Why endometriosis comes back after surgery — the biology excision doesn't touch

What you ran into is not a failure of the operation. It is a limit of what an operation can do. The constraint is mechanistic, not technical, and understanding it changes what you ask about next.

Laparoscopic excision is the reference intervention for symptomatic and deep infiltrating endometriosis, and a good excision does exactly what it promises: it removes the macroscopic lesion, gives histologic confirmation, and produces durable pain relief in a meaningful proportion of women. But the lesion you can see and cut out is not the whole disease. Excision does not remove the peritoneal inflammatory milieu the lesion was sitting in, it cannot reliably find and remove residual microscopic implants scattered across the pelvic lining, and it does nothing to the molecular programs — the transcriptional instructions, in effect — that re-establish disease wherever the conditions still favor it. The recurrence figures reflect this directly. In systematic-review summaries, cumulative recurrence approaches 40 to 50 percent within five years of complete excision, with roughly 20 percent recurrence already by two years, attenuated but not abolished by post-operative suppression. Each repeat operation also carries a cumulative cost to ovarian reserve and pelvic anatomy, which is its own argument for understanding why the disease returns rather than simply cutting again.

Here is the part that the post-operative reassurance tends to skip. The IUD you carry is doing real work, but it is not doing the work you may think it is. The levonorgestrel intrauterine system is, mechanistically, a local antiproliferative and antihemorrhagic agent. It decidualizes and atrophies the eutopic endometrium — the lining inside your uterus — and it reduces menstrual flux. That is why the bleeding becomes manageable and often disappears. But that action is happening inside the uterus, on the orthotopic tissue. It does not abrogate the aromatase activity that ectopic lesions run on the peritoneum, it does not resolve the established fibrosis and adhesions that have already formed, and it does not interrupt the macrophage-driven inflammatory signaling that sustains a lesion. So a woman with stage III-IV disease who underwent excision without hysterectomy, and who carries an IUD, frequently retains biologically active lesions despite genuine symptomatic control. The bleeding and the disease were never the same thing, and quieting one was never going to silence the other.

To see why anatomy-only treatment recurs, it helps to look at what an endometriotic lesion actually is once it has established itself, because it is far more autonomous than the phrase "misplaced tissue" suggests. The lesion behaves like a small, self-sustaining micro-tissue with its own estrogen supply. Whereas the lining inside the uterus expresses negligible aromatase, endometriotic stroma ectopically transcribes CYP19A1 — the gene for aromatase, the terminal enzyme that converts androgens into estrogens — and it simultaneously runs low on the enzyme that would normally inactivate that estrogen. So the lesion both manufactures estradiol locally and holds onto it. That local estradiol drives COX-2 and the production of prostaglandin E2, an inflammatory signal, and prostaglandin E2 in turn is among the most potent inducers of aromatase in this tissue, closing a feed-forward loop in which estrogen begets prostaglandin and prostaglandin begets more estrogen. Because that loop is intracrine — generated inside the lesion — and largely independent of the ovary, lowering the estrogen your ovaries produce leaves it substantially intact. This is the same reason conventional endocrine suppression so often produces a partial response.

There is a quieter consequence of that local estrogen worth naming, because it bears directly on why disease can stay active under an IUD. The intrauterine system works by acting on the lining inside your uterus — the orthotopic tissue, which is rich in progesterone receptors and responds to a local progestin by thinning and quieting. The ectopic lesion is a different tissue with a different receptor profile, often relatively resistant to progesterone, and it is generating its own estrogenic drive rather than waiting on signals from the uterine cavity. So the device's mechanism and the lesion's mechanism barely overlap. The bleeding stops because the uterine lining stops building up; the lesion persists because nothing in that local action reaches the aromatase loop running on the peritoneum. Two tissues, two biologies, one of which the IUD addresses and one of which it largely does not.

The estrogen loop is only one circuit. The lesion also recruits the body's own clean-up cells against it. Peritoneal macrophages are drawn in and pushed into an alternatively activated, permissive state — the kind that tolerates and nourishes tissue rather than clearing it. These macrophages secrete TGF-beta1, which drives the conversion of fibroblasts into myofibroblasts, the deposition of collagen, and the dense adhesions and fibrosis responsible for much of the pain and the infertility. They also help supply the new blood vessels, by way of VEGF, that the implant needs to survive, along with the growth factors that keep its cells proliferating. Layered over all of it is NF-kB, a master inflammation switch that integrates these signals and keeps the lesion's harmful behaviors transcriptionally switched on. None of these circuits lives in the lump a surgeon removes. They live in the environment around it, in the immune cells that traffic through the pelvis, and in the cells' own genetic programming — which is precisely why excision can be anatomically complete and biologically partial at the same time.

Underneath even those circuits sits a metabolic amplifier, and it matters because it is one of the few inputs that lives outside the pelvis entirely. Endometriotic cells favor a glycolytic, survival-oriented metabolism, and systemic hyperinsulinemia together with elevated IGF-1 reinforces it through the PI3K-AKT-mTOR pathway — lowering the threshold at which lesional cells resist apoptosis and amplifying their responsiveness to growth factors. Body fat compounds this from a second direction, because adipose tissue is itself a site of estrogen production outside the ovary, so adiposity quietly raises systemic estrogenic tone at the same time that insulin lowers the binding globulin that would otherwise hold that estrogen inactive. None of this is removed by an operation and none of it is touched by a local intrauterine device. It is the kind of input that keeps the soil favorable for recurrence regardless of how clean the excision was, and it is part of why recurrence is a whole-body phenomenon dressed up as a local one.

The organizing idea that emerges from this is worth holding onto, because it reframes the whole experience of recurrence. The endometriotic lesion is a hormonally autonomous, chronically inflamed, fibrogenic micro-tissue sustained not by a single pathway but by a redundant network of reinforcing signaling circuits. A redundant network is hard to collapse by removing one component, and impossible to collapse by removing the component — the visible lesion — that the network can simply rebuild. That redundancy is the mechanistic reason an operation that succeeds anatomically can still be followed by recurrence, and the reason a device that succeeds at controlling bleeding can leave the disease biologically active beneath the symptom relief.

It is worth being concrete about what biologically active but symptomatically controlled actually looks like, because the phrase can sound like hairsplitting until you have lived it. It means the bleeding is gone but the deep pre-menstrual pull is not. It means imaging may be quiet while pain with sex returns in the same places. It means the disease can be doing its slow fibrogenic work — laying down adhesions, sustaining the inflammatory milieu, holding the aromatase loop open — in a body whose most visible symptom, menstrual flooding, has been switched off by the IUD. Symptom control is real and worth having; it is simply not the same variable as disease activity, and conflating the two is what makes recurrence feel like a betrayal rather than a predictable course. The two can run on separate tracks, and for many women with retained advanced disease they do.

Repeated surgery is the other reason this distinction earns its keep. Because excision cannot reach the milieu or the molecular programs, recurrence often leads back toward the operating room, and each operation extracts a cumulative cost — to ovarian reserve, to pelvic anatomy, to the tissue planes a future surgeon will have to work through. A model in which the disease is understood to regrow from biology rather than from surgical incompleteness reframes that loop. It suggests that the relevant question after recurrence is not always whether to cut again, but whether anything is being done about the conditions that regrew the lesion in the first place — the local estrogen, the fibrosis, the macrophage signaling, the metabolic fuel — alongside whatever surgical decision is made. The biology does not argue against surgery; it argues against expecting surgery alone to be durable when the network that sustains the disease is left running.

This is also where the boundaries matter, because understanding the biology is not the same as having a fix for it. The interventions that engage these circuits directly are a separate and carefully graded subject, and the honest framing is that they are adjuncts and research directions discussed alongside specialist care, not replacements for it. The principal real-world risk in this whole area is not that someone tries a low-risk adjunct; it is that someone with active stage III-IV disease defers effective specialist management because their symptoms are partially controlled. Surgery and supervised endocrine management remain the structural foundation. Nothing about the molecular picture changes that, and the picture is most useful precisely when it sits inside specialist care rather than around it.

What does change is the conversation. If you understand that recurrence is driven by the inflammatory milieu, the microscopic residue, and the lesion's own self-sustaining biology rather than by a surgical mistake, then "the surgery was successful and the IUD is working" stops being a contradiction and starts being an incomplete description. It tells you that the anatomy was addressed and the bleeding was addressed, and it tells you, by omission, what was not — the aromatase loop, the fibrosis, the macrophage signaling, the metabolic fuel. That is the gap recurrence lives in. Knowing where it lives is what lets you walk back into the room and ask the more precise question: not why the surgery failed, because it didn't, but what in the underlying biology is still being left to run, and what, alongside your specialist's care, can be brought to bear on it.