R

Oc

AD

a

ARRAA

KOICCH

C

CcC

UT

h1

Prostaglandins & other Lipid Mediators 116–117 (2015) 104–111

Contents lists available at ScienceDirect

Prostaglandins and Other Lipid Mediators

eview

mega-3 fatty acids and inflammation: A perspective on thehallenges of evaluating efficacy in clinical research

nn C. Skulas-Ray ∗

epartment of Nutritional Sciences, The Pennsylvania State University, University Park, PA, United States

r t i c l e i n f o

rticle history:eceived 24 July 2014eceived in revised form 3 February 2015ccepted 9 February 2015vailable online 16 February 2015

eywords:

a b s t r a c t

Chronic inflammation is a common underpinning of many diseases. There is a strong pre-clinical evi-dence base demonstrating the efficacy of omega-3 fatty acids for ameliorating inflammation and therebyreducing disease burden. Clinically, C-reactive protein (CRP) serves as both a reliable marker for monitor-ing inflammation and a modifiable endpoint for studies of anti-inflammatory pharmaceuticals. However,clinical omega-3 fatty acid supplementation trials have not replicated pre-clinical findings in terms ofconsistent CRP reductions. Methodological differences present numerous challenges in translating pre-

mega-3 fatty acidsnflammation-reactive proteinlinical researchuman endotoxemia model

clinical evidence to clinical results. It is crucial that future clinical nutrition research clearly distinguishbetween the reversal of established inflammation and preventing the development of inflammation.Future clinical studies evaluating the ability of omega-3 fatty acids to attenuate an excessive inflamma-tory response, may be advanced by employing new statistical approaches and utilizing models of inducedinflammation, such as low-dose human endotoxemia.

© 2015 Elsevier Inc. All rights reserved.

ontents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1042. C-reactive protein (CRP): a clinical marker of inflammation and outcome measure in research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1053. Evaluating existing clinical evidence for the efficacy of marine omega-3 fatty acids to reduce plasma CRP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1054. Challenges for translation of pre-clinical efficacy to clinical research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

4.1. Low signal-to-noise ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084.2. Reversing established inflammation instead of preventing its development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084.3. Differences in animal model physiology and duration of intervention. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

5. Essential features of studies designed to evaluate the efficacy of omega-3 fatty acids for preventing the development of human inflammation 1086. Clinical models of induced inflammation to study the effects of omega-3 supplementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1097. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Disclosures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110Acknowledgment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Abbreviations: EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; CRP,-reactive protein; IL-6, interleukin-6; TNF-�, tumor necrosis factor-alpha; CVD,ardiovascular disease; AHA, American Heart Association; hs-CRP, high sensitivity-reactive protein; DPA, docosapentaenoic acid; LPS, lipopolysaccharide.∗ Correspondence to: Department of Nutritional Sciences, The Pennsylvania Stateniversity, 110 Chandlee Lab, University Park, PA 16802, United States.el.: +1 814 863 8622; fax: +1 814 863 6103.

E-mail address: aus164@psu.edu

ttp://dx.doi.org/10.1016/j.prostaglandins.2015.02.001098-8823/© 2015 Elsevier Inc. All rights reserved.

1. Introduction

Inflammation is an essential biological process initiated by theimmune system in response to injury, irritation, or infection. How-ever, when inflammation becomes prolonged or chronic it is akey component in the etiology of several diseases, including car-diovascular disease (CVD), diabetes, rheumatoid arthritis, cancer,

and neurodegenerative diseases such as Alzheimer’s disease. Cur-rent standard-of-care medical therapies for such diseases are verycostly and/or carry significant side effects. As such, there is a need

er Lipid Mediators 116–117 (2015) 104–111 105

tmt

o(oirphoccsdbmmftpeid

h[fsmi

ciicwfissii

2i

umrisc(parAglr[tl

Box 1: Common causes of acute CRP elevation that mayoccur in clinical research participants*

• Viral illness (Ex: cold or flu exposure)• Bacterial infection (Ex: urinary tract infection)• Dental cleanings or procedures• Unusual strenuous physical activity (Ex: marathon race,

downhill running, or implementation of a new weight liftingroutine)

• Exacerbation or “flare” of inflammatory disease unrelated tostudy

• Change in medications (Ex: withdrawal of anti-inflammatorymedication)

• Physical trauma (Ex: cuts, scrapes, contusions, sprains,strains, and broken bones)

• Significant changes in psychosocial stressors (Ex: becominga caretaker, losing a job, or death of a loved one)

• Prolonged sleep deprivation• Significant weight gain or dramatic dietary changes

*It is possible to minimize these inflammatory insults bycounseling participants to avoid certain activities prior to mea-surements and rescheduling when such events are reported.

A.C. Skulas-Ray / Prostaglandins & oth

o identify dietary patterns, lifestyle factors, and alternative phar-acological strategies capable of ameliorating inflammation and

hereby reducing the burden of disease.The pre-clinical evidence base strongly supports the efficacy

f the marine-derived omega-3 fatty acids, eicosapentaenoic acidEPA) and docosahexaenoic acid (DHA), for preventing the devel-pment of inflammation. In observational studies for example,ndividuals who report greater dietary EPA and DHA intake expe-ience less inflammation as they age [1–4]. This relationship is alsoresent when biomarkers of intake are used [5–7]. For instance, inealthy adults with relatively low omega-3 intake, higher quartilesf red blood cell (RBC) DHA were inversely associated with plasmaoncentrations of the inflammatory cytokine TNF-� [8]. This asso-iation between omega-3 fatty acids and inflammation is alsoupported by investigations utilizing in vitro, ex vivo, and animalisease models, which have begun to illuminate the mechanismsy which omega-3 fatty acids may act to prevent excessive inflam-ation [9–14]. In particular, research utilizing the fat-1 transgenicouse [15–17]—which is able to endogenously convert omega-6

atty acids to omega-3 fatty acids—has provided a system in whichhe effects of elevated tissue omega-3 fatty acids can be com-letely isolated from other confounding factors, such as diet. Whenxposed to inflammatory stimuli, these mice experience attenuatednflammatory responses and are more resilient than their omega-3eficient littermates [18].

However, randomized controlled clinical supplementation trialsave yielded mixed results for accepted markers of inflammation19,20]. This creates much difficulty in evaluating whether theseatty acids are efficacious as anti-inflammatory therapies, and ifo, determining critical details in terms of their delivery (e.g. opti-al dose, duration, target condition, and relative efficacy of the

ndividual fatty acid species).This article will discuss many of the unique challenges regarding

linical research evaluating the efficacy of omega-3 fatty for reduc-ng biomarkers of inflammation and suggest potential means ofmprovement for future investigations. Following a discussion oflinical biomarkers, the current clinical research evidence baseill be presented. Further discussion of the methodological dif-culties of evaluating therapeutic interventions for inflammatorytates will be given, specifically with regard to achieving adequateignal-to-noise in clinical studies. Finally, models of human inducednflammation will be introduced as a potential strategy for progressn future clinical research.

. C-reactive protein (CRP): a clinical marker ofnflammation and outcome measure in research

The acute phase product C-reactive protein (CRP) is typicallysed to evaluate systemic levels of inflammation in the manage-ent of chronic inflammatory diseases and assessment of CVD

isk. CRP is produced hepatically in response to elevated plasmanflammatory cytokines (TNF-� and IL-6), which spill over into theystemic circulation from sites of inflammation. The plasma con-entration of these inflammatory cytokines can fluctuate rapidlyhour-to-hour), but CRP serves as a more stable reflection of theirroduction during periods of inflammation. CRP concentrationsre markedly elevated during periods of active pathology, such asheumatoid arthritis and inflammatory bowel disease flares. Themerican Heart Association (AHA) has established evidence-baseduidelines defining cardiovascular risk categories based on CRPevels, where values <1 mg/L represent low risk, values 1–3 mg/L

epresent intermediate risk, and values >3 mg/L represent high risk21]. As values greater than 10 mg/L may represent acute infec-ion, it is recommended that a re-evaluation be performed after ateast 2 weeks to determine whether such elevations are acute or

However, bacterial and viral exposures can be asymptomatic,and many other events go unreported.

chronic. Persistently high CRP levels—such as those that occur withrheumatoid arthritis—signify radically increased risk of CVD eventsand mortality. Clinicians can order either a standard CRP to mon-itor active inflammation or a high sensitivity (also referred to as“cardio”) CRP (hs-CRP) test that can quantify CRP values <10 mg/L(lower limit = 0.2 mg/L). Diagnostic labs also do not typically mea-sure IL-6 and TNF-�.

Although CRP concentrations are more stable than plasmacytokines, they can exhibit large day-to-day variations due to acuteinflammation that is unrelated to either study procedures or thetarget inflammatory conditions. This means that significant treat-ment effects are contingent on a large magnitude of response to anintervention. Additionally, CRP elevations may be present even inthe absence of symptoms because CRP remains elevated for severaldays following the resolution of acute inflammation. Exacerbationsof acute inflammation can be caused by an assortment of factors andmay not always be perceived and/or reported by research partici-pants (Box 1).

In addition to reflecting active inflammation, CRP values can bereduced by anti-inflammatory therapies. Immunosuppressive ther-apies, including anti-TNF biologics, decrease CRP concentrationsand CVD risk in people with rheumatoid arthritis [22–24]. However,these therapies carry a significant risk of side effects, which limitstheir use. In several large trials, including the JUPITER study, statinadministration also reduced CRP values [25]. Such trials demon-strate that CRP can function as a modifiable endpoint for clinicalresearch on anti-inflammatory interventions.

3. Evaluating existing clinical evidence for the efficacy ofmarine omega-3 fatty acids to reduce plasma CRP

Although many clinical trials have assessed whether supple-mental EPA and/or DHA can reduce plasma CRP, it is not typicallythe primary endpoint. A systematic review by Rangel-Huerta et al.identified and evaluated the randomized placebo-controlled clin-ical trials investigating the potential anti-inflammatory effects of

omega-3 fatty acid supplementation [19]. For the present discus-sion, additional recently published studies were added to thoseoriginally reviewed by Rangel-Huerta et al. (Table 1). Only random-ized controlled trials in which CRP was measured are included.

106 A.C. Skulas-Ray / Prostaglandins & other Lipid Mediators 116–117 (2015) 104–111

Table 1Randomized controlled trials evaluating the effects of EPA+DHA on C-reactive protein (CRP).

Reference Study design Population EPA+DHA (g/d), n Placebo oil, n Duration (months) CRP effect

HealthysubjectsCiubotaru et al. [26] Parallel Postmenopausal women on

HRT2.2, n = 10 or 1.1,n = 10

Safflower, n = 10 1.25 ↓ (low dose only)

Madsen et al. [27] Crossover Healthy adults 6.6, n = 60 or 2.0,n = 60

Olive, n = 60 3 –

Vega-Lopez et al. [28] Parallel Healthy adults 1.5, n = 20 NR, n = 20 3 –Geelen et al. [29] Parallel Healthy adults 1.3, n = NSa Sunflower, n = NSa 3 –Fujioka et al. [30] Parallel Healthy adults 0.86, n = 68 Olive, n = 73 3 –Sanders et al. [31] Parallel Healthy adults 1.5 (DHA only),

n = 40Linoleic acid, n = 39 1 –

Yusof et al. [32] Parallel Healthy men 2.1, n = 9 Coconut, n = 11 2 –Nieman et al. [33] Parallel Endurance athletes 2.4, n = 11 Soybean, n = 12 1.5 –Bloomer et al. [34] Crossover Exercise-trained men 4.4, n = 14 Soybean & canola,

n = 141.5 ↓

Dangardt et al. [35] Crossover Obese adolescents 1.2, n = 25 Soybean, n = 25 3 –Ottestad et al. [36] Parallel Healthy adults 1.6, n = 17 High-oleic sunflower,

n = 191.75 –

Root et al. [37] Parallel Young adults 0.58, n = 26 Safflower, n = 25 1 –

CVDriskfactorsChan et al. [38] Parallel Obese men 3.4, n = 12 Corn, n = 12 1.5 –Mori et al. [43] Parallel Type 2 diabetics

w/hypertension3.8 EPA, n = 17 or3.7 DHA, n = 18

Olive, n = 16 1.5 –

Jellema et al. [39] Crossover Obese men 1.1, n = 11 Sunflower, n = 11 1.5 –Browning et al. [40] Crossover Overweight women 4.2, n = 29 Linoleic & oleic acid,

n = 293 ↓

Jones et al. [45] Crossover Hyperlipidemic adults 5.4 (w/sterols),n = 21

Olive & sunflower,n = 21

1 –

Derosa et al. [47] Parallel Dyslipidemic adults 2.6, n = 164 Sucrose, mannitol, andmineral salts, n = 162

6 ↓

Micallef et al. [46] Parallel Hyperlipidemic adults 1.4, n = 15 Sunola, n = 15 0.75 ↓Kelley et al. [49] Parallel Hypertriglyceridemic men 3.0 (DHA only),

n = 17Olive, n = 17 3 ↓*

Skulas-Ray et al. [50] Crossover Hypertriglyceridemic adults 3.4, n = 26 or 0.85,n = 26

Corn, n = 26 2 –

Tajalizadekhoob et al.[52]

Parallel Depressed elderly adults 0.30, n = 32 Coconut, n = 29 6 –

Itariu et al. [41] Parallel Severely obese adults 3.4, n = 27 Butter, n = 28 2 –Gammelmark et al. [42] Parallel Overweight adults 1.1, n = 25 Olive, n = 25 1.5 –Malekshahi Moghadam

et al. [44]Parallel Type 2 diabetics 2.4, n = 42 Sunflower, n = 42 2 –

Derosa et al. [48] Parallel Dyslipidemic adults 2.6, n = 78 Sucrose, mannitol, andmineral salts, n = 79

6 ↓

Koh et al. [51] Parallel Hypertriglyceridemic adults 1.7, n = 50 NR, n = 49 2 –

HeartdiseasepatientsGrundt et al. [53] Parallel Acute MI patients 1.7, n = 125 Corn, n = 130 12 –O’Keefe et al. [55] Crossover CHF patients 0.81, n = 18 Corn & olive, n = 18 4 –Madsen et al. [54] Parallel Post-MI patients 4.3, n = NSb Olive, n = NSb 3 –Zhao et al. [56] Parallel CHF patients 0.60, n = 38 NR, n = 37 3 –Eschen et al. [57] Parallel CHF patients 0.90, n = 69 Olive, n = 69 6 –Mackay et al. [58] Crossover PAD patients 0.85, n = 150 Palm & sunflower,

n = 1501.5 –

RenaldiseasepatientsMadsen et al. [54] Parallel CRF patients 2.4, n = 22 Olive, n = 24 3 –Himmelfarb et al. [60] Parallel HD patients 0.80 (DHA only),

n = 27Sunflower, n = 30 2 –

Bowden et al. [19] Parallel ESRD patients 1.6, n = 18 Corn, n = 15 6 ↓**

Kooshki et al. [61] Parallel HD patients 2.1, n = 17 MCT, n = 17 2.5 –Deike et al. [63] Parallel CKD patients 2.4, n = 17 Safflower, n = 14 2 –

OtherchronicdiseaseorillnessBarbosa et al. [64] Crossover Ulcerative colitis patients 4.5, n = 9 Soybean, n = 9 2 –Sundrarjun et al. [65] Parallel Rheumatoid arthritis patients 3.4, n = 13 NR, n = 13 3 –Freund-Levi et al. [66] Parallel Alzheimer’s patients 2.3, n = 18 Corn, n =17 6 –Mohammadi et al. [67] Parallel Women w/PCOS 1.2, n = 30 Paraffin, n = 31 2 –Finocchiaro et al. [68] Parallel Lung cancer patients

undergoing chemotherapy0.85, n = 13 Olive, n = 14 2 ↓***

HRT, hormone replacement therapy; NR, not reported; NS, not specified; MI, myocardial infarction; CHF, coronary heart failure; PAD, peripheral arterial disease; CRF, chronicrenal failure; HD, hemodialysis; ESRD, end-stage renal disease; CKD, chronic kidney disease; PCOS, polycystic ovarian syndrome; MCT, medium-chain triglycerides.

a Sample size per group not specified, total n = 81.b Sample size per group not specified, total n = 41.* Comparison was made to pre-treatment values and not to placebo group.

** Used ratios for endpoint assessment.*** Placebo group increased and omega-3 group did not, so effect represents attenuation of increase.

er Lipid Mediators 116–117 (2015) 104–111 107

ppwdl[if[cepead[wt4bc(3dliwddDEcolo[at[[nt(batodiopabiwpsdcfrt

mt

Box 2: Best practices for conducting clinical trials andevaluating publications

• Handling skew or outliers and attention to sample size◦ Normality and the presence of outliers must be assessed. If

outliers are present and/or data is not normally distributed,an appropriate method of addressing these issues must beused. This should be clearly explained in the manuscript.

◦ In studies with very small sample sizes, even a single out-lier can impact results. In such instances, authors shouldbe encouraged to show individual data points.

◦ The reader should note the magnitude of error terms. Largeerror terms may be due to the presence of skew and/oroutliers, which may indicate that assumptions of normaldistribution required for the statistical model have beenviolated. In such instances, when large error terms arepresent and statistical tests that assume normal distribu-tion (e.g. t-tests) are used, skepticism is warranted.

◦ Log transformations do not always correct skewed datadue to outliers. Therefore, it may be preferable to set athreshold for acute elevations that cannot feasibly be dueto the intervention (this will depend on the characteris-tics of the study population). For CRP, 3 mg/L may be anappropriate cut-off for healthy participants with habituallylow levels of inflammation. An upper limit of 10 mg/L is aconservative threshold used in other studies [29,54].

• Testing treatment effects◦ Treatment effects should be compared to a valid placebo

control. Comparisons to baseline/pre-treatment values donot account for unanticipated changes in participants overthe course of the study that are unrelated to the treatment.The placebo should be matched for any added antioxidantsand not have adverse effects. Olive, soybean, and cornoil-based placebos contain fatty acids that are abundantlyconsumed in typical Western diets, and do not appear tohave independent effects at the modest doses (∼4 g/d) usedin most supplementation studies.

◦ In some instances, the end-of-treatment values in theomega-3 treatment group may be similar or higher thanbaseline, while those in the placebo group worsen overthe course of the study. It is crucial that the discussion andabstract clarify this effect as preventing or attenuating anincrease in inflammation, as opposed to interpreting thetreatment effect as reducing inflammation.

• Characterization of the intervention◦ A characterization of the intervention and placebo should

be provided, with as much detail about the fatty acid com-position and any additional components as possible.

◦ Although omega-3 capsules are typically compared onthe basis of their EPA and DHA content, other compo-nents of the intervention (e.g. antioxidants and other fattyacids such as docosapentaenoic acid [DPA]) may influenceresponses and should be included in the characterization.

◦ Future research should evaluate the presence of trans fattyacids and oxidative stress products that may be created

A.C. Skulas-Ray / Prostaglandins & oth

The identified trials span a range of doses, duration, and studyopulations (Table 1). These studies were conducted in varyingopulations, ranging from healthy individuals [26–37]; individualsith CVD risk factors, including overweight/obese [38–42], type 2iabetes [43,44], hypertension [43], hyperlipidemia [45,46] or dys-

ipidemia [47,48], hypertriglyceridemia [49–51], and advanced age52]; individuals with established forms of heart disease, includ-ng acute [53] and post-myocardial infarction [54], coronary heartailure patients [55–57], and peripheral arterial disease patients58]; individuals with established forms of renal disease, includinghronic renal failure patients [59], hemodialysis patients [60,61],nd-stage renal disease patients [62], and chronic kidney diseaseatients [63]; and individuals with other established chronic dis-ases or illnesses, including ulcerative colitis [64], rheumatoidrthritis [65], Alzheimer’s disease [66], polycystic ovarian syn-rome [67], and lung cancer patients undergoing chemotherapy68]. The duration of these trials ranged from 3 weeks to 1 year,ith the majority supplementing for relatively short periods of

ime (≤12 weeks). Only 8 of the 43 studies were conducted for months or more [47,48,52,53,55,57,62,66], with only one studyeing conducted for a full year [53]. The majority of trials wereonducted for 2 to 3 months, with many others lasting 4–6 weeksTable 1). The dose of EPA+DHA used in these trials ranged from00 mg [52] to 6.6 g [27] per day. However, only 2 studies usedoses greater than 4.5 g/d [27,45]. Ten of the 43 studies used doses

ess than 1 g/d [30,37,50,52,55–58,60,68]; a 1–2 g/d dose was usedn 13 cases [26,28,29,31,35,36,39,42,46,51,53,62,67]; a 2–3 g/d dose

as used in 11 cases [26,27,32,33,44,47,48,59,61,63,66]; a ≥3–4 g/dose was used in six cases [38,41,43,49,50,65]; and six studies usedoses greater than 4 g/d [27,34,40,45,54,64]. In some instances, soleHA and/or EPA supplements were provided [31,43,46,60] or thePA+DHA was combined with sterols [45,46]. A variety of placeboontrol oils were also used, including: safflower oil [26,37,63],live oil [27,30,42,43,49,54,57,59,68], sunflower oil [29,39,44,60],inoleic acid [31], coconut oil [32,52], soybean oil [33,35,64], high-leic sunflower [36], corn oil [38,50,53,62,66]; butter [41], paraffin67], and combinations of soybean and canola oil [34]; linoleicnd oleic acid [40]; olive and sunflower oil [45]; sucrose, manni-ol, and mineral salts [47,48]; Sunola oil [46]; corn and olive oil55]; palm and sunflower oil [58]; and medium-chain triglycerides61]. It is interesting to note that studies that enrolled a largerumber of participants did not demonstrate an ability of EPA+DHAo reduce plasma CRP, with the exception of two relatively largen ≥ 78/group) 6-month long studies of 2.6 g/d EPA+DHA conductedy Derosa et al. in people with dyslipidemia [47,48]. Smaller studiesre more susceptible to violations of statistical model assump-ions, which in many cases may erroneously increase the chancef observing a significant treatment effect (Box 2). Most studieso not report biomarkers of omega-3 status prior to and follow-

ng supplementation. Utilizing biomarkers of omega-3 status (e.g.mega-3 fatty acids in plasma, RBC membranes/omega-3 index,hospholipids, cholesteryl esters, etc.) may also improve the designnd analysis of omega-3 clinical studies. Measurement of omega-3iomarkers prior to supplementation could better identify deficient

ndividuals who may be more likely to respond to supplementation;hereas, measurement of post-supplementation omega-3 statusrovides an objective measure of compliance with taking studyupplements. As a whole, there has been a wide variety in studyesigns but few that detected significant treatment effects. In manyases, those studies that did report significant reductions in CRPollowing EPA+DHA supplementation have noteworthy features inegards to study design and/or statistical analysis that should be

aken into consideration when interpreting their findings (Table 1).

In addition to systematic reviews, meta-analyses are anothereans of evaluating clinical evidence. By pooling results from mul-

iple clinical studies, they increase sample size and consequently

during refinement and manufacturing processes involvingheat.

heighten sensitivity for detecting significant treatment effects.Unfortunately, such analyses are also forced to make assumptionsabout the data collection and appropriateness of statistical meth-ods used in the original studies—generally without the benefit ofhaving access to the raw data. Thus, even in these large datasets, thestatistical problems introduced by skew and outliers in the original,smaller datasets may remain. Additionally, heterogeneous studies

are typically grouped for pooled analyses, complicating inferencesabout critical aspects related to efficacy (e.g. dose, duration, targetpopulation, etc.). A recent meta-analysis by Li et al. attempted toovercome the skew of original datasets using statistical procedures.

1 er Lipi

TsanlFtia−nft6Kr0i3ptDosifhta

4c

ipc

4

ohtdascpb

4d

tTiwHsacm

08 A.C. Skulas-Ray / Prostaglandins & oth

hey reported a significant 0.2 mg/L reduction in hs-CRP across 44tudies of EPA+DHA supplementation in people with “chronic non-utoimmune disease” [20]. However, their definition of “chronicon-autoimmune disease” spanned healthy individuals with dys-

ipidemia [50] to patients diagnosed with ulcerative colitis [64].urthermore, the clinical implications of a 0.2 mg/L CRP reduc-ion are unknown and would be difficult to detect in a singlenterventional study. Their analysis of populations with “chronicutoimmune disease”, which resulted in a pooled effect size of0.54 mg/L (p < 0.0001), is based on three studies [69–71] and didot include a study of patients with ulcerative colitis [64] that

avored the placebo. The trial by Wright et al. reported a neu-ral effect of 24 weeks of 3.4 g/d EPA+DHA supplementation in0 people with systemic lupus erythematosus [71]. The report byolahi et al. did not evaluate effects relative to the control group,eported null findings, and had broad ranges in CRP values from.1 to 16.5 mg/L [70]. Therefore, it is likely that the significant find-

ngs by Li et al. were largely derived from a 30-day evaluation of00 mg/d krill oil that provided only 80 mg/d of EPA+DHA in peo-le with rheumatoid arthritis, CVD, or osteoarthritis and in whichhe placebo group became more inflamed [69]. This publication byeutsch et al. did not address potential ethical concerns, such asversight by a review board or inclusion of a conflicts of interesttatement. It should also be noted that the pooled analysis of stud-es in which dietary interventions were used to increase omega-3atty acids did not result in a significant CRP reduction. These highlyeterogeneous study designs do not present a clear interpreta-ion and are not sufficient to come to a consensus regarding thenti-inflammatory effect of omega-3 fatty acids.

. Challenges for translation of pre-clinical efficacy tolinical research

The lack of consistent clinical findings regarding the anti-nflammatory efficacy of omega-3 fatty acids—despite the robustreclinical evidence—may be due to intrinsic limitations in howlinical nutrition studies are currently designed.

.1. Low signal-to-noise ratio

A low signal-to-noise ratio is caused by relatively modest effectsf nutrition interventions combined with typical inter-individualuman variability. In relatively healthy subjects, nutrition interven-ions typically result in modest improvements—a low signal that isifficult to detect. When established risk factors or disease statesre involved, the noise introduced by real world pro-inflammatorytimuli is further amplified. This has been especially challenging forlinical studies of diseases that flare and remit in response to com-lex environmental and biological triggers, such as inflammatoryowel diseases and inflammatory arthritis.

.2. Reversing established inflammation instead of preventing itsevelopment

Clinical models generally test disease reversal (i.e. reduc-ion of established risk factors) rather than disease prevention.esting whether foods or supplements can reverse establishednflammation is a different research question than evaluating

hether they can prevent the development of inflammation.owever, an assessment of primary prevention requires exten-

ive investment of resources and years or decades to generatepplicable results—during which time the medical standard-of-are will likely evolve and potentially render such results lesseaningful.

d Mediators 116–117 (2015) 104–111

4.3. Differences in animal model physiology and duration ofintervention

The animal models that inform clinical trial design are not com-plete physiological equivalents of human pathologies. Additionally,the bioavailability and physiological effects of a nutritional inter-vention in humans compared to animal models may differ. Animalmodels can also be exposed to the intervention across a greaterportion (or entirety) of their life span. These differences should beappreciated when extrapolating what effect the same nutritionalcompounds or bioactives may have in a human system.

It is likely that these factors have played a role in complicatingthe translation of compelling preclinical evidence and have limitedthe ability of previous clinical nutrition studies to evaluate the anti-inflammatory efficacy of omega-3 fatty acids. Therefore, it maybe useful to both appreciate the limitations and incorporate theadvantageous features of different study designs to advance futureclinical studies of omega-3 fatty acids. Utilizing such strategies hasthe potential to enhance the research evaluating the efficacy ofomega-3 fatty acids for preventing human inflammation.

5. Essential features of studies designed to evaluate theefficacy of omega-3 fatty acids for preventing thedevelopment of human inflammation

Epidemiological analyses, mechanistic studies, and controlledclinical trials are complementary approaches to nutrition research.Translational clinical models—in vivo clinical studies that incor-porate mechanistic approaches—are an evolving research tool.Evaluating the efficacy of omega-3 fatty acids for preventing exces-sive inflammation requires the combined strengths of each of thesemodel systems (Fig. 1).

The potential anti-inflammatory efficacy of omega-3 fatty acidsrests on a foundation of pre-clinical testing (e.g. observational stud-ies and mechanistic experiments). Epidemiological data provideskey evidence of beneficial dietary attributes but results may be con-founded by other dietary and lifestyle factors that influence health.Cell and animal models provide a highly controlled study systemwith reduced measurement noise in which essential research ques-tions can be interrogated quickly and efficiently. However, thesemechanistic results may not directly translate to humans (e.g. ifsupra-physiologic doses are used or when key aspects of humanphysiology are not reflected in the model system). Additionally,both mechanistic and epidemiological studies benefit from the abil-ity to test the prevention of inflammatory responses. In cell andanimal models, a variety of inflammatory stimuli can be used tomimic immune system responses in human pathology. Althoughnot a controlled system, individuals evaluated in epidemiologicalstudies experience years of cumulative inflammatory stimuli overthe course of their lives.

Another level of vital information is provided by randomizedcontrolled clinical trials that can interrogate the efficacy of anintervention in the context of human physiology. However, theseclinical studies can be very expensive, require many years to com-plete, and are susceptible to issues regarding statistical power.Relatively short term trials also lack the inflammatory stimuluscomponent provided by mechanistic (or observational) studies.Therefore, these trials are likely more effective for evaluating therelatively stronger effects (including side effects) that occur withpharmacotherapy.

The ideal efficacy study design would possess all four features:

(1) having the interventional nutrition compound (e.g. omega-3fatty acids) onboard prior to the inflammatory stimulus, (2) abilityto isolate the effects of the intervention and assess causality, (3) testphysiologically achievable levels of the nutritional compound and

A.C. Skulas-Ray / Prostaglandins & other Lipid Mediators 116–117 (2015) 104–111 109

F a-3 fatty acids for attenuating the development of excessive inflammation. The dasheda ging can serve as an (imprecise) inflammatory stimulus.

iimt

6e

dastpcetiatappStcv

aoi(ippprt

Fig. 2. Intravenous injection of a low dose of endotoxin (0.6 ng/kg) leads to increasedplasma concentrations of C-reactive protein (CRP). Data presented as mean ± SEM

ig. 1. Strengths of research model systems used to evaluate the efficacy of omegrrow for observational studies reflects that day-to-day inflammatory stimuli and a

ts metabolites in vivo, and (4) testing in the context of human phys-ology. This may be best achieved by using a translational human

odel that incorporates an in vivo inflammatory stimulus similaro that utilized in cell and animal-based model systems.

. Clinical models of induced inflammation to study theffects of omega-3 supplementation

A variety of clinical inflammatory challenge systems have beeneveloped, but each model has its own unique set of complicationsnd contexts in which it is most useful. Wound models, such asuction blisters, can be used to evaluate the ability to heal localizedissue injury. Such research has illuminated the important role ofsychosocial factors in optimizing immune responses [72]. Exer-ise models are also used to induce localized trauma such as thatxperienced by muscle during downhill running. However, underhese conditions inflammatory signals may not reflect immunolog-cal derangements as they are required for muscle growth. High fatnd/or high glucose feeding can stimulate a low grade inflamma-ory response but this is difficult to standardize across individualsnd may only be relevant to metabolic inflammation. Sustainedsychological stress can acutely stimulate inflammatory cytokineroduction, but this response is modest and variable (reviewed byteptoe et al. [73]). Vaccination or viral exposure can be used toest immune responses to viral pathogens but does not provide aonsistent inflammatory stimulus as individuals will have highlyariable abilities to clear the pathogen.

The human endotoxemia model may present the best avail-ble induced inflammation model to evaluate the efficacy ofmega-3 fatty acids in preventing the development of systemicnflammation. The low dose (0.6 ng/kg) intravenous endotoxinlipopolysaccharide, LPS) model safely produces a consistent, acutenflammatory response, with peak CRP production occurring 24 host-injection (Fig. 2) [74]. Injection of human subjects with

urified endotoxin has also been used to test the ability of thera-eutic interventions or medications to attenuate the inflammatoryesponse [75]. However, use of this model requires the comple-ion of an Investigational New Drug Application to access clinical

(n = 10).

Reprinted from Mehta et al. [74].

supplies and extensive resources to monitor research volunteers.Infusion of EPA+DHA-containing emulsions attenuated inflamma-tory responses to endotoxin challenge [76], and 8 weeks of oralsupplementation with 3.4 g/d EPA+DHA attenuated fever responsesto low dose endotoxin challenge [77]. These results demonstratethe potential of the human low-dose endotoxemia model for use infuture clinical nutrition studies evaluating the efficacy of omega-3fatty acids for preventing the development of inflammation.

7. Conclusions

Despite robust pre-clinical evidence for the anti-inflammatoryefficacy of omega-3 fatty acids, clinical trials have notprovided compelling evidence that omega-3 supplementationreduces established inflammation. These findings are likely due

1 er Lipi

timfacb(Noovpattcc

D

PB

A

t

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

10 A.C. Skulas-Ray / Prostaglandins & oth

o many factors, particularly those related to study design. Theres no doubt that additional research is needed, and it is likely that

odifications to the current model of clinical nutrition studies willurther our ability to better evaluate the efficacy of omega-3 fattycids in preventing the development of inflammation. Future clini-al research may benefit from the use of statistical approaches thatetter account for other factors that affect levels of inflammatione.g. removing CRP outliers according to pre-specified strategies).ull clinical findings should not be interpreted to mean thatmega-3 fatty acids have no effect in preventing the developmentf inflammation. This is a question that can only be tested byery large, long-term studies, which may not be an efficient orarticularly informative approach in the context of continuouslydvancing medical therapies. Therefore, clinical models that testhe ability of omega-3 fatty acids to attenuate induced inflamma-ory responses may be a key progression in the design of futurelinical nutrition studies, and low-dose intravenous endotoxinhallenge may represent a particularly useful tool in this regard.

isclosures

ACS-R has received research funding support from Pronova Bio-harma and GlaxoSmithKline, as well as fellowship support fromaxter Pharmaceuticals.

cknowledgment

Chesney K. Richter provided valuable input and editorial assis-ance during preparation of the manuscript.

eferences

[1] Zampelas A, Panagiotakos DB, Pitsavos C, et al. Fish consumption among healthyadults is associated with decreased levels of inflammatory markers related tocardiovascular disease: the ATTICA study. J Am Coll Cardiol 2005;46(1):120–4.

[2] Niu K, Hozawa A, Kuriyama S, et al. Dietary long-chain n-3 fatty acids of marineorigin and serum C-reactive protein concentrations are associated in a popu-lation with a diet rich in marine products. Am J Clin Nutr 2006;84(1):223–9.

[3] Pischon T, Hankinson SE, Hotamisligil GS, Rifai N, Willett WC, Rimm EB. Habit-ual dietary intake of n-3 and n-6 fatty acids in relation to inflammatory markersamong US men and women. Circulation 2003;108(2):155–60.

[4] He K, Liu K, Daviglus ML, et al. Associations of dietary long-chain n-3 polyun-saturated fatty acids and fish with biomarkers of inflammation and endothelialactivation (from the Multi-Ethnic Study of Atherosclerosis [MESA]). Am J Car-diol 2009;103(9):1238–43.

[5] Ferrucci L, Cherubini A, Bandinelli S, et al. Relationship of plasma polyunsatu-rated fatty acids to circulating inflammatory markers. J Clin Endocrinol Metab2006;91(2):439–46.

[6] Makhoul Z, Kristal AR, Gulati R, et al. Associations of obesity with triglyceridesand C-reactive protein are attenuated in adults with high red blood cell eico-sapentaenoic and docosahexaenoic acids. Eur J Clin Nutr 2011;65(7):808–17.

[7] Farzaneh-Far R, Harris WS, Garg S, Na B, Whooley MA. Inverse association oferythrocyte n-3 fatty acid levels with inflammatory biomarkers in patientswith stable coronary artery disease: the Heart and Soul Study. Atherosclerosis2009;205(2):538–43.

[8] Flock MR, Skulas-Ray AC, Harris WS, Gaugler TL, Fleming JA, Kris-EthertonPM. Effects of supplemental long-chain omega-3 fatty acids and erythrocytemembrane fatty acid content on circulating inflammatory markers in a ran-domized controlled trial of healthy adults. Prostaglandins Leukot Essent FattyAcids 2014;91(4):161–8.

[9] Calder PC. n-3 fatty acids, inflammation and immunity: new mechanisms toexplain old actions. Proc Nutr Soc 2013;72(3):326–36.

10] Calder PC. Fatty acids and inflammation: the cutting edge between food andpharma. Eur J Pharmacol 2011;668(Suppl. 1):S50–8.

11] Calder PC. Omega-3 fatty acids and inflammatory processes. Nutrients2010;2(3):355–74.

12] Serhan CN, Petasis NA. Resolvins and protectins in inflammation resolution.Chem Rev 2011;111(10):5922–43.

13] Serhan CN. Pro-resolving lipid mediators are leads for resolution physiology.Nature 2014;510(7503):92–101.

14] Ariel A, Serhan CN. Resolvins and protectins in the termination program of

acute inflammation. Trends Immunol 2007;28(4):176–83.

15] Kang JX. Fat-1 transgenic mice: a new model for omega-3 research.Prostaglandins Leukot Essent Fatty Acids 2007;77(5–6):263–7.

16] Kang JX. From fat to fat-1: a tale of omega-3 fatty acids. J Membr Biol2005;206(2):165–72.

[

d Mediators 116–117 (2015) 104–111

17] Kang JX, Wang J, Wu L, Kang ZB. Transgenic mice: fat-1 mice convert n-6 to n-3fatty acids. Nature 2004;427(6974):504.

18] Hudert CA, Weylandt KH, Lu Y, et al. Transgenic mice rich in endoge-nous omega-3 fatty acids are protected from colitis. Proc Natl Acad Sci USA2006;103(30):11276–81.

19] Rangel-Huerta OD, Aguilera CM, Mesa MD, Gil A. Omega-3 long-chainpolyunsaturated fatty acids supplementation on inflammatory biomakers: asystematic review of randomised clinical trials. Br J Nutr 2012;107(Suppl.2):S159–70.

20] Li K, Huang T, Zheng J, Wu K, Li D. Effect of marine-derived n-3 polyunsaturatedfatty acids on C-reactive protein, interleukin 6 and tumor necrosis factor alpha:a meta-analysis. PLOS ONE 2014;9(2):e88103.

21] Pearson TA, Mensah GA, Alexander RW, et al. Markers of inflammation andcardiovascular disease: application to clinical and public health practice:a statement for healthcare professionals from the Centers for DiseaseControl and Prevention and the American Heart Association. Circulation2003;107(3):499–511.

22] Popa C, Netea MG, Radstake T, et al. Influence of anti-tumour necrosis fac-tor therapy on cardiovascular risk factors in patients with active rheumatoidarthritis. Ann Rheum Dis 2005;64(2):303–5.

23] Gonzalez-Juanatey C, Vazquez-Rodriguez TR, Miranda-Filloy JA, et al. Anti-TNF-alpha-adalimumab therapy is associated with persistent improvement ofendothelial function without progression of carotid intima-media wall thick-ness in patients with rheumatoid arthritis refractory to conventional therapy.Mediat Inflamm 2012;2012:674265.

24] Stagakis I, Bertsias G, Karvounaris S, et al. Anti-tumor necrosis factor therapyimproves insulin resistance, beta cell function and insulin signaling in activerheumatoid arthritis patients with high insulin resistance. Arthritis Res Ther2012;14(3):R141.

25] Ridker PM. Moving beyond JUPITER: will inhibiting inflammation reduce vas-cular event rates? Curr Atheroscler Rep 2013;15(1):295.

26] Ciubotaru I, Lee YS, Wander RC. Dietary fish oil decreases C-reactive protein,interleukin-6, and triacylglycerol to HDL-cholesterol ratio in postmenopausalwomen on HRT. J Nutr Biochem 2003;14(9):513–21.

27] Madsen T, Christensen JH, Blom M, Schmidt EB. The effect of dietary n-3 fattyacids on serum concentrations of C-reactive protein: a dose-response study. BrJ Nutr 2003;89(4):517–22.

28] Vega-Lopez S, Kaul N, Devaraj S, Cai RY, German B, Jialal I. Supplementation withomega3 polyunsaturated fatty acids and all-rac alpha-tocopherol alone and incombination failed to exert an anti-inflammatory effect in human volunteers.Metabolism 2004;53(2):236–40.

29] Geelen A, Brouwer IA, Schouten EG, Kluft C, Katan MB, Zock PL. Intake of n-3fatty acids from fish does not lower serum concentrations of C-reactive proteinin healthy subjects. Eur J Clin Nutr 2004;58(10):1440–2.

30] Fujioka S, Hamazaki K, Itomura M, et al. The effects of eicosapentaenoicacid-fortified food on inflammatory markers in healthy subjects—a random-ized, placebo-controlled, double-blind study. J Nutr Sci Vitaminol (Tokyo)2006;52(4):261–5.

31] Sanders TA, Gleason K, Griffin B, Miller GJ. Influence of an algal triacylglyc-erol containing docosahexaenoic acid (22: 6n-3) and docosapentaenoic acid(22: 5n-6) on cardiovascular risk factors in healthy men and women. Br J Nutr2006;95(3):525–31.

32] Yusof HM, Miles EA, Calder P. Influence of very long-chain n-3 fatty acids onplasma markers of inflammation in middle-aged men. Prostaglandins LeukotEssent Fatty Acids 2008;78(3):219–28.

33] Nieman DC, Henson DA, McAnulty SR, Jin F, Maxwell KR. n-3 polyunsaturatedfatty acids do not alter immune and inflammation measures in enduranceathletes. Int J Sport Nutr Exerc Metab 2009;19(5):536–46.

34] Bloomer RJ, Larson DE, Fisher-Wellman KH, Galpin AJ, Schilling BK. Effect ofeicosapentaenoic and docosahexaenoic acid on resting and exercise-inducedinflammatory and oxidative stress biomarkers: a randomized, placebo con-trolled, cross-over study. Lipids Health Dis 2009;8:36.

35] Dangardt F, Osika W, Chen Y, et al. Omega-3 fatty acid supplementationimproves vascular function and reduces inflammation in obese adolescents.Atherosclerosis 2010;212(2):580–5.

36] Ottestad I, Vogt G, Retterstol K, et al. Oxidised fish oil does not influence estab-lished markers of oxidative stress in healthy human subjects: a randomisedcontrolled trial. Br J Nutr 2012;108(2):315–26.

37] Root M, Collier SR, Zwetsloot KA, West KL, McGinn MC. A randomized trial offish oil omega-3 fatty acids on arterial health, inflammation, and metabolicsyndrome in a young healthy population. Nutr J 2013;12(1):40.

38] Chan DC, Watts GF, Mori TA, Barrett PH, Redgrave TG, Beilin LJ. Randomized con-trolled trial of the effect of n-3 fatty acid supplementation on the metabolism ofapolipoprotein B-100 and chylomicron remnants in men with visceral obesity.Am J Clin Nutr 2003;77(2):300–7.

39] Jellema A, Plat J, Mensink RP. Weight reduction, but not a moderate intakeof fish oil, lowers concentrations of inflammatory markers and PAI-1 anti-gen in obese men during the fasting and postprandial state. Eur J Clin Invest2004;34(11):766–73.

40] Browning LM, Krebs JD, Moore CS, Mishra GD, O’Connell MA, Jebb SA. Theimpact of long chain n-3 polyunsaturated fatty acid supplementation on

inflammation, insulin sensitivity and CVD risk in a group of overweight womenwith an inflammatory phenotype. Diabetes Obes Metab 2007;9(1):70–80.

41] Itariu BK, Zeyda M, Hochbrugger EE, et al. Long-chain n-3 PUFAs reduce adiposetissue and systemic inflammation in severely obese nondiabetic patients: arandomized controlled trial. Am J Clin Nutr 2012;96(5):1137–49.

er Lipi

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

A.C. Skulas-Ray / Prostaglandins & oth

42] Gammelmark A, Madsen T, Varming K, Lundbye-Christensen S, Schmidt EB.Low-dose fish oil supplementation increases serum adiponectin without affect-ing inflammatory markers in overweight subjects. Nutr Res 2012;32(1):15–23.

43] Mori TA, Woodman RJ, Burke V, Puddey IB, Croft KD, Beilin LJ. Effect ofeicosapentaenoic acid and docosahexaenoic acid on oxidative stress andinflammatory markers in treated-hypertensive type 2 diabetic subjects. FreeRadic Biol Med 2003;35(7):772–81.

44] Malekshahi Moghadam A, Saedisomeolia A, Djalali M, Djazayery A, Pooya S,Sojoudi F. Efficacy of omega-3 fatty acid supplementation on serum levels oftumour necrosis factor-alpha, C-reactive protein and interleukin-2 in type 2diabetes mellitus patients. Singapore Med J 2012;53(9):615–9.

45] Jones PJ, Demonty I, Chan YM, Herzog Y, Pelled D. Fish-oil esters of plant sterolsdiffer from vegetable-oil sterol esters in triglycerides lowering, carotenoidbioavailability and impact on plasminogen activator inhibitor-1 (PAI-1) con-centrations in hypercholesterolemic subjects. Lipids Health Dis 2007;6:28.

46] Micallef MA, Garg ML. Anti-inflammatory and cardioprotective effects of n-3 polyunsaturated fatty acids and plant sterols in hyperlipidemic individuals.Atherosclerosis 2009;204(2):476–82.

47] Derosa G, Maffioli P, D’Angelo A, et al. Effects of long chain omega-3 fatty acidson metalloproteinases and their inhibitors in combined dyslipidemia patients.Expert Opin Pharmacother 2009;10(8):1239–47.

48] Derosa G, Cicero AF, Fogari E, et al. Effects of n-3 PUFAs on postprandial variationof metalloproteinases, and inflammatory and insulin resistance parameters indyslipidemic patients: evaluation with euglycemic clamp and oral fat load. JClin Lipidol 2012;6(6):553–64.

49] Kelley DS, Siegel D, Fedor DM, Adkins Y, Mackey BE. DHA supplementationdecreases serum C-reactive protein and other markers of inflammation inhypertriglyceridemic men. J Nutr 2009;139(3):495–501.

50] Skulas-Ray AC, Kris-Etherton PM, Harris WS, Vanden Heuvel JP, WagnerPR, West SG. Dose-response effects of omega-3 fatty acids on triglycerides,inflammation, and endothelial function in healthy persons with moderatehypertriglyceridemia. Am J Clin Nutr 2011;93(2):243–52.

51] Koh KK, Quon MJ, Shin KC, et al. Significant differential effects of omega-3 fattyacids and fenofibrate in patients with hypertriglyceridemia. Atherosclerosis2012;220(2):537–44.

52] Tajalizadekhoob Y, Sharifi F, Fakhrzadeh H, et al. The effect of low-dose omega3 fatty acids on the treatment of mild to moderate depression in the elderly: adouble-blind, randomized, placebo-controlled study. Eur Arch Psychiatry ClinNeurosci 2011;261(8):539–49.

53] Grundt H, Nilsen DW, Mansoor MA, Nordoy A. Increased lipid peroxidation dur-ing long-term intervention with high doses of n-3 fatty acids (PUFAs) followingan acute myocardial infarction. Eur J Clin Nutr 2003;57(6):793–800.

54] Madsen T, Christensen JH, Schmidt EB. C-reactive protein and n-3 fatty acids inpatients with a previous myocardial infarction: a placebo-controlled random-ized study. Eur J Nutr 2007;46(7):428–30.

55] O’Keefe Jr JH, Abuissa H, Sastre A, Steinhaus DM, Harris WS. Effects of omega-3fatty acids on resting heart rate, heart rate recovery after exercise, and heart ratevariability in men with healed myocardial infarctions and depressed ejectionfractions. Am J Cardiol 2006;97(8):1127–30.

56] Zhao YT, Shao L, Teng LL, et al. Effects of n-3 polyunsaturated fatty acidtherapy on plasma inflammatory markers and N-terminal pro-brain natri-uretic peptide in elderly patients with chronic heart failure. J Int Med Res2009;37(6):1831–41.

57] Eschen O, Christensen JH, LA Rovere MT, Romano P, Sala P, Schmidt EB.Effects of marine n-3 fatty acids on circulating levels of soluble adhesion

molecules in patients with chronic heart failure. Cell Mol Biol (Noisy-le-grand)2010;56(1):45–51.

58] Mackay I, Ford I, Thies F, Fielding S, Bachoo P, Brittenden J. Effect of Omega-3fatty acid supplementation on markers of platelet and endothelial function inpatients with peripheral arterial disease. Atherosclerosis 2012;221(2):514–20.

[

d Mediators 116–117 (2015) 104–111 111

59] Madsen T, Schmidt EB, Christensen JH. The effect of n-3 fatty acids on C-reactive protein levels in patients with chronic renal failure. J Ren Nutr2007;17(4):258–63.

60] Himmelfarb J, Phinney S, Ikizler TA, Kane J, McMonagle E, Miller G. Gamma-tocopherol and docosahexaenoic acid decrease inflammation in dialysispatients. J Ren Nutr 2007;17(5):296–304.

61] Kooshki A, Taleban FA, Tabibi H, Hedayati M. Effects of marine omega-3 fattyacids on serum systemic and vascular inflammation markers and oxidativestress in hemodialysis patients. Ann Nutr Metab 2011;58(3):197–202.

62] Bowden RG, Wilson RL, Deike E, Gentile M. Fish oil supplementation lowers C-reactive protein levels independent of triglyceride reduction in patients withend-stage renal disease. Nutr Clin Pract 2009;24(4):508–12.

63] Deike E, Bowden RG, Moreillon JJ, et al. The effects of fish oil supplementationon markers of inflammation in chronic kidney disease patients. J Ren Nutr2012;22(6):572–7.

64] Barbosa DS, Cecchini R, El Kadri MZ, Rodriguez MA, Burini RC, Dichi I. Decreasedoxidative stress in patients with ulcerative colitis supplemented with fish oilomega-3 fatty acids. Nutrition 2003;19(10):837–42.

65] Sundrarjun T, Komindr S, Archararit N, et al. Effects of n-3 fatty acids on seruminterleukin-6, tumour necrosis factor-alpha and soluble tumour necrosis fac-tor receptor p55 in active rheumatoid arthritis. J Int Med Res 2004;32(5):443–54.

66] Freund-Levi Y, Hjorth E, Lindberg C, et al. Effects of omega-3 fatty acidson inflammatory markers in cerebrospinal fluid and plasma in Alzheimer’sdisease: the OmegAD study. Dement Geriatr Cogn Disord 2009;27(5):481–90.

67] Mohammadi E, Rafraf M, Farzadi L, Asghari-Jafarabadi M, Sabour S. Effects ofomega-3 fatty acids supplementation on serum adiponectin levels and somemetabolic risk factors in women with polycystic ovary syndrome. Asia Pac JClin Nutr 2012;21(4):511–8.

68] Finocchiaro C, Segre O, Fadda M, et al. Effect of n-3 fatty acids on patientswith advanced lung cancer: a double-blind, placebo-controlled study. Br J Nutr2012;108(2):327–33.

69] Deutsch L. Evaluation of the effect of Neptune Krill Oil on chronic inflammationand arthritic symptoms. J Am Coll Nutr 2007;26(1):39–48.

70] Kolahi S, Ghorbanihaghjo A, Alizadeh S, et al. Fish oil supplementationdecreases serum soluble receptor activator of nuclear factor-kappa B lig-and/osteoprotegerin ratio in female patients with rheumatoid arthritis. ClinBiochem 2010;43(6):576–80.

71] Wright SA, O’Prey FM, McHenry MT, et al. A randomised interventional trialof omega-3-polyunsaturated fatty acids on endothelial function and dis-ease activity in systemic lupus erythematosus. Ann Rheum Dis 2008;67(6):841–8.

72] Kiecolt-Glaser JK, Loving TJ, Stowell JR, et al. Hostile marital interactions, proin-flammatory cytokine production, and wound healing. Arch Gen Psychiatry2005;62(12):1377–84.

73] Steptoe A, Hamer M, Chida Y. The effects of acute psychological stress on cir-culating inflammatory factors in humans: a review and meta-analysis. BrainBehav Immun 2007;21(7):901–12.

74] Mehta NN, Heffron SP, Patel PN, et al. A human model of inflammatory cardio-metabolic dysfunction; a double blind placebo-controlled crossover trial. JTransl Med 2012;10:124.

75] Lowry SF. Human endotoxemia: a model for mechanistic insight and therapeu-tic targeting. Shock 2005;24(Suppl. 1):94–100.

76] Pluess TT, Hayoz D, Berger MM, et al. Intravenous fish oil blunts the phys-

iological response to endotoxin in healthy subjects. Intensive Care Med2007;33(5):789–97.

77] Ferguson JF, Mulvey CK, Patel PN, et al. Omega-3 PUFA supplementation andthe response to evoked endotoxemia in healthy volunteers. Mol Nutr Food Res2014;58(3):601–13.