- Email: [email protected]
Pulsatile luteinizing hormone secretion in hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease during naltrexone treatment
r FERTILITY AND STERILITY
Vol. 57, No.4, April 1992
Copyright ~ 1992 The American Fertility Society
Printed on acid-free paper in U.S.A.
Pulsatile luteinizing hormone secretion in hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease during naltrexone treatment Magda C. Armeanu, M.D., Ph.D.* Gerda M. J. Berkhout, R.N. Joop Schoemaker, M.D., Ph.D.t Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Fertility, Free University Hospital, Amsterdam, The Netherlands
Objective: To determine if chronic treatment with the long-acting oral opioid antagonist naltrexone can increase luteinizing hormone (LH) and follicle-stimulating hormone (FSH) secretion in women with secondary amenorrhea. Design: Prospective. Setting: Large reproductive endocrinology unit of an academic hospital. Patients: Three groups of women with oligomenorrhea or amenorrhea: (1) hypothalamic amenorrhea; (2) anorexia nervosa; and (3) polycystic ovarian disease (PCOD). Intervention: Naltrexone 50 mg every day for 4 days. Main Outcome Measures: Luteinizing hormone pulse pattern, frequency and amplitude, mean LH and FSH levels, measured by serial blood sampling over a 6-hour period before and after naltrexone. Results: Naltrexone caused a significant increase (P < 0.05) of the LH pulse frequency in patients with hypothalamic amenorrhea and in PCOD but not in anorexia nervosa. The mean levels of LH and FSH and LH pulse amplitudes were not significantly changed by naltrexone. The naltrexone nop,responders were underweight either because of simple weight loss or anorexia nervosa and had low levels of estradiol and an LH pulse pattern similar to the luteal one. Conclusion: The luteal LH pulse pattern in weight loss-related amenorrhea is caused by a nonFertil Steril1992;57:762-70 opioid, undernutrition-linked factor. Key Words: Luteinizing hormone, follicle-stimulating hormone, polycystic ovarian disease, naltrexone, hypothalamic amenorrhea, anorexia nervosa, opioid antagonist, gonadotropin-releasing hormone, estradiol
Oligomenorrhea and amenorrhea in women are often associated with changes in the normal cyclic secretion of pulsatile luteinizing hormone (LH). In hypothalamic amenorrhea, a reduced LH pulse frequency was reported by some authors (1, 2) and an
Received May 20, 1991; revised and accepted December 20, 1991. * Present address: Department of Obstetrics and Gynecology, Kennemer Gasthuis Hospital, Deo, Haarlem, The Netherlands. t Reprint requests: Joop Schoemaker, M.D., Ph.D., Department of Obstetrics and Gynecology, Free University Hospital, Post Office Box 7057,1007 MB Amsterdam, The Netherlands.
762
Armeanu et aI. Naltrexone and LH pulse patterns
increased pulse frequency by others (3, 4). Shortterm blockade of the endogenous opioids with intravenous (IV) naloxone caused an increase in the LH pulse frequency and LH level (1, 5) or the pulse frequency alone (2), suggesting an involvement of a high tone of endogenous opioids in women with hypothalamic amenorrhea. However, no response to naloxone was seen in 40% of the patients with hypothalamic amenorrhea (5) and with anorexia nervosa (6). In the latter group, even no response at all was reported by some authors (7). In polycystic ovarian disease (peOD), an increased LH pulse frequency and amplitude (4) or Fertility and Sterility
pulse amplitude alone (8) is found. It was suggested that impairment of the opioid inhibition might be responsible for the increased LH secretion in patients with polycystic ovarian disease, as demonstrated by the fact that most patients with polycystic ovarian disease (9) did not respond to naloxone. The absence of naloxone response in certain patients with oligomenorrhea and amenorrhea might be because of its short action, which might mask the involvement of endogenous opioids. We studied the effects of the long-acting opioid antagonist naltrexone on the pulsatile LH secretion in women with hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease. Naltrexone has a greater potency and increases LH levels in the early follicular phase (10), whereas naloxone has no effect in this phase of the cycle (1). Because naltrexone is an oral preparation, a chronic opioid blockade becomes possible. Our hypothesis was that by stimulating the hypothalamic secretion of gonadotropin-releasing hormone (GnRH), naltrexone will increase LH pulse frequency and the levels of LH and follicle-stimulating hormone (FSH) in patients with hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease. This might have future therapeutic implications for ovulation induction with naltrexone. MATERIALS AND METHODS Subjects
Twenty-five patients between 18 and 40 years old participated in the study. The protocol was approved by the Committee for the Ethics of Research on Human Subjects of our hospital. All patients gave their informed consent. The patients were subdivided into three groups. Group 1 consisted of 10 women with hypothalamic amenorrhea because of different causes, except anorexia nervosa: 4 women with simple weight lossrelated amenorrhea because of dieting and stress, 3 with amenorrhea because of excessive physical exercise, 2 with stress-related amenorrhea, and 1 with primary amenorrhea because of underweight. An initial weight loss varying between 8.3% and 27% body weight occurred in 6 patients. Two patients failed to gain weight during puberty and were for this reason underweight. Group 2 consisted of six patients with secondary amenorrhea because of anorexia nervosa, as defined by the DSM-III R criteria (11). Partial weight recovery had taken place in all patients before the Vol. 57, No.4, April 1992
study. Five ofthe six patients were still underweight and had a disturbed eating pattern, as demonstrated by their eating diaries. Group 3 consisted of nine women with polycystic ovarian disease, defined as oligomenorrhea or amenorrhea associated with high LH levels (>6.5 U/L) and normal FSH levels in peripheral blood. Three of the patients with polycystic ovarian disease were overweight (body mass index [BMI] >25 kg/m2 ). All patients were normoprolactinemic (prolactin [PRL] < 0.70 U/L). Clinical and endocrine data are presented in Table 1. None ofthe patients abused alcohol or drugs. Any medication was discontinued at least 3 months before the study. None of the patients was seeking pregnancy during the naltrexone study. Protocol
Patients received naltrexone (Trexan; Du Pont Pharma, Bad Homburg, Germany) 50 mg/d in the morning before 8:00 A.M. for 4 days (day 1 to day 4). Pulse studies and GnRH tests were performed on days 0 and 4. On these days, a cannula (Abbocath, 20 gauge; Abbott Hospitals Inc., North Chicago, IL) was inserted into a convenient forearm vein at 8:00 A.M. Blood was sampled every 10 minutes during 6 hours (from 8:00 A.M. to 2:00 P.M.) for LH and FSH determinations. The gonadotropin response to GnRH was tested by IV injection of 100 ILg of GnRH at the end of the pulse study. Additional blood samples were taken 30, 60, and 120 minutes after injection also for LH and FSH determination. Plasma levels of estradiol (E 2 ), testosterone (T), androstenedione (A), and PRL were determined in the first blood sample on days 0 and 4. All samples were immediately centrifuged, and serum was stored at -20°C until assayed. Hormone Assays
In all samples LH and FSH were measured in duplicate by a recently developed immunoradiometric assay (IRMA). This method (LH-MAIAclone and FSH -MAIAclone; Serono Diagnostics Ltd., Woking, United Kingdom) uses two monoclonal antibodies, one labeled with 1251 and the other with fluorescein. The First International Reference Preparation (IRP) of human pituitary LH immunoassay standard (68/40) was used as a standard for LH and the Second IRP human pituitary LH/FSH bioassay standard (78/549) as a standard for FSH. The method was described in detail by Wennink et al. (12). The detection limit of both assays was 0.30 Armeanu et al.
Naltrexone and LH pulse patterns
763
r !
Table 1 Patients
Group 1 1 2 3 4 5 6 7 8 9 10 Group 2 11 12 13 14 15 16 Group 3 17 18 19 20 21 22 23 24 25
I
Clinical and Endocrine Characteristics * of the 25 Patients Age
BMI
Amenorrhea
y
kg/m'
y
25 22 28 26 34 27 18 29 25 18
19 16 17 28 21 29 25 20 23 16
1 Primary 2 1 8 10 3 2 2 2
1.5 0.7 2.1 1.6 3.6 0.2 1.5 5.2 2.5 0.5
5.7 3.4 7.5 7.9 7.5 2.1 4.7 5.9 5.7 5.0
38 60 148 216 60 49 152 104 91 56
Simple weight loss Primary underweight Simple weight loss Physical exercise Simple weight loss Simple weight loss Physical exercise Physical exercise Stress Stress
23 35 20 32 35 39
14 18 18 19 18 20
7 16 3 2 13 19
2.9 2.8 3.5 0.5 1.4 0.5
8.4 5.9 10.8 4.2 7.6 4.8
40 117 83 58 64 70
Restrictive anorexia Bulimic anorexia Restrictive anorexia Bulimic anorexia Restrictive anorexia Restrictive anorexia
23 32 30 32 30 31 28 29 35
45 33 21 34 23 25 22 23 19
2 5 3 5 5 Oligomenorrhea Oligomenorrhea Oligomenorrhea Oligomenorrhea
9.4 7.3 9.2 6.8 14.1 6.7 11.1 11.0 8.3
8.9 3.2 7.5 6.9 7.3 4.1 4.0 5.2 5.8
113 251 200 143 122 149 130 169 215
PCOD PCOD PCOD PCOD PCOD PCOD PCOD PCOD PCOD
FSH
LH
E2
Etiology
pmol/L
U/L
* Reference values: LH (follicular phase) 1.8 to 6.5 U/L; FSH (follicular phase) 3 to 12 U/L, and E2 (follicular phase) 110 to 180 pmol/L.
IV jL. The intra-assay SD was assessed from the difference between the duplicates in each assay for LH as well as for FSH. Each subject's LH and FSH concentrations on days 0 and 4, respectively, were estimated in single assays. The interassay and intraassay coefficients of variation (CV) for LH were 12.5% and 10%, respectively, for values < 2 VjL and 3.9% and 2.8%, respectively, for values> 2 VjL. The interassay and intra-assay CVs for FSH were 6.5% and 5%, respectively. Plasma E2 levels were measured by radioimmunoassay (RIA; Baxter Dade AG, Duedingen, Switzerland). The lower limit of detection was 37 pmol/L. The interassay and intra-assay CVs in the lowest and the highest ranges varied from 13.2% to 32.7% and from 4.0% to 7.2%, respectively. Plasma T levels were measured by a solid phase RIA (Coat-A-Count; Diagnostic Products Corp., Los Angeles, CA). The interassay and intra-assay CVs were 10% and 8%, respectively. Plasma A was measured by RIA after extraction (Diagnostic Products Corp.). The interassay CV was 9%; the intra-assay CV varied between 5.1% and 8.6%. 764
Armeanu et at.
Naltrexone and LH pulse patterns
Plasma PRL was measured by IRMA (Medgenix, Belgium). The interassay and intra-assay CVs were 7% and 4%, respectively. Plasma progesterone (P) was measured by RIA (Bioscentia, Mainz, Germany). The intra-assay and interassay CVs were 8% and 10%, respectively. Pulse Analysis
Pulse analysis as described by Lambalk et al. (13) and as modified by Scheele et al. (14) was used. A pulse was defined as a rise in concentration exceeding a minimum amplitude criterion. The amplitude (D) was defined as the difference between a maximum value and the preceding nadir. An LH pulse was defined when two pulse criteria were satisfied. The threshold criterion as described by Lambalk et al. (13) considers an increase to be a pulse when the LH concentration exceeds a threshold value of D = 2. VSpeak2 + Snadi/. The variables Speak and Snadir are the intra-assay CVs ofthe concentration ranges in which the peak and nadir concentrations are situated, respectively. When the difference between nadir and peak exceeds the threshold value D, the peak is considered to be a pulse with a probability Fertility and Sterility
of being a false-positive of 0.02275 (I-tailed probability table for normal distributions (15». The second pulse criterion as described by Scheele et al. (14) is based on the quality of the duplicates of nadir and peak. The amplitude D calculated with the first pulse criterion should exceed three times the difference between the duplicates of either the nadir or the peak, whichever is greater. This method prevents variations, in which the duplicates of nadir and peak are of unacceptable quality, from being detected as pulses. Of each subject, the mean LH and FSH concentration, the mean LH pulse amplitude, and the incidence of LH pulses per 6 hours on day 0 and day 4 were calculated. The individual patients were considered to respond to naltrexone when at least one additional LH pulse and an increase of the mean LH concentration were detected over the 6-hour period on day 4. In groups 1 and 2, only at least a doubling of the mean LH level was considered a positive response to naltrexone to rule out the possibility that minor changes in LH secretion may bias the results. Nadirs preceding the pulses are indicated as marker points in the hormone patterns rather than the pulses themselves. The effect of naltrexone on the LH pulse pattern was studied. Luteinizing hormone pulse patterns were classified following the criteria described by Schoemaker et al. (16). Luteinizing hormone pulse patterns are divided into three types (Fig. 1). The differentiation between the three types is mainly based on the magnitude of the pulse amplitudes, as will be illustrated below. Differences in LH pulse frequency and/or LH level are not used as criteria for this classification. 1. Type 1: solid pattern. The majority of the pulses have an amplitude of more· than twice the threshold pulse criterion. 2. Type 2: regular pattern. Luteinizing hormone secretion is stable. Either the majority of the pulses have an amplitude less than twice the threshold pulse criterion or there are no pulses at all. 3. Type 3: slow pattern. Starting in the early morning, the LH level shows a hyperbolic descent, with either no pulse at all or with large pulse amplitudes and nadir intervals. The maximal LH and FSH increments were taken as a parameter for the response to the GnRH challenge of 100 #lg. Statistical Analysis
Statistical differences between hormone parameters on day 0 and day 4 were calculated using WilVol. 57, No.4, April 1992
TYPE 1
1- LHOAY~
15 12
e:
e
9
:I:
-'
8 3 0
80
0
120
180
2«1
300
380
TIME (min)
TYPE 2
4
1- LHOAY~
e:
a
:I:
--'
2
o+---~~--~----,---~~--~--~ o 60 120 180 2«1 300 380 TIME (min)
TYPE 3
8~====~--------------1
1_
LHDAyol
4
2
o+---~~--~----,---~~--~--~ o 80 120 180 2«1 300 380 TIME (min)
Figure 1 Examples of the three different LH pulse patterns. Data are collected from patient 24 (type 1), patient 4 (type 2), and patient 13 (type 3) on day o.
coxon's signed rank sum test. Correlations between hormone parameters were calculated with Spearman's rank correlation coefficient. Relationships between BMI and the response to naltrexone were calculated with a X2 contingency test. Differences between naltrexone responders and nonresponders were calculated by the Mann-Whitney test. The criterion of significance was set at P < 0.05. RESULTS
The effect of naltrexone on LH pulse characteristics in the three groups is presented in Figure 2. Armeanu et al. Naltrexone and LH pulse patterns
765
c 12f.::==;------.,
A
i 51
. a:
§
2
!i
1
2 2
2 PAnENr GROUPS
PAnENT GROUPS
2 PAnENr GROUPS
Figure 2 Effects of naltrexone on (A) LH pulse frequency, (B) LH pulse amplitude, and (C) mean LH level. Results are presented as means ± SEM. Significant differences in pulse frequency between day 0 and day 4 in groups 1 and 3; *, P < 0.05.
Naltrexone significantly increased the number ofLH pulses in group 1 from 1.6 ± 0.5 to 2.6 ± 0.6 pulses/ 6 hours (P < 0.05) and in group 3 from 6.2 ± 0.5 to 7.1 ± 0.6 pulses/6 hours (P < 0.05) but failed to do so in group 2. There was no significant change in either the LH pulse amplitude or the mean LH level in any of the groups. The distribution of LH pulse pattern types on day 0 and day 4 is presented in Figure 3. In group 1, two patients had a type 1 and two had a type 2 pulse pattern. These patients responded to naltrexone by an increased LH pulse frequency. In one of them, the pulse pattern changed from type 2 into type 3; in the others the pulse pattern remained unchanged. Six patients in group 1 had a type 3 pulse pattern, and this pattern was unchanged by naltrexone. All patients in group 2 had a type 3 pulse pattern. Only one patient responded to naltrexone by an increased pulse frequency and a changed pulse pattern (from type 3 into type 1). In group 3, all patients had type 1 pulse pattern before and after naltrexone administration. The relationship between BMI and the response to naltrexone is shown in Figure 4. Significantly less patients who were underweight (BMI < 20) responded to naltrexone, when compared with normal weight (BMI ~ 20) patients (X 2 = 9.14; P < 0.005; GROUP 1
A 120
IDa
100
B 120
.,0 I
100
d.,4
~:
20
20
C 120
IDII .,0 I .,4
100
i
80 ",80
:.,.40
UI PULSE PAnERN
766
GROUP 2
i
:.,.40
Figure 3 group 3.
n = 25). The underweight nonresponders seemed to have a lower baseline E2 level (median 64, range 38 to 215) than the normal weight naltrexone responders (median 126, range 49 to 216), but this difference was not significant (0.05 < P ::;; 0.1). Eight of the 12 naltrexone nonresponders had a type 3 pulse pattern and were underweight. In contrast, 8 of the 13 naltrexone responders had a type 1, 2 of the 13 had a type 2 pulse pattern, and all these 10 patients had a normal BMI. The baseline E2 level was significantly lower (P < 0.025) in the 8 patients with type 3 pulse pattern who were underweight and did not respond to naltrexone (median 61, range 38 to 150) than in the 10 normal weight responders who had types 1 and 2 pulse patterns (median 126, range 49 to 216). It seems that type 3 pulse pattern tends to be more frequent among underweight, low estrogenic, naltrexone nonresponders. The effects of naltrexone on other endocrine parameters are presented in Table 2. The median FSH level was not changed by naltrexone in any of the groups during the pulse study. Estradiol was significantly increased after naltrexone therapy in group 1 (P < 0.05) but was unaffected in groups 2 and 3. Testosterone was significantly increased (P < 0.05) in group 3 but was unchanged in groups 1 and 2. Naltrexone had no effect on PRL and A concentraGROUP 3
m a
.,4
80
",80
:.,.40 20
UI PULSE PATTERN
UI PUI.SE PAnERN
Effect of naltrexone on the distribution of LH pulse patterns in (A) group 1, (B) group 2, and (C)
Armeanu et al.
Naltrexone and LH pulse patterns
Fertility and Sterility
50
tite) , and one patient reported sleepiness and muscular weakness. No side effects were seen in groups 1 and 2. There were no changes in hepatic and renal functions, hemoglobin, or erythrocyte sedimentation rate in any of the patients. There were no dropouts from the study.
• 40
Q' E
l
•
• •
30
5i
III
20
.,•••••.
:i'•
10
DISCUSSION
•
NXNO
NXYES
Figure 4 Differences of BMI (kg/m2 ) between naltrexone responders (NX YES) and naltrexone nonresponders (NX NO). Data from groups 1, 2, and 3 (n = 25).
tion (results not shown). Progesterone had anovulatory values (0.3 to 0.8 nmol/L) before and after naltrexone in all patients. Naltrexone did not cause any significant change in the maximal LH and FSH response to GnRH (results not shown). The LH response was higher than the FSH response in all patients both on day o and day 4. No prepubertal response was seen in the patients with hypothalamic amenorrhea. In patients from group 1 and 2 who responded to naltrexone, a negative correlation (r = -0.8 and P < 0.05; n = 7, group 1 and 2 together) was found between the E2 level and the LH response to GnRH on day o. No such correlation was found in the nonresponders (r = 0.08; n = 8, group 1 and 2 together). Side Effects
Naltrexone treatment during 4 days in a dosage of 50 mg/d was well tolerated by most of the patients. Minor side effects were noticed in 16% of the patients. Three patients in group 3 reported mild gastrointestinal complaints (nausea and lack of appe-
Table 2
Twelve of the 16 patients in groups 1 and 2 had type 3 pulse pattern. This pattern is normally seen in the luteal phase of the menstrual cycle (1, 5, 16). Administration of naltrexone (17) to women during the luteal phase increased the LH pulse frequency, demonstrating that endogenous opioids are responsible for the typical LH pulse pattern during the luteal phase. Administration of P (18) or a combination of E2 and P (19) caused a similar pulse pattern. These reports demonstrate that the type 3 pattern in the luteal phase is caused by P in the presence of E2 through mediation of endogenous opioids. Our study could not demonstrate that endogenous opioids are also involved in the type 3 LH secretion pattern in hypothalamic amenorrhea because, in the majority of the patients, this pulse pattern was not modified by naltrexone. These results suggest that the type 3 pulse pattern may be mediated by both opioid and nonopioid pathways. In the patients with hypothalamic amenorrhea and anorexia nervosa who did not respond to naltrexone, no correlation was found between E2 level and the LH response to GnRH. This finding demonstrates an impaired negative feedback of gonadal steroids, which may explain the lack of response to naltrexone in these patients. Furthermore, underweight naltrexone nonresponders with type 3 pulse pattern had significantly lower E2 levels in comparison with the naltrexone responders with type 1 and 2 patterns. These results confirm that the respon-
Levels of FSH*, E 2t, and T:j: Before and After Naltrexone FSH
Patients
Day 0
Group 1 Group 2 Group 3
5.7 (2.1 to 7.9) 6.8 (4.2 to 10.8) 5.8 (3.2 to 8.9)
Day 4
Day 0
5.6 (2.9 to 8.2) 7.9 (3.5 to 9.5) 6.1 (3.8 to 8.5)
76 (38 to 216) 67 (40 to 117) 145 (103 to 251)
U/L
Day 4
Day 0
125 (38 to 480) § 63 (50 to 207) 170 (113 to 274)
1.3 (1.0 to 1.7) 1.2 (1.0 to 1.4) 2.0 (1.0 to 4.3)
pmol/L
* Reference values 3 to 12 U /L. t Reference values 110 to 180 pmol/L. :j: Reference values < 2.5 nmol/L.
Vol. 57, No.4, April 1992
T
E2
Day 4 nmol/L
1.5 (1.0 to 2.0) 1.0 (1.0 to 1.9) 3.6 (1.1 to 5.6) II
§ Significant difference (P < 0.05 in E between day 4 and dayO. II Significant difference (P < 0.05) in T between day 4 and dayO. Armeanu et al.
Naltrexone and LH pulse patterns
767
siveness to opioid antagonists depends on the activation of the endogenous opioid system by E2 (20, 21). Obviously, an E2 independent factor, which is not mediated by endogenous opioids, inhibits GnRH secretion in naltrexone nonresponders with hypothalamic amenorrhea and type 3 pulse pattern. A link between this inhibiting factor and underweight is suggested by the fact that 75% of the patients with type 3 pulse pattern were underweight and did not respond to naltrexone. This hypothesis is supported by evidence that inhibitory effects on gonadotropins, caused by nutritional deprivation and by E2 in castrated ewes, are additive (22). The nature of this undernutrition-linked factor still remains unclear. Corticotropin-releasing factor (CRF), which is known to mediate stress and feeding behavior, inhibits GnRH secretion partly by an opioid dependent and partly by an opioid independent pathway (23) and might be involved in the inhibitory action of this undernutrition-linked factor. We might speculate that this undernutrition-linked factor is a remainder of the prepubertal central restraint, which is re-activated by undernutrition in combination with a low-steroid environment. The inhibition of LH secretion seen in the early puerperium, which cannot be antagonized by naltrexone (24) either, might also be caused by the rapid weight loss, occurring in this period and might be mediated by this undernutrition-linked factor too. The fact that patients with anorexia nervosa failed to respond to naltrexone is partly in disagreement with Baranovska et al. (6). These authors observed an increase of LH levels after administration of naloxone in patients with anorexia nervosa in whom the amenorrhea preceded the weight loss. This discrepancy may be explained by a different methodology. The mean LH levels measured by us are based on 36 blood samples obtained during 6 hours after 4 days of naltrexone, whereas Baranovska et al. (6) based their results on single blood samples taken 30 minutes after a bolus naloxone injection. Furthermore, the latter authors did not observe any increase in LH pulse frequency and amplitude when blood sampling was performed during naloxone infusion for 4 hours, agreeing with our findings. Naltrexone caused significant increments of LH pulse frequency and E2 levels in responsive patients with hypothalamic amenorrhea, resulting from a diminished inhibition of GnRH pulse frequency. Based on the previously observed consistency of LH pulse patterns in patients with hypogonadotropic amenorrhea on different sampling occasions (16), we concluded that any changes in LH secretion may 768
Armeanu et al.
Naltrexone and LH pulse patterns
be attributed to naltrexone. However, we realize that interpretation of these results is controversial because of the lack of a randomized, placebo-controlled design of the study. Thus, it cannot be excluded that the changes in LH secretion may at least partly be caused by a placebo effect. Within these limitations of our study design, our results are in agreement with previous reports (1,5) that demonstrated that opioid inhibition is involved in the pathophysiology of hypothalamic amenorrhea in certain patients. In spite of the strong antagonistic activity and the chronic treatment, naltrexone was not able to stimulate LH secretion in all patients. Our results do not rule out the possibility that naltrexone causes subtle increments of GnRH secretion, but the pituitary is not sensitive enough and does not respond with a similar LH secretion in the naltrexone nonresponders. Further experiments with naltrexone administration after GnRH priming of the pituitary may elucidate this aspect. The LH pulse pattern in patients with polycystic ovarian disease is similar to type 1 pulse patterns seen in the follicular phase of the menstrual cycle (16) and does not change after naltrexone treatment, although LH pulse frequency increases. Our results confirm the findings of Berga and Yen (25) who reported an intact opioidergic inhibition in patients with polycystic ovarian disease. These authors demonstrated a response to naloxone after pretreatment with medroxyprogesterone. Our study demonstrates that naltrexone increases LH pulse frequency in patients with polycystic ovarian disease without P pretreatment as well. Contrary to Petraglia et al. (9), who found a link between the response to naloxone and obesity, we found that the naltrexone response also occurred in nonobese patients with polycystic ovarian disease. The differences between naloxone and naltrexone may be because of the longer treatment (4 days) and the stronger activity of naltrexone. The higher LH pulse frequency demonstrates an increased GnRH pulse secretion. The T concentrations in plasma of women with polycystic ovarian disease were significantly increased after 4 days ofnaltrexone therapy. This may be the result of the naltrexone-induced increased LH secretion, as suggested by the significant increase of LH pulse frequency and the small, but not significant, increase of LH levels in group 3 (Fig.2C). The secretion of FSH was not stimulated by chronic treatment with naltrexone in the patients with oligomenorrhea or amenorrhea. Either the increment in GnRH pulsatility induced by naltrexone Fertility and Sterility
is insufficient to cause an increase in pituitary FSH secretion or the FSH secretion may still be inhibited by neurotransmitters and/or neuromodulators other than the endogenous opioids. However, it is possible that small changes in FSH secretion may be present but cannot be detected because of limitations inherent to the assay and/or analysis techniques. In conclusion, the oral treatment with the longacting antagonist naltrexone enabled us to study the effect of a prolonged blockade of the opioid receptors on the LH pulsatility in patients with hypothalamic amenorrhea and polycystic ovarian disease. The response in 60% of the patients with hypothalamic amenorrhea because of simple weight loss, stress, and physical exercise demonstrates the involvement of endogenous opioids in the pathophysiology of this condition. The GnRH inhibition in patients with anorexia nervosa and other weight loss-related forms of secondary amenorrhea is probably caused by an opioid-independent, undernutrition-linked factor that remains to be identified. The response in patients with polycystic ovarian disease demonstrates an intact central opioid activity in these patients. Because 4 days of naltrexone treatment only slightly stimulated LH pulse frequency but had no effect on FSH secretion, it is questionable whether a longer treatment with naltrexone would be effective in ovulation induction in these forms of oligoamenorrhea. Ackrwwledgments. We are grateful to the staff of the Endocrinology Laboratory of the Free University Hospital for performing the hormone assays and to Mr. Tjarco van Wijngaarden from the Audio-Visual-Center of the Free University Hospital for the technical assistance in preparing the illustrations for this manuscript. We thank Du Pont Pharma, Bad Homburg, Germany, for the kindness of providing naltrexone (Trexan) for this clinical study.
REFERENCES 1. Khoury SA, Reame NE, Kelch RP, Marshall JC. Diurnal patterns of pulsatile luteinizing hormone secretion in hypothalamic amenorrhea: reproducibility and response to opiate blockade and an a2-adrenergic agonist. J Clin Endocrinol Metab 1987;64:755-62. 2. Judd S, Stranks S, Michailov L. Gonadotropin-releasing hormone pacemaker sensitivity to negative feedback inhibition by estradiol in women with hypothalamic amenorrhea. Fertil Steril 1989;51:257-62. 3. Genazzani AD, Petraglia F, Fabbri G, Monzani A, Montanini V, Genazzani AR. Evidence of luteinizing hormone secretion in hypothalamic amenorrhea associated with weight loss. Fertil Steril 1990;54:222-6. 4. Burger CW, Korsen T, Van Kessel H, Van Dop PA, Caron F JM, Schoemaker J. Pulsatile luteinizing hormone patterns Vol. 57, No.4, April 1992
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
in the follicular phase of the menstrual cycle, polycystic ovarylike syndrome (PCOD) and non-PCOD secondary amenorrhea. J Clin Endocrinol Metab 1985;61:1126-32. Quigley E, Sheehan KL, Casper RF, Yen SSC. Evidence for increased dopaminergic and opioid activity in patients with hypothalamic hypogonadotropic amenorrhea. J Clin Endocrinol Metab 1980;50:949-54. Baranovska B, Rozbicka G, Jeske W, Abdel-Fattah MH. The role of endogenous opiates in the mechanism of inhibited luteinizing hormone (LH) secretion in women with anorexia nervosa: the effect of naloxone on LH, follicle-stimulating hormone, prolactin and {:1-endorphin secretion. J Clin Endocrinol Metab 1984;59:412-6. Giusti M, Torre R, Traverso L, Cavagnaro P, Attanasio R, Giordano G. Endogenous opioid blockade and gonadotropin secretion: role of pulsatile luteinizing hormone-releasing hormone administration in anorexia nervosa and weight loss amenorrhea. Fertil Steril 1988;49:797-801. Kazer RR, Kessel B, Yen SSC. Circulating luteinizing hormone pulse frequency in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1987;65:233-6. Petraglia F, D' Ambrogio G, Comitini G, Facchinetti F, Volpe A, Genazzani AR. Impairment of opioid control of luteinizing hormone secretion in menstrual disorders. Fertil Steril 1985;43:534-40. Teoh SK, Mendelson JH, Mello NK, Skupny A. Alcohol effects on naltrexone-induced stimulation of pituitary, adrenal, and gonadal hormones during the early follicular phase of the menstrual cycle. J Clin Endocrinol Metab 1988;66: 1181-6. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 3rd ed. Washington D.C.: American Psychiatric Association, 1987. Wennink JMB, Delemarre-Van de Waal HA, Van Kessel H, Mulder GH, Foster JP, Schoemaker J. Luteinizing hormone secretion patterns in boys at the onset of puberty measured using a highly sensitive immunoradiometric assay. J Clin Endocrinol Metab 1988;67:924-8. Lambalk CB, De Koning J, Van Kessel H, Van Rees G, Schoemaker J. Calculation of the intra assay variation per assay and the relevance to LH pulse detection. IRCS Med Sci 1985;13:1183-4. Scheele F, Lambalk CB, Schoemaker J, Van Kessel H, De Koning J, Van Dieten JAM, et al. Patterns of LH and FSH in men during high-frequency blood sampling. J Endocrinol 1987;114:153-60. Armitage P. Statistical methods in medical research. Oxford: Blackwell Scientific Publications, 1983. Schoemaker J, Schoemaker H, Berkhout G, Schoemaker HC. Pulsatile luteinizing hormone secretion in hypothalamic amenorrhea. In: Genazzani AR, Petraglia F, Volpe A, Facchinetti F, editors. Advances in gynecological endocrinology. Carnforth (UK): The Parthenon Publishing Group, 1988: 43-50. GindoffPR, Jewelewicz R, Hembree W, Wardlaw S, Ferin M. Sustained effects of opioid antagonism during the normal human luteal phase. J Clin Endocrinol Metab 1988;66: 1000-4. Buckler HM, Phillips SE, Cameron IT, Healy DL, Burger HG. Vaginal progesterone administration before ovulation induction with exogenous gonadotropins in polycystic ovarian syndrome. J Clin Endocrinol Metab 1988;67:300-7.
Armeanu et aI.
Naltrexone and LH pulse patterns
769
19. Nippoldt TB, Reame NE, Kelch RP, Marshall JC. The role
of estradiol and progesterone in decreasing luteinizing hormone pulse frequency in the luteal phase of the menstrual cycle. J Clin Endocrinol Metab 1989;69:67-76. 20. Reid RL, Quigley ME, Yen SSC. The disappearance of opioider~c regulation of gonadotropin secretion in postmenopausal women. J Clin Endocrinol Metab 1983;57: 1107-11. 21. Chowen JA, Argente J, Vician L, Clifton DK, Steiner RA. Pro-opiomelanocortin messenger RNA in hypothalamic neurons is increased by testosterone through aromatization to estradiol. Neuroendocrinology 1990;52:581-8. 22. Kyle JP, Alexander BM, Moss GE, Hallford DM, Nett TM. Gonadotropin-releasing hormone overrides the negative effect
770
Armeanu et al.
Naltrexone and LH pulse patterns
of reduced dietary energy on gonadotropin synthesis and secretion in ewes. Endocrinology 1991;128:843-9. 23. Williams CL, Nishihara M, ThaiabardJ, Hotchkiss J, Knobil E. Corticotropin-releasing factor and gonadotropin-releasing hormone pulse generator activity in the rhesus monkey. Electrophysiological studies. Neuroendocrinology 1990;52: 133-7. 24. Kremer JA, Borm G, Schellekens LA, Thomas CMG, Rolland R. Pulsatile secretion of luteinizing hormone and prolactin in lactating and nonlactating women and the response to naltrexone. J Clin Endocrinol Metab 1991;72:294-300. 25. Berga SL, Yen SSC. Opioidergic regulation of LH pulsatility in women with polycystic ovary syndrome. Clin Endocrinol (Oxf) 1989;30:177-84.
Fertility and Sterility
Vol. 57, No.4, April 1992
Copyright ~ 1992 The American Fertility Society
Printed on acid-free paper in U.S.A.
Pulsatile luteinizing hormone secretion in hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease during naltrexone treatment Magda C. Armeanu, M.D., Ph.D.* Gerda M. J. Berkhout, R.N. Joop Schoemaker, M.D., Ph.D.t Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Fertility, Free University Hospital, Amsterdam, The Netherlands
Objective: To determine if chronic treatment with the long-acting oral opioid antagonist naltrexone can increase luteinizing hormone (LH) and follicle-stimulating hormone (FSH) secretion in women with secondary amenorrhea. Design: Prospective. Setting: Large reproductive endocrinology unit of an academic hospital. Patients: Three groups of women with oligomenorrhea or amenorrhea: (1) hypothalamic amenorrhea; (2) anorexia nervosa; and (3) polycystic ovarian disease (PCOD). Intervention: Naltrexone 50 mg every day for 4 days. Main Outcome Measures: Luteinizing hormone pulse pattern, frequency and amplitude, mean LH and FSH levels, measured by serial blood sampling over a 6-hour period before and after naltrexone. Results: Naltrexone caused a significant increase (P < 0.05) of the LH pulse frequency in patients with hypothalamic amenorrhea and in PCOD but not in anorexia nervosa. The mean levels of LH and FSH and LH pulse amplitudes were not significantly changed by naltrexone. The naltrexone nop,responders were underweight either because of simple weight loss or anorexia nervosa and had low levels of estradiol and an LH pulse pattern similar to the luteal one. Conclusion: The luteal LH pulse pattern in weight loss-related amenorrhea is caused by a nonFertil Steril1992;57:762-70 opioid, undernutrition-linked factor. Key Words: Luteinizing hormone, follicle-stimulating hormone, polycystic ovarian disease, naltrexone, hypothalamic amenorrhea, anorexia nervosa, opioid antagonist, gonadotropin-releasing hormone, estradiol
Oligomenorrhea and amenorrhea in women are often associated with changes in the normal cyclic secretion of pulsatile luteinizing hormone (LH). In hypothalamic amenorrhea, a reduced LH pulse frequency was reported by some authors (1, 2) and an
Received May 20, 1991; revised and accepted December 20, 1991. * Present address: Department of Obstetrics and Gynecology, Kennemer Gasthuis Hospital, Deo, Haarlem, The Netherlands. t Reprint requests: Joop Schoemaker, M.D., Ph.D., Department of Obstetrics and Gynecology, Free University Hospital, Post Office Box 7057,1007 MB Amsterdam, The Netherlands.
762
Armeanu et aI. Naltrexone and LH pulse patterns
increased pulse frequency by others (3, 4). Shortterm blockade of the endogenous opioids with intravenous (IV) naloxone caused an increase in the LH pulse frequency and LH level (1, 5) or the pulse frequency alone (2), suggesting an involvement of a high tone of endogenous opioids in women with hypothalamic amenorrhea. However, no response to naloxone was seen in 40% of the patients with hypothalamic amenorrhea (5) and with anorexia nervosa (6). In the latter group, even no response at all was reported by some authors (7). In polycystic ovarian disease (peOD), an increased LH pulse frequency and amplitude (4) or Fertility and Sterility
pulse amplitude alone (8) is found. It was suggested that impairment of the opioid inhibition might be responsible for the increased LH secretion in patients with polycystic ovarian disease, as demonstrated by the fact that most patients with polycystic ovarian disease (9) did not respond to naloxone. The absence of naloxone response in certain patients with oligomenorrhea and amenorrhea might be because of its short action, which might mask the involvement of endogenous opioids. We studied the effects of the long-acting opioid antagonist naltrexone on the pulsatile LH secretion in women with hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease. Naltrexone has a greater potency and increases LH levels in the early follicular phase (10), whereas naloxone has no effect in this phase of the cycle (1). Because naltrexone is an oral preparation, a chronic opioid blockade becomes possible. Our hypothesis was that by stimulating the hypothalamic secretion of gonadotropin-releasing hormone (GnRH), naltrexone will increase LH pulse frequency and the levels of LH and follicle-stimulating hormone (FSH) in patients with hypothalamic amenorrhea, anorexia nervosa, and polycystic ovarian disease. This might have future therapeutic implications for ovulation induction with naltrexone. MATERIALS AND METHODS Subjects
Twenty-five patients between 18 and 40 years old participated in the study. The protocol was approved by the Committee for the Ethics of Research on Human Subjects of our hospital. All patients gave their informed consent. The patients were subdivided into three groups. Group 1 consisted of 10 women with hypothalamic amenorrhea because of different causes, except anorexia nervosa: 4 women with simple weight lossrelated amenorrhea because of dieting and stress, 3 with amenorrhea because of excessive physical exercise, 2 with stress-related amenorrhea, and 1 with primary amenorrhea because of underweight. An initial weight loss varying between 8.3% and 27% body weight occurred in 6 patients. Two patients failed to gain weight during puberty and were for this reason underweight. Group 2 consisted of six patients with secondary amenorrhea because of anorexia nervosa, as defined by the DSM-III R criteria (11). Partial weight recovery had taken place in all patients before the Vol. 57, No.4, April 1992
study. Five ofthe six patients were still underweight and had a disturbed eating pattern, as demonstrated by their eating diaries. Group 3 consisted of nine women with polycystic ovarian disease, defined as oligomenorrhea or amenorrhea associated with high LH levels (>6.5 U/L) and normal FSH levels in peripheral blood. Three of the patients with polycystic ovarian disease were overweight (body mass index [BMI] >25 kg/m2 ). All patients were normoprolactinemic (prolactin [PRL] < 0.70 U/L). Clinical and endocrine data are presented in Table 1. None ofthe patients abused alcohol or drugs. Any medication was discontinued at least 3 months before the study. None of the patients was seeking pregnancy during the naltrexone study. Protocol
Patients received naltrexone (Trexan; Du Pont Pharma, Bad Homburg, Germany) 50 mg/d in the morning before 8:00 A.M. for 4 days (day 1 to day 4). Pulse studies and GnRH tests were performed on days 0 and 4. On these days, a cannula (Abbocath, 20 gauge; Abbott Hospitals Inc., North Chicago, IL) was inserted into a convenient forearm vein at 8:00 A.M. Blood was sampled every 10 minutes during 6 hours (from 8:00 A.M. to 2:00 P.M.) for LH and FSH determinations. The gonadotropin response to GnRH was tested by IV injection of 100 ILg of GnRH at the end of the pulse study. Additional blood samples were taken 30, 60, and 120 minutes after injection also for LH and FSH determination. Plasma levels of estradiol (E 2 ), testosterone (T), androstenedione (A), and PRL were determined in the first blood sample on days 0 and 4. All samples were immediately centrifuged, and serum was stored at -20°C until assayed. Hormone Assays
In all samples LH and FSH were measured in duplicate by a recently developed immunoradiometric assay (IRMA). This method (LH-MAIAclone and FSH -MAIAclone; Serono Diagnostics Ltd., Woking, United Kingdom) uses two monoclonal antibodies, one labeled with 1251 and the other with fluorescein. The First International Reference Preparation (IRP) of human pituitary LH immunoassay standard (68/40) was used as a standard for LH and the Second IRP human pituitary LH/FSH bioassay standard (78/549) as a standard for FSH. The method was described in detail by Wennink et al. (12). The detection limit of both assays was 0.30 Armeanu et al.
Naltrexone and LH pulse patterns
763
r !
Table 1 Patients
Group 1 1 2 3 4 5 6 7 8 9 10 Group 2 11 12 13 14 15 16 Group 3 17 18 19 20 21 22 23 24 25
I
Clinical and Endocrine Characteristics * of the 25 Patients Age
BMI
Amenorrhea
y
kg/m'
y
25 22 28 26 34 27 18 29 25 18
19 16 17 28 21 29 25 20 23 16
1 Primary 2 1 8 10 3 2 2 2
1.5 0.7 2.1 1.6 3.6 0.2 1.5 5.2 2.5 0.5
5.7 3.4 7.5 7.9 7.5 2.1 4.7 5.9 5.7 5.0
38 60 148 216 60 49 152 104 91 56
Simple weight loss Primary underweight Simple weight loss Physical exercise Simple weight loss Simple weight loss Physical exercise Physical exercise Stress Stress
23 35 20 32 35 39
14 18 18 19 18 20
7 16 3 2 13 19
2.9 2.8 3.5 0.5 1.4 0.5
8.4 5.9 10.8 4.2 7.6 4.8
40 117 83 58 64 70
Restrictive anorexia Bulimic anorexia Restrictive anorexia Bulimic anorexia Restrictive anorexia Restrictive anorexia
23 32 30 32 30 31 28 29 35
45 33 21 34 23 25 22 23 19
2 5 3 5 5 Oligomenorrhea Oligomenorrhea Oligomenorrhea Oligomenorrhea
9.4 7.3 9.2 6.8 14.1 6.7 11.1 11.0 8.3
8.9 3.2 7.5 6.9 7.3 4.1 4.0 5.2 5.8
113 251 200 143 122 149 130 169 215
PCOD PCOD PCOD PCOD PCOD PCOD PCOD PCOD PCOD
FSH
LH
E2
Etiology
pmol/L
U/L
* Reference values: LH (follicular phase) 1.8 to 6.5 U/L; FSH (follicular phase) 3 to 12 U/L, and E2 (follicular phase) 110 to 180 pmol/L.
IV jL. The intra-assay SD was assessed from the difference between the duplicates in each assay for LH as well as for FSH. Each subject's LH and FSH concentrations on days 0 and 4, respectively, were estimated in single assays. The interassay and intraassay coefficients of variation (CV) for LH were 12.5% and 10%, respectively, for values < 2 VjL and 3.9% and 2.8%, respectively, for values> 2 VjL. The interassay and intra-assay CVs for FSH were 6.5% and 5%, respectively. Plasma E2 levels were measured by radioimmunoassay (RIA; Baxter Dade AG, Duedingen, Switzerland). The lower limit of detection was 37 pmol/L. The interassay and intra-assay CVs in the lowest and the highest ranges varied from 13.2% to 32.7% and from 4.0% to 7.2%, respectively. Plasma T levels were measured by a solid phase RIA (Coat-A-Count; Diagnostic Products Corp., Los Angeles, CA). The interassay and intra-assay CVs were 10% and 8%, respectively. Plasma A was measured by RIA after extraction (Diagnostic Products Corp.). The interassay CV was 9%; the intra-assay CV varied between 5.1% and 8.6%. 764
Armeanu et at.
Naltrexone and LH pulse patterns
Plasma PRL was measured by IRMA (Medgenix, Belgium). The interassay and intra-assay CVs were 7% and 4%, respectively. Plasma progesterone (P) was measured by RIA (Bioscentia, Mainz, Germany). The intra-assay and interassay CVs were 8% and 10%, respectively. Pulse Analysis
Pulse analysis as described by Lambalk et al. (13) and as modified by Scheele et al. (14) was used. A pulse was defined as a rise in concentration exceeding a minimum amplitude criterion. The amplitude (D) was defined as the difference between a maximum value and the preceding nadir. An LH pulse was defined when two pulse criteria were satisfied. The threshold criterion as described by Lambalk et al. (13) considers an increase to be a pulse when the LH concentration exceeds a threshold value of D = 2. VSpeak2 + Snadi/. The variables Speak and Snadir are the intra-assay CVs ofthe concentration ranges in which the peak and nadir concentrations are situated, respectively. When the difference between nadir and peak exceeds the threshold value D, the peak is considered to be a pulse with a probability Fertility and Sterility
of being a false-positive of 0.02275 (I-tailed probability table for normal distributions (15». The second pulse criterion as described by Scheele et al. (14) is based on the quality of the duplicates of nadir and peak. The amplitude D calculated with the first pulse criterion should exceed three times the difference between the duplicates of either the nadir or the peak, whichever is greater. This method prevents variations, in which the duplicates of nadir and peak are of unacceptable quality, from being detected as pulses. Of each subject, the mean LH and FSH concentration, the mean LH pulse amplitude, and the incidence of LH pulses per 6 hours on day 0 and day 4 were calculated. The individual patients were considered to respond to naltrexone when at least one additional LH pulse and an increase of the mean LH concentration were detected over the 6-hour period on day 4. In groups 1 and 2, only at least a doubling of the mean LH level was considered a positive response to naltrexone to rule out the possibility that minor changes in LH secretion may bias the results. Nadirs preceding the pulses are indicated as marker points in the hormone patterns rather than the pulses themselves. The effect of naltrexone on the LH pulse pattern was studied. Luteinizing hormone pulse patterns were classified following the criteria described by Schoemaker et al. (16). Luteinizing hormone pulse patterns are divided into three types (Fig. 1). The differentiation between the three types is mainly based on the magnitude of the pulse amplitudes, as will be illustrated below. Differences in LH pulse frequency and/or LH level are not used as criteria for this classification. 1. Type 1: solid pattern. The majority of the pulses have an amplitude of more· than twice the threshold pulse criterion. 2. Type 2: regular pattern. Luteinizing hormone secretion is stable. Either the majority of the pulses have an amplitude less than twice the threshold pulse criterion or there are no pulses at all. 3. Type 3: slow pattern. Starting in the early morning, the LH level shows a hyperbolic descent, with either no pulse at all or with large pulse amplitudes and nadir intervals. The maximal LH and FSH increments were taken as a parameter for the response to the GnRH challenge of 100 #lg. Statistical Analysis
Statistical differences between hormone parameters on day 0 and day 4 were calculated using WilVol. 57, No.4, April 1992
TYPE 1
1- LHOAY~
15 12
e:
e
9
:I:
-'
8 3 0
80
0
120
180
2«1
300
380
TIME (min)
TYPE 2
4
1- LHOAY~
e:
a
:I:
--'
2
o+---~~--~----,---~~--~--~ o 60 120 180 2«1 300 380 TIME (min)
TYPE 3
8~====~--------------1
1_
LHDAyol
4
2
o+---~~--~----,---~~--~--~ o 80 120 180 2«1 300 380 TIME (min)
Figure 1 Examples of the three different LH pulse patterns. Data are collected from patient 24 (type 1), patient 4 (type 2), and patient 13 (type 3) on day o.
coxon's signed rank sum test. Correlations between hormone parameters were calculated with Spearman's rank correlation coefficient. Relationships between BMI and the response to naltrexone were calculated with a X2 contingency test. Differences between naltrexone responders and nonresponders were calculated by the Mann-Whitney test. The criterion of significance was set at P < 0.05. RESULTS
The effect of naltrexone on LH pulse characteristics in the three groups is presented in Figure 2. Armeanu et al. Naltrexone and LH pulse patterns
765
c 12f.::==;------.,
A
i 51
. a:
§
2
!i
1
2 2
2 PAnENr GROUPS
PAnENT GROUPS
2 PAnENr GROUPS
Figure 2 Effects of naltrexone on (A) LH pulse frequency, (B) LH pulse amplitude, and (C) mean LH level. Results are presented as means ± SEM. Significant differences in pulse frequency between day 0 and day 4 in groups 1 and 3; *, P < 0.05.
Naltrexone significantly increased the number ofLH pulses in group 1 from 1.6 ± 0.5 to 2.6 ± 0.6 pulses/ 6 hours (P < 0.05) and in group 3 from 6.2 ± 0.5 to 7.1 ± 0.6 pulses/6 hours (P < 0.05) but failed to do so in group 2. There was no significant change in either the LH pulse amplitude or the mean LH level in any of the groups. The distribution of LH pulse pattern types on day 0 and day 4 is presented in Figure 3. In group 1, two patients had a type 1 and two had a type 2 pulse pattern. These patients responded to naltrexone by an increased LH pulse frequency. In one of them, the pulse pattern changed from type 2 into type 3; in the others the pulse pattern remained unchanged. Six patients in group 1 had a type 3 pulse pattern, and this pattern was unchanged by naltrexone. All patients in group 2 had a type 3 pulse pattern. Only one patient responded to naltrexone by an increased pulse frequency and a changed pulse pattern (from type 3 into type 1). In group 3, all patients had type 1 pulse pattern before and after naltrexone administration. The relationship between BMI and the response to naltrexone is shown in Figure 4. Significantly less patients who were underweight (BMI < 20) responded to naltrexone, when compared with normal weight (BMI ~ 20) patients (X 2 = 9.14; P < 0.005; GROUP 1
A 120
IDa
100
B 120
.,0 I
100
d.,4
~:
20
20
C 120
IDII .,0 I .,4
100
i
80 ",80
:.,.40
UI PULSE PAnERN
766
GROUP 2
i
:.,.40
Figure 3 group 3.
n = 25). The underweight nonresponders seemed to have a lower baseline E2 level (median 64, range 38 to 215) than the normal weight naltrexone responders (median 126, range 49 to 216), but this difference was not significant (0.05 < P ::;; 0.1). Eight of the 12 naltrexone nonresponders had a type 3 pulse pattern and were underweight. In contrast, 8 of the 13 naltrexone responders had a type 1, 2 of the 13 had a type 2 pulse pattern, and all these 10 patients had a normal BMI. The baseline E2 level was significantly lower (P < 0.025) in the 8 patients with type 3 pulse pattern who were underweight and did not respond to naltrexone (median 61, range 38 to 150) than in the 10 normal weight responders who had types 1 and 2 pulse patterns (median 126, range 49 to 216). It seems that type 3 pulse pattern tends to be more frequent among underweight, low estrogenic, naltrexone nonresponders. The effects of naltrexone on other endocrine parameters are presented in Table 2. The median FSH level was not changed by naltrexone in any of the groups during the pulse study. Estradiol was significantly increased after naltrexone therapy in group 1 (P < 0.05) but was unaffected in groups 2 and 3. Testosterone was significantly increased (P < 0.05) in group 3 but was unchanged in groups 1 and 2. Naltrexone had no effect on PRL and A concentraGROUP 3
m a
.,4
80
",80
:.,.40 20
UI PULSE PATTERN
UI PUI.SE PAnERN
Effect of naltrexone on the distribution of LH pulse patterns in (A) group 1, (B) group 2, and (C)
Armeanu et al.
Naltrexone and LH pulse patterns
Fertility and Sterility
50
tite) , and one patient reported sleepiness and muscular weakness. No side effects were seen in groups 1 and 2. There were no changes in hepatic and renal functions, hemoglobin, or erythrocyte sedimentation rate in any of the patients. There were no dropouts from the study.
• 40
Q' E
l
•
• •
30
5i
III
20
.,•••••.
:i'•
10
DISCUSSION
•
NXNO
NXYES
Figure 4 Differences of BMI (kg/m2 ) between naltrexone responders (NX YES) and naltrexone nonresponders (NX NO). Data from groups 1, 2, and 3 (n = 25).
tion (results not shown). Progesterone had anovulatory values (0.3 to 0.8 nmol/L) before and after naltrexone in all patients. Naltrexone did not cause any significant change in the maximal LH and FSH response to GnRH (results not shown). The LH response was higher than the FSH response in all patients both on day o and day 4. No prepubertal response was seen in the patients with hypothalamic amenorrhea. In patients from group 1 and 2 who responded to naltrexone, a negative correlation (r = -0.8 and P < 0.05; n = 7, group 1 and 2 together) was found between the E2 level and the LH response to GnRH on day o. No such correlation was found in the nonresponders (r = 0.08; n = 8, group 1 and 2 together). Side Effects
Naltrexone treatment during 4 days in a dosage of 50 mg/d was well tolerated by most of the patients. Minor side effects were noticed in 16% of the patients. Three patients in group 3 reported mild gastrointestinal complaints (nausea and lack of appe-
Table 2
Twelve of the 16 patients in groups 1 and 2 had type 3 pulse pattern. This pattern is normally seen in the luteal phase of the menstrual cycle (1, 5, 16). Administration of naltrexone (17) to women during the luteal phase increased the LH pulse frequency, demonstrating that endogenous opioids are responsible for the typical LH pulse pattern during the luteal phase. Administration of P (18) or a combination of E2 and P (19) caused a similar pulse pattern. These reports demonstrate that the type 3 pattern in the luteal phase is caused by P in the presence of E2 through mediation of endogenous opioids. Our study could not demonstrate that endogenous opioids are also involved in the type 3 LH secretion pattern in hypothalamic amenorrhea because, in the majority of the patients, this pulse pattern was not modified by naltrexone. These results suggest that the type 3 pulse pattern may be mediated by both opioid and nonopioid pathways. In the patients with hypothalamic amenorrhea and anorexia nervosa who did not respond to naltrexone, no correlation was found between E2 level and the LH response to GnRH. This finding demonstrates an impaired negative feedback of gonadal steroids, which may explain the lack of response to naltrexone in these patients. Furthermore, underweight naltrexone nonresponders with type 3 pulse pattern had significantly lower E2 levels in comparison with the naltrexone responders with type 1 and 2 patterns. These results confirm that the respon-
Levels of FSH*, E 2t, and T:j: Before and After Naltrexone FSH
Patients
Day 0
Group 1 Group 2 Group 3
5.7 (2.1 to 7.9) 6.8 (4.2 to 10.8) 5.8 (3.2 to 8.9)
Day 4
Day 0
5.6 (2.9 to 8.2) 7.9 (3.5 to 9.5) 6.1 (3.8 to 8.5)
76 (38 to 216) 67 (40 to 117) 145 (103 to 251)
U/L
Day 4
Day 0
125 (38 to 480) § 63 (50 to 207) 170 (113 to 274)
1.3 (1.0 to 1.7) 1.2 (1.0 to 1.4) 2.0 (1.0 to 4.3)
pmol/L
* Reference values 3 to 12 U /L. t Reference values 110 to 180 pmol/L. :j: Reference values < 2.5 nmol/L.
Vol. 57, No.4, April 1992
T
E2
Day 4 nmol/L
1.5 (1.0 to 2.0) 1.0 (1.0 to 1.9) 3.6 (1.1 to 5.6) II
§ Significant difference (P < 0.05 in E between day 4 and dayO. II Significant difference (P < 0.05) in T between day 4 and dayO. Armeanu et al.
Naltrexone and LH pulse patterns
767
siveness to opioid antagonists depends on the activation of the endogenous opioid system by E2 (20, 21). Obviously, an E2 independent factor, which is not mediated by endogenous opioids, inhibits GnRH secretion in naltrexone nonresponders with hypothalamic amenorrhea and type 3 pulse pattern. A link between this inhibiting factor and underweight is suggested by the fact that 75% of the patients with type 3 pulse pattern were underweight and did not respond to naltrexone. This hypothesis is supported by evidence that inhibitory effects on gonadotropins, caused by nutritional deprivation and by E2 in castrated ewes, are additive (22). The nature of this undernutrition-linked factor still remains unclear. Corticotropin-releasing factor (CRF), which is known to mediate stress and feeding behavior, inhibits GnRH secretion partly by an opioid dependent and partly by an opioid independent pathway (23) and might be involved in the inhibitory action of this undernutrition-linked factor. We might speculate that this undernutrition-linked factor is a remainder of the prepubertal central restraint, which is re-activated by undernutrition in combination with a low-steroid environment. The inhibition of LH secretion seen in the early puerperium, which cannot be antagonized by naltrexone (24) either, might also be caused by the rapid weight loss, occurring in this period and might be mediated by this undernutrition-linked factor too. The fact that patients with anorexia nervosa failed to respond to naltrexone is partly in disagreement with Baranovska et al. (6). These authors observed an increase of LH levels after administration of naloxone in patients with anorexia nervosa in whom the amenorrhea preceded the weight loss. This discrepancy may be explained by a different methodology. The mean LH levels measured by us are based on 36 blood samples obtained during 6 hours after 4 days of naltrexone, whereas Baranovska et al. (6) based their results on single blood samples taken 30 minutes after a bolus naloxone injection. Furthermore, the latter authors did not observe any increase in LH pulse frequency and amplitude when blood sampling was performed during naloxone infusion for 4 hours, agreeing with our findings. Naltrexone caused significant increments of LH pulse frequency and E2 levels in responsive patients with hypothalamic amenorrhea, resulting from a diminished inhibition of GnRH pulse frequency. Based on the previously observed consistency of LH pulse patterns in patients with hypogonadotropic amenorrhea on different sampling occasions (16), we concluded that any changes in LH secretion may 768
Armeanu et al.
Naltrexone and LH pulse patterns
be attributed to naltrexone. However, we realize that interpretation of these results is controversial because of the lack of a randomized, placebo-controlled design of the study. Thus, it cannot be excluded that the changes in LH secretion may at least partly be caused by a placebo effect. Within these limitations of our study design, our results are in agreement with previous reports (1,5) that demonstrated that opioid inhibition is involved in the pathophysiology of hypothalamic amenorrhea in certain patients. In spite of the strong antagonistic activity and the chronic treatment, naltrexone was not able to stimulate LH secretion in all patients. Our results do not rule out the possibility that naltrexone causes subtle increments of GnRH secretion, but the pituitary is not sensitive enough and does not respond with a similar LH secretion in the naltrexone nonresponders. Further experiments with naltrexone administration after GnRH priming of the pituitary may elucidate this aspect. The LH pulse pattern in patients with polycystic ovarian disease is similar to type 1 pulse patterns seen in the follicular phase of the menstrual cycle (16) and does not change after naltrexone treatment, although LH pulse frequency increases. Our results confirm the findings of Berga and Yen (25) who reported an intact opioidergic inhibition in patients with polycystic ovarian disease. These authors demonstrated a response to naloxone after pretreatment with medroxyprogesterone. Our study demonstrates that naltrexone increases LH pulse frequency in patients with polycystic ovarian disease without P pretreatment as well. Contrary to Petraglia et al. (9), who found a link between the response to naloxone and obesity, we found that the naltrexone response also occurred in nonobese patients with polycystic ovarian disease. The differences between naloxone and naltrexone may be because of the longer treatment (4 days) and the stronger activity of naltrexone. The higher LH pulse frequency demonstrates an increased GnRH pulse secretion. The T concentrations in plasma of women with polycystic ovarian disease were significantly increased after 4 days ofnaltrexone therapy. This may be the result of the naltrexone-induced increased LH secretion, as suggested by the significant increase of LH pulse frequency and the small, but not significant, increase of LH levels in group 3 (Fig.2C). The secretion of FSH was not stimulated by chronic treatment with naltrexone in the patients with oligomenorrhea or amenorrhea. Either the increment in GnRH pulsatility induced by naltrexone Fertility and Sterility
is insufficient to cause an increase in pituitary FSH secretion or the FSH secretion may still be inhibited by neurotransmitters and/or neuromodulators other than the endogenous opioids. However, it is possible that small changes in FSH secretion may be present but cannot be detected because of limitations inherent to the assay and/or analysis techniques. In conclusion, the oral treatment with the longacting antagonist naltrexone enabled us to study the effect of a prolonged blockade of the opioid receptors on the LH pulsatility in patients with hypothalamic amenorrhea and polycystic ovarian disease. The response in 60% of the patients with hypothalamic amenorrhea because of simple weight loss, stress, and physical exercise demonstrates the involvement of endogenous opioids in the pathophysiology of this condition. The GnRH inhibition in patients with anorexia nervosa and other weight loss-related forms of secondary amenorrhea is probably caused by an opioid-independent, undernutrition-linked factor that remains to be identified. The response in patients with polycystic ovarian disease demonstrates an intact central opioid activity in these patients. Because 4 days of naltrexone treatment only slightly stimulated LH pulse frequency but had no effect on FSH secretion, it is questionable whether a longer treatment with naltrexone would be effective in ovulation induction in these forms of oligoamenorrhea. Ackrwwledgments. We are grateful to the staff of the Endocrinology Laboratory of the Free University Hospital for performing the hormone assays and to Mr. Tjarco van Wijngaarden from the Audio-Visual-Center of the Free University Hospital for the technical assistance in preparing the illustrations for this manuscript. We thank Du Pont Pharma, Bad Homburg, Germany, for the kindness of providing naltrexone (Trexan) for this clinical study.
REFERENCES 1. Khoury SA, Reame NE, Kelch RP, Marshall JC. Diurnal patterns of pulsatile luteinizing hormone secretion in hypothalamic amenorrhea: reproducibility and response to opiate blockade and an a2-adrenergic agonist. J Clin Endocrinol Metab 1987;64:755-62. 2. Judd S, Stranks S, Michailov L. Gonadotropin-releasing hormone pacemaker sensitivity to negative feedback inhibition by estradiol in women with hypothalamic amenorrhea. Fertil Steril 1989;51:257-62. 3. Genazzani AD, Petraglia F, Fabbri G, Monzani A, Montanini V, Genazzani AR. Evidence of luteinizing hormone secretion in hypothalamic amenorrhea associated with weight loss. Fertil Steril 1990;54:222-6. 4. Burger CW, Korsen T, Van Kessel H, Van Dop PA, Caron F JM, Schoemaker J. Pulsatile luteinizing hormone patterns Vol. 57, No.4, April 1992
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
in the follicular phase of the menstrual cycle, polycystic ovarylike syndrome (PCOD) and non-PCOD secondary amenorrhea. J Clin Endocrinol Metab 1985;61:1126-32. Quigley E, Sheehan KL, Casper RF, Yen SSC. Evidence for increased dopaminergic and opioid activity in patients with hypothalamic hypogonadotropic amenorrhea. J Clin Endocrinol Metab 1980;50:949-54. Baranovska B, Rozbicka G, Jeske W, Abdel-Fattah MH. The role of endogenous opiates in the mechanism of inhibited luteinizing hormone (LH) secretion in women with anorexia nervosa: the effect of naloxone on LH, follicle-stimulating hormone, prolactin and {:1-endorphin secretion. J Clin Endocrinol Metab 1984;59:412-6. Giusti M, Torre R, Traverso L, Cavagnaro P, Attanasio R, Giordano G. Endogenous opioid blockade and gonadotropin secretion: role of pulsatile luteinizing hormone-releasing hormone administration in anorexia nervosa and weight loss amenorrhea. Fertil Steril 1988;49:797-801. Kazer RR, Kessel B, Yen SSC. Circulating luteinizing hormone pulse frequency in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1987;65:233-6. Petraglia F, D' Ambrogio G, Comitini G, Facchinetti F, Volpe A, Genazzani AR. Impairment of opioid control of luteinizing hormone secretion in menstrual disorders. Fertil Steril 1985;43:534-40. Teoh SK, Mendelson JH, Mello NK, Skupny A. Alcohol effects on naltrexone-induced stimulation of pituitary, adrenal, and gonadal hormones during the early follicular phase of the menstrual cycle. J Clin Endocrinol Metab 1988;66: 1181-6. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 3rd ed. Washington D.C.: American Psychiatric Association, 1987. Wennink JMB, Delemarre-Van de Waal HA, Van Kessel H, Mulder GH, Foster JP, Schoemaker J. Luteinizing hormone secretion patterns in boys at the onset of puberty measured using a highly sensitive immunoradiometric assay. J Clin Endocrinol Metab 1988;67:924-8. Lambalk CB, De Koning J, Van Kessel H, Van Rees G, Schoemaker J. Calculation of the intra assay variation per assay and the relevance to LH pulse detection. IRCS Med Sci 1985;13:1183-4. Scheele F, Lambalk CB, Schoemaker J, Van Kessel H, De Koning J, Van Dieten JAM, et al. Patterns of LH and FSH in men during high-frequency blood sampling. J Endocrinol 1987;114:153-60. Armitage P. Statistical methods in medical research. Oxford: Blackwell Scientific Publications, 1983. Schoemaker J, Schoemaker H, Berkhout G, Schoemaker HC. Pulsatile luteinizing hormone secretion in hypothalamic amenorrhea. In: Genazzani AR, Petraglia F, Volpe A, Facchinetti F, editors. Advances in gynecological endocrinology. Carnforth (UK): The Parthenon Publishing Group, 1988: 43-50. GindoffPR, Jewelewicz R, Hembree W, Wardlaw S, Ferin M. Sustained effects of opioid antagonism during the normal human luteal phase. J Clin Endocrinol Metab 1988;66: 1000-4. Buckler HM, Phillips SE, Cameron IT, Healy DL, Burger HG. Vaginal progesterone administration before ovulation induction with exogenous gonadotropins in polycystic ovarian syndrome. J Clin Endocrinol Metab 1988;67:300-7.
Armeanu et aI.
Naltrexone and LH pulse patterns
769
19. Nippoldt TB, Reame NE, Kelch RP, Marshall JC. The role
of estradiol and progesterone in decreasing luteinizing hormone pulse frequency in the luteal phase of the menstrual cycle. J Clin Endocrinol Metab 1989;69:67-76. 20. Reid RL, Quigley ME, Yen SSC. The disappearance of opioider~c regulation of gonadotropin secretion in postmenopausal women. J Clin Endocrinol Metab 1983;57: 1107-11. 21. Chowen JA, Argente J, Vician L, Clifton DK, Steiner RA. Pro-opiomelanocortin messenger RNA in hypothalamic neurons is increased by testosterone through aromatization to estradiol. Neuroendocrinology 1990;52:581-8. 22. Kyle JP, Alexander BM, Moss GE, Hallford DM, Nett TM. Gonadotropin-releasing hormone overrides the negative effect
770
Armeanu et al.
Naltrexone and LH pulse patterns
of reduced dietary energy on gonadotropin synthesis and secretion in ewes. Endocrinology 1991;128:843-9. 23. Williams CL, Nishihara M, ThaiabardJ, Hotchkiss J, Knobil E. Corticotropin-releasing factor and gonadotropin-releasing hormone pulse generator activity in the rhesus monkey. Electrophysiological studies. Neuroendocrinology 1990;52: 133-7. 24. Kremer JA, Borm G, Schellekens LA, Thomas CMG, Rolland R. Pulsatile secretion of luteinizing hormone and prolactin in lactating and nonlactating women and the response to naltrexone. J Clin Endocrinol Metab 1991;72:294-300. 25. Berga SL, Yen SSC. Opioidergic regulation of LH pulsatility in women with polycystic ovary syndrome. Clin Endocrinol (Oxf) 1989;30:177-84.
Fertility and Sterility