Is there a role for n-3 long-chain polyunsaturated fatty acids in the regulation of mood and behaviour? A review of the evidence to date from epidemiological studies, clinical studies and intervention trials

K. M. Appleton; P. J. Rogers; A. R. Ness

doi:10.1017/S0954422408998620

Is there a role for n-3 long-chain polyunsaturated fatty acids in the regulation of mood and behaviour? A review of the evidence to date from epidemiological studies, clinical studies and intervention trials

Published online by Cambridge University Press: 01 June 2008

K. M. Appleton ,

P. J. Rogers and

A. R. Ness

Show author details

K. M. Appleton*: Affiliation:
School of Psychology, Queen's University of Belfast, Belfast BT9 5BP, UK
P. J. Rogers: Affiliation:
Department of Experimental Psychology, University of Bristol, Bristol BS8 1TU, UK
A. R. Ness: Affiliation:
Department of Oral and Dental Science, University of Bristol, Bristol BS1 2LY, UK
*: *Corresponding author: Dr K. M. Appleton, fax +44 28 9066 4144, email [email protected]

Article contents

Abstract
Introduction
n-3 Long-chain polyunsaturated fatty acids
n-3 Long-chain polyunsaturated fatty acids in cell membrane structure and function
n-3 Long-chain polyunsaturated fatty acids in mood and behaviour
Depression, depressive illness and suicidal behaviour
Anxiety
Fatigue
Aggression, hostility and anti-social behaviour
Inattention, hyperactivity, impulsivity and attention deficit hyperactivity disorder
Schizophrenic disorders
Other mood and behavioural disorders
n-3 Long-chain polyunsaturated fatty acids in mood and behaviour – evaluation
References

Rights & Permissions

Abstract

Selected biochemical evidence suggests a potential role for n-3 long-chain PUFA (n-3PUFA) in the regulation of mood and behaviour. The present paper reviews the relevant evidence, to date, from epidemiological studies, clinical studies and intervention trials. Most evidence is available investigating a role for n-3PUFA in depression, depressive illness and suicidal behaviour, but work is also available on anxiety and anxiety-related disorders, fatigue and fatigue-related disorders, aggression, hostility and anti-social behaviour, inattention, impulsivity and attention deficit hyperactivity disorder and schizophrenic disorders. For all these aspects of mood and behaviour, the evidence available is currently limited and highly inconsistent, both in terms of study methodology and study findings. There is a clear need for further work in this area.

Keywords

n-3 Long-chain polyunsaturated fatty acids Mood Behaviour

Type: Research Article
Information: Nutrition Research Reviews , Volume 21 , Issue 1 , June 2008 , pp. 13 - 41

DOI: https://doi.org/10.1017/S0954422408998620 [Opens in a new window]
Copyright: Copyright © The Authors 2008

Introduction

Increasing interest surrounds the possibility that n-3 long-chain PUFA (n-3PUFA) may be implicated in the regulation of mood and behaviour. This interest stems from evidence of effects of n-3PUFA on cell membrane structure and function, and early evidence suggesting a role for n-3PUFA in the development and/or treatment of various disorders of mood and behaviour. More recent evidence, however, questions the robustness of this early evidence. The present paper aims to provide a comprehensive review of the evidence to date investigating a role for n-3PUFA in the regulation of mood and behaviour.

n-3 Long-chain polyunsaturated fatty acids

n-3PUFA (also named omega-3 fatty acids) are a family of PUFA, named as such due to the positioning of the first double carbon bond on the third atom from the methyl end of the acyl chain. All members of the family are derived from the parent fatty acid 18: 3n-3 (α-linolenic acid (ALA)), via desaturation and elongation⁽Reference Haag¹^, Reference Ruxton, Calder, Reed and Simpson²⁾, as demonstrated in Fig. 1. Closely related to the n-3PUFA are the n-6 long-chain PUFA (n-6PUFA), named from the positioning of the first double bond on the sixth carbon atom from the methyl end of the acyl chain. n-6PUFA are derived from the parent fatty acid 18 : 2n-6 (linoleic acid), and for synthesis share the same desaturases and elongases as n-3PUFA. The n-3 and n-6PUFA thus compete for synthesis from their parent fatty acids. The parent fatty acids, however, cannot be synthesised by man⁽Reference Haag¹^, Reference Ruxton, Calder, Reed and Simpson²⁾.

Fig. 1 Desaturation and elongation of n-3 long-chain PUFA.

As essential fatty acids, ALA and linoleic acid must be obtained from the diet. Longer-chain n-3 and n-6PUFA can be formed in man, but biological conversion is slow and inefficient, making diet an important source for these fatty acids as well⁽Reference Ma, Folsom, Eckfeldt, Lewis and Chambless³⁾. Dietary sources of ALA include certain nuts and seeds, such as walnuts, flaxseed and rapeseed oil, and dietary sources of the longer n-3PUFA EPA and DHA include fatty fish, some white fish, shellfish and other seafoods such as seaweed, and certain eggs and animal products dependent on the animals' diet⁽Reference Ruxton, Calder, Reed and Simpson²^, ⁴^–Reference Simopoulos⁶⁾. Dietary sources of linoleic acid and all n-6PUFA include plant and vegetable seeds and oils, as found in margarines and the majority of processed foods⁽Reference James, Gibson and Cleland⁵^, Reference Simopoulos⁶⁾. Dietary intakes of n-3 and n-6PUFA, however, have changed dramatically over recent decades. Our traditional diet is thought to have contained approximately equal amounts of energy from n-3PUFA and n-6PUFA⁽Reference Simopoulos⁶⁾. By comparison, a current Western diet is estimated to contain approximately five to twenty times more energy from n-6PUFA than from n-3PUFA⁽Reference Simopoulos⁶^, Reference Gregory, Foster, Tyler and Wiseman⁷⁾. The increasing imbalance between n-3 and n-6PUFA is currently thought to impact on cell membrane structure and function.

n-3 Long-chain polyunsaturated fatty acids in cell membrane structure and function

Both n-3 and n-6PUFA are integral components of all cell membranes as part of the phospholipid bilayer. Within the phospholipid bilayer, n-3PUFA and n-6PUFA can be interchanged, where incorporation into the membrane depends largely on PUFA availability, both from the diet and from chemical synthesis⁽Reference Haag¹^, Reference Ruxton, Calder, Reed and Simpson²^, Reference James, Gibson and Cleland⁵⁾. Incorporation into cell membranes, dependent on availability, has been clearly demonstrated⁽Reference Bourre, Faivre, Dumont, Nouvelot, L'Oudes, Puymirat and Tixier-Vidal⁸^, Reference Bourre, Pascal, Durand, Masson, Dumont and Piciotti⁹⁾. Membrane composition, however, affects membrane function. First, n-3PUFA molecules, due to their size and shape, influence the physical state of the membrane, resulting in increased fluidity and permeability. Ehringer et al. ⁽Reference Ehringer, Belcher, Wassall and Stillwell¹⁰⁾ found DHA administration to result in increased permeability of the phospholipid bilayer in vitro, Tappia et al. ⁽Reference Tappia, Ladha, Clark and Grimble¹¹⁾ found supplementation with fish oil to result in increases in membrane fluidity in rats and Hirashima et al. ⁽Reference Hirashima, Parow, Stoll, Demopulos, Damico, Rohan, Eskesen, Zuo, Cohen and Renshaw¹²⁾ found supplementation with n-3PUFA to result in a decrease in T₂ values (brain water proton transverse relaxation times) indicative of increased membrane fluidity in humans. This increased fluidity may aid cross-cell membrane transport, aiding cell communication and functionality⁽Reference Haag¹⁾.

Second, n-3 and n-6PUFA are also thought to have different effects on surrounding molecules and cell functions. Specifically, n-3 and n-6PUFA are considered important in cell signalling and signal transduction⁽Reference Haag¹^, Reference Ruxton, Calder, Reed and Simpson²^, Reference James, Gibson and Cleland⁵⁾. n-3 and n-6PUFA are thought to impact directly on the activity of a number of enzymes in various neurotransmitter pathways, resulting in changes in the activities of these systems⁽Reference Haag¹⁾. n-3PUFA-deficient diets have been associated with lower levels of serotonin and dopamine in piglets⁽Reference De la Presa Owens and Innis¹³⁾. In rats, n-3PUFA-deficient diets have been found to result in increased serotonin 5-HT₂ receptor density in the frontal cortex⁽Reference Delion, Chalon, Herault, Guilloteau, Besnard and Durand¹⁴^, Reference Delion, Chalon, Guilloteau, Besnard and Durand¹⁵⁾, elevated 5-HT_2A receptor binding density in the prefrontral cortex⁽Reference McNamara, Richtand and Levant¹⁶⁾, lower levels of dopamine in the cortex, hippocampus and striatum⁽Reference Delion, Chalon, Herault, Guilloteau, Besnard and Durand¹⁴^, Reference Delion, Chalon, Guilloteau, Besnard and Durand¹⁵^, Reference Takeuchi, Fukumoto and Harada¹⁷⁾, a decreased density of D₂ receptors in the frontal cortex⁽Reference Delion, Chalon, Herault, Guilloteau, Besnard and Durand¹⁴^, Reference Delion, Chalon, Guilloteau, Besnard and Durand¹⁵⁾, reduced D₂ receptor binding density in the prefrontal cortex⁽Reference McNamara, Richtand and Levant¹⁶⁾, decreased activity in the mesocortical dopamine pathway and increased activity in the mesolimbic dopamine pathway⁽Reference Zimmer, Vancassel, Cantagrel, Breton, Delamanche, Guilloteua, Durand and Chalon¹⁸⁾, and higher levels of noradrenalin in the cortex, hippocampus and striatum⁽Reference Takeuchi, Fukumoto and Harada¹⁷⁾, compared with control diets. Supplementation with n-3PUFA has also been found to result in enhanced 5-HT responsivity⁽Reference Yao, Magan, Sonel, Gurklis, Sanders and Reddy¹⁹⁾ and reductions in noradrenalin⁽Reference Hamazaki, Itomura, Huan, Nishizawa, Sawazaki, Tanouchi, Watanabe, Hamazaki, Terasawa and Yazawa²⁰^, Reference Sawazaki, Hamazaki, Yazawa and Kobayashi²¹⁾ in humans.

n-3 and n-6PUFA are also thought to affect enzymes which result in the release of fatty acids from the phospholipid bilayer to form a number of eicosanoids, prostaglandins and leucotrienes. These compounds can have effects on signal transduction, resulting again in increased activity and increased cell signalling⁽Reference Haag¹^, Reference Ruxton, Calder, Reed and Simpson²^, Reference James, Gibson and Cleland⁵⁾. These compounds can also have either pro- or anti-inflammatory properties depending on their derivation – prostaglandins formed from the n-6PUFA arachidonic acid are typically pro-inflammatory, prostaglandins formed from EPA are typically anti-inflammatory⁽Reference Ruxton, Calder, Reed and Simpson²^, Reference James, Gibson and Cleland⁵^, Reference Calder²²⁾. Supplementation with EPA has been found to result in reduced production of inflammatory cytokines in animals and in humans⁽Reference James, Gibson and Cleland⁵^, Reference Calder²²⁾. Low levels of docosapentaenoic acid n-3 in human erythrocyte membranes have also been associated with high levels of the pro-inflammatory cytokine IL-6⁽Reference Yao, Sistilli and van Kammen²³⁾, and supplementation with ALA has been found to result in reductions in inflammatory cytokines – TNFα and IL- 1β⁽Reference Caughey, Mantzioris, Gibson, Cleland and James²⁴⁾, and reductions in IL-6, C-reactive protein and serum amyloid A⁽Reference Rallidis, Paschos, Liakos, Velissaridou, Anastasiadis and Zampelas²⁵⁾.

n-3 and n-6PUFA have also been found to modulate ion channels important in cell signalling and transmission. Ion transfer is vital for neurotransmission, and n-3PUFA have been associated with an inhibition of enzymes that maintain ion gradients⁽Reference Kearns and Haag²⁶⁾. Deficiencies in dietary ALA have also been found to result in a reduction of neural enzyme activity in the rat brain⁽Reference Bourre, Francois, Youyou, Dumont, Piciotti, Pascal and Durand²⁷⁾, and deficiencies in dietary DHA have been found to result in decreased neuron size in several areas of the rat brain⁽Reference Ahmad, Moriguchi and Salem²⁸⁾, diminished nerve growth factor levels in the hippocampus and increased nerve growth factor in the piriform cortex⁽Reference Ikemoto, Nitta, Furukawa, Ohishi, Nakamura, Fujii and Okuyama²⁹⁾.

The serotonergic, dopaminergic and adrenergic neurotransmitter systems are known to be important in the regulation of mood and behaviour⁽Reference Haag¹⁾. The anti-inflammatory effect of n-3PUFA are also thought to be important in a number of behavioural conditions⁽Reference James, Gibson and Cleland⁵⁾. The biochemical evidence thus suggests a potential role for n-3PUFA in the regulation of mood and behaviour, and has resulted in the development of a number of hypotheses centring around a role for n-3PUFA in a number of mood and behavioural conditions. These hypotheses include the biogenic amine hypothesis of depression⁽Reference Hibbeln and Salem³⁰⁾, the immune/inflammatory hypothesis of psychiatric disease⁽Reference Smith³¹⁾ and the membrane hypothesis of schizophrenia⁽Reference Horrobin³²⁾.

While biochemical work continues to examine these various hypotheses, epidemiological, clinical and trial evidence investigating a role for n-3PUFA in the regulation of mood and behaviour is also available.

n-3 Long-chain polyunsaturated fatty acids in mood and behaviour

To date, studies have investigated the effects of n-3PUFA in relation to various aspects of mood and behaviour. The majority of this work has focused on a role for n-3PUFA in the development and treatment of depression and a variety of depressive illnesses, but work is also available on anxiety and anxiety-related disorders, aggression, hostility and anti-social disorders, inattention, impulsivity and attention deficit hyperactivity disorder (ADHD) and other psychiatric symptoms and disorders, such as schizophrenia. The present review will consider each of these in turn. Where studies measure more than one aspect of mood, these studies are included separately in the discussion of each condition. Aspects of cognitive function, such as vigilance, concentration and disorders of cognitive function, such as Alzheimer's' disease, while closely associated with mood and behaviour, are not covered.

Depression, depressive illness and suicidal behaviour

Depression is characterised by high levels of depressed or low mood, a loss of interest or pleasure in nearly all activities, changes in appetite, weight, sleep or activity, decreased energy, difficulties thinking, concentrating or making decisions, feelings of worthlessness or guilt, and recurrent thoughts of death or suicidal ideation, plans or attempts. Depressive disorders are defined by two consecutive weeks of depressed mood or loss of interest in nearly all activities, plus verification of four additional symptoms⁽³³⁾. Epidemiological, clinical and trial evidence investigating a role for n-3PUFA in depression and depressive illness is available.

Epidemiological evidence

Studies investigating the association between the dietary intake of n-3PUFA and depression are given in Table 1⁽Reference Hibbeln³⁴^–Reference Lovibond and Lovibond⁵⁹⁾. Four ecological studies have found negative linear and non-linear associations between national fish consumption and national prevalence of major depression⁽Reference Hibbeln³⁴^, Reference Peet³⁹⁾, postpartum depression⁽Reference Hibbeln³⁶⁾ and bipolar disorders⁽Reference Noaghiul and Hibbeln³⁷⁾. These studies, however, use crude population (rather than individual) measures of n-3PUFA intake and depressive illness, and few potential confounders of n-3PUFA intake and depression are considered.

Table 1 Epidemiological evidence investigating a role for n-3 long-chain PUFA (n-3PUFA) in depression and depressed mood

DSM, Diagnostic and Statistical Manual of Mental Disorders; EPDS, Edinburgh Postnatal Depression Scale⁽Reference Cox, Holden and Sagovsky⁵⁸⁾; SES, socio-economic status; FOASTAT, Food and Agriculture Organization of the United Nations Statistics; BDI-21, Beck Depression Inventory (twenty-one-item version)⁽Reference Beck and Steer⁵³⁾; SF-36, MOS Short Form Health Survey⁽Reference Ware and Sherbourne⁵⁴⁾; ALA, α-linolenic acid; OTA, octadecatetraenoic acid; ETA, eicosatetraenoic acid; DPA, docosapentaenoic acid; HADS-D, Hospital Anxiety and Depression Scales (depression scale)⁽Reference Zigmond and Snaith⁵⁵⁾; HSCL-25, Hopkins Symptom Checklist (twenty-five-item version)⁽Reference Winokur, Winokur, Rickels and Cox⁵⁶⁾; CES-D, Center for Epidemiologic Studies (depression scale)⁽Reference Radloff⁵⁷⁾; n-6PUFA, n-6 long-chain PUFA; DASS-21, Depression, Anxiety and Stress Scales (twenty-one-item version)⁽Reference Lovibond and Lovibond⁵⁹⁾.

The majority of epidemiological studies investigating n-3PUFA intakes and depression have been conducted within individuals. Studies have used FFQ and diet recalls to record fish consumption, seafood consumption and the whole diet, and all studies measured self-reported depression, although different questionnaires have been used. Of the eleven (within-individual) studies reported in Table 1, eight studies found a negative association between fish or n-3PUFA intake and depression⁽Reference Tanskanen, Hibblen, Hintikka, Haatainen, Honkalampi and Viinamaki⁴¹^–Reference Silvers and Scott⁴³^, Reference Suzuki, Akechi, Kobayashi, Taniguchi, Goto, Sasaki, Tsugane, Nishiwaki, Miyaoka and Uchitomi⁴⁶^–Reference Barberger-Gateau, Jutand, Letenneur, Larrieu and Tavernier⁴⁸^, Reference Appleton, Peters, Hayward, Heatherley, McNaughton, Rogers, Gunnell, Ness and Kessler⁵⁰^, Reference Appleton, Woodside and Yarnell⁵¹⁾, although in two studies, associations were only found in female participants⁽Reference Tanskanen, Hibbeln and Tuomilehto⁴²^, Reference Timonen, Horrobin, Jokelainen, Laitinen, Herva and Rasanen⁴⁷⁾, and in one study associations were found for fish intake but not for fish +n-3PUFA supplement intake⁽Reference Appleton, Peters, Hayward, Heatherley, McNaughton, Rogers, Gunnell, Ness and Kessler⁵⁰⁾, whereas in another study associations were found for calculated total n-3PUFA intake but not for fish intake⁽Reference Suzuki, Akechi, Kobayashi, Taniguchi, Goto, Sasaki, Tsugane, Nishiwaki, Miyaoka and Uchitomi⁴⁶⁾. Three studies found no associations⁽Reference Hakkarainen, Partonen, Haukka, Virtamo, Albanes and Lonnqvist⁴⁴^, Reference Jacka, Pasco, Henry, Kotowicz, Nicholson and Berk⁴⁵^, Reference Miyake, Sasaki and Yokoyama⁴⁹⁾. In seven of the studies that found associations, the relationship between n-3PUFA intake and depression was found following adjustment for confounders⁽Reference Tanskanen, Hibblen, Hintikka, Haatainen, Honkalampi and Viinamaki⁴¹^–Reference Silvers and Scott⁴³^, Reference Suzuki, Akechi, Kobayashi, Taniguchi, Goto, Sasaki, Tsugane, Nishiwaki, Miyaoka and Uchitomi⁴⁶^–Reference Barberger-Gateau, Jutand, Letenneur, Larrieu and Tavernier⁴⁸^, Reference Appleton, Woodside and Yarnell⁵¹⁾. In three of these studies, however, associations were reduced following adjustment for confounders⁽Reference Silvers and Scott⁴³^, Reference Suzuki, Akechi, Kobayashi, Taniguchi, Goto, Sasaki, Tsugane, Nishiwaki, Miyaoka and Uchitomi⁴⁶^, Reference Appleton, Woodside and Yarnell⁵¹⁾, and in one study no association between n-3PUFA intake and depressed mood remained following adjustment for the confounders, age and deprivation⁽Reference Appleton, Peters, Hayward, Heatherley, McNaughton, Rogers, Gunnell, Ness and Kessler⁵⁰⁾.

Clinical evidence

Clinical studies typically investigate associations between n-3PUFA status and depressive illness, and either compare individuals suffering from depressive symptoms with controls, or investigate the continuous relationship between n-3PUFA status and depressive symptom severity. Details of studies investigating differences between those with depressive symptoms and those without are given in Table 2⁽Reference Ellis and Sanders⁶⁰^–Reference Kaiya, Horrobin, Manku and Morse-Fisher⁸³⁾. The majority of studies have compared individuals with diagnosed clinical depression and controls with no depression, although some studies have also included individuals with undiagnosed, sub-clinical depression⁽Reference Maes, Smith, Christophe, Cosyns, Desnyder and Meltzer⁶²^, Reference Tiemeier, van Tuijl, Hofman, Kiliaan and Breteler⁶⁶⁾, and some studies have used self-report measure cut-offs to define high levels of depression as opposed to clinical diagnosis⁽Reference Mamalakis, Tornaritis and Kafatos⁷³^–Reference Mamalakis, Jansen, Cremers, Kiriakakis, Tsibinos and Kafatos⁷⁶⁾. Studies on postpartum depression⁽Reference Otto, de Groot and Hornstra⁷⁷^, Reference Browne, Scott and Silvers⁷⁸⁾, suicide attempt⁽Reference Huan, Hamazaki, Sun, Itomura, Liu, Kang, Watanabe, Terasawa and Hamazaki⁷⁹⁾, bipolar disorder⁽Reference Chiu, Huang, Su, Lu, Huang, Chen and Shen⁸⁰^, Reference Ranjekar, Hinge, Hegde, Ghate, Kale, Sitasawad, Wagh, Debsikdar and Mahadik⁸¹⁾, self-harm⁽Reference Garland, Hallahan, McNamara, Carney, Grimes, Hibbeln, Harkin and Conroy⁸²⁾ and affective and paranoid disorders⁽Reference Kaiya, Horrobin, Manku and Morse-Fisher⁸³⁾ are also available. Studies used a variety of different biological samples for assessment of n-3PUFA status. The majority of studies involving individuals diagnosed with clinical depression show low levels of a number of n-3PUFA and high ratios of n-6PUFA:n-3PUFA in depressed individuals compared with controls. All studies, however, involve assays of a number of fatty acids, where associations are found for some fatty acids and not others, with no consistent patterns emerging for those comparisons that yield associations or those that do not. Low levels of n-6PUFA are also often reported, again in inconsistent patterns, and few studies use adjusted P values to take account of multiple testing. Two studies also demonstrate higher levels of n-3PUFA in depressed individuals compared with controls⁽Reference Ellis and Sanders⁶⁰^, Reference Fehily, Bowey, Ellis and Meade⁶¹⁾. Comparisons of individuals with other depressive illnesses reveal some differences between cases and controls, where cases have lower levels of some n-3PUFA, but again patterns of associations are inconsistent⁽Reference Huan, Hamazaki, Sun, Itomura, Liu, Kang, Watanabe, Terasawa and Hamazaki⁷⁹^–Reference Kaiya, Horrobin, Manku and Morse-Fisher⁸³⁾. Comparisons between individuals with high and low levels of undiagnosed depression reveal few differences between groups⁽Reference Mamalakis, Tornaritis and Kafatos⁷³^–Reference Browne, Scott and Silvers⁷⁸⁾.

Table 2 Clinical evidence investigating a role for n-3 long-chain PUFA (n-3PUFA) in depression and depressed mood: comparisons between depressed cases and non-depressed controls

ALA, α-linolenic acid; DPA, docosapentaenoic acid; LA, linoleic acid; AA, arachidonic acid; C, clinical; PCPG, plasma choline phosphoacylglycerols; ↑ , higher PUFA in depressed cases compared with comparison; x, no association; ↓ , lower PUFA in depressed cases compared with comparison; EEPG, erythrocyte ethanolamine phosphoacylglycerols; ECPG, erythrocyte choline phosphoacylglycerols; SC, sub-clinical; PL, plasma phospholipids; NC, non-clinical; CE, plasma cholesteryl esters; EM, erythrocyte membranes; CES-D, Center for Epidemiologic Studies (depression scale)⁽Reference Radloff⁵⁷⁾; CRP, C-reactive protein; ACS, acute coronary syndromes; MI, myocardial infarction; mol*, concentration calculated with references to the internal standard 18-methylnonadecanoic acid; L, plasma lipids; AT, adipose tissue; ZSDS, Zung Self-Rating Depression Scale⁽Reference Biggs, Wylie and Ziegler⁸⁴⁾; BDI, Beck Depression Inventory⁽Reference Beck and Steer⁵³⁾; HADS-D, Hospital Anxiety and Depression Scales (depression scale)⁽Reference Zigmond and Snaith⁵⁵⁾; GDS-15, Geriatric Depression Scale (fifteen-item version)⁽Reference Sheikh and Yesavage⁸⁵⁾; EPDS, Edinburgh Postnatal Depression Scale⁽Reference Cox, Holden and Sagovsky⁵⁸⁾.

* Other n-3PUFA tested but no associations found.

† Results following adjustment for multiple testing.

‡ Other n-3PUFA tested and associations found.

Studies investigating relationships between n-3PUFA status and depressive symptomatology are given in Table 3⁽Reference Beck and Steer⁵³^, Reference Zigmond and Snaith⁵⁵^, Reference Radloff⁵⁷^, Reference Cox, Holden and Sagovsky⁵⁸^, Reference Fehily, Bowey, Ellis and Meade⁶¹^, Reference Maes, Smith, Christophe, Cosyns, Desnyder and Meltzer⁶²^, Reference Edwards, Peet, Shay and Horrobin⁶⁴^, Reference Maes, Christophe, Delanghe, Altamura, Neels and Meltzer⁶⁵^, Reference Parker, Heruc, Hilton, Olley, Brotchie, Hadzi-Pavlovic, Friend, Walsh and Stocker⁶⁷^, Reference Mamalakis, Tornaritis and Kafatos⁷³^, Reference Mamalakis, Kiriakakis, Tsibinos and Kafatos⁷⁴^, Reference Mamalakis, Jansen, Cremers, Kiriakakis, Tsibinos and Kafatos⁷⁶^, Reference Otto, de Groot and Hornstra⁷⁷^, Reference Huan, Hamazaki, Sun, Itomura, Liu, Kang, Watanabe, Terasawa and Hamazaki⁷⁹^, Reference Garland, Hallahan, McNamara, Carney, Grimes, Hibbeln, Harkin and Conroy⁸²^, Reference Biggs, Wylie and Ziegler⁸⁴^–Reference Beck, Schuyler, Herman, Beck, Resnick and Lettieri⁹⁷⁾. These studies show similar patterns to those above. Negative associations between depressive symptoms and n-3PUFA status, and positive associations between depressive symptoms and n-6PUFA:n-3PUFA balance have been found, but again patterns with individual n-3PUFA are inconsistent and the majority of assays do not find associations.

Table 3 Clinical evidence investigating a role for n-3 long-chain PUFA (n-3PUFA) in depression and depressed mood: associations between n-3PUFA status and depressive symptom severity

ALA, α-linolenic acid; DPA, docosapentaenoic acid; LA, linoleic acid; AA, arachidonic acid; C, clinical; BDI, Beck Depression Inventory⁽Reference Beck and Steer⁵³⁾; PCPG, plasma choline phosphoacylglycerols; +, positive correlation; x, no correlation; HDRS, Hamilton Depression Rating Scale⁽Reference Hamilton⁹³⁾; CE, plasma cholesteryl esters; PL, plasma phospholipids; EM, erythrocyte membranes; − , negative correlation; DMI-18, Depression in the Medically Ill (eighteen-item version)⁽Reference Parker, Hilton, Bains and Hadzi-Pavlovic⁹⁴⁾; NC, non-clinical; ZSDS, Zung Self-Rating Depression Scale⁽Reference Biggs, Wylie and Ziegler⁸⁴⁾; AT, adipose tissue; CES-D, Center for Epidemiologic Studies (depression scale)⁽Reference Radloff⁵⁷⁾; GDS-15, Geriatric Depression Scale (fifteen-item version)⁽Reference Sheikh and Yesavage⁸⁵⁾; EPDS, Edinburgh Postnatal Depression Scale⁽Reference Cox, Holden and Sagovsky⁵⁸⁾; IBS, irritable bowel syndrome; SCL, Symptom Checklist (ninety-item version)⁽Reference Arrindell and Ettema⁹⁵⁾; HADS-D, Hospital Anxiety and Depression Scales (depression scale)⁽Reference Zigmond and Snaith⁵⁵⁾; MADRS, Montgomery–Asberg Depression Rating Scale⁽Reference Montgomery and Asberg⁹⁶⁾; mol*, concentration calculated with references to the internal standard 18-methylnonadecanoic acid; SIS, Suicide Intent Scale⁽Reference Beck, Schuyler, Herman, Beck, Resnick and Lettieri⁹⁷⁾; L, plasma lipids.

* Other n-3PUFA tested but no associations found.

† Other n-3PUFA tested and associations found.

‡ No correlation, but significant negative predictor in regression model.

$ Significant correlation and significant positive predictor in regression model.

A few clinical studies have also investigated the association between n-3PUFA intake and depressive illness in groups of patients. One study again found negative associations between n-3PUFA intake and depression⁽Reference Edwards, Peet, Shay and Horrobin⁶⁴⁾, although no association between n-3PUFA intake and a number of depressive illnesses and behaviours has also been reported⁽Reference Fehily, Bowey, Ellis and Meade⁶¹^, Reference Browne, Scott and Silvers⁷⁸^, Reference Huan, Hamazaki, Sun, Itomura, Liu, Kang, Watanabe, Terasawa and Hamazaki⁷⁹⁾. These studies, however, are typically small, often fail to account adequately for confounding factors as above, and due to the difficulty of accurately measuring usual diet, are far from conclusive.

Trial evidence

Trials measure the effects of n-3PUFA supplementation either compared with placebo (placebo-controlled trials) or with no comparison (open-label trials). A number of open-label trials have investigated the impact of supplementation with n-3PUFA on depression. Studies are typically small (between six and thirty-seven participants), and the majority of studies have found decreases in depression or depressive symptoms. Freeman et al. found decreases in diagnosed major depression⁽Reference Freeman, Hibbeln, Wisner, Watchman and Gelenberg⁹⁸⁾ and diagnosed postpartum depression⁽Reference Freeman, Hibbeln, Wisner, Brumbach, Watchman and Gelenberg⁹⁹⁾ following different doses of EPA+DHA, Osher et al. ⁽Reference Osher, Bersudsky and Belmaker¹⁰⁰⁾ found decreases in bipolar depression following supplementation with EPA, and Wozniak et al. ⁽Reference Wozniak, Biederman, Mick, Waxmonsky, Hantsoo, Best, Cluette-Brown and Laposata¹⁰¹⁾ found decreases in diagnosed childhood bipolar depression following supplementation with EPA+DHA. Wozniak et al. ⁽Reference Wozniak, Biederman, Mick, Waxmonsky, Hantsoo, Best, Cluette-Brown and Laposata¹⁰¹⁾ also found decreases in mania and psychotic symptoms following supplementation and Sagduyu et al. ⁽Reference Sagduyu, Dokucu, Eddy, Craigen, Baldassano and Yildiz¹⁰²⁾ found decreases in mania and bipolar symptoms following supplementation with EPA+DHA. Case reports of treatment with n-3PUFA for depressive disorders have also yielded benefits⁽Reference Chiu, Huang, Shen and Su¹⁰³^, Reference Puri, Richardson and Horrobin¹⁰⁴⁾. Marangell et al. ⁽Reference Marangell, Martinez, Zboyan, Chong and Puryear¹⁰⁵⁾, however, found no benefits of EPA+DHA on depression in a non-clinical sample of women with a history of postpartum depression, and Kaplan et al. ⁽Reference Kaplan, Matar, Kamin and Cohen¹⁰⁶⁾ found no benefits of EPA on depression in a sample diagnosed with post-traumatic stress disorder.

Placebo-controlled trials investigating the effects of n-3PUFA on depression and depressive symptoms are shown in Table 4⁽Reference Hirashima, Parow, Stoll, Demopulos, Damico, Rohan, Eskesen, Zuo, Cohen and Renshaw¹²^, ³³^, Reference Beck and Steer⁵³^, Reference Lovibond and Lovibond⁵⁹^, Reference Llorente, Jensen, Voigt, Fraley, Berretta and Heird⁸⁹^, Reference Hamilton⁹³^, Reference Montgomery and Asberg⁹⁶^, Reference Nemets, Stahl and Belmaker¹⁰⁷^–Reference Coccaro, Harvey, Kupsaw-Lawrence, Herbert and Bernstein¹³⁷⁾. Most studies have involved individuals diagnosed with major depression, but some studies have involved individuals diagnosed with bipolar disorders⁽Reference Hirashima, Parow, Stoll, Demopulos, Damico, Rohan, Eskesen, Zuo, Cohen and Renshaw¹²^, Reference Stoll, Severus, Freeman, Rueter, Zboyan, Diamond, Cress and Marangell¹¹⁴^–Reference Keck, Mintz and McElroy¹¹⁶⁾ and some studies have involved volunteers without diagnosis of depression⁽Reference Llorente, Jensen, Voigt, Fraley, Berretta and Heird⁸⁹^, Reference Ness, Gallacher, Bennett, Gunnell, Rogers, Kessler and Burr¹¹⁷^–Reference Rogers, Appleton, Kessler, Peters, Gunnell, Hayward, Heatherley, Christian, McNaughton and Ness¹¹⁹⁾ or individuals with other psychiatric conditions⁽Reference Behan, Behan and Horrobin¹²⁰^–Reference Hallahan, Hibbeln, Davis and Garland¹²⁶⁾. Of the studies conducted in individuals with diagnosed major depression, two studies found decreases in depression following supplementation with EPA+DHA compared with placebo⁽Reference Su, Huang, Chiu and Shen¹¹⁰^, Reference Nemets, Nemets, Apter, Bracha and Belmaker¹¹²⁾, two studies found decreases in depression following supplementation with EPA ethyl ester (E-EPA) compared with placebo⁽Reference Nemets, Stahl and Belmaker¹⁰⁷^, Reference Peet and Horrobin¹⁰⁸⁾, although Peet & Horrobin⁽Reference Peet and Horrobin¹⁰⁸⁾ only found benefits for 1 g E-EPA, and not for 2 g E-EPA or 4 g E-EPA, and three studies found no differences between treatment and placebo groups following supplementation with DHA⁽Reference Marangell, Martinez, Zboyan, Kertz, Seung Kim and Puryear¹⁰⁹⁾ or EPA+DHA⁽Reference Silvers, Woolley, Hamilton, Watts and Watson¹¹¹^, Reference Grenyer, Crowe, Meyer, Owen, Grigonis, Caputi and Howe¹¹³⁾. Of the studies involving individuals with bipolar disorder, two studies found decreases in depression and in bipolar symptoms following supplementation with EPA+DHA compared with placebo⁽Reference Stoll, Severus, Freeman, Rueter, Zboyan, Diamond, Cress and Marangell¹¹⁴⁾ or E-EPA compared with placebo⁽Reference Frangou, Lewis and McCrone¹¹⁵⁾, although neither study found similar improvements in mania. Two studies, however, found no benefit of supplementation with EPA+DHA⁽Reference Hirashima, Parow, Stoll, Demopulos, Damico, Rohan, Eskesen, Zuo, Cohen and Renshaw¹²⁾ or E-EPA⁽Reference Keck, Mintz and McElroy¹¹⁶⁾ for depression, mania or bipolar symptoms. All four studies involving volunteers with no diagnosis of depression also found no differences between treatment and placebo groups following supplementation with DHA⁽Reference Llorente, Jensen, Voigt, Fraley, Berretta and Heird⁸⁹⁾, EPA+DHA⁽Reference Rogers, Appleton, Kessler, Peters, Gunnell, Hayward, Heatherley, Christian, McNaughton and Ness¹¹⁹⁾, EPA+DHA+other n-3PUFA⁽Reference Fontani, Corradeschi, Felici, Alfatti, Bugarini, Fiaschi, Cerretani, Montorfani, Rizzo and Berra¹¹⁸⁾ or advice to eat fish or supplementation with EPA⁽Reference Ness, Gallacher, Bennett, Gunnell, Rogers, Kessler and Burr¹¹⁷⁾. Of the studies involving individuals with other psychiatric conditions, one study found a beneficial effect of EPA+DHA compared with placebo on depression in chronic fatigue patients⁽Reference Behan, Behan and Horrobin¹²⁰⁾, one study found a beneficial effect of E-EPA compared with placebo on depression in patients with borderline personality disorder⁽Reference Zanarini and Frankenburg¹²⁴⁾ and one study found a benefit of EPA+DHA compared with placebo on depression in individuals who self-harm⁽Reference Hallahan, Hibbeln, Davis and Garland¹²⁶⁾. Hallahan et al. ⁽Reference Hallahan, Hibbeln, Davis and Garland¹²⁶⁾ also found a beneficial effect of EPA+DHA supplementation on the presence or absence of suicidal ideation. However, one study also found no improvement in depression following EPA+DHA supplementation in chronic fatigue patients⁽Reference Warren, McKendrick and Peet¹²¹⁾, one study found no improvement in depression following supplementation with E-EPA in patients with obsessive–compulsive disorder⁽Reference Fux, Benjamin and Nemets¹²⁵⁾ and two studies found no improvements in depression following E-EPA supplementation in schizophrenic patients⁽Reference Fenton, Dickerson, Boronow, Hibbeln and Knable¹²²^, Reference Peet and Horrobin¹²³⁾.

Table 4 Trial evidence investigating a role for n-3 long-chain PUFA (n-3PUFA) in depression and depressed mood: placebo-controlled trials

C, clinical; E-EPA, ethyl ester EPA; HDRS, Hamilton Depression Rating Scale⁽Reference Hamilton⁹³⁾; MADRS, Montgomery–Asberg Depression Rating Scale⁽Reference Montgomery and Asberg⁹⁶⁾; BDI, Beck Depression Inventory⁽Reference Beck and Steer⁵³⁾; GAF, Global Assessment of Functioning⁽³³⁾; HDRS-SF, Hamilton Depression Rating Scale, short form⁽Reference Reynolds and Kobak¹²⁷⁾; CDRS, Children's Depression Rating Scale⁽Reference Poznanski, Cook and Carroll¹²⁸⁾; CDI, Children's Depression Inventory⁽Reference Kovacs, Beck, Schulterbrandt and Raskin¹²⁹⁾; CGI, Clinical Global Impression⁽Reference Guy¹³⁰⁾; YMRS, Young Mania Rating Scale⁽Reference Young, Biggs, Ziegler and Meyer¹³¹⁾; NOS, not otherwise specified; IDS-C, Inventory of Depressive Symptomology⁽Reference Rush, Gullion, Basco, Jarrett and Trivedi¹³²⁾; CGI-BP, Clinical Global Impression – bipolar disorder⁽Reference Spearing, Post, Leverich, Brandt and Nolen¹³³⁾; NC, non-clinical; EPDS, Edinburgh Postnatal Depression Scale⁽Reference Cox, Holden and Sagovsky⁵⁸⁾; SCID-CV, Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders (DSM)-IV axis I disorders – clinical version⁽Reference First, Spitzer, Gibbon and Williams¹³⁴⁾; DSP, Derogatis Stress Profile⁽Reference Derogatis¹³⁵⁾; w-s, within-subjects; POMS, Profile of Mood States (depression question)⁽Reference McNair, Lorr and Droppleman¹³⁶⁾; DASS, Depression, Anxiety and Stress Scales (depression scale)⁽Reference Lovibond and Lovibond⁵⁹⁾; MOAS, Modified Overt Aggression Scale⁽Reference Coccaro, Harvey, Kupsaw-Lawrence, Herbert and Bernstein¹³⁷⁾.

*Advised to eat more fatty fish (mackerel, herring, kipper, pilchard, sardine, salmon, trout) or given EPA capsules; recommended dose is not reported.

Evaluation

Evidence of a role for n-3PUFA in depressive disorders is inconclusive. Epidemiological studies provide some evidence that n-3PUFA intake is associated with depressed mood, but not all studies have found associations, and some studies have found effects that have subsequently disappeared on consideration of confounders. Clinical studies also provide some evidence that depression may be associated with reduced n-3PUFA status, although the results are inconclusive. With some exceptions, the majority of studies are small, conducted on highly selected samples, and studies that find no associations between n-3PUFA or reverse associations are also available.

Evidence on the effects of n-3PUFA supplementation on depression is also inconclusive. Greatest evidence for a beneficial effect of supplementation with n-3PUFA can be found in the studies involving individuals diagnosed with major depression, but even here studies finding no benefit are also available. Evidence from individuals with bipolar disorders and other psychiatric conditions is equivocal. Evidence from the three trials involving individuals with non-diagnosed depression suggests no benefit of n-3PUFA supplementation on depression in these individuals.

Four recent meta-analyses have attempted to evaluate the evidence from placebo-controlled trials investigating a role for n-3PUFA in depression⁽Reference Appleton, Hayward, Gunnell, Peters, Rogers, Kessler and Ness¹³⁸^–Reference Appleton, Rogers, Ness and Heikkinen¹⁴¹⁾. Differences between analyses exist dependent on inclusion criteria, but all four suggest a beneficial effect of n-3PUFA for depressive illness – combined effect sizes range from 0·13 (95 % CI 0·01, 0·25) to 0·61 (95 % CI 0·21, 1·01). All four meta-analyses also report clear heterogeneity between study findings, and the three conducted following a systematic review of the published literature⁽Reference Appleton, Hayward, Gunnell, Peters, Rogers, Kessler and Ness¹³⁸^, Reference Lin and Su¹⁴⁰^, Reference Appleton, Rogers, Ness and Heikkinen¹⁴¹⁾ also suggest considerable publication bias, where small studies reporting positive findings are more likely to be published than small studies showing negative findings⁽Reference Appleton, Hayward, Gunnell, Peters, Rogers, Kessler and Ness¹³⁸⁾. The heterogeneity and publication bias in these analyses argue for caution when interpreting the overall effect sizes. The combined effect size and heterogeneity from the most recent meta-analysis conducted⁽Reference Appleton, Rogers, Ness and Heikkinen¹⁴¹⁾ are clearly demonstrated in the relevant Forest plot (see Fig. 2). A beneficial effect of n-3PUFA compared with placebo was also found by Lin & Su⁽Reference Lin and Su¹⁴⁰⁾ when combining trials investigating a role for n-3PUFA in bipolar disorder (combined effect size 0·69 (95 % CI 0·28, 1·10), although heterogeneity and publication bias were also found. However, no differences between treatment and placebo were found by Appleton et al. ⁽Reference Appleton, Hayward, Gunnell, Peters, Rogers, Kessler and Ness¹³⁸⁾ when combining trials investigating a role for n-3PUFA in individuals with no diagnosis of depressive disorder (combined standardised mean difference − 0·13 (95 % CI − 0·29, 0·02) sd).

Fig. 2 Forest plot for the meta-analysis of all randomised controlled trials investigating the effects of n-3 long-chain PUFA on depressed mood up to September 2007 (taken from Appleton et al. ⁽Reference Appleton, Rogers, Ness and Heikkinen¹⁴¹⁾).

Anxiety

Anxiety is defined as a state of uneasiness or tension caused by apprehension of possible misfortune or danger⁽¹⁴²⁾. Anxiety disorders include panic attacks, phobias, specific anxiety disorders, obsessive–compulsive disorder and generalised anxiety disorder. All are characterised by an intense or persistent apprehension, worry, fearfulness or terror⁽³³⁾.