Tobacco smoking in relation to body fat mass and distribution in a general population sample

Transcription

1 (2004) 28, & 2004 Nature Publishing Group All rights reserved /04 $ PAPER Tobacco smoking in relation to body fat mass and distribution in a general population sample C Bamia 1 *, A Trichopoulou 1, D Lenas 2 and D Trichopoulos 3 1 Department of Hygiene and Epidemiology, University of Athens Medical School, Athens, Greece; 2 Department of Ichthyology and Fisheries, Technological Educational Institute of Epirus, Igoumenitsa, Greece; and 3 Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA OBJECTIVE: To investigate the effects of variable amounts of tobacco smoking on body mass index and waist-to-hip ratio among current smokers. DESIGN: Population-based cohort study. SUBJECTS: A total of apparently healthy men and women, who enrolled in the Greek EPIC cohort, aged years, who had never smoked (14 751) or were current cigarette smokers (7308). MEASUREMENTS: Body mass index and waist-to-hip ratio (by anthropomentry), amount of tobacco smoking and energy expenditure (by an interviewer-administered, lifestyle questionnaire), energy intake and ethanol intake (by an intervieweradministered, validated, semiquantitative, food frequency questionnaire), at enrollment. RESULTS: In comparison to nonsmokers, smokers of the average number of cigarettes have lower body mass index. Among smokers, however, increased amount of smoking tends to be positively associated with body mass index, particularly among men. Waist-to-hip ratio is positively associated with amount of cigarettes smoked, among both men and women. CONCLUSION: Among smokers, tobacco smoking is positively associated with body mass index and waist-to-hip ratio. Our data suggest that the lower body mass index of smokers compared to nonsmokers reflects personality characteristics of those who choose to smoke and that the tendency to gain weight after smoking cessation may have behavioral rather than tobacco-related pharmacological roots. (2004) 28, doi: /sj.ijo Published online 15 June 2004 Keywords: smoking; body mass index; waist-to-hip ratio; energy intake; energy expenditure Introduction In the developed world, tobacco smoking and obesity are generally recognized as the most important modifiable factors responsible for excess mortality at the population level. 1,2 It has been reported that tobacco smokers have lower body mass index in comparison to nonsmokers and that those who stop smoking tend to gain weight, 3 8 the implication being that the benefits from smoking cessation may be partly alternated because of increases in body weight. It has not been established, however, that long-term smokers are heavier than nonsmokers on account of their smoking. It is also possible that those who take up smoking tend to have lower body mass index in comparison to nonsmokers because *Correspondence: Dr C Bamia, Department of Hygiene and Epidemiology, University of Athens Medical School, 75 Mikras Asias st, Athens, Greece. cbamia@nut.uoa.gr Received 12 January 2004; revised 30 March 2004; accepted 11 April 2004; published online 15 June 2004 of personality and lifestyle characteristics and that the weight gain associated with smoking cessation is a reaction to withdrawal symptoms Differentiation between these alternatives requires an examination of the association between tobacco smoking and body mass index among current smokers alone. Such a design also allows a more straightforward examination of the reported positive association between tobacco smoking and waist-to-hip ratio We have investigated the association between tobacco smoking on the one hand and body mass index and waist-tohip ratio on the other, in a large general population sample of adult women and men in Greece. A widely held, but poorly documented, view is that tobacco smoking, in spite its other catastrophic consequences, may reduce body weight and certain obesity indicators. We have excluded from the study ex-smokers among whom uncontrollable confounding may be operating because of the variability of reasons that may have led to smoking cessation. An unusual advantage of the present study was that energy intake and energy expenditure were available for each individual. 17,18

2 1092 Smoking and body fat mass and distribution Methods The European Prospective Investigation into Cancer and Nutrition (EPIC) is a prospective cohort study, conducted in 22 research centers across 10 European countries under the coordination of the International ncy for Research on Cancer (IARC), with the purpose of investigating the role of nutrition and other lifestyle and environmental factors in the etiology of cancer and other chronic diseases. 19,20 In the context of the Greek component of the EPIC study, a total of (men and women) apparently healthy volunteers, years old, were recruited from all regions of Greece during The large preponderance of women was intentional in the design of the EPIC study, because one of the primary objectives was to explore the nutritional etiology of breast cancer. All procedures were in line with the Helsinki declaration for human rights, all volunteers signed informed consent forms and the study protocol was approved by ethical committees at IARC and at the University of Athens Medical School. Each EPIC participant underwent a baseline examination during which extensive questionnaires were administered by specially trained interviewers. Information on various sociodemographic variables, medical history, lifestyle habits and dietary intakes was recorded in the questionnaires. In addition, anthropometry and blood pressure measurements were performed and a blood sample was collected. Information on smoking status was collected at recruitment through a special section of the lifestyle questionnaire (smoking history questionnaire). Subjects were categorized as never smokers (never smoked cigarettes, cigars or pipes), current cigarette smokers and ex-smokers (quit smoking for at least 1 year prior to their enrollment) on the basis of their self-reported status on the of their recruitment. For current and ex-smokers, the number of cigarettes smoked per, as well as, the age at which they have started and (for exsmokers) quitted smoking were recorded. Current cigarette smokers were further divided into four subgroups according to the average number of cigarettes smoked daily: less than 10, 10 19, and 30 or more cigarettes. Very few smokers were exclusive cigar and/or pipe smokers and these were included among those smoking cigarettes. Concerning anthropometry, all measurements were carried out with an inelastic tape and were recorded to the nearest 0.1 cm, while participants were without shoes, lightly clothed, with no restrictive underwear. Measurements were taken at the end of the normal respiration while the participant was standing erect, with the arms at the side and the feet together. Waist circumference was measured around the smallest circumference between the lowest rib and iliac crest, or, for obese subjects with no natural waist, midway between the lowest rib and iliac crest. Hip circumference was measured horizontally at the level of the greatest lateral extension of the hips. Waist-to-hip ratio was calculated using the waist and hip circumferences. Body mass index was calculated as the ratio of weight in kilograms divided by the square of the height in meters. A section of the lifestyle questionnaire addressed the frequency and duration of participation in occupational and leisure time physical activities. 21 A metabolic equivalent index was computed by assigning a multiple of resting metabolic rate to each activity (MET value). 22 Time spent on each of the above activities was multiplied by the MET value of the activity, and all MET-hour products were summed to produce an estimate of daily physical activity, indicating the amount of energy per kilogram of body weight expended during an average. Usual dietary intake over the year preceding enrollment was assessed by a validated, semiquantitative food frequency questionnaire, including approximately 150 foods and beverages commonly consumed in Greece. Standard portion sizes were used for the estimation of consumed quantities, 23,24 and nutrient intakes were calculated using a food composition database modified to accommodate the particularities of the Greek diet. 24,25 For each participant, grams of intake of various food groups and ethanol and, eventually, total energy intake (in kj) were calculated. 26 Smoking history was available for of the study participants. For 765 study participants, however, information was missing regarding the number of cigarettes smoked and/or one or more anthropometric measurements, and these were excluded. We also excluded 5136 participants who were ex-smokers in order to avoid any possible interference of the effect of current smoking with a residual effect of past smoking on body mass index and/or waist-tohip ratio. Thus, in this report, we present results for individuals, 7325 men and women, aged years, who had never smoked (14 751) or were current cigarette smokers (7308) as of the date of recruitment, and for whom complete information on all study variables was available. Statistical analysis Statistical analysis was initially performed through cross tabulations. Subsequently, the data were modeled through multiple linear regression assessing the effect of cigarette smoking (exposure) on body mass index or, alternatively, waist-to-hip ratio (dependent variables), while controlling for a number of potential confounders. Possible confounders were age (categorically: less than 45, 45 54, and 65 or more years), energy intake (continuously, in kcal), energy expenditure (continuously, in met-h /) and average daily intake of ethanol in grams (continuously). In regression models for waist-to-hip ratio, body mass index (continuously, in kg/m 2 ) was also included as a potential confounder. All models were fitted separately for each gender. In all regression analyses the impact of current cigarette smoking on either body mass index or waist-to-hip ratio (controlling for body mass index) was modeled through two variables: one indicator variable denoting smoking status (1 ¼ current cigarette smoker, 0 ¼ never smoker) and one centered (around the mean) continuous variable denoting the number of cigarettes smoked daily (among smokers of

3 the respective gender). In this way the coefficient of the indicator variable evaluates the effect of smoking the mean number of cigarettes vs never smoking, while the coefficient of the continuous centered variable evaluates the effect on the dependent variable of a specified increment in cigarettes smoked. The continuous variable is introduced only as a product term with the indicator variable, whereas the latter is also introduced on its own. 27,28 In all comparisons a level of 5% was used to denote statistical significance. All statistical analyses were performed with the STATA statistical package. 29 Results In the investigated general population sample, there were 7308 current cigarette smokers and never smokers. Their distribution by gender, age and cigarettes smoked per is shown in Table 1. The table demonstrates the wellknown heavy smoking habits of Greek men and the alarming increase of tobacco smoking among successive cohorts of Greek women. The mean number of cigarettes smoked per was 25 (1 90) and 16 (1 70) for male and female smokers respectively. Among men, the mean daily energy intake was kj (2432 kcal) (standard deviation 3075 kj) and the mean daily energy expenditure (per kg) was 35.6 MET-h (standard deviation 6.4 MET-h). Among women, mean daily energy intake was 7874 kj (1881 kcal) (standard deviation 2406 kj) and mean daily energy expenditure (per kg) was 35.4 MET-h (standard deviation 4.3 MET-h). Table 2 shows the mean (and standard deviation) of body mass index and waist-tohip ratio by age, gender and current smoking status in the study population. In these tables no account is taken for energy intake or energy expenditure and the results are inherently confounded. Nevertheless, the data in the table are in line with the expected patterns, in that body mass index increases with age among both genders but more sharply among women, and in that, waist-to-hip ratio Smoking and body fat mass and distribution increases also with age among both genders but it is considerably higher among men. Table 3 shows multiple regression-derived partial regression coefficients, their standard errors and the associated 95% confidence intervals of body mass index and waist-tohip ratio on current cigarette smoking, by number of cigarettes. In comparison to nonsmokers, smokers of the average number of cigarettes have lower body mass index. Among smokers, however, increased amount of smoking tends to be positively associated with body mass index and the association is somewhat stronger among men. After adjusting for body mass index, waist-to-hip ratio is positively associated with amount of cigarettes smoked per, among both men and women. When comparing nonsmokers with smokers of the average number of cigarettes, however, with respect to waist-to-hip ratio smokers have higher mean value among men, whereas among women there is a suggestion that the opposite could be true. We have also modeled the data introducing an interaction term involving tobacco smoking and energy intake, in addition to the main effects of these variables and to other indicated covariates (age, energy expenditure and ethanol intake), using as outcome either body mass index or waist-tohip ratio (controlling for BMI). As expected, the results were in line with those emerging from the analysis outlined in Table 3. Discussion In this large population-based study with control for age, energy intake, ethanol intake and energy expenditure, we found that among smokers, number of cigarettes smoked per is positively associated with body mass index, particularly among men. When smokers of the average number of cigarettes were compared to nonsmokers, however, the former were found to have lower body mass index. The apparent contradiction may be explained by postulating that smokers, on account of their personality and lifestyle characteristics, tend to be lighter rather than heavier in 1093 Table 1 Distribution of 7308 current cigarette smokers and never smokers from the Greek EPIC study, by age, gender and current smoking status Never smokers (%) 1 9 cigarettes (%) cigarettes (%) cigarettes (%) 30+ cigarettes (%) All (%) Males (7325 (33.21%)) r (30.0) 190 (7.1) 288 (10.7) 564 (20.9) 845 (31.3) 2696 (100.0) (30.5) 148 (8.4) 132 (7.5) 361 (20.4) 588 (33.2) 1769 (100.0) (46.3) 112 (7.8) 140 (9.7) 239 (16.6) 282 (19.6) 1439 (100.0) Z (57.6) 146 (10.3) 148 (10.4) 197 (13.9) 112 (7.9) 1421 (100.0) Females (14734 (66.79%)) r (56.2) 479 (12.4) 429 (11.1) 495 (12.8) 289 (7.5) 3866 (100.0) (77.1) 236 (7.1) 180 (5.4) 210 (6.3) 142 (4.2) 3348 (100.0) (93.1) 95 (2.5) 76 (2.0) 68 (1.8) 24 (0.6) 3834 (100.0) Z (97.5) 47 (1.3) 24 (0.7) 19 (0.5) 3 (0.1) 3686 (100.0)

4 Smoking and body fat mass and distribution 1094 Table 2 Mean (7s.d.) of body mass index and waist-to-hip ratio by age, gender and current smoking status among 7308 current cigarette smokers and never smokers from the Greek EPIC Study Body mass index (kg/m 2 ) Waist-to-hip ratio 30+ cigarettes per cigarettes cigarettes Never smokers 1 9 cigarettes per 30+ cigarettes per cigarettes cigarettes Never smokers 1 9 cigarettes per Males (7325 (33.21%)) r (3.48) (3.38) (3.87) (3.65) (4.17) 0.92 (0.07) 0.92 (0.07) 0.92 (0.07) 0.92 (0.08) 0.93 (0.07) (3.35) (3.39) (3.43) (9.74) (4.06) 0.95 (0.06) 0.95 (0.06) 0.95 (0.07) 0.96 (0.07) 0.97 (0.07) (3.93) (4.43) (3.56) (3.81) (3.98) 0.97 (0.06) 0.97 (0.06) 0.97 (0.06) 0.97 (0.06) 0.99 (0.08) Z (3.82) (3.92) (4.15) (4.30) (4.43) 0.97 (0.07) 0.97 (0.06) 0.96 (0.07) 0.98 (0.09) 0.99 (0.07) Females ( (66.79%)) r (5.21) (5.10) (4.60) (5.01) (5.57) 0.77 (0.07) 0.76 (0.06) 0.76 (0.08) 0.77 (0.06) 0.78 (0.07) (5.09) (5.25) (5.05) (4.87) (5.77) 0.81 (0.07) 0.79 (0.07) 0.80 (0.08) 0.80 (0.08) 0.80 (0.07) (5.08) (4.35) (5.11) (5.48) (4.93) 0.84 (0.08) 0.82 (0.07) 0.82 (0.07) 0.83 (0.08) 0.83 (0.09) Z (4.90) (4.90) (6.57) (4.79) (3.44) 0.87 (0.08) 0.83 (0.08) 0.85 (0.07) 0.87 (0.09) 0.84 (0.02) comparison to nonsmokers, whereas smoking per se in not inversely related with body mass index and, particularly among men, tends to be positively associated with it. After controlling for body mass index as well as for age, energy intake, ethanol intake and energy expenditure, tobacco smoking appears to be positively associated with waist-tohip ratio among both men and women. Many studies have evaluated body mass index in relation to tobacco smoking and most of them had indicated that nonsmokers tend to be heavier than smokers. 12,30,31 The results of our study do not contradict this evidence but indicate that the lower average body mass index among smokers may be due to personality and lifestyle characteristics of the smokers rather than to direct effects of smoking, since, among smokers, there is a positive association of amount of smoking with body mass index. This interpretation is not incompatible with existing evidence from other studies and indeed the few investigations that have evaluated dose response of body mass index in relation to tobacco smoking with adequate control for possible confounding have also reported a positive association Fewer studies have evaluated waist-to-hip ratio in relation to tobacco smoking after controlling for body mass index, 13,14,16 and even though in many of them energy intake and expenditure were not accounted for the results were consistent in indicating a positive association. Our results that focus on dose response trends among smokers point to positive associations among both men and women whereas smoking women tend to have somewhat lower waist-to-hip ratio probably on account of personality and lifestyle characteristics. Advantages of our study are its large size and its reliance on a general population sample, control for possible confounding by energy intake and expenditure and the focus on trends among smokers, which accommodates possible selection factors influencing the choice of becoming a smoker. A weakness of the study is that it is of cross-sectional nature, a fact that does not allow the introduction of time elements in the further exploration of the reported associations. The mechanisms of the association between tobacco smoking on the one hand and body mass index and waistto-hip ratio on the other are difficult to identify because tobacco smoke contains many hundreds of compounds, and the regulation of body mass index depends on several physiological processes. Differential effects of tobacco smoking on sex hormones and on adipocytes have been implicated, 13,14,16,35 but, at this stage, documentation of the nature of the associations between tobacco smoking and body mass index parameters; as well; as the public health implications of these associations may take precedence over the clarification of the underlying physiological and pathophysiological mechanisms. In conclusion, in a large population based study we have found evidence that, among smokers, tobacco smoking is positively associated with body mass index and waist-to-hip ratio. Our data indicate that the lower body mass index of

5 Smoking and body fat mass and distribution Table 3 Multiple regression-derived partial regression coefficients (b), their standard errors and the 95% associated confidence intervals of body mass index and waist-to-hip ratio on current cigarette smoking 1095 Males (7325 (33.21%)) Females ( (66.79%)) b (standard error) 95% confidence interval b (standard error) 95% confidence interval Body mass index (in kg/m 2 ) a Smoking the mean number of cigarettes vs nonsmoking (0.108) ( to 0.307) (0.118) ( to 0.939) Among smokers incremental effect per 10 cigarettes/ (0.044) (0.075 to 0.250) (0.082) ( to 0.305) Waist-to-hip ratio b Smoking the mean number of cigarettes vs nonsmoking (0.002) (0.009 to 0.148) (0.002) ( to 0.001) Among smokers incremental effect per 10 cigarettes/ (0.0006) (0.002 to 0.005) (0.001) (0.003 to 0.007) a Adjusted for age (categorically as in Table 1), total energy intake (continuously), intake of ethanol (continuously) and total energy expenditure (continuously). b Adjusted for age (categorically as in Table 1), total energy intake (continuously), intake of ethanol (continuously), total energy expenditure (continuously) and body mass index (continuously). smokers compared to nonsmokers reflects personality characteristics of those who choose to smoke and that the tendency to gain weight after smoking cessation may have behavioral rather than tobacco-related pharmacological roots. From a public health point of view the findings of our study suggest that smokers cannot expect a health benefit from the association of their habit with body mass index and adipose tissue distribution. Acknowledgements The European Prospective Investigation into Cancer and Nutrition (EPIC) is coordinated by the International ncy for Research on Cancer (World Health Organization) and supported by the Europe Against Cancer Program of the European Commission. The Greek segment of the EPIC study is also supported by the Greek Ministry of Health and the Greek Ministry of Education. References 1 Adami HO, Trichopoulos D. Obesity and mortality from cancer. N Engl J Med 2003; 348: Poulter N. Global risk of cardiovascular disease. Heart 2003; 89: ii2. 3 Blitzer PH, Rimm AA, Giffer EE. The effect of cessation of smoking on body weight in women: cross-sectional and longitudinal analyses. J Chronic Dis 1977; 30: Stamford BA, Matter S, Fell RD, Papanek P. Effects of smoking cessation on weight gain, metabolic rate caloric consumption, and blood lipids. Am J Clin Nutr 1986; 43: Albanes D, Jones DY, Micozzi MS, Mattson ME. Associations between smoking and body weight in the US population: analysis on NHANES II. Am J Public Health 1987; 77: Williamson DF, Madans J, Anda RG, Kleinman JC, Giovino GA, Byers T. Smoking cessation and severity of weight gain in a national cohort. N Engl J Med 1991; 324: Froom P, Melamed S, Benbassat J. Smoking cessation and weight gain. Fam Pract 1998; 46: Owen-Smith V, Hannaford PC. Stopping smoking and body weight in women living in the United Kingdom. Br J Gen Pract 1999; 49: Hughes JR. Tobacco withdrawal in self-quitters. J Consult Clin Psychol 1992; 60: Patten CA, Martin JE. Measuring tobacco withdrawal: a review of self-report questionnaires. J Subst Abuse 1996; 8: Piasecki TM, Fiore MC, Baker TB. Profiles in discouragement: two studies of variability in the time course of smoking withdrawal symptoms. J Abnorm Psychol 1998; 107: Barrett-Connor E, Khaw K. Cigarette smoking and increased central adiposity. Ann Intern Med 1989; 111: Shimokata H, Muller DC, Andres R. Studies in the distribution of body fat. III. Effects of cigarette smoking. JAMA 1989; 261: Troisi RJ, Heinold JW, Vokonas PS, Weiss ST. Cigarette smoking, dietary intake, and physical activity: effects on body fat distributionfthe Normative Aging Study. Am J Clin Nutr 1991; 53: Simon JA, Seeley DG, Lipschutz RC, Vittinghoff E, Browner WS. The relation of smoking to waist-to-hip ratio and diabetes mellitus among elderly women. Prev Med 1997; 26: Visser M, Launer LJ, Deurenberg P, Deeg DJ. Past and current smoking in relation to body fat distribution in older men and women. J Gerontol A Biol Sci Med Sci 1999; 54: M293 M Trichopoulou A, Gnardellis C, Lagiou P, Benetou V, Naska A, Trichopoulos D. Physical activity and energy intake selectively predict the waist-to-hip ratio in men but not in women. Am J Clin Nutr 2001; 74: Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean diet and survival in a Greek population. N Engl J Med 2003; 348: Riboli E, Kaaks R. The EPIC project: rational and study design. Int J Epidemiol 1997; 26: S6 S Costacou T, Bamia C, Ferrari P, Riboli E, Trichopoulos D, Trichopoulou A. Tracing the Mediterranean diet through principal components and cluster analyses in the Greek population. Eur J Clin Nutr 2003; 57: Trichopoulou A, Gnardellis C, Lagiou P, Benetou V, Trichopoulos D. Body mass index in relation to energy intake and expenditure among adults in Greece. Epidemiology 2000; 11: Ainsworth B, Haskell WL, Leon AS. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc 1993; 25: Gnardellis C, Trichopoulou A, Katsouyianni K, Polychronopoulos E, Rimm EB, Trichopoulos D. Reproducibility and validity of an extensive semi-quantitative Food Frequency Questionnaire among Greek school teachers. Epidemiology 1995; 6: Trichopoulou A. Composition of Greek Foods and Dishes (in Greek and English). Athens School of Public Health: Athens; Trichopoulou A, Georga K. Composition Tables of Simple and Composite Foods. Parisianos: Athens; Katsouyianni K, Rimm EB, Gnardellis C, Trichopoulos D, Polychronopoulos E, Trichopoulou A. Reproducibility and rela-

6 1096 Smoking and body fat mass and distribution tive validity of an extensive semi-quantitative Food Frequency Questionnaire using dietary records and chemical markers among Greek school teachers. Int J Epidemiol 1997; 26: S118 S Trichopoulos D, Hsieh CC, MacMahon B, Lin TM, Lowe CR, Mirra AP, Ravnihar B, Salber EJ, Valaoras VG, Yuasa S. at any birth and breast cancer risk. Int J Cancer 1983; 31: Miettinen OS. Theoritical Epidemiology: Principles of Occurrence Research in Medicine. Wiley: New York, NY; STATA [computer program]. STATA Corporation: Intercooled Stata 7.0 for Windows 98/95/NT. Texas, USA; Martinez JA, Kearney JM, Kafatos A, Paquet S, Martinez-Gonzalez MA. Variables independently associated with self-reported obesity in the European Union. Public Health Nutr 1999; 2: de Martins VM, Almeida RMRV, Pereira RA, Barros MBA. Factors associated with overweight and central body fat in the city of Rio de Janeiro: results of a two stage random sampling survey. Public Health 2001; 115: Colditz GA, Segal MR, Myers AH, Stampfer MJ, Willett W, Speizer FE. Weight change in relation to smoking cessation among women. J Smoking Relat Disord 1992; 3: Istvan JA, Nides MA, Buist AS, Greene P, Voelker H. Salivary cotinine, frequency of cigarette smoking, and body mass index: findings at baseline in the Lung Health Study. Am J Epidemiol 1994; 139: Liu L, Choudhury SR, Okayama A, Hayakawa T, Kita Y, Ueshima H. Changes in body mass index and its relationships to other cardiovascular risk factors among Japanese population: results from the 1980 and 1990 national cardiovascular surveys in Japan. J Epidemiol 1999; 9: Klesges RC, Eck LH, Isbell TR, Fulliton W, Hanson CL. Smoking status: effects on the dietary intake, physical activity, and body fat of adult men. Am J Clin Nutr 1990; 51: