Fatty liver disease is one of the most common chronic conditions most people have never heard of. Its current name is metabolic dysfunction-associated steatotic liver disease, or MASLD: fat building up in the liver alongside things like high blood sugar, high blood pressure, or excess weight. It is usually silent for years. In a minority of people it progresses to scarring, and eventually to the stage where the liver stops keeping up.
Older adults carry a disproportionate share of that risk, and they are also the group with the least long-term data on GLP-1 medications. A study published this month in Age and Ageing tries to fill part of that gap.
What the study did
Researchers at Houston Methodist, with a co-author from Harvard Medical School, used TriNetX, a network of de-identified electronic health records from U.S. health systems. They found adults aged 65 and older with MASLD, then compared people who started a GLP-1 receptor agonist against people who did not.
To make the groups comparable, they used propensity-score matching: each GLP-1 starter was paired with a non-user who looked similar on age, other conditions, medications, and lab values. After matching, there were 43,930 people in each group. The average age was about 72, and 62% were women. Follow-up ran from 90 days out to as long as ten years.
What it found
Compared with matched non-users, starting a GLP-1 was associated with:
- Hepatic decompensation (the serious complications of advanced liver disease): about 28% lower risk, hazard ratio 0.72
- Death from any cause: about 45% lower risk, hazard ratio 0.55
- Hospitalization: a small reduction, hazard ratio 0.95
- Major cardiovascular events: no meaningful difference, hazard ratio 0.98
The researchers reran the analysis several ways, including starting follow-up at one year instead of 90 days and removing tirzepatide users, and the liver, hospitalization, and mortality associations held up.
The study also reported trade-offs. Osteoporosis and retinopathy diagnoses were more frequent in the GLP-1 group, and gastrointestinal events were more common. Pancreatitis, sarcopenia (age-related muscle loss), and fractures did not differ between groups.
That is a genuinely encouraging signal, in a population that rarely gets studied on its own. It is also exactly the kind of result that needs to be read carefully.
What the study does not establish
It shows association, not cause and effect. The authors say so themselves. This is a retrospective analysis of medical records, not a randomized trial. Matching can balance the things a database records. It cannot balance what it doesn’t record: how engaged someone is with their own care, whether they can afford a newer medication, or why a clinician judged them a good candidate for one in the first place.
The mortality number deserves the most skepticism. A 45% lower risk of death from any cause is a very large effect, larger than what randomized cardiovascular outcome trials of these drug classes have generally reported. When an observational study shows a mortality gap that big, some of it often reflects who gets prescribed a new medication. People who are frail, very sick, or nearing the end of life are less likely to be started on one. That pattern, sometimes called healthy-user bias, can make almost any newly started drug look protective. It doesn’t mean the finding is wrong. It means the size of the number is not the thing to quote.
The comparison group took nothing. Non-users were people not on a GLP-1, not people on some other active treatment. Comparing “started a medication” against “didn’t” tends to exaggerate benefit more than comparing two treatments head to head.
It doesn’t separate the drugs. The analysis grouped GLP-1 receptor agonists together, so it can’t tell you whether one medication did better than another, or what dose anyone was on.
It doesn’t tell you about any one person. A hazard ratio describes a large population on average. Whether an individual 70-year-old with fatty liver would benefit, and whether the bone and eye findings matter for them, depends on their own history. That is a clinical question, not a statistical one.
How it fits with the trial evidence
The observational data doesn’t stand alone. ESSENCE, a randomized, placebo-controlled phase 3 trial of branded semaglutide 2.4 mg in adults with MASH (the inflamed, scarring form of fatty liver disease), reported in 2025 that at 72 weeks about 63% of participants on semaglutide saw their steatohepatitis resolve without worsening fibrosis, compared with about 34% on placebo. Fibrosis improved in about 37% versus 22%.
Those are liver biopsy findings, which are a step removed from outcomes like decompensation or death. ESSENCE is still following participants to see whether the tissue changes translate into fewer of those events. Read together, a randomized trial showing liver tissue improvement and a large records study showing fewer complications point in the same direction, but neither one alone closes the question for older adults.
The part that applies to compounded medications
ESSENCE studied a branded, manufactured semaglutide product made by its manufacturer. The prescriptions in the Houston Methodist database were for manufactured products too.
Compounded semaglutide and tirzepatide are prepared by state-licensed pharmacies for individual patients. Compounded medications are not FDA-approved, and no trial has tested a compounded preparation for fatty liver disease. The evidence above belongs to the branded products it studied and does not transfer to a compounded one. Anyone telling you otherwise is stretching the research past what it covers.
Why this is worth knowing
Fatty liver disease often goes unnoticed until a routine blood test or imaging scan turns it up. If you have been told you have it, or you have elevated liver enzymes on a lab report, that belongs in the conversation with a licensed provider before any GLP-1 decision is made, along with bone health and eye history, especially after 65.
That kind of individual review is the reason a licensed provider independently reviews every intake rather than a checkout form deciding eligibility. And if you have questions about what a provider looks at, our FAQ walks through it.
This post is general information, not medical advice. It does not describe a benefit of any NoBsRx program and it is not a claim that any medication will treat or prevent liver disease. Compounded medications are not FDA-approved. Treatment is never guaranteed. Eligibility and every treatment decision are made by independent licensed providers based on an individual clinical assessment, and availability varies by state. Do not start, stop, or change any treatment based on an article. If you have fatty liver disease or abnormal liver tests, talk to a licensed clinician.