Researchers in Taiwan matched 176,981 women who started tirzepatide against the same number who entered lifestyle weight programs, and over the next two years the tirzepatide group was diagnosed with endometrial cancer about 28 percent less often.

That is the headline from a records study published October 13 in Diabetology & Metabolic Syndrome ↗ by a team at Chi Mei Medical Center. They drew both groups from TriNetX, a network that pools de-identified electronic health records from U.S. health systems, and limited it to women with overweight or obesity who had no cancer at the start. The hazard ratio was 0.717, which in plain terms means that for every 100 endometrial cancers in the lifestyle group, the tirzepatide group had roughly 72. The gap held across most of the subgroups the authors looked at.

Endometrial cancer is the one to watch here. It is the most common gynecologic cancer in wealthy countries, and obesity is its single largest modifiable driver: fat tissue converts other hormones into estrogen, and unopposed estrogen is what pushes the uterine lining toward malignancy. A drug that takes 15 to 20 percent of body weight off should, in theory, pull that risk down. Tirzepatide ↗, the Eli Lilly molecule sold as Mounjaro for diabetes and Zepbound for obesity, is the dual GIP and GLP-1 receptor agonist that produces exactly that kind of weight loss. The same obesity-to-cancer logic applies to semaglutide ↗, Novo Nordisk's GLP-1 drug sold as Wegovy and Ozempic, and to the wider incretin class, but this particular dataset isolated tirzepatide.

So the direction of the finding is biologically unsurprising. The problem is the comparison.

Why the comparator is the catch

The study did not test tirzepatide against a placebo or against another drug. It compared it against lifestyle programs, and those two populations are not interchangeable the way a trial's randomized arms are. Women who are prescribed an injectable incretin and women enrolled in diet-and-exercise counseling differ in income, in how often they see a doctor, in baseline health. Any of those can move a cancer-diagnosis count without the drug touching a single tumor.

Short follow-up makes it harder to read. Two years is a sliver of the time endometrial cancer takes to develop and get caught. What a two-year window mostly captures is when a cancer gets diagnosed, not whether it ever forms. A drug cohort under regular clinical contact can look protected simply because its members are screened differently, or because the cancers being counted are the ones already present and waiting to be found. The authors say as much themselves: the observational, short-term data do not show that tirzepatide prevents the disease.

The other direction of the same question

peptidemodel covered the opposite end of this relationship last month. A single-center study found that women's own GLP-1 levels fell as endometrial tissue turned malignant ↗, a correlation that a cross-sectional design could not turn into cause. Put the two together and the shape is familiar: GLP-1 biology and endometrial cancer travel together in the data, in both the natural hormone and the drug that mimics it, and neither study design can say which way the arrow points.

That is not a reason to ignore the signal. It is a reason to want the trial that settles it. Teams are now enrolling women with endometrial hyperplasia or early endometrial cancer ↗ to test whether tirzepatide can treat or reverse the disease directly ↗, not merely correlate with less of it. Those readouts are the ones that will matter. Until then, 0.72 is a number to file, not a number to act on.