Abstract

To examine whether intakes of wine, beer, spirits, and total alcohol are associated with the risk of common cold, in 1998–1999 the authors analyzed data from a cohort study carried out in a population of 4,272 faculty and staff of five Spanish universities. Usual alcohol intake was assessed at baseline by means of a standardized frequency questionnaire that was validated in a random sample of the population. The authors detected 1,353 cases of common cold. Total alcohol intake and beer and spirits consumption were not related to the occurrence of common cold, whereas consumption of wine was inversely associated with the risk of common cold. When drinkers of >14 glasses of wine per week were compared with teetotalers, the relative risk was 0.6 (95% confidence interval: 0.4, 0.8) after adjustment for age, sex, and faculty/staff status. The association was stronger for red wine. These results remained unaltered after adjustment for total alcohol intake and for other potential risk factors for common cold. Findings suggest that wine intake, especially red wine, may have a protective effect against common cold. Beer, spirits, and total alcohol intakes do not seem to affect the incidence of common cold.

Common cold is one of the most frequent human diseases and, although its evolution is generally benign, causes the loss of at least 30 million days of work each year in the United States (1). Alcohol consumption has been shown to modulate the immune response and therefore could conceivably influence the risk of developing an episode of common cold. An inverse association between alcohol intake and peripheral blood leukocyte count has also been described (2). Chronic and even acute moderate alcohol intake may increase host susceptibility to infections and alter the inflammatory response through impaired cytokine production (3). In fact, it has been suggested that the beneficial effect of red wine intake on other diseases may partially be due to downregulation of the production of cytokines (4). Furthermore, animal studies have shown that the anti-inflammatory activity of some non-alcohol-related wine products, especially resveratrol, is strong (5).

Despite alcohol's immunosuppressive properties, moderate intake was unexpectedly associated with a decreased risk of common cold among nonsmokers in the only available epidemiologic study known (6). However, this study was conducted on volunteers artificially challenged with rhinovirus; therefore, no prospective data are available on the effect of alcohol consumption on the natural occurrence of common cold. Whether any specific type of alcoholic drink may have a particular benefit is also unknown.

In 1998–1999, we carried out a follow-up study to address the question of whether alcohol intake has any effect on the risk of developing a common cold episode. We also investigated whether such an effect depends on the type of alcoholic beverage consumed.

MATERIALS AND METHODS

Study population and data collection

The methods used in this study have been partially explained elsewhere (7). Our study population was established in October 1998, when 4,287 faculty members and administrative staff of five Spanish universities in two regions (Galicia and Canary Islands) responded to a questionnaire about drinking habits, tobacco use, other lifestyle variables, and medical history. Participants were asked to join “an epidemiologic study to determine risk factors of respiratory diseases in our community.” We excluded from follow-up those persons who reported a current common cold episode, history of asthma, or history of chronic obstructive pulmonary disease. Subsequent short questionnaires were sent every 10 weeks for 1 year.

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. Participants with duplicated codes were excluded from follow-up. The study was approved by the institutional review boards of the five participating universities.

Disease assessment

We provided participants with a calendar and instructed them to record, daily during the duration of what they perceived as a common cold episode, the presence of eight symptoms of common cold and to rate their intensity on a four-level scale from none (0) to very intense (3). These symptoms included runny nose, sneezing, nasal congestion, headache, chills, sore throat, cough, and malaise. Every 10 weeks, we mailed a short questionnaire to participants and asked them to transcribe the information from their calendars so we could update our data 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 (refer to the information below). However, our main study results did not change substantially when we used other cutoff points.

Exposure assessment

We determined usual alcohol intake by asking subjects about their weekly consumption of red wine, white wine, beer, and spirits during the 12 months before baseline. For each subject, total intake of alcohol per week was calculated as the product of the ethanol content of a drink of average size, as given in the Spanish Health Survey (8) (9.6 g for wine, 9.6 g for beer, and 13.44 g for spirits), and the average number of drinks of red wine, white wine, beer, and spirits consumed per week.

Validation substudy

Concurrently with the main study, we carried out a validation substudy in a random sample of 69 participants. Our goal was to document the validity and reproducibility of common cold diagnosis and alcohol intake.

We compared our diagnoses with those obtained by using the complete diagnosis method of Beare and Reed (9). This method uses a 20-item checklist of common cold symptoms and physical signs; thus, it requires that a physician examine the patient. Each sign and symptom is rated from 0 to 3. Although all items are weighted the same in the original checklist, we assigned them unequal weights based on the relative frequency of symptoms and signs during common cold episodes reported in a large case series (10). Hence, nasal manifestations contributed 60 percent to the total score compared with 25 percent for lower airways manifestations and 15 percent for general manifestations. A patient was considered to have experienced a common cold episode if his or her score was greater than or equal to 150 points of a total 300 points.

To assess the validity of our measure of usual alcohol intake, we compared our frequency questionnaire with a 12-day diet record. To account for seasonal variation, these 12 days were distributed as 3 days during each season: two nonconsecutive weekdays and one weekend day. To assess reproducibility, we administered the same frequency questionnaire used at the beginning of the study to the 69 participants included in the validation substudy, approximately 1 year later.

Data analysis

Each participant contributed person-time from when he or she returned the initial questionnaire to onset of a common cold episode, termination of the study, or loss to follow-up, whichever occurred first. To take into account the overdispersion due to common cold episodes not being completely independent events, we used negative binomial regression to obtain adjusted incidence rate ratios and their corresponding 95 percent confidence intervals (11). Use of the negative binomial distribution produced broader confidence intervals compared with the Poisson distribution. Secondary analyses in which we did not censor participants after their first common cold episode yielded similar results and are not presented here.

Possible risk factors for common cold considered in the analyses included age, sex, faculty/staff status, university, region, smoking status, contact with children, psychological stress, and vitamin C and zinc intake. Cigarette smoking was assessed by using the standardized World Health Organization questionnaire (12). Smoking status categories were never smoker, former smoker, occasional smoker, and current smoker of 1–19, 20–40, or more than 40 cigarettes a day. Contact with children was measured by three variables: total number of children, number of children less than age 2 years, and number of children who attend kindergarten. Psychological stress was measured by using four independent scales: positive affect, negative affect, stressful events, and perceived stress. Further details on stress assessment have been presented elsewhere (7). Vitamin C and zinc intake were determined by using a food frequency questionnaire and were validated with 12 daily diet records.

To explore the shape of the curve relating wine consumption to occurrence of common cold episodes, we fitted a model with cubic splines (13) and adjusted for potential confounders. We set three knots at the boundaries of the exposure categories (0, 1–7, 8–14, and >14 drinks per week). Other model specifications produced similar curves.

RESULTS

We detected 1,353 cases of common cold occurring in 4,272 subjects between October 1, 1998, and September 30, 1999. Fifteen of the 4,287 subjects who responded to the questionnaire were excluded because of duplicated codes. The overall incidence of common cold was 1.4 per person-year (1.1 among men, 1.7 among women). Participants were aged 21–69 years, 46 percent were female, and 67 percent were faculty members.

Twenty-seven percent of the participants were teetotalers, and 10 percent drank more than 13 alcoholic drinks of any type per week. Drinkers consumed on average 96 g (standard deviation, 104) of pure alcohol and 3.9 glasses (standard deviation, 5) of wine per week.

Men consumed any alcoholic drink more often than women did. Mean age was higher among moderate-to-heavy wine drinkers compared with teetotalers, but no substantial differences regarding age were found for beer and spirits. Table 1 displays the distribution of sex, age, and other factors by intake of total alcohol and alcoholic beverages.

TABLE 1.

Distribution of potential risk factors for common cold, according to intake category of alcohol and alcoholic beverages, in five Spanish universities, 1998–1999

Type and no. of drinks/weeks Mean age (years) Female
 
Faculty member
 
Ever smoker
 
Contact with children
 
Allergy
 
Other diseases
 
No. No. No. No. No. No. 
Wine (red or white or both)              
 0 37 960 58.2 961 58.2 441 26.2 424 25.2 313 18.7 217 13.0 
 1–7 40 802 42.7 1,387 73.2 534 27.8 450 23.4 332 17.3 231 12.1 
 8–14 44 100 23.5 330 77.1 137 31.7 78 18.0 63 14.6 45 10.5 
 >14 46 32 16.8 136 71.2 65 34.2 32 16.8 24 12.8 22 11.6 
Red wine only              
 0 37 351 59.1 360 61.2 170 28.1 144 23.7 118 19.6 87 14.4 
 1–7 40 231 36.7 430 73.4 163 27.4 137 23.0 98 16.6 66 11.2 
 8–14 44 30 23.3 101 80.2 42 32.8 26 20.2 16 12.4 12 9.4 
 >14 45 13 21.3 38 62.3 15 25.0 11 18.0 13.3 11.5 
White wine only              
 0 37 106 60.6 96 53.6 42 23.3 50 27.9 33 18.6 22 12.4 
 1–7 42 30 43.5 51 72.9 21 29.6 15 21.1 14 19.7 12.7 
 8–14 43 24.1 20 69.0 23.3 13.3 20.0 16.7 
 >14 47 11.8 12 70.6 41.2 1.8 5.9 11.8 
Beer              
 0 39 1,167 60.5 1,216 62.8 493 25.0 483 24.5 353 18.1 243 12.5 
 1–7 39 655 34.8 1,359 72.0 519 27.1 429 22.3 328 17.1 231 12.1 
 >7 40 73 21.5 239 70.3 164 48.0 71 20.8 51 15.0 41 12.1 
Spirits              
 0 40 1,451 52.0 1,860 66.2 667 23.4 715 25.1 503 17.7 342 12.1 
 1–7 39 438 34.1 904 70.4 463 35.6 255 19.5 223 17.2 162 12.5 
 >7 41 8.2 51 69.9 47 62.7 14 18.7 8.0 11 14.7 
Total alcohol intake (median, g/week)              
 1st quartile (0.0) 38 689 66.5 585 56.5 223 21.1 293 27.8 194 18.5 129 12.3 
 2nd quartile (19.2) 39 528 51.9 707 68.8 242 23.2 268 25.7 187 18.1 119 11.5 
 3rd quartile (66.0) 40 438 42.2 756 72.6 284 26.7 213 20.0 198 18.6 147 13.9 
 4th quartile (160.4) 41 239 22.7 765 72.3 427 40.1 209 19.6 153 14.6 120 11.3 
Type and no. of drinks/weeks Mean age (years) Female
 
Faculty member
 
Ever smoker
 
Contact with children
 
Allergy
 
Other diseases
 
No. No. No. No. No. No. 
Wine (red or white or both)              
 0 37 960 58.2 961 58.2 441 26.2 424 25.2 313 18.7 217 13.0 
 1–7 40 802 42.7 1,387 73.2 534 27.8 450 23.4 332 17.3 231 12.1 
 8–14 44 100 23.5 330 77.1 137 31.7 78 18.0 63 14.6 45 10.5 
 >14 46 32 16.8 136 71.2 65 34.2 32 16.8 24 12.8 22 11.6 
Red wine only              
 0 37 351 59.1 360 61.2 170 28.1 144 23.7 118 19.6 87 14.4 
 1–7 40 231 36.7 430 73.4 163 27.4 137 23.0 98 16.6 66 11.2 
 8–14 44 30 23.3 101 80.2 42 32.8 26 20.2 16 12.4 12 9.4 
 >14 45 13 21.3 38 62.3 15 25.0 11 18.0 13.3 11.5 
White wine only              
 0 37 106 60.6 96 53.6 42 23.3 50 27.9 33 18.6 22 12.4 
 1–7 42 30 43.5 51 72.9 21 29.6 15 21.1 14 19.7 12.7 
 8–14 43 24.1 20 69.0 23.3 13.3 20.0 16.7 
 >14 47 11.8 12 70.6 41.2 1.8 5.9 11.8 
Beer              
 0 39 1,167 60.5 1,216 62.8 493 25.0 483 24.5 353 18.1 243 12.5 
 1–7 39 655 34.8 1,359 72.0 519 27.1 429 22.3 328 17.1 231 12.1 
 >7 40 73 21.5 239 70.3 164 48.0 71 20.8 51 15.0 41 12.1 
Spirits              
 0 40 1,451 52.0 1,860 66.2 667 23.4 715 25.1 503 17.7 342 12.1 
 1–7 39 438 34.1 904 70.4 463 35.6 255 19.5 223 17.2 162 12.5 
 >7 41 8.2 51 69.9 47 62.7 14 18.7 8.0 11 14.7 
Total alcohol intake (median, g/week)              
 1st quartile (0.0) 38 689 66.5 585 56.5 223 21.1 293 27.8 194 18.5 129 12.3 
 2nd quartile (19.2) 39 528 51.9 707 68.8 242 23.2 268 25.7 187 18.1 119 11.5 
 3rd quartile (66.0) 40 438 42.2 756 72.6 284 26.7 213 20.0 198 18.6 147 13.9 
 4th quartile (160.4) 41 239 22.7 765 72.3 427 40.1 209 19.6 153 14.6 120 11.3 

Total alcohol intake and beer and spirits consumption were not related to the occurrence of common cold, whereas consumption of wine was inversely associated with the risk of common cold (table 2). The adjusted incidence rate ratio for drinkers of >14 glasses of wine per week compared with teetotalers was 0.6 (95 percent confidence interval (CI): 0.4, 0.8). The association was stronger for red wine. These results did not materially change after further adjustment for alcohol, smoking, contact with children, psychological stress, vitamin C and zinc intake, university, and geographic region. The inverse association between wine intake and common cold was also present after we restricted the analysis to nonsmokers (70 percent of participants, table 3).

TABLE 2.

Incidence rate ratios of clinical common cold in five Spanish universities, according to intake of each type of alcoholic beverage, 1998–1999

Type and no. of drinks/week Not alcohol adjusted*,
 
Alcohol adjusted,
 
No. of person-weeks No. of cases 
RR§ 95% CI§ RR 95% CI 
Wine (red or white or both)       
 0 1.00 Reference 1.00 Reference 17,737 571 
 1–7 0.77 0.66, 0.90 0.77 0.65, 0.90 23,165 594 
 8–14 0.71 0.55, 0.92 0.71 0.53, 0.94 5,800 124 
 >14 0.56 0.38, 0.82 0.56 0.36, 0.86 2,949 53 
Red wine only       
 0 1.00 Reference 1.00 Reference 6,318 228 
 1–7 0.66 0.51, 0.86 0.65 0.49, 0.87 6,764 179 
 8–14 0.59 0.37, 0.96 0.58 0.34, 0.98 1,740 35 
 >14 0.41 0.21, 0.81 0.39 0.18, 0.87 950 16 
White wine only       
 0 1.00 Reference 1.00 Reference 1,776 60 
 1–7 0.67 0.33, 1.36 0.77 0.37, 1.58 1,021 21 
 8–14 0.86 0.32, 2.30 1.22 0.42, 3.59 501 11 
 >14 0.33 0.06, 1.74 0.80 0.11, 6.02 364 
Beer       
 0 1.00 Reference 1.00 Reference 22,771 652 
 1–7 1.03 0.88, 1.20 1.05 0.89, 1.23 22,397 599 
 >7 0.96 0.71, 1.28 0.99 0.72, 1.37 4,488 93 
Spirits       
 0 1.00 Reference 1.00 Reference 33,532 918 
 1–7 1.08 0.92, 1.26 1.18 0.97, 1.44 15,436 409 
 >7 1.08 0.58, 2.01 1.64 0.73, 3.68 704 17 
Total alcohol intake (median, g/week)       
 1st quartile (0.0) Reference   11,234 362 
 2nd quartile (19.2) 0.92 0.75, 1.13   12,347 338 
 3rd quartile (66.0) 1.00 0.82, 1.22   12,357 334 
 4th quartile (160.4) 0.98 0.79, 1.20   13,697 308 
Type and no. of drinks/week Not alcohol adjusted*,
 
Alcohol adjusted,
 
No. of person-weeks No. of cases 
RR§ 95% CI§ RR 95% CI 
Wine (red or white or both)       
 0 1.00 Reference 1.00 Reference 17,737 571 
 1–7 0.77 0.66, 0.90 0.77 0.65, 0.90 23,165 594 
 8–14 0.71 0.55, 0.92 0.71 0.53, 0.94 5,800 124 
 >14 0.56 0.38, 0.82 0.56 0.36, 0.86 2,949 53 
Red wine only       
 0 1.00 Reference 1.00 Reference 6,318 228 
 1–7 0.66 0.51, 0.86 0.65 0.49, 0.87 6,764 179 
 8–14 0.59 0.37, 0.96 0.58 0.34, 0.98 1,740 35 
 >14 0.41 0.21, 0.81 0.39 0.18, 0.87 950 16 
White wine only       
 0 1.00 Reference 1.00 Reference 1,776 60 
 1–7 0.67 0.33, 1.36 0.77 0.37, 1.58 1,021 21 
 8–14 0.86 0.32, 2.30 1.22 0.42, 3.59 501 11 
 >14 0.33 0.06, 1.74 0.80 0.11, 6.02 364 
Beer       
 0 1.00 Reference 1.00 Reference 22,771 652 
 1–7 1.03 0.88, 1.20 1.05 0.89, 1.23 22,397 599 
 >7 0.96 0.71, 1.28 0.99 0.72, 1.37 4,488 93 
Spirits       
 0 1.00 Reference 1.00 Reference 33,532 918 
 1–7 1.08 0.92, 1.26 1.18 0.97, 1.44 15,436 409 
 >7 1.08 0.58, 2.01 1.64 0.73, 3.68 704 17 
Total alcohol intake (median, g/week)       
 1st quartile (0.0) Reference   11,234 362 
 2nd quartile (19.2) 0.92 0.75, 1.13   12,347 338 
 3rd quartile (66.0) 1.00 0.82, 1.22   12,357 334 
 4th quartile (160.4) 0.98 0.79, 1.20   13,697 308 
*

Adjusted for sex, age (continuous), and faculty/staff status.

Further adjustment for the following variables did not change the results: contact with children (three variables), psychological stress (four variables), vitamin C and zinc intake, university, and geographic region.

Adjusted for sex, age (continuous), faculty/staff status, and total alcohol intake.

§

RR, incidence rate ratio; CI, confidence interval.

TABLE 3.

Incidence rate ratios of clinical common cold, according to type of alcoholic beverage, among nonsmokers only in five Spanish universities, 1998–1999

Type and no. of drinks/week Nonsmokers only
 
RR* 95% CI* 
Wine (red or white or both)   
 0 1.00 Reference 
 1–7 0.82 0.68, 0.98 
 8–14 0.64 0.46, 0.88 
 >14 0.62 0.38, 1.00 
Red wine only   
 0 1.00 Reference 
 1–7 0.71 0.52, 0.97 
 8–14 0.57 0.31, 1.04 
 >14 0.60 0.28, 1.03 
White wine only   
 0 1.00 Reference 
 1–7 0.89 0.39, 1.99 
 8–14 1.98 0.26, 3.72 
 >14 0.12 0.01, 1.49 
Beer   
 0 1.00 Reference 
 1–7 1.05 0.88, 1.26 
 >7 0.81 0.54, 1.22 
Spirits   
 0 1.00 Reference 
 1–7 0.96 0.79, 1.17 
 >7 1.00 0.36, 2.77 
Type and no. of drinks/week Nonsmokers only
 
RR* 95% CI* 
Wine (red or white or both)   
 0 1.00 Reference 
 1–7 0.82 0.68, 0.98 
 8–14 0.64 0.46, 0.88 
 >14 0.62 0.38, 1.00 
Red wine only   
 0 1.00 Reference 
 1–7 0.71 0.52, 0.97 
 8–14 0.57 0.31, 1.04 
 >14 0.60 0.28, 1.03 
White wine only   
 0 1.00 Reference 
 1–7 0.89 0.39, 1.99 
 8–14 1.98 0.26, 3.72 
 >14 0.12 0.01, 1.49 
Beer   
 0 1.00 Reference 
 1–7 1.05 0.88, 1.26 
 >7 0.81 0.54, 1.22 
Spirits   
 0 1.00 Reference 
 1–7 0.96 0.79, 1.17 
 >7 1.00 0.36, 2.77 
*

RR, incidence rate ratio; CI, confidence interval.

The exposure-effect curve shows this inverse relation between wine intake and the incidence of common cold (figure 1A), which was stronger for red wine intake (figure 1B). Incidence rate ratio estimates for red wine intakes of more than 15 glasses per week, as well as for any intake of white wine (figure 1C), were imprecise because of limited data, as shown by the wide 95 percent confidence intervals in figure 1C and table 2. The curves were nearly identical after adjustment for total alcohol intake.

FIGURE 1.

Incidence rate ratios of common cold, calculated by using splines regression, among study participants in five Spanish universities, according to wine consumption, 1998–1999. A: intake of wine of any type (red or white or both); B: intake of red wine; C: intake of white wine. Incidence rate ratios were adjusted for sex, age, and faculty/staff status. Solid line, point estimates; thin dotted lines, 95% confidence intervals.

FIGURE 1.

Incidence rate ratios of common cold, calculated by using splines regression, among study participants in five Spanish universities, according to wine consumption, 1998–1999. A: intake of wine of any type (red or white or both); B: intake of red wine; C: intake of white wine. Incidence rate ratios were adjusted for sex, age, and faculty/staff status. Solid line, point estimates; thin dotted lines, 95% confidence intervals.

To disentangle the effect of wine from that due to other alcoholic drinks, we restricted our analysis to nondrinkers and to participants who exclusively drank wine but not beer or spirits. The incidence rate ratios were 0.7 (95 percent CI: 0.6, 0.9) for drinkers of 1–7 glasses per week, 0.5 (95 percent CI: 0.3, 0.9) for drinkers of 8–14 glasses per week, and 0.6 (95 percent CI: 0.3, 1.4) for drinkers of >14 glasses per week compared with teetotalers. Very few subjects drank red wine exclusively but did not consume any white wine, beer, or spirits. Therefore, it was not possible to conduct a meaningful analysis of this group.

Both history of allergic rhinitis and history of any other upper respiratory tract disease were associated with common cold. The incidence rate ratios were 2.7 (95 percent CI: 2.3, 3.2) for history of allergic rhinitis and 3.3 (95 percent CI: 2.7, 4.0) for other upper respiratory tract disease. To reduce the possibility of self-report of bouts of these diseases as common cold episodes, we restricted the analyses to participants who had no history of allergic rhinitis or any other upper respiratory tract disease. The results were similar to those reported above.

Twenty-two percent of the participants were lost to follow-up before the study ended, but the distribution of exposure variables (consumption of red wine, white wine, beer, and spirits) and other factors (sex, age, faculty/staff status) was similar between participants with incomplete follow-up and those with complete follow-up. We also recalculated the incidence rate ratios in two extreme situations. First, we assumed that all subjects lost to follow-up developed a cold within the week after dropping out of the study. Second, we assumed that none of the subjects ever developed a cold. Again, the results were not substantially altered.

The sensitivity of our diagnosis of common cold was 94 percent, and its specificity was 84 percent. The Spearman coefficients of correlation between our alcohol frequency questionnaire and the diet records were 0.76 for wine, 0.72 for beer, 0.59 for spirits, and 0.78 for total alcohol intake. The intraclass correlation coefficients for reproducibility were 0.86 for wine, 0.93 for beer, 0.80 for spirits, and 0.82 for alcohol intake. These results indicate that wine intake is measured with some error. However, this error is likely to occur independently of disease status because our information on exposure was collected before disease occurred. To explore the direction and magnitude of the potential bias due to misclassification of exposure, we reanalyzed the data by entering the variable “wine intake” as dichotomous (abstemious/nonabstemious). The resulting incidence rate ratio was 0.7 (95 percent CI: 0.6, 0.9), which suggests that the true association between wine intake and common cold may be even stronger than the one we observed (nondifferential misclassification of a dichotomous variable).

DISCUSSION

The findings from this prospective study support the existence of a strong inverse relation between the consumption of wine, but not other alcoholic beverages, and the incidence of common cold. This inverse association persisted after adjustment for total alcohol intake, smoking, and known risk factors for common cold. Among those participants consuming both red and white wine, the association was even stronger among those consuming red wine exclusively. Because subjects with a high intake of wine were rare in the study population, our results were limited to light-to-moderate wine consumption. Total alcohol intake was not related to the risk of common cold.

A validation substudy showed that the quality of our measurement of alcohol intake was similar to others reported in the literature (14) and 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 (15). The prospective study design ensured that usual alcohol intake was assessed before the diagnosis of common cold was made and thus prevented differential reporting of alcohol consumption between common cold cases and noncases. We also adjusted for known or suspected risk factors for common cold and conducted sensitivity analyses to assess the influence of loss to follow-up, but the results did not change materially. Thus, the inverse association we found is not easily ascribed to confounding, misclassification, or selection bias due to loss to follow-up. Further research may help clarify whether the inverse association between wine intake and common cold can be partially explained by unmeasured variables (e.g., lifestyle habits, drinking patterns).

A protective effect of some nonalcoholic components of wine on the occurrence of common cold is consistent with previous animal and human studies, which showed that some nonalcohol components of wine have strong anti-inflammatory (4, 5), antioxidant (16), and vasorelaxant properties (17). Resveratrol, one nonalcohol component with strong anti-inflammatory activity, is especially abundant in red wine (5). Therefore, wine consumption may downregulate the immune response that leads to the clinical manifestations of common cold.

In addition, red wine is particularly rich in flavonoids, such as quercetin and catechin (18). Flavonoids are polyphenolic antioxidants whose antiviral properties are mediated by their binding to viral protein and interfering with the synthesis of the viral nucleic acid (19). Furthermore, specific activity of synthetic flavonoids against rhinoviruses—considered the major causal agents of the common cold—has been demonstrated in vitro (20). This finding might explain an increased resistance to viral infection among wine drinkers, but the relevance of any of these or other mechanisms to the relation between wine consumption and common cold episodes remains to be established.

Since the public health burden of common cold results from the disability caused by its clinical manifestations, and the economic costs associated with common cold are due to absenteeism and to the drugs used to alleviate its symptoms, our diagnosis of common cold was a clinical one. Previous studies of risk factors for common cold among artificially infected subjects have sometimes required the presence of symptoms and seropositivity to confirm the diagnosis (6, 21). These studies found that only one third of subjects exposed to viruses became infected and that only one third of infected subjects developed a clinical cold (22, 23). Thus, a diagnosis based on seropositivity would include a high proportion of nonclinically detectable common cold episodes (false positives according to our criteria). In summary, our findings suggest that moderate consumption of wine, but not of other alcoholic beverages, may reduce the incidence of clinical common cold.

Correspondence to Dr. Miguel Hernán, Department of Epidemiology, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115 (e-mail: miguel_hernan@post.harvard.edu).

This work was funded by grant FIS 99/0021-01 from the Spanish “Fondo de Investigaciones Sanitarias” (Madrid, Spain) and by a grant from the University of Santiago de Compostela (Santiago de Compostela, Spain).

The authors thank the following persons who participated in data collection: Drs. Carlos Quintas, Rosa Meijide, Lourdes Maceiras, and Milagros Torres.

REFERENCES

1.
Couch RB. The common cold: control?
J Infect Dis
 
1984
;
150
:
167
–73.
2.
Schwartz J, Weiss ST. Host and environmental factors influencing the peripheral blood leukocyte count.
Am J Epidemiol
 
1991
;
134
:
1402
–9.
3.
Szabo G. Consequences of alcohol consumption on host defence.
Alcohol Alcohol
 
1999
;
24
:
830
–41.
4.
Johansen KM, Skorpe S, Olsen JO, et al. The effect of red wine on the fibrinolytic system and the cellular activation reactions before and after exercise.
Thromb Res
 
1999
;
96
:
355
–63.
5.
Jang M, Cai L, Udeani GO, et al. Cancer chemopreventive activity of resveratrol, a natural product derived from grapes.
Science
 
1997
;
275
:
218
–20.
6.
Cohen S, Tyrell DAJ, Russel MAH, et al. Smoking, alcohol consumption, and susceptibility to the common cold.
Am J Public Health
 
1993
;
83
:
1277
–83.
7.
Takkouche B, Regueira C, Gestal-Otero JJ. A cohort study of stress and the common cold.
Epidemiology
 
2001
;
11
:
345
–9.
8.
Encuesta nacional de salud de España 1997. (In Spanish). Madrid, Spain: Ministerio de Sanidad y Consumo, 1999:331.
9.
Beare AS, Reed SE. The study of antiviral compounds in volunteers. In: Oxford JS, ed. Chemoprophylaxis and virus infections. Vol 2. Cleveland, OH: CRC Press, 1977:27–55.
10.
Tyrrell DAJ, Cohen S, Schlarb JE. Signs and symptoms in common colds.
Epidemiol Infect
 
1993
;
11
:
143
–56.
11.
Greenland S. Introduction to regression modeling. In: Rothman K, Greenland S, eds. Modern epidemiology. Philadelphia, PA: Lippincott-Raven, 1998:401–32.
12.
The evaluation and monitoring of public action on tobacco. Smoke-free Europe: 3. Copenhagen, Denmark: World Health Organization, Regional Office for Europe and the Commission of the European Communities, 1987:14.
13.
Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis.
Epidemiology
 
1995
;
6
:
356
–65.
14.
Giovannucci E, Colditz G, Stampfer MJ, et al. The assessment of alcohol consumption by a simple self-administered questionnaire.
Am J Epidemiol
 
1991
;
133
:
810
–17.
15.
Gwaltney JM Jr. The common cold. In: Mandell GL, Benett JE, Dolin R, eds. Principles and practice of infectious diseases. New York, NY: Churchill Livingstone, 1995:561–6.
16.
Maxwell S, Cruickshank A, Thorpe G. Red wine and antioxidant activity in serum.
Lancet
 
1994
;
344
:
193
–4.
17.
Fitzpatrick DF, Hirschfield SL, Coffey RG. Endothelium-dependent vasorelaxing activity of wine and other grape products.
Am J Physiol
 
1993
;
265(2 pt 2)
:
H774
–8.
18.
Soleas GJ, Diamandis EP, Goldberg DM. Wine as biological fluid: history, production and role in disease prevention.
J Clin Lab Anal
 
1997
;
11
:
287
–313.
19.
Formica JV, Regelson W. Review of the biology of Quercetin and related bioflavonoids.
Food Chem Toxicol
 
1995
;
33
:
1061
–80.
20.
Conti C, Mastromarino P, Sgro R, et al. Anti-picornavirus activity of synthetic flavon-3-yl-esters.
Antivir Chem Chemother
 
1998
;
9
:
511
–15.
21.
Cohen S, Tyrell DAJ, Smith AP. Psychologic stress and susceptibility to the common cold.
N Engl J Med
 
1991
;
325
:
606
–12.
22.
Gwaltney JM Jr. Rhinoviruses. In: Evans AS, ed. Viral infections of humans: epidemiology and control. New York, NY: Plenum Medical, 1997:815–39.
23.
Glaser R, Rabin B, Chesney M, et al. Stress-induced immunomodulation: implications for infectious diseases?
JAMA
 
1999
;
281
:
2268
–70.