CiteSeerX — Association of Cigarettes Smoked Daily with Obesity in a General Adult Population

(1)

Obesity in a General Adult Population

Arnaud Chiolero,*† Isabelle Jacot-Sadowski,*† David Faeh,*‡ Fred Paccaud,* and Jacques Cornuz*†

Abstract

CHIOLERO, ARNAUD, ISABELLE JACOT-

SADOWSKI, DAVID FAEH, FRED PACCAUD, AND JACQUES CORNUZ. Association of cigarettes smoked daily with obesity in a general adult population. Obesity.

2007;15:1311–1318.

Objective: We analyzed the cross-sectional association be- tween obesity and smoking habits, taking into account diet, physical activity, and educational level.

Research Methods and Procedures: We used data from the 2002 Swiss Health Survey, a population-based cross-sectional telephone survey assessing health and self-reported health behaviors. Reported smoking habits, height, and weight were available for 17,562 subjects (7844 men and 9718 women) ⱖ25 years of age. BMI was calculated as (self-reported) weight divided by height².

Results: Mean BMI was 25.1 kg/m²for non-smokers, 26.1 kg/m²for ex-smokers, 24.6 kg/m²for light smokers (1 to 9 cigarettes/d), 24.8 kg/m² for moderate smokers (10 to 19 cigarettes/d), and 25.3 kg/m²for heavy smokers (ⱖ20 cigarettes/d) in men and 24.0, 24.1, 22.9, 22.9, and 23.3 kg/m², respectively, in women. Obesity (BMI ⱖ 30 kg/m²) was increasingly frequent with older age, lower physical activity, lower fruits/vegetables intake, and lower educational level. Compared with non-smokers, the odds ratio for obesity vs. normal weight (BMI⫽ 18.5 to 25.0 kg/m²) adjusted for age, nationality, educational level, leisure time physical activity, and fruit/vegetable intake were 1.9 (95% confidence interval: 1.5 to 2.3) for ex-smokers, 0.5 (0.3 to 0.8)

for light smokers, 0.7 (0.4 to 1.0) for moderate smokers, and 1.3 (1.0 to 1.7) for heavy smokers in men and 1.3 (1.1 to 1.6), 0.7 (0.5 to 1.0), 0.8 (0.5 to 1.0), and 1.1 (0.8 to 1.4), respectively, in women.

Discussion: Among smokers, obesity was associated in a graded manner with the number of cigarettes daily smoked, particularly in men. More emphasis should be put on the risk of obesity among smokers.

Key words: weight, risk factors, smoking

Introduction

Obesity and smoking are leading causes of preventable morbidity and mortality worldwide (1,2). Scientific and popular views on the association between smoking and body weight are focused on the common observation that smoking cessation leads to weight gain; for example, in a national U.S. cohort, the weight gain attributable to cessation was estimated to be 2.8 kg in men and 3.8 kg in women (3). A matter of more concern is the fact that a substantial part of smokers fear an increase in body weight and consequently neglect the benefits associated with smoking cessation (4,5).

The association between smoking and body weight is complex. On the one hand, smoking increases energy expenditure (6,7) and might suppress appetite (4). Numerous studies report that smokers have lower mean body weight and lower mean BMI (8,9), and those who quit smoking tend to put on weight (3,10,11). Smoking initiation among girls could be related to weight control (12), and in elderly women, attempts for smoking cessation could be limited by fear of weight gain (5). On the other hand, male and female smokers tend to cumulate other risk behaviors potentially favoring weight gain, e.g., poor diet or low physical activity (13,14); such a clustering is potentially conducive to higher weight among heavy smokers compared with other smokers. There are controversies about the relationship between body weight and the amount of cigarettes smoked daily.

Whereas among smokers, previous studies suggested a U- shaped relationship between weight and daily consumption

Received for review August 9, 2006.

Accepted in final form November 26, 2006.

The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

*Institute of Social and Preventive Medicine (IUMSP), Centre Hospitalier Universitaire Vaudois and University of Lausanne, Lausanne, Switzerland; †University Outpatient Clinic, Lausanne, Switzerland; and ‡Department of Physiology, University of Lausanne, Lausanne, Switzerland.

Address correspondence to Arnaud Chiolero, Prevention Unit, Institute of Social and Preventive Medicine, University of Lausanne, 17, rue du Bugnon, 1005 Lausanne, Switzer- land.

(2)

of cigarettes (15,16), moderate smokers weighing less, re- cent reports suggest that weight increases according to the daily consumption of cigarettes (9,17,18). In a prospective study of adolescents, smoking initiation was followed by an increase of BMI after 1 or 2 years (19).

Using a large sample of the general adult population, our objective was to analyze the association between smoking and BMI, taking into account risk behaviors potentially favoring weight gain (low leisure-time physical activity;

low fruit/vegetable intake) and educational level. A specific attention was devoted to the prevalence of obesity in relation to the number of cigarettes smoked daily.

Research Methods and Procedures

We previously described the Swiss Health Survey (14). It is a cross-sectional, nationwide, population-based telephone survey conducted every 5 years since 1992 by the Federal Statistical Office of Switzerland to track public health trends (14,20,21). It consists of questions on health status and health behaviors asked of a representative sample of adults and adolescents (ⱖ15 years of age) living in Swit- zerland. The study population was chosen by stratified random sampling of a database of all private Swiss households with telephones in 2002. The entire sample (30,824 households) received a letter inviting one household mem- ber to participate in the survey. Sampled households were contacted thereafter by phone. The survey was completed by 19,706 subjects (8909 men and 10797 women), corre- sponding to a participation rate equal to 63.9%. Data on smoking habits, weight, and height among adultsⱖ25 years of age were available for 17,562 persons (7844 men and 9718 women).

Three age categories were considered: 25 to 44, 45 to 64, and ⱖ65 years. Nationality was dichotomized as either Swiss or foreign. Education was categorized as follows (22,23): 1) no education completed, 2) first level (primary school), 3) lower secondary level, 4) upper secondary level, and 5) tertiary level, which included university and other forms of education after the secondary level. We defined

“low education” (categories 1 and 2), “middle education”

(categories 3 and 4), and “high education” (category 5) groups.

Subjects were asked about their current body weight and height. BMI was calculated as weight divided by height². Subjects were considered underweight, normal weight, overweight, and obese if the BMI was ⬍18.5, ⱖ18.5 and

⬍25, ⱖ25 and ⬍30, and ⱖ30 kg/m², respectively (24).

Subjects were categorized as non-smokers if they did not smoke currently and had not smoked regularly for ⬎6 months. Subjects were categorized as ex-smokers if they had ever smoked regularly forⱖ6 months but did not smoke any more. Subjects who smoked cigarettes were divided in three predefined categories according to daily consumption:

light smoker (1 to 9 cigarettes/d), moderate smoker (10 to

19 cigarettes/d), and heavy smoker (ⱖ20 cigarettes/d). Sub- jects smoking only cigars, pipes, or cigarillos were categorized as other smokers. Subjects smoking less than one cigarette daily were considered as non-smokers.

Subjects were asked about the physical activity they performed weekly during leisure time. Low leisure-time physical activity was defined as the absence of any physical activity during leisure time that caused the subject to sweat (25). Intake of fruits, fruit juice, and vegetables was assessed as part of a brief food frequency questionnaire (21).

Low fruit/vegetable intake was defined as either not taking fruits (or fruit juice) every day or not taking vegetables (including lettuce, not including potatoes) every day.

We reported mean BMI and the prevalence of BMI categories by nationality, educational level, smoking status, physical activity, and fruit/vegetable intake, stratified by sex and age. We applied maximum-likelihood multinomial (polytomous) logistic regression to assess the association between smoking categories and body weight categories separately for men and women. Models included smoking category as an independent variable and were adjusted for age, educational level, nationality, physical activity during leisure time, and fruit/vegetable intake. The interaction term of age and smoking categories was also tested. Pylotomous regression models allowed computation of the “relative risk ratio” or “pseudo odds ratio” (26). In this report, the term odds ratio (OR)¹ will be used to ease reading. Analyses were performed with Stata 9.0 software (StataCorp Lp., College Station, TX).

Results

Table 1 shows the general characteristics of subjects.

Approximately every one of two men and one of three women were overweight or obese. One of four participants currently smoked cigarettes. Compared with women, men were more frequently physically active during leisure time (low leisure-time physical activity: 35.7⫾ 0.6% of men vs.

44.6⫾ 0.5% of women; p ⬍ 0.001) but less frequently ate fruits or vegetables (low fruit/vegetable intake: 42.0⫾ 0.6%

vs. 24.1⫾ 0.4%; p ⬍ 0.001).

Excess weight (overweight or obesity) was more frequent with increasing age in both sexes. Participants with excess weight were less often physically active in leisure time and less often ate fruits/vegetables than normal weight participants. Participants who were underweight had more frequent low leisure-time physical activity than normal weight participants. In both sexes and across all ages, overweight and obesity were more frequent with lower educational level. Underweight was equally frequent across educational levels (data not shown).

1Nonstandard abbreviations: OR, odds ratio; CI, confidence interval.

(3)

Mean BMI by sex, age, and smoking status is shown in Table 2. For both sexes, ex-smokers had higher mean BMI compared with both never-smokers (difference of 0.1 to 1.0 unit of BMI) and cigarette smokers (0.4 to 1.3 units of BMI). Male cigarette smokers had a mean BMI similar or lower to that of non-smokers, whereas female cigarette smokers had systematically lower BMIs than non-smokers.

Among male smokers 25 to 44 or 45 to 64 years of age, BMI increased with the number of cigarettes daily smoked in all age groups. In the oldest men, a J-shaped relationship was observed. Among female smokers, a U-shaped relationship was observed in older women (45 to 64 and ⱖ65 years):

light and heavy smokers had higher BMI compared with moderate smokers. This was not true for younger women (25 to 44 years); BMI was higher in moderate and heavy smokers. Overall, mean BMI was higher in heavy smokers compared with light and moderate smokers (difference of

0.3 to 1.0 unit of BMI). Other smokers (cigar, pipe, or cigarillo smokers) had the highest mean BMI compared with non-, ex-, or cigarette smokers.

Prevalence of obesity according to smoking status is shown in Figure 1. Prevalence of obesity was higher in ex-smokers and heavy smokers compared with light smokers. Proportion of obesity increased across categories of male cigarette smokers. In female cigarette smokers, a U-shape or J-shape relationship between smoking status and prevalence of obesity was observed for older age strata (45 to 64 andⱖ65 years). In women 25 to 44 years of age, prevalence of obesity did not increase in heavy smokers compared with light and moderate smokers.

Associations of smoking habits with overweight and obesity after adjustment for various characteristics are shown in Tables 3 and 4 (among all participants, with non-smokers as Table 1. Demographic characteristics, weight categories, and smoking habits of subjects

Men Women

N Percent (SE) N Percent (SE)

Age (years)

25 to 44 3337 42.5 (0.6) 3740 38.5 (0.5)

45 to 64 2783 35.5 (0.5) 3483 35.8 (0.5)

ⱖ65 1724 22.0 (0.5) 2495 25.7 (0.4)

Nationality

Swiss 6832 87.1 (0.4) 8647 89.0 (0.3)

Foreign 1012 12.9 (0.4) 1071 11.0 (0.3)

Educational level

High 2120 27.1 (0.5) 983 10.1 (0.3)

Middle 4816 61.5 (0.5) 6406 66.0 (0.5)

Low 900 11.5 (0.4) 2318 23.9 (0.4)

Weight categories

Underweight 66 0.8 (0.1) 507 5.2 (0.2)

Normal-weight 3853 49.1 (0.6) 6074 62.5 (0.5)

Overweight 3191 40.7 (0.6) 2337 24.1 (0.4)

Obesity 734 9.3 (0.3) 800 8.2 (0.3)

Smoking habits

Non-smokers 3133 39.9 (0.5) 5727 58.9 (0.5)

Ex-smokers 2236 28.5 (0.5) 1760 18.1 (0.4)

Cigarette smokers 1873 23.9 (0.5) 2201 22.7 (0.4)

Light (1 to 9 cigarettes/d) 383 4.9 (0.2) 656 6.8 (0.3)

Moderate (10 to 19 cigarettes/d) 504 6.4 (0.3) 744 7.7 (0.3)

Heavy (ⱖ20 cigarettes/d) 986 12.6 (0.4) 801 5.9 (0.7)

Other smokers 602 7.7 (0.3) 30 0.3 (0.1)

SE, standard error. Education level was defined as: Low, no education completed or first level (primary school); Middle, lower or upper secondary levels; and High, tertiary level, which included university and other forms of education after the secondary level.

(4)

the reference category) and Figure 2 (among smokers, with light smokers as the reference category).

In men, ex-smokers had higher ORs of being overweight or obese with respect to non-smokers (Table 3). The OR of cigarette smokers (all categories confounded) was lower for being overweight but not for being obese. The OR for being obese increased progressively with number of cigarettes daily smoked (Figure 2). Furthermore, heavy smokers had higher ORs for being obese than non-smokers. For those who were underweight, ORs tended to be higher in cigarette smokers, with an OR of 1.8 [95% confidence interval (CI):

0.8 to 2.9], and to increase with the number of cigarettes daily smoked (OR⫽ 1.0, 95% CI ⫽ 0.3 to 3.3; OR ⫽ 1.2, 95% CI⫽ 0.4 to 3.2; and OR ⫽ 2.0, 95% CI ⫽ 1.0 to 4.0;

respectively, in light, moderate, and heavy smokers).

In women, ex-smokers had an increased OR for being obese (Table 4). Similar to men, albeit to a lower extent, the OR for being obese increased with number of cigarettes smoked daily (Figure 2). For being underweight, the OR was higher in female cigarette smokers compared with non-smokers (OR ⫽ 1.4, 95% CI ⫽ 1.1 to 1.8) and increased with the number of cigarettes smoked daily (OR⫽ 1.1, 95% CI⫽ 0.8 to 1.6; OR ⫽ 1.4, 95% CI ⫽ 1.0 to 1.9;

and OR⫽ 1.8, 95% CI ⫽ 1.3 to 2.4; respectively, in light, moderate, and heavy smokers).

Discussion

In this large sample of the general adult population, we found an association between smoking and BMI; although Table 2. Mean BMI according to sex, age, and smoking status

Age (years)

Men Women

N BMI (kg/m²) (SE) N BMI (kg/m²) (SE)

25 to 44

Non-smokers 1561 24.6 (0.1) 2035 22.7 (0.1)

Ex-smokers 538 25.2 (0.2) 603 22.8 (0.2)

Light 213 24.1 (0.2) 322 22.1 (0.2)

Moderate 307 24.6 (0.2) 375 22.6 (0.2)

Heavy 542 24.9 (0.2) 394 22.4 (0.2)

Other smokers 176 25.2 (0.3) 11 20.8 (0.6)

All 3337 24.7 (0.1) 3740 22.6 (0.1)

45 to 64

Non-smokers 930 25.5 (0.1) 1846 24.3 (0.1)

Ex-smokers 928 26.5 (0.1) 759 24.4 (0.2)

Light 122 25.1 (0.3) 250 23.4 (0.3)

Moderate 143 25.1 (0.4) 273 23.1 (0.2)

Heavy 378 25.6 (0.4) 343 24.0 (0.2)

Other smokers 282 27.0 (0.2) 12 24.3 (1.2)

All 2783 26.0 (0.1) 3483 24.1 (0.1)

⬎64

Non-smokers 642 25.9 (0.1) 1846 25.0 (0.1)

Ex-smokers 770 26.3 (0.1) 398 25.5 (0.2)

Light 48 25.2 (0.7) 84 24.5 (0.5)

Moderate 54 24.7 (0.5) 96 23.0 (0.4)

Heavy 66 26.1 (0.4) 64 25.5 (0.6)

Other smokers 144 26.7 (0.4) 7 22.9 (1.0)

All 1724 26.1 (0.1) 2495 25.0 (0.1)

SE, standard error.

(5)

mean BMI tended to be lower among current smokers, the risk for being obese increased with the number of cigarettes smoked daily, irrespective of some confounding factors.

This trend was more pronounced in men than in women. On the other hand, daily cigarette consumption was associated with being underweight, particularly in women.

Our data are consistent with results of other reports. In Danish men 20 to 29 years of age, moderate smokers had a decreased risk of obesity compared with non-smokers (adjusted OR ⫽ 0.4), but heavy smokers tended to have increased risk (OR⫽ 1.3) (27). In the Greek EPIC cohort, in which information was gathered by examination and not by self-report, Bamia et al. (9) showed in a sample of appar- ently healthy 25- to 84-year-old men and women that heavy smokers had higher mean BMI compared with light smok-

ers; as in this study, this association was stronger in men than in women. The association persisted after adjustment for variables reflecting the levels of energy expenditure and intake, which were assessed with a validated questionnaire.

The strength of this study is the large size of this sample of the general adult population. This allowed analyzing associations between body weight categories with various categories of smoking habits, stratified by sex and age.

Moreover, potential confounding factors could be taken into account in our analysis, in particular, educational level, which strongly related to both overweight/obesity and smoking habits in this study, as well as in several others (28,29). It was recently reported that the proportion of U.S.

adults who both smoked and were obese was higher in lower socioeconomic groups (29).

A limitation of this study is that body height and weight were self-reported. Height tends to be overestimated, whereas weight tends to be underestimated, particularly by women (30,31). This leads to an underestimation of BMI, a bias that increases with age (30,32). In addition, it was shown that if BMI is calculated on self-reported measures of height and weight, the association between socioeconomic status and underweight or obesity could be underestimated (33). To our knowledge, bias for reporting height and weight does not differ according to smoking habits. How- ever, it is possible that obese smokers were more prone to underreport the number of cigarettes daily smoked and to underreport their body weight than non-obese smokers. This would lead to an underestimation of the association between smoking and obesity. Another limitation is that a rather rudimentary questionnaire was used to evaluate diet and physical activity (21), likely to induce a non-differential misclassification. The association of overweight/obesity Figure 1: Prevalence of obesity (BMIⱖ 30 kg/m²) according to

smoking status in men and women. Non-smk, non-smokers; Ex- smk, ex-smokers; Cig-L, light cigarette smokers; Cig-M, moderate cigarette smokers; Cig-H, heavy cigarette smokers.

Table 3. Odds ratio for the association of smoking status with overweight and obesity vs. normal-weight in men

Overweight Obese

Non-smoker 1* 1*

Ex-smoker 1.3 (1.2 to 1.5) 1.9 (1.5 to 2.3) Cigarette smoker 0.8 (0.7 to 0.9) 1.0 (0.8 to 1.2) Light 0.8 (0.6 to 1.0) 0.5 (0.3 to 0.8) Moderate 0.7 (0.6 to 0.9) 0.7 (0.4 to 1.0) Heavy 0.9 (0.8 to 1.0) 1.3 (1.0 to 1.7) Other smoker 1.5 (1.2 to 1.8) 2.7 (2.0 to 3.7)

Values are odds ratio (adjusted for age, nationality, educational level, physical activity during leisure time, and fruit/vegetable intake) (95% confidence interval).

* Reference category.

Table 4. Odds ratio for the association of smoking status with overweight and obesity vs. normal-weight in women

Overweight Obese

Non-smoker 1* 1*

Ex-smoker 0.9 (0.8 to 1.0) 1.3 (1.1 to 1.6) Cigarette smoker 0.8 (0.7 to 0.9) 0.8 (0.7 to 1.0) Light 0.6 (0.5 to 0.8) 0.7 (0.5 to 1.0) Moderate 0.7 (0.6 to 0.9) 0.8 (0.5 to 1.0) Heavy 0.9 (0.7 to 1.1) 1.1 (0.8 to 1.4) Other smoker 0.3 (0.1 to 1.3) 1.1 (0.2 to 4.6)

Values are odds ratio (adjusted for age, nationality, educational level, physical activity during leisure time, and fruit/vegetable intake) (95% confidence interval).

* Reference category.

(6)

with smoking habits remained significant after adjusting for risk behaviors favoring weight gain. This could relate to the fact that we used rough estimates for diet and physical activity, leaving space for residual confounding.

However, using precise assessment of energy balance in a cross-sectional study, Bamia et al. (9) found a similar association between daily cigarette consumption and mean BMI.

One third of eligible subjects did not participate. Some were not attainable despite repeated attempts (up to 20 times) to reach them. Most refused because of lack of interest and/or lack of time. This selection might distort prevalence estimates for smoking and obesity. Subjects declining to participate in surveys are more often smokers (34,35) and more often obese (35). Consequently, any association between smoking and obesity is likely to be underestimated if subjects with both conditions refused to participate much more frequently than subjects with one or none of the conditions.

The survey followed a two-stage sampling frame: a direct random sample of households, followed by a direct random sample of one person among those living in the household (not only those present in the households at the time of call).

If the sampled person declined participation, no other person of the household was selected. Selection mechanisms are likely to occur with this procedure (availability at home, size of the household, readiness to answer, etc.). For example, the number of sampled women is larger than expected, probably because they were more easily attainable at home and that they were more willing to participate. In general, it is difficult to consider that a massive systematic error is distorting the results, although this limits the generalizibility of our findings to the whole Swiss population.

We previously reported that ex-smokers had similar patterns of health behaviors compared with non-smokers or light smokers (14). In line with other studies (36,37), this suggests that smoking cessation is associated with positive change in other heath risk behaviors. However, despite these favorable changes, this study showed a higher mean BMI and a higher prevalence of obesity in ex-smokers compared with non-smokers. This is consistent with a dou- ble effect of smoking, i.e., an increased energy expenditure and a decrease of appetite, both being lost in the case of smoking cessation (4). Moreover, because smoking is caus- ally involved in several major chronic diseases conducive to weight loss, this could explain why, in this study, heavy smokers tended to have increased odds of being underweight compared with light smokers.

As the number of cigarettes smoked daily increases, our findings suggest that the weight-stabilizing effect of smoking might be counterbalanced by the clustering with other risk behaviors favoring weight gain, e.g., unhealthy diet and low physical activity (14). These risk behaviors and the related mechanisms are likely to interact throughout the life course and were only superficially captured in this cross- sectional study. Therefore, it could not be expected that adjusting for diet and physical activity would withdraw the association between smoking and obesity. Moreover, the increase in energy expenditure induced by smoking might depend on body weight: the post-smoking increase in resting energy expenditure has been found to be greater in normal weight women compared with obese women (38), which limits the potential weight control effect of smoking among the obese. In addition, repeated attempts to quit smoking and relapses may be more frequent in heavy smokers compared with light smokers; this could induce weight cycling and result in weight increase. Finally, the detrimen- tal effect of smoking on fat distribution should also be considered. Actually, smoking is associated with a central- ization of fat deposition (39), an increased insulin resistance (40), and a higher risk for diabetes (41,42).

In conclusion, our results indicate that cigarette smoking was associated with BMI in a dose-dependent manner, Figure 2: Adjusted OR (95% CI) for the association of number of

cigarettes smoked daily (light: 1 to 9 cigarettes/d; moderate: 10 to 19 cigarettes/d; heavy: ⱖ20 cigarettes/d) with overweight and obesity (vs. normal weight) among smokers (light smokers is the reference category). ORs are adjusted for age, nationality, educational level, physical activity during leisure time, and fruit/vegetable intake.

(7)

resulting in increased prevalence for obesity in heavy smokers, particularly in men. In addition to helping smokers who quit smoking to limit their weight gain, more emphasis should be put on the risk of obesity among smokers.

Acknowledgments

The Swiss Health Survey is funded by the Swiss Federal Office of Statistics. This analysis received no additional funding.

References

1. Mokdad AH, Marks JS, Stroup DF, Gerberding JL. Actual causes of death in the United States, 2000. JAMA. 2004;291:

1238 – 45.

2. World Health Organization. The World Health Report:

2002: Reducing Risks, Promoting Healthy Life. Geneva, Swit- zerland: World Health Organization.

3. Williamson DF, Madans J, Anda RF, Kleinman JC, Giovino GA, Byers T. Smoking cessation and severity of weight gain in a national cohort. N Engl J Med. 1991;324:

739 – 45.

4. Filozof C, Fernandez Pinilla MC, Fernandez-Cruz A.

Smoking cessation and weight gain. Obes Rev. 2004;5:95–

103.

5. Trawdella D, Loew M, Rothenbacher D, Stegmaier C, Ziegler H, Brenner H. The impact of body weight on smok- ing cessation in German adults. Prev Med. 2006;42:109 –13.

6. Hofstetter A, Schutz Y, Jequier E, Wahren J. Increased 24-hour energy expenditure in cigarette smokers. N Engl J Med. 1986;314:79 – 82.

7. Collins LC, Cornelius MF, Vogel RL, Walker JF, Stam- ford BA. Effect of caffeine and/or cigarette smoking on resting energy expenditure. Int J Obes Relat Metab Disord.

1994;18:551– 6.

8. Flegal KM, Troiano RP, Pamuk ER, Kuczmarski RJ, Campbell SM. The influence of smoking cessation on the prevalence of overweight in the United States. N Engl J Med.

1995;333:1165–70.

9. Bamia C, Trichopoulou A, Lenas D, Trichopoulos D. To- bacco smoking in relation to body fat mass and distribution in a general population sample. Int J Obes Relat Metab Disord.

2004;28:1091– 6.

10. Albanes D, Jones DY, Micozzi MS, Mattson ME. Associ- ations between smoking and body weight in the US popula- tion: analysis of NHANES II. Am J Public Health. 1987;77:

439 – 44.

11. Shimokata H, Muller DC, Andres R. Studies in the distri- bution of body fat. III. Effects of cigarette smoking. JAMA.

1989;261:1169 –73.

12. Austin SB, Gortmaker SL. Dieting and smoking initiation in early adolescent girls and boys: a prospective study. Am J Public Health. 2001;19:446 –50.

13. Laaksonen M, Pra¨tta¨la¨ R, Karisto A. Patterns of unhealthy behaviour in Finland. Eur J Public Health. 2001;11:294 –300.

14. Chiolero A, Wietlisbach V, Ruffieux C, Paccaud F, Cornuz J. Clustering of risk behaviors with cigarette consumption: a population-based survey. Prev Med. 2006;42:348 –53.

15. Blitzer PH, Rimm AA, Giefer EE. The effect of cessation of smoking on body weight in 57,032 women: cross-sectional and longitudinal analyses J Chron Dis. 1977;30:415–29.

16. Gordon T, Kannel WB, Dawber TR, McGee D. Changes associated with quitting smoking: the Framingham Study. Am Heart J. 1975;90:322– 8.

17. John U, Hanke M, Rumpf HJ, Thyrian JR. Smoking status, cigarettes per day, and their relationship to overweight and obesity among former and current smokers in a national adult general population sample. Int J Obes Relat Metab Disord.

2005;29:1289 –94.

18. Ali SM, Lindstrom M. Socioeconomic, psychosocial, behav- ioural, and psychological determinants of BMI among young women: differing patterns for underweight and overweight/

obesity. Eur J Public Health. 2006;16:324 –30.

19. Cooper TV, Klesges RC, Robinson LA, Zbikowski SM. A prospective evaluation of the relationships between smoking dosage and body mass index in an adolescent, biracial cohort.

Addict Behav. 2003;28:501–12.

20. Rehm J, Spuhler T. Measurement error in alcohol consump- tion: the Swiss Health Survey. Eur J Clin Nutr. 1993;47(Suppl 2):S25–30.

21. Eichholzer M, Bisig B. Daily consumption for (red) meat or meat products in Switzerland: results of the 1992/93 Swiss health Survey. Eur J Clin Nutr. 2000;54:136 – 42.

22. Organization for Economic Co-operation and Develop- ment. Education in OECD countries 1987– 89. In: A Compen- dium of Statistical Information. Washington, DC: OECD Pub- lications and Information Center; 1990, pp. 1–149.

23. Cavelaars AE, Kunst AE, Geurts JJ, et al. Educational differences in smoking: international comparison. BMJ. 2000;

320; 1102–7.

24. International Obesity Task Force. Available at http://www.

iotf.org/ (Accessed May 18, 2006).

25. U.S. Department of Health and Human Services. Physical Activity and Health: A Report of the Surgeon General. At- lanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention; 1996.

26. Kleinbaum DG, Klein M. Logistic Regression. A Self-Learn- ing Text. 2nd ed. New York: Springer Verlag; 2002.

27. Nielsen TL, Wraae K, Brixen K, Hermann AP, Andersen M, Hagen C. Prevalence of overweight, obesity and physical inactivity in 20- to 29-year-old, Danish men. Relation to sociodemography, physical dysfunction and low socioeco- nomic status: the Odense Androgen Study. Int J Obes Relat Metab Disord. 2006;30:805–15.

28. Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity, diabetes and obesity-related health risk factors 2001.

JAMA. 2003;289:76 –9.

29. Healton CG, Vallone D, McCausland KL, Xiao H, Green MP. Smoking, obesity, and their co-occurrence in the United States: cross sectional analysis. BMJ. 2006;333:25– 6.

30. Niedhammer I, Bugel I, Bonenfant S, Goldberg M, Leclerc A.

Validity of self-reported weight and height in the French GAZEL cohort. Int J Obes Relat Metab Disord. 2000;24:1111– 8.

31. Paccaud F, Wietlisbach V, Rickenbach M. Body mass in- dex: comparing mean values and prevalence rates from tele- phone and examination surveys. Rev Epidemiol Sante Pub- lique. 2001;49:33– 40.

(8)

32. Kuczmarski MF, Kuczmarski RJ, Najjar M. Effects of age on validity of self-reported height, weight, and body mass index: findings from the Third National Health and Nutrition Examination Survey, 1988 –1994. J Am Diet Assoc. 2001;101:

28 –34.

33. Bostrom G, Diderichsen F. Socioeconomic differentials in misclassification of height, weight and body mass index based on questionnaire data. Int J Epidemiol. 1997;26:860 – 6.

34. Jackson R, Chambless LE, Yang K, et al. Differences between respondents and nonrespondents in a multicenter community-based study vary by gender ethnicity. The Ath- erosclerosis Risk in Communities (ARIC) Study Investigators.

J Clin Epidemiol. 1996;49:1441– 6.

35. Van Loon AJ, Tijhuis M, Picavet HS, Surtees PG, Ormel J.

Survey non-response in the Netherlands: effects on prevalence estimates and associations. Ann Epidemiol. 2003;13:105–10.

36. Morabia A, Curtin F, Bernstein MS. Effects of smoking and smoking cessation on dietary habits of a Swiss urban popula- tion. Eur J Clin Nutr. 1999;53:239 – 43.

37. Laaksonen M, Luoto R, Helakorpi S, Uutela A. Associa- tions between health-related behaviors: a 7-year follow-up of adults. Prev Med. 2002;34:162–70.

38. Audrain JE, Klesges RC, Klesges LM. Relationship be- tween obesity and the metabolic effects of smoking in women.

Health Psychol. 1995;14:116 –23.

39. Canoy D, Wareham N, Luben R, et al. Cigarette smoking and fat distribution in 21,828 British men and women: a population-based study. Obes Res. 2005;13:1466 –75.

40. Facchini FS, Hollenbeck CB, Jeppesen J, Chen YD, Reaven GM. Insulin resistance and cigarette smoking. Lan- cet. 1992;339:1128 –30.

41. Wannamethee SG, Shaper AG, Perry IJ. British Re- gional Heart Study. Smoking as a modifiable risk factor for type 2 diabetes in middle-aged men. Diabetes Care. 2001;

24:1590 –5.

42. Hu FB, Manson JE, Stampfer MJ, et al. Diet, lifestyle, and the risk of type 2 diabetes mellitus in women. N Engl J Med.

2001;345:790 –7.