| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Article |
Departments of Endocrinology and Metabolic Diseases (E.W.C.M.v.D., F.R.), General Internal Medicine (A.E.M., H.P.), Clinical Chemistry (M.F.), and Obstetrics, Gynecology and Reproductive Medicine (F.H.H.), Leiden University Medical Center, 2300 RC Leiden, The Netherlands; and Division of Endocrinology, Department of Internal Medicine, General Clinical Research Center, Specialized Cooperative Center for Reproduction Research and Center for Biomathematical Technology, University of Virginia Medical School (J.D.V.), Charlottesville, Virginia 22908
Address all correspondence and requests for reprints to: Hanno Pijl, M.D., Department of Internal Medicine, Leiden University Medical Center, C1-R39, P.O. Box 9600, 2300 RC Leiden, The Netherlands. E-mail: . h.pijl{at}lumc.nl
Abstract
Obesity is associated with considerably reduced plasma GH concentrations, which may contribute to anovulation in (obese) women with polycystic ovary disease (PCOS). This clinical investigation was undertaken to establish whether the GH release process is deranged in obese women with PCOS and, if so, whether the observed anomalies are features of the syndrome or a sequel of body fat accretion. To this end we sampled 24-h plasma GH concentration profiles at 10-min intervals in 15 obese PCOS patients [mean age, 29 yr (range, 20–38); percent body fat, 47 ± 5.2%], 15 equally obese controls with regular menstrual cycles [age, 34 yr (range, 20–44); percent body fat, 48 ± 4.9%], and 15 healthy age-matched lean controls [age, 34 yr (range, 21–45); percent body fat, 29 ± 9.0%]. Compared with lean controls, obese PCOS patients exhibited a greater than 60% reduction in basal and a greater than 75% reduction in pulsatile and total daily GH secretion due to a 2.7-fold attenuation of burst mass and a lesser (1.4-fold) slowing of GH pulse frequency. The mean ± SEM number of statistically significant GH peaks was 13.9 ± 1.2/24 h, the endogenous GH half-life was 14.1 ± 0.4 min, basal GH secretion was 5.0 ± 0.7 mU/liter·24 h, and total secretion was 61.4 ± 9.6 mU/liter·24 h in obese women with PCOS. None of these parameters differed from those in the body mass index-matched controls. The approximate entropy ratio was significantly increased in obese women (both PCOS and controls), indicating greater irregularity of the GH release process. Total GH secretion in patients and the two control groups correlated strongly and negatively with percent body fat (r = -0.775; P < 10-8). Serum concentrations of IGF-I and IGF-binding protein-3 were higher in patients with PCOS than in obese controls (P = 0.03 and P = 0.02, respectively), but the IGF-1/IGF-binding protein-3 ratio was equivalent in all three study groups. In conclusion, the profoundly reduced and irregular GH release in obese women with PCOS appears to be a corollary of body fat accretion and not of the syndrome per se.
POLYCYSTIC OVARY SYNDROME (PCOS) is characterized by inappropriate release of LH, increased ovarian androgen production, and chronic anovulation (1). The mechanism of anovulation in PCOS is uncertain. Follicle development is somehow arrested at an early stage (2). Several (neuro)endocrine hypotheses have been put forward to explain this pathophysiology (2, 3).
Among other considerations, anomalies of plasma GH secretion and/or altered IGF-I concentrations may play a role in the pathogenesis of PCOS (4, 5). Abdominal obesity, which can exacerbate the insulin resistance and reproductive features of the syndrome (6, 7), is associated with profoundly reduced and disorderly GH secretion (8, 9). As in vitro and in vivo evidence supports a stimulatory role for GH in early and later stages of folliculogenesis and ovulation (10, 11, 12), hyposomatotropism may contribute to impaired follicular development and anovulation in PCOS.
Plasma GH concentrations in PCOS have been reported as reduced, normal, or increased (13). The majority of studies in PCOS have evaluated the status of the somatotropic axis on the basis of acute secretagogue-stimulated GH release. Many papers report a blunted GH response to a variety of stimuli in women with PCOS compared with weight-matched controls (14, 15, 16, 17). Relatively few analyses report frequently sampled spontaneous plasma GH concentrations as an index of neuroendocrine dysregulation in PCOS. The first of these papers documents reduced average diurnal (0900–2300 h) GH levels in PCOS women, even after accounting for body mass index (BMI) (4). A second paper described increased 24-h GH levels in lean vs. obese PCOS patients or lean controls (obese controls were not included) (18). However, both studies were hampered by infrequent (hourly or less) blood sampling, which precludes more precise appraisal of pulsatile dynamics. A third more recent study employed the Cluster discrete pulse detection algorithm to monitor a 24-h, frequently (10-min interval) sampled, plasma GH concentration-time series in PCOS vs. weight-matched premenopausal control women. GH pulse amplitude was higher, whereas pulse frequency was similar in lean PCOS vs. lean healthy women (19). In obese PCOS patients the GH pulse amplitude (but not pulse frequency) was considerably reduced compared with that in lean, but not obese, control subjects (19). Unfortunately, only less sensitive GH assays were available at the time of this study, which limits the detection of (low amplitude) pulses. Taken together, the available data on GH neuroregulation in (obese) PCOS patients are inconclusive.
The present study explores the dynamics of spontaneous GH secretion in obese women with PCOS using a high sensitivity immunofluorometric assay, frequent and extended blood sampling, controls matched for BMI, and contemporary techniques of secretory (pulsatile and basal) analysis and regularity [approximate entropy (ApEn)] quantitation. We specifically aimed to establish whether the GH release process is deranged in obese women with PCOS and, if so, whether the observed anomalies are features of the syndrome itself or merely a sequel of body fat accretion.
Subjects and Methods
Subjects
The characteristics of the participating subjects are shown in Table 1
. Fifteen obese women (mean age, 29 yr; range, 20–38) with the diagnosis of PCOS were studied. PCOS women were recruited from the out-patient department for reproductive medicine. The diagnosis of PCOS was based on the presence of infertility due to anovulation, which is not secondary to a specific underlying disease of the pituitary gland, ovaries, thyroid, or adrenal glands, and elevated serum testosterone concentration in the absence of hyperprolactinemia. Late-onset adrenal hyperplasia was excluded by a morning serum 17-hydroxyprogesterone level less than 10 nmol/liter. The control groups comprised 15 age- and BMI-matched women (age, 34 yr; range, 20–44) and 15 age-matched lean women (age, 34 yr; range, 21–45) with regular menstrual cycles. Controls were recruited through advertising in local newspapers. None of the subjects had used a neuroactive drug (including oral contraceptives) for at least 3 months before the study. All women had stable body weight for at least 3 months before the study. The purpose, nature, and possible risks of the study were explained to all subjects, and written informed consent was obtained. The study protocol was approved by the ethics committee of Leiden University Medical Center.
|
Anovulatory or oligoovulatory patients with PCOS were studied on a random day. Control women were studied in the early follicular phase of their menstrual cycle (d 2–7). All subjects were admitted to the Clinical Research Center of Leiden University Medical Center at 0800 h after a 10-h overnight fast.
On the morning of admission body weight was measured. A single venous blood sample was withdrawn at least 30 min after insertion of a forearm intravenous cannula at 0800 h for baseline serum hormone measurements, including total testosterone, estradiol, SHBG, IGF-I, and IGF-binding protein-3 (IGFBP-3) concentrations. At least 1 h after cannulation, blood samples were drawn into heparinized tubes every 10 min for 24 h for determination of the GH concentration. Samples were centrifuged within 1 h of sampling, and plasma was stored at -20 C until assay. All women were studied according to the same protocol. Meals were served at 0900, 1400, and 1900 h. Subjects were allowed to walk around inside the research center during the day. Lights were put out at 2300 h.
Bioelectrical impedance analysis
The percent body fat was measured by bioelectrical impedance analysis (Bodystat Ltd., Douglas, Isle of Man, UK) before each 24-h hormone profile. The impedance measurements were obtained on the morning of the study after an overnight fast, after the subjects had voided and while they were resting in bed.
Assays
Plasma GH concentrations were measured in duplicate using a sensitive, time-resolved, immunofluorometric assay (Wallac, Inc., Turku, Finland) specific for the 22-kDa GH protein. Human biosynthetic GH (Pharmacia Biotech & Upjohn, Inc., Uppsala, Sweden) was used as standard and calibrated against WHO-IRP 80-505, with a detection limit of 0.03 mU/liter and intraassay variation coefficients of 1.6–8.4% at plasma values between 0.25–40 mU/liter (to convert milliunits per liter to micrograms per liter, divide by 2.6). All samples from any subject were run in the same assay.
Serum IGF-I concentrations were determined by RIA (INCSTAR Corp., Stillwater, MN) with a detection limit of 1.5 nmol/liter and an interassay coefficient of variation of less than 11%. Serum IGFBP-3 was measured by RIA (Nichols Institute Diagnostics, San Juan Capistrano, CA), with a detection limit of 0.08 mg/liter and an interassay coefficient of variation less than 6.8% at various plasma levels.
The serum total testosterone concentration was measured by coated tube RIA (Diagnostic Products, Los Angeles, CA). Serum SHBG concentrations were measured by coated tube immunoradiometric assay (Spectria, Espoo, Finland). Estradiol was measured by RIA (Diagnostic Systems Laboratories, Inc., Webster, TX).
Analytical techniques
A largely model-free computerized peak detection algorithm was used to quantify statistically significant GH pulses in relation to dose-dependent measurement error (20). A 2 x 2 test cluster size was used to test for rises and falls in the data, and t statistics of 2.0 were used to test for significant up- and down-strokes of GH peaks.
A multiparameter deconvolution technique was used to estimate relevant measures of GH secretion from the 24-h serum GH concentration profiles, as described previously (21). Initial estimates of the basal GH secretion rate were calculated to model the mean of the lowest 5% of all plasma GH concentrations in the time series. Peak detection entailed application of 95% statistical confidence intervals to two thirds of all GH secretory peaks considered jointly and individual 95% statistical confidence intervals to the remaining one third smaller pulses. These criteria yield a specificity and sensitivity of 90% or more. The following specific measures of GH secretion were estimated: secretory burst frequency (number of statistically significant pulses per 24 h), amplitude (maximal rate of calculated GH secretion attained within a release episode), mass (amount of hormone secreted per burst per unit distribution volume), half-duration (time in minutes elapsing during a calculated secretory episode at half its maximal amplitude), basal GH secretion rate (milliunits per liter per minute) and the endogenous GH half-life (minutes). These values were used to calculate daily pulsatile and total GH secretion.
The sample to sample regularity or serial orderliness of GH secretion was quantitated by the ApEn statistic, a scale- and model-independent technique (22). Normalized ApEn parameters of m = 1 and r = 20% of the intraseries SD were applied, as previously validated for GH concentration-time series of this length (23). This member of the ApEn set is designated ApEn (1,20%). ApEn quantifies the regularity of subordinate (nonpulsatile) patterns in the data and as such yields information complementary to peak detection and deconvolution techniques (24). The ApEn ratio, as reported in this paper, delineates the orderliness of a concentration time series vs. the ApEn score associated with maximal chaos in the same time series as estimated by computer simulation. Thus, an ApEn ratio of 1 reflects maximal chaos in the time series under study, and smaller ApEn ratios denote greater regularity.
Statistical analysis
Results are expressed as the mean ± SEM unless indicated otherwise. Differences between patients and the control groups were estimated by ANOVA and post hoc contrast testing. In addition, regression techniques were applied to evaluate the relation between variables. Data were transformed logarithmically when necessary. Statistical analysis was performed using Systat (release 10, SPSS, Inc., Chicago, IL). Differences were considered significant for P < 0.05.
Results
Cluster analysis
All features of the pulsatile plasma GH concentration profile were reduced in obese compared with normal weight women, whereas none differed between obese women with and without PCOS (Table 2). In addition, the mean intervalley GH concentration was significantly higher in lean controls than those in obese controls and PCOS patients.
|
Representative 24-h plasma GH concentration and secretion profiles for one obese women with PCOS, one obese control, and one lean control subject are shown in Fig. 1. Pulsatile GH secretion was decreased by 4-fold in patients compared with lean controls as a result of the concerted effects of diminished mean GH pulse duration and amplitude (and hence GH pulse mass) with a smaller decrease in pulse frequency. Basal GH secretion in patients and obese controls was about 3-fold lower than that in lean controls (Table 3).
|
|
).
Daily GH secretion rates in patients and controls were not correlated with circulating concentrations of IGF-I and IGFBP-3 or with their ratio (Table 4). However, the daily GH secretion rate was significantly negatively correlated with percent body fat, as illustrated in Fig. 2.
|
|
Although IGF-I and IGFBP-3 were slightly increased in PCOS vs. obese controls, the IGF-I/IGFBP-3 ratio was similar among groups (Table 4).
ApEn
The ApEn ratio was significantly increased in PCOS women compared with lean controls, which reflects a more irregular GH release process (Table 3). However, obese control women maintained an equivalently disorderly GH release pattern as obese PCOS patients.
Discussion
The present study documents a profound reduction of daily basal (62%) and pulsatile (76%) GH release in obese women with PCOS vs. healthy lean controls. The reduction of pulsatile GH secretion could be explained by a marked (2.7-fold) diminution of secretory burst mass and a lesser (1.4-fold) decline in burst frequency. Elevated GH ApEn values also suggest less orderly GH release in obese PCOS women than in lean controls. Despite these differences in GH secretion, the plasma IGF-I concentration and the IGF-I/IGFBP-3 ratio were not different in obese PCOS vs. normal weight control women. Further, all the principal features of GH secretion were similar in the patient cohort and equally obese women without PCOS, and GH secretion correlated strongly (and inversely) with body fat percentage, which indicates that the distinct alterations in GH neuroregulation in obese PCOS patients relate to body fat accretion and not to PCOS per se.
The present findings are apparently inconsistent with a previous paper that reported lower average daily plasma GH concentrations in women with PCOS than in weight-matched controls (4). However, infrequent blood sampling (every 2 h) and different patient cohorts and GH assays limit a direct comparison with the current data (25). A more recent study disclosed similar plasma GH concentration pulse characteristics, as determined by the Cluster peak detection algorithm, in eight obese women with PCOS and equally obese controls, whereas GH pulse amplitude (but not pulse number) were reduced compared with values in lean controls (19). Our results corroborate these observations and provide additional insight into the underlying kinetic mechanism. The reduced plasma GH concentration in obese women with PCOS appears to be fully attributable to diminished secretion, as the calculated plasma GH half-life was not increased. Analogous characteristics of pulsatile GH secretion have been observed in healthy (upper body) obese men and women in several previous studies (9, 26, 27), which further minimizes any contribution of the abnormal sex steroid milieu to diminished GH secretion in PCOS and underlines the concept that hyposomatotropism in obese women with the syndrome is a sequel of body fat storage.
The stage of the menstrual cycle critically affects GH secretion in women (28). Plasma GH concentrations typically double during the late follicular phase as plasma estradiol levels rise (29, 30), which is fully attributable to an enhanced GH secretion rate (31). Our healthy subjects were studied during the early follicular phase of their cycle, rendering their estrogen levels significantly lower than those observed in PCOS. It is conceivable that the obese control subjects would have displayed higher GH secretion rates if they had been studied in a slightly later stage of their cycle (with estrogen levels approaching those in PCOS). Thus, the GH secretion rate in obese women with PCOS may be reduced if studied in the face of plasma estrogen levels similar to those in healthy obese women. However, the above-mentioned studies of GH secretion across the menstrual cycle have all been performed in normal weight women, and the effects of gonadal steroids on GH secretion in obese women remain to be established.
Our data argue against an important role for the somatotropic axis in the pathophysiology of anovulation in PCOS. The mechanism that underlies anovulation in PCOS is unknown. As GH and IGF-I can act as cogonadotropins under some conditions to support orderly folliculogenesis and ovulation (10, 11, 12), and a number of papers reported reduction of somatotropic activity in PCOS (4, 14), it was suggested that anomalies of this neuroendocrine ensemble contribute to the disorganization of ovarian function in PCOS. The results of our investigation suggest that reduced GH secretion in obese women with PCOS is a sequel of body fat accrual rather than part of the syndrome. In view of the fact that anovulation is not a regular feature of obesity per se, it is unlikely that hyposomatotropism significantly contributes to disruption of ovulatory cycles in these women. This inference is in keeping with data showing that adjuvant GH therapy does not affect the dose of gonadotropin required to induce ovulation in PCOS (32). Moreover, normal weight women with PCOS have normal or even increased plasma GH levels (19), which obviously also argues against a primary pathogenetic contribution of hyposomatotropism to anovulation in this syndrome.
Muting of negative feedback restraint increases the serial irregularity of hormone output patterns in a number of neuroendocrine ensembles, including the somatotropic axis (33). In the present study the GH release process was more irregular in obese (PCOS) women compared with their normal weight controls. Viscerally obese women with regular menstrual cycles exhibit similarly irregular GH output (9). Thus, the ApEn statistic of the plasma GH concentration-time series suggests that GH secretion in obese women is blunted in the face of muted negative feedback restraint. As the IGF-I/IGFBP-3 ratio was similar in obese and normal weight women, other negative feedback inputs (such as GH itself) may be reduced to explain the irregularity of GH secretion. If negative feedback restraint is reduced, hyposomatotropism in obese women must be brought about by diminished GHRH/GH-releasing protein/ghrelin feedforward inputs to the pituitary. Thus, our data suggest that the conjoint effects of blunted feedforward drive and negative feedback restraint underlie the anomalous pituitary GH output profiles in abdominal obesity.
One might argue that the obese PCOS patients studied here were not functionally hyposomatotropic because the plasma total IGF-I concentration and IGF-I/IGFBP-3 ratio were normal. However, serum IGF-I levels largely reflect GH’s action on the liver, and GH exerts multiple local tissue effects via direct action (34, 35). Moreover, the plasma IGF-I concentration is influenced by a number of endocrine and metabolic factors other than GH (36, 37). Thus, normal plasma IGF-I levels or a normal IGF-I/IGFBP-3 ratio do not necessarily reflect adequate bioavailability of GH at the level of the ovary.
In conclusion, the present study documents profoundly reduced and disorderly spontaneous basal and pulsatile GH release in obese women with PCOS. As this pattern is also seen in obese controls, the anomaly appears to relate to percentage body fat and not to PCOS per se.
Acknowledgments
We thank Bep Ladan for her help with the experiments, and Eric Gribnau and Sjoukjen Walma for their technical support.
Footnotes
Present address for J.D.V.: Mayo Clinic, St. Mary’s Hospital, Room 5-194, 1216 Second Street SW, Rochester, Minnesota 55905.
Abbreviations: ApEn, Approximate entropy; BMI, body mass index; IGFBP-3, IGF-binding protein-3; PCOS, polycystic ovary disease.
Received December 14, 2001.
Accepted May 10, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  C. Franco, J. D Veldhuis, A. Iranmanesh, J. Brandberg, L. Lonn, B. Andersson, B.-A. Bengtsson, J. Svensson, and G. Johannsson Thigh intermuscular fat is inversely associated with spontaneous GH release in post-menopausal women with abdominal obesity. Eur. J. Endocrinol., August 1, 2006; 155(2): 261 - 268. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Glintborg, R. K. Stoving, C. Hagen, A. P. Hermann, J. Frystyk, J. D. Veldhuis, A. Flyvbjerg, and M. Andersen Pioglitazone Treatment Increases Spontaneous Growth Hormone (GH) Secretion and Stimulated GH Levels in Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., October 1, 2005; 90(10): 5605 - 5612. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Misra, K. K. Miller, J. Bjornson, A. Hackman, A. Aggarwal, J. Chung, M. Ott, D. B. Herzog, M. L. Johnson, and A. Klibanski Alterations in Growth Hormone Secretory Dynamics in Adolescent Girls with Anorexia Nervosa and Effects on Bone Metabolism J. Clin. Endocrinol. Metab., December 1, 2003; 88(12): 5615 - 5623. [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 |
), obese controls (•), and lean controls (
). Data were available for 15 patients, 15 BMI-matched controls, and 8 lean controls. The correlation coefficient was -0.775 (P < 10-8).