intake. 26 In our study, this adjustment was not feasible, because the food frequency questionnaire we used did not cover the complete diet of individuals but focused only on those food items that are relevant for the determination of vitamin C and zinc intake. A food frequency questionnaire that assesses the complete diet would have been too cumbersome to use concomitantly with the other questionnaires of stress and other lifestyle variables used in our study.
The validation substudy also indicated that the diagnosis of common cold was accurate, hence confirming the previous finding that self-diagnosis of common cold is usually correct because the manifestations are typical. 27
We deliberately excluded any serologic or virologic determination from our case definition criteria. We based our diagnosis on clinical criteria that were validated subsequently. We believe that the public health relevance of the common cold results from the disability caused by its clinical manifestations, and the economic costs due to absenteeism and to the drugs used to mitigate its symptoms. Previous studies of risk factors of the common cold among artificially infected subjects have emphasized the use of seropositivity in their diagnosis, 28 and one of them has found that an intake of vitamin C higher than 85 mg per day may prevent the common cold. 29 However, it is remarkable that, in these studies, only one-third of subjects inoculated with viruses became infected and that only one-third of those infected subjects developed a clinical cold. 30,31 Thus, a diagnosis based on seropositivity would have included a high proportion of common cold episodes that were not clinically detectable (false positives according to our criteria).
Differential reporting of vitamin C and zinc intake between cases and noncases was minimized because the determination of exposure was made before the onset of a common cold episode. We also adjusted for known or suspected risk factors for the common cold and conducted sensitivity analyses to assess the influence of loss to follow-up, but the results did not materially change.
A possible explanation for the lack of protective effect in our study is the fact that the amounts of vitamin C and zinc consumed in the regular diet are far below the doses of these micronutrients that were used in those preventive or curative trials that found a beneficial effect. The great majority of the vitamin C trials used doses higher than 1 gm per day, an amount that may cause adverse consequences in healthy people. 32 The zinc trials used doses close to 180 mg per day, which represents more than four times the tolerable upper intake level. 5
In summary, our study found no evidence for a preventive effect of a diet rich in vitamin C and zinc. If an effect is to be found, it is probably at doses that are far higher than those found in a regular unsupplemented diet.
We thank the following persons who participated actively in the data collection: Carlos Quintas, Rosa Meijide, Lourdes Maceiras, and Milagros Torres.
1. Papadopoulos NG, Bates PJ, Bardin PG, et al
. Rhinoviruses infect the lower airways. J Infect Dis 2000; 181: 1875–1884.
2. Monto AS, Ullman BM. Acute respiratory illness in an American community: the Tecumseh study. JAMA 1974; 227: 164–169.
3. Douglas RM, Chalker EB, Treacy B. Vitamin C for preventing and treating the common cold (Review). Cochrane Database Syst Rev, 2000;(2):CD000980.
4. Marshall I. Zinc for the common cold (Review). Cochrane Database Syst Rev, 2000;(2):CD001364.
5. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc. Washington DC: National Academy Press, 2001; 351–398.
6. Dietary Reference Intakes for Vitamin C, Vitamin E, Selenium, and Carotenoids. Washington DC: National Academy Press, 2000; 95–185.
7. Hemilä H. Vitamin C intake and susceptibility to the common cold. Br J Nutr 1997; 77: 59–72.
8. Geist FC, Bateman JA, Hayden FG. In vitro activity of zinc salts against human rhinoviruses. Antimicrob Agents Chemother 1987; 31: 622–624.
9. Hemilä H. Vitamin C supplementation and common cold symptoms: problem with inaccurate reviews. Nutrition 1996; 12: 804–809.
10. Zarembo JE, Godfrey JC, Godfrey NJ. Zinc (II) in saliva: determination of concentrations produced by different formulations of zinc gluconate lozenges containing common excipients. J Pharm Sci 1992; 81: 128–130.
11. Takkouche B, Regueira C, Gestal-Otero JJ. A cohort study of stress and the common cold. Epidemiology 2001; 12: 345–349.
12. Willett W. Food frequency methods. In: Willett W, Nutritional Epidemiology. Oxford: Oxford University Press, 1998:74–100.
13. EPIC-Spain Group. Food Composition Table. 2nd version. Barcelona: EPIC-Spain, 1996.
14. Holland B, Welch AA, Unwin I, et al. McCance and Widdowson’s The Composition of Foods. 5th ed. London: Royal Society of Chemistry, Biochemical Press, 1991.
15. Ministerio de Sanidad y Consumo. Encuesta nacional de salud de España 1997. Madrid: Ministerio de Sanidad y Consumo, 1999; 331.
16. World Health Organization. The Evaluation of Monitoring of Public Action on Tobacco. Smoke-Free Europe: 3. Copenhagen: WHO Regional Office for Europe, 1987; 14.
17. Beare AS, Reed SE. The study of antiviral compounds in volunteers. In: Oxford JS, ed. Chemoprophylaxis and Virus Infections. vol. 2. Cleveland: CRC Press, 1977; 27–55.
18. Tyrrell DAJ, Cohen S, Schlarb JE. Signs and symptoms in common colds. Epidemiol Infect 1993; 11: 143–156.
19. Greenland S. Introduction to regression modeling. In: Rothman K, Greenland S, eds. Modern Epidemiology. Philadelphia: Lippincott-Raven, 1998; 401–432.
20. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology 1995; 6: 356–365.
21. Figueiras A, Cadarso-Suárez C. Application of nonparametric models for calculating odds ratios and their confidence intervals for continuous exposures. Am J Epidemiol 2001; 154: 264–275.
22. Serra Majem L, Ribas L, Sainz de Bustamante P, López F, Barbachano M. Consumo de suplementos vitamínico-minerales en la población española. Nutr Clin 1996; 16: 7–15.
23. Poole C. Low P
-values or narrow confidence intervals: which are more durable? Epidemiology 2001; 12: 291–294.
24. Hemilä H. Vitamin C and common cold incidence: a review of studies with subjects under heavy physical stress. Int J Sports Med 1996; 17: 379–383.
25. Martin-Moreno JM, Boyle P, Gorgojo L, et al
. Development and validation of a Food Frequency Questionnaire in Spain. Int J Epidemiol 1993; 22: 512–519.
26. Willett W. Commentary. Dietary diaries vs
food frequency questionnaires: a case of undigestible data. Int J Epidemiol 2001; 30: 317–319.
27. Gwaltney JM Jr. The common cold. In: Mandell GL, Benett JE, Dolin R, eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone, 1995; 561–566.
28. Cohen S, Tyrell DAJ, Russel MAH, Jarvis MJ, Smith AP. Smoking, alcohol consumption, and susceptibility to the common cold. Am J Public Health 1993; 83: 1277–1283.
29. Cohen S, Doyle W, Skoner DP, Rabin BS, Gwaltney JM Jr. Social ties and susceptibility to the common cold. JAMA 1997; 277: 1940–1944.
30. Gwaltney JM Jr. Rhinoviruses. In: Evans AS, ed. Viral Infections of Humans: Epidemiology and Control. New York: Plenum Medical, 1997; 815–839.
31. Glaser R, Rabin B, Chesney M, Cohen S, Natelson B. Stress-induced immunomodulation: implications for infectious diseases? JAMA 1999; 281: 2268–2270.
32. Levine M, Rumsey SC, Daruwala R, Park JB, Wang Y. Criteria and recommendations for vitamin C intake. JAMA 1999; 281: 1415–1423.