In 180 obese women biopsied for abnormal bleeding at an Ankara hospital, the ones who turned out to have endometrial cancer had the least GLP-1 in their blood. Not a drug. The hormone their own bodies made.
GLP-1 is the natural gut hormone that the weight-loss injections semaglutide ↗ and tirzepatide ↗ were built to imitate. Endometrial cancer, the most common gynecologic cancer in wealthy countries, is tightly bound to obesity, because fat tissue raises estrogen and estrogen pushes the uterine lining to overgrow. A study that measures a woman's own GLP-1 against her biopsy result sits right on the seam between the two things this field argues about most: metabolism and cancer.
The gradient was clean. Published in Gynecologic Oncology ↗, the study sorted 180 obese, non-diabetic women with abnormal uterine bleeding into three groups of 60 by what their biopsy showed: a benign polyp, hyperplasia (a pre-cancerous overgrowth of the lining), or cancer. Fasting GLP-1 averaged 14.6 pg/mL in the polyp group, 11.2 in the hyperplasia group, and 7.4 in the cancer group. The level roughly halved as the tissue went from harmless to malignant.
The authors then report that the hormone told cancer from non-cancer almost perfectly. The area under the ROC curve, a single score where 1.0 is a flawless test and 0.5 is a coin flip, came out at 0.94. A cutoff of 9.12 pg/mL caught 86 percent of the cancers and correctly cleared 92 percent of the women who did not have it. Those are numbers that should make a careful reader lean back rather than forward. Near-perfect separation out of one center with 60 confirmed cancers is exactly the pattern that tends to shrink the moment someone else runs the assay on a different population.
The deeper problem is direction. This was a cross-sectional study: each woman's GLP-1 was measured once, at the time of biopsy. A single blood draw cannot say whether low GLP-1 helped the cancer along or the cancer, and the metabolic state that travels with it, drove GLP-1 down. The adjusted odds (0.744 for every extra 1 pg/mL of GLP-1, meaning more hormone tracked with lower odds of cancer) held up after the authors corrected for body weight, HbA1c, blood pressure, inflammation, and the tumor marker CA-125, which is more scrutiny than most biomarker papers apply. It still says nothing about which came first.
Then there is the trouble with measuring GLP-1 at all. The active hormone lasts about two minutes in circulation before the enzyme DPP-4 clips it apart, and the ELISA kits used to quantify it differ in whether they read the active form, the spent fragment, or both. A fasting GLP-1 value is a snapshot of something that barely holds still. The authors themselves call for prospective multicenter validation before anyone treats 9.12 pg/mL as a real threshold, and a competing-interest statement that mentions a pending patent is a reminder that a biomarker with a cutoff is also a product waiting to be sold.
There is a genuine wrinkle here for anyone watching the weight-loss drugs. Agonists of the GLP-1 receptor ↗ are being studied to lower endometrial cancer risk, mostly by driving the weight loss that cools the estrogen excess an obese endometrium sits in. This study points at the endogenous hormone from the opposite end: the women who already had less of their own GLP-1 were the ones with cancer. Those are not the same claim. Flooding the receptor with a long-acting drug is a different act from carrying a healthy baseline level of the natural peptide, and a correlation in 180 women is not a reason to check anyone's GLP-1 before a biopsy. It is a reason to run the larger study.