Omega-3 Fatty Acids Therapy in Children with Nonalcoholic Fatty Liver

Disease: A Randomized Controlled Trial

Wojciech Janczyk, MD1, Dariusz Lebensztejn, MD, PhD2, Aldona Wierzbicka-Rucinska3, Artur Mazur, MD, PhD4,
Joanna Neuhoff-Murawska1, Pawe1 Matusik, MD5, and Piotr Socha, MD, PhD1

Objective To evaluate the efficacy and safety of omega-3 fatty acid supplementation in children with nonalcoholic
fatty liver disease (NAFLD).
Study design Overweight/obese children with NAFLD (n = 76; median age, 13 years; IQR, 11.1-15.2 years) were
eligible to participate in the study. The diagnosis of NAFLD was based on elevated alanine aminotransferase (ALT) to
$30% of the upper limit of normal (ULN) and liver hyperechogenicity on ultrasound. Patients were randomized to
receive omega-3 fatty acids (docosahexaenoic acid and eicosapentaenoic acid, 450-1300 mg/day) or placebo
(omega-6 sunflower oil). The primary outcome was the number of patients who demonstrated decreased ALT
activity by $0.3 times the ULN. Secondary outcomes included alterations in liver function tests, liver hyperechoge-
nicity, insulin resistance, and other metabolic markers after 6 months of intervention.
Results Out of 76 enrolled patients, 64 completed the trial and were analyzed. After 6 months, we found no sig-
nificant differences between the omega-3 and placebo groups in the number of patients with decreased ALT by
$0.3 times the ULN (24 vs 23) or in median (IQR) ALT activity (48.5 [31-62] U/L vs 39 [27-55] U/L), liver hyperecho-
genicity, insulin resistance, or serum lipid levels. However, patients in the omega-3 group had lower levels of
aspartate aminotransferase (28 [25-36] U/L vs 39 [27-55] U/L; P = .04) and gamma-glutamyl transpeptidase (26
[17.5-36.5] U/L vs 35 [22-52] U/L; P = .04), and significantly higher levels of adiponectin.
Conclusion Omega-3 fatty acid supplementation did not increase the number of patients with decreased ALT
levels and it did not affect liver steatosis on ultrasound, but it improved aspartate aminotransferase and gamma-
glutamyl transpeptidase levels in children with NAFLD compared with placebo. (J Pediatr 2015;-:---).
Trial registration Registered with ClinicalTrials.gov: NCT01547910.

See editorial, p 

T
he diagnosis of nonalcoholic fatty liver disease (NAFLD), a common liver abnormality in children aged 2-19 years,1 is
based on histological criteria of liver biopsy specimens.2 Liver disease may progress to cirrhosis. The principle of treat-
ment is amelioration of risk factors, such as obesity/overweight and insulin resistance. Previous systematic reviews have
addressed the effectiveness of pharmacotherapy (including dietary supplements) in treating NAFLD.3 An analysis of 15 ran-
domized controlled trials identified metformin, YHK, and carnitine in high doses (3 g/day) as promising agents, demonstrating
a significant effect on normalizing alanine aminotransferase (ALT) levels. In a meta-analysis of randomized controlled trials for
the treatment of NAFLD, Musso et al4 found that thiazolidinediones, such as pioglitazone or rosiglitazone, could improve
insulin sensitivity, steatosis, and inflammation of the liver. However, the use of thiazolidinediones may be associated with
excessive weight gain and hepatotoxicity, and the duration of therapeutic effect is poor.
Dietary supplements, such as polyunsaturated fatty acids (PUFAs), appear
promising. In a randomized controlled trial, Nobili et al5 investigated the effect
of supplementation with docosahexanoic acid (DHA) on liver fat content and
liver tests in children with NAFLD. Two groups received either DHA (250 or 1
From the Department of Gastroenterology, Hepatology
and Nutrition Disorders, The Childrens Memorial Health
500 mg) or placebo daily for 6 months. DHA in both doses was found to reduce 2
Institute, Warsaw, Poland; Department of Pediatrics,
Gastroenterology, and Allergology, Medical University of
liver fat (as assessed by abdominal ultrasound) and insulin resistance; however, Bialystok, Bialystok, Poland; Department of 3

no effect was seen on ALT or body weight. Biochemistry and Radioimmunology, The Childrens
Memorial Health Institute, Warsaw, Poland; Medical 4

Faculty, University of Rzeszow, Rzeszow, Poland; and

5
Department of Pediatrics, Pediatric Endocrinology and
Diabetes, Medical University of Silesia, Katowice, Poland
Supported by the Polish Ministry of Science and Higher
Education (3180/B/P01/2007/33). Omega-3 and placebo
ALT Alanine aminotransferase GGTP Gamma-glutamyl transpeptidase capsules were manufactured and blinded by Nepentes
S.A. and Hasco-Lek Polska S.A. The electronic case
AST Aspartate aminotransferase LC Long chain report form and randomization process was conducted
BMI Body mass index NAFLD Nonalcoholic fatty liver disease in cooperation with Medical Network, Poland. The
authors declare no conflicts of interest.
DHA Docosahexanoic acid PUFA Polyunsaturated fatty acid
eCRF Electronic case report form ULN Upper limit of normal 0022-3476/$ - see front matter. Copyright 2015 Elsevier Inc.
EPA Eicosapentaenoic acid WC Waist circumference All rights reserved.
http://dx.doi.org/10.1016/j.jpeds.2015.01.056

1
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. -, No. -

In the present randomized controlled trial, we evaluated In addition, all patients were regularly instructed by an
the efficacy and safety of omega-3 fatty acid supplementation experienced dietician in each center to comply with an indi-
in children with NAFLD. vidually prescribed diet, which, in combination with
increased physical activity, was aimed at producing a slow
Methods reduction in body weight (approximately 0.5 kg/week).
Extensive dietary data were collected using a validated food
This multicenter, randomized, double-blind, placebo- frequency questionnaire. The level of physical activity and
controlled clinical trial was conducted in 4 Polish pediatric sedentary lifestyle were assessed using the modified Interna-
departments. The main aim of this study was to assess the ef- tional Physical Activity Questionnaire.
ficacy of omega-3 long chain (LC)-PUFAs in children with The study protocol was approved by the Childrens
NAFLD. Full details of the trial protocol, study design, and Memorial Health Institutes Bioethical Committee. Written
rationale have been described previously.6 informed consent was obtained from all participants. The
The inclusion criteria were: (1) age >5 and <19 years; Consolidated Standards of Reporting Trials flowchart is pre-
(2) overweight or obesity (according to International sented in the Figure (available at www.jpeds.com).
Obesity Task Force body mass index (BMI) charts7); (3)
ALT activity $1.3 times the upper limit of normal Randomization
(ULN); (4) presence of hyperechogenic liver on ultrasound The list of random treatment assignments was generated us-
or liver histology consistent with NAFLD/nonalcoholic stea- ing StatsDirect version 2.7.9 (StatsDirect, Altrincham,
tohepatitis; and (5) written consent obtained from the pa- United Kingdom). Randomization was fixed and balanced
tient and/or a legal representative. The exclusion criteria in blocks of 4 individuals, and stratified according to center,
were: (1) age <5 or >19 years; (2) history of significant to 2 arms: omega-3 and placebo. The treatment allocation
alcohol consumption; (3) any known pathological condi- was done using a central randomization. Investigators sent
tions affecting the liver eg, hepatitis B or C virus infection, their randomization requests by fax to the central randomi-
chronic or acute liver failure, cholestasis, metabolic diseases, zation center responsible for the randomization process. In
such as a1-antitripsin deficiency, Wilson disease, diabetes return, they received a random number that was subse-
mellitus, and hypothyroidism; (4) treatment with vitamin quently allocated to the choice of one of the products (desig-
E, statins, antihypertensives, ursodeoxycholic acid, probiot- nated 100 or 101). Patient data were entered and organized in
ics, or metformin within 3 months before randomization; an electronic case report form (eCRF). Both the eCRF and the
(5) history of parenteral nutrition; and (6) unlikely to coop- randomization process were monitored by the external clin-
erate with the study regime. The patients recruited for the ical research organization.
study received dietary counsling in their previous therapy,
which proved ineffective. Follow-Up Visits
Study visits were scheduled at 12 weeks and 24 weeks after
Treatment Groups randomization. At the baseline visit (week 0) and final visit
The eligible patients were randomized into blocks of 4 indi- (week 24), patients underwent physical examination,
viduals, stratified by center, to either fish oil containing including anthropometric measurements (height, weight,
omega-3 LC-PUFA (DHA and eicosapentaenoic acid [EPA] and waist circumference [WC]); heart rate and blood pres-
in a 3:2 proportion [450-1300 mg/day]) or placebo (sun- sure; laboratory tests, including hematology, ALT, aspartate
flower oil, containing omega-6 LC-PUFA) for 24 weeks. aminotransferase (AST), gamma-glutamyl transpeptidase
The study drug and placebo were administered orally twice (GGTP), bilirubin, international normalized ratio, fasting
a day in the same dose and formulation of brown, oval- glucose and insulin; fasting lipid profile; adipokines and cy-
shaped capsules. The dose of omega-3 was dependent on tokines; abdominal ultrasound; dietary assessment (using
patient weight (Table I). the food frequency questionnaire); and physical activity
The omega-3 fatty acids, obtained from marine algae, and assessment (using the International Physical Activity Ques-
placebo were supplied and blinded by Hasco-Lek (Wroclaw, tionnaire).
Poland) from the composition of Incromega DHA 500 TG The visit at week 12 included a routine physical examina-
SR (Croda Poland, Krakow, Poland) and Incromega EPA tion with vital signs, simple laboratory tests (ALT, AST,
500 TG SR (Croda Poland). GGTP, bilirubin, international normalized ratio, fasting
glucose and insulin) and anthropometry, as well as dietary
and physical activity assessments.6
Liver biopsy evaluation was not mandatory in this study;
Table I. Dosage of omega-3 LC-PUFAs however, it was proposed for clinical reasons to selected
Body weight, Approximate patients who presented with elevated ALT persisting for
kg total omega-3, mg/d DHA, mg/d EPA, mg/d longer than 6 months during weight reduction therapy.
<40 450 267 177.5 Any potential adverse events were reported at each visit
40-60 900 534 355
>60 1300 800 532.5
throughout the study. Data for all investigations were entered
into the eCRF and stored there anonymously.
2 Janczyk et al
- 2015 ORIGINAL ARTICLES

Primary and Secondary Outcomes terms of final ALT activity, reduction of steatosis on
The primary study endpoint was the number of patients in ultrasound, fasting glucose level, fasting insulin level,
whom the ALT level decreased by $0.3 times the ULN after homeostasis model assessment for insulin resistance, BMI z-
6 months of treatment with omega-3 compared with the pla- score, or WC z-score (Tables III and IV). AST and GGTP
cebo group. Secondary outcomes included the following vari- levels were significantly lower and adiponectin level was
ables compared in the omega-3 and placebo groups: ALT, markedly higher in the omega-3 group. There was no
AST, GGTP, and other liver function tests; insulin resistance difference in the degree of steatosis between the 2 groups;
markers, cholesterol and lipid profiles; liver steatosis on the omega-3 group included 12 children with grade 0, 10
ultrasound; basic hematology and biochemical laboratory children with grade 1, and 8 children with grade 2, whereas
tests; selected adipokines; anthropometric measurements; the control group comprised 8 children with grade 0, 11
and caloric intake and dietary content. Further imaging and children with grade 1, 6 children with grade 2, and 3
biochemical results were analyzed in subgroups of patients. children with grade 3. Energy and nutrient intake did not
differ between the 2 groups.
Statistical Analyses BMI and WC z-scores did not differ significantly between the
We estimated that a sample of 64 patients would be suffi- 2 groups; however, BMI and WC z-scores were significantly
cient to reveal a difference in treatment effect of 30% lower only in the omega-3 group in the secondary post hoc
(45% in the experimental group vs 15% in the control analysis of baseline vs end-of-study results (Table V;
group), considering a = 0.05 and a power (b) of 80%. available at www.jpeds.com). In both groups, energy intake
Because many of the studied variables did not show a decreased significantly during the course of the study.
normal distribution, the results are presented as absolute We did not observe any significant between-group differ-
and percent values or as median with IQR. The final results ences in the studied variables at the 12-week visit (data not
were compared using the c2 test or Mann-Whitney U test. shown).
Changes in variables during the course of the study were We also compared patients who lost $5% of their body
tested with the Wilcoxon matched-pairs test. Statistica for weight with the others irrespective of group assignment.
Windows version 10 (Statsoft, Tulsa, Oklahoma) was used ALT, AST, and GGTP levels were significantly decreased in
for descriptive statistics, the Mann-Whitney U test, and these patients in both groups (data not shown).
the Wilcoxon matched-pairs test; the c2 test was performed Compliance with therapy was controlled based on the
using StatsDirect version 2.7.9. number of capsules received and given back by the patients.
All patients who completed the study fulfilled the protocol
Results criteria for compliance, and the final compliance rate was
95.5% (IQR, 91.2%-99.7%).
We recruited 76 overweight/obese (54 overweight, 22 obese) The only adverse event recorded was mild abdominal
children (65 boys, 11 girls) with NAFLD with a median age of discomfort in 1 patient in the omega-3 group and in 1 patient
13 years (IQR, 11.1-15.2 years) between July 2008 and April in the placebo group.
2011. The patients presented with moderately increased ALT,
AST, GGTP, and homeostasis model assessment for insulin Discussion
resistance values; however, there were no significant baseline
differences between omega-3 and placebo groups. Baseline In children with NAFLD, omega-3 therapy did not have a signif-
anthropometric and laboratory data are presented in icant effect on the primary outcome measure, a decrease in ALT
Table II (available at www.jpeds.com). level. AST and GGTP levels were significantly lower and serum
All children presented with features of fatty liver on ultra- adiponectin levels were higher in the omega-3 group
sound evaluated according to the criteria of Saverymuttu (Table III). Keep in mind, however, that all of our patients
et al.8 The omega-3 group included 18 children with grade presented with mild fatty liver disease and only moderately
1, 19 children with grade 2, and 2 children with grade 3; elevated liver function test values at baseline (Table II). Patie-
the placebo group included 19 children with grade 1, 17 chil- nts in both groups showed improved liver tests and caloric
dren with grade 2, and 1 child with grade 3. intake, which can be explained by the effect of participation in
Among the 76 patients enrolled into the study, 64 finished the study, with regular visits and dietary counseling.
the study according to the protocol and were available for Patients with NAFLD typically present with mildly or
analysis. We did not perform an intention-to treat analysis, moderately elevated ALT levels, but the exact cutoff points
because data on 12 patients were not available for analysis have not been defined.9,10 Ultrasound is considered the
of outcomes (Figure). Final anthropometric, laboratory, first-line imaging technique for revealing fatty liver and is
and dietary data are presented in Table III. widely available, although it has been shown to have lower
For the primary outcome, the number of patients with a sensitivity when the liver fat content is <30%.11,12 To increase
decrease in ALT levels by $0.3 times the ULN after 6 months the specificity for diagnosing NAFLD, we used several inclu-
of therapy was similar in the 2 groups, and no significant dif- sion criteria of proximate markers of NAFLD: 1.3 times the
ference was seen (Table IV). For the secondary outcomes, ULN of ALT, liver steatosis on ultrasound, and the presence
omega-3 fatty acid therapy was not superior to placebo in of overweight or obesity.
Omega-3 Fatty Acids Therapy in Children with Nonalcoholic Fatty Liver Disease: A Randomized Controlled Trial 3
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. -, No. -

Table III. End-of-treatment characteristics after 24 weeks

Characteristics Placebo (n = 34) Omega-3 (n = 30) P value, placebo vs omega-3
Anthropometric measurements
BMI z-score 2.4 (1.9-3.4) 2.4 (1.4-3.2) .40
WC z-score 2.3 (2.2-2.6) 2.4 (2-2.6) .83
Hematology
Erythrocytes 4.9 (4.7-5.1) 5.0 (4.7-5.1) .52
Leukocytes 7.1 (6.2-8.2) 7.1 (6-8.3) .55
Hemoglobin 14 (13.5-14.7) 14.2 (13.5-14.8) .81
Platelets 252 (219-318) 262.5 (214.5-299) .94
Serum biochemistry
ALT, U/L 53.5 (39-99) 48.5 (31-62) .13
AST, U/L 39.0 (27-55) 28.0 (25-36) .04
GGTP, U/L 35.0 (22-52) 26.0 (17.5-36.5) .04
Total bilirubin, mg/dL 0.8 (0.5-1.3) 0.6 (0.5-0.7) .20
INR 1.0 (1.0-1.1) 1.0 (1-1.1) .45
Lipids
Triglycerides, mg/dL 97.5 (72.5-118.5) 91.0 (72-120) .74
Cholesterol, mg/dL 167.5 (159-193.5) 183.0 (149-202.5) .53
HDL-C, mg/dL 39.5 (35.5-44.5) 44.0 (36.5-48.5) .22
LDL-C, mg/dL 109.5 (94-136) 122.0 (90.5-144.5) .59
VLDL-C, mg/dL 19.0 (14-22.5) 17.5 (14-22.5) .79
Metabolic factors
Fasting serum glucose, mg/dL 85 (79-89) 85.0 (79-91) .72
Fasting serum insulin, mU/mL 18.2 (12.9-22.4) 16.4 (12.9-26.2) .84
HOMA-IR 3.8 (2.9-4.9) 3.70 (2.6-5.8) .74
High-sensitivity C-reactive protein, mg/mL 0.3 (0.3-0.5) 0.34 (0.3-0.4) .97
Oxy-LDL, mg/mL 182.0 (124.5-249) 163.0 (136-267) .81
Adiponectin, mg/mL 12.0 (5.3-28.3) 19.8 (7.7-29.9) .02
Leptin, mg/mL 5.3 (1.84-8.0) 7.1 (5.4-10.6) .34
Dietary record*
Energy, kcal 1654.3 (1434-2272.2) 1533.8 (1283.3-1999.6) NS
Energy/kg 21.9 (16.7-32.8) 20.1 (17.6-30) .84
% energy from proteins 16.9 (15-19.2) 16.5 (13.9-18.5) .67
% energy from carbohydrates 51.2 (44.5-55.1) 52.9 (48.7-56.1) .21
% energy from lipids 32.4 (26.1-37.7) 29.9 (26.9-33.5) .62

HOMA-IR, homeostasis model assessment for insulin resistance; INR, international normalized ratio; NS, not significant.
Values are expressed as median (IQR) unless specified otherwise.
*Obtained from the food frequency questionnaire.

A wide range of omega-3 doses (0.8-13.7 g/day) has been dren aged 6-19 years, and thus the anthropometric variables
tested in clinical trials in patients with NAFLD; however, varied greatly. In previous studies, omega-3 was supple-
the optimal dose and content of omega-3 LC-PUFA formulas mented for various indications in a weight-dependent
for treating NAFLD and other indications remain un- manner (per range and per kilogram).15,16 The duration of
known.5,13 Our omega-3 formula combined the potential treatment was similar to that in other trials using supple-
positive metabolic effects of EPA and the anti- ments such as omega-3 PUFAs to provide reliable therapeutic
inflammatory action of DHA. The ratio of EPA to DHA results and satisfactory compliance.3,14
(3:2) and the dose were determined based on a review of Although we failed to demonstrate an effect of omega-3
the results from available open-label studies.14 Furthermore, LC-PUFA therapy on the primary study outcome, we found
we used weight-dependent dosing of omega-3 and placebo, significant effects on the secondary outcomes. GGTP level,
which may be controversial, but the participants were chil- which was decreased in the omega-3 group, has been

Table IV. Selected primary and secondary outcomes

Outcomes Placebo (n = 34) Omega-3 (n = 30) Total (n = 64) P value, placebo vs omega-3
Primary outcome
ALT reduction by $0.3 times ULN, n (%) 23 (67.7) 24 (80) 47 (73.4) .26
Secondary outcomes
ALT normalization, n (%) 3 (8.8) 5 (16.7) 8 (12.5) .28*
Improvement in insulin resistance, n (%) 18 (52.3) 17 (56.7) 35 (54.7) .76
Absence of liver steatosis on ultrasound, n (%) 8 (23.5) 12 (40) 20 (31.3) .16
Improved liver steatosis on ultrasound, n (%) 10 (29.4) 12 (40) 22 (34.4) .37
Weight reduction $5%, n (%) 7 (20.6) 5 (16.7) 12 (18.8) .69
BMI reduction $5%, n (%) 5 (14.7) 12 (40) 17 (26.6) .02
*The Fisher exact test was used instead of the c2 test, because the expected values were too small.
According to the Saverymuttu scoring system.8

4 Janczyk et al
- 2015 ORIGINAL ARTICLES

proposed as a marker of metabolic syndrome and a predictor The negative results regarding the primary outcomes in
of endothelial dysfunction and atherosclerosis in patients our study do not allow us to formulate recommendations
with NAFLD.17,18 Serum GGTP level also has been suggested regarding the use omega-3 LC-PUFA supplements to treat
to be a sensitive measure for the presence of NAFLD19 and a NAFLD in children. Omega-3 fatty acids may be considered
component of low-grade subclinical inflammation.20 There- useful in patients with NAFLD to prevent other complica-
fore, a positive effect of omega-3 LC-PUFA therapy may be tions, such as cardiovascular disease, but this application
considered a prognostic factor of cardiovascular disease in awaits confirmation in further studies. n
patients with NAFLD.
Omega-3 LC-PUFA therapy did not influence insulin The following investigators assisted with the trial: Childrens Memorial
resistance, a correlate of liver injury in NAFLD. In the liver, Health Institute, Warsaw, Poland: Piotr Socha, MD, PhD, Wojciech
adiponectin, through its anti-inflammatory action, can Janczyk, MD, Aldona Wierzbicka, Joanna Neuhoff-Murawska, Katar-
induce hepatic fatty acid oxidation, inhibit fatty acid synthe- zyna Szott, Elzbieta Dzik, MD, Elzbieta Jurkiewicz, Prof, Roman Janas,
Prof; Medical University of Bialystok, Poland: Dariusz Lebensztejn,
sis, and reduce TNF-a synthesis, thereby protecting against MD, PhD, Eugeniusz Tarasow MD, PhD; Medical University of Sile-
the progression of NAFLD to nonalcoholic steatohepatitis.21 sia, Katowice: Pawel Matusik, MD, PhD; and University of Rzeszow,
Our results suggest that prolonged supplementation with Poland: Artur Mazur, MD, PhD.
omega-3 LC-PUFAs may increase adiponectin levels.
In this study, BMI z-scores and WC z-scores did not differ Submitted for publication Aug 21, 2014; last revision received Dec 9, 2014;
significantly between the groups (Table III). However, BMI accepted Jan 30, 2015.

and WC z-scores were significantly reduced in the omega-3 Reprint requests: Piotr Socha, MD, PhD, Department Gastroenterology,
Hepatology, and Nutrition Disorders, The Childrens Memorial Health Institute,
group when comparing baseline data with end-of-study Al Dzieci Polskich 20, 04-730 Warsaw, Poland. E-mail: [email protected]
results (Table V).
To assess the potential effect of weight reduction on liver
tests, we compared those patients who lost $5% of body References
weight (defined in many studies as the minimum weight
1. Schwimmer JB, Deutsch R, Kahen T, Lavine JE, Stanley C, Behling C.
loss required to improve steatosis and metabolic syndrome) Prevalence of fatty liver in children and adolescents. Pediatrics 2006;
with all of the other patients, irrespective of study group. 118:1388-93.
Liver test results decreased significantly in both groups, 2. Vajro P, Lenta S, Socha P, Dhawan A, McKiernan P, Baumann U, et al.
which may indicate that the improvement in liver tests was Diagnosis of nonalcoholic fatty liver disease in children and adolescents:
position paper of the ESPGHAN Hepatology Committee. J Pediatr Gas-
not mediated by weight loss alone (data not shown). Ran-
troenterol Nutr 2012;54:700-13.
domized controlled trials,22,23 including pediatric patients 3. Socha P, Horvath A, Vajro P, Dziechciarz P, Dhawan A, Szajewska H.
with NAFLD,5 have produced no convincing evidence that Systematic review: pharmacological interventions for non-alcoholic fatty
omega-3 LC-PUFA therapy can cause weight reduction. liver disease in adults and in children. J Pediatr Gastroenterol Nutr 2009;
The effects of omega-3 LC-PUFA supplementation in our 48:587-96.
study confirm the results of previous studies. A systematic re- 4. Musso G, Gambino R, Cassader M, Pagano G. A meta-analysis of ran-
domized trials for the treatment of nonalcoholic fatty liver disease. Hep-
view and meta-analysis of omega-3 LC-PUFA supplementa- atology 2010;52:79-104.
tion in adults with NAFLD14 included 9 studies (5 5. Nobili V, Bedogni G, Alisi A, Pietrobattista A, Rise P, Galli C, et al.
randomized controlled trials) involving a total of 355 individ- Docosahexaenoic acid supplementation decreases liver fat content in
uals. The duration of therapy ranged from 8 weeks to children with non-alcoholic fatty liver disease: double-blind randomised
controlled clinical trial. Arch Dis Child 2011;96:350-3.
12 months. LC-PUFA was given in doses from 0.25 g/day to
6. Janczyk W, Socha P, Lebensztejn D, Wierzbicka A, Mazur A, Neuhoff-
13.7 g/day. Omega-3 fatty acids significantly decreased AST Murawska J, et al. Omega-3 fatty acids for treatment of non-alcoholic
level (effect size, 0.97; 95% CI, 0.13-1.82; P = .02), but not fatty liver disease: design and rationale of randomized controlled trial.
ALT level. In contrast, the reduction in liver fat (using liver bi- BMC Pediatr 2013;13:85.
opsy, ultrasound, or 1H-magnetic resonance spectroscopy) 7. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard defi-
was beneficial (effect size, 0.97; 95% CI, 0.58-1.35; P < .001). nition for child overweight and obesity worldwide: international survey.
BMJ 2000;320:1240-3.
In one randomized controlled trial, children with NAFLD 8. Saverymuttu SH, Joseph AE, Maxwell JD. Ultrasound scanning in the
were supplemented with 250 or 500 mg DHA or placebo for detection of hepatic fibrosis and steatosis. BMJ 1986;292:13-5.
6 months.5 DHA in both doses reduced liver fat (assessed by 9. Wang CL, Liang L, Fu JF, Zou CC, Hong F, Xue JZ, et al. Effect of lifestyle
abdominal ultrasound) and insulin resistance. There was no intervention on non-alcoholic fatty liver disease in Chinese obese chil-
beneficial effect of omega-3 therapy on ALT or body weight. dren. World J Gastroenterol 2008;14:1598-602.
10. Grnbk H, Lange A, Birkebk NH, Holland-Fischer P, Solvig J,
A limitation of the present study is the use of ALT level as a Hrlyck A, et al. Effect of a 10-week weight loss camp on fatty liver dis-
primary outcome. In other studies, ALT level was considered ease and insulin sensitivity in obese Danish children. J Pediatr Gastroen-
the primary outcome because there are insufficient data in chil- terol Nutr 2012;54:223-8.
dren to allow sample size calculations for histological variables. 11. Dasarathy S, Dasarathy J, Khiyami A, Joseph R, Lopez R, McCullough A.
Validity of real time ultrasound in the diagnosis of hepatic steatosis: a
Moreover, diagnostic criteria were not based on liver histology,
prospective study. J Hepatol 2009;51:1061-7.
because indications for liver biopsy in children are limited. 12. Saadeh S, Younossi ZM, Remer EM, Gramlich T, Ong JP, Hurley M, et al.
Another limitation is the relatively small cohort of patients The utility of radiological imaging in nonalcoholic fatty liver disease.
studied. Gastroenterology 2002;123:745-50.

Omega-3 Fatty Acids Therapy in Children with Nonalcoholic Fatty Liver Disease: A Randomized Controlled Trial 5
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. -, No. -

13. Parker HM, Johnson NA, Burdon CA, Cohn JS, OConnor HT, George J. 18. Arinc H, Sarli B, Baktir AO, Saglam H, Demirci E, Dogan Y, et al. Serum
Omega-3 supplementation and non-alcoholic fatty liver disease: a sys- gamma-glutamyl transferase and alanine transaminase concentrations
tematic review and meta-analysis. J Hepatol 2012;56:944-51. predict endothelial dysfunction in patients with non-alcoholic steatohe-
14. Capanni M, Calella F, Biagini MR, Genise S, Raimondi L, Bedogni G, patitis. Upsala J Med Sci 2013;118:228-34.
et al. Prolonged n-3 polyunsaturated fatty acid supplementation amelio- 19. Franzini M, Fornaciari I, Fierabracci V, Elawadi HA, Bolognesi V,
rates hepatic steatosis in patients with non-alcoholic fatty liver disease: a Maltinti S, et al. Accuracy of b-GGT fraction for the diagnosis of non-
pilot study. Aliment Pharmacol Ther 2006;23:1143-51. alcoholic fatty liver disease. Liver Int 2012;32:629-34.
15. Lloyd-Still JD, Powers CA, Hoffman DR, Boyd-Trull K, Lester LA, 20. Lee DH, Jacobs DR Jr. Association between serum gamma-
Benisek DC, et al. Bioavailability and safety of a high dose of docosahex- glutamyltransferase and C-reactive protein. Atherosclerosis 2005;178:
aenoic acid triacylglycerol of algal origin in cystic fibrosis patients: a ran- 327-39.
domized, controlled study. Nutrition 2006;22:36-46. 21. Musso G, Gambino R, Durazzo M, Biroli G, Carello M, Faga E, et al. Adi-
16. Panchaud A, Sauty A, Kernen Y, Decosterd LA, Buclin T, Boulat O, et al. pokines in NASH: postprandial lipid metabolism as a link between adi-
Biological effects of a dietary omega-3 polyunsaturated fatty acids sup- ponectin and liver disease. Hepatology 2005;42:1175-83.
plementation in cystic fibrosis patients: a randomized, crossover 22. DeFina LF, Marcoux LG, Devers SM, Cleaver JP, Willis BL. Effects of
placebo-controlled trial. Clin Nutr 2006;25:418-27. omega-3 supplementation in combination with diet and exercise on
17. Banderas DZ, Escobedo J, Gonzalez E, Liceaga MG, Ramrez JC, weight loss and body composition. Am J Clin Nutr 2011;93:455-62.
Castro MG. Gamma-glutamyl transferase: a marker of non-alcoholic 23. Munro IA, Garg ML. Dietary supplementation with long chain omega-3
fatty liver disease in patients with the metabolic syndrome. Eur J Gastro- polyunsaturated fatty acids and weight loss in obese adults. Obes Res
enterol Hepatol 2012;24:805-10. Clin Pract 2013;7:e173-8.

6 Janczyk et al
- 2015 ORIGINAL ARTICLES

Figure. Study CONSORT flowchart.

Omega-3 Fatty Acids Therapy in Children with Nonalcoholic Fatty Liver Disease: A Randomized Controlled Trial 6.e1
THE JOURNAL OF PEDIATRICS  www.jpeds.com Vol. -, No. -

Table II. Baseline characteristics of patients by treatment group

Characteristics Placebo (n = 39) Omega-3 (n = 37) Total (n = 76)
Demographic factors
Age, y 12.8 (10.9-14.9) 13.2 (11.2-15.9) 13 (11.1-15.2)
Female sex, n (%) 6.0 (15.4) 5.0 (13.5) 11.0 (14.5)
Anthropometric measurements
Weight, kg 73.0 (62.8-96.7) 77.7 (64.5-93.0) 76.9 (62.9-95.9)
Height, cm 162.0 (154.0-174.5) 165.7 (154.8-172.4) 163.9 (154.3-173.5)
BMI 28.86 (25.6-32.0) 28.6 (26.3-32.1) 28.7 (26.0-32.0)
BMI z-score 2.7 (2.0-3.8) 3.0 (2.2-4.0) 2.8 (2.1-4.0)
WC, cm 94.6 (90.1-106.0) 100.2 (92.7-106.0) 98.4 (90.3-106.0)
WC z-score 3.7 (3.2-4.4) 4.0 (3.3-5.0) 3.8 (3.2-4.6)
Hematology
Erythrocytes, millions/mL 4.9 (4.7-5.1) 4.8 (4.7-5.1) 4.9 (4.7-5.1)
Leukocytes, thousands/mL 7.4 (6.6-9.3) 7.2 (6.1-8.4) 7.4 (6.5-8.8)
Hemoglobin, g/dL 13.9 (12.8-14.8) 13.8 (13.3-14.6) 13.9 (13.1-14.8)
Platelets, mL/dL 278.0 (233.0-318.0) 287.0 (246.0-350.0) 283.0 (236.0-325.0)
Serum biochemical levels
ALT, U/L; n <26 80.0 (64.0-125.0) 79.0 (60.0-92.0) 79.0 (61-108)
AST, U/L; n <33 48.0 (39.0-55.0) 42.0 (35.0-55.0) 45.0 (37.0-55.0)
GGTP, U/L; n <29 48.0 (28.0-63.0) 32.0 (25.0-49.0) 37.0 (26.0-58.0)
Total bilirubin, mg/dL; n <1 0.7 (0.5-1.0) 0.6 (0.5-0.7) 0.6 (0.5-0.9)
INR 1.0 (0.9-1.1) 1.0 (1.0-1.1) 1 (1.0-1.1)
Lipids
Triglycerides, mg/dL; n: 50-150 98.0 (79.0-127.0) 111.5 (79.5-149.0) 108.0 (79.0-140.0)
Cholesterol, mg/dL; n <190 186.0 (158.0-206.0) 177.5 (161.0-217.5) 185.0 (160.0-217.0)
HDL-C, mg/dL; n: 40-80 42.0 (36.0-46.0) 40.5 (34.5-46.0) 41.0 (35.0-46.0)
LDL-C, mg/dL; n: 96-130 127.0 (100.0-150.0) 118.5 (100.0-140.5) 122.0 (100.0-148.0)
VLDL-C, mg/dL; n: 5-20 19.0 (14.0-26.0) 20.5 (15.0-28.0) 20.0 (15.0-26.0)
Metabolic factors
Fasting serum glucose, mg/dL 84.0 (79.0-89.0) 86.0 (79.0-91.0) 85.5 (79.0-91.0)
Fasting serum insulin, mU/mL 20.4 (15.0-26.4) 21.8 (16.9-28.3) 21.4 (15.5-28.3)
HOMA-IR 4.3 (3.1-6.0) 4.9 (3.5-5.9) 4.7 (3.2-5.9)
High-sensitivity C-reactive 0.4 (0.2-0.6) 0.4 (0.3-0.6) 0.4 (0.3-0.6)
protein, mg/mL
Oxy-LDL, mg/mL 196 (114-296) 193 (157-223) 193 (123-265)
Adiponectin, mg/mL 4.3 (1.4-8.0) 6.1 (2.7-8.4) 5.7 (1.7-8.2)
Leptin, mg/mL 14.7 (6.5-24.9) 17.6 (4.9-27.3) 16.5 (5.0-25.4)
Dietary record*
Energy, kcal 2606.2 (2018.5-3403.6) 2502.3 (1879.5-3301.5) 2600.0 (1912.0-3302.0)
Energy/kg body weight 21.9 (16.7-32.8) 20.1 (17.6-30.0) 31.6 (23.9-41.7)
% energy from proteins 15.3 (13.2-16.6) 13.8 (12.0-16.1) 14.8 (12.7-16.4)
% energy from carbohydrates 50.6 (45.9-53.4) 54.1 (46.3-58.2) 51.4 (46.1-56.8)
% energy from lipids 34.2 (30.5-37.6) 30.0 (25.7-37.7) 33.8 (28.7-37.6)

HDL-C, high-density lipoprotein cholesterol; HOMA-IR, homeostasis model assessment for insulin resistance; INR, international normalized ratio; LDL-C, low-density lipoprotein cholesterol; VLDL-C,
very-low-density lipoprotein cholesterol.
Values are expressed as median (IQR) unless specified otherwise.
*Obtained from the food frequency questionnaire.

6.e2 Janczyk et al
- 2015 ORIGINAL ARTICLES

Table V. Differences between baseline and end-of-study

visits
Omega-3, P value Placebo, P value
BMI .002255 .804212
BMI z-score .001197 .241555
WC .108642 .303754
WC z-score .022534 .556602
ALT, U/L .001074 .001196
AST, U/L .001426 .038577
GGTP, U/L .016968 .030356
Total bilirubin, mg/dL .336553 .202744
INR .851925 .338196
Triglycerides, mg/dL .035174 .646864
Total cholesterol, mg/dL .913907 .147019
HDL-C, mg/dL .230140 .901782
LDL-C, mg/dL .690262 .138997
HOMA-IR* .205889 .082035
Adiponectin, mg/mL .130135 .316372
Energy, kcal .000326 .000024
Protein, total intake .006833 .000763
Carbohydrates, total intake .000063 .000013
Lipids, total intake .004273 .000117
*Statistical analysis performed with the Wilcoxon test.

Omega-3 Fatty Acids Therapy in Children with Nonalcoholic Fatty Liver Disease: A Randomized Controlled Trial 6.e3