Sleeve patients lost three times more weight than semaglutide patients, and liver stiffness fell about the same

A German research group at the University of Lübeck followed 92 adults with obesity and fatty liver disease for 72 weeks and compared three routes: intensive lifestyle treatment alone, lifestyle treatment plus semaglutide, and lifestyle treatment plus a sleeve gastrectomy. The results were published online in the Journal of Clinical Endocrinology and Metabolism on August 27, 2026.

The condition studied is metabolic dysfunction-associated steatotic liver disease, usually shortened to MASLD. It means fat has built up in the liver alongside metabolic problems such as obesity, high blood sugar or high blood pressure. It used to be called non-alcoholic fatty liver disease. In some people the fat drives inflammation and then scarring, which is the part that eventually threatens the liver.

Nobody in this study had a liver biopsy. The researchers used a FibroScan, an ultrasound-based device that gives two readings. The controlled attenuation parameter, or CAP, estimates how much fat is in the liver and is reported in decibels per meter. Liver stiffness measurement, or LSM, estimates how stiff the liver is and is reported in kilopascals. Stiffness is used as a stand-in for scarring. They also calculated two scores, FIB-4 and FAST, which combine blood test results with other measures.

The liver fat reading fell furthest after surgery, by 83.7 decibels per meter

Over 72 weeks, the CAP reading dropped by 17.9 in the lifestyle group, 46.0 in the semaglutide group and 83.7 in the surgery group, all in decibels per meter. The overall comparison was statistically significant at p equals 0.002.

Weight loss tracked the same order, and the gaps were large. The lifestyle group lost 4.2% of body weight, the semaglutide group 10.7%, and the surgery group 31.6%. Semaglutide was given at an average of 1.0 mg per week, which is below the 2.4 mg weekly dose licensed for weight management.

Liver stiffness dropped in both treatment groups, and the adjusted gap between them was not significant

Stiffness fell by 0.8 kilopascals with lifestyle alone, 2.8 with semaglutide and 4.9 with surgery. The overall comparison was significant at p below 0.001. But when the researchers compared the relative reduction in semaglutide against surgery, the difference was not statistically significant, at p equals 0.428.

That is the finding the authors flag, and they are careful about it. Surgery produced three times the weight loss and still did not produce a clearly larger drop in liver stiffness. They call this hypothesis-generating and say it warrants further investigation into weight-independent mechanisms.

One analysis supports that idea. In the surgery group, the improvement in stiffness tracked how much weight a person lost, at p equals 0.027. In the semaglutide group it did not, at p equals 0.508. That is consistent with the drug doing something to the liver beyond simply making people smaller, though a single observational cohort cannot establish it.

One blood-based fibrosis score moved upward after surgery

The FAST score fell significantly in all three groups. FIB-4, however, rose modestly after surgery, by 0.29 with p below 0.001, while staying stable in the semaglutide and lifestyle groups.

The authors did not conclude that surgery worsened liver scarring, and their overall conclusion is that both semaglutide and sleeve gastrectomy improved these markers while lifestyle treatment alone did comparatively little. The FIB-4 result is reported as an observation, not as evidence of harm. It is the kind of discordance between markers that is a good reason not to read any single number as the answer.

Ninety-two people at one clinic, and nobody was randomly assigned

This was a single-center prospective observational cohort, not a randomized trial. People ended up in the lifestyle, drug or surgery group through the ordinary clinical process, which means the groups differed at the start in ways that a randomized trial would have evened out. The authors note that baseline liver stiffness already differed between the groups.

The groups were small, at 30, 30 and 32 people. Seventy-two weeks is a short window for a disease measured in decades, and liver scarring in particular changes slowly. FibroScan readings are estimates, not tissue samples, and they can be affected by factors including body size, which is itself changing dramatically in a weight-loss study.

Nothing here supports choosing one treatment over another for liver disease. It supports asking the question in a randomized trial.