GLP-1s and endometriosis — the metabolic lever and the early signal

That observation — repeated now by enough people to be more than anecdote, though still far short of proof — is what makes the incretin drugs interesting in a conversation about endometriosis. These are the GLP-1 agents: semaglutide, which activates the GLP-1 receptor; tirzepatide, a dual GLP-1/GIP receptor agonist; and retatrutide, a tri-agonist that adds glucagon-receptor activity and is not yet approved for any use. They were built to treat metabolic disease, and the reason they keep surfacing in discussions of endometriosis is that endometriosis turns out to have a metabolic engine, one that these drugs are unusually well placed to throttle. Before going further, the necessary frame: the use of incretin agonists in endometriosis is investigational. It is not an approved indication. What follows is mechanism and an early human signal, not a settled treatment.

To see why the mechanism is coherent, it helps to understand what an endometriotic lesion actually runs on. It is not simply misplaced tissue waiting to be cut out. It behaves like a small, self-sustaining micro-organ that makes its own estrogen, generates its own inflammation, builds its own blood supply, and is fed by the body's metabolic state. One of the most underappreciated fuel lines into that micro-organ is systemic insulin. When circulating insulin runs high, as it does in insulin resistance, it amplifies a growth-signaling cascade — the IGF-1 receptor feeding into the PI3K–AKT–mTOR pathway — that lowers the threshold at which lesional cells would otherwise die and heightens their responsiveness to growth factors. In plain terms, high insulin and high IGF-1 act like fertilizer on tissue that is already inclined to proliferate.

The first and most central thing the incretin drugs do is pull that fertilizer back. By enhancing glucose-dependent insulin secretion, slowing gastric emptying, and reducing appetite through the brain, they lower circulating insulin over time and reduce body fat. As insulin and IGF-1 drive recede, less signal reaches the IGF-1R/PI3K–AKT–mTOR machinery that the lesion has been exploiting. This is the metabolic route, and it is the core action. It does not target the lesion's anatomy at all; it changes the systemic conditions the lesion depends on.

The second route runs through estrogen, and it is the one most directly relevant to a disease defined by estrogen sensitivity. Body fat is not metabolically inert. Adipose tissue is the dominant site of extragonadal aromatization — the conversion of androgens into estrogens outside the ovaries — which means that the more fat mass a person carries, the more peripheral estradiol their body manufactures independent of the ovary. Reducing fat mass therefore lowers that peripheral estrogen production. Conceptually, this is a systemic counterpart to what aromatase inhibitors do pharmacologically inside the lesion: both withdraw estrogen from a tissue that is fundamentally estrogen-driven, one by blocking the enzyme directly and the other by shrinking the reservoir of tissue that houses it. It is worth being precise here — this is the proposed mechanistic logic, the kind of thing researchers may track and study, not a demonstrated clinical endpoint in endometriosis.

The third route is the one that does not depend on weight loss at all, and it is mechanistically the most surprising. GLP-1 receptors sit on immune and endothelial cells, and when they are activated, intracellular cyclic AMP rises. That rise has consequences for inflammation: it attenuates assembly of the NLRP3 inflammasome and the maturation of IL-1β, two engines of the inflammatory signaling that characterizes active endometriosis. It also appears to nudge macrophages away from the permissive, lesion-tolerating M2-like state and toward a posture that lowers overall inflammatory tone. This immunometabolic action is what makes the incretin story more than a weight-loss story — it suggests the drugs may be touching the inflammatory biology of the disease through a channel that operates whether or not the scale moves much. Again, in endometriosis specifically this is studied and proposed, not established.

The reason any of this is more than elegant theory is a neighboring disease where the evidence is genuinely controlled: polycystic ovary syndrome. PCOS is metabolically and inflammatorily analogous to endometriosis — it runs on the same insulin/IGF-1 amplifier and the same adipose-estrogen reservoir — and in PCOS, incretin therapy has well-documented controlled-trial benefit. GLP-1 and dual GIP/GLP-1 agonists reduce weight, improve insulin sensitivity, lower circulating androgens, and help restore ovulation. That body of evidence does something important: it proves that incretin therapy can favorably reset the exact insulin/IGF-1 and adipose-estrogen axis that the endometriosis model implicates. It is the strongest read-across available. But it is read-across, not direct endometriosis evidence, and the distinction matters — a mechanism proven in one condition is a reason to study it in a related one, not a license to assume the result transfers.

Which brings the conversation to what is actually known in endometriosis itself, and here the honesty has to be deliberate. The most cited human observation is a 2025 patient-reported survey of incretin users with confirmed or suspected endometriosis. In that survey, roughly two-thirds of users reported improvement in at least one symptom, about one-third reported complete resolution of at least one symptom, and respondents described meaningful reductions in pelvic pain, bloating, and low-back pain. Those figures are striking, and they are worth taking seriously as a signal. They are not controlled evidence, and the gap between the two is not a technicality. A self-reported survey selects for people motivated to respond — often those who improved. It relies on recall, which distorts. It cannot separate a true drug effect from a placebo effect, which in pain conditions is substantial. Many respondents were doing other things at the same time — changing diet, adding supplements, adjusting hormonal therapy — so benefit cannot be cleanly attributed to the incretin. And the products themselves, particularly when sourced outside standard pharmacy channels, vary in quality. What the survey establishes is that a meaningful number of patients perceive benefit, that the perceived benefit is frequent enough to drive rapid adoption, and that the pattern of reported improvement converges with the predicted mechanism. The correct inference from that is to prioritize formal trials, not to substitute the survey for them.

Two further threads sit alongside the survey. In preclinical, hormone-sensitive tissue models, semaglutide shows synergy with progestins — a finding that, if it holds, would matter because progestins are already a mainstay of endometriosis suppression and a metabolic agent that potentiates them is mechanistically attractive. And a controlled trial pairing an incretin agonist with a levonorgestrel intrauterine device is in progress, which is precisely the kind of structured evaluation the survey signal calls for: it would begin to test, under controlled conditions, whether the metabolic lever adds anything to the local hormonal control an IUD already provides. Until trials like it report, the endometriosis indication remains exactly where it sits today — investigational, promising on mechanism and early signal, unproven on outcome.

There is a real-world risk that has to be named plainly, because the appeal of a drug that might quiet symptoms is exactly what makes it dangerous in the wrong context. Endometriosis at its more advanced stages — deeply infiltrating disease, large endometriomas, bowel or ureteral involvement — is a structural problem that effective, established management exists to address. The principal hazard of an investigational metabolic approach is not the drug's known side-effect profile; it is the possibility that a person feeling somewhat better defers the surgical and specialist care that active stage III–IV disease actually requires, allowing the structural disease to progress while the symptom is masked. A metabolic lever that withdraws fuel from a lesion is not the same as a surgeon removing it, and feeling better is not the same as the disease being controlled.

What the incretin story changes, then, is less a treatment than a frame. For years endometriosis has been understood almost entirely as tissue to be excised and estrogen to be suppressed at the ovary. The incretin angle insists on a third axis — that the disease is fed by systemic metabolism, by insulin and by fat-derived estrogen and by inflammatory tone, and that those inputs are, in principle, modifiable. That reframing is the actual contribution here. Whether semaglutide or tirzepatide or an eventual approved tri-agonist earns a defined place in endometriosis care will be decided by the trials now beginning, not by the survey or the mechanism. But the recognition that a hormonally autonomous lesion is still, in part, a metabolically dependent one is the kind of shift that tends to outlast any single drug — and it suggests the question worth bringing to a prescribing provider is not whether a weight-loss drug can be borrowed for pain, but whether a person's own metabolic biology is quietly underwriting their disease.