A STEP UP MRI subgroup showed lower liver fat during semaglutide treatment. Read the results, study limits and implications beyond scale weight.
On 2026-10-06, Pharmacy Times reported liver-fat findings from a STEP UP analysis involving semaglutide. The results were presented at the European Association for the Study of Diabetes (EASD) Annual Meeting 2026 in Milan. Researchers used MRI measurements to examine a health outcome beyond the change on a scale.
The central observation was straightforward: mean liver fat was lower at follow-up in the semaglutide-treated group assessed by imaging. Among participants who began above the reported liver-fat threshold, most were below it by the end of follow-up.
That makes this a useful story about how treatment outcomes are measured. Weight change is one outcome; a measurement inside the liver is another. The value of this report lies in reading those results together without treating them as interchangeable.
The biological issue in this analysis is fat within the liver. Researchers used MRI-PDFF, an imaging method that quantifies liver fat. In plain language, they used scans to track an organ-specific measurement during treatment.
A liver-fat percentage and a body-weight percentage answer different questions. The first describes the liver measurement. The second describes a change in total body weight. A liver-fat reading is not the percentage of body weight a person has lost.
This distinction matters because the word “fat” can make different outcomes sound like the same thing. Here, the researchers were not simply relabeling a weight-loss result. They were looking at a separate measurement within the body.
That is why the result belongs in a metabolic-health discussion rather than only a weight-loss discussion. The report invites consideration of a change within an organ alongside the familiar scale outcome. Neither measure needs to become a judgment about effort or personal worth.
Our coverage of using imaging to assess body composition takes up the same broader question: what information does a particular measurement add? The relevant point here is the distinction between an overall weight change and a specific organ-level outcome.
The Pharmacy Times headline says “Semaglutide normalized liver fat in most adults with obesity.” Read quickly, that can sound like a result established for the wider population. The underlying report supports a narrower interpretation.
Here are the boundaries that matter:
The findings also describe an outcome rather than establish the biological pathway behind it. The report is a reason to investigate the liver-fat signal further, not to assume that a particular mechanism or broad liver-disease benefit has been proved.
There is no need to choose between dismissing the observation and treating it as settled clinical evidence. A meaningful change in a measured outcome can justify further research while remaining an early finding. In this case, the wording should stay attached to the population and measurement actually studied.
In the semaglutide-treated MRI subgroup, mean liver fat fell from 8.8% at baseline to 3.1% at week 72, according to Pharmacy Times.
Among the 26 participants whose liver fat began above 5%, 23 (88.5%) were below 5% at week 72. The report describes a value below 5% as normal for this measurement.
Those findings describe two views of the data. The mean summarizes the subgroup’s liver-fat readings. The threshold result counts how many participants with an initially higher reading moved below the stated cutoff.
The threshold calculation begins with the participants who were above the cutoff at baseline. It should not be read as though every person in the MRI subgroup started with elevated liver fat. Likewise, the subgroup’s average starting value does not describe the starting value of every participant.
STEP UP randomized 1,407 adults with obesity and without diabetes to semaglutide 7.2 mg, semaglutide 2.4 mg or placebo. Treatment was accompanied by a lifestyle intervention over 72 weeks.
In its separate report on STEP UP’s weight findings, Pharmacy Times described the mean body-weight changes under the reported treatment-policy estimand:
These are group averages from the trial, not targets for a reader. They provide context for the liver result while remaining a different outcome.
The practical lesson is to broaden the health conversation, not to pursue more extreme weight loss. A qualified clinician can discuss individual liver-risk assessment and how it fits with treatment decisions. The report is not a reason to alter medication, diet or exercise independently.
A useful clinical discussion would distinguish the reason for a liver assessment from the expected benefits of a weight-management treatment. It would also distinguish a test result from a diagnosis. Those are personal care questions, rather than conclusions a reader can draw from a news headline.
Our coverage of medication alongside lifestyle support provides related context for that broader care discussion. WeightRestart’s focus is research-led guidance that supports sustainable habits without shame. The aim is better-informed decisions, not turning a subgroup result into a personal performance standard.
What to watch next is whether larger studies make MRI-measured liver fat a planned outcome and include an appropriate comparison group. Reporting each semaglutide dose separately would help clarify the pooled result. Researchers also need to assess whether liver-fat changes persist and are accompanied by improvements in inflammation or fibrosis and by better longer-term clinical outcomes. Before these findings change standard practice, the evidence needs to connect the scan result with benefits that matter in ongoing care.
WeightRestart shares research-led guidance on weight loss, metabolism, nutrition, strength, appetite, sleep and recovery. Our goal is to make complex health information clear, practical and useful for people building progress they can maintain.
Learn how to build a weight-management approach around better information, realistic expectations and habits you can keep using.
read the blog
