| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Department of Gynecological Endocrinology, University of Brescia, Brescia 25125, Italy
Address all correspondence and requests for reprints to: Dr. Alessandro Gambera, Via XXV Aprile, n°10, 24058 Romano di Lombardia (BG), Italy.
Abstract
This study evaluated the prognosis of functional hypothalamic amenorrhea (FHA) and the predictive factors of recovery, through a long-term follow-up. Ninety-three women affected by FHA underwent a follow-up for an average period of 8.1 yr (range 7–9 yr). At the end of the follow-up, 65 (70.7%) patients recovered. Statistical analysis showed that there was no association between recovery and anamnestic causes of FHA or with the echographic ovarian morphology but identified the predictive factors of recovery as the basal body mass index (BMI), the basal cortisol, and androstenedione plasma levels. A higher basal BMI and A, and lower cortisol values are positive prognostic factors for the recovery. Also the BMI, acquired during the follow-up, is important for FHA resolution: in fact, in recovered women the BMI increased or remained stable, whereas in nonrecovered women it decreased or remained stable. At the end of the follow-up, 52 (74.3%) patients treated with hormone replacement therapy and 8 (80%) with no therapy recovered, but only 5 (41.7%) with oral contraceptive pills recovered.
FUNCTIONAL HYPOTHALAMIC AMENORRHEA (FHA) is an endocrine disorder due to a deficiency of pulsatile GnRH secretion, not related to hypothalamus-pituitary organic lesions, to endocrine (hyperprolactinemia, hypo/hyperadrenal activity, hypo/hyperthyroidism) or systemic diseases (1, 2). Based on the degree of GnRH suppression, these patients can present different clinical profiles that vary from an inadequate luteal phase, to anovulation with various menstrual irregularities, to hypothalamic amenorrhea (2).
Patients with FHA have low or normal gonadotropin and estrogen levels (i.e. similar to those seen in the follicular phase of a normal menstrual cycle), low or normal PRL levels, increased cortisol secretion and a normal imaging evaluation of the Sella Turcica. Medroxyprogesterone acetate administration (10 mg/d orally for 5 consecutive days) does not usually induce menstrual bleeding (2).
In literature, the incidence of FHA ranges from 15% (3) to 48% (4) of the secondary amenorrheas. In clinical practice, FHA is often associated with stressors (metabolic, physical, or psychological), with weight loss due to decreased caloric intake or intensive physical exercise (1). In fact, several studies have established how competitive or prolonged physical exercise, weight reduction, and psycho-social stress may be responsible, through different mechanisms, for the decrease in GnRH and gonadotropin pulsatile secretion (5, 6, 7).
Studies on the secretory patterns of gonadotropins, especially LH, in women with FHA, have shown either an increase of frequency with small amplitude (1), or a reduction of frequency and amplitude (8), or a reduction of frequency with variable amplitude at the beginning of this dysfunction (9).
The mechanisms responsible for FHA have not been definitively determined. Numerous neuroendocrine abnormalities seem to be implicated in the pathogenesis of FHA, such as the hypothalamus-pituitary-adrenal axis hyperfunction with an increased secretion of CRH, ACTH, cortisol (3, 10), and endogenous opioids (11), an increase of the dopaminergic tone (12) and of the nocturnal melatonin secretion (13). Some studies have shown that insulin and GH/IGFs/IGF binding protein (IGFBP) system could have a role in the pathogenesis of FHA, through an increase of IGFBP-1 and a decrease of IGFs and insulin activity (6). Furthermore, in women with FHA due to a reduction of the body weight, especially caused by physical exercise, low average leptin levels during 24 h were documented (14, 15). A reduction of 10% in body weight is associated with a 53% reduction in serum leptins (16). There is evidence that leptin levels are regulated by the energetic balance, independently from body fat reserves (17). Leptin reductions, due to the increase of NPY, could increase ACTH and cortisol levels, thus interfering with GnRH pulsatility (18). Furthermore, weight loss itself induces a cortisol increase (19) and cortisol can suppress gonadotropin secretion from the pituitary (20).
The aim of our study is to evaluate, through a long-term follow-up, the prognosis of FHA and to identify the predictive factors of recovery.
Materials and Methods
Among patients admitted to the Department of Gynecological Endocrinology of the University of Brescia from 1990–1992, 93 were found to have FHA, and all were included in our study. FHA was diagnosed by clinical (amenorrhea for a period 
6 months), hormonal findings (Table 1
), and the negativity of the radiological evaluation of the Sella Turcica. Medroxyprogesterone acetate administration (10 mg/d orally for 5 consecutive days) that was carried out before admission did not induce menstrual bleeding in any patient.
|
No patient had polycystic ovary syndrome, premature ovarian failure, hyperprolactinemia, hypo/hyper adrenal activity, hypo/hyperthyroidism, anorexia nervosa, a history of drug abuse, or had used any medication in the 6 months preceding admission. Anorexia nervosa was excluded because no patient weighed less than 85% of their ideal body weight, or had a BMI less than 17, and none responded to the Diagnostic and Statistical Manual of Mental Disorders (DSM IV) criteria (21).
All women underwent a complete history to individualize the possible causes of FHA, a clinical examination, ematochemical and hormone assays, an ultrasonographic investigation and a radiological evaluation of the Sella Turcica.
Hormonal assays were assessed in fasting conditions between 0800 h and 0900 h in duplicate in the same assay for 2 consecutive days, except for PRL assayed in supine position at 0930 and 1030 h.
All patients underwent the following assays: LH, FSH, PRL, androstenedione (A), total (T) and free (FT) testosterone, 17ß-E2, cortisol, dehydroepiandrosterone sulfate (DHEAS), insulin (I), IGF-I, TSH, free T3 (FT3) and T4 (FT4). Furthermore, the sex hormone-binding globulins (SHBG) and IGFBP-3 were assessed.
Transabdominal or transvaginal pelvic ultrasonographic investigation was carried out using Ultrasound System and 6.5 MHz vaginal transducer with 160-degree sector angle focused at 3 cm. Ovarian morphology was evaluated according to Adams’ criteria (22). All patients gave written informed consent. The study was approved by the Ethical Committee of the University of Brescia.
After the diagnosis, during a physician-patient personalized interview, we furnished guidelines to correct the nutritional status either qualitatively or quantitatively, and/or to moderate the patients physical exercise, to equilibrate the energetic balance.
The 93 patients underwent a clinical follow-up, for an average period of 8.1 yr (range 7–9 yr), with a check-up every 6–12 months. During the follow-up, the body weight, the BMI, the compliance to the therapy and the possible appearance of spontaneous menstrual cycles and/or pregnancies were evaluated. The last check-up was performed in February 2000. During the follow-up, one patient died in a road accident and was thus excluded from the casistic.
Patients underwent two kinds of therapy, taking into consideration their contraceptive desires:
1) Hormone replacement therapy (HRT): the estrogens used were conjugated estrogens (Premarin, Wyeth Lederle, Milan, Italy) at a dosage of 0.625 mg/d for 21 d. The progestogens used were dydrogesterone 10 mg/d or medroxyprogesterone acetate 5 mg/d in the last 11 d (Dufaston, Solvay Pharmaceuticals, Inc., Torino, Italy or Provera G, Pharmacia \|[amp ]\| Upjohn, Inc., Milan, Italy, respectively). A pause of 7 d was observed after the 21 d. Both dydrogesterone and medroxyprogesterone acetate are progestins 21-carbon derivates, and at a dosage of 10 mg/d and 5 mg/d respectively, are estimated comparably (23).
2) Oral contraceptive pills (OC): monophasic combination of ethinyl-estradiol 0.030 mg/d and desogestrel 0.150 mg/d (Practil 21, Organon, Pomezia, Italy) or Gestodene 0.075 mg/d (Ginoden, Schering, Milan, Italy). Both of them belong to the third generation of OCs.
Ten (10.9%) women (BMI 19.1 ± 2.3 kg/m2) refused both hormone therapies; therefore they were included in a group defined no therapy.
HRT was administered to 70 (76.1%) patients (BMI 19.8 ± 2.3 kg/m2) for 6 consecutive cycles, followed by a pause of 2 months. OC was administered to 12 (13.0%) women (BMI 20.6 ± 2.3 kg/m2) for 12 consecutive cycles and then suspended for 2 months. During the pause, observed after each cycle of therapy, we checked the eventual recovery, that is the possible appearance of spontaneous ovulatory menstrual cycles and/or pregnancies. Also patients with restored spontaneous menstrual cyclicity or pregnant underwent the follow-up until February 2000.
The control group for hormones and for the SHBG and IGFBP-3 examination was made up of 40 women with a mean age of 23.5 ± 3.7 yr, mean BMI of 22.6 ± 2.7 kg/m2 and with regular ovulatory menstrual cycles. Samples were carried out in the early follicular phase of the menstrual cycle (fourth to seventh day of the cycle).
Statistical analysis
All values were expressed as mean ± SD. Student’s unpaired t test was used to compare the hormone levels of the controls with the women with FHA, and of recovered and nonrecovered patients with FHA. The paired t test was used to compare BMI levels at baseline and at the end of the follow-up. The Spearman test was used to verify the association between LH and cortisol levels.
The association between recovery and causes of FHA, echographic ovarian morphology, BMI variation or therapy was analyzed using the 
-square test. To find the most important prognostic factors on recovery, we performed an univariate logistic regression, as a first step, using the following parameters as independent variables: age, period of amenorrhea prediagnosis, causes of FHA, echographic ovarian morphology, basal BMI and BMI at the end of the follow-up, therapies, all hormones and carrier proteins. Then, the statistically significant variables were used in the multivariable logistic regression. The selection of the most important predictors for recovery was carried out using the Akaike Information Criterion.
A P value < 0.05 was considered statistically significant. The statistical package STATA (STATA Corp., College Station, TX) was used to perform the statistical analysis.
Hormonal assays
Plasma LH, FSH, and SHBG levels were determined using an immunoradiometric assay method (Radim, Pomezia, Rome, Italy), while the remaining hormones were tested using RIA commercial kits from: Diagnostic Products Corp. (Los Angeles, CA) (A; T; FT; cortisol, I); Ortho-clinical Diagnostics, Amersham, UK (PRL, E2, TSH, FT3, FT4); Immunotech, Marseille, France (DHEAS); plasma IGF-I and IGFBP-3 levels were tested using a radioisotopic assay method (Nichols Institute Diagnostics, San Juan Capistrano, CA).
The average intra-interassay coefficients of variation were: 2.6% and 3.8% for LH, 7.1% and 7.2% for FSH, 8.1% and 8.7% for PRL, 4.3% and 6.0% for A, 3.1% and 7.0% for T, 6.2% and 7.8% for FT, 7.1% and 7.8% for E2, 6.8% and 7.1% for cortisol, 7.6% and 7.8% for DHEAS, 4.6% and 5.3% for I, 5.2% and 5.7% for IGF-I, 5.0% and 5.5% for TSH; 3.1% and 4.2% for FT3; 3.0% and 3.8% for FT4, 5.0% and 7.5% for SHBG, 3.8% and 6.3% for IGFBP-3, respectively.
Results
Table 1
shows the basal endocrine profile of patients with FHA compared with normal women (control group). Patients with FHA have significantly lower gonadotropins, PRL, E2, and TSH levels and higher cortisol levels than controls. A, T, FT, DHEAS, I, IGF-I, FT3, FT4, SHBG, and IGFBP-3 are not statistically different.
Anamnestic factors associated with the development of FHA (Table 2) are: psycho-social stress, psychological stress with secondary weight loss, voluntary weight loss by dieting, physical stress with weight loss, and competitive sporting activity. In particular, in this last group, two patients were professional cyclists, two were long-distance runners, and one was a professional gymnast. It is evident that stress and weight loss are the principal factors associated with FHA in our casistic. At the end of the follow-up, there was a recovery in 65 (70.7%) FHA patients. The data analysis showed that there was no association between recovery and anamnestic causes (P = 0.733).
|
indicates that age and period of amenorrhea preceding diagnosis (indicated as number of months) did not significantly differ between recovered and nonrecovered patients, whereas the basal BMI was higher in the group of the recovered women (P < 0.05).
|
. Only cortisol and A plasma levels were significantly different within the 2 groups: cortisol is higher in nonrecovered patients (576.6 ± 82.8 vs. 518.7 ± 93.8 nmol/liter) whereas A in the recovered (5.9 ± 1.4 vs. 5.2 ± 1.4 nmol/liter). Furthermore, it is possible to note that the highest cortisol levels correlate with the lowest LH levels (r = -0.658; P < 0.01) and that the absolute values of I and IGF-I are higher in the recovered women.
|
). In the group of the 65 recovered women, 33 (50.8%) increased their BMI, 32 (49.2%) maintained it unmodified, and nobody showed a decrease. In the 27 nonrecovered patients, 4 (14.8%) reduced their BMI, 23 (85.2%) maintained it unmodified and nobody showed an increase. Overall, the BMI in recovered women varied from 20.2 ± 2.4 kg/m2 to 21.4 ± 1.6 kg/m2 at the follow-up, with an average weight increase of 6.1 kg (range 3–13 kg). In nonrecovered patients, the BMI went from 19.0 ± 2.1 kg/m2 to 18.6 ± 2.3 kg/m2, with an average weight decrease of 5.8 kg (range 3–9 kg). The basal BMI, separated from the other variables, has a notable impact on the prognosis of FHA (Table 6): each increase of 1 kg/m2, increases the recovery probability by 24.7% (Univariate Logistic Regression Model). In fact, we have seen that the subjects who improve their own BMI during the follow-up, have about twice the probability of recovering than those who do not.
|
|
At the end of the follow-up, in the group of patients receiving HRT, 52 (74.3%) women recovered, in the group of OCs 5 (41.7%) women recovered, whereas among the group no therapy, 8 (80%) spontaneously recovered.
In the recovered patients, HRT has been meanly administered for 15.6 (range 6–56) cycles, OC for 21.6 (range 12–48) cycles.
The mean recovery time was of 25.2 ± 17.8 months for patients treated with HRT, 33.6 ± 15.6 months for those with OC and 16.5 ± 8.9 months for patients with no therapy.
Although it is not possible to establish in this study a causality between recovery and therapy prescribed, we found that patients in the group of OC have a recovery probability of 8% compared with those in the HRT group (odds ratio = 0.079; P = 0.002). Women who were in the group no therapy have the same recovery probability than those in the group of HRT (odds ratio = 3.148; P = 0.291).
Among the recovered patients during follow-up, 14 spontaneous pregnancies developed in 12 women: 2 patients had 2 pregnancies and 2 patients a twin pregnancy. Thirteen pregnancies developed in the group treated with HRT and one in the group of no therapy.
The final Multivariable Logistic Regression Model is shown in Table 7. Contemporarily considering all variables together, thus influencing each other, the factors predictive of recovery are the basal BMI, the basal cortisol and A plasma levels. A higher BMI and A, and lower cortisol values are positive prognostic factors for the resolution of FHA. In particular, each increase of 1 kg/m2 of the BMI, increases the recovery probability by 34%, each cortisol level reduction of 27.6 nmol/liter increases the recovery probability by 24%, and each increase of 0.35 nmol/liter of A, within the range of normality, is associated with a higher recovery probability of 22%.
|
shows the relationship between the basal BMI and the recovery probability (adjusted to cortisol, A, and therapy).
|
Our study demonstrated that FHA is reversible, with recovery in 70.7% of cases over an average period of 8.1 yr. In basal conditions, recovered patients presented lower cortisol (518.7 ± 93.8 vs. 576.6 ± 82.8 nmol/liter) and higher A plasma levels (5.9 ± 1.4 vs. 5.2 ± 1.4 nmol/liter) than nonrecovered patients. Furthermore, the basal BMI of recovered women was significantly higher than nonrecovered women (20.2 ± 2.4 vs. 19.0 ± 2.1 kg/m2).
At the end of the follow-up, 52 (74.3%) patients treated with HRT, 5 (41.7%) with OC and 8 (80.0%) with no therapy recovered. During the follow-up, overall, the 65 recovered patients presented an increase of BMI (from 20.2 ± 2.4 to 21.4 ± 1.6 kg/m2), whereas the 27 nonrecovered decreased their BMI (from 19.0 ± 2.1 to 18.6 ± 2.3 kg/m2). In particular, 33 recovered patients showed a significant increase in their BMI (from 18.6 ± 1.8 to 21.0 ± 1.3 kg/m2) and 4 nonrecovered showed a significant decrease (from 19.5 ± 1.3 to 17.0 ± 1.8 kg/m2). The BMI was not modified in the remaining 32 recovered and 23 nonrecovered patients. The recovery or the persistence of FHA in these patients could be attributed either to a significantly different basal BMI (21.8 ± 1.7 vs. 18.9 ± 2.2 kg/m2), or to qualitative variation in the diet with an amelioration of the energetic balance. Also the basal nutritional indices (FT3, IGF-I), in terms of absolute values, were better in recovered than nonrecovered women (Table 4).
On the basis of our results and due to the different numbers of the 3 samples, it is not easy to establish if the therapies have influenced the BMI variations during the follow-up. In fact, in the recovered patients, both the HRT (18.5 ± 1.6 vs. 20.9 ± 1.2 kg/m2; P < 0.001) and the no therapy (18.8 ± 2.6 vs. 21.7 ± 1.8 kg/m2; P < 0.005) are associated with an increase in BMI. From the data of Table 5 the concept emerges that in recovered women the BMI increases or remains stable, whereas in nonrecovered subjects BMI never increases. Thus, it is evident that the basal BMI and/or the BMI acquired during the follow-up plays a fundamental role in the resolution of FHA. The final multivariable logistic model (Table 7) shows that a greater recovery probability corresponds to a greater value of basal BMI (odds ratio = 1.34).
Our findings confirm the importance of the CRH-ACTH-cortisol axis in FHA (3, 10). Independently from the causes associated with FHA, the highest cortisol levels are inversely proportional to the recovery probability (Odds ratio = 0.76). The hypothesis is strengthened in that different plasma cortisol levels are associated with a different degree of alteration of the hypothalamus-pituitary-ovarian axis and of the pulsatility of GnRH and gonadotropins, especially LH. Several studies have shown that CRH produces a dose-dependent decrease in GnRH release from the mediobasal hypothalamus in vitro and, in rats, the intraventricular administration of CRH leads to a reduction of LH plasma levels (24, 25).
Saketos et al. (20) demonstrated that the administration of glucocorticoids in normal women can suppress gonadotropin secretion from the pituitary, independently from CRH action on GnRH in the hypothalamus.
Our study shows that in FHA, the basal PRL plasma levels are significantly lower than controls. It is possible that the increase of the central dopaminergic tone could constitute another neuroendocrine mechanism responsible for GnRH pulse frequency inhibition (26). The hypoprolactinemia could also be due to an increase in the opiod tone, to the decreased hypothalamic TRH and to the hypoestrogenic state (27, 28).
In contrast with the majority of authors (1, 5, 29), we observed a significant reduction in TSH (-18.2%), without significant FT3 (-7.5%) and FT4 (-6.7%) variations in patients with FHA. These findings could be explained by the causes associated with FHA in our casistic, with an elevated incidence of psychological causes and low incidence due to physical activity.
It is not easy to explain the behavior of A: its plasma levels do not differ from controls, but they are significantly higher in recovered patients than those with persistent FHA. In literature, the behavior of A in FHA is unclear: some authors have reported a parallel increase of A along with cortisol (12), others a decrease of A due to the inhibition of ovarian steroidogenesis (30). We think that the higher A levels in the recovered patients, associated with the lowest cortisol and the highest LH levels, could signify an ovarian origin.
In conclusion, FHA is often induced by psychological stress or by increased physical activity that, along with improper diets, may cause alterations of the energetic balance and BMI variations.
It is evident that metabolic-endocrine peripheral impulses (leptin, IGF-I), able to modify the central neuronal functions (NPY, glucagon-like peptide-1 and pro-opiomelanocortin systems, CRH, galanin, etc.) and the reproductive axis, can play a fundamental role in FHA onset (31).
It is possible that for the resolution of FHA, the elimination of stressor factors and the correction of the energetic balance, either through a qualitative and quantitative modification of the diet, or through a reduction of physical activity may be important.
Acknowledgments
We gratefully thank Prof. Giovanni Parrinello, of the Chair of Medical Statistics and Biometry of the University of Brescia, for his precious statistical assistance and Prof. Sturgeon Blanche, of the Chair of Medical English of the University of Brescia, for her linguistic assistance.
Footnotes
Abbreviations: A, Androstenedione; BMI, body mass index; DHEAS, dehydroepiandrosterone sulfate; FHA, functional hypothalamic amenorrhea; FT, free testosterone; FT3, free T3; FT4, free T4; HRT, hormone replacement therapy; IGFBP, IGF binding protein; OC, oral contraceptive pills; SHBG, sex hormone-binding globulins; T, total testosterone.
Received April 16, 2001.
Accepted October 19, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  S. F. Abraham, B. Pettigrew, C. Boyd, and J. Russell Predictors of functional and exercise amenorrhoea among eating and exercise disordered patients Hum. Reprod., January 1, 2006; 21(1): 257 - 261. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Andrico, A. Gambera, C. Specchia, C. Pellegrini, L. Falsetti, and E. Sartori Leptin in functional hypothalamic amenorrhoea Hum. Reprod., August 1, 2002; 17(8): 2043 - 2048. [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 |
