Intermittent pulse cyclophosphamide is widely used in renal [6, 14-17] and major extrarenal complications of lupus erythematosus [18-20]. Because of its more favorable balance of efficacy and toxicity, intermittent pulse cyclophosphamide therapy is considered an acceptable alternative to daily oral cyclophosphamide for the treatment of proliferative lupus nephritis. In addition to lupus, pulse cyclophosphamide therapy has been used with variable results for the treatment of various immune-mediated rheumatic [21-28], renal [29-31], neurologic [32, 33], and hematologic diseases [34] or their complications (reviewed in [20]).

Because pulse cyclophosphamide therapy is used for women of child-bearing age who have immunologically mediated diseases, accurate information about ovarian toxicity rates with this therapy is critical. Even though several studies exist of the gonadal toxicity of alkylating agents in patients with malignancy and autoimmune diseases, the doses and treatment schedules in these studies [3-13] differ from those of pulse cyclophosphamide as currently used in immunologically mediated disorders [14-34]. Moreover, the concomitant use of adjunctive oncologic therapies that may be toxic to the ovaries also limits the applicability of some of these data [9-13] to patients with autoimmune diseases.

To address some of these issues more directly, we evaluated the risk for amenorrhea in women with systemic lupus erythematosus treated with pulses of cyclophosphamide according to duration of treatment (number of doses) and the age at the initiation of therapy. Our data suggest that intermittent pulse cyclophosphamide therapy is associated with secondary amenorrhea and that duration of therapy and age are independent risk factors. The toxicity rates provided by our study should be useful to physicians and patients before they decide whether to use pulse cyclophosphamide therapy.

Methods

Top Methods Results Discussion Author & Article Info References

Selection of Patients

For the purpose of this study, we defined a short course of cyclophosphamide (Cytoxan, Bristol Myers Oncology, Princeton, New Jersey) as 7 monthly pulses of intravenous cyclophosphamide (short-CY) and a long course as 15 or more pulses (long-CY). Criteria for eligibility included women who were 40 years old or younger. Patients with other causes of secondary amenorrhea (including end-stage renal disease) were excluded from the analysis. Amenorrhea was defined as lack of menses for at least 4 months. Sustained amenorrhea was defined as amenorrhea not resolving within 12 months after cessation of pulse cyclophosphamide therapy.

Patients included in this analysis participated in two different prospective therapeutic trials for lupus nephritis at the National Institutes of Health from 1973 to 1990 [6, 17] and in a retrospective study for neuropsychiatric lupus [19]. For the first protocol, patients with severe proliferative nephritis [defined as impaired renal function alone, very active renal histology, or both] were randomly assigned to one of the following three treatment groups: 1) methylprednisolone pulses at 1.0 g/m² of body surface area, monthly for 6 months [total of nine doses, n = 25]; 2) a short course of cyclophosphamide pulses (short-CY) at 0.5 to 1.0 g/m², monthly for 6 months only [total of seven doses, n = 20]; or 3) a long course of cyclophosphamide pulses (long-CY) given monthly for 6 months followed by quarterly pulses for an additional 2 years (total of 15 doses, n = 20). Sixteen patients from the methylprednisolone (Medrol, The Upjohn Company, Kalamazoo, Michigan) group; 13 patients from the short-CY group; and 14 patients from the long-CY group were eligible by age, sex, and renal function criteria for our analysis.

In the second protocol [6], one group of patients with active lupus nephritis was randomly assigned to receive quarterly pulse cyclophosphamide (0.5 to 1.0 g/m² [n = 20]). Patients from this protocol who received at least 15 doses (range, 15 to 24 doses) of cyclophosphamide were included in our analysis (n = 9). Finally, three patients with neuropsychiatric lupus treated with a short course of cyclophosphamide [19], analogous to the protocol of short-CY for lupus nephritis, were included. Patients were followed for at least 4 years after the cessation of therapy. All patients had a thorough gynecologic evaluation after the occurrence of amenorrhea. Serum gonadotrophin levels were available for 7 of 11 patients who developed sustained amenorrhea.

Statistical Analysis

The distribution of clinical features among the treatment groups at study entry was analyzed using the Kruskal-Wallis and chi-square tests. Two-tailed tests were used to estimate the P values. The proportion of patients developing sustained amenorrhea was compared according to duration of cyclophosphamide therapy and patient age at study entry. For these comparisons, statistical analysis was done using the Fisher exact test and the chi-square test for trend where appropriate.

Results

Top Methods Results Discussion Author & Article Info References

Pertinent data on the patients evaluated for amenorrhea are shown in Table 1. The distributions of demographic and laboratory features were not statistically different among the methylprednisolone and cyclophosphamide treatment groups.

The risk for sustained amenorrhea among patients treated with Long-CY was most evident in patients older than 25 years (see Table 2). Among patients 26 years of age or older, 2 of 12 treated with Short-CY developed sustained amenorrhea, whereas 7 of 11 patients treated with Long-CY developed sustained amenorrhea (the Fisher exact test, P = 0.03).

Discussion

Top Methods Results Discussion Author & Article Info References

Patients with systemic lupus erythematosus now survive longer. With the decreased morbidity and longer life expectancy of these patients, gonadal toxicity is becoming a problem. Our study addresses the ovarian toxicity of pulse cyclophosphamide therapy, a commonly used intensive therapy for the major manifestations of lupus and other immune-mediated diseases.

Ovarian Toxicity

The mechanism of ovarian toxicity from cyclophosphamide has been studied in animal models [35-38]. A single intraperitoneal injection of 100 mg/kg of cyclophosphamide decreases the number of small follicles in ovaries of mice by about 63% [35]. The pool of growing follicles (medium to large) appears to be more vulnerable to the cytotoxic effect of cyclophosphamide than the small follicles [37]. After intraperitoneal injections of cyclophosphamide in immature rats primed with pregnant mare gonadotrophins, serum estradiol levels and the number of granulosa cells expressed from each ovary were decreased [37]. Cross-links in DNA in granulosa cells continue up to and probably beyond 24 hours, suggesting that the effects of cyclophosphamide on granulosa cells are prolonged. In addition to DNA, other macromolecules (enzymes, proteins) in the granulosa cells may be alkylated by cyclophosphamide metabolites. Decreased serum estrogen levels lead to up-regulation of follicle-stimulating hormone secretion, thus accelerating further follicular recruitment into the developing cyclophosphamide-sensitive pool of follicles. This mechanism perpetuates a vicious cycle, whereby cyclophosphamide destroys the developing follicles by attacking rapidly dividing granulosa cells, reducing their steroid secretion, and leading to increased pituitary gonadotrophin production, which enhances further recruitment of follicles into the pool of maturing follicles susceptible to cyclophosphamide [37]. These early events eventually result in accelerated depletion of ovarian follicles as shown in histologic sections from patients treated with cyclophosphamide that are devoid of follicles [3, 9, 13]. This general pathogenetic scheme has led to various strategies in an effort to protect the ovary by suppressing ovarian function using oral contraceptives [39-41].

The menstrual cycle has resumed after cyclophosphamide therapy in some patients [7]. This result emphasizes the need for long follow-up before declaring amenorrhea to be permanent. In our study, all episodes of transient amenorrhea were observed in the Short-CY group. It is possible that early removal of cyclophosphamide may prevent permanent damage to the ovaries. Thus, shorter courses of cyclophosphamide therapy in selected patients may minimize the risk for ovarian toxicity.

Cyclophosphamide Therapy and Fertility

The impact of pulse cyclophosphamide therapy on fertility is an important consideration for many patients. Byrne and colleagues [42] examined fertility rates in men and women, compared with sibling controls, in a retrospective cohort study of 1232 patients treated between the ages of 15 and 20 years for various malignancies. Chemotherapy and alkylating agents, with or without radiation to sites below the diaphragm, were associated with problems with fertility in about 60% of the men. Among the women, the adjusted relative fertility was similar in treated compared with untreated women. Unfortunately, only 55 patients received an alkylating agent alone, and neither the type of drug nor the dose were specifically reported.

Among the patients in our report, there was a total of 10 pregnancies. After methylprednisolone therapy, two pregnancies (one miscarriage and one elective abortion) occurred. After Short-CY therapy, four full-term pregnancies with delivery of normal babies and 4 elective abortions occurred. No pregnancies occurred in the Long-CY group. Our study cannot address adequately the effect of pulse cyclophosphamide on fertility. However, our data suggest that women who do not develop amenorrhea during pulse cyclophosphamide therapy should be considered fertile, and they may want to consider using birth control.

Another equally important issue is whether patients who were treated with pulse cyclophosphamide therapy and who maintain normal menstrual cycles will have premature menopause. One of our patients in the Short-CY group had apparent menopause at age 34, 3 years after cessation of pulse cyclophosphamide therapy. Because cyclophosphamide may accelerate the loss of follicles that naturally occurs with aging, it is possible that it may cause menopause earlier than expected in some patients. Longitudinal studies will be necessary to address these questions.

Conclusions

The optimal intensity and duration of pulse cyclophosphamide therapy are not yet clearly defined. A long course of cyclophosphamide for patients with severe nephritis did not statistically reduce the risk for doubling serum creatinine levels compared with those treated with a short course [17]. However, patients treated with a short course of pulse cyclophosphamide therapy had a statistically higher probability of exacerbations of lupus nephritis than did patients treated with a long course [17]. This retrospective analysis shows that, despite its efficacy, intermittent pulse cyclophosphamide therapy is associated with a risk for amenorrhea that is related to age and duration of treatment. Taken together, these findings emphasize the need for defining subsets of patients with active lupus in whom shorter courses of pulse cyclophosphamide therapy may be adequate. Finally, we hope that alternative therapeutic modalities and effective strategies to minimize the risk for the ovarian toxicity from pulse cyclophosphamide therapy can be found.