Skip to content
This GLP
← Research
Evidence

Liver stiffness: the sleeve took more weight and no more stiffness

Sleeve gastrectomy took off more weight than semaglutide and improved liver stiffness no more than the drug did. Ninety-two people, nobody randomized.

Ruth Alvarez7 min read
Over 72 weeks — liver fat signal, then stiffnesslifestyle17.9 dB/m0.8 kPasemaglutide46.0 dB/m2.8 kPasurgery83.7 dB/m4.9 kPaStiffness: no significant difference between drug and surgery, p = 0.428.

The interesting line in this study is the one where the bigger intervention stopped buying a bigger result. Surgery took off more weight than the drug did and the liver stiffness improved about the same, which is either a measurement limit or a hint that the liver effect is not purely a weight effect — the question the trial evidence has been circling.

What was compared

Ninety-two adults at one center, watched prospectively for 72 weeks. [1] Thirty on a lifestyle program, thirty on that plus semaglutide at 1.0 ± 0.4 mg a week, and thirty-two who had a laparoscopic sleeve gastrectomy. Everyone had hepatic steatosis at baseline.

Nobody was randomized. People who choose an operation differ from people who choose a weekly injection, in ways no covariate list holds.

Fat first

The controlled attenuation parameter, which estimates how much fat is in the liver, fell by 17.9 dB/m on lifestyle, 46.0 on semaglutide and 83.7 after surgery, overall p = 0.002. That ordering is what anybody would expect, and it tracks the weight loss — fat comes out of a liver when weight comes off a body, which is why a marker close to the mechanism moves most readily and proves least.

Then stiffness

Liver stiffness fell 0.8, 2.8 and 4.9 kPa across the same three groups, overall p < 0.001. In relative terms the semaglutide and surgery groups did not differ significantly, p = 0.428.

The authors are careful about this and so is the page: they call it hypothesis-generating and say it warrants investigation into weight-independent mechanisms. It is 92 people, it is not randomized, and a non-significant difference between two small groups is a weak instrument for claiming they are equivalent.

The result nobody explained

FIB-4, a score built from age, platelet count and two liver enzymes, rose modestly after surgery — up 0.29, p < 0.001 — while staying stable in both other groups.

A fibrosis score moving in the wrong direction after an operation that improved every other marker is the sort of thing worth an explanation, and the published abstract does not offer one. This desk is not going to invent a mechanism for it either. It is on the page because a reader deserves to know the results did not all point one way.

The dose is not the dose you are buying

One milligram a week is a type 2 diabetes dose. The sellers on this roster price 2.4 mg for weight, and most price something different again as the dose climbs.

So this study does not tell you what the dose you would actually pay for does to a liver, and it does not tell you that a smaller dose is sufficient either. It tells you what 1.0 mg did to 30 people over 72 weeks, which is the kind of specificity observational work outside a trial usually loses in the retelling.

Frequently asked

Did surgery beat semaglutide for the liver?
On liver fat, yes — 83.7 against 46.0 dB/m. On stiffness the two did not differ significantly in relative reduction, p = 0.428, despite surgery producing more weight loss.
What dose of semaglutide was used?
1.0 ± 0.4 mg a week, a type 2 diabetes dose rather than the 2.4 mg sold for weight management.
Were the liver findings from biopsies?
No. All outcomes are non-invasive markers — FibroScan measurements and scores calculated from blood tests — which correlate with biopsy findings without being the same thing.
Did anything get worse?
FIB-4 rose by 0.29 after surgery while staying stable in the other groups. The published abstract does not explain why.

Sources

  1. [1] Lünswilken P, et al. (2026). Effects of Semaglutide versus Bariatric Surgery on Noninvasive Markers of Hepatic Steatosis and Fibrosis in Obesity with MASLD The Journal of Clinical Endocrinology and Metabolism. PMID 42657770

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence