Article Text

Download PDFPDF

Associations between fruit and vegetable consumption and depressive symptoms: evidence from a national Canadian longitudinal survey
  1. Mila Kingsbury1,
  2. Gabrielle Dupuis1,
  3. Felice Jacka2,
  4. Marie-Hélène Roy-Gagnon1,
  5. Seanna E McMartin1,
  6. Ian Colman1
  1. 1Department of Epidemiology and Community Medicine, University of Ottawa, Ottawa, Ontario, Canada
  2. 2IMPACT Strategic Research Centre, Deakin University, Geelong, Australia
  1. Correspondence to Dr Ian Colman, Department of Epidemiology and Community Medicine, University of Ottawa, 451 Smyth Road, Room 3230C, Ottawa, Ontario K1H 8M5, Canada; icolman{at}uottawa.ca

Abstract

Background Several cross-sectional studies have demonstrated associations between diet quality, including fruit and vegetable consumption, and mental health. However, research examining these associations longitudinally, while accounting for related lifestyle factors (eg, smoking, physical activity) is scarce.

Methods This study used data from the National Population Health Survey (NPHS), a large, national longitudinal survey of Canadians. The sample included 8353 participants aged 18 and older. Every 2 years from 2002/2003 to 2010/2011, participants completed self-reports of daily fruit and vegetable consumption, physical activity, smoking and symptoms of depression and psychological distress. Using generalised estimating equations, we modelled the associations between fruit and vegetable consumption at each timepoint and depression at the next timepoint, adjusting for relevant covariates.

Results Fruit and vegetable consumption at each cycle was inversely associated with next-cycle depression (β=−0.03, 95% CI −0.05 to −0.01, p<0.01) and psychological distress (β=−0.03, 95% CI −0.05 to −0.02, p<0.0001). However, once models were adjusted for other health-related factors, these associations were attenuated (β=−0.01, 95% CI −0.04 to 0.02, p=0.55; β=−0.00, 95% CI −0.03 to 0.02, p=0.78 for models predicting depression and distress, respectively).

Conclusions These findings suggest that relations between fruit and vegetable intake, other health-related behaviours and depression are complex. Behaviours such as smoking and physical activity may have a more important impact on depression than fruit and vegetable intake. Randomised control trials of diet are necessary to disentangle the effects of multiple health behaviours on mental health.

  • DIET
  • MENTAL HEALTH
  • DEPRESSION

Statistics from Altmetric.com

Introduction

Depression is a major cause of disability, affecting more than 350 million people worldwide.1 Dietary patterns are among the behavioural factors associated with risk of depression. In particular, adherence to a diet emphasising whole foods and adequate consumption of fruits and vegetables has been associated with reduced risk of depression2 ,3 while unhealthy ‘western’ diets are associated with an increased likelihood of depression.4–7 Two recent trials have also offered initial support for dietary improvement as a useful preventive intervention for depression8 ,9 although more studies are necessary to confirm these findings.

Fruits and vegetables are nutrient dense foods, high in fibre and rich in essential nutrients such as folate, vitamin B6, magnesium and selenium. Dietary inadequacy of such nutrients has been implicated in the development of depression.10–12 Several cross-sectional studies have demonstrated an association between greater intake of unprocessed fruits and vegetables and lower risk of depression,13–17 and this association persists even when adjusting for potential confounders such as body mass index and total caloric intake.13 ,15

However, fewer studies have examined the longitudinal associations between fruit and vegetable intake and depressive symptoms. One study examining the prospective relationship between fruit and vegetable consumption and depression in older Taiwanese adults reported that vegetable consumption was protective against depressive symptoms 4 years later.13 Other studies have demonstrated prospective associations between broader dietary patterns and depressive symptoms.2 ,6 ,18 ,19

Other lifestyle factors, such as physical activity and smoking status, have also been associated with depression. For example, higher levels of leisure time physical activity have been associated with reduced risk of depression,20 whereas sedentary behaviour has been associated with elevated risk.21 Similarly, smoking has been implicated as an independent risk factor for depression.22 As dietary patterns tend to be associated with these other health-related behaviours,23 it is difficult to disentangle the relations between fruit and vegetable intake, other lifestyle factors and mental health.

Moreover, the relationship between lifestyle factors and depression appears to be bidirectional.24 Depression may lead to unhealthy dietary habits (eg, ‘emotional eating’), including lower intake of fruits and vegetables.25 However, few studies have explicitly examined reverse causation. One longitudinal study assessed the association between healthy diet (including adequate fruit and vegetable consumption and minimal processed, fried or high-sugar foods) and adolescents’ psychological well-being 1 year later.18 Healthy diet predicted better mental health at follow-up, and changes in diet over time were mirrored by changes in psychological well-being. However, results did not support the reverse model—that is, mental health at baseline did not predict diet quality at time 2. Some longitudinal studies in adults have yielded similar findings, with sensitivity analyses not supporting the reverse causality hypothesis.4 ,19 ,26

The purpose of the present study was to examine the associations between fruit and vegetable consumption and depression across multiple timepoints, while taking into account the time-varying nature of the outcome and exposure variables and covariates, as well as autocorrelations across timepoints for each variable. We also sought to test the reverse model: the association between depression at each timepoint and subsequent fruit and vegetable consumption. We also tested associations with psychological distress, a construct comprising non-specific symptoms of depression and anxiety.

Method

Study population

The study used data from the National Population Health Survey (NPHS). The NPHS is a large, national longitudinal survey of Canadians of all ages. Participants were interviewed every 2 years, beginning in 1994/1995, with the most recent data collection in 2010/2011 (cycle 9). Complete information about the survey design and methodology is available from Statistics Canada.27 The present sample consisted of n=8353 individuals aged 18 and over with at least one assessment of fruit/vegetable intake and depression/distress between 2002/2003 (cycle 5) and 2010/2011 (cycle 9).

Exposure and outcome variables

Fruit and vegetable consumption

Fruit and vegetable intake was assessed in the NPHS every 2 years starting in 2002/2003 (cycle 5). The daily frequency of fruit and vegetable consumption was assessed using the following questions, based on the fruit and vegetable module in the Behavioural Risk Factor Surveillance System of the USA Centers for Disease Control and Prevention:28 “Not counting juice, how often do you usually eat fruit?”; “How often do you (usually) eat green salad?”; “How often do you (usually) eat carrots?”; “Not counting carrots, potatoes, or salad, how many servings of other vegetables do you usually eat?”. Total fruit and vegetable intake was assessed by combining the responses to questions regarding daily consumption frequency of fruits (not counting juice), green salad, carrots and other vegetables (excluding potatoes). This provides a score representing how many servings of fruits and vegetables an individual consumed on a daily basis. Juices and potatoes were not included, due to their strong association with unhealthy foods and lack of nutritional value.16

Depression

Major depression in the study cohort was measured by use of the Composite International Diagnostic Interview Short Form (CIDI-SF).29 For each individual, the CIDI-SF produces a predictive probability of a major depressive episode in the past 12 months. An individual with a predictive probability of 90% or higher was considered to have major depression (a score of 5 or higher on a 0–8 scale).29 Depression was treated as a binary outcome variable in the analyses.

Distress

Distress was measured in the NPHS using the Kessler Psychological Distress Scale (K6),30 which assesses the frequency of six non-specific symptoms of anxiety and depression (eg, sadness, hopelessness) during the previous month. Items are rated on a 5-point scale, ranging from 0 (‘none of the time’) to 4 (‘all of the time’). Items are summed to create a total score with a possible range of 0–24, with higher scores indicating greater distress. Distress score was treated as a continuous variable in the analyses.

Covariates

Covariates included participants’ age and gender, history of major depression, education, income, chronic illness, smoking status and physical activity.

History of depression was assessed using the CIDI-SF from cycles 1–4. Individuals with scores of 5 or higher on the 0–8 scale at any one timepoint were considered to have major depression at that timepoint. Individuals who were depressed at any cycle from cycles 1–4 were considered to have a history of major depression.

Participants self-reported their highest level of education at each cycle. For the analyses, education was dichotomised (at least some postsecondary vs none), with no postsecondary as the reference category.

Socioeconomic status (SES) at each cycle was calculated using a ratio of household income to the ‘low-income cut off’ (LICO), which considers an individual's income relative to family size and area of residence.31 For the present analyses, participants were divided into quartiles based on this ratio, with the lowest quartile selected as the reference category.

Chronic illness

Participants self-reported a variety of medical conditions at each cycle. In the present study, individuals with any of the following conditions were considered to have a chronic illness: diabetes, heart disease, hypertension, cancer. These illnesses were chosen due to their association with lifestyle factors.

Binge drinking was defined as reporting at least one occasion of binge drinking (more than 5 alcoholic drinks on one occasion) per month in the past year.

Obesity status was calculated based on self-reported height and weight at each cycle. Participants with body mass index values of 30 or above were considered obese.

Social support was self-reported at each cycle using the Medical Outcomes Study Social Support Scale (MOSSS).32 Participants reported on 19 items assessing functional social support, including informational/emotional support, tangible support, positive social interaction and affection.

Smoking status was self-reported at each cycle. For these analyses, three categories of smoking were compared: ‘daily’ and ‘occasional’ smoking were compared to the reference category, ‘non-smoker’.

Physical activity was self-reported at each cycle. Participants reported on the frequency with which they engaged in various formal and informal physical activities (eg, ‘walking’, ‘aerobics’ and ‘gardening’) during leisure time, as well as the usual duration of each activity. Based on this information, a variable was derived representing each individual's average daily energy expenditure, expressed as kilocalories expended per kilogram of body weight (kcal/kg/day).27

Statistical analysis

Analyses were conducted with SAS software V.9.2 using PROC GENMOD.33 Generalised Estimating Equations (GEE) regression models were used to model the associations between fruit and vegetable consumption at each cycle and depression status (binary) and distress scores (continuous) at the next cycle, while accounting for the intraindividual correlations between repeated measurements of the variables over time. Models testing the inverse associations (ie, depression/distress at one cycle and fruit and vegetable consumption at the next cycle) were also fitted. Models were fitted using an autoregressive correlation structure—that is, assuming that repeated measures closer in time will be more highly correlated than more temporally distant measures. However, coefficients obtained from GEE modelling are unbiased even when the correlation structure is misspecified.34 To account for the complex survey design, we included normalised longitudinal sampling weights in the analysis.

Results

Prevalence of depression and mean distress scores were relatively stable throughout the follow-up period, while fruit and vegetable intake increased over time (table 1).

Table 1

Descriptive statistics (means and percentages are weighted using appropriate longitudinal weights) by cycle, Canada, 2002/2003–2010/2011

Results of GEE models predicting major depression (table 2) indicated that greater fruit and vegetable consumption during a given cycle (5–8) was associated with lower risk of depression at the next cycle (6–9). This association persisted once the model was adjusted for age, gender, history of depression, education, SES and chronic diseases. However, once obesity was included in the model, the association between fruit and vegetable consumption and depression was attenuated. Several covariates (age, sex, history of depression, SES, education, obesity, social support and smoking) were significantly associated with risk of depression.

Table 2

Results of GEE models predicting major depression, Canada, 2002/2003–2010/2011

Similar results were observed in models predicting distress scores. Greater fruit and vegetable consumption during a given cycle was associated with lower distress scores at the next cycle (table 3). Associations between fruit and vegetable consumption and distress remained significant once the model was adjusted for age, gender, history of depression, education, SES, chronic diseases, binge drinking and obesity. However, when social support, smoking and physical activity were added to the model, the association between increased fruit and vegetable consumption and distress was attenuated. Several covariates (age, sex, history of depression, SES, chronic disease, binge drinking, social support, smoking and physical activity) were significantly associated with distress.

Table 3

Results of GEE models predicting continuous distress scores, Canada, 2002/2003–2010/2011

With respect to the inverse associations, after adjusting for age, gender, history of depression, education, SES, chronic disease, binge drinking, and obesity, depression/distress at one cycle (5–8) predicted lower consumption of fruits and vegetables at the next cycle (6–9; tables 4 and 5). These associations were attenuated once social support, smoking and physical activity were added to the model.

Table 4

Results of GEE models predicting fruit and vegetable intake from major depression, Canada, 2002/2003–2010/2011

Table 5

Results of GEE models predicting fruit and vegetable intake from distress scores

Discussion

Recent research suggests that a healthy diet, including adequate consumption of fruits and vegetables, is associated with better mental health.4 ,5 Our findings, from a nationally representative longitudinal study of more than 6000 Canadians, showed that increased intake of fruits and vegetables at any timepoint was associated with decreased distress, and lower risk of depression 2 years later. The associations between fruit and vegetable consumption and indicators of distress and depression remained significant after adjusting for the effects of important confounders, including previous history of depression, sex, education, income, chronic illness and binge drinking. The association with psychological distress remained significant after obesity was taken into account. These findings are consistent with a number of other studies reporting significant inverse associations between fruit and vegetable intake and depression.13–16

However, in this study the associations between fruit and vegetable consumption and indicators of distress and depression were attenuated after adjustment for social support, physical activity and smoking. Several studies of the links between diet and mental health have accounted for the effects of physical activity and smoking.2 However, most of these studies have adjusted for these confounders at a single timepoint. Recent longitudinal research using the NPHS confirms that multiple health behaviours, such as smoking, physical activity and alcohol use, are interrelated, and tend to covary over time.35

In particular, smoking has been associated with lower consumption of fruits and vegetables.35 Psychosocial factors, including personality characteristics and lack of health literacy, have been implicated in this link.36 Nicotine may also affect taste receptors, making some healthy foods less palatable to smokers.37 Smoking is also associated with increased risk of depression22 and increased likelihood of recurrence of depression over time.38 The causal mechanisms underlying this link are less clear, but evidence for neurobiological, genetic and psychosocial factors has been presented.38–40 Smoking may cause oxidative stress,36 leading to increased risk of depression.41 Some evidence suggests that genetic factors associated with dopamine receptor functioning are associated with smoking progression.39 Low endogenous levels of dopamine may make some individuals particularly sensitive to the rewarding effects of nicotine, and also more prone to depressive symptoms.39 ,40

Moreover, individuals who regularly engage in physical activity tend to consume a healthier diet, including more fruits and vegetables.23 ,42 This association may be due in part to psychosocial factors. For example, mastery (a sense of efficacy or control over one's life) has been associated with both greater fruit and vegetable intake and more frequent engagement in physical activity,42 and is also associated with better long-term prognosis of depression.38 Leisure time physical activity is also associated with better mental health, including lower likelihood of depression.20 ,21 ,23 Physical activity may reduce imbalances in the hypothalamic-pituitary-adrenal (HPA) axis, reducing responses to stress21 and has been associated with increased levels of endogenous opioids, as well as several neurotransmitters and neurotrophic factors implicated in depression.21 ,42 Some evidence also suggests that increases in self-efficacy and positive affect following physical activity may garner protective effects,20 and likely act in concert with biological mechanisms.43

The associations between smoking, physical activity, diet and mental health may be a source of confounding in studies assessing associations between fruit and vegetable consumption and depressive symptoms. Much of the previous literature suggests that the association between fruit and vegetable consumption and depression remains significant even when taking other behavioural factors into account.2 ,14 ,17 However, longitudinal studies assessing these factors together are rare. One study demonstrated that the negative association between vegetable intake and depressive symptoms at follow-up persisted when adjusting for physical activity and smoking—but these confounders were assessed only at a single timepoint.17 Consistent with previous research20–23 ,36–40 ,42 ,43 these lifestyle factors showed strong associations with both diet and mental health in the present sample. Once smoking and physical activity were taken into account in our time-sensitive approach, the relationships between fruit and vegetable intake and indicators of mental health were attenuated, suggesting that these other health-related behaviours may have a more important impact on depressive symptoms than the consumption of fruits and vegetables.

Most studies of the associations between diet and mental health have not adjusted for social support.2 Low social support is an independent risk factor for depression,44 and has also been associated with poor diet including low fruit and vegetable consumption.45 Results of the present study supported these findings, showing strong associations between social support, distress, depression and fruit and vegetable consumption. Echoing the findings regarding smoking and physical activity, once social support was included as a covariate, associations between fruit and vegetable intake and depression/distress were attenuated.

Further complicating matters, lifestyle factors seem to show a bidirectional association with mental health,24 ,25 making it difficult to infer causality. Some longitudinal studies of diet and mental health have explicitly tested both forward and reverse models, or conducted sensitivity analyses, and found that the reverse causality hypothesis did not explain the observed relationships between diet and mental health.4 ,19 ,26 For example, a study in adolescents suggested that whereas healthy diet at baseline predicted improved mental health 1 year later, the reverse was not true—improvements in psychological well-being did not predict diet quality at follow-up.18 Similarly, in two large longitudinal cohort studies in adults, depression scores did not predict dietary patterns at a later timepoint.4 ,26 In the present study, the same pattern of results emerged for the forward and reverse models—that is, indicators of mental health at one timepoint significantly predicted fruit and vegetable consumption 2 years later. The inclusion of confounding lifestyle factors (smoking, physical activity, social support) in the model rendered the associations between depression/distress and later fruit and vegetable non-significant.

Taken together, these findings suggest that the relations between fruit and vegetable intake, other health behaviours and depressive symptoms are complex and bidirectional. A healthy diet, including adequate consumption of fruits and vegetables may engender improvements in mental health, possibly due to the impact of essential nutrients such as folate, magnesium, selenium and vitamin b6.5–7 Following a healthy diet may also facilitate engagement in other positive health behaviours, such as exercise and quitting smoking,23 ,34 which may in turn contribute to improved mental health.20–21 ,23 ,38–40 However, symptoms of depression and distress may also make it more difficult to engage in positive health behaviours.24 ,25 In order to disentangle the effects of multiple health behaviours and assess causality, future research should ideally employ randomised control trials of diet; one such trial is currently underway.46

A strength of the current study was the use of a large, nationally-representative longitudinal survey with multiple assessments of depressive symptoms, fruit and vegetable intake and other health-related behaviours over time. Nonetheless, results should be interpreted in light of several limitations. As a longitudinal survey, the NPHS was subject to participant attrition over time. The cumulative attrition rate between 2002/2003 and 2010/2011 (the cycles used in the present study) was 26%,27 which may introduce bias if less healthy Canadians are more likely to drop out over time. According to Statistics Canada, the longitudinal sample is still considered representative of the Canadian population,27 and other research on health behaviours assessed in the NPHS has found few differences in findings for the longitudinal sample versus imputed data.21 The NPHS assessed only leisure time energy expenditure. Although recent research suggests that leisure time physical activity may be more strongly associated with depression than occupational physical activity,20 future researchers may wish to examine physical activity in other domains. Finally, although we were able to adjust for the effects of many potential confounding factors, including obesity, problematic alcohol use, exercise and smoking, there are numerous other factors that may be important to consider. It is likely that other aspects of diet (eg, fish consumption, total energy intake) also contribute to the complex relationships between diet, other lifestyle factors and mental health.

Conclusion

Identifying modifiable behavioural factors that may contribute to mental health is crucial to planning effective interventions. A healthy diet, including consumption of whole fruits and vegetables, is one behavioural factor that has been associated with improvements in mental health. Our results suggest that the longitudinal associations between mental health, fruit and vegetable intake and other lifestyle factors are complex and difficult to disentangle. We suggest that the only way to overcome this inherent complexity is by conducting randomised control trials to assess the unique causal effects of fruit and vegetable consumption on mental health.

What is already known on this subject

Eating a healthy diet, including adequate amounts of fruit and vegetables, is associated with better mental health. However, few studies have examined this association longitudinally, while taking other lifestyle factors into account.

What this study adds

The longitudinal associations between fruit and vegetable consumption and depression may be driven by other lifestyle factors such as social support, physical exercise and smoking. Randomised control trials are needed to disentangle the complex relationships between diet, other lifestyle factors and mental health.

References

View Abstract

Footnotes

  • Contributors IC conceived the study. MK, M-HR-G and SEM designed the analytical plan. MK performed the data analysis and drafted the manuscript. GD performed the literature review. MH-R-G provided statistical consultation. FJ provided expert consultation. All authors commented on the implications of the results and critically reviewed the final manuscript.

  • Funding This research was undertaken, in part, thanks to funding from the Canada Research Chairs programme for IC. The analysis presented in this paper (or in this presentation) was conducted at the COOL RDC, which is part of the Canadian Research Data Centre Network (CRDCN). The services and activities provided by the COOL RDC are made possible by the financial or in-kind support of the SSHRC, the CIHR, the CFI, Statistics Canada, Carleton University, the University of Ottawa and the Université du Québec en Outaouais. The views expressed in this paper do not necessarily represent the CRDCN's or that of its partners.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Data for this study is accessible through Statistics Canada's Research Data Centre programme.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.