Premature ovarian insufficiency is rising in Ontario, with endometriosis the top risk

Premature ovarian insufficiency (POI) — the loss of normal ovarian function before age 40 — has a cause that remains unknown in roughly 70% of non-iatrogenic cases (cases not caused by medical treatment like chemotherapy or surgery), and existing incidence estimates have been limited by small samples and recall bias. A population-based cohort study published this month in Human Reproduction, covering nearly the entire province of Ontario over 24 years, provides the most comprehensive incidence data yet [s1].

The design

This population-based cohort study included all women aged 39 or younger residing in Ontario, Canada, between 1995 and 2019 — 3,635,702 eligible women in total [s1]. POI was identified using administrative health data, defined as a diagnosis before age 40 via either surgical menopause or at least one physician consultation coded for menopause [s1]. Follow-up continued until POI diagnosis, age 40, death, loss of provincial health insurance, or the end of the study period [s1]. POI cases were classified into four categories: non-iatrogenic, surgical, post-chemotherapy and/or radiotherapy, or combined surgical with chemotherapy/radiotherapy [s1]. Age-adjusted modified Poisson regression estimated incidence rate ratios for associated clinical factors [s1].

What it found

Among the 3.6 million eligible women, 168,173 developed POI from any cause during the study period, an overall incidence rate of 389.4 per 100,000 person-years [s1]. Median follow-up was 11 years, and the median age at POI diagnosis was 31 [s1]. Non-iatrogenic POI was by far the most common category, at 351.4 per 100,000 person-years, compared with 30.2 for surgical POI, 5.5 for post-chemotherapy/radiotherapy POI, and 2.3 for the combined surgical-plus-treatment category [s1].

The incidence trend over time showed a distinct pattern: overall POI incidence declined between 2002 and 2008, then rose steadily and sustained through 2019, an increase largely attributable to non-iatrogenic POI specifically [s1]. Meanwhile, surgically induced POI incidence fell substantially over the same period, from 60 per 100,000 person-years in 1995 to 20 in 2019, while the combined surgical-with-chemotherapy/radiotherapy category increased steadily [s1]. Among clinical risk factors, endometriosis carried by far the strongest association: incidence rate ratio 3.45 (95% CI 3.39–3.52) [s1]. Cancer treated with chemotherapy or radiotherapy carried an incidence rate ratio of 2.01 (95% CI 1.89–2.14), and drug or tobacco use 1.59 (95% CI 1.51–1.67), with chromosomal disorders also showing an elevated, though not fully specified in the available summary, association [s1].

Why the declining-surgical, rising-non-iatrogenic split matters

The opposite-direction trends within this data — falling surgically induced POI alongside rising non-iatrogenic POI — likely reflect two separate underlying shifts happening simultaneously: evolving surgical practice, with fewer procedures now resulting in both ovaries being removed in younger women compared with decades past, alongside whatever is driving the rise in POI without an identifiable surgical or treatment-related cause. The study doesn't identify what's behind that non-iatrogenic increase specifically — it documents the trend without explaining its cause, which the 70%-unknown-etiology figure for non-iatrogenic POI cases underscores as a genuine, unresolved gap in the field.

Why the endometriosis association is the study's strongest, most actionable finding

An incidence rate ratio of 3.45 for endometriosis dwarfs the other risk factors identified in this analysis, including cancer treatment itself (2.01) — making endometriosis history the single strongest clinical predictor of POI risk found in this cohort. That's a clinically significant finding for a condition, endometriosis, that already affects a substantial share of reproductive-age women and is itself frequently underdiagnosed or diagnosed with significant delay — suggesting POI risk should be part of the broader conversation clinicians have with endometriosis patients about their reproductive health trajectory.

What this doesn't establish

This is an administrative health-data cohort study — POI diagnosis relied on physician coding practices and administrative records rather than a standardized research diagnostic protocol applied uniformly across the entire 24-year period, meaning some variation in diagnostic thresholds or coding practices over that long timespan is possible and could partly explain trends. The study establishes association between endometriosis and POI risk, not the underlying biological mechanism connecting the two conditions. The Ontario, Canada population and healthcare system may not fully generalize to other countries with different demographics or diagnostic practices.

What to watch

Whether the rising non-iatrogenic POI trend continues, and whether research into its underlying causes — the 70% of cases still unexplained — narrows that gap. This article is not medical advice.

Sources

1. Incidence and factors associated with premature ovarian insufficiency: a population-based cohort study — Human Reproduction, 15 July 2026