A comprehensive review of the literature was done to assess the effect of adjuvant chemotherapy for operable breast cancer on ovarian function, fertility, and birth defects. Data were limited. Cyclophosphamide, an alkylating agent, is the major cause of amenorrhea, which is due to primary ovarian failure. Ovarian dysfunction is related to age, dose, and duration of treatment. In women less than 35, pregnancy following adjuvant chemotherapy is possible. However, data are limited regarding the impact of subsequent pregnancy on the results of breast cancer. There appears to be no increased risk of teratogenesis in offspring exposed to chemotherapy after the first trimester of pregnancy. Prospective data on women who have subsequent pregnancies and their offspring are very limited. Formation of a registry for long-term follow-up of young women detailing reproductive potential and follow-up of offspring is needed. [Monogr Natl Cancer Inst 16:125-129, 1994]

In 1948, the first description of cytotoxic-induced gonadal dysfunction was reported as caused by the alkylating agent, nitrogen mustard (1). Over the ensuing years, reports of the effects of chemotherapy on gonadal function have appeared, most frequently in the context of treatment for hodgkin's disease, childhood leukemia, and testicular cancer, since these diseases commonly affect young people, with the majority of reports focusing on testicular rather than ovarian function. Testicular biopsy and semen analysis have provided an accurate estimation of male reproductive potential. However, the relative paucity of data regarding female reproductive potential has been due to the lack of a reliable and accessible animal model and the relative inaccessibility of the ovary to biopsy. The recording of menstrual history is a crude measure of ovarian function. Amenorrhea is not synonymous with sterility, just as menstruation does not mean that a woman is fertile. Direct comparisons can not be made between male and female reproductive physiology, since spermatogenesis is continuous from the time of puberty and the full complement of primary oocytes is present at birth. The systemic consequences of the underlying disease may also affect males and females differently, as in the case of hodgkin's disease in which pretreatment testicular dysfunction has been reported in as many as 30% of young males, whereas ovarian function is not clearly compromised (2). In breast cancer, the disease state in itself does not appear to impact on gonadal function. Endocrine consequences of chemotherapy also differ, depending on the cytotoxic agents, treatment dose, schedule, and duration as well as the patient's age. Extrapolations cannot be made from the experience generated from the treatment of other malignancies that affect young individuals with the effects of currently used adjuvant chemotherapy regimens administered to young women with operable breast cancer. Since more young women are offered adjuvant treatment at earlier stages of disease, the treating physicians need to be able to make accurate assessments of risks and benefits of treatment. The complex physiologic and psychosocial issues related to ovarian function in premenopausal women need to be better defined.

Normal Ovarian Physiology A review of normal ovarian physiology is essential to an understanding of the impact of chemotherapy on ovarian function (3,4). Approximately 3 weeks after conception, the primordial germ cells arise from the endodermal yolk sac and begin migrating to the developing ovaries, where the ovaries begin to form during the fifth week of intrauterine life. During early fetal life, a complex series of cellular transformations change the primordial germ cells to oogonia and then to primary oocytes. From late in fetal life until puberty, the oocytes remain in the prophase stage of the first meiotic division. At puberty, the oocytes enlarge and the surrounding follicular layer changes to form primary follicles. From puberty until menopause, follicular growth occurs as a continuous process with ovulation in a cyclical fashion. The surrounding granulosa cells proliferate, follicular fluid accumulates, and the ovum completes its first meiotic division and is arrested in metaphase of the second meiotic division to become a secondary oocyte. This secondary, or Graafian follicle, continues to enlarge until the time of ovulation.

A single dominant follicle is selected by days 5-7 in the ovarian menstrual cycle, with ovulation of a fertilizable egg at day 14. All other follicles that were recruited in that cycle undergo atresia. As the dominant follicle emerges, there is increasing production of estrogen that stimulates the mid-cycle luteinizing hormone (LH) and follicle-stimulating hormone (FSH) surge. The abrupt rise in LH leads to meiotic maturation of the oocyte before ovulation, release of the oocyte by the ovulating process, and formation of the corpus luteum by luteinization of the granulosa and theca cells of the dominant follicle.

No primary oocytes form after birth in contrast to the continuous production of spermatocytes in the male after puberty. At its peak, that occurs at approximately 7 months postconception, the number of oocytes is estimated at 5-7 million. Thereafter the number of oocytes decrease steadily by atresia. At birth there are 2 million follicles, 200,000 remain at puberty and roughly 400 are left at menopause. Generally, only one ovum is liberated during each menstrual cycle, and since the reproductive life of a woman lasts about 30-40 years, only 300-400 oocytes mature and are extruded by ovulation. All other follicles undergo atresia.

Monthly ovulation is regulated by an intricate feedback mechanism between the hypothalamic-pituitary-ovarian axis. The periodicity of ovulation requires cyclic secretion of the gonadotropins, FSH and LH, which in turn requires pulsatile hypothalamic stimulation via gonadotropin-releasing hormone (GnRH) (4). However, gonadotropin cyclicity is predominantly influenced by ovarian estrogen and progesterone production. The normal menstrual cycle has three phases of steroid production, the follicular phase, ovulation, and the luteal phase. Serum levels of LH, FSH, and estradiol also have characteristic patterns during each stage of reproductive life: prepubertal, normally menstruating, and postmenopausal.

Fertility experts use both FSH and chronologic age to predict human reproductive potential and success rate of new methods used in infertility management. Ovarian reserve is partially defined by the FSH concentration on day 3 of the menstrual cycle. Briefly, in the normal situation, FSH produced by the pituitary signals the ovary to develop follicles. In turn the ovary produces estradiol, causing feedback inhibition of the pituitary that results in an FSH concentration less than 10 mIU/ml. In ovarian failure or menopause, the ovary is depleted of all follicles, and there is no feedback inhibition of the pituitary, resulting in FSH levels greater than 40 mIU/ml. In menstruating women with diminished ovulation reserve, the day 3 FSH levels range is between 18 and 30 mIU/ml.

Chemotherapy and Ovarian Function Combination chemotherapy has resulted in long-term cures in the treatment of Hodgkin's disease, leukemia, lymphoma, testicular cancer, ovarian cancer, and breast cancer. Regimens have been designed with special care to avoid overlapping acute toxic effects. However, the long-term consequences on reproductive potential were not anticipated, most notably, when the MOPP (mechlorethamine, vincristine, procarbazine, and prednisone) regimen was used for Hodgkin's disease (2,5). The major ovarian toxicity of cell-cycle specific agents appears to be directed at the process of follicular growth and maturation. Alkylating agent cause cytotoxicity independent of the cell cycle and affect the resting oocyte. Ovarian biopsies in patients undergoing cyclophosphamide-based treatment reveal complete absence of ova or small numbers of inactive ova with fibrosis and no evidence of follicular maturation (2,6,7). In many cases, the histology is similar to postmenopausal ovaries. With the destruction of ovarian follicles, there is a loss of steroid-producing cells. The reduction of the serum estrogen levels results in the interruption of the normal cycling of gonadotropins. Ovarian dysfunction is manifested by irregular menses, with eventual amenorrhea and development of menopausal symptoms.

Adjuvant Chemotherapy and Ovarian Function For the purposes of this review, the chemotherapy agents that are frequently used in the adjuvant setting will be discussed in detail. These include cyclophosphamide (C), methotrexate (M), and fluorouracil (F), CMF and its variants, and doxorubicin (A), commonly combined as CAF, or AC. The alkylating agent, cyclophosphamide, is most frequently at the center of adjuvant regimens and appears to be the major cause of ovarian failure (8). Amenorrhea and infertility have not been associated with the cell-cycle specific antimetabolites, methotrexate and fluorouracil, when given in the dose ranges used for adjuvant treatment. The data regarding the anthracycline, doxorubicin, is even more limited and its affects on infertility has not been well defined. When administered, however, as part of the ABVD regimen (Adriamycin, bleomycin, vinblastine, dacarbazine) for Hodgkin's disease, there is less gonadal dysfunction than with the MOPP (mechlorethamine, vincristine, procarbazine, and prednisone) regimen (9). Clearly, the impact of the adjuvant regimens, as they are currently administered, warrants further study.

Endocrine hormone profiles obtained from premenopausal treated with adjuvant chemotherapy who develop drug-induced amenorrhea are consistent with primary ovarian failure (7,10-14). Adjuvant regimens studied include "standard" CMF with oral cyclophosphamide, FAC, continuous oral cyclophosphamide with and without mitomycin, fluorouracil, and L-phenylalanine mustard. Estradiol and progesterone levels remain persistently low and cease to show their normal cyclic changes, While the pituitary gonadatropins, FSH and LH, are elevated to postmenopausal levels. Serum dehydroepiandrosterone and prolactin levels are not affected, which is consistent with intact adrenal and pituitary functions, respectively. Similar findings have been reported in Hodgkin's disease after MOPP chemotherapy and after cyclophosphamide for glomerulonephritis (2,5,6). Coincident with the endocrine profile of treatment-induced ovarian failure, patients may develop menopausal symptoms, such as hot flashes, decreased libido, changes in sleep patterns, irritability, vaginal discharge and/or dryness and dyspareunia.

The risk for developing treatment-induced ovarian failure is dependent on patient age and the total cumulative dose of drug. Generally, younger patients are able to tolerate larger cumulative doses of chemotherapy before developing amenorrhea and have a greater likelihood of resumption of menses after therapy is discontinued. Older patients who already have a depleted number of follicles at the time of exposure are more susceptible to the effects of chemotherapy.

Koyama et al. (7) demonstrated the relationship of age and dose among 18 premenopausal patients who received single-agent daily oral cyclophosphamide at 100 mg/day as adjuvant chemotherapy for operable breast cancer. All of 13 patients who were at least 40 years old developed permanent amenorrhea, and the average cumulative dose at the onset of amenorrhea was 5.2g (range, 1.4-8.4g). Four to five patients in their 30s developed amenorrhea that was permanent in three, and the mean dose of cyclophosphamide at the onset of amenorrhea was 9.3g (range, 7.0-11.1g). Five patients in their 20s (mean age, 25.4 years) received cyclophosphamide with mitomycin and/or chest wall irradiation, and three developed amenorrhea; the mean dose of cyclophosphamide at onset was 20.4g (range, 14-24.5g).

Dnistrian et al. (14) evaluated 26 patients, treated with "standard" CMF +/- levamisole adjuvant chemotherapy, which included 14 premenopausal women, of whom 12 developed amenorrhea. The onset of amenorrhea was accompanied by a hormonal profile similar to that as seen in the untreated postmenopausal or surgically castrated control patients. There was an inverse relationship between age and the duration of treatment required to induce ovarian suppression with this regime. Patients who were at least 40 years old developed amenorrhea within 2 to 4 months of treatment, while younger patients required larger cumulative doses of cytotoxic drugs to induce ovarian dysfunction. in fact, two patients under age 30 showed no evidence of ovarian suppression after 24 cycles (2 years) of therapy.

Bonadonna et al. (15) reported the incidence of CMF-induced amenorrhea among 549 menstruating women. Amenorrhea occurred in 54% who were less than 40 years old and was reversible in 23%. In women greater than 40, the incidence of amenorrhea was 96% and reversible in 4%. There was no difference in the duration of treatment, 6 vs 12 months, on the incidence of amenorrhea in women greater than 40 years. However, among younger women, the patient's age at treatment was the most important factor in determining both amenorrhea incidence and reversibility.

Pregnancy After Adjuvant Chemotherapy Approximately 50% of the patients younger than age 35 resume normal menses after completion of cytotoxic chemotherapy, and thus, may be capable of becoming pregnant and having offspring (8,15,16). Sutton et al. (17) retrospectively reviewed the records of 227 patients who received adjuvant chemotherapy with FAC who were 35 years old or younger to determine the frequency and effect of pregnancy on disease outcome. Of the 128 patients whose menstrual histories were known, 59% continued to menstruate after chemotherapy, 32% experienced temporary amenorrhea, and 9% experienced permanent amenorrhea. Of the 25 patients who became pregnant, 64% continued to menstruate regularly during and after chemotherapy, 32% experienced temporary amenorrhea, and this information was not available for the remaining 4%. In this series, 33 pregnancies occurred in 25 patients, whose median age was 28 years (range 22-33 years) and who had received chemotherapy for a median of 7 months (range, 2-24 months). The median interval between the last dose of chemotherapy and pregnancy was 12 months (range, 0-87 months). ten pregnancies were terminated on the physician's advice or the patient's wishes (including four terminations for conceptions during chemotherapy), two patients had spontaneous abortions, 19 pregnancies resulted in full-term deliveries, and two patients were pregnant at the time of the report. There were no fetal malformations in the 19 offspring. These data illustrate that a significant number of patients retained reproductive potential after chemotherapy. These authors suggest that pregnancy did not impact adversely on the clinical course of breast cancer, nor did the offspring of these patients demonstrate teratogenic effects. Due to limited data, the authors were unable to assess the relationship of estrogen receptor status of the tumor and pregnancy. However, the reproductive potential of the patient population was not assessed because there were no data regarding the use of contraceptives or attempts at conception.

Reports of pregnancy after cytotoxic chemotherapy have suggested that there is no increased incidence of fetal wastage and malformations over that of the general population. However, this must be interpreted with caution just as the suggestion that disease outcome is not affected by subsequent pregnancy (18-23). Data are available for a small number of patients, a population that may not adequately represent the population at large. Moreover, since the occurrence of breast cancer during pregnancy is rare, it takes many years to accrue a series, during which time both adjuvant treatment regimens and staging systems used change. Since pregnancy is not usually case-coded as a disease in the record room, the physician's memory is often the basis for retrospective reviews. Data generated from accrual-based surveys may bear the consequence of faulty or selective memory on the part of the physicians. Physician's warnings to patients with good and poor prognosis might have been different and resulted in differing attempts at reproduction and pregnancy rates in these groups. The effects of the underlying disease and pregnancy results are unknown in patients who died or were lost to follow-up. Clearly, information needs to be generated for large populations of women with similar disease and treatment profiles with regard to menstrual history, hormone levels, contraception and attempts at reproduction, and subsequent pregnancy and results.

Pregnancy During Adjuvant Chemotherapy and Fetal Results Data are limited regarding the teratogenic and mutagenic potential of cytotoxic chemotherapy delivered during chemotherapy. Review of the literature reveals anecdotal experiences, numerous reports with single-agent chemotherapy, many of which employed doses that would be considered suboptimal by today's standards. Moreover, there are very limited data regarding the long-term follow-up of offspring who were exposed to chemotherapy in utero or conceived after chemotherapy. This would reflect germ-cell exposure, with possible direct germ-cell mutagenicity and/or toxicity. Beyond obvious malformations present at birth, special attention needs to be directed to the possible subtle as well as the delayed consequences that include impaired physical growth, intellectual and neurological function, gonadal function and reproductive capacity, transplacental carcinogenesis, transplacental mutagenesis of germ-line tissue, and secondary carcinogenesis. The karyotypic analysis of offspring might provide useful information. Recessive mutations may take several generations to appear. Thus, the risk of the gene pool has not been assessed.

The probability of teratogenesis during pregnancy is influenced by the trimester of exposure, the agents given, drug dose, and schedule (23-28). During the first trimester, the period of organogenesis, exposure may result in congenital malformations and/or fetal demise. During the second and third trimesters, chemotherapy may affect fetal growth and functional development, especially that of the brain, but rarely causes congenital malformations. In a comprehensive review of chemotherapy given during pregnancy, Doll et al. (24) summarized the reported cases of fetal malformations associated with treatment during the first versus the second and third trimesters of pregnancy. During the first trimester, there were 24 malformations among 139 (17%) exposed to single agents versus seven malformations among 45 (16%) exposed to combination chemotherapy. However, when chemotherapy with folate antagonists and/or concurrent radiation was excluded, the incidence for single-agent associated malformations declined to 6%. In addition, four of the seven malformations seen with combinations and one of the 24 associated with single agents were exposed to procarbazine, as part of MOPP. Of note, the incidence for major congenital malformations in the general population is 3% of all births (29). In contrast, chemotherapy given during the second and third trimesters has not been associated with an increased risk of teratogenesis (24). Antimetabolites, especially methotrexate, should be avoided throughout pregnancy. Alkylating agents appear to be less teratogenic , with six fetal malformations reported among the 50 patients exposed when given as a single agent. The effects of anthracyclines, in particular, doxorubicin, has not been clearly defined, although one case of fatal myocardial necrosis was reported with daunorubicin (30).

Mulvihill et al. (31) reported the results of a retrospective review of pregnancy outcome by postal survey among female patients treated on Cancer and Leukemia Group B trials for a variety of advanced malignancies that included four patients with breast cancer. there were 133 pregnancies among 66 patients; 43 pregnancies ended before therapy (group 1); therapy was given at conception or during pregnancy in 32 (group 2), and 58 pregnancies occurred after chemotherapy (group 3). The frequencies of abnormal results were similar in groups 1 and 2, eight of 37 versus five of 22, respectively. There were eight normal offspring among 10 first trimester exposures. In group 3, conceptions occurred at a mean of 27 months after treatment (range, 2-104 months). there was an unexpected increase in low birth weight, stillborn, and premature termination of pregnancy, with no excess of congenital anomalies in the first year after chemotherapy. This reflects dysfunction in the milieu required to maintain pregnancy rather than damage to oocytes.

There are very little data regarding the effects of adjuvant chemotherapy administered during pregnancy. It has been demonstrated that terminating the pregnancy does not impact on the prognosis (20). The risks and benefits of treatment must be considered on an individual basis, with attention to the patient's personal desires and the urgency to begin chemotherapy and/or radiation therapy, which can cause potential fetal harm. if feasible, and if the patient wishes to continue the pregnancy, chemotherapy should be delayed until after the first trimester in view of the increased risk of teratogenicity during the first trimester. Chemotherapy may be cautiously administered during the second and third trimester. However , there may be an increased risk of intrauterine growth retardation, premature labor and birth, and spontaneous abortion. Clearly, decision-making is complex in this setting. If chemotherapy is given during pregnancy, delivery should be planned when the maternal hematopoietic profile is optimal, and the fetus' blood counts should be monitored. Breast feeding is contraindicated since drugs may reach significant levels in maternal milk.

Possible Benefit of Premature Menopause on Breast Cancer Prognosis In young women with breast cancer, disease-free and overall survival depend on the disease stage at presentation and the efficacy of adjuvant treatment. The effect of currently used adjuvant chemotherapy regimens involving higher doses for shorter duration have not been adequately assessed in terms of impact on ovarian function and reproductive potential. The issue of the possible beneficial effect of chemically induced ovarian failure on long-term survival in premenopausal patients remains controversial. The recent meta-analysis of adjuvant therapy trials indicated that this may be desirable (32).Current adjuvant protocols in premenopausal patients are addressed in this issue with the randomized inclusion of LH-RH agonists and/or tamoxifen. Thus, the proposed strategies to protect ovarian tissue from cytotoxic-induced damage, including short term use of LH-RH agonists, may not be relevant in the setting of breast cancer (33,34). This subject is discussed in greater detail elsewhere in this monograph.

Infertility Management: Are There Options? Advances in infertility management may provide alternative methods of reproduction, including embryo cryopreservation before to cancer therapy and ovary donation (35). There is no option similar to sperm banking available for ova storage. The effect of hormone replacement for the maintenance of pregnancy as part of infertility management regimens is not known. These emerging technologies present new medical, psychosocial, ethical, and legal dilemmas.

All women who have undergone treatment for breast cancer and wish to bear offspring must face the possibility that they may not live to see their children grow up. Traditionally, physicians have recommended that a woman wait 2-5 years after treatment before considering pregnancy to allow for recurrence among the majority of women who are so destined. For many women with breast cancer who already delayed childbearing, this warning might preclude realization of their reproductive potential. This recommendation might not apply equally to all women without consideration of initial stage of disease, prognostic variables, and systemic adjuvant treatment. Several series have reported that subsequent pregnancy does not impact unfavorably on disease outcome (18-23). However, typically follow-up is short. We are not certain whether the pregnancy-associated hormonal changes will have an undesirable effect in the population with hormone-receptor positive disease, a group who would otherwise be considered to have a more favorable prognosis. As mentioned earlier, data are limited regarding the relationship between pregnancy and estrogen receptor status of the tumor.

As the incidence of breast cancer in young women continues to rise, data regarding the impact of adjuvant treatment on reproductive potential are needed. Hopefully, improvements in early detection and adjuvant treatment will translate into increased disease-free survival. Formation of a registry for the long-term follow-up of young women detailing reproduction, disease outcome, and the outcome of offspring is needed.

K.B. Green, Department of Hematology/Oncology, New York Hospital-Cornell Medical Center. Correspondence to: Bonnie S. Reichman, M.D., Strang-Cornell Breast Center, 428 E. 72nd St., New York, NY 10021.