The dramatic change in human life style, in particular in dietary habits, might be one of the reasons for the raise of civilization diseases, such as obesity, cardiovascular diseases, cancer, inflammatory, and autoimmune diseases, but also mental disorders such as depression, given that in modern societies the nutritional environment does not match the human genetic constitution (13). The modern Western dietary habits have changed in an unprecedented way since the industrial revolution with a dramatic shift from a balanced omega-3 to omega-6 ratio towards an excessive intake of omega-6 fatty acids. Nowadays, there are hardly any processed foods without an excess of omega-6 fatty acids due to the widespread industrial use of vegetable oils (e.g. sunflower oil). Furthermore, the intake of fish and other sources of omega-3 long chained polyunsaturated fatty acids (LC-PUFAs) have decreased, in particular in urban areas.

Epidemiological studies provide some evidence that high intake of fish seems to be a protective factor against the development of MDD (14), in particular in females (15). Case-control studies confirm that a high fish intake reduces the relative risk for child and adolescent depression (16). Reduced levels of omega-3 fatty acids in red blood cells of depressed patients provide further evidence supporting a link between altered PUFA metabolism and depression (17). Interestingly, such a reduction in PUFAs is only found in patients experiencing a current depressive episode, but not in remitted patients (18, 19), suggesting a direct and reversible relationship between omega-3 fatty acids deficiency and depression.

Meta analyses summarizing the results of various randomized placebo-controlled clinical trials on the efficacy of omega-3 fatty acids treatment in MDD reported standardized mean differences ranging from 0.11 to 0.56 (an overview of all RCTs is given in SI Table 1 ), reaching different conclusions of the clinical significance of omega-3 fatty acids as an antidepressant (AD) treatment (20–30). A recent rigorous Cochrane review encompassing 20 RCTs in MDD suggests a small-to-modest benefit for depressive symptomology with an SMD = −0.32 (27). Some of the inconsistencies between studies may also be explained by methodological differences, such as the inclusion/exclusion criteria, study subgroup selection, and choice of drug composition. For example, omega-3 fatty acids have only been found to be effective in populations with a clinical diagnosis of depression, but not in subclinical depressed individuals (23, 28). In addition, formulations rich in eicosapentaenoic acid (EPA) seem to be more effective than formulations rich in docosahexaenoic acid (DHA) (26), although it is unclear whether the effects depend on a specific ratio of EPA to DHA (24) or solely on a higher EPA dose irrespective of the DHA content (28, 29). Furthermore, omega-3 fatty acids might be especially effective in depressed patients with no comorbid disorders (30), and some studies have shown that it is also effective as an augmentation to another AD therapy, possibly even more so (28, 29, 31). Nearly all meta-analyses come to the conclusion that, at present, we do not have sufficient evidence to support or rule out that omega-3 PUFAs are potential treatments for MDD or could reduce dosages of selective serotonin reuptake inhibitors (SSRI) medication in MDD. Adequately powered RCTs are, therefore, warranted.

In contrast to the abundance of studies conducted in adults, the evidence of a link between omega-3 fatty acids and depression in children and adolescents is relatively sparse. In a Japanese study including 6,700 adolescents aged 12 to 15 years daily fish intake was inversely associated with depressive symptoms, but only in college boys and not girls (16). Red blood cell fatty acid levels have also been shown to be lower in depressed adolescents than in healthy controls (32) although the highest effects were found in other fatty acids than EPA or DHA.

The few randomized controlled trials conducted in minors showed at least partly encouraging results (see SI Table 1 ). In a small study including 20 depressed children aged 6 to 12 years, omega-3 fatty acid supplementation showed a large advantage over placebo (33). Given the small sample size, though, these results need to be interpreted with caution. Similarly, an RCT including 23 young adults around 20 years of age also showed significant reduction in depression severity after 3 weeks of omega-3 fatty acids treatment (34). In another small scale study including 35 patients with half of them getting a daily dose of 2,400 mg omega-3 fatty acids and the other half getting the same amount of omega-6 fatty acids (35), a significant treatment effect was found for patients with a major depressive disorder. Patients with a mixed anxiety depressive disorder, however, did not benefit from the omega-3 fatty acid treatment. The largest study to date included 72 youth with pediatric depression aged 7 to 14 years which underwent omega-3 monotherapy, individual-family psychoeducational psychotherapy (PEP) or combinations, respectively, for a trial duration of 12 weeks (36). Omega-3 treatment showed a small to medium effect compared to placebo. In an interesting study including 14 adolescents with SSRI resistant MDD, half of the patients received a very high dose of 16.2 g omega-3 fatty acids per day and the other half a low dose of 2.4 g per day over 10 weeks (37). In the high-dose group, 100% of patients showed symptoms remissions compared to 40% in the low dose group. In contrast though, no beneficial effects of omega-3 fatty acids were found in an RCT including 51 adolescents with MDD who were treated for a period of 10 weeks (38). In summary, preliminary data point towards beneficial effects of omega-3 fatty acids on pediatric depression but large-scale studies are needed to validate these results.

Cognitive deficits are often observed in depressed individuals reflected in reduced concentration, attention, and impairments in executive and memory functions, among others (39). Even more so, residual cognitive deficits may persist after remission of the depressive episode affecting academic performance (40). The role of omega 3 fatty acids in improving cognitive functions has been extensively studied in a variety of populations ranging from infants to the elderly, and from healthy individuals to patients with psychiatric, neurodegenerative or neurodevelopment disorders. For example, omega 3 fatty acids supplementation has been shown to improve depressive symptoms and verbal fluency in elderly with mild cognitive impairments (41), although the effects might depend also on genetic variations of the patients (42). However, other studies reported no or only small beneficial effects (43, 44), or only on specific cognitive subdomains such as reaction time (45).

The evidence of a beneficial effect of omega-3 fatty acids on cognition in children and adolescents remains inconclusive. In a Danish study assessing the impact of a healthy diet on school performance in third and fourth grade children, the dietary intervention improved school performance and reading comprehension. Post hoc analysis revealed that about 20% of the intervention effect might be attributed to an increase in EPA and DHA levels due to the intake of fish (46). Similarly, end-term grades and vocabulary were found to be higher in the group of 700 Dutch high school students who met the national guidelines of fish consumption compared to the group who never ate fish (47). In healthy adolescents aged 13 to 15 years, the blood level ratio between omega-3 and omega-6 correlated with scores on the letter Digit Substitution Test, indicating a higher information processing speed (48).

A recent meta-analysis confirmed a beneficial effect of omega-3 fatty acids in infants up to 18 months of age but not in children or adults (49), although the inclusion of a wide range of assessment instruments and age stages might raise questions regarding the validity of the result (50). Even if omega-3 fatty acids might not improve cognition in healthy developing children, they might alleviate cognitive deficits in adolescents with psychiatric disorders. For example, omega-3 fatty acid supplementation for 16 weeks improved working memory but no other cognitive functions in children diagnosed with attention-deficit/hyperactivity disorder (ADHD) (51). Similarly, the level of DHA was lower in ADHD children with learning difficulties than in those without, and higher DHA levels predicted better word reading skills (52). In a recent meta-analysis Chang and colleagues (53) concluded that omega-3 fatty acids supplementation improves cognitive functioning associated with attention in ADHD children. Surprisingly, only one study to date assessed directly whether omega-3 fatty acid supplementation will not only improve mood symptoms in pediatric depression but also associated cognitive deficits. Vesco and colleagues (54) found improved executive functioning after a 12-week omega-3 fatty acid intervention. It remains to be established whether omega-3 fatty acids have beneficial effects on other cognitive domains that are often affected in pediatric depression, such as working memory or attention. Given the high prevalence of cognitive deficits in depressed youth and their impact on academic performance and school work, more studies using a set of different cognitive tests are necessary to evaluate differential effects of omega-3 fatty acids on cognitive functioning in depressed children and youth.

Preclinical and clinical data point toward several mechanisms that most likely act in concert (55). For example, omega-3 fatty acids have shown to attenuate the exaggerated, persistent elevation of the stress response in animal models with depressive features (56–58) and humans (59, 60). Several lines of evidence support an altered immune-modulation in the pathophysiology of depression. Inflammatory mediators, like cytokines, play an important role in the stress system by influencing the activity of the hypothalamic pituitary adrenal (HPA) axis (61). Chronic stress, elicits a neuroinflammatory response, releasing inflammatory mediators, such as interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) (62), two cytokines that are inhibited by omega-3 fatty acids. Chronic stress also alters activation of the immune system in the periphery, which might account for the state of chronic inflammation observed in depressed patients (63). Omega-6 fatty acids such as arachidonic acid (AA) and omega-3 fatty acids such as EPA and DHA are both converted in bioactive molecules, also called eicosanoids. The eicosanoids derived from omega-3 and omega-6 fatty acids have partially opposing functions, with the first ones being mainly pro-inflammatory and the latter ones acting anti-inflammatory (64, 65). Both compete with each other for incorporation into cell membrane phospholipids (66–68). Omega-3 fatty acids inhibit the release of pro-inflammatory cytokines, such as IL-1β, IL-2, IL-6, interferon-γ, and TNFα (69). Given the link between depression and an increased production of pro-inflammatory cytokines (70), this inhibition might explain some of the AD effects of omega-3 fatty acids. Furthermore, cytokines also lower neurotransmitter precursor availability, and neurotransmitter metabolism (71), which might be important in the pathogenesis of mood disorders.

Omega-3 fatty acids are also crucially important for brain development. Myelination and synaptic pruning are core processes during normal pubertal brain development (72). The regulation of PUFA metabolism is crucial for both processes (73, 74). Of particular interest is a preclinical study investigating cognition and behavior across different developmental stages in rats. Omega-3 fatty acids deficient diets across consecutive generations produced a modality-selective and task-dependent impairment in cognitive and motivated behavior in adolescent rats distinct from the deficits observed in adult rats (75, 76). Omega-3 fatty acids supplementation was able to attenuate such depression-like animal behaviors during critical periods of brain development (77). Furthermore, omega-3 fatty acids have a preventive and neurotrophic effect against structural hippocampal alterations in animal models with depression- and anxiety-like behaviors (78–81). In addition, supplementation with omega-3 fatty acids might increase synaptic plasticity through an increase of brain-derived neurotrophic factor (BDNF), acting neuroprotective (71). Omega-3 fatty acids might ameliorate symptoms of depression through the protection of cells and their positive effects on neurogenesis, and counteract the shrinkage of the hippocampus (82). Omega-3 fatty acids deprived rats also provide evidence for an increase in serotonin 2 (5-HT2) and a decrease in dopamine 2 (D2) receptor density in the frontal cortex, as well as an increased serotonin turnover in the prefrontal cortex and decreased midbrain tryptophan hydroxylase-2 expression (83–89). In humans, omega-3 intake is associated with an increase in cerebrospinal fluid 5-5-hydroxyindoleacetic acid (5-HIAA) release (90, 91).

Omega-3 fatty acids are essential components of intracellular und neuronal cell membranes, affecting cell membrane integrity and fluidity (69, 92, 93). Through their effect on the efficiency of membrane functioning, omega-3 fatty acids play a role in a variety of biological mechanisms, such as enzyme and receptor activity, ion channels functioning, and the production and activity of neurotransmitters (85, 94). An increase in membrane fluidity results in a more flexible membrane and facilitates transmission (95).

While supplementation of omega-3 fatty aims to increase concentrations of fatty acids and their derived bioactive lipids that are available to the organism, omega-3 fatty acid metabolism depends on a variety of factors, including genetic variations. Therefore, omega-3 fatty concentrations will be monitored over the course of the study with blood sampling procedures carried out at baseline, after 3 months, and at the end of the study. Remaining blood is stored in a biobank allowing for further analyses of various parameters associated with omega-3 fatty acid balance and depression as well as the impact of genetic makeups on fatty acid metabolism.

As summarized above, there is growing evidence for the efficacy of omega-3 fatty acids as a treatment for major depression, in particular in pediatric populations. The generalizability of the results, though, is hampered by the conduction of small-scale studies, which do not allow for an accurate estimation of effect sizes. Therefore, the implementation of a well-designed large-scale clinical trial is warranted.

The present omega-3-pMDD trial recruits 220 patients with clinically diagnosed pediatric depression to assess the efficacy and safety of omega-3 fatty acids treatment. Given the knowledge gained from trials conducted in adults, formula rich in EPA will be used. In addition, comorbid disorders are assessed systematically for each patient, and concomitant medication is recorded throughout the 36-week trial. In case no concomitant medication is given, the efficacy of omega-3 fatty acid supplementation as a first line treatment is evaluated. However, ethical considerations do not allow for all patients to be treated solely with omega-3 fatty acids or placebo over a period of 36 weeks. In cases of severe or treatment resistant depression, additional pharmacological therapies might be indicated. In a recent survey, 40% of 85,000 depressed adolescents received AD medication during the course of their illness (96). This number appears relatively high given the major concern of increased suicidal ideations in the first weeks of treatment, which has been linked to the intake of ADs in minors (97, 98). In 2004, the advisory board of the Food and Drug Administration (FDA) decided to issue a black box warning indicating an increased risk of suicidal ideation in children receiving AD medication, especially SSRIs. Although further meta-analyses debated the increase of suicidal ideation after use of these drugs (99), it has contributed to a lot of insecurity in affected individuals and their families, as well as in the professional community (100).

When we designed the Omega-3 pMDD study, we evaluated several possible study designs. If we had decided to do a pure augmentation trial, the likelihood to lose those families opposed to AD medication (quite common in pediatric populations) would have been large. If we had decided to do a pure monotherapy omega-3 PUFAs RCT, we were concerned that clinicians would only refer less severe cases raising the risk of a ceiling effect due to the natural recovery rates of less severely depressed patients. Furthermore, omega-3 fatty acids seem to have better efficacy in severely depressed patients (23), potentially because more severe states of depression may be associated with stronger inflammatory processes. Both options were therefore considered less favorable to investigate the primary aim that omega-3 PUFAs have AD properties in moderately to severely depressed children and adolescents. For all these reasons, we decided that concomitant treatment with AD medication is permitted to not compromise the representativeness of the sample and to ensure that patients receive the optimal therapy for their disorder. Our design will allow us to directly compare the effects of omega-3 fatty acids as a monotherapy as well as adjunct to other AD drugs. In addition, statistical methods including multiple imputations in those who are put on ADs will be employed to control for the effects of AD medications on the trials’ results. Without controlling for AD use, we would run the risk to compare omega-3 fatty acids with ADs instead of placebo. In the proposed additional intention-to-treat analysis, we will consider the start of AD medication as a dropout criterion and impute the consecutive assessment time points of the main outcome measurements based on the values of those subjects that are not put on an AD. By considering the beginning of an AD as a treatment failure, we minimize the risk to mask a positive effect of omega-3 fatty acid supplementation by starting ADs more frequently or earlier in the placebo arm than in the omega-3 fatty acid arm. With this unique und novel way of analyzing the primary outcome data, the omega-3-pMDD trial aims to answer a wide range of questions regarding the treatment with omega-3 fatty acids, such as their efficacy and safety, the influence of omega-3 fatty acid supplementation on cognitive deficits associated with depression, and the emergence of putative biomarkers for omega-3 fatty acid treatment with and without the confounding effects of ADs.

The Omega-3-pMDD study is a Swiss multicenter, randomized, double blind, placebo controlled clinical trial, enrolling a sample of 220 individuals aged 8 to 17 years who have a present primary diagnosis of a major depressive disorder of at least moderate symptom severity. The study design incorporates a 7- to 10-day lead in phase and a 36-week double blind placebo-controlled treatment phase. The study was approved by the local ethics committees and the regulatory affairs and is registered at ClinicalTrials.gov protocol No NCT03167307. The primary outcome is changes in the Children’s Depression Rating Scale (CDRS-R) total score (101, 102). The categorical co-primary outcome includes response rates at 6 weeks and remission rates at 12 and 36 weeks, and rates of recovery defined by the absence of major depression for > 4 months at 36 weeks. Secondary outcome measures are a variety of psychopathological, neuropsychological, and biological measures, including questionnaires, cognitive testing, and biological markers. In addition, potential response predictors are assessed, such as inflammatory mediators in serum, red blood cell acids, and bioactive lipid mediators.

The background treatment across both groups is standardized based on the evidence and consensus-based German S3 Guidelines for the treatment of depression in children and adolescents (105). The core element is based on the cognitive behavioral therapy (CBT) method that includes individual CBT as well as psycho-educational family sessions. All participant sites are trained accordingly. In severe or treatment resistant cases of depression, the prescription of AD drugs might be indicated by experienced senior clinicians of the participating centers. Medicated patients can still participate in the trial. However, the start time and dose of the AD medication are recorded and are taken into account in the statistical analyses, as described further below.

The data are stratified by sex (male, female), age group (≥ 13 years old, < 13 years old), center, and whether the patient belongs to an in- or outpatient unit. In addition, the value of the high sensitive C-reactive protein (hsCRP) forms another stratification parameter (low < 1 mg/L; average 1.0–3.0 mg/L; high > 3.0 mg/L). Previous studies in adults (107, 108) and adolescents (109) have shown serum hsCRP to be an independent risk factor in depressive disorders, highlighting the role of low grade inflammation in the pathogenesis of mood disorders. Given the anti-inflammatory mechanisms of omega-3 fatty acids, a balanced allocation of patients with and without low-grade inflammation in the treatment arms are warranted. Randomization is carried out using the in-built functions of the clinical trial software SecuTrial^®. SecuTrial^® assigns the patient to the subgroup which leads to least imbalance within all strata. The software then provides a randomization number from a list implemented previously by a data administrator not otherwise involved in the study. Given that not always the next randomization number is selected by the computer, the study team is unable to guess what the next number will be. The research psychologist hands out the correct trial medication, which are stored in identical bottles and with the number as only identification.

A flowchart of the study schedule is depicted in Figure 1 . After the consent process, patients and their families undergo a screening interview in order to assess inclusion criteria, such as the diagnosis and severity of the depressive disorder and exclusion criteria, such as comorbidities that might prevent participation. Sociodemographic data and information about medical and family history and current medication use is also collected. If a patient is considered eligible for the trial, he or she enters the 7 to 10 days single blinded placebo lead-in phase. All patients receive placebo capsules for at least 7 days prior to randomization in order to assess any difficulties in swallowing the medication and measure compliance. Furthermore, the lead-in phase allows the identification of fast placebo responders. The baseline visit takes place at the end of the placebo lead-in phase, during which the CDRS and the subscale depressive disorder of the K-SADS-PL will be administered for a second time. If the patient still fulfills the diagnosis of a major depressive disorder with at least moderate severity, and no other exclusion criteria have emerged, the patient is randomized. Four follow-up assessments at 6, 12, 24, and 36 weeks are performed, consisting mainly out of the same measurements taken during the lead-in phase. Only four follow-up measurements are scheduled given that the number of research assessments has been identified as a predictor for placebo response (141, 142). The timing of the assessments was based on the design of the “Treatment for Adolescents with Depression” study (143), allowing us to make direct comparisons of illness trajectories between the trials. Intelligence is assessed only once at the 6-week appointment, and hair sample are taken at the 6-week appointment and at the end of the trial. Unused study drugs are counted at each appointment with compliance rates below 60% resulting in the discontinuation of the trial.

Descriptive statistics include mean and standard deviation for continuous variables, as well as number and percentages of total for categorical variables. Ordinal scaled variables will be reported as median and interquartile range. Descriptive statistics will be reported within treatment groups.

Primary analysis is performed based on the intention to treat (ITT) principle, i.e. including all randomized patients. For the continuous primary outcome (total CDRS-R score), we will apply a linear random intercept regression model. For the co-primary endpoint, i.e., the binary outcome measure for treatment response, we will apply a logistic regression model, and treatment group as independent variable will be used. Results will be presented as effect estimates — beta-coefficients or odds ratios — with 95% confidence intervals for primary and co-primary endpoints, respectively.

However, the analysis of treatment efficacy of the omega-3 fatty acids needs to address that some study participants already receive ADs prior to study inclusion or will be put on ADs in the due course of the study. The additional use of ADs might not be uniformly distributed between the two treatment arms, which might introduce biased results if not accounted for. To account for this major confounding factor, we propose to use two different analytic approaches. (1) We will perform an additional ITT analysis of the primary and secondary outcomes defining the beginning of an AD as a treatment failure and include treatment failure/additional use of AD as time-dependent covariate in the random effects model. (2) Using 20-fold multiple imputation, depression scores that a patient would have had without AD prescriptions are estimated, based on consecutive data of subjects without AD, and on earlier observations in the same subjects. To address that the missingness generating mechanism is considered to be “missing not at random” (MNAR), the “drawn indicator method” proposed by Jolani et al. (147) will be applied. A subsequent sensitivity analysis of the results will reveal the dependence of the results on specific assumptions in the two different analytic approaches.

A pre-specified subgroup analysis will address whether there is a differential treatment effect between patients with or without the pre-specified variable (e.g. AD therapy at baseline). If there is evidence for an interaction (as quantified with the p-value of the corresponding interaction test), the treatment effect will also be reported in those subgroups.

Furthermore, we’ll run following analyses to investigate the impact of AD use based on a priori working hypotheses: Firstly, we will test differences in novel AD prescriptions between the treatment groups by means of the two proportions z-Test using the ITT sample (assuming that ADs are more frequently prescribed in the placebo arm). We will also calculate retention rates between the two groups by means of the two proportions z-Test prior of “dropping out” those subjects put on ADs (assuming that the retention rate is longer in the active compared to the placebo arm).

Further analyses on primary and secondary outcome measures will be run as detailed in the study protocol, which is available on the clinical.trials.gov website.

Omega-3 fatty acids are usually well-tolerated, but according to a recent meta-analysis by Chang et al. (148), they might elicit mild adverse effects such as gastrointestinal symptoms in some patients. In a meta-analysis including 148 omega-3 fatty acid studies with dosage up to 6 g per day, 6.6% of the active group relative to 4.3% in the placebo group reported gastrointestinal complaints. Only one study reported an increase in bleeding time (64). Thus, any medical conditions experienced during the trial are recorded. In addition, patients fill in the ASEC at each visit to measure other putative side effects. No doses changes of active/placebo are permitted during the trial, but patients and their families are informed of the voluntary nature of the study and that consent can be rescinded anytime. While no serious side effects are expected due to omega-3 fatty acids intake, the current patient population warrants for a tight safety control due to the risk of suicide. Therefore, suicidal tendencies will be closely monitored by the study team by asking directly about suicidal ideations during the clinical interviews in addition to the administration of the Suicidal Ideation Questionnaire (SIQ-Jr). Furthermore, a data monitoring committee (DMC) is set up consisting of an expert in child and adolescent psychiatry, an expert in ethical considerations, and an expert in biostatistics. All members have access to the unblinded data due to permissions associated with their study function in the database SecuTrial. In case of a serious adverse event, the members of the DMC are automatically notified through an inbuilt function of the software in order for them to continuously monitor the safety of the patients. Furthermore, after 60 enrolled patients, an interims analysis on the safety of the treatment is was performed by a member of the DMC.

The main outcome of the trial is published irrespective of the results. A publications committee establishes publication guidelines and coordinates the publication of secondary outcome measures, also giving recommendations to the timing of different abstracts, reviewing all publications for their appropriateness and scientific merit. No later than five years after closure of the data base, a completely de-identified data set will be provided for sharing purposes.