What Does Research Say About Mood Changes in Perimenopause?

You know the week I mean. Nothing has gone wrong, and yet the smallest thing lands hard, or the flat gray feeling stays through a whole Sunday, and you start to wonder whether it's you or something in your body that has shifted. It is a reasonable thing to ask whether the research has anything to say about it. It does, and it's messier than any headline lets on.

I read the abstracts of the main studies. Below I say what kind of study each one was, how many women were in it, and who they were, because those details change how much weight a result deserves.

What the large studies found

The Harvard Study of Moods and Cycles (Cohen and colleagues, 2006, Archives of General Psychiatry) was a prospective cohort study of 460 premenopausal women in the Boston area, aged 36 to 45, none of whom had ever had major depression. Women who entered perimenopause were twice as likely to develop significant depressive symptoms as women who stayed premenopausal, after adjusting for age and stressful life events. The risk was somewhat higher in women who also reported hot flashes or night sweats.

The same year, the Penn Ovarian Aging Study followed 231 women (Freeman and colleagues, 2006, Archives of General Psychiatry), also with no history of depression at the start. The study followed each woman over eight years and compared her with herself. High score on a standard depression symptom questionnaire was more than four times as likely (odds ratio 4.29) during the menopausal transition as when she was premenopausal, and a diagnosed depressive disorder was 2.5 times as likely.

Then there is the Study of Women's Health Across the Nation, known as SWAN. One analysis (Bromberger and colleagues, 2011, Psychological Medicine) interviewed 221 women, aged 42 to 52 and premenopausal at the start, at one Pittsburgh site. They were two to four times more likely to have a major depressive episode when perimenopausal or early postmenopausal. A prior history of major depression was a strong predictor throughout.

A 2024 systematic review and meta-analysis (Badawy and colleagues, Journal of Affective Disorders) pooled seven prospective cohort studies with 9,141 women, drawn from 17 studies and 16,061 women in the full review. Perimenopausal women had higher odds of depressive symptoms or diagnoses than premenopausal women (odds ratio 1.40, 95 percent confidence interval 1.21 to 1.61). Postmenopausal women did not show a significantly higher risk. I read only the abstract of this one.

A 2018 clinical guideline from The Menopause Society (then called NAMS) and the National Network of Depression Centers (Maki and colleagues) calls the perimenopause a window of vulnerability for depressive symptoms. It adds that most midlife women who have a major depressive episode in perimenopause have had one before. Again, I read the abstract only.

Where the data disagree or run out

The direction is fairly consistent across these studies. The size is not. Estimates range from 1.4 times the odds in the pooled analysis to more than four times in one cohort, and the studies measured depression in different ways: questionnaires in some, structured interviews in others. The 2024 authors say themselves that different ways of classifying menopausal stages and measuring depression added inconsistency, and that they could not compare perimenopause with postmenopause directly because longitudinal studies of that are lacking.

The reasons are less settled still. In the Penn study, swings in estradiol and rises in the hormones FSH and LH were associated with low mood. In the large SWAN analysis of 3,302 women aged 42 to 52 (Bromberger and colleagues, 2010, Archives of General Psychiatry), estradiol and FSH were not associated with high depressive symptoms. Higher testosterone was, modestly (odds ratio 1.15). Menopausal status itself stayed associated, whatever the hormones did. Stressful life events, low social support, and hot flashes each showed an independent link as well. So hormones are one candidate in the picture, not an established cause.

The trials that tested hormone therapy point in different directions, and I am describing them, not advising on them. A 2018 randomized trial of 172 women aged 45 to 60 with no depression at the start (Gordon and colleagues, JAMA Psychiatry) gave them a transdermal estradiol patch plus intermittent oral progesterone, or placebo, for 12 months. Of those on placebo, 32.3 percent had a questionnaire score in the clinically significant range at least once, versus 17.3 percent on hormones. The benefit showed up in the early transition, not in the late transition or after menopause. A 2015 systematic review (Rubinow and colleagues, Depression and Anxiety) looked at 24 studies and found only five randomized trials in women who were depressed, only two in samples made up entirely of perimenopausal women. It concluded there was little evidence that estradiol improves mood in women who are not depressed, and some evidence of an antidepressant effect in perimenopausal but not postmenopausal women. Those two sources are not quite asking the same question, which is part of why this field is hard to summarize.

One more caution. A small 2015 NIH trial of 56 postmenopausal women (Schmidt and colleagues, JAMA Psychiatry) found that women with past perimenopausal depression who were switched from estradiol to placebo had symptoms return, while those who stayed on it did not. It was short (three weeks of blinded treatment), small, and in a selected group, so I treat it as a direction rather than a finding to build on.

What I could not find: a trial of a daily habit, such as journaling or a set bedtime, aimed at mood specifically in perimenopausal women, with enough participants for me to lean on. The guideline lists psychotherapy, exercise, and natural health products among the options studied. The evidence behind each one is its own question, and I have not covered it here.

What this can look like in practice

None of this studied a Tuesday morning, so what follows is what the research gives me to work with and one small habit of mine, not a protocol. The risk numbers describe groups of women over years. They do not predict any single woman's week. What stood out most to me is that low mood and the physical symptoms travel together. Hot flashes, poor sleep, and mood overlap in the research, and the guideline says they complicate each other. If you want to read about the first two, I wrote about hot flashes and sleep and about what is known about sleep in perimenopause.

What I do is write one plain line at the end of each day, with a rough note on my mood, how I slept, and where I am in my cycle. I'm not testing anything. It just means that when I describe my month to someone, I am not reconstructing it from a feeling. The guideline names the menopausal stage and the other symptoms as part of what a clinician looks at, so a record like that may be useful in a conversation, but no trial tested it, and I won't pretend one did.

If you want something to keep that record in, I put together the Daily Check-In Journal on Etsy. It is a place to notice patterns in your own days. It doesn't treat anything, and I can't promise any day will feel different.

What the research doesn't back

It does not support reading your own week as a diagnosis or reading these odds as a verdict. It does not show that any habit or supplement shifts mood in perimenopause, and I would not start or stop anything because of an article. If low mood is heavy or lasting, that is a conversation for a clinician, not an article.