Longitudinal trajectories of menopausal symptoms in women receiving endocrine therapy for breast cancer compared with natural menopause.

Save

Background

Endocrine therapy (ETx) for hormone receptor–positive breast cancer frequently induces menopausal symptoms that may compromise treatment adherence. However, longitudinal symptom trajectories compared with natural menopause remain unclear.

Methods

This longitudinal study included 5734 women from two prospective cohorts: 2583 women with breast cancer from the Breast Cancer Information Grand Round for Survivorship cohort and 3151 women without breast cancer from the Kangbuk Samsung Health Study. Menopausal symptoms were assessed using the Menopause Rating Scale (MRS). Linear mixed-effects models evaluated longitudinal changes by breast cancer status, menopausal status, and ETx use.

Difference-in-differences (DiD) analyses designated premenopausal women without breast cancer as the reference group.

Results

Among premenopausal women, ETx was associated with greater increases in total MRS scores than in women without breast cancer (DiD, 3.30; 95% CI, 2.04–4.58 at 1 year). Within the breast cancer cohort, premenopausal women receiving ETx experienced sustained increases in menopausal symptoms over 3 years, with the largest increases observed among those receiving regimens including gonadotropin-releasing hormone agonists. Among postmenopausal women, ETx was associated with modest changes in total MRS scores but a significant increase in urogenital symptoms (DiD 0.71; 95% CI, 0.26–1.16).

Conclusions

Compared with women without breast cancer, women receiving ETx experienced greater increases in menopausal symptoms during the first year of follow-up. Within the breast cancer cohort, elevated symptom burden persisted over 3 years, particularly among premenopausal women receiving ovarian function suppression. These findings highlight the importance of proactive symptom monitoring and supportive care during long-term endocrine therapy.

Introduction

Endocrine therapy (ETx) is a central component of treatment for hormone receptor-positive breast cancer but frequently causes menopausal symptoms that can persist throughout treatment [1], [2], [3]. Women receiving ETx frequently report vasomotor symptoms, musculoskeletal pain, sleep disturbance, cognitive complaints, and urogenital symptoms [4], all of which negatively affect quality of life and daily functioning [5], [6], [7].

Premenopausal women appear particularly vulnerable because ETx induces a rapid transition to a hypoestrogenic state [8], and ovarian function suppression is associated with more severe menopausal symptoms. In the Suppression of Ovarian Function Trial (SOFT), patient-reported outcomes showed greater hot flushes and sleep disturbance during the early treatment period and persistent vaginal dryness among women receiving tamoxifen plus ovarian function suppression compared with tamoxifen alone [9].

Despite the high prevalence of these symptoms, most previous studies have relied on cross-sectional designs. A longitudinal cohort of 1462 Chinese patients with breast cancer followed for up to 5 years reported no substantial increase in the overall prevalence of severe menopausal symptoms over time [10]. However, that study did not stratify outcomes by menopausal status at treatment initiation, an important determinant of menopausal symptom burden.

In addition, because no non-cancer comparison group was included, treatment-associated symptom changes could not be distinguished from changes accompanying the natural menopausal transition. To address these gaps, we analyzed two prospective cohorts with complementary characteristics. The Breast Cancer Information Grand Round for Survivorship (BIG-S) cohort provided detailed treatment information and repeated symptom assessments in women with breast cancer, whereas the Kangbuk Samsung Health Study (KSHS) followed women without breast cancer through the natural menopausal transition.

Materials and methods

Study participants

This longitudinal cohort study evaluated changes in menopausal symptoms among 5734 women according to menopausal status and, among women with breast cancer, endocrine therapy. The study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.

Study participants were derived from two prospective cohort studies to compare menopausal symptoms between women with and without breast cancer (Fig. 1). Women with breast cancer were recruited from a complementary prospective cohort at Samsung Medical Center, the BIG-S [11], which has consecutively enrolled women newly diagnosed with histologically confirmed breast cancer since November 2018.

Eligible participants were women aged ≥20 years without prior malignancy or distant metastasis who were scheduled to undergo curative-intent surgery. Among these patients, 2583 women with primary breast cancer and available baseline information on endocrine therapy and menopausal symptoms were included in this analysis.Fig. 1Flow diagram for the selection of the study population.

Women without breast cancer were recruited from a cohort study of Korean women designed to evaluate physiological and psychosocial changes across menopausal stages. Participants were enrolled between 2014 and 2018 from the KSHS, an ongoing cohort of Korean adults undergoing regular health examinations at the Kangbuk Samsung Hospital Total Healthcare Center clinics [12]. The eligibility criteria at enrollment were women aged 42–52 years with no history of oophorectomy, hysterectomy, hormone therapy, or chronic diseases that could affect menstrual cycles (e.g., cancer, renal failure, hypothyroidism, or hyperthyroidism), and who had experienced at least one menstrual period within the previous 3 months.

Menopausal symptoms had been assessed since cohort enrollment (2014–2018) using the Menopause-Specific Quality of Life Questionnaire. Beginning in January 2024, the MRS was additionally introduced as a standardized measure of menopausal symptoms. For the present analysis, baseline was defined as each participant's first available MRS assessment between January 2024 and November 2025, irrespective of the original cohort enrollment date, and follow-up time was calculated from this individual MRS baseline.

Among these participants, 3151 women who completed the menopausal symptom survey were included.

All participants provided written informed consent, and the study was approved by the Institutional Review Board of Samsung Medical Center (SMC-2019-08-009-020) and of Kangbuk Samsung Hospital (KBSMC 2023-05-036).

Measurement

Menopausal symptoms

Menopausal symptoms were assessed using the existing Korean-language version of the Menopause Rating Scale (MRS), a self-administered questionnaire comprising 11 items patient-reported instrument designed to evaluate the presence and severity of menopause-related symptoms as a self-administered questionnaire [13], [14]. The MRS comprises three domains. The somatic/vasomotor domain includes hot flushes and sweating, heart discomfort, sleep problems, and joint or muscular discomfort.

The psychological domain includes depressive mood, irritability, anxiety, and physical or mental exhaustion. The urogenital domain includes sexual problems, bladder problems, and vaginal dryness [15]. Each item is rated on a 5-point Likert scale ranging from 0 to 4, with 0 indicating that the symptom is absent and 4 indicating that it is very severe.

Domain scores are calculated by summing the scores of the corresponding items. The somatic/vasomotor and psychological domain scores each range from 0 to 16, and the urogenital domain score ranges from 0 to 12. The total MRS score is calculated as the sum of all 11 item scores and ranges from 0 to 44, with higher scores indicating a greater menopausal symptom burden.

The same MRS questionnaire and scoring procedure were applied to all participants, irrespective of breast cancer status or baseline menopausal status. Thus, premenopausal women without breast cancer rated the presence and severity of all 11 symptoms using the same 0–4 scale as the other study groups. Menopausal status was classified separately and did not affect the scoring of individual MRS items.

In the present study, total and domain-specific MRS scores were analyzed as continuous outcomes. Participants were not categorized into mild, moderate, or severe symptom groups. In the BIG-S cohort, MRS scores were measured at baseline and at 1, 2, and 3 years after diagnosis.

In the KSHS, MRS scores were assessed at baseline and at 1-year follow-up.

Although no universally accepted minimal clinically important difference has been established for the MRS, relative changes of approximately 20% from baseline have been regarded as perceptible in previous patient-reported outcome research [16]. In the present cohort, given the relatively low mean baseline MRS scores, such relative changes correspond to absolute differences of approximately 2–3 points. These values were used to aid interpretation of symptom changes rather than to define a prespecified clinical threshold.

Menopausal status

Menopausal status was classified according to the information available in each cohort. In the KSHS, menopausal stages were categorized according to the Stages of Reproductive Aging Workshop (STRAW) +10 criteria as follows: premenopause, early menopausal transition, late menopausal transition, and postmenopause. Premenopause was defined as regular menstrual cycles without meeting criteria for transition stages.

Early menopausal transition was defined as a persistent difference of ≥7 days in consecutive cycle length. Late menopausal transition was defined as amenorrhea lasting ≥60 days. Postmenopause was defined as amenorrhea lasting ≥12 months [17].

For the present analysis, menopausal status was consolidated into three groups representing the natural menopausal trajectory: (1) premenopausal women, including those in early transition; (2) late menopausal transition; and (3) postmenopausal women. This classification reflects evidence that menopausal symptoms, particularly vasomotor symptoms, increase during late transition and remain elevated after menopause, whereas premenopausal and early transition women generally maintain preserved ovarian function [18]. In the BIG-S cohort, menopausal status was clinically classified as premenopausal or postmenopausal at breast cancer diagnosis because the information required for detailed STRAW+10 staging was not available.

For the primary analysis, participants were classified according to menopausal status, breast cancer status, and ETx use. Women without breast cancer were categorized according to menopausal status to characterize symptom changes occurring during the natural menopausal transition. Women with breast cancer were further categorized according to ETx exposure because ETx may induce or exacerbate menopausal symptoms beyond changes associated with the natural menopausal transition.

Thus, this classification was intended to distinguish changes associated with the natural menopausal transition from those additionally associated with ETx among women with breast cancer, rather than to create an equal number of case and control groups. The reference group consisted of premenopausal women without breast cancer. The remaining analytic subgroups included women in late menopausal transition without breast cancer, premenopausal women with breast cancer who did not receive ETx, premenopausal women with breast cancer who received ETx, postmenopausal women without breast cancer, postmenopausal women with breast cancer who did not receive ETx, and postmenopausal women with breast cancer who received ETx.

ETx exposure was determined from treatment records. Women without a documented prescription for ETx were classified as the non- ETx group.

In secondary analyses restricted to women with breast cancer, longitudinal symptom trajectories were examined according to chemotherapy exposure. In sensitivity analyses among premenopausal women with breast cancer, endocrine therapy regimens were categorized as no ETx, tamoxifen monotherapy, tamoxifen combined with a gonadotropin-releasing hormone (GnRH) agonist, and aromatase inhibitor (AI) combined with a GnRH agonist.

Covariates

Baseline sociodemographic information (age, marital status, and education), lifestyle factors (physical activity, smoking status), antidepressant use, cancer stage, and treatment modalities (chemotherapy, radiotherapy, and endocrine therapy) were collected in both cohorts through electronic surveys and medical record review. Body mass index (BMI) was assessed using multi-frequency bioelectrical impedance analysis in the BIG-S cohort (InBody 770, InBody Co., Ltd., Seoul, Korea) [19]. In the KSHS, body composition was measured using the InBody 720 (Biospace Inc., Seoul, Korea), which also employs multi-frequency BIA technology.

Statistical analysis

Baseline characteristics were summarized as means with standard deviations for continuous variables and frequencies with percentages for categorical variables. The differences of characteristics were assessed using ANOVA for continuous variables and chi-square tests for categorical variables.

Longitudinal changes in menopausal symptoms were analyzed using linear mixed-effects models with random intercepts to account for within-individual correlations across repeated measurements. Models included random intercepts for participants and fixed effects for time, study group, and their interaction. Time was treated as a categorical variable corresponding to each follow-up visit.

Participants were not required to complete every scheduled follow-up assessment. Women who missed one or more intermediate visits but subsequently returned for later assessments contributed all available observations to the linear mixed-effects models. Linear mixed-effects models incorporated all available repeated observations without requiring complete follow-up for every participant and provided valid estimates under the missing-at-random assumption.

To evaluate potential attrition bias, baseline characteristics were compared between participants who completed the 3-year follow-up and those who did not using independent t-tests for continuous variables and χ2 tests for categorical variables. Standardized mean differences were also calculated to assess the magnitude of baseline differences. For the main analysis, adjusted differences from baseline and difference-in-differences (DiD) estimates were calculated to compare changes in MRS scores across the seven predefined groups, using premenopausal women without breast cancer as the reference group.

All models were adjusted for age and BMI. As a sensitivity analysis, models were further adjusted for marital status, education levels, smoking status, physical activity, and antidepressant use in addition to age and BMI. These analyses were conducted using a complete-case approach, excluding observations with missing data on any of the covariates included in the models.

The corresponding numbers of participants and repeated observations are reported in Supplementary Table 3.

In secondary analyses restricted to women with breast cancer, models were further stratified by chemotherapy status to evaluate whether symptom trajectories differed according to chemotherapy exposure. Sensitivity analyses were conducted among premenopausal women with breast cancer to examine the additional effects of GnRH agonists by comparing ETx subgroups.

All statistical analyses were conducted using Stata version 17.0 (Stata Corp LP, College Station, TX) and R Statistical Software (version 4.2.2; R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as a two-sided P value less than 0.05.

Results

Study population

A total of 5734 women were included in the study, comprising 2631 premenopausal women and 3103 postmenopausal women. Among premenopausal participants, 466 were premenopausal controls, 679 were in late menopausal transition, 234 were breast cancer patients not receiving ETx, and 1252 were breast cancer patients receiving ETx. Among premenopausal breast cancer patients receiving ETx, 700 (55.91%) received tamoxifen monotherapy, 328 (26.2%) received tamoxifen plus a gonadotropin-releasing hormone (GnRH) agonist, and 221 (17.65%) received an AI plus a GnRH agonist.

Among postmenopausal women, 2006 were controls and 1097 were breast cancer patients. Of the postmenopausal breast cancer patients receiving ETx, 217 (28.0%) received tamoxifen and 533 (68.8%) received AI therapy. The majority of women in the non-ETx group had hormone receptor-negative disease, for whom adjuvant endocrine therapy is generally not indicated (Table 1).

Attrition analysis showed comparable baseline total and domain-specific MRS scores between participants who completed the 3-year follow-up and those who did not, despite differences in age and several demographic characteristics (Supplementary Table 1).Table 1Baseline characteristics of the study population (N = 5734).CharacteristicsPremenopausal women (N = 2631)Postmenopausal women (N = 3103)p-valueaWithout breast cancer (N = 1145)With Breast cancer (N = 1486)Without breast cancer (N = 2006)With Breast cancer (N = 1097)Premenopausal (N = 466)Transition late (N = 679)No ETx (N = 234)ETx (N = 1252)No ETx (N = 322)ETx (N = 775)Age, years, mean (SD)50.56 (1.67)51.53 (1.88)42.98 (6.23)44.42 (5.16)54.30 (2.62)56.37 (5.88)57.10 (6.11)<0.001Body mass index, kg/m2, mean (SD)22.22 (2.90)22.46 (3.16)22.98 (3.33)22.52 (3.33)22.49 (2.96)23.58 (3.14)24.02 (3.39)<0.001Marital status at diagnosis, n (%)<0.001 Unmarried22 (4.72)13 (1.91)29 (12.39)165 (13.18)33 (1.65)11 (3.42)28 (3.61) Married429 (92.06)629 (92.64)194 (82.91)1007 (80.43)1876 (93.52)265 (82.30)637 (82.19) Divorced/Separated/Widowed1 (0.21)8 (1.18)11 (4.70)78 (6.23)17 (0.85)46 (14.29)108 (13.94) MissingEducation, ≥university, n (%)402 (86.27)575 (84.68)162 (69.23)934 (74.60)1605 (80.01)128 (39.75)306 (39.48)<0.001Physical activity, n (%)<0.001 Inactive107 (22.96)172 (25.33)84 (35.90)407 (32.51)517 (25.77)84 (26.09)226 (29.16) Minimally/moderate active276 (59.23)384 (56.55)74 (31.62)406 (32.43)1082 (53.94)95 (29.50)235 (30.32) Active83 (17.81)123 (18.11)16 (6.84)130 (10.38)404 (20.14)50 (15.53)130 (16.77) Missing0 (0.00)0 (0.00)60 (25.64)309 (24.68)3 (0.15)93 (28.88)184 (23.74)Smoking status, n (%)<0.001 Never408 (87.55)590 (86.89)204 (87.18)1115 (89.06)1766 (88.04)298 (92.55)720 (92.90) Ever56 (12.02)87 (12.81)30 (12.82)137 (10.94)225 (11.22)22 (6.83)53 (6.84) Missing2 (0.43)2 (0.29)0015 (0.75)2 (0.62)2 (0.26)Antidepressant drug usage, n (%)7 (1.50)9 (1.33)0 (0.00)2 (0.16)26 (1.30)0 (0.00)1 (0.13)<0.001Stage, n (%)<0.001 DCIS25 (10.68)173 (13.63)27 (8.39)80 (10.32) Stage I84 (35.90)572 (45.69)117 (36.34)368 (47.48) Stage II37 (15.81)374 (29.87)44 (13.66)242 (31.23) Stage III8 (3.42)76 (6.07)8 (2.48)45 (5.81) Stage Ⅳ01 (0.08)01 (0.13) ypCRb53 (22.65)33 (2.64)98 (30.43)22 (2.84) Missing27 (11.54)23 (1.84)28 (8.70)17 (2.19)Treatment modality, n (%) Chemotherapy87 (37.18)329 (26.28)127 (39.44)213 (27.48)<0.001 Radiation therapy186 (79.49)912 (72.84)262 (81.37)623 (80.39)<0.001 Endocrine therapy, n (%)<0.001 None234 (100.0)322 (100.0) Tamoxifen700 (55.91)217 (28.00) Tamoxifen + GnRH328 (26.20)15 (1.94) Aromatase Inhibitors0533 (68.77) Aromatase Inhibitors + GnRH221 (17.65)7 (0.90) others3 (0.24)3 (0.39)Hormone receptor status, n (%)<0.001 Negative207 (88.46)26 (2.08)305 (94.72)19 (2.46) Positive22 (9.40)1223 (98.23)16 (4.97)755 (97.92) Unknown5 (2.14)3 (0.24)1 (0.31)1 (0.13)Values are presented as n (%) or mean (standard deviation); Others include medroxyprogesterone, megestrol, nomegestrol or only GnRH.aThe p-values represent overall comparisons across the seven analytic subgroups.bPatients achieving pathologic complete response (ypCR) after neoadjuvant therapy were presented as a separate category.

Changes in MRS by breast cancer status and endocrine therapy

At baseline, total MRS scores were higher in postmenopausal women than in premenopausal women overall. However, within the premenopausal group, baseline total MRS scores were similar between women with and without breast cancer (7.93 and 7.98 vs 7.68, respectively), and no statistically significant differences were observed according to endocrine therapy use at diagnosis among women with breast cancer. Similarly, among postmenopausal women, baseline total MRS scores did not substantially differ according to breast cancer status or endocrine therapy use at diagnosis (Supplementary Table 2).

During longitudinal follow-up, total MRS scores increased among premenopausal women with breast cancer who received ETx. Compared with premenopausal breast cancer patients not receiving ETx, ETx use was associated with greater increases in total MRS scores across follow-up. At 1 year after diagnosis, premenopausal breast cancer patients receiving ETx showed an increase of 3.30 points (95% confidence interval [CI], 2.04 to 4.58) relative to the reference group.

Among postmenopausal breast cancer patients, ETx use was also associated with increases in total MRS scores, however, the change for total MRS was modest and did not reach statistical significance (DiD = 1.17; 95% CI, −0.11 to 2.48). In contrast, domain-specific analyses demonstrated a significant excess increase in urogenital symptoms (DiD = 0.71; 95% CI, 0.26 to 1.16), whereas somatic/vasomotor and psychological domains did not show significant incremental differences (Table 2).Table 2Estimated differences in menopausal symptom changes between women with and without breast cancer (N = 5734).Premenopausal women (N = 2631)Postmenopausal women (N = 3103)Without breast cancer (N = 1145)With Breast cancer (N = 1486)Without breast cancer (N = 2006)With Breast cancer (N = 1097)Premenopausal (N = 466)Transition late (N = 679)No ETx (N = 234)ETx (N = 1252)No ETx (N = 322)ETx (N = 775)Total menopausal symptomsEstimated difference from baseline−0.05 (−1.26, 1.16)0.94 (0.01, 1.88)0.82 (0.11, 1.52)3.25 (2.96, 3.59)0.06 (−0.54, 0.67)0.41 (−0.27, 1.09)1.12 (0.67, 1.57)Difference in difference0.0 (Reference)0.99 (−0.39, 2.37)0.87 (−0.54, 2.27)3.30 (2.04, 4.58)0.11 (−1.24, 1.46)0.46 (−0.93, 1.85)1.17 (−0.11, 2.48)Somatic/vasomotor symptomsEstimated difference from baseline0.01 (−0.47, 0.49)0.51 (0.14, 0.88)0.02 (−0.27, 0.31)1.15 (1.02, 1.28)0.01 (−0.23, 0.25)−0.05 (−0.33, 0.23)0.24 (0.06, 0.43)Difference in difference0.0 (Reference)0.5 (−0.05, 1.05)0.01 (−0.56, 0.57)1.14 (0.64, 1.64)0 (−0.53, 0.54)−0.06 (−0.61, 0.5)0.23 (−0.28, 0.75)Psychological symptomsEstimated difference from baseline−0.07 (−0.6, 0.47)0.18 (−0.22, 0.59)0.4 (0.07, 0.73)0.98 (0.83, 1.12)−0.12 (−0.39, 0.14)0.08 (−0.23, 0.4)0.18 (−0.03, 0.39)Difference in difference0.0 (Reference)0.25 (−0.37, 0.86)0.47 (−0.16, 1.09)1.04 (0.49, 1.59)−0.06 (−0.65, 0.53)0.15 (−0.47, 0.76)0.24 (−0.33, 0.82)Urogenital symptomsEstimated difference from baseline0 (−0.42, 0.41)0.27 (−0.05, 0.59)0.41 (0.15, 0.67)1.19 (1.08, 1.31)0.19 (−0.01, 0.39)0.4 (0.15, 0.65)0.71 (0.54, 0.87)Difference in difference0.0 (Reference)0.27 (−0.21, 0.76)0.41 (−0.08, 0.91)1.19 (0.75, 1.62)0.19 (−0.27, 0.66)0.4 (−0.09, 0.89)0.71 (0.26, 1.16)ETx, endocrine therapy.Results were estimated using linear mixed-effects models and adjusted for age and body mass index.

In sensitivity analyses further adjusting for marital status, education levels, smoking status, physical activity, and antidepressant use, the overall direction of the estimates was generally consistent with the primary analyses, although several associations were attenuated (Supplementary Table 3).

Changes in MRS by endocrine therapy among breast cancer

Among women with breast cancer who did not receive chemotherapy, model-estimated total MRS scores differed according to menopausal status and ETx use (Fig. 2). Premenopausal women receiving ETx showed the greatest increase in total MRS scores, with significant DiD estimates of 2.64 (95% CI, 1.65–3.63), 2.89 (95% CI, 1.82–3.96), and 2.15 (95% CI, 0.90–3.40) at 1, 2, and 3 years, respectively, compared with premenopausal women who did not receive ETx (Table 3).

Among women with breast cancer who received chemotherapy, model-estimated total MRS scores also increased over time, with the largest increases observed among premenopausal women receiving ETx (Fig. 3). Compared with premenopausal women who did not receive ETx, the corresponding DiD estimates were 2.13 (95% CI, 0.78–3.49), 2.11 (95% CI, 0.70–3.52), and 3.02 (95% CI, 1.22–4.82) at 1, 2, and 3 years, respectively (Table 3).

Similar patterns were observed across all three MRS domains, regardless of chemotherapy status (Supplementary Table 4). Psychological symptoms increased during the first 1 to 2 years and attenuated by year 3. Among premenopausal women receiving chemotherapy, the larger year-3 DiD in total MRS reflected both persistent symptoms in the ETx group and a greater decline in the no-ETx group, and should not be interpreted as renewed symptom worsening in the ETx group.Fig.

2Model-estimated mean trajectories of total and domain-specific menopausal symptoms among women with breast cancer who did not receive chemotherapy.Table 3Estimated differences in menopausal symptom changes by endocrine therapy among women with breast cancer (N = 2583).Premenopausal women (N = 1, 486)Postmenopausal women (N = 1097)No ETx (N = 234)ETx (N = 1252)No ETx (N = 322)ETx (N = 775)Without chemotherapy (N=1827)N = 147N = 923N = 195N = 5621 year afterEstimated difference from baseline0.49 (−0.43, 1.41)3.13 (2.75, 3.51)0.1 (−0.75, 0.95)1.24 (0.69, 1.8) Difference in difference0.00 (Reference)2.64 (1.65, 3.63)−0.39 (−1.64, 0.86)0.75 (−0.32, 1.82)2 years afterEstimated difference from baseline−0.42 (−1.42, 0.58)2.47 (2.07, 2.87)−0.38 (−1.39, 0.63)0.85 (0.27, 1.43) Difference in difference0.00 (Reference)2.89 (1.82, 3.96)0.03 (−1.38, 1.45)1.27 (0.12, 2.41)3 years afterEstimated difference from baseline0.05 (−1.11, 1.21)2.2 (1.68, 2.73)−1.75 (−3.11, −0.4)0.83 (0.08, 1.57) Difference in difference0.00 (Reference)2.15 (0.9, 3.4)−1.8 (−3.58, −0.03)0.77 (−0.59, 2.14)With chemotherapy (N=756)N = 87N = 329N = 127N = 2131 year afterEstimated difference from baseline1.36 (0.15, 2.57)3.49 (2.87, 4.11)1 (−0.27, 2.27)1.3 (0.45, 2.14) Difference in difference0.00 (Reference)2.13 (0.78, 3.49)−0.36 (−2.1, 1.39)−0.06 (−1.53, 1.4)2 years afterEstimated difference from baseline0.95 (−0.32, 2.23)3.06 (2.43, 3.7)0.51 (−0.81, 1.82)1.47 (0.57, 2.38) Difference in difference0.00 (Reference)2.11 (0.7, 3.52)−0.45 (−2.26, 1.37)0.52 (−1.02, 2.07)3 years afterEstimated difference from baseline−0.59 (−2.23, 1.04)2.42 (1.61, 3.23)0.79 (−0.81, 2.39)0.68 (−0.56, 1.92) Difference in difference0.00 (Reference)3.02 (1.22, 4.82)1.38 (−0.89, 3.65)1.27 (−0.76, 3.3)ETx, endocrine therapy.Results were estimated using linear mixed-effects models and adjusted for age and body mass index.Fig. 3Model-estimated mean trajectories of total and domain-specific menopausal symptoms among women with breast cancer who received chemotherapy.

Changes in MRS according to endocrine therapy regimen among premenopausal women with breast cancer

Regimens incorporating GnRH agonists were associated with greater increases in total MRS scores compared with no endocrine therapy, regardless of chemotherapy exposure (Table 4). Among women who did not receive chemotherapy, treatment with tamoxifen plus GnRH agonist and AI plus GnRH agonist was associated with greater increases in total MRS scores compared with no ETx. At 2 years after diagnosis, the DiD was 2.89 points (95% CI, 1.61 to 4.18) for tamoxifen plus GnRH agonist and 4.74 points (95% CI, 3.29 to 6.20) for AI plus GnRH agonist.

At 3 years after diagnosis, the corresponding change estimates were 1.87 points (95% CI, 0.33 to 3.40) and 2.80 points (95% CI, 0.81 to 4.80), respectively (Table 4). Among women who received chemotherapy, treatment with tamoxifen plus GnRH agonist and AI plus GnRH agonist was also associated with significant increases in total MRS scores compared with no ETx, with DiD estimates of 2.47 points (95% CI, 0.28 to 4.66) and 3.93 points (95% CI, 1.65 to 6.21) at 3 years, respectively (Table 4). Tamoxifen monotherapy was also associated with sustained increases in total MRS scores.

Among women who did not receive chemotherapy, the DiD estimates were 2.09 points (95% CI, 1.03 to 3.15) at 1 year, 2.41 points (95% CI, 1.27 to 3.54) at 2 years, and 2.07 points (95% CI, 0.74 to 3.41) at 3 years. Among women who received chemotherapy, the 3-year DiD was 2.78 points (95% CI, 0.67 to 4.90). In direct comparisons between tamoxifen plus GnRH agonist and tamoxifen monotherapy, a significantly greater increase in total MRS scores was observed at 1 year among women who did not receive chemotherapy (adjusted difference in changes, 1.28 points; 95% CI, 0.37 to 2.19; P = 0.006).

However, the differences were not statistically significant at 2 or 3 years. Among women who received chemotherapy, no statistically significant differences between the two regimens were observed at any follow-up time point (Supplementary Table 5).Table 4Estimated differences in menopausal symptom changes by endocrine therapy types among premenopausal women (N = 1483).No ETxTamoxifen monotherapyTamoxifen +Aromatase inhibitor + GnRH agonist (N=221)(N = 234)(N = 700)GnRH agonist (N = 328)Without chemotherapy (N=1070)N=147N=574N=217N=1301 year afterEstimated difference from baseline0.3 (−0.65, 1.24)2.38 (1.89, 2.88)3.66 (2.89, 4.43)4.02 (3.02, 5.03) Difference in difference0.00 (Reference)2.09 (1.03, 3.15)3.37 (2.15, 4.58)3.72 (2.35, 5.1)2 years afterEstimated difference from baseline−0.77 (−1.8, 0.25)1.63 (1.11, 2.16)2.12 (1.32, 2.92)3.97 (2.92, 5.02) Difference in difference0.00 (Reference)2.41 (1.27, 3.54)2.89 (1.61, 4.18)4.74 (3.29, 6.2)3 years afterEstimated difference from baseline−0.49 (−1.68, 0.71)1.59 (0.91, 2.26)1.38 (0.37, 2.39)2.32 (0.7, 3.94) Difference in difference0.00 (Reference)2.07 (0.74, 3.41)1.87 (0.33, 3.4)2.8 (0.81, 4.8)With chemotherapy (N=416)N=87N=126N=111N=911 year afterEstimated difference from baseline1.22 (−0.04, 2.47)2.82 (1.77, 3.86)3.08 (1.98, 4.18)4.56 (3.35, 5.76) Difference in difference0.00 (Reference)1.6 (−0.02, 3.22)1.87 (0.21, 3.53)3.34 (1.61, 5.07)2 years afterEstimated difference from baseline0.66 (−0.67, 1.99)1.76 (0.7, 2.83)2.79 (1.69, 3.89)4.26 (2.99, 5.54) Difference in difference0.00 (Reference)1.1 (−0.57, 2.78)2.13 (0.43, 3.83)3.6 (1.79, 5.41)3 years afterEstimated difference from baseline−0.99 (−2.7, 0.71)1.79 (0.46, 3.12)1.48 (0.04, 2.92)2.94 (1.36, 4.51) Difference in difference0.00 (Reference)2.78 (0.67, 4.9)2.47 (0.28, 4.66)3.93 (1.65, 6.21)ETx, endocrine therapy; GnRH, gonadotropin-releasing hormone.Results were estimated using linear mixed-effects models and adjusted for age and body mass index.

In domain-specific analyses, somatic/vasomotor symptoms increased early after treatment initiation and remained elevated throughout follow-up. Psychological symptoms increased during the first 1 to 2 years and showed partial attenuation by year 3. In contrast, urogenital symptoms demonstrated a more sustained increase, with higher scores observed at 3 years (Supplementary Table 6).

Baseline antidepressant use was uncommon and was included in the sensitivity analysis. Because longitudinal use of antidepressants and other symptom-directed treatments was unavailable, their contribution to changes in psychological symptoms could not be directly evaluated.

Discussion

In this longitudinal cohort study of 5734 women, we compared 1-year changes in menopausal symptom trajectories between breast cancer patients receiving ETx and women without breast cancer. Using women without breast cancer, as a reference group, we quantified the excess symptom burden attributable to endocrine therapy beyond natural menopausal progression. The non-cancer cohort provided a reference for background menopausal symptom changes during the first year of follow-up, whereas persistence of symptoms through 3 years was evaluated within the breast cancer cohort.

We further examined 3-year symptom trajectories within the breast cancer cohort according to baseline menopausal status, ETx type, and the use of GnRH agonists. In this study, we have three key findings. First, premenopausal women at diagnosis receiving ETx experienced greater increases in menopausal symptoms at 1 year compared to women undergoing natural menopause.

Second, although the overall difference in total MRS score was not statistically significant among postmenopausal women, urogenital symptoms showed significantly greater increases compared with women without breast cancer, whereas the other domains did not show consistent between-group differences. Third, within the breast cancer cohort, premenopausal patients receiving regimens that included ovarian function suppression demonstrated the greatest and most persistent symptom burden over 3 years, particularly among those who also received chemotherapy.

In this study, women receiving ETx experienced greater increases in menopausal symptoms over time than women without breast cancer. Importantly, by incorporating women without breast cancer as a natural menopause comparator, we were able to distinguish treatment-related symptom exacerbation from background age-related changes. In our cohort, where baseline total scores were approximately 8 to 10 points, an absolute differences of 2 to 3 points corresponded to this relative threshold and may therefore represent a clinically meaningful change.

Although menopausal symptoms progressed over time in the general population, symptom trajectories among breast cancer patients with ETx showed more increases from baseline, suggesting that ETx contributes additional morbidity. Even in women who are already postmenopausal at treatment initiation, ETx was associated with significant increases in urogenital symptoms. This persistent symptom burden is biologically plausible.

Although ovarian estrogen production ceases after menopause, peripheral tissues continue to synthesize low levels of estrogen through aromatization of androgens [20]. In contrast, ETx suppresses this residual production, resulting in estradiol levels substantially lower than those observed in natural menopause [21]. Unlike the gradual hormonal decline during physiologic menopause, ETx induces a rapid and profound estrogen deprivation, which may limit adaptive mechanisms and contribute to persistent urogenital symptoms [22], [23].

In postmenopausal women, the symptom monitoring and counseling also needed before and during ETx.

Within the breast cancer cohort, longitudinal analyses over 3 years demonstrated sustained symptom trajectories. A previous 5-year longitudinal cohort study of Chinese breast cancer survivors reported no marked increase in the overall prevalence of severe menopausal symptoms [10]. In contrast, our analysis stratified participants according to baseline menopausal status and demonstrated that symptom trajectories differed substantially between premenopausal and postmenopausal women.

Premenopausal women experienced greater increases in total MRS scores than postmenopausal women, particularly in the somatic/vasomotor and psychological domains. Moreover, symptom elevations persisted through 3 years of follow-up. While some reports suggest that treatment-related menopausal symptoms may attenuate over time [24], our data indicate sustained excess symptom burden, especially among women receiving ovarian suppression.

Considering, younger women have consistently shown higher rates of early discontinuation [9], this prolonged trajectory has important clinical implications. For premenopausal women initiating ETx, clinicians should anticipate symptom escalation beyond natural menopausal progression and incorporate structured early monitoring and anticipatory counseling into care pathways.

Among premenopausal women, menopausal symptom burden was observed across endocrine therapy regimens. Tamoxifen monotherapy was associated with sustained increases in total MRS scores, indicating that symptom burden was not confined to regimens containing ovarian function suppression. Previous prospective studies have similarly reported persistent menopausal symptoms during tamoxifen treatment [25].

Regimens incorporating GnRH agonists, however, were associated with the greatest and most sustained increases in MRS scores over 3 years. The excess burden was observed both in women who received chemotherapy and in those who did not. Domain-specific analyses revealed distinct temporal patterns.

Vasomotor symptoms remained persistently elevated throughout follow-up, psychological symptoms were more prominent during the first 1 to 2 years, and urogenital symptoms became more evident over time, particularly by year 3. These findings indicate that ovarian suppression produces not only a higher overall symptom burden but also evolving domain-specific trajectories. These findings extend prior clinical trial data demonstrating increased menopausal toxicity with ovarian function suppression [26], [27], [28].

A cohort study reported higher MRS scores at 6 and 12 months in women receiving tamoxifen plus ovarian suppression compared with tamoxifen alone, with a later symptom plateau in the combination group [29]. Our 3-year longitudinal analysis demonstrates that excess symptom burden with ovarian suppression persists well beyond the early treatment period and remains detectable at the population level. The biological mechanism is consistent with profound estrogen deprivation.

In premenopausal women, ETx, particularly with ovarian suppression, induces a rapid decline in circulating estrogen levels that contrasts sharply with the gradual hormonal transition of natural menopause [30]. Acute estrogen withdrawal may drive early neurovascular and psychological symptoms [31], whereas progressive estrogen depletion and mucosal atrophy may contribute to the later emergence and persistence of urogenital symptoms [32]. This differential physiological response may explain the domain-specific temporal divergence observed in our study.

These findings have important implications for long-term ETx management. When considering ovarian function suppression, the magnitude and duration of expected toxicity should be explicitly integrated into shared decision-making, particularly for patients also receiving chemotherapy, and proactive adherence-support strategies should be implemented for high-risk patients.

Several limitations should be acknowledged. First, menopausal symptoms were assessed using self-reported MRS scores. Although self-reported outcomes are inherently subjective, validated patient-reported outcomes are the standard and recommended approach for menopausal symptom evaluation.

Second, the exact timing of ETx initiation was not available, and symptom trajectories were measured from breast cancer diagnosis. Variation in treatment sequencing may therefore have introduced heterogeneity in the timing of symptom assessment relative to ETx initiation. Information on ETx discontinuation, treatment adherence, and reasons for missed follow-up assessments could not be reliably ascertained from the available follow-up data.

Consequently, we were unable to evaluate whether worsening menopausal symptoms contributed to treatment discontinuation or non-adherence. Although mixed-effects models used all available repeated measurements and baseline symptom scores were similar according to 3-year follow-up status, informative loss to follow-up cannot be completely excluded. Baseline antidepressant use was included in the sensitivity analysis; however, longitudinal changes in antidepressant and other symptom-directed treatments were not available, and their contribution to changes in psychological symptoms during follow-up could not be directly evaluated.

Similarly, longitudinal information on local vaginal therapies and non-hormonal treatments for urogenital symptoms was not available, and their potential influence on urogenital symptom trajectories could not be evaluated. Third, participants were drawn from two cohorts. Although this design introduces potential differences between the BIG-S and KSHS cohorts, including participant selection, health-screening behavior, and the level of detail in menopausal classification, the inclusion of a non-cancer reference group provided a reference for background menopausal symptom changes during the first year of follow-up.

All analyses were adjusted for age and BMI to reduce baseline imbalance. Fourth, a formal minimal clinically important difference for the MRS has not been universally defined. Therefore, interpretation of absolute score changes should consider both statistical significance and relative percentage change from baseline.

However, our study has several strengths. First, it integrates a natural menopausal reference group, including women in late transition, allowing separation of treatment effects from physiologic aging. Second, the difference-in-differences analytic framework strengthens causal inference by accounting for background symptom progression.

Third, extended 3-year follow-up provides insight into symptom persistence beyond the initial treatment phase. Fourth, detailed stratification by endocrine regimen, including ovarian suppression, enables granular evaluation of toxicity profiles.

Conclusion

In conclusion, women receiving endocrine therapy for breast cancer experienced greater increases in menopausal symptoms during the first year of follow-up compared with women without breast cancer. Within the breast cancer cohort, menopausal symptom burden remained elevated over 3 years, particularly among premenopausal women receiving ovarian function suppression. These findings provide population-level evidence relevant to long-term endocrine therapy management and may inform individualized counseling regarding treatment-related morbidity.

CRediT authorship contribution statement

Se Kyung Lee: Writing – review & editing, Writing – original draft, Data curation, Conceptualization. Hye Rin Choi: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Conceptualization. Danbee Kang: Writing – review & editing, Writing – original draft, Methodology, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Yoonyoung Jang: Writing – review & editing, Data curation. Seungho Ryu: Writing – review & editing, Data curation. Jeong Eon Lee: Writing – review & editing, Data curation.

Seok Jin Nam: Writing – review & editing, Data curation. Seok Won Kim: Writing – review & editing, Data curation. Byung Joo Chae: Writing – review & editing, Data curation.

Jai Min Ryu: Writing – review & editing, Data curation. Yoosoo Chang: Writing – review & editing, Supervision, Funding acquisition, Data curation. Jonghan Yu: Writing – review & editing, Supervision, Conceptualization.

Ethics declaration

Written informed consent to take part in the study and to publish the article has been obtained from all participants or their legal representatives. The privacy rights of participants have been observed.

This study was performed in compliance with relevant laws, regulatory frameworks and guidelines where the research took place. This study was approved by the Samsung Medical Center, Kangbuk Samsung Hospital. (Approval No.

SMC-2019-08-009-020, KBSMC 2023-05-036)

Data availability

Data supporting the findings of this study are available from the corresponding author upon reasonable request.

Funding

This research was supported by the 10.13039/100006180Bio&Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (10.13039/501100014188MSIT) (No.RS-2024-00440881). This research was also supported by the National Institute of Health (NIH), South Korea (Project No. 2020ER710200, 2020ER710201, 2020ER710202, 2023ER060500, 2023ER060501, 2023ER060502, and 2026ER070500).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Related articles

PubMed

Factors Associated With Health Literacy Among Women in Menopause in Western Iran: A Cross-Sectional Study.

Background: There is a lack of information regarding health literacy in homemaker menopausal women in the western Iran. The aim of this study was to investigate factors associated with health literacy among women in menopause who are homemakers in western Iran. Methods: This cross-sectional study was conducted from February 5, 2025 to August 20, 2025, among 350 women in menopause attending comprehensive health service centers in Hamadan, western Iran. The participants were recruited via syste...

PLOS

Supporting engagement, behavior change, and equitable self-management in menstrual and menopausal health: A multi-agent digital-health framework

Menstrual-cycle–related conditions and, later, menopausal symptoms are highly prevalent across the female life course and impose a large, under-recognized burden on quality of life and work, yet digital tools to support self-management are fragmented, hard to sustain, and unevenly reached. We propose HERA (Health Engagement and Reflective Agent architecture), a conceptual, hypothesis-generating framework for supporting engagement, behavior change, and equitable self-management in menstrual and m...