Review Article

The Management of Paediatric Nonalcoholic Fatty Liver Disease

E. B. Mitchel; J. E. Lavine


Aliment Pharmacol Ther. 2014;40(10):1155-1170. 

In This Article

Abstract and Introduction


Background Paediatric nonalcoholic fatty liver disease (NAFLD) is a major public health concern given the recent increase in its prevalence and link to obesity and other metabolic comorbidities. Current treatment strategies involve lifestyle changes. Other surgical and pharmacologic interventions have been proposed; however, limited randomised controlled trials (RCTs) in the paediatric population restrict their use.

Aim To review the current management of paediatric NAFLD, including lifestyle and pharmacologic interventions, and to formulate recommendations for study design for future studies.

Methods A MEDLINE, Pubmed and Cochrane Review database search used a combination of keywords, including NAFLD, nonalcoholic steatohepatitis (NASH), paediatric, treatments, lifestyle changes, bariatric surgery, orlistat, metformin, thiazolidinediones, vitamin E, cysteamine bitartrate, ursodeoxycholic acid (UDCA), probiotics, omega-3 fatty acids, pentoxyfylline, farnesoid X receptor agonist and toll-like receptor modifiers. The articles were selected based on their relevance to the review.

Results Lifestyle interventions involving diet and exercise remain first-line treatment for paediatric NAFLD. Bariatric surgery, orlistat, insulin sensitisers and UDCA have been evaluated but are not recommended as first or second-line therapy. Medications such as cysteamine bitartrate, probiotics, polyunsaturated fats and pentoxyfilline share beneficial effects in trials, however, there is a paucity of adequately powered RCTs in which liver histology is evaluated. Vitamin E has been shown to be effective and safe in improving NASH histology in children.

Conclusions Lifestyle intervention should be first-line treatment for paediatric NAFLD. Vitamin E should be considered for those with biopsy-proven NASH or borderline NASH failing first-line therapy. Other therapeutics show promising results but require larger RCTs with convincing endpoints. Improved screening techniques, objective validated inclusion criteria and outcome measures as well as rigour in study design are necessary for propelling therapeutic discovery.


Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease in children living in developed nations.[1] Its steep rise in prevalence is closely related to the increase in obesity in the paediatric population.[2] Epidemiologic studies demonstrate that paediatric NAFLD predominates in obese peri-pubertal males with other features of metabolic syndrome including hyperlipidaemia, insulin resistance and central obesity. It is generally more severe in indigenous American and certain Hispanic and Asian populations.[1,3,4]

Determination of NAFLD prevalence is limited by the utilisation of noninvasive but nonspecific surrogate serologic markers and ultrasound imaging as the primary methods for diagnosis. However, liver biopsy with confirmation on histology is considered the gold standard for diagnosis. One large coroner's study examining liver histology found NAFLD prevalence among all children between 2 and 19 years old to be 9.6% after adjustment for age, gender, race and ethnicity and up to 38% in obese children.[5]

As in adults, paediatric NAFLD is defined as macrovesicular steatosis in more than 5% of hepatocytes in the absence of other aetiologies or disorders that may lead to fat deposition in the liver.[5,6] NAFLD exists within a spectrum of liver pathology that ranges from isolated steatosis to steatosis with inflammation and cellular injury (nonalcoholic steatohepatitis – NASH) and cirrhosis.[5] A significant subset of paediatric NAFLD cases demonstrates histological changes that are not found in adults. This pattern is characterised by pronounced steatosis and inflammation located in the portal and periportal areas with no or minimal lobular inflammation, and portal fibrosis.[5,7,8] Whether this pattern, termed Borderline Zone 1 or Type 2 NAFLD, evolves into adult NASH remains unclear.[4] NAFLD must be distinguished from other underlying liver diseases such as autoimmune hepatitis, Wilson disease, alpha-1-Antitrypsin deficiency, metabolic disorders, glycogen storage disease and hepatitis C as well as many other childhood fat-infiltrating diseases.[9]

Paediatric NAFLD has become a major public health concern due to the striking increase in prevalence, the limited data on aetiology and pathogenesis, and the implications for future development of liver dysfunction, diabetes and cardiovascular disease. Current therapeutic strategies aim to decrease the incidence of known risk factors to prevent progression of liver disease and NAFLD-associated comorbidities.[10,11] There are a limited number of randomised controlled trials (RCTs) that have evaluated the effectiveness of treatments for paediatric NAFLD. Weight loss through caloric restriction, improved diet and physical activity are considered the cornerstone of treatment in paediatric NAFLD.[12,13] The current literature on the efficacy of both lifestyle and pharmacologic interventions for the treatment of paediatric NAFLD will be discussed in this review. In areas where paediatric studies are lacking, adult studies will be used to discuss the implications of these interventions. Recommendations for proper study design and outcomes monitoring for future studies will also be assessed.

Lifestyle Changes

Lifestyle changes are the mainstay of therapy in NAFLD, demonstrated to improve liver histology and attendant surrogate markers of improvement.[14] Children with biopsy-proven NAFLD have been shown to be more sedentary and consume a Western diet, characterised by high levels of total fat, saturated fat and refined sugar, when compared to both obese and lean peers without liver disease.[15,16] Thus, provision of information on needed lifestyle change and guidance as to how to invoke change is of primary concern.

Weight Loss. Weight reduction decreases delivery of free fatty acids to the liver, increases extrahepatic insulin sensitivity by means of improving peripheral glucose utilisation, promotes a reduction in reactive oxygen substances and reduces adipose inflammation.[17] NAFLD in adults remits with weight reduction and similar success has been reported in obese children and adolescents.[18–21] Loss of 3–5% of total body weight decreases hepatic steatosis, although loss of 10% or more may be required for reduction in inflammation and progression of NASH.[13,14] This weight loss should be achieved gradually at about 500 g/week. Rapid weight loss can lead to onset of severe metabolic disorders and promote further liver damage.[22,23]

Evidence to support the utility of weight loss in the treatment of paediatric NAFLD includes a prospective study of 84 children with biopsy-proven NAFLD over 1 year. Lifestyle intervention incorporating low-calorie diet (25–30 cal/kg/day; 50–60% carbohydrate, 23–30% fat, 15–20% protein, 2/3 saturated and 1/3 unsaturated fatty acids and 4:1 omega-6/omega-3 ratio) with weekly nutritional counselling and aerobic exercise (30–45 min/day at least three times per week) resulted in a significant decrease in BMI [baseline BMI: 25.9 kg/m2 ±3.8 vs. 1 year post-intervention BMI: 23.8 ± 2.7 (P < 0.0001)] with improvement in fasting glucose, insulin levels, serum lipid levels and aminotransferases as well as diminished liver steatosis on ultrasound.[24] These results were confirmed in a longitudinal randomised controlled trial of lifestyle intervention over 2 years in 152 obese children with NAFLD diagnosed by ultrasound. Greater degrees of BMI reduction resulted in greater reduction in serum aminotransferase.[18] In Denmark, 117 obese children were enrolled into a 10-week 'weight loss camp' with subsequent dietary and exercise interventions for 1 year. Average weight loss was 7.1 ± 2.7 kg. The 71 children completing the study improved their insulin sensitivity, serum aminotransferases and liver steatosis. Only 24% of the children maintained this weight loss at follow-up 12 months later.[25]

While these studies relate weight loss to improvement in NAFLD as assessed by laboratories and liver imaging, few trials utilise liver histology to assess the effect of weight loss. For studies that assess histology, NAFLD Activity Score (NAS) is often used as an endpoint. NAS can range from 0 to 8 [NAS = steatosis (0–3) + lobular inflammation (0–3) + hepatocyte ballooning (0–2)]. This system was developed and validated by the NASH Clinical Research Network Pathology Committee and was designed to semi-quantitate features deemed meaningful in categorising NAFLD severity.[8] In one trial, 31 obese adults with NASH randomised to either intensive lifestyle change or scripted basic instruction showed that the intensive arm had a 9.3% weight loss (vs. 0.2% with counselling alone) and showed significant reduction in NAS (−2.4 vs. −1.4, P = 0.05) and steatosis score (−1.1 vs. −0.3, P = 0.02). A higher proportion of participants in the intervention group had a NAS of ≤2 at the end of the study as compared to controls (P = 0.02). The subset of patients with >7% weight loss showed improvement in steatosis (−1.4 vs. −0.4, P < 0.001), lobular inflammation (−0.8 vs. −0.2, P = 0.03), ballooning (−1.3 vs. −0.5, P = 0.03) and NAS (−3.5 vs. −1.2, P < 0.001) compared to those who lost less.[26] Similar results were obtained when adults with NASH losing >5% body weight showed only improvement in steatosis, as compared to individuals with >9% weight loss who had significant improvement in steatosis, lobular inflammation, ballooning and NAS.[27] Significant improvement in steatosis, inflammation and ballooning after 2 years of lifestyle intervention was reported in one paediatric study.[28]

Physical Activity. Lifestyle interventions for obese children with NAFLD incorporate increased physical activity. The extent to which exercise in isolation impacts liver disease is unknown. Exercise may mediate progression of fatty liver by reducing fatty acid uptake, decreasing oxidative stress and inflammation and improving insulin sensitivity.[29–31] Vigorous and sustained exercise programmes can improve serum ALT, insulin resistance and NASH severity independent of weight loss.[32–34] Resistance training improves insulin sensitivity, lowers fasting glucose and decreases abdominal fat independent of weight loss.[35–37] Obese adolescents undergoing a 12-week resistance exercise regimen for 1 h twice weekly showed significant increases in body strength, hepatic insulin sensitivity and decreased hepatic fat-fraction.[38] Current recommendations for children entail daily aerobic physical exercise for 30–60 min with moderate to vigorous intensity, in conjunction with age-appropriate muscle and bone-strengthening at least 3 days weekly.[39]

Diet. The optimal dietary composition for the treatment of NAFLD is derived from observational studies. Dietary recommendations for children must be balanced to allow for sufficient statural growth and should be individualised based on recognised nutritional requirements.[40] Appropriate dietetic measures include reduction in total fat, trans-fat, cholesterol, sucrose and fructose. In a prospective cross-sectional study, 149 children with biopsy-proven NAFLD were surveyed about diet. There was no significant difference found between children with steatosis compared to those with steatohepatitis for fraction of calories consumed from fat, carbohydrates and protein. While reported sweetened-drink consumption did not correlate with histological features, uric acid, a surrogate marker for fructose intake was significantly increased in patients with definite NASH (P = 0.008). Vitamin E was insufficient and significantly lower in children with a higher grade of steatosis.[41] In a dietary intervention, 46 children between 6 and 14 years were provided meals with decreased total fat, carbohydrates and protein for 1 year. Liver fat fraction declined by >9% (P < 0.001) concomitant with decreases in serum aminotransferases (AST P = 0.027, ALT P = 0.024).[42]

The relationship between total fat consumption and hepatic fat show that the amount of fat consumed directly correlates with changes in liver fat and fasting serum insulin levels.[43] Low carbohydrate diets have been shown to decrease ALT levels, improve insulin sensitivity and reduce intrahepatic fat, independent of weight loss.[44–46] Other studies evaluated the association of fructose consumption with development and severity of NAFLD. Patients with NAFLD consume two to three times the amount of fructose compared to controls. These patients express increased hepatic mRNA for fatty acid synthase, an enzyme required for lipogenesis.[47] In a dietary intervention, children and adolescents with NAFLD as well as healthy controls underwent treatment for 6 months with a FRAGILE diet (low fructose/low glycaemic index/low glycaemic load). Absolute fructose intake related strongly to plasma aminotransferase, systolic blood pressure, per cent body fat, insulin resistance and cholesterol level, independent of weight loss. This suggests that modest reduction in fructose intake and glycaemic index improves liver dysfunction and cardiometabolic risk.[48] The role of cholesterol in diet and its association to NAFLD remains indefinite. Some studies demonstrate a relationship to liver disease and other key metabolic parameters while others do not.[49–53]

While lifestyle modifications seem to be beneficial in the treatment of NAFLD in children and adults, it is difficult to achieve and maintain in practice due to obstacles in lifestyle commitments, adherence and perseverance. Understanding the specific underlying issues that lead to noncompliance can help healthcare providers decrease barriers to care. In one study of obese adolescents and their parents, the most commonly cited reasons for noncompliance included feeling frustrated and hopeless. Patients reported poor comprehension of programme goals as well as lack of time and other practical barriers to care. Parents reported patient effort, unhappiness and increased conflict in the household as issues.[54] Adherence may be improved if treatment and research demands are minimised, adolescent motivation is enhanced, sessions are more engaging and treatment is convenient and scheduled around family demands.[25,54,55]

Bariatric Surgery

Bariatric surgery has become an important treatment option for morbidly obese adults who have failed lifestyle changes.[56,57] Given the relation of obesity to NAFLD and the benefit of weight loss in this population, bariatric surgery has been proposed as a potential therapy. Three forms of bariatric surgery are used, restrictive procedures (gastric banding or sleeve gastrectomy), malabsorptive procedures (bilopancreatic diversion) and the combination procedure [Roux-en-Y gastric bypass (RYGB)].[58] A reduction in hepatic injury occurs through the loss of fat mass and systemic inflammation, improvement in insulin resistance, increase in beneficial adipokines and modification of the intestinal microbiome.[58]

Several adult studies, utilising all three methods of bariatric surgery, suggest that surgery results in improvement or resolution of NAFLD.[56,57,59–61] Long-term follow-up data are limited since many patients do not undergo post-surgical liver biopsy and among those that do, the interval to biopsy varies.[13] Mathurin et al., demonstrated significant improvement in prevalence and severity of steatosis and hepatocyte ballooning at 1 and 5 years post-surgery. In those with NASH at baseline, there was significant improvement in steatosis, ballooning and NAS within the first year after surgery and this did not recrudesce between 1 and 5 years. Insulin resistance improved with histology and in the multivariate analysis, refractory insulin resistance predicted persistence of steatosis and ballooning after 5 years.[60] The lack of randomised trials limits the use of bariatric surgery as a therapeutic tool to treat NASH in adults.

While bariatric surgery has been shown to successfully induce weight loss in adolescents, there are no reports on outcomes of patients with NAFLD. Children with the diagnosis of NAFLD tend to be younger and less obese than the adolescents that generally undergo bariatric surgery.[5,62] Among the different bariatric surgeries, there are risks and benefits that are important to consider in paediatric cases. Roux-en-Y gastric bypass in adolescent patients has been shown to be the most efficacious but with an increased risk of nutrient deficiency and protein malnutrition. Gastric banding has the least apparent risk and is reversible and may be a more suitable procedure for the paediatric population; however, long-term results are variable.[63] Official guidelines for bariatric surgery in younger age groups are not standardised and studies on long-term safety and efficacy in the paediatric population remain understudied.[63] In one study, 41 obese adolescent patients (14–17 years) who were followed over 2 years after laparoscopic banding showed decreases in serum aminotransferases [AST (P < 0.001), ALT (P < 0.001)], fasting insulin (P < 0.001), HbA1c (P < 0.001) and lipids [triglycerides (P < 0001), HDL (P < 0.001)].[64] These promising results warrant further confirmation before bariatric surgery can be recommended for children and adolescents as a treatment for NAFLD in the context of obesity and failed lifestyle intervention.


Orlistat. Orlistat is the only FDA-approved drug labelled for weight loss in children. Orlistat inhibits pancreatic lipase to induce fat malabsorption. However, it also interferes with absorption of fat-soluble vitamins and provokes undesirable gastrointestinal symptoms.[40,65] It induces modest weight loss in children in combination with lifestyle changes. In one study only 26% of children treated with orlistat achieved >5% weight loss.[65]

The use of orlistat to promote weight loss in an attempt to improve hepatic steatosis, inflammation and fibrosis, aminotransferase levels and insulin resistance has been explored in adults. This treatment yielded variable results, thus limiting current consideration for its development as an agent to treat NASH in children (Table 1).[27,66]

Insulin Sensitising Agents. Insulin resistance is present in most children with NASH. Thus, pharmacologic interventions to diminish insulin resistance have been pursued as a potential therapeutic strategy (Table 1).[67–70]

Metformin: Metformin reduces hepatic glucose production and increases insulin sensitivity in patients with type 2 diabetes. Several studies in adults failed to show sufficient benefit on hepatic insulin sensitivity, aminotransferase levels or liver histology.[71–76] A recent meta-analysis including three randomised trials of patients with biopsy-proven NASH on metformin with changes in histology as endpoints showed no difference in histological response, ALT or BMI between those on metformin relative to placebo.[77] Metformin has insufficient effect on liver histology to recommend it as a treatment for NASH in adults.[13]

Fewer studies in children are available. One open-label trial evaluated metformin (500 mg taken orally twice a day for 24 weeks) in 10 nondiabetic obese children with biopsy-proven NASH. Serum ALT improved from baseline (P < 0.01) with normalisation of ALT in 40% of patients. Liver fat normalization quantitiated by MR spectroscopy decreased in 90% of patients despite absence of weight loss. The change in liver fat percentage was inversely correlated to the amount of metformin normalised to body mass.[69] The randomised, placebo-controlled TONIC trial evaluated 173 patients ages 8–17 years receiving metformin, vitamin E or placebo for 96 weeks. Forty-one percent of patients treated with metformin showed improvement in hepatocyte ballooning compared to 21% in the placebo group (P = 0.02), independent of improvement in insulin sensitivity. No other histological changes were observed in those treated with metformin compared to the placebo group.[78] Given the paucity of evidence for the use of metformin in adult trials fortified by the absence of meaningful histological changes in children, it is not considered a viable therapeutic option (Table 1).

Thiazolidinediones: Thiazolidinediones such as pioglitazone and rosiglitazone have been evaluated in RCTs to treat NASH in adult patients. Studies show that pioglitazone effectively resolves NASH.[70,79,80] The largest randomised, multi-centre prospective study in adults included 247 nondiabetic adult patients who were administered either pioglitazone (30 mg/day), alpha-tocopherol (800 IU/day) or placebo for 96 weeks. The primary end point, reduction of NAS by ≥2 points without worsening of fibrosis, was achieved in 34% of the pioglitazone group as compared to 19% of the controls (P = 0.04, considered not significant due to multiple comparisons). Significant improvement in hepatic steatosis (P < 0.001) and lobular inflammation (P = 0.004) was found comparing those on pioglitazone to placebo. Resolution of NASH was achieved in a higher number of patients receiving pioglitazone than those receiving placebo (47% vs. 21%, P = 0.001) despite the treated group gaining significant weight relative to controls.[81] Pioglitazone receives support in Practice Guidelines for NASH treatment in adults. However, long-term safety and efficacy of pioglitazone in patients with NASH is not established. Concerns surrounding safety of this class of drugs limits use in children (Table 1).[13]

Anti-oxidants. Increased oxidative stress is regarded as an initiator of hepatocellular injury in a fat-laden liver.[82–85] Thus, measures to reduce oxidative stress or augment anti-oxidant defenses are being pursued (Table 1).

D-alpha-tocopherol (Vitamin E): The randomised controlled PIVENS trial, showed improvement in the primary outcome, designated by changes in liver histology, in 43% of adult patients on vitamin E (800 IU/day) over 96 weeks as compared to 19% of placebo group (P = 0.001).[81] Improvement in ALT levels was predictive of histological improvement and the rate of ALT decline was significantly more rapid with vitamin E treatment compared to placebo (39 vs. 7% at 24 weeks; 48% vs. 16% at 96 weeks). Patients with weight gain were less likely to improve on vitamin E and discontinuation of vitamin E was followed by a relapse and loss of ALT response in 42% of patients.[86] Vitamin E is considered first-line therapy in nondiabetic, noncirrhotic adult patients with biopsy-proven NASH. However, evidence is insufficient to promote its use at this time in those without demonstration of NASH by biopsy.[13]

In children with NAFLD, an early pilot study showed that vitamin E (400–1200 IU/day) given for 2–4 months reduced serum aminotransferase levels independent of changes in BMI and liver echogenicity on ultrasound.[87] The ensuing randomised, placebo-controlled TONIC trial reported significant improvements in histological outcomes in subjects provided oral vitamin E 400 IU twice a day for 96 weeks. This included improvements in hepatocellular ballooning (vitamin E −0.5, 95% CI −0.8 to −0.3 vs. placebo 0.1, 95% CI −0.2 to 0.3) and NAS (vitamin E −1.8, 95% CI −2.4 to −1.2 vs. placebo −0.7, 95% CI −1.3 to −0.2). Fifty-eight percent of patients treated with vitamin E showed resolution of NASH as compared to 28% of the placebo group.[78] Treatment with vitamin E improved serum and urinary markers for oxidative stress and cardiometabolic indicators as well as ALT in children with NAFLD.[88,89]

Paediatric studies using vitamin E have not recognised any serious adverse events associated with vitamin E. In adult studies of vitamin E for treatment of nonliver related conditions, there has been concern that doses of 800 IU/day or higher may increase all-cause mortality and prostate cancer.[90,91] Thus, vitamin E dosed at 800 IU/day offers histological benefits to children with biopsy-proven NASH or borderline NASH although confirmatory studies on clinical outcomes, safety and efficacy are warranted (Table 1).

Cysteamine Bitartrate: Cysteamine bitartrate is an anti-oxidant used to treat cystinosis and acts by scavenging reactive oxygen intermediates and increasing intracellular glutathione.[92,93] In children, the use of cysteamine bitartrate as a treatment for NAFLD is under investigation. An open-pilot study evaluated the effects of enteric-coated cystamine bitartrate on serum ALT over 24 weeks in children with biopsy-proven NAFLD. Seven of 11 subjects completing the study showed a designated sustained response of ≥50% decrease in ALT from baseline. There was a significant increase in serum adiponectin and reduction in leptin and cytokeratin 18 fragments without accompanying changes in BMI (Table 1).[94,95] A fully enrolled multi-centre trial is near completion by the NASH Clinical Research Network investigating the effect of delayed-release cystamine bitartrate in 168 paediatric patients with NAFLD. The primary outcome from this trial is change in NAS over 52 weeks of treatment, comparing treatment to placebo.[96]

Ursodeoxycholic Acid. Ursodeoxycholic acid is a bile acid with demonstrated anti-apoptotic, cytoprotective, anti-oxidative and immuno-modulatory functions. Overall, adult trials indicate that UDCA offers no histological benefit compared to placebo.[97] The largest and most recent randomised controlled trial compared high-dose UDCA (23–28 mg/kg/day) to placebo in adults with biopsy-proven NASH over 18 months. There was no significant change in overall histology on repeat biopsy.[98] These findings were confirmed in a Cochrane Review by Orlando et al., which concluded that there was no clear evidence to support or refute the effect of bile acids in adult patients with NAFLD.[99]

A single randomised controlled trial in children confirmed the lack of efficacy found in adults (Table 1). Thirty-one obese children with NAFLD based on surrogate measures were randomised to no intervention, lifestyle intervention, UDCA (10–12.5 mg/kg/day) without lifestyle intervention or UDCA with lifestyle intervention. The UDCA without lifestyle intervention group showed no significant decrease in aminotransferase level or diminution in liver echogenicity, whereas the group subjected to both interventions showed a significant improvement in ALT and liver ultrasound with weight loss. There were no significant differences in ALT level or liver US change between the lifestyle intervention groups as compared to the UDCA with lifestyle intervention group.[100]

Probiotics and Prebiotics. Studies indicate that gut microflora plays a significant role in the pathogenesis of NASH.[101,102] Small intestinal bacterial overgrowth (SIBO) is present in a large percentage of patients with NAFLD and is associated with steatosis severity.[103,104] Treatment with certain antibiotics reduces the severity of hepatic steatosis in animal models and human subjects.[105,106] Gut microbiota has also been shown to affect fat storage and energy harvesting, inducing insulin resistance, a major risk factor in the development of NAFLD.[107,108]

Thus, the role of gut microbiota in the pathogenesis of the disease indicates that manipulation of gut microbial ecology may ameliorate NAFLD.[107] Probiotics show promise as a new therapeutic agent in NAFLD, as they exhibit anti-inflammatory properties, inhibit bacterial adherence to intestinal mucosa and epithelial invasion, enhance intestinal permeability and diminish bacterial translocation from the gut.[109] Animal models demonstrate that probiotics are useful in mitigating features of NAFLD such as reduction in aminotransferase levels, improvement in insulin resistance, diminution of inflammatory markers and improvement in liver histology, specifically attenuation of steatosis and fibrosis.[110–114]

Probiotics have been assessed as a treatment for NASH in children and adults. There have been multiple promising RCTs assessing either histological endpoints or surrogate serum markers. A double-blind randomised controlled trial in adults with biopsy-proven NASH showed that Lactobacillus bulgaricus and Streptococcus thermophilus reduced liver aminotransferases significantly without changes in other anthropometric or cardiometabolic parameters.[115] Subsequently, 55 adults were randomised to either a mixture of probiotic with prebiotic (Bifidobacterium with fructose oligosaccharides 2.5 g) and lifestyle intervention or lifestyle intervention alone for 24 weeks. Patients in the treatment group showed reduction in AST (−69.6 vs. −45.9 IU/mL, P < 0.05), insulin resistance (HOMA-IR −1.1 vs. −0.6, P < 0.001), LDL cholesterol (−0.84 mmol/L vs. −0.18, P < 0.001), TNF-alpha (−0.45 vs. −0.12 ng/mL, P < 0.001), CRP (−2.9 vs. −0.7 mg/L, P < 0.05) and serum endotoxin (−45.2 vs. −30.6 pg/mL, P < 0.001). There was a decrease in NAS and steatosis (both P < 0.05).[116] Another study assessed change in intrahepatic triglyceride content by proton magnetic resonance spectroscopy, aminotransferase levels and metabolic profile over a 6-month treatment period with Lepicol© probiotic formula containing intestinal bacteria including Lactobacillus plantarum,Bifidobacterium bifidum and Lactobacillus deslbrueckii. Intrahepatic triglyceride decreased in the probiotic group compared to placebo (P = 0.034) with six patients demonstrating >30% reduction. There was also greater reduction in serum AST compared to placebo (P = 0.008). There were no changes in BMI, waist circumference, glucose or lipid levels.[117]

In the paediatric population, a double-blind randomised controlled trial was performed in 20 children with persistently elevated aminotransferases and echogenic livers. Subjects were treated with Lactobacillus rhamnosus strain GG (12 billion CFU/day) or placebo for 8 weeks. The treatment group showed improvement in serum ALT (P = 0.03) and anti-peptidoglycan polysaccharide antibody levels (P = 0.03), a surrogate marker for SIBO. These results were independent of changes in BMI.[118] Histological confirmation of benefit with probiotics needs to be assessed in RCTs.

A meta-analysis summarised these four completed RCTs listed above and concluded that the use of Lactobacillus,Bifdobacterium and Streptococcus significantly decreases ALT, AST and lipid profiles as well as serum insulin levels.[119] Given the limited experience, appropriately designed trials with justified power analyses, long-term follow-up and histological assessment will be necessary to justify the type, amount and duration of therapy with probiotics for paediatric NAFLD (Table 1).

Prebiotics are nondigestible food ingredients that stimulate growth and modify the metabolic activity of microflora. Prebiotics have also been shown to augment the action of probiotics.[120] Animal models have shown that prebiotics alter the composition and normalise the gut microbiota, improve glycaemic control, hepatic cholesterol and triglyceride accumulation.[121] Prebiotics represent another therapeutic option to be tested in children (Table 1).

Polyunsaturated Fat Supplements. Omega-3 fatty acids are essential fatty acids mediating inflammation. Omega-3 polyunsaturated fatty acids are efficacious in the prevention and therapy of cardiovascular disease, dyslipidaemia and metabolic syndrome.[122–124] Individuals with NAFLD have a lower intake of polyunsaturated fat as compared to controls and given the link between NAFLD, cardiovascular disease and metabolic syndrome, it is thought that supplementation with omega-3 fatty acids could have therapeutic utility in NAFLD.[51,124–126] Studies in animal models and adults with NAFLD demonstrate the anti-inflammatory and insulin sensitising properties of omega-3 with improvement in hepatic steatosis, serum aminotransferases, triglycerides and measures of insulin sensitivity.[127–133] However, while these results are promising, adult studies lack RCTs of adequate size and duration with histological end points. Evidence remains premature to recommend omega-3 fatty acids for the treatment of NAFLD.[134]

Paediatric trials have used supplementation with docosahexaenoic acid (DHA), an omega-3 fatty acid, for the treatment of NAFLD. A randomised controlled trial of 60 children with biopsy-proven NAFLD were treated with either 250 mg DHA, 500 mg DHA or placebo over a 6-month period. Liver steatosis was assessed by ultrasound. Improvement was reported for both DHA-treated groups (OR = 0.01, 95% CI 0.002–0.11, P < 0.001; OR = 0.04, 95% CI 0.002–0.46, P = 0.01, respectively) with no dosage-related differences. There was no effect on serum aminotransferases or BMI.[135] The effects reportedly persisted for the 24-month study duration.[136] DHA was well tolerated. In a follow-up study, liver histology in 20 children before and after treatment with DHA for 18 months was assessed. Improvement in NAS (P < 0.01), ballooning (P < 0.001), steatosis (P < 0.001) and lobular inflammation (P < 0.05) were reported; however, there was no significant change in fibrosis.[137] Absence of a placebo group in this study should prompt future studies to examine the effects of DHA in RCTs (Table 1).

Pentoxyfylline. Pentoxifylline (PTX) is a phosphodiesterase inhibitor that antagonises effects of TNF-α, a pro-inflammatory cytokine implicated in NASH pathogenesis.[138] Zein et al. randomised 55 adults with NASH to treatment with PTX (1200 mg/day) or placebo for 1 year. Fifty percent of patients treated with PTX showed reduction in NAS by ≥2 points compared to 15% of placebo (P = 0.01) with improvement in steatosis (P < 0.001), lobular inflammation (P = 0.02) and fibrosis (P = 0.04).[139] A smaller randomised controlled trial was unable to show a significant difference between the PTX and placebo group in regard to aminotransferase level and histology; however, patients who received PTX did show a significant decrease in ALT and AST, steatosis, and cellular ballooning from baseline (all P ≤ 0.05).[140] A recent meta-analysis of five RCTs that utilised PTX evaluated efficacy in improving NASH. Overall, it was found that patients showed a significant decrease in body weight (P = 0.04), serum aminotransferases (P < 0.001), lobular inflammation (P < 0.001) and TNF-α (P = 0.007). On histology, NAS and lobular inflammation improved (each P < 0.001); however, steatosis, ballooning and fibrosis did not. One limitation to this meta-analysis was inclusion of trials comparing PTX to UDCA and not just to placebo. Thus, the results may underestimate the potential therapeutic effects of PTX.[138] Given the experience in adults, RCTs of PTX in paediatric NAFLD patients should be considered (Table 1).

Farnesoid X Receptor Agonists. The nuclear bile acid receptor, farnesoid X receptor (FXR), is expressed in the bowel and liver and mediates control over lipid and glucose homoeostasis, bacterial flora growth and inflammation.[141,142] Research has been primarily focused on animal models. Studies using the FXR-α agonist, obeticholic acid (OCA), in humans will soon be reported. In a phase II clinical trial, 60 diabetic patients with NAFLD, diagnosed by US, underwent treatment with OCA (25 mg or 50 mg) for 6 weeks. Treated subjects showed significant improvement in insulin resistance, weight loss and serum GGT relative to controls. No serious adverse events were found.[142] A Phase IIb-III double-blind randomised controlled multi-centre trial is now complete, modified after being found to meet a pre-designated interim primary endpoint demonstrating efficacy of treatment over placebo (P ≤ 0.003). Patients received OCA 25 mg orally per day over 72 weeks. The primary endpoint was response in the active treatment group compared to placebo, with response defined as a decrease in NAS by ≤2 points with no progression in fibrosis.[143] Given the novelty of FXR agonists as a therapeutic class, potential trials in children await demonstration of safety and efficacy in adult studies (Table 1).

Toll-like Receptor Modifiers. Toll-like receptors (TLR) have been implicated in the pathogenesis of NAFLD.[144] TLR stimulation by TLR ligands, such as lipopolysaccharide, in the setting of bacterial overgrowth and augmented intestinal permeability results in activation of the innate inflammatory response.[145–147] TLR agonists that mitigate the inflammatory response may be useful. Studies in both animals and humans are needed.

Use of FXR and TLR agonists are anticipated to provide more specific and tailored therapeutic options that could change the management of NAFLD (Table 1).

Current Issues and Future Directions

A uniformly effective treatment strategy for children with NAFLD failing appropriate lifestyle changes is lacking. Ideal therapy will ultimately abate NASH and other obesity-related comorbidities, be affordable, easily tolerated and easy to administer. Therapeutic discovery could be accelerated with improved techniques for screening that requires development of noninvasive diagnostics, increased awareness of NASH as a comorbidity of obesity and development of evidence-based screening guidelines. Hindrances to development include objective validated inclusion criteria and outcome measures. Rigour in study design will facilitate adoption of promising therapeutics.

Study Design and Patient Selection. A standardised approach to subject inclusion and exclusion criteria should be a focus in future studies as current patient selection is highly variable. Patients are often defined as having NAFLD based on elevated liver transaminases or liver ultrasound without a histological diagnosis of NAFLD. These diagnostic tools are not as reliable and may overestimate or underestimate therapeutic effects. Even when NAFLD histology is used as inclusion criteria, patients often have a wide-range of NAFLD histology from simple steatosis to steatohepatitis. These histological gradations likely respond differently to treatment and may be a confounding factor in studies. Studies that use surrogate markers should not be compared to those that use histological endpoint, as occurs in many meta-analyses. In addition, a specific range for BMI should be considered as an inclusion criteria; obese patients with NAFLD are very different from patients with a normal BMI and likely have a different response to lifestyle intervention and pharmacologic treatment.

Trial design often differs across studies, especially in children where study size is small. Open-label and uncontrolled studies lack the scientific rigour that exists in randomised clinical trials. While randomised trials are imperative, the specific type of trial utilised is important. Phase IIb trials which assess the efficacy of a pharmacologic therapy and show a direct dose–response relationship in patients as well as Phase III trials which may be of longer duration and include a larger clinical population should be used.

In addition, length of treatment often varies among studies, making it difficult to interpret results. The time by which it takes for histological damage to be reversed in NAFLD is unknown, thus, in many studies lack of change may be due to under treatment. The optimal dosage that is both efficacious and safe is not known for the therapeutics being studied, adding to the variability in studies and perhaps under treatment of patients. Also, standardised definitions of side effects and clinical phenotypes of toxicity would allow safety data from multiple trials to be combined. An understanding of the adverse events that occur at specific dosages would enable optimisation of therapy. Compliance issues, especially in the adolescent population, make intervention efficacy harder to interpret.

Outcome Measures. Measuring outcome after treatment is of great interest. Currently a variety of outcome measures are used in studies. Determination of a measure that is of high sensitivity and specificity while balancing the degree of invasiveness, expense and practicality is important.

Liver biopsy is considered the gold standard for NAFLD diagnosis and histological change is used as an outcome measure in treatment trials. Liver biopsy allows for the exclusion of other causes of liver damage and an estimation of the severity. However, along with these advantages, biopsy is invasive, expensive, associated with procedural risks, not representative of the entire liver parenchyma and up to the subjective interpretation of a pathologist.[8,148,149] There are various classification scales used in studies, some of which are not appropriate for the histological pattern that exists in paediatric patients. In addition, what is defined as 'histological improvement' after therapy is often variable across studies.[150]

Alternative non-invasive biomarkers for outcome measures are under study. The Paediatric NAFLD Fibrosis Index which uses age, waist circumference and triglyceride level as well as the Enhanced Liver Fibrosis scale which looks at three serologic measures of the extracellular matrix (hyaluronic acid, amino-terminal propeptide of type III collagen and TIMP1) have both been shown to have a fairly high degree of sensitivity and specificity in predicting fibrosis.[13,151,152] When these two measures are combined, liver fibrosis is predicted more accurately than the use of either test alone.[153] One serum marker that has been under study is caspase-generated cytokeratin-18 fragment. It measures apoptosis level and thereby predicts fibrosis stage and NAFLD severity. Studies in children have shown that cytokeratin-18 is elevated in patients with NAFLD as compared to controls and higher depending on the degree of NAFLD severity (i.e. NASH vs. NAS <3) and fibrosis.[154] Although it has been thought to be an accurate predictor of NAFLD, in a recent study, it was shown to be no better than alanine aminotransferase at predicting histological improvement.[155–157] Other markers for oxidative stress and inflammation are under study.[158–160]

Imaging as a diagnostic tool and outcome measure has also been the subject of study. Liver ultrasound is often used in studies. It is inexpensive, widely available and useful in quantifying steatosis. However, it is user-dependent and interpretation is subjective. Estimated sensitivity and specificity is 60–94% and 84–100%, respectively. However, sensitivity decreases when the liver contains less than 30% fat or when the BMI is over 40 kg/m2. In addition, a negative scan cannot rule-out the presence of steatohepatitis or fibrosis.[161–163] While CT has high sensitivity and specificity, it has limited utility due to radiation exposure, especially in the paediatric population.[161,164] MRI has the greatest accuracy, but its high cost outweighs potential benefit.[10]

Transient elastography (TE) uses a transducer probe that emits a low-frequency vibration and calculates the speed of the propagating mechanical wave through the liver, directly relating to tissue stiffness and degree of fibrosis.[97,165] Studies in children have shown that it is an accurate and reproducible variable for measuring fibrosis.[166–169] In one study by Lee et al., TE was superior for predicting advancing fibrosis as compared to other biomarkers. Even combination of biomarkers and TE was not superior to TE alone.[170] TE is difficult to perform in obese patients as subcutaneous fatty tissue attenuates elastic shear wave, and may not be sufficient in monitoring the progression of liver disease in children.[171] Recently, vibration-controlled TE has been studied, a new technology which accounts for the attenuation effect in obese patients; however, this technique requires further investigation prior to clinical use.[172,173] Magnetic resonance elastography has also started to be used in NAFLD detection. Studies have suggested that it can better discriminate between mild and moderate to severe fibrosis than TE.[174,175] It has also been shown to discriminate between simple steatosis and NASH with high accuracy.[176]

Genetic Testing. Understanding the role of individual genetic susceptibility in NAFLD likely will inform future personalised therapeutics. In recent years, genetic studies have demonstrated that select single nucleotide polymorphisms in genes involving lipid metabolism (Lipin 1 – LPIN1; patatin-like phospholipiase domain containing-3 – PNPLA3), oxidative stress (superoxide dismutase 2 – SOD2), insulin signalling (insulin receptor substrate-1 – IRS-1) and fibrogenesis may predict risk of development and progression of NAFLD (Kruppel-like factor 6 – KLF6).[177] In particular the adiponutrin gene, patatin-like phospholipase domain containing protein 3 (PNPLA3), was linked to early onset and greater histological severity in patients with nonalcoholic fatty liver disease.[178,179] Two hundred twenty-three children were included in a genome-wide association study, which showed a high-risk allele (rs738409G) in the PNPLA3 gene associated with steatosis, portal fibrosis, lobular inflammation, NAS and fibrosis as well as early onset of NAFLD.[178] In another study, the same high-risk allele was associated with greater susceptibility to hepatic steatosis in obese children.[179]

Nobili et al. evaluated the combined effect of four genetic polymorphisms and found that multi-SNP analysis may be useful in predicting development of NASH in obese children with elevated transaminases.[180] Future identification of individual genes that dictate predisposition, outcome and treatment response will ultimately play a role in pharmacogenomic selection.