Key words

FormalPara Key Points

  • Over the past 30 years the implications of long-chain omega-3 fatty acids (n-3 LC-PUFAs) in the development and for the treatment of mental illnesses such as psychotic disorders have been explored.

  • The emerging body of evidence indicates the possible relevance of a dysregulation in polyunsaturated fatty acid (PUFAs) homeostasis in the pathophysiology of these disorders, with a convincing biological plausibility based on the involvement of PUFAs in underlying structural and functional neuronal mechanisms related to mental illnesses.

  • Although there is a lack of primary efficacy as an augmenting agent in chronic schizophrenia, longer-term effects in later stages, such as metabolic or extra-pyramidal side effects, might be positively influenced by PUFAs.

  • Most promising, however, is the increasing evidence pointing towards the potential of n-3 LC-PUFAs for indicated prevention, i.e., delaying transition to psychosis in high-risk populations.

Introduction

Schizophrenia and other psychoses are heterogeneous and severe psychiatric disorders, with a worldwide prevalence of schizophrenia of 1–2% (1). Psychotic disorders are characterized by positive (e.g., delusions, hallucinations), negative (e.g., affective flattening, alogia, avolition), and cognitive symptoms (e.g., disorganized thinking, low concentration), all of which largely affect psychosocial well-being and functioning.

Their etiology appears complex, including both genetic and environmental factors, and is not yet fully understood. The vast majority of research has focused on the role of the dopaminergic system, and the efficacy of dopamine D2-receptor blocking antipsychotics seemed to validate this hypothesis. However, this hypothesis does not necessarily serve to explain all findings and observations in individuals with psychotic disorders. Therefore, alternative hypotheses, in particular with relation to brain development, brain structural abnormalities, and structural remodelling, have been postulated. With the increasing awareness of the crucial role of polyunsaturated fatty acids (PUFAs) for brain development and functioning, implications of PUFA alterations in the development and for the treatment of psychotic disorders, such as schizophrenia and ­schizophrenia-spectrum disorders, have been explored over the past 30 years. Meanwhile, the emerging body of evidence indicates the possible relevance of a dysregulation in PUFA homeostasis in the pathophysiology of these disorders. It has been suggested that PUFA dysregulation and the dopamine hypothesis are merely two different facets of the same underlying etiopathology (2).

The available controlled PUFA supplementation studies are often small and seem to yield conflicting results (3, 4) across the different stages of psychotic illnesses: Whereas the efficacy as an augmenting agent in chronic schizophrenia might seem almost futile at first (5), epidemiological data, as well as some recent controlled studies in emerging psychosis, point towards possible preventive effects of PUFAs in early and very early stages of psychotic disorders (6–8). Furthermore, potential secondary or tertiary beneficial long-term effects in later, more chronic stages, in particular for metabolic or extrapyramidal side effects have been explored.

Physiology of PUFAs

What Are PUFAS and How Do Genes and Environment Influence PUFA Physiology?

Cholesterol and phospholipids are the main components of neuronal membrane bilayers. Phospholipids consist of a glycerol or sphingosine backbone with three carbon molecules. A phosphate molecule and a hydrophilic head group, such as choline, ethanolamine, inositol, or serine are attached to the third position. Fatty acids are the hydrocarbon tails ranging from 16 to 24 carbon molecules attached to the first and second positions. The number of double bonds defines their degree of saturation. A higher number of double bonds corresponds to a higher membrane fluidity. Since mammals have lost the ability to introduce a double bond beyond carbon 9, humans cannot synthesize these fatty acids de novo, making the parent molecules of n-3 and n-6 PUFAs, alpha-linolenic acid (ALA, 18:3 n-3) and linoleic acid (LA, 18:2 n-6), essential fatty acids (EFAs). EFAs serve as substrates for further conversion to long-chain PUFAs (LC-PUFAs) with higher numbers of carbon atoms and double bonds. Desaturases and elongases are involved in the multi-stage process of LC-PUFA synthesis from the parent substances, whereby n-3 PUFAs and n-6 PUFAs compete for the same enzymes. Desaturases are responsible for the introduction of a double bond in the acyl chain, whereas the alternating elongases are the rate-limiting steps of the chain-extension process. In humans, only membrane-bound fatty acid desaturases and fatty acid elongases are found and thus determine the structure of fatty acids (9).

Since the conversion rates from ALA and LA are limited in humans, some LC-PUFAs, such as arachidonic acid (AA, 20:4 n-6) and eicosapentaenoic acid (EPA, 20:5 n-3), need to be ingested with AA- and EPA-rich foods. Nonessential fatty acids, in particular the nonessential omega-9 fatty acids, replace EFA in times of unavailability, which, however, changes the phospholipid structure (10, 11). A number of factors, which are also associated with psychotic disorders, further decrease conversion rates to LC-PUFAs; among these are very young and old age (12), catecholamines and glucocorticoids (13), viral infections (14), and male sex (15, 16).

Nutritional sources of LC-PUFAs such as EPA and docosahexaenoic acid (DHA, 22:5, n-3) include fatty fish (e.g., salmon), certain white fish, shellfish and other seafood, as well as some animal products depending on the animals’ diet. Sources of ALA include nuts and seeds, such as walnuts, flaxseed, and rapeseed oil (17, 18). For the infant, breast milk also constitutes a source of intake (19). Plant and vegetable seeds and oils, such as margarines and those included in processed foods, provide LA and all other n-6 PUFAs. Over the hominid evolution and even more rapidly over the last century, dietary intakes of n-3 and n-6 PUFAs have most probably changed significantly. Traditionally, the human diet was composed by approximately equal amounts of energy deriving from n-3 and n-6 PUFAs and much lower amounts of saturated fats (20), whereas the current Western diet is high in n-6 PUFAs as well as saturated fats (21).

In addition to dietary intake, genetic factors further regulate PUFA status. Recently, a genome-wide association study found associations between plasma levels of omega-3 and omega-6 fatty acids and three desaturases genes (FADS1–3) on chromosome 11. The strongest association was found with a specific SNP (rs174537) near FADS1 that has the highest association with AA metabolism; it was also associated with levels of eicosadiaenoic acid (EDA) and EPA. Another region with a strong association was an elongase-encoding region on chromosome 6 (ELOVL2), which was associated with EPA (22). Several further candidate gene studies confirmed associations between FADS gene variations and PUFA levels (23, 24). In particular, the association of FADS1 gene variations (rs174556, rs174561, and rs3834458) with serum phospholipid PUFAs seems to explain a large degree of variation in AA levels and membrane PUFA levels (25, 26).

PUFAs and the Brain

The human brain’s grey matter is enriched with AA and DHA, which together account for 20% of the membrane fatty acids. In particular, AA and DHA concentrate in larger amounts in synaptic membrane compared to other membranes (27). DHA is also abundant in the retina (28). EPA and DHA exert different physiological effects on neuronal cell membranes. EPA acts on the brain via its role as a precursor for eicosanoids, whereas DHA itself has been shown to alter densities of dopamine, serotonin, and muscarinic receptors (29, 30).

The availability of these PUFAs determines phospholipid structure, which in turn impacts on functioning, making AA, EPA, and DHA particularly important for the specific properties and functions of neuronal cells. n-3 and n-6 PUFAs have also been shown to directly affect enzymes in neurotransmitter pathways with resulting changes in the latters’ activities (31). Animal studies have shown that diets deficient in n-3 PUFAs result in altered monoaminergic transmission in regions of the brain related to schizophrenia, such as the prefrontal cortex and the mesolimbic system (32). In humans, supplementation with n-3 PUFAs has been shown to enhance serotonin (5-HT) responsivity and decrease noradrenalin (33, 34).

Brain development (35), synaptic functioning (36), regulation of corticotrophin-­releasing hormone (37), prevention of neuronal apoptosis (38, 39), neurite growth (40), gene expression (41), as well as neuronal migration, pruning, and synaptic plasticity (36) are further processes regulated by n-3 PUFAs (42). Furthermore, phospholipids and thus their fatty acid components are crucial to many cell-signalling systems such as protein kinases (43) as well as mechanisms related to calcium, diacylglycerol, and cyclic nucleotide (44, 45), all of which are linked to monoaminergic neurotransmission in the brain.

The tissue fatty acid composition changes during brain development. The main body of LC-PUFAs is localized in the diacyl glycerophospholipids (GPL), mainly in phosphatidylethanolamine (PE) and phosphatidylserine (PS). Alkenylacyl GPLs (plasmalogens), which are crucial to brain myelogenesis, also include a bulk of LC-PUFAs in the mature brain. Paralleling the mid-to-late gestational increase of DHA in the human fetal brain, oleic acid (18:1 n-9), an unsaturated fatty acid predominant in phosphatidylcholine and PS, and adrenic acid (22:4 n-6) increase in plasmalogens during the myelogenesis of the fetal brain. In fact, oleic acid and its products, particularly nervonic acid (24:1 n-9), are the main components of sphingolipids—cerebrosides, sulfatides, and sphingomyelin. Hence, the rise in adrenic, oleic, and nervonic acid within fatty acid composition might be a marker for brain myelination. Indeed, white matter alterations may represent an early correlate of an increased vulnerability for psychosis (144). More detailed information on the development of n-3 fatty acid status is available in a different chapter of this book.

Effects of omega-3 fatty acids on the brain can be measured in vivo with magnetic resonance imaging (MRI) by measuring brain water proton transverse relaxation times (T2), an index of the water in neuronal tissues. Reduced neuronal health has been shown to be associated with T2 increases (46). As a matter of fact, studies in patients with bipolar disorders (47) and first episode psychosis (FEP) using T2 relaxation time (46) support preclinical findings on neuroprotective properties of EFAs.

Another way of measuring the effects of n-3 fatty acids on the brain is proton magnetic resonance spectroscopy (1H-MRS), which gives estimates of regional concentrations of different brain metabolites. The underlying idea is that alterations in N-acetylaspartate (NAA) reflect changes in neuronal integrity; furthermore, variations in glutamate/glutamine (GLX) are linked to excitotoxicity, and finally, that changes in glutathione (GSH) are associated with oxidative stress or apoptosis (48, 49). Glutathione acts as an essential defense mechanism against excess nitric oxide (NO), which in turn is a mediator of inflammation; GSH further serves to protect glial and neuronal mitochondria (50, 51). Indeed, in the context of an E-EPA augmentation study in FEP, 1H-MRS findings suggest that EPA augmentation results in a 30% increase in glutathione and a 20% increase in glutamate/glutamine, suggesting that EPA supplementation has a major impact on GSH availability and GLX cycle modulation and may be mediated via neuro- and gliaprotective effects of EPA (52).

What Do We Learn from Epidemiology?

Fish Intake and Mental Health

PUFAs play a crucial role in neuronal systems and accordingly have the ability to interfere with neural development and functioning (36). Ecological and epidemiological studies provide some insight into possible effects on diverse aspects of mental health. Studies using crude population measures of n-3 intake found inverse correlations between national fish consumption and prevalence of major depression (53, 54), ­postpartum depression (55), and bipolar disorders (56). The large majority of the positive studies could maintain their findings on inverse correlation even after adjustment for ­confounders (57–64). A similar negative association with n-3 intake was found in an epidemiological study of hostility and DHA intake (65). More detailed information on this subject is discussed in a different chapter of this book.

Fish Intake and Psychotic Disorders

Few epidemiological studies regarding possible associations between n-3 PUFA intake and schizophrenia are available. Christensen and Christensen investigated the relationship between course and outcome of schizophrenia and the amount of fat in the average national diets discriminated for different sources of fats in eight countries. They report that a higher intake of fat from vegetables, fish, and seafood is associated with a more favorable course and outcome of schizophrenia, whereas the incidence of schizophrenia was not associated with dietary factors (66). These findings may point towards a modulating effect of n-3 deficiencies on the course of illness, but may also just be an epiphenomenon associated with underlying factors such as an unhealthy lifestyle, urbanicity, stress, or migration. Similar results seemed to be found in an ecological study by Peet (67); however, multiple regression analyses showed that refined sugar was the dietary constituent that was most strongly associated with a worse outcome of schizophrenia. In a non-epidemiological pilot study the latter group reported an association of the severity of psychotic symptoms with intake of PUFAs that was, unlike sugar consumption, independent of medication effects (68). Noaghiul and Hibbeln, however, found no association between seafood intake and prevalence rates of schizophrenia in 14 different countries (56).

Recently, a Swedish group published data regarding the prevalence of psychotic-like symptoms in a cohort of over 33,000 women from the general population with respect to dietary intake of PUFAs (69). The risk of having so-called “high-level psychotic-like symptoms” (i.e., frequent, ≥ “often” or “almost always” answers) was approximately 50% lower in women who ate fish three to four times per week compared to women who never ate fish. Specifically, the risk was lower for women with a high intake of n-3 and especially n-6 PUFAs compared to those with a lower intake of PUFAs. Interestingly, the strongest inverse relationship was found for an intermediate fish intake of one to three times per week. Women with a high intake of fatty fish (>5 times per week) had an increased risk of psychotic-like symptoms. The explanations offered by the authors for this intriguing finding are possible unhealthy constituents of fatty fish, such as environmental pollutants, and the idea of a therapeutic window for n-3 LC-PUFAs regarding their neuroprotective role (70, 71). Hedelin et al. also found that the risk-reducing effect of PUFA intake on the prevalence of psychotic-like symptoms was most pronounced for n-6 PUFAs. Also, intermediate, but not higher, doses of EPA increased levels of AA in red cell membranes, and the EPA-induced rise in AA was associated with clinical improvement. These findings suggest that the balance between the n-3 and n-6 PUFA ratio might be more important than absolute values alone.

Fish Intake and Early Brain Development

During mid- to late gestation the fetal brain size increases alongside an accumulation of DHA (72). In such, low maternal n-3 intake has been shown to correlate with lower verbal IQ, and impaired social and verbal development and fine motor functions (75). However, similar to the findings of Christensen and Christensen described previously, high DHA maternal serum levels (i.e., n-3 PUFA rich marine foods) also are associated with low birthweight and impaired myelination (73, 74). The potential consequences of both low and high gestational DHA levels resemble those often observed in schizophrenia. Regarding the importance of PUFAs for brain development, maternal fish and seafood consumption of less than 340 g per week during pregnancy has been reported to increase the risk for a low IQ and deficient neurodevelopmental outcomes in childhood (75). The threshold of 340 g per week was chosen on the basis of US federal advice that was issued for pregnant women in 2004 to restrict seafood consumption to 340 g per week to avoid fetal exposure to methylmercury (76). Some (77, 78) but not all (79–81) studies have yielded findings of protective effects of EFAs in pregnancy and breastfeeding against schizophrenia and related disorders. One study reported that the duration of breastfeeding was positively associated with age of onset of schizophrenia (82). Some of the negative findings were population-based studies, whereas the positive ones investigated high-risk populations with family histories of schizophrenia, very premature birth, birth complications, viral infections, or extreme malnutrition. Possibly, the protective role of EFAs in breast milk might be relevant only for risk populations.

Markers of Altered PUFA Metabolism

From Clinical Observations to the Membrane-Phospholipid Hypothesis

A range of clinical observations is suggestive that the modulation of inflammatory mediators and its precursors such as LA, AA, EPA, and DHA have important clinical implications for schizophrenia and related disorders. A negative association between the prevalence of rheumatoid arthritis—an illness associated with excessive production of AA and its derivatives—and mental disorders, specifically schizophrenia, has been shown (83). Furthermore, impaired sensitivity to pain and temperature (84) and symptomatic improvement after development of fever (85) led to the formulation of the prostaglandin-deficiency hypothesis of schizophrenia (86) comprising the idea of a deficit in AA-derived PGs. In a further elaboration of this original idea, David Horrobin formulated the “membrane-phospholipid hypothesis” (MPH) of schizophrenia in 1994 (87), later also introduced as a possible biochemical basis of the neurodevelopmental concept of schizophrenia (88). The MPH proposes that brain phospholipid metabolism is altered because of an increased loss of AA, EPA, and DHA from phospholipids, and that this increased loss leads to changes in the functioning of membrane-associated proteins and of cell-signalling systems. As a matter of fact, a review of 15 published studies reports that reductions, particularly in AA levels but also in LA and DHA levels in red cell and also in thrombocyte and fibroblast membranes, have been replicated in patients with schizophrenia (89). Although this depletion has been found in both drug-free and medicated patients (90, 91), most studies report data of medicated patients and results have not been entirely consistent (92). (In this context, it is noteworthy that primate studies confirm that EFA levels in red blood cell membranes reflect levels in nerve cell membranes (93).) A recent study in patients suffering from schizophrenia-spectrum disorders and being treated with antipsychotic agents indicates a bimodal distribution of PUFAs: one group presenting with high PUFAs similar to healthy subjects, and another group showing low PUFAs, especially low DHA levels (94). Nevertheless, the interpretation of findings in antipsychotic-free patients also remains somewhat inconclusive, some reporting depletions of AA, DHA, and DPA as part of the pathogenetical cascade leading to schizophrenia (95), and others postulating that differences in fatty acid levels and composition might be an epiphenomenon reflecting interactions between abnormalities of fatty acid and phospholipid metabolism and environmental factors such as diet and medication (96). Interestingly, recent data in patients with established schizophrenia suggest that the composition of PUFAs seems to be associated with psychotic state and symptoms; also, PUFA dysregulations returned to levels comparable to healthy individuals after antipsychotic treatment (90, 97). Studies in drug-naïve patients at the onset of psychosis provide further evidence for lowered AA and DHA levels (98, 99) and reflect the idea that fatty acid metabolism might indeed be an underlying factor at the onset of psychosis that is affected by duration of untreated psychosis, antipsychotic treatment, diet, or other exogenous factors. Lower membrane AA and DHA levels possibly predate the onset of a psychotic disorder, and treatment with antipsychotic agents seems to increase these reduced PUFA levels, pointing towards the possibility of an essential contribution of PUFA metabolism to the manifestation of illness (90). As a matter of fact, alterations in phospholipid metabolism have been measured in vivo by means of phosphorus magnetic resonance spectroscopy (31P-MRS). The most recurrent findings are a decrease in the level of phosphomonoesters (PME) and an increase in the level of phosphodiesters (PDE) and nucleotide triphosphate in the frontal lobe of medicated and unmedicated patients suffering of schizophrenia (100–102), but also in adolescent offspring at risk for psychosis (103). These findings indicate an increased breakdown of neuronal membrane phospholipids and further support the MPH.

Postmortem Findings of Altered PUFA Metabolism

To our knowledge, five studies of postmortem investigations of fatty acid composition of brain tissue from patients with schizophrenia are available (104–108). Horrobin et al. found lower concentrations of DHA and AA in the frontal cortex of patients with schizophrenia (107). Yao and colleagues’ study replicated the lower levels of AA but not of DHA (106), whereas Landen et al. could not replicate any of the mentioned alterations in individual fatty acid concentrations in postmortem cingulate cortex of patients (108). The most recent study investigated total fatty acid composition of postmortem orbitofrontal cortex from age-matched drug-free and antipsychotic-treated patients and controls (105). Significantly lower DHA concentrations and elevated AA:DHA ratios were found in patients compared to controls. However, these differences were found in male patients only, were most pronounced in drug-free patients, and partially normalized in patients treated with antipsychotic agents. Similar effects of antipsychotic treatment had been published earlier (98, 99, 109). A possible overlap with results of postmortem analyses of fatty acid composition in patients with other psychiatric disorders can be observed: deficits in DHA and AA levels have been found in patients with bipolar disorder (110), with numerically greater deficits in DHA in those with a history of psychosis, whereas patients with major depressive disorder exhibited deficits in DHA only (111). While supporting the idea of a general disturbance of fatty acid metabolism in psychiatric syndromes, the findings also reflect possible differences in terms of the affected fatty acid family and the affected brain region in specific disorders. So far it remains unclear whether the deficits found in schizophrenia are region-specific. A recent postmortem study did not find significant differences in n-3 PUFA composition in the hippocampus of patients with schizophrenia or bipolar disorder and healthy controls. However, there was a selective decrease in DPA without affecting DHA in both patients groups, suggesting altered lipid metabolism, particularly in terms of n-6 PUFA (104).

Phospholipases in Psychotic Disorders

Of the five different classes of phospholipases—the enzymes that hydrolyze phospholipids into fatty acids and other lipophilic derivates—phospholipases A2 (PLA2) and phospholipases C (PLC) are the most important ones in the brain. The superfamily of PLA2 isoenzymes constitutes three major enzyme groups: (1) secretory (extracellular) Ca2+-dependent PLA2 (sPLA2), (2) cytosolic Ca2+-dependent PLA2 (cPLA2), and (3) Ca2+-independent PLA2 (iPLA2) (112–114). The latter two subtypes constitute the ­intracellular phospholipases A2 (inPLA2), corresponding to the genetically determined subgroups IIa, Iva, and VIa, all important for the regulation of PUFA (AA) release (112). In fact, PLA2GIVa has a 50-fold preference for phospholipids containing AA over any other PUFA, and initiates AA release and eicosanoid production, whereas PLA2GVIa and PLA2GIIa show no fatty acid preference, and the latter enhances AA production. PLA2GVIa cleavage of phospholipids primarily serves the purpose of cellular membrane remodelling, by altering phospholipids–fatty acid ratios and modifying membrane fluidity. Loss of PLA2GVIa function leads to significant reductions in the amount of AA incorporated into the cell membrane (29). Thus, the inPLA2 group of enzymes controls the breakdown of membrane lipids in the context of damage (e.g., increased lipid peroxidation and release of highly reactive molecules) by producing lysophospholipids and releasing free PUFAs (mainly AA). Hence, inPLA2 enzymes provide the first step to deactivate cytotoxic intermediate metabolites (“housekeeping role” of inPLA2) (115) while delivering precursors to eicosanoid pathways. Increased activity of inPLA2 activity is one of the most replicated findings in the field of biochemical schizophrenia research. InPLA2 activity was found to be increased in blood plasma, blood serum, and platelet membranes, and in postmortem brain tissue of patients ­compared to healthy controls (116–123). To our knowledge there was only one published negative study, probably due to methodological problems (124). Accordingly, an increase in PLA2GIVa enzyme protein concentration was also found in schizophrenia patients (125). Furthermore, red blood cell membrane lipid precursors were found to be decreased (126), whereas membrane degradation products were increased (127). An initial report of increased peripheral calcium-dependent PLA2 activity in schizophrenia (128) has not been replicated (118, 120, 129). An analysis of postmortem brain tissue found calcium-dependent PLA2 activity to be decreased, rather than increased (121).

Interestingly, patients with a first psychotic episode rather than those with chronic schizophrenia show inPLA2 hyperactivity (122). Higher inPLA2 activity has been correlated to greater severity of psychopathology in some (120), but not all, studies (117, 123).

In a very recent study in a group of early onset first-episode schizophrenia patients, serum inPLA2 activity correlated positively with negative symptom scores and inversely with levels of social functioning. The higher the inPLA2 activity before treatment, the better patients responded to 12 weeks of treatment with modern antipsychotic agents (130). The increased inPLA2 activity in schizophrenia was shown to be even greater in a subgroup of patients insensitive to niacin skin stimulation (a vitamin B3-based, prostaglandin D2-mediated skin flush test), which relates to the biochemical link between inPLA2 activity, AA release, and prostaglandin availability. Eight weeks of antipsychotic treatment reduced abnormally high inPLA2 activity. Furthermore, those patients who converted to being niacin-sensitive showed further normalization of inPLA2 activity (123). Another methodically combined study disclosed associations between inPLA2 activity and dynamic brain structural changes in the left prefrontal cortex (PFC), bilateral anterior cingulated cortex (ACC), thalamus, caudate nucleus, left hippocampus, and the anterior cerebellum in first-episode schizophrenia patients (131). These findings do not only support the involvement of inPLA2 in brain lipid metabolism; they also emphasize the close connection between lipid biochemistry and focal brain structural abnormalities in first-onset schizophrenia.

There are different ways of explaining the link between inPLA2 alterations and psychosis. Using different animal models, Gattaz et al. were able to show inhibition of dopaminergic activity in vivo by intracerebral PLA2 application. Integrating results of disturbed phospholipid metabolism (from 31P-MRS) and of hypodopaminergic activity (hypofrontality hypothesis) in the PFC of schizophrenia patients, they speculated that increased PLA2 activity in the PFC of schizophrenic patients might contribute to both abnormalities (132). The idea of a genetical link between inPLA2 activity and monoaminergic neurotransmission was substantiated by the co-localization of monoaminergic receptors with inPLA2 (PLA2GIVa, PLA2GIIa) and eicosanoid-synthesizing enzymes (29). Therefore, the initial membrane lipid hypothesis (87), assuming PLA2 effects to be causative for transmitter deregulation and structural abnormalities, fueled association studies between PLA2 genes and schizophrenia (carefully summarized and discussed by (29)). The following studies on gene variations at gene loci of PLA2GVIa, PLA2GIVa, and PLA2GIIa revealed conflicting results due to methodical problems, and did not substantiate a causative role of genetic aberrations at PLA2 gene loci within the cascade leading to schizophrenia. More recently, the findings regarding dynamic changes of inPLA2 activity during the course of illness with most pronounced increase in acute stages (122) and the stage-dependent association with brain structural changes (131) have initiated the notion of inPLA2 alterations as a general reflection (and result) of pathology of membrane plasticity (e.g., changes due to oxidative stress, or excitotoxicity). The current assumption is that of a primary molecular pathology leading to increased membrane turnover/remodelling and to increased inPLA2 activity. The latter, and the potential of inPLA2 activity to serve as a possible dynamic marker of this pathology, are subjects of current research in this field.

Electroencephalography

Clinical trials conducted in psychiatry most often rely on subjective measurements when assessing the effects of a treatment. Electroencephalography (EEG) is a powerful technique that can be used to objectively measure changes in the activity of the brain. The speed of the recorded EEG waveforms can be measured following spectral analyses and classified into well-established frequency bands (133), allowing researchers to quantify brain activity. Two recent meta-analyses have shown that no single pattern of EEG abnormality has been identified in chronic schizophrenia or FEP patients (134, 135). However, both reviews claim that the most consistent observation obtained from spectral analysis in medicated as well as in unmedicated patients with schizophrenia is the increase in slow wave activity (mainly delta), particularly in the frontal region. Increased beta activity and decreased alpha power have also been observed, but with little consistency. Recently, a first study was published on quantitative resting EEG in individuals at risk for psychosis (136). This research showed that EEG spectral power itself cannot predict transition to psychosis, while negative symptoms scores in combination with power in the delta, theta, beta-1, and beta-2 bands predicted transition.

The association between resting-state EEG activity and fatty acid levels in schizophrenia or other psychiatric disorders is largely unknown. One study measured resting-state EEG following omega-3 PUFA treatment in adolescent boys with attention deficit hyperactivity disorder (ADHD) (137). They observed a positive correlation between the alpha frequency band and the plasma levels of DHA. This study also found that activity in the theta band was positively associated with EPA plasma levels. Based on these findings, the authors suggested EEG activity as an objective biomarker of neural function associated with long-chain omega-3 fatty acids in ADHD. Another study used EEG as an objective measurement of the effect of fish oil supplementation in healthy participants (138). Once again, an increase in both theta and alpha bands was recorded following omega-3 intervention, concomitant with a reduction in the fast frequency band beta 2. Considering the EEG findings in people with schizophrenia and the clinical effectiveness of omega-3 supplementation in individuals at-risk of psychosis (6), future research should determine the associations between EPA and DHA plasma levels and brain activity in different clinical stages of psychotic disorder.

Treatment of Psychotic Disorders with PUFAs: Myth or Fact?

PUFAs as Add-On Therapy

With respect to the efficacy of add-on treatments of schizophrenia or related disorders with n-3 PUFAs, the currently available data seem inconclusive. Eight heterogeneous trials are available (7, 8, 70, 139–143), four of which are double-blind and placebo-controlled (7, 8, 140, 141) and only one was performed solely in patients with a first psychotic episode (7). Only one study compared EPA, DHA and placebo (8), and one study investigated different doses of EPA compared to placebo (Table 9.1) (70).

Table 9.1 Indicated prevention, monotherapy, and add-on treatment studies with n-3 long-chain fatty acids in patients with psychosis and/or schizophrenia-spectrum disorders
Full size table

In order to better understand the available data it might be useful to break down the results from different perspectives.

Five studies showed an improvement of psychotic symptoms with add-on treatment with EPA (8, 139, 140, 142, 143). However, three of the positive studies were not placebo-controlled: one added EPA/DHS to the diet of patients with schizophrenia (142). The other assigned patients to a supplementation with a EPA/DHA combination and had a healthy control comparison group (139). The third added EPA/DHA and antioxidant vitamins in an open-label manner to existing treatment with haloperidol (143). Furthermore, the study by Emsley and colleagues (140) reported that the psychopathological improvement might at least partly be due to an improvement of dyskinesia.

The three studies that showed negative findings with respect to the primary outcome measure “improvement of psychopathology” consist of one trial that included both patients with schizophrenia and schizoaffective disorder (141), one trial in FEP (7), and one dose-finding study comparing 1, 2, and 4 g/day EPA vs. placebo (70).

In a recent meta-analysis, the RCT studies were shown to indicate no overall beneficial effect of EPA augmentation of antipsychotic medication on positive psychotic symptoms in established schizophrenia (144).

PUFAs in Monotherapy

Currently, only one study has investigated the effect of n-3 PUFA monotherapy as primary treatment in patients with an established schizophrenic disorder (8). Thirty drug-free patients were randomized to either 3 months of 2 g/day EPA or corn oil placebo. Nine patients were drug-naïve, and the others were off medication for at least 2 weeks; however, medication was permitted on the basis of clinical judgement. At the end of the 3-month study period all patients (n  =  12) in the placebo group required antipsychotic medication, but only 6 of the 14 in the treatment group required antipsychotic medication. Furthermore, patients in the treatment group spent significantly fewer days on medication and had significantly lower Positive and Negative Syndrome Scale (PANSS) scores, especially on the positive symptoms subscale, than those who were on placebo (145).

The singularity of the study and its small sample size prevent us from being able to draw any conclusions.

PUFAs in Indicated Prevention

Much work has been done over the last two decades to define a specific and prospective mental state of clearly increased risk of development of psychosis. Paralleling clinicopathological staging in somatic medicine, the clinical staging model for psychosis (146) provides a framework for this concept and defines three risk stages, two of which are symptomatic risk stages—one with nonspecific, another with more specific symptoms. The latter has been developed in Australia as the ultrahigh risk (UHR) concept (147). The UHR concept comprises help-seeking and symptomatic individuals, differentiating between a group with attenuated subthreshold psychotic symptoms, a group with brief limited intermittent psychotic symptoms (BLIPS) subsiding within 1 week, and a group with a genetic risk and a clear functional decline. The idea behind the UHR concept is to go beyond simple genetic high risk due to family history of psychosis and define more specific symptomatic stages of UHR. About 10–40% of the individuals defined as UHR will transition to a psychotic disorder within 12 months; an incidence rate which is much higher than the one of schizophrenia in the general population (147–151). Evolving clinical and neurobiological knowledge on high-risk stages seems to support the clinical staging model. PUFAs and their effects on, amongst others, brain maturational processes, regulation of synaptic plasticity and myelination as well as their anti-inflammatory and metabolic effects might play different roles within distinct stages of the illness.

The concept of indicated prevention in psychosis comprises strategies for individuals at a clearly increased risk of developing a disorder with specific but subthreshold symptoms as potential precursors of the illness. This concept allows focusing on preventive intervention strategies in individuals with UHR that do not fulfil criteria for diagnosis of a psychotic disorder within the current classification systems. The aim of indicated prevention strategies is to delay or prevent the onset of psychosis. Most initial early intervention strategies in UHR individuals adapted existing treatment concepts for manifest psychosis (152). Data on five randomized controlled trials (RCT) with transition rates as primary outcome are currently available (6, 153–157).

A recent meta-analysis concluded that non-pharmacological focused treatments using cognitive behavioral therapy (CBT) or intensive community care were effective in reducing the risk of transition to psychosis over a time period of 12 months. Interventions that were provided over a limited time were less effective over a longer observation period, although they still managed to delay onset of psychosis (152). Interestingly, two of the RCT-interventions based on antipsychotic medication had a higher proportion of individuals developing psychosis when discontinuing treatment compared to the non-pharmacological interventions: 16.1% for olanzapine and 19.3% for risperidone vs. 5.7% for cognitive therapy. Also, trials using antipsychotic medication have not yielded significantly better results than non-pharmacological treatments and bear the risk of short- and long-term side effects. Thus, in the UHR population more benign treatments might be more appropriate with respect to efficacy and side effects (158).

The most recent trial of indicated prevention in UHR groups was based on the concept of potential neuroprotective properties of n-3 PUFAs (6, 159). Eighty-one individuals with UHR for psychosis were randomized to 12 weeks of either 1.2 g/day of n-3 PUFAs or placebo with a follow-up period until 12 months. The study was conducted in a double-blind manner. At 12-months follow-up 4.9% (n  =  2/41) of patients in the n-3 PUFA-group and 27.5% (n  =  11/40) in the placebo group had transitioned to a psychotic disorder; this difference was significant. In the n-3 PUFA group there was further a significant reduction of positive, negative, and general symptoms as well as improved functioning compared to the placebo group. There was no significant difference in the incidence of adverse effects between the two groups. Regarding fatty acid composition there was not only an expected significant increase of n-3 fatty acids relative to n-6 fatty acids in the treatment group after the intervention, but pre- to posttreatment change in the n-6 to n-3 ratio was significantly related to functional improvement. This trial yielded a number needed to treat (NNT) for n-3 PUFAs of 4, which corresponds to the NNT of the two antipsychotic trials in UHR groups (153, 154). The most interesting finding of this study is the fact that treatment effects were sustained even after cessation of the actual 12-weeks intervention. This effect might indeed be due to potential neuroprotective, anti-apoptotic, and antioxidant properties of n-3 PUFAs that might be effective in very early stages of illness development. However, substantiation of this interpretation and replications of these recent clinical results are still outstanding (180).

PUFAs and Tolerability

Treatment with antipsychotic drugs but also schizophrenia-spectrum illnesses themselves are associated with a range of side effects and other, mainly metabolic disturbances, which often impede treatment adherence and increase illness-related mortality. Addressing these side effects and somatic problems is an important challenge in the treatment of patients with psychosis and should be done proactively (Table 9.2).

Table 9.2 Summary of findings on polyunsaturated fatty acids and tolerability in patients with schizophrenia-spectrum disorders
Full size table

PUFAs and Extrapyramidal Side Effects and Tardive Dyskinesia

With the emergence of atypical antipsychotic agents, extrapyramidal side effects (EPS) and tardive dyskinesia (TD) have become more seldom. Nevertheless, mostly in a dose-related way, many patients do exhibit EPS even when on atypical antipsychotics. Eight heterogeneous studies have looked at the effects of n-3 supplementation on EPS either as primary outcome or while assessing side effects (7, 70, 140–143, 160, 161). Their results seem inconclusive. In FEP patients, EPA augmentation of antipsychotics led to a decreased incidence of EPS; furthermore, those patients with rigidity at baseline improved with the addition of EPA (7). Three other studies report significant decreases in motor side effects, mainly dyskinesia scores, with n-3 PUFA supplementation, one being placebo-controlled with 3 g/day EPA (140), the other a nutritional supplementation of 10 g EPA/DHA (142), and the third an open label study with a combination of EPA/DHA and antioxidant vitamins in haloperidol-treated patients (143). The former also report that decreases in dyskinesia were associated with an increase in ALA (alpha-linolenic acid) levels (162). In another placebo-controlled trial of 2 g/day EPA, albeit no group differences, there was a trend that those patients with more recent development of TD had greater improvements of these symptoms with EPA supplementation (160). Three groups (70, 141, 161) report no changes. None of the studies investigated long-term effects of n-3 PUFA augmentation with respect to TD.

Finally, one study aimed to assess the relationship between TD and levels of n-3 and n-6 fatty acids in RBC membranes and plasma in patients with schizophrenia or schizoaffective disorders over a follow-up period of 4.5 years (163). The only consistent findings within a great EFA variability were lower levels of LA and higher levels of dihomogamma-linoleic acid (DGLA, 20:3 n-6) in the patient population.

Summarizing, there seems to be a tendency for n-3 supplementation to improve tolerability of antipsychotic medication with respect to EPS.

PUFAs and Sexual Dysfunctions

Sexual dysfunction is a common side effect of psychopharmacological medication and can be strongly disturbing even more so in adolescents or young adults before or at the onset of psychosis. One study in FEP patients reported a significant positive effect of EPA augmentation on orgastic dysfunction and a trend for ejaculatory dysfunction (7).

PUFAs and Metabolic and Cardiovascular Risks Factors

The “endocrine psychosyndrome” (i.e., the association of metabolic dysfunction and psychiatric disorders) has been described by Eugen Bleuler and his son around 60 years ago (164). Alongside a possible effect of the psychiatric illness per se, treatment with many antipsychotic agents may increase the risk for metabolic abnormalities (165).

There is strong evidence that n-3 LC-PUFAs may positively affect lipid metabolism, such as reduction of elevated triglyceride (TG) levels (166–168). However, studies that examined these effects in patients with schizophrenia are sparse.

One trial in patients with chronic schizophrenia and concomitant TD receiving 12 weeks of EPA vs. placebo with a 40-week open extension period specifically investigated metabolic effects of 2 g/day EPA in this population (169). They found a significant, but clinically barely relevant, increase in Body Mass Index (BMI) in the EPA group. Furthermore, cholesterol levels were significantly decreased, but so were HDL (high-density lipoprotein) levels. There were no changes in LDL (low-density lipoprotein) or in TG levels. However, the authors report a trend level increase in fasting blood glucose. Although not substantial, the increases in fasting blood glucose, BMI, the decrease in HDL, and the lack of TG reduction are noteworthy in a population that is at increased risk for metabolic syndrome and should be monitored carefully. Furthermore, the results regarding TG are rather surprising; it has been suggested that the hypotriglyceridemic potential of EPA might only be effective in hypertriglyceridaemic states (170).

Two studies provide specific results in clozapine-treated patients, who are at even higher risk of metabolic dysfunction. In the dose ranging study of EPA augmentation by Peet et al., elevated TG levels were significantly reduced with doses of 2 and 4 g/day EPA in the clozapine-treated patient group (70). In an open-label study using 3 g/day EPA/DHA over 4 weeks in patients with schizophrenia and schizoaffective disorder, Caniato and colleagues (171) demonstrated significant improvement of lipid abnormalities with a reduction of elevated serum TG level, increase in total cholesterol levels, in LDL but also in HDL levels, as well as a decrease in VLDL (very low density lipoprotein).

The decrease in HDL levels in Emsley’s study and the increase in LDL levels in Caniato’s study are somewhat worrisome since the relationship between increased LDL and decreased HDL levels and cardiovascular disease is long established. These findings, albeit originating from only two trials, should reinforce the necessity of metabolic monitoring in patients with psychotic illnesses. n-3 LC-PUFA supplementation has been proposed to have a beneficial effect against atherosclerosis and even more so against fatal arrhythmias, wherefore 1 g/day of EPA/DHA is recommended by the American Heart Association for secondary cardiovascular disease prevention (172, 173). Omega-3 fatty acids may exhibit their cardioprotective effects by affecting gene expression of factors involved in atherogenesis, inflammation, and angiogenesis (174).

With a high prevalence of metabolic syndrome, due to possible disease-inherent factors, medication (metabolic dysbalance and QT prolongation), and an unhealthy lifestyle, patients with schizophrenia are at increased risk for cardiovascular disease. Thus, supplementation with n-3 LC-PUFAs might not only be a useful strategy in emerging and incident psychosis with respect to their potential in indicated prevention and antipsychotic-sparring effects but also possibly to prevent metabolic and cardiovascular detriments, although further evaluation of these effects is needed (175).

Do PUFAs Have Adverse Events?

As the interest in n-3 LC-PUFAs in the field of psychiatric disorders, and specifically in emerging psychosis, increases, the question of potential adverse events must be addressed.

n-3 LC-PUFAs have been reported as being associated with prolonged bleeding in anecdotal reports. Although a statement by the FDA (176) reported that fish oils can significantly prolong bleeding time and a dose of up to 3 g/day EPA/DHA was considered as safe, clinical trial evidence does not support relevant increases in bleeding time, not even in patients treated with antiplatelet or antithrombotic agents (177). In the study in chronic schizophrenic patients performed by Emsley et al. (169) the increase in bleeding time approached significance, but was not associated with clinically relevant bleeding events. As some psychopharmacological drugs have also been anecdotally associated with epistaxis, mindfulness with respect to potential interactions seems useful (178, 179). It is also noteworthy that in a population with increased risk of metabolic and cardiovascular disease a clinically not significant increased bleeding time might even be discussed as being beneficial against cardiovascular mortality.

Other common but clinically rarely relevant side effects of n-3 fatty acid supplementation include mild nausea, mild diarrhea, fishy eructation, and exacerbation of asthma in aspirin-sensitive patients (89). Only two available treatment trials in patients at UHR, with a FEP or chronic schizophrenia, report differences in adverse effects with n-3 LC-PUFA supplementation (70, 141): they report gastrointestinal symptoms and upper respiratory infections.

Generally, tolerability of n-3 LC-PUFAs can currently be assumed to be very good.

Conclusion

The current stage of knowledge does not support the notion that n-3 PUFAs, either purified EPA or combinations of EPA and DHA, are more efficacious than antipsychotic medication alone in reducing positive psychotic symptoms. There is some evidence that n-3 PUFAs may be beneficial in early stages of psychosis, supporting a faster response to lower doses of antipsychotic treatment. In chronic schizophrenia some patients (e.g., particularly those treated with clozapine) may benefit from n-3 PUFAs augmentation. The most clear-cut positive finding was found in indicated prevention of psychosis in UHR adolescents. However promising, these data still need replication findings and more precise neurobiochemical interpretation. Future studies should also investigate whether PUFAs may increase tolerability of antipsychotic agents in several aspects, particularly their long-term metabolic and cardiac side effects, as well as their long-term effects in secondary prevention after a first psychotic episode.