Depression is one of the relatively common mental disorders in today’s society. Based on the evidence from the WHO, more than 264 million people are affected by depression worldwide(Reference James, Abate and Abate1,Reference Musazadeh, Zarezadeh and Faghfouri2) . Recent evidence has also shown that the prevalence of depression has increased in the current context of the corona epidemic(Reference Ettman, Abdalla and Cohen3). The most obvious symptoms of depression include sadness, dissatisfaction with life, changes in appetite, low self-esteem, lack of motivation, anhedonia, low energy level, discomfort without a clear cause, changes in sleep, suicidal ideation, etc(Reference Turegeldieva and Amitov4,Reference Yang, Wang and Hu5) . It has been shown that nutrition, genetic, environmental, immunologic and endocrine factors and neurogenesis have major roles in the pathogenesis of depression(Reference Jesulola, Micalos and Baguley6). The levels of biogenic amines including serotonin, melatonin, dopamine,γ-aminobutyric acid and glutamate are altered in these patients(Reference Nobis, Zalewski and Waszkiewicz7). Several mechanisms have been proposed for depression. As a known aetiopathological mechanism of depression, excessive release of glutamate under stressful conditions causes astrocyte apoptosis through overstimulation of the N-methyl-D-aspartic acid (NMDA) receptor in the hippocampus. Depression is highly attributed to the overactivation of NMDA receptor in the brains(Reference Jiang, Zhang and Guo8).
While long-term concerns about their effectiveness and safety exist, antidepressants have been used for years to treat depression(Reference Kirsch, Deacon and Huedo-Medina9–Reference Jakobsen, Katakam and Schou11). Furthermore, non-responsiveness to a wide range of pharmaceutical treatments(Reference Fava12) suggests that other mechanisms are involved in the pathogenesis of depression, including abnormalities in the neuroendocrine, immunological, neurotrophic and metabolic systems(Reference Strawbridge, Young and Cleare13). Despite these challenges, new complementary depression treatments are needed.
Dietary long-chain fatty acids (n-3 PUFA), namely EPA(C20:5n-3), DHA (22:6n-3) and α-linolenic acid (18:3n-3), have many beneficial effects like anti-inflammatory, pro-apoptotic, anti-tumour and anti-catabolic activities(Reference Santos, Price and Bueno14,Reference Musazadeh, Kavyani and Naghshbandi15) . DHA, as an essential and structural n-3 fatty acid (FA), improves neurogenesis and repairs the myelin membrane(Reference Weiser, Butt and Mohajeri16). n-3 FA suppress neuroinflammation via inhibition of expression of some of IL genes and cyclooxygenase enzymes(Reference Krawczyk and Rybakowski17–Reference Kavyani, Musazadeh and Fathi19). Also, studies reported that n-3 FA have therapeutic potential in patients who show resistance to common medications(Reference Krawczyk and Rybakowski17). On the other hand, Liao and et al. in a meta-analysis highlighted that composition of PUFA intake is important as EPA is more effective than DHA in the treatment of depression(Reference Liao, Xie and Zhang20).
There are many clinical trial and meta-analysis articles investigating the relation between n-3 FA and depression. Most studies have shown the beneficial effects of n-3 FA on improving depressive symptoms(Reference Liao, Xie and Zhang21–Reference Suradom, Suttajit and Oon-Arom26). Although various meta-analyses have been performed over the past years, they have not yet reached a clear conclusion about the effect of n-3 on treating depression and the results are still conflicting in some cases. Therefore, we tried to summarise and classify the effects of various unsaturated fatty acids (linolenic, EPA, DHA) on different symptoms of depression by conducting an umbrella review on published meta-analyses in this field to provide a comprehensive perspective.
Methods
Search strategy and study selection
PICO criteria for the current study were as follows: Population/Patients (P: adults of 18 > years of age, who were supplemented with n-3 PUFA); Intervention (I: n-3 PUFA); Comparison (C: control or placebo group) and Outcome (O: depression symptoms). A comprehensive systematic search of the international scientific databases including PubMed, Scopus, Embase, Web of Science and Cochrane Central Library was carried out up to June 2021. The search strategy was developed using the following MeSH terms and keywords: (depress* OR dysthymi OR ‘affective disorder’ OR ‘affective symptom’ OR ‘mood disorder’ OR ‘mental health)’ AND (‘DHA’ OR docosahex OR eicosapent OR ‘EPA’ OR ‘fatty acid’ OR fish OR linolenic OR omega OR ‘n-3’ OR ‘ω-3’ OR ‘PUFA’ OR ‘cod liver oil’) AND ‘systematic review’ OR ‘meta-analysis’) (online Supplementary file 1). The wild-card term ‘*’ was used to increase the sensitivity of the search strategy. The articles in English language only were included in the study. The protocol of the study has been registered on PROSPERO (registration code: CRD42021282975).
Inclusion and exclusion criteria
Meta-analysis studies examining the effects of n-3 FA supplementation on depression symptoms while reporting the effect sizes (ES) and corresponding CI were included in the current umbrella meta-analysis. In vitro, in vivo and ex vivo studies, case reports, observational studies, quasi-experimental studies, clinical trials and also the studies that did not achieve the least quality score were excluded from this meta-analysis of meta-analyses.
Quality assessment
Two reviewers (D.Q, J.M) independently assessed the methodological quality of the eligible articles using the AMSTAR questionnaire(Reference Shea, Grimshaw and Wells27). The eleven questions in the questionnaire are completed with yes, no, can’t answer or not applicable and has a maximum of eleven points. Articles with a score of over 7 were considered as high quality.
Data extraction
Articles were screened based on the eligibility criteria by two independent reviewers (D.Q, J.M). At first, the abstract and title of the articles were reviewed. Then, the full text of screened papers was evaluated to ascertain the suitability of the study to include in the umbrella meta-analysis. The judgement of the third author (Y.L) resolved any disagreement.
The first authors’ name, year, sample size, study location, the dose and duration range of supplementation, ES and CI for depression symptoms were extracted from the selected studies.
Results
Systematic review
The flow diagram of the literature search process is summarised in Fig. 1. The initial search identified 101 records of which sixty were duplicates. After screening titles and abstracts, fifty-one studies were considered potentially relevant and included for full-text evaluation. Twenty-nine articles after a full-text review were excluded and at the end, twenty-two studies published between 2007 and 2021 were included in the current analysis. The characteristics of eligible studies are described in Table 1. Five studies were performed in UK(Reference Appleton, Sallis and Perry23,Reference Appleton, Hayward and Gunnell28–Reference Appleton, Rogers and Ness31) , five in China(Reference Liao, Xie and Zhang21,Reference Bai, Bo and Wu32–Reference Yang, Han and Qiao35) , three in Netherlands(Reference Mocking, Steijn and Roos22,Reference Jans, Giltay and Van der Does36,Reference Mocking, Harmsen and Assies37) , two in USA(Reference Bloch and Hannestad38,Reference Hallahan, Ryan and Hibbeln39) ,one in Australia(Reference Sarris, Mischoulon and Schweitzer40), one in Columbia(Reference Sublette, Ellis and Geant41), two in Korea(Reference Bae and Kim24) and Italy(Reference Grosso, Pajak and Marventano42) and three in Thailand(Reference Suradom, Suttajit and Oon-Arom26), Germany(Reference Wolters, von der Haar and Baalmann43) and Taiwan(Reference Lin and Su44). The duration of interventions varied from 4 to 160 weeks in included studies. Six meta-analyses were conducted exclusively on women(Reference Suradom, Suttajit and Oon-Arom26,Reference Jans, Giltay and Van der Does45–Reference Zhang, Zou and Li49) and the remaining studies on both sexes. The pooled number of participants per meta-analysis ranged from 201 to 10 297. Two meta-analyses reported two ES, each(Reference Sublette, Ellis and Geant50,Reference Hallahan, Ryan and Hibbeln51) . Included meta-analyses have used Cochrane Risk of Bias Tool(Reference Higgins, Altman and Gøtzsche52) and Jadad scores(Reference Clark, Wells and Huët53) to assess the quality of clinical trials. Most of the meta-analyses included high-quality trials. The quality of the original randomized controlled trial (RCT) studies is presented in Table 1.
Fig. 1. Flow diagram of study selection.
Table 1. Study characteristics of included studies. (95 % confidence intervals)
BDI, Beck Depression Inventory; HDRS, Hamilton Depression Rating Scale; MADRS, Montgomery-Asberg Depression Rating Scale; IDS-C, Inventory of Depressive Symptomatology; DSP, Derogatis Stress Profile; HDRS-SF, HDRS short form; POMS, Profile of Mood States; CDRS, Children’s Depression Rating Scale; CDI, Children’s Depression Inventory; CGI, Clinical Global Impression; IDS-C, Inventory of Depressive Symptomatology; MMSE, Mini-Mental State Examination; EPDS, Edinburgh Postnatal Depression Scale; NPI, Neuropsychiatric Inventory; DASS, Depression, Anxiety and Stress Scales; GHQ, General Health Questionnaire; CES-D, Center for Epidemiologic Studies–Depression scale; GDS-15, Geriatric Depression scale (15 items); LEIDSR, Leiden Index of Depression Sensitivity; PPBQ, Postpartum Blues Questionnaire; PGWB, Psychological General Well-Being Index; HSCL-D-20, Hopkins Symptom Checklist Scale (20 items); ALA, α-linolenic acid.
Risk of bias assessment
Almost all of the included meta-analyses had high quality. On average, the meta-analyses achieved a mean of 10 scores. The results of quality assessment of meta-analyses were done according to the AMSTAR questionnaire and summarised in Table 2.
Table 2. Results of assessing the methodological quality of meta-analysis
The result of assess the methodological quality using AMSTAR: each item for included studies (?: can’t answer; -: means no; +: means yes).
Effects of n-3 PUFA on depression
The results of five meta-analyses reported standardized mean difference (SMD) ranging from −0·61 to −0·94 that n-3 PUFA supplementation significantly reduced depressive symptoms(Reference Wei-Hong, Cheng-Gui and Peng-Fei46,Reference Zhang, Zou and Li49,Reference Yang, Han and Qiao54–Reference Bae and Kim56) . The results also indicated that a more pronounced improving effect was observed on depression symptoms when EPA with DHA were supplemented, average dose of n-3 PUFA was > 3 g/d and intervention period was > 15 weeks. The findings from eleven other meta-analyses showed that n-3 PUFA supplementation significantly improved depression symptoms, although the clinical significance was negligible(Reference Liao, Xie and Zhang21,Reference Appleton, Rogers and Ness31,Reference Grosso, Pajak and Marventano42,Reference Sublette, Ellis and Geant50,Reference Hallahan, Ryan and Hibbeln51,Reference Appleton, Hayward and Gunnell57–Reference Sarris, Mischoulon and Schweitzer62) . The remaining meta-analyses reported no significant effect of n-3 supplementation on depression symptoms, as almost half of the included RCT had low quality and with high significant between-study heterogeneity(Reference Suradom, Suttajit and Oon-Arom26,Reference Bloch and Hannestad38,Reference Jans, Giltay and Van der Does45,Reference Mocking, Steijn and Roos47,Reference Sublette, Ellis and Geant50,Reference Bai, Bo and Wu63,Reference Martins, Bentsen and Puri64) . Seven meta-analyses had large heterogeneity (I2 > 50 %)(Reference Liao, Xie and Zhang21,Reference Appleton, Rogers and Ness31,Reference Bloch and Hannestad38,Reference Grosso, Pajak and Marventano42,Reference Hallahan, Ryan and Hibbeln51,Reference Appleton, Sallis and Perry58,Reference Mocking, Harmsen and Assies60) , and seven had very large heterogeneity (I2 > 75 %)(Reference Mocking, Steijn and Roos47,Reference Zhang, Zou and Li49,Reference Yang, Han and Qiao54,Reference Bae and Kim56,Reference Appleton, Hayward and Gunnell57,Reference Martins59,Reference Wolters, von der Haar and Baalmann61) . In addition, three meta-analyses did not report information on heterogeneity between studies(Reference Sublette, Ellis and Geant50,Reference Bai, Bo and Wu63,Reference Martins, Bentsen and Puri64) . The quality assessment process of the RCT included in the five meta-analyses was not explained in detail, which could affect their overall quality(Reference Sublette, Ellis and Geant50,Reference Hallahan, Ryan and Hibbeln51,Reference Lin and Su55,Reference Mocking, Harmsen and Assies60,Reference Martins, Bentsen and Puri64) . The results of included meta-analyses are presented in Table 1.
Discussion
Assessment of twenty-two meta-analyses including eighty-three RCT in the current umbrella systematic review revealed that n-3 PUFA (EPA, DHA and combination of these FA) supplementation has a significant improvement effect on depression indices in most studies. However, some contradictory results have also been reported(Reference Mocking, Steijn and Roos22,Reference Suradom, Suttajit and Oon-Arom26,Reference Martins, Bentsen and Puri30,Reference Bai, Bo and Wu32,Reference Jans, Giltay and Van der Does36,Reference Bloch and Hannestad38,Reference Hallahan, Ryan and Hibbeln39,Reference Sublette, Ellis and Geant41) . The first important factor leading to these contradictory results is the degree of depression. Some studies have indicated that n-3 PUFA supplementation led to a non-significant or small effect on depression indicators in people with mild symptoms of depression or normal conditions. While in people with major depressive disorder, this supplement has a major positive effect(Reference Mocking, Steijn and Roos22,Reference Appleton, Hayward and Gunnell28,Reference Martins29,Reference Appleton, Rogers and Ness31,Reference Appleton, Sallis and Perry58) . The next item is the supplementary dose. Most of the studies performing subgroup and meta-regression analysis based on dose have pointed out that there is a direct relationship between dose and ES on depression(Reference Bai, Bo and Wu32,Reference Bloch and Hannestad38) . However, Lin et al. reported that the association between n-3 PUFA and depression is not dose-dependent(Reference Lin and Su44). The next important point is the control group. Appleton et al. who compared the effects of n-3 PUFA with placebo or antidepressants showed that n-3 PUFA have a better effect than placebo, although, compared with antidepressants, they do not have a superior effect(Reference Appleton, Sallis and Perry58). In addition to the mentioned factors, the type of n-3 FA is important. Some studies have indicated that EPA had superior effect on depression indices compared with DHA(Reference Liao, Xie and Zhang20,Reference Martins29,Reference Zhang, Zou and Li34,Reference Hallahan, Ryan and Hibbeln39,Reference Sublette, Ellis and Geant41) . Moreover, the substantial inefficacy of n-3 PUFA supplementation on depressive disorders in some studies might be related to their studied participants. Grosso et al. reported that n-3 PUFA was not effective in alleviating of depressive symptoms in patients suffering from depression as a secondary outcome(Reference Grosso, Pajak and Marventano42).
There are the broad types of scales to measure the depression in the investigated studies including Montgomery-Asberg Depression Rating Scale (MADRS), Beck Depression Inventory (BDI), Depression Anxiety Stress Scale (DASS), Hospital Anxiety and Depression Scale (HADS), Hamilton Depression Rating Scale, Edinburgh Postnatal Depression Scale (EPDS), Leiden Index of Depression, Geriatric Depression Scale (GDS), Profile of Mood States and Quick Inventory of Depressive Symptomatology (QIDS). Following purposes have been defined for these scales: screening, diagnosis, differentiated diagnosis, severity, treatment effect judgement and processing observation(Reference Lee, Kim and Shin65). Regarding inconsistency between these questionnaire, QIDS and BDI dedicated less than 10 % to ‘Mood and Outlook’ while the MADRS, DASS, EPDS and GDS dedicated 15 %+ to this category. Regarding ‘Confidence and Self Judgement’, BDI and GDS dedicated most and lowest to this category, respectively. The other categories such as sleeping, appetite, anxiety, self-harm and distress were even more heterogeneously distributed across the questionnaires(Reference Fried66). Consequently, the conflicting results in some studies could be due to the different scales used in the clinical trials they analysed.
It must be noted that the baseline levels of EPA and DHA and levels achieved have been traditionally ignored in intervention trials. Therefore, positive effect of n-3 PUFA on depression in included meta-analysis studies might be an underestimated result. It has been reported that DHA synthesis from ingested α-linolenic acid is typically < 1 % of the oral α-linolenic acid dose(Reference Domenichiello, Kitson and Bazinet67). Therefore, many studies ignore the endogenous biosynthesis of n-3 PUFA. Many studies have recorded n-3 PUFA intake as self-reports, which can be inaccurate in some cases. Therefore, it is recommended that future studies measure the serum or tissue level of n-3 PUFA to accurately determine their positive effects.
Both sexes benefited from n-3 PUFA to improve depression. However, major depression is more prevalent among females. It may be related to hormonal changes, particularly oestrogen in women during the different periods of their life such as prior to menstruation, puberty, following pregnancy and at perimenopause(Reference Albert68). Gordon et al. suggested that hormone replacement therapy especially could be effective in the treatment of premenopausal depression(Reference Gordon and Girdler69). In addition, improving efficacy of n-3 PUFA on maternal depression has been investigated in some studies(Reference Zhang, Zou and Li49,Reference Kaviani, Saniee and Azima70,Reference Hsu, Tung and Chen71) . In the male brain, oestrogen could be produced from testosterone via endogenous aromatase (CYP19). Oestrogen could exert its protective properties against depression via oestrogen receptors expressed in the male brain(Reference Gillies and McArthur72). Therefore, improving effect of n-3 PUFA on the females might be related to their beneficial effects on hormonal changes among them(Reference Ouladsahebmadarek, Khaki and Khanahmadi73).
The mechanism of action of n-3 FA on depression has been explained in many studies. n-3 PUFA have been proven to influence the activity and structure of brain through affecting adult neurogenesis and synaptogenesis by increasing the signalling factors involved in neurogenesis, such as brain-derived neurotrophic factor, cAMP response element-binding protein or calcium/calmodulin-dependent protein kinase II(Reference Cutuli74). It must be added that DHA is a component of neural membrane phospholipids(Reference Tanaka, Farooqui and Siddiqi75). Moreover, n-3 PUFA boost the blood flow of brain(Reference Howe, Evans and Kuszewski76). Therefore, decrease or increase in their intakes could affect the brain functions. n-3 PUFA through different pathways could improve depression. In one pathway, n-3 PUFA might increase the expression of dopamine receptor, thereby increasing the dopamine activity as a neurotransmitter in restoring the mood(Reference Sousa and Santos77). In addition, n-3 PUFA were effective in balancing the low level of brain-derived neurotrophic factor among depressed patients(Reference Paduchová, Katrenčíková and Vaváková78). Brain-derived neurotrophic factor could boost the action of antidepressant agents(Reference Yu and Chen79). In addition, n-3 PUFA act as an potent anti-inflammatory agent to supress the production of inflammatory cytokines such as IL-1 and TNF-α, which are involved in the pathogenesis of the depression(Reference Wani, Bhat and Ara80). In addition, n-3 PUFA were inversely associated with depression in patients with the increased levels of oxidative stress biomarkers(Reference Bigornia, Harris and Falcón81). Koponen et al. reported that depressed patients had impaired glucose metabolism, and this could result in pathological effects on the brain(Reference Koponen, Kautiainen and Leppänen82). Yeum et al. showed that n-3 PUFA tended to improve the metabolism of glucose(Reference Rafraf, Mohammadi and Asghari-Jafarabadi83). In depressive disorders, the hypothalamic–pituitary–adrenal axis tended to be up-regulated. Besides, negative feedback controls of this axis were down-regulated in depression. This led to increased secretion of corticotropin-releasing factor from the hypothalamus and subsequently adrenocorticotropic hormone from the pituitary and cortisol from the adrenal cortex(Reference Keller, Gomez and Williams84). Excessive cortisol desensitises cortisol receptors, thereby increasing the activity of macrophage as well as the level of proinflammatory cytokines. Moreover, this desensitisation led to disturbances in noradrenaline and serotonin transmission(Reference Silverman and Sternberg85). Kim et al. reported that n-3 PUFA could reduce the circulating adrenocorticotropin hormone, corticosterone and corticotropin-releasing factor, as well as elevate serotonin and brain-derived neurotrophic factor levels in a rat model(Reference Kim, Choi and Kim86).
Present umbrella systematic review is the first umbrella study that could collect and evaluate various and inconsistent results of published meta-analyses and reach to a decisive conclusion. Second, high quality of included meta-analyses and their included RCT helped us to obtain more reliable results. However, it had some limitations too that must be noted. First, subgroup analyses were limited in terms of dose, duration of supplementation, sex, depression screening tool and population in the included studies. As a result, no definite conclusions can be drawn about these terms. Second, the heterogeneity of some studies was high (Table 1), which can reduce the validity of the results. Therefore, results should be interpreted with precaution. Third, none of the studies has done a GRADE assessment. As a result, no conclusion can be drawn about the certainty of the evidence.
Conclusion
n-3 PUFA (EPA, DHA and combination of these FA) supplementation has a significant improvement effect on depression indices in most studies. This result was more pronounced in major depressive disorder. There is a direct relationship between dose of n-3 PUFA and ES on depression. n-3 PUFA have not a superior effect on depression compared with anti-depressant drugs. However, EPA has more anti-depressive effects than DHA. In conclusion, n-3 PUFA supplementation could be considered as an effective therapeutic adjuvant approach in relieving depression symptoms.
Acknowledgements
This study was supported by Shandong Province Science and Technology Development Project of Traditional Chinese Medicine (No. 2019-0527): Research on the relativity of emotion theory in the diagnosis and treatment of bipolar disorder. Ethics Committee of Shandong Mental Health Center, Approved June 21, 2019, no. (2019) Ethics review no. (R06).
Y. L.: contributions to the concept/design, data analysis/interpretation, critical revision of the manuscript and approval of the article. D. Q.: drafting the manuscript, acquisition of the data and approval of the article. G. M.: drafting the manuscript, acquisition of the data and approval of the article.
There are no conflicts of interest.
Supplementary material
For supplementary material referred to in this article, please visit https://doi.org/10.1017/S000711452300226X
