Effects of the insulin sensitizing drug metformin on ovarian function, follicular growth and ovulation rate in obese women with oligomenorrhoea

I.R. Pirwany1, R.W.S. Yates, I.T. Cameron and R. Fleming

University Department of Obstetrics and Gynaecology, Glasgow Royal Infirmary, 10 Alexandra Parade, Glasgow G31 2ER, UK


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Hyperinsulinaemic insulin resistance is commonly associated with hyperandrogenaemia, and menstrual dysfunction. The aim of this study was to examine the effects of the insulin sensitizing drug, metformin, on ovarian function, follicular growth, and ovulation rate in obese women with oligomenorrhoea. Twenty obese subjects with oligomenorrhoea [polycystic ovarian syndrome; (PCOS)] were observed longitudinally for 3 weeks prior to and for 8 weeks during treatment with metformin (850 mg twice per day). Fifteen patients completed the study. The frequency of ovulation was significantly higher during treatment than before treatment (P = 0.003). A significant decline in both testosterone and luteinizing hormone concentrations was recorded within 1 week of commencing treatment. Patients with elevated pretreatment testosterone concentrations showed the most marked increase in ovulation rate (P < 0.005), and significant reductions in circulating testosterone from 1.02 to 0.54 ng/ml (P < 0.005) after only 1 week of treatment. However, the sub-group with raised fasting insulin showed less marked changes, and the sub-group with normal testosterone concentrations showed no effect of treatment. Metformin had a rapid effect upon the abnormal ovarian function in hyperandrogenic women with PCOS, correcting the disordered ovarian steroid metabolism and ovulation rate; however, there appeared to be no effect in cases where the circulating androgen concentration was normal.

Key words: hyperandrogenism/metformin/ovulation/polycystic ovarian syndrome


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Polycystic ovarian syndrome (PCOS) affects between 6 and 10% of women of reproductive age (Franks, 1995Go). Hyperinsulinaemic insulin resistance is a common finding in lean and obese women with PCOS (Chang et al., 1983Go; Dunaif et al., 1989Go). Insulin appears to promote androgen biosynthesis, both in vitro (Barbieri et al., 1986Go; Cara and Rosenfield, 1988Go; Bergh et al., 1993Go; Nahum et al., 1995Go) and in vivo (Poretsky and Kalijn, 1987; Barbieri et al., 1988Go). It may act by augmenting luteinizing hormone (LH) activity through stimulation of ovarian receptors of insulin and insulin-like growth factors (Bergh et al., 1993Go; Nahum et al., 1995Go; Willis and Franks, 1995Go).

Weight loss in women with PCOS corrects the hyperinsulinaemia (Kiddy et al., 1992Go), while the pharmacological approach using diazoxide (Nestler et al., 1989Go), or troglitazone (Dunaif et al., 1996Go; Ehrmann et al., 1997aGo) also appears to improve insulin resistance and may normalize ovarian steroidogenesis in some cases.

The use of metformin, another insulin sensitizing agent, in PCOS has provided conflicting evidence of its beneficial effect upon insulin resistance (Acbay et al., 1996; Ehrmann et al., 1997bGo). Studies in PCOS that have shown improvement in insulin sensitivity during metformin treatment have shown an apparent beneficial effect on ovarian function assessed by menstrual frequency (Nestler and Jakubowicz, 1996Go; Velazquez et al., 1997aGo). Although its precise mode of action is unclear, it has been suggested that the decreased peripheral insulin resistance may be related to the weight loss that often accompanies its use (Crave et al., 1995Go; Morin-Papunen et al., 1998Go). Weight loss alone can result in improved insulin sensitivity, and ovarian function (Kiddy et al., 1992Go). However, a population of obese, hyperandrogenic women with oligomenorrhoea who were randomized to receive metformin or placebo before clomiphene induction of ovulation showed suppression of the insulin response to a glucose challenge after only a month of metformin treatment (Nestler et al., 1998Go). Infrequent blood sampling suggested that there was an increase in the spontaneous ovulation rate in the metformin treated group, although the sampling regime used in this study did not allow definitive assessment of ovarian function. However, the results did suggest that weight loss might not be an important component of the responses. The above was one of few studies that have directly addressed the issue of ovarian function during metformin treatment in obese women with PCOS, and no study has examined follicular growth and ovulation in a longitudinal investigation during treatment with metformin. Cross-sectional studies have shown changes in sex hormone binding globulin (SHBG) and ovarian steroids (Velazquez et al., 1994Go, 1997bGo) that may influence ovarian function indirectly.

In this longitudinal study, details of the short-term responses and longer term effects of metformin on reproductive hormone changes, follicular development and ovulation in obese women with PCOS are reported. Since serum androgen concentrations may be normal in 30–50% of women with oligomenorrhoea and polycystic ovaries (De Vane et al., 1975; Robinson et al., 1992Go), it is important to determine whether patients with normal androgen profiles may also benefit from this approach.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Patients
Twenty obese women [body mass index; (BMI) >29 kg/m2] with oligomenorrhoea, normal circulating follicle stimulating hormone (FSH) concentrations and polycystic ovaries observed by ovarian ultrasonography were recruited from the infertility or the gynaecological endocrine clinics of Glasgow Royal Infirmary. No patient had taken any medication in the 2 months preceding the study. Oligomenorrhoea was diagnosed as fewer than six menstrual periods a year or inter-menstrual interval >41 days. Polycystic ovaries were diagnosed according to previously established criteria (Adams et al., 1986Go). The patients were defined as having PCOS on these criteria, irrespective of clinical or biochemical evidence of hyperandrogenism. Non-insulin-dependent diabetes mellitus and adult onset congenital adrenal hyperplasia were excluded by fasting glucose and 17{alpha}-hydroxyprogesterone estimations respectively. The study was approved by the local hospital ethics committee, and all patients gave written informed consent to participate in the study.

Study protocol
At the baseline visit, which took place at least 2 weeks following a menstrual period and outside the luteal phase (confirmed by plasma progesterone assays), all patients underwent a transvaginal ultrasound examination of the ovaries, using a 7.5 MHz vaginal transducer (Siemens Sonoline® SI-400 scanner). Anthropometric measurements were made prior to treatment and at the end of treatment by the same observer using standard World Health Organization (WHO) techniques. Body weight was measured to within 100 g in light clothes using digital scales (Seca®, Germany); height was measured barefoot to within 0.5 cm. Circumferences were measured to within 1 mm using flexible tape in the standing position. Waist circumference was measured mid-way between the lowest rib margin and the iliac crest at the end of gentle expiration, and hip circumference at the widest level of the greater trochanters.

Analyses
Concentrations of glucose and insulin were measured in plasma and serum samples respectively obtained after an overnight fast and hormone concentrations were estimated by measurement in duplicate on patient specific batches in frozen (–20°C) stored samples. Fasting insulin (FI) concentrations were recorded only once during the treatment part of the study (at 6 weeks).

Patients were instructed to maintain their usual diet and activities throughout the study period, and ß-human chorionic gonadotrophin (ß-HCG) tests were performed at regular intervals (14 days) during treatment.

Serum progesterone, testosterone and insulin concentrations were measured by specific radioimmunoassay (Coat-A-Count®, DPC, Los Angeles, CA, USA); serum oestradiol, LH, FSH and sex hormone binding globulin (SHBG) concentrations were determined by fluoroimmunoassays (Delphia® Ltd, Wallac, UK). The testosterone assay showed low cross-reactions with other androgen analytes, except for 5{alpha}-dihydrotestosterone (9.1%).

Monitoring of ovarian activity and treatment
Following the baseline visit, blood samples were obtained twice weekly to monitor follicle growth and to check for ovulation for a minimum of 3 weeks prior to starting treatment. Sampling began more than 2 weeks after a menstrual period and not during a luteal phase, so that the window of observation would be able to detect spontaneous ovulations taking place after a follicular phase of a minimum duration of 5 weeks, depending on when the sampling was started relative to the previous menstrual period.

Metformin treatment was started in the first 5 days following a menstrual period subsequent to the observation period, or where this was not possible, it was commenced after confirmation that the patient was not in the luteal phase of cycle. Treatment was commenced at a dose of 850 mg once a day for 1 week, and increased to 850 mg twice daily for a further 7 weeks. This approach to treatment was recommended because of the gastric irritation that can occur at the start of treatment. During the treatment period, patients provided twice weekly blood samples for monitoring ovarian function. Ovulation was considered to have occurred when the serum progesterone was >3.5 ng/ml following a surge of LH, indicating a luteal phase. Ovulation rates were calculated as: (luteal weeks/observation weeks)x100. Thus, in a woman with normal menstrual rhythm having a cycle of 2 follicular and 2 luteal weeks, the maximum ovulation rate possible was 50%. Other hormone variables were examined pretreatment and at weeks 1, 2, 4, 6 and 8 weeks (post-treatment). The last blood sample (week 8) was taken immediately (1–3 days) after cessation of treatment to determine if the effects of treatment would be maintained for any length of time.

Statistics
Distributions of the data were examined for normality by the Shapiro–Wilks test. Results are reported as mean ± SD where normal distributions were confirmed and geometric mean with ranges when log-transformation was required. Within a group, results before treatment were compared with those after treatment, using Student's two-tailed paired t-test. Comparisons between groups were made by Student's two-tailed unpaired t-test.

Subgroup analyses were performed to study the variables that had the most significant influence on the ovulation rate. LH concentrations were considered elevated in subjects with serum LH>9.0 IU/l; testosterone was considered to be elevated when >0.9 ng/ml, and patients were considered to be hyperinsulinaemic if the FI concentration was >14 mIU/ml. These upper limit values were 2 SD above the follicular phase means of women of reproductive age in our laboratory.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Of the 20 subjects, 15 completed the study, and are included in the analyses. Two patients conceived during treatment and withdrew from the study. One of these patients delivered a healthy child at term, while the other miscarried in the first trimester. Diarrhoea and vomiting were the commonest side-effects of treatment, which caused the other three patients to withdraw from the study.

The mean age of study subjects was 29.4 years (SD 3.5 years).

Table IGo shows that there was no change recorded in the BMI, waist circumference, or waist:hip ratio (WHR) of patients after 8 weeks of metformin treatment. Mean FI was high at the baseline visit and showed an apparent increase in the fasting sample taken immediately after discontinuation of treatment (within 3 days), possibly indicating a rebound effect. However, the FI results were variable and inconsistent (as indicated by the high SD) and so they were disregarded for the purposes of this study.


View this table:
[in this window]
[in a new window]
 
Table I. Effects of 8 weeks metformin treatment upon anthropometric criteria and fasting insulin concentrations in 15 oligomenorrhoeic women. Values are means (SD) unless otherwise indicated
 
Ovulation
Figure 1Go shows that significant increases in ovulation rate were observed in the whole group and sub-groups of subjects during treatment with metformin compared to their pretreatment observations. During the pretreatment observation period, there were 6 luteal weeks (in four patients) in 46 observation weeks for the whole group, while during treatment, there was a significant increase in the proportion of luteal weeks (46 in 120 observation weeks; {chi}2 = 8.74; P = 0.031).



View larger version (30K):
[in this window]
[in a new window]
 
Figure 1. Ovulation rates in all 15 patients and the different sub-groups (see Table IIGo), before (PRE) and during 8 weeks of metformin treatment (Rx). NS = not significant.

 
Figure 1Go also shows that the sub-group of patients with elevated testosterone before treatment (testosterone >0.9 ng/ml: n = 9) showed a significant improvement in ovulation rate. While only 4 luteal weeks were recorded in 27 observation weeks before treatment, 28 luteal weeks were observed in 72 observation weeks during treatment ({chi}2 = 7.784; P = 0.005). This latter ratio was close to the theoretical maximum ratio that would be seen in women with normal menstrual rhythm, compared with a very low rate of corpus luteum formation observed prior to treatment. Patients with normal pretreatment testosterone concentration (n = 6) did not show a significant increase in ovulation rate, perhaps because they showed a higher rate of ovulation before treatment.

The sub-group of nine patients who were hyperinsulinaemic (FI > 14 mIU/l) before treatment showed a moderate increase in ovulation rate from 4 luteal weeks in 32 observation weeks before treatment, to 28 luteal weeks in 71 observation weeks ({chi}2 = 6.268; P = 0.012) during treatment. Patients with normal fasting insulin concentrations showed no significant increase in ovulation rate.

Endocrine profiles
All patients
Table IIGo shows that the serum testosterone concentrations in all patients declined from 0.82 to 0.60 ng/ml within the first week of treatment with metformin (P < 0.05). Thereafter, the mean concentrations remained within the normal range throughout the treatment period. However, there was a rapid return to the pretreatment values immediately (samples taken 3 days after discontinuation) after treatment (0.87 ng/ml ± 0.31 SD: P < 0.05). Table IIGo also shows that patients with normal testosterone prior to treatment showed no changes in any parameter measured.


View this table:
[in this window]
[in a new window]
 
Table II. Circulating concentrations of testosterone and luteinizing hormone (LH) seen before and after 7 days of metformin treatment in all subjects treated and in the subgroups defined by their pretreatment testosterone concentrations. Values are means (SD)
 
Although the concentration of LH did not decline significantly in the first week of treatment, there was a general decline observed through treatment. However, these data were complicated by the inclusion of LH surges at the relevant sample points during the treatment period as the ovulation rate increased.

Table IIIGo shows that at the week 6 assessment point, there was no change in SHBG or androstenedione concentrations in the whole population. The decline in testosterone combined with no change in the SHBG represents a real decline in the bioavailable androgens in these patients during treatment.


View this table:
[in this window]
[in a new window]
 
Table III. Circulating concentrations of androstenedione and sex hormone binding globulin (SHBG) seen before and after 6 weeks of metformin treatment in all subjects treated and in the subgroups defined by their pretreatment testosterone concentrations. Values are means (SD)
 
Patients with elevated testosterone
Figure 2Go shows the rapid and steep decline in circulating testosterone that was observed in the patients with elevated baseline concentrations of testosterone within 1 week of starting metformin, from a baseline concentration of 1.02–0.58 ng/ml at week 1 (P < 0.005, Table IIGo). The lowest concentrations were observed after 4 weeks of treatment (0.48 ± 0.17 ng/ml), but this was not different from the value seen at week 1. At discontinuation of treatment, circulating testosterone increased rapidly (within 3 days), such that immediate post-treatment values were no different from pretreatment concentrations. Figure 2Go also shows that LH declined significantly from a pretreatment concentration of 12.1 to 8.5 IU/l (P = 0.02) at week 1 in the hyperandrogenic sub-group. However, at the end of treatment, the LH concentration profile did not mirror the changes in testosterone. Thus, the post-treatment rise in testosterone concentrations preceded any subsequent change in LH.



View larger version (18K):
[in this window]
[in a new window]
 
Figure 2. Changes in the mean plasma testosterone and luteinizing hormone (LH) concentrations in response to metformin treatment in the sub-groups with elevated testosterone and normal testosterone. At the end of treatment (post) the LH remained at the concentrations seen during treatment, while the testosterone returned to the elevated pretreatment values (*P < 0.001). For the SD on each point, see Table IIGo.

 
Figure 2Go also shows that there was no change in the circulating testosterone or LH concentrations amongst the group with normal testosterone values prior to treatment.

Table IIIGo shows that although there was no change in SHBG concentrations during treatment, there was a decline in the androstenedione concentrations recorded at week 6 in the hyperandrogenic sub-group.

Patients with raised FI
The sub-group of patients with high FI (n = 9) showed a mean baseline testosterone in the normal range 0.72 (0.32) ng/ml that decreased to 0.44 (0.29) ng/ml (P = 0.04) within 1 week of commencing metformin treatment. The decline was followed by a return to post-treatment concentration of 0.80 (0.28) ng/ml, which was not significantly different from the baseline value (P = 0.54). The mean profiles for testosterone in the normo-insulinaemic sub-group showed similar changes, but the data points were not statistically significant. Although the mean concentration of LH decreased in both hyper- and normo-insulinaemic subjects, the changes were not significant at any time point (data not shown). Concentrations of SHBG were similar in both groups at baseline, and did not change during treatment.

The hyperinsulinaemic group showed other features often associated with high fasting serum insulin. They had a higher BMI [41.8 (4.5) kg/m2 versus 28.9 (7.4) kg/m2; P = 0.002] and waist circumference [113.0 (6.8) cm versus 83.0 (10.1) cm; P < 0.003] compared to normo-insulinaemic subjects. Neither BMI nor waist circumference was significantly different at the end of treatment.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
These results show that metformin administration was associated with a significant increase in biochemically defined ovulation rates in obese women with PCOS. The improvements in ovulation rate were most marked in subjects with high baseline testosterone, but were less marked in the hyperinsulinaemic subgroup, suggesting that responses to metformin are not determined solely by the existence of hyperinsulinaemia. The improvements in ovarian function and ovulation rates occurred after a short duration of treatment, and were independent of changes in weight or anthropometric criteria. The hyperandrogenic sub-group showed a rapid decline in circulating testosterone (within 1 week) that was maintained throughout the treatment period, and at treatment discontinuation, it rapidly returned to pretreatment concentrations (within 3 days); this event preceded any rise in LH. The suppressive effect of treatment upon androstenedione is further evidence supporting an effect at the ovarian level as this hormone has been shown to correlate with the volume of ovarian stroma in women with PCOS (Kyei-Mensah et al., 1998Go). The data indicate that the effects of metformin upon abnormal ovarian androgen metabolism in women with hyperandrogenic anovulation do not require protracted exposure to normalized insulin activity, and they suggest that there may be a direct tissue-sensitizing effect of the drug at the ovarian level.

The most significant improvement in ovulation rate and hormone profile was observed in the subgroup of subjects who had high testosterone or high LH concentrations at baseline. Patients with high baseline testosterone concentrations were not significantly more hyperinsulinaemic at the baseline visit than the group with normal testosterone concentrations, and they showed similar BMI, and waist circumferences. During treatment this group exhibited a significant decrease in circulating testosterone and LH concentrations within the first week, and the testosterone values reverted to pretreatment concentrations soon after discontinuing treatment. These data suggest that metformin alters the complex of follicular steroid metabolism, that is aberrant in women with PCOS when the diagnosis includes hyperandrogenism, and its effects may not be mediated by directly influencing hyperinsulinaemia. Rather, metformin may act by modulating the sensitivity of ovarian follicles to circulating insulin, perhaps by augmenting the post-receptor mechanism of action of insulin within the follicular cells.

Metformin may selectively modulate the action of insulin on steroidogenesis in follicular cells in PCOS, which is thought to be abnormally regulated (Gilling-Smith et al., 1994Go, 1997Go). This could explain the 20% increase in serum testosterone upon discontinuation of metformin seen without a simultaneous increase in LH. The mechanism of this, however, remains unclear. Interestingly, patients in the high androgen sub-group showed a rapid return of testosterone to pretreatment concentrations after treatment discontinuation. This rise preceded the rise in LH, arguing against the hypothesis that PCOS involves a primary disorder of LH hypersecretion as a cause of the hyperandrogenaemia (Yen et al., 1970Go; Rebar et al., 1976Go; Schoemaker, 1991Go; Willis et al., 1996Go). However, the LH responses observed in this study should be interpreted in the light of the increased follicular development and ovulation rate that followed metformin treatment, eliciting LH surge activity at a higher frequency than in the pretreatment period.

It was not possible to demonstrate a simple association between hyperinsulinaemia and hyperandrogenaemia in our group of obese subjects with PCOS. The hyperinsulinaemic patients demonstrated other stigmata of insulin resistance, as they had significantly higher BMIs and waist circumferences (Cigolini et al., 1991Go; Pasquali et al., 1993Go; Edwards et al., 1994Go). LH and testosterone concentrations were however in the normal range and were similar to the normo-insulinaemic subjects.

The differential responses seen in the hyperandrogenic sub-group in this study population of obese women with oligomenorrhoea, may indicate that hyperandrogenic anovulation is a distinct syndrome with a definable aetiology (Franks et al., 1997Go). On the other hand, it provides no insight into the causes of abnormal ovarian function in the remainder of the population, except that peripheral insulin resistance may not be a relevant component.

In conclusion, this study provides clear indications that treatment with metformin results in improved ovulation profiles in hyperandrogenic women with PCOS, by a mechanism that is independent of changes in body weight. The mechanism is probably also independent of FI. This is being further investigated in a placebo-controlled trial with bigger sample sizes currently ongoing. The rapid changes in sex steroids at the start and at discontinuation of treatment indicate that the effects of metformin on ovarian function are rapid and short acting, and may be direct at the follicular level. Metformin treatment had no effect on ovarian function in women with normal testosterone. These results imply that the mechanisms of oligo- or anovulation are different in patients with PCOS with elevated testosterone or LH compared with those with normal concentrations of these hormones. Metformin may thus play an important role in restoring ovulatory function in a sub-group of oligomenorrhoeic women with PCOS.


    Acknowledgments
 
The authors wish to thank Professor I.A.Greer and Dr Naveed Sattar for their constructive criticism during the preparation of this manuscript and Serono UK Ltd for their financial assistance for this study.


    Notes
 
1 To whom correspondence should be addressed at: Bellshill Maternity Hospital, North Road, Bellshill, Lanarkshire ML4 3JN, UK Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Acbay, O. and Gundogdu, S. (1996) Can metformin reduce insulin resistance in polycystic ovary syndrome? Fertil. Steril., 65, 946–949.[ISI][Medline]

Adams, J., Polson, D.W. and Franks, S. (1986) Prevalence of polycystic ovaries in women with anovulation and idiopathic hirsutism. Br. Med. J. Clin. Res. Ed., 293, 355–359.[Medline]

Barbieri, R.L., Makris, A., Randall, R.W. et al. (1986) Insulin stimulates androgen accumulation in incubations of ovarian stroma obtained from women with hyperandrogenism. J. Clin. Endocrinol. Metab., 62, 904–910.[Abstract]

Barbieri, R.L., Smith, S. and Ryan, K.J. (1988) The role of hyperinsulinemia in the pathogenesis of ovarian hyperandrogenism. Fertil. Steril., 50, 197–212.[ISI][Medline]

Bergh, C., Carlsson, B., Olsson, J.H. et al. (1993) Regulation of androgen production in cultured human thecal cells by insulin-like growth factor I and insulin. Fertil. Steril., 59, 323–331.[ISI][Medline]

Cara, J.F. and Rosenfield, R.L. (1988) Insulin-like growth factor I and insulin potentiate luteinizing hormone-induced androgen synthesis by rat ovarian thecal-interstitial cells. Endocrinology, 123, 733–739.[Abstract]

Chang, R.J., Nakamura, R.M., Judd, H.L. et al. (1983) Insulin resistance in nonobese patients with polycystic ovarian disease. J. Clin. Endocrinol. Metab., 57, 356–359.[Abstract]

Cigolini, M., Seidell, J.C., Charzewska, J. et al. (1991) Fasting serum insulin in relation to fat distribution, serum lipid profile, and blood pressure in European women: the European Fat Distribution Study. Metab. Clin. Exp., 40, 781–787.[ISI][Medline]

Crave, J.C., Fimbel, S., Lejeune, H. et al. (1995) Effects of diet and metformin administration on sex hormone-binding globulin, androgens, and insulin in hirsute and obese women. J. Clin. Endocrinol. Metab., 80, 2057–2062.[Abstract]

DeVane, G.W., Czekala, N.M., Judd, H.L. et al. (1975) Circulating gonadotropins, estrogens, and androgens in polycystic ovarian disease. Am. J. Obstet. Gynecol., 121, 496–500.[ISI][Medline]

Dunaif, A., Segal, K.R., Futterweit, W. et al. (1989) Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes, 38, 1165–1174.[Abstract]

Dunaif, A., Scott, D., Finegood, D. et al. (1996) The insulin-sensitizing agent troglitazone improves metabolic and reproductive abnormalities in the polycystic ovary syndrome. J. Clin. Endocrinol. Metab., 81, 3299–3306.[Abstract]

Edwards, K.L., Austin, M.A., Newman, B. et al. (1994) Multivariate analysis of the insulin resistance syndrome in women. Arterioscler. Thromb., 14, 1940–1945.[Abstract]

Ehrmann, D.A., Schneider, D.J., Sobel, B.E. et al. (1997a) Troglitazone improves defects in insulin action, insulin secretion, ovarian steroidogenesis, and fibrinolysis in women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab., 82, 2108–2116.[Abstract/Free Full Text]

Ehrmann, D.A., Cavaghan, M.K., Imperial, J. et al. (1997b) Effects of metformin on insulin secretion, insulin action, and ovarian steroidogenesis in women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab., 82, 524–530.[Abstract/Free Full Text]

Franks, S. (1995) Polycystic ovary syndrome. N. Engl. J. Med., 333, 853–861.[Free Full Text]

Franks, S., Gharani, N., Waterworth, D. et al. (1997) The genetic basis of polycystic ovary syndrome. Hum. Reprod., 12, 2641–2648.[Abstract]

Gilling-Smith, C., Willis, D.S., Beard, R.W. et al. (1994) Hypersecretion of androstenedione by isolated thecal cells from polycystic ovaries. J. Clin. Endocrinol. Metab., 79, 1158–1165.[Abstract]

Gilling-Smith, C., Story, H., Rogers, V. et al. (1997) Evidence for a primary abnormality of thecal cell steroidogenesis in polycystic ovary syndrome. Clin. Endocrinol., 47, 93–99.[ISI][Medline]

Kiddy, D.S., Hamilton-Fairley, D., Bush, A. et al. (1992) Improvement in endocrine and ovarian function during dietary treatment of obese women with polycystic ovary syndrome. Clin. Endocrinol., 36, 105–111.[ISI][Medline]

Kyei-Mensah, A.A., LinTan, S., Zaidi, J. et al. (1998) Relationship of ovarian stromal volume to serum androgen concentrations in patients with polycystic ovary syndrome. Hum. Reprod., 13, 1437–1441.[Abstract]

Morin-Papunen, L.C., Koivunen, R.M., Ruokonen, A. et al. (1998) Metformin therapy improves the menstrual pattern with minimal endocrine and metabolic effects in women with polycystic ovary syndrome. Fertil. Steril., 69, 691–696.[ISI][Medline]

Nahum, R., Thong, K.J., and Hillier, S.G. (1995) Metabolic regulation of androgen production by human thecal cells in vitro. Hum. Reprod., 10, 75–81.[Abstract]

Nestler, J.E. and Jakubowicz, D.J. (1996) Decreases in ovarian cytochrome P450c17 alpha activity and serum free testosterone after reduction of insulin secretion in polycystic ovary syndrome. N. Engl. J. Med., 335, 617–623.[Abstract/Free Full Text]

Nestler, J.E., Barlascini, C.O., Matt, D.W. et al. (1989) Suppression of serum insulin by diazoxide reduces serum testosterone levels in obese women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab., 68, 1027–1032.[Abstract]

Nestler, J.E., Jakubowicz, D.J., Evans, W.S. et al. (1998) Effects of metformin on spontaneous and clomiphene-induced ovulation in the polycystic ovary syndrome. N. Engl. J. Med., 338, 1876–1880.[Abstract/Free Full Text]

Pasquali, R., Casimiri, F., Cantobelli, S. et al. (1993) Insulin and androgen relationships with abdominal body fat distribution in women with and without hyperandrogenism. Horm. Res., 39, 179–187.[ISI][Medline]

Poretsky, L. and Kalin, M.F. (1987) The gonadotropic function of insulin. Endocr. Rev., 8, 132–141.[Abstract]

Rebar, R., Judd, H.L., Yen, S.S. et al. (1976) Characterization of the inappropriate gonadotropin secretion in polycystic ovary syndrome. J. Clin. Invest., 57, 1320–1329.[ISI][Medline]

Robinson, S., Rodin, D.A., Deacon, A. et al. (1992) Which hormone tests for the diagnosis of polycystic ovary syndrome? Br. J. Obstet. Gynaecol., 99, 232–238.[ISI][Medline]

Schoemaker, J. (1991) Neuroendocrine control in polycystic ovary-like syndrome. Gynecol. Endocrinol., 5, 277–288.[ISI][Medline]

Velazquez, E.M., Mendoza, S., Hamer, T. et al. (1994) Metformin therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin resistance, hyperandrogenemia, and systolic blood pressure, while facilitating normal menses and pregnancy. Metab. Clin. Exp., 43, 647–654.[ISI][Medline]

Velazquez, E.M., Mendoza, S.G., Wang, P. et al. (1997a) Metformin therapy is associated with a decrease in plasma plasminogen activator inhibitor-1, lipoprotein(a), and immunoreactive insulin levels in patients with the polycystic ovary syndrome. Metab. Clin. Exp., 46, 454–457.[ISI][Medline]

Velazquez, E.M., Acosta, A. and Mendoza, S.G. (1997b) Menstrual cyclicity after metformin therapy in polycystic ovary syndrome. Obstet. Gynecol., 90, 392–395.[Abstract/Free Full Text]

Willis, D. and Franks, S. (1995) Insulin action in human granulosa cells from normal and polycystic ovaries is mediated by the insulin receptor and not the type-I insulin-like growth factor receptor. J. Clin. Endocrinol. Metab., 80, 3788–3790.[Abstract]

Willis, D., Mason, H., Gilling-Smith, C. et al. (1996) Modulation by insulin of follicle-stimulating hormone and luteinizing hormone actions in human granulosa cells of normal and polycystic ovaries. J. Clin. Endocrinol. Metab., 81, 302–309.[Abstract]

Yen, S.S., Vela, P., and Rankin, J. (1970) Inappropriate secretion of follicle stimulating hormone and luteinizing hormone in polycystic ovarian disease. J. Clin. Endocrinol. Metab., 30, 435–442.[ISI][Medline]

Submitted on May 14, 1999; accepted on September 13, 1999.