| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Merck Research Laboratories (D.P., Y.M., J.W., M.L., D.S.), Rahway, New Jersey 07065-0900; Clinical Research Center (S.M.), San Antonio, Texas 78229; Department of Obstetrics and Gynecology, University of Louisville (S.N.), Louisville, Kentucky 40292; Department of Obstetrics and Gynecology, Fallon Clinic (E.P.), Worcester, Massachusetts 01605; Department of Obstetrics and Gynecology, University of Massachusetts Memorial Health Care (E.P.), Worcester, Massachusetts 01605; Department of Obstetrics and Gynecology, Baystate Medical Center (R.B.), Springfield, Massachusetts 01106; Department of Gynecology and Obstetrics, Henry Ford Hospital (D.R.), Detroit, Michigan 48202; and Department of Obstetrics and Gynecology, University of Medicine and Dentistry of New Jersey (N.S.), Newark, New Jersey 07103-2757
Address all correspondence and requests for reprints to: Dr. Diane Plotkin, Department of Clinical Research, Endocrinology and Metabolism, 126 East Lincoln Avenue, P.O. Box 2000, RY34-A236, Rahway, New Jersey 07065-0900. E-mail: . diane_plotkin{at}merck.com
Abstract
In this double-blind, randomized, placebo-controlled study, normally cycling women (n = 86) with elevated low density lipoprotein cholesterol (LDL-C) levels were studied over six menstrual cycles. At the end of the screening phase, participants received placebo for the second menstrual cycle and subsequently were randomized to receive either placebo or simvastatin (40 mg/d) for the next four cycles. The second and sixth menstrual cycles were considered baseline and treatment cycles, respectively. Participants kept a menstrual diary throughout the study and provided daily first-void urine samples during cycles 2 and 6. Urine samples were assayed for LH and pregnanediol glucuronide (PdG). The primary end point was change in luteal phase duration as defined by the day of the urinary LH peak to the day preceding the onset of menstruation. Treatment with simvastatin (40 mg/d) effectively lowered LDL-C by 34.3% (P < 0.001). Simvastatin was generally well tolerated, and no meaningful difference in adverse event profile was observed between treatment groups. Compared with the placebo group, simvastatin did not have clinically relevant effects on luteal phase duration, peak PdG concentration, or integrated luteal phase PdG concentration. The results of this study demonstrate that treatment of healthy premenopausal women for approximately 4 months with simvastatin (40 mg/d) lowers LDL-C without adversely affecting reproductive gonadal function. Simvastatin should not be used during pregnancy or by nursing mothers.
SIMVASTATIN, a 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitor, is effective in the treatment of hypercholesterolemia (1). The use of simvastatin has been shown to dramatically reduce the risk of coronary events, stroke, and mortality in controlled secondary prevention trials (1, 2, 3). The primary benefit conferred by this class of drugs results from their ability to reduce low density lipoprotein cholesterol (LDL-C), but beneficial effects on other lipoproteins have also been noted, such as raising high density lipoprotein cholesterol (HDL-C) and lowering triglycerides (TG) (1).
Recent cholesterol management guidelines issued by the National Cholesterol Education Program emphasize aggressive lipid-lowering intervention to further reduce morbidity and mortality associated with atherosclerotic disease (4). The call for intensive lipid management in high risk groups, such as individuals with diabetes mellitus, may well lead to increasing numbers of premenopausal women receiving lipid-lowering therapies, a cohort underrepresented in prior clinical trials. As cholesterol is the precursor of all steroid hormones, the impact of lipid-lowering therapy on gonadal function in this population is worthy of study. It should be noted that as specified in the simvastatin label, this drug should be used by women of child-bearing potential only when such patients are highly unlikely to conceive.
Clinically significant adverse effects of HMG-CoA reductase inhibitors on testicular function have not been shown in any controlled clinical trials; however, small variable effects on testosterone have been noted (5, 6, 7, 8, 9). Although menstrual dysfunction has not been reported in women enrolled in clinical trials, the effects of statin therapy on the menstrual cycle has not been specifically evaluated in a formal, carefully controlled manner. As the luteal phase of the menstrual cycle is the final common pathway for normal folliculogenesis and ovulation, we examined luteal phase duration to assess whether simvastatin adversely affects female gonadal function.
Subjects and Methods
Patient population
The protocol and patient consent form were approved at all sites by the local human subjects committees, and signed informed consent was obtained from each subject before enrollment. Eighty-six reproductive-aged women (25–41 yr old) were recruited for this study. All subjects met the following criteria: 1) screening serum LDL cholesterol between 130–250 mg/dl; 2) history of regular menstrual cycles between 25–35 d in length for the previous year by self-report and during the screening period; 3) normal PRL, TSH, and midluteal progesterone levels; 4) no excessive exercise (defined as more than five 45-min sessions per wk or jogging more than 20 miles per wk); 5) weight 90% or greater and 130% or lower normal weight for height; and 6) and nonsmoking status.
Study design
This was a randomized, double-blind, parallel, placebo-controlled, multicenter study. The study spanned six consecutive menstrual cycles. The second and sixth menstrual cycles were monitored hormonally and considered baseline and treatment cycles, respectively. A window of +1 menstrual cycle was permissible for cycles 2 and 6 if the participant knew she would not be able to reliably collect daily urine samples for the specified cycle. Participants kept a menstrual diary throughout the study, recording onset and duration of menstrual bleeding, and had six clinic visits to monitor menstrual cycles, study compliance, and side effects. Plasma lipid and safety laboratory assays, including liver function tests and creatine kinase (CK) measurements, were performed at each clinic visit.
The first menstrual cycle (cycle 1) served as a screening period for verification of normal menstrual cycle length and evaluation of hormonal entry criteria. The second menstrual cycle (cycle 2) comprised a single-blind placebo treatment baseline phase in which all participants were administered placebo. This baseline placebo phase continued for one menstrual cycle, during which menstrual cycle length and hormonal descriptors were monitored. The active treatment phase started on the first day of the third menstrual cycle (cycle 3), at which time study participants were stratified by age into two groups (21–35 and 36–40 yr) and randomized 1:1 to receive either simvastatin 40 mg or matching placebo. Participants continued taking this medication for the next four menstrual cycles. During cycle 6, daily urine hormonal monitoring was undertaken as described below.
Participants were instructed to collect their first morning voided urine upon awakening each day into a clean container, provided by the investigators, containing glycerol as a preservative. They were instructed to freeze the tubes within 2 h of collection in their home freezer. Urine collection commenced on the second day of the menstrual cycle and continued until the first day of menstrual bleeding in cycle 3 or cycle 7. Urine samples were assayed for LH and pregnanediol glucuronide (PdG), the chief urinary metabolite of progesterone, and normalized for creatinine (Cr).
Laboratory assays
Assays of daily urinary LH and PdG were performed at the University of Medicine and Dentistry of New Jersey (Newark, NJ). Urine was assayed in duplicate for PdG, the chief urinary metabolite of progesterone, using an ELISA using antisera and conjugate provided by William Lasley (Davis, CA) (10, 11, 12). LH was measured in duplicate using a two-site, immunofluorometric assay (DELFIA, Pharmacia Biotech, Gaithersburg, MD) with a sensitivity of less than 0.1 mIU/ml (13). LH was measured against WHO IRP 68/40 human pituitary LH standard. All hormonal data were corrected for glycerol (7%) and normalized for Cr (14, 15).
Plasma samples from fasting patients were shipped overnight to MEDLAB Clinical Trials Laboratory (San Antonio, TX). The Friedewald equation was used to calculate LDL-C for each patient in relation to total cholesterol (total C), HDL-C, and TG (16). For patients with TG less than 400 mg/dl, LDL-C can be estimated using the following equation: LDL-C = total C - [(0.20 x TG) + HDL-C]. LDL-C was not calculated for patients with TG greater than 400 mg/dl.
End point definitions and study power
The primary end point was change from baseline in luteal phase length. The study was designed to have 80% power to detect a mean difference of 3.8 d in luteal phase duration between the two treatment groups (
= 0.05, by two-tailed test). A period of 4 days was judged to be a clinically important difference, as normal luteal phase length is approximately 14 d, and less than 10 d is generally considered evidence of luteal dysfunction (17). This calculation assumed an SD of 5.2 for pre- and on-treatment periods based upon unpublished data (unpublished observations); however the present analysis showed much smaller variability (see Discussion).
Luteal phase length was defined as the day from the LH surge to subsequent menstruation. The first, predefined rule used to identify the LH surge was the first day with LH levels of 7 mIU/mg Cr or more, followed by a PdG rise to at least 3 µg/mg Cr for at least 3 d. This primary definition was based upon a previous analysis of 15 normal cycles (13). When the blinded hormonal assays were nearly completed, it became clear that LH surges demonstrated considerably more variability and frequently were less than 7 mIU/mg Cr, although ovulatory PdG rises were detected. To minimize the potential for bias in selecting cycles available for luteal phase analysis, a decision was made while the study data were still blinded to use an inferred day of the LH surge, defined using a stepwise algorithm. The purpose of designing the stepwise algorithm was to assign an LH surge day to as many cycles as possible with available urinary measurements. If the peak LH was 10 mIU/mg Cr or greater, then this day was assigned the LH surge day. However, if the peak LH was 10 mIU/mg Cr or less, but the patient had a normal PdG rise to more than 4 µg/mg Cr, then the day with the highest LH value preceding the PdG rise was considered the LH surge day. The higher LH cut-off value at the first step (10 mIU/mg Cr) forced inspection of PdG levels to confirm that ovulation occurred at some point during the menstrual cycle in cases when the peak LH concentration was lower. This less restrictive definition allowed for the inclusion of a greater number of subjects in analyses of primary and secondary end points. Analyses using both definitions for the LH surge day are presented in this report.
Secondary end points included gonadal function measurements and a comparison of the safety (adverse experiences) and efficacy (plasma lipids) of simvastatin treatment vs. placebo. Secondary gonadal function end points were 1) peak urinary luteal PdG concentration (defined as the maximum PdG value for a patient during the luteal phase), 2) integrated urinary luteal PdG (measured as the area under the concentration curve using the trapezoidal rule from d -2 relative to the urinary LH surge through the last day before onset of menstruation), 3) proportion of participants with low peak urinary luteal PdG (PdG concentrations <4 µg/mg Cr), 4) proportion of participants with anovulatory cycles [defined as the absence of both an LH surge (LH 
7 mIU/mg Cr) and a peak PdG 4 µg/mg Cr], and 5) proportion of women with abnormal cycle lengths (<25 or >35 d).
Approaches to analyses
A per-protocol approach was used in the analyses of all gonadal function end points. A set of prespecified rules was used to determine per-protocol violators, including excessive weight loss (consistent average weight loss of 1.25 lb/wk throughout the study period), excessive exercise, and drug noncompliance (<75% compliance). Only patients with baseline (cycle 2) and on-treatment (cycle 6) urine samples were evaluated, and patients with the aforementioned violations were excluded from the analysis. An additional per-protocol rule was adopted for the analysis of integrated luteal PdG concentration that excluded patients with missing urine data (PdG measurements missing for 2 consecutive d, or 5 d total).
An intention to treat approach was used for safety [the proportion of patients experiencing greater than 3-fold increases in liver function tests or more than 10-fold CK increases in the upper limit of normal (ULN); median changes in alanine aminotransferase (ALT), aspartate aminotransferase, and CK) and efficacy (mean percent change in lipids) analyses]. All randomized patients with baseline data and at least one on-treatment measurement were included in these analyses, and if data from the final visit were missing, the last observed measurement in the treatment period was carried forward to impute the missing value.
Statistics
An ANOVA model was used to analyze change from baseline in luteal phase length, peak urinary luteal PdG, and integrated urinary luteal PdG. P < 0.050 was considered statistically significant. All ANOVA models included factors for treatment, study center, and age stratum. The least squares mean for the between-treatment difference and the 95% confidence interval (CI) were provided. The between-group difference in luteal phase duration was not considered clinically significant if the lower boundary of the CI did not include -4, indicating a 4-d decrease in length. A Cochran-Mantel-Haenszel analysis, stratified by baseline status and age stratum, was used for all categorical gonadal function end points (proportion of patients with abnormally low peak PdG, anovulatory cycles, and abnormal cycle lengths) (18). The Mantel-Haenszel estimate compared the probability of gonadal function abnormalities in the placebo vs. simvastatin groups. If the 95% CI included 1.0, the relative risk was judged as not statistically significant. Proportions of patients with clinically important changes in liver function tests, such as ALT and aspartate aminotransferase, and CK, were compared using Fisher’s exact test.
Results
Baseline demographics for the two treatment groups are provided in Table 1
. The treatment groups were generally comparable with regard to baseline parameters, including age, race, and gonadal function characteristics. A total of 86 women were recruited and randomized to receive either placebo (n = 44) or simvastatin (n = 42). Of these patients, 82 (95.3%) completed the study. Four subjects were dropped from the study, 3 because of protocol violations and 1 who was lost to follow-up. Clinic visits were scheduled based on the individual’s menstrual cycle, and the total treatment duration was 4 cycles, or approximately 4 months (16 wk). The mean duration of treatment was 120 d.
|
The effects of four menstrual cycles of treatment with simvastatin compared with baseline on LDL-C, total C, HDL-C, and TG are shown in Table 2. Patients receiving simvastatin (40 mg/d) had a mean reduction from baseline of 34.3% in LDL-C (P < 0.001), and significant reductions in total C (23.7%; P < 0.001) and TG (14.7%; P < 0.001). Compared with baseline, HDL-C increased in the simvastatin group by 4.9% (P = 0.001). Treatment with placebo did not produce significant effects compared with baseline for any of the lipid parameters measured. The between-group comparisons were highly significant for LDL-C, total C, and TG (P < 0.001) and borderline significant for HDL-C (P = 0.082).
|
No patients discontinued therapy due to clinical or laboratory adverse experiences. Eight patients in the placebo group had clinical adverse experiences reported as possibly, probably, or definitely drug-related by the investigator compared with five patients in the simvastatin group; no differences in type or frequency were observed between groups (Table 3). Two placebo and three simvastatin patients reported laboratory adverse experiences during the study; none was rated as drug-related by the investigator (placebo group: one patient had separate ALT and CK elevations, one woman had an abnormal PAP smear; simvastatin: one patient each had CK or ALT elevation, and one woman had an abnormal PAP smear). No patients in either treatment group experienced liver enzyme elevations greater than 3 times the ULN, the traditional boundary for a clinically significant change. Additionally, no patients experienced CK increases greater than 5 times the ULN, and no patients experienced myopathy, defined as muscle symptoms and CK elevation to 10 times or more the ULN.
|
Forty placebo and 35 simvastatin patients were included in the stepwise analysis of luteal phase duration (Table 4). A total of 11 patients were excluded from this analysis, of whom 4 patients discontinued the study, 2 patients had unusable urine samples, 1 patient had a missing urine sample for cycle 6, and 4 patients had an unidentifiable LH surge day by the stepwise definition. In the placebo group, the mean luteal phase lengths were 14.9 and 13.9 d at baseline and in cycle 6, respectively, with a least squares mean change of -1.3 d from baseline (P < 0.05). The mean luteal phase lengths for the simvastatin group were 14.5 and 14.9 d at baseline and in cycle 6, respectively, representing a nonsignificant mean change of 0.4 d. The between-group comparison was 1.4 d (P < 0.05) with a 95% CI of 0.2, 2.5 d, driven by the change in the placebo group. These results were similar to those obtained using the primary rule for the LH surge day, in that the simvastatin group demonstrated slightly longer luteal phase durations relative to the placebo group (Table 4). The lower boundary of the 95% CI did not include -4 d by either analysis; therefore, the between-group difference in luteal phase duration was not deemed clinically relevant by the predefined definition (see Subjects and Methods).
|
Table 5 contains the secondary continuous gonadal function measurements, peak luteal PdG (n = 80) and integrated urinary luteal PdG concentration (n = 71). There were no significant differences in luteal phase peak or mean integrated PdG concentrations, either within or between groups. The placebo and simvastatin-treated groups also had similar PdG profiles across the luteal phase of the menstrual cycle (Fig. 1). The rate of rise in PdG and the midcycle peak PdG concentration did not differ between the placebo and simvastatin groups. Furthermore, the peak urinary PdG occurred approximately 7–8 d after the LH surge (d 0) in both treatment groups at baseline and on-treatment. Other secondary end points, including analyses of the proportion of patients with abnormally low peak PdG concentrations, anovulatory cycles, and abnormal menstrual cycle lengths (Table 6), revealed no significant within- or between-treatment group differences.
|
|
|
The results of this study demonstrate that treatment of premenopausal women, aged 25–41 yr, for four menstrual cycles (
4 months) with simvastatin (40 mg/d) lowers LDL-C without adversely affecting reproductive gonadal function, as measured by the luteal phase length. We chose luteal phase length as the primary end point of this study for several reasons. Menstrual cycle physiology is dependent upon complex, integrated endocrine systems, involving hypothalamic, pituitary, ovarian, and uterine mechanisms. Although it is not feasible to effectively measure the function of each individual system, the presence of normal progesterone secretion and luteal phase length indicates that the processes of follicular recruitment, ovulation, and corpus luteum function have proceeded normally. The dominant hormone of the luteal phase is progesterone, and therefore, by measuring excretion of its chief metabolite, PdG, we were able to evaluate the quality of the luteal phase. This information, combined with urinary LH measurements, allowed us to evaluate luteal phase duration.
The participants in this study had an average baseline luteal length of 14.5 and 14.9 d in the simvastatin and placebo groups, respectively. There was a small nonsignificant increase in luteal length from 14.5 to 14.9 d among women receiving simvastatin. Luteal length was significantly (P < 0.05) reduced from baseline in placebo-treated women resulting in between-group differences of approximately 1 d, which, although statistically significant, were not considered clinically relevant.
Luteal phase duration is fairly consistent in women with regular menstrual cycles. In one study, the 95% lower confidence limit for normal luteal length was 11.3 d (17). In the present study the SD for the change from baseline of luteal phase duration was unexpectedly low, on the order of 1–3 d depending on the LH surge definition. This value is considerably less than that found in a population-based sample of women undergoing similar urinary testing (unpublished observations), in which an SD of more than 5 d was observed. The difference probably reflects the stringent inclusion criteria for this study and for other, nonpopulation-based studies. Study eligibility criteria excluded women who may be prone to menstrual cycle abnormalities, such as women on weight loss programs, performing excessive exercise, or with PRL or thyroid disorders, a history of infertility, or abnormal menstrual cycle length. Furthermore, participants were screened for evidence of ovulatory cycles by assaying for midluteal serum progesterone levels. These more stringent criteria probably resulted in the reduction in luteal length variability observed in this study.
Progesterone biosynthesis in the corpus luteum is dependent upon receptor-mediated uptake of LDL-C from plasma (19, 20). Thus, one concern before this study was initiated was that HMG-Co A reductase inhibitors could decrease the synthesis and secretion of this critical hormone, which, in turn, may result in inadequate maturation of the endometrium. Because of its limited serum binding and relatively short serum half-life, serum progesterone has been found to be pulsatile in nature, with concentrations ranging from as low as 5 ng/ml to as high as 35 ng/ml within 6–8 h (21). Consequently, blood sampling every 20 min over a 12- to 24-h period is necessary to determine progesterone concentrations with sufficient precision (22). We attempted to reduce some of the variation in progesterone measurement by measuring overnight urinary pregnanediol excretion. Collection over a window of time serves to dampen the extremes seen with serum sampling and reduce variability, rendering this a more sensitive measure (23). Even with the use of daily PdG measurements, we did not detect significant differences in pregnanediol concentrations between the simvastatin (40 mg/d) and placebo groups.
Although the results of this study indicate that simvastatin is generally well tolerated by premenopausal women, caution should be used when prescribing this drug to women of child-bearing potential. As specified in the simvastatin label this drug should not be used during pregnancy or by nursing mothers (24). Simvastatin should be administered to premenopausal women only when such patients are highly unlikely to conceive. In conclusion, the efficacy and tolerability of simvastatin (40 mg/d) in premenopausal women observed in this study is consistent with the extensive experience in adult populations using simvastatin (1, 25, 26). Treatment of premenopausal women with simvastatin (40 mg/d) resulting in reductions of LDL-C by approximately 35% did not disrupt the menstrual cycle, demonstrating that simvastatin can be safely used in reproductive-aged women. This information is especially important to clinicians in light of the most recent clinical guidelines for cholesterol testing and management that recommend cholesterol-lowering treatment for all at-risk patients regardless of age and gender (4).
Acknowledgments
The reported clinical study was entirely supported by Merck Research Laboratories, Merck \|[amp ]\| Co., Inc. Diane Plotkin, Yale Mitchel, Joanne Waldstreicher, Minzhi Liu, and Deborah Shapiro are employees of and hold stock in Merck \|[amp ]\| Co., Inc. We thank Dr. William Lasley (University of California, Davis, CA) for providing unpublished hormonal data, Dr. Gordon Saperia (Division of Cardiology, Fallon Clinic, Worcester, MA) for his contribution to study design and conduct, and Dr. Amy Johnson-Levonas (Merck \|[amp ]\| Co., Inc.) for assistance in preparing this manuscript for publication.
Footnotes
This work was supported Merck Research Laboratories, Merck \|[amp ]\| Co., Inc.
Present address for N.S.: Department of Obstetrics, Gynecology, and Women’s Health, Division of Reproductive Endocrinology, Albert Einstein School of Medicine, Bronx, New York 10461.
Abbreviations: ALT, Alanine aminotransferase; CI, confidence interval; CK, creatine kinase; Cr, creatinine; HDL-C, high density lipoprotein cholesterol; HMG-CoA, 3-hydroxy-3-methylglutaryl-coenzyme A; LDL-C, low density lipoprotein cholesterol; PdG, pregnanediol glucuronide; total C, total cholesterol; TG, triglycerides; ULN, upper limit of normal.
Received December 14, 2001.
Accepted March 17, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  S. B. Clauss, K. W. Holmes, P. Hopkins, E. Stein, M. Cho, A. Tate, A. O. Johnson-Levonas, and P. O. Kwiterovich Efficacy and Safety of Lovastatin Therapy in Adolescent Girls With Heterozygous Familial Hypercholesterolemia Pediatrics, September 1, 2005; 116(3): 682 - 688. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Santoro, B. Lasley, D. McConnell, J. Allsworth, S. Crawford, E. B. Gold, J. S. Finkelstein, G. A. Greendale, J. Kelsey, S. Korenman, et al. Body Size and Ethnicity Are Associated with Menstrual Cycle Alterations in Women in the Early Menopausal Transition: The Study of Women's Health across the Nation (SWAN) Daily Hormone Study J. Clin. Endocrinol. Metab., June 1, 2004; 89(6): 2622 - 2631. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
