Intravenous cyclophosphamide remains an important therapy for patients with severe systemic lupus erythematosus—including lupus nephritis, which primarily affects women in their reproductive years. As prognosis improves, the chronic toxicity of this therapy assumes greater importance. This article reviews cyclophosphamide use, its effect on gonadal function, and protection of gonadal reserve during therapy. Egg, embryo, or gonadal tissue cryopreservation and alternative therapeutic strategies are considered.
Intravenous cyclophosphamide (IVC) remains an important therapy for patients with severe systemic lupus erythematosus (SLE) including lupus nephritis (LN). Given the striking female predominance in SLE (9:1 in reproductive years) and greater risk of nephritis with younger age at onset, women with LN in their reproductive years comprise the largest group of rheumatic disease patients receiving this therapy. As prognosis in SLE improves, the chronic toxicity of this therapy assumes greater importance. Among the devastating complications of IVC therapy for young patients is the high rate of gonadal failure. The risk of premature ovarian failure with sterility, menopausal symptoms, and increased long-term risks of osteoporosis and coronary artery disease may delay physicians and young women with SLE from undertaking IVC therapy, despite the risks of poor long-term renal outcome.
Cyclophosphamide use in LN
Despite widespread use of IVC therapy for proliferative LN, few studies have formally evaluated the ovarian toxicity of this regimen in humans without malignancy. The National Institutes of Health (NIH) IVC regimen evolved to include 6 monthly doses of IVC followed by quarterly doses to complete 2 years; patients with inadequate response could “recycle” for additional monthly dosing as needed. The ovarian toxicity of this regimen has been reported in several retrospective studies. One was a retrospective, nested cohort study of 39 women with either LN or neuropsychiatric lupus receiving either short (7 doses) or long (at least 15 doses) courses of pulse IVC as compared with women receiving pulse intravenous methylprednisolone. Subjects were followed for up to 4 years and those with cessation of menses had a full gynecologic evaluation. As shown in Table 1 , there was an increased risk of sustained amenorrhea in women who were older at the time of treatment and in those receiving a higher cumulative dose of IVC. No patients treated with methylprednisolone had amenorrhea. This study is limited by its small sample size, lack of information regarding use of gonadotropin-releasing hormone agonist (GnRHa), and relatively short duration of follow-up. Ioannidis and colleagues found that 21 out of 67 (31%) receiving pulse IVC had sustained amenorrhea and, in women greater than 32-years of age, the risk of amenorrhea increased rapidly and linearly with the adjusted cumulative dose of IVC. The dose at which 50% of the women age 32 or more developed amenorrhea was about 8gm/m 2 , and the dose at which 90% had amenorrhea was 12gm/m 2 . They also noted that there was an increased prevalence of premature amenorrhea in women with longer prior SLE disease duration (over 5 years), anti-Ro antibodies, and anti-U1RNP antibodies. Mok and colleagues demonstrated a similar trend regarding increasing age and risk of ovarian failure, but also noted that women with IVC-induced amenorrhea had significantly fewer severe flares of SLE during the 5-year follow-up period as compared with those who continued to menstruate. While no studies were designed to assess fecundity after IVC treatment, Wang and colleagues observed that 14 of the 23 (60%) IVC-treated women who desired conception after cessation of treatment were able to conceive, resulting in 20 live births and 2 abortions. Unfortunately, the current data does not demonstrate an absolute threshold dose of IVC by age to maintain ovarian function, nor does it provide data on fecundity in women who do maintain menstrual function after IVC treatment.
|Age||All Subjects||Short-course IVC||Long-course IVC|
|<25||2/16 (12)||0/4 (0)||2/12 (17)|
|26–30||4/15 (27)||1/8 (12)||3/7 (43)|
|>31||5/8 (62)||1/4 (25)||4/4 (100)|
|All ages||11/39 (28)||2/16 (12)||9/23 (39)|
Effect of cyclophosphamide on ovarian function
Traditionally it was believed that women are born with their lifetime supply of oocytes, with approximately 2 million resting oocytes present at birth. At menarche, there are approximately 300,000 to 400,000 oocytes, and there was thought to be a steady decline in the number of oocytes, both via ovulation and atresia, throughout a woman’s reproductive life. In a series of studies in mice, Tilley and colleagues recently challenged the dogma that there is a fixed and nonrenewing pool of oocytes, demonstrating a return of fertility after bone-marrow transplantation in mice treated with chemotherapy. These studies suggest that a bone marrow transplant may restore lost or damaged germ cells in the ovary, though further research needs to be completed to test this hypothesis and assess these results in a human population. Neither rat nor human granulosa cells metabolize IVC in vitro under varying concentrations of luteinizing hormone. Whereas previous research suggested that IVC exerts its greatest toxicity to primordial follicles in the rat and human ovary, more recently researchers have found greater toxicity to growing or antral follicles in rat models. In a study with immature rats treated with IVC (100 mg/kg) as compared with control rats, the number of granulosa cells decreased, the mean follicular diameter decreased, and the granulosa nuclear size increased. In another study with mature cycling rats, there was a significant reduction in the number of ovarian follicles (especially medium and large follicles) after a 21-day course of IVC (5 mg/kg/d).
Cyclophosphamide Effects on Human Gonadal Function
Chemotherapy can affect a woman’s reproductive function, though the exact mechanism is unclear. One theory is that chemotherapy causes injury to rapidly dividing granulosa cells that normally provide hormonal support to developing follicles and oocytes, and that damage to these cells causes loss of ovarian function. Another theory is that oocytes, both in the resting and developing stages, are damaged by chemotherapy, perhaps because of their unique situation in which the chromosomes are arrested in meiosis I and perhaps more vulnerable to damage. The degree of damage to the ovary by chemotherapy depends on several factors, including the agent used, cumulative dose, and the age of the patient. Unfortunately, to date, nearly all of the prior studies related to chemotherapy and ovarian function use premature ovarian failure (POF) as the primary outcome. One of the largest studies looks at POF in childhood cancer survivors. This study involves a retrospective cohort of 5-year survivors of childhood cancer diagnosed before the age of 21 who were given a survey about late adverse outcomes, including a menstrual history. Of a total of 3390 survivors studied, 215 developed POF (incidence of 6.7%) with a slightly older mean age of diagnosis in the POF group versus the nonaffected survivors (9.8 ± 6 years vs 8.3 ± 6 years). Multivariate logistic regression found that the independent risk factors included older age at diagnosis, exposure of the ovaries to radiotherapy, and exposure to IVC and procarbazine. The risk of POF with IVC was only significant in older girls (age 13–20) as compared with the cohort treated before the age of 12.
Ovarian pathology in prepubertal girls receiving IVC in combination with other chemotherapeutic regimens reveals ovarian tissue depleted of ova at any stage. In contrast, studies evaluating ovarian outcome following treatment of childhood disorders employing IVC as a single chemotherapeutic agent suggest maintained ovarian function with high rates of normal pubertal development and onset of menses. However, the primary concern for many patients is the effect of chemotherapy on future fertility potential—there have been no studies to date on risk of infertility or subfertility after chemotherapy in women who continue to have menstrual cycles.
Traditionally, prepubertal children have been thought to be at relatively low risk of gonadal toxicity from IVC. Unfortunately, boys receiving IVC treatment before puberty are not protected from gonadal dysfunction. Prepubertal age has been classically characterized as a quiet period. However, during this period, there is active proliferation of Sertoli and Leydig cells in the testes, which may account for the damage from IVC. Unlike females, the risk of infertility after IVC in boys is independent of age and pubertal stage. Cyclophosphamide is among the most damaging chemotherapeutic agents impacting the testes; prolonged or permanent azoospermia has been seen in boys who received IVC.
Recently, 248 adult male long-term survivors of childhood cancer were assessed for IVC gonadal toxicity. Approximately 70% of male patients who had received less than 7.5 g/m 2 (median 4.1 g/m 2 ) of IVC regained fertility, but only 10% recovered when doses exceeded 7.5 g/m 2 . Almost all survivors who received IVC of 10 g/m 2 or more were at risk for severe spermatogenic dysfunction. Testosterone levels were slightly lower than the controls; however, only 7% of the survivors had low testosterone values and all patients had achieved adult secondary sex characteristics. In a group of 33 young, male survivors of childhood cancer, IVC treatment caused severe gonadal failure with reduced testicular size, low sperm counts, and impairment of Leydig cell function. Testicular biopsy specimens from 19 of these patients showed germinal aplasia in all cases.
In a pilot study of 15 males receiving IVC, 5 patients who were randomized to receive depot testosterone therapy showed higher sperm counts and testosterone levels after chemotherapy than the remaining 10 patients. Given the lack of research that supports the use of a concurrent protective therapy in males, young men should continue to be counseled about gamete storage before IVC treatment.
Techniques to store gametes before chemotherapy are both more readily available and more technically successful for spermatozoa than for oocytes. Unfortunately, prevention of sterility by sperm banking is not possible in prepubertal boys, as active spermatogenesis does not occur at this age. Artificial procedures including electroejaculation methods may be considered. New techniques, including intrauterine insemination, intracytoplasmic sperm injection, and testicular sperm extraction have improved fertility outcomes for men with low sperm counts or motility defects, both of which can be caused by IVC.
Protecting ovarian reserve during IVC therapy
Use of GnRHa to Protect Ovarian Function?
Strategies to prevent ovarian toxicity from chemotherapeutic agents focus on preventing ovulation, decreasing ovarian metabolic activity and blood flow, hence IVC dose to the ovary. GnRHa act to suppress pituitary release of gonadotropins, therefore creating a temporary “prepubertal” state. In theory, this prepubertal state could help protect ovarian function during IVC therapy. Proposed mechanisms of action include centrally mediated suppression of gonadotropins, direct suppression of gonadotropin receptors on ovaries, or reduction of biologic activity of gonadotropins.
Data from animal models support this premise. Ataya and colleagues found that rats treated with IVC and luteinizing hormone-releasing hormone agonist (LHRHa) had an increased number of small follicles and overall follicles as compared with rats treated with IVC alone. They theorize that treatment with LHRHa inhibits the recruitment of small follicles into the pool of medium or large follicles that are more susceptible to damage by IVC, thus minimizing the number of follicles damaged by chemotherapy. In another study, female rats treated with both IVC and LHRHa increased the future pregnancy rate as compared with rats treated with IVC alone (9:10 vs 4:11, P <.05). Of course, there are differences between rats and humans that make it difficult to extrapolate these findings to humans. For example, it is possible that human ovaries have a different susceptibility to damage from chemotherapy and radiation as compared with rats. In addition, rats have a very different reproductive cycle compared with women. Therefore, studies in nonhuman primates and humans are crucial to confirm the efficiency of GnRHa before widespread use is recommended.
The majority of human studies that evaluate the efficacy of GnRHa with respect to protecting ovarian reserve are nonrandomized observational studies with premature ovarian failure as the primary outcome. A recent meta-analysis concluded that the use of a GnRHa during chemotherapy was associated with a 68% increase in the rate of preserved ovarian function compared with women not receiving a GnRHa (summary relative risk = 1.68, 95% CI 1.34–2.1). However, only 2 of 9 studies included in the meta-analysis were randomized-controlled trials and these two trials included only 37 patients collectively. Also, the authors note that included studies have varying intensities and types of chemotherapy, making comparisons between studies difficult. There are very few reported studies of GnRHa use in women with rheumatologic disease. One was a prospective, nonrandomized trial involving women receiving IVC for LN and a depot GnRHa was associated with a significant reduction of POF; 5% in the GnRHa-treated group compared with 30% of controls. The study included add-back estradiol therapy, demonstrating that the benefit was not the result of a hypoestrogenemia. Recently, Blumenfeld and Eckman reported in a prospective nonrandomized study that less than 7% of 125 young women exposed to gonadotoxic chemotherapy for malignant or nonmalignant diseases while receiving GnRHa therapy developed POF. However, the use of GnRHa to prevent POF was shown ineffective in a recent cancer trial. Differences in age distribution and the doses of IVC employed may explain the widely differing results observed. Given the conflicting results regarding efficacy of GnRHa at preventing diminished ovarian reserve, it is difficult to advise patients about the use of this treatment. Large randomized controlled trials are currently underway and will, hopefully, provide definitive data about this treatment.
Potential Adverse Effects of Gonadotropin-releasing Hormone Suppression During Cyclophosphamide
In addition to the routine risks to young women receiving GnRHa, 2 potentially disastrous risks are important in the LN population. If GnRHa are started in the follicular phase of the cycle, the initial ovarian upregulation within the first days following therapy increases estrogen levels. The majority of LN patients are hypertensive and nephrotic and a substantial minority may have antiphospholipid antibodies—all increased risks for potential thrombotic events. Physicians should evaluate the individual patient’s risk for thrombosis to assess the need to treat with anticoagulation until downregulation is documented. The risk for pregnancy with failure to employ effective contraception is a concern, especially during the first month of therapy and if ovarian suppression is interrupted during therapy. The potential for a multiple pregnancy is increased if conception occurs during initial upregulation following dosing.
General risks from GnRHa include menopausal symptoms such as hot flashes, decreased libido, emotional lability, headaches, acne, decreased breast size, or vaginal dryness. Ovarian hyperstimulation with GnRHa has been described: a woman receiving the medication during hemodialysis and a woman receiving subcutaneous GnRHa. The hyperstimulation was attributed to possible intermittent dosing and has not been described with intramuscular depot administration. In each case, hyperstimulation resolved with continued therapy. Dual energy x-ray absorptiometry performed to assess bone mineral density in women receiving GnRHa for treatment of endometriosis has raised concern for bone loss.
GnRHa is associated with an increase in blood cholesterol and triglycerides in 9% and 12% of women receiving this therapy, however, HDL/LDL ratios are not altered. The effects of GnRHa are short term and resolve in 4 to 12 weeks after stopping the drug. In addition to potential central nervous system effects of high-dose corticosteroids that may be given at the same time, GnRHa has potential risk of mood alteration or depression.