Serum androgen and gonadotropin levels decline after progestogen-induced withdrawal bleeding in oligomenorrheic women with or without polycystic ovaries*

Vol. 58, No.4, October 1992

FERTILITY AND STERILITY

Printed on acid-free paper in U.S.A.

Copyright e 1992 The American Fertility Society

Serum androgen and gonadotropin levels decline after progestogeninduced withdrawal bleeding in oligomenorrheic women with or without polycystic ovaries·

Leena Anttila, M.D.t:j: Pertti Koskinen, M.D.§ Hanna-Leena Kaihola, Ph.D.§

Risto Erkkola, M.D.t Kerttu Irjala, M.D.§ Kristiina Ruutiainen, M.D.tll

University of Turku, Turku, Finland

Objective: To examine the effect of short-term progestogen treatment on androgen, gonadotropin, and sex hormone-binding globulin (SHBG) levels in oligomenorrheic women. Design: Comparative study of changes in hormonal parameters in patients with or without ultrasonographically diagnosed polycystic ovarian disease (PCOD). Setting: Open patient clinic of reproductive endocrinology at University Central Hospital of Turku, Finland. Patients: Seventy-five oligomenorrheic women with (n = 51) or without (n = 24) PCOD. Main Outcome Measures: Serum concentrations of testosterone (T), androstenedione (A), dehydroepiandrosterone sulfate, luteinizing hormone (LH), follicle-stimulating hormone (FSH), andSHBG. Results: The levels ofT, A, LH, and the LH:FSH ratios decreased significantly after oral treatment with medroxyprogesterone acetate (10 mg/d for 10 days) in non-PCOD women and in women with PCOD decreasing the frequencies of pathological laboratory findings, in particular elevated levels of LH:FSH ratio and A in PCOD women and of LH:FSH ratio in non-PCOD women. The levels ofT, A, and LH as well as the LH:FSH ratio were significantly higher in women with PCOD. Obesity was associated with high free androgen indices, low LH:FSH ratios, and low concentrations of LH, A,andSHBG. Conclusions: The serum samples for hormonal analyses used as an aid in diagnosing PCOD should be obtained without pretreatment with progestogen because it masks the biochemical findings of PCOD. Fertil Steril1992;58:697-702 Key Words: Polycystic ovarian disease, oligomenorrhea, androgens, gonadotropins, medroxyprogesterone acetate

Polycystic ovarian disease (PCOD) is a common cause of amenorrhea, oligomenorrhea, and infertility (1, 2). The etiology and pathophysiology of this Received April 22, 1992; revised and accepted June 24, 1992.

* Supported by grants from the Paulo Foundation, Helsinki, Finland, and the Foundation of the University of Turku, Turku, Finland. t Department of Obstetrics and Gynecology. :j: Reprint requests: Leena Anttila, M.D., Department of Obstetrics and Gynecology, Turku University Central Hospital, Kiinamyllynkatu 4-8, SF-20520 Turku, Finland. § Central Laboratory, Turku University Central Hospital. II Present address: Department of Obstetrics and Gynecology, The University of Maryland, Baltimore, Maryland. Vol. 58, No.4, October 1992

clinical entity are still incompletely understood. Likewise, several diagnostic criteria based on ovarian morphology (2-4) or biochemical findings have been presented for PCOD (5-7). In most studies on androgen metabolism in eumenorrheic women, the sampling for hormonal analyses has been carried out in the early follicular phase. In amenorrheic and oligomenorrheic subjects, the results of the hormonal analyses are presumed to become more comparable with the normal early follicular phase values by taking the samples after a progestogen-induced withdrawal bleeding (8). However, it may be questioned whether such treatment would not change the androgen and gonadoAnttila et al.

MPA effect on hormones in oligomenorrhea

697

tropin status of a patient with menstrual irregularities, thereby masking diagnostic hormonal findings. Cycle-regulating progestagens (e.g., medroxyprogesterone acetate [MPA)) are known to suppress testosterone (T) (9), luteinizing hormone (LH) (911), and sex hormone-binding globulin (SHBG) levels (12). The aim of the present study was to investigate the effect of progestogen treatment on the levels of serum androgens, gonadotropins, and SHBG in oligomenorrheic women with or without morphologically diagnosed PCOD. MATERIALS AND METHODS Subjects

The study group consisted of 75 oligomenorrheic women. Oligomenorrhea was defined as menstrual cycles of >6 weeks during the previous 6 to 12 months. Patients with hyperprolactinemia or thyroid dysfunction were excluded. All subjects attending the study bled in response to a progestogen challenge indicating an estrogen (E)-primed endometrium. The diagnosis of PCOD was based on ovarian morphology assessed by the same investigator (L.A.) using a 7.5-MHz vaginal ultrasonographic vaginal probe (Combison 310; Kretztechnik, Zipf, Austria) and diagnostic criteria by Adams (3). The body mass index (BMI, weight [kg]/height [m]2) was measured. Hair growth was assessed using a modification of the criteria of Ferriman and Gallwey (13). In hirsute and hyperandrogenic women, a dexamethasone suppression test was performed to exclude Cushing's syndrome. Adrenocorticotropin stimulation test was performed when needed to exclude late onset adrenal hyperplasia. Polycystic ovaries fulfilling diagnostic criteria (enlarged ovaries with multiple small subcortical follicles and increased stroma) were found in 51 women. Their age ranged from 14 to 40 years (27.9 ± 5.5 years, mean ± SD) and the BMI from 17 to 56 kg/m2 (28.8 ± 7.7 kg/m 2). The remaining 24 women had normal (n = 18) or multicystic (n = 6) ovaries on ultrasound. The age of the non-PCOD women ranged from 19 to 40 years (27.4 + 5.2 years) and the BMI from 18 to 48 kg/m2 (28.4 ± 9.2 kg/m2). Thirty-one women with and 21 without PCOD were obese (BMI > 25 kg/m 2). Hirsutism (total score> 9) were found in 11 women with PCOD. None of the women without PCOD were hirsute. Hormone Sampling

The blood samples were collected after an overnight fast between 7 and 9 A.M. The first samples 698

Anttila et at.

MPA effect on hormones in oligomenorrhea

were taken during oligomenorrhea. The sampling was repeated after an MPA-induced (MPA 10 mg/d for 10 days) withdrawal bleeding (days 3 to 7 after the onset of induced menstrual flow). Hormone and SHBG Assays

Serum T and androstenedione (A) were measured with in-house radioimmunoassays (RIAs) as described previously (14). Dehydroepiandrosterone sulfate (DHEAS) was measured with an RIA (1251_ DHEAS RIA kit; Baxter Dade AG, Dudingen, Germany). Luteinizing hormone and follicle-stimulating hormone (FSH) were measured with time-resolved immunofluorometric assays (Delfia; Wallac Oy, Turku, Finland). Sex hormone-binding globulin was determined with a ligand-binding assay (14). The free androgen index was calculated as the quotient T (nmoljL)/SHBG (nmoljL) X 1,000. The early follicular phase reference intervals for all the analytes were determined in 40 healthy female volunteers (14). Statistical Methods

Paired differences were evaluated with the paired t-test. The effects of PCOD and obesity in the entire study group and hirsutism in the subgroup with PCOD on the hormone and SHBG levels were assessed with ANCOVA. The changes in the distribution were analyzed with McNemar's X2test. RESULTS

The serum concentrations of T, A, and LH as well as the LH:FSH ratios decreased significantly after treatment with MPA in women with PCOD as well as in non-PCOD women (Table 1, Fig. 1). The levels ofDHEAS and the free androgen indices decreased significantly only in PCOD women (Table 1, Fig. 1). The levels of T, A, and LH as well as the LH: FSH ratios and the free androgen indices were higher in women with PCOD than in non-PCOD women both in the oligomenorrheic state and after MPA-induced bleeding (Table 1). The PCOD and non-PCOD women did not differ with respect to serum DHEAS concentrations (Table 1). Figure 2 shows the distributions of the women according to normal (outside the appropriate circle) or elevated (inside the circle) LH:FSH ratios, free androgen indices, and A concentrations in oligomenorrhea and after MPA-induced bleeding. The overlaping areas present those with more than Fertility and Sterility

Table 1

Gonadotropins, Androgens, and SHBG in the Study Groups* Cycle phase

LH (lUlL) LH:FSH ratio T (nmol/L) SHBG (nmol/L) Free androgen index A (nmol/L) DHEAS (ltmol/L)

Oligomenorrhea After MPA Oligomenorrhea After MPA Oligomenorrhea After MPA Oligomenorrhea After MPA Oligomenorrhea After MPA Oligomenorrhea After MPA Oligomenorrhea After MPA

PCODwomen 10.8 7.4 2.4 1.7 3.1 2.5 35.2 34.8 1Z0 98 13.3 10.3 7.0 7.6

± 5.5 ± 4.5 ± 1.1 ± 0.8 ± 0.9 ± 0.7 ± 20.5 ± 22.8 ±83 ± 64 ± 3.9 ± 3.4 ± 3.4 ± 4.1

(1.2 to 29):1: (1.4 to 28):1: (0.9 to 5.6):1: (0.4 to 4.2):1: (1.9 to 5.9):1: (1.3 to 4.7):1: (7 to 91) II (6 to 121) II (35 to 475):1: (26 to 300):1: (7.0 to 22.6):1: (5.1 to 20.1):1: (2.2 to 16.3) § (2.0 to 16.6) §

Probabilityt <0.01 <0.01 <0.05 <0.05 <0.001 <0.001 NS1f NS <0.001 <0.01 <0.001 <0.001 NS NS

Non-PCOD 7.3 4.9 1.9 1.1 2.0 1.8 42.0 38.4 56 55 7.6 6.7 5.8 6.0

± ± ± ± ± ± ± ± ± ± ± ± ± ±

5.5 2.8 1.1 0.6 0.4 0.3 22.6 19.0 24 24 2.5 1.4 2.9 2.3

(1 to 18) § (1.4 to 12) § (0.3 to 3.8) § (0.3 to 2.1) § (1.5 to 3.3) § (1.2 to 2.4) § (18 to 94) II (19 to 88) II (20 to 111) II (19 to 100) II (3.8 to 15.9) § (4.9 to 10.8) § (1.7 to 12.5) II (1.6 to 9.7) II

* Values are means ± SD with ranges in parentheses. t The difference between PCOD and non-PCOD women. :I: The difference between the results obtained during oligomenorrhea and after MPA-induced bleeding, P < 0.001.

§ The difference between the results obtained during oligomenorrhea and after MPA-induced bleeding, P < 0.01. II The difference between the results obtained during oligomenorrhea and after MPA-induced bleeding is not significant. 1f NS, not significant.

one abnormal finding. Treatment with MP A decreased the frequency of pathological laboratory findings, in particular elevated levels of LH:FSH ratio (from 66% to 38%, P < 0.01) and A (from 66% to 34%, P < 0.001) in PCOD women and elevated levels of LH:FSH ratio (from 57% to 13%, P < 0.05) in non-PCOD women. Seventy-seven percent of PCOD women had a combination of more than one pathological finding. The frequency decreased to 51 % after MP A in the PCOD group (P < 0.01). Among the non-PCOD group, elevated LH:FSH ratio together with elevated free androgen index was found in 24 % of subjects in the nonPCOD group during oligomenorrhea. Only one of the non-PCOD women had high A during amenorrhea (Fig. 1). Free androgen index was elevated in 83% of PCOD patients and 59% of non-PCOD patients, in whom mainly obese subjects had elevated free androgen indicies (Fig. 2). No significant difference was found in BMI between PCOD and non-PCOD women. Obesity was associated with high levels of free androgen index (P < 0.001, Fig. 3) and low levels of LH (P < 0.01), LH:FSH ratio (P < 0.001, Fig. 3), A (P < 0.05, Fig. 3), and SHBG (P < 0.001) in oligomenorrheic state. The levels of SHBG were significantly lower (29.7 ± 18.3 nmoljL versus 42.2 ± 21.1 nmoljL, P < 0.05) and the levels of free androgen index significantly higher (132.8 ± 73.9 versus 89.7 ± 45.0, P < 0.05) in hirsute than in nonhirsute PCOD women. No significant difference was found in BMI between the hirsute and nonhirsute PCOD groups.

DISCUSSION

Vol. 58, No.4, October 1992

We have previously shown that the timing of the hormone sampling is important when exploring the androgen and particularly the gonadotropin status in regularly menstruating women (14). During a normal menstrual cycle, T and SHBG values are greater in the luteal phase, but despite the cyclic variation, serum androgen concentrations and especially free androgen index are confined within narrow limits (14). However, the levels of the LH: FSH ratio as measured by novel fluoroimmunoassays vary more in the luteal phase (range 0.3 to 3.5; coefficient variation [CV] 60%) than in the early follicular phase (range 0.2 to 1.7; CV 30%) (14). Identifying excessive androgen production resulting in hyperandrogenism and altered gonadotropin secretion resulting in a high LH:FSH ratio has widely been used in the biochemical diagnosis of PCOD (5-7). To get adequate information of gonadotropin and androgen status is problematic in patients with occasional ovulations and/or periods, as often found in PCOD patients. It has been recommended that the hormone samples should be taken in amenorrheic and oligomenorrheic patients after a progestogen-induced bleeding (8). This view relies on the assumption that progestogen administration would mimic the luteal phase, rendering the results more comparable with the early follicular phase values. In this study, we found highly significant declines in the levels of LH and androgens after a short-term Anttila et al.

MPA effect on hormones in oligomenorrhea

699

6

....

5

%

4

2

II)

u.

:::

2

%

...J

~

2

• '" 0

0

M 0

M 0

le.n

ob •••

M 0

ob.,.

I•• n

M 0 M M 0 obi •• I••n

nonhirlute

"onhinut.

hinute peDD

25

non-PCOO

::: "0 E

20

.E ~ 0

~

15

e "«"

10

~"

. ~

5

II)

0

0

M 0 I.,n

ob... 0 0

500

~

,

M 0 M 0 M 0 ob... I.. n

M 0

ob.,.

nonhinut.

hinut.

M

le.n "oRhinut. non-PCOD

peDD

::: 400 "0

.§ Cl

300

'"

.

:I: II)

:::

200

"0

.§ ....

100

~

l;

'"

0

~

~~~

--. ~

ment suppresses T (9), LH (9), and SHBG (12) levels. Homburg and his co-workers (10) found an intramuscular (1M) short-term MPA (50 mgjd for 5 days) to markedly influence the gonadotropin release. Furthermore, 7 of 10 clomiphene citrate (CC)resistant PCOD women responded to CC with ovulation after such pretreatment (10). Petsos et al. (11) found MPA treatment (5 mg two times daily for 14 days) to suppress LH and FSH levels. In this work, even a short-term MPA treatment masked the biochemical findings of PCOD in several patients (Fig. 2). The rather long half-life ofMPA (24 to 50 hours) may explain the changes in hormonal levels even though the samples for endocrinological evaluation were collected approximately 6 to 10 days after the last MP A dose. Despite the mean levels of LH, the LH:FSH ratios' and all other androgen levels except DHEAS were significantly higher in the PCOD group, we found considerable heterogeneity of biochemical findings within PCOD patients (Fig. 2). This finding is in accordance with the current concept of the heterogeneous etiology of PCOD. Fox et al. (20) have recently suggested that a positive progestogen challenge test (5 mg oral MPA for 5 days) alone can predict PCOD in oligoamenorrheic subjects. Moreover, they found that the combination of LH and free androgen index was the best biochemical test to predict PCOD in oligoamenorrheic subjects. In this study, all subjects had a positive progestogen challenge test, indicating an E-primed endometrium. Within the non-PCOD group, 57% of

Ssil

peeD

non-PeDD

Free Androgen Index

Figure 1 Serum LH:FSH ratios, free androgen indices, and A concentrations in women with PCOD and non-PCOD women before (0) and after (M) MPA-induced withdrawal bleeding. Rectangles indicate the reference intervals.

Androstenedione

700

Anttila et a1.

MPA effect on hormones in oligomenorrhea

Free Androgen Index

Androstenedione

2

In oligomenorrhea

In oligomenorrhea

lHfFSH

lH/FSH

Free Androgen Index

oral treatment with MPA (Table 1, Fig. 1). The abnormal gonadotropin secretion in the PCOD patients is presumed to be reversible because of chronic deprivation of progesterone (P) (15). Progesterone acts at hypothalamic levels, decreasing LH pulse frequency (16). Similarly, vaginal P supplementation decreases LH pulse frequency in anovulatory PCOD women (17). Furthermore, in PCOD women ovulation is followed by a temporary normalization of the LH:FSH ratio (18, 19). Long-term MPA-treat-

LH/FSH

LH/FSH

M M 0 M 0 M 0 M 0 M 0 0 obese I•• n obese lean obese I••n nonhinut. "onhinut. hirsute

Androstenedione

Afler MPA-inducedWithdrawal bleeding

PCOD

Free Androgen Index

After MPA-InducedWithdrawal bleeding

non-PCOD

Figure 2 The distribution (%) ofPCOD and non-PCOD women according to elevated (inside the appropriate circle) or normal (outside the circle) levels of serum A (upper limit of the reference interval 11 nmol/L), free androgen index (T/SHBG X 1,000, upper limit ofthe reference interval 55) and LH:FSH ratio (upper limit of the reference intervall. 7). Percentages of subjects are shown in the states of oligomenorrhea and after MPA-induced withdrawal bleeding. The overlaping areas present those women with more than one pathological finding.

Fertility and Sterility

6

500 0

5

2: I

en

4

l<-

~3

0

•

00

o o~

I

• i II!::J

E

.s C)

m

en "2 200

0

•• 0 0 10

20

30

0

40

.,c

•

2 10

•

c9

I-

E

0

C

•

100

000~

E

o

2

CD

en

00c9

2

•

0 0

CO

5

o

0

•

• • --.. '""-

~c9c§J~0

«

E

0

°db~

!! "0

.s

0

cg• :.

c .2 15 "0

(;

ODD· 0l1li

., en

o·

.s .,

•

•

I

0 ...

2

300

o

••

E

0

0

•

~20

~400

s

0. 0

-' E 2

25

•

0

8

0

0

ClIO

0 00

o cJI 0

0

c7l 50

60

0 10

20

BMI (kg/m2)

30

40

50

60

o

10

20

60

BMI (kg/~)

Figure 3 Relationship between BMI and LH:FSH ratios, free androgen index (T/SHBG X 1,000), and A in hirsute (_) and nonhirsute women (D) with PCOD and non-PCOD oligomenorrheic women (0).

the women had elevated LH:FSH ratio during oligomenorrhea; thus the elevation of LH is not a pathognomonic feature of PCOD. Furthermore, free androgen index was elevated in 59% of non-PCOD patients and, in accordance with earlier reports (21, 22), elevated free androgen index was associated with obesity within non-PCOD patients (Figs. 1 and 3). The combination of elevated free androgen index together with high LH:FSH ratio was found in 24% of non-PCOD patients. Within PCOD women, elevated levels of free androgen index were found in both obese and lean subjects (together 83%). Highest free androgen indexes and lowest SHGB values were found in hirsute PCOD women. Among the PCOD women, 77% had a combination of more than one pathological finding. Interestingly, elevated A concentration, presumably reflecting the ovarian overproduction, was found only in the PCOD group (66%). We recently have demonstrated that obesity is related to low/normal bioactive LH concentrations within PCOD women (23). In this study, we measured the gonadotropins with highly sensitive fluoroimmunoassays, which have a good correlation with the bioassay (24). Our results concerning the association of obesity to lower LH concentrations within PCOD women is in agreement with that previously stated by Laatikainen and his co-workers (25), who found normal LH levels in massively obese PCOD patients. Furthermore, obesity has significant effect on free androgen indexes, LH:FSH ratios, and A concentrations. Thus, the common practice to compare the hormonal findings simply between lean (BMI < 25) and obese (BMI > 25) subjects may obscure the effects of varying degrees of obesity. The ultrasonographically diagnosed PCOD group showed considerable heterogeneity as concerning the Vol. 58, No.4, October 1992

biochemical findings. Which of the biochemical changes is the most efficient diagnostic analyte for the disease obviously depends on the pathogenetic mechanism that primarily prevails behind each individual case of PCOD. In conclusion, the pretreatment with MP A lowered significantly androgen and LH levels both in PCOD and non-PCOD patients. Thus the samples for hormonal analyses used as an aid in diagnosing PCOD should be obtained in native state because progestagen treatment masks biochemical findings of PCOD. Acknowledgments. We are indebted to Katariina Korkeila, M.D., and Jaana Lepisto, M.D., for their help in collecting the data.

REFERENCES 1. Hull MGR. Epidemiology of infertility and polycystic ovarian disease: endocrinological and demographic studies. Gynecol Endocrinol 1987;1:235-45. 2. Adams J, Polson DW, Franks S. Prevalence of polycystic ovaries in women with anovulation and hirsutism. Br Med J 1986;293:355-9. 3. Adams J, Polson DW, Abdul Wahid N, Morris DV, Franks S, Mason HD, et al. Multifollicular ovaries: clinical and endocrine features and responses to pulsatile gonadotropin releasing hormone. Lancet 1985;1:1375-9. 4. Ardaens Y, Robert Y, Lemaitre L, Fossati P, Dewailly D. Polycystic ovarian disease: contributions of vaginal endosonography and reassessment of ultrasonic diagnosis. Fertil SteriI1991;55:1062-8. 5. Yen SSC. The polycystic ovary syndrome. Clin Endocrinol (Oxf) 1980;12:177-82. 6. Rebar R, Judd HL, Yen SSC, Rakoff J, Vandenberg G, Naftolin F. Characterization of the inappropriate gonadotropin secretion in polycystic ovary syndrome. J Clin Invest 1979;57: 1320-9. 7. Lobo RA, Grager L, Gobelsmann U, Mishell DR Jr. Elevations of unbound serum estradiol as a possible mechanism

Anttila et al.

MPA effect on hormones in oligomenorrhea

701

for inappropriate gonadotropin secretion in women with PCOD. J Clin Endocrinol Metab 1981;52:156-62. 8. Heineman MJ, Thomas CMG, Doesburg WH, Rolland R. Hormonal characteristics of women with clinical features of the polycystic ovary syndrome. Eur J Obstet Gynecol Reprod Bioi 1984;17:263-71. 9. Wortsman J, Singh KB, Murphy J. Evidence for the hypothalamic origin of the polycystic ovary syndrome. Obstet Gynecol 1981;58:137-41. 10. Homburg R, Weissglas L, Goldman J. Improved treatment for anovulation in polycystic ovarian disease utilizing the effect of progesterone on the inappropriate gonadotropin release and clomiphene response. Hum Reprod 1988;3:285-8. 11. Petsos P, Ratcliffe WA, Anderson DC. Effects of medroxyprogesterone acetate in women with polycystic ovary syndrome. Clin Endocrinol (OxO 1986;25:651-60. 12. Wortsman J, Khan MS, Rosner W. Suppression of testosterone-estradiol binding globulin by medroxyprogesterone acetate in polycystic ovary syndrome. Obstet GynecoI1986;67: 705-9. 13. Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 1961;21:1440-5. 14. Anttila L, Koskinen P, Irjala K, Kaihola H-L. Reference intervals for serum sex steroids and gonadotropins in regularly menstruating women. Acta Obstet Gynecol Scand 1991;70: 475-81. 15. Berga SL, Yen SSC. Opioidergic regulation of LH pulsatility in women with polycystic ovary syndrome. Clin Endocrinol (OxO 1989;30:177-84. 16. Soules MR, Steiner RA, Clifton DK, Cohen NL, Aksel S, Bremner W J. Progesterone modulation of pulsatile luteinizing secretion in normal women. J Clin Endocrinol Metab 1984;58: 378-83. 17. Bruckler HM, Phillips SE, Cameron IT, Healy DL. Vaginal progesterone administration before ovulation induction with exogenous gonadotropins in polycystic ovarian syndrome. J Clin Endocrinol Metab 1988;67:300-6.

702

Anttila et al.

MPA effect on Iwrmones in oligomenorrhea

18. Eden JA, Place J, Carter GD, Alaghand-Zadeh J, Pawson WE. The role of chronic anovulation in the polycystic ovary syndrome: normalization of sex-hormone-binding globulin levels after clomifen-induced ovulation. Clin Endocrinol (OxO 1989;30:323-32. 19. Blankstein J, Rabinovici J, Goldenberg M, Shaley J, Mehta A, Serr DM, et al. Changing pituitary reactivity to folliclestimulating hormone and luteinizing hormone-releasing hormone after inducted ovulatory cycles and after anovulation in patients with polycystic ovarian disease. J Clin Endocrinol Metab 1987;65:1164-7. 20. Fox R, Corrigan E, Thomas PA, Hull MGR. The diagnosis of polycystic ovaries in women with oligo-amenorrhea: predictive power of endocrine tests. Clin Endocrinol (OxO 1991;34:127-31. 21. Hosseinian AH, Kim MH, Rosenfield RL. Obesity and oligomenorrhea are associated with hyperandrogenism independent of hirsutism. J Clin Endocrinol Metab 1976;42: 765-9. 22. Grenman S, Riinnemaa T, Irjala K, Kaihola H-L, Griinroos M. Sex steroid, gonadotropin, cortisol and prolactin levels in healthy, massively obese women: correlation with abdominal fat cell size and effect of weight reduction. J Clin Endocrinol Metab 1986;63:1257-61. 23. Anttila L, Ding Y -Q, Ruutiainen K, Erkkola R, Irjala K, Huhtaniemi 1. Clinical features and circulating gonadotropin, insulin, and androgen interactions in women with polycystic ovarian disease. Fertil Steril 1991;55:1057-61. 24. Jaakkola T, Ding Q-Y, Kellokumpu-Lehtinen P, Valavaara R, Martikainen H, Tapanainen J, et al. The ratios of serum bioactive/immunoactive luteinizing hormone and folliclestimulating hormone in various clinical conditions with increased and decreased gonadotropin secretion: reevaluation of highly sensitive immunometric assay. J Clin Endocrinol Metab 1990;70:1496-1505. 25. Laatikainen T, Tulenheimo A, Andersson B, Kiirkkiiinen J. Obesity, serum steroid levels and pulsatile gonadotropin secretion in polycystic ovarian disease. Eur J Obstet Gynecol Reprod Bioi 1983;15:45-53.

Fertility and Sterility