Summary
A six-month study of 356 perimenopausal women in China found that microbiome changes during hormone therapy differed by body site. Microbial patterns present before treatment were associated with differences in hormonal and symptom responses.
A six-month study of 356 perimenopausal women in China found that microbiome changes during menopausal hormone therapy varied across body sites, while microbial patterns present before treatment were associated with differences in response. The strongest measured restructuring was in vaginocervical samples; the gut microbiome changed little overall.
The study followed women aged 40–55 with menopausal symptoms as they received oral estradiol and dydrogesterone as part of routine care. Researchers collected samples at the start of treatment and at weeks 4, 12 and 24, then used shotgun metagenomic sequencing to profile microbes in vaginocervical secretions, urine, tongue coating and stool. Of the 356 women enrolled, 295 completed the 24-week regimen.
Microbiome shifts depended on body site
During treatment, the median modified Kupperman Index—a measure of menopausal symptom burden—fell by 63%. Other symptom scores also improved over the study period, including sleep quality and sexual function. These changes occurred during treatment in a before-and-after study, rather than a comparison with an untreated or placebo group.
Microbial changes were not uniform. The vaginocervical community showed the largest association with treatment over time (PERMANOVA R² = 0.008), followed by urinary (R² = 0.006) and tongue (R² = 0.004) communities. The gut community remained largely stable. In vaginocervical samples, the relative abundance of Lactobacillus crispatus rose from 15.1% at baseline to 26.0% at week 24, while Gardnerella vaginalis declined.
The researchers also grouped baseline vaginocervical profiles by their dominant community pattern. Women with G. vaginalis-dominated profiles showed the most microbial remodeling, while those with more diversified profiles showed less. Higher baseline L. crispatus was associated with greater subsequent L. crispatus abundance and larger improvement in Female Sexual Function Index scores. These are associations within the study, not evidence that changing a person’s microbiome would itself improve symptoms.
Baseline gut profiles tracked with treatment response
Although the gut microbiome changed little during treatment, its baseline community type was associated with differences in response. Women with Phocaeicola vulgatus-dominated gut profiles had greater increases in oestradiol and declines in follicle-stimulating hormone than women with Prevotella copri-dominated profiles. Rapid symptom remission at four weeks—defined as a modified Kupperman Index score of 7 or lower—was reported in 38.7% of the former group and 23.5% of the latter.
A model combining baseline symptoms and gut microbial features predicted rapid response in an independent test set of 70 participants, with an area under the curve of 0.793. For comparison, the clinical-feature-only model scored 0.701 and the microbiome-only model 0.665. The results suggest that microbial profiles could add information to symptom measures when studying who responds sooner, but the model is a research finding rather than an established clinical tool.
The study was a longitudinal, self-controlled before-and-after investigation: treatment was not assigned or changed by the research team, and there was no placebo group. Its design therefore cannot determine whether the observed microbial patterns caused differences in symptom or hormonal response. The findings also concern one oral estradiol/dydrogesterone regimen in women recruited across 14 centres in China. The authors propose further validation and mechanistic research before microbiome-guided care can be established.