Psychiatry has, to date, lacked a clear evidence base regarding modifiable risk factors for depression and, as a consequence, has been unable to adequately address primary prevention strategies. The discipline has thus focused its attentions on the treatment of established disorders and neglected preventive public health approaches targeting the role of modifiable risk factors. Yet the imperative to increase efforts to prevent depression at a population level is clear.
Although unipolar depression is not officially categorised as a non-communicable disease, the World Health Organization has identified it as the illness accounting for the largest burden of disease in middle- and high-income countries, exceeding that of ischaemic heart disease. However, although lifestyle determinants of other non-communicable diseases have been recognised for decades, surprisingly, the link between environment and lifestyle and risk for psychopathology is a very recent discovery (e.g. Akbaraly et al, Reference Akbaraly, Brunner, Ferrie, Marmot, Kivimaki and Singh-Manoux1 Pasco et al, Reference Pasco, Williams, Jacka, Ng, Henry and Nicholson2 Lucas et al Reference Lucas, Mekary, Pan, Mirzaei, O'Reilly and Willett3 ).
Vitamin D and depression
One potentially influential factor is 25-hydroxyvitamin D (25(OH)D). There is now a substantial body of evidence linking low levels of serum 25(OH)D to depressive symptoms, paralleling the other health risks from low vitamin D status, such as osteoporosis and cancer. Given that vitamin D is the most prevalent deficiency in developed societies, the report by Kjærgaard and colleagues in this issue of the Journal is of considerable public health importance. Reference Kjærgaard, Waterloo, Wang, Almås, Figenschau and Hutchinson4 The study had two aims: first, to examine the relationship between those with low or high serum 25(OH)D levels regarding depressive symptoms; and second, to study, in the subgroup of individuals with low levels of 25(OH)D, the impact of supplementation with 40 000 IU vitamin D3/week for 6 months on depressive symptoms. The study succeeded in demonstrating that individuals with low baseline 25(OH)D levels had higher depression scores than those with higher 25(OH)D levels. However, supplementation was of no benefit in addressing depressive symptoms.
Vitamin D has many functions that overlap with the known pathophysiology of depression, which supports the plausibility of a causal role. Vitamin D 25-hydroxylase and 25D-1α-hydroxylase are expressed in brain areas such as the hypothalamus, cerebellum, substantia nigra and retina. Vitamin D has a role in sleep and circadian rhythms, and circadian disruption is well documented in depression. Reference Turek5 In animal models there is cross-talk between glucocorticoids and vitamin D in the hippocampus; Reference Obradovic, Gronemeyer, Lutz and Rein6 dysregulated glucocorticoid signalling is also a core component of depression. Vitamin D also influences neuronal growth, cell proliferation in the developing brain and embryogenesis. Reference Eyles, Burne and McGrath7
Evaluating nutritional treatments
Well-designed negative studies are important to publish, as they may put to bed hot clinical questions. Since three studies have found a significant effect of high-dose vitamin D supplementation on depressive symptoms, Reference Lansdowne and Provost8–Reference Vieth, Kimball, Hu and Walfish10 while two others were negative, Reference Kjærgaard, Waterloo, Wang, Almås, Figenschau and Hutchinson4,Reference Sanders, Stuart, Williamson, Jacka, Dodd and Nicholson11 further studies are needed to establish whether there is a causal relation between vitamin D status and symptoms of depression, what confounding influence physical symptoms might have and, finally, to determine whether correcting vitamin D status makes any difference. There are substantive differences between treatment and pathophysiology of symptoms and threshold clinical disorders, and this is likely to hold true for vitamin D as well; biological treatments tend to be more useful in threshold disorders and studies will need to examine interventions independently in each group.
Depression therapy is a methodologically treacherous sea to navigate, and Kjærgaard et al's trial illuminates many of the difficulties inherent in the quest. Addressing a single risk factor, which individually accounts for a small percentage of the variance of a complex disorder with many risk factors, allows for a limited scope for change. The use of antidepressants may be a marker of illness acuity; excluding those with prior treatment may leave only a mildly ill group who, based on our understanding of placebo effects in other clinical trials, would bias the sample to have a high placebo response rate and a low likelihood of response to active therapy. That this study examined a non-clinical sample, where treatment effects are less likely, may have reduced the likelihood of success. Last, a cut-off 25(OH)D level of 55 nmol/l is quite high, and may not reflect overt clinical deficiency. A signal, if there is any, may lie in individuals with more severe deficiency. As an example, a recent intervention study of folic acid and vitamin B12 for the prevention of late-life depression excluded those with severe depression and/or clinically significant vitamin deficiencies on ethical grounds. Reference Walker, Mackinnon, Batterham, Jorm, Hickie and McCarthy12 Although understandable, this is methodologically problematic, as one would not expect to see an impact of vitamin supplementation on those who were not deficient to start with.
The difficulties in evaluating the impact of vitamin D supplementation on depression apply to all nutritional supplementation in a clinical context. An important consideration when inferring causality in any research examining nutritional status against disease outcomes are the metabolic and behavioural effects of acute illness; in individuals with mental illness, nutrient deficiencies can be caused by the disorder itself. Clearly appetite and self-care are altered in a depressed state. Germane to vitamin D, illness can reduce physical activity and hence sun exposure. Reference Bauer, Glenn, Alda, Andreassen, Ardau and Bellivier13 Moreover, stress and acute illness result in significant alterations in nutrient homeostasis, and reduced concentrations of particular nutrients in serum or tissue are observed in those with acute depressive illnesses in the absence of differences in dietary consumption.
In adults with acute illnesses, such as mood disorders, there is a decrease in concentrations of vitamin E, C and A, retinol-binding protein, total lipids, pyridoxal-5′-phosphate and albumin, alongside a simultaneous increase in C-reactive protein levels. The inflammatory response is accompanied by an increase in oxidative stress, due to either increased free radical production or inefficient antioxidant systems, which in turn leads to increased lipid peroxidation. Increased oxidative stress is a pathway by which levels of lipids in membranes may be reduced in the presence of depressive illness, amplifying any pre-existing deficiency. This may very well account for the consistent reports of omega-3 fatty acid deficiency in individuals with acute depressive illness. This is supported by the evidence that omega-3 fatty acids are only useful in those with severe depression, where levels are likely to be reduced as a result of illness, and are of no evident protective utility in the general population. Decreases in serum zinc and folate are also seen in patients with major depression, potentially being secondary to the inflammatory response. Indeed, some studies have reported that the resolution of a depressive episode is accompanied by an increase in serum zinc, in the absence of supplementation. What remains to be established is the efficacy of manipulation of the whole diet, as opposed to supplementation of dietary insufficiency. Although there is some, albeit weak, evidence for folate and omega-3 supplementation in depression, the use of other supplements remains unclear. The relative benefits of supplements as opposed to sun exposure are not established.
Lifestyle and depression
Although many somatic illnesses increase the risk for depression by increasing pain and disability, it is now clear that there are also shared pathophysiological and behavioural risk factors that directly contribute to both somatic and depressive illnesses, including inflammation and oxidative stress. Reference Pasco, Nicholson, Williams, Jacka, Henry and Kotowicz14 Lifestyle plays an important role in determining levels of inflammation; consumption of a Mediterranean-style diet, rich in antioxidants, vitamins, minerals and fibre, is associated with reduced systemic inflammation, whereas unhealthy dietary patterns, now common in both low- and middle-income countries and high-income countries, are associated with increased systemic inflammation. The synergistic interactions between the multitudinous components of diet mean that focusing only on individual nutrients may obscure the true associations between diet and mental health. Thus, an examination of the whole diet has greater utility in examining the nutrition–mental health connection. Reference Akbaraly, Brunner, Ferrie, Marmot, Kivimaki and Singh-Manoux1,Reference Jacka, Pasco, Mykletun, Williams, Hodge and O'Reilly15,Reference Sanchez-Villegas, Delgado-Rodriguez, Alonso, Schlatter, Lahortiga and Majem16 Similarly, physical activity is associated with reduced markers of systemic inflammation and may have direct anti-inflammatory effects, while smoking potently increases inflammation and oxidative stress. Another consequence of poor lifestyle practices is obesity, which is a pro-inflammatory state. Obesity and depression share a bidirectional relationship, with obesity potentially contributing to depression via increasing the level of circulating pro-inflammatory cytokines, and depression predisposing to the accumulation of excess adipose tissue. Thus, many of the same pathways whereby poor lifestyle practices contribute to the high-prevalence non-communicable somatic illnesses also appear to influence the risk and progression of common psychiatric illnesses.
While not discounting the possibility of other shared factors, such as genetics, this rapidly developing evidence base suggests that mental and physical illnesses and lifestyle form a triad, with lifestyle as a common denominator. This opens the door to common and integrated treatment approaches based on these shared pathways. It also affords the potential for novel preventive and treatment strategies for these disorders that are the major contributor to the global burden of disease. Given the success of prevention strategies in reducing cardiovascular disease in many countries, this is the next threshold for mental health. Suggested priority actions overlap with those for non-communicable diseases and include government policies to improve the food and built environments and greater funding for prevention programmes.