How Long Perimenopause Lasts — and What Changes When It Ends

Quick answer: Two numbers answer this question and they measure different things. The transition itself — the stretch of changing cycles leading to your final period — lasts about four years on average, with a documented range of two to eight. Frequent hot flashes and night sweats last longer: a median of 7.4 years, of which a median of 4.5 years falls after the final period. Neither figure is wrong. They are answers to two different questions. And once it is over, most of this does ease. The exceptions are vaginal and urinary symptoms, bone, and, depending on which cohort you read, mood.

That distinction does most of the work here. Search this question and you will be told four years, or two to eight, or seven to nine, or up to 14, sometimes in adjacent paragraphs on one page. The figures conflict because they measure different things and get quoted as though they measured one. What almost nobody answers is the question that follows: whether you feel better once it ends.

The short answer: two clocks, not one

The transition. The Office on Women's Health answers the question directly: perimenopause "can last between two and eight years before your periods stop permanently," and "for most women, this transition to menopause lasts about four years" (OWH). Cleveland Clinic gives the same central figure — "the average length of perimenopause is about four years, but it can last up to eight years" (Cleveland Clinic). This clock is set by your menstrual cycles.

The symptoms. The best evidence here is SWAN, the Study of Women's Health Across the Nation, which followed 3,302 women at seven US sites from 1996 to 2013. Among the 1,449 reporting frequent hot flashes or night sweats — at least six days in the previous two weeks — median total duration was 7.4 years, and among the 881 whose final period could be dated, symptoms persisted a median of 4.5 years past it (Avis et al., 2015). This clock is set by symptoms, and it keeps running after the first one stops.

If you want one number to plan around, take the one that answers the question you are actually asking.

Why sources say four years, seven to nine, and up to 14

Seven figures are in wide circulation. Every one of them is traceable to a real source. None of them means the same thing.

Three examples show how the confusion propagates. Cleveland Clinic's page — the strongest consumer reference here, and the one ranking first for most of these searches — states that "perimenopause begins about eight to 10 years before menopause" and, further down the same page, that "the average length of perimenopause is about four years" (Cleveland Clinic, last updated 29 July 2024). Both sentences are quotable in isolation. Read together, they don't reconcile.

The second is subtler. The National Institute on Aging uses the same two-to-eight-year span but attaches it to something else: "symptoms related to menopause can last for between two and eight years" (NIA). Same span, same federal government, different measurement — one is timing cycles, the other symptoms.

The third is a definitional difference rather than a mistake, and it is the one worth learning to spot. UK contraception guidance defines perimenopause as "the transition phase preceding menopause and ending 1 year after the last menstrual period (LMP)" and says it "typically starts in a woman's mid-to-late-40s and lasts 4–5 years", while a later section of the same document puts "the average duration of perimenopause" at "around 4 years but this can vary greatly" (FSRH, cover dated August 2017, amended May 2025). A count that runs to twelve months past the last period is necessarily about a year longer than one that stops at it. Whether that definitional year is what separates the guideline's own two figures, it does not say — but it is the same extra year STRAW+10 adds, and it is the commonest reason two careful sources end up one apart.

One pair does agree, which is worth noticing in a field this noisy. The NHS puts symptom duration at "7 to 9 years, sometimes longer" (NHS, last reviewed 19 May 2026). SWAN measured a median of 7.4 years for frequent hot flashes and night sweats. A national health service summarising clinical experience and a seventeen-year American cohort study arrived at the same place.

How long the transition itself lasts

About four years, two to eight — but note exactly where that clock stops. The Office on Women's Health measures it "before your periods stop permanently" (OWH), which means the count ends at the final period.

The clinical definition runs 12 months longer. STRAW+10, the staging system specialists use for reproductive aging, defines perimenopause as beginning at Stage −2 and ending "12 months after the FMP" — the final menstrual period (Harlow et al., 2012). The entire first postmenopausal year sits inside it, which is where perimenopause and postmenopause are most often confused. Counting by that definition, add a year to the familiar four. Our guide to what perimenopause actually is sets out the full staging framework.

STRAW+10 also puts a figure on the final stretch: the late menopausal transition, which begins with your first gap of 60 days or longer without a period, is "estimated to last, on average, 1 to 3 years" (Harlow et al., 2012). For anyone already skipping periods that is the most useful number on this page, because it is the one you can locate yourself on. Which of the two stages inside perimenopause you are in turns on those cycle thresholds rather than on your age.

One limitation is worth stating plainly: none of this is knowable in real time. Menopause "can't be confirmed until one year after a woman's final menstrual period" (NIA), so the period that turns out to have been your last is only identifiable a year later. Nobody knows they have finished until they have been finished for months. For the other half of the timeline, our guide to when perimenopause starts and when symptoms usually peak covers both ends of the clock.

How long the symptoms last

SWAN assessed hot flashes and night sweats at each visit, roughly annually, across a median of 13 visits per woman. Of 3,302 participants, 1,449 reported frequent symptoms at one or more visits, and their median total duration was 7.4 years (Avis et al., 2015). That is a median, so half the sample had symptoms for longer.

Then the finding that most changes what a reader should expect. The strongest determinant of how long symptoms lasted was when they started. In the authors' words: "the best single predictor of both total VMS duration and post-FMP persistence was beginning VMS at an earlier MT stage."

Figures from Avis et al., 2015. "More than 11.8 years" is written that way in the paper because more than half that group still had symptoms when observation ended, so no exact median could be calculated. The 3.4-year figure appears in the abstract as the shortest total duration and in the results section in terms of duration after the final period; we quote it as published.

Read the top row against the intuition it contradicts. Symptoms arriving early, before your cycles have visibly changed, are commonly taken for a false start. In the data they predict the longest course — a median of more than 11.8 years, more than nine of them after the final period.

That is unwelcome information for exactly the woman most likely to be reading this, so here is the other half of it in the same breath. SWAN was measuring untreated symptoms: duration was censored at the point a participant started hormone therapy, and women already using it when their frequent symptoms began were excluded from the analysis altogether (Avis et al., 2015). These figures describe what happens when nothing is done — a natural course, not a prescription for how yours has to go. The authors drew that conclusion themselves, writing that a median of 7.4 years "highlights the limitations of guidance recommending short-term HT use." Which symptoms are actually attributable to the transition is covered in our guide to perimenopause symptoms.

Duration by race and ethnicity

SWAN recruited deliberately across racial and ethnic groups at its seven sites, and median symptom duration differed significantly between them (P < .001).

Avis et al., 2015. Every pairwise difference was statistically significant except Chinese versus Japanese.

More than five years separate the top and bottom rows. We are not going to explain that gap, because we found no verified explanation to offer. The NIA states the pattern without attributing a mechanism either: "compared with White women, Black women are more likely to have an early onset of menopause, experience more menopausal symptoms, and have symptoms for longer, while Asian women are less likely to suffer from menopausal symptoms" (NIA).

Two caveats belong with the table: this is a US cohort recruited at specific American sites and does not automatically describe populations elsewhere, and these are group medians, which say nothing predictive about any individual. The reason to publish them anyway is that a woman told to expect "a few years" is being handed an average that fits some groups considerably worse than others.

What actually makes it longer or shorter

Four factors here have a clear documented basis, and this list is deliberately shorter than the ones published elsewhere.

• When symptoms started — the strongest predictor in the data, as above: earlier onset, longer course (Avis et al., 2015).
• Smoking. The NIA states that smoking "is associated with an early onset of menopause and more severe menopausal symptoms" (NIA); SWAN found the duration half independently, with significantly longer total symptom duration in ever-smokers (Avis et al., 2015).
• Race and ethnicity, per the table above — an observed difference, not an explained one.
• Surgical menopause. Removal of both ovaries produces a different course: "if you have surgery to remove your ovaries and are not taking hormones, you will stop having periods and may experience symptoms of menopause immediately" (NIA). No number on this page describes that situation — SWAN excluded the 330 participants who had a hysterectomy or bilateral oophorectomy during the study (Avis et al., 2015).

The SWAN paper reports a longer list of characteristics statistically associated with duration. We have not converted them into things to change: associations measured in a single observational cohort are not established causes, and presenting them as levers a woman could have pulled would misrepresent what the study shows.

After perimenopause: what gets better, and what doesn't

Most of it gets better. The Office on Women's Health says so in one line: "After menopause, most women get relief from period problems or menopause symptoms" (OWH). That is the honest headline, and it is also where most pages stop, which leaves out the part a woman planning the next decade needs. The improvement is uneven, some of it is slow, and two items on the list move the other way.

The NHS makes the unevenness point in a sentence that deserves more attention than it gets. Symptoms "can change" over the years they last, so "you may get hot flushes and night sweats for a while, but they may improve and be replaced by other symptoms such as low mood and anxiety" (NHS, last reviewed 19 May 2026). Improvement in one symptom is not improvement overall.

Hot flashes: the slowest item on the list

SWAN's 4.5-year median persistence after the final period has a second cohort standing behind it. The Penn Ovarian Aging Study assessed 255 women annually from a premenopausal baseline and found moderate or severe hot flashes "reaching a peak of 46% in the first 2 years after the final menstrual period (FMP)", after it rather than before, with a mean duration past that point of "4.6 (2.9) years" (Freeman et al., 2014). A median of 4.5 years in one cohort and a mean of 4.6 in another are not strictly the same statistic, but they point at the same stretch of time.

The tail is longer than either average suggests. In the same cohort, "One third of women at 10 years or more after menopause continued to experience moderate/severe hot flashes." Not mild ones, and not a negligible fraction.

Memory and concentration: the one that rebounds

The clearest good news on this page is about cognition, and it is more specific than "brain fog lifts".

SWAN tested 2,362 women repeatedly over four years on processing speed, verbal memory and working memory. What it tracked was not raw ability but whether scores improved with practice, the way they normally do when the same test is taken again. Premenopausal, early perimenopausal and postmenopausal women "scored higher with repeated SDMT administration" (the processing-speed test), while late perimenopausal women did not, and delayed verbal recall "climbed during premenopause and postmenopause" but stalled in between. The authors concluded that "perimenopause was associated with a decrement in cognitive performance", that "Improvement rebounded to premenopausal levels in postmenopause", and that "menopause transition-related cognitive difficulties may be time-limited" (Greendale et al., 2009).

Two limits belong with that. It is four years of observation in one cohort rather than a promise to any individual, and an observational design cannot separate the transition from everything else happening in a woman's forties. What it counts against is the reading the symptom usually invites: that the fog is the opening stage of a permanent decline. In this cohort it was not.

Sleep: the one that mostly doesn't change

A SWAN analysis modelled night-waking patterns in 1,285 women who reached menopause naturally and found four distinct trajectories, from a low-prevalence group of 487 women to a high-prevalence group of 236. Sleep problems became more common across the cohort overall, but the final period was not the turning point: "awakenings were stable from pre-FMP to post-FMP" (Kravitz et al., 2017).

What predicted waking that carried on into postmenopause was trouble falling asleep, early-morning waking and frequent hot flashes and night sweats. That is a more useful finding than it first looks, because each of those is treatable in its own right rather than something to outlast.

Mood: where two good studies disagree

This is the one question here the evidence does not settle, and splitting the difference would misrepresent both sides.

The Penn cohort followed 203 women through what its authors call "a 14-year period around menopause", modelling depressive symptoms "from 10 years before to 8 years after the final menstrual period". Its result was clean and directional: "Relative to the FMP, the risk of depressive symptoms was higher in the years before and lower in the years after the FMP", with prevalence falling by "approximately 15% of baseline per year (odds ratio, 0.85; 95% CI, 0.81-0.89)" (Freeman et al., 2014).

SWAN, following 1,551 women for a median of nineteen years, found the opposite on the far side of the final period. Compared with being premenopausal, women had higher odds of a high depressive-symptom score while postmenopausal, an odds ratio of 1.49 (95% CI 1.09-2.04), with no significant difference during perimenopause itself, and the authors concluded that "postmenopause remains a period of increased risk for higher depressive symptoms" (Kravitz et al., 2022).

The two are not quite measuring the same thing. Penn asks how a woman's risk moves relative to her own final period; SWAN compares menopausal stages against premenopause in a sample seven times larger, drawn from seven US sites. A third analysis, from SWAN's Pittsburgh site, sits in between: its 221 participants were "two to four times more likely to experience a major depressive episode" while peri- or early postmenopausal than while premenopausal, and that held after adjustment for prior depression, upsetting life events, psychotropic medication, vasomotor symptoms and reproductive hormone levels (Bromberger et al., 2011). That one counts diagnosed episodes rather than symptom scores. None of the three is a mistake, and we are not going to average them into a reassurance.

What they do agree on is the part you can act on. How a woman felt before the final period was the strongest predictor of how she felt after it. Women on SWAN's consistently-high pre-menopause symptom trajectory had nearly seven times the odds of high scores afterwards (odds ratio 6.88; 95% CI 4.72-10.02) (Kravitz et al., 2022), and in the Penn cohort a history of depression carried risk "more than 13 times greater overall and 8 times greater after menopause compared with women with no depression history" (Freeman et al., 2014).

So mood is not a symptom to wait out on the strength of a date in the future. Low mood that persists, or any thought of harming yourself, is a reason to contact a clinician now.

Vaginal and urinary symptoms: the exception nobody mentions

One group of symptoms runs the other way, and the professional position statement devoted to them says so in a single sentence: "In contrast to vasomotor symptoms (VMS) that usually improve over time, GSM is generally progressive without effective therapy" (The Menopause Society, 2020).

GSM is the genitourinary syndrome of menopause — vaginal dryness, pain during sex, burning, urinary urgency and recurrent urinary infections. It affects "approximately 27% to 84% of postmenopausal women", and in one report of more than 900 women at routine examinations it "was identified in 84% of women 6 years after menopause". The Office on Women's Health puts the same direction of travel plainly: "Vaginal dryness may be more common post-menopause" (OWH).

The reason to say this out loud on a page about duration is that expecting improvement works against getting treated. In a survey of 1,858 US postmenopausal women with genitourinary symptoms, 50% "had never used any therapy for this problem", and the same statement pairs that figure with the absence of spontaneous improvement. These symptoms are treatable and worth raising with a clinician; waiting is the one approach the evidence does not support. Bleeding after twelve months without a period is a separate matter and needs assessment whatever else is going on.

Bone: the change that produces no symptoms at all

Bone loss announces nothing, and it has the tightest timing of anything here.

SWAN scanned 862 women across a ten-year window, five years either side of the final period, and found that loss "began 1 year before the FMP" and ran fast for roughly three years — from one year before to two years after. Most of the decade's loss happened inside that window: 7.38% at the lumbar spine out of 10.6% across the full ten years, and 5.8% at the femoral neck out of 9.1% (Greendale et al., 2012). After it, loss "decelerated (but did not cease)".

Two of those three years have already gone by the time twelve months without a period confirm that menopause has been reached, which is an argument for raising bone health with a clinician during the transition rather than after it.

Can it be stopped, reversed or delayed?

No. Perimenopause is a transition rather than a condition, so there is nothing to cure and nothing on offer that turns it back; treatment acts on the symptoms, not on the process underneath them. UK contraception guidance draws the same line about fertility, noting that hormone therapy "does not restore fertility" — a point it makes about women who are already past menopause (FSRH). That distinction matters most when something is being marketed as a reversal. Our treatment guide covers what treatment can and can't change, including the clinic services sold as "ovarian rejuvenation" and what reproductive-medicine guidance says about them.

Does it have to feel like this the whole time?

That is a different question with a different answer, and it is where duration figures stop being useful. Everything above describes what these symptoms do when left alone: a natural history, not a forecast you are obliged to accept.

Hormone therapy "remains the most effective treatment for vasomotor symptoms and the genitourinary syndrome of menopause and has been shown to prevent bone loss and fracture," per the 2022 position statement of The North American Menopause Society, now The Menopause Society, which adds that risks differ by type, dose, duration, route and timing of initiation, so the decision is individualised (NAMS, 2022). Non-hormonal options with good and consistent evidence exist too: the same body's 2023 nonhormone statement recommends cognitive-behavioural therapy, clinical hypnosis, SSRIs and SNRIs, gabapentin and fezolinetant at Level I — its strongest evidence grade — specifically for vasomotor symptoms (NAMS, 2023). One update post-dates that statement and should travel with it: on 16 December 2024 the FDA added a boxed warning, its most prominent warning, to fezolinetant (Veozah) for rare but serious liver injury (FDA).

Which of those fits is a conversation with a clinician, not a decision to make from an article. The point for a page about duration is narrower: a median of 7.4 years is a long time to wait out something treatable, and early-onset symptoms — the ones predicting the longest course — are the strongest argument for having that conversation sooner rather than at the end of a decade.

Common questions about how long perimenopause lasts (FAQ)

How long does menopause usually last? Menopause itself is a single point in time — the day twelve months have passed since your last period — so it has no length. What has a length is the transition before it and the symptoms around it, and they are measured separately. The transition runs about four years on average, with a range of two to eight (OWH). Symptoms run longer: the NHS puts them at "7 to 9 years, sometimes longer" (NHS), and among 1,449 women in SWAN with frequent hot flashes or night sweats the median total duration was 7.4 years (Avis et al., 2015). Conflating the two is why published answers disagree.

Do you feel better after perimenopause? Mostly, yes. The Office on Women's Health states that "After menopause, most women get relief from period problems or menopause symptoms" (OWH), but the improvement is uneven. Hot flashes fade over years rather than months, and one in three women in the Penn Ovarian Aging Study still reported moderate or severe ones ten or more years after menopause (Freeman et al., 2014). Vaginal and urinary symptoms run the other way: The Menopause Society describes them as "generally progressive without effective therapy" (The Menopause Society, 2020). And mood is unsettled, with one cohort finding risk falls after the final period (Freeman et al., 2014) and SWAN finding higher odds of high depressive-symptom scores in postmenopause (Kravitz et al., 2022). Cognition is the clearest improver: in SWAN, "Improvement rebounded to premenopausal levels in postmenopause" on tests of processing speed and verbal memory (Greendale et al., 2009).

How long do hot flashes last? Both readings of that question have an answer. An individual episode lasts between 30 seconds and 10 minutes (NIA). Across the whole course, SWAN found a median of 7.4 years of frequent hot flashes or night sweats, 4.5 of them after the final period (Avis et al., 2015), and the Penn Ovarian Aging Study found a mean of 4.6 years after the final period, with moderate or severe flashes peaking at 46% in the first two years afterwards (Freeman et al., 2014).

How do I know when perimenopause is over? Only in hindsight. Menopause cannot be confirmed until 12 months have passed with no period or spotting (NIA), and under STRAW+10 perimenopause ends at that 12-month mark rather than at the final period (Harlow et al., 2012). Pregnancy remains possible throughout, which is why the NIA advises continuing birth control for the whole of those 12 months (NIA; OWH). Any bleeding after those twelve months needs checking by a clinician.

Can perimenopause be reversed? No. Perimenopause is the transition to menopause rather than a condition that can be cured, and no treatment turns it back; what treatment does is act on the symptoms while they last. UK contraception guidance draws the same line about fertility, noting that hormone therapy "does not restore fertility" in women who are already past menopause (FSRH). Services marketed as reversing or rejuvenating the ovaries are covered on our treatment page, along with what reproductive-medicine guidance says about them.

Medical disclaimer: This article is for informational purposes only. Always consult a qualified healthcare provider before making changes to your health routine.

Sources

1. Office on Women's Health, US Department of Health and Human Services. Menopause basics — Government health information. https://womenshealth.gov/menopause/menopause-basics
2. National Institute on Aging (NIH). What Is Menopause? — Government health information. https://www.nia.nih.gov/health/menopause/what-menopause
3. Cleveland Clinic. Perimenopause — Health system disease reference. https://my.clevelandclinic.org/health/diseases/21608-perimenopause
4. Avis NE, Crawford SL, Greendale G, et al. Duration of Menopausal Vasomotor Symptoms Over the Menopause Transition. JAMA Internal Medicine, 2015;175(4):531–539 (SWAN) — Peer-reviewed journal article via PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC4433164/
5. Harlow SD, Gass M, Hall JE, et al. Executive summary of the Stages of Reproductive Aging Workshop +10: addressing the unfinished agenda of staging reproductive aging. Menopause, 2012;19(4) — Peer-reviewed journal article via PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC3340903/
6. The North American Menopause Society. The 2022 hormone therapy position statement of The North American Menopause Society. Menopause, 2022;29(7):767–794 — Professional position statement (PDF). https://menopause.org/wp-content/uploads/professional/nams-2022-hormone-therapy-position-statement.pdf
7. The North American Menopause Society. The 2023 nonhormone therapy position statement of The North American Menopause Society. Menopause, 2023;30(6):573–590 — Professional position statement (PDF). https://menopause.org/wp-content/uploads/professional/2023-nonhormone-therapy-position-statement.pdf
8. US Food and Drug Administration. FDA adds warning about rare occurrence of serious liver injury with use of Veozah (fezolinetant) for hot flashes due to menopause — Drug Safety Communication, 16 December 2024. https://www.fda.gov/drugs/drug-safety-communications/fda-adds-warning-about-rare-occurrence-serious-liver-injury-use-veozah-fezolinetant-hot-flashes-due
9. NHS. Menopause and perimenopause — Symptoms — Government health information, page last reviewed 19 May 2026. https://www.nhs.uk/conditions/menopause-and-perimenopause/symptoms/
10. Faculty of Sexual & Reproductive Healthcare. Contraception for Women Aged Over 40 Years — Clinical guideline (PDF), August 2017, amended May 2025 per the cover page. https://www.cosrh.org/Common/Uploaded%20files/documents/fsrh-guideline-contraception-for-women-aged-over-40-years.pdf
11. The North American Menopause Society. The 2020 genitourinary syndrome of menopause position statement of The North American Menopause Society. Menopause, 2020;27(9):976–992 — Professional position statement (PDF). https://menopause.org/wp-content/uploads/professional/The_2020_genitourinary_syndrome_of_menopause.5.pdf
12. Greendale GA, Huang MH, Wight RG, et al. Effects of the menopause transition and hormone use on cognitive performance in midlife women. Neurology, 2009;72(21):1850–1857 (SWAN) — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/19470968/
13. Kravitz HM, Janssen I, Bromberger JT, et al. Sleep Trajectories Before and After the Final Menstrual Period in The Study of Women's Health Across the Nation (SWAN). Current Sleep Medicine Reports, 2017;3(3):235–250 — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/28944165/
14. Freeman EW, Sammel MD, Boorman DW, Zhang R. Longitudinal pattern of depressive symptoms around natural menopause. JAMA Psychiatry, 2014;71(1):36–43 — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/24227182/
15. Kravitz HM, Colvin AB, Avis NE, et al. Risk of high depressive symptoms after the final menstrual period: the Study of Women's Health Across the Nation (SWAN). Menopause, 2022;29(7):805–815 — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/35796553/
16. Bromberger JT, Kravitz HM, Chang YF, et al. Major depression during and after the menopausal transition: Study of Women's Health Across the Nation (SWAN). Psychological Medicine, 2011;41(9):1879–1888 — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/21306662/
17. Freeman EW, Sammel MD, Sanders RJ. Risk of long-term hot flashes after natural menopause: evidence from the Penn Ovarian Aging Study cohort. Menopause, 2014;21(9):924–932 — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/24473530/
18. Greendale GA, Sowers M, Han W, et al. Bone mineral density loss in relation to the final menstrual period in a multiethnic cohort: results from the Study of Women's Health Across the Nation (SWAN). Journal of Bone and Mineral Research, 2012;27(1):111–118 — Peer-reviewed journal article. https://pubmed.ncbi.nlm.nih.gov/21976317/

Sources 1–8 accessed 3 August 2026 and re-checked 16 October 2026 where reachable; sources 9–18 accessed 16 October 2026. The National Institute on Aging page (source 2) returned an HTTP error to every request on 16 October 2026, so its wording could not be re-verified on that date; the quotations from it are unchanged from the original 3 August 2026 check.

Claire Whitfield

Women's Health Specialist

A women's health specialist who checks what a number is measuring before repeating it.

View profileView Xing profile