Defining endpoints in clinical studies in NAFLD
Ian A Rowe MRCP PhD
University Clinical Academic Fellow and Honorary Consultant Hepatologist
i.a.c.rowe@leeds.ac.uk
Leeds Institute of Data Analysis (LIDA) & Leeds Institute for Medical Research
University of Leeds
Liver Unit, St. James’s Hospital
The principal aim of a treatment for NASH is that it reduces liver-related morbidity and mortality
78 patients (12%) with F3 or F4
Hagstrom, Nasr, et al, J Hepatology 2017
78 patients (12%) with F3 or F4
Hagstrom et al, J Hepatology 2017
17 liver related deaths (7.9%)
193 deaths
17 due to liver disease
1 liver transplant
Angulo et al, Gastroenterology 2015
Non-cirrhotic NASH
Cirrhotic NASH
NASH
F2/F3
Histological endpoints
Accelerated approval
Clinical endpoints
Full marketing approval
Phase 3, randomized placebo controlled
Phase 4, confirmatory continuation study
12-18 months
Event driven
3-5 years
NASH
cirrhosis (F4)
Clinical endpoints
Full marketing approval
Phase 3, randomized placebo controlled
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Clinical endpoints
Endpoints that impact on how a person with NASH feels, functions, or survives
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
Liver histology
Progression to cirrhosis
Clinical outcomes
Reduction in decompensating events (variceal bleed, ascites, encephalopathy, etc)
Increase in MELD score
Reduction in death
Reduction in need for liver transplant
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
These endpoints are unvalidated
There are no prospective data to verify that resolution of steatohepatitis is associated with a meaningful change in clinical outcomes
Improvements in liver injury and fibrosis, associated with treatment for hepatitis B and hepatitis C, are associated with improved outcomes
Activity
Fibrosis
Marcellin, et al, Lancet, 2013
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
These endpoints are unvalidated
There are no prospective data to verify that resolution of steatohepatitis is associated with a meaningful change in clinical outcomes
Improvements in liver injury and fibrosis, associated with treatment for hepatitis B and hepatitis C, are associated with improved outcomes
There is no way to know how histologic changes will modify clinical outcomes
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
Rinella & Noureddin. J Hepatol, 2020
Rinella et al. J Hepatol, 2019
Davison et al, J Hepatology, 2020
Rowe & Parker, Accepted CG&H, 2021
In large randomized trials including patients with F2/3 and treated with placebo there is no net progression in fibrosis
Seems implausible, unless NASH is very easy to treat
Rowe & Parker, Accepted CG&H, 2021
Sampling variability
Observer variability
Regression to the mean
Net fibrosis progression
Fibrosis improvement
“placebo response”
Fibrosis progression
Factors determining the “placebo response” are largely unmodifiable and challenge the use of liver biopsy for trial endpoint assessment in NASH
Observer variability
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
These endpoints are fragile and unvalidated
There is no way to know how histologic changes will modify clinical outcomes
Liver histology
Progression to cirrhosis
96% of “clinical” events in STELLAR-3 were progression to cirrhosis
Sampling variability probably accounts for a substantial proportion of apparent progressors
Rowe & Parker, Accepted CG&H, 2021
Clinical endpoints
Endpoints that impact on how a person with NASH feels, functions, or survives
Clinical outcomes
Reduction in decompensating events (variceal bleed, ascites, encephalopathy, etc)
Increase in MELD score
Reduction in death
Reduction in need for liver transplant
Critical for clinical decision-making
There is an urgent need to understand what the patient relevant benefits to treatment with NASH will be
Per protocol population
Progression in disease reduced by approximately 30%
Rowe, Hagström, et al, submitted
UK reference population
US reference population
Model outcomes account for deaths from common (and potentially modifiable) competing events
Treatment outcomes in persons with NASH aged 55y, modelled 30% reduction in liver-related mortality with new treatment
The principal aim of a treatment for NASH is that it reduces liver-related morbidity and mortality, but…
Treatment outcomes in persons with NASH aged 55y, modelled 30% reduction in liver-related mortality and 12% reduction in CVD mortality with new treatment
Treatment outcomes in persons with NASH aged 55y, modelled 30% reduction in liver-related mortality and 12% reduction in CVD mortality with new treatment
Where treatment has a strong impact on metabolic multimorbidty, what is the indication for treatment?
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Clinical endpoints
Endpoints that impact on how a person with NASH feels, functions, or survives
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
Liver histology
Progression to cirrhosis
Clinical outcomes
Reduction in decompensating events (variceal bleed, ascites, encephalopathy, etc)
Increase in MELD score
Reduction in death
Reduction in need for liver transplant
Safety
Stage | Age (years) | Treatment effect | 10-year outcomes | Mean life years gained | ||
Liver | CVD | Liver-related mortality (%) | CVD-related mortality (%) | |||
F2 | 55 | - | - | 1.3 | 6.2 | - |
-30% | - | 0.8 | 6.2 | 0.12 | ||
-30% | 5% | 0.8 | 6.4 | -0.05 | ||
-30% | 10% | 0.8 | 7.2 | -0.25 | ||
F3 | 55 | - | - | 4.4 | 8.2 | - |
-30% | - | 3.4 | 8.2 | 0.29 | ||
-30% | 5% | 3.2 | 9.1 | 0.12 | ||
-30% | 10% | 3.2 | 9.3 | 0.00 | ||
F4 | 55 | - | - | 16.2 | 10.5 | - |
-30% | - | 12.3 | 10.5 | 0.91 | ||
-30% | 5% | 11.9 | 11.3 | 0.80 | ||
-30% | 10% | 11.6 | 12.1 | 0.68 | ||
Treatment outcomes in persons with NASH aged 55y
Treatment carrying additional CVD liability risks overall treatment benefit
A modest increase in CVD risk in early-stage disease may result in net harm from treatment
Non-cirrhotic NASH
Cirrhotic NASH
NASH
F2/F3
Histological endpoints
Phase 3, randomized placebo controlled
12-24 months
NASH
cirrhosis (F4)
Clinical endpoints
Full marketing approval
Phase 3, randomized placebo controlled
FDA Webinar, January 2021
Sufficient participants to satisfy safety requirements
Surrogate endpoints
Endpoints that are reasonably likely to predict clinical benefit
Clinical endpoints
Endpoints that impact on how a person with NASH feels, functions, or survives
Liver histology
Resolution of steatohepatitis and no worsening of fibrosis
or
Improvement in liver fibrosis (greater than one stage) and no worsening of steatohepatitis
or
Both resolution of steatohepatitis and improvement in fibrosis
Liver histology
Progression to cirrhosis
Clinical outcomes
Reduction in decompensating events (variceal bleed, ascites, encephalopathy, etc)
Increase in MELD score
Reduction in death
Reduction in need for liver transplant
Safety
Acknowledgments
University of Leeds (LIMR)
Anna Roskilly
Eleanor Taylor
Robert Driver
Jessica Shearer
Thazin Min
University of Leeds (LIDA)
Amy Downing
Leeds Teaching Hospitals
Richard Parker
Rebecca Jones
Jennifer Spencer
Amy Hicks
Carys Lippiatt
Karolinska Institutet
Hannes Hägstrom
Linköping
Mattias Ekstedt
University of Oxford
Eva Morris