Chemotherapy-Related Amenorrhea and Quality of Life Among Premenopausal Women With Breast Cancer

Key Points Question Which factors are associated with chemotherapy-related amenorrhea, and what is its association with quality of life among breast cancer survivors? Findings In this cohort study of 1636 premenopausal women, 416 of 729 women (57.1%) included in the quality of life analysis reported persistent amenorrhea after chemotherapy, although late recovery patterns were observed. Amenorrhea was more frequent among older women and those receiving adjuvant tamoxifen and was significantly associated with worse insomnia, systemic therapy–related adverse effects, and sexual dysfunction. Meaning These findings can facilitate the treatment of premenopausal women with early breast cancer undergoing chemotherapy and experiencing chemotherapy-related amenorrhea, by informing risk communication, personalized counseling, and early supportive care referrals.


Introduction
Breast cancer is the most commonly diagnosed tumor in women, and approximately 20% of women with breast cancer are younger than 50 years at diagnosis. 1 Compared with older women with breast cancer, younger survivors have higher risk of cancer-related symptoms and quality of life (QOL) deterioration 2,3 and report worse depressive symptoms, higher level of stress, more weight gain, and greater difficulty managing the adverse effects of therapy. 4In particular, treatment-related symptoms linked to the menopausal transition (ie, vasomotor symptoms and sexual problems) represent an important source of distress during and after treatment, highlighting a need to monitor and address survivorship-related problems that are specific to this population. 5,6emotherapy-related amenorrhea (CRA) is among the possible adverse effects induced by treatment for breast cancer. 7,8There is no consensus definition of CRA across studies.According to the International Consensus Guidelines for Breast Cancer in Young Women, 9 CRA can be empirically defined as absence of menses for at least 2 years, a persistent postmenopausal profile, and vaginal ultrasonography indicating that the ovaries are no longer functioning.4][15] Rates of CRA also may vary according to type of chemotherapy regimen 16 and duration. 14Although some retrospective studies 13,[17][18][19] did not show a significant difference in the incidence of CRA with the addition of taxanes to an anthracycline-based regimen, some more recent and larger prospective cohort studies suggested increased rates of CRA among women receiving combination of a taxane and an anthracycline, 11,20 irrespective of type of taxane (ie, docetaxel or paclitaxel). 21CRA may also have downstream outcomes on QOL.CRA is indeed often accompanied by symptoms associated with premature menopause and diminished ovarian reserve, including hot flashes, night sweats, vaginal dryness, concerns about infertility, and decreased sexual interest and enjoyment. 4,22ere are still unanswered questions about CRA.There is a lack of longitudinal data regarding long-term trajectories of CRA and mid-term to long-term recovery patterns of menses, making it difficult for clinicians to give accurate estimations of the likelihood of recovery over time according to age range at diagnosis of breast cancer.Studies looking at treatment-related determinants of CRA in the era of currently standard combination chemotherapies including an anthracycline and a taxane or comparing different taxanes are also lacking.In addition, the association of persistent CRA with long-term patient-reported QOL has not been comprehensively explored.Using a large cohort with longitudinal patient characteristics and patient-reported outcomes, the present study aimed to (1)   assess the prevalence of CRA and associated factors among women receiving standard anthracycline-based or taxane-based chemotherapy or a combination of both an anthracycline and a taxane and (2) evaluate the association of CRA with long-term QOL up to 4 years after breast cancer diagnosis.

JAMA Network Open | Oncology
Chemotherapy-Related Amenorrhea and Quality of Life Among Women With Breast Cancer

Data Source
This study used data from CANTO (Cancer Toxicities Study), a French prospective, longitudinal, multicenter cohort study of women with stage I to III breast cancer. 23Briefly, patients are assessed at the time of breast cancer diagnosis (ie, before the start of any treatment), and then regular follow-up evaluations occur annually for up to 5 years.For the present analysis, relevant data were available at diagnosis and then at year 1 (Y1), year 2 (Y2), and year 4 (Y4) after diagnosis.A comprehensive assessment of clinical (including gynecological history, comorbidities, and behavioral factors), socioeconomic, tumor, and treatment data are available, as well as patient-reported outcomes assessing physical, psychological, and social domains of QOL.Patients who experience breast cancer recurrence (other than local), second cancers, or death exit the CANTO study at the time of the event.Enrollment covered 2012 to 2017.This study was approved by the French ethics committee and health authority.All patients provided written informed consent.All data were already available at the time of analysis, and no additional collection was performed.The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for cohort studies. 24

Study Cohort
We included women younger than 50 years 25 with known premenopausal status at diagnosis, who received adjuvant and/or neoadjuvant chemotherapy.Women who (1) received ovarian function suppression using luteinizing hormone-releasing hormone agonists as part of adjuvant endocrine therapy, (2) underwent hysterectomy or oophorectomy either before or after diagnosis, and (3) had missing information about posttreatment menses status during all follow-up visits (Y1 and Y2 and Y4 after diagnosis) were excluded from the analysis.For the QOL analysis, women who had posttreatment menses data at all follow-up visit (Y1, Y2, and Y4 after diagnosis) were included (eFigure in Supplement 1).

Variables of Interest Primary Analysis
For the primary analysis, the outcome of interest was CRA, based on patient-reported menses status, dichotomized as the presence vs absence of menses.Exposure variables (all assessed at diagnosis) included age, comorbidities, health behaviors, sociodemographic characteristics, and tumor and treatment characteristics listed in Table 1.

QOL Analysis
For the secondary analyses, we focused on patient-reported outcomes assessed longitudinally at diagnosis (baseline), Y1, Y2, and Y4 (primary outcome) using the validated European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire (QLQ)-C30 [26][27][28] and QLQ-BR23.Worse QOL is reflected by a lower score for functional domains and by a higher score for symptom domains (score range, 0-100).Overall QOL was evaluated by the QLQ-C30 summary score (ie, obtained using the mean of 13 QLQ-C30 domains, excluding global health and financial impact, and scored as a functional domain). 29e exposure variable was CRA, dichotomized as never recovery of menses vs recovery at any time after diagnosis (either by Y1 [early recovery], or Y2 [late recovery], or Y4 [very late recovery]).
Variables previously listed as exposure variables for the primary analysis served as covariates in the QOL analysis.

JAMA Network Open | Oncology
Chemotherapy-Related Amenorrhea and Quality of Life Among Women With Breast Cancer

Statistical Analysis Primary Analysis
Descriptive statistics summarized cohort characteristics and rates of CRA over time, overall, and by age at breast cancer diagnosis.We then fit models to correlated responses by generalized estimating equations to explore the association of baseline exposure variables with the outcome of reporting CRA over time-that is, absence of menses at Y1, Y2, or Y4 after diagnosis.An exchangeable working correlation matrix structure was specified to account for within-patient correlations.The variance function for the binomial distribution and the logit-link function were used (binary response data).
The specific contribution of chemotherapy type (ie, docetaxel vs paclitaxel) to the odds of CRA was investigated among patients receiving a taxane.f All patients in this analysis were prescribed adjuvant tamoxifen as adjuvant endocrine therapy.Women not receiving any endocrine therapy either had hormone receptor-negative breast cancer or they refused or had contraindications to adjuvant tamoxifen.

QOL Analysis
Descriptive statistics summarized menses status after diagnosis (never recovery of menses vs recovery by Y1, Y2, or Y4 after diagnosis).Mean estimates and changes in QOL outcomes (as a continuous variable) with respective 95% CIs were obtained using multivariable random-effect mixed models, adjusted by menses status group (vs recovery by Y1, Y2, or Y4 as a reference), time, menses status-by-time interactions, and covariates.Separate models were constructed for distinct QOL outcomes, each also including the respective QOL domain assessed at baseline.

Cohort Characteristics
The analytic cohort included 1636 women for the primary analysis, of whom 1497 had menses data available at the Y1 evaluation, 1323 at Y2, and 906 at Y4 (eFigure in Supplement 1 Compared with women who recovered menses by Y4, those who did not recover scored similarly for overall QOL (QLQ-C30 summary score mean difference, −1.9 points; 95% CI, −4.7 to 0.9 points; P = .18);however, they reported worse symptoms and more impaired functionality across some specific domains (Figure 2).In particular, never recovering menses was associated with more insomnia (mean difference, 9.9 points; 95% CI, 3.2 to 16.5 points; P = .004),more systemic therapyrelated adverse effects (defined as having a number of chemotherapy-related adverse effects that include dry mouth, dysgeusia, hot flashes, headaches, and alopecia; mean difference, 3.0 points; 95% CI, 0.2 to 5.8 points; P = .04),and worse sexual functioning (mean difference, −9.2 points (95%

JAMA Network Open | Oncology
Chemotherapy-Related Amenorrhea and Quality of Life Among Women With Breast Cancer Overall, we observed high rates of amenorrhea in this premenopausal cohort (83.0% at Y1, 72.5% at Y2, and 66.1% at Y4). Persistent amenorrhea was reported by 57.1% of women, with great age-related heterogeneity.Age represents the primary factor associated with the risk of premature ovarian failure and CRA. 12 In the younger age groups of our analysis, many patients resumed menses by Y2, confirming prior results. 14,30,31Nevertheless, our longer-term data are also important and somehow reassuring.Even when a recovery is not observed by Y2, a large proportion of younger women seem to recover menses very late, between 2 and 4 years after diagnosis.These observations have several clinical implications.Before treatment inception, premenopausal women should be Point estimates represent model-based mean estimates with respective 95% CI error bars obtained using multivariable random-effect mixed models (adjusted by menses status group, time, menses status-by-time interactions, and covariates).A 95% CI not crossing the vertical zero line indicates a statistically significant between-group difference at year 4.

JAMA Network Open | Oncology
Chemotherapy-Related Amenorrhea and Quality of Life Among Women With Breast Cancer made aware of the risks associated with chemotherapy-related premature ovarian failure and persistent CRA and receive systematic oncofertility counseling. 32Among women younger than 39 years at the time of breast cancer diagnosis, 36% have not yet completed their family plans, 14 and many of them may wish to have children after completion of treatment.Persistent or permanent CRA will necessarily cause delays or force these plans to change.In addition, in light of data showing possible late recoveries, contraceptive options should also be clearly discussed.Previous studies reported suboptimal contraceptive information for women not desiring pregnancy, 33 with high rates of unintended pregnancies among breast cancer survivors. 34Dedicated gynecological counseling may help patients who have an inaccurate perception of infertility due to previous exposure to chemotherapy and long-term absence of menses. 33,35late menses recovery pattern was also observed in older age groups in our analysis, with 8.3% of women older than 45 years reporting very late recovery between Y2 and Y4.This aspect can be challenging when it comes to the choice of the optimal adjuvant endocrine treatment.The absence of menses after completion of chemotherapy should not be used as a proxy for permanent transition to menopause, because it does not represent a reliable surrogate of gonadotoxicity. 9Adjuvant endocrine treatment choices should be based on a more thorough and comprehensive evaluation, combining absence of menses, assessments of circulating hormone levels, and gynecological ultrasonographic imaging. 9is study also helps identifying treatment-related determinants of CRA.Confirming some previous studies, we report higher rates of CRA among women treated with a combination of a taxane plus an anthracycline chemotherapy 11 and/or with tamoxifen, 14 and we found no difference in the odds of CRA according to type of taxane used, consistent with a previous study. 21[38][39] Furthermore, our study can inform personalized care pathways targeting patients at higher risk of QOL deterioration associated with a permanent menopausal transition.Women who never recovered menses had worse QOL, including more insomnia and other systemic therapy adverse effects at Y4 after diagnosis, similar to previous studies reporting difficulty sleeping and worse physical symptoms among women with CRA. 40Occurrence or exacerbation of menopausal symptoms caused by chemotherapy or endocrine therapy were suggested as a key precipitating factor for sleep problems and QOL deterioration. 2,3,41In particular, vasomotor symptoms have been positively correlated with sleep complaints in breast cancer survivors. 42,43In addition, we found substantial differences in patient-reported sexual function among women with persistent CRA compared with those who had recovered menses, as previously reported. 22Other analyses performed in the CANTO cohort also highlighted that sexual problems already may have been present before breast cancer treatment and seemed to either persist or worsen following primary therapy. 44

Limitations
This study is one the largest evaluating CRA prospectively in a contemporary cohort of premenopausal women with early breast cancer and also focusing on long-term QOL using internationally validated self-reported instruments, including a baseline point before treatment.
However, some limitations should be acknowledged.We did not have access to the exact information on chemotherapy dose and scheduling.Therapeutic algorithms have recently evolved, including adding neoadjuvant carboplatin and immunotherapy for triple-negative breast cancer, which is common among younger women. 45Similarly, the impact of novel targeted agents, such as poly-ADP ribose polymerase inhibitors, pertuzumab, trastuzumab emtansine, or cyclin-dependent kinase 4 or 6 inhibitors, in the adjuvant setting could not be explored.Levels of antimullerian hormone or other markers of treatment-induced gonadotoxicity were not available, and CRA was self-reported.
Ovarian function may resume while receiving tamoxifen even in the absence of menses, and this is difficult to determine without biochemical correlation. 46Adherence to endocrine therapy may also

JAMA Network Open | Oncology
Chemotherapy-Related Amenorrhea and Quality of Life Among Women With Breast Cancer  47,48 The study may not fully capture other potential confounding variables, including gender and ethnicity (the latter cannot be collected by French law).Generalizability of findings to patient populations other than ours (ie, premenopausal women aged <50 years, of whom most were aged 40-50 years, not receiving pharmacological ovarian function suppression) may be limited.As common in longitudinal studies, some dropout was observed, reducing the sample at later time points (particularly for the QOL analysis) and possibly introducing some selection bias toward patients experiencing a more favorable long-term disease course.However, the remaining cohort characteristics were balanced, and overall QOL scores were similar between the 2 groups.Given the hypothesis-driven approach of these analyses, we did not correct for multiple testing.Furthermore, some of the statistical differences in QOL that we found may correspond only to a small-to-moderate clinical between-group difference, based on previous studies using European Organization for Research and Treatment of Cancer QLQ-C30. 49

Conclusions
In conclusion, our findings can be used to inform clinical practice in several ways.Risk and duration of CRA, including potential late resumption of menses and its downstream implications for QOL, should be approached using a coordinated biopsychosocial model addressing multiple dimensions of physical, psychological, and social health.Proactive management of premenopausal women with early breast cancer undergoing chemotherapy should also include adapted strategies for risk communication, as well as personalized counseling and early supportive care referrals.

Figure 2 .
Figure 2. Longitudinal Evolution of Quality of Life From Diagnosis Through Year 1, Year 2, and Year 4 After Diagnosis

Figure 3 .
Figure 3. Mean Differences Between Quality of Life Scores at Year 4 Among Patients Reporting No Menses Recovery vs Recovery at Any Time After Diagnosis

Table 1 .
Patient and Tumor Characteristics at Baseline

Table 1 .
Patient and Tumor Characteristics at Baseline (continued) a Missing data were not included in calculations of percentages.b Body mass index is calculated as weight in kilograms divided by height in meters squared.c Women with missing data (n = 185) were considered without children.Association of this variable with chemotherapy-related amenorrhea was not tested in subsequent models.d Overall, 1571 of 1636 women (96.0%) received cyclophosphamide.The median number of chemotherapy cycles received was 4 in the taxanebased chemotherapy group, 5 in the anthracyclinebased chemotherapy group, and 6 in the combination chemotherapy group.e Overall, 1238 of 1636 women (75.7%) received docetaxel, 192 of 1636 (11.7%) received paclitaxel, and 70 of 1636 (4.3%) received a sequence of the 2 agents.

Table 2 .
There was no difference in the odds of CRA according to type of taxane received (adjusted OR for docetaxel vs paclitaxel, 1.00 [95% CI, 0.67-1.49];1430patients).Chemotherapy-Related Amenorrhea and Quality of Life Among Women With Breast Cancer Y4.A total of 416 patients (57.1%) reported having never recovered menses, whereas 313 (42.9%) had done so by the end of 4 years of follow-up.Specifically, across all age groups, 101 of 729 women (13.9%) had early menses recovery within Y1 after diagnosis, 133 of 628 women (21.2%) did so between Y1 and Y2, and 79 of 495 women (16.0%) recovered menses between Y2 and Y4.We then looked at the distribution of menses recovery rates by age group.Menses recovery by QOL Analysis: Association of CRA With Long-Term QOLAmong 729 women with menses data available at all time points, cohort characteristics were consistent with the primary analysis (eTable 2 in Supplement 1).Characteristics of women not included in this analysis are available in eTables 3 and 4 in Supplement 1.Of note, 172 of 907 women not included had experienced recurrence, second cancer, or death and exited the CANTO study by JAMA Network Open | Oncology JAMA Network Open.2023;6(11):e2343910. doi:10.1001/jamanetworkopen.2023.43910(Reprinted) November 16, 2023 6/15 Downloaded from jamanetwork.comby guest on 12/08/2023

Table 2 .
Factors Associated With Chemotherapy-Related Amenorrhea in a Multivariable Generalized Estimating Equation Model Figure 1.Menses Recovery Rate by Age Group Among 729 Women With Menses Status Available at All Time Points a Adjusted for all factors in the table.
CI, −14.3 to −4.1 points; P < .001) at Y4. Between-group estimate differences for the remaining domains at Y4 are reported in Figure3.
JAMA Network Open.2023;6(11):e2343910. doi:10.1001/jamanetworkopen.2023.43910(Reprinted) November 16, 2023 10/15 Downloaded from jamanetwork.comby guest on 12/08/2023 Pistilli B, Ferreira AR, Combarel D, et al.Longitudinal evaluation of serum assessed non-adherence to tamoxifen (TAM) among premenopausal patients (pts) in the prospective multicenter CANTO cohort.Ann Oncol.2020;31:S308.doi:10.1016/j.annonc.2020.08.289 49.Cocks K, King MT, Velikova G, Martyn St-James M, Fayers PM, Brown JM.Evidence-based guidelines for determination of sample size and interpretation of the European Organisation for the Research and Treatment of Cancer Quality of Life Questionnaire Core 30.J Clin Oncol.2011;29(1):89-96.doi:10.1200/JCO.2010.28.0107Factors Associated With Chemotherapy-Related Amenorrhea (CRA) in a Multivariable Generalized Estimating Equation Model Among Women Receiving Docetaxel or Paclitaxel (N=1430) eTable 2. Patient and Tumor Characteristics at Baseline Among Women With Menses Data Available at All Time Points and Included in the Quality of Life Analysis (N=729) eTable 3. Patient and Tumor Characteristics at Baseline Among Women Excluded From the Quality of Life Analysis (N=907) eTable 4. Description of Quality of Life Domain Scores Among Women With Menses Data Available at All Time Points and Included in the Quality of Life Analysis (n=729) and Those Excluded From the Quality of Life Analysis eTable 5. Menses Recovery Rate by Age Group Among Women With Menses Status Available at All Time Points 48.