Oiling the Brain: A Review of Randomized Controlled Trials of Omega-3 Fatty Acids in Psychopathology across the Lifespan

Around one in four people suffer from mental illness at some stage in their lifetime. There is increasing awareness of the importance of nutrition, particularly omega-3 polyunsaturated fatty acids (n-3 PUFA), for optimal brain development and function. Hence in recent decades, researchers have explored effects of n-3 PUFA on mental health problems over the lifespan, from developmental disorders in childhood, to depression, aggression, and schizophrenia in adulthood, and cognitive decline, dementia and Alzheimer’s disease in late adulthood. This review provides an updated overview of the published and the registered clinical trials that investigate effects of n-3 PUFA supplementation on mental health and behavior, highlighting methodological differences and issues.


Introduction
Mental health is described as "a state of well-being in which every individual realizes his or her own potential, can cope with the normal stresses of life, can work productively and fruitfully, and is able to make a contribution to her or his community" [1]. About 450 million people worldwide do not enjoy this state of wellbeing, and it is estimated that one in four people will suffer from mental illness at some stage in their lifetime, causing enormous suffering, personal, social and financial burden, as

OPEN ACCESS
well as associated physical illness [2]. Identification of preventable risk factors and effective treatments therefore constitutes an international priority.
Nutrition has an important role in mental health as the brain relies on both macro-and micronutrients for development and functioning [3][4][5][6]. In particular, suboptimal levels of omega-3 polyunsaturated fatty acids (n-3 PUFA) are emerging as a potentially modifiable risk factor for mental illness. Since their discovery in the 1970s as key components of brain tissue [7], long-chain n-3 PUFA have been postulated to serve critical roles in brain development, and function and a lack of these fatty acids has been implicated in a number of mental health conditions over the lifespan, from developmental disorders and mental retardation in childhood, to depression, bipolar disorder, schizophrenia and borderline personality disorder, stress, hostility and aggression in adulthood, and cognitive decline, dementia and Alzheimer's disease in late adulthood [8].
Although the etiology of these disorders is clearly multifactorial and subject to individual variation, the number of clinical trials addressing suboptimal n-3 PUFA levels as a possible underlying biological contributor to associated symptoms, and importantly symptoms that often co-occur and overlap between these disorders, is steadily growing. Following an overview of n-3 PUFA, we will provide a review of randomized controlled trials [8], and an update of recent trials located via the World Health Organization's International Clinical Trials Registry Platform Search Portal [9], including preliminary results from our own recent studies, focusing on the role of n-3 PUFA in alleviating symptoms of psychopathology from childhood throughout adulthood. Trial identification numbers of registered trials not included in the original review were searched on PubMed to check if they had been published. Published and registered trials are summarized in Tables 1 and 2, respectively.

Overview
The omega-6 (n-6) and n-3 families of PUFA have numerous biological roles in cellular structure and function. Alpha linolenic acid (ALA; 18:3n-3), parent n-3 PUFA, and linoleic acid (LA; 18:2n-6), parent n-6 PUFA, cannot be synthesized by humans and must be provided via the diet. They are therefore termed essential fatty acids (EFA). Humans can obtain ALA in their diet via plant sources. EFA have 18 carbon atoms with double carbon bonds. These chains undergo further processes of elongation and de-saturation (insertion of additional double bonds) to form highly unsaturated long chain (LC) PUFA. Our bodies can perform these conversions, although not always very efficiently. Long chain (LC) derivatives of ALA, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), are made by phytoplankton and transferred via the food-chain to marine organisms, particularly coldwater fish and marine animals. Therefore, humans can also obtain these elongated n-3 PUFA directly via algal or fish sources.
Structurally, PUFA are key components of phospholipids, comprising cellular and intracellular membranes. They govern growth and vitality through oxidation (metabolism of food to produce energy required for cellular processes), chemical activities and transportation. As well as constructing membranes and generating and propagating electrical impulses, LC-PUFA are required for synthesis of eicosanoids, important signaling hormones with numerous complex functions. Those derived from n-3 LC-PUFA are generally anti-inflammatory, anti-thrombotic and vasodilatory, balancing and counteracting pro-inflammatory, vasoconstrictor actions of eicosanoids derived from n-6 LC PUFA [10]. The cardiovascular benefits of n-3 PUFA [11] are largely attributed to these eicosanoid properties, and they may also have important roles in brain function.

Attention Deficit Hyperactivity Disorder (ADHD)
Increasing attention has been given to the role of EFA in childhood developmental disorders. Given that brain development continues throughout childhood [42,43], PUFA supplementation may also have a beneficial effect during developmental milestones, particularly in children with developmental disorders that might be related to problems with PUFA metabolism and/or deficiency.
Early intervention trials were reported to show little or no improvement in alleviating hyperactive symptoms of ADHD with evening primrose oil, which contains the n-6 PUFA GLA [44,52]. Following these studies and with increasing awareness of the importance of n-3 PUFA, 8 randomised, doubleblind, placebo-controlled PUFA supplementation trials have been completed with children displaying ADHD-related symptoms, with varying results. These could be attributed to methodological differences such as variations in dosage, PUFA formulation, period of supplementation, selection criteria and outcome measures.
In the first study, supplementation with 345mg of DHA per day over four months in 6-12 year old children diagnosed with ADHD did not significantly improve objective or subjective measurements of symptoms, even though there was a significant increase in RBC levels of DHA in the DHA group [53]. In this study, stimulant medication ceased 24 hours before laboratory measurements. However, parents completed the subjective scales (Child Behavior Checklist and Conners' Rating Scales) while the children were still receiving stimulant medication. This may have influenced the results by masking their symptoms and making it difficult to detect any improvements; closer inspection confirms that tscores on parent rating scales were in the normal range before supplementation. Moreover, children who had experienced ineffective treatment with stimulant medication or had a previous diagnosis of mood, anxiety, thought or bipolar disorder were excluded from the study. It is presumed that the rationale for this was to select a group of children with 'pure' ADHD symptoms. This may not be entirely practical due to the high co-morbidity prevalent in ADHD and, as discussed by Richardson [54], the heterogeneity of ADHD and its overlap with other neurodevelopmental disorders needs to be explored within a phospholipid framework.
Another study using both DHA and EPA (supplemented in fish oil-enriched bread, supplying 3600mg DHA and 700mg EPA per week, or 514mg DHA and 100mg EPA per day) [55] also found no significant treatment effects on ADHD symptoms, in a two month placebo-controlled, double-blind trial with 40 children aged 6-12 who were mostly drug-free (34/40). The placebo bread contained olive oil. Paradoxically, the control group had significant improvements in visual short-term memory and errors of commission that were not seen in the treatment group. Blood samples were not taken, so it is not clear whether this sample had low n-3 PUFA levels at baseline. Given that the study was conducted in Japan, known to have high fish consumption, it is likely that baseline levels were relatively high. It is also possible that two months may have been insufficient time for effects to become observable, as it may take longer to elevate the fatty acid composition of neural membranes [56,57].
Five other interventions reported improvements in symptoms with a combination of n-6 and n-3 PUFA, including both EPA and DHA. In a pilot study, 41 British children aged 8-12 years with ADHD-related symptoms and specific learning disabilities (dyslexia) were given a PUFA supplement or olive oil placebo over 12 weeks [58]. The PUFA supplement in this study provided a daily dose of 186mg EPA, 480mg DHA, 864mg LA and 42mg AA, as well as α-tocopherol (Vitamin E). These children all had above average scores on parent ratings of ADHD symptoms, and average general ability yet low reading achievement. Baseline measures on learning and behaviour did not differ between the groups, but the PUFA group showed improvements on global ADHD indices and subscale ratings of cognitive problems/inattention, hyperactivity, restlessness/impulsivity, psychosomatic complaints and anxiety/shyness on Conners' Parent Rating Scale following 12 weeks of supplementation, with medium effect sizes reported.
Similarly, another study investigated effects of PUFA supplementation, providing a daily dose of 80mg EPA, 480mg DHA, 40mg AA, 96g GLA and 24mg α-tocopherol, or olive oil placebo, over four months in American children with ADHD-like symptoms, selected for thirst and skin problems indicative of fatty acid deficiency [59]. Plasma proportions of DHA, EPA and α-tocopherol increased in the intervention group following supplementation and there was a significant drop in n-6 to n-3 ratios, although the olive oil group demonstrated a small increase in ALA. Significant treatment effects were found for only 2 of 16 outcome measures when compared to placebo: parent ratings of conduct and teacher ratings of attention. Oppositional defiant behavior scores significantly improved from clinical to non-clinical levels, and significant relationships were reported between the change in RBC PUFA and the magnitude of improvements in outcome measures. Increased levels of EPA and DHA were associated with decreases in teacher ratings of inattention and increased levels of EPA with parent ratings of reduced disruptive behaviour. Methodological issues in this study included increased n-3 plasma levels in the placebo (olive oil) group and small sample size (N = 50 plus 17 dropouts).
Two larger double-blind, placebo-controlled trials investigated effects of PUFA on ADHD symptoms. The first study [60] included 117 children diagnosed with dyspraxia, or Developmental Coordination Disorder (DCD), and a third of the sample had scores  2 SD above the population mean on the DSM-IV ADHD subscale of Conners' Teacher Rating Scales (CTRS). The children were functioning on average about a year behind their chronological age in reading and spelling. The active supplement provided a high ratio of EPA:DHA with a daily dose of 552 mg EPA, 168 mg DHA, 60mg DGLA and 9.6 mg α-tocopherol. Although no improvements were detected on the core DCD symptom of poor motor function after three months, highly significant improvements were reported in reading and spelling in the treatment group compared to placebo over three months and significant treatment effects were observed for teacher ratings of oppositional behaviour, cognitive problems/inattention, hyperactivity, anxiety/shyness, and global/DSM-IV ADHD. Following a one-way, uncontrolled crossover to active treatment for a further three months, the placebo group demonstrated similar improvements in mean scores, while the original treatment group's scores continued to improve.
The second large intervention included 132 Australian children, all with ADHD symptoms > 90 th percentile and therefore in the clinical range for a diagnosis [61]. Children received the same treatment as above (a daily dose of 552 mg EPA, 168 mg DHA, 60mg DGLA and 9.6 mg α-tocopherol) with or without a micronutrient supplement or placebo for 15 weeks, and then all groups crossed over to active treatment for a further 15 weeks. Significant improvements were observed on parent ratings of inattention, hyperactivity and impulsivity in both PUFA groups over 15 weeks, which were then mirrored in the placebo group after switching to active treatment for 15 weeks. Objective improvements were found in children's ability to switch and control their attention, and on vocabulary scores, and these mediated parent-reported improvements in hyperactivity, impulsivity and inattention [62]. This study also found that the fatty acid deficiency symptoms used as a selection criterion by Stevens et al. [59] improved in both the treatment and placebo groups over 15 weeks, and concluded that these symptoms are more likely to be an indicator of n-6 rather than n-3 deficiency [63].
A more recent study was conducted in Sweden with a similar supplement and study design as the previous two large trials, but including 75 children aged 8-18 years with ADHD [64]. They did not detect significant improvements overall; however, when dividing the sample into subgroups, they found that children with the inattentive subtype of ADHD and associated neurodevelopmental disorders, i.e. autistic symptoms and learning difficulties, showed similar clinical improvements as the previous large studies.
Therefore, there is evidence that children with ADHD-related symptoms may respond to PUFA supplementation and that those with the attention subtype and/or learning difficulties may be more likely to respond. Interestingly, studies using a combination of n-3 and n-6 fatty acids have been more effective than those with DHA or GLA (DGLA precursor) alone. However, the relative importance of EPA, DHA and GLA remain to be determined due to methodological differences between the studies. Our group has recently completed a 12 month trial with 7-12 year old children with ADHD in Adelaide and Brisbane, Australia, in a 3 x 3 crossover design comparing EPA-rich and DHA-rich oils with an LA placebo. At baseline higher DHA was associated with improved reading (although not quite significant after controlling for confounders) and higher AA with poorer reading, spelling, vocabulary and attention. Interestingly, children with learning difficulties (behind age level on literacy; n = 30 with bloods) had lower DHA and total n-3 and higher n-6 PUFA levels, although only the DHA relationship remained significant after controlling for differences in age and parent-reported health [65]. Preliminary results indicate that, although there were no significant between-group differences over 4-months, in children who consented to have blood taken increased erythrocyte DHA was associated with improved parent ratings of oppositional behaviour, anxiety/shyness, divided attention and reading (n = 45). In the subgroup with learning difficulties (n = 16 with bloods at both time points) increased erythrocyte DHA had even stronger associations with improved parent ratings of oppositional behaviour, hyperactivity, restlessness/impulsivity, divided attention, reading and spelling [66]. These preliminary data indicate that DHA supplementation may be associated with improved ADHD symptoms and that children with learning difficulties may be more likely to benefit, supporting a phospholipid spectrum hypothesis that suggests these symptoms occur on a continuum of overlapping and related developmental difficulties [54]. However, the latter sample size was small due to difficulties with recruitment/eligibility and need to be investigated further in larger samples.
Other randomized controlled trials with ADHD and n-3 PUFA are underway in Israel, the UK and France, as listed in Table 2.

Other Developmental Disorders
PUFA have also been implicated in other, often related developmental disorders such as autistic spectrum disorder [67,68] and dyslexia [69][70][71][72][73][74]. There is some evidence to suggest that autism may involve functional deficiencies or imbalances in n-3 PUFA. Studies on the profiles of PUFA in children with autism that have found defects of PUFA and phospholipids, the major constituent of cell membranes, have had conflicting results. Vancassel et al. examined plasma phospholipid fatty acids in a population of children with autism compared to intellectually disabled controls [75]. They reported a 23% reduction of DHA levels in the children with autism compared with controls. Reduced RBC levels of n-3 PUFA and increased levels of saturated fatty acids (SFA) have also been reported in individuals with autism [67,68]. Additionally, there is evidence that n-3 PUFA can assist in inflammatory bowel disorders [76], which are often found in children with autism and may play a role in its etiology [77].
Few studies have investigated effects of n-3 PUFA supplementation in autism. A case report noted improvements in symptoms of a boy with autism following EPA supplementation that was increased from 1g per day to 3g per day over four weeks, with improvements continuing after an eight-month follow-up [78]. An uncontrolled open-label study completed by 18 out of 20 children reported significant within-group improvements after three months of combined n-3, n-6 and n-9 PUFA supplementation (1g/day) [79]. Only one randomized controlled trial (RCT) trialing n-3 PUFA supplementation in autism has been published [80]. This six-week pilot RCT in 13 children with autism reported a trend for improvements in hyperactivity and stereotypy in the group receiving n-3 PUFA. Menhaden fish oil was used with daily doses of 840 mg EPA and 700 mg DHA.
A three-month crossover RCT was conducted in Durham by Dr Madeleine Portwood, and is yet to be published. The daily doses of PUFA used were 558 mg EPA and 174 mg DHA. Results regarding outcome measures were inconclusive and postulated to be due to failure to reach a ceiling effect of the PUFA over three months. After a one-way crossover, however, improvements were reportedly observed after six months of active n-3 PUFA supplementation. Therefore it was concluded that further studies examining active versus placebo treatment over six months are warranted. As with previous large studies that extended the treatment period in a one-way crossover [60,61], this study indicates the importance of the length of supplementation relative to dosage for improvements to be observed. Considering the complexity of autism, n-3 PUFA may also be best investigated as an adjunctive therapy. Therefore, studies are limited and results are inconclusive. A study is currently underway in the US investigating n-3 PUFA for treatment of hyperactivity in children with autism and another registered study aiming to investigate n-3 PUFA supplementation in children and adolescents with autistic spectrum disorder aimed to complete in 2007 but does not appear to have published results (see Table 2).

Mental Retardation
To date, fatty acid researchers have not focused specifically on the heterogeneous group of disorders that characterize mental retardation. Many people with mental retardation are likely to display co-morbid symptoms including hyperactivity, impulsivity, aggression, anxiety, and DSM-IV diagnosed psychiatric disorders such as bipolar disorder, schizophrenia [81] and ADHD [82]. It has been suggested that the high rate of psychiatric disorders in children and adolescents with mental retardation might be a major problem in their functioning and adaptation [83]. The impact of PUFA for the amelioration of learning, cognition and behavior control problems in developmental disorders like ADHD may have implications for the treatment of these same difficulties in people with mental retardation. This is important, particularly given that there are indications that some forms of mental retardation might be caused by defects in the structure and functioning of neural synapses [84], which are rich in the n-3 PUFA, DHA. For a more in-depth discussion see Sinn and Wilson [85].

Depression and Suicide Ideation
PUFA deficiency is also implicated in depression. Along with cardiovascular disease (CVD), depression seems to be more prevalent in countries where fish consumption is lower, after controlling for demographic variables and potentially confounding factors such as alcohol intake [86][87][88][89][90]. It is well established that depression often co-occurs with CVD, which is associated with elevated cholesterol and lower PUFA levels, and studies have confirmed that depression does not result from CVD as it has been found to exist before symptoms of CVD set in [91]. Indicative of a common underlying biological component, higher serum AA:EPA ratios were reported in depressed post-myocardial infarction patients [92] and lower total n-3 PUFA and DHA levels in depressed patients recovering from acute coronary syndromes [93] than in similar patients without depression. Shared PUFA abnormalities might further account for the depression that co-occurs with stress, post-natal depression, autoimmune diseases, multiple sclerosis, cancer, diabetes, aging and osteoporosis [91]. A number of studies have found lower n-3 PUFA levels or higher AA/EPA ratios in people with depression compared with controls [21,[94][95][96][97][98][99], including a postmortem comparison of 15 patients with major depressive disorder (MDD) versus 27 controls -specifically lower DHA levels in the frontal cortex [100].
A case study reported improvement in treatment-resistant depression when EPA was added to the conventional treatment, which was also associated with reduced neuronal phospholipid turnover and structural brain changes over the nine months of treatment [101]. Clinical trials have been previously reviewed [41,102], although new studies have since been published. Six trials used high EPA supplements. In a small dose-ranging study, treatment with 1g/day of ethyl-EPA but not 2 or 4g/day was effective in improving depressive symptoms in adults with ongoing depression whilst continuing their standard medication [103]. However, another small study reported positive treatment effects with 2g ethyl-EPA as an adjunctive therapy compared with placebo over four weeks [104]. This group also conducted a study with 6-12 year old children with major depressive disorder [105]. Only five of the twenty children were medicated, and following sixteen weeks of 400mg EPA plus 200mg DHA per day, the treatment group showed significant improvements in symptoms on all outcome measures.
It should be noted that blood samples were not collected in the latter studies. After eight weeks of supplementation with large doses of n-3 PUFA per day (4.4g EPA, 2.2g DHA) as an adjunctive therapy [106], depressed patients in the n-3 PUFA group had significantly lower depression scores than the control group (p < 0.001). Importantly, this team did collect blood samples and reported average baseline blood DHA levels of 2.5% of fatty acids in the placebo group and 2.4% in the treatment group (the latter increased to 5.8% in the DHA group after treatment). According to the Omega-3 Index for cardiovascular health the baseline levels correspond with the risk category for CVD (EPA + DHA: 4% of total membrane fatty acids) [11] and are therefore indicative of suboptimal n-3 PUFA levels. There is some recent evidence that 1 g/day EPA may have similar therapeutic effects to 20 mg/day fluoxetine as a treatment for major depression over eight weeks and that using the two treatments in combination is superior to either of them alone [107]. A larger recent study (N = 120) found no improvements in depressive symptoms following 1.5g ethyl-EPA plus 0.5g ethyl-DHA supplementation in middle-aged women with psychological distress over eight weeks [109], although they reported improvements in women who did not meet criteria for major depressive episode at baseline. A more recent study investigated treatment with 2.2g EPA and 1.2g DHA in Menhaden fish oil as a monotherapy for major depressive disorder in pregnant women and found significant improvements in the treatment group on all outcome measures compared with placebo [109]. This study is interesting because n-3 PUFA are preferentially transported to the growing fetus during pregnancy, which can deplete n-3 PUFA levels in pregnant mothers. An uncontrolled study also reported reductions in postpartum depression following n-3 PUFA supplementation over eight weeks [110].
Four studies using DHA or DHA-rich oils did not report improvements following n-3 supplementation. One of these studies supplemented patients with 2g DHA for only six weeks as a monotherapy [111]; one supplied 0.6g EPA + 2.4g DHA as an adjunctive therapy for 12 weeks [112] and the largest study gave a similar dosage as an adjunctive therapy for 16 weeks [113]. It should be noted though that two of these studies collected blood samples and revealed normal baseline n-3 levels; Marangell et al. [111] reported that the DHA group and placebo group had 4.15% and 3.78% DHA as a percentage of total fatty acids at baseline, respectively, and Grenyer et al. [113] reported baseline DHA as 4.2% of total RBC fatty acids. Although in the latter study reductions in depression scores did not differ between the groups, both treatment and placebo groups had received counseling, which is likely to account for improved mood in the whole sample. Furthermore, these authors later reported that increased DHA levels were significantly associated with improvements in depression scores [114], warranting further research into DHA and depression. Therefore studies using high EPA supplements should also determine to what extent the EPA has been converted to DHA and which of these is associated with improvements.
A larger study investigated depressive symptoms in a cohort of elderly people with mild-moderate Alzheimer's disease who were living in their own homes. This study (cognitive results reported further below) found no benefits of treatment on neuropsychiatric symptoms overall, although improved mood was detected in the treatment group compared with placebo in non apoE-4 allele carriers and reduced agitation in apoE-4 carriers [115]. Given the incidence of depression in Alzheimer's disease, and increased risk of depressed people developing dementia, this merits further investigation. Another recently published study investigated additive effects of supplementation with ethyl ester EPA and DHA on depressive symptoms in conjunction with sertraline in patients with major depression and coronary heart disease, finding no benefits of treatment over placebo [116].
In one of two further, larger population based studies, Rogers et al. randomized a community sample of 190 individuals with mild to moderate depression to a combined EPA + DHA supplement (1.5g per day) or olive oil placebo for 12 weeks [117]. They did not report any favourable effects of the supplement on depressive symptoms or a range of outcomes and concluded that this supports a negligible effect of n-3 on depression. A group from the Netherlands also found no effect of n-3 PUFA supplementation on mood in 302 independently living adults [118]. These studies did not use the Hamilton Depression Rating Scales (HDRS), which has been generally used as a primary outcome measure by the other studies, thereby reducing comparability. Although Rogers et al. took blood samples, they measured levels in plasma rather than RBC, and it should be noted that other studies reporting improvements in depression scores tended to use inpatients with more severe depression, many of whom were treatment resistant. Finally, some studies used a placebo lead-in phase to eliminate placebo responders [106], and others used patients who had depression scores indicating impaired function despite ongoing treatment with medication [103]. These selection criteria also could have influenced the outcomes: eliminating placebo responders may be pertinent to clinical trials in depression because often depressive symptoms can lift due to the attention associated with taking part in a clinical trial; non-responders to medication may be a subgroup more likely to respond to n-3 PUFA. Interestingly, Carney et al. did use a two-week placebo lead in phase but did not include nonresponders to medication, and improvements were seen in both treatment and placebo groups, possibly in response to the sertraline administered as an adjunctive therapy [116].
In summary, 13 randomized controlled trials that investigated effects of n-3 PUFA on depressive symptoms were identified, ranging from 4-26 weeks with sample sizes ranging from N = 20-302, in wide-ranging populations including psychiatric patients, coronary patients, pregnant women, elderly people with Alzheimer's disease and non-depressed community cohorts, one with children, some as adjunctive therapy to existing medication, using variable n-3 PUFA dosages, ratios of EPA:DHA and outcome measures. Five studies reported improvements overall, and two in subgroup analyses. Clearly more studies are needed with larger sample sizes and attention to selection criteria and likely responders, baseline n-3 PUFA levels, erythrocyte blood samples, length of supplementation and comparison of EPA vs. DHA-rich oils. A large number of randomized controlled trials are currently underway around the world investigating n-3 PUFA in depression, including two US studies that are comparing EPA and DHA -see Table 2 for a summary.

Bipolar Disorder
In addition to finding lower DHA levels in the frontal cortex of patients with MDD, McNamara et al. also identified lower DHA levels in the postmortem frontal cortex of 18 patients with bipolar depression compared with 19 age-matched controls [119]. One of the pioneering intervention trials reported on n-3 fatty acids and psychopathology was conducted by Stoll et al. [120] in bipolar disorder patients. The placebo-controlled, double-blind clinical trial produced a significantly greater response to treatment (64%) in the n-3 PUFA group than in those given olive oil placebos (18.8%). In this study, patients received fish oil capsules containing a total of 6.2g EPA and 3.4g DHA daily over four months, with no adverse effects overall. Since then another five studies have been conducted with bipolar patients. An open label study (N = 12) provided 1.5-2g ethyl EPA for 24 weeks as an adjunctive therapy [121]. Although there was no control group, at one month 7/10 of these previously treatment-resistant patients showed ≥ 50% reduction in HDRS scores and scores in five completers were dramatically reduced at six months. The other placebo-controlled studies provided ethyl EPA as an adjunctive therapy, supplying 1-2 g for 12 weeks [122] and 6g for 16 weeks [123]. Keck et al. reported no improvements in symptoms but Frangou et al. [122] reported significant improvements with 1g and 2g/day ethyl EPA over placebo in two out of three outcome measures. None of these studies took blood samples.
Frangou et al. [124] also reported increased brain levels of N-acetyl-aspartate (NAA), a presumed marker of neuronal integrity with 2g ethyl-EPA in 14 female outpatients compared with placebo, although this was not associated with improvements in symptoms. Treatment with n-3 PUFA in an uncontrolled study with a small group of women with bipolar disorder found no significant improvement over four weeks but also reported alterations in a marker of membrane fluidity in the n-3 groups compared with bipolar patients not taking n-3 PUFA and healthy controls [125]. These alterations in markers of brain function should be explored further, and perhaps in the context of multimodal approaches to treatment.
Depression and bipolar disorder are often associated with suicidal ideation. The relationship between PUFA, increased suicide risk and related disorders is considered in a review by Brunner and colleagues [126], who report an association between low cholesterol and higher risk of suicide. A recent study investigated n-3 PUFA supplementation, in addition to standard psychiatric care, on psychometric measures in adults with recurrent non-fatal self harm [127]. Daily supplementation with n-3 PUFA (1.2 g EPA and 0.9 g DHA) for 12 weeks improved psychopathology, specifically scores for depression, suicidality and daily stresses, compared with placebo in 49 participants.
In summary, two controlled studies have improved bipolar symptoms with a large dose combination of EPA and DHA over four months and 1-2g/day ethyl-EPA over 12 weeks, and there is some evidence from another study that this may impact on suicidal ideation. Again, further investigation of supplement formulation and blood analyses of erythrocyte n-3 PUFAs are needed to investigate relative efficacy of EPA vs. DHA vs. a combination of the two. Randomized controlled trials in bipolar depression are currently underway in Germany and the US (see Table 2).

Schizophrenia
There are a number of indications that phospholipid abnormalities may explain biological underpinnings of schizophrenia [26,[128][129][130][131]. Low n-3 fatty acid levels in diets or RBC have consistently been found in people with schizophrenic symptoms compared to controls [132][133][134][135][136]. Clinical trials with schizophrenia have reported mostly positive results. In an apparently open-label trial with 20 schizophrenic patients, significant improvements in symptoms were reported after daily supplementation with 10g fish oil over six weeks, and improvements were reported to correlate with increased n-3 PUFA blood levels [137]. Four double-blind controlled trials with EPA supplementation have produced some significant, positive outcomes [138][139][140] although in Peet et al. 2002 [140] only with 2g (vs. 1 and 4g) ethyl-EPA per day in the subgroup of patients taking clozapine and improvements were also associated with increased levels of AA.
Another trial found no differences between schizophrenic patients given ethyl EPA and those given placebo over 16 weeks [141]. Horrobin [142] proposed a number of reasons why an effect might not have been obtained, including the possibility that ethyl EPA may not have therapeutic benefits. Hibbeln et al. [143] investigated possible confounding variables using the same data set, and found that PUFA status in schizophrenic patients was predicted by smoking status, gender, and dietary intake of n-3 fatty acids. They concluded that phospholipid abnormalities may be an artefact of these variables. However, Emsley et al. [138] found significant differences between active and placebo groups on the Positive and Negative Syndrome Scale after controlling for age, gender, length of illness and dietary n-3 intake; further analyses revealed that these improvements were accounted for by improved dyskinesia.
In another recent study [132], schizophrenic patients were matched with healthy controls on a number of factors known to influence PUFA status including age, gender, dietary patterns, smoking and substance abuse. Schizophrenic patients still had significantly lower levels of EPA and DHA, and higher levels of AA than controls. They then gave patients in their study a combination of EPA, DHA and antioxidants (Vitamins E and C) over four months; there was no placebo group. The researchers reported significant within-group improvements on various measures of psychopathology between preand post-treatment, which continued for four months after the treatment was discontinued. These results need to be interpreted with caution given equivalent placebo to treatment effects found in other trials of both depression and schizophrenia. Two more recently published studies focused on firstepisode psychosis. The first study was with young adults who were provided 2g/day of ethyl-EPA or placebo along with psychotic medication (the majority had been previously unmedicated) [144]. Initial analyses found no treatment effects over 12 weeks; however did find improvements in patients with non-affective psychosis and in those who had had longer duration of untreated psychosis. Treatment effects were significant at 4-6 weeks then matched the placebo group at 12 weeks, indicating that the ethyl-EPA may have improved response to psychotic medication and/or the effects of the medication matched the ethyl-EPA by 12 weeks. The second study aimed to investigate whether 1.3g/day n-3 PUFA supplementation over 12 weeks could reduce the transition of sub-threshold psychosis to firstepisode psychotic disorder in adolescents and young adults over 12 months [145]. There was a significantly reduced transition to psychosis in the treatment group (2/41) compared with the placebo group (11/40). A couple of further trials in schizophrenia, one a multicentre intervention involving the latter research group, are currently underway (see Table 2).

Stress, Aggression, Hostility, Impulsivity, Criminal and Antisocial Behavior
A number of studies have indicated a biological basis for aggression and criminal-related behaviour. Low concentrations of cerebrospinal fluid 5-hydroxyindoleacetic acid (CSF 5-HIAA), a marker of cerebral serotonin turnover, have consistently been associated with impulsive, violent, suicidal, hostile and aggressive behaviors, as well as early onset of alcohol dependence and personality disorders [31,146]. Low PUFA status has also been associated with deficient serotonergic neurotransmission [28,29] and it is possible that low CSF 5-HIAA reflects this deficiency.
A small number of clinical trials have investigated the impact of PUFA supplementation on aggressive behaviour. A double-blind, placebo-controlled study measured effects of daily supplementation with oil containing 49.3% (1.5g) DHA, 6.7% EPA, 7.3% oleic acid, 3.3% AA and other oils for three months on cognitive outcomes and aggression in a group of 41 healthy Japanese students from two universities [147]. The placebo capsules contained small amounts of LA and ALA. No differences between groups were observed on cognitive tests of executive function, but a significant difference was reported in aggression towards others ('extraggression') at the end of the study, during a highly stressful final exam period. Extraggression scores significantly increased in the control groups and decreased (though not significantly) in the DHA supplemented groups. Serum blood analyses confirmed a significant increase in EPA and DHA levels in the treatment groups. Therefore three months' supplementation with n-3 PUFA seemed to assist students in controlling aggression towards others on a psychological test during a period of high stress. Drawing from research on stress-induced hostility and aggression, an Australian group investigated a possible adaptive role of DHA in perceived stress levels of university staff, finding no differences between treatment and placebo but positive effects in the treatment compared with a control group [148].
Another double-blind study tested effects of supplementation with 1g ethyl-EPA daily or placebo over eight weeks on moderate borderline personality disorder, characterized by emotional reactivity and impulsive aggression, in a group of women [149]. The treatment group had a significantly greater decrease in depression and aggression scores at the end of the study.
A landmark randomized, placebo-controlled trial was undertaken with 231 young adult prisoners to investigate effects of supplementation with micronutrients and essential fatty acids over four-five months on antisocial behavior [150]. The participants were given multi-vitamin/mineral capsules and fatty acid supplement providing n-3 and n-6 fatty acids (1260mg LA, 160mg GLA, 80mg EPA and 44mg DHA daily) or placebo. During the trial, the treatment group received significantly fewer disciplinary actions, and the greatest effect was noted for violent behavior. These studies add an important dimension to research with PUFA in ADHD, which is commonly associated with antisocial behavior symptomatic of conduct disorder and oppositional defiant disorder in childhood [151][152][153] and criminality and antisocial behavior in late adolescence and adulthood [154,155]. This work is currently being continued on a large scale with prisons in the UK, investigating effects of micronutrients and PUFA on anti-social behaviour in a randomized controlled trial over 4 months ('PINUP'; see Table 2).
Recently, a small randomized controlled trial in 24 men with substance abuse found that supplementation with 3g/day of n-3 PUFAs for three months improved self-reported scores of anger and anxiety compared to a soybean oil placebo [156]. The active group received five capsules per day each containing 450 mg of EPA, 100 mg of DHA and 50 mg of n-3 PUFAs ALA and DPA. Although five participants in both the placebo and active groups were on antidepressant medication during the study, their doses remained stable. Interestingly, the anger ratings remained lower in the active group for three months after supplementation ceased. Each participant's n-3 PUFA intake was calculated from a diet questionnaire covering the month preceding the study and it was reported that the participants had a daily intake of 148mg LC n-3 PUFAs, only 30% of the ISSFAL recommended intake of 500 mg/day. Furthermore, those participants who reported a history of violent behaviour had a significantly lower LC n-3 PUFA intake compared to those who had no history of this behaviour. After the 3 month treatment with LC n-3 PUFAs, a low anxiety score was associated with an increase in EPA but not DHA, whereas a low anger score was associated more strongly with DHA. This indicates possible variation between the two LC n-3 PUFAs in their effect on psychiatric conditions and warrants comparison of them in further research.
A randomized controlled trial is currently investigating the role of n-3 PUFA in alleviating psychological stress in university students undertaking exams, and another study in Spain is assessing n-3 PUFA supplementation for symptoms of borderline personality disorder (Table 2).

Dementia and Alzheimer's Disease
Free radical activity and phospholipid membrane degeneration have been associated with neurodegenerative disorders. With aging, neural membrane fluidity is generally compromised due to an increase in cholesterol, reduced activity level of desaturase enzymes, impaired phospholipid metabolism and increased oxidative stress [20]. Such derangements may be associated with dementia and Alzheimer's Disease (AD), e.g., autopsied brains of patients who suffered AD showed significantly higher saturated fat and lower PUFA content, particularly DHA, in the hippocampus and frontal lobes compared to aged controls [157], which is consistent with reports of decreased hippocampus size and function in AD patients [20].
Prospective cohort/population studies have indicated that higher fish consumption is associated with reduced risk of dementia/AD [158][159][160][161][162][163][164][165]. A single case report of an elderly dementia patient noted clinical improvement over several months of increased fish consumption, which led the author to hypothesise that n-3 PUFA contributed to his improved functioning, although it could have been attributed to numerous other factors associated with admittance to a nursing home [166]. Investigation of n-3 PUFA status in AD patients did find significantly reduced EPA and DHA levels compared to controls, particularly DHA levels which were consistently less than half those in the control group [167].
Relatively few controlled studies have investigated effects of n-3 PUFA supplementation on dementia to date. In a short four-week double blind trial with 100 AD patients, a number of improvements in symptoms were reported in 49 of the 60 patients who received an n-3/n-6 supplement comprising 0.5g of ALA and LA in a 1:4 ratio [168].
Supplementation with 4.32g DHA daily for one year improved dementia scores of 10 elderly volunteers with vascular dementia compared with 10 controls who did not receive capsules [169]. Another pilot study in 19 participants with mild to moderate AD treated with 1g/day ethyl-EPA for 12 weeks showed no improvement in cognitive functioning, possibly due to the short period of supplementation and/or the nature of the supplement formulation [170]. Supplementation with 240mg/day of DHA and AA improved immediate memory and attention in 12 adults with mild cognitive impairment (MCI) but not eight adults with AD compared to an olive oil MCI placebo group [171]. In the largest study to date with 174 volunteers with diagnosed AD, only those with very mild AD (n = 27) showed reductions in cognitive decline after adjunctive supplementation with 1700 mg/day DHA and 600 mg/day EPA for six months and the placebo group improved after switching to n-3 PUFA for a further 6 months [172]. Similarly, a study recently published online also reported improvements relative to placebo in MCI but not AD patients following supplementation with 1.8g/day n-3 PUFA (1080mg EPA, 720mg DHA) over 24 weeks [173]. One population based study investigated the effect of DHA and EPA on cognitive performance of 302 elderly people > 65 years without dementia and detected no treatment effects over 26 weeks, although it did report improvements in attention in apoE-4 carriers and males [174]. The prior studies indicate that early stages of cognitive decline may be optimal for intervention with n-3 PUFA. MCI is considered a risk factor for AD, with an estimated 16-41% of people with MCI converting to a diagnosis of AD each year and up to 64% after two years. It has been suggested therefore that MCI is possibly the earliest stage of detectable AD [175]. Furthermore, major and minor depression often co-occur in MCI [176] and increase the risk of progressing to AD [177].
Further investigation of the role of PUFA in dementia is needed, focusing on larger sample sizes, the nature of the supplement (e.g., EPA versus DHA), dosage, length of supplementation, baseline RBC n-3 PUFA proportions, response in ApoE ε4 allele carriers, longer-term outcomes and therapeutic implications. Finally, cardiovascular risk factors for AD have led to recent suggestions that lifestyle factors such as inadequate physical activity and/or n-3 PUFA intake may contribute to reduced cerebral blood flow and blood-brain barrier integrity [37], supporting the notion that one of the primary functions of n-3 PUFA on mental health may be improved circulatory function [36]. This should also be explored in future research.
We are currently collecting data for a six-month study comparing EPA with DHA-rich supplements versus an LA placebo on mood and cognition in elderly people with MCI. Other controlled trials are underway to assess n-3 PUFA supplementation and cognition in healthy older adults and those with dementia (see Table 2 for details).

Conclusions
There is growing evidence that suboptimal intakes of n-3 PUFA may be associated with psychopathology over the lifespan and include highly prevalent disorders that present a growing public health concern. Most clinical trials have been conducted with major depressive illness in adulthood and in childhood disorders; there are growing numbers of interventions with cognitive decline in older adulthood. Although the causes of these mental health problems are complex and multifactorial, even from a nutritional perspective alone [e.g., 40,178], dietary and lifestyle factors including n-3 PUFA present modifiable risk factors that can be accessed relatively easily by individuals.
However, findings of clinical trials have been inconsistent and in many cases inconclusive; methodological differences between studies need to be critically evaluated before drawing conclusions about the efficacy or otherwise of n-3 PUFA in alleviating symptoms. These include sample size, selection criteria, the dosage of the supplement, the nature of the fatty acids (i.e. EPA-or DHA-rich or combined oil; inclusion of n-6 PUFA), and length of supplementation. Given the widespread use of ethyl-esthers, short-and long-term investigation of bioavailability compared with naturally occurring fatty acids in fish oil is warranted. The role of n-6 PUFA also needs to be clarified as some studies found improved and some worse outcomes associated with AA, and successful studies with children included evening primrose oil in the treatment. Some studies in depression have included inpatients with clinical depression despite ongoing medication. Non-responders to other treatments are a possible future area of attention for n-3 PUFA and/or micronutrient supplementation because suboptimal nutrient levels may explain this lack of response to other treatments. Some depression studies have conducted a placebo-lead in phase to eliminate placebo responders and this may also assist in clarifying true treatment effects. Importantly, studies need to identify baseline n-3 PUFA levels in erythrocyte membranes because this may be a major contributor to different findings between studies. This will also assist researchers and clinicians in identifying likely responders.
Outcome measures also need consideration; as can be seen in Table 1, a great variety of psychological/cognitive outcome measures have been used across these studies and their sensitivity to detecting effects of n-3 PUFA supplementation is varied. Of particular interest are two large studies in the UK [60,150]. In most studies with children, parent and teachers have reported improvements; however cognitive assessments detected varied (mostly no) improvements. In the Durham trial [60], reading and spelling were included as outcome measures with highly significant, and meaningful, improvements in these literacy indicators. In the prison-based trial with criminal offenders [150], a range of psychological assessments were used that did not detect any response to micronutrient/n-3 PUFA supplementation; however prison disciplinary actions were recorded in the treatment and placebo groups and detected significantly less reprimands in the treatment group, particularly in violent behaviour. Therefore selection of outcome measures is an area in serious need of attention.
Finally, only one study appears to have conducted a follow-up; four months after treatment ceased, anger levels remained low [156]; whether this was due to retention of n-3 PUFA in cell membranes or a result of newly learned behavioral changes remains to be determined and also needs to be investigated by future researchers. In conclusion, as well as identifying likely responders, future research investigating effects of n-3 PUFA supplementation on mental health and behavior needs to define meaningful and long term outcomes and optimal supplementation strategies including combinations with other lifestyle risk factors and nutrients that are consumed suboptimally.

Notes added in proof
After writing this review another study was identified by a group in Israel (Raz et al., 2009). They published a randomised controlled trial with 7-13 year old unmedicated children with ADHD that reported no treatment effects on parent and teacher questionnaires or a computerised continuous performance task. However the PUFA supplement contained relatively small amounts of LA (480mg) and ALA (120mg), the latter of which is likely to have had minimal conversion to EPA and DHA, and they were only supplemented for 7 weeks, whereas at least 8 and preferably a minimum of 12 weeks tend to be required to show any improvements overall, and with larger doses of long-chain n-3 PUFAs. Interestingly they did report some improvements in both groups which could be a placebo effect although the placebo contained vitamin C and was therefore not a non-active compound. Raz, R.; Carasso, R. L.; Yehuda, S. The Influence of Short-Chain Essential Fatty Acids on Children with Attention-Deficit/Hyperactivity Disorder: A Double-Blind Placebo-Controlled Study. J Child Adolesc Psychopharmaco, 2009, 19, 167-177.