What is the evidence for an association between diet quality and the common mental disorders?

In 2010, in a population-based study of over 1,000 adult women, Jacka et al. [15] reported that women who had a diet characterized by a higher intake of unhealthy western-type foods were more likely to have either major depression or dysthymia, while those whose diets were higher in recognized healthy foods, such as vegetables, lean meats, fish and wholegrains, were less likely to have clinical depressive or anxiety disorders. Similar associations were observed between diet quality and bipolar disorder in the same cohort [16]. Subsequent investigations in a study of more than 5,000 population-based Norwegians demonstrated that adults with better quality diets were less likely to be depressed, while a higher intake of processed and unhealthy foods was associated with increased anxiety [17]. These studies are complemented by prospective studies from other groups: data from the SUN Cohort study in Spain, based on more than 10,000 middle-aged university graduates, demonstrated a prospective inverse association over approximately four and a half years between the level of adherence to a Mediterranean dietary pattern and the risk for incident depression, and a positive association between the consumption of fast foods and commercial baked goods and depression risk [18, 19]. In the Whitehall II cohort study of nearly 3,500 public servants, the authors observed an increased risk for self-reported depression for those on a diet higher in western-style foods, and reduced risk for those on a diet higher in whole foods over approximately five years of follow-up [20]. In 529 women participating in the RHEA mother and baby cohort study, women with better quality diets during pregnancy were at reduced risk for postpartum depression [21]. Numerous recent cross-sectional studies from research groups in Japan [22, 23], China [24], the US [25–28] and elsewhere have supported these findings in adults. It is worth noting that these studies regarding dietary intake are also supported by studies showing that serum folate concentrations, which are markers of dietary quality, predict depression risk over time [29].

Given that adolescence is the age of risk for the onset of these disorders, it is noteworthy that similar data exist in population-based child and adolescent studies: in a 2009 study of more than 1,300 adolescents, Oddy et al. [30] reported that those eating a diet characterized by higher intakes of takeaway foods, red meat and confectionary had increased internalizing and externalizing behaviors, whilst the intake of fruits and vegetables showed the opposite association. In a Chinese study of more than 5,000 young adolescents, healthy dietary habits were associated with reduced psychological symptoms, whilst unhealthy dietary habits had a positive relationship with such symptoms [31]. Similarly, a study of 7,000 young population-based adolescents reported inverse relationships between measures of diet quality and the likelihood of adolescent depression, with higher scores on an unhealthy dietary measure associated with increased symptomatic depression, and higher scores on a healthy dietary measure with decreased symptomatic depression [32]. A subsequent prospective study reported that diet quality was associated with adolescent mental health over a two-year period, even after taking mental health at baseline into account [33]. Moreover, improvements in diet quality were mirrored by improvements in mental health, while reductions in diet quality were associated with declining psychological functioning over the follow-up period. More recently still, a cross-sectional study of German children participating in both the GINIplus and LISAplus cohort studies reported that higher diet quality was associated with fewer emotional symptoms, while increased confectionary consumption had a positive relationship with emotional symptomatology [34].

How strong is the evidence for a causal effect of diet on common mental disorders?

Although the existing observational data are consistent, it is still an unresolved question as to whether the observed associations are a result of causal relationships between diet and mental health, or whether the associations between diet and common mental disorders are due to residual confounding by socioeconomic status or healthy lifestyle behaviors in general, or confounding by unrecognized third factors. Further, reverse causality is not entirely ruled out. The same methodological challenges apply to studies of an eventual effect of physical activity and smoking on mental health. The picture is further complicated by the recognition of the bidirectional relationship between diet and mental health, wherein appetite changes may be a result of mental health symptoms rather than a causative factor. Although the large majority of the previous observational studies have addressed confounding, and the prospective studies have assessed the reverse causality hypothesis, there are obvious limitations to observational studies as regards determining causality. Moreover, apart from one report of no relationship between measures of healthy diet and depression in male, but not female, college students [35] and another study of 865 women reporting no relationship between dietary patterns and postpartum depression [36], there is a dearth of negative studies. There are no published papers examining the issue of publication bias in the diet-common mental disorders literature, thus we cannot exclude publication bias in favor of positive findings.

On the other hand, it appears that the published studies on the diet-mental health association do satisfy most of the Bradford Hill criteria for causality [37]: that of consistency, with concordant findings and effect sizes across cultures, genders and age groups, with multiple methods used to assess diet quality and mental health; biological gradient [15, 18, 19, 21, 24, 32, 33]; temporality [18–21, 33]; biological plausibility [38, 39]; and coherence of the findings with what we already know about the impact of habitual diet on noncommunicable illnesses. However, we only have very preliminary experimental evidence at this time [40].

In contrast to physical activity-depression research, randomized controlled trial evidence for the diet-common mental disorders association is lacking. However, such trials, when conducted, will present myriad methodological problems, as often seen in the field of depression research, such as difficulties with blinding and the problems inherent in conducting randomized controlled trials in patients with depression [41]. Beyond such trials in clinical samples, proving the effectiveness of universal prevention is also problematic, given the substantial statistical power required [42]. However, the reduction in smoking in the general population in many western societies is clearly the result of primary prevention at a population level [43], despite the lack of evidence from randomized controlled trials. Consequently, we would argue that primary prevention holds considerable promise for reducing the burden of depression and anxiety in the population, in concert with selected and indicated prevention, as well as treatment.

Mechanisms

Diet, exercise and smoking may influence mental health through a number of different pathways, including via modification of neurotrophins critical to depression [44, 45]. Diet, smoking and exercise also influence systemic inflammation [46, 47], which is thought to play a role in the genesis of depression [48]. Inflammation is a common pathophysiological mechanism thought to, at least in part, underpin the bidirectional relationships between depression and noncommunicable illnesses such as obesity [49], cardiovascular disease and type 2 diabetes [50]. This suggests immune dysregulation as a key shared pathophysiological pathway by which poor lifestyle practices contribute to both the common mental disorders and the noncommunicable disorders [51].

Prevention: taking a population health approach to the common mental disorders

The 2010 World Health Organization Global Status Report highlights the impact of the changing environmental determinants of the noncommunicable disorders, including poor dietary and physical activity behaviors that have been driven by changes to the global food supply systems, policy, and marketing, as well as increasing urbanization [52]. It outlines and promotes population-wide preventive interventions targeting lifestyle behaviors, primarily aimed at reducing cardiovascular diseases, cancers, diabetes and chronic lung diseases. Although mental health disorders are mentioned as part of the 'broader scope of the noncommunicable conditions,' the prevention of mental health problems using these targeted lifestyle strategies is not specifically addressed. However, we believe that it is both feasible and timely to use our emerging understanding of the impact on mental health of lifestyle behaviors to begin to develop effective, sustainable, population-level prevention initiatives for the common mental disorders. The challenge is how to expand from the nascent epidemiological evidence to develop and evaluate a prevention system for lifestyle-related mental ill health.

A strategy for moving forward

There is agreement that decreasing lifestyle-related risk must address a complex and interconnected mix of etiological factors, from individual behaviors to cultural norms and policy environments [53]. To begin to develop universal prevention approaches for the common mental disorders, it is necessary to link policy options with epidemiology and biological sciences. As such, we would argue that there is a continuing need to generate quality epidemiological evidence to address the identified shortcomings in the literature outlined above. Further research that attempts to disentangle the multitudinous interacting social, medical, behavioral, cultural and biological factors that impact upon both diet quality and mental illness risk is required, while systematic reviews of the diet-mental health observational research, with best evidence syntheses, would also be a valuable and timely contribution to this developing field.

Epidemiological studies should also gather biomarker data that elucidate the role of biological factors, such as oxidative stress, inflammation and neurotrophic mechanisms, as mediators in the relationship between lifestyle factors and the high prevalence mental illnesses. While there are data to show that diet quality, physical activity and mental health are each independently associated with neurotrophin levels and inflammatory and oxidative processes, to date there have been few studies directly addressing mediation via biological mechanisms. Such studies are necessary to either support or refute the case for causal biological relationships and to attempt to explicate the fundamental processes by which lifestyle may influence common mental disorders. There is also a pressing need for evidence from methodologically rigorous randomized controlled trials of dietary improvement as a treatment intervention in those with existing depressive illness. Although not addressing the question of prevention, such evidence - if positive - would be further support for direct causal relationships.

The next strategy advocated is to leverage on existing intervention studies in the field of obesity prevention research to assess the impact of lifestyle improvement on mental health outcomes, particularly in young people. The majority of mental health problems have their onset in childhood and adolescence [54] and this is where the potential for prevention is likely to be greatest. The recent Cochrane Review of effective strategies to prevent obesity in children found strong evidence to support beneficial effects of child obesity prevention programs on weight gain, particularly for programs targeted to children aged six to twelve years [55]. Given that the strategies employed to reduce the prevalence of obesity in children and adolescents overlap substantially with the objective of reducing the prevalence of mental health problems in children and adolescents through lifestyle improvement, the potential to leverage on existing and planned studies in this field and apply the findings to mental health prevention is clear. In this way, mental health research can avoid the failures of previous approaches to obesity prevention and build on the successes. In parallel, economic evaluations of such interventions are necessary to afford comparisons of various prevention models and the projection of the impact and cost-effectiveness of different policy and environmental changes at a population level.

Previous analyses of the costs and benefits of various approaches to public health, including the Assessing Cost-Effectiveness Prevention study [56], have identified that policy approaches to prevention are more cost-effective than health promotion or clinical interventions. As public health policy is the key responsibility of government, government must be at the forefront of prevention initiatives. Suggested priority actions overlap with those for noncommunicable diseases and include policies to improve the food and built environments and greater funding for prevention programs, keeping always in mind the impact on social equity [53].

It is also necessary to incorporate these proposed prevention strategies with attempts to address other risk factors. Childhood abuse and neglect are critical risks, however they are arguably better addressed via the justice and social welfare systems, where active programs are in play. Similarly, the issue of how society should address the regulatory and therapeutic framework of drug and alcohol abuse is of core salience to risk and prevention. Western societies are increasingly nuclear, fragmented, mobile and anomic. The lack of social connectedness is a risk for depression, and similarly merits attention. Thus, policy efforts aimed at prevention, as well as public health messages and educational programs developed in partnership with public health and advocacy groups, will need to integrate these new data regarding lifestyle with those previously understood. From a clinical perspective, support for lifestyle improvement should be routinely provided to all patients with depression and incorporated into treatment guidelines [57].

What sort of impact can we expect to have?

It is important to acknowledge the likely limitations of lifestyle modification in the prevention of the common mental disorders in individuals. In the observational studies, diet and physical activity do not account for a large proportion of the variance in psychological symptomatology. However, there are no identified factors, other than genetics, that account for any more than a small percentage of the variance in depression. Given the size of the burden of illness, the benefits of even a minimal impact on the prevalence of depression will be substantial, as even modest benefits at the individual level accrue to the entire population. Finally, it should be acknowledged that the evidence base for lifestyle as a true causal risk factor for common mental disorders is still developing. However, the benefits to physical health of improvements in lifestyle behaviors are well-established and the precautionary principle provides justification for beginning the development of the outlined programs in the absence of unequivocal evidence.