The Effect of Omega-3 Fatty Acids on Biomarkers of Inflammation: A Rapid Evidence Assessment of the Literature

Transcription

1 MILITARY MEDICINE, 179, 11:2, 2014 The Effect of Omega-3 Fatty Acids on Biomarkers of Inflammation: A Rapid Evidence Assessment of the Literature Raheleh Khorsan, PhDc* ; Cindy Crawford, BA ; John A. Ives, PhD ; Avi R. Walter, MSc ; Wayne B. Jonas, MD ABSTRACT Introduction: Previous studies of omega-3 fatty acids report improved outcomes where inflammation is a key factor. The objective of this systematic review is to evaluate effects of omega-3s on inflammatory biomarkers. Methods: Randomized clinical studies that measured the influence of omega-3 fatty acids on inflammatory biomarkers were identified using a comprehensive search. Eligible studies were rated with the American Dietetic Association Evidence Analysis Manual and Grading of Recommendations, Assessment, Development and Evaluation (GRADE) process to examine study quality and risk/benefit. Results: 112 studies were included. Over 65% reported statistically significant effects. The majority were scored as low risk of bias (high quality) and scored strong (cardiac populations and critically ill) to weak (Alzheimer s Disease, hypertriglyceridemia/diabetes, and obesity) on the risk/benefit ratio evidence for modulation of inflammatory biomarkers. There was inadequate data to determine a GRADE for inflammatory biomarker studies for some conditions (healthy individuals, rheumatoid arthritis, metabolic syndrome, renal disease, pregnancy, or children). Conclusion: Clinical literature on the effects of omega-3 fatty acids on inflammatory biomarkers contains mostly small sample sizes, is neutral to high quality, and report mixed effects. Larger studies examining dose and delivery are needed. INTRODUCTION Omega-3 Fatty Acids In 1946, the first major study of essential fatty acids (EFA) identified vitamin-like properties of linoleic acid (18:2n-6) (omega-6) and linolenic acid (18:3n-3) (omega-3) in human infants. 1 Today, there is much debate over assisting the public in identifying and preventing the current imbalanced intakes of omega-3 and omega-6 nutrients that cause preventable clinical disorders and continue escalating healthcare costs. 2 Indeed, studies continue to assess the importance of the omega-6/omega-3 fatty acid ratio in cardiovascular disease and other chronic diseases. 3 Current studies advocate lower intake of omega-6 compared to greater intake of omega-3 fatty acids in reducing the risk of some of the chronic diseases of high prevalence that are associated with inflammation. 1,3 Omega-3 fatty acids (also known as n-3 or w-3) include a-linolenic acid (ALA), eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and docosapentaenoic acid (DPA). Among fatty acids (FAs), long-chain omega-3 polyunsaturated FAs (PUFAs) possess the most potent immunomodulatory activity, 4 and among the omega-3 PUFAs, those from fish oil (FO) are rich sources of the bioactive long-chain n-3 *Samueli Institute, 2101 East Coast Highway, Suite 300, Corona del Mar, CA Department of Planning, Policy and Design, School of Social Ecology, University of California Irvine, Irvine, CA Samueli Institute, 1737 King Street, Suite 600, Alexandria, VA The views opinions and/or findings contained in this work are those of the author(s) and should not be construed as an official Department of the Army position, policy, or decision unless so designated by other documentation. doi: /MILMED-D including EPA and DHA. The majority of DHA and EPA are found in diets containing cold-water fatty fish, FOs, flax, range-fed poultry, eggs and farm animals, and in supplements and prescription formulations. Other less biologically potent FAs include plant-based ALA with n-3 PUFA. 4 The major plant omega-3 ALA is found in walnuts, soybean, canola oil, and flaxseed. Epidemiological and experimental studies on the benefits of FOs and omega-3 FA consumption has increased over the past few decades. 5 Diets enriched with high amounts of both plant- and marine-derived omega-3 PUFAs seem to be associated with a lower incidence of coronary heart disease (CHD) 6,7 and a reduction in cardiovascular events. 8 In addition, both animal and clinical studies support the use of omega-3 FAs for inflammatory 9 and autoimmune disease management. 10 There have been a number of clinical trials assessing the benefits of dietary supplementation with FOs in treating several inflammatory and autoimmune diseases among humans, 4 including rheumatoid arthritis (RA), 11,12 Crohn s disease, 13,14 inflammatory bowel disease, 13,15 17 as well as other inflammatory conditions. 18,19 Many of the placebo-controlled trials of FO with chronic inflammatory diseases report significant benefit, including decreased disease activity and a lowered use of anti-inflammatory drugs. 4,20 In 2012, a systematic review (SR) of randomized controlled trials (RCTs) on omega-3 FAs and inflammatory biomarkers in healthy and clinical populations was conducted utilizing PubMed and LILACS databases. 21 This study selected 26 RCTs published over the preceding 10 years and written in the English language that evaluated omega-3 diet supplementation or dietary manipulation. Studies were excluded if a supplement was administered that could potentially confound the effects of omega-3 FAs or if there was 2

2 no ethical approval. Specifically, they excluded five studies in which arginine and omega-3 FA were used together as supplement sources. Twenty-six articles met their inclusion criteria. The quality of each publication was determined by using the JADAD scale and the CONSORT checklist, and combining the results to come up with an aggregate total score. The outcome of the SR was that omega-3 supplementation may be beneficial for lowering levels of inflammation and endothelial activation in cardiovascular disease and other chronic and acute diseases, including chronic renal disease, sepsis, and acute pancreatitis. 21 However, numerous studies evaluating the effects of omega-3 in which inflammatory biomarkers were considered as outcomes were omitted from their review. Moreover, in the interim, there have been additional clinical studies published. The SR reported here builds on the Rangel-Huerta et al SR 21 and reviews the potential of omega-3 FAs to ameliorate inflammatory conditions, especially those of military relevance. In support of a conference in 2009, titled Nutritional Armor for the Warfighter Workshop, 47 speakers noted that omega-3 FAs have the potential to ameliorate inflammatory conditions in warfighters. As a consequence of a conference where speakers discussed the pros and cons for omega supplementation in warfighters, the authors were asked to evaluate the effects of omega-3 FAs on biomarkers of inflammation in clinical studies. The purpose of this report is to provide an updated and more extensive analysis of the available evidence on markers of inflammation, clinical efficacy, and safety of n-3 PUFA in healthy and clinically ill humans, as compared to standard (non n-3 enriched) care. Objective The purpose of this review is to examine the quality and quantity of clinical research using the REAL SR approach and to assess the evidence for the effects of omega-3 FA interventions on inflammatory biomarkers for clinical conditions associated with inflammation and in studies of healthy humans. The specific objectives of this review were to (1) survey the available literature on omega-3 FAs for conditions related to inflammation in the human population using a REAL ; (2) assess the literature for quantity, quality, and effectiveness of omega-3 FAs on inflammatory biomarkers that used RCTs study design; and (3) identify gap areas and suggest next steps for the research field. METHODS The following databases were searched from inception through February 2014: Pubmed, CINAHL, PsycINFO, all of Cochrane EBM Databases, and the National Agricultural Library Online. The initial search terms were EPA, DHA, omega-3 PUFA, or Omega, coupled with inflammation terms or those conditions known as inflammatory conditions (i.e., RA, etc.). Medical subject headings vocabulary was utilized in relevant databases where applicable. The search was restricted to only English language, peer-reviewed published literature, and experiments involving human subjects. Reference lists were pearled of all relevant studies to capture additional reports not found through traditional searching of databases, as well as conversing with subject matter experts (SMEs). Eligibility Criteria Articles were included in this REAL if they met the following criteria: (1) an inflammatory condition and/or inflammation itself was being assessed; (2) the intervention involved omega-3 FA supplementation; (3) a control/ comparator intervention was assessed at same time; and (4) the outcome assessed inflammatory biomarkers as primary or secondary outcomes, and the study design was a RCT or SR published in the English language. Articles were excluded if the report was on a combination treatment, such as omega-3 FAs plus Vitamin E or the nutritional package included omega-3 FAs as one of the components where the effect of omega-3 FAs alone could not be directly assessed, or if the study was unable to control for confounding effects of the add-on therapy; however, studies that administered combination treatments of FAs only, such as omega-3 FAs and omega-6 FAs or if the effect of omega-3 FAs alone could be directly assessed were included. Any study not assessing inflammatory biomarkers was excluded. Three investigators (CC, ARW, and RK) independently screened titles and abstracts for relevance based on the inclusion criteria. Any disagreements about including a study were resolved through discussion and consensus with the review team, or by SMEs (JAI and WBJ). Assessment and Data Extraction Methodological quality of the included studies was assessed independently by three reviewers (CC, RK, and AW). The individual studies were all RCTs and were evaluated for study quality and bias using Section 1 and 2 of the American Dietetic Association (ADA) Evidence Analysis Manual Criteria Checklist for Primary Research. 52 This checklist was developed to conduct evidence analysis on nutrition and dietary supplement topics of research and is well-accepted in the field. 53 Each study, once assessed, received a score of positive/high (), neutral (0), or negative/ low ( ) according to the quality criteria for ADA. Data extraction was conducted to describe each study included by population, whether a power calculation was conducted and achieved, the intervention (diet, supplement, etc.), dose of the omega-3 FA product, nature of controls, dropout rates among groups, relevant outcomes and results for the inflammatory biomarkers assessed, author s main conclusions, and any related adverse events reported. Once the quality assessment of the individual studies was completed, the SMEs with the reviewers conducted a quality assessment of the overall 3

3 literature pool for each condition using a modification of the international standard for Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. 54 The modified GRADE evaluates (1) the confidence in the estimate of the effect; (2) the magnitude of the effect size; and (3) safety; to be able to assess an overall recommendation for the intervention. All screeners, reviewers, as well as SMEs were fully trained in the methodologies employed and were able to attain a consistent Kappaabove90%. Study Selection Conditions and s Eligible studies included participants 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 clinical characteristics. This SR grouped the studies included in the analysis into the following categories that emerged from the literature: (1) healthy populations; (2) cardiac patients; (3) RA patients; (4) functional conditions, which includes populations with an active inflammatory process that impacts ongoing function including asthma, bone marrow transplant patients, Crohn s disease, metabolic syndrome (MetS), periodontitis, and ulcerative colitis patients; (5) long-term organic conditions, which captures those populations with AD, hypertriglyceridemia and diabetes, obesity, peripheral arterial occlusive disease, and renal disease; (6) emulsion, which mainly included critically ill patients, primarily those with septic shock patients; (7) children; and (8) pregnant women. Proinflammatory vs. Anti-Inflammatory Markers as Outcomes Inflammation is characterized by interplay between pro- and anti-inflammatory cytokines. Major anti-inflammatory cytokines include interleukin (IL)-1 receptor antagonist, IL-4, IL-6, IL-10, IL-11, IL-13, and alpha-interferon (IFN-a). Tumor necrosis factor-alpha (TNF-a), gamma-interferon (IFN-g), IL-12, IL-18, and granulocyte-macrophage colonystimulating factor (GM-CSF) are well characterized as proinflammatory cytokines. 55,56 Also, a number of biomarkers can be either pro- or anti-inflammatory depending on their dose, location of action, and the conditions of the system. For example, endothelial nitric oxide synthase plays a role in inflammation and can regulate the expression of the proinflammatory molecules factor- kb (NF-kB) and cyclooxygenase Alternatively, proinflammatory cytokines can modulate the expression of endothelial nitric oxide synthase. Molecules of this sort can be classified as inflammation modulators. Any and all pro- or anti-inflammatory biomarkers were of interest to this SR and included for analysis. The majority of inflammatory biomarkers included in this SR review were (1) the cytokines including chemokines (CCL2, CCL3, CCL5, CCL11), interferons (IFN-a and IFN-b), ILs (IL-1a, IL-1b, IL-8, IL-10, IL-12, IL-13), lymphokines (IL-2, IL-3, IL-4, IL-5, IL-6, GM-CSF, IFN-g), tumor necrosis factor (TNF-a, TNF-b), transforming growth factor; (2) acute phase reactant proteins; (3) eicosanoids (PGE1, PGE2, LTB4, F2-isoprostanes); (4) adhesion molecules (svcam-1, sicam-1, se-selectin, sp-selectin); and (5) 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, a1-ntichymotrypsin, fibrinogen, PAI-1, orosomucoid AGP). 21 RESULTS Included Studies From the total of 355 articles screened according to eligibility criteria, a total of 112 articles 22 38,40 42,44 46, were included in the present analysis. These included articles were cross-checked against the previous 2012 SR discussed above. 21 Two articles, 39,43 which were included and scored for bias in the Rangel-Huerta et al 21 SR were excluded from this review as we were unable to (1) locate an inflammatory biomarker in the Engler et al 39 study or (2) confirm the Perunicic-Pekovic et al 43 was a RCT design (Fig. 1). Healthy s Twenty-one studies 22 28,38,45,58 69 included in the REAL SR analysis were on healthy populations where omega-3 FAs were introduced and inflammatory biomarkers were assessed (Table I). Most of the subjects were given capsules containing omega-3 FAs, but in some studies, they were provided as part of their diets. 58,61 Two studies tested omega-3 fortified drink while continuing to consume their usual diet. 25,68 Three studies were emulsion studies administering omega-3 via infusion (intravenously) to healthy FIGURE 1. Flow chart. 4

4 participants and are presented in the emulsion section (Table II). 36,126,127 Of the twenty-one studies included, fourteen were scored as high quality () with primarily mixed results or no effect. Six studies scored a neutral quality, of which two showed statistically significant results. 22,67 The remaining study scored low quality ( ), which showed an effect in favor of omega The doses provided, the duration of administration, and age range of participants varied widely across studies (Table I). In addition, most studies had more male participants (about 71% male participants completed) compared to female participants. In fact, eight studies included only male participants. 12,22,24,60,61,67 69 Cardiac Patients Twelve studies 24,30,32,41,70 77 examined cardiac populations of which eight were scored as high quality (), six of which (75%) were statistically significant in favor of omega-3 in reducing inflammation across all biomarkers assessed such as ICAM, sv-cam-1, CRP, TNF, IL-6, or IL-18, 30,41,71,74,76,77 with the remaining having at least one biomarker showing favorable reduction in inflammation. Four studies scored as neutral quality (0), with primarily conflicting results across these studies (Table III). There were no poor quality studies in this category. Similar to healthy subjects, most of the cardiac study subjects were given capsules containing omega-3 FAs, but in some studies, diet was manipulated. 71,76 Dose for omega-3, EPA, and DHA varied across studies. Also like in the healthy subject studies, the majority of cardiac studies reported outcomes based on male subjects. In total, cardiac studies had about 78% of male participants complete the intervention. Three studies did not report gender as a demographic. 32,71,72 Rheumatoid Arthritis (RA) Patients Eleven studies examined RA patients. Four scored high quality () with primarily conflicting results across studies ,83 The other two studies showed nonsignificant effects on C-reactive protein (CRP) when given FO supplements. 79,83 There were five neutral quality (0) studies, 78,84,86 88 with one showing positive effects across biomarkers, 84 three with at least one biomarker with an effect in favor of omega-3, 78,86,87 and one showing no effect across any biomarkers assessed. 88 In addition, there were two low quality ( ) studies showing a reduction across at least one of the several proinflammatory biomarkers assessed 82,84 (Table IV). Again, like most of the cardiac study subjects, RA patients were given capsules containing omega-3 FAs rather than a fatty fish diet. In one study, patients received FO as part of their diets intravenously. 79 This single study is discussed both in this section of the review as well as the emulsion section (Table II). Unlike the healthy participants and cardiac patient omega-3 studies, RA patients were primarily female (about 78% females completed interventions). Two studies did not describe gender. 82,86 Functional and Other Conditions Eleven studies 42,59,89 97 of varying conditions (asthma, Crohn s disease, MetS, periodontitis, ulcerative colitis) were included in our functional category (Table V). Seven studies (64%) scored high quality () and four studies (36%) scored neutral in quality (0) according to ADA criteria. Forty-two percent (42%) of the completed participants in the functional condition group were male. A few studies involved dietary interventions of omega-3 92,94 compared to omega-3 supplement capsules. Two studies administered omega-3 via powder drinks. 91,92 Asthma Two studies 89,90 involving asthmatic patients with statistically significant results scored as high quality (). The first study involved 60 patients with atopic asthma who received lipid extract of New Zealand green-lipped mussel or a matching placebo for a period of 8 weeks and reported a significant decrease in daytime wheeze and the concentration of exhaled H 2 O The second study involved 23 dust mite allergic asthmatics receiving PUFA-enriched fat blend or placebo for 5 weeks and reported significantly lower exhaled NO (eno) levels in the omega 3 PUFAsupplemented group. 90 These two studies measured what we are terming modulators and not necessarily pro or anti-inflammatory biomarkers. Crohn s Disease Two studies 91,92 examined Crohn s disease, one scoring high quality () and the other neutral (0). One of these studies involved 31 patients who received either omega-3 or omega-6 Impact Powder (IP) and showed a significant difference in concentration of IL-1b and in concentration of monocyte chemoattractant protein (MCP-1) in week 9 for omega-3 supplementation. 92 The second study was neutral in quality (0) and had nonstatistically significant results. These patients were provided with omega-3 or omega-6 IP and both groups showed a decrease in CRP; however, there was no statistically significant difference between the two groups. 91 Metabolic Syndrome Four studies 42,93 95 included patients with MetS, two of which scored high quality 94,95 and two scored neutral in quality (0). 42,93 Of those studies that were high quality (), one study 94 examined the effect of four different diets (including omega-3) on inflammation, reporting no statistically significant dietary effects between the pre and postintervention on inflammatory status in fasting MCP-1, IL-6, or IL-1b. The other high-quality score study, 95 a four-arm research design, found statistically significant decreases 5

5 between treatment groups ( p < 0.05) in DBP in the kinako group after 90 days when compared to the results obtained from the FO and kinako groups. Both studies, which scored neutral in quality (0), did not find significant differences between groups. 42,93 Periodontitis Two studies 59,96 on periodontitis populations assessed modulators as their inflammatory biomarkers instead of pro- or anti-inflammatory biomarkers. One study involved participants who were asked to stop brushing their teeth for a period of time and then provided with omega-3 PUFA or olive oil (OO) for 8 days. This study scored high quality according to ADA and showed a tendency toward improvement in plaque and gingival index (GI), but no statistically significant between-group differences were found. 59 The other study involved participants taking borage oil, EPA, or a combination of both or a placebo oil where the use of borage oil showed better results than EPA for plaque index (PI) and GI, 96 and received a neutral ADA score. Ulcerative Colitis The final study in this category included 18 patients with ulcerative colitis who received FO capsules or a vegetable oil placebo for 4 months. This study scored high quality () and reported statistically significant results for reductions in rectal dialysate leukotriene B4 levels, improvements in histologic findings, and weight gain for the FO group. 97 Long-Term Organic Conditions We categorized twenty-seven studies 29,31,33 35,46, as long-term organic conditions (Table VI). These conditions included AD, hypertriglyceridemia and diabetes, overweight and obesity, peripheral arterial occlusive disease, renal disease, and cancer and bone marrow transplant. Nineteen studies (70%) scored high quality () and eight studies (30%) scored neutral in quality (0) according to ADA criteria. Fifty-four percent (54%) of the participants in the long-term organic conditions were male participants. Alzheimer s Disease Three studies 34,35,98 assessed AD patients, all of which scored high quality (). One of these studies showed statistically significant results with DHA-rich omega-3 FAs supplementation, which increased plasma concentrations of DHA (and EPA), and were associated with reduced release of IL-1, IL-6, and granulocyte colony-stimulating factor (G-CSF) from peripheral blood mononuclear cells (PBMCs) for patients receiving omega-3 FAs vs. placebo oil capsules. 35 This is one of the few studies with a relatively large sample size (N = 361) and that measured blood levels of FA. In the other two studies, patients received DHA and EPA or a placebo oil capsules for 6 months, and showed no influence on inflammatory or other measured biomarkers in CSF or plasma. 34,98 Hypertriglyceridemia and Diabetes There were eight studies on hypertriglyceridemia 33, or diabetes. 31,103,104 Most of the studies (75%) 31,99, in this group scored high quality () according to ADA criteria. Of the six studies that scored high quality, most reported reductions in inflammatory markers Of the two studies 33,100 that scored neutral (0) according to ADA criteria, one study examined subjects with hypertriglyceridemia that were otherwise healthy individuals and showed statistically significant results with DHA supplementation in the absence of EPA, which resulted in a reduction in the number of circulating neutrophils and serum CRP and GM-CSF and an increase in the concentration of matrix metalloproteinase (MMP-2) according to ADA criteria. 100 The other study compared two different doses of DHA plus EPA and found that neither dose statistically improved inflammatory markers (IL-1b, IL-6, TNF-a, and high-sensitivity CRP) or the expression of inflammatory cytokine genes in isolated lymphocytes. 33 There were three studies scoring high quality () involving patients with type 2 diabetes with or without hypertension consuming EPA, DHA, or placebo capsules. 31,103,104 Although the serum IL-2 and TNF-a levels were reduced in the treatment group serum, CRP levels varied across these studies. 31,103,104 Overweight and Obesity There were ten studies 29, ,115 on overweight/obesity patients, five scoring high quality () and five scoring neutral quality (0). For the five high-quality studies, four (80%) reported nonsignificant statistical results for TNF, IL-6, and CRP inflammatory biomarkers in patients receiving omega-3 or a placebo 29, (Table VI). For the neutral quality studies, one showed statistically significant results across various inflammatory biomarkers when patients consumed flaxseed flour vs. a placebo. 106 The other neutral quality studies showed nonsignificant results for inflammatory biomarkers when patients consumed omega-3 long-chain PUFA-enriched versions of typical processed foods or placebo. 105,111,112 Peripheral Arterial Occlusive Disease One study addressed patients with peripheral arterial occlusive disease where participants received canola oil or sunflower oil, as placebo, added to their usual diets for 8 weeks. This study was scored as high quality and there were no significant differences for either group for CRP inflammatory biomarkers. 116 Renal Disease Three studies assessed omega-3 among patients with renal disease. 46,113,114 Two of the three studies (67%) scored high quality () according to ADA criteria. 46,113 For all three studies, patients received pills of omega-3 FA or placebo pills. The two high-quality studies concluded that consuming omega-3 FA lowers CRP and IL-1b significantly more than placebo. 46,113 6

6 Cancer and Bone Marrow Transplant Two studies reported on omega-3 interventions for cancer and bone marrow transplant patients. 117,118 Both studies scored as high quality () according ADA criteria. The first study where 16 patients underwent allergic bone marrow transplant received EPA orally and showed significantly lower levels of TNF-a, IFN-g, and IL-10 compared with the non-epa group. 117 The second study investigated the effects of an oral nutritional supplement containing n-3 PUFAs on nutritional status and inflammatory markers where the EPA plus DHA group had greater weight loss and lower IL-6 production after 5 week, though not statistically significant (B = 227.9; p = 0.08). Parenteral Omega-3 (Emulsion Studies) Our SR retrieved twenty-seven studies 36,37,40,44,79, on parenteral nutrition (PN) of omega-3 on participants to reduce inflammation. Nineteen (68%) of these studies were published since Twenty-three studies (85%) scored high quality () and four studies (15%) scored neutral in quality (0) according to ADA criteria. Emulsion studies were done for healthy participants, patients with RA, and critically ill patients. Four studies were categorized as healthy participants and patients with RA. The three healthy participant emulsion studies are referenced in Table II and of the single RA emulsion study is referenced in Tables II and IV. The majority of the studies (n = 23) were on critically ill patients or surgical patients. In fact, 16 of the 27 studies (59%) were surgery-related studies and six studies (22%) included patients in intensive care units (ICUs). Studies of Parenteral Omega-3 in Healthy and RA Patients Four studies, all assessed as high quality (), 36,79,126,127 addressed inflammatory responses and the composition of platelets in healthy subjects and patients with RA given FO intravenously or with usual treatment (soybean oil lipid emulsion). These studies reported that the use of omega-3-enriched emulsions increases the serum concentration of alphatocopherol and the percentage content of EPA and DHA from start of study to end of study. In addition, these studies also reported that omega-3 emulsion resulted in a greater reduction in IL-6 and TNF-a, though most of these finding were not significant. Studies of Parenteral Omega-3 in Critically Ill Patients Twenty-three studies (85%) were included in our critical illness population category receiving PN (Table II). Six studies that met our inclusion criteria involved patients in the ICU and three with systemic inflammatory response syndrome (SIRS). Nineteen (83%) of these studies were scored high quality () according to ADA criteria. Four (17%) scored natural (0) according to ADA criteria. Of those that scored high quality (), one high-quality study demonstrated nonsignificant effects for the inflammatory biomarker assessed (IL-6), showing that supplementation with FO did not affect this inflammatory biomarker. 121 Two other high-quality studies on septic patients who were given enteral nutrition received a standard soybean oil-based emulsion or an emulsion containing FO (OmegavenÒ) for 5 40 or days. Within 2 days of FO infusion, free n-3 FAs increased, and the n-3/n-6 ratio was reversed and the generation of proinflammatory cytokines by mononuclear leukocytes was markedly elevated during n-6 and was suppressed during n-3 lipid application. 40 The authors reported that both these studies establish that infusion of long-chain n-3 PUFA for patients with sepsis can modulate inflammatory mediator production and related inflammatory processes. 147 Of the four studies that scored neutral in quality (0) according to ADA criteria, two studies compared the effects of a medium- and long-chain triglyceride fat infusion on systemic inflammation and hepatic steatosis in ICU patients. The first study 44 found that plasma cytokine, PGE2, LTB4, IL-1b, IL-6, IL-10, and TNF-a concentrations decreased over time in both groups ( p < 0.05). The second study included sepsis and SIRS patients who received PN employing the MCT/LCT fat infusion or the FO-based fat infusion over 7 days in the ICU. Groups receiving the MCT/LCT PN had higher proinflammatory cytokine (TNF-a, IL-1, and IL-6) values and lower anti-inflammatory cytokine (IL-10) values than patients fed with FO, but statistical significance was seen only at the middle and end of the study periods in the septic groups. The results of this study may support that FO could be more effective than MCT/LCT fat emulsion, especially in patients with infectious conditions. 136 Taken together, these studies indicate that inclusion of FO in PN regimens for critically ill and surgical patients modulates the generation of inflammatory eicosanoids and cytokines. It seems there is a shift in the generation of leukotrienes toward the leukotriene-5 series and an ameliorated LTB5: LTB4 ratio, thereby reducing the incidence of systemic inflammatory response. 148 Furthermore, the inclusion of FO in PN regimens, may help to counter the surgery-induced decline in antigen presenting cell activity 139 and T-lymphocyte cytokine production. 135 Importantly, these studies do not reveal any deleterious effects of FO infusion in these patients. 147(p567) Pediatric s This review retrieved four studies 128, on the effect on omega-3 FAs in children. Two scored high quality () and two studies scored neutral in quality (0) according to ADA criteria. One study, neutral in quality, was unable to find significant effect on biomarkers for inflammation for DHA supplementation compared to corn and soy oil (control). 141 The other remaining three studies reported mixed results. A study on lipid omega-3 emulsion was administered ininfantsfordays1to4before open-heart surgery, and 10 days after surgery found that the plasma TNF-a concentration was significantly lower ( p = 0.003) in the treatment compared to the control group. In infants without 7

7 sepsis, plasma TNF-a did not differ according to treatment; however, when sepsis developed, mean plasma TNF-a was significantly lower in treatment compared to control ( p = ). 128 The two remaining studies in this group of studies reported similar findings. In neonates that have undergone surgery and are fed omega-3 FAs through a nasogastric tube, it was found that although decreases of cytokines during hospitalization were similar in both groups, there was a greater decrease of IL-1b in the DHA group than in the OO group after adjusting for confounders ( p = 0.028) 142 (Table VII). Pregnant Women The role of omega-3 on inflammation during pregnancy and fetal development is not well studied. We retrieved three studies on the effect of these EFAs on pregnancy inflammation outcomes. During pregnancy, the immune tolerance between the fetus and the mother is mediated by a number of complex processes at the materno fetal interface. 149 Omega-3 FAs may play a part in a number of these pathways. Two studies scored high quality () and one scored neutral in quality (0) according to ADA criteria. Of the two highquality studies, one study 145 on pregnant women (<19-week gestation) asked that the control group to continue their habitual diet until they gave birth, whereas the women in the treatment group were asked to incorporate 2 portions of farmed salmon (150 g/portion) into their diet per week from study entry (week 20) until birth. Afterward, the infants were clinically evaluated for immune responses at 6monthsofage(n = 48 in the salmon group; n = 38 in the control group). Intakes of salmon during pregnancy, which effectively increased the n-3 long-chain polyunsaturated fatty acids (LCPUFA) content of cord plasma, affected neonatal cytokine production. To assess adaptive memory responses, this study stimulated cord blood mononuclear cells (CBMCs) with both inhalant Dermatophagoides pteronyssinus allergen 1 (Derp1) and food ovalbumin (OVA) and salmon parvalbumin (Sals1) allergens. CBMC IL-10 responses to Derp1 ( p = 0.022), OVA ( p = 0.004), Sals1 ( p =0.002), and mitogen (phytohemagglutinin [PHA]) ( p = 0.044) were all significantly lower in the salmon group than in the control group. However, in infants at age 6 months, the immune differences identified at birth were no longer significant. Therefore, the long-term effect of omega-3 on the neonates was not established (Table VIII). 145 The second high-quality study 144 in this group compared the plasma inflammatory and vascular homeostasis biomarker status of pregnant women consuming a diet that included 2 portions of salmon per week from week 20 of pregnancy until delivery. This study found that the dietary intervention with oily fish in this study population had no effect on plasma cytokine profiles. The last study 146 in this group scored neutral in quality (0) according to ADA criteria. This study found that maternal serum/plasma phospholipid FA proportions were altered by omega-3 supplementation (capsules) compared to soybean oil during pregnancy. (Table VIII). 146 Assessment of Overall Literature Pool by Conditions (GRADE) The quality assessment of the overall literature pool in this REAL was conducted by SMEs and a review team (WBJ, JAI and RK, CC) across the various categories of conditions using the GRADE criteria (Table IX). 54 A GRADE analysis was conducted for each condition category if there were 3 or more studies per category. In general, supplementation with omega-3 FAs had limited effect on inflammatory biomarkers among healthy individuals in most high-quality studies and mixed effects in lower quality studies. Three studies in this group had a large number of participants (>100). In total, there were twentyone studies and 1,191 participants in this group. Therefore, we believe that further research is very unlikely to change our confidence in the estimate of effect. Thus, we are unable to recommend omega-3 to reduce inflammation in healthy participants. In contrast, most high-quality studies in patients with established cardiovascular disease had improved inflammatory status (lower proinflammatory and higher anti-inflammatory) biomarkers compared to controls. Three studies in this group had a large number of participants (>100). In total, there were 12 studies and 1,121 participants in this group. This category was dominated by one study with a large sample size and good score. 77 In this study, levels of IL-18 were decreased by diet ( 10.5% vs. baseline, p = compared with no diet) and by omega-3 PUFA supplementation ( 9.9% vs. baseline, p = compared with placebo). Levels of IL-18 were also significantly reduced in each treatment group compared with the control group ( p = for both), with a potentially additive effect of the 2 intervention principles ( p < 0.001). CRP, IL-6, and TNF-a were reduced vs. baseline in all study groups, but with no between-group differences. 77 Our review would recommend omega-3 for the reduction of inflammation in cardiac patients. There were eleven studies on the effect of omega-3 on inflammation in patients with RA. These studies had mixed results and have had a small number of participants (total n = 453). Our study is unable to make a recommendation in favor of using Omega-3s in this group. Currently, any estimate of effect is very uncertain. Long-term organic conditions (AD, hypertriglyceridemia/ diabetes, overweight/obesity, and renal disease) generally showed favorable inflammatory biomarker modulator effects, but almost all studies were small in size. This group of studies was dominated by one Alzheimer s study by Freund-Levi et al, 34 which was reasonably large and high quality. It reported no effects on any biomarker of inflammation in serum or CSF. Therefore, use of omega-3 for the reduction of inflammation in AD patients, hypertriglyceridemia and type 2 diabetes patients, and overweight and obese patients requires further research and is likely to have an important impact on our confidence in 8

8 the estimate of effect and may change the current estimate. We can provide a weak recommendation for omega-3 for these patients. More research is needed for this study group. We are unable to make a recommendation for omega-3 for the reduction of inflammation in renal disease patients. The lack of studies and mixed results of the few existing studies does not allow for conclusive recommendation for omega-3emulsions in healthy populations and RA patients. In contrast, the quality and quantity of research on inflammation from omega-3 emulsions for surgical and critically ill patients is significant. Omega-3 emulsions for the critically ill reported very few and mostly minor adverse events. In fact, intravenous lipid emulsions (mainly n-6 PUFAs including soybean oil) have been established for many years as an integral part of parenteral nutrition due to their high energy density and low osmolarity. 148 The mediators reported from n-6 PUFA infusions (mainly arachidonic acid [AA]) seem to favor an inflammatory response, whereas mediators stemming from infusions of EPA and DHA seem to exert less proinflammatory actions. 148 The high quality and positive results of these studies suggest that omega-3 is an appropriate form of PN for critically ill patients or those undergoing surgery. Our review retrieved three studies measuring inflammation as a biomarker in pregnant women and four studies on children. These studies had small numbers of participants and the results were mixed. Therefore, further research is very likely to have an important impact on our confidence in the estimate of effect for these populations, and we are unable to recommend omega-3 in reducing inflammation among pregnant women or for the use of children. All of the healthy or condition categories scored either a B (moderate) or a C (low) on confidence in the estimate of the effect size. This score is interpreted to mean that in the current body of literature, conclusions could easily be altered by a single large, high-quality study in almost any patient category. Most studies which addressed specific illness conditions had small sample sizes and we found only a few studies in each condition. Therefore, at the current time, any new study with sufficient power could change our estimation of the direction of the outcome for the conditions included in Table IX. As a result of this low level of confidence, all conclusions should be considered tentative. The authors are not able to comment on the magnitude of the effect since none of the studies reported effect sizes for the relevant outcomes captured by GRADE. The safety grade was marked as 0 for most studies as the majority of studies did not report on any adverse effects. In fact, most studies make no mention of adverse event assessment at all. Adverse Events Almost half of the RCTs included in this study (48%) made no report of anything related to adverse events, leaving the reader unsure if any were measured or occurred. Without this information, it is difficult to make clear recommendations for use of FAs. Some studies reported that there were no adverse events associated with interventions (26%). Additionally, about 26% reported minor adverse events with omega-3, including diarrhea, rash, nausea, constipation, gastrointestinal complaints, bloating, prolonged bleeding, headache, muscle ache, depression, fishy taste, vomiting, rictus, metallic taste, and abdominal cramps. No major adverse event was reported. Adverse events considered to be moderate or severe did not appear to be related to the study medication, i.e., two wound infections, urologic infections, respiratory tract infection, intestinal infection, 37 myocardial infarction, 69 cellulitis, transient ischemic attack, and sustained fractures after falling. 83 A small elevation of alanine aminotransferase was detected in in a study among the omega-3 group (n = 23). 107 DISCUSSION This article reports the evidence from and the quality of research of omega-3 FA supplementation on biomarkers of inflammation in healthy individuals and inflammatory conditions in human RCTs. The quality of the individual studies is impressive overall. In addition, the overall evidence is significant for several conditions. Supplementation with long-chain n-3 PUFA produced no demonstrable effects on inflammation for healthy participants, as measured by inflammatory biomarkers. At first glance, the outcome of these clinical studies might suggest thatomega-3fasupplementation has no benefits in healthy groups. However, these studies reported induced elevation of plasma n-3 and positive affects other health-related outcomes that were not directly related to inflammation (i.e., inflammatory biomarkers). In this respect, some consideration should be given to the fact that healthy individuals may not exhibit inflammation as compared to diseased participants. In general, the 1980s were a period of pioneering expansion in research about PUFAs in general and omega-3 FAs in particular. 4 However, about 80% of the studies in this review were published in the last decade. The popular interest around omega-3 in nutrition to reduce inflammation may have caused this increase. To date, there are no official recommendations for intake of n-3 PUFAs, particularly EPA and DHA. Although there are many international authorities and expert groups that have issued dietary recommendations for them, doses vary greatly. 150 In 1994, the United Kingdom Committee on Medical Aspects of Food Policy guidelines recommended that the average n-3 PUFA intake should be increased from 0.1 to 0.2 g/d. 151 In 2002 and 2005, the Institute of Medicine (IOM) report concluded there was insufficient evidence to set a dietary reference intake for EPA and DHA. 152,153 Although the omega-3 intakes were not specified by the IOM, they have been estimated (n-3 PUFA [ALA]) at 0.5 g/d (infants 0 12 months), 0.7 to 0.9 g/d (children 1 3 years and 4 8 years, respectively), 1.2 g/d (male children 9 13 years), 1.6 g/d (male years), 1.0 g/d (female children 9 13 years), 1.1 g/d (female years), 1.4 g/d (pregnant women), and 1.3 g/d (lactating woman). 152,153 The IOM recommendation for EPA plus DHA is about 140 mg/d. 154 The 9

9 International Society for the Study of Fatty Acids and Lipids report on specific recommendations for the intake of polyunsaturates (PUFA): (1) an adequate linoleic acid (LA) intake as 2 energy %; (2) a healthy intake of ALA as 0.7 energy %; and (3) for cardiovascular health, a minimum intake of EPA and DHA combined as 500 mg/d. 155 In 2010, the Report of the Dietary Guidelines Advisory Committee on the Dietary Guidelines for Americans recommends 8 oz of seafood per week to contribute an average of 250 mg/d of long-chain omega-3 FAs, for all Americans. 156 Others recommend a dietary strategy for achieving the 500-mg/d recommendation is to consume 2 fish meals per week (preferably fatty fish). Foods enriched with EPA and DHA or FO supplements are a suitable alternate to achieve recommended intakes and may be necessary to achieve intakes of 1 g/d. 154 In addition, when discussing dosage the concurrent dietary intakes of LA, a-lna, AA, EPA, and DHA should be considered in predicting final tissue proportions of n-3 FAs in long-chain FAs. 157 Thus, a healthy dietary allowance for n-3 long-chain FAs intake must take into account concurrent intakes of LA, a-lna, and AA. In the United States per capita, a healthy dietary allowance of 3.5 g EPA DHA/d could be reduced to one-tenth of that amount if the intake of n-6 fatty acids, in particular LA, can be lowered to <2% of energy. 158(p1489S) In short, the dosage and dietary recommendation for omega-3 varies greatly. Similarly, the dosage for omega-3 lipid infusions also varies. Currently, there are several lipid emulsions that include FO as a component (i.e., Omegaven (Fresenius Kabi, Bad Homberg, Germany, 100 g lipid/l); Lipoplus (also known as Lipidem; B. Braun, Melsungen, Germany, 200 g lipid/l); and SMOFLipid (Fresenius Kabi, Bad Homberg, Germany,(200 g lipid/l). 147 To date, there are no guidelines or recommendation for omega-3 emulsions. Omega- 3 use among military personnel is not well documented. A recent survey reported categories of CAM used by U.S. active duty military personal and high-dose megavitamins (8.4%) are among the top ten most popular CAM therapies. 159 A 2011 study evaluated deficiencies of highly unsaturated omega-3, in particular DHA, to the increased risk of suicide death among a large random sample of active duty U.S. military. This study found the U.S. military population had a very low and narrow range of omega-3 status. They reported that, among men, risk of suicide death was 62% greater with low serum DHA status (adjusted odds ratio [OR] =1.62, 95% CI , p < 0.01, comparing DHA below 1.75% [n = 1,389] to above [n = 141]). 160 Another study reported that improving omega-3 status through an increase in omega-3 intake may improve neurocognitive performance and confer an element of resilience to poor sleep. 161 There have been no published studies on the impact of omega-3 in inflammatory biomarkers among U.S. military personnel. Emulsion studies of omega-3 may be especially important to military health care. Warriors injured in combat may benefit from omega-3 PN. Currently, most lipids used in PN are soybean oil, in which n-6 linoleic acid comprises about 50% of the FAs. There has been some concern about the potential harm of n-6 emulsions. These concerns are mainly because of the proinflammatory and immune-suppressive aspect of n-6 PUFA. In our study, omega-3 lipid emulsions for parenteral applications were found to reduce inflammation, increase immune response, improve clinical outcomes, and reduce cardiovascular risk. 18,147 Lastly, some emulsion studies have found a lower rate of infection and shorter lengths of stay in the ICU. A decrease in length of hospital stay would have important cost and policy implications for the use of FO infusion in ICU and perioperative patients. 147 Limitations and Challenges It is clear that the inflammatory system is extremely complicated and studies continue to reveal this complexity, often with contradictory findings. For example, the proinflammatory molecule TNF-a has shown therapeutic promise against RA while failing against SIRS, leading researchers to devise complex mathematical models to account for this behavior. 162 This complexity is not only an issue for individual biomarkers but also evident in the complex dynamics of pro- and antiinflammatory interactions in living systems in general. As a consequence, it was often a challenge to classify whether some of the inflammatory biomarkers were proinflammatory, anti-inflammatory, or modulators of both processes. For the most part, this study reviewed RCTs where the effects of omega-3 FAs were consumed in isolation via supplementation. Omega-3 FAs are not usually consumed in isolation. Indeed, clinical dietary guidelines advocate the consummation of omega-3 FAs in foods, the properties of which will be influenced by the food matrix and its overall nutrient composition. 163(p166) Another limitation of this study is that we only examined the inflammatory biomarker in clinical studies involving humans, whether this was a primary outcome of the study or not. Therefore, for those studies where the inflammatory biomarkers captured were part of a secondary analysis of the study, they did not represent the complete picture of the original research. None of the studies reported on an effect size for the inflammatory biomarkers of interest to our review and therefore, we were not able to GRADE effect size for the overall literature pool or conduct a meta-analysis. It is difficult to assess whether omega-3 FA changes the underlying state of inflammation, especially given the lack of evidence in this review. Overall, the RCTs in this review had small patient groups. Larger cohort studies or longitudinal research were not assessed in our review. Therefore, we are unable to address the long-term effects of omega-3 FA response given the lack of larger population studies. Finally, in part, because this is a rapidly changing field but mostly because inflammatory/anti-inflammatory regulation is both patently complex and just as obviously not well understood, the existing literature leaves much to be desired. Most of the studies in healthy populations asked the question 10