Does a GLP-1 Cause Liver Damage? The Evidence on NAFLD, MASH, and Why the Answer Is Usually the Opposite

No — GLP-1 receptor agonists do not typically cause liver damage, and in the randomized trials that have looked specifically at the liver, semaglutide and tirzepatide do something closer to the opposite: they reduce liver fat, lower ALT and AST, and in dedicated phase 2 trials improve the histology of metabolic dysfunction-associated steatohepatitis (MASH, formerly called NASH). Newsome and colleagues showed semaglutide resolved NASH in a significantly higher fraction of patients than placebo^[1], and the 2024 SYNERGY-NASH trial reported that tirzepatide improved both steatohepatitis resolution and fibrosis stage^[2]. There is a small postmarketing drug-induced liver injury signal, but it is rare, and the dominant hepatic story for GLP-1s is benefit, not harm. This article walks through the evidence, the rare adverse signals, and how to monitor.

Why patients ask this question

The fear of liver damage on a GLP-1 usually comes from three places. First, the Wegovy^[7] and Zepbound^[8] labels carry a pancreatitis precaution in Section 5, and many patients conflate “pancreas” with “liver” (they sit adjacent to each other under the right ribs and both are abdominal organs that weight-loss patients worry about). Second, almost every drug class has some association with drug-induced liver injury (DILI) in the public imagination — statins, acetaminophen, antibiotics — and patients reasonably ask whether injectables are any different. Third, patients who carry obesity often already know they have “fatty liver” from a prior ultrasound or an elevated ALT on routine labs, and they worry that adding a new drug will make it worse.

The honest answer, grounded in the trial data, is that the GLP-1 class is among the most thoroughly studied drug classes for liver safety in the obesity population, and the studies that have looked at the liver directly have shown improvement, not harm.

The fatty liver context: MASLD and MASH

In 2023 a multisociety Delphi consensus^[3] replaced the old terms “nonalcoholic fatty liver disease” (NAFLD) and “nonalcoholic steatohepatitis” (NASH) with two new names: MASLD (metabolic dysfunction-associated steatotic liver disease) and MASH (metabolic dysfunction-associated steatohepatitis). The rationale was that calling the disease “nonalcoholic” defined it by what it was not, and the new nomenclature instead names the metabolic drivers. The diseases themselves are the same: MASLD is fat in the liver (steatosis) in a patient with metabolic risk factors; MASH is MASLD plus inflammation and hepatocyte injury (ballooning), which can progress to fibrosis, cirrhosis, and hepatocellular carcinoma.

The prevalence of MASLD in adults with obesity is very high — most cohort estimates put it above half of adults with a BMI over 30, and roughly a fifth of those have progressed to MASH. This matters for the liver-damage question in two ways. First, most patients who start a GLP-1 already have some degree of fatty liver disease before they ever take an injection, so the background rate of abnormal liver enzymes and hepatic findings is high. Second, this is exactly the population in which GLP-1s have been shown to benefit the liver: the metabolic-hepatic axis is what these drugs target.

The trial evidence: semaglutide in NASH (Newsome 2021)

The pivotal trial establishing semaglutide as having activity in steatohepatitis is Newsome and colleagues, published in the New England Journal of Medicine in 2021^[1]. This was a 72-week phase 2 randomized, placebo-controlled trial of once-daily semaglutide (0.1, 0.2, or 0.4 mg) versus placebo in 320 patients with biopsy-confirmed NASH and stage F1-F3 fibrosis. The primary endpoint was NASH resolution without worsening of fibrosis on repeat biopsy at 72 weeks.

At the 0.4 mg dose, 59% of semaglutide patients achieved NASH resolution versus 17% on placebo — a large and statistically significant effect. ALT and AST both fell substantially, body weight fell by about 13%, and HbA1c improved in the patients with diabetes. The catch: the prespecified fibrosis endpoint (improvement in fibrosis stage without worsening of NASH) did not reach statistical significance at 72 weeks. Fibrosis is a slower process, and the trial was likely underpowered and too short to detect improvement in the scar tissue itself. NASH resolution is still clinically meaningful — it removes the inflammatory driver — but the trial did not prove semaglutide reverses established fibrosis.

The trial evidence: tirzepatide in MASH (SYNERGY-NASH 2024)

The 2024 SYNERGY-NASH trial^[2], published in NEJM, took the next step. This was a 52-week phase 2 trial of tirzepatide (5, 10, or 15 mg weekly) versus placebo in 190 patients with biopsy-confirmed MASH and stage F2-F3 fibrosis. Tirzepatide hit both endpoints: a significantly higher proportion of patients on each active dose achieved resolution of steatohepatitis without worsening of fibrosis, and a significantly higher proportion achieved improvement in fibrosis stage by at least one level without worsening of MASH. At the 15 mg dose, roughly 62% achieved steatohepatitis resolution versus 10% on placebo, and roughly 55% achieved fibrosis improvement versus 30% on placebo.

This is the first GLP-1 class drug to show both inflammatory and fibrotic improvement in a randomized trial. The effect size on fibrosis is the bigger deal clinically, because fibrosis is what drives long-term outcomes (cirrhosis, liver cancer, liver transplant). A phase 3 trial is underway.

The STEP and SURMOUNT liver enzyme sub-analyses

Outside the dedicated NASH trials, the large weight-loss registration trials have reported on liver enzymes as exploratory endpoints. In STEP-1^[5], the 68-week phase 3 trial of weekly semaglutide 2.4 mg in 1,961 adults with obesity, ALT and AST both fell modestly in the semaglutide arm relative to placebo, and the proportion of patients with elevated ALT at baseline who normalized by week 68 was higher on semaglutide. In SURMOUNT-1^[6], the 72-week phase 3 trial of weekly tirzepatide in 2,539 adults with obesity, the liver enzyme and FibroScan sub-analyses similarly showed reductions in ALT, AST, and controlled attenuation parameter (CAP — a proxy for hepatic steatosis) on tirzepatide versus placebo.

These are not dedicated liver-outcome endpoints and should not be over-read, but they tell a consistent story: in the weight-loss trials that were not designed to study the liver, the liver got better on the drug, not worse.

Rezdiffra and what it tells us

In March 2024, the FDA approved resmetirom (brand name Rezdiffra) for MASH with stage F2-F3 fibrosis — the first drug ever approved specifically for steatohepatitis^[4]. This matters as context for the GLP-1 liver question for two reasons. First, it establishes a regulatory pathway: MASH is now a disease the FDA will approve drugs for, and the expected trial designs and endpoints are clear. Tirzepatide, with its positive phase 2 data, is the most likely GLP-1 to follow. Second, Rezdiffra is not a GLP-1 — it is a thyroid hormone receptor beta agonist — so the GLP-1s and Rezdiffra are potentially complementary, not competitive. Patients with MASH who are also obese and diabetic may end up on both classes in the future.

Adverse hepatic events in the registration trials

Across the STEP^[5], SURMOUNT^[6], and Newsome^[1] trials, the reported rates of hepatic adverse events on semaglutide and tirzepatide were low and generally similar to placebo. A small numerical excess of transiently elevated ALT was reported in a subset of patients in some trials, but these elevations were mild, did not meet the threshold for drug-induced liver injury, and did not lead to drug discontinuation at meaningfully higher rates than placebo. Critically, no Hy's law cases — the pattern of combined ALT elevation and bilirubin elevation that is the strongest clinical signal of serious drug-induced liver injury — were reported in the major registration trials. Hy's law is the standard the FDA uses to flag a drug as hepatotoxic, and GLP-1s did not trip it.

Postmarketing pharmacovigilance — what the FAERS signal says

Once a drug reaches the broader population, the FDA Adverse Event Reporting System (FAERS) collects spontaneous reports. For semaglutide and tirzepatide, FAERS does contain a small number of drug-induced liver injury reports. Interpreting them is difficult because the patients on these drugs are the exact population with the highest background rate of MASLD and abnormal liver enzymes — distinguishing a DILI event from a pre-existing liver disease flare is genuinely hard. The signal is modest, has not led either the FDA or the EMA to add a specific DILI warning to the labels, and is orders of magnitude smaller than the pancreatitis signal or the gallbladder signal. The Wegovy^[7] and Zepbound^[8] labels do not carry a hepatotoxicity warning.

Who should be cautious about a GLP-1 for liver reasons

• Active acute hepatitis (viral, autoimmune, alcoholic, or drug-induced) — any acute liver injury should be fully worked up and managed before adding a new drug with even a small hepatic signal. This is general prescribing prudence, not a specific GLP-1 contraindication.
• Decompensated cirrhosis (Child-Pugh C) — the Wegovy^[7] and Zepbound^[8] labels note that clinical experience in patients with severe hepatic impairment is limited. Neither label contains an absolute contraindication, but prescribing in a Child-Pugh C patient should be done only with hepatology input.
• Unexplained persistent ALT or AST elevations before starting — these should be worked up (viral panel, autoimmune markers, iron studies, imaging) before starting a GLP-1, not after, so that any future enzyme changes can be interpreted against a known baseline.

Who specifically benefits

The patient group with the most upside from a GLP-1 on the liver axis is the one where the metabolic, hepatic, and glycemic problems all sit together: adults with obesity, type 2 diabetes, and MASLD or MASH. In this group, a GLP-1 addresses all three simultaneously — weight comes down, glycemia improves, and liver fat and inflammation both fall. The SYNERGY-NASH^[2] data specifically support tirzepatide in this population, and while tirzepatide is not yet approved by the FDA for MASH (that approval is pending further phase 3 data), its approval for obesity and type 2 diabetes makes it routinely prescribable for patients who meet those primary indications and happen to have MASH as well.

What to monitor

A reasonable monitoring approach for a patient starting a GLP-1 who already has metabolic risk factors for liver disease:

• Baseline labs before the first injection: ALT, AST, alkaline phosphatase, total and direct bilirubin, albumin, INR, and a hepatitis panel if not done recently. A baseline FibroScan or imaging-based steatosis and fibrosis assessment is reasonable in higher-risk patients (diabetics, BMI over 35, or patients with known abnormal enzymes).
• Recheck at 3 to 6 months, then annually while on therapy. The expectation in most patients is that ALT and AST will fall as weight comes down.
• Triggered workup for new symptoms: right-upper-quadrant pain, jaundice (yellowing of the eyes or skin), dark urine, pale stool, severe fatigue, or nausea that worsens weeks into therapy (as opposed to the typical titration nausea in the first month) should prompt immediate liver labs and evaluation. These symptoms are not expected on a GLP-1 and should never be dismissed as “just the drug.”

Bottom line

• GLP-1s do not typically cause liver damage. In the trials that specifically studied the liver, semaglutide and tirzepatide improved NASH/MASH histology.
• Newsome 2021 showed semaglutide resolved NASH in 59% of patients vs 17% on placebo; SYNERGY-NASH 2024 showed tirzepatide improved both MASH resolution and fibrosis stage.
• STEP-1 and SURMOUNT-1 sub-analyses showed ALT and AST fall on both drugs in the weight-loss population.
• No Hy's law cases have been reported in the major registration trials, and neither the Wegovy nor Zepbound label carries a hepatotoxicity warning.
• A rare postmarketing DILI signal exists in FAERS but is small and has not prompted regulatory label changes.
• Baseline liver labs before starting and rechecking at 3 to 6 months is a reasonable, conservative monitoring approach in the metabolic population.

Related research and tools

• Does a GLP-1 cause cancer? MTC, pancreatic, and the BMJ signal — the sibling patient-anxiety article on cancer risk
• 17 GLP-1 side effect questions answered — every common patient concern with trial-data context
• SELECT trial cardiovascular benefits in non-diabetics — the largest non-diabetic semaglutide outcomes trial
• FLOW trial: semaglutide in kidney disease — another major end-organ outcomes trial
• GLP-1 drug interaction checker — search any medication, including hepatically metabolized drugs, for interaction with your GLP-1

Important disclaimer. This article is educational and does not constitute medical advice. Any patient with known liver disease, persistent unexplained enzyme elevations, or symptoms of acute hepatic injury (jaundice, dark urine, right-upper-quadrant pain, severe fatigue) should be evaluated by a clinician before starting or continuing a GLP-1. Weight Loss Rankings does not provide medical advice, diagnosis, or treatment recommendations. If you experience signs of possible liver injury while on a GLP-1, contact your prescriber promptly.