Factors to Consider in Development of Drugs for Pediatric Nonalcoholic Fatty Liver Disease

Nonalcoholic fatty liver disease (NAFLD) has quickly become the most common chronic liver disease in children in many countries around the world.^1–5 Estimates in Europe range from 1.3% to 22.5% of children being affected.^2,6–9 In the United States, 9.6% of all children and 38.0% of obese adolescents are estimated to have NAFLD.^10,11 In China, 45% of obese adolescents are estimated to have NAFLD.¹² Certain populations have notable predisposition; for example, in the United States, the highest prevalence is found in Mexican Americans.^11,13 Children have the full range of disease severity from mild steatosis alone to steatohepatitis with fibrosis to end-stage cirrhosis. In both the United States and Europe, fibrosis was observed in approximately 70% of biopsied pediatric NAFLD cases.^14,15 Early stage fibrosis (stages 0–1) was observed frequently and advanced fibrosis (stages 2–4) was found in 16%–31% of children with NAFLD who underwent liver biopsy.^14,15

This overview is a product of The Liver Forums’ Pediatric Working Group (Appendix A), a pediatric focused group within a multi-stakeholder collaboration of academic, industry, patient, and regulatory experts. Part of the Forum for Collaborative Research, the mission of the Liver Forum is to accelerate drug development for NAFLD by increasing the quality, efficiency, and output of clinical trials. At the time of this writing, there were no approved pharmaceutical treatments for either adult or pediatric NAFLD, resulting in an unmet need for medical treatments. NAFLD is rapidly increasing as an indication for liver transplant, particularly for people <40 years of age.¹⁶ Without effective treatments, children with nonalcoholic steatohepatitis (NASH) are at risk of developing cirrhosis¹⁶ and liver-related mortality in early adulthood. The overall prevalence of NASH is even higher in patients with associated comorbidities such as type 2 diabetes.¹⁷

The Mandate for Pediatric NAFLD Trials

The ethical basis for the US Food and Drug Administration (FDA) required initial pediatric study plan (iPSP)¹⁸ and the European Medicines Agency (EMA) equivalent pediatric investigation plan (PIP) is that the inclusion of children in research promotes their safety and well-being. When there are approved medications in adults for a disease that occurs in children but no clinical trials to support pediatric dose considerations, efficacy, or safety, physicians often prescribe the medications based on data collected in adults. Such off-label use may ultimately prove harmful in children owing to delays in clinical trial development of appropriate treatments, as well as inappropriate dosing and unknown safety profile of unapproved treatments.¹⁹

In the United States, congress has enacted the Pediatric Research Equity Act (PREA) and Best Pharmaceuticals for Children Act (BPCA) that are intended to promote the development of safe and efficacious drugs in children. PREA requires all applications submitted to FDA under section 505 of the Act (21 U.S.C. 355) for a new active ingredient, new indication, new dosage form, new dosing regimen, or new route of administration to contain a pediatric assessment unless the applicant has obtained a waiver or deferral in an iPSP agreed-upon by FDA’s Pediatric Review Committee. PREA applies to drugs or biological products developed for diseases that occur in both adults and children. Products intended for pediatric-specific indications are subject to the PREA requirement only if they are initially developed for a subset of the relevant pediatric population. PREA does not apply to drugs that have been granted orphan designation for a particular indication. Under the FDA Safety and Innovation Act, signed into law on July 9, 2012, PREA includes a provision that requires manufacturers of drugs subject to PREA to submit an iPSP early in the drug development process. The intent of the iPSP is to identify needed pediatric studies early in drug development and begin planning for these studies. Sponsors must submit their iPSP before submission of the required pediatric assessments and no later than 60 calendar days after the date of the end-of-phase II meeting.²⁰

The BPCA was enacted on January 4, 2002. The BPCA²¹ reauthorized and amended the pediatric exclusivity incentive program by granting an additional 6 months of marketing exclusivity and creating funding mechanisms for pediatric studies that had not been conducted voluntarily under previously approved new drug applications.

In Europe, there are also regulatory requirements for promoting safe and efficacious drug development in children. In 2006, the European Commission passed the Pediatric Regulation No 1901/2006 that was later amended to 1902/2006 and required pediatric studies for drug development.²² The European Pediatric Regulation requires that pharmaceutical companies submit a PIP for all new products and any line extensions for existing products. Typically, the PIP provides the design and the timelines for the studies.²² The PIP is submitted to the Pediatric Committee, an expert committee of the EMA that reviews PIPs, waivers, and deferrals of pediatric studies. Once the PIP is agreed upon by the Pediatric Committee, it is binding and issued by the EMA. The PIPs are intended to cover all children from birth to 18 years of age; however, a waiver from the requirement for certain age groups can be granted when the product is likely to be ineffective or if the disease does not occur in pediatric patients in an age group.²² Similar to iPSPs, PIPs are intended to identify the need for pediatric studies early in the development process and increase availability of safe and effective medications for children. Although the scope and procedures for the regulations in the United States and Europe differ, the agencies collaborate to be able to share results from pediatric development and limit repetition of studies in children.

Most ethical guidelines for research posit that children are a vulnerable population, with a higher benefit–risk burden than adults. Although pediatric plans for NAFLD therapeutics are required, there are also additional safeguards required for children by both the FDA and EMA. For children, studies with greater than minimal risk may be approved when there is the prospect of direct benefit to the individual subjects (Department of Health and Human Services, FDA, 21 CFR Sec50.52). If there is greater than minimal risk but no prospect of direct benefit to the subjects, a study may be ethical for children only if the risk represents only a minor increase over minimal risk, or the risk is from an intervention or procedure that is commensurate with those inherent to their disease/situation, or the intervention or procedure is likely to yield generalizable knowledge about the disease that is vital for the understanding or amelioration of the disease. These regulations are in place to protect children while balancing the need for studies to advance effective therapeutics for children’s diseases.

Defining At-Risk Populations

Risk factor identification in disease progression (especially fibrosis progression) is important because NAFLD (in both adults and children) has a wide phenotypic range at presentation and variable progression. Long-term nationally representative data in US adults having severe hepatic steatosis as well as elevated alanine aminotransferase (ALT), aspartate aminotransferase or gamma-glutamyl transferase were associated with increased liver disease mortality.²³ In adults, type 2 diabetes and cardiovascular disease are also frequent adverse clinical outcomes associated with NAFLD²⁴ and there are few data suggesting that increased incidence of type 2 diabetes is the most rapid adverse clinical outcome^17,25 in children. One of the most challenging knowledge gaps is the lack of longer term natural history information, which complicates benefit–risk determinations in the development of therapeutics for pediatric NAFLD. Available shorter term data from the Treatment of NAFLD in Children (TONIC)²⁶ trial showed that approximately 60% of children maintained or progressed in hepatic fibrosis stage over 24 months in all treatment groups, including placebo. In a small clinical cohort, 39% of children (7/18) had progression of fibrosis after an average of 28 months.²⁷ There are 2 longer term studies. Sixty-six children with NAFLD (mean age, 13.9 ± 3.9 years) were followed for ≤20 years. Five underwent repeat liver biopsies showing progression of fibrosis in 4 of the children. Two children died and 2 underwent liver transplantation.²⁸ In another retrospective study of 44 patients diagnosed with NAFLD as children, development of type 2 diabetes was the most frequent adverse clinical outcome, occurring in 30%.²⁵ Further long-term longitudinal cohort studies are needed to assess 5-year, 10-year, and longer outcomes in pediatric patients with NAFLD and NASH and to identify risk factors for progression to associated comorbidities such as type 2 diabetes, cardiovascular disease, cirrhosis, hepatic decompensation, and hepatocellular carcinoma.

Histopathologic Phenotypes of Pediatric NAFLD

Similar to adults, NAFLD in children has a wide range of histologic phenotypes. In 2005, Schwimmer et al published the first description of a particular subset of pediatric histology in NAFLD and described it as a periportal pediatric pattern rarely found in adults.²⁹ In adults, the pattern of injury is typically centered in zone 3, with the combination of steatosis, parenchymal inflammation, ballooning and perisinusoidal fibrosis. A zone 1 pattern (often called the pediatric pattern) is more prevalent in children <12 years of age, whereas teenagers are more likely to have patterns similar to adults.³⁰ Because of the absence of classical ballooning hepatocytes in the pediatric pattern, pediatric biopsies often do not fulfill the classical criteria for adult NASH and are characterized as borderline NASH zone 1. Although the natural history of the pediatric pattern is not yet understood, it does not appear to be less severe. In the literature, there is some overlapping use of the terminology regarding phenotypes and histopathologic descriptions; these definitions are listed in Table 1.

Table 1.

Histologic Phenotypes of NAFLD

Phenotypes	Definitions
NAFLD	Inclusive term referring to the full spectrum of disease
	Indicates fatty infiltration of the liver in the absence of significant alcohol, genetic diseases or medications that cause steatosis
	Fatty infiltration is typically defined as fat ≥5% of the liver
NAFL	Steatosis without specific changes to suggest steatohepatitis
NASH	Hepatic steatosis with inflammation, with or without fibrosis, further subdivided as: Zone 3 centered injury pattern or confluent pattern with hepatocyte ballooning injury. Also called type 1. Zone 1 (portal) centered injury pattern with periportal to panacinar steatosis, portal inflammation, with or without hepatocyte ballooning injury. Also called type 2. Pediatric NASH – umbrella category that includes zone 3 NASH, zone 1 NASH and borderline NASH found in a child (age ≤17 years)
NAFLD with fibrosis	NAFL or NASH with periportal, portal or sinusoidal fibrosis
NAFLD with cirrhosis	NAFL or NASH with bridging fibrosis with architectural distortion and nodule formation
Cryptogenic Cirrhosis	Bridging fibrosis with architectural distortion and nodule formation in the setting of obesity and the absence of any other identified liver disease

NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis.

The clinical phenotypes of NAFLD in children remain not well-defined. This area requires additional research, including an exploration of whether future clinical adverse outcomes can be predicted by pediatric histologic phenotypes.

Clinical Trial Designs in Previous Pediatric NAFLD Trials

Studies in pediatric NAFLD have ranged from small pilot studies to large, randomized and controlled clinical trials.^26,31–34 Although the outcomes and endpoints used in these studies varied, many of the early phase trials focused on rapidly changing and easily measured biomarkers such as serum ALT.

To facilitate assessment of histologic disease severity, a NAFLD Activity Score (NAS) was developed by the NASH Clinical Research Network and is now commonly used for evaluating the treatment response through assessment of histological change over time in NAFLD clinical trials,^35,36 including pediatric studies.^26,37 In NAS, 3 components are scored (steatosis 0–3, lobular inflammation 0–3, and hepatocyte ballooning 0–2) for a possible total score of 8. In children, hepatocyte ballooning is less common, particularly in younger children and early teenagers. In younger children, the portal pattern of inflammation predominates and lobular inflammation scores are lower compared with adult studies. For example, in the FLINT cohort study.³⁸ the mean for lobular inflammation in adults is 1.8, compared with1.6 in the pediatric TONIC^26,38 trial, although the eligibility criteria were different. Further, in these trials, hepatocyte ballooning in adults was 1.3 ± 0.7 and in the pediatric cohort 0.6 ± 0.8.

A frequently used primary outcome for early phase clinical trials in NAFLD is a reduction of NAS score of ≥2 with no worsening of fibrosis.^39,40 In the adult FLINT trial, the mean NAS of the participants, who were required to have NASH for inclusion, was 5.1 at baseline, compared with 4.6 in children in TONIC. In Cysteamine Bitartrate Delayed-Release for the Treatment of NAFLD in Children (CyNCh), the mean NAS was 4.7.⁴¹ For CyNCh, having a NAS of ≥4 was an inclusion criterion and the primary outcome was a decrease in NAS of ≥2 with no worsening of fibrosis.⁴¹

The somewhat lower baseline severity and lower frequency of ballooning in children could make a NAS change of ≥2 difficult to achieve in pediatric trials. For example, if a child’s histology has no ballooning (which is a common occurrence), a 2-point improvement in NAS can only be achieved through a decrease in either steatosis or lobular inflammation or by a 1-point change in both. Given that not all drugs have mechanisms of action that impact steatosis (reduction of liver fat is thought to be less likely to predict clinical outcomes⁴²), the 2-point improvement in NAS would need to be achieved in the lobular inflammation category. This change is unlikely in children who have a portal predominant inflammation pattern and minimal lobular inflammation. Hence, further research is needed to determine the utility of NAS in pediatric trials and to test the relationship of “change in NAS” to long-term clinical outcomes. In TONIC, although the primary endpoint of a change in ALT was not met, an improvement of NAS by ≥2 points without worsening of fibrosis was achieved in the vitamin E treatment arm compared with placebo (P = .02).²⁶

At this time, neither the United States nor the EU accepts change in NAS alone as an endpoint for trials to support a marketing/registration application for NASH. There are many different endpoints recommended for early phase proof-of-concept and dose-ranging trials that should be primarily based on the drug mechanism of action and target population.

Looking Forward: Clinical Trial Participants and NAFLD Subpopulations

NAFLD versus Non-NAFLD

The diagnosis of NAFLD is a clinical determination based on exclusion of other chronic liver diseases and confirming the presence of steatosis. At a minimum, the diagnosis of NAFLD requires ruling out chronic alcohol use, medications known to cause steatosis, other chronic conditions causing fat deposition (uncontrolled diabetes, autoimmune hepatitis, starvation, celiac disease), metabolic diseases (cystic fibrosis, mitochondrial diseases, lysosomal disorders, Wilson’s disease, alpha-1 antitrypsin deficiency), hepatitis C, and documenting the presence of steatosis in the liver.⁴³ The diagnosis of NASH requires ruling out these diseases and a histopathologic confirmation of steatohepatitis.

Noninvasive modalities may be used to assess NAFLD, but most imaging modalities, such as ultrasound examination or magnetic resonance elastography, although cleared by the FDA, are not approved for the diagnosis or staging of specific disease indications. A combination of noninvasive imaging methods along with biochemical and clinical assessments are often used in clinical practice as well as early phase clinical trials to establish a probable or likely diagnosis of NAFLD/NASH. However, at this time, noninvasive imaging methods to assess inflammation remain in the early phases of development.

Histopathology remains the current reference standard for NAFLD and is used as a surrogate endpoint for clinical trials in NASH along with clinical benefit endpoints (ie, improvements in mortality, transplant, and decompensation rates). Liver biopsy is generally considered to have the same risk in children as in adults and has the benefit of assessing the severity of inflammation and fibrosis as well, although the procedure has the potential of increased risks to the patient compared with noninvasive imaging. Steatosis on biopsy is traditionally assessed by counting the percent of hepatocytes with macrovesicular fat and if >5% is deemed abnormal.^44,45 Noninvasive methods, including magnetic resonance imaging (MRI), novel ultrasound-based techniques, and other imaging and non–imaging-based tests, are in development and are being tested in pediatric cohorts. Accumulating literature evidence for MRI assessment of protein density fat fraction suggests precise and accurate measurement of liver fat content.^42,46 The upper limit of normal for MRI fat fraction is debated⁴⁷ and varies based on the technique used and population studied. Unlike in adults, imaging or histologic findings have not been shown to be prognostic in children, partly because of the lack of available natural history data.

Further research to validate noninvasive biomarkers is urgently needed for adult and pediatric liver diseases. Clinical trials in adult and pediatric populations are encouraged to consider the feasibility of biopsies performed concurrently with noninvasive tests for assessing a potential correlation of results obtained using both modalities.

Pediatric NAFLD Subpopulations

By Histology.

Histologic categories in pediatric NAFLD are well-defined and described in Table 1. At this time, NASH can only be determined with liver histopathology. For practical reasons, in this commentary, all of the NASH categories for children have been combined into one and called pediatric NASH. Children may present with fibrosis owing to NAFLD at any age and it may be advanced at presentation. At this time, fibrosis can only be accurately staged with liver histopathology.

There are few long-term data on the natural history of NAFLD progression in children; however, data in adult patients demonstrate that fibrosis is the strongest predictor of long-term outcomes.^48,49 The pathophysiology of other liver diseases in children also support that fibrosis is the main driver of clinical outcomes; it is likely that the cohort of children who have inflammation and fibrosis owing to NAFLD may benefit the most from the development of pharmacotherapies.

By Biochemical Liver-Associated Enzymes.

Serum chemistries to assess liver dysfunction demonstrate that ALT is the most sensitive marker for detecting NAFLD, although many studies have shown that aspartate aminotransferase and gamma-glutamyl transferase are better predictors of the presence of inflammation and fibrosis. Although ALT is typically elevated in the setting of NASH, it is not sufficiently specific to predict clinical outcomes. Various cutoffs for ALT minimum levels have been used as inclusion criteria in pediatric NAFLD trials ranging from 45 to 60 IU. Importantly, ALT levels for adolescents in pediatric studies must be selected using the upper limit of normal based on the revised sex-specific recommendations (23 for girls and 26 for boys, ages 12–17 years).⁵⁰

For early phase, proof-of-concept studies, ALT is a reasonable outcome and the study should focus on participants with a baseline ALT value high enough to show significant decreases. However, a recent randomized clinical trial evaluating liver biopsy and ALT contemporaneously found that changes in ALT were significant, whereas dichotomous changes in liver histology after 1 year were not.⁴¹ Some of this is due to the complexity in histologic change and some is due to comparing a dichotomous change to a continuous variable. The TONIC study in children was longer (2 years) and found that ALT change did predict histology change after 2 years.⁵¹ Thus, reduction of serum ALT seems to be useful in early phase studies, although later phase warrant evaluation of histologic endpoints.

By Imaging.

Categorizing patients with NAFLD based on the volume of fat present in the liver could be useful for studies in which the target is steatosis reduction. Some phase I and II trials have used a decrease in steatosis as the primary endpoint; however, there are inadequate data, at this time, to support that steatosis reduction leads to a clinically meaningful benefit in phase III trials. The amount of fat on MRI can also be classified as mild, moderate, or severe, similar to the scoring of steatosis on a liver biopsy or can be reported as a percent of the total liver fat. What is not established is the amount of change in hepatic fat that is clinically relevant in predicting and improving long-term outcomes in NAFLD.

Noninvasive measurement of fibrosis is rapidly advancing and use of several modalities are being explored in pediatric NAFLD including sheer wave elastography, transient elastography, and magnetic resonance elastography. Further validation of these tests in children is needed before relying on these to assess change in fibrosis in clinical trials.

Associated Comorbidities

Diabetes.

NAFLD in children is associated with prediabetes and type 2 diabetes and there seems to be a strong relationship in some racial/ethnic groups. In biopsy-proven NAFLD (defined as ≥5% steatosis on histology), the estimated prevalence was 23% prediabetes and 6.5% type 2 diabetes.¹⁷ Although patients with diabetes are sometimes excluded from NAFLD clinical trials, it may be that the coexistence of NAFLD and diabetes is a specific phenotype of disease that should be studied as a subgroup in clinical trials.

Hypertension.

Arterial hypertension can also be associated with NAFLD and should be assessed during clinical trials.

Other Considerations for Pediatric NAFLD Clinical Trials

Age.

NAFLD is rarely diagnosed in very young children (<6 years) and this subgroup is generally excluded from pediatric NAFLD trials. The inclusion criterion for age in pediatric NAFLD trials generally ranges from 7 years to <18 years.^32,37,41 Children <13 years of age are more likely to have a periportal pattern of inflammation on histology, the presence of which becomes rare by late adolescence (>17 years). Thus, it is likely that pediatric trials will benefit from age group stratification such as 7–11 and 12–17 years inclusive because these subgroups may exhibit different histologic characteristics and potential differential response to treatment.

Sex.

NAFLD is diagnosed more often in boys than girls (2.5 to 1.0). However, the pathophysiology of NAFLD in boys and girls seems to be similar and no difference in treatment response was observed in TONIC²⁶ or CyNCh.⁴¹ Both sexes are typically included and assessed as subgroups to examine differential response in clinical trials.

Puberty.

Pubertal hormones may affect NAFLD; one study suggests that NAFLD is less severe in children who are peripubertal and postpubertal compared with prepubertal.⁵² Puberty stage (Tanner stages) should be documented carefully in pediatric clinical trials to allow subgroup analysis of the potential for puberty stage to impact treatment response.

Drug Development in Pediatric NAFLD

Pharmacokinetics.

An important consideration for development of pediatric NAFLD therapeutics is pharmacokinetics. In children, growth and development, in addition to weight, may impact the pharmacokinetics of the investigational agent.

When children with obesity were given recommended doses of vancomycin,⁵³ the mean volume of distribution and the mean clearance was reduced by 81% and 80% respectively, relative to controls (children with normal weight). The clearance of fentanyl (lipophilic drug) is slower in children with obesity relative to non-obese children.⁵⁴ In contrast, metformin clearance was found to be increased in children with obesity.⁵⁵ Therefore, obesity may be an important consideration for dose selection in pediatric NAFLD clinical trials; hence, formal pharmacokinetic studies in children with NAFLD are needed to establish evidence-based dosing.

Early Phase I and II Trials

In early phase trials, endpoints that allow rapid observable changes (over days/months), are measurable with low risk and are reasonable in burden to participants are preferable. Because of this, almost all early phase trials in NAFLD use biomarkers in place of histology.

Biomarkers of improvement in NAFLD used in previous pediatric NAFLD trials include sustained reduction in ALT to ≤50% of baseline or <40 U/L, and normalization of ALT. Although a change in ALT has not been validated as a predictive biomarker, it remains a reasonable biomarker for early phase pediatric NAFLD trials to assess efficacy, in part because of the long clinical tradition of using ALT levels to assess treatment response in many liver diseases, including NAFLD.

Changes in imaging and other novel circulating biomarkers may be proposed for early phase trials; to date, these biomarkers have largely been used as exploratory endpoints. As the level of evidence grows for imaging and novel biomarkers, validation is a possibility validation would require data generated from multiple clinical trials using different investigational agents, that is, drugs of different mechanism of action that demonstrate changes in the same biomarkers, which subsequently demonstrates improvement in clinical endpoints (death, liver transplant, etc). If the biomarker is demonstrated to predict clinical outcomes, then it could ultimately serve as a primary endpoint in late phase clinical trials intended to support marketing/registration applications.

Late Phase IIb and III Trials

The best predictor of clinical outcomes in adults with NAFLD currently is fibrosis. Although still in debate, the presence of NASH in the absence of fibrosis may also be clinically relevant for future trials, especially in children.⁴² Histologic measurement of liver fibrosis has substantial limitations, including small sample size (sampling error owing to micro-heterogeneity of liver biopsies) and semiquantitative staging of fibrosis (ie, assessment of architectural changes but no measurement of the amount or the degree of fibrosis). The following biopsy-based surrogate endpoints are considered by the FDA and EMA⁵⁶ as “reasonably likely to predict clinical outcomes” in adult NASH trials for the purposes of drug approval using the accelerated approval (21 CFR Part 314, Subpart H & E) pathway:

• Resolution of steatohepatitis and no worsening of fibrosis

  And (EMA)/or (FDA)

• At least a 1-point improvement in fibrosis stage with no worsening of steatohepatitis (ie, no increase in steatosis, ballooning or inflammation).

However, these histologic endpoints are currently not validated and therefore cannot support a full/traditional approval. Confirmatory trial(s) in adults intended to verify clinical benefit should be underway at the time of accelerated approval. In children with histopathologic patterns similar to adults (NASH including ballooning), late phase pediatric trials would likely use these or similar surrogate endpoints. For trials evaluating children who have a periportal pattern of inflammation without hepatocyte ballooning, the histologic endpoint would need to account for this.

The development and validation of histologic and noninvasive tools for the assessment of pediatric NASH with fibrosis are needed. None of the available modalities are yet recommended as surrogate endpoints for pediatric trials to support marketing/registration applications. FDA evaluates bio-marker(s) or patient reported outcome endpoints within individual drug development programs as well as through a formal “qualification” program to allow public adoption of new biomarkers by a scientific community to advance public health by encouraging efficiencies and innovation in drug development. The biomarker qualification program may be useful for broadening the armamentarium for use in pediatric NAFLD trials. Rapidly evolving technologies to measure fibrosis noninvasively (magnetic resonance elastography, sheer wave elastography imaging, acoustic radiation force imaging, and transient elastography) or to measure changes in the liver function in response to fibrosis are undergoing evaluation and have potential to support future trials.^57–59 Certain serum metabolomic markers may also hold promise. In the United States, the acceptability of a surrogate endpoint(s) beyond those currently recommended by the FDA for phase III clinical trials is determined on a case by case basis (available: www.fda.gov/downloads/Drugs/Guidances/UCM358301.pdf).

There are no phase III or IV trials (to confirm clinical benefit of a drug approved via accelerated approval) in pediatric NAFLD at this time. Endpoints deemed clinically relevant in the adult NASH population such as liver-related (transplant, death, progression to cirrhosis, liver decompensation events, etc) and non–liver-related (death, occurrence of diabetes, cardiac outcomes, malignancy, etc) outcomes are unlikely to occur in a reasonable timeframe for phase IV studies in children. There is the potential for extrapolation of adult clinical benefit to pediatrics; however, better characterization of the natural history of pediatric NAFLD is needed to support this extrapolation.

Future Directions and Important Research Needs

Progress in the NAFLD field over the past decade has been rapid. Lessons learned from pediatric NAFLD clinical trials of recent past provide helpful foundations for future study design. Many questions and areas of uncertainty remain that will require global collaboration amongst patients, the scientific community, drug developers and regulators to attain the bright future of effective therapies for pediatric NAFLD (Table 2).

Table 2.

Summary Table of Recommendations

Therapeutics are needed for children with NAFLD and regulations in place protect children that balance the need for pediatric studies. Long-term longitudinal studies are essential to better define the natural history of NAFLD with early onset. While change in NAS of 2 points has been used in pediatric NAFLD trials, other histologic criteria may be considered in pediatric trials owing to the unique portal-predominant pattern in many children. Early phase and dose-ranging trial endpoints should be based on the mechanism of action of the drug and the target population. Clinical trials designed with liver biopsies should have concurrent non-invasive tests to advance validation of non-invasive biomarkers. Children with NAFLD who have inflammation and fibrosis may benefit the most from pharmacotherapies and thus are preferred for inclusion in trials. Formal pharmacokinetic studies are recommended to establish evidence-based dosing. For early phase studies, reduction of elevated serum ALT is a reasonable primary outcome. Although steatosis can be measured accurately with MRI, there is inadequate data to support that steatosis reduction will lead to clinically meaningful benefit or changes in other pertinent features related to NASH. Endpoints “reasonably likely to predict clinical outcomes” by the regulatory authorities for adults are as follows, and pediatric trials may use similar endpoints in those with NASH: FDA: Biopsy based resolution of steatohepatitis and no worsening of fibrosis or at least a 1-point improvement in fibrosis with no worsening of steatosis, ballooning or inflammation. EMA: Biopsy based resolution of steatohepatitis and no worsening of fibrosis and at least a 1-point improvement in fibrosis with no worsening of steatosis, ballooning or inflammation. Histology endpoints in children should recognize and account for the unique periportal pattern of inflammation without ballooning in a significant subset of pediatric NAFLD patients.

ALT, alanine aminotransferase; EMA, European Medicine Agency; FDA, US Food and Drug Administration; MRI, magnetic resonance imaging; NAFL, nonalcoholic fatty liver; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis.

Appendix. Liver Forum Pediatric Working Group Members

Nathalie Adda, Enanta Pharmaceuticals, Inc.

William Baldyga, Patient Representative

Rajarshi Banerjee, Perspectum Diagnostics Ltd.

Cynthia Behling, University of California, San Diego

Sherif Boulos, Resonance Health

Gary Burgess, Vectura Limited

Dania Calboli, Novartis Pharma AG

Edgar Charles, Bristol-Myers Squibb

Rose Christian, Bristol-Myers Squibb

Claude Cohen-Bacrie, SuperSonic Imagine

Doina Cosma-Roman, Teva Pharmaceuticals

Claus-Peter Danzer, Novartis Pharma AG

Ingrid Delaet, Intercept Pharmaceuticals, Inc.

Mark Delegge, QuintilesIMS

Lara Dimick-Santos, U.S. Food and Drug Administration

Nicholas DiProspero, Janssen Research and Development

Kathleen Donohue, U.S. Food and Drug Administration

Laurent Fischer, Allergan

Emer Fitzpatrick, King’s College London School of Medicine at King’s College Hospital

Michael Fried, University of North Carolina

David Hagerty, Conatus Pharmaceuticals, Inc.

Paula Hale, Novo Nordisk

Keri Hildick, Perspectum Diagnostics Ltd.

Dean Hum, GENFIT SA

Khurram Jamil, Mallinckrodt Pharmaceuticals

Lijuan Jiang, Enanta Pharmaceuticals, Inc.

Saul Karpen, Emory University

Matt Kelly, Perspectum Diagnostics Ltd.

David E. Kleiner, National Cancer Institute

Rohit Kohli, Children’s Hospital Los Angeles

Kattayoun Kordy, Novartis Pharmaceuticals

Nancy Krieger, Novartis Pharma AG

Joel Lavine, Columbia University

Lois Lee, Intercept Pharmaceuticals

Eric Lefebvre, Allergan

Patricia Lopez, Novartis Pharma AG

Erica Lyons, U.S. Food and Drug Administration

Laura Malahias, TARGET PharmaSolutions

Sophie Megnien, GENFIT SA

Ruby Mehta, U.S. Food and Drug Administration

Peter Mesenbrink, Novartis Pharmaceuticals

Veronica Miller, Forum for Collaborative Research

Pansy Minnick, Janssen Research and Development

Christine Murray, Raptor Pharmaceuticals

Tien Nghiem, Afimmune Limited

Nikki Nicholson, TARGET PharmaSolutions

Stephanie Noviello, Bristol-Myers Squibb

Stephanie O. Omokaro, U.S. Food and Drug Administration

Wenjie Pang, Resonance Health Ltd

Lisa Percival, Bristol-Myers Squibb

Dan Peres, Immuron Limited

Margaret Powell, TARGET PharmaSolutions

Dragos Roman, U.S. Food and Drug Administration

Mark Root, Novo Nordisk

Claire Sampson, Novo Nordisk

Arun Sanyal, Virginia Commonwealth University

Elmer Schabel, Bundesinstitut für Arzneimittel und Medizinprodukte

Kathleen Schwarz, Johns Hopkins University School of Medicine

Jeffrey B. Schwimmer, University of California, San Diego and Rady Children’s Hospital San Diego

Star Seyedkazemi, Allergan

David Shapiro, Intercept Pharmaceuticals, Inc.

Reshma Shringarpure, Intercept Pharmaceuticals

Debra Silberg, Takeda Pharmaceuticals

Edward Smith, Conatus Pharmaceuticals Inc

Piotr Socha, The Children’ s Memorial Health Institute

Robert Squires, Children’s Hospital of Pittsburgh of UPMC

Peter Szitanyi, General University Hospital, First Faculty of Medicine, Charles University in Prague

Johannes Taminiau, European Medicines Agency

Richard Torstenson, Novo Nordisk A/S Denmark

William Treem, Takeda Pharmaceuticals

Pamela Vig, Allergan

Miriam Vos, Emory University

Mason Yamashita, Conatus Pharmaceuticals, Inc.

Michael Zemel, NuSirt Biopharma