GLP-1 With Cirrhosis & Portal Hypertension: Safe? (2026)

If you have cirrhosis or portal hypertension and take — or want to take — a GLP-1 like Ozempic, Wegovy, Mounjaro or Zepbound, the honest answer depends almost entirely on where you are on the liver-disease spectrum. For early-to-moderate fatty-liver disease (MASH/MASLD without cirrhosis), the evidence is now genuinely strong and even positive: in the phase 3 ESSENCE trial, semaglutide resolved metabolic dysfunction–associated steatohepatitis without worsening fibrosis in 62.9% of patients with stage F2–F3 fibrosis versus 34.3% on placebo (Sanyal 2025^[1]). But for compensated cirrhosis the data are mixed (a fibrosis benefit was not proven), and for decompensated cirrhosis and clinically significant portal hypertension the data are essentially absent — those patients were excluded from every major trial, so the right word is caution, not endorsement. GLP-1 drugs are not contraindicated in liver disease and need no dose change for hepatic impairment on the label, but “not contraindicated” is not the same as “studied and safe.” This article walks the line carefully. For the broader liver-safety question, see does a GLP-1 cause liver damage?

The honest summary

• Pre-cirrhotic MASH (F2–F3): strong, positive evidence. In ESSENCE, semaglutide 2.4 mg weekly resolved steatohepatitis without worsening fibrosis in 62.9% vs 34.3% on placebo, and improved fibrosis without worsening MASH in 36.8% vs 22.4% (Sanyal 2025^[1]). This is the part of the spectrum where GLP-1s shine.
• Compensated cirrhosis (F4, no decompensation): mixed. A dedicated phase 2 trial of semaglutide in NASH-related compensated cirrhosis did not meet its fibrosis-improvement endpoint, though there were no decompensating events or deaths and metabolic parameters improved (Loomba 2023^[2]).
• Decompensated cirrhosis & significant portal hypertension: no trial data — caution. Patients with ascites, variceal bleeding, hepatic encephalopathy, or Child-Pugh C disease were excluded from the major trials. There is no efficacy or safety evidence to support starting a GLP-1 in this group.
• The label allows it — but doesn't study it. The Wegovy/semaglutide label lists no hepatic contraindication and recommends no dose adjustment for hepatic impairment; however, experience in severe hepatic impairment is limited (Wegovy FDA label^[6]).
• GI side effects are the practical danger in portal hypertension. Nausea, vomiting, and reduced intake can worsen volume depletion, electrolyte shifts, and the muscle loss (sarcopenia) that already drives bad outcomes in advanced cirrhosis.
• This is a hepatologist decision. If you have cirrhosis of any kind, do not start, stop, or change a GLP-1 without a liver specialist who knows your fibrosis stage and decompensation status.

Why the disease stage is the whole question

Fatty-liver disease is a ladder. It starts as simple fat accumulation (steatosis / MASLD), can progress to inflammation and ballooning (steatohepatitis / MASH), then to scarring graded F0 (none) through F4 (cirrhosis). Cirrhosis itself splits into two very different worlds: compensated cirrhosis, where the liver is scarred but still doing its jobs, and decompensated cirrhosis, marked by complications of portal hypertension — ascites (fluid in the belly), variceal bleeding, hepatic encephalopathy, or jaundice. Portal hypertension is the high pressure in the portal venous system that scarring creates, and clinically significant portal hypertension is what produces those complications. Every claim about GLP-1 “safety in liver disease” has to be pinned to one of these rungs, because the evidence is wildly different at each.

Why does it matter so much? Two reasons. First, the trials that built the positive GLP-1 liver story deliberately enrolled people on the lower rungs and excluded the higher ones. Second, the biology changes: in decompensated cirrhosis the liver and kidneys are precariously balanced, blood volume is dysregulated, and patients are often sarcopenic and frail. A drug whose main side effects are appetite loss, nausea, and vomiting behaves very differently in someone with F2 fibrosis and obesity than in someone with ascites and muscle wasting.

Pre-cirrhotic MASH (F2–F3): the strong, positive evidence

This is where GLP-1 receptor agonists have earned real evidence. The story began with Newsome 2021, a phase 2 NEJM trial in which once-daily semaglutide produced NASH resolution in up to 59% of patients at the highest dose versus 17% on placebo, without worsening of fibrosis (Newsome 2021^[3]). It culminated in ESSENCE, the phase 3 trial of once-weekly semaglutide 2.4 mg in 1,197 patients with biopsy-confirmed MASH and stage F2 or F3 fibrosis. At 72 weeks, 62.9% on semaglutide achieved resolution of steatohepatitis without worsening of fibrosis, versus 34.3% on placebo; 36.8% improved liver fibrosis by at least one stage without worsening MASH, versus 22.4% (Sanyal 2025^[1]). On the strength of these data, the American Association for the Study of Liver Diseases issued updated practice guidance in November 2025 endorsing semaglutide for eligible MASH patients with F2–F3 fibrosis (Bansal 2025^[4]).

The crucial caveat sits right in the enrollment criteria: ESSENCE studied F2–F3 fibrosis — pre-cirrhotic stages — and excluded patients with cirrhosis. So the headline 63% number, and the FDA pathway it supports, apply to people who have not yet developed cirrhosis. That is exactly the population most likely to benefit and least likely to be harmed. For the full breakdown of that approval pathway, see the Wegovy MASH ESSENCE trial and FDA pathway, and for the staging and action plan, the GLP-1 fatty-liver patient guide.

Compensated cirrhosis (F4): the evidence is mixed

Once scarring reaches cirrhosis, the picture changes. The most direct evidence is Loomba 2023, a dedicated randomized phase 2 trial of semaglutide 2.4 mg weekly in 71 patients with NASH-related compensated cirrhosis (F4). After 48 weeks, the trial did not meet its primary endpoint: an improvement in fibrosis without worsening of NASH occurred in only 11% on semaglutide versus 29% on placebo, and NASH resolution was not significantly different (Loomba 2023^[2]). In plain terms, the histologic benefit seen earlier on the ladder did not materialize once cirrhosis was established — possibly because at F4 the scarring is harder to reverse, or the trial was too small and too short.

The reassuring part: in that compensated-cirrhosis population, semaglutide was generally tolerated, hepatic and renal function stayed stable, and there were no decompensating events or deaths over the trial. Metabolic parameters (weight, glycemic control, lipids) improved as expected. So for a patient with compensated cirrhosis and a strong metabolic reason to take a GLP-1 — obesity, type 2 diabetes, cardiovascular risk — the drug is not off the table, but the expectation should be metabolic benefit and weight loss, not proven reversal of liver scarring. This is a nuanced, individualized call that belongs with a hepatologist.

Decompensated cirrhosis & portal hypertension: no data, real caution

Here the honest answer is the simplest: there is essentially no trial evidence. Patients with decompensated cirrhosis — current or prior ascites, variceal bleeding, hepatic encephalopathy, Child-Pugh class C, or clinically significant portal hypertension — were excluded from ESSENCE, from Loomba 2023, and from the obesity and cardiovascular trials that anchor GLP-1 approvals. The European liver and diabetes societies' MASLD guidelines make the same point: the role of these agents in decompensated disease is unestablished and they are not recommended in that setting (EASL-EASD-EASO 2024^[5]). Absence of data is not proof of harm — but in a YMYL setting it is a strong reason to avoid starting one of these drugs in decompensated cirrhosis outside a clinical trial or a hepatologist's explicit, monitored plan.

What the FDA label actually says about liver disease

Many patients are surprised to learn that liver disease is not a contraindication to semaglutide. The Wegovy label lists only personal/family history of medullary thyroid carcinoma, MEN 2, and prior serious hypersensitivity as contraindications — cirrhosis is not among them. On dosing, the label states that no dose adjustment is recommended for patients with hepatic impairment, because pharmacokinetic studies showed no clinically relevant change in semaglutide exposure across degrees of hepatic impairment (Wegovy FDA label^[6]).

That sounds permissive, and it is — but read it precisely. “No dose adjustment needed” is a pharmacokinetic statement: the drug is cleared by enzymatic breakdown, not the liver, so blood levels do not spike in hepatic impairment. It is not a safety endorsement, and the experience in severe hepatic impairment is explicitly limited. The label permits use; the clinical-trial program never tested decompensated cirrhosis. The label and the evidence base are answering two different questions — and for your decision, the evidence base (or its absence) is the one that should drive caution.

Other liver-disease angles worth knowing

Two related findings round out the picture. First, GLP-1 receptor agonists have been associated with a reduced risk of alcohol-associated cirrhosis in patients with type 2 diabetes and alcohol use disorder, possibly via effects on both metabolism and alcohol intake (Hwang 2026^[9]) — a hint that the benefit may extend beyond metabolic liver disease, though this is observational. Second, the broad cohort and meta-analytic data on hepatic-event prevention (Passos 2026^[8]; Kuo 2025^[7]) consistently favor GLP-1–containing regimens for preventing progression in non-cirrhotic and compensated metabolic liver disease. The throughline: the earlier you intervene on the fibrosis ladder, the stronger the case; the more decompensated the cirrhosis, the thinner the evidence and the higher the caution.

Bottom line

The safety of a GLP-1 with cirrhosis is a tale of two answers. For pre-cirrhotic MASH (F2–F3), the phase 3 ESSENCE data are strong and positive, semaglutide is now guideline-endorsed for that group, and the liver signal is protective (Sanyal 2025^[1]; Bansal 2025^[4]). For compensated cirrhosis (F4), a dedicated trial failed to show a fibrosis benefit but found no decompensation or deaths — so it is a tolerable, metabolically useful option in selected patients, not a proven liver therapy (Loomba 2023^[2]). For decompensated cirrhosis and clinically significant portal hypertension, there is no trial evidence at all; those patients were excluded everywhere, the GI side effects carry real volume-depletion and sarcopenia risks, and the appropriate posture is caution and specialist oversight (EASL-EASD-EASO 2024^[5]). The label permits GLP-1 use in hepatic impairment without a dose change, but “permitted” is not “studied” (Wegovy FDA label^[6]). If you have cirrhosis of any stage, this is a hepatologist's decision — anchored to your exact fibrosis stage and decompensation status, not to a headline number.

This article is educational and is not medical advice. Every claim above is sourced to a peer-reviewed trial, society clinical practice guidance, or the FDA-approved label, verified against the live PubMed database and DailyMed before publication. Decisions about GLP-1 use in cirrhosis or portal hypertension must be made with your hepatologist.