| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Departments of Obstetrics, Gynecology, and Women’s Health (A.J., A.J.P., G.Z., S.F., B.I., N.S.) and Medicine (D.R., H.N.P.), Albert Einstein College of Medicine, Bronx, New York 10461; Department of Obstetrics and Gynecology (S.L.B., T.L.), Emory University, Atlanta, Georgia 30322; and Department of Surgery (K.G.), Montefiore Medical Center, Bronx, New York 10467
Address all correspondence and requests for reprints to: Nanette Santoro, M.D., Division of Reproductive Endocrinology, Department of Obstetrics, Gynecology, and Women’s Health, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, New York 10461. E-mail: glicktoro{at}aol.com.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objective and Design: To elucidate the reproductive defects associated with obesity, we sampled 18 eumenorrheic (nonpolycystic ovary syndrome) women with a mean ± SEM body mass index of 48.6 ± 1.4 kg/m2 with daily, first morning voided urine collections, seven of whom also had early follicular phase 12-h, every 10-min blood sampling to assess LH pulses. Daily hormones were compared with 11 eumenorrheic, normal-weight controls. A separate control group of 12 eumenorrheic, normal-weight women was used for the LH pulse studies.
Main Outcome Measures: Assays for LH (serum and urine) and FSH, and estradiol and progesterone metabolites (estrone conjugate and pregnanediol glucuronide; urine) were performed. Daily hormones were meaned and normalized to a 28-d cycle length. LH pulsations were determined using two objective methods. Group means were compared using t tests.
Results: Reduced whole-cycle mean, normalized pregnanediol glucuronide was observed in obese (38.2 ± 2.1 µg/mg creatine) compared with normal-weight women (181.3 ± 35.1 µg/mg creatine; P = 0.002), without significant differences in LH, FSH, or estrone conjugate. Early follicular phase LH pulse frequency did not differ from normal-weight women, but both amplitude and mean LH were dramatically reduced in obese women (0.8 ± 0.1 and 2.0 ± 0.3 IU/liter) compared with controls (1.6 ± 0.2 and 3.4 ± 0.2 IU/liter; P < 0.01).
Conclusions: A novel defect in the amplitude but not the frequency of LH pulsations appears to underlie the reproductive phenotype of obesity. The deficit in pregnanediol glucuronide appears to exceed the deficit in LH. The patterns of hypothalamic-pituitary-ovarian axis function unique to the obese state differ from other abnormal reproductive states.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
We hypothesized that the reduced reproductive hormones in ovulatory, regularly cycling obese women are attributable to a deficient central neural reproductive drive. We studied morbidly obese (baseline BMI of >35 kg/m2) women scheduled to undergo bariatric surgery before weight loss and compared menstrual cycle urinary hormone excretion and serum LH secretory patterns using frequent blood sampling to infer the characteristics of GnRH secretion. In this study, we compared our findings in the high-BMI women with normal-weight historical controls.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Nineteen participants were recruited through a weight-loss surgery support group at the Montefiore Medical Center and Beth Israel Medical Center (New York, NY). Participants were aged 35–50 yr at enrollment and had to meet the following criteria: 1) BMI of at least 35 kg/m2; 2) history of regular menses every (q) 25–40 d; 3) presence of a uterus and at least one ovary; 4) no evidence of renal, hepatic, or systemic disease that might affect gonadotropin or sex steroid production or clearance; and 5) no exogenous hormones for at least 3 months before study entry. Participants were screened with a fasting glucose, prolactin, TSH, free T4, total T3, reverse T3, and urine dipstick analysis. Patients with diabetes, alcoholism, or any screening laboratory test abnormality were excluded. Dietary habits were recorded and periodically reviewed to assess nutritional intake. Six of the seven high-BMI women (one participant declined) who underwent frequent blood sampling studies also had a transvaginal ultrasound examination immediately before the 12-h study to rule out the presence of polycystic ovaries. All ultrasounds were done by one of the coauthors (A.J.) using a dedicated Aloka (Tokyo, Japan) machine at the General Clinical Research Center.
Controls
Two control groups of normal-weight, eumenorrheic women were used. Eleven normal-weight (BMI of 21.3 ± 0.4 kg/m2) women aged 19–38 yr (mean of 31.5 ± 1.5 yr) who underwent the same daily urine collection methods and whose data have been reported previously were used as controls for the daily urine studies (5, 6). These women were normoprolactinemic, free of systemic illness, and had no history of excessive exercise (>4 h/wk). Previously published data from 12 women (BMI of 20.8 ± 0.5 kg/m2) who had LH pulsatility assessments using q15-min blood sampling and an identical assay to the one we used for LH were provided by the laboratory of coauthors (S.L.B., T.L.) (7). These women were aged 20–33 yr (mean of 24 ± 1.5 yr), were ovulatory by luteal phase progesterone, had no systemic disease, and had no history of excessive exercise. This protocol was approved by the internal review board, and all participants provided their informed consent before participation.
Protocols
Baseline assessments were performed before bariatric surgery. First-morning voided urine was collected over an entire menstrual cycle and assayed for LH, FSH, estrone conjugate (E1c), and pregnanediol glucuronide (Pdg). Details on specimen collection procedures have been published previously (8). Women were instructed to collect urine daily with the onset of their menstrual period and to continue collection up to the onset of their next menstrual period. A subset of patients underwent early follicular phase (cycle d 2–5), q10-min, frequent blood sampling for LH pulse assessment. Participants were admitted to the General Clinical Research Center, a large-bore IV was inserted into a forearm vein, and 2 ml of blood were withdrawn q10 min over a 12-h period from 0800 to 2000 h. Isovolumetric fluid was replaced with heparinized normal saline. Procedures have been described previously (9). The use of daytime 12-h frequent sampling has been shown to have adequate power to detect differences in pulse frequency with sample sizes of eight patients (10).
Hormone assays
LH and FSH were measured using a solid-phase, two-site specific fluoroimmunometric assay (DELFIA; PerkinElmer, Turku, Finland) and were validated in our laboratory using previously described methods (6, 7). For urinary LH, interassay coefficient of variation (CV) and intraassay CVs were 13.7 and 5.0%, respectively, and 16.4 and 7.6% for FSH, respectively. E1c and Pdg were measured in duplicate by ELISA using antibodies and conjugate tracers provided by Dr. Bill Lasley (University of California-Davis, Davis, CA). Interassay CV for E1c was 10.1% and intraassay CV was 8.4%. Corresponding CVs for Pdg were 15.0 and 14.0%, respectively. Hormone concentrations were adjusted for glycerol and normalized to creatinine (11). Serum LH for the frequent blood sampling studies was measured using the same reagents as the urine assay, with interassay and intraassay CVs being 5.5 and 2.3%, respectively.
Data analysis
The presence or absence of luteal activity was determined by a sustained increase in Pdg concentrations, and the day of ovulation was determined using a previously published algorithm validated in other studies (4). Before analysis, cycle length was normalized to 28 d, after determination of individual whole-cycle mean hormones (12). Cycle days were aligned to the day of ovulation, which was set as d 0. Transverse means of all hormones, normalized to a 28-d cycle, were compared using t tests. P < 0.05 was considered statistically significant.
For frequent blood sampling studies, serum was aliquoted and assayed for LH. Two objective criteria were used to determine LH pulses. In the first method, an increment of 20% from the previous nadir for two consecutive points was used to define a pulse (13). Because the data series were not both taken at the same interval (q10 min in the high-BMI group and q15 min in the normal-weight group), we imposed an additional criterion that pulses in the q10 min series had to consist of two consecutive points in which the 20% criterion was met. The second pulse detection method, PulseFit (Kushler-Brown), uses nonparametric data smoothing and a statistical definition of pulses (14). Additional validation of this methodology was accomplished by examining a noise series of 72 specimens at a level of approximately 1 IU/liter. When both pulse identification methods were applied, only a single pulse with an amplitude of 0.5 IU/liter was observed by PulseFit. The statistical outcomes of the data analysis were identical regardless of the pulse detection technique applied. We report the findings using the 20% criterion for pulse detection, because it best duplicated the visual data. We did not make a priori assumptions about the direction of LH pulse frequency or amplitude changes in obese preoperative women and therefore used two-tailed testing for all outcomes.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
).
|
Participant and control group characteristics are shown in Table 1
for whole-cycle urine. A significantly greater mean BMI was seen in the obese group compared with the normal-weight control groups. Table 2 (top, Urine metabolites) shows the whole-cycle hormone concentrations. Gonadotropins did not differ significantly between the groups. Overall E1c excretion was similar in both groups; however, mean Pdg excretion was dramatically reduced in the high-BMI obese cohort (181.3 ± 35.1 vs. 38.2 ± 2.1 µg/mg creatine; P = 0.002) (Fig. 1).
|
|
Seven women with a mean BMI of 47.2 ± 2.2 kg/m2 completed frequent sampling studies. Mean pulse frequencies over 12 h were similar in both the high-BMI and normal-weight women. However, pulse amplitude was significantly reduced by 50% (0.8 ± 0.1 vs. 1.6 ± 0.2 IU/liter; P = 0.001), and mean LH level was lower (2.0 ± 0.3 vs. 3.4 ± 0.2 IU/liter; P = 0.006) in the high-BMI group compared with the normal-weight controls (Table 2, bottom, LH pulsatility; Fig. 2). Mean estradiol levels for the control and study groups were found to be 0.11 ± 0.015 vs. 0.13 ± 0.01 nmol/liter (P = 0.451; 30.4 ± 4.0 and 34.2 ± 2.7 pg/ml).
|
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In 1975, Sherman and Korenman (2) reported on a small group of regularly cycling obese women (n = 6) who underwent daily blood sampling and were compared with women of normal body weight. The cycles of the overweight women were found to have longer follicular phases, decreased FSH and LH, and lower luteal phase progesterone (2). Similar findings were reported by another group studying very obese women (BMI of >35 kg/m2) (3). The SWAN Daily Hormone Study, which evaluated 836 cycles of reproductive age women, found a progressive association of increasing body weight with reduced luteal Pdg, LH, and FSH but not E1c. These relationships persisted after controlling for ethnicity, smoking, age, menopausal status, and creatinine (which is higher in obese women) (4). In our study, which included a sample much smaller than the SWAN Daily Hormone Study, daily hormone assessments demonstrated a large deficit in Pdg excretion, but reduced LH secretion was only observable in the frequent blood sampling studies and not in the daily assessments. This finding implies that the deficit in Pdg exceeds the deficit in LH and suggests that more than one mechanism may be responsible for our findings. To ensure that the lower Pdg was not attributable to differences in creatinine excretion, the data were reanalyzed after correcting for the higher creatinine in the high-BMI women, and the relationship remained highly statistically significant (P = 0.004).
Obesity has a direct association with abnormal menstrual cycles (15), and weight reduction can restore ovulatory cycles in obese, anovulatory women (16). A large body of data has elucidated the pathophysiology of anovulation in association with obesity as it relates to polycystic ovarian syndrome (PCOS). The gonadotropin dysfunction seen in PCOS may also be the result of hypothalamic and pituitary mechanisms as seen by increased LH pulse frequency and often elevated LH or LH/FSH ratio (17). A relationship with adiposity similar to what we described has also been seen in PCOS. With increasing body weight, LH pulse amplitude is reduced in women with PCOS, although pulse frequency is increased compared with normally cycling women in both lean and obese PCOS (18). In women with PCOS, response to exogenous GnRH and to submaximal GnRH receptor blockage with an antagonist have demonstrated relatively normal dynamics, suggesting that inadequate secretion of hypothalamic GnRH is not the cause of this finding (19). Our non-PCOS, high-BMI sample demonstrated normal LH pulse frequency but a decreased amplitude, resulting in overall lower LH levels. These data imply that a selective reduction in LH secretion may be a result of increased BMI or chronic overnutrition. We do not believe that this finding was related to any acute reduction in consumption before surgery based on preoperative food recall assessments.
There are several mechanisms by which increased BMI might reduce peripheral LH and progesterone. First, these findings could simply be the result of selective partitioning of secreted LH into adipose tissue. Although data are limited, several aspects of our findings argue against this interpretation. The observation that circulating concentrations of LH and FSH are confined within a narrow range for most of the menstrual cycle, save the LH surge, suggests that women are in a steady state for most of the cycle and that adipose should be saturated with LH, should partitioning occur. The finding that circulating FSH did not differ between high-BMI and normal-weight women in our study, although LH was reduced by 50%, argues against adipose partitioning of gonadotropins. There are virtually no data that address partitioning of gonadotropins, although sex steroid uptake of progesterone and, to a lesser extent, estrogen has been reported by others (20). This latter finding may help to explain the marked reduction in Pdg that we observed, because the bulk of progesterone secretion is confined to the luteal phase of the menstrual cycle, and selective adipose partitioning of progesterone could reduce its bioavailability in the circulation. The recent discovery of an LH receptor mRNA interfering factor, mevalonate kinase (21), could explain the profound deficit in Pdg excretion on the basis of impaired LH action on the corpus luteum.
It is also possible that the pituitary response to endogenous GnRH is attenuated by obesity. A weight-related decrement in response to exogenous GnRH given in a physiological-range dose of 10 µg was observed in a sample of women with PCOS but not in normally cycling controls in one study (22). The presence of insulin receptors on pituitary cells implicate insulin as a potential factor influencing LH secretion; however, an inverse relationship between insulin and LH levels has not been identified (19). Leptin, a product of adipose tissue, has been shown to have a stimulatory effect on LH and FSH secretion (23). The severe leptin resistance associated with obesity could reduce this stimulation and result in less LH output. However, one would expect that FSH would also be reduced by such a mechanism, and that is not consistent with our findings. Other adipokines, such as TNF-
and IL-1ß, are secreted in excess with increasing fat mass, and both have been shown to exert negative effects on the pituitary production of LH when injected intracerebroventrically into rats (24). Anti-TNF-antibody neutralization has been demonstrated to reverse the hypothalamic-pituitary axis suppression in a similar model (25). TNF-, but not IL-6, has also been shown to reduce GnRH-induced LH release from the pituitary gland (26). Inflammatory adipocytokines may also inhibit end-organ response of the corpus luteum. In vitro production of sex steroids has been shown to be reduced in bovine (27) and rat (28) granulosa cells treated with TNF-. IL-6, but not TNF-, reduced progesterone production by luteinized granulosa cells in the monkey (29). Leptin directly inhibits human granulosa cells in vitro as evidenced by decreased progesterone by luteinized granulosa cells in response to human chorionic gonadotropin (30) and inhibition of IGF-I-mediated enhancement of FSH-stimulated estradiol synthesis (31). In addition, leptin may potentially interrupt normal oocyte maturation (32) and may serve as a marker of poor implantation potential (33). If leptin were to interfere with gonadotropin action, one would have to hypothesize that leptin resistance, a well-known central phenomenon in obesity (34), does not occur in the periphery.
This study carries several limitations. As an interim analysis to a larger study that will eventually provide postoperative weight-loss data, the comparison groups for this study are historical and may not represent the ideal control group. The use of urinary metabolites for gonadotropin and sex steroid assessments across the menstrual cycle has been validated as a useful proxy for serum. However, urinary measurement is probably least accurate for estradiol assessment because both estrone and estradiol are measured (8). Nonetheless, we supported our frequent sampling study equivalency of ambient estradiol by measuring it directly. Second, we are unable to account for possible adipose tissue uptake and storage of steroid hormones, which may have contributed to lower detected levels of serum or urinary hormones. However, we did not observe an overall dilutional effect on all hormones, a fact that argues against a simple dilutional effect. Finally, our ability to detect subtle differences in pulse frequency associated with obesity may be compromised by our relatively small sample size. The decreased frequency of sampling (q15 min) in the controls would have resulted in a lower frequency of pulse detection in the group and would have amplified the finding of a rapid pulse frequency in the high-BMI women.
In summary, we support and extend previous observations to provide a description of a novel reproductive phenotype of obesity in ovulatory, cycling women. We propose that impaired amplitude of LH pulsatility results in inadequate luteal stimulation and, in possible combination with peripheral factors, causes additional reductions in urinary Pdg. Although the primary cause of this phenotype is not presently known, available data implicates peripheral adipose signaling as responsible for the reproductive dysfunction. As the postoperative (after weight loss) phase of this study is completed, a better understanding of the acute and chronic effects of reduced adiposity will be possible.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Data were presented in part at the 88th Annual Meeting of The Endocrine Society, Boston, Massachusetts, 2006.
N.S. has served as an ad hoc medical advisor and/or speaker on the topics of menopause hormone therapy for Wyeth, Pfizer, and Berlex. During the past 2 yr, S.L.B. has served as an ad hoc medical advisor and/or speaker on the topics of menopause hormone therapy for the following: Foundation for Better Health, Med Pro Communications, The Wright Resource, P-Value Communications, Med Think Communications, Promedica Communications, Haymarket Medical, Medical Marketing Services, Medsite CME, Berlex, Merck, Parexel, PhRMA, QuatRx Pharmaceuticals, Solvay Pharmaceuticals, and Wyeth. All remaining authors have nothing to disclose.
First Published Online April 17, 2007
Abbreviations: BMI, Body mass index; CV, coefficient of variation; E1c, estrone conjugate; PCOS, polycystic ovarian syndrome; Pdg, pregnanediol glucuronide; q, every.
Received October 18, 2006.
Accepted April 6, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
mediates endotoxin induced suppression of gonadotropin-releasing hormone pulse generator activity in the rat. Endocr J 44:141–148[Medline] and interleukin-6 in the hypothalamic-pituitary gonadal axis. J Neuroendocrinol 13:296–301[CrossRef][Medline] (TNF) inhibits progesterone and estradiol 17ß production from cultured granulose cells: presence of TNF receptors in bovine granulosa and theca cells. J Reprod Dev 49:441–449[CrossRef][Medline] inhibits follicle-stimulating hormone-induced differentiation in cultured rat granulose cells. Biochem Biophys Res Commun 163:1038–1046[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  R. J. Chang Obesity and the Emergence of Sleep-Wake Gonadotropin Secretion in Girls during Early Pubertal Development J. Clin. Endocrinol. Metab., April 1, 2009; 94(4): 1094 - 1096. [Full Text] [PDF]
|
|
|
|
|
|
B. Bordini, E. Littlejohn, and R. L. Rosenfield Blunted Sleep-Related Luteinizing Hormone Rise in Healthy Premenarcheal Pubertal Girls with Elevated Body Mass Index J. Clin. Endocrinol. Metab., April 1, 2009; 94(4): 1168 - 1175. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Santoro, S. L. Crawford, W. L. Lasley, J. L. Luborsky, K. A. Matthews, D. McConnell, J. F. Randolph Jr., E. B. Gold, G. A. Greendale, S. G. Korenman, et al. Factors Related to Declining Luteal Function in Women during the Menopausal Transition J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1711 - 1721. [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 |
