The current evidence linking dietary patterns to depression and risk
We have lots of observational and epidemiological studies suggesting a correlation between your dietary patterns and both your risk for developing depression as well as your current depressive symptoms. This blog will talk about the current evidence to date from these studies as well as some of the issues regarding randomized controlled trials to see how much diet quality and patterns really matter when it comes to depression.
Evidence from Observational Studies
Observational studies allow us to determine associations or correlations between one thing and another. In this case, is a “healthy” dietary pattern associated with symptoms and risk for developing depression?
Typically, what is defined as a healthy dietary pattern whether it is Mediterranean, vegetarian, or Tuscan, is this: one that includes plenty of fibrous rich foods including fruits, vegetables, whole grains, nuts and seeds, as well as mostly plant based sources of protein. In addition, it is defined by a limited intake in saturated fats as well as salt and added sugars.
One of the troubles in trying to reconcile this question with a confident answer is that there are many differences between the available studies including having differing methodologies, as well as different definitions of outcomes and measurements and scales used for collecting data, and different sample characteristics. These differences make it difficult to make conclusive statements about the diet and depression relationship. For example, the outcome measurements used to measure and quantify depression are varied and may include self report measures (Beck Depression Inventory or the PHQ-9), or indication of the use of antidepressants, or the presence of a clinical diagnosis by a qualified mental health professional. This matters because the outcome (depression) may be more sensitive to certain analyses than others. Self report measures, while valid, sometimes may only indicate the presence of symptoms which may not reflect a clinical diagnosis of depression, but rather reflect the presence of subclinical symptoms, while a mental health professional would be able to quantify a person’s symptoms from a broader perspective taking into account other variables and important clinical information that would indicate true clinical depression. Overall, the differences that exist between studies are important because when we are looking at results from something like a meta-analysis, we need to keep these contextual differences in mind when trying to extract clinical meaning from the data. Another issue that complicates understanding the diet and depression relationship is that altered eating patterns are associated with depression and it’s onset. To date, the cross sectional studies investigating the diet and depression relationship could not differentiate to what extent diet or dietary behavior is a risk factor, a consequence, or a concomitant phenomenon of depression, leading to further research with different designs to allow us to understand this complicated relationship.
Let’s first take a look at some of the prospective cohort studies that assess dietary pattern and depression risk in healthy adults (not diagnosed with clinical depression) to help us determine if indeed a relationship exists.
One prospective study using the SUN cohort looked to compare adherence to three different dietary score patterns and incidence of depression at 10 year follow up. Specifically, researchers compared the association of three separate diet quality scores (Mediterranean Diet Score, Pro-vegeterian Dietary Pattern and the Alternative Healthy Eating Index-2010) with depression independently of their micro/macronutrients composition to establish the type of relationship between these patterns and the risk of developing a depressive disorder. Dietary intake was assessed at baseline and after 10 years of follow-up with a validated semi-quantitative food-frequency questionnaire. Both the baseline assessment and follow up information was gathered through web based or postal questionnaires collected every two years. With regards to assessing the outcome, incident cases of depression were defined as participants who were free of depression and not using antidepressant treatment at baseline (not using antidepressant medication), who in the follow up questionnaires indicated yes to the following question “Have you ever been diagnosed with depression by a medical doctor?” or who reported the use of antidepressant drugs. The authors report that a statistically significant association exists between adhering to all three healthy dietary patterns and having a reduced risk of depression at 10 year follow up and that these results stayed significant even after controlling for other covariates (sex, employment status, smoking, physical activity, the use of vitamins, as well as prevalence and history of CVD, and other medical conditions). Interestingly, they state that a moderate adherence to the Mediteranean Diet Score at baseline was associated with a reduction in the risk of developing depression during follow up as compared to minimum adherence. And that, participants in the 2nd to 5th quintile of adherence showed a 25-30% relative risk reduction. More specifically, authors “found a suggestion of L-shaped associations, indicating that moving from low to moderate adherence to these diet quality scores should be responsible for a reduction in the risk of depression.” Additionally, there was a threshold effect that was observed, such that no extra benefit was found beyond moderate adherence compared to high or very high adherence versus moderate adherence. Which means that moderate adherence to diet quality scores showed the strongest inverse association with depression compared to low adherence. From this study, the authors conclude that moderate or high adherence to diet quality scores such as the ones investigated in this study, could be effective for the reduction of depression risk.
A more recent 2018 meta-analysis assessing diet quality and depression risk evaluated 24 prospective cohort studies, pooling data spanning over 1,000,000 person years of observations found that the highest category of adherence to a high-quality diet, whether it was conceptualized as healthy/prudent-, Mediterranean-, pro-vegetarian or Tuscan, was associated with lower depression incidence when compared to the lowest category of adherence. And here is why subgroup analyses and exploring the details of the studies are important: Adherence to healthy dietary patterns or healthy food groups was not associated with depression incidence in studies in which a formal diagnosis was used as the outcome and in studies that controlled for depression severity at baseline. This means that the association between high-quality dietary exposures and a lower depression incidence was only present in studies that used symptom severity scales of depression as an outcome. These specific measures only assess the presence of symptoms of depression (I.e. fatigue, restlessness, impaired focus, etc.), not whether the full diagnostic criteria for depression is fulfilled, which would then warrant a clinical diagnosis. This detail is important to understand as it influences our inferences about this relationship. A person may indicate the presence of certain symptoms but they may not indicate the core symptoms of depression on these measures, which may elevate their score but not indicate the true presence of depression.
Another reason this is important is because this means that depression severity at the start of a study significantly influences depression incidence and risk over time INDEPENDENTLY. There’s a lot of clinical significance here because we know that past history of depression and depressive episodes influences future depressive episodes and risk. In fact, lots of factors alone independently affect depression risk, which is why it is so important to try and control for these factors when studies are analyzing data assessing diet and depression risk. Other factors that independently affect depression risk include other psychiatric and medical comorbidities (I.e. anxiety and type 2 diabetes), socioeconomic status, sex, familial mental health history, as well as race and the experience of oppression and trauma.
What about a specific dietary pattern like the Mediterranean style diet and that having an effect on depression risk? A systematic review from 2019 analyzed 7 cohort studies (both cross sectional and cohort methodologies) published between 2013 and 2017 including 3 from Australia, 2 from the U.S. and 1 each from France and Spain, to answer this question. The 7 studies included 50,588 healthy people aged greater than 45 years with 2,550 incident cases of depression during the follow-up period of 3 to 12.6 years. With regards to outcome assessment, 5 studies had used the Center for Epidemiologic Studies Depression Scale,1 had used the Kessler Psychological Distress Scale, and 1 had used physician diagnosis and/or the use of antidepressant drugs. Again, looking at the details of the review is where the clinical significance and importance is. When the authors analyzed the cross sectional data, a significant inverse association was found between adherence to the Mediterranean diet and odds of depression. But when the authors analyzed the cohort studies, no significant association was observed between adherence to the Mediterranean diet and risk of depression. Again, the details here are important. Cross sectional studies are useful in helping us discover a potential relationship but they lack the ability to help us determine causation or help us determine the strength of this relationship. This is why prospective cohort studies are useful and important to assess in the larger context of the existing body of evidence. These findings suggest that when we look at overall depression risk in a healthy population (free of a clinical diagnosis or history of such), adhering to a “healthy” dietary pattern barely reduces depression risk, and if it does, it may only do so in a small way.
Now let’s take a look at some of the randomized controlled trials (RCTs) to assess if dietary patterns and quality affect depression symptoms and risk.
Evidence from Randomized Controlled Trials
Randomized Controlled Trials (RCTs) allow us to determine if one variable causes an effect in another. Now that we have some evidence that tells us that there is a correlation and association between dietary patterns and depression, the question is, do dietary modifications that include a “healthy” pattern cause depressive symptoms to decrease?
A recent meta-analysis reviewed 16 RCT’s that included over 40,000 participants in order to determine the strength or magnitude of the effect that dietary interventions have on depressive symptoms and anxiety. Of the 16 studies, only one study in the meta-analysis examined the effects of a dietary intervention in a sample with a primary diagnosis of clinical depression. The rest of the studies included healthy participants who did not have a diagnosis of depression. The authors report an overall effect size of .28, indicating a small effect size and state that dietary interventions “significantly reduce depressive symptoms in comparison to control conditions” (no change in diet). Specifically, the pooled effect size on depressive symptoms across 10 dietary interventions that compared dietary interventions with habitual diet alone (control group/no change in diet) indicates a small to moderate significant effect. Effect sizes are important because they help us determine the strength between two variables (the larger the effect size, the stronger the relationship). In this case, how strong the relationship is between changing one’s diet and that modifying depressive symptoms. These results indicate that compared to not modifying your diet, a dietary intervention affects depressive symptoms in a small to moderate way (.28). This may or may not translate to clinical significance, meaning, a difference in symptoms of depression that would create a meaningful difference in a person’s life (i.e. sleeping better, feeling more energized, increased focus, etc.). With regards to anxiety, no overall effect of dietary interventions on anxiety compared with control conditions.
This meta-analysis is unique in that the authors further outline other factors influencing the magnitude of effect in dietary intervention trials, which are extremely important to consider. One of these influences included the kind of control group used for a comparator. Whether the dietary intervention was compared to an active (participants received a comparable treatment) vs. an inactive control group (participants received no comparable treatment), influenced the results such that the the effect size for dietary intervention compared to active control was .174 vs. the effect size for dietary intervention compared to active control which was .306. This suggests that a dietary intervention is only beneficial for outcomes of depression and anxiety depending on the kind of control group it is compared to. In addition, whether or not the dietary intervention was delivered by a nutrition professional influenced the results such that the effect size for using a nutrition professional was .329 while the effect size for not using a nutrition professional was .124. This suggests that who delivers the intervention also matters with respect to the results. The authors go on to state that “studies specifying the involvement of a nutritional professional (e.g., dietitians or nutritionists) in the delivery of dietary interventions observed a significant effect on depressive symptoms, whereas those that were delivered without dietitian/nutritionist professional involvement had no greater effects than control conditions.” This is significant because having this additional support allowed for greater adherence to the dietary intervention, leading to greater improvements in depressive scores. For details about this specific study and what we can take away from it, make sure you read it here.
Another recent 2019 RCT investigated whether young adults with elevated depression symptoms would comply with a brief 3-week diet intervention to see if this would improve symptoms of depression and whether compliance to the diet was associated with improvement in depression symptoms. With regards to methodology, researchers enacted a 3-week, parallel group, single blind RCT using participants who were recruited from an undergraduate psychology course and participated for course credit. Participants were eligible if they were aged 17–35, had a score ≥7 on the Depression, Anxiety and Stress Scale-21 Depression subscale (DASS-21-D), and a score > 57 on the Dietary Fat and Sugar Screener (DFS), with scores >57, suggesting a poor diet that at the time did not comply with the Australian Guide to Healthy Eating.
Participants in the Diet Change group received the diet intervention instructions from a registered dietician via a 13-minute video and had access to this video to watch as needed throughout the study duration. Participants were given sample meal plans and recipes, including “a handout answering frequently asked questions and troubleshooting solutions.” With specific regards to troubleshooting, the authors indicate that a particular “focus was given to potential problems such as cost-saving and limited time for food preparation.” They also received a small hamper of food items including olive oil, natural nut butter, nuts and seeds (walnuts, almonds, pepitas, sunflower seeds) and spices (cinnamon, turmeric) and were instructed to keep their shopping receipts in order to receive a $60 gift card as reimbursement for study foods. Additionally, participants in the diet change group received a brief, 5-minute phone call on Day 7 and Day 14 to trouble shoot any difficulties regarding diet adherence. Participants in the habitual diet group were given no instructions regarding diet, and were asked to return after 3 weeks for follow up.
As far as outcomes, the Centre for Epidemiological Studies Depression scale-Revised (CESD-R) and the Depression Anxiety Stress Scale-21 (DASS-21) were used to assess depression. An important note about these two validated scales: The CESD-R is able to detect at risk individuals for depression with a score of 16 or greater indicating that an individual is at risk for developing clinical depression while the DASS-21 is able to detect and quantify the core symptoms of depression and anxiety with scores of 7 plus indicating at least moderate depression and anxiety. Here’s what the authors found: “significant reductions in depressive symptoms on the primary outcome measure, the CESD-R, for those in the diet change group compared to the habitual diet control group, with a moderate effect size of 0.65. Similarly, significant improvements were observed for the secondary outcome measure of depressive symptoms, the DASS-21 depression subscale, with a moderate effect size of 0.75.” Sounds pretty good, right? Let’s take a closer look at some of the raw scores to help us determine if there was a clinically significant difference pre to post treatment.
Level of severity on the DASS-21 depression subscale at baseline for those in the diet change and habitual diet groups.
Severity labels: Normal(0–4) Mild(5–6) Moderate(7–10) Severe(11–13) Extremely Severe (14+) Diet Change Group n 12 9 5 7 5
% 31.6% 23.7% 13.2% 18.4% 13.2%
Habitual Diet Group n 16 8 4 5 6 % % 41.0% 20.5% 10.3% 12.8% 15.4%
Let’s take a look at these raw scores and how severity of depression is distributed in these two groups. You can see that the majority of participants in both the intervention group and the control group were in the “normal” and “mild” range for depression. This is important because this does not represent typical scores of a more “clinically depressed” population. If participants are already starting with subclinical or mild depression scores, there’s only so much room for “improvement.” This is something to consider when we are interpreting the results.
If we look at the secondary measure which is the DASS-21, the mean baseline score in the diet change group is 7.18 and the mean post treatment score is 4.37. There are a few things I want to highlight here using this as an example. As mentioned previously, the DASS-21 is able to demonstrate whether any of these identified issues are having a significant effect on the person’s life at present. That’s one thing to keep in mind. The next is the scoring system and interpretation of scores for the DASS-21. Here is the breakdown of scores with regards to how they get interpreted: 0-9 indicates “normal”, while 10-13 indicates “mild”, 14-20 indicates “moderate”, and 21-27 indicates “severe” depression state. The 3 point decrease in scores pre to post treatment don’t necessarily reflect a clinically significant difference with regards to the state of depression. This is an example of how statistical significance doesn’t necessarily reflect clinical significance. The same here goes for their primary measure, the CESD-R. The mean baseline score of the diet change group for the CESD-R was 20.56 while the scores post treatment was 14.62. This is important here too as a score equal to or above 16 indicates a person at risk for clinical depression (subthreshold for clinical depression). This 6 point difference may be slightly more reflective of a possible clinically significant difference when it comes to the state of depression.
Here are some additional issues and confounding variables that I think are very much at play here when it comes to RCT’s and the diet change intervention groups. The intervention groups are getting a ton of support usually from a designated dietician that not only provides them with education and instruction on how to change their diet, but also, helps them troubleshoot any issues related to adherence to a diet. This within itself is an intervention (other than changing one’s diet). And this is not being adequately accounted for when it comes to analyzing the data. Take the 2019 RCT discussed above. Removing significant barriers that typically interfere with an at risk depressed person’s ability to engage in adaptive behaviors, like giving participants recipes for low cost meals that take limited time to prepare, I think as an intervention within itself is significant. Tailoring the diet to fit the financial means of the individual effectively removes a major social determinant (SES) obstacle for individuals who typically struggle with clinical depression since depression rates tends to be more prevalent in lower SES populations. Additionally, the researchers considered the time it would take to prepare food which could address a potentially significant clinical symptoms of depression which is withdrawn behavior and low motivation to engage in adaptive behaviors. Providing at risk depressed individuals with short meals they can prepare that are low cost can drastically increase motivation to cook these meals and adhere to it.
Beyond these factors, there’s lots of additional follow up in the intervention group as well in these RCTs, again indicating that there’s a lot of social support and careful problem solving that is needed when it comes to helping people change and adhere to a specific dietary intervention. I think it’s more likely that any statistical significance (and therefore possible clinical significance), is coming more from this element of intentional support which effectively is helping someone engage in a positive behavior (which within itself helps to reduce depression). And we know this very well because of the behavioral intervention for depression called Behavioral Activation. It’s a behavioral strategy that increases a person’s engagement in positive and pleasant activities, thus helping to increase reinforcement, while decreasing punishing behaviors like social isolation, which tends to worsen depression. Behavioral activation independently has really strong effect sizes when we look at this as an independent intervention for depression. And so I ask: what is the difference between behavioral activation as an intervention itself and using a dietician to provide multiple contacts of support to help someone modify their diet?
If you’re interested in these additional details, make sure to check out this blog post for more.
Conclusion
As far as the research is concerned, there’s a small to moderate association between dietary patterns and depression risk and depressive symptoms. And even within this small to moderate effect, context here is important. It seems that if you are someone who experiences mild or subclinical symptoms of depression, lacks a significant history of mental illness and a familial history of mental illness, doesn’t have any other physical or medical comorbidities, potentially modifying your diet in a way that adheres to a “healthy” pattern, may help. And to understand the nuance even further, it seems that modifying your diet may have a greater effect on your mental health in general if your baseline diet is already devoid of nutrient dense foods. Going from a poor dietary pattern to a moderate adherence of a “healthy” one, seems to produce greater effects on depression (given the other individual context and circumstances).
Remember, diet is a tool that you can try and modify to help support your mental health, just like other lifestyle interventions and evidence based behavioral practices. And if it is having a positive effect, what matters most seems to be your overall dietary pattern, not individual nutrients or foods.
Depression is extremely complex and there are many different etiologies and causes of depression. Not everyone’s depression will respond to dietary modifications alone. In fact, a more comprehensive treatment plan using many different forms of evidence based psychological interventions is likely best practice with the addition of using lifestyle interventions as an adjunct. Make sure you seek a professional mental health provider if you are struggling with feelings of depression.
Resources
I have an E-Book that includes meals and recipe ideas for when you are struggling with your mental health. It’s your resource for when your overloaded and need quick nutrient dense meals to get you through it. Grab it here.
