Honest Criticisms and Limitations of the MAESTRO-NASH Trial

At a glance

Field	Detail
Trial	MAESTRO-NASH (Phase 3)
N	966 randomized
Intervention	Resmetirom 80 mg or 100 mg daily
Comparator	Placebo
Duration	52 weeks (primary biopsy endpoint)
Primary endpoint	NASH resolution with no worsening of fibrosis; fibrosis improvement ≥1 stage with no worsening of NAS
Key result	Both doses met both co-primary endpoints vs placebo (Harrison et al., NEJM 2024)

The Enrollment Filter: Who Was (and Wasn't) Studied

MAESTRO-NASH enrolled adults with biopsy-confirmed NASH and fibrosis stages F1B through F3. The practical consequence is that patients with compensated cirrhosis (F4), the group at highest near-term risk of decompensation, transplant listing, and death, were excluded from the key efficacy analysis. The FDA label for Rezdiffra reflects this gap: the indication specifies moderate-to-advanced fibrosis (consistent with F2-F3) without extending to cirrhosis.

Additional exclusion criteria narrowed the population further. Patients with decompensated liver disease, uncontrolled diabetes (HbA1c >9%), bariatric surgery within five years, or significant alcohol use were removed. Each exclusion is defensible for internal validity, but collectively they sculpt a trial cohort that is healthier, more metabolically stable, and less complex than the real-world MASH population presenting to hepatology clinics.

Race and ethnicity data from the published trial show the cohort was predominantly White (approximately 95%). Hispanic/Latino patients, who carry disproportionate MASH burden in U.S. epidemiologic data, were underrepresented relative to disease prevalence.

The 52-Week Biopsy Window: Too Short for Fibrosis Conclusions?

Liver fibrosis progresses and regresses slowly. A 52-week interval between biopsies is standard for regulatory acceptance in MASH trials, but it compresses the timeline for detecting meaningful structural remodeling. The primary endpoint, a ≥1-stage fibrosis improvement on the NASH CRN scale, captures early directional change. Whether that change translates to reduced portal hypertension, fewer varices, or lower hepatocellular carcinoma incidence over five to ten years is unknown.

The AASLD practice guidance on MASLD, published in Hepatology (2023), explicitly notes that no pharmacotherapy trial in MASH has yet demonstrated improvement in clinical outcomes such as liver-related mortality or transplant-free survival. MAESTRO-NASH's accelerated approval was granted on the basis of histological surrogates, with a confirmatory outcomes trial (MAESTRO-NASH-OUTCOMES) still ongoing.

The HealthRX.com Surrogate-to-Outcome Gap Framework

To contextualize the limitation above, we apply a structured assessment of how far the trial's endpoints sit from hard clinical outcomes.

The entire regulatory case rests on Layer 2 (histological). This is not unusual for accelerated approval, but prescribers should understand that "fibrosis improvement" on biopsy does not yet equal "fewer patients dying of liver disease." The confirmatory trial will need to close that gap, or the approval could theoretically be withdrawn.

Biopsy Variability and Central Reading

Liver biopsy is subject to sampling error. A single 15-20 mm core captures roughly 1/50,000th of total liver volume. Fibrosis distribution in MASH is patchy, so a repeat biopsy at week 52 might sample a different zone than the baseline core. MAESTRO-NASH used central pathology reading by two independent hepatopathologists, with adjudication for discordant reads. This is better than local reading, but it does not eliminate the spatial sampling problem inherent to needle biopsy.

The NAS scoring system itself is semiquantitative. A one-point change in ballooning or lobular inflammation can shift a patient across the "NASH resolution" threshold. In the trial's primary analysis, the absolute difference in NASH resolution between resmetirom 100 mg (29.9%) and placebo (9.7%) was approximately 20 percentage points. That margin is clinically meaningful, but the responder definition is binary (resolved vs. not), which obscures the continuous biology underneath.

The Placebo Response Problem

A 9.7% placebo NASH-resolution rate at 52 weeks is not trivial. Placebo responses in MASH trials historically range from 10% to 25%, driven partly by lifestyle counseling provided to all participants and partly by regression to the mean on repeat biopsy. In the REGENERATE trial of obeticholic acid, the placebo fibrosis improvement rate was 23% at 18 months.

MAESTRO-NASH's placebo rate was lower than some predecessors, which strengthened the relative treatment effect. But the variability in placebo response across MASH trials makes cross-trial comparisons unreliable. Claims that resmetirom "outperforms" other MASH drug candidates based on indirect comparison of responder rates from different trials are statistically unsound without a head-to-head design.

Conflict-of-Interest and Sponsor Involvement

Madrigal Pharmaceuticals funded MAESTRO-NASH, and the company's employees were involved in trial design, data collection, statistical analysis, and manuscript preparation. The corresponding author (Stephen Harrison) has disclosed consulting relationships with Madrigal and multiple other pharmaceutical companies active in the MASH space.

This does not invalidate the findings, but it is a structural conflict that readers should weigh. Post-marketing pharmacovigilance and independent replication will be important. The FDA's approval review noted the histological benefit was "reasonably likely to predict clinical benefit," a deliberately hedged phrase that signals regulatory confidence is conditional.

Statistical Design: Multiplicity and Alpha Spending

MAESTRO-NASH had two co-primary endpoints (NASH resolution and fibrosis improvement), each tested at both doses (80 mg and 100 mg). The trial used a hierarchical testing procedure to control the family-wise Type I error rate. Both doses met both endpoints, so the multiplicity adjustment did not create interpretive problems in this case. Had only one dose or one endpoint succeeded, the hierarchy would have determined which findings could be reported as statistically significant, a subtlety that press coverage routinely ignores.

The trial also had numerous secondary and exploratory endpoints (liver fat by MRI, lipid panels, liver enzymes, patient-reported outcomes). These were not adjusted for multiplicity and should be interpreted as hypothesis-generating rather than confirmatory, regardless of their p-values.

What Post-Publication Commentary Raised

Several editorials and letters following the NEJM publication highlighted recurring concerns:

• Thyroid axis effects. Resmetirom is a selective thyroid hormone receptor beta agonist. While it avoids THR-alpha-mediated cardiac effects in short-term data, long-term consequences of chronic THR-beta activation on bone metabolism, lipid flux, and hepatic mitochondrial stress remain under investigation.
• LDL reduction as a confound. Resmetirom lowered LDL cholesterol by roughly 14-22%. Some commentators questioned whether part of the histological improvement reflects improved metabolic milieu (lower lipotoxicity) rather than direct antifibrotic action, which has implications for combination therapy design.
• Dropout and missing biopsies. Approximately 6-8% of randomized patients did not have evaluable paired biopsies. While sensitivity analyses using multiple imputation supported the primary findings, any missing-data pattern in a biopsy-dependent trial introduces potential bias.

Generalizability to Global MASH Populations

MAESTRO-NASH was conducted across sites in the United States, Europe, and select other regions. Patients from East Asia, South Asia, and Sub-Saharan Africa, areas where MASH prevalence is rising rapidly and where lean MASH phenotypes are more common, were largely absent. The lean MASH phenotype (BMI <25 kg/m²) may have different fibrosis biology and drug response profiles. No subgroup analysis in the published data addresses this phenotype specifically.

Patients with MASH and concurrent type 2 diabetes were included but were not the majority. Given that type 2 diabetes is the strongest independent predictor of fibrosis progression in MASH (Rinella et al., Hepatology 2023), the trial's mixed diabetic/non-diabetic population may dilute or obscure differential treatment effects in the highest-risk subgroup.

What the Confirmatory Trial Must Answer

The ongoing MAESTRO-NASH-OUTCOMES trial is designed to evaluate whether resmetirom reduces liver-related clinical events (decompensation, transplant, liver-related death) and all-cause mortality over a longer follow-up period. Until those data mature, the risk-benefit profile of resmetirom rests on histological surrogates, favorable but incomplete safety data through 52 weeks, and a biological rationale grounded in THR-beta agonism.

If the outcomes trial is negative or equivocal, the accelerated approval could be revisited. This is not hypothetical: the FDA withdrew the accelerated approval of obeticholic acid for primary biliary cholangitis indication discussions after confirmatory trial concerns. MASH drug development has a history of late-stage failures (elafibranor, cenicriviroc, selonsertib), and each failure was preceded by encouraging surrogate endpoint data.

Frequently asked questions

›Why was the MAESTRO-NASH biopsy window only 52 weeks?

Regulatory agencies accept 52-week paired biopsies as standard for accelerated approval in MASH. Longer biopsy intervals (e.g., 18-24 months) would delay approval timelines. The trade-off is that 52 weeks may be too brief to capture durable fibrosis remodeling vs. early anti-inflammatory effects.

›Were patients with cirrhosis included in MAESTRO-NASH?

No. The trial enrolled patients with fibrosis stages F1B through F3. Patients with compensated cirrhosis (F4) were excluded from the primary efficacy analysis, leaving the highest-risk fibrosis group without controlled data.

›How reliable is liver biopsy as an endpoint?

Liver biopsy remains the reference standard for fibrosis staging, but it samples a tiny fraction of the liver. Interobserver variability and spatial sampling error can cause misclassification. MAESTRO-NASH mitigated this with central reading by two pathologists, though the fundamental sampling limitation persists.

›Does NASH resolution on biopsy mean the disease is cured?

Not necessarily. "NASH resolution" is defined by absence of ballooning and minimal inflammation on a single biopsy sample. It does not guarantee that the underlying metabolic drivers (insulin resistance, adipose dysfunction) are resolved, and histological relapse after treatment cessation is plausible.

›What was the placebo response rate in MAESTRO-NASH?

Approximately 9.7% of placebo patients achieved NASH resolution at 52 weeks. This is lower than some prior MASH trials but still clinically significant, driven by lifestyle counseling and biopsy regression to the mean.

›Did Madrigal Pharmaceuticals fund the entire trial?

Yes. Madrigal funded the trial and had employees involved in design, data collection, analysis, and manuscript writing. This is standard for industry-sponsored Phase 3 trials but represents a structural conflict of interest that independent replication should address.

›Is resmetirom safe for long-term use?

Safety data through 52 weeks showed resmetirom was generally well tolerated. Diarrhea, nausea, and mild liver enzyme elevations were the most common adverse events. Long-term safety over years of use, particularly regarding thyroid axis and bone metabolism effects, awaits data from the ongoing outcomes trial.

›Can MAESTRO-NASH results be compared to other MASH drug trials?

Cross-trial comparisons are unreliable because MASH trials differ in patient populations, biopsy timing, placebo response rates, and endpoint definitions. Only a head-to-head randomized trial can establish relative efficacy between resmetirom and other candidates.

›What happens if the confirmatory outcomes trial fails?

The FDA could withdraw resmetirom's accelerated approval. This has precedent in hepatology. Prescribers and patients should understand that current approval is conditional on future demonstration of clinical outcome benefit.

›Were diverse racial and ethnic populations represented?

The trial cohort was approximately 95% White. Hispanic/Latino and Black patients, as well as populations from East and South Asia where lean MASH phenotypes are more prevalent, were underrepresented relative to global disease burden.

References

1. Harrison SA, Bedossa P, Guy CD, et al. A Phase 3, Randomized, Controlled Trial of Resmetirom in NASH with Liver Fibrosis. N Engl J Med. 2024;390(6):497-509. PubMed
2. Rinella ME, Lazarus JV, Ratziu V, et al. A multisociety Delphi consensus statement on new fatty liver disease nomenclature. Hepatology. 2023;78(6):1966-1986. PubMed
3. Younossi ZM, Ratziu V, Loomba R, et al. Obeticholic acid for the treatment of non-alcoholic steatohepatitis: interim analysis from a multicentre, randomised, placebo-controlled phase 3 trial. Lancet. 2019;394(10215):2184-2196. PubMed
4. U.S. Food and Drug Administration. Rezdiffra (resmetirom) prescribing information. 2024. FDA Label
5. Loomba R, Hartman ML, Lawitz EJ, et al. Tirzepatide for Metabolic Dysfunction-Associated Steatohepatitis with Liver Fibrosis. N Engl J Med. 2024;391(4):299-310. PubMed