Do Dietary and Supplementary Intakes of Antioxidants

International Journal of Epidemiology O International Epldemlotogtcal Association 1996

Vol. 25, No. 1 Printed In Great Britain

Do Dietary and Supplementary Intakes of Antioxidants Differ with Smoking Status? KRINA T ZONDERVAN,** MARGA C OCKE,* HENRIETTE A SMIT* AND JACOB C SEIDELL* Zondervan K T (Department of Chronic Diseases and Environmental Epidemiology, National Institute of Public Health and Environmental Protection, PO Box 1, 3270 BA Birthoven, The Netherlands), Ocke M C, Smit H A and Seidell J C. Do dietary and supplementary intakes of antioxidants differ with smoking status? International Journal of Epidemiology 1996;

25: 70-79.

of some diseases associated with smoking.7 Several studies have investigated dietary patterns in different smoking groups compared to non-smokers,8"13 usually showing that smokers have diets that are high in (saturated) fat intake and low in fruit and vegetable intake. It has been proposed that antioxidants have a protective effect by scavenging oxidants in blood plasma and on tissue level,14 which has been confirmed in some epidemiological studies13 but not in all.16 Since cigarette smokers are exposed to considerably more oxidants, it is generally assumed that they would need a higher intake of antioxidants to compensate for this increased exposure.17 In previous studies that estimated intake of different antioxidants by smoking status either no distinction was made between intensity of current smoking in the number of cigarettes smoked per day or ex-smokers were not considered separately,8-10"13'18 or no information was given on supplement use. 8 1 0 1 9

Smoking is known to be a major risk factor for several chronic diseases, including cardiovascular diseases, pulmonary diseases, and cancer.1 It has been suggested that reactive oxidants present in and released by cigarette smoke 2 react with biological membranes by causing membrane lipid oxidation, eventually leading to cardiovascular disease. 1 Exposure to oxidants may also result in DNA damage which may lead to the initiation of carcinogenesis. 3 In addition, inflammatory reactions in the lung may result from high levels of oxidants and may lead to pulmonary diseases.4 Components of the diet may modify the effect of smoking on oxidative damage to tissues. This modification depends on the type of nutrients considered, which can cause either an increase 5 ' 6 or a decrease in the risk

• Department of Chronic Diseases and Environmental Epidemiology, National Institute of Public Health and Environmental Protection, PO Box 1, 3720 BA Bilthoven, The Netherlands. * University of Leiden, The Netherlands Reprint requests to: Jacob C Seidell.

This observational study of over 5000 Dutch men and women aged 20-59 years examines dietary and 70

Downloaded from ije.oxfordjournals.org by guest on July 13, 2011

Background. Differences in dietary and supplementary intake of antioxidants were determined between different categories of smokers and never-smokers. Methods. Data from a large, cross-sectional, population-based study were used. Subjects (n = 4244) were divided into five smoking categories according to the number of cigarettes smoked per day. Differences in intake of antioxidants or frequency of supplement use were assessed using multiple linear regression analysis and multiple logistic regression analysis, adjusting for potential confounders such as age, body mass index, educational level, alcohol intake, and total energy intake. Results. Men who smoked >20 cigarettes/day had significantly lower intakes of (J-carotene and especially ascorbic acid compared to those who never smoked, resulting from an almost 60% lower fruit Intake. Moderate and heavy smoking women also had lower ascorbic acid and fruit Intake but differences were not as large as In men. A higher percentage of female heavy smokers compared with never-smokers consumed vitamin C (21.1% versus 14.1%), vitamin E (5.6% versus 1.8%), and multivitamin supplements (18.5% versus 12.2%). Among men only the moderate smokers differed significantly from never-smokers in supplement Intake, in the sense that male moderate smokers had a higher percentage of multtvttamin use (15.3% versus 12.2%) compared to never-smokers. Conclusions. Male heavy smokers not only have a lower dietary antioxidant intake than never-smokers, but additionally seem to use supplementation relatively infrequently. Keywords: smoking, antioxidants, ascorbic acid, [^carotene, vitamin E, vitamin supplements

ANTIOXIDANT INTAKE AND SMOKING

supplementary intakes of antioxidants in three different current smoking groups and ex-smokers in contrast to never-smokers, adjusting for factors that may confound the relationship between diet and smoking, such as age, body mass index (BMI) and alcohol intake.

In 1993 a total of 5155 subjects agreed to participate in the study (response rate: 48.3%). Since the presence of certain diseases and conditions may have led to altered dietary patterns and smoking status, 542 subjects who reported having had a heart attack, cerebrovascular accident, coronary-artery bypass grafting, or heart catheterization in the previous 5 years, those who had (a history of) cancer, and who were diabetic were excluded from analysis. In addition, those who were pregnant or who did not have Dutch nationality were also excluded from analysis, along with 146 people who followed a diet on the advice of a doctor or dietician. Since no information on the intensity and frequency of pipe or cigar smoking was collected in the questionnaire only cigarette smoking was considered and noncigarette smoking, pipe or cigar smokers were excluded from analysis (n = 28). Occasional smokers who smoked 1 cig/day), light smokers (1-9 cig/day), moderate smokers (10-19 cig/day), and heavy smokers (S21 cig/day). Ex-smokers, stratified into four groups according to duration since quitting, did not differ significantly in baseline characteristics, nutrient, or food intake and were therefore considered as one group. Statistical Analysis Differences in characteristics between the smoking groups were analysed using univariate Pearson's x 2 tests and analysis of variance. Multiple linear regression analysis was used to determine the mean nutrient and food group intakes by smoking status, adjusted for age (years), body mass index (kg/m2), educational level (low, medium, high), town (Amsterdam, Maastricht, Doetinchem), total energy intake (kJ/day), and alcohol intake (no/low, ^ 1 glass/week; intermediate, between 1 glass/week and 3 glasses/day (women) or 4 glasses/ day (men); high, > 3 glasses/day (women) or > 4 glasses/day (men)). Since all nutrient and food group distributions were approximately log-normally distributed, the analyses were performed on log-transformed values. In all analyses total P-values were determined and pairwise comparisons were made in relation to never-smokers using two-tailed least significance difference tests, f-values below 0.05 were considered to be statistically significant. Values of the adjusted means and 95% confidence intervals were backtransformed to antilog values (geometric means) for presentation purposes.


SUBJECTS AND METHODS The data used in the present study were collected in 1993 in the Monitoring Project on Risk Factors for Chronic Diseases in the Netherlands (MORGEN). The general purpose of the project is to determine the prevalence of risk factors for several chronic diseases: cardiovascular diseases, cancer, diabetes, migraine, and chronic non-specific lung diseases. The project was approved by the Medical Ethical Committee of TNO, the Netherlands. Randomized samples of men and women aged 2 0 59 and stratified according to gender and 5-year age categories were taken in three towns: Amsterdam, Maastricht, and Doetinchem. Those selected received full information on the project along with an invitation to participate. Participants received two questionnaires to complete at home, a general and a dietary questionnaire, and were invited to attend a medical examination at the Muncipal Health Service where blood samples were collected, blood pressure was measured, and anthropometric measurements were taken. Height was measured to the nearest 0.5 cm and weight to the nearest 0.1 kg, without shoes or heavy clothes and pockets emptied. The general questionnaire collects data on demographic characteristics and risk factors for chronic diseases (including smoking habits, alcohol consumption, and physical activity). The dietary questionnaire is a semi-quantitative food frequency questionnaire that was developed to assess the intake of total energy and the most important nutrients and food groups, with an emphasis on vegetables and fruits. In the questionnaire, information on the average consumption frequency and portion sizes of 80 main food items over the past year is collected. Vegetables and fruits, as major sources of antioxidants, were classified into subgroups in order to test whether results for the subgroups were different from those for the main groups. Nutrient intake was calculated by use of an extended version of the Dutch Food Composition Table. 20 In 1991-1992 the questionnaire was validated in a pilot study and slightly adjusted after this study (Personal communication: Ocke' M et al. The EPIC food frequency questionnaire in the Netherlands. I Description of the questionnaire and validity for food groups. II Validity and reproducibility for nutrients).

71

72

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

TABLE 1 Characteristics of the study population used in nutritional analysesfrysmoking status (n - 4244) Smoking status never

ex

light (1-9 c/d)"

moderate (10-19 c/d)

129 39.3(12.1) 24.9(3.1)

233 40.5(11.0) 24.7 (3.2)f

heavy (>20 c/d)

/>-valueb

Men 630 39.0(10.9) 25.4 (3.3)

N age in years (SD) Body mass index in kg/m2 (SD) Town' % Amsterdam % Maastricht % Doetinchem Educational level1 % low % intermediate %high Alcohol intake' % no/low % intermediate %high

924 42.1 (12.3) 25.1 (4.2)

659 46.7 (9.4)' 26.4 (3.3)*

265 42.9 (10.0)' 25.8 (3.8)

< 0.001 < 0.001

27.8 44.3 27.9

22.9 42.3 34.8

34.9 34.9 30.2

30.9 30.9 38.2

32.8 37.8 29.4

< 0.001

37.1 29.2 33.7

47.2 28.4 24.4

43.4 32.6 24.0

54.5 28.3 17.2

58.3 24.2 17.4

< 0.001

18.7 75.1 6.2

13.5 75.4 11.1

11.6 76.7 11.6

15.9 70.0 14.2

13.6 63.0 23.4

< 0.001

Women 614 207 40.7(11.1) 43.5 (9.6)' 24.2 (4.2)' 25.2 (3.6)

313 41.7(10.7) 24.4 (3.7)'

270 40.7 (9.6) 24.6 (4.3)

0.002 0.001

25.1 41.0 33.9

26.4 39.3 34.3

38.7 39.6 21.7

24.3 48.2 27.5

39.6 43.3 17.1

< 0.001

56.2 24.1 19.7

60.1 21.7 18.2

47.3 28.5 24.2

67.1 21.1 11.8

60.7 27.8 11.5

< 0.001

49.9 48.5 1.6

30.0 63.5 6.5

24.6 71.0 4.4

38.0 55.2 6.7

31.1 57.0 11.9

< 0.001

*Cigarettes/day. b Total P-value of analysis of variance or x 2 test c Not tested for pairwise significances of difference from never-smokers. ' Significantly different (P < 0.05) from never-smoken in least significant difference test.

Since quantitative amounts of supplement intake could not be determined, frequencies of intake were assessed. Subjects who reported to have used a supplement 'now and then', 'only in winter', or 'throughout the year' during the past year, were defined as supplement users. The number of supplement users was too small to consider these frequencies separately, i.e. 44, 14, and 48 men and 64, 22, and 68 women reported use throughout the year of vitamin C, vitamin E, and multivitamin supplements, respectively. Differences in frequencies of supplement use between smoking categories were analysed in a logistic regression model, adjusting for age, BMI, town, level of education, alcohol intake, and total

energy intake. All statistical analyses were performed using SAS release 6.07.

RESULTS The sample of 4244 subjects used in these analyses included 1916 men and 2328 women, and approximately one-third of both groups consisted of current smokers. Table 1 shows some characteristics of men and women by smoking status. Educational level and alcohol intake differed significantly across the smoking groups in men and women, with heavy smokers generally having a higher proportion of subjects with a


N age in years (SD) Body mass index in kg/m2 (SD) Townc * Amsterdam % Maastricht % Doetinchem Educational level c %low % intermediate %high Alcohol intake' % no/low % intermediate %high

73


TABLE 2 Geometric mean (95% confidence interval) daily nutrient intake by smoking status for men, adjusted for age, body mass index, town, level of education, alcohol intake, and total energy intake, unless stated otherwise (analyses performed on log-transformed data) Smoking status

Total energy (MJ)C Total protein (g) Total fat (g) saturated monounsaru rated polyunsaturated Cholesterol (mg) Total carbohydrate (g)

Iron (mg) Retinol (mg) fV-Carotene (mg) Ascorbic acid (mg) Vitamin E (mg)

moderate (10-19 c/d) (n = 233)

11.0 (10.8-11.3) 96.2' (95.1-97.2) 106.3* (105.0-107.6) 39.6 (39.1^0.2) 39.3' (38.8-39.9) 21.5' (21.0-22.0) 284 (279-289) 285.1* (282.0-288.3) 20.9 (20.5-21.3) 15.2 (15.0-15.4) 0.61 (0.58-0.63) 2.16 (2.09-2.23) 116.3 (112.6-120.2) 17.0 (16.7-17.3)

11.0 (10.6-11.5) 98.1 (95.8-100.3) 105 8 (103.1-108.7) 39.7 (38.5-^0.9) 39.2 (38.0-40.4) 21.1 (20.1-22.2) 288 (276-299) 286.5 (279.5-293.5) 20.4 (19.6-21.3) 14.9 (14.5-15.3) 0.60 (0.57-0.65) 2.15 (2.01-2.30) 119.0 (111.1-127.6) 17.1 (16.4-17.9)

11.4' (11.0-11.7) 94.3* (92.7-96.1) 105.3 (103.2-107.3) 39.0 (38.1-40.0) 39 2* (38.3-40.1) 21.3 (20.6-22.1) 283 (275-291) 286.2 (280.9-291.5) 20.3* (19.7-21.0) 14.9 (14.6-15.2) 0.63 (0.59-0.67) 1.98' (1.88-2.08) 112 4 (106.6-118.5) 17.0 (16.4-17.5)

ex

(n = 63O)

10.7 (10.5-10.9) 97.9 (96.8-99.0) 103.9 (102.7-103.2) 39.2 (38.7-39.8) 38.0 (37.5-38.6) 20 5 (20.0-21.0) 281 (276-286) 290.7 (287.4-294.1) 21.2 (20.8-21.6) 15.3 (15.1-15.4) 0.61 (0.60-0.62) 2.13 (2.07-2.20) 118.0 (114.3-121.9) 16.6 (16.2-16.9)

heavy (>20 c/d) (n - 265)

11.0 (10.7-11.4) 92.7* (91.2-94.3) 107.1* (105.2-109.2) 40.0 (39.2-40.9) 40.1* (39.2^*1.0) 21.4* (20.7-22.2) 290' (282-299) 281.3' (276.4-286.3) 18.5' (18.0-19.0) 14.5' (14.2-14.7) 0.65 (0.61-0.69) 1.96' (1.87-2.06) 95.8' (91.2-100.7) 16.8 (16.3-17.4)

P-valueb

0.07 < 0.001 0.04 0.47 0.0O1 0.08 0.34 0.03 < 0.001 < 0.001 0.23 0.002 < 0.001 0.43

"Cigarettes/day. b Total P-value of multilinear regression analysis, testing for differences across smoking categories. c Not adjusted for total energy intake. ' Significantly different from never-smokere (P < 0.05) in least significant difference test

high alcohol intake and a low educational level. Male heavy and ex-smokers and female moderate and heavy smokers also were significantly less active in sports, as estimated by the method of Baecke et al.21 (results not shown). Tables 2 and 3 show the geometric mean nutrient intakes per day by smoking status for men and women, adjusted for age, town, level of education, alcohol intake and total energy intake. Generally, in men large differences in nutrient intake compared with neversmokers were mainly observed for heavy smokers, while in women these differences were observed for both moderate and heavy smokers. Male heavy smokers and female moderate and heavy smokers had lower intakes of ascorbic acid than never-smokers. Male

heavy and moderate smokers had a significantly lower p-carotene intake than never-smokers, while the intake of this antioxidant was significantly higher in female ex-smokers. No significant differences were found for the intake of vitamin E according to smoking status. Tables 4 and 5 show the geometric mean food group intakes for men and women, adjusted for age, BMI, town, level of education, alcohol intake, and total energy intake. The largest differences in food group intake occurred for both male and female moderate and heavy smokers compared with never-smokers, but differences were most pronounced for male heavy smokers. They had significantly lower intakes of cruciferous vegetables, carotene-rich vegetables, and fruit/vegetable juice and they consumed almost 60%


Dietary fibre (g)

(n = 659)

light (1-9 c/d)' (n = 129)

never

74

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

TABLE 3 Geometric mean (95% confidence interval) daily nutrient intake by smoking status for women, adjusted for age, body mass index, town, level of education, alcohol intake, and total energy intake, unless stated otherwise (analyses performed on log-transformed data) Smoking status

Total energy (MJ)C Total protein (g) TotaJ fat (g) saturated monounsatu rated polyunsatu rated Cholesterol (mg) Total carbohydrate (g)

Iron (mg) Retinol (mg) ^-Carotene (mg) Ascorbic acid (mg) Vitamin E (mg)

ex (n = 614)

light (1-9 c/d)1 (n = 207)

moderate (10-19 c/d) (n = 313)

(n = 924)

8.3 (8 1-8.4) 77.2 (76.4-78.0) 79.7 (78.8-80.6) 29.7 (29.4-30.1) 29.1 (28.8-29.5) 16.2 (15.9-16.6) 220 (216-223) 225.6 (223.4-227 7) 17.9 (17.6-18.1) 12.3 (12.2-12.4) 0.43 (0.41-0.44) 2.17 (2.11-2.24) 131 5 (127.8-135.4) 14.2 (13.9-14.4)

8.4 (8.3-8.6) 77.0 (76.0-77.9) 81.9* (80.9-83.0) 30.2 (29.7-30.6) 30.2 f (29.8-30.6) 17.0* (16.6-17.4) 220 (216-224) 222.1 r (219.6-224.8) 18.4+ (18.1-18.8) 12.6* (12.4-12.7) 0.42 (0.40-0 44) 2.27* (2.19-2.36) 131.6 (127.1-136.5) 14.5 (14 2-14.8)

8.3 (8.0-8.6) 76.7 (75.0-78.3) 81.4 (79.7-83.2) 30.3 (29.5-31 1) 29.8 (29.1-30.6) 16.6 (15.9-17.4) 223 (216-230) 221.1 (217.0-225 2) 178 (17.3-18 3) 12.4 (12.1-12.6) 0.41 (0 38-0.43) 2.23 (2.11-2.37) 126.9 (119.7-134.6) 14 3 (13.8-14.9)

8.2 (8.0-8.5) 75.3* (74.0-76.6) 83.8* (82.4-85.4) 31.5* (30.9-32.2) 30.8* (30.1-31.4) 17.0* (16.4-17.7) 230* (224-236) 214.8* (211.5-218.3) 17.0* (16.6-17.4) 12.2 (12.0-12.4) 0.47* (0.45-0.50) 2.08 (1.98-2.19) 114.3* (108.8-120.2) 14.4 (14.0-14.9)

heavy (*20 c/d) (n o 270)

8.5 (8.3-8.8) 74.0* (72.6-75.4) 84 9* (83.3-86.6) 31.9* (31.2-32.6) 31.4* (30.8-32.1) 17.1* (16.5-17.8) 235* (229-242) 212.0* (208.5-215.7) 16.8* (16.3-17.2) 12.2 (11 9-12.4) 0.46* (0.44-0.49) 2.06 (1.95-2.17) 115.1* (109.2-121.4) 14.3 (13.9-14.8)

/>-valueb

0 38 0.001 < 0.001 < 0.001 < 0.001 0.02 < 0.001 < 0.001 < 0.001 0.001 < 0.001 0.004 < 0.001 0.47

"Cigarettes/day. Total P-value of multilinear regression analysis, testing for differences across smoking categories. Not adjusted for total energy intake. * Significantly different from never-smokers (P < 0.05) in least significant difference lest.

b c

less fruits (citrus and non-citrus) than never-smokers, while moderate smokers consumed 25% less fruit than never-smokers. Female moderate and heavy smokers also had a lower intake of total fruits (citrus and non-citrus) and fruit/vegetable juice, though not as markedly as male heavy smokers. The higher intake of total vegetables in female heavy smokers was due to a higher potato (product) intake. Figures 1-3 show the crude frequencies of total and specific supplement use in men and women. Results for vitamin E intake are not shown, since frequencies of intake were low (in women ranging from 1.8% in nonsmokers to 3.4% in heavy smokers, and in men ranging from 1.4% in non-smokers to 5.6% in heavy smokers).

After adjusting for age, BMI, town, educational level, alcohol intake, and total energy intake, a higher percentage of women compared to men used supplements in all smoking categories. In men supplement intakes of vitamin C, vitamin E, and multivitamins were slightly more frequent in heavy smokers than in neversmokers but differences were not significant. Only male moderate smokers differed significantly from never-smokers with a higher percentage taking multivitamins (P = 0.006). In women the differences with never-smokers were much larger: a higher percentage of light and heavy smokers used vitamin C (P = 0.03 and P = 0.04, respectively), ex-, moderate, and heavy smokers more frequently used vitamin E (P = 0.01, P = 0.05, P = 0.005, respectively), and light and heavy


Dietary fibre (g)

never

75


TABLE 4 Geometric mean (95% confidence interval) food group intake (g/day) by smoking status for men, adjusted for age, body mass index, town, level of education, alcohol intake, and total energy intake (analyses performed on log-transformed data) Smoking status

Total vegetables cruciferous carotene-rich fibre-nch potato (products) Total fruit citrus non-citrus

moderate (10-19 c/d) (n = 233)

328.0 (318.6-337.3) 25.0 (23.5-26.7) 58.6 (56.0-61.3) 13.1 (12.1-14.2) 123.0 (117.0-129.5) 94.7 (85.8-104.5) 21.6 (193-23.9) 71 0 (64.5-78.1) 42.1* (37.6-47.1)

319.2 (300.0-339.3) 24 9 (21.7-28.6) 56.6 (51.3-62.6) 12.4 (10.4-14.8) 110.3 (98.7-123.2) 99.7 (80.5-123.5) 21.9 (17.5-27.3) 74.1 (60.2-91.2) 53.2 (41.7-68.0)

328.7 (313.9-344.5) 22.8 (20.6-25.4) 55.0 (51.1-59.3) 128 (11.3-14.6) 126.6 (116.4-137.7) 76.8* (65.3-90.3) 18.1* (15.3-21.4) 57.3* (49.0-67.1) 37.2* (30.9-44 8)

ex

(n = 630)

322.7 (313.6-332.0) 24.2 (24.1-24.4) 56.5 (54.0-59.1) 12.4 (11.4-13.4) 123.2 (117.0-129.7) 102.1 (92.5-112.8) 23.4 (21.1-25.9) 76.5 (69.4-84.3) 52.0 (46.4-58.3)

heavy (»20 c/d) (n = 265)

316.0 (302.5-330.0) 20.5* (18.6-22.7) 50.5* (47.0-54.1) 11.9 (10.6-13.5) 130.3 (120.4-141.0) 41.9* (36.(M8.8) 9.5' (8.2-11.2) 32.2* (27 7-37.3) 34.7* (29.1-41.3)

/"-value11

0.60 0.01 0.01 0.75 0.18