A systematic review and meta-analysis of randomized controlled trials to evaluate plant-based omega-3 polyunsaturated fatty acids in nonalcoholic fatty liver disease patient biomarkers and parameters.

Nonalcoholic fatty liver disease (NAFLD) is an important noncommunicable disease (NCD) that poses a major challenge to public health. 1 , 2 The global prevalence of NAFLD is approximately 25% in the world population, although there is wide geographical variation, and reaching 30% in Middle Eastern and South American countries. 2 , 3 The disease is recognized as a spectrum of conditions, ranging from ectopic fat accumulation in the liver 4 or simple steatosis (accumulation of lipid droplets within >5% of hepatocytes) 5 to nonalcoholic steatohepatitis (NASH), and ultimately cirrhosis. 6 The estimated prevalence of NAFLD is 25–30% in British and European populations, with younger generations increasingly affected. 5 , 7 With increasingly high prevalence in the United Kingdom and Western Europe, obesity is one of the main contributary factors for NAFLD. The latest statistics for England published in 2020 show that approximately two-thirds of adults are above a healthy weight and half are living with obesity, 8 of whom 90% will have some degree of NAFLD. 2 , 9 Furthermore, obesity accounted for over 1 000 000 UK hospital admissions in 2019/2020, resulting in annual National Health Service costs in excess of £6 billion. 8 , 10 The European Health Interview Survey shows a similar pattern, with 53% of adults over 18 years living with overweight or obesity in Europe 11 and over 4 million people died globally in 2017 due to being overweight or obese. 12 Lifestyle changes, usually aimed at weight loss, are a primary aim of recommended therapeutic strategies, while drug approaches have generally yielded disappointing results so far regarding liver histology outcomes. 13 Omega-3 (n-3) fatty acids offer well-evidenced benefits to NAFLD biomarkers, 14 and diets low in n-3 have been identified as a risk factor for NAFLD mortality using the Global Burden of Disease 2019 data. 15 , 16 Previous systematic reviews and meta-analyses have shown that regular marine-based n-3 consumption predominantly in the form of fish or krill oils through diet and/or supplementation offers significant ( P  <   0.05) benefits to liver fat, alanine aminotransferase (ALT), aspartate aminotransferase (AST), γ-glutamyl transferase (GGT), steatosis score, plasma/serum triglycerides (TGs), insulin resistance, and glycemic control biomarkers. 17–22 These marine-based n-3 fatty acids (predominantly consisting of long-chain n-3 eicosapentaenoic acid [EPA] and docosahexaenoic acid [DHA]) can contain relatively high concentrations of marine pollutants and pose sustainability issues. Fish oil is thought to be the main vector carrying marine pollutants to human consumers. 23 Exposure to heavy metals in adolescence is thought to lead to alterations in energy homeostasis and excessive adiposity in later life. 24 In terms of sustainability, there is increasing demand for highly purified EPA and DHA for the preparation of ethyl ester–based pharmaceutical products, which incurs production losses of 90–95%. 25 In addition, climate change has resulted in the decline of marine microalgae from which oily fish species obtain their EPA and DHA, with a reduction in DHA content for oily fish of 58% predicted by 2100. 26 Marine sources of n-3 by their nature are unsuitable for vegetarians and vegans, and public health organizations are increasingly recommending a plant-based dietary approach for health and sustainability. 27 , 28 Plant-based n-3 sources have been shown to improve markers of NAFLD-related comorbidities, including cardiovascular disease (CVD) and the metabolic syndrome. 29–31 Commercial applications have recently successfully developed direct sources of EPA- and DHA-rich oil from microalgal sources, which provide a plant-based bioequivalent source of n-3 to those found in marine sources. 32 Algal sources of EPA and DHA are freely available food-based supplements, widely recommended as a suitable supplement for non–fish eaters, 25 , 33 with safe, tolerable, nonclinical dosages of up to 3 g/day of EPA/DHA. 34–36 Walnut, flax, chia, canola, hemp, echium, and perilla seed oils are all good plant-based sources of n-3 alpha-linolenic acid (ALA) and offer various health benefits. These include improvements in insulin sensitivity, inflammation, hepatic steatosis, and CVD risk factors, although the health benefits of ALA are not as well established as those attributed to EPA and DHA. 34 Although the benefits of marine n-3 consumption on NAFLD are well established, the effects of plant-based n-3 fatty acids are yet to be examined in a systematic review and meta-analysis. Therefore, the aim of this systematic review and meta-analysis was to evaluate the potential benefits of plant-based n-3 supplementation on NAFLD surrogate biomarkers and parameters.

The current systematic review and meta-analyses follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines 37 and additional nutrition-based systematic review and meta-analysis guidance 38 and is registered in the International Prospective Register of Systematic Reviews (PROSPERO CRD42021251980). 39 The searches for data and articles for inclusion in this systematic review with meta-analysis were carried out until March 2022. Relevant trials were identified through systematic searches of Medline (EBSCO), CINAHL (EBSCO), PubMed, and Google Scholar databases along with Cochrane Central Register of Controlled Trials and the International Clinical Trials Registry Platform to identify randomized controlled trials (RCTs) published between January 1970 and March 2022. The search used the following keywords: “Flax* oil” OR “Linseed oil” OR “Chia seed oil” OR “Rapeseed oil” OR “Echium seed oil” OR “Hemp seed oil” OR “Perilla seed oil” OR “Alga* oil” in the article title or abstract OR “Omega*3” OR “Alpha-linolenic acid” OR “Stearidonic acid” OR “Eicosapentaenoic acid” OR “Docosahexaenoic acid” in the title or abstract AND “liver” (MeSH) OR “hepat*” OR “steat*” OR “fibrosis” OR “cirrhosis” OR “fatty liver” OR “nonalcoholic” OR “Non-alcoholic fatty liver disease” OR “NAFLD” OR Nonalcoholic fatty liver disease” OR “Non-alcoholic fatty liver” OR “Non-alcoholic steatohepatitis” OR “NASH” OR “non-alcoholic steatohepatitis” in the title, abstract, or key words. The full search strategy can be found in Supplementary Material S1 ( please see the Supporting Information online ). To further minimize the effects of publication bias, a snowball method, characterized by manual checking of references from retrieved articles, was applied to ensure complete collection. Identified studies were directly exported into Endnote (version X9; Thomson Reuters) bibliography software for criterion application, whereby a reference lists of eligible studies and review articles were screened for eligibility. Publication alerts were set to identify any studies published after the date of the literature search. Randomized controlled trials evaluating dietary interventions of at least 2 weeks’ duration were included. Studies were excluded on the basis that 1 or more of the following criteria applied: (1) RCTs that only evaluated marine oil n-3 fatty acid sources, (2) RCTs that did not utilize an n-3 fatty acid–rich plant-based source, (3) studies that did not provide data on the levels of the outcomes of interest at baseline and/or at the end of trial, (4) studies that did not indicate the specific n-3 fatty acid or dosage, (5) nonhuman studies, (6) observational studies, (7) cross-sectional studies, (8) case reports, (9) studies not published in the English language, and (10) studies without full access to published findings. Please see the Participant, Interventions, Comparisons, Outcomes and Study Design (PICOS) criteria in Table 1 . PICOS criteria Population Adults diagnosed with NAFLD Adolescents diagnosed with NAFLD Children diagnosed with NAFLD Participants with NAFLD-related pathology (MetS, type 2 diabetes, overweight, and obesity) Population Diagnosed with a serious unrelated pathology other than NAFLD/NASH, MetS, type 2 diabetes, overweight, and obesity (such as cancer, kidney disease, cystic fibrosis, myocardial infarction, organ transplant) Altered endocrinological state (such as pregnancy, lactation, endometriosis, or polycystic ovary syndrome) Intervention Plant-based n-3 supplementation Lifestyle intervention Intervention Marine-based n-3 supplementation without a plant-based arm Comparisons Placebo or nonintervention control (treatment as usual/standard-care group, lifestyle modification to lose 7–10% of total body weight by diet and exercise is the first-line standard-care treatment for NAFLD 5 ) Marine and other oils as a comparator to plant-based n-3 fatty acids Comparisons Studies using animals Single-arm intervention studies without a control group Outcomes Compares effect of plant-based n-3 fatty acid supplementation NAFLD surrogate biomarkers including ALT, AST, and GGT Lipid profiles (TC, HDL cholesterol, LDL cholesterol), plasma/serum TGs Inflammatory markers (Hs-CRP); effect on glycemic-control biomarkers (blood glucose, insulin, HOMA-IR) Body-composition markers (BMI, WC, weight loss) Study design Randomized controlled trials evaluating plant-based n-3 interventions of at least 2 weeks’ duration Study design Studies that did not indicate the specific n-3 fatty acid or dosage Observation studies Cross-sectional studies Case reports Other Full paper English language Other Protocol papers Abstract only Abbreviations : ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; GGT, γ-glutamyl transferase; HDL, high density lipoprotein; HOMA-IR, homeostatic model of insulin resistance; Hs-CRP, high-sensitivity C-reactive protein; LDL, low density lipoprotein; MetS, metabolic syndrome; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; PICOS, Participant, Interventions, Comparisons, Outcomes and Study Design; TC, total cholesterol; TG, triglyceride; WC, waist circumference. Inclusion criteria: Adult, adolescent, and infant populations diagnosed with NAFLD or NASH using any recognized diagnostic criteria. Exclusion criteria: Participants diagnosed with a serious unrelated pathology other than NAFLD/NASH, metabolic syndrome, type 2 diabetes, overweight, and obesity (such as cancer, kidney disease, cystic fibrosis, myocardial infarction, organ transplant) and participants with altered endocrinological state (such as pregnancy, lactation, endometriosis, or polycystic ovary syndrome). Intervention: Plant-based n-3 supplementation using good sources of n-3 fatty acids (with “good” defined as typically >23% of total fatty acids). 34 , 40 , 41 Comparator: Placebo or nonintervention control (treatment as usual/standard-care group, lifestyle modification to lose 7–10% of total body weight by diet and exercise is the first-line standard-care treatment for NAFLD). 5 Marine and other oils as a comparator to plant-based n-3 fatty acids. Studies using animals were excluded. Single-arm intervention studies without a control group were excluded. Primary outcome: A comparison of the effects of plant-based n-3 fatty acid supplementation on recognized NAFLD-related surrogate biomarkers, including liver enzyme biomarkers (ALT, AST, and GGT). Lipid profiles (total cholesterol [TC], high-density-lipoprotein [HDL] cholesterol, and low-density-lipoprotein [LDL] cholesterol), plasma/serum TGs, inflammatory markers (high-sensitivity C-reactive protein [Hs-CRP]), and effects on glycemic-control biomarkers (blood glucose, insulin, homeostatic model of insulin resistance [HOMA-IR]) were also evaluated along with body-composition markers (body mass index [BMI], waist circumference [WC], and weight loss). 17–22 Screening was conducted with 2 independent researchers (E.M. and K.E.L.) in duplicate. Differences in opinion were resolved by group consultation (E.M., K.E.L., I.G.D., and I.P.) until consensus was reached. Extracted data included author, study year, country, study design, intervention source/dose, number of participants and intervention allocation, participant health status, intervention time frame, and main findings. Data were captured for baseline, endpoint, and changes in surrogate biomarkers measured in routine NAFLD patient care and monitoring. 42 Biomarkers included ALT, AST, GGT, TC, HDL cholesterol, LDL cholesterol, TGs, Hs-CRP, blood glucose, insulin, HOMA-IR alongside BMI, and WC, all of which can be elevated in patients with NAFLD, 14 , 43 and weight loss, the most effective treatment for NAFLD. 42 The SI units were included for each outcome measure. Risk of bias of RCTs was evaluated independently by 2 investigators (E.M. and K.E.L.). The assessment was performed at the study level with the revised Cochrane risk-of-bias tool, which grades the risk of selection, performance, attrition, detection, and reporting biases. 44 This tool assesses whether a study has a low risk of bias, some concerns, or a high risk of bias. Differences in opinion were resolved by group consultation (E.M., K.E.L., and I.G.D.) until consensus was reached. All data were collected as means ± SDs. The SD of the unreported mean difference (MD) of the change was calculated by the following formula: SD 2 = [(SD baseline 2 + SD final 2) − (2 × R × SD baseline × SD final )], 45 correlation coefficients ( R ) were computed for each outcome from study data. 46 , 47 Mean differences with 95% confidence intervals (CIs) of changes with forest plots were estimated with the DerSimonian and Laird random-effects model. 48 Heterogeneity between studies was assessed using Cochrane’s Q test ( P  <   0.10) and I 2 statistic (low heterogeneity: 75%), 49 and visual inspection of the Galbraith plot ( see Supplementary Material S2 [ Figures  1.1–1.14 ] in the Supporting Information online ). A sensitivity analysis was conducted to evaluate the effect of each study with significant clinical heterogeneity on overall effect size using the “leave-one-out method,” 50 by omitting 1 study at a time to ensure findings were not driven by any single study. STATA statistical software (version 17; StataCorp) was used for the whole process of meta-analysis. A 2-tailed P value <0.05 was considered to be statistically significant.

Figure 1 presents the study selection process in line with the PRISMA flow chart; the PRISMA checklist can be found in Supplementary Materials S3 in the Supporting Information online . The systematic database search collected a total of 986 articles, 566 from Medline, 263 from PubMed, and 153 from CINAHL. A total of 3 articles were located through additional manual searches using Google Scholar and 1 study was registered with the International Clinical Trials Registry Platform. A total of 360 duplicate articles were removed, leaving 626 articles eligible for screening. Once titles and abstracts had been screened, a total of 566 articles were excluded based on the eligibly criteria. A total of 60 articles remained eligible for full-text screening. A total of 52 articles were excluded on the basis that one of the following applied: (1) participants with elevated NAFLD risk markers but not officially diagnosed, (2) marine oil study only, (3) healthy participants, (4) not n-3, (5) n-3 source/dosage unclear, and (6) serious illness/disease. In summary, a total of 8, 46 , 47 , 51–56 journal articles (showing findings from 6 RCTs) with 362 participants (210 adults, 152 children/adolescents) remained eligible for quantitative synthesis. A full summary table of study design and findings are presented in Table 2 , 46 , 47 , 51–56 including author, study year, country, study design, intervention source/dose, number of participants and intervention allocation, participant health status, intervention time frame, and main findings. All included studies incorporated hypocaloric diets and/or lifestyle interventions aiming to increase physical activity for participants in both the intervention and control/placebo groups. PRISMA flow diagram.   Abbreviations : CCRCT, Cochrane Central Register of Controlled Trials; ICTRP, International Clinical Trials Registry Platform; NAFLD, nonalcoholic fatty liver disease; PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Summary of study design and findings Abbreviations : ALA, alpha-linolenic acid; ALT, alanine aminotransferase; DB, double-blinded; DHA, docosahexaenoic acid; HD, hypocaloric diet; HSC, hepatic stellate cells; IL-6, interleukin-6; LA, linoleic acid; LM, lifestyle modification; MRI, magnetic resonance imaging; NAFLD, nonalcoholic fatty liver disease; NAS, NAFLD Activity Score; OL, open label; PA, physical activity; RCT, randomized controlled trial. A total of 3 included studies recruited adult patients with NAFLD and were carried out in Iran. 46 , 47 , 51 , 52 All of these were RCTs, one of which was double-blinded, 51 another was open-label, 47 , 52 and the third was not disclosed. 46 With regard to intervention, Yari et al 46 introduced ground whole flaxseed powder (WF; 30 g/d; 6.9 g/d ALA, 112.5 mg lignans), 57 hesperidin (HS; 1 g/d) and flaxseed and hesperidin combined (HWF; 30 g/d; 6.9 g/d ALA) hesperidin; 1 g/d, 112.5 mg lignans). 57 Similarly, Yari et al 47 introduced a brown milled flaxseed (WF) intervention (30 g/d; 6.9 g/d ALA, 112.5 mg lignans). 57 Rezaei et al 51 intervened with flaxseed oil (20 g/d; 9.96 g/d ALA), alongside the active control, sunflower oil (20 g/d). All patients consumed a hypocaloric diet and were recommended to perform 30 to 40 minutes of moderate physical activity per day. The mean dosage of ALA for the adult studies was 7.665 g/d and all interventions ran for a period of 12 weeks. 46 , 47 , 51 , 52 The remaining 3 studies recruited children and adolescent patients with NAFLD and were completed in Italy and used algal oil containing 39% DHA. 53–56 The Pacifico et al 53 study intervened with algal oil dosages of 250 mg/d DHA; Nobili et al 54 , 55 used study arms with dosages of 250 mg/d DHA and 500 mg/d DHA, whereas Della Corte et al 56 included 500 mg/d DHA from algal oil and 800 IU vitamin D. A total of 2 studies introduced placebo wheat germ oil (290 mg/d; omega-6 [n-6] linolenic acid), 53–55 as an active control; and 1 study included a hypocaloric diet and lifestyle modifications with no other intervention. 56 Algal oil intervention durations ranged from 24 weeks 53 , 56 to 24 months. 54 , 55 Details of the quality assessment can be found in Table 3 46 , 47 , 51–56 and Figures 2 and 3 . 46 , 47 , 51–56 The child/adolescent algal oil studies by Nobili et al 54 and the adult flaxseed study Pacifico et al 53 showed an overall low risk of bias. The remaining RCTs 46 , 47 , 51 , 52 , 56 showed some concerns in at least 1 domain, including randomization process and measurement of outcome, but were not deemed to be at high risk for any domain. 44 Risk-of-bias graph: authors’ judgments for each risk-of-bias item are presented as percentages across all included studies. Risk-of-bias summary for the included studies. Cochrane risk-of-bias tool for randomized trials—quality assessment of all studies

When combined with lifestyle interventions to increase physical activity and a calorie-controlled diet, plant-based n-3 fatty acid supplementation leads to improvements in ALT enzyme biomarkers, TGs, BMI, WC, and weight loss, therefore offering a potential adjunct to the first-line treatment of lifestyle modification (ie, losing 7–10% of weight by diet and exercise). Further well-designed human studies are needed to better understand the beneficial effects offered by plant-based n-3 fatty acids, with and without lifestyle and dietary interventions and with a specific focus on the different effects of ALA, and plant-based EPA and DHA using larger numbers of patients with NAFLD and longer study durations.

The present systematic review and meta-analysis evaluated RCTs investigating the effect of plant-based n-3 fatty acids on NAFLD-related surrogate biomarkers, including liver enzymes, glycemic-control biomarkers, lipid profiles, inflammatory markers, and body composition. To the authors’ knowledge, this is the first study to evaluate plant-based n-3 fatty acids using NAFLD surrogate markers. The current novel meta-analysis shows that plant-based n-3 fatty acid supplementation has beneficial effects on ALT, TGs, and body-composition markers, including BMI, WC, and total weight. Three studies recruited adult patients with NAFLD and were carried out in Iran where NAFLD prevalence is estimated at 27.88% in men and 30.17% in women. 58 The remaining 3 studies recruited children and adolescents in Italy where NAFLD prevalence is 22.5–27.0% in the general population. 59 This study aimed to evaluate the benefits of plant-based n-3 fatty acid supplementation and did not make direct comparisons to marine-based n-3 studies. The meta-analysis showed a significant ( P  =   0.01) pooled reduction of ALT with plant-based n-3; however, there was no effect on AST or GGT. This suggests that plant-based n-3 does not offer improvements to AST and GGT; however, it should be noted that 2 of the child/adolescent studies 53–55 did not measure AST and GGT, so the meta-analysis predominantly represents the adult ALA studies for these biomarkers. The adult flaxseed studies had a relatively short 12-week duration, dosages of 6.9 46 , 47 or 9.96 g/d 51 ALA, and small sample sizes, which may have resulted in null findings for AST and GGT. It should also be noted in some of the study arms, baseline ALT levels were within the upper-normal limits for NAFLD. 60 , 61 Although this is the first systematic review and meta-analysis of plant-based n-3 fatty acid supplementation and NAFLD-related liver enzyme biomarkers, findings on ALT are similar to recent marine oil n-3 meta-analyses. Yan et al 18 showed significant ( P  <   0.05) improvements in ALT in 18 RCTs evaluating fish-oil supplementation in 1424 patients. The study also showed benefits to AST and GGT but with a larger number of studies and patients. Similarly, Lee et al 62 showed significant improvements in ALT ( P < 0.001) and AST ( P < 0.001) in a systematic review and meta-analysis of 22 marine/fish-oil RCTs. NAFLD is thought to be a manifestation of metabolic syndrome 63 and insulin resistance is a common factor that connects obesity, hypertension, dyslipidemia, type 2 diabetes, and NAFLD progression. 63 , 64 In the present article, the meta-analysis showed that plant-based n-3 supplementation offered no significant improvements in glycemic-control biomarkers and glucose levels were within healthy ranges for some of the study arms. 61 The marginal significance obtained for blood glucose ( P  =   0.09) and HOMA-IR ( P  =   0.07) could result from a small sample size and short-term study durations and, although not statistically significant, should be interpreted with caution. The findings for this study are in agreement with a systematic review and meta-analyses of 7 RCTs by He et al, 21 who showed the efficacy of marine n-3 including fish oil was unclear in relation to glycemic-control markers. The findings of this study agree with the growing consensus in recent literature showing that, despite improvements in insulin sensitivity in rodent studies, the vast majority of human intervention studies fail to demonstrate beneficial effects of n-3 fatty acids in glycemic-control biomarkers in type 2 diabetes or insulin resistance. 65 , 66 The meta-analyses showed a significant reduction in plant-based n-3 from flaxseed oil and algal oil on TGs ( P  =   0.01) and a trend in lowering LDL-C ( P  =   0.05). Patients with NAFLD have a substantially increased risk of mortality from CVD, 67 , 68 and lipid profiles form an important part of CVD risk management. 69 Moderate increases in LDL-C in patients with NAFLD can increase the risk of a cardiovascular event. 70 The significant TG reductions are important as TG measurement represents a useful noninvasive biomarker for diagnosis, prognosis, and monitoring of NAFLD progression. 71 This study’s findings relating to plant-based n-3 supplementation and changes in lipid profiles partially agree with recently published literature. Meta-analyses of flaxseed oil supplementation studies and CVD risk by Hadi et al 29 and Masjedi et al 31 showed flaxseed supplementation significantly lowered TGs ( P  <   0.05). In contrast to the findings of the present review, significant reductions were also shown for TC ( P  =   0.01). However, both meta-analyses had higher numbers of studies/participants and therefore greater statistical power than the present study. In a meta-analysis of 47 RCTs with 1305 individuals, Yue et al 72 showed that dietary intake of ALA significantly ( P  <   0.05) reduced lipid profile biomarkers. A meta-analysis by Naghshi et al 73 showed that a high intake of ALA (defined as ≥3 g/d) was significantly ( P  <   0.05) associated with a lower risk of death from CVD and coronary heart disease compared with low intake (defined as ≤0.35 g/d). The meta-analyses showed significant improvements in body-composition measures, including BMI (MD: −1.83 kg/m 2 ; 95% CI: −2.99, −0.68; P  =   0.00; I 2 = 76.37%), WC (MD: −3.21 cm; 95% CI: −4.92, −1.50; P  =   0.00; I 2 = 20.32%), and weight (MD: −4.68 kg; 95% CI: −8.33, −1.04; P  =   0.01; I 2 = 0.00%). Weight loss is widely recognized as the most effective treatment for NAFLD, 63 , 74 with current treatment guidelines recommending lifestyle interventions aiming to lose 7–10% of total weight in patients with NAFLD. 5 , 63 Improvements in body composition compare with previously published lifestyle intervention studies. A systematic review and meta-analyses by Katsagoni et al 75 showed that diet interventions and exercise decreased BMI and WC from 20 RCTs with 1073 patients with NAFLD. Improvements in body composition also translated to liver enzyme biomarkers, which showed significant decreases in ALT and AST (both P  <   0.05). Long-chain n-3 polyunsaturated fatty acids are categorized by the presence of a double bond at the third carbon atom from the methyl end of the carbon chain. 4 Where direct marine-based n-3 sources are scarce or not present in the diet, n-3 ALA (18:3; which was used in the 3 adult RCTs in the present review) 46 , 47 , 51 must undergo conversion to the more biologically active longer chain highly unsaturated EPA (20:5) 76 and DHA (22:6) 77 in the n-3 metabolic pathway. High dietary n-6 ratios, which are indicative of low n-3 intakes, can be proinflammatory and limit the conversion of ALA to EPA and DHA in the n-3 metabolic pathway. 34 , 78 Those consuming low intakes of n-3 coupled with a high n-6 to n-3 dietary fatty acid ratio are at increased risk of developing NAFLD. 15 , 79 Incorporating a direct source of DHA, which was used in the child/adolescent RCTs identified in the present review, 53 , 54 , 56 is the most effective way to obtain DHA in the diet for low/non–fish consumers. 34 , 41 The molecular mechanisms causing the initiation and progression of NAFLD have yet to be fully recognized and understood; however, inflammation and oxidative stress induced by reactive oxygen species are likely to be significant mechanisms that can lead to hepatic cell death and tissue injury. 80 NAFLD is usually believed to have a mutual and bidirectional nexus with metabolic syndrome and its individual components and, therefore, to be a noncommunicable disease. However, Helicobacter pylori infection is also believed to play a role in the development of NAFLD, as does the gut microbiota. 81–84 Dietary n-3 fatty acids are thought to have beneficial effects in bioactive metabolites involved in inflammatory pathways and liver metabolism, including the reduction in hepatic TG accumulation and regulation of hepatic lipid metabolism and adipose tissue function. 4 , 47 Furthermore, dietary n-3 fatty acids have been shown to regulate various H. pylori –associated gastric diseases. 85 However, further studies are required to evaluate the underlying mechanisms involved. 4 , 19 , 79 , 85 To the authors’ knowledge this is the first systematic review and meta-analysis to evaluate the effect of plant-based n-3 fatty acid interventions on NAFLD biomarkers and parameters. However, there are limitations that should be considered. The systematic search criteria identified 6 RCTs with 362 patients with NAFLD, representing a small number of participants and short study durations compared with similar systematic reviews and meta-analysis that have evaluated fish/marine-based oil n-3 supplementation. As the search criteria identified that there have only been 6 relevant clinical trials investigating the topic since 1970, this highlights an overall scarcity of plant-based n-3 intervention studies in this area of research. The use of liver enzyme measurements on the whole may be a limitation as up to 78% of patients with NAFLD have liver enzyme biomarkers within normal ranges. 2 The relatively short duration of studies may limit inferences on the discordance between results of significance between ALT, AST, and GGT. The AST-to-ALT ratio can be a useful indicator of fibrosis in NAFLD 86 ; however, the study by Rezaei et al 51 was the only RCT identified in the present study to use this measure, which showed no significant differences between the active treatment control and intervention groups. Outcomes in this review (biomarkers and parameters) are surrogate endpoints, which might not directly translate to patient benefit or predict improvements in clinical outcomes for patients. 5 , 87 The measurement of surrogate biomarkers in the identified studies with no histological confirmation of the diagnosis and outcomes does not address American Association for the Study of Liver Diseases and European Association for the Study of the Liver NAFLD/NASH accepted endpoints, including changes in liver steatosis, hepatic inflammation, and fibrosis. 84 , 88 A number of publications have used magnetic resonance imaging for the noninvasive measurement of adult and adolescent liver steatosis 19 , 89 , 90 ; however, none of the studies identified in this review used magnetic resonance imaging measurements to monitor the effectiveness of interventions. The study findings concerning body composition represent a key limitation as it was not possible to delimit and separate the beneficial effects of plant-based n-3 and weight loss, the first-line standard-care treatment for NAFLD 5 in the meta-analysis. However, it is noteworthy that the control/placebo groups in 3 of the 6 included studies by Pacifico et al 53 (active treatment control: wheat germ oil), Yari et al, 46 and Della Corte et al 56 (inactive treatment control) showed significant reductions to BMI in plant-based n-3 intervention groups ( P  <   0.05) but not in the control/placebo groups despite the same lifestyle intervention. Significant heterogeneity was found in the present meta-analyses. The effect of plant-based n-3 supplementation in young and adult populations may be affected by type, duration, dosage, and further effects of weight loss. However, due to the limited number of relatively short-term studies focusing on this area, and translational small patient cohorts included in the meta-analysis, it was not possible to conduct further subgroup analysis to identify potential sources of heterogeneity; hence, the results need to be interpreted with caution. Some of the null findings could be related to the dosages administered or the samples sizes used. The search did not identify any studies that made direct comparisons between marine and plant-based n-3 fatty acid sources in surrogate NAFLD biomarkers. Further consideration must be given to wheat germ oil, 91 which was used as a placebo/control in RCTs by Pacifico et al 53 and Nobili et al 54 and represents a good source of vitamin E and n-6 (α-tocopherol ∼1179 mg/kg and β-tocopherol ∼398 mg/kg; linolenic acid: 49–60% [n-6]). 91 Vitamin E supplementation shows beneficial effects in patients with NAFLD, both with and without weight loss, as evidenced by several systematic reviews and meta-analyses 92–94 and the Pioglitazone versus Vitamin E versus Placebo for the Treatment of Nondiabetic Patients with Nonalcoholic Steatohepatitis (PIVENS) trial, 95 posing a potential limitation for the studies using a wheat germ oil placebo/controls. In the present review, neither of the 2 RCTs 53 , 54 specified the exact quantity of wheat germ oil used, stating a dose of 290 mg/d linolenic acid (n-6), representing 49–60% of fatty acids in wheat germ oil composition. 91 A cross-sectional study in 6693 adults in the United States showed that both dietary n-3 and n-6 fatty acids may offer benefits in NAFLD; however, further studies are needed to verify this and to identify the associated mechanisms. The summary findings in Table 2 show that there were no significant improvements between baseline and endpoint for any of the biomarkers included in the meta-analysis for the comparator groups in studies using wheat germ oil as a placebo/control. A further limitation of this meta-analysis is that the 3 included studies of adult populations were all conducted in 1 country, Iran. As such, these studies may be affected by environmental factors such as local diet that may not be applicable in other countries. This may also limit the generalizability of the results to other, diverse populations and/or ethnicities. The use of whole ground flaxseed by Yari et al 46 , 47 in 2 of the 3 identified adult studies may also represent a confounding effect as fiber intake is known to ameliorate NAFLD 96 ; this may have added to the heterogeneity between studies. Due to the limited number of studies in this area, there is a lack of consensus surrounding effective doses of ALA for the amelioration of NAFLD. The adult RCTs identified in the review used flaxseed interventions and the child/adolescent RCTs used algal oil comprising ALA and DHA, respectively. The child studies used doses of either 250 or 500 mg/d DHA, which is in line with dietary recommendations based on cardiovascular risk considerations for European adults; again, there is a lack of consensus surrounding effective DHA doses for NAFLD in children and adolescents due the shortage of published studies. It was not possible to separate the effect for the 2 different n-3 sources; although the identified studies included study arm demographics for males and females, none of them conducted separate statistical analyses by sex and reproductive status, which can account for differences in prevalence, risk factors, fibrosis, and clinical outcomes in NAFLD. 97 Research investigating the distinct effects of n-3 supplementation in separate male and female populations is warranted. This highlights a gap in the evidence for DHA-based adult studies and ALA-based child/adolescent studies. Both sources have shown improvements in cardiometabolic risk biomarkers. 34 , 72 However, their effects on NAFLD parameters have yet to be extensively evaluated. Furthermore, none of the RCTs identified in the present review examined plant-based EPA or EPA and DHA combined (mainly found in algal oils), 34 which are thought to be the most health advantageous sources of n-3. 98 Further RCTs are now needed to investigate the potential benefits of ALA from flaxseed oil and EPA/DHA- and DHA-rich algal oil on NAFLD biomarkers using appropriate placebo/controls, larger numbers of patients with NAFLD, and longer study durations to address these limitations. In addition to the analysis of NAFLD surrogate markers, provision for additional noninvasive markers such as magnetic resonance imaging should be included in future studies.

The pooled estimate from all 6 studies of the effects of plant-based n-3 fatty acids on ALT from interventions including 83 adults and 63 child/adolescent patients showed a significant ( P  =   0.02) pooled reduction (MD: −8.04 IU/L; 95% CI: −14.7, −1.38; I 2 = 48.61%) 46 , 51–55 ( Table 4 46 , 47 , 51–56 and Figure 4A 4 6 , 47 , 51–56 ). Forest plot of differences in alanine aminotransferase between n-3 and placebo arms.   Abbreviation : CI, confidence interval. Liver enzyme baseline and endpoint biomarkers Abbreviations : ALT, alanine aminotransferase; AO, algal oil; AO+D, algal oil + vitamin D; AST, aspartate aminotransferase; C, control; FO, flaxseed oil; GGT, γ-glutamyl transferase; HS, hesperidin; HWF, hesperidin and whole flaxseed; NS, not significant; WF, whole flaxseed; WGO, wheat germ oil. *ANOVA ( P < 0.05) between baseline and 12 months. Statistically significant difference ( P  < 0.008) compared with the Control and Flax group, respectively. The pooled estimate on the effects of plant-based n-3 fatty acids on AST and GGT from 4 studies showed no significant pooled effect (AST—MD: −2.29 U/L; 95% CI: −6.76, 2.18; I 2 = 34.61%; GGT—MD: −3.32 U/L; 95% CI: −10.24, 3.30; I 2 = 70.62%). All 3 of the adult flaxseed studies were included (83 patients) in the meta-analysis 46 , 47 , 51 , 52 along with the child/adolescent algal oil study (18 patients) by Della Corte et al 56 ( Table 4 , Figure 4B ) 46 , 47 , 51 , 56 and ( Figure 4C ). 46 , 47 , 51 , 56 Forest plot of differences in aspartate aminotransferase between n-3 and placebo arms. Abbreviation : CI, confidence interval. Forest plot of differences in γ-glutamyl transferase between n-3 and placebo arms. Abbreviation : CI, confidence interval. The pooled estimate on the effects of plant-based n-3 fatty acids on blood glucose and insulin from 4 studies showed a marginally significant ( P  =   0.09) pooled effect for glucose (MD: −1.80 mg/dL; 95% CI: −3.85, 0.25; I 2 = 75.92%) and was nonsignificant for insulin (MD: −0.50 µU/mL; 95% CI: −1.59, 0.59; I 2 = 16.12%) for the included studies, which accounted for 2 adult flaxseed ALA interventions 46 , 51 (58 patients) and 2 child/adolescent (43 patients) DHA trials 53 , 56 ( Table 5 , 46 , 47 , 51–56 Figure 5A , 46 , 51 , 53 , 56 and Figure 5B 46 , 47 , 51 , 53 , 54 , 56 ). Forest plot of differences in blood glucose between n-3 and placebo arms. Abbreviation : CI, confidence interval. Forest plot of differences in insulin between n-3 and placebo arms. Abbreviation : CI, confidence interval. Glycemic-control biomarker baseline and endpoint measures Abbreviations : AO, algal oil; AO+D, algal oil + vitamin D; C, control; FO, flaxseed oil; HOMA-IR, homeostatic model of insulin resistance; HS, hesperidin; HWF, hesperidin and whole flaxseed; WF, whole flaxseed; WGO, wheat germ oil. Statistically significant difference ( P  < 0.008) compared with the Control group. The pooled estimate on the effects of plant-based n-3 fatty acids on HOMA-IR for all 6 of the included studies 46 , 51–55 showed a marginally significant ( P  =   0.07) pooled effect (MD: −0.22; 95% CI: −0.44, 0.01; I 2 = 0.00%) at the intervention follow-up for 63 adults and 83 child/adolescent patients ( Table 5 46 , 47 , 51–56 and Figure 5C 46 , 47 , 51 , 53 , 54 , 56 ). Forest plot of differences in homeostatic model of insulin resistance between n-3 and placebo arms. Abbreviation : CI, confidence interval. The pooled estimate on the effects of plant-based n-3 fatty acids on TC and HDL cholesterol for the 5 included studies 46 , 47 , 51 , 53 , 56 showed no significant pooled effect (TC—MD: −4.91 mg/dL; 95% CI: −18.14, 8.32; I 2 = 59.26%; HDL cholesterol—MD: 1.054 mg/dL; 95% CI: −3.20, 5.31; I 2 = 89.39%) at the intervention follow-up in 43 child and adolescent patients and 83 adults ( Table 6 , 46 , 47 , 51–56 Figure 6A , 46 , 47 , 51 , 53 , 56 and Figure 6B 46 , 47 , 51 , 56 ). Forest plot of differences in total cholesterol between n-3 and placebo arms. Abbreviation : CI, confidence interval. Forest plot of differences in high-density-lipoprotein cholesterol between n-3 and placebo arms. Abbreviation : CI, confidence interval. Lipid profile biomarkers—total cholesterol, HDL cholesterol, and LDL cholesterol baseline and endpoint measures Abbreviations : AO, algal oil; AO+D, algal oil + vitamin D; C, control; FO, flaxseed oil; HDL, high density lipoprotein; HS, hesperidin; HWF, hesperidin and whole flaxseed; LDL, low density lipoprotein; TC, total cholesterol; WF, whole flaxseed; WGO, wheat germ oil. The pooled estimate on the effects of plant-based n-3 fatty acids on LDL-C from 4 studies showed a nonsignificant pooled reduction effect (MD: −8.75 mg/dL; 95% CI: −17.53, 0.04; I 2 = 0.00%). All of the adult flaxseed studies were included (83 adult patients) in the meta-analysis, 46 , 47 , 51 , 52 along with the child/adolescent algal oil study (18 patients) by Della Corte et al 56 ( Table 6 and Figure 6C 46 , 47 , 53 , 56 ). Forest plot of differences in low-density-lipoprotein cholesterol between n-3 and placebo arms. Abbreviation : CI, confidence interval. The pooled estimate on the effects of plant-based n-3 fatty acids on plasma/serum TGs from 4 studies showed a significant ( P  =   0.01) pooled reduction (MD: −44.51 mg/dL; 95% CI: −76.93, −12.08; I 2 = 69.93%) 46 , 47 , 53–56 ( Table 7 46 , 47 , 51–56 and Figure 6D 46 , 47 , 53 , 54 , 56 ). All of the child/adolescent algal oil studies 53 , 54 , 56 were included (63 patients) in the meta-analysis alongside 2 adult ALA studies (49 patients). 46 , 47 Forest plot of differences in triglycerides between n-3 and placebo arms. Abbreviation : CI, confidence interval. Triglycerides, high-sensitivity C-reactive protein, and tumor necrosis factor baseline and endpoint measures Abbreviations : AO, algal oil; AO+D, algal oil + vitamin D; C, control; FO, flaxseed oil; HS, hesperidin; Hs-CRP, high-sensitivity C-reactive protein; HWF, hesperidin and whole flaxseed; TG, triglyceride; TNF, tumor necrosis factor; WF, whole flaxseed; WGO, wheat germ oil. Statistically significant difference ( P  < 0.008) compared with the Flax group. The pooled estimate on the effects of plant-based n-3 on Hs-CRP showed no significant pooled effect for the 3 included studies (MD: −1.03 ng/dL; 95% CI: −7.20, 5.15; I 2 = 0.00%) at the intervention follow-up in 25 child 53 and adolescent patients and 49 adults 46 , 47 , 52 ( Table 7 and ( Figure 7 46 , 47 , 53 ). Forest plot of differences in high-sensitivity C-reactive protein between n-3 and placebo arms. Abbreviation : CI, confidence interval. Plant-based n-3 fatty acids showed favorable effects for all 3 body-composition markers. With respect to BMI, the pooled estimate on the effects of plant-based n-3 showed a significant ( P  <   0.001) pooled reduction effect (MD: −1.83 kg/m 2 ; 95% CI: −2.99, 0.68; I 2 = 76.37%) for all 6 of the included studies at the intervention follow-up including 83 adult 46 , 47 , 52 and 63 child/adolescent patients 53–56 ( Table 8 46 , 47 , 51–56 and Figure 8A 46 , 47 , 51–56 ). Similar results we observed for WC, with a significant ( P  <   0.001) pooled reduction effect (MD: −3.21 cm; 95% CI: −4.92, −1.50; I 2 = 20.32%) for 5 studies including 63 children/adolescents 53 , 54 and 45 adult 46 , 47 , 51 patients ( Table 8 and Figure 8B 46 , 47 , 51 , 53 , 56 ). Finally, weight loss showed a significant ( P  =   0.01) pooled reduction effect (MD: −4.68 kg; 95% CI: −8.33, −1.04; I 2 = 0.00%) for 3 studies including 25 child/adolescent patients 53 and 59 adults 46 , 51 ( Table 8 and Figure 8C 47 , 51 , 53 ). Forest plot of differences in body mass index between n-3 and placebo arms. Abbreviation : CI, confidence interval. Forest plot of differences in waist circumference between n-3 and placebo arms. Abbreviation : CI, confidence interval. Forest plot of differences in weight loss between n-3 and placebo arms. Abbreviation : CI, confidence interval. Body-composition measures—body mass index, waist circumference, and weight baseline and endpoint measures Abbreviations : AO, algal oil; BMI, body mass index; C, control; D, vitamin D; FO, flaxseed oil; HS, hesperidin; HWF, hesperidin and whole flaxseed; WC, waist circumference; WF, whole flaxseed; WGO, wheat germ oil.

Click here for additional data file.