Adjuvant chemotherapy prolongs disease-free and overall survival of patients with breast cancer. In premenopausal women, cytotoxic chemotherapy can induce temporary or permanent ovarian dysfunction 1 2 3 4 . The incidence of chemotherapy-induced amenorrhea is directely related to age and varies with the type of chemotherapeutic agent used 5 6 , as well as its dose and schedule 1 . Data on ovarian function are widely available for certain regimens, such as cyclophosphamide, methotrexate and 5-fluorouracil (CMF) polychemotherapy 6 7 and anthracycline-based treatments 8 9 , but fewer studies have been conducted on taxane-based regimens and they unfortunately show contradictory results 10 11 12 13 . Here, we report the results of a retrospective analysis evaluating the incidence of reversible amenorrhea in women with early-stage breast cancer undergoing adjuvant anthracycline-based chemotherapy with or without docetaxel.

The rate of chemotherapy-related amenorrhea varies according to the diagnostic criteria adopted and the duration of follow-up 1 , but inconsistencies exist in the way it is reported. Some authors report the incidence of amenorrhea immediately upon completion of chemotherapy, while others select various time points after the start and end of chemotherapy 1 5 . The time point most commonly encountered in the literature is 12 months after the end of chemotherapy 2 . Chemotherapy-related amenorrhea is generally linked to the patient's age as well as treatment protocol 1 2 7 8 9 11 12 13 . In our study, in women over 40 years of age, reversible amenorrhea was more frequently observed in patients treated with 3FEC/3D than 6FEC. Moreover, our data suggest that DFS was better in patients with persistent amenorrhea in the 3FEC/3D group.

Chemotherapy-related amenorrhea depends considerably on the regimen used. Cyclophosphamide is a cytotoxic agent that has been extensively investigated and is widely known to induce amenorrhea 6 7 . According to Goldhirsh 7 , 61 % of women under 40 years of age experienced amenorrhea after receiving CMF, while 95 % of women over 40 became amenorrheic after this regimen. Amenorrhea associated with anthracycline-based therapy nevertheless shows significant variation from study to study. Bines et al 1 found a 35 % amenorrhea rate with the 4-cycle AC regimen (adriamycin 60 mg/m², cyclophosphamide 600 mg/m², 4 courses every 3 weeks), Hortobagyi et al 8 59 % with the 6-cycle FAC regimen (5-fluorouracil 500 mg/m², doxorubicin 50 mg/m², cyclophosphamide 500 mg/m²), and Levine et al 9 51 % with the 6-cycle CEF regimen (cyclophosphamide 75 mg/m² orally on days 1 through 14, epirubicin 60 mg/m² IV on days 1 and 8, 5-fluorouracil 500 mg/m² on days 1 and 8). The results observed in our study with 6FEC 100 are not different from those observed in the literature 1 2 3 8 9 .

There are fewer published results on taxane regimens and the are somewhat contradictory 11 12 13 . Moreover, some of the literature is reported in abstract form 15 16 17 18 . In the BCIRG 001 study 15 16 18 , 6TAC was compared to 6FAC (TAC: docetaxel 75 mg/m², doxorubicin 50 mg/m², cyclophosphamide 500 mg/m²; FAC: 5-fluorouracil 500 mg/m², doxorubicin 50 mg/m², cyclophosphamide 500 mg/m²). TAC gave better disease-free and overall survival than FAC, but increased the rate of amenorrhea (66 % versus 54 %) (p = 0.008) by contrast, in our study, the 3 FEC/3D regimen induced less definitive amenorrhea than the 6FEC 100 regimen. Unfortunately, the method and time of evaluation of amenorrhea were not reported.

In the study by Fornier, 166 very young patients were reviewed. All patients were treated with AC (doxorubicin at a dose of 60 mg/m2 + cyclophosphamide at a dose of 600 mg/m2 for 4 cycles followed by a taxane). The majority of patients were given AC followed by paclitaxel at a dose of 175 mg/m2 for 4 cycles adminitered at 2- or 3-week intertreatment intervals.

Only 7 patients received docetaxel (100 mg/m2).

In this cohort, long-term amenorrhea was defined as the absence of menstruation ≥ 12 months after the completion of all chemotherapy. No hormone values were available and the conclusions of this study were that addition of a taxane did not appear to produce a higher rate of chemotherapy-related amenorrhea, compared to historical controls.

In the study by Davis, 159 premenopausal patients were reviewed. As initial chemotherapy, 102 women received AC (doxorubicin/cyclophosphamide), 39 received CMF (cyclophosphamide/methotrexate/5-fluorouracil) and 18 received CAF (cyclophosphamide/doxorubicin/5-fluorouracil).

Following the initial regimen, 53 patients received additional adjuvant chemotherapy, generally with a taxane for 12 weeks (paclitaxel in 32 patients and docetaxel in 19 patients).

The conclusions of this study were similar to those of Fornier. Sequential addition of taxanes did not appear to increase the risk of chemotherapy-induced amenorrhea, when added to non-taxane regimens. Moreover, the author did not find any impact of the type of initial chemotherapy administered. The definition of chemotherapy-related amenorrhea was the same as that used by Fornier and our team. As in Fornier's study, no hormone values were available.

More recently, Tham published a study involving 191 patients (158 patients ≤ 40 years of age at the start of chemotherapy). The patients received 4 cycles of AC alone or followed by a taxane (there was no stratification between paclitaxel and docetaxel).

The definition of chemotherapy-related amenorrhea was a little different in this study. Indeed, it was defined as cessation of menses within 1 year of starting chemotherapy and lasting ≥ 6 months.

As demonstrated in all these studies, age and type of chemotherapy regimen were independently related to the rate of chemotherapy-induced amenorrhea.

No hormone values were available.

In the subgroup of younger patients (≤ 40 years), addition of a taxane resulted in a higher incidence of chemotherapy-related amenorrhea (61 versus 44 %). In women over 40 years of age, amenorrhea rates were high in both the group of AC alone and the group of AC followed by a taxane (81 versus 85 %). No statistically significant difference was observed between the two groups.

However, in this study, the patients treated with a taxane received more chemotherapy cycles and this could have had an impact on the results.

In our study, women over 40 years of age were at much greater risk of developing definitive amenorrhea than those under 40, highlighting the importance of patient age 19 20 . Milan's regimen, using 6 CMF with or without doxorubicin, resulted in an amenorrhea rate of 4 % in women under 30 years of age, 50 % in women aged between 36 and 40 years, 86 % in women aged between 41 and 45 years, and 100 % in women over 45 years of age 20 . Similar variations were observed with epirubicin-containing regimens 8 9 and taxanes 11 12 13 . Our study has the advantage of providing hormone profiles of 2/3 of the patients, but the limitations are similar to those encoutered, in other studies, i.e. the retrospective design and the small number of patients.

Whether or not induction of amenorrhea by cytotoxic chemotherapy is a prognostic factor in the treatment of premenopausal women is still controversial. A positive impact on DFS has been found by some 21 22 23 24 25 26 27 , but not confirmed by others 17 28 . Del Mastro et al 21 conducted a review of 13 studies involving 3929 patients undergoing CMF-based regimens, with follow-up ranging from 3 to 20 years. A statistically significant association was found between the development of chemotherapy-related amenorrhea and DFS. In the majority of cases, overall survival was found to be associated with amenorrhea (in 3 out of 5 studies reviewed). In a study recently published by Parulekar 21 , similar results were observed with intensive CEF (cyclophosphamide, epirubicin, fluorouracil) therapy, which induced a higher rate of amenorrhea than the classic CMF protocol, but overall survival was also better.

Another important question relates to the duration of amenorrhea. In Trial VI (study by the International Breast Cancer Study Group (IBCSG)), cessation of menses, even for a limited time period, appeared to be beneficial, especially in patients with ER-positive breast tumors. In this study 29 , however, the greatest effect was observed in patients receiving suboptimal treatment with only three initial CMF courses. In the PACS 01 trial 10 , a survival advantage in favor of the 3 FEC/3D arm was observed only for women aged over 50 years, but not for the younger population. The reason for this is unclear but the impact of reversible amenorrhea in this context needs to be investigated further, since our small retrospective analysis suggests that amenorrhea was correlated with DFS in the 3FEC/3D group.

In our population, 73 % of patients received tamoxifen in the 6FEC arm, and 74 % in the 3FEC/3D arm. The ovarian stimulation effect of tamoxifen in premenopausal women is well documented 30 , but its effects on the duration of chemotherapy-induced amenorrhea are not well defined or understood. In a study by Tham 13 , tamoxifen was not shown to influence the incidence of amenorrhea, but the impact of tamoxifen on the hypothalamo-pituitary axis is complex and sequential administration of tamoxifen after chemotherapy does not simplify the data. It is therefore important to monitor a variety of hormonal tests to assess the consequences of sequential administration of antiestrogen treatment after chemotherapy.

Gonadal toxicity resulting from chemotherapy was first reported almost 60 years ago and the gonadotoxic effects of chemotherapeutic agents are well documented. Although the prevailing mechanisms are not fully understood, three clinical studies have suggested that chemotherapy induces apoptotic changes in pregranulosa cells that subsequently develop into follicles 31 . However, these findings have not yet been confirmed in human studies 3 . Chemotherapy also provokes a reduction in the number of available oocytes. These changes are similar to those observed in natural postmenopausal ovaries. This cytotoxic damage appears to be progressive and irreversible in the ovary, as jump cells are limited in number and cannot be regenerated 32 33 .

Our retrospective study confirms that chemotherapy-related amenorrhea is highly complex, requiring in-depth analysis, especially with the advent of new chemotherapeutic agents. A number of questions need to be elucidated in the coming years. Is achieving amenorrhea a desirable side effect? Is temporary amenorrhea as effective as definitive amenorrhea? What are the implications of endocrine treatment for patients who recover ovarian function? Is it necessary to propose ovarian function suppression associated with tamoxifen or aromatase inhibitors? Some of these questions are currently being addressed in large randomized trials (PERCHE (stopped), SOFT, TEXT). Besides the prognostic potential of reversible amenorrhea and therapeutic impact of the endocrine approach, other issues also need clarification. What are the implications of reversible amenorrhea on both sexual function and preservation of fertility potential? All these quesions are often a major concern for young premenopausal breast cancer patients. On the basis of our findings, we believe it is crucial to carefully monitor ovarian function after chemotherapy for at least 12 months following the end of treatment, with separate analysis of results in women with hormone-sensitive disease. The impact of reversible amenorrhea on prognosis needs to be investigated prospectively, and ovarian function followed clinically, as well as by means of biochemical testing (large randomized trials).

ER, estrogen receptor; PR, progesterone receptor; N, no; Y, yes

HR has been a consultant with Sanofi-Aventis in the past five years. The rest of the authors have no competing interests to declare.

MB was responsible for the 2^nd Belgian study, as well as drafting the manuscript. FD was responsible for the French data and made an important contribution to the analysis and interpretation of data. NM was responsible for the collection of all data and was involved in the analysis and interpretation. AV was responsible for one Belgian center (Cliniques Ste Elisabeth) and was involved in revising the manuscript. PP performed the statistical analysis. HR was the principal investigator of the PACS 01 study and was involved in revising the content of the manuscript. JD was involved in the initial drafting and revision of the manuscript. MS was involved in the design of the PACS 01 study and critically revising the manuscript. JK made an important contribution to the analysis and interpretation of data. JPM was involved in drafting and critically revising the manuscript.

All authors read and approved the final manuscript.