Menopausal Depression: What 70 Trials Say About Treatment

For women in perimenopause with a diagnosed depressive disorder, the best-ranked treatment in a large network meta-analysis was fluoxetine combined with oral hormone therapy—not fluoxetine alone and not hormone therapy alone. The estimate was large: standardized mean difference (SMD) −1.59 versus placebo, with a 95% confidence interval of −2.69 to −0.50. But the result has strict boundaries. It applied to a specific subgroup, and the same analysis found no pharmacologic or hormonal treatment that beat placebo in women more than one year past their final menstrual period.

That distinction matters more than a catchy treatment ranking. Menopausal symptoms, clinical depression during perimenopause, and depression that begins after menopause are overlapping problems, not interchangeable ones.

The treatment ranking came with a stage-specific warning

Tseng et al.'s 2023 network meta-analysis pooled 70 randomized controlled trials involving 18,530 women, with a mean age of 62.5 years. Unlike a standard meta-analysis that compares one treatment with one comparator at a time, a network meta-analysis uses direct and indirect trial comparisons to place several treatments on a common scale. Here, the network included antidepressants, hormone therapies, and combination regimens for depressive symptoms in peri- and postmenopausal women.

Among perimenopausal women with a confirmed clinical diagnosis of depression, fluoxetine plus oral hormone therapy ranked highest and outperformed placebo (SMD −1.59; 95% CI −2.69 to −0.50). Because the confidence interval stayed below zero, this comparison was statistically significant. An SMD of that size would usually count as large, but the interval is unusually wide. The true effect could be closer to −0.50 than −1.59, and the spread points to uncertainty from differences among the underlying trials.

The analysis does not justify telling every woman with low mood during midlife to add hormones to an SSRI. Women without a formal depression diagnosis showed no treatment superiority over placebo. Neither did women classified as postmenopausal, defined as having more than one year of amenorrhea. Those null findings are not footnotes; they define who the result does and does not describe.

The study also does not settle whether transdermal estrogen would match oral hormone therapy, which progestin is best, how long treatment should continue, or what the safety profile is for the specific fluoxetine-plus-hormone combination. Those decisions require an individual discussion of symptoms, depression history, uterine status, clotting and cardiovascular risk, breast cancer risk, and other medications.

Why estrogen could change an antidepressant response

The proposed biology is plausible, though it does not prove that this mechanism caused the trial result. Estradiol influences serotonin synthesis, serotonin-transporter activity, and 5-HT2A receptor density, according to a 2025 review of sex steroids and serotonin. During perimenopause, estrogen levels can fluctuate sharply rather than simply decline in a smooth line. That instability may affect mood regulation in people who are susceptible to it.

This offers a reasonable explanation for why a serotonergic antidepressant and hormone therapy could work better together during the transition. It remains an explanation, not a license to treat mechanism diagrams as clinical proof. The serotonin review is a narrative review, not a randomized test of estrogen augmentation.

Current reviews also describe a shift toward considering hormone therapy for perimenopausal psychological symptoms. Langhe et al. and Crockett et al. discuss NICE guidance that positions HRT as a first-line option for perimenopausal mood disturbance. Guideline context can help frame a clinical conversation, but it does not erase the uncertainty around formulation, duration, and combination safety.

After menopause, the ranking stops being a prescription

It would be tempting to turn the network meta-analysis into a simple rule: depression around menopause needs hormones, and depression after menopause does not. The data are narrower than that. In the postmenopausal subgroup, none of the pharmacologic or hormonal interventions in this analysis outperformed placebo. That is a null result within this evidence base, not proof that every postmenopausal depressive episode is biologically identical or untreatable.

Several reasons call for restraint. The pooled trials had a mean participant age of 62.5 years, so the perimenopausal group may have made up only a fraction of the full sample. The leading combination estimate had a broad confidence interval. The published information available here does not provide the complete treatment hierarchy, individual comparisons for each SSRI or hormone regimen, or trial-level safety outcomes. A ranking is useful; it is not a treatment plan with the labels peeled off.

Exercise has a better case for sleep than for head-to-head depression treatment

Sleep problems often magnify mood symptoms, and exercise offers a non-drug option with menopausal-specific comparative evidence. Zhou et al.'s 2025 network meta-analysis evaluated 31 randomized trials and ranked five exercise approaches for sleep quality in perimenopausal and postmenopausal women: aerobic exercise, resistance training, yoga, combined exercise, and mind-body practices.

The analysis reported larger sleep-quality improvements among women aged 55 or younger (SMD −2.00; 95% CI −2.70 to −1.31), women with sleep disorders (SMD −2.29; 95% CI −2.79 to −1.78), and participants in Asia (SMD −2.50; 95% CI −3.62 to −1.38). It also assigned differing certainty ratings across modalities, including low certainty for flexibility and aerobic exercise and moderate certainty for combined exercise. The available results do not provide a single overall winning modality or its effect size, so claiming one would be a neat answer with no evidence behind it.

For mood, a 2025 systematic review and meta-analysis of randomized trials found that aerobic exercise improved depression and anxiety outcomes in postmenopausal women. It also assessed cardiovascular measures. A separate 2024 meta-analysis of randomized trials found aerobic training improved cardiometabolic risk factors in postmenopausal women. Exercise therefore has practical appeal when sleep, mood, physical function, and cardiometabolic health all need attention. Still, these reviews did not compare aerobic exercise directly with fluoxetine-plus-hormone therapy for perimenopausal clinical depression. They answer a different question.

Probiotics are an adjunct hypothesis, not menopause treatment

The gut-brain story has enough human trial evidence to be interesting and too little menopause-specific evidence to be promoted as a replacement for established care. Cheng et al.'s 2025 meta-analysis included 12 randomized trials with 1,258 patients and found that mixed probiotics containing Lactobacillus reuteri improved depressive symptoms versus control (SMD −0.44; 95% CI −0.72 to −0.16). Acceptability did not differ significantly between groups (odds ratio 1.04; 95% CI 0.75 to 1.45).

Broader evidence points in the same direction. Pan et al.'s 2025 meta-analysis of microbiome-targeted treatments for depression reported an SMD of −0.26 (95% CI −0.32 to −0.19; I² = 54%) across eligible trials, with acceptable safety. That is a modest pooled effect with moderate heterogeneity, not a substitute for assessment and treatment of major depression.

Neither analysis establishes efficacy specifically in perimenopausal or postmenopausal women. Menopause changes both hormonal signaling and the gut microbiome, so a result in a general adult depression population may not transfer cleanly. A probiotic may be a reasonable clinician-approved adjunct for some people; it is not evidence-based hormone replacement in a capsule.

What the evidence supports now

• For perimenopausal women with diagnosed depression, ask a qualified clinician whether hormone therapy belongs in the treatment discussion, especially if an SSRI is already under consideration.
• Do not extend the fluoxetine-plus-oral-HRT result to nonspecific mood symptoms, self-treatment, or depression arising after menopause. The primary analysis did not show superiority in those groups.
• Use exercise as a credible non-hormonal tool for sleep and general health, with evidence that aerobic exercise can also improve depression and anxiety outcomes in postmenopausal women.
• Treat probiotics as an emerging adjunct with general depression evidence, not as a menopausal-depression therapy.

The most useful finding is not that one combination topped a ranking. It is that treatment effects appeared to depend on reproductive stage and a confirmed depression diagnosis.

That makes the next clinical question more specific than "Which antidepressant?" It is: what exactly is being treated, and where is the person in the menopausal transition? The available evidence gives a meaningful answer for one defined group. Outside that group, certainty falls quickly.