Study analysis · The lancet. Gastroenterology & hepatology · 2023

The blockbuster weight-loss drug semaglutide failed to reverse liver scarring in a groundbreaking cirrhosis trial — but it did melt liver fat and improve metabolic health.

In people with liver scarring from NASH, weekly semaglutide shots for a year did not significantly improve scarring, but they reduced liver fat and helped weight and blood sugar.

Reading level
Moderate certainty
Level 1b · Individual RCTAssociation, not causationNo causal claims

Overview

What the study found

The study in plain English — the bottom line, every takeaway we extracted, and what to do with them.

In simple terms

This was a randomized controlled trial, which is one of the best ways to test if a medicine works. It compared semaglutide to a placebo in people with severe liver scarring (cirrhosis). The main result was that semaglutide did not significantly improve liver scarring or inflammation, so we cannot say it fixes the liver. It did help with weight loss and blood sugar, which are important for overall health.

What’s the bottom line?

A 48-week trial tested weekly semaglutide 2.4 mg against placebo in 71 people with NASH-related cirrhosis. Semaglutide did not significantly improve liver scarring or clear NASH, but it did reduce liver fat and improve weight and blood sugar. Corrections/errata have been published; check the notices for updated information.

How strong is this study?

The study was well-designed because it was randomized and double-blind, meaning neither patients nor doctors knew who got the real medicine, which helps prevent bias. But the study was very small (only 71 people), so it may not have had enough power to detect a real difference in liver scarring. Also, the main goal was changed during the study, which can make results harder to trust.

Reporting

0 / 100

  • COI disclosureconflicts of interest not disclosed
  • Data availabilitydata not shared
  • Code availabilitycode not shared
Methodology

82 / 100

  • Randomization+20/20
  • Blinding+15/15
  • Control group+15/15
  • Sample size (n=71)+6.0/20
  • Follow-up+10/10
Publication

100 / 100

Statistical

100 / 100

  • P-values+15/15
  • Effect size+20/20
  • Confidence intervals+15/15
  • Pre-registration+15/15

Each component is scored out of 100 and then capped by the study design — a case series cannot reach the ceiling a randomised trial can, however well it is reported.

Where it sits

RCT reviews

Max 100

Randomized Trials

Max 90

Reviews of Cohort Studies

Max 85

Cohort Studies

Max 72

Reviews of Case-Control Studies

Max 63

Case-Control Studies

Max 58

Cross-Sectional & Case Series

Max 50

Expert Opinion

Max 5
StrongerWeaker
Randomized Trials
Level 1b
74

74 / 100

Probability of being correct

Participants are randomly assigned to treatment or control groups, minimizing bias. The gold standard for testing whether an intervention causes an effect.

This design cannot establish causation — the findings describe an association, not a cause. Although the study is a randomized, double-blind, placebo-controlled trial (Level 1b), it is underpowered for its revised primary endpoint, the confidence interval for the primary outcome is wide and crosses 1 (OR 0.28, 95% CI 0.06–1.24), and the primary endpoint was not statistically significant. Therefore, this trial cannot establish that semaglutide causes improvement in liver fibrosis or NASH resolution. Significant secondary outcomes (weight loss, HbA1c, liver fat, liver enzymes) may support causal inference with caution, but they are exploratory and not adjusted for multiplicity.

Major COI

Major conflicts that significantly reduce study credibility

Not Disclosed

The study likely tests semaglutide, a Novo Nordisk product, and is sponsored by industry, but no funding or conflict-of-interest statement is provided in the excerpt. This constitutes a major potential conflict because the funder has a direct financial interest in the study drug.

Industry Funded
Undisclosed — Suspicious

Funders

Novo Nordisk (implied by semaglutide, not explicitly named in text)

Independent Analysis Safeguards

  • Double-blind, placebo-controlled design
  • Central randomisation via interactive web response system
  • Patients, investigators, trial site staff, and sponsor remained blinded
  • Outcome assessed by a central pathologist blinded to visit, patient characteristics, and treatment
  • MRI scans analysed by a central imaging supplier blinded to treatment

The provided excerpt lacks a conflict-of-interest or funding statement. Semaglutide is a Novo Nordisk product, making industry funding highly likely, but the funder is not named in the text. The primary endpoint was changed from MRE to liver biopsy after protocol completion but before unblinding, which could indicate sponsor or regulatory influence. The negative result reduces concern that the study serves as marketing, but the undisclosed industry sponsorship remains a major COI concern.

Key takeaways

  1. 01

    Primary: 11% vs 29% had liver fibrosis improvement (relative odds ratio 0.28; p=0.087; absolute difference about 18 fewer per 100 people, not statistically significant).

  2. 02

    NASH resolution: 34% vs 21% (relative odds ratio 1.97; p=0.29; absolute difference about 13 more per 100, not statistically significant).

  3. 03

    Liver fat ≥30% reduction: 49% vs 13% (relative odds ratio 6.58; p=0.0037; absolute difference about 36 more per 100).

  4. 04

    Weight: mean relative change -8.83% vs -0.09%; absolute difference -8.75 percentage points.

  5. 05

    HbA1c in diabetes: -1.39% vs +0.24%; absolute difference -1.63 percentage points.

  6. 06

    ALT drop ≥17 U: 40% vs 8%; absolute difference about 32 more per 100.

  7. 07

    For the main liver scarring endpoint, the result was not statistically significant, so semaglutide did not clearly improve scarring; the absolute difference was about 18 fewer per 100 but could be chance.

  8. 08

    The liver fat and metabolic improvements were statistically significant, with about 36 more per 100 achieving ≥30% liver fat reduction and an average 8.75 percentage point greater weight loss.

  9. 09

    The trial was small and underpowered, and absolute risks for clinical outcomes like death or decompensation were not reported.

  10. 10

    Published corrections/errata exist; check them for updated information.

Surprising findings

  • Placebo numerically outperformed semaglutide on the primary endpointIt is counterintuitive that a placebo group would have more than double the rate of fibrosis improvement (29% vs 11%), even if not statistically significant. This may reflect baseline imbalance or biopsy sampling variability.
  • Liver fat plummeted without fibrosis improvementMany assume reducing liver fat automatically reverses scarring; this trial shows a dramatic fat reduction (49% vs 13%) with no significant scar benefit.
  • No liver decompensation despite advanced cirrhosisGLP-1s cause GI side effects and weight loss, which could be risky in cirrhosis, but no decompensating events or deaths occurred.

Practical takeaways

If you have NASH cirrhosis, don't expect semaglutide alone to reverse scarring based on this trial.

Small, underpowered phase 2; secondary endpoints improved; check published corrections/errata.

low-moderate for primary endpoint confidence

Semaglutide may still improve liver fat, weight, blood sugar, and ALT; discuss with your hepatologist.

Secondary endpoints, not clinical outcomes; GI side effects common.

moderate confidence

Monitor for GI side effects like nausea, diarrhea, vomiting if prescribed.

Most were mild-moderate and transient, but some led to dose reduction/discontinuation.

high confidence

Don't use liver enzyme improvements as proof of cirrhosis reversal.

ALT improved significantly but fibrosis did not; need biopsy or validated non-invasive tests.

moderate confidence

Why this study matters

Primary endpoint miss: no significant fibrosis improvement

In 71 adults with NASH-related compensated cirrhosis, semaglutide 2.4 mg weekly did not significantly improve liver fibrosis by ≥1 stage without worsening NASH: 11% (5/47) vs 29% (7/24) with placebo; relative odds ratio 0.28 (95% CI 0.06–1.24; p=0.087); absolute difference about 18 fewer per 100 patients, not statistically significant. NASH resolution also did not differ significantly: 34% vs 21%; relative OR 1.97 (95% CI 0.56–7.91; p=0.29); absolute difference about 13 more per 100.

GLP-1 drugs are hyped as miracle treatments; this trial shows they may not reverse advanced liver scarring, challenging expectations for cirrhosis.

Liver fat melted, but scars stayed

Semaglutide significantly reduced liver fat: ≥30% reduction in 49% (23/47) vs 13% (3/24) with placebo; relative odds ratio 6.58 (95% CI 1.63–39.31; p=0.0037); absolute difference about 36 more per 100. MRI-PDFF estimated treatment ratio 0.67 (95% CI 0.51–0.88; p=0.0042).

It suggests liver fat and liver scarring are not the same thing; you can improve one without fixing the other, which matters for how we measure treatment success.

Weight loss was substantial

Mean relative weight change was -8.83% with semaglutide vs -0.09% with placebo; absolute treatment difference -8.75 percentage points (95% CI -12.41 to -5.09; p<0.0001). 62% on semaglutide lost ≥5% body weight and 40% lost ≥10%, vs 25% and 0% on placebo.

Weight loss is the mainstay of NASH treatment, and semaglutide delivered it even in people with cirrhosis, but we don't know if it was healthy muscle or fat loss.

Blood sugar control improved in diabetes

Among patients with type 2 diabetes, HbA1c changed by -1.39% with semaglutide vs +0.24% with placebo; absolute treatment difference -1.63 percentage points (95% CI -2.20 to -1.06; p<0.0001). 75% of trial participants had diabetes.

Diabetes worsens liver outcomes, so improving glucose control could have indirect liver benefits even if scarring didn't reverse.

Liver enzymes dropped

ALT improved significantly: estimated treatment ratio 0.76 (95% CI 0.61–0.93; p=0.0090). A clinically significant ≥17 U decrease occurred in 40% (19/47) vs 8% (2/24) with placebo; absolute difference about 32 more per 100 (p=0.0057). AST and GGT also improved.

ALT is a common liver blood test; patients may see improvements and assume scarring is reversing, but this trial shows that may not be true.

Safety: GI side effects, but no liver decompensation

Adverse events were similar overall (89% vs 79%) and serious AEs (13% vs 8%). Nausea (45% vs 17%), diarrhea (19% vs 8%), and vomiting (17% vs 0%) were more common with semaglutide. There were no decompensating events or deaths.

Cirrhosis patients are fragile; the absence of liver decompensation is reassuring, but GI side effects could hurt quality of life.

Want the whole report?

Detailed mode opens the full scientific breakdown — every score component, the methodology, conflicts of interest, the evidence analysis behind each claim, and the raw study data.

Standing

The people behind it

The researchers who wrote the study this analysis is built on.

Authored by

11 researchers

If this is your work, this is how we attribute it on Fit Body Science. Rohit Seth Loomba is listed as the lead author.