The same finding means different things at different ages

Adolescence

Irregular cycles in the first year or two after menarche represent normal maturation of the hypothalamic–pituitary–ovarian axis, not a disorder. Cycles in the years after menarche are commonly longer than adult cycles, and the widely used adolescent thresholds are: any single cycle over 90 days beyond the first year; cycles under 21 or over 45 days from one to three years post-menarche; cycles under 21 or over 35 days beyond three years. Primary amenorrhea — no period by age 15, or more than three years after breast development began — requires evaluation.

The abnormality that must not be dismissed at this age is heavy bleeding from the very first period, which is the classic presentation of an undiagnosed bleeding disorder.

The puberty sequence, the thelarche-to-menarche interval, what the first period is actually like, and the full list of findings that need checking are covered on the menarche page. To apply all of it to one particular girl, use the First Periods Checker — it uses the adolescent ranges rather than the adult ones.

Reproductive years

Here the priorities are pregnancy (test first, always, including in women who believe they cannot be pregnant), then structural causes and ovulatory dysfunction. New intermenstrual or postcoital bleeding warrants cervical assessment regardless of screening history.

Perimenopause

Cycle change is expected — that is what the transition is. What is not expected, and what still requires evaluation rather than attribution to "just perimenopause": bleeding heavy enough to cause anaemia or flooding, bleeding between periods, and any cycle that has become both irregular and much heavier. Anovulatory cycles in this window mean prolonged unopposed estrogen exposure, which is precisely the setting in which endometrial hyperplasia develops.

After menopause

Any bleeding after menopause is abnormal. One episode. Spotting counts. In a meta-analysis of 129 studies, postmenopausal bleeding was present in about 91% of women with endometrial cancer, while about 9% of women presenting with postmenopausal bleeding were found to have endometrial cancer — approximately 7% among hormone therapy users and 12% when hormone therapy users were excluded.

Read those two numbers together. Nine out of ten women with postmenopausal bleeding will not have cancer, so this is not a reason to panic. But nine out of ten women who do have endometrial cancer are found because of exactly this symptom — which is why it is always investigated, and why it is investigated promptly.The complementary framing matters for counselling: high sensitivity of PMB as a presenting symptom for endometrial carcinoma, with a pre-test probability of roughly 9% that supports universal evaluation while permitting realistic risk communication. Risk stratification by hormone therapy exposure and geography accounted for significant heterogeneity in the pooled estimate.

Clarke MA, Long BJ, Del Mar Morillo A, Arbyn M, Bakkum-Gamez JN, Wentzensen N. JAMA Intern Med. 2018;178(9):1210–1222.