Dose-dependent positive association between cigarette smoking, abdominal obesity and body fat: cross-sectional data from a population-based survey

We analyzed the baseline data from the CoLaus study, a cross-sectional, population-based study of 6,123 participants. The details of the CoLaus study have been previously described in detail [18, 19]. Briefly, the CoLaus Study was designed to investigate the prevalence and genetic determinants of risk factors for cardiovascular disease. The survey started in 2003 and was approved by the Institutional Ethics Committee of the University of Lausanne. All subjects between 35 and 75 years of age living in Lausanne (Switzerland) were identified from a city register, and a random sample of 19,830 subjects (35% of the overall population) was invited to participate by mail. Inclusion criteria included providing written informed consent, being 35-75 years of age, and of Caucasian origin. The last inclusion criterion was chosen for the genetic arm and analysis of the study. Of the initial 19,830 subjects sampled, 54 subjects were considered as non-eligible before contact and 15,109 (76%) responses were obtained. A total of 4,667 subjects did not respond. Among responders, 6,189 (41%) subjects refused to participate in the study and 799 (5%) were considered as non-eligible. The sample of 8,121 subjects who agreed to participate represented 41% of the initially sampled population, 54% of all responders and 57% of all eligible responders. As there were more eligible participants than requested for the initial study, the last 1,383 subjects were not included in the study; a further 549 non-Caucasians were also excluded, and one participant initially included withdrawn from the study. Further, for our analysis, we included only participants with complete data for the main variables of interest. We excluded an extra 65 participants because of missing values for BMI (n = 1), WC (n = 1), BF (n = 57), smoking status (n = 2), education (n = 5) and marital status (n = 2) (Figure 1). Subjects interested in participating were then contacted by telephone and sent the first questionnaire by mail, which recorded information on demographic data, socioeconomic status, and several lifestyle factors, namely tobacco and alcohol consumption. All eligible participants were then asked to attend an outpatient clinic in the morning after an overnight fast, and data were collected by trained field interviewers during a single visit. The first questionnaire mailed with the appointment letter and completed by the participant prior the visit was reviewed. A second questionnaire, focused on personal and family history of disease as well as cardio-vascular risk factors and treatment, was given during the interview. Trained nurses measured body height, weight, waist and hip circumferences and BF during this visit.

Data on smoking included the subject's previous and current smoking status as well as the amount of tobacco smoked and age at which the subject began smoking, and, if they were former smokers, when they stopped smoking. Participants were categorized as non-smokers if they had never smoked, former smokers if they had quit smoking at the time of the interview, as current smokers if they were currently smoking ≥1 cigarette per day, and as "other smokers" if they were currently smoking pipe, cigars, or cigarillos. Based on current smoking categories, current smokers were divided into 3 predefined categories according to daily consumption; light (1 to 9 cigarettes/day), moderate (10 to 19 cigarettes/day), and heavy smokers (≥20 cigarettes/day). For light smokers, most recent studies used a cutoff of 10 cigarettes/day [20, 21]. We considered people smoking ≥ 20 cigarettes/day as heavy smokers because this corresponds to the quantity of cigarettes contained in a standard pack in Western countries and other studies have also used this cutoff [21, 22]. Data on alcohol consumption was collected, including past and current drinking habits as well as the amount of alcoholic beverage units consumed the week prior to the interview. At-risk alcohol consumption was defined as an intake of more than 14 drinks/week for men <65 years of age, and more than 7 drinks/week for women all ages or men ≥65 years of age [23]. Body weight and height were measured with the participants standing barefooted in light indoor clothing. Body weight was measured in kilograms to the nearest 100 g. Height was measured to the nearest 5 mm. BMI, defined as weight/height², was calculated and subjects were classified as underweight (BMI <18.5 kg/m²), normal (BMI ≥18.5 and <25 kg/m²), overweight (BMI ≥25 and <30 kg/m²), or obese (BMI ≥30 kg/m²) according to WHO criteria [24]. Waist and hip circumferences were measured with a non-stretchable tape. The waist was measured over the abdomen but under the clothing at a level midway between the lower rib margin and the iliac crest, and the measurement was rounded to the nearest centimetre. Two measurements were taken and the mean was used for analyses. Abdominal obesity was defined according to the literature (Table 1) [25]. BF and fat-free mass were assessed by electrical bioimpedance using the Bodystat^® 1500 body mass analyser (Bodystat Ltd, Isle of Man, England) [26]. Subjects had to fast for at least 8 hours, not engage in strenuous physical activity during the previous 12 hours, and abstain from consuming caffeine or alcohol-containing beverages for 24 hours before the examination. All metallic objects were removed from the body and clothing, and the measurement was performed after a 10-minute rest in the supine position. BF was expressed as a percentage of total body weight. Excess BF was defined if BF mass (expressed in percent) was superior or equal to the 95^th percentile for the Swiss population according to Kyle et al [27] (Table 1).

The baseline characteristics of the 6,123 participants (3,211 women and 2,912 men) are shown in Table 2. About a third of the men and a quarter of the women were current smokers. Men were more frequently obese or overweight than women, while the prevalence of abdominal obesity and excess BF was higher in women. Smoking and obesity were more frequent among less educated participants (data not shown).

Mean age-adjusted WC, BF, and BMI were calculated for non-smokers, current and former smokers. For current smokers, data were stratified according to number of cigarettes smoked per day (Figure 2, 3, 4). Non-smokers had a lower WC compared with former smokers but a higher WC compared with current smokers (Figure 2). Non-smokers had higher BF levels compared with smokers and about the same BF levels compared with former smokers. (Figure 3). Non-smokers had a higher BMI than current smokers and a lower BMI compared with former smokers (Figure 4). Multi-adjusted analyses (adjusted for alcohol consumption and educational level) were similar to the age-adjusted results (data not shown).

Among current smokers, gradients of higher WC, BF, and BMI levels with increasing numbers of cigarettes smoked per day were found. The trends were stronger for WC and BF than for BMI (Figure 2, 3 4). The results were similar in analyses using pack-years instead of number of cigarettes smoked per day (data not shown).

Among current female smokers, heavy smoking was associated with increased odds of having abdominal obesity and excess BF compared with light smoking (Table 3). The results of the age-adjusted and multi-adjusted analyses were similar. Interestingly, we observed a trend such that the odds in either sex of having abdominal obesity increased with the number of cigarettes smoked daily. After adjustment for age, education level, and alcohol consumption this trend was statistically significant for abdominal obesity in both genders (P = 0.03 in men and P < 0.01 in women). Among men, we found that moderate and heavy smoking were associated with increased odds of having a medium WC compared with light smoking. No significant association between the number of cigarettes smoked daily and obesity (as defined as BMI) was found.