Endpoints for Clinical Development in NASH

The lack of accurate, reproducible, and easily applied methods to assess NASH creates major limitations not only for drug development, but also in the clinical management of NASH patients.

Liver biopsy remains the gold standard for the diagnosis of NASH, but it has several limitations. There is always a risk that the biopsy taken might not be representative for the amount of fibrosis in the whole liver. Increasing the length of liver biopsy decreases the risk of sampling error. In general, except for cirrhosis, for which microfragments may be sufficient, a 25-mm-long biopsy is considered an optimal specimen for accurate evaluation, although 15 mm is considered sufficient. Not only the length but also the caliber of the biopsy needle is important in order to obtain a piece of liver of adequate size for histological evaluation: A 16-gauge needle is considered appropriate for percutaneous liver biopsy (Sanyal et al., 2011; Papastergiou et al., 2012). Interobserver variation is another limitation, which is related to the discordance between pathologists in biopsy interpretation. This can be as high as 25%, but the variation is less pronounced when the biopsy assessment is done by specialized liver pathologists (Papastergiou et al., 2012).

Besides the technical problems with biopsy, the procedure is unpleasant for the patient, costly, and carries a risk of rare but potentially life-threatening complications - this limits the use of liver biopsy for mass screening. Due to these limitations, the fact that NASH is usually asymptomatic and that patients usually have normal liver aminotransferases, most NASH patients are undiagnosed. It has been reported that only a minority (less than 25%) of academic gastroenterologists and hepatologists in the United States routinely perform liver biopsies in patients with presumed NASH (Lominadze et al., 2014). This level of undiagnosed patients and the need of two liver biopsies within a 48- to 72-week period establish a significant hurdle in the enrollment of patients in clinical trials.

Endpoints in Clinical Trials

In spite of all these limitations, liver histology currently offers the best shortterm method not only for grading and staging but also for tracking the progression of NASH.

The primary objective of the treatment for NASH is to prevent liver-related morbidity and mortality, due mainly to the development of cirrhosis, which generally takes more than 10-20 years to develop. Due to this long natural history, there is a need of surrogate markers of avoidance of cirrhosis and thus liver-related mortality. The main predictor of disease progression is increasing fibrosis. On the other hand, patients with steatohepatitis are more likely to have a progressive disease compared to patients with isolated fatty liver. Therefore, complete resolution of NASH (i.e., absence of ballooning with no or minimal inflammation) with no worsening of fibrosis, and actual improvement in fibrosis are recommended as "surrogate endpoints, reasonably likely to predict clinical benefit in progression to cirrhosis and liver-related death" (Sanyal et al., 2014) (Table 14.1). Ideally, a co-primary endpoint of two composite endpoints, one-complete resolution of NASH with no worsening of fibrosis and, two-at least one-point improvement in the fibrosis stage with no worsening of steatohepatitis, should be demonstrated in the marketing authorization trial. However, the clinical outcome study aiming to demonstrate a reduction in progression to cirrhosis and portal hypertension/ cirrhosis-related events needs to be demonstrated.

For dose-ranging trials, the histology endpoint of improvement in activity as assessed by a reduction in at least two points in NAS (including at least one point in ballooning or inflammation) is an acceptable surrogate marker of improvement. It is important to note that although NAS has proven useful for comparative analyses and interventional studies, it does not provide information about fibrosis or the location of lesions. Therefore, the reduction in NAS must be associated with a lack of progression in fibrosis.

TABLE 14.1

Endpoints and Population in Clinical Trials in NASH


Primary Endpoint

Target Population

Trials to support a marketing application

Composite endpoint: Complete resolution of steatohepatitis and no worsening of fibrosis Composite endpoint: At least one-point improvement in fibrosis with no worsening of steatohepatitis (no increase in steatosis, ballooning, or inflammation) Clinical outcome underway by the time of submission: Histopathologic progression to cirrhosis MELD score change by >2 points or MELD increase to >15 in population enrolled with <13

  • • Death
  • • Transplant
  • • Decompensation events
  • • Hepatic encephalopathy - West Haven > grade 2
  • • Variceal bleeding - requiring hospitalization
  • • Ascites - requiring intervention
  • • Spontaneous bacteria peritonitis

Biopsy-confirmed NASH patients with moderate/ advanced fibrosis (F2/F3)

Dose ranging/ phase II

Improvement in activity (NAS)/ballooning/ inflammation without worsening of fibrosis can be acceptable

Include a subpopulation with modernte/advanced fibrosis (F2/F3) to inform phase III

Biopsy-proven NASH (NAS > 4) Include patients with NASH and liver fibrosis with any stage of fibrosis Include patients with NASH

and > fibrosis stage 2 to inform phase III

Early-phase trials/РОС

Endpoints should be based on the mechanism of drug Consider using improvement in NAS (ballooning and inflammation) and/or fibrosis Reduction in liver fat with a sustained improvement in transaminases

Ideal to use patients with biopsy-proven NASH, but acceptable to use patients at high risk for NASH (fatty liver + T2DM, the metabolic

syndrome, and high transaminases are acceptable)

Source: Filozof et al. (2017).

Biopsy-driven endpoints are, in general, not feasible in a 12- to 24-week РОС trials. In short-term РОС studies, which are designed mainly to assess tolerability of new drugs and to look for futility signals to direct decisions regarding further development, an improvement of hepatic steatosis, as determined by magnetic resonance technology, might be suitable since improvement in steatohepatitis is generally associated with a reduction in liver fat (Sanyal et al., 2011). Improvement in liver aminotransferase and other noninvasive biomarkers of insulin sensitivity, inflammation, apoptosis, and fibrosis could be helpful to evaluate the efficacy of the compound and support decision-making. However, it is important to note that the use of noninvasive biomarker methods is still considered experimental and there are no validated noninvasive biomarkers.

Target Population

Prevention of cirrhosis and demonstrating a positive effect on the well-defined liver outcomes are key clinical goals when considering a NASH drug development program. Therefore, for trials aiming to support marketing application, it is important that subjects with the greatest risk of progression to cirrhosis are enrolled (Table 14.1). Among individual features, liver fibrosis has proven the best independent association with liver-related mortality. Patients with NASH develop progressive fibrosis in 25%-50% of individuals over 4-6 years, while 15%-25% of individuals with NASH can progress to cirrhosis (Musso et al., 2011). In another study, with a mean follow-up of 13 years, 13.3% of NASH patients with mild to moderate fibrosis (stages 1-2) and 50% of patients with fibrosis stage 3 at inclusion developed cirrhosis (Ekstedt et al., 2006). Since in patients with NASH and advanced fibrosis (F2-F3), the probability of developing cirrhosis is much higher than in patients with early fibrosis (FI), this population is recommended for the long-term outcome trials in order to enhance the chances of demonstrating a benefit within a reasonable time frame. The enrollment of patients with moderate/advanced fibrosis for the evaluation of long-term outcomes, including progression to cirrhosis, should ensure that an expected number of events, calculated based on the progression rate for each fibrosis stage, are obtained based on the literature (see, e.g., Ekstedt et al., 2006; Argo et al., 2009; Pagadala and McCullough, 2012; Angulo et al., 2015; Singh et al., 2015). In patients with NASH and advanced fibrosis (F2-F3), this progression rate can be estimated at 8% per year for fibrosis stage 3, and 6% per year for fibrosis stage 2. Since the progression rate in some patients with mild fibrosis with additional risk factors of progression (e.g., presence of T2DM, the metabolic syndrome, high transaminases) might be fast, it is worth exploring this subgroup of patients, as an additional exploratory group.

A broad population of NASH patients, including those with mild fibrosis, is acceptable in dose-ranging (phase II) trials. However, it is recommended that a sufficient number of patients with moderate and severe fibrosis are enrolled in order to get preliminary data to inform the trial(s) to support marketing application.

Ideally, in early РОС trials, the target population should also be patients with biopsy-confirmed NASH. However, patients at high risk of NASH, namely, patients with fatty liver and diabetes and/or the metabolic syndrome with or without high liver enzymes, can be acceptable at this stage. Noninvasive serum biomarkers of imaging can be used to enrich a population in а РОС trial (EA, 2015).

Adaptive Design Approach in NASH

Statistical Considerations

One of the most commonly considered adaptive designs is probably a multiple-stage seamless adaptive design that combines several independent studies into a single study that can address the study objectives of the intended individual studies. For two-stage adaptive designs, four categories have been reported (Chow and Tu, 2008; Chow and Lin, 2015) depending upon their study objectives and study endpoints at different stages. These categories include (i) designs with the same study objectives and study endpoints at different stages, (ii) designs with the same study objectives but different study endpoints at different stages, (iii) designs with different study objectives but the same study endpoints at different stages, and (iv) designs with different study objectives and different study endpoints at different stages. Details regarding different categories of two-stage adaptive seamless designs can be found in Chapter 11. The use of the adaptive design has several advantages and some limitations (Table 14.2). Though in the 2010 FDA draft guidance (see also, FDA, 2018), the two-stage seamless adaptive design was emphasized as a less well-understood design because the valid statistical methods were not

TABLE 14.2

Relative Merits and Limitations of Two-Stage Adaptive Design in NASH


Two Independent Trials

Two-Stage Adaptive Design


90% 90% (81%)


Sample size

n = n + n2

N < N, + N2-

Operational bias


Moderate to severe

Data analysis

By study analysis

Combined analysis


6-12 months of lead time between studies

Reduced lead time between trials

Flexibility/ long-term follow-up

New study design based on previous data. New patients are enrolled

Adaptations based on IA: Stop one or more study arms/randomize more patients, etc. Continue follow-up

Regulatory aspects

Standard practice

Requires buy-in by global authorities prior to initiation

Statistical perspective

Valid statistical methods are well established

Evolving statistical methods





Source: Filozof etal. (2017). Note:

a Depends on adaptations.

N, is the sample size for trial 1. N2 is the sample size for trial 2.

fully developed at that time (Chow and Corey, 2011), the two-stage adaptive design has gradually become well understood due to the increasing number of regulatory submissions utilizing this design (see, e.g., Lu et al., 2009,2010,2012).

Utility Function for Endpoint Selection

We recommend a utility function be adopted to link all NASH endpoints at different stages (Chow and Lin, 2015). We propose a therapeutic index function for the analysis of two-stage adaptive designs with distinct study objectives and study endpoints at each stage under the assumption that even though the two endpoints are not the same, there is a well-established relationship between them.

A therapeutic index function is defined for each one of the endpoints. It takes different endpoints with pre-specified criteria into consideration, and it is based on a vector of therapeutic index function rather than the individual endpoints. The vector of therapeutic index model allows the investigator to accurately and reliably assess the treatment effect in a more efficient way (see Appendix: Utility Function and Statistical Tests). The direct application of standard statistical methods leads to the concern that the obtained p-value and confidence interval for the assessment of the treatment effect may not be correct or reliable. Most importantly, the sample size required for achieving a desired power obtained under a standard group sequential trial design may not be sufficient for achieving the study objectives under the two-stage seamless adaptive trial design, especially when the study objectives and/or study endpoints at different stages are different.

< Prev   CONTENTS   Source   Next >