Changes in androgens during treatment with four low-dose contractptives

ELSEVIER

Changes in Androgens During Treatment with Four Low-Dose Contraceptives C.M.H. Coenen,*

C.M.G. Thomas,*t

G.F. Borm,$ J.M.G. Hollanders,*

The aim of the present study was to compare changes in the endogenous androgen environment in healthy women while on low-dose oral contraceptives (OCs). One-hundred healthy women were randomized to receive one of four OCs during six months: 21 tablets of Cilest@, Femodeen@, Marvelon@, or Mercilon@. During the luteal phase of the pretreatment cycle, body weight and blood pressure were recorded and the following parameters were measured: sex hormone-binding globulin (SHBG), corticosteroid-binding globulin (CBG), testosterone (T), free testosterone (FT), 5~ dihydrotestosterone (DHT), androstenedione (A), dehydroepiandrosterone-sulphate (DHEA-S) and 17cY-hydroxyprogesterone (17OHP) while also the free androgen index (FAI) was calculated. Measurements were repeated during the 3rd week of pill intake in the 4th and the 6th pill month. There were no differences on body mass and blood pressure with the use of the four OCs. The mean serum DHEA-S decreased sigplificantly in all groups though less in the Mercilon@ group when compared to CilesP and Marvelon@ (approximately 20% vs 45%). Mean serum SHBG and CBG increased significantly in all four groups approximately 250% and lOO%, respectively. In each group CBG also increased significantly but less in women taking Mercilon@ (-75%) as compared to the others (-100%). Current low-dose OCs were found to have similar impact on the endogenous androgen metabolism with significant decreases of serum testosterone, DHT, A, and DHEA-S. They may be equally beneficial in women with androgen related syndromes such as acne and hirsutism. CONTRACEPTION

1996;53: 171-176 KEY

WORDS:

androgens, oral contraceptives

‘Department of Obstetrics and Gynecology, tEndocrinology and Reproduction Laboratory, *Department of Blostatistics, University Hospital Nijmegen St. Radboud, Nijmegen, the Netherlands, and §IVF/Reproductive Endocrinology, King Fahad National Guard Hospitul, P.O. Box 22490, Riyadh 11426, Saudi Arabia Name and address for corsespondence: C.M H. Coenen, M.D., Department of Obstetrics and Gynecolog University Hospital Nijmegen St. Radboud. P.O. Box 9101 6500 HB Nijmege the Netherlands. Tel: 31-24-3614748; Fax: 3124-3541 194 Submitted for publication \ gust 22, 1995 Revised December 12,199 Accepted for publication D cember 18, 1995

0 1996 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, ,New York, NY 10010

and R. Rolland*§

Introduction ral contraceptives (OCs) have been described as a risk factor for the development of cardiovascular disease (CVD). At first this phenomenon was attributed to the estrogenic component of the OC.‘j’ Parallel with dose reduction of both the estrogenic and the progestogenic component and the introduction of new progestins, the incidence of CVD decreased significantly. At present there is doubt whether or not modern OCs increase the risk for cardiovascular disease.3t4 The mechanism of a possible influence on CVD has also been attributed to the intrinsic androgenic potential of progestins used in OCs. This can at least in part be theoretically attributed to the influences on lipid metabolism.5 The androgen environment in individual women taking oral contraceptives depends on the eventual androgenic effect of the utilized progestin, the antigonadotropic effect of the estrogen-progestin combination altering ovarian androgen production and the effect of the OC on plasma proteins with sex steroidbinding properties. The androgenic effect of progestins may be exerted in two different ways: first, through stimulating the binding of androgens to the androgenic receptors of the target organ and, second, to their affinity to sex hormone-binding globulin (SHBG) and other androgen binding proteins, thus displacing testosterone (T) from SHBG binding and consequently increasing the proportion of free testosterone (FT).6 Therefore, the SHBG level is a useful parameter in the evaluation of the androgenicity of oral contraceptives. Literature data examining the new progestins norgestimate, desogestrel and gestodene concentrate on their possible influence on lipid metabolism and clotting. Surprisingly, few reports systematically assess the changes in sex steroids exerting androgenic properties during treatment with OC. The aim of the present study was to compare the effects of four modern low-dose OCs on sex steroids with androgenic properties and their binding proteins

0

ISSN OOlO-7824/96/$15.00 PII SOOlO-7824(96)00006-S

172

Coenen

et al.

Contraception 1996:53:171-176

in serum. Therefore, the present study specifically evaluated changes brought about in plasma concentrations of SHBG, CBG, albumin (Alb), T, FT, dihydrotestosterone (DHT), androstenedione (A), dehydroepiandrosterone-sulphate (DHEA-S) and 1701hydroxyprogesterone (I 70HP).

Material

and Method

Study Design In an open randomised study, 100 female volunteers were equally assigned to one of four different OC preparations during six cycles on therapy (Table 1). The healthy volunteers ranging in age between 18 and 38 years, were not pregnant, not lactating, and had delivered at least six months before inclusion into the study. Their menstrual cycles had to be within a duration between 2.5 and 36 days. Obese women were excluded [Body Mass Index (BMI)= weight in kg/height in m2 > 28.01.’ The use of steroid-containing medication or medication possibly interfering with steroid metabolism was discontinued for at least two months prior to treatment and the use of medication known to alter liver enzyme function was not allowed. Also, the well known relative contraindications to the use of OCs such as a history of thromboembolic disorder, hyperlipoproteinemia, known or suspected estrogen-dependent neoplasia, liver tumors and undiagnosed abnormal genital bleeding were reasons for exclusion. Finally, the volunteers were not allowed to smoke more than 10 cigarettes a day. Each subject was seen before therapy during a control cycle between days 24-27 (luteal phase) and while on therapy during pill cycle 4 and 6 (between days 18-21 of pill intake). At every visit, body weight and blood pressure were recorded and blood collected for measurements of SHBG, CBG, Alb, T, FT, DHT, A, DHEA-S and 170HP. Also, the free androgen index (FAI=Tx 100/SHBG)8 was calculated. All blood samples were drawn between 08.00 and 10.30 am after a fasting period of at least 12 h. The blood was allowed to clot at room temperature, serum was col1. Composition tives tested

Table

PreDaration

Cilest@ Femodeenm Marvelon@ MerciIon@

of the four monophasic contracep-

Ethinylestradiol Der Tablet 35 1.18 30 !-% 30 I%

20 1.18

Progestagen Der Tablet

lected after centrifugation at 2000xg and kept frozen at -20°C until assayed. The volunteers were instructed to take the first tablet on the first or second day of the menstruation following the control cycle visit. Then one tablet was taken daily at a fixed time for 21 days followed by 7 tablet-free days. The study was approved by the ethical committee of the hospital beforehand and was conducted according to the guidelines for Good Clinical Practice.’ Written informed consent was obtained from all subjects at the beginning of the study. Assays SHBG was measured by a commercially available non-competitive coated-tube immunoradiometric assay (IRMA, Farmos Diagnostica, Orion Corporation, Turku, Finland) based on the method of Hammond et al. lo The within-assay and between-assay coefficients of variation (CV, and CV,) ranged from 2.6% to 2.9% and from 3.5% to 4.6%, respectively. Serum concentrations of CBG were measured with a commercially available direct competitive liquid-phase RIA (Techland SA, Liege, Belgium); the CV, ranged from 3.2% to 4.7% and the CV, from 8.6% to 8.8%. Serum T was measured after diethylether extraction by a specific charcoal RIA as described by Dony et al.“; the CV, and the CVb were 5.6% and 6.9%, respectively. Serum FT concentrations were assayed with a commercially available direct competitive solid-phase RIA (Diagnostic Products Corporation, Los Angeles, CA, USA); CV, and CV, were 6.4% and 12.1%, respectively. DHT was measured with a specific in-house charcoal RIA. Purification of serum specimens by diethylether extraction and column chromatography with Sephadex LH20 were as described by Van Duren12; CV, and CV, calculated from a serum pool with a mean DHT level of 1.8 nmol/L were 6.5% and 11.7%, respectively. Androstenedione was measured after diethylether extraction and Sephadex LH20 chromatography by a charcoal RIA as described by Van Duren12; CV, was 3.6% and CVI, was 7.0%. DHEA-S concentrations were measured with a highly specific charcoal RIA as described by Reijnders13; CV, and CV, were 3.7% and 6.9%, respectively. Concentrations of 170HP in serum were determined after diethylether extraction followed by Sephadex LH20 chromatography with a specific charcoal RIA as described by Dony et al.“; CV, and CV, were 3.7% and 6.9%, respectively.

250 pg norgestimate

75 pg gestodene 150 pg desogestrel 150 pg desogestrel

Statistics All values are given as mean * sd. P-values less than 0.05 were considered significant. In order to remove

Androgenicity

Contraception 1996;53:171-176

of Four Low-Dose

Contraceptives

173

Table 3. Drop-outs according to pill cycle, subject number, pill type and complaint(s)

heteroscedasticity and skewness, all parameters were log-transformed. In the primary analysis, changes from baseline in the four groups were compared using the Welch multiple range test. l4 In order to control1 the overall error rate per parameter, t&s was done in a stepwise manner: first the changes over 6 months were evaluated, and only if these were (nominally) significant, the changes over the first 4 months were also evaluated (close testing procedure15). This method corresponds to the intuitive idea Ehat one first investigates if there are any long-term differences after 6 months and then, only if this is the case, one investigates whether the differences already occur after 4 months. In a secondary analysis, the changes from baseline within the groups were evaluated. In order to control the overall error rate,, a pooled t-test was carried out: the changes were evaluated for the four groups simultaneously in an AN0VA model with no intercept. Again changes over 4 months were tested, only if those over 6 months were found significant.

Pill Subject Number Type*

Pill Cycle

15

Ci

1

85

Fe

2t 2t

41 43

Me Me

2t

68

Ci

3

76

Fe

3t 4 4 4t 4t

81 3

Ci Fe Ci Ma Me Me

16

Unknown ‘Ci-CilesP; tDrop-out

Results

after

58 6 98 Fe-FemodeerP; pill day 21.

Complaint(s) Headache of unknown origin Spotting/breakthrough bleeding MastopatGy Tachycardia/hyperventilation Body weight and mood change Headache during pill-free week Pruritis vaginalis Sinusitis: antibiotics Hepatitis A Protocol violation Increase in migraine Poor compliance

Ma-Maw&n”;

Me-Mercilon@.

the pill cycles 4 and 6 are shown in Table 4. None of these parameters significantly changed throughout the study period.

Population Characteristics The characteristics of the study groups are given in Table 2.

Hormones and Binding Proteins All the steroidal serum parameters tested (T, FT, DHT, A, DHEA-S, 170HP, Alb) significantly decreased while on medication during 6 pill cycles (ratio of decrease between 1.3 and 3.0), irrespective of the OC preparation used (Tables 5 and 6). All changes were observed already after 4 pill cycles. Comparing the decreases of the above mentioned variables after 6 months between the 4 pill groups, the only significant difference observed concerned DHEA-S: the mean decrease of DHEA-S was 21+18% in the Me-group, significantly different compared to 43+ 18 % and 44* 18 %, respectively, in the Ci- and Magroup, while the Fe-group was in between with a decrease of 34+14%. Significant increases were observed in the levels of steroid binding proteins. Both SHBG and CBG increased significantly during pill-intake in all four

Dropouts Two women became pregnant during the control cycle and, therefore, dropped out prior to treatment. They were replaced according to the protocol. There were twelve true dropouts, the reasons for which are summarized in Table 3. Subject number 8 1 (Ci-group) was not available during pill cycle 4, which is the reason why the second visit was planned during pill cycle 3. Subject number 85 (Fe-group) was not available during the third, fourth and fifth pill cycle; she was unable to attend the second visit. Hence, a total of 88 volunteers completed the study according to the protocol. Vital Signs Changes in body weight and blood pressure (systolic and diastolic] registared in the control cycle and in Table 2. Sample characteristics (mean * sd) Oral Contraceptive

I

Cilest@ Femodeenm Marvelon@ Mercilona Overall

26.9 26.3 27.1 27.3 26.9 I

‘Body

mass index.

Age (Years) (4.2) (4.9) (5.1) (5.3) (4.8)

Cycle Length (Days) 28.3 28.8 28.6 29.0 28.7

(1.5) (2.1) (2.1) (2.2) (2.0)

Menses Pays) 5.0 4.7 5.0 5.4 5.0

(0.8) (1.2) (1.0) (1.1) (1.1)

BMI’ (kg/m*1 22.2 22.0 22.0 22.4 22.1

(2.3) (2.4) (2.4) (2.2) (2.3)

174

Coenen

et al.

Contraception 1996;53:171-176

4. Changesin body massand systolic and diastolic blood pressure of the four preparations tested after six pill cvcles (mean * sd)

Table

CilesP FemodeexP Marvelon@ Mercilona

Change in Body Mass WI

Change in Systolic Blood Pressure (%I

Change in Diastolic Blood Pressure (%I

+0.5 (3.9) +1.8 (3.9) +l.O (2.8) +l.l (2.81

+1.6 (9.8) -1.4 (6.6) +O.O(7.6) -0.5 17.7)

+3.9 +0.5 -2.4 -1.0

intrinsic androgenic properties. Many studies have deed indicated that they have no negative effect lipid metabolism. In this study the suppression plasma hormone concentrations in individuals four different OCs has been evaluated.

All androgen parameters investigated in the present study decreased while taking the new OCs regardless of the type. At the same time, the androgen binding proteins increased. The net effect of this was an even more substantial decrease of biologically active testosterone. The concentration of the most active androgen in the human, 5ol-dihydrotestosterone, is directly related to that of free testosterone. Therefore, during intake of these OCs the endogenous androgen environment changed in the direction of hypoandrogenism. It is speculated that even if the applied progestins exert an intrinsic androgen action, this is negligible as concerned to the changes in the endogenous environment. It seems safe to postulate that all these

groups. At six months on therapy, the mean increase of SHBG was 263% (*119%) and 94%(*26%) in the case of CBG. Again, there were no differences between the different pill groups except for CBG: this protein shows a significantly less increase in women taking Mer-

four OCs are equally beneficial with regard to sebor-

cilon@ (74+21%) as compared to the three other contraceptives, Cilest @ 96*31%, Femodeen@ 101+21%

and Marvelon@

inon of on

rhoea, acne and hirsutism, but the clinical effects of the four preparations regarding to the skin have not been investigated in our study. They should also give rise to a favourable change in lipid metabolism and an increase in the high density/low density lipids (HDL/ LDL) ratio has indeed been reported for all of them.‘9,20 Norgestimate, desogestrel and gestodene all bind to progesterone and androgen receptors. The binding to the androgen receptor is significantly reduced, however, as compared to the older progestins. In view of our findings and the reports on HDL/LDL, therefore, a further attempt to reduce the progestin content of these OCs does not seem necessary. It may even be unwarranted, as a less effective cycle control may be the only achieved result. 21 All OCs tested in the present study decreased the concentration of DHEA-S

102*22 % .

Discussion Progestins, especially those with intrinsic androgenic activity, have an opposite effect on lipoprotein metabolism as compared to estrogens. They especially stimulate hepatic lipase, thereby accelerating metabolism and clearance of high density lipids (HDL) and lowering HDL levels. Low HDL levels are thought to increase the risk for CVD.i6 The binding of a given progestin to SHBG is an important determinant of its androgenicity. The progestins used in this study may, in this respect, however, differ from others. “t18 These new progestins have been developed because of their small or absent

5. Serum concentrations of albumin @lb), sex hormone-binding globulin (SHBG), and corticosteroid-binding globulin (CBG): values are given as mean * sd Table

Serum Variable Alb (g/L1

CBG b&-l

different

Pill Cycle 4 (n = 90)

Pill Cycle 6 (n = 88)

47.40 (2.63) 47.76 (2.79) 46.80 (2.53) 47.60 (2.45) 55 (20) 51(17)

42.73 (2.73) 43.95 (2.50) 42.56 (1.76) 43.00(2.56) 173 (48) 163 (49)

43.10 (1.97) 43.41 (2.91) 42.50 (2.02) 43.43 (2.60) 174(64) 162 (47)

61 (31)

201 (68)

217 (88)

Ci Fe Ma Me Ci Fe Ma Me Ci Fe Ma Me

SHBG (nmol/L)

‘Significantly

Control Cycle (n = 100)

from

MercilmP

with

48 (20) 44.0 (5.6) 43.8 (4.5)

157 85.5 86.2 82.7 76.7

41.2 (5.5) 43.6 respect

to change

from

baseline

(3.0) [closed

testing

procedure).15

(52) (7.9)' (8.1)* (6.4)' (8.1)

156(49) 84.5 (7.9)* 88.3 (7.9)* 82.4 (9.6)* 75.7 (8.6)

Androgenicity

Contraception 1996;53:171-176

of Four Low-Dose

Contraceptives

Table 6. Serum concentrations of testosterone (T), free testosterone (FT), free androgen index (FAI), So-dihydrotestosterone (DHT), androstenedione (A), dehydroepiandrosterone-sulphate (DHEA-S), 17a-hydroxyprogesterone (170HP); values

175

are

given as mean f sd

Control Cycle (n = 100)

Serum Variable Ci Fe Ma Me Ci Fe Ma Me Ci Fe Ma Me Ci Fe Ma Me Ci Fe Ma Me

T (nmol/L)

FT (pmol/L)

FAI

DHT (nmoI/L)

A (nmol/L)

DHEA-S (nmol/L)

Ci Fe Ma Me

170HP

Ci Fe Ma Me

*Significantly

(nmol/L)

different

from

MercilmP

with

2.26 2.43 2.48 2.50 5.79 6.57 7.08 7.12 4.74 5.16 4.73 6.19 0.89 0.91 1.00 0.86 5.97 7.18 7.72 7.38 7.88 7.98 8.42 7.23 4.81 5.71 5.60 6.23 respect

to change

from

Pill Cycle 4 (n = 90)

(0.85) (0.53) (0.85) (1.01) (2.57) (2.11) (2.46) (2.92) (2.41) (1.67) (2.04) (3.55) (0.36) (0.30) (0.39) (0.44) (2.38) (2.61) (3.10) (3.59) (2.90) (2.35) (2.57) 12.49) (2.89) (2.39) (2.99) (2.80)

baseline

significantly. This decrease is probably brought about by a suppression of the adrenocorticotropic hormone (ACTH).22 This effect is not only effected by dose and type of progestin but also by the ethinylestradiol (EE) content: Ma and Me differ only by their EE content and the effect of Me on DHEA-S was significantly less as compared to Ma. The lower EE content of Me explains, in our view, the significantly lower increase in CBG. From a clinical point of view, the difference in Me on DHEA-S as compared to the others is possibly of little significance: :DHEA-S is a weak androgen and the change seems minor when compared to the major decrease of free testosterone (and hence Sor-dihydrotestosterone), although this hypothesis needs further study. A profound decrea e in both total and free testosterone, and the abs rice of any negative effects on clinical variables su h as blood pressure and body mass are similar foi all four studied preparations. These OCs may therefore be especially indicated for women with some s/igns of hyperandrogenism. The strong androgen-lowering effect may also be disad-

1.81 1.94 2.05 1.91 2.59 2.70 3.02 2.95 1.14 1.30 1.09 1.36 0.61 0.61 0.73 0.64 3.99 4.15 4.63 4.29 4.78 5.17 5.09 5.10 1.29 1.52 1.55 1.59

[closed

testing

(0.61) (0.44) (0.66) (0.58) (0.80) (1.01) (1.63)

(1.11) (0.53) (0.50) (0.39) (0.68) (0.32) (0.22) (0.29) (0.26) (1.58) (1.94) (1.92) (2.04) (2.48) (2.14) (2.32) (2.20) (0.87) (1.41) (0.98) (1.30)

Pill Cycle 6 (n = 88) 1.70 (0.49) 1.88(0.41) 1.98 (0.63) 1.96 (0.64) 2.61 (1.17) 2.66 (0.76) 2.88 (1.07) 2.89 (1.09) 1.28 (1.30) 1.27(0.52) 0.98 (0.36) 1.36 (0.58) 0.55 (0.22) 0.61 (0.19) 0.73 (0.28) 0.66 (0.29) 3.92 (1.24) 3.55 (1.36) 4.64(1.78) 4.17 (2.14) 4.71 (2.65)' 5.42 (1.87)' 4.90 (2.22)' 5.8212.53) 1.15 (0.66) 1.49 (1.02) 1.51 (0.82) 1.36(1.03)

procedure).‘”

vantageous. Although thus far no data are reported in the literature, low androgen levels in the female may negatively interfere with libido and self-esteem.= We therefore recommend that future research on new OC formulations should build into their protocols means to evaluate the tendency to hyperandrogenism and eventual effects on libido and self-esteem.

Acknowledgments The authors thank the staff of the Endocrinology and Reproduction Laboratory for expert technical assistance.

References 1. Inman WI-N, Vessey MP, Westerhom B, et al. Thromboembolic disease and the steroidal content of oral contraceptives. A report to the Committee on Safety and Drugs. Br Med J 1970;2:203-9. 2. Meade TW, Greenberg G, Thompson SG. Progestens and cardiovascular reactions associated with oral contraceptives and a comparison of the safety of 50- and

176 Coenen et al.

30-pg oestrogen preparations. Br Med J 1980;280:115761. 3. Thoregood M, Vessey MP. An epidemiologic survey of cardiovascular disease in women taken oral contraceptives. Am J Obstet Gynecol 1990;163:274-81. 4. Lidegaard 0. Oral contraception and risk of cerebral thromboembolic attack: Results of case-control study. Br Med J 1993;306:956-63. 5. Ellis J. Low-dose oral contraceptives: Progestins potency, androgenicity and atherogenic potential. Clin Therapeutics 1986;8:607-18. 6. Philips A, Hahn DW, Mcguire JL. Preclinical evaluation of norgestimate, a progestin with minimal androgenic activity. Am J Obstet Gynecol 1992;167:1191-6. 7. Valkenburg HA, Hofman A, Klein F, Groustra FN. Een epidemiologisch onderzoek naar risico-indicatoren voor hart-en vaatziekten (EPOZ). Bloeddruk, serumcholesterolgehalte, Quetelet-index en rookgewoonten in een open bevolking van vijf jaar en ouder. Ned T Geneesk 1980;124:183-9. 8. Thomas CMG, Berg van den RJ, Rolland R. Measurement of serum testosterone: Results of the “Farmos” SHBG IRMA kit and the “Coat-A-Count” free testosterone kit compared with salivary testosterone. Clin Chem 1986;32;702. 9. EEC Note for Guidance: Good Clinical Practice for trials on medicinal products in the European community, CPMP Working Party on Efficacy of Medical Product. Pharmacol Toxic01 1990;67:361-72. 10. Hammond GL, Langley M, Robinson PA. A liquid phaseimmuno radiometric assay(IRMA) for human sex hormone-binding globulin. J Steroid Biochem 1985j86: 405-14. 11. Dony JMJ, Smals AGH, Rolland R, Fauser BJCM, Thomas CMG. Effect of aromatase inhibition by 6-ltestolactone on basal and luteinizing hormonereleasing hormone-stimulated pituitary and gonadal

Contraception 1996:53:171-176

12. 13. 14. 15.

hormonal function in oligospermic men. Fertil Steril 1985;43:787-92. Van Duren BPJ. The infertile couple: some cycle and sperm characteristics. Academic Thesis, Catholic University of Nijmegen, 1988. Reijnders FJL. The influence of 17c+hydroxyprogesterone carproate on early pregnancy. Academic Thesis, Catholic University of Nijmegen, 1988. Welch RE. Stepwise multiple comparison procedures. J Am Stat ASSOC1977;72:566-75. Bauer P. Multiple testing in clinical trials. Stat Med 1991;10:871-90.

16. Krauss RM, Burkman RT. The metabolic impact of oral contraceptives. Am J Obstet Gynecol 1992;167:117784.

17. Kloosterboer HJ, Vonk-Noordegraaf CA, Turpijn EW. Selectivity of progesteroneand androgenreceptor binding of progestogensusedin oral contraceptives. Contraception 1988;38:32532. 18. Runnebaum B, Rabe T. New progestogensin oral contraceptives. Am J Obstet Gynecol 1987;157:1059-63. 19. Kafrissen ME. A norgestimate-containing oral contraceptive: review of clinical studies. Am J Obstet Gynecol 1992;167:1196-202. 20. Krauss RM, Burkman RT Jr. The metabolic impact of oral contraceptives. Am J Obstet Gynec 1992;167:117784.

21. Collins DC. Sex hormone receptor binding, progestin selectivity, and the new oral contraceptives. Am J Obstet Gynecol 1994;170:1508-13. 22. Murphy AA, Cropp CS, Smith BS, Burkman RT, Zacur HA. Effect of low doseoral contraceptive on gonadotropins, androgens, and sex hormone-binding globulin in nonhirsute women. Fertil Steril 1990;53;35-9. 23. Adams DB, Gold AR. Rise in female-initiated sexual activity at ovulation and its suppressionby oral contraceptives. New Engl J Med 19783229:1145-50.