http://orcid.org/0000-0003-3450-4189
Figures
Abstract
The analyses of geographic variations in the prevalence of major chronic conditions, such as overweight and obesity, are an important public health tool to identify “hot spots” and inform allocation of funding for policy and health promotion campaigns, yet rarely performed. Here we aimed at exploring, for the first time in Luxembourg, potential geographic patterns in overweight/obesity prevalence in the country, adjusted for several demographic, socioeconomic, behavioural and health status characteristics. Data came from 720 men and 764 women, 25–64 years old, who participated in the European Health Examination Survey in Luxembourg (2013–2015). To investigate the geographical variation, geo-additive semi-parametric mixed model and Bayesian modelisations based on Markov Chain Monte Carlo techniques for inference were performed. Large disparities in the prevalence of overweight and obesity were found between municipalities, with the highest rates of obesity found in 3 municipalities located in the South-West of the country. Bayesian approach also underlined a nonlinear effect of age on overweight and obesity in both genders (significant in men) and highlighted the following risk factors: 1. country of birth for overweight in men born in a non-European country (Posterior Odds Ratio (POR): 3.24 [1.61–8.69]) and women born in Portugal (POR: 2.44 [1.25–4.43]), 2. low educational level (secondary or below) for overweight (POR: 1.66 (1.06–2.72)] and obesity (POR:2.09 [1.05–3.65]) in men, 3. single marital status for obesity in women (POR: 2.20 [1.24–3.91]), 4.fair (men: POR: 3.19 [1.58–6.79], women: POR: 2.24 [1.33–3.73]) to very bad health perception (men: POR: 15.01 [2.16–98.09]) for obesity, 5. sleeping more than 6 hours for obesity in unemployed men (POR: 3.66 [2.02–8.03]). Protective factors highlighted were: 1. single marital status against overweight (POR: [0.60 (0.38–0.96)]) and obesity (POR: 0.39 [0.16–0.84]) in men, 2. the fact to be widowed against overweight in women (POR: [0.30 (0.07–0.86)], as well as a non European country of birth (POR: 0.49 [0.19–0.98]), tertiary level of education (POR: 0.34 [0.18–0.64]), moderate alcohol consumption (POR: 0.54 [0.36–0.90]) and aerobic physical activity practice (POR: 0.44 [0.27–0.77]) against obesity in women. A double burden of environmental exposure due to historic mining and industrial activities and past economic vulnaribility in the South-West of the country may have participated to the higher prevalence of obesity found in this region. Other demographic, socioeconomic, behavioural and health status covariates could have been involved as well.
Citation: Samouda H, Ruiz-Castell M, Bocquet V, Kuemmerle A, Chioti A, Dadoun F, et al. (2018) Geographical variation of overweight, obesity and related risk factors: Findings from the European Health Examination Survey in Luxembourg, 2013-2015. PLoS ONE 13(6): e0197021. https://doi.org/10.1371/journal.pone.0197021
Editor: C. Mary Schooling, CUNY, UNITED STATES
Received: April 24, 2017; Accepted: April 25, 2018; Published: June 14, 2018
Copyright: © 2018 Samouda et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: There are both legal and ethical restrictions on sharing our data publicly. The informed consent signed by the participants to the study stipulates that the collected data will be exclusively analysed by the Luxembourgish Public Research Centre for Health and the European coordination centre of the survey (EHES). The Public Research Centre for Health (CRP-Santé) is the former name of our institution, currently Luxembourg Institute of Health (LIH). Researchers and/or institution interested in analyzing our data can contact the Principal Investigator of EHES-LUX2013-2015, Dr Maria Ruiz-Castell (Maria.Ruiz@lih.lu) and fill out a request form. Data requests may also be sent to Dr Laetitia Huiart, Head of the Population Health Department, LIH (Laetitia.Huiart@lih.lu). For researchers requesting access to this data, there would not be any barriers to access beyond standard ethical approval and data access protocols for confidential data.
Funding: This work was supported by the Luxembourg Institute of Health, Luxembourg; Ministry of Culture, Higher Education and Research, Luxembourg; and the Ministry of Health, Luxembourg. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Overweight and obesity are regarded as a public health priority, both because of their pandemic progression with more than 2 billions people impacted worldwide, namely about 28% of the world population, and owing to the potentially relatedburden of diseases [1–3].
The latest systematic analysis performed in the frame of the the Global Burden of Disease Study estimated the prevalence of obesity around 23.7% for men and 26% for women in Luxembourg. Overweight prevalence estimation was about 34.3% in men and 18.4% in women (data collected before 2013) [1].
Obesity is a complex disease, with a multiple etiology, including foetal programmation, genetic predisposition, physiological mechanisms, psychological factors, food production and behavior, physical activity pattern, environmental and social context [4–6].
Demographic, behavioural and socioeconomic risk factors such as increased age, lower education level and employment, unbalanced dietary habits and/or physical inactivity have previously been highlighted as correlates of obesity in Luxembourg and worldwide [7–11].
Overweight and obesity have also been linked to environmental risk factors such as the urban area of residence, in spite of existing controversies especially between urban and rural areas [12–16]. A north-south decreasing gradient of obesity has also been underlined in several countries [17]. In Luxembourg, Tchikaya et al. found a higher likelihood of obesity amongst people living outside the capital, in particular in the North and the West areas of the country [11]. Westernisation has also been suggested as an additional driver of the obesity increase in several countries [18, 19]. Recent evidence has also suggested the influence of factors such as environmental chemicals on body weight disorders [20]. Nevertheless, individual intrinsic and environmental extrinsic risk factors tend to be separately considered in the literature, in spite of their close interconnection [7–13, 17–19].
Geographic variation analyses address the challenge to combine prevalence of overweight and obesity with both individual and environmental risk factors through an integrative spatial approach. Geographic variation analyses are mainly based on Bayesian statistical approaches and, therefore, also investigate both observed and random data [21–23]. Through the geographic variation analyses based on Bayesian approaches, certain authors called into question some concepts previously underlined by the logistic statistical approaches usually used, such as the north-south decreasing gradient of obesity [17, 21]. Analyses of geographic variations in the prevalence of major chronic conditions, such as overweight/obesity, are an important public health tool to identify “hot spots”, which would be of a very high concern especially in Luxembourg. Indeed, in spite of its small surface area of approximately 2 586 km2, Luxembourg shows a complex territory organization with no less than 106 administrative municipalities, which, despite their organization in 3 main districts (Diekirch, Grevenmacher and Luxembourg), remain differently influenced by the geographic location. Several municipalities could indeed be influenced by more than one neighbouring country. Some municipalities located in the district of Diekirch could be influenced by both Belgian (North-West side) and German (North-East side) genetic predisposition, environment and behaviours. Others located in the disctrict of Luxembourg could be impacted by both Belgian and French neighbourhood, respectively situated on the West- and South-side. Municipalities located in Grevenmacher district could be influenced by German (Centre-East side) and/or French specificities (South-Est side).
The aim of the present paper was to investigate the geographical variation of overweight and obesity risk in Luxembourg, adjusted for several demographic, socioeconomic, behavioural and health status characteristics.
We used the most recent population-based data, from the European Health Examination Survey, conducted in Luxembourg between 2013 and 2015 [24, 25].
Materials and methods
European Health Examination Survey in Luxembourg (EHES-LUX 2013–2015)
Data came from EHES-LUX2013-2015 conducted amongst the 25–64 years old noninstitutionalized resident population of Luxembourg, from February 2013 to January 2015 [24–26]. 1529 participants were invited to take part to the survey following a random selection (One-stage sampling procedure from the national population register managed by the “Inspection générale de la sécurité sociale” in Luxembourg, with age-, gender- and district of residence stratification) [26, 27].
For the purpose of this analysis, data of 24 participants having a BMI <18.5 kg/m2 and 21 pregnant women were excluded [28]. We analysed the data of 1484 subjects, men (N = 720) and women (N = 764). Participants gave written informed consent (S1 File: Consent Form). National Ethics of Research and Data Protection Committees authorized the study.
The European Health Examination Survey (S2 File: EHES Questionnaire) includes: a questionnaire (on demographic, socioeconomic, behavioural, geographic and health status items) and a clinical examination (including weight and height measurement).
Overweight and obesity definition
In the clinical examination part, weight and height were measured by trained nurses, based on a standardised protocol developed in the framework of EHES-LUX [25].
Body Mass Index (BMI) was calculated as individual’s weight divided by the square of the height (kg/m2). Overweight (BMI of 25–29.9 kg/m2) and obesity (BMI ≥ 30 kg/m2) prevalence were calculated, based on the population of the 25–64 years-old residents in Luxembourg in 2014 [29].
Demographic, socioeconomic and geographic determinants
We analysed demographic, socioeconomic, behavioural, geographic and health status data from Ehes-Lux2013-2015: age and gender, administrative municipality of residence, administrative district area of residence (North: Diekirch. Centre: Luxembourg. East: Grevenmacher), country of birth (according to the political division, Luxembourg and Portugal being the two most prevalent birth places for Luxembourg residents), marital status (single: “never married nor in civil partnership”, “married and/or in partnership”, “divorced and/or in dissolved civil partnership”, “widowed and/or surviving to partner death”), employment status (“employed” carrying out a job or profession and “unemployed”) and the household total net monthly income [24–26].
Behavioural aspects
Data related to the daily consumption of fruits and vegetables were analysed (number of portions and frequency of consumption: “once or more a day” or “less than once a day” [25]). We assessed physical activity, based on 4 indicators, according to EHIS-PAQ questionnaire: 1.WRPA (work-related physical activity): “percentage of people mostly physically active when working” [30]; 2.TRPA (transport-related physical activity): “metabolic equivalent per minute (MET/minute) and per week of walking and cycling to reach the workplace, expressed in quintiles” [30]; 3.HEPA (Health-enhancing physical activity), which includes: -APA (Aerobic PA guideline compliance): “percentage of people performing at least 150 min aerobic physical activity per week” [30] and -MSPA (muscle strengthening physical activity): “percentage of people practicing a muscle strengthening physical activity at least 2 days per week” [30].
To evaluate the weekly standard alcoholic drink consumption, we developed a 3-category indicator based on the product of “the weekly frequency of alcoholic drink” by “the number of standard drinks per day defined as 30 centilitres (cl) of beer, 15cl of champagne, 15cl of wine, 4cl of spirits, 8cl of aperitif cocktails, 4cl of shots and/or 1 tall glass of spirits or soft drink”:—Category 1: ‘No drinker or seldom’ (< once a week consumption).—Category 2: ‘Low drinker’ (1 to 6 drinks a week).—Category 3: ‘High drinker’ (> 6 drinks a week). As both variables “weekly frequency” and “daily number of standard drinks” contained ranges as choices, the median values were used, 6 drinks being the median value of the alcoholic weekly consumption [24, 25].
Self-perceived health, physical and mental health condition
Self-perceived health was evaluated through the question “how is your health in general?”, with answers separated in 3 categories: “Very good to good”, “fair” and/or “bad to very bad” health. People were also asked to document “the intensity of physical pain they had suffered during the past 4 weeks”: “from low intensity to no pain”, “moderate pain” or “severe or very severe pain” [25]. Employed participants were enquired about their “usual sleep duration per night for nights preceding a work day” [24, 25]. For unemployed participants, the following question was asked: “how many hours do you sleep normally at night when you do not work the next day?”. We created a 4-categories indicator gathering sleep duration and labour status:—Employed participants with sleep duration ≤ 6h;—Employed participants with sleep duration > 6h;—Unemployed participants with sleep duration ≤ 6h;—Unemployed participants with sleep duration > 6h. Presence of symptoms of depression was defined by PHQ-9 (Patient Health Questionnaire) score ≥ 5, based on the daily frequency to have “little interest to do things”, “depressive or hopeless feelings”, “hypersomnia or insomnia”, “tiredness”, “eating disorders”, “failure feeling”, “concentration problems”, “slow speaking and moving or fidgeting” and/or “dead or self-harming ideas” [25, 31].
The one hundred six municipalities of residence were used to investigate geographic distribution of overweight and obesity.
Statistical analyses
Descriptive analyses were expressed in percentages, tertiles, means, standard deviations, minimum and/or maximum values. Chi-square (χ2) and Mann–Whitney U-tests were performed to assess the statistical significance of overweight/obesity prevalence and risk factors. To estimate the likelihood of overweight or obesity, we used logistic regression analyses where continuous variables were centred. Selection of variables was based on a literature review and on statistical criteria (variables showing p< 0.20 in the univariate logistic analyses were entered in the multivariable logistic models). Odds Ratio (OR) were assessed. Multicollinearity between explanatory variables for BMI was checked. Interactions were verified. To account for the stratified random sampling method used to recruit the participants, weighted statistical methods were applied to produce nationally representative estimates. A sampling weight equal to the inverse probability of unit selection was allocated to each participant from the same stratum. This stratum sampling weight was defined as the ratio between the population and the observed sample stratum sizes (SAS version 9.4, SAS Institute Inc., Cary, NC, USA).
Moreover, we investigated the geographical variation of overweight/obesity prevalence, adjusted for demographic, socioeconomic, behavioural and health status characteristics at the municipality level and in comparison with normal weight. To take into account the potential nonlinear effects of overweight/obesity risk factors, we used a geo-additive semi-parametric mixed model and a Bayesian approach based on the Markov Chain Monte Carlo (MCMC) techniques for inference and model checking [21, 26, 27, 32]. This approach investigate both observed and random data [21, 23]. The response variable was defined as yi = 1 if overweight or obesity and yi = 0 for normal weight. The standard measure of effect was the posterior odds ratio (POR) and 95% credible region [21, 26, 32]. Multivariate Bayesian geo-additive modelisation was performed in order to assess POR significance calculated for the fixed, nonlinear and spatial effects. Results were considered to be significant at the 5% critical level (CR) (p<0.05). Bayesian analyses took into account the effect of the neighbouring municipalities (also called communes in Luxembourg) in the models developed. POR were represented on the map of Luxembourg, for each municipality. Red coloured areas represented the municipalities at high risk of overweight/obesity, while green coloured areas illustrated those at low risk. The significance of the spatial effect was reported on a second map: black coloured areas illustrated significant positive spatial effect at the municipalities level. White coloured areas represented significant negative spatial effect and grey coloured areas concerned no significant effect. Bayes X software package, version 2.0.1, was used in order to perform the Bayesian analyses.
Results
Prevalence of overweight and obesity between 2013 and 2015 in Luxembourg
High rates of overweight (37.3%) and obesity (20.6%) were observed in Ehes-Lux2013-2015, which, once extended to the general adult population residing in Luxembourg between 2013 and 2015 [29], would represent a prevalence of overweight of approximately 38%, and a prevalence of obesity of about 20,2% (Table 1).
According to EHES-LUX2013-2015 Survey.
General characteristics are detailed in Table 2.
Demographic, socioeconomic, behavioural and health status correlates
Univariate analyses are detailed in S1 and S2 Tables.
Overweight.
Table 3 in men and Table 4 in women detail the OR (logistic regressions) and POR (Bayesian modelisation) of overweight and obesity according to their respective covariates.
Women born in Portugal (OR: 1.98 [1.08–3.61]) and men born in a non European country (OR: 2.19 [1.15–4.15]) and perceiving their health as bad or very bad (OR: 4.99 [1.31–19.10]) showed a higher risk to develop overweight. These associations were confirmed through the Bayesian analysis (POR, Women: 2.44 [1.25–4.43]; POR, Men: 3.24 [1.61–8.69] and 4.45 [1.02–36.90], respectively). There were also significant relationships, in men, with age (S1 Fig), the fact to have never been married nor in civil partnership (Protective effect, POR: [0.60 (0.38–0.96)] and with a low educational level (secondary or below) [Deleterious effect: POR: 1.66 (1.06–2.72)], and, in women, with the fact to be widowed (Protective effect, POR: [0.30 (0.07–0.86)] (Table 3, Table 4).
Obesity.
In men, obesity seems to increase with age (OR: 1.04 [1.01–1.07]), a fair (OR: 2.35 [1.12–4.92]) bad or very bad self-perceived health (OR: 9.74 [2.04–46.55]), sleeping less than 6 h for employed people (OR: 2.70 [1.51–4.83]) and decrease for men who have never been married nor have been in civil partnership (OR: 0.45 [0.23–0.87]).
In men, Bayesian analyses showed that obesity increases with age (S2 Fig), a fair (POR: 3.19 [1.58–6.79], respectively bad or very bad (POR: 15.01 [2.16–98.09]) self-perceived health, sleeping more than 6 hours for unemployed men (POR: 3.66 [2.02–8.03]) and a low educational level (secondary or below) (POR:2.09 [1.05–3.65]); and decreases amongst single men (POR: 0.39 [0.16–0.84]).
In women, both multivariable logistic and Bayesian models showed that obesity decreases with higher education (OR: 0.33 [0.16–0.68], POR: 0.34 [0.18–0.64]), more than 150 minutes of physical activity practice per week (OR: 0.50 [0.29–0.84], POR: 0.44 [0.27–0.77]) and low alcohol consumption (6 drinks or less a week, OR: 0.58 [0.34–0.99], POR: 0.54 [0.36–0.90]), and increases with poor self-perceived health (OR: 2.10 [1.23–3.60], POR: 2.24 [1.33–3.73]). In women, Bayesian analyses showed a protective effect against obesity of a Non European country of birth (POR: 0.49 [0.19–0.98]) and a negative effect of being never been married nor in civil partnership (POR: 2.20 [1.24–3.91]) (Table 4).
Geographic variation of overweight and obesity between 2013 and 2015 in Luxembourg
Total residual spatial effects for overweight and obesity risk by communes/municipalities adjusted for demographic, socioeconomic, behavioural and health covariates and geographical location, are shown in Figs 1–4.
Left: Adjusted total residual spatial effects for men’s overweight risk, at municipalities-level in Luxembourg in 2013–2015. Shown are the posterior odds ratios. Right: Corresponding posterior probabilities at 80% nominal level (EHES, 2013–2015). Red coloured–high risk. Green coloured–low risk. Black coloured–significant positive spatial effect. White coloured- significant negative spatial effect. Grey coloured–no significant effect.
Left: Adjusted total residual spatial effects for men’s obesity risk, at municipalities-level in Luxembourg in 2013–2015. Shown are the posterior odds ratios. Right: Corresponding posterior probabilities at 80% nominal level (EHES, 2013–2015). Red coloured–high risk. Green coloured–low risk. Black coloured–significant positive spatial effect. White coloured- significant negative spatial effect. Grey coloured–no significant effect.
Left: Adjusted total residual spatial effects for women’s overweight risk, at municipalities-level in Luxembourg in 2013–2015. Shown are the posterior odds ratios. Right: Corresponding posterior probabilities at 80% nominal level (EHES, 2013–2015). Red coloured–high risk. Green coloured–low risk. Black coloured–significant positive spatial effect. White coloured- significant negative spatial effect. Grey coloured–no significant effect.
Left: Adjusted total residual spatial effects for women’s obesity risk, at municipalities-level in Luxembourg in 2013–2015. Shown are the posterior odds ratios. Right: Corresponding posterior probabilities at 80% nominal level (EHES, 2013–2015). Red coloured–high risk. Green coloured–low risk. Black coloured–significant positive spatial effect. White coloured- significant negative spatial effect. Grey coloured–no significant effect.
Higher but not statistically significant posterior odds ratio (POR) of overweight were observed in the South-West for men and in the North- and South-East for women (Figs 1 and 3).
Regarding obesity, the highest and significant POR was located in the South-West for both men (municipalities of Petange and Steinfort) and women (municipality of Differdange) at 80% credible region (CR) (Figs 2 and 4).
Discussion
The present manuscript represents the first attempt to characterise geographical variation of overweight and obesity in Luxembourg, adjusted for several demographic, socioeconomic, behavioural and health status characteristics, using the most recent data from a nationwide representative sample, with respect to the municipality of residence. In Luxembourg, geographic variation of cardiovascular risk factors has previously been investigated, using data from 2007, and showing in particular a higher overweight/obesity prevalence in some North-East and Centre-East areas [33]. However, the Bayesian approach used by the Authors were only adjusted for gender and age, did neither took into account detailed demographic, socioeconomic, behavioural and health status characteristics, nor distinguished between overweight and obesity which are not especially linked with the same potential health issues [33].
To our knowledge, EHES-LUX2013-2015 is the first survey exploring these aspects, also at the European level.
In the present study, the differences observed between logistic (Odd Ratios) and posterior models (Posterior Odd Ratios) showed that as we hypothesised, there is a spatial effect that is not captured by conventional logistic regression models.
In recent years, there has been increasing attention to overweight and obesity mapping adjusted for demographic and socioeconomic characterisics in Africa [21].
Spatial analysis of overweight/obesity performed by Kandala et al. [21] at the state-level in Nigeria underlined a highest prevalence of overweight/obesity amongst women in 5 states in the south-eastern areas, as well as in one eastern and one northern states, rather than an exclusive decreasing north-south gradient as showed in some previous studies [17]. These findings were explained by the ongoing urbanisation, westernisation and lifestyle transition process towards more processed high-calorie food and sedentary habits in these states [21].
In our study, after adjustment for demographic, socioeconomic, behavioural and health characteristics, we observed significant geographic variation in the likelihood of developing obesity in both men and women in the South-West of the country, which is in line with the previous work from the same study by Ruiz-Castell et al.[26], which focused on hypertension. The authors suggested that the high rates of hypertension in this area may actually be due to the “historical mining and industrial activities, followed by the 1970s’ steel industry crisis which has critically affected the region” [26, 34]. We do agree with Ruiz-Castell et al.[26] on the fact that the inhabitants of that geographic area may have been exposed to the double burden of environmental pollutant exposure due to mining and industrialisation and economic vulnerability, both exposures being described as major risk factors for overweight and obesity [35, 36].
Moreover, additional demographic, socioeconomic, behavioural and health-related factors might explain the observed geographic patterns of obesity risk in the present study.
Despite its limited size, Luxembourg shows a very heterogeneous population given the large percentage of immigrants. Two types of immigration exist in the country. There are the highly qualified foreign professionals who live in socioeconomically favourable areas, in particular in the capital, Luxembourg City. There are also less qualified immigrants, living in socioeconomically disadvantaged peripheral regions. The former mining region located in the South-West of Luxembourg, the peripheral region highlighted in the present study, gathers the most economically vulnerable citizens of the country [37].
Socio-cultural factors could also have been involved. Specifically, the process of acculturation related to immigration could play a role in the observed geographic patterns amongst the less qualified immigrants [38–40]. Immigration from low and/or middle towards higher income areas seems to fuel overweight and obesity insofar, as through acculturation, migrant groups may adopt “obesogenic lifestyle habits” in the host country [38]. The process may be the result of several factors such as a rapid epidemiological transition of overnutrition, low physical activity, along with low economic status for migrants and/or easy access to low-quality and high-calorie food [38, 41]. Acculturation process may differ according to gender, which may explain the fact that, in the present survey, men born in Portugal were not impacted [38, 41].
Our findings are in line with the literature regarding the negative impact of a poor self-perceived health on weight gain. Poor health perception has previously been found to be associated with higher rates of cardiometabolic conditions, earlier mortality and, more recently with overweight and obesity, a relationship which seems to be mediated by aging, lack of physical activity, smoking and/or low socioeconomic status [42].
Regarding aging, similar results were observed by Tchicaya et al. in 2007 amongst 45–64 years old men in Luxembourg, as well as amongst 45–54 and more than 65 years-old women [11]. Aging is a major risk factor of overweight/obesity, owing to numerous physiological factors such as the menopausal process in women, and the decline of physical activity, muscle mass and metabolic rate in older people [43].
Moreover, the likelihood to develop obesity appeared to be attenuated amongst people living alone. Possible explanations include weight stigmatization in sentimental relationships, also called “the selection hypothesis” and/or “being in the marriage market concept”. The idea behind is that in the Western culture, to get married, single people have to be attractive, therefore slender. Marriage related weight-stigma is mostly experienced by women in the litterature. Our data confirmed this finding in men, but not in women. This obligation disappears after marriage and/or is replaced with “the social obligation hypothesis” advocating a harmonious family life where food is wealthy and caloric [44–46].
Several studies also showed that individuals with lower levels of education are at higher risk of obesity especially in developed countries, which is in line with our findings in women. In middle and high income countries, wealthy women who reached a high level of education show a higher probability to have normal weight [47], while wealth reverses the protective effect of education amongst women living in low income countries [48]. This is possibly linked to a better health education, greater opportunities to access to healthy and non-obesogenic environments, and/or less stressful lifestyle than in socioeconomically deprived areas. Our findings are in line with these studies performed in other countries, and also with the work of Tchicaya et al. conducted in 2007 in Luxembourg [11].
Other investigations have pointed out physical activity and low-calorie food consumption as protective factors against weight gain [49]. Unfortunately, people in Westernized countries are regularly exposed to two concomitant phenomena: a plethora of processed foods, available at all times and everywhere, and a sedentary lifestyle resulting of large-scale urbanisation, a quasi-exclusive motorized transportation, less field sports and more screen time. Physical activity is becoming mostly practised as a hobby, under a voluntary approach, with often a busy life leaving little time to be physically active. Eating behaviour has evolved towards a reduction of fruits, vegetables, water and complex carbohydrates consumption, in favour of high fat, sugar-based and high calorie processed food [50, 51]. In the present work, only aerobic physical activity of at least 150 min/week, seems to play a protective role against obesity in women. Inverse relationship between aerobic physical activity (bicycling, dancing, jogging, walking, swimming) and obesity has been previously highlighted in the literature [52, 53]. After adjustment for other risk factors, fruit and vegetable consumption does not seem to impact BMI in our population. Conversely, alcohol consumption restricted to no more than 6 drinks a week appears to attenuate the likelihood to develop obesity in women. In this respect, light to moderate alcohol drinking has previously been shown to reduce the likelihood of obesity [54]. This protective effect seems to be mediated by isohumulones and/or polyphenols found in beer and wine, which may decrease body fat absorption, adipocyte size and weight gain [54].
Short sleep duration was associated with increased obesity in our study, in the subpopulation of employed men. Sleeping less than 6 hours per night has previously been linked to an increased likelihood of obesity [55], and led to recommend a longer sleep duration in recently published guidelines [56].In contrast, we found that unemployed men sleeping more than 6 hours a night appear to be at increased risk of obesity. This is in line with the literature and could be caused by reduced energy expenditure during longer sleeping periods [55].
In contrast with our findings, Alkerwi et al. [33] showed an increased overweight/obesity prevalence in the North-East and the Centre-East of the Grand-duché of Luxembourg. These findings were not in line neither with our findings nor with Ruiz-Castell et al. regarding hypertension [26]. This may be explained by the fact that Bayesian models used in Alkerwi et al. study [33] were only adjusted for gender and age and did not distinguished between overweight and obesity, while our analysis took into account overweight, obesity as well as several demographic, socioeconomic, behavioural and health characteristics. The Authors also performed a secondary analysis of data collected in 2007 and the situation (e.g. demographic, socioeconomic changes, migration) might have changed the last eight years [33].
Finally, this paper provides a comprehensive update of overweight and obesity prevalence amongst adults in Luxembourg, using an objective and direct assessment of BMI. Overweight and obesity impact the vast majority of the population of Luxembougr aged 25–64 (2013–2015) (67.8% in men and 47.7% in women), namely about 182413 people, if extrapolated to the entire population [29]. This places Luxembourg amongst the countries with the highest rates in Europe and worldwide, and implies that the United States of America no longer have such exclusiveness [1]. Our data suggest a much higher prevalence than that published for Luxembourg by the Global Burden of Disease (GBD): i.e. 58% in men and 44.4% in women [1]. This may be caused by methodological differences, and may be related with the fact that BMI data provided by the GBD group were obtained by self-assessment/declaration, which may lead to an under-estimation of body weight [57].
Similar results were found by Tchicaya et al. [11] in a survey on “household income and living conditions” in Luxembourg, who also used self-reported BMI. As compared to Tchicaya et al. data [11], we found higher prevalences for both overweight (46.77% vs 43.9%) and obesity (21.05% vs 17.9%) in men, and for obesity in women (19.3% vs 17.7%) [11]. The latter discrepancy may also be linked to historical trends over time of overweight and obesity prevalence in Luxembourg (2007 versus 2015). Conversely, obesity rates observed in the present Ehes-Lux2013-2015 (20.6%) were close to those observed in the Oriscav-Lux survey (20.9%) performed between 2007 and 2009 in a representative sample of adults living in Luxembourg, which also used a direct assessment of BMI [58].
Conclusion
The analyses of geographic variations in the prevalence of major chronic conditions, such as overweight and obesity, are an important public health tool to identify “hot spots” and inform allocation of funding for policy and health promotion campaigns. Here we aimed at exploring, for the first time in Luxembourg, potential geographic patterns in overweight/obesity prevalence in the country, adjusted for several demographic, socioeconomic, behavioural and health status characteristics.
Large disparities in the prevalence of overweight and obesity exist between municipalities, with the highest rates of obesity intriguingly found in 3 municipalities situated in the South-West of the country. We hypothesize that a double burden of environmental exposure due to historic mining and industrial activities and past economic vulnaribility may have participated to the higher prevalence of obesity found in these areas, as previously suggested regarding hypertension burden [26]. Other demographic, socioeconomic, behavioural and health covariates could have been involved as well: age, immigration, bad health self-perception, single life, low level of education, low physical activity practice and high alcohol consumption.
EHES-LUX2013-2015 showed also high overweight and obesity prevalence, placing Luxembourg amongst the countries with the highest rates. Our findings contribute to the development of context-specific public health policies to tackle the prevalence, correlates and geographic variation of overweight and obesity in Luxembourg.
Supporting information
S1 Table. Univariate logistic predictive models of overweight and obesity in men, in relation to normal weight (EHES-LUX2013-2015).
WRPA: Work-related physical activity. TRPA: Transport-related physical activity. APA: Aerobic physical activity. MSPA: Muscle- strengthening physical activity.
https://doi.org/10.1371/journal.pone.0197021.s001
(DOCX)
S2 Table. Univariate logistic predictive models of overweight and obesity in women, in relation to normal weight (EHES-LUX2013-2015).
WRPA: Work-related physical activity. TRPA: Transport-related physical activity. APA: Aerobic physical activity. MSPA: Muscle- strengthening physical activity.
https://doi.org/10.1371/journal.pone.0197021.s002
(DOCX)
S1 Fig.
Estimated nonparametric trend of women’s and men’s overweight risk by women’s age cohort (left) and men’s age cohort (right) in Luxembourg. Shown is the posterior mean within 80% credible regions [EHES, 2013–2015].
https://doi.org/10.1371/journal.pone.0197021.s003
(TIF)
S2 Fig.
Estimated nonparametric trend of women’s and men’s obesity risk by women’s age cohort (left) and men’s age cohort (right) in Luxembourg. Shown is the posterior mean within 80% credible regions [EHES, 2013–2015].
https://doi.org/10.1371/journal.pone.0197021.s004
(TIF)
References
- 1. Ng M, Fleming T, Robinson M, Thomson B, Graetz N, Margono C, et al. Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2014;384(9945):766–81. pmid:24880830.
- 2.
Worldometers. Population mondiale actuelle 2016 [13th June 2016]. Available from: http://www.worldometers.info/fr/population-mondiale/.
- 3. G. B. D. Risk Factors Collaborators, Forouzanfar MH, Alexander L, Anderson HR, Bachman VF, Biryukov S, et al. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015. pmid:26364544.
- 4.
Conscienhealth. 57 Varieties of Obesity: Moving Past One Size Fits All. Adapted from Kaplan L., presentation at the 29th Blackburn Course in Obesity Medicine, June 2, 2016. [16th June 2016]. Available from: http://conscienhealth.org/2016/06/57-varieties-obesity-moving-past-one-size-fits/.
- 5.
The Obesity Society Infographic Task Force. Potential Contributors to Obesity 2015 [15th June 2016]. Available from: http://www.obesity.org/obesity/resources/facts-about-obesity/infographics/potential-contributors-to-obesity.
- 6. Stryjecki C, Alyass A, Meyre D. Ethnic and population differences in the genetic predisposition to human obesity. Obes Rev. 2017. pmid:29024387.
- 7. Van Dyck D, Cerin E, De Bourdeaudhuij I, Hinckson E, Reis RS, Davey R, et al. International study of objectively measured physical activity and sedentary time with body mass index and obesity: IPEN adult study. Int J Obes (Lond). 2015;39(2):199–207. pmid:24984753; PubMed Central PMCID: PMC4282619.
- 8. Rouhani MH, Haghighatdoost F, Surkan PJ, Azadbakht L. Associations between dietary energy density and obesity: A systematic review and meta-analysis of observational studies. Nutrition. 2016;32(10):1037–47. pmid:27238958.
- 9. Devaux M, Sassi F, Church J, Cecchini M, Borgonovi F. Exploring the Relationship Between Education and Obesity. OECD Journal: Economic Studies. 2011;2011(1).
- 10. Bonauto DK, Lu D, Fan ZJ. Obesity prevalence by occupation in Washington State, Behavioral Risk Factor Surveillance System. Preventing chronic disease. 2014;11:130219. pmid:24406093; PubMed Central PMCID: PMC3887052.
- 11. Tchicaya A, Lorentz N. Socioeconomic inequality and obesity prevalence trends in Luxembourg, 1995–2007. BMC Res Notes. 2012;5:467. pmid:22931792; PubMed Central PMCID: PMC3494539.
- 12. Atek M, Traissac P, El Ati J, Laid Y, Aounallah-Skhiri H, Eymard-Duvernay S, et al. Obesity and association with area of residence, gender and socio-economic factors in Algerian and Tunisian adults. PLoS One. 2013;8(10):e75640. pmid:24116063; PubMed Central PMCID: PMC3792975.
- 13. Patterson KA, Gall SL, Venn AJ, Otahal P, Blizzard L, Dwyer T, et al. Accumulated exposure to rural areas of residence over the life course is associated with overweight and obesity in adulthood: a 25-year prospective cohort study. Annals of epidemiology. 2017;27(3):169–75 e2. pmid:28317611.
- 14. Amugsi DA, Dimbuene ZT, Mberu B, Muthuri S, Ezeh AC. Prevalence and time trends in overweight and obesity among urban women: an analysis of demographic and health surveys data from 24 African countries, 1991–2014. BMJ Open. 2017;7(10):e017344. pmid:29079606; PubMed Central PMCID: PMC5665233.
- 15. Wen M, Fan JX, Kowaleski-Jones L, Wan N. Rural-Urban Disparities in Obesity Prevalence Among Working Age Adults in the United States: Exploring the Mechanisms. American journal of health promotion: AJHP. 2017:890117116689488. pmid:29214811.
- 16. Rundle AG, Heymsfield SB. Can Walkable Urban Design Play a Role in Reducing the Incidence of Obesity-Related Conditions? JAMA. 2016;315(20):2175–7. pmid:27218628.
- 17. Scarborough P, Allender S. The North-South gap in overweight and obesity in England. Br J Nutr. 2008;100(3):677–84. pmid:18257942.
- 18. Wang Y, Wang L, Xue H, Qu W. A Review of the Growth of the Fast Food Industry in China and Its Potential Impact on Obesity. International journal of environmental research and public health. 2016;13(11). pmid:27834887; PubMed Central PMCID: PMCPMC5129322.
- 19. Gomez-Arbelaez D, Camacho PA, Cohen DD, Rincon-Romero K, Alvarado-Jurado L, Pinzon S, et al. Higher household income and the availability of electronic devices and transport at home are associated with higher waist circumference in Colombian children: the ACFIES study. International journal of environmental research and public health. 2014;11(2):1834–43. pmid:24514426; PubMed Central PMCID: PMCPMC3945571.
- 20. Tang-Peronard JL, Andersen HR, Jensen TK, Heitmann BL. Endocrine-disrupting chemicals and obesity development in humans: a review. Obes Rev. 2011;12(8):622–36. pmid:21457182.
- 21. Kandala NB, Stranges S. Geographic variation of overweight and obesity among women in Nigeria: a case for nutritional transition in sub-Saharan Africa. PLoS One. 2014;9(6):e101103. pmid:24979753; PubMed Central PMCID: PMC4076212.
- 22. Clements AC, Lwambo NJ, Blair L, Nyandindi U, Kaatano G, Kinung'hi S, et al. Bayesian spatial analysis and disease mapping: tools to enhance planning and implementation of a schistosomiasis control programme in Tanzania. Tropical medicine & international health: TM & IH. 2006;11(4):490–503. pmid:16553932; PubMed Central PMCID: PMC2202922.
- 23. Best N, Richardson S, Thomson A. A comparison of Bayesian spatial models for disease mapping. Statistical methods in medical research. 2005;14(1):35–59. pmid:15690999.
- 24.
European Health Examination Survey. [cited 2016 19.02.]. Available from: http://www.ehes.info/.
- 25.
Tolonen H (Ed.). EHES Manual. Part B. Fieldwork procedures. Helsinki, Finland: National Institute for Health and Welfare, 2013 78-952-245-843-8 Contract No.: 19.02.2016.
- 26. Ruiz-Castell M, Kandala NB, Kuemmerle A, Schritz A, Barre J, Delagardelle C, et al. Hypertension burden in Luxembourg: Individual risk factors and geographic variations, 2013 to 2015 European Health Examination Survey. Medicine. 2016;95(36):e4758. pmid:27603374; PubMed Central PMCID: PMCPMC5023897.
- 27. Ruiz-Castell M, Kandala NB, Perquin M, Bocquet V, Kuemmerle A, Vogele C, et al. Depression burden in luxembourg: Individual risk factors, geographic variations and the role of migration, 2013–2015 European Health Examination Survey. Journal of affective disorders. 2017;222:41–8. Epub 2017/07/03. pmid:28668715.
- 28. Samouda H, Ruiz-Castell M, Karimi M, Bocquet V, Kuemmerle A, Chioti A, et al. Metabolically healthy and unhealthy weight statuses, health issues and related costs: Findings from the 2013–2015 European Health Examination Survey in Luxembourg. Diabetes Metab. 2017. Epub 2017/12/19. pmid:29249613.
- 29.
Le Portail des Statistiques. Grand-Duché de Luxembourg. Population par âge et sexe au 1er janvier 2001–2016 Documentation du tableau. 2014 [cited 2016 07 June]. Available from: http://www.statistiques.public.lu/stat/TableViewer/tableView.aspx.
- 30. Finger JD, Tafforeau J, Gisle L, Oja L, Ziese T, Thelen J, et al. Development of the European Health Interview Survey—Physical Activity Questionnaire (EHIS-PAQ) to monitor physical activity in the European Union. Archives of public health = Archives belges de sante publique. 2015;73:59. pmid:26634120; PubMed Central PMCID: PMCPMC4667448.
- 31. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. Journal of general internal medicine. 2001;16(9):606–13. pmid:11556941; PubMed Central PMCID: PMC1495268.
- 32. Fahrmeir L, Lang S. Bayesian Inference for Generalized Additive Mixed Models Based on Markov Random Field Priors. Applied Statistics (JRSSC). 2010;50:201–20.
- 33. Alkerwi A, Bahi IE, Stranges S, Beissel J, Delagardelle C, Noppe S, et al. Geographic Variations in Cardiometabolic Risk Factors in Luxembourg. International journal of environmental research and public health. 2017;14(6). pmid:28621751; PubMed Central PMCID: PMCPMC5486334.
- 34.
Oumeziane AA. Reconversion économique et construction d'un territoire transfrontalier: l'Agglomération Transfrontalière du Pôle Européen de Développement des Trois Frontières (Belgique-France-Luxembourg)//Economic reconversion and construction of a transfrontier territory: The Agglomération Transfrontalière du Pôle Européen de Développement des Trois Frontières (Belgium-France-Luxembourg) Annales de Géographie [Internet]. 2000 21.03.2017; 109(611):[65–83 pp.]. Available from: http://www.persee.fr/doc/geo_0003-4010_2000_num_109_611_1904.
- 35. Grun F, Blumberg B. Endocrine disrupters as obesogens. Mol Cell Endocrinol. 2009;304(1–2):19–29. pmid:19433244; PubMed Central PMCID: PMCPMC2713042.
- 36. Zhang Q, Wang Y. Socioeconomic inequality of obesity in the United States: do gender, age, and ethnicity matter? Social science & medicine. 2004;58(6):1171–80. PubMed pmid:14723911.
- 37. Lord S, Gerber P. IMMIGRATION, DYNAMIQUES SOCIO-ÉCONOMIQUES TERRITORIALES ET MOUVEMENTS RÉSIDENTIELS. QUELLES PERSPECTIVES POUR LES RÉSIDENTS DU LUXEMBOURG? Annales de géographie. 2013;2 (n° 690):175 à 99.
- 38. Delavari M, Sonderlund AL, Swinburn B, Mellor D, Renzaho A. Acculturation and obesity among migrant populations in high income countries—a systematic review. BMC Public Health. 2013;13:458. pmid:23663279; PubMed Central PMCID: PMC3654930.
- 39. Toselli S, Gualdi-Russo E, Boulos DN, Anwar WA, Lakhoua C, Jaouadi I, et al. Prevalence of overweight and obesity in adults from North Africa. Eur J Public Health. 2014;24 Suppl 1:31–9. pmid:25107996.
- 40. Maqoud F, Vacca E, Tommaseo-Ponzetta M. From Morocco to Italy: How Women's Bodies Reflect their Change of Residence. Coll Antropol. 2016;40(1):9–15. PubMed pmid:27301231.
- 41. Chang SJ, Im EO. Testing a Theoretical Model of Immigration Transition and Physical Activity. Res Theory Nurs Pract. 2015;29(3):177–88. PubMed pmid:26502554.
- 42. Shields M, Shooshtari S. Determinants of self-perceived health. Health reports. 2001;13(1):35–52. PubMed pmid:15069807.
- 43.
Mayo Clinic Staff. Obesity risk factors [cited 2017 03.01.]. Available from: http://www.mayoclinic.org/diseases-conditions/obesity/basics/risk-factors/con-20014834.
- 44. Kark M, Karnehed N. Weight status at age 18 influences marriage prospects. A population-based study of Swedish men. BMC Public Health. 2012;12:833. pmid:23020864; PubMed Central PMCID: PMCPMC3505734.
- 45. Boyes AD, Latner JD. Weight stigma in existing romantic relationships. J Sex Marital Ther. 2009;35(4):282–93. pmid:19466667.
- 46. Sobal J, Rauschenbach BS, Frongillo EA Jr. Marital status, fatness and obesity. Social science & medicine. 1992;35(7):915–23. PubMed pmid:1411692.
- 47. Aitsi-Selmi A, Chandola T, Friel S, Nouraei R, Shipley MJ, Marmot MG. Interaction between education and household wealth on the risk of obesity in women in Egypt. PLoS One. 2012;7(6):e39507. pmid:22761807; PubMed Central PMCID: PMCPMC3384649.
- 48. Aitsi-Selmi A, Bell R, Shipley MJ, Marmot MG. Education modifies the association of wealth with obesity in women in middle-income but not low-income countries: an interaction study using seven national datasets, 2005–2010. PLoS One. 2014;9(3):e90403. pmid:24608086; PubMed Central PMCID: PMCPMC3946446.
- 49. Gast J, Campbell Nielson A, Hunt A, Leiker JJ. Intuitive eating: associations with physical activity motivation and BMI. American journal of health promotion: AJHP. 2015;29(3):e91–9. pmid:24459999.
- 50. Adams J, White M. Characterisation of UK diets according to degree of food processing and associations with socio-demographics and obesity: cross-sectional analysis of UK National Diet and Nutrition Survey (2008–12). Int J Behav Nutr Phys Act. 2015;12:160. pmid:26684833; PubMed Central PMCID: PMCPMC4683717.
- 51. Burgoine T, Forouhi NG, Griffin SJ, Brage S, Wareham NJ, Monsivais P. Does neighborhood fast-food outlet exposure amplify inequalities in diet and obesity? A cross-sectional study. The American journal of clinical nutrition. 2016;103(6):1540–7. pmid:27169835; PubMed Central PMCID: PMCPMC4880999.
- 52. Mangeri F, Montesi L, Forlani G, Dalle Grave R, Marchesini G. A standard ballroom and Latin dance program to improve fitness and adherence to physical activity in individuals with type 2 diabetes and in obesity. Diabetol Metab Syndr. 2014;6:74. pmid:25045404; PubMed Central PMCID: PMCPMC4082296.
- 53. Brien SE, Katzmarzyk PT, Craig CL, Gauvin L. Physical activity, cardiorespiratory fitness and body mass index as predictors of substantial weight gain and obesity: the Canadian physical activity longitudinal study. Canadian journal of public health = Revue canadienne de sante publique. 2007;98(2):121–4. PubMed pmid:17441535.
- 54. Monteiro R, Soares R, Guerreiro S, Pestana D, Calhau C, Azevedo I. Red wine increases adipose tissue aromatase expression and regulates body weight and adipocyte size. Nutrition. 2009;25(6):699–705. pmid:19268535.
- 55. Lin CL, Lin CP, Chen SW, Wu HC, Tsai YH. The association between sleep duration and overweight or obesity in Taiwanese adults: A cross-sectional study. Obesity research & clinical practice. 2016. pmid:27520850.
- 56. H M., Whiton K, Albert SM, Alessi C, Bruni O, DonCarlos L, et al. National Sleep Foundation’s sleep time duration recommendations: methodology and results summary. Sleep Health. 2015;1:40–3. pmid:29073412
- 57. Connor Gorber S, Tremblay M, Moher D, Gorber B. A comparison of direct vs. self-report measures for assessing height, weight and body mass index: a systematic review. Obes Rev. 2007;8(4):307–26. pmid:17578381.
- 58. Alkerwi A, Sauvageot N, Donneau AF, Lair ML, Couffignal S, Beissel J, et al. First nationwide survey on cardiovascular risk factors in Grand-Duchy of Luxembourg (ORISCAV-LUX). BMC Public Health. 2010;10:468. pmid:20698957; PubMed Central PMCID: PMC2925827.