Overall, there is evidence that certain dietary patterns may influence the development and progression of anxiety disorders. The diets associated with lower anxiety include “healthy” diet patterns, the Mediterranean diet, traditional diets, the anti-inflammatory diet, and diets with increased variety. All of these diet patterns share common elements such as an emphasis on vegetables, fruit, limited sugar and refined grains, and greater consumption of minimally processed foods. However, interpretation of dietary patterns studies is somewhat hindered by the dissimilar definitions used for dietary patterns. For example, in the studies delivering “Western” style high fat/high sugar diets meant to induce obesity, a variety of dietary fats were used. The fats used to supplement some of the high fat diets included lard [67], fish oil [68], soybean oil [69] while other studies specified the percentage fat in the diet but not the type of fat that was added to achieve this amount [70,71,72]. Because the impact of different fatty acids on health outcomes can be highly different [73], categorizing diets as low or high in fat may result in heterogenous findings as a result of the type of fat delivered. Similarly, definitions of “healthy” diets have changed over time [74], potentially contributing to the heterogeneity of study results. Unfortunately, many studies lacked clear definitions of “healthy” or “unhealthy” diet patterns and interventions.

The outcomes associated with the vegetarian or vegan diet were generally positive although somewhat mixed and limited by being largely observational in design. The mixed findings may be due to a variety of factors. There is documentation of the adoption of a vegetarian or vegan diet following the development of an eating disorder [75]. Given the association between eating disorders and anxiety disorders [76], this may explain the association between vegetarianism and higher anxiety found in some observational studies. Furthermore, vegan diets, without adequate supplementation, may lack certain essential nutrients shown to play a role in anxiety disorders such as vitamin B12 [77] and long chain omega-3 fatty acids (EPA and DHA) [78]. Bioavailability of certain nutrients such as iron differ between plant and animal sources, possibly limiting nutrient absorption in vegan diets.

There is a significant lack of human intervention studies involving participants with anxiety disorders or elevated baseline anxiety symptoms. Many studies employed dietary patterns that were indicated for the other medical concerns of the study participants. For example, studies delivered the Low FODMAP diet to participants with irritable bowel syndrome and hypocaloric diets to participants with obesity and assessed changes in anxiety as secondary outcomes. The mechanisms by which dietary patterns impact anxiety symptoms may be the result of a combination of the mechanistic factors discussed in the following sections.

The findings of the carbohydrate studies suggest that high intake of sugar and refined carbohydrates may contribute to anxiety symptoms; however, a large proportion of trials are cross-sectional in design, preventing conclusions about causation. There is a need for intervention studies that assess the impact of differing levels of carbohydrate intake in participants with anxiety disorders.

With respect to mechanism, there is evidence that healthy blood sugar regulation is an important factor in mental wellbeing [79]. This relationship may explain the associations seen in the present review between factors that improve blood sugar regulation and lower levels of anxiety symptoms. These include lower intake of sugar and refined carbohydrates, higher fiber intake, regular meals, and caloric restriction.

The evidence related to the role of protein in anxiety symptoms is preliminary. There is some evidence suggesting that adequate dietary protein and, in particular, adequate tryptophan, may be important in improving anxiety symptoms. Amino acids serve as the building blocks for neurotransmitter synthesis, with tryptophan needed for the production of serotonin [80]. The established role of serotonin in the pathogenesis of anxiety disorders [81] may explain the potential harm associated with inadequate dietary protein and tryptophan. This evidence is strengthened by the involvement of many participants with diagnosed anxiety disorders in the intervention studies included in the present review. The human experimental studies used doses of tryptophan ranging from 250 mg per day from a food source (squash seeds) to 3 g per day as a supplement. Although these doses are considered to be below the level associated with side-effects [82], the trial that administered 3 g per day reported side effects such as itching, nausea and urinary changes [83]. Tryptophan supplements should not be used in combination with serotonergic medications such as SSRI/SNRI due to the possible risk of precipitating serotonin syndrome [84]. Food sources of tryptophan include egg, soy, seeds, fish and meat [85].

Overall, there is significant animal and human evidence that adequate or supplemental omega-3 fatty acids may have anti-anxiety effects. There is early evidence, predominantly from animal studies, that diets high in total fat, cholesterol, or trans fat may have an anxiogenic effect. With respect to a possible mechanism, there is evidence that inflammation plays an important role in the pathogenesis of psychiatric disorders [86], including anxiety [87], and that dietary fats can influence levels of inflammation [88]. Through their effects on enzyme pathways involved in the production of anti-inflammatory cytokines, omega-3 fatty acids contribute to lower levels of inflammation [89]. Conversely, omega-6 fatty acids increase levels of inflammation through increased pro-inflammatory cytokine production. Additionally, there is evidence that omega-3 fatty acids impact oxidative stress [90], neurotransmission [91], and neuroplasticity [92], which are known or hypothesized mechanisms for their use in the treatment of anxiety disorders [93,94]. Dietary omega-3 sources include fish and seafood, as well as flax seeds, chia and hemp seeds.

One somewhat inconsistent finding that has been observed in the present review is the impact of high fat diets. A large number of animal studies (39 of 63) reported a worsening of anxiety symptoms in response to intake of a high-fat diet. In contrast, the studies assessing the ketogenic diet, a diet that is very low in carbohydrates and generally high in fat content, suggest a possible therapeutic benefit. While these findings may be considered conflicting, it is speculated that the type of dietary fats used may have differed, with the high-fat diet delivered to the animals being composed of more omega-6, saturated and trans fatty acids. As such, the type of dietary fat, may be a significant factor in addition to the quantity of fat consumed.

There is significant animal data suggesting an anxiolytic effect of several vitamins and minerals as well as supplemental formulas which deliver a combination or broad range of micronutrients. Given the presence of micronutrients in whole, unprocessed foods such as vegetables, fruit, and whole grains, these findings add evidence to the importance of eating a healthy diet containing a variety of unprocessed foods. Intake of foods that provide a rich source of zinc (oysters, crustaceans, meat, organ meat, leafy and root vegetables [95]), and selenium (Brazil nuts, seafood, meat, beans, and lentils [85]) could be prioritized.

Micronutrients such as zinc and selenium are necessary as coenzymes in the synthesis and regulation of neurotransmitters and neurotrophic factors [96] which may explain their importance in maintaining mental wellbeing. Additionally, B vitamins and folic acid contribute to the methylation balance which is hypothesized to be relevant to the pathophysiology of psychiatric illnesses [97].

There is fairly consistent evidence that vegetables, fruit, and plant constituents may exert anti-anxiety actions; however, the majority of the evidence comes from animal studies. Caffeine on its own or added to energy drinks appears to be associated with increased anxiety. Whole foods containing caffeine such as coffee, teas and cacao may have beneficial or equivocal impacts on anxiety, likely due to the co-occurrence of caffeine with other beneficial phytochemicals. Vegetables and fruit contribute to lower levels of inflammation and oxidative stress through their phytochemical and antioxidant constituents [98].

The body of evidence related to the connection between food allergies and anxiety symptoms is limited and the majority of the evidence pertains to the presence of elevated anxiety symptoms among individuals with celiac disease and the anti-anxiety effects of implementing a gluten-free diet in this population. The presence of neuropsychiatric symptoms in celiac disease is established, with hypothesized mechanisms including micronutrient deficiency due to malabsorption and hyperhomocysteinemia [99]; however, it is unclear how the findings of these studies may apply to anxious individuals unaffected by celiac disease.

Preliminary evidence suggests that the intake of beneficial microorganisms and prebiotic fiber may be beneficial in the treatment of anxiety. Habitual diet strongly influences the composition of the gut microbiome, thus adding more rationale for the inclusion of fruits, vegetables, fiber, and fermented foods in the diet [100]. Potential mechanisms for the impact of the microbiome on psychiatric wellbeing include the modulation of the production of gut peptides involved in the gut-brain axis [101] and neurotransmitter synthesis [102].

Strengths of the present review include an extremely rigorous search strategy intended to capture the full range of publications presenting data on this topic. A priori inclusion criteria and duplicate screening decreased the risk of bias. Completion of the project by an interdisciplinary team including clinicians and researchers contributed a range of perspectives and expertise.

The very large scope of this review was both a strength and limitation. Due to the very large volume of articles included in the review, in-depth analysis of individual articles was not possible. The results of the review may include over-simplification of the findings and a lack of attention to evaluating study quality, assessing study or publication bias, or providing contextual information (e.g., dose). In our data extraction and analysis we did not evaluate of the methods used for assessing participant anxiety symptoms or disorders, Anxiety symptoms can be assessed through a variety of methods including clinician- or self-administered questionnaires, or interviews which may utilize a range of diagnostic criteria. These different methods differ in their reliability as well as the exact nature of the symptoms or disorders that they assess. As a result, the studies included in this review report on relationships between food and a heterogeneous group of outcomes including the presence or absence of different anxiety disorders and a range of anxiety symptoms. The decision to include this heterogenous collection of research was an effort to capture a broad range of data related to this topic.

Another limitation of the present review is the unclear relevance of experimental studies which assessed the impact of high dose supplements of a dietary constituent. Doses of some of the nutrients delivered in trials as dietary supplement, such as zinc and omega-3 fatty acids, can be achieved through dietary modification; however, some of the nutrients, such as vitamin B6 (50 mg/day) were delivered in doses that cannot be achieved with food alone.

Another limitation was the exclusion of studies that failed to report changes in anxiety separately from other outcomes. Several studies that were not included in the present analysis reported ‘psychological distress’ as a composite of anxiety and depression symptoms but did not report anxiety results alone [103,104]. These studies were excluded from the present analysis as the purpose of this project was to identify research reporting anxiety outcomes specifically; however, it is noted that this resulted in the exclusion of a number of articles (n = 22).

The ability to draw conclusions from the data is also limited by a number of factors related to the methodology used in the included studies. This scoping review included a large number of animal studies which may have unclear applicability to humans. There are well established tests designed to measure changes in anxiety levels in animals through monitoring their behavior in a variety of experimental settings [105]; however, the applicability of these results to the human experience of anxiety is inherently limited. The benefits of animal research include the ability to manipulate dietary factors in a highly controlled environment, the ability to observe effects rapidly as a result of the animals’ reduced lifespan and the ability to withhold potentially beneficial nutrients. There were also a large number of observational studies, mostly cross-sectional in nature. This type of study cannot draw conclusions about causality. The association between diet and mental health is known to be highly complex and bidirectional. While there is robust evidence that dietary patterns impact the likelihood of developing mental illness [14], there is also evidence that mental illness impacts eating behaviors [106]. This occurs through changes in motivation and appetite that can results from mental illness [107] and metabolic changes, increased appetite and cravings, and gastrointestinal distress [108] that can occur as a results of psychiatric medications [109]. Additionally, confounding factors such as eating disorders may be responsible for associations that are present. Given this bidirectional relationship, the findings of cross-sectional studies have limited ability to answer the question of how food impacts anxiety. While a small number of prospective observational studies were identified in the present review, additional prospective studies are needed in order to accurately assess the impact of dietary patterns on the development of mental disorders, particularly the avoidance of potentially beneficial foods and increased intake of potentially harmful foods, which cannot be studied using an experimental design for ethical reasons.

Another important consideration when interpreting the study findings is the potential for difference between short- and long-term impacts of food on anxiety symptoms. As previously mentioned, it is known that the relationship between mental health symptoms and diet choices is bidirectional; emotional symptoms may drive eating behavior because of their immediate effects on the mitigation of emotional symptoms. The term “comfort eating” has been used to describe the phenomenon where individuals consume foods, especially those higher in calories, sugar, and fat, in response to negative affect [110]. Evidence from mechanistic studies suggest that corticosterone, a stress hormone, positively influences an animal’s intake of a sweet beverage [111] and that consumption of comfort food decreases mRNA production of hormones related to the stress response in animals [110]. It has been hypothesized that comfort eating is a behavior that decreases the stress-response during the experience of anxiety [110]. This phenomenon might explain some of the mixed finding of the present study. When considering the studies using “unhealthy”, “cafeteria”, or Western diets in animal models, 17 studies reported a decrease in anxiety symptoms and 21 reported an increase in anxiety symptoms. In contrast, of the 17 human observational studies assessing the relationship between unhealthy diet patterns and anxiety symptoms, 15 reported an association with more anxiety, and two reported no association. The mixed findings among the animal studies may be due to the duration of the experiment. Many of these studies assessed the impact of three to four weeks of the diet exposure on animal behavior and many of the studies reporting benefit assessed the impact of the diet on animals experiencing stress. The reported beneficial effects may be capturing the short-term stress-reducing effect of foods high in sugar and fat. In contrast, the human observational studies may have been capturing the effects of chronic consumption of unhealthy diets.

Another limitation that impacts the ability to draw clear conclusions from the present data is the enormous complexity of studying nutritional science. When considering the role of macronutrients (carbohydrates, protein, and fat), it is necessary to consider both the amount and type of the nutrient consumed. As highlighted previously, studies which categorized dietary patterns as high or low in macronutrients such as fat or carbohydrates may not have considered the types of fat or carbohydrates being consumed. Given the highly different health impacts of complex and refined carbohydrates, significant attention should be given to the studies differentiating these rather than those assessing total carbohydrates only.

Only a small number of intervention studies involved participants with anxiety disorders, many involved healthy participants or individuals with medical illnesses such as irritable bowel syndrome, diabetes, and cardiovascular disease. This has several implications. First, many of these studies were designed to assess cardiometabolic outcomes primarily and the studies may have not been adequately powered to detect changes in mental health symptoms. Secondly, the participants recruited to participate in these studies related to physical illness may have had low baseline levels of anxiety symptoms making it difficult to detect statistically significant changes in symptoms or becoming more susceptible to other scale attenuation effects (e.g., floor and ceiling effects). Thirdly, the impact of a nutritional intervention on a healthy or non-anxious individual may not be relevant to understanding how the intervention might impact individuals with clinically significant anxiety disorders. There is a clear need for intervention studies enrolling participants with anxiety disorders or elevated anxiety symptoms. Similarly, studies designed with changes in mental health symptoms as the primary outcomes are needed.