Clinical Nutrition xxx (2017) 1e11

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Meta-analyses

A meta-analysis of n-3 polyunsaturated fatty acids effects on

circulating acute-phase protein and cytokines in gastric cancer
Michel C. Mocellin*, Ricardo Fernandes, Thayz R. Chagas, Erasmo B.S.M. Trindade
polis, Santa Catarina, Brazil
Graduate Program of Nutrition, Nutrition Department, Federal University of Santa Catarina, Floriano

a r t i c l e i n f o s u m m a r y

Article history: Background & aims: Chronic inflammation is related with cancer and leads to worsening prognosis in
Received 15 November 2016 cancer patients. n-3 polyunsaturated fatty acids (PUFAs) supplementation has been proposed as adjuvant
Accepted 8 May 2017 treatment in cancer due anti-inflammatory properties. In the present meta-analysis, we pooled ran-
domized clinical trials (RCTs) assessing the effects of n-3 PUFAs (from fish oil isolated or added in an
Keywords: immunonutrition formula) on inflammatory markers in gastric cancer.
Inflammation
Methods: A comprehensive literature search was performed in Medline, Scopus, Cochrane library, Sci-
Eicosapentaenoic acid
ence Direct and Web of Science, besides GOOGLE Scholar and a hand searching of reference lists, through
Docosahexaenoic acid
Stomach neoplasm
July 2016. We pooled the effect size from individual studies using a random-effect model and carried out
heterogeneity and sensitivity analyses.
Results: Nine trials (698 patients) fulfilled the entry criteria and were included in the synthesis of the
systematic review. Eight were carried out in surgical patients and one in patients that received
chemotherapy. Four used only fish oil as intervention and five used an immunonutrition formula. Global
meta-analysis demonstrated higher albumin (7 studies, SMD 0.28; 95% CI 0.07, 0.48) and prealbumin
(4 studies, SMD 0.56; 95% CI 0.12, 1.00) concentrations, and lower IL-6 (2 studies, SMD
0.71; 95%
CI
1.15,
0.27) and TNF-a (2 studies, SMD
0.92; 95% CI
1.58,
0.26) concentrations in patients of the
intervention group as compared to control group. However, total protein, transferrin and CRP concen-
trations were not improved by n-3 PUFAs supplementation.
Conclusion: This study provides evidence that n-3 PUFAs supplementation from fish oil or added an
immunonutrition formula has favorable effects on inflammatory markers in gastric cancer patients
undergoing surgical procedures.
© 2017 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction
The association between cancer and inflammation is recognized
more than a century ago and well discussed in scientific literature.
In gastric cancer the chronic inflammation has a key role in carci-
nogenesis, and most tumours, if not all, secrete molecules that
activate inflammatory cells promoting an inflammatory status [1,2].
Abbreviations: CI, Confidence Interval; CRP, C-Reactive Protein; DHA, Docosa-
hexaenoic acid; EPA, Eicosapentaenoic acid; GPR, G-Protein Receptor; IFN, Inter-
feron; IL, Interleukin; NF-kB, Nuclear factor kappa B; PPAR, Peroxisome Proliferator-
Activated Receptor; PUFA, polyunsaturated fatty acids; RBP, Retinol Binding Protein;
RCT, Randomized Clinical Trial; SD, Standard Deviation; SMD, Standardized Mean
Difference; TNF, Tumor Necrosis Factor.
* Corresponding author. Graduate Program of Nutrition, Nutrition department,
Health Science Center, University Campus e Trindade, Federal University of Santa
Catarina, Florianopolis, Santa Catarina, 88040-900, Brazil. Fax: þ55 48 3721 9542.
E-mail address: michel.mocellin@hotmail.com (M.C. Mocellin).
Cytokines such as IL-6, TNF-a and IL-1b can serve as growth and
survival factors that act on malignant cells [3e5]; induce an acute-
phase protein response by hepatocytes and changes in metabolism
[6]; stimulate angiogenesis, tumour progression and metastasis
[4,5]; cause chemoresistance and increase the toxicity to anti-
cancer drug or radiation therapy [5]; contribute with cachexia-
related to cancer [6]; and also, maintain tumour-promoting
inflammation in a kind of vicious-cycle [4].
The n-3 polyunsaturated fatty acids (PUFAs), especially eicosa-
pentaenoic (EPA) and docosahexaenoic (DHA), have been linked to
numerous beneficial effects on cancer such as anti-inflammatory,
pro-apoptotic, antineoplastic and anti-catabolic properties [7e10].
Regarding their influences on inflammation related to cancer, there
is a previous meta-analysis showing a reducing action of n-3 PUFAs
on several inflammatory markers in colorectal cancer [11]. How-
ever, in gastric cancer, there are many published clinical trials
assessing the effects of n-3 PUFAs on inflammatory response with

http://dx.doi.org/10.1016/j.clnu.2017.05.008
0261-5614/© 2017 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
2 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11
discrepant results. Therefore, the aim of the present study was to
systematically review the evidence on the effects of n-3 PUFAs
supplementation on inflammatory markers in gastric cancer pa-
tients, through a pooled effect size from all studies and from spe-
cific subgroups.
2. Materials and methods
We planned and performed this systematic review and meta-
analysis in accordance with Cochrane Handbook for Systematic
Reviews of Intervention [12] and Preferred Reporting Items for
Systematic Review and Meta-analysis statement [13] to answer the
following question: does n-3 PUFAs supplementation in gastric
cancer alter circulating biomarkers related to inflammatory
response? The protocol of this study was registered on PROSPERO
database (registration number: CRD42016037283).
2.1. Literature search
A systematic search of published literature was conducted in
Cochrane Central Register of Controlled Trials (CENTRAL), Medline,
Scopus, Web of Science and Science Direct to identify all relevant
trials until July, 14th 2016. We used search terms, keywords or
medical subject headings relating to the groups of PICO acronym
(patient, intervention, comparison and outcome): Patient (P) e
patients diagnosed with gastric cancer: (Gastric OR stomach OR
gastrointestinal) AND (cancer OR neoplasm* OR malignan* OR
tumour OR tumor); Intervention (I) e n-3 PUFAs: (“fish oil” OR “n-3”
OR “unsaturated fatty acids” OR “eicosapentaenoic acid” OR EPA OR
“docosahexaenoic acid” OR DHA OR “marine oil” OR “krill oil” OR
omega3 OR “omega-3” OR “omega 3” OR PUFA OR “u-3”); Compar-
ison (C): standard diet or placebo/control supplement, or none
(omitted in the search strategy); Outcomes (O): circulating acute
phase proteins (“acute phase protein” OR albumin OR “prealbumin”
OR transferrin OR “retinol binding protein” OR RBP OR CRP OR “C-
reactive protein” OR ceruloplasmin OR fibrinogen) and/or others
biomarkers of inflammatory status (Cytokin* OR interleukin* OR
inflammat* OR “inflammatory markers” OR “inflammatory mediators”
OR eicosanoid* OR prostaglandin* OR leukotriene* OR chemokine* OR
TNF OR “tumor necrosis factor” IFN OR interferon).
We combined the search terms using Boolean operators (OR was
used to combine search terms from the same PICO's group and,
AND, to combine search terms from different groups). Also, we used
truncation symbol (*) to find all multiple word endings and spell-
ings, and, quotation markers (“ ”) to force the database match a
term exact enclosed by this symbol (search for exact phrases). No
filter was used and in the Science Direct database we added the
term “clinical trial” to refine the search.
Additionally, reference lists of eligible studies and topic-related
reviews were also manually searched to find others relevant trials.
GOOGLE Scholar was also searched with a reduced strategy
[(“omega-3” OR “n-3 PUFA” OR “fish oil” OR EPA OR DHA) AND
(“gastric cancer” OR “gastric neoplasm” OR “stomach tumor”)] and
only the first 10 pages of results were reviewed. The number of
articles reviewed in this database were not counted in the total
number of records retrieved, because the search mechanism not
permit a combination of all variations of the intervention, clinical
setting and outcomes terms performing a precise search strategy.
2.2. Entry criteria
Trials were included if they met the following criteria: 1)
controlled or randomized clinical trial carried out in humans; 2) full
articles published (published abstracts were not included); 3)
language of publication in English; 4) study sample composed only
Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
by patients diagnosed with gastric cancer by an histological tech-
nique; 5) n-3 PUFAs offered in capsules (EPA and/or DHA isolate or
fish oil) or liquid diet enriched with these fatty acids (cocktails with
n-3 PUFA plus antioxidants, glutamine, arginine or nucleic acids
were not excluded); 6) control group that not received in parallel n-
3 PUFAs supplementation or the amounts of n-3 PUFAs was much
less than intervention group (we included studies which control
group received a placebo without n-3 PUFAs, i.e. a supplement of
the paraffin; or, a control/placebo formula with nutritional char-
acteristics similar to the intervention formula with the exception of
the amount of n-3 PUFAs present); 7) assessment the effect of n-3
PUFAs on any biomarkers of our interest (acute phase protein, cy-
tokines and/or eicosanoids).
2.3. Data selection and extraction
Records retrieved by search strategy in databases, except
GOOGLE Scholar, were exported to a reference manager software
and those repeated were quantified and excluded. One investigator
(MM), initially screened the records by title analyzing. If it was very
unlikely to be relevant, the abstract was not assessed. Records with
titles denoting relevance had their abstract read and, those
potentially eligible, full texts were assessed to confirm entry
criteria. This selection process was performed two times by the
same investigator. When eligibility was not clear by data presented
in the article, original authors were contacted by e-mail to obtain
more information. If there were doubts about the eligibility and we
did not have return from authors clarifying our doubts, the article
was not included in this systematic review and meta-analysis.
The following information about the methodology and contin-
uous data for specific outcomes were extracted independently from
the eligible articles by two investigators (MM, TC) and were
registered in a predesigned data extraction table: study identifi-
cation (authors and publication year); country; blinding; random-
ization technique; number of patients in each group that were
randomized, dropped and malnourished; number of male and fe-
male; mean age; anti-cancer treatment; n-3 PUFAs supplement
data (dose, formula, time and route of administration); n-6:n:3
fatty acid ratio of control and intervention supplements; antibiotic
use during the study period; and, results of inflammatory bio-
markers concentrations (mean and standard deviation). Divergence
data from two investigators were discussed after new consult to the
original article and if no reached consensus, a third investigator
(ET) resolved it.
To obtain data for meta-analysis from articles that mentioned
effects on inflammatory markers without reporting numerical data,
we contacted the authors by e-mail. For continuous data presented
in median and interquartile range or minimum and maximum
values, we estimated their mean and standard deviation by the
equation proposed by Hozo et al. (2005) [14] (if the original author
did not provide it by e-mail). Studies were excluded of the meta-
analysis (only used in systematic review synthesis) when data
were not possible to estimate and the data could not be obtained
from the authors.
2.4. Risk of bias assessment
Risk of individual study bias was assessed using the Cochrane
Collaboration's tool, version 5.1.0 [12]. The tool comprises seven
domains to evaluate the risk of six known bias and other unknown:
1) adequacy of sequence generation and, 2) allocation concealment
(selection bias); 3) blinding of participants and researchers (per-
formance bias); 4) blinding of outcome assessment (detection
bias); 5) incomplete outcome data (attrition bias); 6) selective
outcome reporting (reporting bias); and, 7) other potential sources
 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11 3

of bias (others bias). Two independent investigators judged the risk

of bias into each domain as low (þ), high (
) or unclear (?) risk
(unclear was attributed when the article did not provide sufficient
information to conclude about the presence or absence of bias).
Disagreements between investigators on judgement of bias risk
were resolved by discussion after consulting the conflicting infor-
mation in the article, until consensus was reached.

2.5. Statistical analysis

We conducted independent meta-analysis for those outcomes

more investigated in the studies included (albumin, prealbumin,
transferrin, total protein, CRP, TNF and IL-6) using STATA vs. 13.0
(StataCorp LP, Texas, USA) from mean and respective SD quantified
after supplementation ended (if the study was designed to measure
the outcomes at various moments after supplementation, we used
the measured value on the day that supplementation ceased or the
first value measured after the end of supplementation). Effect sizes
between treatment and control groups for outcomes were calcu-
lated based on standardized mean differences (SMD), since the
measurement units for each biomarker varied within studies, using
Hedge's g statistics and pooled by random-effects meta-analysis
models (this model was applied due the differences in the design
and patients characteristics of included studies). We considered
significant a p-value for SMD test <0.05. For interpretation of
the effects sizes from SMD we adopted the thresholds proposed
by Rosenthal (1996) [15]: 0.2 small, 0.5 medium, 0.8 large and 1.3
very large.
Heterogeneity was considered significant when p < 0.1 on the
chi-square test or I2 > 50%. Subgroup analysis was carried out to
assess potential sources of heterogeneity, and, sensitivity analysis,
to verify the influence of each study on global SMD (made only for
pooled analysis from each outcome). Publication bias analysis was
Fig. 1. Flow chart of the literature search, screening, and selection process for eligible
not performed because the minimum number of studies for the trials.
implementation of this test (10) was not achieved [12].

3. Results
3.1. Search results
From the literature search, 2057 records were retrieved (71 from
CENTRAL, 690 from Scopus, 881 from Science Direct, 210 from Web
of Science and 205 from Pubmed). After exclusion of repeated ar-
ticles, 1631 records were screened and five met the entry criteria.
Additionally, four studies were found from other sources (search in
GOOGLE scholar and reviewing the references of important origi-
nals articles and reviews about the theme), totalling nine studies
included in our systematic review and meta-analysis. The complete
flowchart is shown in Fig. 1.
3.2. Trials characteristics
An overview of study design and demographic parameters of
patients enrolled in the eligible nine studies [16e24] is shown in
Table 1. All were randomized clinical trials (RCTs) published from
2005 and conducted in Europe [17,19e21] or Asia [16,18,20,22e24].
In total, 355 patients were allocated to intervention group and 343
to control group. The sample size (excluded the losses of follow-up)
ranged from 26 [20] to 231 [18] patients. Four studies [17e19,24]
had losses of participants during follow-up: in three [17e19] the
proportion of losses was balanced between intervention and con-
trol groups (main causes were: death [17], surgical and feeding
complications [17,19], withdrew of consent [17,19], not fulfilled
criteria of permanence in the study [18] and randomization of non-
eligible patients [18]); and, in one [24], losses occurred only in
control group (the reason reported by authors to dropout 23.1% of
patients from this group was “incomplete data”). None of the
studies applied intention to treat analysis.
Two RCTs were double-blind [17,22] and a third [20] reported
that only researchers were blinded. All comprised patients of both
sexes and a pool of cancer stages. Mean age of participants was
similar among the RCTs. Only one [22] was performed with patients
undergoing chemotherapy treatment, while the eight remaining
RCTs [16e21,23,24] were carried out in surgical settings.
Details of n-3 PUFAs supplementation protocol used in each RCT
are described in Table 2. Five trials [16e18,21,23] used an immu-
nomodulatory diet enriched with n-3 PUFAs, arginine and antiox-
idants (besides these immunonutrients, one [16] added glutamine
and four [17,18,21,23], ribonucleic acids). n-3 PUFA from fish oil
without others immunonutrients was used in four trials
[19,20,22,24]. Equal number of RCTs (n ¼ 3) administered n-3
PUFAs orally [18,22,23], enterally [16,17,21] and parenterally
[19,20,24]. In studies where n-3 PUFA was offered orally, amounts
of EPA and DHA were similar for all patients, while in studies where
n-3 PUFAs were offered parenterally or enterally, dosages of EPA
and DHA varied among patients within the same RCT because it was
established according to the individual body weights or energy
requirements.
For those studies with surgical patients, in six [16,17,19e21,24]
n-3 PUFA was offered at postoperatively period (5e7 days), and
in two [18,23], pre-operatively (5 or 7 days). Use of antibiotics
during the study period was reported in five [17,18,21,23,24] of
eight RCTs in surgical patients.

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
4 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11

Table 1
Basic characteristics of the randomized clinical trials included in the meta-analysis.

Trial (Year) Country Blinding Randomization technique n randomized: n n male: n female n malnourished Mean age,
analyzed (% dropouts) (%)a (years)

Chen et al. (2005) [16] China NR NR 20: 20 (0.0) NR NR NR

20: 20 (0.0)
Farreras et al. (2005) [17] Spain Double Random number table 33: 30 (9.1) 17: 13 5 (16.7) 66.7¥
33: 30 (9.1) 15: 15 8 (26.7) 69.2
Klek et al. (2005) [19] Poland NR NR 36: 30 (16.7) NR 16 (51.6) NR
35: 30 (14.3) 16 (48.5)
Okamoto et al. (2009) [23] Japan NR Block randomization 30: 30 (0.0) 20: 10 NR 66.9
30: 30 (0.0) 22: 8 70.9
Makay et al. (2011) [20] Turkey Researchers Computer-derived 14: 14 (0.0) 7: 7 NR 59.5
block randomization 12: 12 (0.0) 10: 2 61.7
Fujitani et al. (2012) [18] Japan NR Randomization by 127: 120 (5.5) 97: 30 4 (3.1) 64b
minimization with 117: 111 (5.1) 84: 33 1 (0.9) 65b
an algorithm
Marano et al. (2013) [21] Italy NR NR 54: 54 (0.0) 34: 20 33 (61.0) 66.6
55: 55 (0.0) 37: 18 30 (54.0) 65.1
Wei et al. (2014) [24] China NR NR 26: 26 (0.0) 15: 11 NR 50.5b
26: 20 (23.1) 11: 9 59.0b
Nemati et al. (2015) [22] Iran Double NR 15: 15 (0.0) NR NR NR
15: 15 (0.0)

Abbreviation: NR e No reported; Bold and italic type corresponds to control group. No italic and non bold type corresponds to intervention group.
¥
Significant difference between control and intervention group (p < 0.05).
a
Criteria for diagnostic of malnutrition: Klel et al.: any of these parameters: weight loss (last 3e6 months)  10%, body mass index  18 kg/m2, serum albumin
concentration  3.2 g/dL; serum prealbumin concentration  15 mg/dL; total lymphocytes counts  1200/mm3; Farreras et al.: >10% of lost body weight; Marano et al.: no
reported; Fujitani et al.: subjective global assessment.
b
Values are median.

A placebo supplement without n-6 or n-3 PUFAs was used in
one study [22]. In this trial, Nemati and cols (2015) [22]offered to
the control group a supplement containing paraffin. In another RCT
[18], control group did not receive any supplement. The remaining
seven studies [16,17,19e21,23,24], the formula offered to inter-
vention group containing high dose of n-3 PUFA was replaced by a
formula rich in n-6 PUFAs that was offered to control group. We
emphasize that in Chen's RCT [16] the intervention group received a
supplement with more n-6 PUFAs than n-3 (n-6:n-3 ratio ¼ 3.4:1),
while in all other studies, intervention group received a formula
with more n-3 than n-6 PUFAs (n-6:n-3 ratio  0.9:1).
3.3. Risk of bias assessment
The assessment of risk of bias in the included studies using the
Cochrane tool is shown in Fig. 2. None of RCTs included was free of
potential bias. Although they are randomized trials, only three
[17,18,20] reported the technique used to generate the random
sequence. Regarding allocation concealment, none provided clear
information to judge as adequate. Double-blinding design was re-
ported in two studies [17,22], but was not mentioned who was
blinded (participants, researches or statistician). In Makay's trial
[20] only researchers were blinded. Low risk of reporting bias was
attributed for all trials, except for Okamoto's RCTs [23], where re-
sults for some outcomes were not provided (albumin, prealbumin,
transferrin and retinol binding protein). In the long-period of
supplementation in Nemati's trial [22] (6 weeks) the compliance on
n-3 PUFAs supplements consumption (10 tablets/day) was impor-
tant, but was not assessed. In these same study, the nutrients
consumption from foods and beverages was assessed during the
trial and shown a significant lower consumption for control group
at the end of week 6 for energy, carbohydrate, fat, protein and the
majority of vitamins and minerals, what could explain the positive
effects found in intervention group for the outcomes assessed.
In RCTs performed with surgical patients, due to infusion/con-
sumption monitoring of the diet enriched with n-3 PUFAs during
the hospital stay by researchers, the compliance probably was
secured. However, only one [17] of the eight surgical RCT's shown
results about the nutrients intake/infused during the follow-up
period. Still, in these RCTs, four [16,19,20,24] did not report
important information about the surgery procedures that could
influence the outcomes. Baseline characteristics (demographic and
clinical) of both groups investigated were not provided in four RCTs
[16,19,22,23], which does not guarantee the comparability of the
groups into the RCTs.
3.4. Effect of n-3 PUFA on biomarkers of inflammatory status
Considering the individual studies, we observed a small number
of studies that described statistical differences in circulating con-
centrations of albumin [22], prealbumin [16,19], transferrin [16,22],
total protein [17] and retinol binding protein [23] between study
groups after n-3 PUFAs supplementation (Table 2).
One study for albumin [23], prealbumin [23], total protein [23]
and RBP [23], and, two for transferrin [22,23], did not present in
the article the mean values and/or respectively standard deviation
for these outcomes and we were unable to contact authors to
obtain these data. Thus, meta-analysis was performed with seven
RCTs for albumin; four for prealbumin and transferrin; and three
for total protein.
Global meta-analysis of data from included RCTs showed a sig-
nificant increase in albumin levels (SMD ¼ 0.28; 95% CI ¼ 0.07, 0.48;
p ¼ 0.008; Fig. 3 and Supplemental Table 1) and in prealbumin
(SMD ¼ 0.56; 95% CI ¼ 0.12, 1.00; p ¼ 0.013; Fig. 3 and
Supplemental Table 2) with presence of heterogeneity (Chi2
test ¼ 0.063; I2 ¼ 59.0%), in favor to intervention group, while for
transferrin and total protein, there were no significant results
(Fig. 3; Supplemental Tables 3 and 4), but with substantial het-
erogeneity. When the RCTs that assessed albumin were stratified
there were significant effects for: RCTs with surgical patients with
n-3 supplementation at post-operative period (SMD ¼ 0.24; 95%
CI ¼ 0.02, 0.45; p ¼ 0.030; Supplemental Table 1); in double-blind
RCTs (SMD ¼ 0.59; 95% CI ¼ 0.16, 1.01; p ¼ 0.007; Supplemental
Table 1); in studies with lower than 50 patients enrolled
(SMD ¼ 0.36; 95% CI ¼ 0.02, 0.69; p ¼ 0.037; Supplemental Table 1);
and in RCTs that offered n-3 PUFAs from fish oil alone (SMD ¼ 0.34;

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11 5

Table 2
Characteristics of the n-3 PUFAs supplementation protocol and outcomes assessed in included clinical trials.

Trial (year) Antineoplastic EPA þ DHA, g/d Time and period of Formulation Route of Content of Outcomes
treatment supplementation administration n-6:n-3 in control
formulaj
intervention
formulae

Chen et al. Surgical NRa 7 postoperative days Immunonutrient- Enteral 5:1 j 3.4:1 [ prealbumin,
(2005) [16] enriched and high IL-2, transferrin
protein liquid diet 4 albumin
(arginine, Y IL-6, TNF-a
glutamine and n-3
PUFAs)
Farreras et al. Surgical 4.2b 7 postoperative days Immunonutrient- Enteral 6.8:1 j 0.7:1d [ total protein
(2005) [17] enriched liquid diet 4 prealbumin,
(arginine, n-3 PUFAs albumin, transferrin
and RNA)
Klek et al. Surgical 0.2 g/kgc 7 postoperative days Fish oil Parenteral 6.6:1 j 0.1:1 4 albumin
(2005) [19] [ prealbumin
Okamoto et al. Surgical 3.1 7 preoperative days Immunonutrient- Oral 3:1 j 0.7:1 4 prealbumin,
(2009) [23] enriched liquid diet transferrin,
(arginine, n-3 PUFAs albumin,
and RNA) total protein
[ retinol binding protein
Makay et al. Surgical 0.2 g/kgc 5 postoperative days Fish oil Parenteral 7:1 j 0.1:1 4 albumin
(2011) [20]
Fujitani et al. Surgical 3.4 (2 g EPA þ 5 preoperative days Immunonutrient- Oral 0:0 j 0.8:1 4 CRP
(2012) [18] 1.4 g DHA) enriched liquid diet
(arginine, n-3 PUFAs
and RNA)
Marano et al. Surgical NRa 7 postoperative days Immunonutrient- Enteral 6.6:1 j 0.7:1 4 albumin,
(2013) [21] enriched liquid diet transferrin,
(arginine, n-3 PUFAs total protein
and RNA)
Wei et al. Surgical 0.2 g/kgc 6 postoperative days Fish oil Parenteral 7.4:1 j 0.2:1 4 albumin,
(2014) [24] transferrin, total
protein, prealbumin
and retinol binding
protein, CRP,
VEGF, IGF-1
Y IL-1b, IL-6, TNF-a
Nemati et al. Chemotherapy 3 (1.8 g EPA þ 6 weeks Fish oil Oral 0:0 j <0.1:1 [ albumin, transferrin
(2015) [22] 1.2 g DHA)

Abbreviations: NR e Not reported; RNA e ribonucleic acids; 4 no significant difference between the intervention and control groups after intervention; Y e significantly
lower than control group after intervention; [ e significantly higher than control group after intervention.
a
Total amount of infused diet was determined according to the weight of the participant considering 35 kcal/kg in the study of Marano et al. and 30 non-protein kcal/kg in
the study of Chen et al..
b
The amount corresponds to total omega-3 of the diet considering the volume administered (mean) from the third postoperative day.
c
The amount corresponds to fish oil only, not to n-3 PUFAs.
d
Control group received a diet containing n-3 PUFAs. The amount of these fatty acids provided by the formula from the third postoperative day was 1.7 g/day.
e
Estimated content.

95% CI ¼ 0.03, 0.65; p ¼ 0.032; Supplemental Table 1). No hetero-
geneity was observed for all subgroup analyzes.
Meta-analysis for prealbumin showed an important heteroge-
neity in pooled effect sizes. This heterogeneity may be attributed
to: number of patients enrolled in each study; different formulas
containing n-3 PUFAs offered; route of n-3 PUFAs administration;
and report the use or not of antibiotic during the study period.
Interesting, in subgroups without heterogeneity (RCTs with 50
patients; that offered an enriched-diet with n-3 and others
immunonutrients; enterally; and trials that did not report the use
of antibiotics) the combined effect size was significant in favor to
intervention group to increase prealbumin levels (Supplemental
Table 2).
Regarding the effects of n-3 PUFAs on transferrin and total
protein concentrations, both global and subgroups analysis did not
show significant effect sizes, but subgroup analysis suggest that
heterogeneity from RCTs that assessed transferrin can be explained
by the number of participants and by the report the use of antibiotic
(Supplemental Table 3), while in RCTs that assessed total protein,
the heterogeneity can be explained by blinding and random
sequence generation (Supplemental Table 4).
Sensitivity analysis for albumin, transferrin and total protein
meta-analysis showed that removing a single study included in
pooled analysis for each parameter, there are no significant changes
in combined effect sizes. This does not apply to prealbumin meta-
analysis, where after removing the RCTs of Chen et al. [16], Farre-
ras et al. [17] or Klel et al. [19], the effect size loses significance
(Supplemental Fig. 1).
Only two RCTs assessed the effects of n-3 PUFAs on cytokines
[16,24] and on CRP [18,24] (Table 2). All studies were conducted in
surgical patients, but others characteristics (dose, route of admin-
istration, control formula, time and moment of n-3 PUFAs supple-
mentation) were different among them. Meta-analysis was
performed only for IL-6, TNF-a e CRP (Fig. 4). The findings showed a
significant reduction in favor to intervention group for IL-6
(SMD ¼
0.71; 95% CI ¼
1.15,
0.27; p ¼ 0.001) and TNF
(SMD ¼
0.92; 95% CI ¼
1.58,
0.26; p ¼ 0.006), without sig-
nificant heterogeneity.

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
6 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11
Fig. 2. Summary of risk of bias by Cochrane tool. (þ) Low risk; (
) High risk; (?)
Unclear risk.
Adverse effects on n-3 PUFAs intake were not observed or re-
ported in all trials. Unexpected outcomes related to n-3PUFAs
supplementation were observed in one RCT [21]: Marano et al.
demonstrated a lower total counts of CD4þ T-cell in intervention
group than control at the end of the study. Despite this finding, they
did not observe a significant impact on immune response, since
complications and infections after surgery were less observed in
intervention group than control.
4. Discussion
The present meta-analysis from published RCTs provided evi-
dence that n-3 PUFAs supplementation had a significant improving
effect on concentrations of albumin, prealbumin and cytokines
(TNF-a and IL-6) in gastric cancer patients.
Our findings are consistent with other meta-analysis that
demonstrated reduction in pro-inflammatory biomarkers after n-3
PUFAs supplementation in diverse clinical conditions: CRP, IL-6 and
TNF-a in chronic non-autoimmune diseases and non-obese sub-
jects [25]; IL-6, IL-1 and TNF-a in chronic heart failure [26];
thromboxane B2 in subjects with high risk of cardiovascular disease
and leukotriene B4 in unhealthy subjects and in rheumatoid
arthritis [27]; and, CRP in hemodialysis patients [28]. In cancer, is of
our knowledge only one meta-analysis (performed by our group)
assessing the effects of n-3 PUFAs on inflammation in colorectal
cancer. In this study, we demonstrated lower concentrations of IL-6
(global analysis) and CRP (subgroup of studies performed in pa-
tients undergoing chemotherapy) and higher concentration of al-
bumin (global analysis) in patient's group that received n-3 PUFAs
Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
supplementation in comparison with control [11]. Regarding the
action of n-3 PUFAs in improving the concentrations of negative
acute phase proteins, we did not find other meta-analysis that
demonstrated a beneficial effect on these outcomes.
The improving effect of n-3 PUFAs on albumin, prealbumin, IL-6
and TNF-a has a biological basis. The mechanism may involve an
inhibitory action of these PUFA, especially EPA and DHA, on acti-
vation of Nuclear factor kappa B (NF-kB) e pathways mediated by
activation of peroxisome proliferator activated receptor (PPAR)-g
and binding with G-protein coupled cell membrane GPR120 from
n-3 PUFAs intake and incorporation in membrane phospholipids
[29]. NF-kB is one of the most important transcription factor
involved in up-regulation of the genes encoding inflammatory cy-
tokines (e.g. TNF, IL-1 and IL-6), and is activated in a signalling
cascade triggered by extracellular inflammatory stimuli [30]. In
addition, TNF-a, IL-1b and IL-6 play an important role in acute
phase inflammatory response, inducing an increase on hepatic
production of proteins designated as positive acute phase proteins
(e.g. CRP, serum amyloid A and fibrinogen) and decreasing the
production and circulating concentrations of proteins designated as
negative acute phase proteins (e.g. albumin, prealbumin, trans-
ferrin) [31]. Therefore, the findings of our study are theoretically
plausible, because the reduction on TNF-a and IL-6 with n-3 PUFAs
supplementation may be the causes of increase on albumin and
prealbumin concentrations.
Two other reasons that could explain our findings are: 1) n-3
PUFAs reduces oxidative status compared to other fatty acids and
consequently, reduces the NF-kB activation from this pathway [32];
and, 2) n-3 PUFAs, especially EPA, prevents muscle and weight loss,
contributing with maintenance of serum albumin levels in cancer
patients [8,10].
Controversially, conclusions from meta-analysis were not
consistent with those from individual analysis of the included
studies (systematic review synthesis), where few studies showed
significant differences on biomarkers concentrations between
intervention and control groups. The traditional way to perform
statistical analysis for find differences between study groups after
an intervention (based in means comparisons) not provide the real
practical and clinical significance of the finding. For denote with
more visibility the practical and clinical applicability of an inter-
vention have been encouraged the clinical trials makers to calculate
the “effect sizes” of the finding. Even if there is no statistical dif-
ference considering the mean and its variance, a certain treatment
may have better results than the control which may reflect in a
small, but significant effect size when the RCTs are combined in a
meta-analytic model (we observed this fact in meta-analysis for
albumin and prealbumin). Other reason for report effect size in
RCTs, besides provide the practical consequences of the findings for
daily life, it allows researchers to present the magnitude of the
effect in a standardized metric which can be understood regardless
of the scale used to measure the outcome [33].
Results from open-label or single-blind RCTs included in this
study showed a nil effect of n-3 PUFAs on outcomes, while results
from double-blind RCTs demonstrated a significant pooled effect
size (Supplemental Tables 1e4). In the same way, trials with
adequate random sequence generation demonstrated a significant
effect size when combined in meta-analysis, while those with
“unclear” sequence generation, did not show it. These facts enhance
the claim that the findings found in this study are due to n-3 PUFAs
supplementation and not to bias related to the methodological
design e at least, was not due to performance, detection and se-
lection bias. Blinded studies (the ideal is to blind the greatest
number of individuals as possible e participants, researchers and
statisticians) provide the best estimates of an intervention/treat-
ment impact [34]. In addition, an adequate sequence generation
 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11 7

Fig. 3. Forest plot showing standard mean difference (SMD) and 95% CIs for the impact of n-3 PUFAs supplementation on albumin, prealbumin, transferrin and total protein. The p-
value presented in the figure refer to Chi2 test for heterogeneity.

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
8 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11

Fig. 4. Forest plot showing standard mean difference (SMD) and 95% CIs for the impact of n-3 PUFAs supplementation on IL-6, TNF-a and C-reactive protein. The p-value presented
in the figure refer to Chi2 test for heterogeneity.

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11
and concealment of allocation are other essentials approaches that
minimize the risk of bias in RCTs. In the current meta analyses/re-
view, only three described an adequate technique to generate the
randomization sequence and none described concealing in the
allocation. Although not evident in our analysis, some studies
demonstrated that the beneficial effect from an intervention was
exaggerated in accordance with methodological characteristics:
open label studies have a range of 13e17% exaggeration of inter-
vention effect estimates, on average, than double-blind RCTs;
adequate/unclear sequence generation have 5e11% exaggeration
than adequate; and inadequate/unclear allocation concealment
have 7e30% exaggeration than adequate [35e37]. For these
empirical evidences, we alert to researchers for their future studies
be well designed to avoid bias, as well as, provide sufficient
methodological detail (in the article or in trial register) to ensure
reproducibility and judgment of adequacy on processes adopted.
In sensitivity analysis for prealbumin, removing any of the
studies included in this analysis, we observed a significant change
on the pooled effect size. When the RCTs of Chen et al. (2005) [16],
Farreras et al. (2005) [17] or Klek et al. (2005) [19] were removed, the
combined effect lost significance, and when Wei's RCT was removed,
the effect went from a medium size to large, and the heterogeneity
disappeared. This change occurred due mainly for influences of the
trial of Wei et al. in the combined effect size (we can observe in Fig. 3
that Wei's RCT was the only study that presented a lower mean for
prealbumin in intervention than in control group, pulling the
combined effect for the left side of the graph e “intervention re-
duces prealbumin”). We identified two causes that could explain
this discrepant effect for prealbumin from this trial: losses occurred
only in the control group, justified by “incomplete data” (Table 1);
and, of the four studies combined in this meta-analysis, all were
carried out in surgical patients, but the RCT of Wei et al. was the only
that reported the use of antibiotic at post-operative period.
We can cite some reasons that may confuse the conclusion of
our study: 1) use combined of n-3 PUFAs with arginine, glutamine
and antioxidants in most of included RCTs (these immunonutrients
have anti-inflammatory action that could amplify the effects of n-3
PUFAs on inflammation [38]); 2) use of control supplement rich in
n-6 PUFAs in seven RCTs (previous studies have demonstrated that
n-6 PUFAs play an opposing role with n-3 PUFAs in the process of
inflammation that could overestimate the effect size arising n-3
PUFAs supplementation [39]); 3) use of antibiotic in a moment of
the study period (when it was used, the trials did not show sig-
nificant results from n-3 PUFAs supplementation e it appears to be
an effect canceler); and, 4) different anti-cancer treatment applied
(surgical and chemotherapy). We emphasize that the biggest effect
size for n-3 PUFAs effects on albumin was provided by Nemati's RCT
that not used control formula containing n-6 PUFAs (it was the only
RCT with more than one week of n-3 PUFAs supplementation, be-
sides to be the only RCT carried out in patients receiving chemo-
therapy), and the patients enrolled in the control group consumed
significant lower amounts of energy, carbohydrate, fat and protein,
besides the majority of minerals and vitamins during the follow-up,
important facts that confuse the actual effect of the intervention
tested. However, removing this study of the analysis, the effect size
was reduced but remain significant in favor to n-3 PUFAs
supplementation.
Other limitations include: low number of studies assessing the
effects on n-3 PUFAs on cytokines and CRP (despite the large effect
size found for IL-6 and TNF-a meta-analysis, we suggest caution in
its interpretation); some studies did not provide numerical data
(mean and SD) to include in a meta-analysis (we tried to contact the
authors of these articles, but neither returned); all studies pre-
sented risk of bias; and, the publication bias was not assessed due
the low number of included studies.
Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
9
The basal nutritional status of the patient could influence the
concentration of some biomarkers assessed in this meta-analysis
after n-3 PUFAs supplementation. Thus, it would be necessary to
separate the malnourished patients from those well-nourished,
and so, to evaluate the impact of supplementation in each group.
This analysis can not be done because the originals articles not
provided results considering the stratification by nutritional status.
Of the nine studies included in this work, four presented the
number of malnourished patients. We can observe (Table 1) that
the proportion of malnutrition in control and intervention groups
was similar in each of these four studies, what suggest that the
results found in this work may not be influenced by nutritional
status of the patients.
The nutrients consumption could be another potential
confounder of the results found in this work. But, eight of nine
studies included were performed in a hospital setting, where the
dietetic ingestion is standardized.
Strength points also can be observed in the present study:
extensive literature search; we estimated the numerical data (mean
and SD) in one article [18] from a validated equation (when
possible) to avoid losing it in the meta-analysis; subgroups analysis
for methodological characteristics that could explain heterogeneity
among included studies; sensitivity analysis to assess the effect of
individual studies on pooled effect size, that demonstrated that
none of the studies when removed altered the conclusion of the
analysis (exception for prealbumin e discussed earlier).
In conclusion, the supplementation of n-3 PUFAs in gastric
cancer has a potential effect on increasing the levels of albumin and
prealbumin, and decreasing the pro-inflammatory cytokines IL-6
and TNF-a. In accordance with the studies published until the
moment, it seems prudent to advise the use of n-3 PUFAs from fish
oil isolated or from immune-enriched diet in replacement a for-
mula containing large amounts of n-6 PUFAS at post-operative
period for patients submitted to gastric resection surgery for
improve these parameters. It appears that offering n-3 PUFA orally
or enterally could result in better concentrations of some
biochemical parameters evaluated than parenterally, however,
there is a low evidence to support this claim. Nevertheless, other
well-designed clinical trials need to be conducted, especially in
chemotherapy patients, to confirm our findings.
5. Directions for future research
Based on limitations of the studies included in this systematic
review and meta-analysis we suggest that researchers take careful
in adopt and describe with more details several procedures to
improve the methodological quality and ensure the reproducibility
of studies regarding the use of n-3 PUFAs as intervention in surgical
patients:
a) Describe the method used to generate the random sequence;
b) Describe the method used to ensure the allocation
concealment;
c) Described the procedures of blinding (if applicable);
d) Implement a control group that not received any type of fatty
acids, specially n-6 PUFAs, or standardize the profile of lipids
on both intervention and control formulas where the n-3
PUFAs amounts is the unique difference;
e) Collect and describe data about the surgical procedures such
as surgical time and blood loss during operation;
f) Describe the use of any medicine used before, during, and
after the surgery, especially anti-inflammatory and antibi-
otics drugs;
g) Implement a technique to ensure the compliance in the n-3
PUFAs consumption such as to determine the fatty acids
10 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11
profile of the plasma, especially in studies conducted in an
open setting;
h) Stratify the analysis by grade of tumor stage, nutritional
status and localization of the tumor (if performed with
several types of cancer);
i) In enterally or parenterally studies that use a control formula,
is necessary to standardize the quantities of antioxidants;
j) For include the study in meta-analysis, it is essential describe
and express all numerical data in mean and standard devi-
ation (or standard error) from basal and final moments.
Avoid mean differences, interquartile range and confidence
interval;
k) Describe all clinical and demographic variables as possible at
the baseline, and compare them between the study groups
(apply appropriate statistical tests);
l) Avoid not show data. Present all data of outcomes even in
supplementary material;
m) Take careful with statistic test applied. If study groups are
comparable, use appropriate and simple statistic tests. For all
outcomes or characteristics variables of participants, when
comparing groups, described the p-value of the difference
test.
Conflict of interest
The authors declare that they have no conflict of interest.
Author's contributions
Michel Carlos Mocellin: study concept and design; search in
databases; study selection; data extraction, analysis and interpre-
tation; quality assessment; statistical analysis; drafting of the
manuscript.
Ricardo Fernandes: quality assessment; critical revision of the
manuscript for important intellectual content.
Thayz Rodrigues Chagas: data extraction; critical revision of the
manuscript for important intellectual content.
Erasmo Benicio Santos de Moraes Trindade: critical revision of
the manuscript for important intellectual content; study
supervision.
Acknowledgement
We are grateful to the Graduate Program of Nutrition at Federal
University of Santa Catarina and Coordination for the Improvement
of Higher Education Personnel (CAPES e Brazil) for the scholarship
granted to first three authors.
Funding sources
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.clnu.2017.05.008.
References
[1] Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation.
Nature 2008;454:436e44.
[2] Fox JG, Wang TC. Inflammation, atrophy, and gastric cancer. J Clin Investig
2007;117:60e9.
Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008
[3] Grivennikov SI, Karin M. Inflammatory cytokines in cancer: tumour necrosis
factor and interleukin 6 take the stage. Ann Rheum Dis 2011;70(Suppl. 1):
i104e8.
[4] Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet
2001;357:539e45.
[5] Sethi G, Shanmugam MK, Ramachandran L, Kumar AP, Tergaonkar V. Multi-
faceted link between cancer and inflammation. Biosci Rep 2012;32:1e15.
[6] Tsoli M, Robertson G. Cancer cachexia: malignant inflammation, tumorkines,
and metabolic mayhem. Trends Endocrinol Metab 2013;24:174e83.
[7] D'Eliseo D, Velotti F. Omega-3 fatty acids and cancer cell cytotoxicity: impli-
cations for multi-targeted cancer therapy. J Clin Med 2016;5.
[8] Pappalardo G, Almeida A, Ravasco P. Eicosapentaenoic acid in cancer improves
body composition and modulates metabolism. Nutr (Burbank, Los Angeles
County, Calif) 2015;31:549e55.
[9] Wendel M, Heller AR. Anticancer actions of omega-3 fatty acidsecurrent state
and future perspectives. Anticancer Agents Med Chem 2009;9:457e70.
[10] Murphy RA, Mourtzakis M, Mazurak VC. n-3 polyunsaturated fatty acids: the
potential role for supplementation in cancer. Curr Opin Clin Nutr Metab care
2012;15:246e51.
[11] Mocellin MC, Camargo CQ, Nunes EA, Fiates GM, Trindade EB. A systematic
review and meta-analysis of the n-3 polyunsaturated fatty acids effects on
inflammatory markers in colorectal cancer. Clin Nutr (Edinburgh, Scotland)
2016;35:359e69.
[12] Higgins JPT, Green S. Cochrane Handbook for systematic reviews of in-
terventions. The Cochrane Collaboration; 2011.
[13] Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The
PRISMA statement for reporting systematic reviews and meta-analyses of
studies that evaluate healthcare interventions: explanation and elaboration.
Bmj 2009;339:b2700.
[14] Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the
median, range, and the size of a sample. BMC Med Res Methodol 2005;5:13.
[15] Rosenthal JA. Qualitative descriptors of strength of association and effect size.
J Soc Serv Res 1996;21:37e59.
[16] Chen DW, Fei ZW, Zhang YC, Ou JM, Xu J. Role of enteral immunonutrition in
patients with gastric carcinoma undergoing major surgery. Asian J Surg
2005;28:121e4.
[17] Farreras N, Artigas V, Cardona D, Rius X, Trias M, Gonzalez JA. Effect of early
postoperative enteral immunonutrition on wound healing in patients un-
dergoing surgery for gastric cancer. Clin Nutr (Edinburgh, Scotland) 2005;24:
55e65.
[18] Fujitani K, Tsujinaka T, Fujita J, Miyashiro I, Imamura H, Kimura Y, et al.
Prospective randomized trial of preoperative enteral immunonutrition fol-
lowed by elective total gastrectomy for gastric cancer. Br J Surg 2012;99:
621e9.
[19] Kłek S, Kulig J, Szczepanik AM, Jedrys J, Kołodziejczyk P. The clinical value of
parenteral immunonutrition in surgical patients. Acta Chir Belg 2005;105:
175e9.
[20] Makay O, Kaya T, Firat O, Sozbilen M, Caliskan C, Gezer G, et al. Omega-3 fatty
acids have no impact on serum lactate levels after major gastric cancer sur-
gery. J Parenter Enter Nutr 2011;35:488e92.
[21] Marano L, Porfidia R, Pezzella M, Grassia M, Petrillo M, Esposito G, et al.
Clinical and immunological impact of early postoperative enteral immuno-
nutrition after total gastrectomy in gastric cancer patients: a prospective
randomized study. Ann Surg Oncol 2013;20:3912e8.
[22] Nemati A, Nachvak SM, Djafarian K, Faizi-Khankandi I. Effect of omega-3 fatty
acid supplementation on nutritional status in patients with gastric cancer
during chemotherapy. J Nutr Sci Diet 2015;1:7.
[23] Okamoto Y, Okano K, Izuishi K, Usuki H, Wakabayashi H, Suzuki Y. Attenua-
tion of the systemic inflammatory response and infectious complications after
gastrectomy with preoperative oral arginine and u-3 fatty acids supple-
mented immunonutrition. World J Surg 2009;33:1815e21.
[24] Wei Z, Wang W, Chen J, Yang D, Yan R, Cai Q. A prospective, randomized,
controlled study of omega-3 fish oil fat emulsion-based parenteral nutri-
tion for patients following surgical resection of gastric tumors. Nutr J
2014;13:25.
[25] Li K, Huang T, Zheng J, Wu K, Li D. Effect of marine-derived n-3 poly-
unsaturated fatty acids on C-reactive protein, interleukin 6 and tumor ne-
crosis factor alpha: a meta-analysis. PLoS One 2014;9:e88103.
[26] Xin W, Wei W, Li X. Effects of fish oil supplementation on inflammatory
markers in chronic heart failure: a meta-analysis of randomized controlled
trials. BMC Cardiovasc Disord 2012;12:1e11.
[27] Jiang J, Li K, Wang F, Yang B, Fu Y, Zheng J, et al. Effect of marine-derived n-3
polyunsaturated fatty acids on major eicosanoids: a systematic review and
meta-analysis from 18 randomized controlled trials. PLoS One 2016;11:
e0147351.
[28] He L, Li MS, Lin M, Zhao TY, Gao P. Effect of fish oil supplement in maintenance
hemodialysis patients: a systematic review and meta-analysis of published
randomized controlled trials. Eur J Clin Pharmacol 2016;72:129e39.
[29] Calder PC. Omega-3 polyunsaturated fatty acids and inflammatory processes:
nutrition or pharmacology? Br J Clin Pharmacol 2013;75:645e62.
[30] Sigal LH. Basic science for the clinician 39: NF-kappaB-function, activation,
control, and consequences. J Clin Rheum: Pract Rep Rheum Musculoskelet Dis
2006;12:207e11.
[31] Moshage H. Cytokines and the hepatic acute phase response. J Pathol
1997;181:257e66.
 M.C. Mocellin et al. / Clinical Nutrition xxx (2017) 1e11 11

[32] Richard D, Kefi K, Barbe U, Bausero P, Visioli F. Polyunsaturated fatty acids as
antioxidants. Pharmacol Res 2008;57:451e5.
[33] Lakens D. Calculating and reporting effect sizes to facilitate cumulative sci-
ence: a practical primer for t-tests and ANOVAs. Front Psychol 2013;4:863.
[34] Karanicolas PJ, Farrokhyar F, Bhandari M. Blinding: who, what, when, why,
how? Can J Surg 2010;53:345e8.
[35] Page MJ, Higgins JP, Clayton G, Sterne JA, Hrobjartsson A, Savovic J. Empirical
evidence of study design biases in randomized trials: systematic review of
meta-epidemiological studies. PLoS One 2016;11:e0159267.
[36] Savovic J, Jones H, Altman D, Harris R, Juni P, Pildal J, et al. Influence of re-
ported study design characteristics on intervention effect estimates from
randomised controlled trials: combined analysis of meta-epidemiological
studies. Health Technol Assess (Winchester, England) 2012;16:1e82.
[37] Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence of bias. Di-
mensions of methodological quality associated with estimates of treatment
effects in controlled trials. Jama 1995;273:408e12.
[38] Pierre JF, Heneghan AF, Lawson CM, Wischmeyer PE, Kozar RA, Kudsk KA.
Pharmaconutrition review: physiological mechanisms. JPEN J Parenter Enter
Nutr 2013;37:51Se65S.
[39] Schmitz G, Ecker J. The opposing effects of n-3 and n-6 fatty acids. Prog Lipid
Res 2008;47:147e55.

Please cite this article in press as: Mocellin MC, et al., A meta-analysis of n-3 polyunsaturated fatty acids effects on circulating acute-phase
protein and cytokines in gastric cancer, Clinical Nutrition (2017), http://dx.doi.org/10.1016/j.clnu.2017.05.008