点击显示 收起
1 From the Department of Nutritional Science and Dietetics, University of Nebraska, Lincoln, and the Department of Nutrition, Arizona State University, Tempe.
2 Address reprint requests to NM Betts, Department of Nutritional Science and Dietetics, University of Nebraska, 202 Ruth Leverton Hall, Lincoln, NE 68583-0806. E-mail: nbetts1{at}unl.edu.
ABSTRACT
Background: Cigarette smoking is a major risk factor for several chronic oxidative diseases that can be ameliorated by antioxidants.
Objectives: This study identified the typical dietary intakes and the major food group contributors of the antioxidants ß-carotene, vitamin C, and vitamin E by smoking status.
Design: The 1994–1996 Continuing Survey of Food Intakes by Individuals (CSFII) provided the current sample (n = 6749), who were categorized as non- (n = 3231), former (n = 1684), and current (n = 1834) smokers. In the CSFII, individuals' food intakes were estimated with two 24-h dietary recalls. Data were analyzed by using a chi-square test with a simultaneous Fisher's z test, analysis of variance with Scheffe's test, multivariate analysis of covariance, and analysis of covariance with Bonferroni adjustment for multiple comparisons.
Results: The sample consisted of 3707 men and 3042 women. Current smokers tended to be younger with less education and lower incomes than nonsmokers and former smokers. The average body mass index (in kg/m2) of current smokers was 25.8, the lowest of the 3 groups. Current smokers had the lowest dietary antioxidant intake. Fatty foods such as luncheon meats, condiments and salad dressings, and ground beef contributed more to the antioxidant intakes of current smokers than to those of the other 2 groups, whereas fruit and vegetables contributed less. Current smokers consumed the fewest numbers of servings of all nutrient-bearing groups in the food guide pyramid, except the meat group.
Conclusion: Future interventions should target the clustering of cigarette smoking and other unhealthy lifestyle habits, eg, an imprudent diet.
Key Words: Antioxidants • cigarette smoking • ß-carotene • vitamin E • vitamin C • food categories • Continuing Survey of Food Intakes by Individuals • humans
INTRODUCTION
Cigarette smoking alone has been implicated as a major risk factor for several chronic diseases, including cardiovascular disease, pulmonary disease, and cancer (1). However, smoking tends to cluster with other lifestyle risk factors, such as alcohol abuse, physical inactivity, and an imprudent diet (2, 3). As a result, smoking along with other risky lifestyle habits places individuals at higher risk for the diseases that cause most American deaths (4–6).
Cigarette smoke contains high amounts of free radicals and other oxygen-derived species (7, 8). Free radicals may be the most critical factors triggering plasma antioxidant depletion, lipid peroxidation, and protein modification (8–10). Products arising from lipid peroxidation and protein modification can in turn react with cigarette smoke constituents, creating additional toxic products. The toxic products resulting from both the direct and secondary reactions with cigarette smoke are thought to activate inflammatory immune responses, which may themselves be altered by cigarette smoke constituents and play an influential role in smoking-related oxidative tissue damage (8). The subsequent metabolic and molecular changes are thought to induce the onset of cigarette smoking–related chronic diseases (11–13). Researchers have proposed that antioxidants such as -tocopherol, ascorbate, and carotenoids have a protective effect by scavenging oxidants, thus preventing lipid peroxidation (8, 11, 14).
Previous studies have shown lower plasma concentrations of antioxidants in smokers than in nonsmokers and former smokers (8, 15, 16). Lower antioxidant concentrations in smokers may be due to low intakes of foods rich in antioxidant nutrients (17, 18), a greater utilization of available antioxidants to protect against the burden of free radicals induced by smoking (5, 19), or both. On the basis of these findings, smokers have been advised to obtain sufficient antioxidants in their diets (7, 17, 18).
This study had 3 purposes: 1) to identify the typical dietary intake of the 3 most studied antioxidant vitamins—ß-carotene, vitamin E, and vitamin C—by smoking status and sex; 2) to identify the major food contributors of antioxidants in the diets of men and women in different smoking categories; and 3) to examine the differences in the numbers of servings of the 5 nutrient-bearing US food guide pyramid (FGP) food groups by smoking status and sex.
SUBJECTS AND METHODS
The data were derived from a series of national food consumption surveys conducted by the US Department of Agriculture in 1994, 1995, and 1996 [ie, the Continuing Survey of Food Intakes by Individuals (CSFII)]. The 1994–1996 CSFII, the most recently released series, was conducted between January 1994 and January 1997. In the CSFII, data were collected from a stratified area probability sample of men, women, and children residing in households in the 50 states and the District of Columbia. Two nonconsecutive days of food intake data for 15303 individuals of all ages (50.7% males and 49.3% females) were collected 3–10 d apart during in-person interviews by trained interviewers using the 24-h recall method. In the 1994–1996 CSFII, 3 technical databases (ie, the Food Coding Database, the Nutrient Database, and the Recipe Database) were used to code the food data collected and to calculate the nutritive value of the foods. The Survey Nutrient Database is maintained specifically for use with nationwide food surveys and includes values for food energy and 28 nutrients and food components, including ß-carotene [in retinol equivalents (RE)], vitamin C, and vitamin E (in mg). Most of the values for major contributors of nutrients were supported by laboratory analyses, and assigned data were used for nutrient values not available from such analyses. Multicomponent foods were disaggregated into their ingredients by using the Recipe Database, and nutrient values of each ingredient were computed. Detailed information on the methods used in the CSFII is provided elsewhere (20).
Adults aged 19 y were of interest in the current research. To minimize potential bias due to special dietary intakes, the following exclusion criteria were used: pregnancy, lactation, vegetarianism, and food allergies. This study focused on the dietary sources of antioxidants; therefore, nutrients derived from supplement intakes were not under investigation. Only those respondents who reported 2 d of food intake were included and the average intakes on these 2 d were used in the analyses because 2-d averages provide better estimates of usual intake than do data for a single day. The resulting sample included 6758 adults (3713 men and 3045 women). In the CSFII, smoking status was determined from the answers to 2 questions: "Have you smoked 100 cigarettes during your entire life?" and "Do you smoke cigarettes now?" On the basis of the responses to these 2 questions, the sample was divided into 3 smoking categories: nonsmokers (n = 3231), former smokers (n = 1684), and current smokers (n = 1834). Nine individuals did not report their smoking status and were dropped from the analysis, resulting in a final sample size of 6749.
All analyses were performed by using SPSS for WINDOWS (version 8.0, SPSS, Inc, Chicago). Data were tabulated and demographic variables were compared across smoking groups by using chi-square with simultaneous Fisher's z tests and analysis of variance with Scheffe's method. With energy intake, income, education, and age adjusted for, multivariate analysis of covariates examined potential differences in body mass index (BMI; in kg/m2), percentage of energy as fat, mean antioxidant vitamins (ß-carotene and vitamins E and C), as well as the numbers of servings from the 5 nutrient-bearing food groups in the FGP. The interaction effect of sex and smoking status as well as the main effect of each were examined. Bonferroni's adjustment was applied when multiple comparisons were made.
The US Department of Agriculture assigned each food and beverage item an 8-digit code number and categorized them into food groups. These food groups were found to be relatively broad for the current purpose. To identify the major contributors of each antioxidant in each category classified by sex and smoking status, the reported food items were regrouped and categorized into 58 major food groups on the basis of previous research (21). Mixed dishes were classified by their primary ingredients (eg, chicken parmigiana was classified into the poultry category). The subdivisions of the vegetable group and the fruit group were intended to show contributions from the subgroups of vegetables and fruit. The vegetable group was subgrouped as follows: potatoes and potato products, dark-green vegetables, cruciferous vegetables, tomatoes, carrots, and other vegetables. The major contributors to "other vegetables" were lettuce, green beans, sweet peas, and corn. The following subgroups of fruit were defined: citrus fruit, fruit juices, dried fruit, cherries and berries, melons, and other fruit. The major contributors to "other fruit" were apples, pears, peaches, and grapes. The contribution to total consumption of an antioxidant per food group was calculated by using the following formula: (sum of antioxidants from all foods in a food group)/(sum of antioxidants from all 58 food groups). Analysis of covariance with Bonferroni-adjusted (for age, income, education, and energy intake) multiple comparisons examined differences in the average daily amount of each food group among the nonsmokers, former smokers, and current smokers.
RESULTS
The demographic characteristics of the sample by smoking status are shown in Table 1. Chi-square analysis indicated that there were more male former and current smokers than female former and current smokers in either category. As examined by analysis of variance with Scheffe's method, nonsmokers and current smokers were significantly younger on average than were former smokers. Current smokers had significantly lower mean income and education levels than nonsmokers and former smokers.
View this table:
TABLE 1. . Demographic characteristics of subjects by smoking status1
Differences in BMI, percentage of energy as fat, and average antioxidant intakes (adjusted for energy intake, income, education, and age) in response to the interaction effect of sex and smoking status and the main effect of each are shown in Table 2. The interaction effect of sex and smoking status was significant only for the percentage of energy as fat. In men, nonsmokers consumed significantly less energy as fat than did both former and current smokers. The same trend held true when the main effect of smoking status was examined. For women, both nonsmokers and former smokers consumed significantly less energy as fat than did current smokers. For the main effect of sex, significant differences were seen for BMI, ß-carotene, and vitamin C intakes, the effect being greater in women than in men. For the main effect of smoking status, current smokers had a significantly lower BMI, a lower intake of all 3 antioxidant vitamins, but a significantly higher percentage of energy as fat than the other 2 smoking groups.
View this table:
TABLE 2. . Differences in selected variables across smoking groups and by sex1
The major contributing food groups to total ß-carotene intakes for both sexes are listed in Table 3. For all subjects, carrots, other vegetables, dark-green vegetables, and tomatoes provided 6% of the total ß-carotene intake, which outnumbered the contributions from other food groups. The contribution to total ß-carotene intakes of the listed food groups was only modestly different between the 2 sexes. Butter, margarine, and oils; pizza; and cheese and cheese products made more important contributions to the ß-carotene intakes of current smokers than to those of the other 2 smoking groups. Both nonsmokers and former smokers obtained their ß-carotene mainly from carrots, other vegetables, dark-green vegetables, tomatoes, pasta dishes, fruit juices, other fruit, soups, and vegetable soups.
View this table:
TABLE 3. . Major contributing food groups to total ß-carotene intakes in the 3 groups by sex1
As seen in Table 4, fruit juices, potatoes and potato products, and other vegetables made up the primary sources of total vitamin C for both sexes. The food groups contributing to total vitamin C intakes were similar in men and women in all 3 smoking groups. Compared with former smokers and nonsmokers, current smokers consumed more vitamin C from potatoes and potato products, soft drinks, and cruciferous vegetables but less from breakfast cereals, dark-green vegetables, and citrus fruit. Slight variations in the major contributing food groups to vitamin C intakes were observed between nonsmokers and former smokers.
View this table:
TABLE 4. . Major contributing food groups to total vitamin C intakes in the 3 groups by sex1
The major food groups contributing to total vitamin E intakes for both sexes are listed in Table 5. For men and women in all 3 smoking categories, desserts were the major source of total vitamin E, with contributions ranging from 7% to 9%. Poultry, eggs, breads, condiments and salad dressings, fish, and salty snacks provided a considerable portion of the total vitamin E intake in both sexes. Of these food groups, condiments and salad dressings and salty snacks made greater contributions to the women's vitamin E intake, whereas eggs and ground beef made greater contributions in men. Few differences were observed among the 3 smoking groups; however, "nuts and seeds" was listed as one of the major contributors to vitamin E intakes in current and former smokers but not in nonsmokers.
View this table:
TABLE 5. . Major contributing food groups to total vitamin E intakes in the 3 groups by sex1
Variations in the average daily intakes of the food groups across the 3 smoking categories were examined by analysis of covariance adjusted for age, income, education, and energy intake. Compared with male nonsmokers, former smokers, or both, male current smokers had significantly lower intakes of other vegetables, dark-green vegetables, fruit juices, and other fruit, whereas they had higher intakes of soft drinks, condiments and salad dressings, fish, salty snacks, ground beef, luncheon meats, crackers, alcoholic drinks, coffee, and tea (data not shown). Male nonsmokers and male former smokers had similar consumption patterns. Fewer differences were found for women than for men. Female former smokers had the highest consumption of other vegetables, fruit juices, other fruit, and citrus fruit. In contrast, female current smokers had the highest consumption of vegetable soups, soft drinks, luncheon meats, condiments and salad dressings, alcoholic drinks, coffee, and tea.
Differences in the numbers of servings from the 5 nutrient-bearing FGP food groups by multivariate analysis of covariance, adjusted for age, income, education, and energy intake, are shown in Table 6. Significant interaction effects of sex and smoking status were discerned only for the fruit group. Male nonsmokers reported a significantly higher fruit intake than did male former smokers; however, male former smokers had a significantly higher fruit intake than male current smokers. In women, fruit intake was significantly greater in both nonsmokers and former smokers than in current smokers. As for the main effect of sex, women consumed significantly more fruit and dairy products than did men; in contrast, men reported a significantly higher number of servings from the meat group than did the women. As for the main effect of smoking status, current smokers consistently reported the fewest numbers of servings from all the food groups, except for meat. Current smokers consumed significantly more servings of meat than did the other 2 smoking groups.
View this table:
TABLE 6. . Differences in the numbers of servings from the 5 food guide pyramid groups across smoking groups and by sex1
DISCUSSION
Our data confirm the results obtained from populations in France (19) and the United Kingdom (3, 22), suggesting that nonsmokers, former smokers, and current smokers have distinctive socioeconomic characteristics. In agreement with Thornton et al (3) and Marangon et al (19), former smokers are more likely to be older males than are nonsmokers and current smokers. Compared with the other 2 smoking groups, current smokers tended to be younger men with lower education and income levels.
Antioxidants are of particular importance to smokers because antioxidants are believed to scavenge free radicals, which are found in large quantities in tobacco smoke (7) and have been implicated in the development of coronary heart disease, cancer, and other diseases (23). Studies have reported that an increased intake of a natural complex source of antioxidants from fruit and vegetables had a protective effect on the susceptibility of LDL to oxidation, which offers protection against oxidative diseases (8, 24). Unfortunately, a considerable body of evidence has shown that smokers have lower dietary intakes and plasma concentrations of multiple antioxidants than do nonsmokers and former smokers (1, 5, 6, 25, 26). After comparing the consumption of the 3 commonly studied antioxidant vitamins—ß-carotene, vitamin E, and vitamin C—by smoking status, we discovered congruent findings with previous research. On average, current smokers consistently consumed lower amounts of antioxidants than did nonsmokers and former smokers, whose intakes were more similar.
We need to understand why current smokers tend to consume fewer dietary antioxidants than do nonsmokers and former smokers. Our study attempted to answer this question by identifying the major contributing food sources of the antioxidants for each of the 3 smoking groups and by comparing the amounts (in g) of these contributing foods. The results showed that the types of food groups from which people obtained their daily antioxidants varied only modestly by smoking category. For all subjects, ß-carotene was obtained mainly from carrots, other vegetables, dark-green vegetables, and tomatoes. Food groups such as fruit juices, potatoes and potato products, and other vegetables provided a significant portion of total vitamin C intakes. Desserts, poultry, eggs, breads, condiments and salad dressing, fish, and salty snacks were the primary sources of total vitamin E. In one of the few previous studies that focused on food sources of nutrients, Chug-Ahuja et al (27) examined food sources of ß-carotene in 19–50-y-old respondents to the 1986 CSFII. They found, as we did, that carrots provided the most ß-carotene, followed by tomatoes, dark-green and yellow leafy vegetables, soups, tomato-based sauce, and fruit. We saw relatively few differences in the types of foods providing antioxidants among the 3 smoking groups. This finding suggests that recommendations to increase antioxidant consumption may be more effective if they focus on specific foods rather than on servings of food groups or amounts of nutrients.
Although types of foods were similar, notable differences by smoking status were seen in the importance of the major contributing food groups for each antioxidant. In contrast with nonsmokers and former smokers, current smokers generally acquired more antioxidants from food groups that had a relatively high fat content, such as cheese and cheese products; pizza; nuts and seeds; and butter, margarine, and oils. In addition, current smokers acquired fewer antioxidants from fruit juices, citrus fruits, other fruit, vegetable soups, and breakfast cereals, which are all commonly perceived as healthy food items.
Comparisons of the average daily intakes (in g) of each food group further confirmed that current smokers preferred foods rich in fat to low-fat foods such as fruit and vegetables. Fewer significant differences in food group intakes by smoking status were seen among women than among men. However, overall, current smokers reported a relatively high consumption of luncheon meats, condiments and salad dressings, ground beef, and soft drinks, whereas the intakes of a few specific fruit and vegetables were lower than those of nonsmokers and former smokers. A clustering of smoking and unhealthy dietary habits was documented in many studies (3, 5, 22, 28). Smokers' unhealthy diet preferences may be attributed to the adverse effect of smoking on taste perception, although data are not conclusive (29). We also speculated that the relatively poor socioeconomic status of current smokers (eg, low education and income levels) may be another reason for their unsound dietary selections. Furthermore, a high consumption of alcohol, coffee, and tea was consistently reported in our study by the current smokers of both sexes, which agrees with the findings of previous studies (2, 28).
The numbers of servings from the 5 nutrient-bearing food groups in the FGP were examined by sex and smoking status. The results indicated that smoking status played a significant role in the numbers of servings. Compared with the other 2 smoking groups, current smokers consumed fewer servings of grain products, fruit, and vegetables—the 3 food groups located on the base of the FGP—as well as of dairy products. On the other hand, current smokers had the highest number of servings from the meat group of the 3 smoking groups. Regardless of smoking status, the average number of servings from the fruit group fell markedly short of the minimum recommended 2 servings/d, whereas meat consumption averaged 5–6 servings/d compared with the recommended 2–3 servings/d. These findings are consistent with previous research based on the 1989–1991 CSFII (30, 31).
Throughout the study, few significant differences were found between nonsmokers and former smokers. Of note, however, a significantly higher percentage of total energy intake from fat was observed in male former smokers and male current smokers than in male nonsmokers. Research has documented (32) that a change in smoking habits without a change in lipid intakes is probably one explanation for the weight gain commonly associated with smoking cessation (33–35). Our study found that former smokers of both sexes had significantly higher mean BMIs than current smokers, who reported the lowest BMIs. It has been proposed that changes in metabolism, especially in fat oxidation, and lifestyles may also account for weight fluctuations related to smoking and abstaining (2, 32).
An inherent limitation of the current study was that dietary intake was assessed on the basis of individual recalls. With many of the dietary recall methods, there is an inherent tendency to underreport intakes, which undoubtedly leads to underestimation of intakes (36). Unfortunately, there has been no gold standard dietary assessment method against which other methods can be validated. The effect of underreporting may have been somewhat attenuated in our study by the use of 2-d averages per individual and the adjustment for energy consumption in the statistical comparisons. Additionally, Cole (36) concluded that a dietary assessment as simple as the 1-d dietary recall method in a large population sample is more effective than are more complex and expensive methods when the purpose of the study is to compare group means, which was the case in the present study. There is no published evidence regarding whether food intakes are differentially underreported according to a person's smoking status. Further investigation is warranted.
The primary focuses of the current study were the dietary sources of the 3 selected antioxidants and differential intakes by smoking status. Many significant differences were discovered in food group and dietary antioxidant intakes among the 3 smoking groups. Unfortunately, data that would shed light on physiologic differences were not available. Whether the small but statistically significant differences in dietary intakes seen in our study translate to meaningful biologically significant differences (eg, in plasma concentrations of antioxidants) is an important question. However, Yeum et al (37) concluded that carotenoids in human plasma can be elevated in a relatively short period of time by consuming more fruit and vegetables. More research is needed to examine the bioavailability of each antioxidant and to determine the relation between dietary increases in and elevations in plasma concentrations of antioxidants.
In summary, our findings suggest that future smoking cessation campaigns need to focus more specifically on middle-aged men with low income and education levels. Although these smokers have been difficult to reach, some success has been seen with theoretically based smoking-cessation interventions (38), and similar methods might be transferable to dietary interventions. Intervention programs are unlikely to be successful unless they also target other unhealthy lifestyle habits usually clustered with cigarette smoking, such as an imprudent diet and heavy alcohol consumption. As for the smokers' dietary patterns, greater emphasis needs to be placed on providing information about the health benefits and the major food sources of antioxidants. Two of the most crucial concerns are increasing consumption of fruit and vegetables and decreasing total fat consumption. For those who are concerned about weight gain after smoking cessation, interventions should stress ways to decrease energy intake and increase physical activity levels.
REFERENCES