Caroline Méjean

The Nutrinet-Santé Study: a web-based prospective study on the relationship between nutrition and health and determinants of dietary patterns and nutritional status

BackgroundNutrition-related chronic diseases such as cardiovascular diseases and cancer are of multiple origin, and may be due to genetic, biologic, behavioural and environmental factors. In order to detangle the specific role of nutritional factors, very large population sample cohort studies comprising precisely measured dietary intake and all necessary information for accurately assessing potential confounding factors are needed. Widespread use of internet is an opportunity to gradually collect huge amounts of data from a large sample of volunteers that can be automatically verified and processed. The objectives of the NutriNet-Santé study are: 1) to investigate the relationship between nutrition (nutrients, foods, dietary patterns, physical activity), mortality and health outcomes; and 2) to examine the determinants of dietary patterns and nutritional status (sociological, economic, cultural, biological, cognitive, perceptions, preferences, etc.), using a web-based approach.Methods/designOur web-based prospective cohort study is being conducted for a scheduled follow-up of 10 years. Using a dedicated web site, recruitment will be carried out for 5 years so as to register 500 000 volunteers aged ≥ 18 years among whom 60% are expected to be included (having complete baseline data) and followed-up for at least 5 years for 240 000 participants. Questionnaires administered via internet at baseline and each year thereafter will assess socio-demographic and lifestyle characteristics, anthropometry, health status, physical activity and diet. Surveillance of health events will be implemented via questionnaires on hospitalisation and use of medication, and linkage with a national database on vital statistics. Biochemical samples and clinical examination will be collected in a subsample of volunteers.DiscussionSelf-administered data collection using internet as a complement to collection of biological data will enable identifying nutrition-related risks and protective factors, thereby more clearly elucidating determinants of nutritional status and their interactions. These are necessary steps for further refining nutritional recommendations aimed at improving the health status of populations.

Sociodemographic Factors and Attitudes toward Food Affordability and Health Are Associated with Fruit and Vegetable Consumption in a Low-Income French Population

Determinants of fruit and vegetable consumption, including affordability and attitudes, have been poorly investigated, especially in European deprived populations. Our objective was to analyze various determinants of low consumption of fruits and vegetables in disadvantaged participants. Our participants were randomized into 2 groups, 1 which received nutritional advice alone and 1 that also received vouchers that were exchangeable for fruits and vegetables during a 12-mo period. Socioeconomic characteristics, food insufficiency, affordability, and motivation for eating fruits and vegetables were assessed. A short FFQ was administered. Determinants of consumption of <1 fruit or vegetable/d were analyzed using multivariate logistic regression. A total of 295 participants were included (mean age 44.8 y; 133 men, 162 women). At baseline, mean daily consumption of fruits and vegetables was 2.13 +/- 1.57 times/d. Nearly 30% of the sample did not eat fruits and vegetables every day. Determinants of low fruit and vegetable consumption were: age younger than 55 y, education level lower than tertiary, and absence of financial means for buying fruits and vegetables daily. Other determinants were affordability (whether fruits and vegetables are affordable, lack of money in preventing healthy diet) and attitudes (whether ones own diet is healthy, whether or not fruits and vegetables improve health, whether eating fruits and vegetables is a pleasure). Thus, determinants of inadequate consumption of fruits and vegetables in this deprived French population are numerous. The impact of financial difficulties is crucial, as is the perception of affordability of fruits and vegetables.

Profiles of organic food consumers in a large sample of French adults: results from the Nutrinet-Santé cohort study.

Background Lifestyle, dietary patterns and nutritional status of organic food consumers have rarely been described, while interest for a sustainable diet is markedly increasing. Methods Consumer attitude and frequency of use of 18 organic products were assessed in 54,311 adult participants in the Nutrinet-Santé cohort. Cluster analysis was performed to identify behaviors associated with organic product consumption. Socio-demographic characteristics, food consumption and nutrient intake across clusters are provided. Cross-sectional association with overweight/obesity was estimated using polytomous logistic regression. Results Five clusters were identified: 3 clusters of non-consumers whose reasons differed, occasional (OCOP, 51%) and regular (RCOP, 14%) organic product consumers. RCOP were more highly educated and physically active than other clusters. They also exhibited dietary patterns that included more plant foods and less sweet and alcoholic beverages, processed meat or milk. Their nutrient intake profiles (fatty acids, most minerals and vitamins, fibers) were healthier and they more closely adhered to dietary guidelines. In multivariate models (after accounting for confounders, including level of adherence to nutritional guidelines), compared to those not interested in organic products, RCOP participants showed a markedly lower probability of overweight (excluding obesity) (25≤body mass index<30) and obesity (body mass index ≥30): −36% and −62% in men and −42% and −48% in women, respectively (P<0.0001). OCOP participants (%) generally showed intermediate figures. Conclusions Regular consumers of organic products, a sizeable group in our sample, exhibit specific socio-demographic characteristics, and an overall healthy profile which should be accounted for in further studies analyzing organic food intake and health markers.

The contribution of diet and lifestyle to socioeconomic inequalities in cardiovascular morbidity and mortality.

BACKGROUND The role of differences in diet on the relationship between socioeconomic factors and cardiovascular diseases remains unclear. We studied the contribution of diet and other lifestyle factors to the explanation of socioeconomic inequalities in cardiovascular diseases. METHODS We prospectively examined the incidence of coronary heart disease (CHD) and stroke events amongst 33,106 adults of the EPIC-NL cohort. Education and employment status indicated socioeconomic status. We used Cox proportional models to estimate hazard ratios ((HR (95% confidence intervals)) for the association of socioeconomic factors with CHD and stroke and the contribution of diet and lifestyle. RESULTS During 12 years of follow-up, 1617 cases of CHD and 531 cases of stroke occurred. The risks of CHD and stroke were higher in lowest (HR=1.98 (1.67;2.35); HR=1.55 (1.15;2.10)) and lower (HR=1.50 (1.29;1.75); HR=1.42 (1.08;1.86)) educated groups than in the highest. Unemployed and retired subjects more often suffered from CHD (HR=1.37 (1.19;1.58); HR=1.20 (1.05;1.37), respectively), but not from stroke, than the employed. Diet and lifestyle, mainly smoking and alcohol, explained more than 70% of the educational differences in CHD and stroke and 65% of employment status variation in CHD. Diet explained more than other lifestyle factors of educational and employment status differences in CHD and stroke (36% to 67% vs. 9% to 27%). CONCLUSION The socioeconomic distribution of diet, smoking and alcohol consumption largely explained the inequalities in CHD and stroke in the Netherlands. These findings need to be considered when developing policies to reduce socioeconomic inequalities in cardiovascular diseases.

The Associations between Emotional Eating and Consumption of Energy-Dense Snack Foods Are Modified by Sex and Depressive Symptomatology

In recent years, emotional eating (EmE) has incited substantial research interest as an important psychologic determinant of food intake and overweight. However, little is known about factors that might modulate its relations with dietary habits. The objective of this study was to examine the association between EmE and consumption of energy-dense snack food and assess the 2-way interaction of EmE with sex and depressive symptoms. A total of 7378 men and 22,862 women from the NutriNet-Santé cohort (France, 2009-2013) who completed ≥6 self-reported 24-h food records were included in this cross-sectional analysis. EmE was evaluated via the revised 21-item Three-Factor Eating Questionnaire. Depressive symptoms were assessed by the Center for Epidemiologic Studies Depression Scale. The associations between EmE and energy-dense food consumption were assessed by multivariable logistic and linear regression models adjusted for sociodemographic and lifestyle factors. Higher EmE was associated with higher consumption of energy-dense snacks and, in particular, with consumption of sweet-and-fatty foods across most categories studied. However, these associations were stronger in women with depressive symptoms (e.g., high consumption of chocolate, OR: 1.77, 95% CI: 1.43, 2.20; cakes/biscuits/pastries, OR: 1.81, 95% CI: 1.45, 2.26) compared with those without depressive symptoms (e.g., high consumption of chocolate, OR: 1.52, 95% CI: 1.36, 1.69; cakes/biscuits/pastries, OR: 1.44, 95% CI: 1.29, 1.61). In contrast, the significant positive associations observed in men without depressive symptoms (e.g., high consumption of chocolate, OR: 1.33, 95% CI: 1.16, 1.52; cakes/biscuits/pastries, OR: 1.28, 95% CI: 1.11, 1.48) were not found in men with depressive symptoms. In conclusion, in women, EmE was positively associated with consumption of energy-dense snack food, particularly in those with depressive symptoms. For men, the relation between EmE and energy-dense snack foods was found only in those without depressive symptoms. These findings call for consideration of the psychologic state when targeting unhealthy dietary habits, especially in women. This trial was registered at eudract.ema.europa.eu as 2013-000929-31.

Sex and dieting modify the association between emotional eating and weight status

BACKGROUND Unlike other psychological correlates of weight status, emotional eating (EmE) has received relatively limited attention in the literature. OBJECTIVES We aimed to examine the association between the EmE score and weight status and the influence of sex and dieting on this association. DESIGN A total of 8580 men and 27,061 women aged ≥18 y who participated in the NutriNet-Santé cohort study were selected in this cross-sectional analysis. Self-reported weight and height, EmE scores of the revised 21-item version of the Three-Factor Eating Questionnaire, and dieting status data were collected. The relation between EmE and weight status was estimated by using multiple linear and logistic regression models adjusted for sociodemographic and lifestyle factors. Interactions of EmE with sex and dieting history were assessed. RESULTS Median EmE scores were greater in women than in men and in former or current dieters than in subjects without a history of dieting. Strong associations appeared between the EmE score and weight status in most categories of sex × dieting status. The strongest associations between EmE and weight status were observed in women, particularly in never dieters [body mass index slope (95% CI): 2.61 (2.43, 2.78); overweight OR (95% CI): 5.06 (4.24, 6.05)]. CONCLUSIONS These observations support the existence of an association between EmE and weight status. The effect modification of sex and dieting on the EmE score and on the association of the EmE score with weight status should be taken into account in obesity prevention.

Trends in the prevalence of obesity in employed adults in central-western France: a population-based study, 1995-2005.

OBJECTIVE The overall trend of obesity prevalence has increased during the last decades, even in France which has one of the lowest prevalence in Europe. The aim of this study was to assess, according to socioeconomic status (SES), whether a shift in the obesity prevalence trends could be observed since the French National Nutrition and Public Health Program was implemented in 2001. METHODS Standardised cross-sectional repeated population-based data from the French Social Security Health Examination Centers in the central-western region of France (n=339,882). We examined regression slopes (95% CI) of overall and abdominal obesity from 1995 to 2005 according to SES. We also compared slopes within each SES between 1995-2001 and 2001-2005. RESULTS After standardisation to the French age distribution, 6.9% of men and 6.4% of women were obese in 1995 and 8.9% and 8.6% in 2005, respectively. Abdominal adiposity concerned 5.6% of men and 8.5% of women in 1995 and 9.5% and 14.3% in 2005. Obesity prevalence regression slopes between 1995 and 2005 increased in all SES categories, except management professionals in both genders and office/service personnel male who were stable. Significant regression slopes before 2001 became non-significant afterward in office/service personnel males for obesity prevalence; and for abdominal obesity in manual workers women and office/service personnel (p=0.05, in men). CONCLUSIONS Our data confirm the overall epidemic rise in the obesity prevalence trends during the last decade, except in management professionals and office/service personnel men. Obesity prevalence trends in office/service personnel and manual workers women for abdominal obesity were also observed to stabilise since 2001. We could hypothesize that the National Nutrition and Public Health Program may partly be involved in this decreasing trend among office/service personnel mainly. Policymakers should take into account these data to assess the effectiveness of obesity prevention public-health strategies in the future.

Associations between weight status and liking scores for sweet, salt and fat according to the gender in adults (The Nutrinet-Santé study)

Background/objectives:As taste preferences may be associated with obesity, the present study investigated whether obese subjects presented heightened liking for the sensations of sweet, salt and fat.Subjects/Methods:Liking scores were determined by a questionnaire including 83 items on liking for sweet or fatty foods, and the preferred extent of seasoning with salt, sweet or fat. Data from 46 909 adults included in the French web-based observational cohort of the Nutrinet-Santé study were collected and weighted according to the national population census. Relationships between liking scores and body mass index (BMI) as categorical or linear explanatory variable were assessed separately by gender using covariance and linear regression analyses, adjusted for age, education level, living area, smoking and alcohol.Results:Overall liking scores for salt and fat were linearly positively linked to BMI in men and women (P⩽0.001) and were higher in obese than in normal-weight individuals. The score difference between BMI categories was greater in women for fat liking only. For sweet liking, results differed between gender and compounding factors. Liking for added sugar and sweet foods was positively linked to BMI in women unlike in men; liking for natural sweetness was negatively linked to BMI in both genders.Conclusions:This study demonstrates that the relationship between liking and BMI differs according to the gender in its magnitude for fat and in its nature for sweet, unlike that for salt. Liking for sweet and fat may be linked to overconsumption of the corresponding foods, especially in women. This warrants further investigation.

Contribution of Organic Food to the Diet in a Large Sample of French Adults (the NutriNet-Santé Cohort Study)

In developed countries, the demand for organic products continues to substantially increase each year. However, little information is available regarding the level of consumption of organic food and its relative share of the whole diet. Our aim was to provide, using individual consumption data, a detailed description of organic food consumption among French adults. Conventional and organic intakes were assessed using an organic food frequency questionnaire administered to 28,245 French adults participating in the NutriNet-Santé study. P values of Student t-test or Chi-square for the difference between genders were reported. Less than 12% of the respondents reported never consuming organic food in the past year. Women consumed on average 20% organic food in their whole diet per day while men consumed an average of 18%. The proportion of vegetables consumed that came from organic sources was 31% among women and 28% among men. Overall, the estimate of the contribution of organic food from products of plant origin was higher than that from products of animal origin. Our study provides a framework for the exploration of organic consumption and its correlates and can serve as a basis for future studies investigating relationships between the level of organic food consumption and health outcomes.

Development and Validation of an Individual Dietary Index Based on the British Food Standard Agency Nutrient Profiling System in a French Context

BACKGROUND Nutrient profiling systems could be useful public health tools as a basis for front-of-package nutrition labeling, advertising regulations, or food taxes. However, their ability beyond characterization of foods to adequately characterize individual diets necessitates further investigation. OBJECTIVES The objectives of this study were 1) to calculate a score at the individual level based on the British Food Standard Agency (FSA) food-level nutrient profiling system of each food consumed, and 2) to evaluate the validity of the resulting diet-quality score against food group consumption, nutrient intake, and sociodemographic and lifestyle variables. METHODS A representative sample of the French population was selected from the NutriNet-Santé Study (n = 4225). Dietary data were collected through repeated 24-h dietary records. Sociodemographic and lifestyle data were self-reported. All foods consumed were characterized by their FSA nutrient profile, and the energy intake from each food consumed was used to compute FSA-derived aggregated scores at the individual level. A score of adherence to French nutritional recommendations [Programme National Nutrition Santé guideline score (PNNS-GS)] was computed as a comparison diet-quality score. Associations between food consumption, nutritional indicators, lifestyle and sociodemographic variables, and quartiles of aggregated scores were investigated using ANOVAs and linear regression models. RESULTS Participants with more favorable scores consumed higher amounts of fruits [difference Δ = 156 g/d between quartile 1 (less favorable) and quartile 4 (most favorable), P < 0.001], vegetables (Δ = 85 g/d, P < 0.001), and fish, and lower amounts of snack foods (Δ = -72 g/d, P < 0.001 for sugary snacks); they also had higher vitamin and mineral intakes and lower intakes of saturated fat. Participants with more favorable scores also had a higher adherence to nutritional recommendations measured with the PNNS-GS (Δ = 2.13 points, P < 0.001). Women, older subjects, and higher-income subjects were more likely to have more favorable scores. CONCLUSION Our results show adequate validity of the FSA nutrient profiling system to characterize individual diets in a French context. The NutriNet-Santé Study was registered in the European Clinical Trials Database (EudraCT) as 2013-000929-31.