Expert Panel Review to Compare FDA and EMA Guidance on Drug Development and Endpoints in Nonalcoholic Steatohepatitis

Nonalcoholic steatohepatitis (NASH) has become a focus for drug development owing to its increasing prevalence and health care burden. The global prevalence of NASH is estimated to range from 1.5% to 6.5%,¹ with a higher prevalence of 12% to 14% among middle-aged US populations.^2,3 The increasing prevalence of nonalcoholic fatty liver disease (NAFLD) further exacerbates the evolving shortages of livers for transplantation by negatively affecting the pool of suitable donors.^4,5 This issue and the rising burden of end-stage liver disease due to NASH and NASH-related hepatocellular carcinoma (HCC) are major public health concerns.⁶ However, there are currently no approved therapies, representing a critical unmet need in the field of liver disease.

Regulatory approval of therapies for NASH-related fibrosis has been hindered by clinical trial challenges, including population heterogeneity, slow disease progression, and the requirement for invasive liver biopsies for diagnosis and to assess fibrosis and treatment response.⁷ Issues with liver biopsies including patient burden, morbidity, and sampling, as well as intra- and interreader variability, may lead to inappropriate study entry or exclusion, affecting the ability to detect potential treatment effects.^8–10 Centralized pathology reading, the addition of artificial intelligence to digital pathology, standardized training for pathologists, and adjudication procedures are currently proposed solutions to these issues. To help define and address the clinical trial challenges, the US Food and Drug Administration (FDA) and the European Medicines Agency (EMA) published draft guidance on drug development for noncirrhotic^11,12 and cirrhotic NASH.^11,13 We compared the FDA and EMA guidance to facilitate further discussion and potential alignment between agencies upon finalization of their recommendations.

Review of Noncirrhotic Nonalcoholic Steatohepatitis Guidance

Points of agreement and discordance regarding noncirrhotic NASH are summarized in Table 1.

Table 1.

Comparison of the FDA and EMA Guidance for Noncirrhotic NASH

Noncintiotic NASH trial features	Points of agreement	Points addressed by one agency but not by the other		Points of divergence
	FDA and EMA	Only FDA	Only EMA	FDA	EMA
Long-term endpoint (full market approval)	Composite endpoint includes all-cause mortality, cirrhosis, hepatic decompensation events, and MELD score	Composite endpoint includes progression to liver transplant	—	Change in MELD score from ≤12 to >15 (approximates listing for liver transplant)	MELD score above a threshold defining a high risk of liver-related death (ie, >14)
Intermediate/ surrogate endpoints (accelerated/early approval)	Resolution of NASH and/or improvement of fibrosis	Sponsor may propose degree of histologic improvement based on mechanism of action	—	Requirement for resolution of NASH or improvement in fibrosis or both	Requirement for coprimary endpoints including resolution of NASH and improvement in fibrosis
Study design	Placebo-controlled studies to assess efficacy Patients with an NAS of ≥4, with ≥1 point each for inflammation and ballooning, and a NASH-CRN fibrosis stage of 2 or 3 Sponsors may propose validated scores (or those with scientific justification) as alternatives to the NAS and NASH-CRN score for inclusion Stable doses of therapies for comorbidities and stable body weight required Stratify by presence/absence of T2DM	Additional inclusion and exclusion criteria	Inclusion of patients with NAS ≥ 5 Fibrosis stage 1 patients may be included for exploratory purposes only	Duration (phase 2): 12–18 months depending on drug mechanism of action Enrollment based on biochemical, imaging evidence of steatosis, steatohepatitis, or fibrosis or known NASH diagnosis/risk factors is sometimes acceptable No body weight changes of ≥5% for ≥3 months	Duration (phase 2): 2-year interim evaluation; 5-year final evaluation Enrollment mainly reliant on liver biopsy Failure to lose weight at least once
Safety	Cardiovascular risk in NASH patients is acknowledged Specific approach needed for monitoring liver function in patients with abnormal function/underlying liver disease Sponsors are encouraged to establish an expert committee to adjudicate cases for major adverse cardiac events (reflection paper on cardiovascular safety is considered applicable to NASH) and for liver injury
Pediatric considerations	Histologic characteristics of NASH may differ between adults and children Pediatric trials should only start once long-term information regarding disease progression, dosing, safety, and efficacy in adults are available	Future guidance will specifically address drug development for the pediatric population	—	—	—
Additional considerations	Sponsors are encouraged to identify noninvasive imaging biomarkers to replace liver biopsies	Clinical development considerations for phase 3 trials specified	—	—	—

Points of Agreement

The FDA and EMA are in agreement regarding several areas of noncirrhotic NASH drug development.^11,12 To determine the efficacy of a drug, both agencies recommend using a long-term composite endpoint, with components including all-cause mortality, histologic cirrhosis diagnosis, hepatic decompensation events, and model for end-stage liver disease (MELD) score assessments. Based on the time required to assess long-term endpoints, both agencies advise earlier evaluation of intermediate histologic endpoints to accelerate regulatory approval, given the unmet medical need for NASH therapies.

Placebo-controlled study designs for NASH drug development are required by both agencies in the current absence of an approved therapy for NASH.^11,12 Inclusion criteria comprise a diagnosis of noncirrhotic NASH, defined as an NAFLD activity score (NAS) of ≥4, with ≥1 point each for lobular inflammation and ballooning, and a NASH Clinical Research Network (NASH CRN) fibrosis stage of 2 or 3. The NAS and NASH CRN histologic scoring systems are endorsed as the standard for histologic assessment in phase 2b and 3 trials but have recognized limitations, with inter- and intraobserver variability.^8,14 Therefore, the FDA and EMA have kept the door open to permit the use of alternative validated scoring systems with appropriate justification as the science evolves. Because changes in body weight or treatment for comorbidities could confound the therapeutic effects of a NASH agent, both agencies recommend stabilization of these factors at study entry and stratification by the presence or absence of type 2 diabetes mellitus (T2DM). Any effects of concomitant treatments or lifestyle interventions that are imbalanced between study arms could elevate placebo or drug effects and either limit or exaggerate the ability of a drug to demonstrate therapeutic gains.¹⁵ Appropriate control of these features may help improve the predictability of the background response rates observed in trials.

With the recognized associations with obesity and cardiovascular risk in NASH, the FDA and EMA believe that cardiovascular safety should be monitored in clinical trials.^11,12,16 Both agencies believe that a specific liver monitoring approach is needed in patients with abnormal liver chemistry test results (eg, increased liver aminotransferase levels) and highlight the importance of the adjudication of cardiovascular events and drug-induced liver injury (DILI) by an independent expert adjudication panel.

Given the challenges with liver biopsy, both agencies encourage the development of noninvasive biomarkers to potentially replace liver histology in future trials.^11,12 These biomarkers should be at least as predictive of clinical benefit as liver histology and are an important priority for drug developers to reduce patient burden, allow for safer histologic evaluation, and improve accuracy and statistical power in clinical studies.^11–13 Currently available noninvasive biomarkers include serum markers such as enhanced liver fibrosis score and N-terminal propeptide of type III collagen (PRO-C3); ultrasound-based liver stiffness measurement by vibration-controlled transient elastography; and imaging-based liver stiffness measurement by magnetic resonance elastography, magnetic resonance imaging iron-corrected T1, and magnetic resonance imaging proton density fat fraction.^17,18 Associations between changes in biomarkers and changes in histologic features (eg, fibrosis) remain unclear.

The development of noninvasive biomarkers is considered an even higher priority in children because of the ethical and procedural challenges with repeated liver biopsies in this population.¹¹ Based on the unknown clinical significance of histologic differences between adults and children with NASH, the agencies agree that data on the natural history of pediatric NAFLD are needed.^11,12 However, long-term efficacy and safety data in adults are required before pediatric trials are initiated.

In addition to clinical outcomes, both agencies have guidance on the use of systematically developed and validated patient-reported outcomes to reflect patients’ perspectives and experiences with treatment.^19–21 Assessments of both clinical and NASH-specific patient experience outcomes will ensure the replication of efficacy data from clinical trials in effectiveness data from real-world practices.

Points Addressed by One Agency but Not by the Other

Liver transplantation is recommended by the FDA, but not the EMA, as a component of the long-term composite endpoint.^11,12 We note that liver transplantation may not be entirely reflective of the clinical severity of NASH because the patient selection criteria for organ allocation varies by geographic location and includes factors beyond severity, such as comorbidities, social-psychiatric circumstances, and finances.²² Alcoholic hepatitis is another rising indication for transplantation and is typically associated with higher MELD scores, leading to a MELD-related disadvantage for transplant candidates with cirrhotic NASH despite the clinical severity of their disease.

Regarding entry criteria, the FDA requires the cessation or stabilization of vitamin E and pioglitazone for 6–12 months because of the potential use of these treatments for NASH. Only the EMA includes an NAS of ≥5 in the diagnostic criteria for noncirrhotic NASH and mentions the inclusion of patients with stage 1 fibrosis for exploratory purposes because of their minimally increased risk of disease progression.²³ Only the FDA provides clinical development considerations for phase 3 trials and mentions plans to develop future guidance for pediatric noncirrhotic NASH trials.

Points of Divergence

There was divergence between the agencies in selected endpoint definitions, phase 2 study durations, and the need for liver biopsies in phase 2 trials.^11,12 The MELD score component of the long-term composite endpoint is defined as a change from ≤12 to ≥15 by the FDA and as reaching a score of >14 by the EMA. For the 2 intermediate composite endpoints, the FDA recommends the evaluation of either resolution of NASH without worsening of fibrosis or a ≥1-stage improvement in fibrosis without worsening of NASH (or both), whereas the EMA requires that both endpoints independently demonstrate efficacy in a coprimary fashion.

The body weight stabilization entry criterion is defined as no body weight changes of ≥5% for ≥3 months by the FDA and as a failure to lose weight from ≥1 documented standard weight loss attempt by the EMA (to avoid the enrollment of patients with no lifestyle counseling before receiving it in the trial). The FDA believes that the duration of an early phase 2 trial will depend on the mechanism of action and anticipated time to effect of a drug, with late phase 2 trials requiring a duration of at least 12–18 months unless scientific justification is provided for a shorter trial. The EMA believes that 2-year interim and 5-year final evaluations are appropriate but also suggests that the duration may depend on the trial size or anticipated activity of a drug. Study durations require further verification for evaluating histologic improvement. The FDA strongly encourages liver biopsy in late phase 2 trials to optimize dose selection but permits enrollment into early phase 2 trials according to known NASH risk factors and biochemical or imaging evidence of fibrosis with or without liver biopsy. The EMA requires liver biopsy and believes that long-term metabolic syndrome features should be used only to identify potential trial participants.

Review of Cirrhotic Nonalcoholic Steatohepatitis Guidance

Points of agreement and discordance regarding cirrhotic NASH are summarized in Table 2.

Table 2.

Comparison of the FDA and EMA Guidance for Cirrhotic NASH

Cirrhotic NASH trial features	Points of agreement	Points addressed by one agency but not by the other		Points of divergence
	FDA and EMA	Only FDA	Only EMA	FDA	EMA
Long-term endpoint (full market approval)	All-cause mortality as a component of composite endpoint	—	Endpoints for decompensated cirrhosis	Includes time from randomization to any one of the following events: complication of ascites, variceal hemorrhage, worsening in the MELD score to ≥15, and liver transplantation as part of a composite endpoint	Includes liver decompensation as part of a composite endpoint
Intermediate endpoint/surrogate (accelerated/early approval)	—	Drugs for the treatment of compensated NASH cirrhosis likely evaluated under the traditional approval pathway	Includes intermediate endpoint of cirrhosis reversal for compensated NASH, defined as an improvement of liver cirrhosis to noncirrhotic liver disease (≥1-point improvement in fibrosis stage) Requires support by additional secondary outcomes (noninvasive markers) and data on decompensation events, liver transplantation, death Specified endpoints for trials with relevant proportions with advanced cirrhosis; occurrence of decompensation events or other endpoints (eg, lowering of MELD score below a certain threshold or of the HVPG below 10 mm Hg) based on specific justification	—	—
Study design	Placebo-controlled studies to assess efficacy are required Stable doses of therapies for comorbidities required Stratify by presence/absence of T2DM Include patients with compensated cirrhosis diagnosis supported by histology (ie, fibrosis score of 4)	Sponsors may propose other histologic and nonhistologic criteria for diagnosis of compensated NASH cirrhosis, if scientifically acceptable Additional exclusion criteria Requires exclusion of patients with decompensated cirrhosis	Additional criteria for including patients with compensated cirrhosis: NASH diagnosis based on (1) NAS > 5 or (2) NAS ≥ 4 with scores for all components ≥1 Historical histology proving NASH, with noninvasive tests pointing to NASH (serologic markers, imaging), and relevant comorbidity risk factors are acceptable for inclusion Study duration of 2 years (interim) and 5 years (long-term endpoint) Provides guidance on inclusion criteria for decompensated cirrhosis	—	—
Safety	Acknowledgement of cardiovascular risk in NASH patients Sponsors encouraged to monitor cardiovascular safety during the clinical trial (reflection paper on cardiovascular safety is considered applicable to NASH) Specific approach needed for monitoring liver function in patients with abnormal function/underlying liver disease Expert committee recommended to adjudicate liver injury cases	—	—	—	—
Additional considerations	Sponsors encouraged to identify noninvasive imaging biomarkers to replace liver biopsies	—	—	—	—

HVPG, hepatic venous pressure gradient.

Points of Agreement

The FDA and EMA believe that cirrhotic NASH trials should use a placebo control, stabilize therapies for comorbidities, and stratify patients by the presence of T2DM, similar to noncirrhotic NASH trials.^11,13 The agencies also maintain recommendations for cardiovascular evaluations and DILI adjudication and the importance of noninvasive biomarkers in cirrhotic NASH trials.^11,13,16 Specific to cirrhotic NASH, both agencies believe that the long-term composite endpoint should include all-cause mortality and that enrolled patients should have histologic documentation of cirrhosis (ie, fibrosis score of ≥4 as de fined by the NASH CRN or the steatosis, activity, and fibrosis score).

Points Addressed by One Agency but Not by the Other

Regarding intermediate endpoints for compensated cirrhosis, the EMA provides more extensive recommendations for evaluating a histologic reversal in cirrhosis (eg, a ≥1-point improvement in fibrosis stage), whereas both agencies acknowledge the insufficient evidence of any associations at the cirrhotic stage between reversed fibrosis stage and improved clinical outcomes.^11,13 The EMA indicates that any claim of cirrhosis reversal requires (1) an accompanying substantiated claim that the prognosis of reversed cirrhosis to an earlier fibrosis stage is similar to that of untreated earlier fibrosis stages in progressive disease (eg, from other diseases such as hepatitis B or C); (2) supportive secondary outcomes (ie, noninvasive markers); and (3) available descriptive data on decompensation events, liver transplantation, and death. Given that a histologic surrogate endpoint is not established in cirrhosis, the FDA expects that the traditional approval pathway will be applicable but considers proposals to support accelerated approval if scientifically justified. We note that recent data have suggested an association between the regression of fibrosis and a reduction in liver-related complications for patients with compensated NASH-related cirrhosis.²⁴

Specific EMA-recommended inclusion criteria comprise (1) an NAS of >5, or an NAS of ≥4 and all components ≥1; (2) historical biopsy-confirmed NASH; (3) noninvasive markers of probable NASH (eg, imaging biomarkers); and (4) relevant comorbid conditions such as obesity and T2DM. The latter 3 criteria are required for patients initially diagnosed with cryptogenic cirrhosis. The FDA does not provide diagnostic inclusion criteria beyond fibrosis stage but permits scientifically justified proposals and discussion of other criteria. Only the FDA recommends the following exclusion criteria: (1) alanine aminotransferase or aspartate aminotransferase elevations of >5 times the upper limit of normal; (2) living-related or orthotopic liver transplantation; (3) a baseline MELD score of >12; and (4) a history of HCC or HCC treatment.

Only the EMA specifically defines recommendations for advanced and decompensated cirrhosis, recognizing that considerable data on a potential agent may be needed before including this vulnerable subgroup in clinical trials.^11,13 All-cause mortality or survival is recommended by the EMA as the long-term endpoint for decompensated cirrhotic trials, with liver-related death and transplantation suggested as supportive endpoints. Intermediate endpoints recommended by the EMA for advanced cirrhosis include decompensation events, a MELD score reduction below a threshold of 14, and a hepatic venous pressure gradient reduction below 10 mm Hg. The EMA-recommended inclusion criteria for decompensated cirrhosis comprise the presence of historical biopsy-confirmed unequivocal NASH and cirrhosis as well as decompensation symptoms. The FDA, however, focuses on compensated cirrhotic NASH (and not on the complications of cirrhosis) and requires protocol-specified criteria to exclude patients with decompensated cirrhosis, which can include, but are not limited to, the following criteria: evidence of portal hypertension (eg, low platelet counts, esophageal varices, ascites, history of hepatic encephalopathy, splenomegaly), elevated bilirubin, and elevated international normalized ratio or prolonged prothrombin time. The reversibility of decompensated cirrhosis is unclear because viral eradication in chronic hepatitis C with decompensated cirrhosis was previously shown not to improve liver function for some patients.²⁵ The inclusion of such patients who may be unlikely to respond to medical intervention could affect clinical trial results.

Points of Divergence

The FDA and EMA are divergent regarding the long-term composite endpoint for cirrhotic NASH, with only all-cause mortality being recommended as a component by both agencies.^11,13 The FDA recommends evaluation of the time from randomization to any of the following components: complication of ascites, variceal hemorrhage, hepatic encephalopathy, worsened MELD score to ≥15, and liver transplantation. The EMA recommends liver decompensation as a component of the composite endpoint in situations where the cirrhotic population is enriched with patients with advanced cirrhosis. The FDA recommendation for the long-term composite endpoint in cirrhotic NASH trials may be preferred because of its granularity.

Recommendations for Harmonization

The following points of potential future alignment between the FDA and EMA guidance documents could help accelerate drug development in NASH:

• histologic endpoints for demonstrating improvement in fibrosis without worsening of NASH and/or NASH resolution without worsening of fibrosis for conditional approval in noncirrhotic NASH trials;

• uniform plan for liver histologic assessment across all phase 3 programs in NASH;

• “must-have list” of noninvasive biomarkers to be measured and reported in all future NASH clinical trials and clear guidance on their assessment to replace liver biopsy as an endpoint;

• NASH-specific approaches to patient-reported outcome assessments;

• appropriate study durations for noncirrhotic and cirrhotic trials;

• consideration of updated evidence supporting associations between reversal of fibrosis and improvement in clinical outcomes in cirrhotic NASH;

• long-term composite endpoint of cirrhotic NASH trials, for which the more granular FDA recommendation may be preferred;

• consistent DILI monitoring plan in cirrhosis trials;

• whether to include cardiovascular and liver-related events as composite endpoints in NASH trials, given the associated risk of mortality in NASH; and

• endpoints in pediatric and decompensated cirrhosis populations.

Conclusions

The FDA and EMA guidance documents reflect their agreement on the importance of strategies to overcome the challenges encountered in clinical trials of noncirrhotic and cirrhotic NASH. Trials in NASH should control for confounding factors such as comorbid conditions and monitor for cardiovascular events and DILI. Noncirrhotic and cirrhotic NASH trials should also contribute to the development of noninvasive biomarkers to potentially replace the burden and sometimes unreliability of liver biopsies in clinical trials. Further efforts are needed to harmonize some of the recommendations.