GLP-1s and Hepatic Outcomes: The MASLD Evidence Base Through Mid-2026

Metabolic dysfunction–associated steatotic liver disease is the condition formerly discussed under the NAFLD label, renamed in 2023 to put the metabolic driver in the name. It is also the area where GLP-1 evidence has moved fastest between 2023 and 2026. This is a review of the trial base as it stands.

Why the endpoints are difficult

Liver disease trials are slow and expensive because the outcomes that matter most take decades to accrue. Regulators therefore accept histologic surrogates: a biopsy at baseline and a biopsy at the end of the treatment period, read by blinded pathologists, scored on established systems.

Two co-primary endpoints have become standard in MASH trials:

1. Resolution of steatohepatitis with no worsening of fibrosis. The inflammatory and ballooning components resolve, and scarring does not progress.
2. Improvement in liver fibrosis with no worsening of steatohepatitis. Scarring regresses by at least one stage, and inflammation does not get worse.

Both endpoints depend on paired biopsy, which introduces sampling variability — a needle samples a tiny fraction of the organ — and reader variability. Well-run trials mitigate this with central blinded reading, but the noise does not disappear.

The ESSENCE trial

ESSENCE evaluated once-weekly semaglutide in participants with biopsy-confirmed MASH and fibrosis, with results published in the New England Journal of Medicine in 2025. It reported that a significantly greater proportion of semaglutide-treated participants achieved resolution of steatohepatitis without worsening of fibrosis compared with placebo, and a greater proportion achieved fibrosis improvement without worsening of steatohepatitis.

The trial is the most substantial piece of evidence for a GLP-1 receptor agonist on histologic liver endpoints to date. Its design — biopsy-confirmed entry, paired biopsy at the analysis timepoint, central reading — is the appropriate one for the question.

The confounding that has not been resolved

Semaglutide produces substantial weight loss. Weight loss independently improves hepatic steatosis, inflammation and, at sufficient magnitude, fibrosis — this was established in the lifestyle-intervention literature well before GLP-1s entered the field. The pivotal interpretive question is therefore how much of the hepatic effect is drug-specific and how much is a downstream consequence of weight reduction that any equivalent weight loss would have produced.

Current trials cannot fully separate the two, because the drug and the weight loss are not independently assignable. Approaches that partially address it include weight-matched comparator arms, mediation analysis within trial datasets, and testing agents that improve liver histology without substantial weight change. Each has limitations, and none has settled the question.

From a patient's perspective the distinction may be academic — if the liver improves, the mechanism is secondary. From a scientific and formulary perspective it is not, because it determines whether the hepatic indication should be tied to this drug class specifically or to weight reduction generally.

Non-invasive markers and their limits

Because biopsy is impractical outside trials, clinical practice relies on non-invasive tests: FIB-4 calculated from routine labs and age, transient elastography for liver stiffness, and MR elastography or MRI-PDFF where available. Trials increasingly report these alongside histology.

The interpretive trap is that some non-invasive markers move for reasons unrelated to fibrosis regression. Liver stiffness measurement is affected by hepatic inflammation and congestion, so reduced inflammation can lower a stiffness reading without any change in scarring. Improvement in a non-invasive score is encouraging but is not equivalent to demonstrated fibrosis regression.

What remains unanswered

• Hard outcomes. Progression to cirrhosis, decompensation events, hepatocellular carcinoma incidence and liver-related mortality all require far longer follow-up than the histologic trials provide.
• Durability after discontinuation. Given what the discontinuation literature shows about weight regain, whether hepatic gains persist off treatment is a live and under-studied question.
• Advanced fibrosis and cirrhosis. Trials in compensated cirrhosis are a separate and harder problem than trials in pre-cirrhotic fibrosis.
• Combination approaches. Whether pairing an incretin agent with a mechanistically distinct hepatic agent produces additive benefit is an active area rather than a settled one.