Introduction
Inflammation is an ordered sequelae of events engineered to maintain tissue and organ homeostasis. The timely release of mediators and expression of receptors is essential to complete the programme and restore tissues ad integrum. Inflammation is characterised by redness, swelling, heat, pain and numerous physiological changes, including increased blood flow and permeability across blood capillaries. These changes allow large molecules to leave the bloodstream and cross the endothelial wall, increasing the migration of leukocytes from the bloodstream into the surrounding tissue. Inflammation results from immunological processes in response to injury, invading pathogens, allergens and toxins and leads to repair of the damaged tissue. The first cells detected at inflamed sites are granulocytes, monocytes and macrophages, which have a role in killing pathogens and clearing cellular and tissue debris as well as in tissue repair; lymphocytes appear at a later stage(Reference Romano1–Reference Calder3).
Inflammation can be classified into two types: a) acute inflammation, with a rapid course (from minutes to a few days), in which the most important events are oedema and the migration of leukocytes, mainly granulocytes and monocytes; and b) chronic inflammation, characterised by a long time course, the presence of lymphocytes and macrophages and the proliferation of blood vessels and connective tissue. Hence, an exaggerated inflammatory response can cause local tissue damage and remodelling, which may induce significant and chronic injury. Inflammation underlies and/or accompanies numerous prevalent diseases, including cardiovascular diseases, obesity, diabetes, cancer, inflammatory bowel diseases, neurodegenerative diseases, rheumatoid arthritis and asthma(Reference Seki, Tani and Arita4, Reference Mataix, José and Gil5).
The inflammatory response is regulated by a complex network of mediators. Active mediators modify tissues and organs and specifically adapt their response to each inflammatory inducer signal; these signals increase the production of inflammatory mediators(Reference Medzhitov6).
Chemical mediators of inflammation (CMFs) have been described as biomarkers of inflammation for both healthy individuals and patients with chronic disease(Reference Medzhitov6, Reference Singer, Shapiro and Theilla7). There are eight CMF groups: acute phase protein reactants (C-reactive protein [CRP]), vasoactive amines, vasoactive peptides, fragments of complement components (C3a, C4a; C5a), lipid mediators (eicosanoids, docosanoids and platelet-activating factors), cytokines, chemokines and proteolytic enzymes(Reference Gil, María Aguilera, Gil-Campos and Cañete8). Reactive free radicals, particularly reactive oxygen species (ROS) are major mediators of various acute and chronic inflammatory conditions(Reference Basu9)
Bacterial endotoxins can directly activate monocytes and macrophages, inducing them to form: cytokines, e.g., tumour necrosis factor (TNF-α), interleukins (IL-1, IL-6, IL-8); chemokines, e.g., monocyte chemotactic protein 1 (MCP-1) and macrophage inflammatory protein 1 alpha (MIP-1α); eicosanoids, e.g., PGE2 and LTB4; nitric oxide; matrix metalloproteinases (MMPs) and other mediators, which also regulate body response to infection and injury. In addition to eicosanoids, a number of long-chain polyunsaturated fatty acid (LC-PUFA) derivatives are involved in resolving the inflammatory process, notably lipoxins, resolvins and protectins, recently described as major biomarkers of inflammation(Reference Gil, María Aguilera, Gil-Campos and Cañete8, Reference Serhan, Yacoubian and Yang10, Reference Calder11).
When TNF-α, IL-1 and IL-6 are present in high concentrations, they are especially destructive and implicated in pathological responses in endotoxic shock, in acute respiratory distress syndrome, and in chronic inflammatory diseases, e.g., rheumatoid arthritis and inflammatory bowel disease. Chronic overproduction of TNF-α and IL-1 can damage adipose tissue and muscle and produce loss of bone mass. Anti-inflammatory cytokines (e.g., IL-10) and receptor antagonists (e.g., IL-1R) oppose the cascade of inflammatory mediators initiated by inflammatory cytokines(Reference Calder3–Reference Charo and Ransohoff12).
In addition to CMFs, upregulation of adhesion molecules on the endothelial cell surface allows leukocyte binding and subsequent diapedesis and is a well-known step of inflammation in some chronic diseases, e.g., cardiovascular diseases; these molecules include intercellular adhesion molecule 1 (ICAM-1), vascular cell adhesion molecule 1 (VCAM-1) and E-Selectin(Reference Calder2, Reference Charo and Ransohoff12–Reference Giugliano, Ceriello and Esposito15).
The ability of dietary omega-3 LC-PUFA to limit inflammation has been demonstrated in numerous studies of animals and humans under different conditions and using varied doses. Its intake is associated with reduced concentrations of CRP, proinflammatory eicosanoids, cytokines, chemokines and other inflammation biomarkers(Reference Schwab and Serhan15). However, only some of these studies were properly designed to provide evidence of the influence of dietary omega-3 intake on specific inflammatory biomarkers. The aim of the present review was to systematically review evidence on the effect of dietary omega-3 fatty acid (FA) supplementation on inflammatory biomarkers in healthy subjects and in patients with chronic and acute inflammatory diseases.
Methods
The research question in this systematic review was “Does omega-3 LC-PUFA affect biological levels of inflammatory biomarkers?” It included all randomised controlled trials (RCTs) in humans, both healthy and ill subjects, in which the effect of omega-3 LC-PUFA on fluid (mainly plasma or serum and cerebrospinal fluid) and cellular levels of inflammatory biomarkers was compared with that of placebo; inflammatory biomarkers were considered as primary outcomes. RCTs with prospective, parallel or crossover designs were all considered. The papers, or at least the abstract, had to be in English, but there was no restriction on publication type or sample size. We did not include any cohort, ecological or case-control study.
Inclusion and exclusion criteria
For a study to be considered, dietary supplementation or a specific diet had to be administered, excluding those on dietary recommendations or self-reporting alone. Studies were also excluded if a supplement was administered that could potentially leads to confounding the effects of omega-3 fatty acids (FA) and if there was no ethical approval. Particularly, five studies in which arginine and omega-3 FA were used together as supplement sources were excluded.
Participants
Eligible participants were individuals of all ages, either healthy or with acute or chronic disease. There was no restriction on the basis of gender, ethnicity, study setting or other characteristics.
Types of intervention
LC-PUFA treatments were selected, including eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA) or docosapentaenoic acid (DPA), either individually or in combination with each other or with another pharmacological treatment (or vitamin supplementation) if the study design allowed the effect of omega-3 FA to be isolated. There were no restrictions with regard to dosage or dose regimen.
Primary outcome measures
These had to include changes in the following inflammatory biomarkers: acute phase reactant proteins (PCR), eicosanoids (PGE1, PGE2, LTB4, F2-isoprostanes), cytokines (IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-13, TNF-α, interferon-δ), chemokines (CCL2, CCL3, CCL5, CCL11), adhesion molecules (sVCAM-1, sICAM-1, sE-selectin, sP-selectin) and related biofactors (sIL-1RII; sIL-6R, sTNF-Rs 1 and 2, CD11b, CD18, CD49, CCR2 and CCR5, MMP-7, MMP-9, MMP-12, TIMP-2, α1-ntichymotrypsin, fibrinogen, PAI-1, orosomucoid AGP), comparing measurements between baseline and the end of the intervention. Studies that did not evaluate inflammatory biomarkers were excluded.
Main stages of the systematic review
The main stages in the classification of studies for this systematic review are depicted in Fig. 1. Studies were identified on the PubMed database on February 11, 2011, without restriction on publication date until the search date, using the following equation: #1 “Fatty Acids, Omega-3”[MeSH] AND “Inflammation”[MeSH], which yielded 365 different articles. Next, the equation was modified to: #2 “Fatty Acids, Omega-3”[MeSH] AND “Inflammation”[MeSH] AND “humans”[MeSH Terms], yielding 172 results. A more specific search was then conducted for RCTs, using:#3 “Fatty Acids, Omega-3”[MeSH] AND “Inflammation”[MeSH] AND “humans”[MeSH Terms] NOT “Review “[All Fields] AND “Randomised Controlled Trial” [ptyp]. This search generated 58 original articles, out of which 26 met all requirements for inclusion in the review. The LILACS database was also searched on February 11 2011 but yielded no results.
Fig. 1 Main stages of the systematic review.
A tool, known as SYSCOLLAB (see Chapter 1), was developed and used for the selection, classification and validation of these articles, following the procedures described below.
Study Selection
Abstracts of references obtained by the search were examined by four reviewers (AG, CMA, MDM, ODR), who eliminated all references evidently ineligible for inclusion. Each reviewer studied all of the remaining full texts independently and made a selection. In cases of disagreement, a consensus was reached on the final list after discussion among all four reviewers.
Data extraction
Two reviewers (AG, ODR) introduced the data into a database, and discrepancies between them were resolved by a third reviewer (CMA). Missing data were obtained from the authors whenever possible, solicited on specific forms to avoid bias.
Assessment of quality
The quality of each publication was validated by applying the JADAD scale and the CONSORT checklist for clinical trials(Reference Jadad, Moore and Carroll16, Reference Schulz, Altman and Moher17)JADAD establishes the quality level according to five items(Reference Jadad, Moore and Carroll16) and the use of the CONSORT checklist ensures that publications follow quality guidelines. We developed a quantitative scale by assigning a value to each item in the list, as shown in Table 1, giving a quality score ranging from 0 to 30 (low to excellent).
Table 1 Validation Scale based on CONSORT test
Results
Table 2 lists the 26 publications selected for inclusion in the review, grouped according to study population (healthy, cardiovascular disease, other chronic diseases and acute disease). It also indicates the sample size, type of study, type and dose of omega-3 FA administered, major inflammatory biomarkers determined and main outcomes.
Table 2 Characteristics of selected studies(Reference Pot, Brouwer and Enneman18–Reference Wang, Li and Li40)
Footnote: NA, Not Available. Exptal, Experimental. EPA, Eicosapentaenoic Acid. DHA, Docosahexaenoic Acid. PGE1, Prostaglandins E1. PGE2, Prostaglandins E2. LTB4, Leukotriene B4. TNF-α, Tumour Necrosis Factor alpha. IL, Interleukin. CRP, C-Reactive Protein. ICAM, Inter Cellular Adhesion Molecule. VCAM, Vascular Cell Adhesion molecule. NA, Not Available. Exptal, Experimental. EPA, Eicosapentaenoic Acid. DHA, Docosahexaenoic Acid. PGE1, Prostaglandins E1. PGE2, Prostaglandins E2. LTB4, Leukotriene B4. TNF-α, Tumour Necrosis Factor alpha. IL, Interleukin. CRP, C-Reactive Protein. ICAM, Inter Cellular Adhesion Molecule. VCAM, Vascular Cell Adhesion molecule, NZGLM, New Zealand Green Lipped Mussel. NA, Not Available. Exptal, Experimental. EPA, Eicosapentaenoic Acid. DHA, Docosahexaenoic Acid. PGE1, Prostaglandins E1. PGE2, Prostaglandins E2. LTB4, Leukotriene B4. TNF-α, Tumour Necrosis Factor alpha. IL, Interleukin. CRP, C-Reactive Protein. ICAM, Inter Cellular Adhesion Molecule. VCAM, Vascular
Healthy individuals
Out of the 26 selected articles, 10 reported data on the effect of dietary omega-3 FA on inflammatory biomarkers in healthy participants(Reference Pot, Brouwer and Enneman18–Reference Wang, Li and Li40). Quality scores for these studies ranged from 3 to 5 for the JADAD scale, and from good to excellent (G-E) for the CONSORT scale. All 10 studies were in healthy adults (sample sizes from 12 to 161): three in young adults (19–40 years), six in the middle-aged (40–65 years) and one in the elderly (>65 years). Two of these studies investigated the effects of omega-3 dietary supplementation on exercise-induced markers of inflammatory mediators in exercise-trained men(Reference Bouwens, van de and Dellschaft19, Reference Nieman, Henson and McAnulty25) (Table 2). In eight of these studies, omega-3 was delivered in fish oil capsules at daily doses of 0·12–2·22 g of EPA and 0·26–0·56 g of DHA, alone or in combination, for 2–26 weeks. One study supplied 0·6 g EPA and 0·2 g DHA daily in supplemented soybean milk for 12 weeks, while another provided 2 ml/day of New Zealand Green Lipped Mussel (97 mg EPA and 72 mg DHA) and 2 ml/day of fish oil (87 mg EPA and 50 mg DHA), administered using a dropper. Only one study investigated the effect of parenteral lipid infusion (4·55–9·8 g EPA +4·9–10·9 g DHA), for a period of 48 h and with cross-over repetition of the infusion after 3 months.
Inflammation outcome variables in the above studies were serum markers, including CRP, cytokines (IL-1α, IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-13, TNF-α, interferon-γ), chemokines (CCL2, CCL3, CCL5, CCL11) and soluble forms of cell adhesion molecules (ICAM-1, VCAM-1, E-Selectin). Four of the studies investigated the effects of omega-3 supplementation on peripheral blood cell levels of inflammatory markers induced by lipopolysaccharides (LPS) or endotoxins, including eicosanoids (PGE1, PGE2, LTB4) and cytokines (IL1 β, IL 6, IL 8, IL10 and TNF-α). Only one study measured the expression of adhesion molecules in monocytes (CD11b, CD18, CD49, CCR2 and CCR5).
Five studies determined the effect of omega-3 supplementation on serum CRP levels in healthy individuals, with one report detecting no effect in healthy volunteers without active inflammation, whereas another found lower exercise-induced serum CRP levels after EPA/DHA supplementation in trained men(Reference Bloomer, Larson and Fisher-Wellman26).
Omega-3 supplementation did not significantly affect the serum concentration of any of the cytokines and chemokines analysed in healthy individuals, with the exception of a reduction in exercise-induced TNF-α(Reference Bloomer, Larson and Fisher-Wellman26).
Plasma soluble forms of cell adhesion molecules were measured in three of these studies(Reference Pot, Brouwer and Enneman18, Reference Yusof, Miles and Calder20, Reference Thies, Miles and Nebe-von-Caron24), finding no effects except for a decrease in sICAM-1(Reference Yusof, Miles and Calder20) and sVCAM-1(Reference Thies, Miles and Nebe-von-Caron24) after supplementation with 1·8 g EPA plus 0·3 g DHA (8 weeks) and 0·72 g EPA plus 0·28 g DHA (12 weeks), respectively. This fish oil supplementation significantly decreased the release of PGE1 and LTB4 alone after ex vivo LPS stimulation(Reference Schubert, Kitz and Beermann21), whereas parenteral omega-3 lipid emulsion significantly suppressed the generation of TNF-α, IL-1, IL-6 and IL-8 by monocytes in response to LPS, although their surface expression of relevant adhesive molecules was unchanged(Reference Mayer, Meyer and Reinholz-Muhly39). Finally, only the study by Bouwens et al. (Reference Yusof, Miles and Calder20) examined the effect of omega-3 intake on gene expression in peripheral blood mononuclear cells (PBMCs), finding that EPA+DHA intake reduced the expression of genes involved in inflammatory-related pathways, such as nuclear transcription factor κB (NF- κB) signalling and eicosanoid synthesis.
Cardiovascular disease
Eight of the articles selected in this review addressed the effect of omega-3 LC-PUFA on inflammatory molecules as biomarkers of CVD risk(Reference Cawood, Ding, Napper and Young27–Reference Skulas-Ray, Kris-Etherton and Harris33, Reference Nilsen, Almdahl and Svensson41), with quality scores ranging from 3 to 5 for the JADAD scale and from good to excellent (G-E) for the CONSORT scale. Seven of these studies were in adults (>18 years) and only one was in hyperlipidaemic children and young people (9–19 years)(Reference Engler, Engler and Malloy31). The study populations in the eight selected studies comprised: chronic heart failure patients with presence or high risk of CVD (Reference Zhao, Shao and Teng28); patients awaiting carotid endarterectomy(Reference Cawood, Ding, Napper and Young27); patients undergoing coronary bypass surgery(Reference Nilsen, Almdahl and Svensson41), patients with metabolic syndrome(Reference Tulk and Robinson29) or type 2 diabetes(Reference Mori, Woodman and Burke32); overweight individuals(Reference Browning, Krebs and Moore30) and people with moderate hyperlipidaemia but otherwise healthy(Reference Skulas-Ray, Kris-Etherton and Harris33).
All studies in this section used capsules to supply EPA and DHA, alone or combined, except for a postprandial study that administered an oily supplement (1 g fat per kg body weight) containing 2·25 % EPA plus 0·98 % DHA in a high omega-3 PUFA blend or 0·37 EPA plus 0·16 DHA in a low omega-3 PUFA blend(Reference Tulk and Robinson29). When combined, EPA and DHA doses ranged from 0·6 g/d (0·36 g/d EPA plus 0·24 g/d DHA)(Reference Zhao, Shao and Teng28) to 6 g/d(Reference Nilsen, Almdahl and Svensson41) for 12 weeks. When administered alone, the regimen was 3·68 g/d DHA for 13 or 6 weeks and 3·84 g/d of EPA for 6 weeks(Reference Mori, Woodman and Burke32).
Outcome variables in the studies in this group were cytokines (IL-1β, IL-2, IL-6 and IL-6 soluble receptor, IL-8, IL-10, TNF-α, TNF-R1, TNF-R2); serum markers, including CRP, α1-antichymotrypsin (ACT), fibrinogen, tissue factor pathway inhibitor (TFPI), tissue plasminogen activator inhibitor (PAI-1), thrombin-antithrombin III (TAT) complexes, α1-acid glycoprotein (AGP) and sialic acid; MMP2; soluble adhesion molecules (ICAM-1, VCAM-1, E-Selectin), pro-brain natriuretic peptide (NTproBNP), serum amyloid A (SAA), granulocyte colony stimulating factor (G-CSF) and granulocyte monocyte-colony stimulating factor (GM-CSF). There were also assessments of the expression of IL-6, IL-10, TNF-α, ICAM-1, MMP3, MMP7, MMP8, MMP9, MMP12, MMP13, TIMP1, TIMP2 and 36B4 in plaque cells(Reference Cawood, Ding, Napper and Young27) and of IL-1β, IL-6 and TNF-α in mononuclear cells(Reference Skulas-Ray, Kris-Etherton and Harris33).
Omega-3 supplementation did not significantly affect the serum concentration of cytokines and chemokines with the exception of IL-6 and TNF-α. IL-6 levels, which were lower after omega-3 PUFA in two of the studies(Reference Zhao, Shao and Teng28, Reference Browning, Krebs and Moore30). Cawood et al. (Reference Cawood, Ding, Napper and Young27) reported a decreased expression of IL-6 mRNA in plaque cells, whereas Mori et al. (Reference Mori, Woodman and Burke32) and Tulk et al. (Reference Tulk and Robinson29 )found no significant effect on this cytokine. Lower TNF-α plasma levels were reported by Zhao et al. (Reference Zhao, Shao and Teng28) after omega-3 PUFA, but this reduction was not observed by Mori et al. (Reference Mori, Woodman and Burke32).
A decrease in CRP was reported after 12 weeks of omega-3 supplementation(Reference Browning, Krebs and Moore30), but no such effect was found by Mori et al. (Reference Mori, Woodman and Burke32), Zhao et al. (Reference Zhao, Shao and Teng28), or Skulas-Ray et al. (Reference Skulas-Ray, Kris-Etherton and Harris33). Zhao et al. (Reference Zhao, Shao and Teng28) observed an association between omega-3 LC-PUFA supplementation and decreased plasma levels of high sensitive CRP (hsCRP), especially among smokers, while another study found the postprandial CRP response to be similar after intake of a low or high omega-3 PUFA dose(Reference Tulk and Robinson29). With regard to adhesion molecules, one study showed lower plasma ICAM-1 levels(Reference Zhao, Shao and Teng28). Finally, analyses of cell gene expression showed that supplementation with EPA plus DHA significantly reduced mRNA levels of MMP7, MMP9, MMP12, IL-6, ICAM-1 and TIMP-2 in plaque cells(Reference Cawood, Ding, Napper and Young27) but had no effect on mononuclear cell gene expression of IL-1β, IL-6 or TNF-α(Reference Skulas-Ray, Kris-Etherton and Harris33).
Other chronic diseases
Effects of omega-3 FA on inflammatory markers were studied in chronic renal disease patients undergoing haemodialysis in two of the selected articles(Reference Bowden, Wilson and Deike34, Reference Perunicic-Pekovic, Rasic and Pljesa35) and in patients with Alzheimer's disease (AD) in another two(Reference Freund-Levi, Hjorth and Lindberg36, Reference Vedin, Cederholm and Freund Levi37). The quality scores for these articles ranged from 3 to 5 for the JADAD scale and from very good to excellent (V-E) for the CONSORT scale.
In the Bowden et al. study(Reference Bowden, Wilson and Deike34), eligible patients undergoing chronic haemodialysis maintenance were randomly assigned to an omega-3 (experimental, n 18) or corn oil (control, n 15) group. The experimental group received 960 mg EPA/d plus 600 mg DHA/d orally for 6 months, which had a significant effect on CRP levels not observed in the placebo group. Perunicic-Perovic et al. (Reference Perunicic-Pekovic, Rasic and Pljesa35) examined 42 haemodialysis patients and 16 healthy controls, administering the patients with daily supplements of 2·4 g omega-3 LC-PUFA (180 mg EPA and 120 mg DHA) for 2 months, and reported a significant decrease in inflammatory markers (IL-6 and TNF-α) in the supplemented group.
Vedin et al. (Reference Vedin, Cederholm and Freund Levi37) carried out an RCT in 174 AD patients who received 1·7 g DHA plus 0·6 g EPA or placebo for 6 months. The omega-3 group showed significant decreases in IL-6, IL-1 and G-CSF secretion after LPS stimulation of PBMCs. In addition, increases in the concentrations of DHA and EPA were negatively associated with IL-1 and IL-6 release in all of the patients. However, no decrease in the secretion of TNF-α, IL-8, IL-10, or GM-CSF was observed. Freund-Levy et al. (Reference Freund-Levi, Hjorth and Lindberg36) found that omega-3 FA treatment for 6 months did not influence inflammatory biomarkers in cerebrospinal fluid (CSF) or plasma of patients with mild to moderate AD. In their study, 35 patients were randomly assigned to a daily intake of 2·3 g omega-3 FA or placebo for 6 months, analysing CSF and plasma levels of IL-6, TNF-α and soluble IL-1 receptor type II at baseline and 6 months.
Acute diseases
Three of the selected articles studied the effects of omega-3 FA on inflammatory biomarkers in septic patients(Reference Mayer, Gokorsch and Fegbeutel23, Reference Barbosa, Miles and Calhau38–Reference Mayer, Meyer and Reinholz-Muhly39) and one investigated them in patients with severe acute pancreatitis(Reference Wang, Li and Li40). Their quality scores ranged from 3 to 5 for the JADAD scale and from very good to excellent (V-E) for the CONSORT scales, respectively.
The studies in septic patients were randomised but single blind(Reference Barbosa, Miles and Calhau38) or open label(Reference Mayer, Gokorsch and Fegbeutel23, Reference Mayer, Meyer and Reinholz-Muhly39). The treatment was short (5–7 days) in all three studies, and the omega-3 LC-PUFA dose ranged from 6·4 to 23·6 g/d. In the study by Barbosa et al. (Reference Barbosa, Miles and Calhau38), parenteral administration of fish oil emulsion significantly decreased IL-6 levels, with a parallel increase in IL-10, and improved the gas exchange, finding a tendency to a shorter hospital stay in patients receiving this supplementation.
Mayer et al. (Reference Mayer, Gokorsch and Fegbeutel23) studied 21 patients with sepsis requiring parenteral nutrition, who were randomised to receive an omega-3 lipid emulsion rich in EPA and DHA or a conventional omega-6 lipid emulsion for 5 days. Generation of proinflammatory cytokines by mononuclear leukocytes was markedly amplified by omega-6 lipid intake and suppressed by omega -3 intake. Mayer et al. (Reference Mayer, Meyer and Reinholz-Muhly39) also studied ten patients with septic shock and eight healthy controls. Patients requiring Parenteral Nutrition (PN) received intravenously an omega-3 or omega-6 lipid emulsion (five per group) for a 10-day period. With the omega-6 lipid infusion regimen, the impairment of neutrophil functions persisted or worsened, whereas LTB5 and LTB4 appeared upon neutrophil stimulation in the omega-3 lipid group and neutrophil function was significantly improved.
Wang et al. (Reference Wang, Li and Li40) performed an RCT in 47 severe acute pancreatitis patients randomly assigned to receive PN for 5 days with identical amounts of amino acids, glucose and fat but different lipid compositions. The control group received a soybean oil based emulsion and the omega-3 FA group was supplemented with 0·15–0·2 g/kg/d fish oil. Fish oil-treated patients had significantly lower CRP levels and blood cell counts after 5 days of PN.
Discussion
Omega-3 FA supplementation is known to lower inflammation biomarkers in the plasma and cells of patients with acute and chronic diseases, but its effect on healthy subjects is less well documented. Numerous authors have studied the combined effects of EPA and DHA but only a few have administered EPA or DHA alone. In addition, very few studies have investigated the effects of these FA on in vivo cell inflammatory biomarkers, on ex vivo cytokine release or on the expression of genes related to inflammation cascade signalling.
Healthy subjects
In the studies of healthy participants selected in this review, omega-3 supplementation generally had no effect on inflammatory biomarkers, which may be due to their low levels in serum, minimizing the possibility of their reduction through the intake of fish oil.
Fish oil supplementation was found to have only isolated effects on plasma soluble forms of cell adhesion molecules, reducing sICAM-1(Reference Yusof, Miles and Calder20) and sVCAM-1(Reference Thies, Miles and Nebe-von-Caron24) levels. The stronger association of DHA changes than EPA changes with sICAM described by Yusof et al. (Reference Yusof, Miles and Calder20) suggested that DHA may be more anti-inflammatory than EPA. However, this hypothesis was ruled out by the neutral effect on sICAM found by Thies et al. (Reference Thies, Miles and Nebe-von-Caron24), who used a very similar dosage of DHA (0·28 g vs. 0·3 g) alone or in combination with 0·72 g EPA during a longer period (12 vs. 8 weeks) and only found an influence on sVCAM-1 levels. The different effects of omega-3 on sICAM-1 and sVCAM-1 concentrations may suggest that VCAM-1 and ICAM-1 are regulated by different mechanisms and are not equally sensitive to n-3 PUFA.
The release of cytokines after ex vivo LPS stimulation was not affected by oral administration of omega-3(Reference Schubert, Kitz and Beermann21, Reference Thies, Miles and Nebe-von-Caron24–Reference Nieman, Henson and Gross42). However, in a study of healthy volunteers with a cross-over design, a decrease in LPS after parenteral administration of omega-3 lipid emulsion(Reference Mayer, Meyer and Reinholz-Muhly39) induced the production of cytokines (TNF-α, IL-1, IL-6 and IL-8), demonstrating that a short (48 h) infusion with omega-3 lipids is sufficient to significantly suppress monocyte proinflammatory cytokine generation. PN studies in healthy volunteers are rare and ethically questionable, making it impossible to draw conclusions about the anti-inflammatory effect of omega-3 in this situation.
Findings of the pioneering study by Bouwens et al. (Reference Bouwens, van de and Dellschaft19) opened up a new approach for research into the effects of omega-3, demonstrating the potential of PBMC gene expression profiling for determining the effects of nutrition on human health status. However, this study was performed in elderly individuals, who are likely to express a slightly more proinflammatory gene expression profile, given that aging is associated with a chronic, low-grade elevation of inflammatory activity(Reference Bruunsgaard, Pedersen and Pedersen43). Therefore, the results cannot be generalised to the whole population, and further research is required to confirm these findings in other age groups.
One way to investigate the effect of omega-3 supplementation on inflammatory markers in healthy individuals is to stimulate the inflammatory response by exercise. Transient inflammation and immune dysfunction are part of the normal human response to prolonged and intense exercise and are characterised in part by the production of inflammatory cytokines, acute-phase proteins and oxidative stress from reactive oxygen species, as well as by alterations in innate and adaptive immunity(Reference Nieman, Henson and Gross42, Reference Toft, Thorn and Ostrowski44). However, conflicting results were published by two studies of inflammation. One found decreased levels of exercise-induced CRP, TNF-α and IL-6 after omega-3 supplementation(Reference Bloomer, Larson and Fisher-Wellman26), whereas the other observed no change in these levels(Reference Nieman, Henson and McAnulty25). Discrepancies may be due to differences in the training status of participants, the duration of supplementation, the dose and FA composition of the n-3 PUFA administered, and the exercise protocol used. Thus, the two aforementioned studies(Reference Nieman, Henson and McAnulty25, Reference Bloomer, Larson and Fisher-Wellman26) differed in the total daily dosage of DHA (2208 mg vs. 400 mg) and exercise protocol (3 h/d cycling for 3 days vs. treadmill walking with weighted backpack for 60 min). The omega-3 supplementation was unable to counteract the acute inflammatory response induced in the former study, while it had no effect on the very weak inflammatory response induced by exercise in the latter, with the exception of decreased CRP and TNF-α levels at rest. In conclusion, the exercise-related omega-3-supplementation studies included in this review do not provide strong evidence of anti-inflammatory effects in trained or untrained participants.
Cardiovascular diseases
Results of selected articles on the effect of omega-3 LC-PUFA on inflammatory molecules as biomarkers of CVD risk remain inconclusive, and further studies are required. The evidence of clinical efficacy is not strong enough to make final recommendations for a specific dose or duration of supplementation.
The only inflammatory biomarkers affected by omega-3 PUFA in multiple studies were IL-6 and CRP, while TNF-α, sICAM-1 and GM-CSF were decreased in single studies. No effects on plasma biomarkers were found when omega-3 PUFA was administered for less than 12 wks, even at high doses, as shown by Nilsen et al. (Reference Nilsen, Almdahl and Svensson41) and Mori et al. (Reference Mori, Woodman and Burke32). However, shorter treatments may possibly induce changes in the mRNA expression of selected inflammatory genes by specific cells without affecting plasma levels, as observed by Cawood et al. (Reference Cawood, Ding, Napper and Young27)
There is no consensus on the doses required to exert an anti-inflammatory effect, given that limited positive effects have been observed after low(Reference Zhao, Shao and Teng28) or high(Reference Browning, Krebs and Moore30) doses. It is therefore not possible to identify the dose needed to obtain an anti-inflammatory effect in patients with CVD risk.
In relation to the type of FA, most studies administered a combination of EPA plus DHA, and only Mori et al. (Reference Mori, Woodman and Burke32) compared the effects of EPA versus DHA supplementation (for 6 weeks), reporting that neither had a significant effect. Hence, there is inadequate evidence to establish whether cardiovascular inflammation can be reduced by EPA or DHA or by the synergic effects of both.
Two of the studies showing an impact on IL-6 after omega-3 supplementation were reported to have inadequate power to detect a significant effect on inflammatory status, due to their small sample sizes (n 32, Browning et al., 2006, and n 38, Zhao et al., 2009)(Reference Zhao, Shao and Teng28, Reference Browning, Krebs and Moore30). The same limitations apply to the effects on CRP, whose reduction was reported in the same studies. Specifically, Zhao et al. (Reference Zhao, Shao and Teng28) found omega-3 LC-PUFA to be associated with a decrease in plasma CRP levels in smokers but not in non-smokers. Although there is no consensus on the reproducible effects of cigarette smoke, it appears evident that smoking can increase the inflammatory response(Reference Smith, Paszkiewicz and Hutson45), and the results obtained by Zhao et al. (Reference Zhao, Shao and Teng28) suggest that smokers, with an increased inflammatory status, may take greater advantage of the anti-inflammatory effects of omega-3 PUFA.
Finally, the postprandial inflammatory response after an oral fat tolerance test was not modified by increasing omega-3 PUFA content, indicating that omega-3 PUFA does not have an acute effect and must be incorporated into cell membranes before exerting an anti-inflammatory effect.
Based on the results related to CVD risk examined in this review, there is inadequate evidence to establish a relationship between the consumption of omega-3 LC-PUFA and beneficial inflammatory biomarker outcomes, probably due to discrepancies in protocols, doses and treatment times. Further research is necessary to standardize protocols and determine suitable doses and regimens.
Other chronic diseases
Only two RCTs studied inflammatory biomarkers in AD, with contradictory results. Omega-3 FA affected the release of cytokines from blood mononuclear leukocytes but did not influence inflammatory biomarkers in CSF(Reference Vedin, Cederholm and Freund Levi37), although the correlation between soluble IL-1 receptor II (sIL-1RII) and amyloid peptide Aβ1-42 may reflect reciprocal interactions between IL-1 and Aβ peptides(Reference Freund-Levi, Hjorth and Lindberg36). However, it remains to be established whether omega-3 FA attenuate the release of cytokines from brain cells in AD.
CVD is a comorbid condition in patients with end-stage renal disease (ESRD) and is associated with a high mortality rate. The CVD risk in patients undergoing chronic haemodialysis can be ascribed to elevated serum lipid levels and blood pressure, although additional factors (e.g., acute phase reactants) may lead to a risk profile that is 10- to 20-fold higher in comparison to apparently healthy individuals and is responsible for the very high (50 %) annual mortality in ESRD patients(Reference Vernaglione, Cristofano and Chimienti46).
The use of omega-3 LC-PUFA has been proposed to control endothelial dysfunction and thereby control CRP levels. Additionally, ESRD patients do not usually consume enough omega-3 FA in their diet(Reference Saifullah, Watkins and Saha47). The consumption of 960 mg/d of EPA and 600 mg/d of DHA can lower CRP in ESRD patients, even though there is no clinically significant change in their triglyceride levels(Reference Bowden, Wilson and Deike34). Furthermore, ESRD patients on haemodialysis who received 2·4 g of n-3 LC-PUFA per day for 2 months exhibited a significant decrease in inflammatory markers (IL-6 and TNF-α)(Reference Perunicic-Pekovic, Rasic and Pljesa35). However, compliance was not assessed. Further studies in ESRD patients are required to identify a complete set of biomarkers involved in inflammation.
Acute diseases (Critically ill patients)
In the cascade of inflammatory response during the acute phase of severe acute pancreatitis, PN supplemented with omega-3 FA is a treatment option for patients with systemic inflammatory response syndrome (SIRS) or sepsis complicated by multiple organ failure(Reference Martin and Stapleton48). After 5 days of PN plus omega-3 FA, severe acute pancreatitis patients had significantly lower CRP levels and blood white cell counts and showed improved respiratory function and shortened continuous renal replacement therapy time, suggesting that the systemic response to pancreatic and organ injury is attenuated by omega-3 LC-PUFA(Reference Wang, Li and Li40).
Omega-3 FA are thought to suppress inflammation and ameliorate the course of infection by reducing proinflammatory eicosanoids and cytokines. However, only one RCT assessed CRP changes, and further experimental studies and clinical trials are needed to establish the appropriate doses of omega-3 LC-PUFA and the specific effects of EPA and DHA on inflammatory biomarkers.
Novel potent lipid mediators derived from LC-PUFA, arachidonic acid (AA), EPA and DHA, namely lipoxins, resolvins, and protectins, have been identified in resolving exudates over recent years, and their biosynthetic pathways and actions have been described. Their activity is likely to account in part for some of the immunoregulatory and disease-modifying actions attributed to dietary omega-3 LC-PUFA(Reference Bannenberg and Serhan49). However, no controlled studies in humans have evaluated the influence of omega-3 FA on these new lipid mediators.
Dietary recommendations of n-3 LC-PUFA in health and disease
A recent report of the FAO has established that for adult males and non-pregnant/non-lactating adult females 0·250 g/d of EPA plus DHA is recommended, with insufficient evidence to set a specific minimum intake of either EPA or DHA alone(50). The upper value of acceptable macronutrient distribution range for EPA + DHA consumption has been set at 2 g/d due to experimental evidence indicating that high supplement intakes of n-3 LC-PUFA may increase lipid peroxidation and reduce cytokine production. However, it is acknowledged that higher consumption values, as high as 3 g/d reduce some cardiovascular risk factors and have not had adverse effects in short- and intermediate term randomised trials.
The American Heart Association (AHA) recommended 0,5 to 1 g/d of omega-3 EPA and DHA for individuals with borderline fasting triglyceride levels (150 to 199 mg/dl), 1 to 2 g/d for individuals with high fasting-triglyceride levels (200–499 mg/dl), and 2 to 4 g/d for individuals with very high fasting-triglyceride levels (>500 mg/dl)(Reference Kris-Etherton51).
Based in our study of RCT, a moderate consumption of n-3 LC-PUFA (0·9 to 2 g/d) does not change inflammatory biomarkers in healthy subjects(Reference Pot, Brouwer and Enneman18–Reference Mayer, Gokorsch and Fegbeutel23). However, about 4 g/d of EPA/DHA supplementation decreases resting levels of inflammatory biomarkers in exercise-trained men, but it does not appear necessary for exercise-induced attenuation in either inflammation or oxidative stress(Reference Nieman, Henson and McAnulty25). n-3 PUFA intake of 1·2 to 6 g/d decrease plasma levels of the pro-inflammatory cytokines and is associated with a reduced number of foam and T cells in atherosclerotic plaques(Reference Cawood, Ding, Napper and Young27 )and with reduced levels of some inflammatory biomarkers in plasma(Reference Zhao, Shao and Teng28). However, there is no consensus on the doses required to exert an anti-inflammatory effect(Reference Zhao, Shao and Teng28–Reference Tulk and Robinson29). Regarding the effects of n-3 LC-PUFA on inflammatory biomarkers in neurodegenerative diseases, AD patients treated with DHA exhibit a reduced release of IL-1β, IL-6 and granulocyte colony-stimulating factor from PBMCs(Reference Vedin, Cederholm and Freund Levi37). Furthermore, in patients with chronic renal disease, the intake of 1·5 to 2·4 g of n-3 LC-PUFA decreases CRP and other cytokines(Reference Bowden, Wilson and Deike34, Reference Perunicic-Pekovic, Rasic and Pljesa35).
Omega-3 LC-PUFA are able to suppress inflammation and ameliorate the course of infection by reducing proinflammatory eicosanoids and cytokines in critically ill patients (Reference Wang, Li and Li40, Reference Martin and Stapleton48) but further research is needed before definitive recommendations can be made on doses and on the routine use of omega-3 FA in the critically ill.
Despite n-3 fatty acids reduce the risk of cardiovascular disease there has been concern that these fatty acids may increase lipid peroxidation. However, post-intervention urinary F2-isoprostanes were decreased by EPA and DHA in type 2 diabetic patients(Reference Mori, Woodman and Burke32). In addition, it has been reported that supplementation with n-3 fatty acids decreases non-enzymatic free radical–catalyzed isoprostane formation in healthy humans(Reference Nälsén, Vessby, Berglund, Uusitupa, Hermansen and Riccardi52. Likewise, plasma free F2-isoprostane concentrations were lower after fish-oil supplementation than after sunflower-oil supplementation in postmenopausal women(Reference Higdon, Liu, Du, Morrow, Ames and Wander53). In conclusion, 3–4 g of n-3 LC-PUFA intake does not have an apparent effect on lipid peroxidation.
Acknowledgements
ODR and AG contributed to the planning of the search of the literature, designed the analysis and results presentation and created the tool for assessing the quality of the articles. All authors were involved in the analyses of the articles. ODR wrote the introduction and methodology of the first manuscript and CMA, MDM and AG wrote the results section; CMA focused on healthy subjects, MDM on cardiovascular patients and AG on other chronic diseases and critically ill patients. All authors discussed and revised all drafts and approved the final manuscript.
There are no conflicts of interest to report.
This work has been funded in part by the Instituto de Salud Carlos III del Ministerio de Ciencia e Innovación Red SAMID, RETIC n. RD08/0072.
