Endometriosis Isn’t One Disease. Researchers Now Think It Could Be Four.
Researchers analyzed health records from more than 22,000 women with endometriosis and found something that upends the standard picture of the disease: the patients did not cluster around a single set of symptoms. They clustered around four. The analysis, drawn from the NIH’s All of Us database, identified groups defined not just by pelvic pain, but by patterns that included gastrointestinal problems, migraines, anxiety, and depression.
Endometriosis affects roughly one in ten women of reproductive age worldwide. It still takes four to eleven years to diagnose on average, a delay researchers have attributed to symptoms being dismissed as ordinary period pain. But if the disease actually represents four distinct biological presentations, the dismissal problem runs deeper than anyone thought.
Women who also had adenomyosis, a related condition where tissue grows into the uterine wall, showed markedly more severe disease across the board. That overlap matters for what comes next in treatment, and for who gets taken seriously first.
Keep reading. It’s free.
Join 6,000+ readers getting women’s health research, translated clearly.
The practical implication of subtyping is significant. Right now, a woman with endometriosis and debilitating migraines may never have those two conditions connected in a clinical chart. A woman whose primary symptom is gastrointestinal pain may spend years in gastroenterology before anyone considers a reproductive cause. Subtyping creates a framework for connecting those dots at the point of first complaint, rather than after a decade of misrouted care.
The study does not claim these subtypes are biologically fixed or that they predict treatment response. What it offers is a new vocabulary for a condition that has been treated as one thing for far too long. For patients, it means the symptom picture you are describing is not unusual. It may simply match a subtype that medicine has not had the language to name until now.
The research was published in July 2026 via EurekAlert, based on analysis of the NIH All of Us Research Program database.