The association of plasma inflammatory markers with omega-3 fatty acids and their mediating role in psychotic symptoms and functioning: An analysis of the NEURAPRO clinical trial

Background: There is increasing evidence that dysregulation of polyunsaturated fatty acids (FAs) mediated membrane function plays a role in the pathophysiology of schizophrenia. Even though preclinical findings have supported the anti-inflammatory properties of omega-3 FAs on brain health, their biological roles as antiinflammatory agents and their therapeutic role on clinical symptoms of psychosis risk are not well understood. In the current study, we investigated the relationship of erythrocyte omega-3 FAs with plasma immune markers in a clinical high risk for psychosis (CHR) sample. In addition, a mediation analysis was performed to examine whether previously reported associations between omega-3 FAs and clinical outcomes were mediated via plasma immune markers. Clinical outcomes for CHR participants in the NEURAPRO clinical trial were measured using the Brief Psychiatric Rating Scale (BPRS), Schedule for the Scale of Assessment of Negative Symptoms (SANS) and Social and Occupational Functioning Assessment Scale (SOFAS) scales. The erythrocyte omega-3 index [eicosapentaenoic acid (EPA) + docosahexaenoic acid (DHA)] and plasma concentrations of inflammatory markers were quantified at baseline (n = 268) and 6 month follow-up (n = 146) by gas chromatography and multiplex immunoassay, respectively. In linear regression models, the baseline plasma concentrations of Interleukin (IL)-15, Intercellular adhesion molecule (ICAM)-1 and Vascular cell adhesion molecule (VCAM)-1 were negatively associated with baseline omega-3 index. In addition, 6-month change in IL-12p40 and * Corresponding authors. E-mail addresses: subashrajsusai@rcsi.ie (S.R. Susai), drcotter@rcsi.ie (D. Cotter). 1 Equal contributors.

TNF-α showed a negative association with change in omega-3 index. In longitudinal analyses, the baseline and 6 month change in omega-3 index was negatively associated with VCAM-1 and TNF-α respectively at follow-up. Mediation analyses provided little evidence for mediating effects of plasma immune markers on the relationship between omega-3 FAs and clinical outcomes (psychotic symptoms and functioning) in CHR participants. Our results indicate a predominantly anti-inflammatory relationship of omega-3 FAs on plasma inflammatory status in CHR individuals, but this did not appear to convey clinical benefits at 6 month and 12 month follow-up. Both immune and non-immune biological effects of omega-3 FAs would be resourceful in understanding the clinical benefits of omega-3 FAs in CHR papulation.

Introduction
Schizophrenia is a mental disorder with a multifactorial etiology. Most patients with schizophrenia experience a prodromal phase of nonspecific psychiatric and subthreshold psychotic symptoms (Yung and McGorry, 1996). The clinical high-risk (CHR) paradigm provides operational criteria to define a subpopulation at high risk of psychosis and offers the opportunity to intervene early to improve prognosis or even prevent transition to a psychotic disorder. CHR individuals have a 22% risk of developing psychosis in the first year after ascertainment and increases at least 38% after 16 years (Yung et al., 2007;Fusar-Poli et al., 2012;Mei et al., 2021;Beck et al., 2019;Salazar de Pablo et al., 2021;Yung et al., 2005). A growing body of evidence has indicated the involvement of inflammation and oxidative stress from an early stage in psychotic disorders such as schizophrenia (Muller et al., 2015;Müller, 2018;English et al., 2018;Föcking et al., 2016;Föcking et al., 2021;Madireddy, 2020;Perkins et al., 2020;Hagen et al., 2020;Wezenberg et al., 2020). Several inflammatory mediators that participate in neuronal development and synaptic pruning are consistently found to be involved in the early stages of psychosis, including Interleukin (IL)-1β, IL-6, IL-8, IL-12p40 and Tumor Necrosis Factor-α (TNF-α) (Muller et al., 2015;Müller, 2018;Föcking et al., 2016;Delpech et al., 2015;Yirmiya and Goshen, 2011;Perkins et al., 2015;Kelsven et al., 2020). Currently available antipsychotic medications still broadly target dopaminergic neurotransmission and are inadequate in treating negative and cognitive symptoms and can cause serious adverse events (Moncrieff et al., 2020). Hence, there is a need for investigation of further interventions targeting alternative mechanistic pathways to delay or prevent the development of psychosis. In fact, neuroinflammation and antioxidative defense seem to represent potential candidates for those pathways.
In humans, the anti-inflammatory effects of omega-3 FAs have been observed in neurodegenerative disorders, cerebrovascular ischemic disorders, and psychiatric disorders such as depression and bipolar disorder (Marginȃ et al., 2020;Román et al., 2019;Abbas et al., 2021;. However, in patients with established schizophrenia, clinical trials of omega-3 FA supplementation have not produced consistent results. For instance, five studies showed improved prognoses with greater efficacy in omega-3 group than the placebo whereas two studies found no difference and one study reported poor prognosis following omega-3 supplementation (Hsu et al., 2020;Sommer et al., 2016). This may be due to factors such as late onset of omega-3 FA supplementation, confounding effects of age-related metabolic changes and varying regimens of anti-psychotics (Hsu et al., 2020;Sommer et al., 2016). Since inflammation and disruption of membrane architecture (especially in neurons) have been noticed from an early stage of psychosis, intervening at an earlier phase with omega-3 FA has been suggested as an option for preventing or delaying the development of psychosis (Muller et al., 2015;Müller, 2018;Hsu et al., 2020;Mongan et al., 2020;. A previous clinical trial, the Vienna High Risk (VHR) study, found that omega-3 FA supplementation reduced the risk of transition to psychosis in CHR participants (Amminger et al., 2015). The potential role of omega-3 FAs in modulating inflammation was investigated in a subsequent analysis . The authors analyzed the relationship of baseline erythrocyte omega-3 FA levels with three immune mediators including interleukin 6 (IL-6), the soluble alpha (Tac) subunit of the interleukin 2 receptor (sIL-2r), and the circulating soluble form of the intercellular adhesion molecule one (sICAM-1) and found no significant association with membrane omega-3 FAs at baseline . While previous animal and human studies have identified various molecular pathways of inflammatory reactions through which omega-3 FAs influence inflammatory cytokines, the clinical benefit in early psychosis is not yet clearly established (Calder, 2015;Calder, 2017;Susai et al).
The aims of the current study were to investigate the relationship of omega-3 FAs on plasma immune markers in CHR participants and to evaluate whether inflammatory cytokines mediated previously described relationships between omega-3 FAs and clinical outcomes in CHR patients. We used plasma and erythrocyte samples from the NEU-RAPRO trial which is an omega-3 FA based randomized controlled trial aimed to prevent the onset of psychosis in CHR participants. The participants had little or no exposure to antipsychotic medication that can potentially influence inflammatory cytokines McGorry et al., 2017). In this study we intend to answer the two questions: i) whether Omga-3 FAs exert any relationship with plasma inflammatory markers in CHR individuals measured at baseline and 6 month follow-up, and ii) whether plasma immune markers mediate the relationship between omega-3 FAs and clinical outcomes in particular psychotic symptoms and social functioning. Considering that in the NEURAPRO clinical trial no direct clinical association was observed with omega-3 FA supplementation (Berger et al., 2019), we mainly focus on the indirect effect through omega-3 FAs and the overall effect ( Fig. 1) of erythrocyte membrane omega-3 FAs level on psychopathology in CHR participants. Based on evidence from preclinical studies, we hypothesized that omega-3 FAs would be negatively associated with plasma inflammatory biomarkers in CHR participants and would mediate associations between omega-3 FAs and clinical outcomes.

The NEURAPRO clinical trial
The primary outcome of the NEURAPRO clinical trial was to assess the efficacy of omega-3 FA supplementation preventing transition to psychosis in CHR participants. The NEURAPRO clinical trial was registered with the Australian New Zealand Clinical Trial Registry as ACTRN 12608000475347. The study was conducted between March 2010 and the end of September 2014, in accordance with the Declaration of Helsinki and consistent with the International Council for Harmonization of Good Clinical Practice with appropriate ethical approval obtained from each site before the trial commenced. Ethical approval for the biomarker analysis presented in this study was obtained from the research ethics committee of the Royal College of Surgeons in Ireland [REC-No. 1699].

Study participants
A total of 285 out of 304 participants aged 18.97 ± 4.49 years (mean ± SD) who met CHR criteria and who had valid baseline and follow-up clinical data were considered for the plasma biomarkers analysis study. The exclusion criteria were: history of psychotic episodes of seven days or longer; any current symptoms of intoxication, organic brain disease or developmental disorder; abnormal coagulation profile; thyroid abnormalities; physical illness with psychotropic effect, if not stabilized; current treatment with any mood stabilizers or recreational use of ketamine; past antipsychotic exposure equivalent to a total lifetime haloperidol dose of greater than 50 mg; a diagnosis of a serious developmental disorder; premorbid IQ less than 70; current acute suicidality/self-harm or aggression/dangerous behavior; pregnancy; or intake of more than 4 weeks of supplementation with omega-3 FAs .

Exposure: Erythrocyte omega-3 FAs measures
The study participants in the clinical trial were randomized to receive either omega-3 FA supplementation or placebo. The study medication contains a dose of four gelatin capsules taken daily throughout the 6-month treatment period. Participants were provided with dispensed bottles of capsules containing either omega-3 FAs or placebo and participants were instructed to take two capsules in the morning and two at night. The omega-3 group received ~2.8 g of marine fish oil containing approximately 840 mg of EPA and 560 mg of DHA per day, whereas the placebo group received an equivalent dose of paraffin oil per day for 6 months. In addition, all participants could receive up to 20 sessions of cognitive behavioral case management (CBCM) as a cointervention. The full description of treatments participants received during the trail is provided by McGorry et al . Assessing adherence, McGorry et al., reported that nearly 58% of the participants were non-adherent to the study intervention in the NEU-RAPRO trial (Schlogelhofer et al., 2018). To address this, we considered the erythrocyte omega-3 FA levels as a measure of exposure irrespective of the study arms (omega-3/placebo) as this is accepted to be an accurate biological marker for dietary intake of omega-3 FAs and closely reflects the omega-3 FA content of neuronal membranes (HARRIS, 2007;Makhoul et al., 2011;Pierigè et al., 2008;Owen et al., 2004). Fasting plasma samples were collected at baseline and 6 months following the intervention. The molecular percentage of total fatty acid levels of EPA, DHA and n3-index (EPA + DHA) were measured using gas chromatography (McLaverty et al., 2021). The Phosphatidyl-ethanolamine (PE) fraction was used to determine the omega-3 FA content, because of their high abundancy in the lipid raft (Kim et al., 2016;Smesny et al., 2014;Alqarni et al., 2018).

Clinical outcome: Psychotic symptoms and functioning
Positive psychotic symptoms of participants were measured the Brief Psychiatric Rating scale-psychotic score (BPRS-Psychotic) which comprises combined scores of suspiciousness, hallucination, bizarre thoughts & unusual thoughts. Negative symptoms were assessed using the Scale for the Assessment of Negative Symptoms (SANS). Functional status was measured using the Social and Occupational Functional Assessment Score (SOFAS) recorded at baseline, 6 month and 12 month follow-up.

Statistical analysis
The erythrocyte membrane levels of EPA, DHA and the omega-3 index were calculated as percentage of total erythrocyte membrane fatty acid content. The plasma concentrations of inflammatory markers that were within the detection limits of the assay with a coefficient of variance (CV) of at most 20% were taken forward for analysis. Missing concentration of inflammatory markers were imputed with average values and converted to Z scores for the analysis.
Statistical analysis was performed using IBM® SPSS® statistics version 26 and STATA IC ® version 16. The association between erythrocyte omega-3 levels and plasma biomarkers were evaluated using linear regression models. The cross-sectional and longitudinal analyses were conducted for baseline, follow-up and for change of omega-3 values with corresponding inflammatory marker concentrations. For cross-sectional models, the results were adjusted for age, sex, and body mass index (BMI) and for the longitudinal analysis the results were also adjusted for the baseline omega-3 FA levels. The inflammatory markers that showed significant association with change in omega-3 FA level were considered for the mediation analysis.
The mediation analysis was carried out to evaluate the direct effect of change in omega-3 index (exposure) on clinical outcome and the possible mediating role of plasma immune markers on an association between omega-3 index and clinical outcomes (indirect effect) at 6 and 12-month follow-up (Hayes, 2017). Mediation analysis was performed in IBM® SPSS® using the PROCESS platform. Regression coefficients were constructed using conventional mediation analysis model (model 4) with a bootstrap sample size of 5000 and with 95% confidence interval (Fig. 1). The mediation analysis was adjusted for age, sex, BMI and baseline omega-3 index levels. The results were corrected using a false discovery rate (FDR) of 5%, as described by Benjamini-Hochberg (Benjamini and Hochberg, 1995).

Participant characteristics
Erythrocyte membrane FA and plasma immune marker concentrations were available for 268 participants at baseline and for 146 participants at both time-points (baseline and 6-month follow-up) ( Table 1). Plasma levels of IL-6, IL-8, IL-10, IL-12p40, IL-15, TNF-α, ICAM-1 and VCAM-1 were selected as they were described in the literature to be associated with omega-3 FAs in animal and human studies (Calder, 2015;Calder, 2017). The details of clinical symptoms of psychosis include BPRS-psychotic and SANS score with functional scores at baseline, 6-and 12-month follow-up are presented in Table 2.

Relationship between 6-month change in omega-3 FAs and 6-month change in cytokine levels
In a linear regression model, omega-3 index changed inversely with changes in plasma levels of TNF-α such that those with increasing levels of TNF-α had decreasing levels of Omega-3 with a beta coefficient of 0.06 (FDR-corrected p = 0.032). A similar inverse association was noticed for EPA and DHA levels individually with plasma cytokines. An increase in DHA levels associated with decrease in plasma TNF-α and ICAM-1 (β-coef = − 0.09 & 0.09; FDR-corrected p = 0.02 & 0.02, respectively) at 6 months follow-up. Similarly, change in EPA showed an inverse association with change in plasma levels of IL-12p40, IL-15 and TNF-α (β-coef = − 0.16, − 0.12 & − 0.14; FDR-corrected p = 0.048, 0.048 & 0.048, respectively) such that increasing EPA score was associated with decreasing plasma immune marker levels (Table 5).

Table 4
Cross-sectional relationship between erythrocyte omega-3 FAs and inflammatory cytokines at follow-up.  Results are presented from models adjusting for age, sex, BMI and baseline omega-3 measures. The underlined markers were used for mediation analysis. B-H-Benjamini Hochberg correction (FDR = 0.05%), EPA-eicosapentaenoic acid, DHA-docosahexaenoic acid, IL-Interleukin, TNF-Tumour necrosis factor, ICAMintercellular adhesion molecule, VCAM-Vascular cell adhesion molecule.

Direct and indirect effects of omega-3 FAs on clinical outcome (mediation analysis)
Changes in the plasma inflammatory markers IL-12p40 and TNF-α were considered for the mediation analysis, as they were significantly associated with change in omega-3 index. However, mediation analyses did not provide evidence of significant direct or indirect effects (via plasma immune markers) on psychotic symptoms (BPRS-psychotic and SANS) or functional outcome (SOFAS) at 6-month (Table 8) or at 12month (Table 9) follow-up.

Discussion
We examined the biological relationship of erythrocyte omega-3 FAs with plasma inflammation markers in a group of CHR participants from the NEURAPRO trial. We hypothesized that plasma inflammatory markers would be negatively associated with omega-3 FAs and at least partially mediate the association between omega-3 FAs and clinical outcomes. The results supported our first hypothesis as plasma inflammatory markers associated inversely with omega-3 index, but did not support our second hypothesis as they did not indicate any mediating effect on psychotic symptoms or functioning. The main results are as follows: Firstly, in a cross-sectional analysis, the omega-3 index at baseline was inversely associated with IL-15 and endothelial immune markers ICAM-1 and VCAM-1. Secondly, increase in omega-3 measures was significantly associated with decrease in TNF-α over the 6-month interval. Thirdly, in the longitudinal assessments, higher baseline omega-3 index and DHA predicted lower plasma levels of VCAM-1 at follow-up. Finally, 6-month change in omega-3 index expressed similar inverse association with TNF-α at follow-up. In the mediation analysis, omega-3 FA associated changes in plasma inflammatory markers did not exert any significant mediation role on psychotic or functional outcome of CHR participants.
At baseline, the vascular endothelial markers VCAM-1 displayed a strong negative association with omega-3 index and DHA levels both cross-sectionally and longitudinally. In addition, ICAM-1 showed a negative association with erythrocyte omega-3 index cross-sectionally among baseline samples. The vascular adhesion molecules ICAM-1 and VCAM-1 belong to the immunoglobulin super family that are synthesized chiefly by leukocytes and endothelial cells (Legein et al., 2013). Studies have found varying levels of endothelial immune markers in schizophrenia patients compared to healthy controls (Nguyen et al., 2018;Schwarz et al., 1999;Stefanović et al., 2016;Thomas et al., 2004;Kavzoglu and Hariri, 2013;Cai et al., 2020). To understand the relative contribution of these endothelial markers in schizophrenia patients, Nguyen et al. developed a composite measure called "Vascular endothelial index" (VEI). The VEI was based on the linear combination of endothelial markers that differed most between the groups, and VEI was found to be increased in schizophrenia patients compared to healthy controls (Nguyen et al., 2018;Berger et al., 2016). Our findings of an inverse association between endothelial markers and omega-3 FAs were supported by Baker et al (Baker et al., 2018). In this review, the authors identified cellular mechanisms through which EPA and DHA inhibit the synthesis of endothelial markers at various levels. At the molecular level, omega-3 FAs decrease the expression of messenger RNAs responsible for coding endothelial cytokine synthesis (but increase the level of arachidonic acid as the main precursor of cytokines such as prostaglandins). At the cellular level, omega-3 FAs inhibit the adhesion and migration of leukocytes across the endothelium (Baker et al., 2018). Apart from such immune roles, endothelial markers are also involved in mechanisms such as disruption of the blood brain barrier (BBB), neuronal apoptosis and age-related impairments in neuronal precursor cells (Zhang et al., 2015;Yousef et al., 2019). Considering that these mechanisms have already been related to the pathophysiology of schizophrenia (Ðorđević et al., 2011;Jarskog et al., 2005;Inta et al., 2017), our results of negative associations between omega-3 FAs and endothelial immune markers provide preliminary evidence of an immune modulating effect of omega-3 FAs in early psychosis (CHR).
Our study also identified an inverse relationship between EPA and IL-

Table 7
Longitudinal relationship between 6 month change in erythrocyte omega-3 FAs and inflammatory cytokines at follow-up.  Mediation analysis with regression co-efficient (95% confidence interval). EPAeicosapentaenoic acid, DHA-docosahexaenoic acid, IL-Interleukin, TNF-Tumour necrosis factor, BPRS-Brief Psychiatric Rating Scale, SANS-Scale for the Assessment of Negative Symptoms, SOFAS-Social and Occupational Functioning Assessment Scale.
12p40 which is a common subunit of cytokines IL-12 and IL-23 and exerts a pivotal agonistic role in early inflammatory reaction (GATELY et al., 1996;Gillessen et al., 1995;Cooper and Khader, 2007). In psychosis, a meta-analysis showed elevated levels of the pro-inflammatory cytokine IL-12 in schizophrenia patients compared to healthy controls (Miller et al., 2011). Moreover, our group previously observed increased plasma levels of IL-12p40 distinguishing CHR subjects who transitioned to psychotic disorder from who did not (Föcking et al., 2016). A similar strong negative association was found for omega-3 FAs with proinflammatory cytokines IL-15 and TNF-α. The biological evidence relating the acute phase inflammatory state cytokine TNF-α with omega-3 FAs has been extensively reviewed and altered regulation of TNF-α and IL-15 have been consistently observed in psychosis Joffre et al., 2019;Myint et al., 2009;Mittleman et al., 1997;de Witte et al., 2014). Here, for the first time, we report an association of omega-3 FAs with TNF-α in those potentially in as early stage of a psychotic disorder. In a placebo-controlled randomized trial, a decrease in the omega-3:omega-6 ratio showed a positive association with IL-6 and TNF-α production, suggesting an anti-inflammatory role of omega-3 FAs on peripheral cytokines (Kiecolt-Glaser et al., 2011). The same research group also observed anti-inflammatory properties of omega − 3 FAs rich diet in an adult population (Kiecolt-Glaser et al., 2012). Moreover, an extensive interaction between TNF-α and IL-15 has been observed at the blood brain barrier (BBB) since TNF-α enhances IL-15 synthesis and IL-15 in turn regulates TNF-α signaling at the level of the BBB (Pan et al., 2013;Pan et al., 2009). Although the current study did not consider the interactions of these pro-inflammatory mediators, the findings of significant associations between omega-3 FAs and these cytokines indicate a possible anti-inflammatory role on BBB by omega-3 FAs in CHR participants.
In the NEURAPRO clinical trial, although omega-3 FA supplementation was not found to be effective in the prevention of transition to psychosis, baseline erythrocyte omega-3 FAs levels have been shown to be associated with improvement of clinical symptoms in CHR individuals. At baseline, the n-3 index (EPA + DHA) was negatively correlated with general psychopathology, psychotic, depressive and manic symptoms, while the n-6/3 PUFA ratio was positively correlated with general psychopathology and depressive symptoms (Berger et al., 2019). In addition, 6-month increase in omega-3 FAs levels predicted less severe psychopathology and better functioning at 6-month and 12month follow-up (Amminger et al., 2020). While these results suggested possible therapeutic effects of omega-3 FAs in CHR, the current study, which investigates the mechanistic (rather than the predictive) role of omega-3 FAs, found no clinical effect. We suspect that the absence of an association found between omega-3 FAs and clinical outcome (directly or indirectly through inflammatory mediators) could be due to the following reasons: (i) sampling bias which occurred due to the drop in sample numbers at follow-up as comparison of baseline characteristics of samples with and without follow-up indicated significant difference in biological and clinical parameters (Supplementary Table A), (ii) consideration of a different biomarker for the omega-3 index which was derived from gas chromatography utilizing a specific phospholipid fraction, which has a 3-4 fold higher magnitude compared to the mass spectrometric measure of total membrane fatty acids used in the previous analysis, or (iii) it could be a real effect with no association between omega-3 FAs and clinical outcomes in UHR state as observed previously .
The strengths of our study include well-characterized CHR participants and the availability of baseline and follow-up biological and clinical data. This enabled us to understand the long-term influence of omega-3 FAs on immune status of the CHR participants in a unique manner. The multiplex assay provided the opportunity to analyze the biological effects of omega-3 FAs on a broad array of plasma immune markers in CHR participants.
Our study also has limitations. Firstly, the participants of the clinical trial displayed a low adherence to study intervention, which limited our ability to evaluate group difference between the omega-3 FA and placebo study arms (Schlogelhofer et al., 2018). We overcame this limitation by considering the erythrocyte omega-3 FA levels. Secondly, the number of follow-up samples was only 55% of the baseline sample size. This drop-out may have resulted in some attritional bias as this affected the statistical power. Thirdly, the interaction within different immune mediators was not considered in the mediation analysis. Finally, the multiplex assay performed to study a broad array of immune biomarkers simultaneously, comes with its own limitations of a possible crossreactivity within the assays (abcam, xxxx).
In conclusion, our results showed an inverse relationship between omega-3 FAs and plasma immune markers that are involved in the pathophysiology of schizophrenia in this CHR sample. In the crosssectional analysis, erythrocyte membrane omega-3 FAs were inversely associated with pro-inflammatory cytokines IL-15, IL-12p40, TNF-α, endothelial markers ICAM-1 and VCAM-1 and in the longitudinal analysis a similar negative association was found with TNF-α and VCAM-1. The predominant negative associations observed with several proinflammatory mediators are in keeping with known immune actions of omega-3 FAs and suggest that omega-3 FAs may reduce inflammatory load in CHR individuals. While the existence of an inflammatory subtype of schizophrenia is still under investigation, we speculate that omega-3 FAs could be more clinically beneficial in those who have high inflammatory load at baseline. However, no overall clinical benefits of omega-3 FAs, related to cytokine measures, were observed after 6-month of follow-up in CHR individuals. This raises the possibility that nonimmune function of omega-3 FAs, such as recovering synaptic membrane activity in the brain, modulation of microbiota-gut-brain axis or production neuroprotective metabolites may impact on clinical outcome in early psychosis (Hsu et al., 2020;Eckert et al., 2013) Future randomized control trials with multiple follow-up time points would be beneficial in understanding the possible long-term biological benefits of omega-3 in (those at risk for) psychosis (abcam, 1998).

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence Table 9 Mediating role of change in cytokines (IL-15, IL-12p40 and TNF-α) on the association between change omega-3 index and clinical outcome at 12th month (Covariates: Age, Sex, BMI and Baseline n-3 Index).

Mediators
Outcome ( Mediation analysis with regression co-efficient (95% confidence interval). EPAeicosapentaenoic acid, DHA-docosahexaenoic acid, IL-Interleukin, TNF-Tumour necrosis factor, BPRS-Brief Psychiatric Rating Scale, SANS-Scale for the Assessment of Negative Symptoms, SOFAS-Social and Occupational Functioning Assessment Scale.
the work reported in this paper.