CORDIS
EU research results

CORDIS

English EN

Multi-country cOllaborative project on the rOle of Diet, FOod-related behaviour, and Obesity in the prevention of Depression (MooDFOOD)

Final Report Summary - MOODFOOD (Multi-country cOllaborative project on the rOle of Diet, FOod-related behaviour, and Obesity in the prevention of Depression (MooDFOOD))

Executive Summary:
Executive summary
MooDFOOD is a ‘Multi-country cOllaborative project on the rOle of Diet, Food-related behaviour, and Obesity in the prevention of Depression’ involving 14 organisations in eight European countries. The project was coordinated by the Vrije Universiteit Amsterdam; it was funded by the European Commission and lasted from April 1, 2014 until March 31, 2019. Depression is one of the most prevalent, severe and disabling disorders in the European Union (EU) and places a heavy burden on individuals and families. A large proportion of the EU population is overweight, which increases depression risk. Targeting food-related behaviours and nutritional status of people prone to overweight and obesity offers opportunities to prevent depression. The MooDFOOD consortium combined expertise in nutrition, consumer behaviour, psychiatry and preventive health psychology and used a unique integrative approach. Existing high quality data from longitudinal prospective European cohort studies were combined with new data from surveys in order to gain scientific knowledge on the bidirectional links of food intake, nutrient status, food-related behaviours and obesity with depression. In addition, short-term experiments were conducted to investigate the feasibility of improving food-related behaviour through a web-based online course. Also, a multi-country, randomized controlled depression prevention trial was conducted among 1025 participants to investigate the effectiveness of two nutritional strategies (nutrient supplements and food-related behavioural activation therapy) in preventing clinical depression. The scientific knowledge obtained from the studies conducted in the project, already resulting in 37 scientific publications, was integrated with the evidence from the current scientific literature in order to develop novel and evidence-based nutritional strategies to prevent depression. The MooDFOOD project identifies these key nutritional recommendations: 1) Following a healthy dietary pattern consistent with national dietary guidelines may help prevent the development of depressive episodes; 2) Weight loss among people with obesity can help prevent depression; 3) There is no evidence to support taking nutritional supplements for the prevention of depression. In close collaboration with stakeholders and experts, MooDFOOD transformed the identified nutritional strategies into guidelines and practical tools to guide policy at EU and Member State levels and to support health professionals (dieticians, general practitioners and psychologists). All conclusions and materials can be found on our website www.moodfood-vu.eu.
Project Context and Objectives:
MooDFOOD:

1.1 Burden of depression
Depression is one of the most prevalent mental disorders worldwide. An estimated 322 million people (4.4% of the global population) suffered from depression in 2015 (WHO 2017). Lifetime prevalence of major depression varies from 17% in women to 9% in men (Alonso et al. 2004). Currently, depression is globally the second leading cause of disability by years with disability (Ferrari et al. 2013). Depression is predicted to rise to a first place ranking with regard to global disease burden by 2030 (Mathers & Loncar 2006).

1.2 Nutrition as preventive strategy
Although the exact aetiology remains unclear, depression is clearly a multi-factorial disorder (Sullivan et al. 2000): its risk is determined by a complex interplay of social, environmental and biological factors, including stressful experiences, genetics, hormonal action in the brain, and lifestyle factors – including nutrition. Observational studies suggest that persons with a healthy diet are less likely to have depression or develop depression, but it remained unclear whether healthy dietary patterns and/or food-related behaviours are causally linked to the development of depression and whether other factors, such as the social environment or obesity, may (partly) explain the association.

2.1 MooDFOOD consortium and funding
On 1 January 2014, the European Commission-funded MooDFOOD project, led by professors Marjolein Visser and Ingeborg Brouwer of the Department of Health Sciences from the Vrije Universiteit Amsterdam, The Netherlands, commenced. MooDFOOD stands for ‘Multi-country cOllaborative project on the rOle of Diet, FOod-related behaviour, and Obesity in the prevention of Depression’ (www.moodfood-vu.eu). The 5-year MooDFOOD project was undertaken by a multidisciplinary consortium, involving 14 organisations in eight European countries, using a unique integrative approach, which combines expertise in nutrition, consumer behaviour, psychiatry and preventive health psychology. MooDFOOD is one of the largest consortia investigating depression, and the first to address the role of nutrition in its prevention (Cabout et al., 2017).

2.2 Objectives and conceptual model
The core aim of the MooDFOOD project was to contribute to the prevention of depression in EU citizens. To reach this aim, the multi-disciplinary international consortium pursued the following two objectives. (1) To gain a better understanding of the psychological, lifestyle and environmental pathways underlying the multi-faceted, bidirectional links of food intake, nutrient status, food-related behaviours and obesity with depression. (2) To develop and disseminate innovative evidence-based, feasible, effective and sustainable nutritional strategies for the prevention of depression in EU citizens.
A conceptual model was created and used as a basis for the MooDFOOD project (Figure 1). The model illustrates the different pathways underlying the bidirectional link of food intake, nutrient status and food-related behaviours with depression. The model is based on the strong bidirectional link of food intake and food-related behaviour with depression (pathway 1). Food intake and food-related behaviours include food choice, patterns of food intake and psychological eating behaviours (e.g. mindful eating, emotional eating, external eating and restrained eating). Nutrient status refers to measurements of nutrients in the blood (e.g. vitamin D in plasma). Depression includes clinical depression, with consideration being given to its subtypes, duration and severity in the MooDFOOD project, as well as the presence of depressive symptoms. The link between food intake/food-related behaviours and depression can be direct (pathway 1) and can be explained by the role of nutrients or combinations of nutrients in the pathophysiology of depression. Indirect pathways mediated by obesity (pathway 2) and/or the social environment (pathway 3) are also likely to exist. Obesity was measured by bodyweight/body mass index (BMI) and waist circumference, and changes were assessed over time. The social environment will be influenced by socioeconomic status, education, income, ethnicity, gender and demographic characteristics.

Figure 1. Conceptual model forming the basis of the MooDFOOD project (see attachment)

The MooDFOOD project generated fundamental insights into the mechanisms underpinning nutrition and the development of depression. The MooDFOOD project delivered epidemiological evidence for the bidirectional link between food intake, nutrient status and obesity with depression and identified underlying mechanisms using data from six prospective cohort studies. In addition, a targeted and extensive survey was conducted in two environmentally distinct regions (Denmark and Spain) to collect additional data. MooDFOOD also performed a web-based experiment in order to investigate the feasibility of improving food-related behaviour through an intensive web-based course. When feasible, such a course could potentially have a large impact by reaching many EU citizens. In addition, a multi-centre, randomized controlled trial was conducted in four countries to investigate the feasibility and effectiveness of two nutritional strategies for the prevention of depression. This trial was conducted to identify whether previously observed associations between nutrition and depression are indeed causal. The results of the MooDFOOD project have been reported, or will be soon, in 48 international, peer-reviewed scientific journals. A summary of these scientific results is provided in the paragraphs below, followed by the overall project conclusions.

Project Results:
3.1 MooDFOOD cohorts and survey
Data from six prospective cohort studies were used, representing diversity across Europe both in terms of population characteristics and differences in diet as well as the social, political and economic contexts. The following cohorts studies were included in the MooDFOOD project: the Longitudinal Aging Study Amsterdam (LASA) (Hoogendijk et al. 2016), the NEtherlands Study on Depression and Anxiety (NESDA) (Penninx et al. 2008), the HEalthy Life In an Urban Setting study (HELIUS) (Stronks et al. 2013), the Whitehall II study (WHITEHALL II). (Marmot & Brunner 2005), the Invecchiare in Chianti study (INCHIANTI) (Ferrucci et al. 2000), and the Age, Gene/Environment Susceptibility study (AGES) (Harris et al. 2007). These cohorts were specifically selected as they provide high quality, repeated information on food intake, nutrient status and/or obesity and involve a valid assessment of (change in) depressive symptoms and/or prevalent and/or incident clinical depression. As part of the MooDFOOD project, new measures of food intake and food-related behaviour were added to two of the above cohorts (LASA and NESDA).

In June and July 2014, a targeted survey was conducted in Denmark and Spain with Qualtrics – a panel service agency. The panel service agency invited a randomly selected representative sample of their panellists to fill out an online questionnaire. In Denmark 1522 persons participated (analytical sample of 1472 persons) and in Spain 1512 persons participated (analytical sample of 1488 persons).

3.2 Dietary patterns, food products and depression
Within the MooDFOOD project, the association between several existing, a priori dietary pattern scores (e.g. the Mediterranean Diet Score (MDS), the Alternative Healthy Eating Index (AHEI) or the Dietary Approaches to Stop Hypertension (DASH) score) and depressive symptoms and clinical depression has been examined. In addition, the association between several a posteriori (driven by the underlying data dietary patterns (e.g. an inflammatory dietary pattern) were derived using reduced rank regression and were subsequently linked to depressive symptoms. Moreover, the role of individual foods groups such as fruits, vegetables, and fish intake as well as the role of sugar in depression were also examined. Finally, studies were conducted to investigate the mechanism underlying the diet-depression relationship.

3.2.1 A priori dietary patterns
The cross-sectional associations between depressive and anxiety disorders, and their clinical characteristics such as disorder type, severity, chronicity and clinical subtype, and two dietary patterns scores (MDS and AHEI) were studied in 1634 participants from the NESDA study (Gibson-Smith et al. 2018a). Diet quality was significantly worse among patients with a current disorder than among healthy controls. Participants with concurrent depressive and anxiety disorders had the lowest MDS and AHEI diet quality scores. Higher chronicity (MDS) and severity (MDS and AHEI) showed a dose-response association with poorer diet quality. Participants with different depression subtypes had a similar diet quality. Overall, the associations were stronger for diet quality using the MDS than the AHEI.

To what extent current depressive symptoms, their short-term changes and the long-term history of depressive symptoms is associated with diet quality cores (MDS, AHEI and DASH) was examined in the LASA study (Elstgeest et al., in press). High current depressive symptoms were associated with lower MDS in men but not in women. Short-term chronic or recurrent depressive symptoms were associated with lower MDS scores, and a trend for lower AHEI scores, compared to no depressive symptoms. Having a history of high depressive symptoms in the past 15 years was associated with lower MDS and AHEI scores compared to having never having high depressive symptoms score in men, but not in women. No associations were found with the DASH score.

3.2.2 A posteriori dietary patterns
In four different studies we applied the Reduced Rank Regression (RRR) method to derive dietary patterns and subsequently investigated the association between these patterns and depression. Using the RRR method, the patterns are derived in an explorative way using the whole diet, similar to for example principle component analysis, while also using available knowledge on a potential pathway linking diet to depression, such as specific nutrients or biomarkers (the so called response variables).

In a first prospective study (with a follow-up of 9 years), a dietary pattern was derived in the In Chianti study using the intake of the nutrients EPA+DHA, folate, magnesium and zinc as response variables (Vermeulen et al. 2016). These nutrients have been previously linked to depression. The derived dietary pattern was rich in vegetables, olive oil, grains, fruit, fish and moderate in wine and red and processed meat and was labelled as 'typical Tuscan dietary pattern'. After full adjustment, an inverse association was observed between this dietary pattern and depressive symptoms at baseline. Moreover, a higher score on this dietary pattern was associated with lower depressive symptoms over 9 years.
Using the same response variables as the study described above, a dietary pattern was developed in the multi-ethnic HELIUS study and linked to current depressive symptoms (Vermeulen et al. 2017a). The dietary pattern was characterized by milk products, cheese, whole grains, vegetables, legumes, nuts, potatoes and red meat. A higher score on this pattern was again associated with lower depressive symptoms in the whole population and in the South-Asian Surinamese subgroup. No statistically significant associations were found among the other ethnic subgroups.

Using data from the In Chianti study, RRR was also used with several inflammatory markers as response variables to derive dietary patterns (Vermeulen et al. 2018a). The association between the derived patterns and repeatedly assessed depressive symptoms over 9 years was investigated. The first dietary pattern was related to the inflammatory markers C-reactive protein, interleukin-6 and tumor necrosis factor α. Higher scores indicated high intakes of sweet snacks, refined grains, pasta, rice and sauce and low intakes of bread, game, shellfish, other alcoholic beverages, other vegetables, wine, dairy products, olive oil and fish. This patterns was not associated with depressive symptoms. The second dietary pattern was related to c-reactive protein, interleukins 18 and 1β, and interleukin-1 receptor antagonist and higher scores indicate high intakes of pasta, sugar-sweetened beverages, processed meat, chocolate and sweets, sauce, other alcoholic beverages and low intakes of dairy products, fruit, added sugars, olive oil, butter, fish, coffee and tea and vegetable oil. Again, no association between this dietary pattern and depressive symptoms over time was observed, although low scores were associated with a lower odds for ‘high depressive symptoms’.

In two additional studies, RRR was used to derive a high-fat high-sugar dietary pattern based on the intake of mono-saccharides, saturated fat and total fat as response variables. In the HELIUS study, the derived dietary pattern included chocolates, red meat, added sugars, high-fat dairy products, fried foods an creamy sauces (Vermeulen et al. 2017b). High scores on this pattern were cross-sectionally associated with higher depressive symptoms and higher odds for ‘high depressive symptoms’. Interestingly, when in a second study the same strategy was applied to a different dataset, the Whitehall II study, no associations between higher high-fat high-sugar dietary pattern scores and depressive symptoms were observed (Vermeulen et al. 2018b).

In a final study, healthy dietary patterns were derived in participants of the AGES-Reykjavik study. Current dietary intake was assessed, but uniquely also dietary information obtained in adolescence and midlife age was available for these participants (Birgisdottir et al., in preparation). Healthy dietary patterns emerged in all age groups, adolescence (fish, fruits and vegetables), mid-life (vegetables, fruits, fish, fish liver oil, unprocessed meat) and old age (fresh fruits, vegetables, skyr (sour milk product), whole wheat bread, oatmeal, rye bread, fish liver oil, fish). A healthy dietary pattern in late life was inversely associated with depressive symptoms. Those scoring high in healthy pattern at all three time points (i.e. throughout life, 13% of study sample) had the absolute lowest depressive symptoms compared to groups scoring low on healthy diet.

3.2.3 Food groups
Previous studies have linked higher diet quality to lower depressive symptoms or less clinical depression. Using data from two cohorts, we examined which food groups are in particular related to depression. In the first study we investigated the bidirectional associations between repeated assessments of intake of 13 food groups and depressive symptoms (assessed at baseline and after 3, 6 and 9 years) in 1058 Italian participants of the InChianti study (Elstgeest et al., in press). Higher fish intake was associated with lower subsequent depressive symptoms, while higher sweet foods intake was associated with higher symptoms. In the other direction, higher depressive symptoms were associated with lower intakes of vegetables, lower intakes of red and processed meat, higher intakes of dairy and higher intakes of savoury snacks. Fruits, nuts and legumes, potatoes, whole grain bread, olive oil, sugar-sweetened beverages, and coffee and tea were not significantly associated in either direction.

In the NESDA study, the association between the overall MDS and clinical depression or anxiety was far stronger than the associations of single food groups (Gibson-Smith et al., manuscript under revision). A higher consumption of non-refined grains was associated with a lower odds of current depression/anxiety disorders versus healthy controls (but not with a lower odds of remitted anxiety/depression) and with a lower depression and anxiety severity. A higher consumption of vegetables was associated with a lower depression and anxiety severity. A higher consumption of fruits and vegetables was associated with a lower fear severity.

A final study focussed on the specific role of sugar intake from sweet food and beverages in common mental disorder (CMD) and depression using data from the Whitehall II study (Knüppel et al. 2017). Sweet food and beverage intake was measured with 15 items and sugar intake was calculated by multiplying the intake by the sugar content. Men in the highest tertile of sugar intake from sweet food and beverages had a 23% increased odds of incident CMD after 5 years (95% CI: 1.02 1.48) independent of health behaviours, socio-demographic and diet-related factors, adiposity and other diseases. The association was not observed in women. Neither CMD nor depression predicted changes in sugar intake, suggesting that reverse causation is not likely.

3.2.4 Mechanism linking diet to depression
Several studies examined the potential underlying mechanism that could explain the observed association between dietary patterns and depressive symptoms. In the HELIUS cohort it was examined whether BMI mediates the cross-sectional association between dietary patterns and depressive symptoms (Vermeulen et al., in press). In the 4,969 participants, four a posteriori dietary patterns were identified using principle component analyses (Western, healthy, mixed and Surinamese pattern) and one using reduced rank regression (high-fat high-sugar pattern). Although most diets were associated with depressive symptoms, no mediating effect of BMI was observed between these diets and depressive symptoms.

In the NESDA study it was examined whether the association between diet and depression was mediated by eating styles (Paans et al. 2019). Emotional, external and restrained eating were examined. Four dietary measures were investigated: energy intake (kcal/d), MDS score, sweet foods intake (g/d) and snack/fast food intake (g/d). Current depression diagnosis and severity were associated with lower MDS and higher intakes of sweet foods and snack/fast foods. The associations between current depression diagnosis and depression severity with intake of snack/fast food were mediated by external eating.

In conclusion, our observational studies on the link between diet and depression suggest that current depression as well as chronicity, severity and history of depression is associated with a poorer diet quality based on a priori dietary patterns. Of the three a priori diet quality scores used, the results were strongest and most consistent for the MDS. Regarding a posteriori dietary patterns, the results suggest that a dietary pattern linked to a high nutrient intake (i.e. EPA+DHA, folate, magnesium and zinc) were consistently associated with lower depressive symptoms in two studies. In contrast, dietary patterns linked to higher inflammation seemed not to be associated with depression, and the results for dietary patterns linked to a high-fat high-sugar intake were inconsistent between studies. Our results on food groups suggest that a higher fish, non-refined grains and vegetables intake, and a lower sweet foods intake might be associated with lower depressive symptoms and support mental health. Finally, our results indicate that the observed link between dietary patterns and depressive symptoms might not be mediated by BMI. Thus, the quality of the diet and BMI may act independently on depressive symptoms. The association between a high snack/fast food intake, but not high sweet food intake or lower MDS score, and clinical depression were mediated by external eating.

3.3 Nutrients and depression
In MooDFOOD, the potential relationship between blood nutrient status and depression was examined for three nutrients: vitamin B12, vitamin D and omega-3 fatty acids.

3.3.1 Vitamin B12
Whether serum vitamin B12 was associated with current depressive symptoms or the course of depressive symptoms over 16 years was examined in Dutch older participants of the LASA study (Elstgeest et al. 2017). Vitamin B12 was neither cross-sectionally nor prospectively associated with depressive symptoms (adjusted β for CES-D score over time, lowest versus highest quartile -0.04; 95% confidence interval -0.15-0.06). We also found no association with incident depression, except for a higher risk of depression over time in younger participants.

3.3.2 Vitamin D
The cross-sectional association of serum vitamin D status (25-hydroxyvitamin D, 25(OH)D) with depressive symptoms and clinical depression was studies among 5006 participants from the AGES-Reykjavik study who live at northern latitudes (Imai et al. 2015). Men and women with deficient (<30 nmol/l) versus adequate (≥50 nmol/l) vitamin D status had more depressive symptoms (difference 0.7 Geriatric Depression Scale points (95 % CI 0.4-0.9) and 0.4 (0.1-0.6) respectively). Deficient men were more likely to have current clinical depression (adjusted OR 2·51; 95 % CI 1·03, 6·13) compared with men with adequate vitamin D status, but this association was not observed in women.
Using data of the LASA cohort, cross-sectional associations between serum vitamin D levels and depressive symptoms were not significant after adjustment for confounders (de Koning et al. 2018). Longitudinally, women in the older LASA cohort (age 65-88 y) with baseline 25(OH)D concentrations lower than 75 nmol/L experienced more depressive symptoms in the following 6 years, compared with women with higher baseline concentrations. Reduced physical performance partially mediated this relationship. In men and in the younger-old cohort (age 55-65 y), no significant associations were observed.
In a second analysis in this same cohort, it was investigated whether change in serum vitamin D levels was associated with parallel change in depressive symptoms (Elstgeest et al. 2018a). Change was assessed over 13 years in the older LASA cohort and 6 years in the younger LASA cohort. In the older cohort, change in vitamin D status was not associated with change in depressive symptoms. In the younger cohort, associations were observed in those with lower baseline vitamin D concentrations (<58.6 nmol/l) - an increase in vitamin D over 6 years was associated with a decrease in depressive symptoms - but not in those with higher baseline vitamin D concentrations.

3.3.3 Omega-3 fatty acids
In the MooDFOOD project, two studies were conducted on the role of omega-3 fatty acids status and depression. In the first study, data from the AGES-Reykjavik cohort were used to examine the association between omega-3 fatty acid concentration and depression in an older sample with frequent intake of food high in omega-3 fatty acids, such as fish and cod liver oil (Imai 2015). Participants in the highest versus lowest tertile of total long chain omega-3 polyunsaturated fatty acids (PUFA) (EPA + DHA) had on average a lower Geriatric Depression Scale score (-0.16 (95% CI -0.41-0.09) but this was non-significant, and also had a lower risk of having clinical depression (OR: 0.45 (95% CI: 0.17-1.21). The latter association seemed to be largely driven by DHA. Other fatty acids were not significantly associated with clinical depression.

In the second study, conducted using data from the NESDA cohort, omega-3 PUFA status was compared between five diagnosis groups: controls, remitted anxiety or depressive disorder patients, current pure anxiety disorder patients, current pure depressive disorder patients, and current comorbid anxiety and depressive disorder patients (Thesing et al. 2018). The results show that lower omega-3 PUFA levels are only observed in patients with a current depressive disorder, especially in the more severe group with comorbid anxiety. Furthermore, lower omega-3 PUFA levels are associated in a dose-response fashion with higher depressive symptoms severity. Finally, within patients with current disorders, several clinical characteristics (such as depressive symptoms severity) were significantly associated with a higher omega-3:FA or lower omega-6:FA ratio. No specific pattern of association was detected for absolute omega-6 PUFA levels and no association between PUFAs and pure anxiety disorders was found.

These observational studies do not show consistent associations between (change in) blood nutrient status and (change in) depressive symptoms, with the exception of omega-3 fatty acids. The results therefore provide no strong evidence for the presence of a relationship of vitamin B12 and D with depression. The potential role of omega-3 fatty acid status in depression needs to be further examined.

3.4 Obesity and depression
In the HELIUS cohort the cross-sectional association between obesity (based on BMI and waist circumference) and depressive symptoms was investigated in 22,165 Dutch adults from six ethnic origins (Gibson-Smith et al. 2018b). The association was stronger for participants of Dutch and African Surinamese origin, and weaker for those of Ghanian, South-Asian Surinamese, Turkish and Moroccan origin. These differences could not be explained by differences in health behaviours or somatic health, and were more pronounced when using BMI instead of waist circumference.

In the Whitehall II cohort, the bidirectional association between psychological distress and categorical change (5-year lag periods) in objectively measured weight and waist circumference was examined (Knüppel et al. 2018). Weight gain and loss were associated with increased odds for incident psychological distress, assessed by the General Health Questionnaire, in models with and without time-lag. Waist circumference changes were only associated with psychological distress in models without time-lag. Reversely, psychological distress was associated with gain in body weight and waist circumference over the subsequent 5 years but not the second 5-year period. Thus, weight change (both gain and loss) increased the odds for psychological distress compared with stable weight over subsequent 10 years.

Whether being overweight or obese in childhood (based on objective measurements) is associated with depressive symptoms more than 60 years later, independent of late-life BMI, was examined in 889 participants of the AGES-Reykjavik study (Gibson-Smith et al., manuscript submitted). Being overweight or obese at age 8 and 13 was not associated with depressive symptoms during late-life, irrespective of late-life BMI. However, being overweight or obese at age 8 was associated with an increased risk of lifetime MDD (OR 4.03 (95%CI 1.16-13.96)) with a trend for age 13 (2.65; 0.69-10.26) although these estimates were based on small numbers.
Using clinical diagnosis data from the NESDA study, we examined the longitudinal relationship of BMI and waist circumference with the incidence and persistence of MDD over time (Gibson-Smith et al. 2016b). Both higher BMI and waist circumference increased the odds of developing MDD over a 6-year period, even after adjustment for health and lifestyle variables (OR per SD increase 1.17; 95%CI 1.00-1.3 and OR 1.20; 1.00-1.43)). However, in already depressed patients, there was no relationship of BMI or waist circumference with the persistence of MDD.

The same NESDA study was also used to determine whether major depressive disorder leads to subsequent weight change (Gibson-Smith et al. 2016a), and compared 2-year weight changes (categorized as weight loss (> 5% loss), weight stable (within 5% weight loss or gain), and weight gain (> 5% gain)) between patients with current MDD, patients with remitted MDD, and healthy controls. Current, but not remitted MDD, was associated with both weight gain and weight loss over a 2-year period after adjustment ((OR 1.67; 95% confidence interval 1.37-2.03) and 1.27; 1.01-1.61)). Although antidepressant use was associated with weight gain, the association between current MDD and weight gain was independent of antidepressant use.

Potential biological mechanisms linking body weight to depression was examined in two studies. In the NESDA cohort (Milaneschi et al. 2015), higher leptin concentrations in plasma were observed in persons with current and remitted atypical MDD subtype versus healthy controls, and this association was stronger in those with higher BMI. Leptin dysregulation (resistance) may thus be an underlying mechanism of the association between higher BMI and MDD with atypical features. Furthermore, underlying genetic profiles may underlie the observed association between BMI and MDD (Milaneschi et al. 2016). In a genome-wide association study, atypical MDD was associated with BMI and triglycerides genomic profile risk scores generated from meta-analysis results of large international consortia, suggesting that the association between BMI and atypical depression stems from the same pathophysiologic mechanism. Potential behavioural mechanisms linking body weight to depression are described in the next section.

These observational studies suggest that obesity is associated with higher depressive symptoms or clinical depression, which may even start in childhood. However, the association differs between ethnic groups. Future studies are needed to determine whether differential social-cultural based normative values or underlying pathophysiology across ethnic groups might explain these observed ethnic differences in the association between obesity and depression. The studies also showed that 2-year weight changes are bi-directionally linked to clinical depression, independent of antidepressant use. A current clinical depression can also lead to subsequent weight loss. Finally, our results suggest that the link between obesity and atypical clinical depression might be explained by leptin dysregulation or the shared underlying genetic profile for BMI and depression.

3.5 Food-related behaviour and depression
The MooDFOOD project also focused on the role of food-related behaviours. First, the association between psychological eating styles and clinical depression was examined in two studies. In addition, a series of studies was conducted to identify potential pathways linking obesity to depression and focused on psychological eating styles, body image, personality traits and biopsychosocial variables. Another study was conducted on food provisioning as an underlying mechanism for the link between nutrition and depression. Finally, in five studies the role of mindful eating in depressive symptoms was investigated.

3.5.1 Psychological eating styles and depression
Persons with depression have been found to present disturbances in eating styles, but it is unclear whether eating styles are different in subgroups of depressed patients. The association between depressive disorder, severity, course and specific depressive symptom profiles and unhealthy eating styles in 1060 remitted depressed patients, 309 currently depressed patients and 381 healthy controls from the NESDA study (Paans et al. 2018b). Remitted and current depressive disorders were significantly associated with higher emotional eating and higher external eating. Higher severity and longer symptom duration were also associated with more emotional and external eating. Little differences in eating styles between depression course groups were observed. Neuro-vegetative depressive symptoms contributed relatively more, while mood and anxious symptoms contributed relatively less to emotional and external eating. No depression associations were found with restrained eating.

In a second study, baseline data of 990 participants of the MooDFOOD depression prevention study were used to examine the associations of history of major depressive disorder and depression severity with psychological eating styles (Paans et al. 2018c). Depression history and severity were associated with more emotional and uncontrolled eating and with less cognitive restrained eating. Mood, somatic, and cognitive symptom clusters were also associated with more emotional and uncontrolled eating, and with less cognitive restrained eating. The somatic depressive symptoms "increased appetite" and "increased weight" were more strongly associated to eating styles compared to other symptoms. These associations were consistent between the four European countries.

3.5.2 Mechanisms linking depression to body weight
Depression is often associated with weight gain but underlying mechanisms are unclear. The potential mediating role of three psychological eating styles (emotional eating, external eating and restrained eating) in the association between depressive symptoms and 5-year weight gain was examined in a sample of 298 fathers and 294 mothers (van Strien et al. 2016a). The overall association between depressive symptoms and weight gain was statistically non-significant in both sexes, but there was a causal chain between depressive symptoms, emotional eating and weight gain in women. Depressive symptoms were related to higher emotional eating, and emotional eating predicted a greater increase in BMI independently of depressive symptoms. This mediation effect was found to be independent of 5-HTTLPR genotype. No such mediation was observed in the men. External eating and restrained eating did not act as mediators in either sex.
Potential replication of the mediation of the association between depressive symptoms and BMI by emotional eating was examined using data from the MooDFOOD survey conducted in Denmark and Spain (van Strien et al. 2016b). In both countries, emotional eating acted as a mediator between depressive symptoms and BMI. In Denmark, but not in Spain, mediation was stronger for participants with increased appetite (characteristic for atypical depression) and for females than for participants with decreases/no change in appetite and for males.
Another pathway through which depression might be linked to body weight was body image. Cross-sectional data from the NESDA study were used to explore this (Paans et al. 2018a). Higher BMI was associated with larger perceptual body size and with higher body image dissatisfaction. Independent of this, depression severity contributed to larger perceptual body size, and both current and remitted depression diagnosis as well as depression severity contributed to higher body image dissatisfaction. In general, both depression and higher BMI contributed independently to a larger body size perception as well as higher body image dissatisfaction.
Two other potential pathways through which depression might be linked to body weight are through personality traits (neuroticism, extraversion, conscientiousness) and cognitive reactivity (hopelessness, aggression, rumination, anxiety sensitivity). Depression and anxiety disorders are associated with specific psychological vulnerabilities, like personality traits and cognitive reactivity, that may also be associated with BMI. In the NESDA study, personality traits were not consistently related to BMI. In patients, higher hopelessness and aggression reactivity and higher depression and anxiety symptoms were associated with higher BMI (Paans et al. 2016). In contrast, in healthy individuals lower scores on hopelessness, rumination, aggression reactivity and anxiety sensitivity were associated with higher BMI. These results suggest that, particularly in people with psychopathology, cognitive reactivity may contribute to obesity.

A final potential pathway linking depression to body weight is through biopsychosocial variables. Which biopsychosocial variables contribute to weight gain over a 4-year period in persons with clinical depression or high depressive symptoms was investigated in the NESDA study (Paans et al. 2017). Twenty-one baseline psychological (e.g. rumination, mastery), lifestyle (i.e. smoking, alcohol consumption and physical activity) and biological (e.g. autonomic nervous system measurements, inflammation markers) variables and antidepressant use were considered as potential contributing variables. Clinical depression and depressive symptoms were associated with subsequent weight gain. None of the biopsychosocial variables or antidepressants was associated with weight gain, thus did not contribute to the observed increased weight gain risk in depression, except for alcohol intake and tricyclic antidepressant use.

3.5.3 Food provisioning and meal patterns
Using survey data from Denmark and Spain, food provisioning practices were associated with both diet quality and depressive symptoms. Tendency of impulse buying was linked to have more food-on-go, store energy-dense foods at home and cook less often meals from basic ingredients at home and thereby lower quality of diet. In turn, higher diet quality was associated with lower depressive symptoms, but impulse buying had an additional direct link to higher depressive symptoms. On the whole, food provisioning in home and outside home were interlinked and form a net that is associated with diet quality and depressive symptoms. Results are similar in DK and SP with some minor differences. These results are first ones that combine food provisioning patterns happening both outside and within households with reported diet quality, and whether these patterns have a role in depressive symptoms that goes beyond the impact on depressive symptoms (Broman-Toft et al., in preparation).
The meals in both DK and SP showed three types of meal patterns: those who ate regular meals and snacked rarely (52/54%), those who ate regular meals but snacked often (29/31%) and those with irregular meals snacking occasionally (19/15%). The pattern with irregular meals was associated with lower diet quality indicator, higher depressive symptoms, but not higher BMI.

3.5.4 Mindful eating
First, a scale to measure mindful eating was developed (Winkens et al. 2018b). This new scale measures mindful eating in common situations, and, in contrast to previous mindful eating scales, is independent from psychological eating styles such as emotional, external and restrained eating. The Mindful Eating Behavior Scale (MEBS) consists of four domains: Focused Eating (5 items), Eating in response to Hunger and Satiety Cues (5 items), Eating with Awareness (3 items), and Eating without Distraction (4 items). Because of low interfactor correlations between these domains, a total score combining the four domains should not be computed. The MEBS showed good internal consistency and preliminary convergent validity in the LASA dataset.

Second, associations between mindful eating and depressive symptoms were studied. In three European samples from Denmark, Spain, and the Dutch LASA study, higher scores on the mindful eating domains Focused Eating, Eating with Awareness, and Eating without Distraction were associated with a lower level of depressive symptoms and a lower likelihood of having depression (based on high symptoms and/or antidepressant use and/or psychological treatment) (Winkens et al. 2018a). In contrast to expectations, a lower score on the mindful eating domain Eating in response to Hunger and Satiety Cues, which was only measured in the Dutch sample, was significantly associated with a lower level of depressive symptoms but not with depression. In a second study conducted using data from the LASA study, prospective associations between mindful eating and 3-year change in depressive symptoms were investigated (Winkens et al., manuscript submitted). Higher baseline scores on Focused Eating, Eating with Awareness, and Eating without Distraction were related to a decrease in depressive symptoms. Eating in response to Hunger and Satiety Cues was not related to change in depressive symptoms.

Third, two possible underlying mechanisms in the associations between mindful eating and depressive symptoms were examined: psychological eating styles and food intake. The eating styles emotional, external, and restrained eating were tested as mediating factors in multiple mediation models of the associations between mindful eating domains with change in depressive symptoms (Winkens et al. 2019). External eating was a mediator in the associations of Eating in response to Hunger and Satiety Cues, Eating with Awareness, and Eating without Distraction with change in depressive symptoms. Emotional and restrained eating did not mediate the associations. Post-hoc analyses showed that Eating with Awareness and Eating without Distraction were also related to change in depressive symptoms through external eating preceded by emotional eating (serial mediation). These results could imply that people with higher mindful eating are less susceptible to attractive food cues, either directly due to mindful eating or through decreased emotional eating. Quality of food intake (using the MDS score) and quantity of food intake (in kcals/d) were also tested as mediating factors in the associations of mindful eating domains with change in depressive symptoms (Winkens et al., manuscript submitted). Diet quality was not mediating the association between mindful eating domains and change in depressive symptoms. Higher scores on Eating with Awareness and Eating without Distraction were associated with a decrease in depressive symptoms through lower total energy intake. Controlling these analyses for physical activity and BMI did not change the conclusions. Diet quantity was no mediator in associations of Focused Eating with depressive symptoms.

3.6 MooDFOOD meta-analysis
The aim of this meta-analysis was to examine the association between adherence to three a priori healthy dietary patterns and depressive symptoms (Nicolaou et al., manuscript submitted). We developed a protocol in which definitions of dietary exposures and confounders were harmonized across cohorts in order to minimize potential heterogeneity in meta-analysis. We included six population-based studies that measured both diet and depressive symptoms: The Invecchiare in Chianti (InCHIANTI); the Longitudinal Aging Study Amsterdam (LASA); Netherlands Study of Depression and Anxiety (NESDA); Healthy Life in an Urban Setting (HELIUS); the Australian Longitudinal Study on Women's Health (ALSWH) and Whitehall II. These cohorts included data of over 23,000 persons representing varied populations in terms of age, sex, ethnicity and risk/vulnerability of depression. The included dietary patterns were the Mediterranean Diet Score (MDS) based on the methodology described by Panagiotakos et al 2007, the Alternative Healthy Eating Index 2010 (AHEI) as described by Chiuve et al 2012, and the Dietary Approaches to Stop Hypertension (DASH) using the method described by Fung et al 2008.
In cross-sectional analyses based on all six cohorts, a higher MDS score (β -0.065; 95% CI -0.094--0.036) the DASH diet (-0.061; -0.092--0.030) and AHEI score (-0.045; -0.066--0.024) was associated with lower depressive symptoms. Effect sizes were small, as both dependent and independent variables were standardised; this indicates a 0.04 to 0.06 standard deviation (SD) lower depressive symptoms score per 1 SD higher dietary pattern score. The MDS score seemed to have a stronger association with high depressive symptoms (OR 0.87; 95% CI =0.84-0.91) than the AHEI (0.93; 0.88-0.98) and DASH diet (0.94; 0.86-1.02). Thus, a 1 SD higher dietary pattern score was associated with an approximately 10% lower odds of high depressive symptoms.
Prospective analyses conducted in three cohorts with a follow-up time of 5-6 years, showed that baseline dietary pattern was associated with subsequent change in depressive symptoms at follow-up for the DASH diet only (-0.030; -0.047--0.013). The results for the association between dietary patterns at baseline and subsequent incidence of high depressive symptoms showed that higher MDS (0.88; 0.81-0.94) the AHEI (0.95; 0.85-1.04) and the DASH (0.90; 0.84-0.97) scores were associated with a lower incidence. There was no consistent pattern of interaction by age, sex and ethnicity.
An inverse association between higher adherence to three different a priori defined dietary patterns and depressive symptoms was observed in cross-sectional as well as prospective analyses over a time frame of five to seven years. Although this study is observational, these results show that greater adherence to a healthy dietary patterns is associated with fewer depressive symptoms and a lower risk of developing depressive symptoms over time. Promotion of a healthy diet, compatible with national dietary guidelines that are developed to prevent chronic diseases, may thus have additional benefits for mental health.

4 MooDFOOD web-based intervention
4.1 Aim and design
The aim of this study was to investigate the feasibility of an 8-week unguided online mindful eating course. First, we aimed to investigate differences in characteristics of participants who finished the intervention and participants who dropped out during different phases of the intervention. We also aimed to investigate user experience. As we do not know whether the developed online course on mindful eating could increase the targeted behaviour, we also aimed to investigate changes from pre-test to post-test in mindful eating behaviours, self-efficacy for mindful eating, and goal attainment for mindful eating. The information of this study can be used to develop an online intervention for mindful eating that is adopted to the needs of the users, that reaches large groups of people from the general population, that is perceived as easy, and that is capable of changing the targeted behaviours.

A randomized clinical trial was conducted in the Netherlands and Denmark in which the mindful eating course was compared to a non-food sustainability course that served as the control group. Mindful eating was assessed by the Mindful Eating Behavior Scale (MEBS), a scale developed in the MooDFOOD project (Winkens et al., 2018b). The MEBS consists of 17 items that make up four domains: Focused Eating, Eating in response to Hunger and Satiety Cues, Eating with Awareness and Eating without Distraction.

4.2 Intervention
Participants had access to the online course for 8 weeks after creating an account. Each course consisted of three modules with a specific topic and these modules consisted of three submodules on which the advice was focused. Each submodule consisted of the elements (1) recommendations, (2) practical tips and (3) tools/exercises on the topic of the submodule. Each course included the use of several diaries which could be filled out and viewed online and in an app to support behaviour change. Participants were allowed to choose which (sub)modules they used within their intervention arm. To improve engagement and continuation, automatic email reminders were sent to participants to complete modules as well as complete the online questionnaires for research purposes. A detailed description of the course can be found elsewhere (Winkens et al., manuscript in preparation)

4.3 Feasibility
The intention-to-treat sample (ITT) consisted of 350 randomized participants, and the per protocol sample of 217 participants who visited at least 1 intervention element. The total completion rate was low (29.3%) and attrition rates did not differ between the intervention and control course. The number of intervention elements visited was higher in the Netherlands compared to Denmark. No other differences were found between participants who finished or dropped out during the intervention. The mindful eating course was perceived as more difficult, but scored higher on user satisfaction. More people set goals within the mindful eating course, but a lower number of goals was achieved compared to the control group. No differences were found in number of times logged in and intervention elements visited between the two courses, except for the participants who visited at least one intervention element: they logged in more often in the mindful eating course. Participants in the mindful eating group showed a significant decrease in the mindful eating domain Focused Eating and an increase in the mindful eating domain Eating without Distraction compared to the control group. No differences were found for the other two mindful eating domains, self-efficacy, and goal attainment for mindful eating. Overall, an unguided 8-week online self-help intervention on mindful eating seems feasible for the general population in terms of user experience and increasing mindful eating behaviours.

5 MooDFOOD depression prevention trial
5.1 Aim and design
The MooDFOOD prevention trial examines the feasibility and effectiveness of two different nutritional strategies (multi-nutrient supplementation and food-related behavioural change therapy (FBC)) to prevent depression in individuals who are overweight and have elevated depressive symptoms but who are not currently or in the last 6 months meeting criteria for an episode of major depressive disorder (MDD). The randomized controlled prevention trial has a two-by-two factorial design and a total of four intervention groups were be created: (1) placebo supplement group; (2) placebo supplement FBC group; (3) multi-nutrient supplementation group; (4) multi-nutrient supplementation + FBC group. Interventions lasted 12 months. In total 1025 participants aged 18–75 years with body mass index between 25–40 kg/m2 and with a Patient Health Questionnaire-9 score ≥ 5 were recruited at four study sites in four European countries (Germany, Spain, United Kingdom and The Netherlands). Baseline and follow-up assessments took place at 0, 3, 6, and 12 months. Primary endpoint was the onset of an episode of MDD, assessed according to DSM-IV based criteria using the MINI 5.0 interview. Depressive symptoms, anxiety, food and eating behaviour, and health related quality of life were secondary outcomes. The design of the trial is described in Roca et al. 2016.

5.2 Intervention
Participants were randomized with equal probability to the four intervention groups, stratified by study site and participants’ lifetime history of depression status. Patients received either multi-nutrient supplements (1412 mg of eicosapentaenoic and docosahexaenoic omega-3 poly unsaturated fatty acids (ratio 3:1), 30 μg selenium, 400 μg folic acid, and 20 μg vitamin D3 coupled with 100 mg calcium) or placebo, each provided in two pills per day, taken daily for one year. FBC consisted of a protocol-based intervention that incorporated standard behavioural activation approaches in order to improve mood by changing dietary habits, food-related behaviours (e.g. snacking), increasing positive behaviours and emphasising a Mediterranean-style diet. FBC was provided in up to 21 sessions (15 individual, 6 group) for one year.

5.3 Feasibility
Among 1025 participants (mean age 46.5 years; 772 (75%) women; mean BMI 31.4 kg/m2) a total of 779 (76%) completed the trial. Missing rates for the primary outcome did not differ between the four intervention groups.
Of those randomized to FBC, 71% attended ≥8 of 21 sessions. A median of 14 individual sessions were attended (interquartile range 6-15), and a median of 0 group sessions (0-4) were attended. The majority of attendance came from individual sessions (47% attended all individual sessions, 53% attended no group sessions). FBC was highly acceptable to users (e.g. 84% of respondents very or somewhat satisfied; 85% reporting worth doing, 85% would recommend to a friend). Participants rated individual sessions more positively than group sessions. Participants in FBC reported attempting recommended food-related behaviour strategies and changing to the recommended Mediterranean diet at a high frequency (e.g. 87% limiting snacks; 85% paying attention to habitual snacking; 82% avoiding emotional eating; 83% eat 300-400 g of vegetables a day; 81% eat 2-3 pieces of fruit a day; 81% limit intake of processed food and drinks) (Grasso et al., manuscript submitted). Strategies were generally reported as successful (90% somewhat or very): most frequently reported as very successful were planning shopping, eating 3 regular main meals and controlling impulses when shopping. FBC resulted in significantly increased consumption of vegetables, fruit, fish, pulses/legumes and whole grains and decreased intake of sweets/extras relative to no FBC. There was no evidence that gender or history of depression moderated compliance with FBC. Complier Average Causal Effect (CACE) analysis found that when accounting for rates of those achieving a pre-specified minimum dose of FBC, FBC did significantly reduce incidence of major depression over 1 year in overweight individuals with sub-syndromal depression (OR 0.78; 95% CI 0.64-0.95).

Of all 12-month follow-up participants, 77% of participants had adherence of >70% to the supplements/placebo. A longitudinal analysis of self-reported compliance indicated 60% had high levels of compliance at all-time points. The most frequently reported reason for not taking the supplements was forgetting. Consistent with good compliance to supplements, in the random sample whose blood samples were analysed, levels of selenium, folic acid, vitamin D and eicosapentaenoic acid significantly increased in those receiving supplements but not in those receiving placebo. Nutritional supplements were significantly less acceptable than FBC (e.g. only 44% would recommend to a friend, 29% somewhat or very satisfied and 62% neutral, 23% thought helpful), but easy to use (75%). Men were more compliant with nutritional supplements than women. There was no effect of compliance on nutritional supplements versus placebo.
In conclusion, both interventions had reasonable compliance and significantly impacted on expected nutritional/dietary outcomes. Only F-BA was judged acceptable and helpful by users and individual FBC sessions were well-attended. FBC was beneficial in those who were more compliant. A complete description of the effectiveness of the intervention is described elsewhere (Watkins et al., manuscript in preparation)

5.4 Effectiveness
During 12 month follow-up, 105 (10%) developed MDD. There was no significant difference in episodes of major depressive disorder over 1 year follow-up with multinutrient supplementation versus placebo (54 (10.5%) vs 51 (9.9%)) or with food-related behavioural activation therapy versus no therapy (48 (9.4%) vs 57 (11.1%)). Neither supplements (odds ratio (OR) 1.06; 95%-confidence interval CI 0.87-1.29) F-BA (0.93; 0.76-1.13) nor their combination (0.93; 0.76-1.14 p for interaction=0.48) affected MDD onset.

There were no significant supplement-by-FBC interactions for any of the secondary outcomes. FBC was related to lower anxiety Generalized Anxiety Disorder-7 (GAD) scores at 12 months follow-up (adjusted mean difference -0.48 95% CI -0.84 to -0.12 p=0.01) but not on other secondary outcomes. There was a significant effect of supplements on overall follow-up measures of the PHQ (0.65; 0.25-1.06) Inventory of Depressive Symptomatology 30-SR scores (1.20; 0.29-2.10) and GAD (0.50; 0.16-0.84) scores, and on Patient Health Questionnaire-9 (PHQ) scores at 12 months follow-up (0.56; 0.11-1.01) showing less improvement in depressive and anxiety symptoms relative to placebo.
The effect of FBC on PHQ at 12 months follow-up was more favourable (larger reduction) when baseline PHQ depression severity was higher, and the effect of FBC on health utility scores at 12-month follow-up was larger in the United Kingdom compared to the Netherlands. Use of supplements resulted in higher follow-up anxiety scores when baseline anxiety severity scores were higher. The results of Complier Average Causal Effect (CACE) analyses, which accounts for treatment adherence, again showed no effect of supplements on the primary outcome, and similar effects for secondary outcomes. However, CACE analyses found a significant effect of FBC on the primary outcome MDD (OR 0.78; 95% CI 0.64-0.95).
In conclusion, among overweight or obese adults with depressive symptoms, multi-nutrient supplementation compared with placebo and food-related behavioural activation therapy compared with no therapy did not reduce episodes of major depressive disorder during 1 year. These findings do not support the use of these interventions for prevention of major depressive disorder. A complete description of the effectiveness of the intervention is described elsewhere (Bot et al. 2019).

6 MooDFOOD project conclusions
The conclusions based on MooDFOOD research were integrated with evidence from the scientific literature in order to develop evidence-based nutritional strategies for the prevention of depression. The integration process was performed using three scientific integration meetings at which both researchers and relevant stakeholders were invited. A paper describing the MooDFOOD integration and dissemination process is in preparation (Pullar et al). The level of evidence of each conclusion was categorized into four groups: 1) Strong evidence: when the food related factor shows a consistent causal impact on depression or depressive symptoms based on evidence from meta-analysis of RCTs, 2) Limited evidence: when the food related factor shows a consistent association with depression or depressive symptoms based on evidence from meta-analysis of prospective cohort studies, 3) Ambiguous evidence: when there was a high level of heterogeneity or conflicting results, and 4) Too few studies: when there were too few studies to draw inferences on the relationship.

Conclusions were developed regarding the prevention of depression and are described in section 6.1. As a service to the field, and in particular to patients who are diagnosed with clinical depression and their health professionals, the MooDFOOD consortium also developed evidence-based conclusions for the role of diet, supplements and obesity in the treatment of depression in patients who have been diagnosed with clinical depression. These conclusions were based on the scientific literature only (as the studies in the MooDFOOD project focused on the prevention of depression) and are described in section 6.2.

6.1 Prevention of depression
In this section the conclusions regarding the prevention of depression are listed by topic: nutrition, food-related behaviour, supplements and obesity. These conclusions are relevant for the general public.

Nutrition and the prevention of depression

• A limited level of evidence from meta-analyses of prospective cohort studies shows that eating a healthy dietary pattern may help to reduce depressive symptoms in the general population.
• The MooDFOOD prevention trial is the first trial to directly test the impact of a food-related behavioural activation therapy on the prevention of depression. The trial found no evidence in primary analysis that food-related behavioural activation therapy reduced depressive symptoms, or the incidence of MDD.
• There is a limited level of evidence that regular intake of fish, fruits and vegetables may help to reduce depressive symptoms.
• A limited level of evidence shows that unhealthy dietary patterns seem not associated with the development of MDD.
• Although there are some indications that sugar and refined grains, as well as junk/fast food, may have an impact on the onset of depression, currently the results of available studies are ambiguous.

Food-related behaviour and the prevention of depression
• There is a complex interplay between food-related behaviours, psychological eating styles, mindful eating and depression. Food behaviour and resulting dietary intake needs to be regarded as an interconnected system. Though there are currently too few studies to draw firm conclusions, there are promising indications that food-related behavioural strategies, such as mindful eating, may help reduce depressive symptoms in the general public.

Supplements and the prevention of depression
• The MooDFOOD prevention trial is the first trial to directly test the impact of a multi-nutrient supplement containing omega-3, selenium, folate, and vitamin D plus calcium on the prevention of depression. The supplement did not reduce depressive symptoms, or the incidence of MDD.
• There is a strong level of evidence that vitamin D, vitamin B12 in combination with folate, and multi-nutrient supplements are not effective in reducing depressive symptoms in the general population.
• There are too few studies to conclude whether individual supplements of magnesium, calcium, selenium, folate, vitamin B6, vitamin B12, omega-3 and zinc are effective in reducing depressive symptoms in the general population.
• There are too few studies to conclude whether reversal of vitamin deficiencies with nutrient supplementation is effective in reducing depressive symptoms in the general population.

Obesity and the prevention of depression
• Evidence from weight loss trials as well as bariatric surgery trials provides a strong level of evidence that weight loss reduces depressive symptoms in people with obesity.

6.2 Treatment of depression
The conclusions regarding the treatment of depression are listed by the following topics: nutrition, supplements and obesity. These conclusions are relevant for persons diagnosed with clinical depression.
Nutrition and the treatment of depression
• Two trials, HELFIMED and SMILES, have investigated the potential effect of diet on depressive symptoms in patients with MDD. Both have found a reduction in depressive symptoms as a result of the healthier dietary pattern.
Supplements and the treatment of depression
• There is a strong level of evidence that omega-3 supplementation (≥1g/day of EPA and DHA) has a small effect on reducing depressive symptoms in patients with clinical depression using antidepressants.
• There is a strong level of evidence that vitamin B12 and folate supplements are not effective in reducing depressive symptoms in patients with clinical depression.
• There are too few studies to conclude whether reversal of vitamin deficiencies with supplementation is effective in reducing depressive symptoms in patients with clinical depression.

Obesity and the treatment of depression
• There are too few studies to determine whether weight loss can help reduce depressive symptoms in patients with MDD and obesity.

7 List of MooDFOOD publications

1. Birgisdottir BE et al. Healthy dietary pattern at various life trajectories in relation to depression in old age – AGES-Reykjavik study. Manuscript submitted.
2. Bot M, Milaneschi Y, Penninx BW, Drent ML. Plasma insulin-like growth factor I levels are higher in depressive and anxiety disorders, but lower in antidepressant medication users. Psychoneuroendocrinology 2016;68:148-55.
3. Broman-Toft M, Kulikovskaja V, Pedersen S, Stancu C, Tudoran AA, Winkens L, van Strien T, Lähteenmäki L. The role of food provisioning practices in diet quality and association with depressive symptoms. Manuscript in preparation.
4. Cabout M, Brouwer IA, Visser M. The MooDFOOD project: prevention of depression through nutritional strategies. Nutrition Bulletin 2017;42:94–103.
5. De Koning EJ, Elstgeest LEM, Comijs HC, Lips P, Rijnhart JJM, van Marwijk HWJ, Beekman ATF, Visser M, Penninx BWJH, van Schoor NM. Vitamin D status and depressive symptoms in older adults: a role for physical functioning? Am J Geriatr Psychiatry 2018;26:1131-43.
6. Elstgeest LEM, Brouwer IA, Penninx BWJH, van Schoor NM, Visser M. Vitamin B12, homocysteine and depressive symptoms: a longitudinal study among older adults. Eur J Clin Nutr 2017;71:468-475.
7. Elstgeest LEM, de Koning EJ, Brouwer IA, van Schoor NM, Penninx BWJH, Visser M. Change in serum 25-hydroxyvitamin D and parallel change in depressive symptoms in Dutch older adults. Eur J Endocrinol 2018a;179:239-49.
8. Elstgeest LEM, Visser M, Penninx BWJH, Colpo M, Bandinelli S, Brouwer IA. Bidirectional associations between food groups and depressive symptoms: longitudinal findings from the Invecchiare in Chianti (InCHIANTI) study. Br J Nutr 2018b;Dec 27:1-12.
9. Elstgeest LEM, Winkens LHH, Penninx BWJH, Brouwer IA, Visser M. Associations of depressive symptoms and history with three a priori diet quality indices in middle-aged and older adults. Journal of Affective Disorders, in press.
10. Gibson-Smith D, Bot M, Brouwer IA, Visser M, Giltay EJ, Penninx BWJH. Association of food groups with depression and anxiety disorders. Manuscript in revision.
11. Gibson-Smith D, Bot M, Brouwer IA, Visser M, Penninx BWJH. Diet quality in persons with and without depressive and anxiety disorders. J Psychiatr Res 2018a;106:1-7.
12. Gibson-Smith D, Bot M, Milaneschi Y, Twisk JW, Visser M, Brouwer IA, Penninx BW. Major depressive disorder, antidepressant use, and subsequent 2-year weight change patterns in the Netherlands Study of Depression and Anxiety. J Clin Psychiatry 2016a;77:e144-51.
13. Gibson-Smith D, Bot M, Paans NP, Visser M, Brouwer I, Penninx BW. The role of obesity measures in the development and persistence of major depressive disorder. J Affect Disord. 2016b;198:222-9.
14. Gibson-Smith D, Bot M, Snijder M, Nicolaou M, Derks EM, Stronks K, Brouwer IA, Visser M, Penninx BWJH. The relation between obesity and depressed mood in a multi-ethnic population. The HELIUS study. Soc Psychiatry Psychiatr Epidemiol 2018b;53:629-38.
15. Gibson-Smith D, Halldorsson TI, Bot M, Brouwer IA, Visser M, Torsdottir I, Birgisdottir BE, Gudnason V, Eiriksdottir G, Launer LJ, Harris TB, Gunnarsdottir I. Children with overweight and obesity and the risk of depression across the lifespan. Manuscript submitted.
16. Grasso AC, Olthof MR, van Dooren C, Roca M, Gili M, Visser M, Cabout M, Bot M, Penninx BWJH, van Grootheest G, Kohls E, Hegerl U, Owens M, Watkins E, Brouwer IA. Effect of food-related behavioral activation therapy on food intake and the environmental impact of the diet: Results from the MooDFOOD prevention trial. Manuscript submitted.
17. Imai CM, Halldorsson TI, Aspelund T, Eiriksdottir G, Launer LJ, Thorsdottir I, Harris TB, Gudnason V, Brouwer IA, Gunnarsdottir I. Associations between proportion of plasma phospholipid fatty acids, depressive symptoms and major depressive disorder. Cross-sectional analyses from the AGES Reykjavik study. J Nutr Health Aging 2017:doi:10.1007/s12603-017-0929-9.
18. Imai CM, Halldorsson TI, Eiriksdottir G, Cotch MF, Steingrimsdottir L, Thorsdottir I, Launer LJ, Harris T, Gudnason V, Gunnarsdottir I. Depression and serum 25-hydroxyvitamin D in older adults living at northern latitudes - AGES-Reykjavik Study. J Nutr Sci 2015;4:e37.
19. Knüppel A, Shipley MJ, Llewellyn CH, Brunner EJ. Sugar intake from sweet food and beverages, common mental disorder and depression: prospective findings from the Whitehall II study. Scientific Reports 2017;7:6287.
20. Knüppel A, Shipley MJ, Llewellyn CH, Brunner EJ. Weight change increases the odds of psychological distress in middle age: bidirectional analyses from the Whitehall II Study. Psychol Med 2018;Nov 20:1-10.
21. Milaneschi Y, Lamers F, Bot M, Drent ML, Penninx BW. Leptin dysregulation is specifically associated with major depression with atypical features: evidence for a mechanism connecting obesity and depression. Biol Psychiatry 2017;81:807-814.
22. Milaneschi Y, Lamers F, Peyrot WJ, Abdellaoui A, Willemsen G, Hottenga JJ, Jansen R, Mbarek H, Dehghan A, Lu C; CHARGE inflammation working group, Boomsma D, Penninx BW. Polygenic dissection of major depression clinical heterogeneity. Mol Psychiatry 2016;21:516-22.
23. Nicolaou M, Colpo M, Vermeulen E, Elstgeest LEM, Cabout M, Gibson-Smith D, Knüppel A, Sini G, Schoenaker DAJM, Mishra GD, Lok A, Penninx BWJH, Bandinelli S, Brunner E, Zwinderman AH, Brouwer IA, Visser M. Association of a priori dietary patterns with depressive symptoms: a harmonized meta-analysis of observational studies. Manuscript submitted.
24. Owens M, Bunce H, Winfield O, Roca M, Brouwer IA, Bot M, Penninx BWJH, Kohls E, Hegerl U, Cabout M, Gili M, van Grootheest G, Visser M, Watkins E. The relationship between healthy and unhealthy habit strength and depressive symptoms in the MooDFOOD Prevention Trial. Manuscript submitted.
25. Paans NP, Bot M, Gibson-Smith D, Van der Does W, Spinhoven P, Brouwer I, Visser M, Penninx BW. The association between personality traits, cognitive reactivity and body mass index is dependent on depressive and/or anxiety status. J Psychosom Res 2016;89:26-31.
26. Paans NPG, Bot M, Brouwer IA, Visser M, Penninx BWJH. Contributions of depression and body mass index to body image. J Psychiatr Res 2018a;103:18-25.
27. Paans NPG, Bot M, Gibson-Smith D, Spinhoven P, Brouwer I, Visser M, Penninx BWJH. Which biopsychosocial variables contribute to more weight gain in depressed persons? Psychiatry Res 2017;254:96-103.
28. Paans NPG, Bot M, van Strien T, Brouwer IA, Visser M, Penninx BWJH. Eating styles in major depressive disorder: Results from a large-scale study. J Psychiatr Res 2018b;97:38-46.
29. Paans NPG, Bot M, Brouwer IA, Visser M, Roca M, Kohls E, Watkins E, Penninx BWJH. The association between depression and eating styles in four European countries: The MooDFOOD prevention study. J Psychosom Res 2018c;108:85-92.
30. Paans NPG, Gibson-Smith D, Bot M, Van Strien T, Brouwer IA, Visser M, Penninx BWJH. Depression and eating styles are independently associated to dietary intake. Appetite 2019;134: 103-10.
31. Pullar J, Cabout M, Brouwer IA, Visser M, Nicolaou M. Disseminating results of the MooDFOOD research project: A case study of the implementing the European Commission’s Communicating Research and Innovation Framework. Manuscript in preparation.
32. Roca M, Kohls E, Gili M, Watkins E, Owens M, Hegerl U, van Grootheest G, Bot M, Cabout M, Brouwer IA, Visser M, Penninx BWJH. Prevention of depression through nutritional strategies in high-risk persons: rationale and design of the MooDFOOD prevention trial. BMC Psychiatry 2016:16;192.
33. Thesing CS, Bot M, Milaneschi Y, Giltay EJ, Penninx BWJH. Omega-3 and omega-6 fatty acid levels in depressive and anxiety disorders. Psychoneuroendocrinology 2018;87:53-62.
34. Van Strien T, Konttinen H, Homberg JR, Engels RCME, Winkens LHH. Emotional eating as a mediator between depression and weight gain. Appetite 2016a;100: 216–24.
35. Van Strien T, Winkens L, Broman Toft M, Pedersen S, Brouwer I, Visser M, Lähteenmäki L. The mediation effect of emotional eating between depression and body mass index in the two European countries Denmark and Spain, Appetite 2016b;105:500-8.
36. Van Strien T, Beijers R, Smeekens S, Winkens LHH. Duration of breastfeeding is associated with emotional eating through its effect on alexithymia in boys, but not girls. Appetite 2019;132:97-105.
37. Vermeulen E, Brouwer IA, Stronks K, Bandinelli S, Ferrucci L, Visser M, Nicolaou M. Inflammatory dietary patterns and depressive symptoms in Italian older adults. Brain Behav Immun 2018a;67:290-8.
38. Vermeulen E, Karien Stronks, Marjolein Visser, Ingeborg A. Brouwer, Marieke B. Snijder, Roel J.T. Mocking, Eske M. Derks, Aart H. Schene, Mary Nicolaou. Dietary patterns derived by reduced rank regression and depressive symptoms in a multi-ethnic population: The HELIUS Study. Eur J Clin Nutr 2017a;71:987-94.
39. Vermeulen E, Knüppel A, Shipley MJ, Brouwer IA, Visser M, Akbaraly T, Brunner EJ, Nicolaou M. High-sugar, high-saturated-fat dietary patterns are not associated with depressive symptoms in middle-aged adults in a prospective study. J Nutr 2018b;148:1598-1604.
40. Vermeulen E, Stronks K, Snijder MB, Schene AH, Lok A, de Vries JH, Visser M, Brouwer IA, Nicolaou M. A combined high-sugar and high-saturated fat dietary pattern is associated with more depressive symptoms in a multi-ethnic population: the HELIUS study. Public Health Nutr 2017b;20:2374-82.
41. Vermeulen E, Stronks K, Visser M, Brouwer IA, Schene AH, Mocking RJ, Colpo M, Bandinelli S, Ferrucci L, Nicolaou M. The association between dietary patterns derived by reduced rank regression and depressive symptoms over time: the Invecchiare in Chianti (InCHIANTI) study. Br J Nutr 2016;115:2145-53.
42. Vermeulen E, Winkens L, Stronks K, Brouwer IA, Visser M, Nicolaou M. The mediating role of BMI on the association between dietary patterns and depressive symptoms. Nutrients, in press.
43. Watkins E, Gili M, Visser M, Brouwer IA, Bot M, Penninx BWJH, Kohls E, Roca M, Hegerl U, Owens M. Feasibility of nutritional strategies to prevent depression. Manuscript in preparation.
44. Winkens LHH, van Strien T, Brouwer IA, Penninx BWJH, Visser M, Lähteenmäki L. Associations of mindful eating domains with depressive symptoms and depression in three European countries. Journal of Affective Disorders 2018a;228:26-32.
45. Winkens LHH, van Strien T, Barrada JR, Brouwer IA, Penninx BWJH, Visser M. The Mindful Eating Behaviour Scale: Development and psychometric properties in a sample of Dutch adults aged 55 years and older. J Acad Nutr Diet 2018b;118:1277-90.
46. Winkens LHH, van Strien T, Brouwer IA, Penninx BWJH, Visser M. Mindful eating and change in depressive symptoms: Mediation by psychological eating styles. Appetite 2019;133:204-11.
47. Winkens LH, Brouwer IA, van Strien T, Penninx BWJH, Elstgeest LEM, Visser M. Associations between mindful eating and change in depressive symptoms: mediation by diet quality and energy intake. Manuscript submitted.
48. Winkens LHH, Lähteenmäki L, Bektas G, van Strien T, Brouwer IA, Visser M. Feasibility of an 8-week online course to increase mindful eating: a randomized-controlled trial. Manuscript in preparation.

Potential Impact:
8 Impact
8.1 Strategic impact
MooDFOOD was executed in the framework of topic KBBE 2013.2.1-01 which called for projects that would reveal the impact of food and nutritional behaviour, lifestyle, and the socioeconomic environment on depression and calls for proposed remedial actions. At the outset of the project it was anticipated that MooDFOOD would provide fundamental knowledge and practical tools that addressed these main objectives of the call.

Update:
MooDFOOD has indeed provided fundamental knowledge and practical tools that reveal the impact of food and nutritional behavior, and the socioeconomic environment on depression. MoodFOOD had the unique opportunity to combine scientific output from six different European cohorts. This resulted in a large number of scientific papers and a high quality and standardized meta-analysis which made clear that health dietary patterns are indeed associated with lower depressive symptoms and a lower risk of developing depression. Furthermore, MooDFOOD performed short feasibility experiments and a multi-country depression prevention trial to investigate the causality of the link between nutrition and depression and the feasibility of the required changes. The main outcome of the prevention trial is published in a high impact journal (Bot et al. JAMA 2019). MooDFOOD also investigated the impact of the advised dietary changes on the environment. Based on the outcomes of these studies and on the state of the art of the field, MooDFOOD developed sustainable nutritional strategies to support mental health. During several integration meetings we received active input from stakeholders which led to the translation of the outcomes to practical tools for several stakeholder groups.

By spreading the tools among the relevant stakeholder groups MooDFOOD will contribute importantly towards improving the diet of all EU citizens in a sustainable way and help to support physical and mental health of EU citizens. This will provide enormous emotional and physical benefits. This will subsequently also lead to economic and social benefits.

The results of MooDFOOD are delivered as scientific publications, publications for professionals, folders and web based materials and different sorts of classical and non-classical dissemination materials that are specifically tailored to different interest groups and stakeholders. We distinguish amongst others:
• Standardized information packages in eleven official major European languages containing nutritional guidelines for EU citizens, mental health professionals, general practitioners and dietitians focusing on the diet itself but also on food-related behaviour and practices
• Policy recommendations for improving food intake and targeting food-related behaviour and practices to change food intake in a sustainable way

All these dissemination materials are evidence-based on the results of MooDFOOD and other European and national projects and provide advice, suggestions and reference to best practices. Stakeholders are addressed as individuals, or small teams of professionals, but also as organizations or their regional/national representatives. MooDFOOD partners that represent patients and health professionals have led the task of developing and disseminating the materials and tools for the general public and health professionals.

The major deliverables from MooDFOOD in relationship to the impact of the call are the following:
Expected impact in call KBBE.2013.2.1-01:
1. Filling existing gaps in the understanding of the link between nutritional aspects like food intake, food composition, nutritional behaviour, conditions such as anorexia or obesity and unipolar depression against the background of changes and trends in food production, lifestyle factors, and wider social determinants.

MooDFOOD deliverables for 1:
- Scientific knowledge on the bi-directional relation of food intake, food patterns, nutrient status, food-related behaviour and obesity with depression, and the underlying psychological, lifestyle, and environmental pathways has been delivered in more than 35 scientific publications in peer reviewed scientific journals, and more scientific publications are expected to follow after the finalization of the project.
- Information on the identification of food-related behaviours that explain the association between food intake and depression, which changes are feasible and result in a healthier diet in non-depressed and depressed consumers has been collected in surveys in Denmark and Spain and comes from small experimental studies on the feasibility of implementing positive food-related behavior and from the MooDFOOD depression prevention trial.
- MooDFOOD developed feasible, effective and environmentally sustainable evidence-based nutritional strategies to support mental health.
- Scientific knowledge on the influence of the social environment, including socioeconomic status, education level, income, age, gender, ethnicity, culture, on the above has been gathered in several European cohorts. The results have been published in peer reviewed scientific journals.

Expected impact in call KBBE.2013.2.1-01:
2. A list of proposed remedial actions and support to guiding policy at EU- and Member State levels, relevant stakeholders and practitioners as well as citizens in dealing with depression and taking preventative measures.

MooDFOOD deliverables for 2:
- Translation of the identified nutritional strategies into practical tools for stakeholders, included:
o Development of treatment guidelines and practical tools in 11 languages for health professionals targeting food intake and food-related behaviour to support mental health in non-depressed and depressed consumers;
o Development of evidence-based nutritional guidelines for EU citizens; as it is clear that multi-nutrient supplement use did not prevent depression in the MooDFOOD prevention trial, EU citizens are not advised to use nutrient supplements for the prevention of depression.
o Development of policy advice and recommendations to promote healthy eating and support mental health for individual Member States and the EU.
- Promotion of the implementation of the practical tools by classical and non-classical dissemination as well as dissemination through large European associations (e.g. EASO, Gamian, EAAD, EFAD) and their extensive networks.

8.2 Academic impact
MooDFOOD promised to contribute to the scientific community in four ways:
1. MooDFOOD will obtain scientific knowledge on the link between diet and depression at much higher level of evidence than most previous studies.
2. MooDFOOD will unravel the causal mechanisms underlying the link between food and depression.
3. Standardized data on dietary intake and food-related behaviour will be collected and added to several large cohort studies.
4. The network of six European cohort studies within MooDFOOD and the proposed research strategy will allow PhD students and postdoctoral students to address their specific research question in different cohort studies.

Update:
First: MooDFOOD has indeed obtained scientific knowledge on the link between diet and depression at much higher level of evidence than most previous studies. The MooDFOOD prevention trial is to date the largest and the first multi-country study to provide extensive scientific evidence on the causal relationship between two different nutritional strategies and the prevention of depression. The study shows that nutrients containing omega-3, selenium, folic acid and vitamin D plus calcium does not prevent depression. This is based on high level evidence (level A of evidence-based evidence). MooDFOOD shows that there is scientific evidence that European citizens who are currently not suffering from a depression should not be advised to take multi-nutrient supplements to prevent depression. The primary analysis of the MooDFOOD prevention trial also did not show any preventive effect of the food-related behavioral activation therapy on the onset of depression. The therapy in the MooDFOOD prevention trial consisted of a behavioural therapy that was focused on changing nutritional behaviour and intake. The study did show that those subjects that regularly attended the sessions did benefit from the food-behavioural therapy.

Second: The rich datasets of the cohorts, combined with the short-term experiments and the repeated measurements of potential intermediate variables within the prevention trial have provided an enormous amount of new insights and knowledge. Though the mechanisms have not been fully unravelled due to the multifactorial causes of depression, the more than 35 manuscripts of the MooDFOOD consortium, published in peer reviewed scientific journals, have importantly extended our knowledge on the mechanisms and pathways underlying the link between food and depression. Data from the observational studies have been collectively presented in a harmonized meta-analysis, which shows the main relationships between nutrition and the onset of depression. The data collected within the framework of MooDFOOD are available for future research and will provide the opportunity to researchers from different disciplines, different sectors, and from all across Europe to use these data to answer research questions that remained unanswered in the current project. Especially the data obtained in the depression prevention trial allow future work on the underlying mechanisms between diet and depression.

Third: in MooDFOOD standardized data on dietary intake and food-related behaviour have been added to two cohorts. This enrichment of the cohorts made it possible to perform a harmonized meta-analysis over all cohorts within MooDFOOD. This was even extended with an Australian cohort. The added data, as well as other cohort data, are available to other researchers though the regular data use procedures of these cohorts. The enriched cohorts can therefore also be used to address future research questions regarding the role of food-related behaviour in the link between food and other nutrition-related diseases and regarding the development of interventions that target food-related behaviour as a way to improve diet quality.

Fourth: Seven PhD students and one post-doctoral student have addressed their specific research questions in different cohort studies within MooDFOOD. The students have travelled to the different partners of the project for short working visits and took advantage of the unique attributes of each cohort, the local scientific expertise and the interaction with each other and with multiple senior experts of MooDFOOD. The young investigators themselves have organized several meetings to share data and experiences. This approach has led to an excellent European research network of young, talented researchers in which research expertise and results have been actively shared. Furthermore, this has led to 7 PhD theses; five of these have already been successfully awarded a PhD degree and two more are expected to follow in the near future. The MooDFOOD PhD students have benefited from the multinational training programs within MooDFOOD, presented their work at (inter)national congresses, and have published in respected peer reviewed scientific journals. Most of the (former) students are currently continuing a career in scientific research.

The dissemination plan of MooDFOOD established links to major scientific conferences. MooDFOOD organized a very well attended symposium during the largest and most important nutrition conference worldwide, namely the conference of the International Union of Nutritional Sciences (IUNS). This conference takes place once in every four years. The 21st International Conference of Nutrition was held in 2017 in Buenos Aires, Argentina. The symposium was led by the MooDFOOD coordinators and several researchers presented their work during the symposium and also during poster sessions. Furthermore, MooDFOOD researchers presented their work during several other relevant conferences, such as meetings of the International Society for Nutritional Psychiatry Research (ISNPR) in Amsterdam, the Netherlands (2015) and Bethesda, USA (2017). Two MooDFOOD symposia were also held at the annual meeting of the international Society of Behavioral Nutrition and Physical Activity in 2016 (Cape Town, South Africa) and during the ECO Congress in Porto, Portugal (2017). In October 2019 an invited keynote lecture on MooDFOOD results will be given by the coordinators at ISNPR in London MooDFOOD researchers have also been invited to present the project during several national and international conferences and scientific meetings. Besides, MooDFOOD has organized specific scientific meetings within the project and with external participants. These actions together have ensured appropriate dissemination of MooDFOOD results within the academic world. For a full list of all scientific MooDFOOD publications see section 7 above.

8.3 Economic impact
At the outset of the project the estimated costs of depression for the European society were high and nutritional strategies were expected to be very cost effective.

Update:
In 2012 Olesen et al. estimated that 30.3 million European citizens were suffering from major depression and the estimated cost per case per year were €3034. The yearly cost of major depression in Europe was €91.9 billion. A figure that is still on the rise (Olesen et al. 2012). Around 26% of these costs are health care costs (hospital, drugs, outpatient visits), where 74% are indirect costs (sick leave, early retirement, and loss of workforce due to early death) (Olesen et al. 2012). The MooDFOOD prevention trial and other studies within and outside the consortium suggest that subjects that adhere to the suggested nutritional therapy may have a lower risk of developing major depression. Preventing depression through nutritional strategies is still likely to be very cost effective. The MooDFOOD prevention trial shows that Intake of nutritional supplements containing omega-3, selenium, folic acid and vitamin D plus calcium does not prevent the onset of depression, and should therefore not be advised.

In 2004 the Dutch RIVM estimated that an unhealthy diet is responsible for 10% of the total yearly mortality. The estimation for overweight was 5% (Van Kreijl et al. 2004). This would mean a reduction in the mean life expectancy of a 40-year old by 1.2 years. In 2018 it was estimated that 8.2 % of the Disability Adjusted Life Years (Daly’s) in the Netherlands can be contributed to an healthy diet. This is only slightly lower than the estimation for smoking. The estimated additional costs for the society caused by unhealthy nutrition in the Netherlands are 6 billion Euro (Volksgezondheid en Toekomstverkenning 2018). It can be expected that other European countries will show similar figures. The nutritional strategies as proposed by MooDFOOD and translated for health professionals and European citizens in tools available on the website (www.moodfood-vu.eu) will contribute to a healthier diet, and thus lengthen life-expectancy and lower health care costs.

In Europe, obesity and overweight are responsible for 1.9% to 4.7% of the total annual health care costs (with regional differences) and for 2.8% of annual hospital costs (Von Lengerke & Krauth, 2011) It is clear that weight problems and obesity are still increasing in most EU countries. Obesity not only increases the risk of chronic diseases, but it is also linked to psychological problems, such as depression. This has enormous consequences on social resources and on healthcare costs (Eurostat, February 2019). Researchers from Germany made clear that in their country overweight and obese people had significantly higher odds of utilization of health care and loss of productivity than people with normal weight. The more obese people were the higher the costs for use of medication and visits to the general practitioner and absence from work. The people in the highest category of obesity had an odds of 2.04 (1.40-2.97) for higher direct costs and an odds of 1.99 (1.20-3.30) for higher indirect when compared to normal weight individuals (Yates et al. 2016). Thus, the prevention of obesity and weight loss in those with obesity will lead to a substantial decrease of health care and related other costs. The MooDFOOD project conclusions recommend weight loss in obese persons to prevent depression and to reduce depressive symptoms in patients with clinical depression. Furthermore, by supporting a heathy diet in the general population to support mental health, MooDFOOD will also benefit support a healthy body weight. The MooDFOOD recommendation are therefore expected to result in decreased costs.

Before the start of the project it was hypothesized that a multi-nutrient supplement strategy could also reduce direct and indirect costs. However, MooDFOOD showed that intake of the multi-nutrient supplement did not provide any beneficial effects. This strategy is therefore not recommended and will not influence direct or indirect costs.

8.4 Impact on industry
MooDFOOD was expected to boost European competitiveness in the field of nutrition and health and to increase innovation among the supplement producing industry. Finally, the food industry could apply knowledge from MooDFOOD to increase the sustainability of their products.

Update:
The MooDFOOD depression prevention trial provided the highest level of evidence possible regarding the causal relationship between the two nutritional strategies and the onset of depression. The trial made clear that the multi-nutrient supplement will not provide any beneficial effect on the onset of depression of depressive symptoms. In the primary analysis the food-related behavioural therapy was also not statistically significantly effective. However, study participants who attended more than eight sessions did benefit from the food-related behavioural therapy. Furthermore, other MooDFOOD studies, including our meta-analysis, show that people who have healthier dietary patterns experience less depressive symptoms. Thus, MooDFOOD shows that following a healthy dietary pattern helps to support mental health. This knowledge provides an additional argument for the European food industry to focus on the production of innovative foods that fit in a healthy dietary pattern. This may not only help to prevent depression, but also depressive symptoms and thereby keep people in better general health.

The MooDFOOD project provides some evidence, but not conclusive, that eating a diet low in (added) sugar may help to prevent depression. People consuming diets that are lower in sugar experience a lower onset of depression than those consuming diets that are higher in sugar. In the general population there is a tendency to be willing to lower the intake of added sugar. Currently, the soda and food industry already started to develop healthier alternatives. The MooDFOOD findings provide, on top of the argument of preventing overweight and obesity, an additional argument for the European soda and food industry to develop novel products with lower levels of added sugar.

Finally, MooDFOOD shows that a healthy dietary pattern does not automatically lead to a more sustainable dietary pattern. However, it is clear that recommendations such as ‘do not eat too much’ and ‘limit your meat intake’ will have positive effects on the environmental foot print. These measures are not at stake with the dietary guidelines that are developed by MooDFOOD. The food industry still has ample opportunities to increase the sustainability of their products and avoiding (food) waste without jeopardizing mental health of European citizens.

8.5 Societal impact
At the start of MooDFOOD the epidemic of depression, unhealthy eating and obesity had enormous societal impact and MooDFOOD would provide evidence-based and practical tools to tackle these challenges.

Update:
The epidemic of depression is still a major problem for health care professionals and health care policy makers. Health care policy makers are insufficiently aware of the possibilities to prevent depression and the role nutrition can have in supporting mental health. MooDFOOD has, based on its own research results and other results that are available from the scientific literature, developed feasible and effective nutritional strategies to support mental health. These guidelines include nutritional guidelines for EU citizens on the whole diet, including tips on healthy food-related behaviour. MooDFOOD also makes clear that there is currently no evidence to advise EU citizens to take additional nutritional supplements to support mental health. These guidelines are translated for policy makers in evidence-based practical tools to be used in the battle against the depression epidemic. The information in the tools provides an additional argument, namely the supportive function of nutrition for the prevention of depression, to follow national dietary guidelines.

Currently, in the Western world depression is already the second leading contributor towards the number of years with disability (Global Burden of Disease Study 2013 Collaborators, 2015). Depression is associated many negative consequences, such as a lower quality of life, stigma and discrimination. It is also associated with loss of work force, sick leave, unemployment and early retirement. Furthermore, it puts a high strain on family members and care takers. For the patient him- or herself it also increases the risk of chronic diseases, such as cancer, diabetes and cardiovascular disease. MooDFOOD study results clearly show that persons with clinical depression or having higher depressive symptoms are at higher risk of subsequent weight gain and eating less healthy, which again might lead to higher risks of chronic diseases. Thus, the nutritional guidelines to support mental health, as provided by MooDFOOD, can be expected to not only have a substantial beneficial impact on the health and well-being of the European citizens that are liable to depression, but they will also have positive impact on family members, the work force and the society as a whole.

It is well-established that an unhealthy nutritional pattern leads to more chronic diseases. However, it is not easy to change people’s dietary intake. Therefore, MooDFOOD also investigated the feasibility of changes in food-related behaviour. Although results still need to be confirmed in future research, the first research outcomes indicate that for example mindful eating may be a viable strategy to improve food-related behaviour and dietary patterns. The guidelines as developed by MooDFOOD provide indications on how to change food-related behaviour and improve dietary patterns. These patterns will not only support mental health, but also lead to longer life with a good quality.

MooDFOOD also determined that a nutritional supplement containing nutrients that were expected to relieve depressive symptoms based on the existing literature was not effective. Intake of a nutritional supplement containing omega-3 fatty acids, selenium, folic acid and vitamin D plus calcium did not effectively prevent depression nor prevent depressive symptoms in the MooDFOOD depression prevention trial. Based on all currently available evidence European citizens should not be advised to take nutritional supplements for the prevention of depression or depressive symptoms. As deficiencies of some nutrients are associated with chronic diseases, some nutritional supplements may be advised to specific population groups to prevent these diseases. National guidelines for the use of nutrient supplements should be followed by these specific population groups.
MooDFOOD has not specifically developed strategies to lose weight, but if the MooDFOOD dietary strategies are followed prevention of overweight and obesity will be supported. The focus of the strategy is on good food-related behavior and a healthy dietary pattern, which in the long-term will positively affect body weight. Small reductions in body weight and the prevention of extreme overweight and obesity will undoubtedly lead to reductions in risks of developing chronic diseases, such as heart disease and diabetes. It will increase quality of life and life expectancy and lead to less health care related costs.

Although in the MooDFOOD prevention trial subjects were clearly steered into the direction of a healthier diet and their diet quality did improve, this did not directly lead to a lower pressure of the diet on the environment. This is likely to be due to the fact that most subjects made healthier choices, but did not change their meat intake considerably, although this was included in the advice. This suggests that more and more specific support is needed to induce a lower meat intake. Furthermore, intake of fish was stimulated although this also puts a high strain on the environment. Thus, to make sure that the strategy not only benefits European’s current health, but also increase the sustainability of the diet and thereby take care of a healthy future of Europe, strategies should also emphasize a lower intake of red meat and a moderate intake of fish.

The evidence-based MooDFOOD dietary guidelines and practical tools for health professionals and European citizens are spread over Europe using the network and high-quality dissemination program of MooDFOOD. This will support mental health of European citizens.

9 Other used references

In order of appearance in the text:
1. World Health Organization. Depression and other common mental disorders: global health estimates. WHO 2017.
2. Alonso J, Angermeyer MC, Bernert S et al. (2004) Prevalence of mental disorders in Europe: results from the European Study of the Epidemiology of Mental Disorders (ESEMeD) project. Acta Psychiatrica Scandinavica 109(Suppl. 420): 21–27.
3. Ferrari AJ, Charlson FJ, Norman RE et al. (2013) Burden of depressive disorders by country, sex, age, and year: findings from the global burden of disease study 2010. PLoS Med 10:e1001547.
4. Mathers CD & Loncar D (2006) Projections of global mortality and burden of disease from 2002 to 2030. PLoS Medicine 3: e442.
5. Sullivan PF, Neale MC & Kendler KS (2000) Genetic epidemiology of major depression: review and meta-analysis. American Journal of Psychiatry 157: 1552–1562.
6. Hoogendijk EO, Deeg DJ, Poppelaars J et al. (2016) The Longitudi-nal Aging Study Amsterdam: cohort update 2016 and major findings. European Journal of Epidemiology 31: 927–945.
7. Penninx BW, Beekman AT, Smit JH et al. (2008) The Netherlands Study of Depression and Anxiety (NESDA) Rationale, objectives and methods. International Journal of Methods in Psychiatric Research 17: 121–140.
8. Stronks K, Snijder MB, Peters RJ et al. (2013) Unravelling the impact of ethnicity on health in Europe: the HELIUS study. BMC Public Health 13: 402.
9. Marmot MG & Brunner EJ (2005) Cohort profile: the Whitehall II study. International Journal of Epidemiology 34: 251–256.
10. Ferrucci L, Bandinelli S, Benvenuti E et al. (2000) Subsystems contributing to the decline in ability to walk. Journal of the American Geriatrics Society 48: 1618–1625.
11. Harris TB, Launer LJ, Eiriksdottir G et al. (2007) Age, Gene/Environment Susceptibility-Reykjavik Study: multidisciplinary applied phenomics. American Journal of Epidemiology 165: 1076–1087.
12. Panagiotakos DB, Pitsavos C, Arvaniti F, Stefanadis C (2007) Adherence to the Mediterranean food pattern predicts the prevalence of hypertension, hypercholesterolemia, diabetes and obesity, among healthy adults, the accuracy of the MedDietScore. Prev Med (Baltim) 44: 335–40.
13. Chiuve SE, Fung TT, Rimm EB et al. (2012) Alternative dietary indices both strongly predict risk of chronic disease. J Nutr 142:1009-18.
14. Fung TT, Chiuve SE, McCullough ML, Rexrode KM, Logroscino G, Hu FB (2008) Adherence to a DASH-style diet and risk of coronary heart disease and stroke in women. Arch Intern Med 168: 713–20.
15. Global Burden of Disease Study 2013 Collaborators (2015) Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study, 2013.
Lancet 386: 743–800.
16. Yates N, Teuner CM, Hunger M et al. (2016) The Economic Burden of Obesity in Germany: Results from the Population Based KORA Studies. Obes Facts 9:397–409.
17. Von Lengerke T, Krauth C. (2011) Economic costs of adults obesity: a review of recent European studies with a focus on sub-group-specific costs. Maturitas 69: 220–9.
18. Olesen J, Gustavsson A, Svenssond M et al. on behalf of the CDBE2010 study group* and the European Brain Council (2011) The economic cost of brain disorders in Europe. European Journal of Neurology 19:155–62.
19. Eurostat. https://ec.europa.eu/eurostat/statistics-explained/index.php/Overweight_and_obesity_-_BMI_statistics February 2019.

List of Websites:
Key Contacts:
Project Website: www.moodfood-vu.eu
Contact MooDFOOD project Office: moodfood.po@vu.nl

Coordinators:
Vrije Universiteit, Amsterdam, Netherlands
Prof. Ingeborg Brouwer PhD - ingeborg.brouwer@vu.nl
Prof. Marjolein Visser PhD - m.visser@vu.nl

Postal address:
Vrije Universiteit
Faculty of Earth and Life Sciences
Department Nutrition and Health
Boelelaan 1085, 1081 HV Amsterdam
The Netherlands

Project information

Grant agreement ID: 613598

Status

Closed project

  • Start date

    1 January 2014

  • End date

    31 December 2018

Funded under:

FP7-KBBE

  • Overall budget:

    € 11 454 337,83

  • EU contribution

    € 8 930 530

Coordinated by:

STICHTING VU