Tirzepatide vs Semaglutide: Liver Outcomes Are Level

A study published in Obesity on 19 August 2026 put tirzepatide and injectable semaglutide head to head on hard liver endpoints in adults with overweight or obesity and type 2 diabetes. Over roughly 17 months, the two came out level: 4.05 against 4.04 major adverse liver events per 1000 person-years.

In the same cohort, tirzepatide produced the larger reduction in BMI, by about 1.1 kg/m², with a p value below 0.001.

Those two sentences sit awkwardly together, and that is the whole point of the paper.

A framing note. Both compounds are approved medicines and this is a study of patients. Peptra Labs supplies tirzepatide as reference material for laboratory research only, semaglutide is not in our catalogue at all, and nothing below is guidance about human use.

How the tirzepatide comparison was built

The authors, working from Kolkata and from the Royal Cornwall Hospital, used the TriNetX global network and framed the analysis as a target trial emulation.

That phrase is worth unpacking, because it is not a synonym for observational study. A target trial emulation starts by writing down the randomised trial you would run if you could: eligibility, the moment of assignment, the follow-up window, the outcome. Then it builds that structure out of existing records, matching people who were and were not exposed. The discipline it imposes is mostly about time. A great many database studies quietly compare people from the moment they became identifiable rather than from the moment they started the drug, which lets prevalent users who already tolerated it look healthier than they are.

Here the population was new users of tirzepatide and of injectable semaglutide. Propensity score matching balanced baseline covariates, and the primary outcome was time to first major adverse liver outcome, a composite of cirrhosis, decompensated events and hepatocellular carcinoma.

What the analysis found

Three analyses, three intervals comfortably containing 1. This is not a study that failed to find a difference because it was underpowered and vague. It is a study that looked three ways and got the same answer each time.

The MASLD subgroup is the one to read closely. Event rates there run roughly two and a half times higher than in the full cohort, which confirms the subgroup is genuinely higher risk and therefore the place a real difference between agents would show up first. It did not.

The check that makes the null believable

A null result is only as good as the study’s ability to detect a real one. The authors addressed this with internal validation: they showed that both tirzepatide and semaglutide significantly outperformed sitagliptin on the same outcome, in the same data.

That is the move that converts “we found nothing” into “we found nothing between these two, and here is proof the method finds something when something is there”. Without it, an equal-outcomes headline is indistinguishable from a data set too noisy to separate anything.

Weight moved, the liver did not

The finding that deserves the attention is the dissociation.

Tirzepatide achieved the greater BMI reduction, by about 1.1 kg/m², and the difference was statistically clear. If weight loss were the mechanism carrying liver benefit in this population, the arm that lost more should have shown fewer cirrhosis, decompensation and hepatocellular carcinoma events. Over this follow-up, it did not.

Several readings survive that observation and the study cannot separate them. One: liver benefit in this class runs through something other than weight, and the two agents supply it equally. Two: weight matters, but 1.1 kg/m² is too small a gap to register on outcomes this hard. Three: 17 months is far too short. Cirrhosis and hepatocellular carcinoma develop over years, and a median follow-up measured in months captures the early tail of that process rather than its middle.

The third reading is the most likely and the authors say as much, limiting their conclusion to short and medium term follow-up.

It also rhymes with something we covered earlier this month. In a 52-week MASH biopsy trial, a combination arm beat monotherapy on every histological number and still missed its primary endpoints. Histology moved, the endpoint did not. Here weight moved, the outcome did not. Two different studies, both showing that the intermediate measure and the thing you actually care about can come apart.

Three limits worth naming

A composite hides its components. Cirrhosis, decompensated events and hepatocellular carcinoma are combined into one outcome. They differ in frequency, in how reliably they are coded, and in how long they take to appear. An overall hazard ratio of 1.04 does not rule out divergence in one component that another offsets.

Codes are not biopsies. TriNetX records what clinicians entered. Liver disease is under-diagnosed generally, and whether a patient receives the code depends on whether anyone looked, which in turn depends on the care setting.

Emulation is not randomisation. Propensity score matching balances measured covariates. Alcohol intake, diet and the reason a prescriber chose one agent over the other are largely not in the record.

Where this leaves the class

The pattern across the last few weeks is consistent. In a meta-analysis of blood pressure related adverse events the endpoint was recorded events rather than measured physiology. In the liver work above, it is coded diagnoses. Real-world evidence for tirzepatide is accumulating quickly and almost all of it measures what was written down.

The next question the field will ask is whether a more aggressive agent changes the answer. Retatrutide, the triple GLP-1, GIP and glucagon agonist, produces larger weight reductions still, and glucagon receptor agonism has direct hepatic effects that neither compound in this study has. If the dissociation reported here holds, more weight loss will not by itself deliver better liver outcomes, and the glucagon arm of the molecule will be doing the work if anything is. Our triple-agonist research guide covers what the published record currently supports.

For laboratories sourcing these compounds, our European research buyer guide covers procurement and documentation, and everything is supplied on a research use only basis.

References

• Banerjee M, Roy A, Sharma VM, Das A. Major Adverse Liver Outcomes With Tirzepatide and Injectable Semaglutide in Adults With Overweight or Obesity and Type 2 Diabetes. Obesity (Silver Spring), 19 August 2026. doi 10.1002/oby.70277
• Ostrominski JW, Ortega-Montiel J, Wexler DJ, et al. Comparative Effectiveness of Tirzepatide Versus Dulaglutide or Semaglutide on Major Cardiovascular Events in Type 2 Diabetes. Diabetes Care, 1 May 2026
• Scola G, Chis Ster A, Bean D, et al. Implementation of the trial emulation approach in medical research: a scoping review. BMC Medical Research Methodology, 16 August 2023

The products referenced on this site are supplied for laboratory research use only. They are not medicines and are not intended for human or veterinary use. This article summarises published research for informational purposes and is not medical advice. Statements about third-party studies belong to their authors.