0021-972X/90/7201-0116$02.00/0 Journal of Clinical Endocrinology and Metabolism Copyright© 1991 by The Endocrine Society

Vol. 72, No. 1 Printed in U.S.A.

Gonadotropin, Steroid, and Inhibin Levels in Women with Incipient Ovarian Failure during Anovulatory and Ovulatory Rebound Cycles* H. M. BUCKLERt, C. A. EVANS, H. MAMTORA, H. G. BURGER, AND D. C. ANDERSON Departments of Medicine (Endocrinology) and Radiology (H.M.), Hope Hospital (University of Manchester School of Medicine), Salford, United Kingdom; and the Medical Research Centre, Prince Henry's Hospital (H.G.B.), Melbourne, Australia

ABSTRACT. We have identified a group of women with infertility and regular menses who have persistently raised FSH levels and probable incipient ovarian failure (IOF). Thirteen such women (19 cycles) had serum samples taken for RIA of LH, FSH, estradiol, and progesterone (P) 3 times a week over 1 menstrual cycle. Sixty infertile women with normal ovulatory cycles (as determined by hormones and ultrasound scan) served as controls. Overall, the FSH was higher (P < 0.01) on all days of the cycle in the IOF group, serum LH was raised on days-14 to-5 before and days 5-11 after the LH surge. There was no difference between estradiol and P levels in the two groups. Ultrasound scanning showed failure of normal ovulation in the IOF group. Inhibin, measured by RIA in 9 cycles in the IOF group was lower (P < 0.01) during the follicular phase than in 43 normal cycles. The highest inhibin level was seen in the luteal

phase, as in normal cycles, but levels were still lower (P < 0.01) in the IOF group. Inhibin was inversely correlated with FSH (P < 0.05) during the follicular and luteal phases and was correlated with P during the luteal phase (P < 0.05) in the IOF group. After 3 weeks of suppression (39 cycles) with an estrogenprogestogen preparation in the IOF group, LH and FSH fell to normal values. Ovulation occurred in 22 cycles on withdrawal of suppression in the presence of high FSH levels and low inhibin levels. No pregnancies occurred. These findings are consistent with the suggestion that diminished ovarian inhibin secretion may contribute to the elevated FSH levels of IOF and indicate that ovulation in the rebound cycle after suppression occurs in the presence of high FSH and low inhibin levels. Such cycles, however, still appear to be subfertile. (J Clin Endocrinol Metab 72: 116-124, 1991)

T

HE TIME preceding the menopause is characterized by variability in menstrual cycle length (1, 2) and is normally referred to as the perimenopausal or transitional phase. Early endocrine studies suggested that it was associated with normal urinary estrogen levels and raised levels of gonadotropins, especially LH and, to a lesser extent, FSH (3, 4). However, Sherman et al. (5) demonstrated lower levels of serum estradiol (E2) during the early follicular, midcycle, and luteal phases in perimenopausal than in young women. They also found raised levels of FSH, particularly in the early follicular phase, in the presence of normal LH levels. Serum FSH levels have been found to increase with age, and significantly higher concentrations were found in a 40- to 50yr-old group than in a 20- to 29-yr-old group (6). LH levels did not show a similar increase with age, although

a follicular phase rise in LH was found in the older group. In this study ovarian E2 production and ovulation were maintained. It appears that women over 40 yrs of age experience a monotropic rise in FSH as the first detectable endocrine manifestation of reproductive aging (5, 7), and this can occur while regular menstruation continues (6-8). It is not known whether women approaching a premature menopause exhibit hormone changes similar to those of women approaching the normal menopausal transition. We have identified a group of infertile women with increased FSH concentrations and fairly regular menses who may be suffering from incipient premature ovarian failure (IOF). We have attempted to induce rebound ovulation in this group of women by suppressing the inappropriately raised FSH levels with an estrogen and progestogen preparation (9). The rebound cycle is usually ovulatory, although, unfortunately, this treatment does not appear to restore fertility in such women. We here report the levels of sex steroids and gonadotropins in 19 untreated and 22 rebound cycles and changes in inhibin in a subgroup of 9 basal and 10 rebound cycles.

Received August 28, 1989. Address requests for reprints to: Dr. H. M. Buckler, Department of Medicine, Clinical Science Building, Hope Hospital, Eccles Old Road, Salford, M6 8HD United Kingdom. * This work was supported by the National Health and Medical Research Council (Australia). t Recipient of a Wellcome Research travel grant.

116

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 23:11 For personal use only. No other uses without permission. . All rights reserved.

INHIBIN AND INCIPIENT OVARIAN FAILURE

Our findings throw further light on the perimenopausal state and the role of inhibin in it. Materials and Methods Subjects An assessment of ovulation was made in women attending our infertility clinic by a combination of ultrasound and detailed hormone analysis. Serial ultrasound scans were carried out with a Technicare Autosector real time scanner with a 5mHz probe or an ALOKA SSD-T56 linear array real time scanner with a 3.5-mHz probe. Scans were started from around day 7 of each cycle and then carried out on alternate week days. Scanning was not stopped until after the demonstration of presumptive ovulation. Serum levels of LH, FSH, E2, and progesterone (P) were measured three times weekly over the whole menstrual cycle. The criteria of normal (ovulatory) menstrual cycles is as follows: cycle length of 26-32 days, the presence of a well defined midcycle LH peak (at least a 2-fold rise from follicular phase levels), luteal phase length of 13 days or more, a midluteal P concentration greater than 25 nmol/L, good follicular growth, and ovulation detected ultrasonically by either disappearance or collapse of the preovulatory follicle within 48 h of the recorded LH peak. We have previously reported our initial experience of using this method of assessment and have determined a normal reference range for different days of the cycle (10). Many abnormalities have been identified, one of which is the combination of anovulation, inappropriately raised FSH levels, and regular or nearly regular menstrual bleeding. Women with the above abnormality compared with our previous normal reference range were recruited into the study. All patients gave informed consent, and the study was approved by the Salford Ethics Committee. The clinical details of the 13 women in 19 study cycles are shown in Table 1. All women had long-standing unexplained infertility, fairly regular cycles (menses occurring every 3-6 weeks), and postcoital tests, laparoscopy, and PRL measurements within the normal range. Semen analyses (also normal) were carried out on 2 occasions on each male partner, and the lower limits of normal accepted were: volume, 2 mL; sperm count, 20 X 106/mL; motility, 70%; and normal forms, 80%. All women in the study underwent detailed assessment of ovulation, as described above, on at least 1 occasion (19 cycles). Inhibin was measured by RIA from samples taken 3 times weekly over 1 menstrual cycle from 9 of the above cycles selected randomly. Control groups Sixty infertile women with normal ovulatory cycles (as determined by hormones and ultrasound scanning) served as a control group for LH, FSH, E2, and P levels. Daily venous blood samples were collected from 33 normally ovulating North American women, aged 25-35 yr, and were measured for inhibin by RIA. Daily venous blood samples were also collected from 10 other normal women who were judged to be ovulating normally on the basis of their menstrual calendars and luteal phase P concentrations (>25 nmol/L). Data from these subjects

117

have been presented previously (11). Comparison of the inhibin data from these 2 groups showed no difference, so data from all of the subjects were combined to give a control group of 43 subjects for comparison of inhibin levels with the study group. An attempt was made to suppress the raised FSH levels in the study group using various estrogen/progestogen preparations. Thirteen patients received either Microgynon [ 30 fig ethinyl estradiol (EE2) and 150 ^g levonorgestrel], Marvelon (30 fig EE2 and 150 fig desogestrel), or 30 /ig EE2 and 10 mg medroxyprogesterone acetate (MPA). Ten women received EE2 and MPA for 14 cycles, and 6 of these women and 3 others were allocated at random to receive Marvelon (8 cycles), Microgynon (9 cycles), and EE2 and MPA (8 further cycles). The estrogen-progestogen preparation was initially administered for 4 weeks (n = 20). In the later stages of this study this was reduced to 3 weeks (n = 19). The preparation was started on days 1-3 of menstrual bleeding. LH, FSH, E2, and P were measured weekly during treatment. The cycle after stopping treatment (rebound cycle) was assessed by detailed hormone and ultrasound measurements, as previously described. In 10 rebound cycles (selected at random) inhibin was also measured by RIA. RIA Inhibin RIA. Serum inhibin was measured by a previously described heterologous RIA (11, 12), except that a new inhibin antiserum (As 1989) was used, raised in a rabbit by immunization initially against crude bovine follicular fluid, with subsequent boosting to purified 31-kDa inhibin. Inhibin was undetectable using this RIA in the serum of castrate men and postmenopausal women, and its sensitivity was enhanced compared to that of the previous method. This antiserum showed less than 0.5% cross-reactivity to purified bovine activin, Mullerian inhibitory substance, and both human and porcine transforming growth factor-/?. Briefly, the assay uses 125I-labeled pure bovine 31-kDa inhibin as tracer and a serum standard consisting of a pool from women undergoing ovarian hyperstimulation for in vitro fertilization. The standard was assigned a potency of 3.32 U/mL against the previously used partially purified human follicular fluid standard (11). The interassay coefficient (CV) of variation was 8% in 12 assays. The CVs in the upper, mid-, and lower regions of the standard curves were 8%, 5%, and 6%, respectively. The within-assay CV, obtained from the mean index of precision of 12 assays, was 0.064. Gonadotropin RIA. LH and FSH were assayed by in-house RIA, using a solid phase double antibody separation method and reference standards 68/40 and 78/549, respectively. Specimens from individual patients were assayed within one batch. Between-batch precision over the working range of the assays was 12%. Sex steroid assays. P was assayed by a direct method (13). E2 was assayed using the Steranti kit (Steranti Research Ltd., St. Albans, Herts, United Kingdom). Between-batch precision over the working range was 6.8% and 10% for P and E2, respectively.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 23:11 For personal use only. No other uses without permission. . All rights reserved.

118

BUCKLER ET

AL.

JCE & M • 1991 Vol 72 • No 1

TABLE 1. Summary of patient characteristics: baseline cycles Cycle/patient

Age (yr)

Infertility0

Cycle length (days)

Ka)

38

S-6

27

LUF R. ovary: 18-mm follicular structure on day 9, which was 22-mm on day 2 L. ovary: 14-mm follicular on day 0, which increased to 18-mm on day 2

Kb)

38

S-6

19

LUF R. ovary: 30-mm cyst present on day 0. This was present up to day 6 and then disappeared L. ovary: cystic structure of 18-mm seen on day 0, progressively increased to 25 mm by day 8, then shrank.

Kc)

38

S-6

43

Abnormal follicular development A cystic structure grew to 36 mm in the follicular phase but had disappeared by day of LH surge.

32

P-7

32

Inconclusive R. ovary: 2 follicles of 20 mm present on day -12; 2 further follicles grew to 24 and 25 mm by day — but then shrank; 1 follicle of 15 mm on day 0. L. ovary: cystic structure of 14 mm on day —4 andO

3(a)

29

S-6

35

Absence of follicular growth

3(b)

29

S-6

30

Abnormal follicular development Max. follicular diameter reached was 13 mm

27

P-4

25

Bilateral immature follicules, max. 14 mm diameter

30

P-2

37

Abnormal follicular development R. ovary: single follicle reaching 10 mm only L. ovary: Follicle reached 22 mm on day —12, but shrank gradually to 5 mm by day 0

33

S-4

27

Bilateral immature and multiple follicles

38

S-6

34

Abnormal follicular development L. ovary: follicle of 11-mm, which was still present in luteal phase

38

S-6

27

Normal ovulation from R. ovary; follicle diameter, 24 mm; follicle at 10 mm on L. present throughout cycle

37

P-4

30

LUF R. ovary: 1 follicle grew to 26 mm by day 0, but remained unchanged until day 7

9(a)

Main ultrasound findings

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 23:11 For personal use only. No other uses without permission. . All rights reserved.

INHIBIN AND INCIPIENT OVARIAN FAILURE

119

TABLE I—Continued

Cycle/patient

Age (yr)

Infertility"

Cycle length (days)

Main ultrasound findings

9(b)

37

P-4

27

Possible ovulation R. ovary: 1 follicle, 5 mm on day, 0. L. ovary: 3 follicles, 17, 6, and 5 mm on day —2, and 1 follicle, 20 mm, on day 0. Nil visible in either ovary on day +3

9(c)

37

P-4

24

Absence of normal follicular maturation R. ovary: 1 follicle, max, 5 mm L. ovary: 1 follicle, max, 8 mm

10(a)

34

P-7

26

10(b)

34

P-7

27

11

36

P-7

26

12

36

S-3

28

13

29

P-8

31

27.9 27-38

4.4 2-8

29.2 19-43

Absence of normal follicular maturation L. ovary: 1 follicle of 11 mm, which persisted in luteal phase Multiple small follicles Multiple small follicles, max, 12 mm Insufficient ultrasound examination Absence of normal follicular maturation

Mean Range " P, Primary; S, secondary.

Statistics The observations during all cycles were synchronized to the day of the LH peak (day 0). To overcome the problem of small

numbers of observations on some days of the cycles, results from days 1-3, 4-6, 7-9, 10-12, and 13-15 before or after the LH peak were grouped together and were called days 2, 5, 8, 11, and 14, respectively. The mean and 95% confidence limits of the five hormones were calculated on values after logarithmic transformation. Paired and unpaired Student's t tests were employed for statistical analysis. The correlation coefficient was used to test the relationships between two variables.

Results Baseline cycles

FSH levels were raised in the IOF group (P < 0.001), and this was particularly evident in the early follicular phase (Fig. 1). Serum LH was raised on days —14 to —8 (P < 0.001), day - 5 (P < 0.01), day 5 (P < 0.01), day 8 (P < 0.001), and day 11 (P < 0.01) relative to the LH surge. There was no difference in E2 and P levels between the study and control groups. Inhibin levels were lower in the IOF group throughout the cycle (P < 0.001). However, they showed the same pattern over the cycle as the normal group, with the highest levels in the luteal phase, secreted in a pattern similar to that for P secre-

tion. There was an inverse relationship between FSH and inhibin during the follicular phase (days -14 to —2; r = —0.54; n = 48; P < 0.05) and during the luteal phase

(days 2-14; r = -0.52; P < 0.05). Inhibin was positively correlated with P during the luteal phase (r = 0.67; n = 48; P < 0.01). Various ultrasound abnormalities were seen (Table 1); these included abnormal follicular development, multiple development of immature follicles, luteinized unruptured follicles, and development of luteal cysts. Seventeen cycles were anovulatory. One basal cycle was ovulatory (Table 1, no. 8), one showed possible ovulation (Table 1, no. 9b), and in another there was insufficient ultrasound data (Table 1, no. 12). Eleven of the 13 women have developed amenorrhoea and raised gonadotropin levels since completing the study. We feel that they now have established premature ovarian failure. Estrogen/progestogen treatment All three estrogen/progestogen combinations produced a significant fall in gonadotropin levels (P < 0.01) by 1 week of treatment, which continued to fall until 3 weeks of treatment. No further suppression of gonadotropins was produced by continuing treatment for a further week. A greater fall in LH and FSH occurred with Microgynon

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 20 November 2015. at 23:11 For personal use only. No other uses without permission. . All rights reserved.

BUCKLER ET AL.

120

JCE & M • 1991 Vol 72 • No 1

30-

B

2724-

10090-

LH Levels (IU/L)

21-

80-

FSH Levels

70-

(IU/L)

1815-

60-

12-

5040-

9-

30-

6-

203-

100-14

-11

-8

-5

-2

0

+2

+5

+8

+11

0-14

+14

-11

-8

-5

-2

0

+2

+5

+8

+11

+14

+5

+8

+11

+14

Day of the Cycle

Day of the Cycle

D

2400

3000-

2100

2700-

1800

24002100Inhibin 1800Levels (U/L) 1500-

E2 Levels

1500

(pmol/L)

1200 900

1200-

600

900bOO-

300

3000

0-8

-5

E

+8

-2 0 +2 Day of the Cycle

+11

-14

+14

-11

-5

-2

0

+2

Day of the Cycle

150135120105-

P Levels

90-

(nmol/L)

75604530150-

-8

-5

-2

0

+2

+5

+8

+11

+14

Day of the Cycle

FIG. 1. LH (A), FSH (B), inhibin (C), E2 (D), and P (E) levels in all incipient ovarian failure cycles (•). The geometric mean and 67% confidence limits are shown. The geometric mean (O) and the 95% confidence limits of the normal control group are also shown in the shaded area. All data are log transformed.

and Marvelon than with EE2 and MPA. Microgynon and Marvelon produced a fall in E2 by 1 week, which continued until the third week. Again, no further decrease was seen by continuing for a further week. By 3 weeks of treatment, the E2 level was below castrate levels (