Stopping HRT: what your own data can and cannot show

Most women who take hormone therapy for symptoms find those symptoms return after stopping — in one Swedish population-based survey, the large majority of former users who had vasomotor symptoms before starting reported them again afterward, though usually less frequent and less bothersome than before. So “will they come back” is close to answered at the group level. The question your own record can actually help with is different: how this stretch compares with the weeks before you stopped, and whether it is still moving.

Why “will they come back” is the wrong thing to watch for

The base rates are not subtle. In that Swedish study, a validated questionnaire went to every woman aged 53 to 54 in one city; among former hormone therapy users who had experienced vasomotor symptoms before starting, roughly seven in eight reported recurrence after they stopped. In the much larger Women’s Health Initiative survey — more than eight thousand women, contacted eight to twelve months after the estrogen-plus-progestin arm was halted — over half of those who had begun therapy for symptoms reported vasomotor symptoms again, and the former treatment group also reported more pain and stiffness.

Read those together and the yes-or-no question loses most of its value. If you are bracing for a return, the research is largely on the side of your worry, and you are not being pessimistic.

What both studies also report, quietly, is the interesting part: the returning symptoms were commonly described as milder than the ones that led to treatment. Milder is not a number you can look up. It is a comparison, and it needs two records of your own.

The measurement almost everyone skips

Attention usually arrives after the fact. Something comes back, and only then does the tracking start — which leaves the new stretch with nothing to be measured against except memory of a year that has since blurred.

The fix is unglamorous and takes two to four weeks: record ordinary nights while you are still on therapy. Not a special protocol — just the nights you were going to have anyway, with the watch worn consistently, symptoms logged in the same words each day, and context noted when it is unusual (illness, alcohol, travel, a hot bedroom, a bad week).

That before-window is small and noisy, and it proves nothing on its own. What it does is give the after-window a reference band. Without it, “my sleep is worse” has no denominator; with it, the sentence becomes “my nights broke up roughly twice as often as they did in the fortnight before, across three weeks” — which is the kind of thing a ten-minute appointment can use.

Taper or stop: what the trial evidence actually found

Whether and how to come off hormone therapy is a decision for you and your clinician, who know your history, your risks and your reasons. The research adds one narrow fact and one honest gap.

The fact: a randomized study in women treated for vasomotor symptoms compared tapering down with stopping outright, and tapering did not reduce the recurrence of flashes or the likelihood of resuming therapy. The gap: a 2025 systematic review of women’s and clinicians’ experiences of discontinuation concluded that the best approach to stopping is still an open research question, not a settled protocol.

There is a small reading consequence worth knowing in advance. A gradual reduction spread over weeks blurs the edge between “on” and “off” in your own timeline, so a change that begins somewhere in the middle is harder to place. That is not a reason to choose one route over another — it is a reason to write down the dates your clinician gives you, whichever route you take.

You are not returning to the person who started

The years on therapy did not pause the underlying transition. In the SWAN cohort, the median total duration of frequent vasomotor symptoms was about 7.4 years, and it ran considerably longer for women whose symptoms began early, before their periods became irregular. Someone who started therapy at 49 and stops at 54 is not rewinding to 49 — she is meeting wherever that trajectory has since arrived.

This is why “milder than before” turns up so often in the survey data, and why comparing today against a memory of your worst pre-treatment month tends to mislead in both directions. It can make a genuinely difficult stretch feel like failure because it is not as bad as 2021. It can also let a slow, real drift pass unnoticed because nothing matches the old drama.

The comparison that carries information is with your own recent band — the weeks just before you stopped, then the weeks after — not with a population average and not with a version of yourself from several years ago. That is the whole of the idea we keep returning to: what a rising resting heart rate can and cannot mean works through the same logic on a single signal.

What your watch can hold, and what it cannot

An Apple Watch can carry four things through this period that memory carries badly: how continuous your nights were, your overnight wrist temperature band on supported models, your resting heart rate band across weeks, and your respiratory rate. All of them are estimates, not lab measurements, and all of them move for reasons that have nothing to do with hormones.

What none of them can do is label an episode. Research that physiologically monitored vasomotor symptoms found a fall in cardiac vagal control — one component of heart rate variability — around measured events at the group level; that is a real finding and still not a detector, because the same component falls with alcohol, a short night, a cold coming on, or a late heavy meal. The wrist temperature signal page sets out the same limits for temperature, and night sweats and wrist temperature covers what an overnight reading can and cannot support. If broken sleep is the thing that returned first, why sleep breaks in perimenopause goes further into that.

So the honest ceiling is this: your record can show that something changed, roughly when it began, and whether it persisted. Whether stopping is the reason is a clinical judgment, made with the rest of your history in view.

How Perigee reads it

Perigee does not decide whether stopping was right for you, and it never nominates hormone therapy as the cause of a change. It compares each available signal with your own recent baseline, keeps the dates you recorded beside the symptoms you logged, and looks for the relationship that keeps repeating across weeks rather than the one that looked dramatic on a single night.

When one lead is better supported than the others, it says which dates support it and what the comparison days looked like. When the record is too thin, or two explanations fit equally well, it says that instead of picking one. For the wider setup, how to track perimenopause with an Apple Watch covers which readings are worth having and how to keep them consistent; the research behind these signals lays out what the studies do and do not support. If you are earlier in the arc, starting HRT: what your Apple Watch can show and patch versus pill take the other two ends of the same question.

One small thing

If a stop date is on your calendar — even a provisional one — start the before-window tonight rather than on the day. One line a day is enough: how the night went, which symptoms showed up, anything unusual. Two weeks of that, recorded while things are still ordinary, will do more for the conversation you have in three months than any amount of careful attention paid after the fact, when there is no longer anything to compare against.