Hum. Reprod. Advance Access published September 30, 2005

Transcription

1 Human Reproduction Page 1 of 8 Hum. Reprod. Advance Access published September 30, 2005 doi: /humrep/dei312 Rosiglitazone and ethinyl estradiol/cyproterone acetate as single and combined of overweight women with polycystic ovary syndrome and insulin resistance A.Lemay 1, S.Dodin, L.Turcot, F.Déchêne and J-C.Forest Départements d Obstétrique Gynécologie et de Biologie Médicale, Centre de Recherche, Hôpital St-François d Assise, CHUQ, Université Laval, Québec, Canada 1 To whom correspondence should be addressed at: Hôpital St-François d Assise (CHUQ), 10 rue de l Espinay, Québec P.Q., Canada G1L 3L5. andre.lemay@ogy.ulaval.ca BACKGROUND: Few studies have evaluated insulin sensitizers in comparison/association with oral contraceptives (OC) in women with polycystic ovary syndrome (PCOS) with insulin resistance (IR). This study assessed the effects of a thiazolidinedione versus an anti-androgenic estrogen progestin followed by their sequential combinations in overweight PCOS women. METHODS AND RESULTS: Twenty-eight candidates in whom elevated insulin was not normalized after 4 months of diet were randomly assigned to 6 months of rosiglitazone 4 mg/day or to ethinyl estradiol 35 g/cyproterone acetate 2 mg (EE/CPA: 21/28 days cycle). Each group then received both medications for another 6 months. Rosiglitazone reduced insulin, IR indices [homeostasis model assessment (HOMA) and quantitative sensitivity check index (QUICKI)] and the insulin area under the curve in response to an oral glucose tolerance test (OGTT), but had limited effect on lipids, androgens and hirsutism. EE/CPA did not modify insulin and OGTT response but increased high-density lipoprotein cholesterol and triglycerides and decreased androgens and hirsutism. Similar changes occurred during combined s. End results were highly significant in combined groups without noticeable side-effects or changes in safety parameters. CONCLUSIONS: In obese PCOS women with high insulin not corrected by diet, the combination of rosiglitazone and EE/CPA may be used to achieve complementary beneficial effects on endocrine metabolic anomalies and clinical symptoms. Key words: insulin resistance/insulin sensitizer/oral contraceptives/polycystic ovarian disease Introduction The polycystic ovary syndrome (PCOS), as recently redefined by the association of two out of three anomalies (oligoovulation/anovulation, hyperandrogenaemia and polycystic ovaries at ultrasound), is rather prevalent and frequently associated with insulin resistance (IR) (ESHRE ASRM, 2004). The proportion of PCOS subjects with IR has been estimated to be 50% in either obese or non-obese women (ESHRE-ASRM, 2004). Recent data indicate a higher prevalence reaching 95% in obese PCOS patients depending on methods to assess IR (Carmina and Lobo, 2004; Legro et al., 2004). In cases with IR, insulin sensitizers are being used in PCOS women to lower insulin which plays a strategic role in anovulation, hyperandrogenaemia, weight gain and dyslipidaemia (Prelevic, 1997; Livingstone and Collison, 2002). According to a Cochrane review, metformin, which acts mainly through inhibition of glucose production by the liver, is considered effective and safe to induce ovulation and to favour pregnancy (Lord et al., 2003). However, data gathered from infertility studies concerning androgens, lipids and hirsutism are limited and indicate little impact on endocrine metabolic anomalies of the syndrome (Lord et al., 2003). By binding to the peroxisome proliferator-activated receptor (PPAR-γ), thiazolidinediones are implicated in the transcription of several factors involved in the regulation of glucose and lipid metabolism, mainly in the adipose and muscle tissues (Camp et al., 2000; Hsueh and Law, 2003). Troglitazone, rosiglitazone and pioglitazone have also been shown effective in restoring fertility and appear promising in improving endocrine and metabolic abnormalities (Dunaif et al., 1996; Hasegawa et al., 1999; Azziz et al., 2001; Romualdi et al., 2003; Shobokshi and Shaarawy, 2003; Brettenthaler et al., 2004; Guido et al., 2004a; Sepilian and Nagamani, 2005). The usual medication for PCOS women not seeking a pregnancy is the use of an oral contraceptive (OC) to inhibit excess ovarian androgen and hirsutism and to control endometrial growth and uterine bleeding. In normal women the intake of various estrogen progestin formulations has been associated with a decrease in insulin sensitivity which is mainly related to the estrogen content of the formulation and the weight of the The Author Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. 1 For Permissions, please journals.permissions@oupjournals.org

2 A.Lemay et al. subject (Godsland et al., 1990). Heterogeneous results have been reported concerning the evaluation of IR in PCOS subjects taking an OC (Prelevic et al., 1990; Korytkowski et al., 1995; Escobar-Morreale et al., 2000; Morin-Papunen et al., 2000, 2003; Cagnacci et al., 2003; Harborne et al., 2003; Guido et al., 2004a; Cibula et al., 2005). Most recent data indicate that low dose OC oppose or restrict the effect of metformin and anti-obesity agents in improving IR in PCOS (Elter et al., 2002; Ibanez and de Zegher, 2003, 2004a,b; Sabuncu et al., 2003; Mitkov et al., 2004; Cibula et al., 2005). This study was designed to evaluate the efficacy and possible interaction of a thiazolidinedione versus an estrogen progestin having a direct anti-androgen action on the endocrine and metabolic anomalies of the syndrome. In overweight PCOS women having high insulin not corrected by a preparatory diet, the insulin sensitizer rosiglitazone and an OC containing estradiol cyproterone acetate (EE/CPA) was initially used alone for 6 months and in reciprocal combination for another 6 months to assess respective actions of each medication and their eventual interactions. The results of this protocol indicate that when used as single agent, rosiglitazone mainly reduces IR whereas EE/CPA normalizes androgens and improves hirsutism. The combination of both medications adds complementary benefits without antagonizing effects, noticeable side-effects or changes in safety parameters. Materials and methods Study population This research was approved by the Ethics Committee of the St- François d Assise Hospital of the Centre Hospitalier Universitaire de Quebec. The volunteers were recruited through local newspaper advertising and referral from physicians. The following criteria were used at screening: age between 18 and 45 years old with at least two previous menses and without menopausal symptoms, evidence of PCOS according to the presence of two out of three criteria as revised by a recent consensus on PCOS (ESHRE-ASRM, 2004): (i) oligomenorrhoea (<8 uterine bleedings/year) or amenorrhoea ( 2 uterine bleedings/year); (ii) elevated levels of androgens (testosterone >1.5 nmol/l and/or androstenedione >10 nmol/l); and (iii) presence of micro-cysts ( 12 follicles measuring 2 9 mm in diameter) surrounding otherwise enlarged ovaries (>10 cm 3 ) (Balen et al., 2003). A high fasting insulin (>90 pmol/l) with normal glucose (<6 mmol/l)) was also required as evidence of IR. The threshold levels were based on reference values for a normal population of women aged years old as estimated in our institution for testosterone (90th percentile) and for a general population of women by the manufacturers of the assays for androstenedione and insulin. The exclusion criteria were: actual desire for pregnancy, hysterectomy, abnormal endometrial biopsy if abnormal bleeding in the last 6 months, clinical evidence of Cushing s syndrome, congenital adrenal hyperplasia (17-OH progesterone >10 nmol/l), excessive androgens suspicious of a tumour, prolactin levels >50 µg/l, severe renal or hepatic disease, gastrointestinal condition interfering with drug absorption, previous breast, uterus, ovary or liver neoplasia, previous use of a drug to lower glucose, lipid or insulin or an oral contraceptive in the last 2 months, previous use of diuretic, β-blocker, corticoid, hormonal replacement therapy in the last 3 months, depo-medroxyprogesterone acetate injection in the last year and a research drug in the last 2 months, alcohol intake >40 g/day and smoking >10 cigarettes/day. 2 Study design Admissible candidates were started on diet low in refined sugar for 4 months before randomization to drug. Subjects were also encouraged to exercise daily whenever possible (Figure 1). Subjects who failed to normalize their fasting insulin (<90 pmol/l on two assessments two weeks apart) at the end of the 4th month were randomized to take first the thiazolidinedione or the estrogen progestin. Computergenerated allocation blocks of four to six subjects were used for randomization which was done by phone call the morning of starting the medication. After 6 months of single, the two medications were administered together for 6 more months. Diet and physical activity recommendations were maintained throughout the study. Medical Diet recommendations were explained by a dietitian to correct for dietary habits when appropriate and to maintain a desirable weight while minimizing refined sugars. At the beginning of the study and at each visit, subjects were encouraged to do a physical activity for 30 consecutive minutes whenever possible each day. The oral formulation containing 35 µg of ethinyl estradiol and 2 mg of cyproterone acetate (Diane 35: Berlex Canada, Montreal, Québec, Canada) was taken daily 21 out of 28 days. Oral rosiglitazone (Avandia: GlaxoSmithKline, Montreal, Québec, Canada) was taken at a daily dose of 4 mg before breakfast. Clinical assessment Visits were scheduled at months 4 and 2 during the pre-medication diet phase, at baseline (month 0) and every 3 months during the phase (months 3, 6, 9 and 12) for clinical evaluation and blood sampling. For measurement of glucose, insulin and lipid profile, the blood sample was drawn in a sitting position after a 12 h fast. A 75 g oral glucose tolerance test (OGTT) was also done at months 0, 6 and 12. The feeding habits were evaluated by a 3 day food diary record including a weekend day. Compliance to the diet was assessed by repeated 3 day food diaries at the scheduled visits. Detailed analysis of food intake could then be done using a computer program calculating proportions of carbohydrates, proteins, lipids, small nutrients and alcohol based on nutrient values of Canadian common foods (Canada Health Services and Promotion Branch, 1988). Physical activity was evaluated and scored by asking the patient at scheduled visits how many times they practised a physical activity for min in the last 3 months. Scores given were 1 for no activity or baseline daily activity, 2 for less than once per month, 3 for once per month, 4 for two or three times per month, 5 for one or two times per week and 6 for three times or more per week (Gionet and Godin, 1989). Figure 1. Diagram of sequential phases during the protocol.

3 Rosiglitazone EE/CPA in obese PCOS with high insulin Clinical symptoms were evaluated using a monthly diary card for occurrence of uterine bleedings, side-effects, intake of drugs and other possible medications. Uterine bleeding was induced by the intake of medroxyprogesterone acetate (10 mg 10 days) whenever there was no spontaneous bleeding after 8 weeks during the diet preparatory phase and the administration of rosiglitazone alone in group A. Hirsutism was assessed every 3 months using the Ferriman Gallwey score (Ferriman and Gallwey, 1961). For clinical safety evaluation, patients were asked at each visit for the monthly occurrence of a series of clinical symptoms usually associated with side-effects (fatigue, dizziness, hot flashes, sudation, sugar craving, headache, diarrhoea, breast tenderness, nausea, vomiting and fluid retention or any peculiar sign or symptom). Safety laboratory tests at each 3- monthly visit included: hepatic glutamic oxalacetic transaminase and glutamic pyruvic transaminase, creatine kinase (CK) and the coagulation factors fibrinogen and plasminogen activator inhibitor type-1 (PAI-1). Laboratory analyses Thyroid-stimulating hormone, prolactin, LH, FSH, estradiol and progesterone were measured by electrochemiluminescence immunoassay (ECLIA) on an Elecsys 2010 immunoassay analyser using commercial kits (Roche Diagnostics, Laval, Québec, Canada). Total testosterone, androstenedione and dehydroepiandrosterone sulphate (DHEA-S) were measured by solid-phase 125 I radioimmunoassay in unextracted serum using specific antibody immobilized to the wall of a polypropylene tube (Diagnostic Products Corporation, Inter Medico, Markham, Ontario, Canada). 17α-Hydroxyprogesterone was estimated by double antibody 125 I radioimmunoassay (Medicorp Inc., Orangeburg, New York). Sex hormone-binding globulin (SHBG) concentrations were measured by an immunoradiometric method using monoclonal antibodies (Diagnostic Products Corporation, Inter Medico, Markham, Ontario, Canada). All these assays were performed as previously reported (Lemay et al., 1995). The free androgen index (FAI) was calculated using the following equation: FAI = [testosterone (nmol/l) 100]/SHBG (nmol/l). Insulin was determined by an homologous radioimmunoassay kit using an antibody having low cross-reactivity with pro-insulin (<0.2%) (Linco Research, Inc., St-Charles, Missouri, USA) as described previously (Lemay et al., 2002). In this assay 1 miu/ml is equivalent to 6 pmol/l of insulin. The normal range of values for adults is 5 15 miu/ml or pmol/l. Insulin sensitivity was determined by homeostasis model assessment (HOMA) = (fasting insulin in miui/l fasting glucose in mmol/l)/22.5 and by the quantitative sensitivity check index (QUICKI) = 1/(log fasting insulin in miu/ml + log fasting glucose in mg/dl) as used and validated in PCOS (Carmina and Lobo, 2004). For OGTT, blood samples were collected at 0 min after a h fast and at 30, 60, and 180 min after ingestion within 5 min of a 300 ml water solution containing 75 g of glucose. Glucose and insulin were assayed in all samples. The area under the curve (AUC) was calculated by the trapezoidal method. Total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C) and triglycerides (TG) concentrations were measured by enzymatic colorimetric reactions on a chemistry analyser Hitachi 917 using reagents from Roche Diagnostics, Laval, Québec, Canada. Low-density lipoprotein cholesterol (LDL-C) was calculated according to the equation of Friedewald et al. (1972). Serum apolipoprotein A-1 (ApoA) and Apo B-100 (ApoB) were measured with commercial reagents (Dade-Behring, Mississauga, Canada) on BN- 100 Nephelometer (Dade-Behring, Marburg, Germany) (Lemay et al., 2001). Fibrinogen was measured fresh after centrifugation. Plasminogen activator inhibitor-1 (PAI-1) was assayed with the spectrolyse tpa/ PAI activity kit of Organon Teknika Inc. (Scarborough, Ontario, Canada). Statistical analysis All statistical analysis was done with SAS statistical software (SAS Institute Inc., Cary, NC, USA). Student s t-test was used for comparisons between the two study groups at randomization, at the end of the single phase (month 6) and at the end of combined s (month 12). Paired t-test was used for within-group comparisons of differences between the beginning and end of each intervention phase (month 0 to month 6, month 6 to month 12) and between baseline and end of the study (month 0 to month 12) using the Bonferroni correction for multiple comparisons. P < 0.05 was considered statistically significant. Mean ± SD was used in the tables for baseline characteristics while mean ± SEM was used in tables for assessment of effects. Results Number of participants during the protocol Sixteen women (36%) normalized their fasting insulin and were not eligible for (Figure 2). A total of 28 subjects were then randomized, either to group A (n = 15) or B (n = 13). The reasons for dropping out during the intervention periods were loss of interest or non-compliance (n = 6), intolerance to the estrogen progestin (n = 3) and the occurrence of another medical problem (n = 2). The anthropometric and biochemical characteristics of women dropping out were not Figure 2. Diagram of patients flow at each step of the protocol. 3

4 A.Lemay et al. different from those completing the study except for older age in three women dropping out from group B. Baseline characteristics At randomization there were no differences in the mean age of both groups ranging from 18 to 40 years of age. However, at completion of the study, the subjects in group B (Table I) were younger than those of group A (P < 0.003), the women dropping out being older in group B than in group A. Otherwise, clinical and biochemical characteristics of the PCOS women were similar in both groups. At baseline, after the initial 4 months of diet and before starting either of the two drugs, both groups had achieved a reduction in the intake of calories ( 12.4 ± 8.3%, P = in group A; 17.0 ± 8.4%, P = in group B). This caloric restriction was related to a decrease in carbohydrate intake ( 19 ± 10%, P = in group A; 22 ± 9%, P = in group B). The proportions of lipids, carbohydrates, proteins and the content in dietary cholesterol indicated adequate dietary habits for a person having appropriate intake of calories (Table I). Improvement in dietary habit was maintained as assessed by repeated evaluation at each visit (data not shown). The baseline score for physical activity was moderate for both groups corresponding to one or two periods per week (Table I). There were no changes in physical activity score throughout the study (data not shown). Glucose and insulin metabolism Both PCOS groups had normal levels of glucose throughout the study. However, insulin levels were high at baseline (month 0, Table II). After 6 months of single, rosiglitazone (group A) decreased insulin, HOMA and QUICKI. The initial intake of EE/CPA for 6 months (group B) did not modify insulin. The addition of the thiazolidinedione to the OC and of EE/ CPA to rosiglitazone (month 12) produced similar changes in insulin as observed during either single drug administration. As also shown in Table II, the AUC for glucose was normal in both PCOS groups. It did not change during. On Table I. Clinical baseline characteristics (means ± SD) for subjects randomized to group A started on rosiglitazone before addition of ethinyl estradiol/cyproterone acetate (EE/CPA) or to group B started on EE/CPA before addition of rosiglitazone Parameter Group A Group B No. of women 10 7 Age (years) 26.8 ± ± 1.5** Hirsutism score 16.7 ± ± 2.0 Acne (%) Ovarian microcysts (%) Weight (kg) 95.9 ± ± 12.9 Height (cm) ± ± 6.0 Body mass index (kg/m 2 ) 34.6 ± ± 4.6 Waist (cm) ± ± 10.2 Waist:hip ratio 1.00 ± ± 0.04 Blood pressure (mmhg) Systolic 118 ± ± 9 Diastolic 78 ± 7 72 ± 7 Intake of calories/day 1577 ± ± 196 Physical activity score 5.70 ± ± 1.86 **P < Table II. Metabolic profile (means ± SEM) at specified time intervals during the protocol Baseline Month 0 Mono Month 6 Combined Month 12 Group A Glucose (mmol/l) 5.1 ± ± ± 0.2 Insulin (pmol/l) 150 ± ± 12** 113 ± 13 HOMA 5.74 ± ± 0.47* 4.01 ± 0.48 QUICKI ± ± 0.006** ± AUC glucose (mmol/l.h) 5.8 ± ± ± 1.3 AUC insulin (pmol/l.h) 1646 ± ± 155** 1236 ± 195 Group B Glucose (mmol/l) 5.0 ± ± ± 0.1 Insulin (pmol/l) 164 ± ± ± 18, HOMA 6.20 ± ± ± 0.73, QUICKI ± ± ± AUC glucose (mmol/l.h) 4.4 ± ± ± 0.9 AUC insulin (pmol/l.h) 2365 ± ± ± 146 In group A, rosiglitazone alone until month 6 was followed by the addition of ethinyl estradiol/cyproterone acetate (EE/CPA) till month 12, whereas in group B, EE/CPA alone until month 6 was followed by addition of rosiglitazone till month 12. The data were analysed within groups by the paired t-test with the Bonferroni correction for multiple comparisons. *P <0.05 month 6 versus month 0; **P <0.01 month 6 versus month 0; P <0.05 month 12 versus month 6; P <0.01 month 12 versus month 6. P <0.05 month 12 versus month 0; P <0.01 month 12 versus month 0. HOMA = homeostasis model assessment; QUICKI = quantitative sensitivity check index; AUC = area under the curve. the other hand, the AUC for insulin at month 0 is high in both groups. In group A, rosiglitazone (month 6) significantly reduced insulin. The increase in AUC for insulin following the addition of EE/CPA during months 6 to 12 is not significant (P = 0.23). In group B, the decrease of AUC for insulin did not reach statistical significance following the sequential addition of EE/CPA (month 6: P = 0.54) and of rosiglitazone (month 12: P = 0.12). Serum lipid profile At baseline (month 0, Table III) TG was >1.0 nmol/l and HDL-C was <1.3 mmol/l in both PCOS groups. After 6 months of single, rosiglitazone (group A) did not induce changes in lipids. The initial intake of EE/CPA for 6 months (group B) increased HDL-C and Apo A. The addition of the thiazolidinedione to the OC and of EE/CPA to rosiglitazone (month 12) produced similar changes in lipids as obtained during either single drug administration. Serum androgen levels At baseline (month 0, Table IV), both PCOS groups had elevated levels of testosterone and androstenedione with a high FAI related to low levels of SHBG. Rosiglitazone caused a decrease of androstenedione and FAI and a limited increase of SHBG. These improvements were similar whether rosiglitazone was administered before EE/CPA (group A, months 0 to 6) or after EE/CPA (group B, months 6 to 12). On the other hand, EE/CPA induced a marked increase in SHBG and suppressed FAI. These changes were also similar whether EE/CPA was administered before rosiglitazone (group B, months 0 to 6) or

5 Rosiglitazone EE/CPA in obese PCOS with high insulin Table III. Lipid profile (means ± SEM) at specified time intervals during the protocol Baseline Month 0 Mono Month 6 Combined Month 12 Group A Triglycerides (mmol/l) 1.39 ± ± ± 0.25, Total cholesterol (mmol/l) 4.23 ± ± ± 0.26 LDL-cholesterol (mmol/l) 2.41 ± ± ± 0.21 HDL-cholesterol (mmol/l) 1.19 ± ± ± 0.14 Apolipoprotein A (g/l) 1.29 ± ± ± 0.10, Apolipoprotein B (g/l) 0.83 ± ± ± 0.06 Group B Triglycerides (mmol/l) 1.14 ± ± ± 0.37 Total cholesterol (mmol/l) 4.67 ± ± ± 0.32 LDL-cholesterol (mmol/l) 2.94 ± ± ± 0.29 HDL-cholesterol (mmol/l) 1.20 ± ± 0.11** 1.64 ± 0.11 Apolipoprotein A (g/l) 1.28 ± ± 0.11** 1.75 ± 0.11 Apolipoprotein B (g/l) 0.94 ± ± ± 0.09 In group A, rosiglitazone alone until month 6 was followed by the addition of ethinyl estradiol/cyproterone acetate (EE/CPA) till month 12 whereas in group B EE/CPA alone until month 6 was followed by addition of rosiglitazone till month 12. The data were analysed within groups by the paired t-test with the Bonferroni correction for multiple comparisons. *P < 0.05 month 6 versus month 0; **P < 0.01 month 6 versus month 0; P < 0.05 month 12 versus month 6; P <0.01 month 12 versus month 6. P < 0.05 month 12 versus month 0; P < 0.01 month 12 versus month 0. HDL = high-density lipoprotein; LDL = low-density lipoprotein. Table IV. Endocrine profile (means ± SEM) at specified time intervals during the protocol Baseline Month 0 Mono Month 6 Combined Month 12 Group A Testosterone (mmol/l) 2.29 ± ± ± 0.22 Androstenedione (mmol/l) 12.2 ± ± ± 1.0 DHEA-S (pmol/l) 6.72 ± ± ± 0.78 SHBG (mmol/l) 18 ± 2 23 ± ± 27, Free androgen index 13.2 ± ± ± 0.5, Group B Testosterone (mmol/l) 1.93 ± ± ± 0.37 Androstenedione (mmol/l) 11.2 ± ± 1.1* 8.4 ± 2.0 DHEA-S (pmol/l) 7.37 ± ± ± 0.81 SHBG (mmol/l) 13 ± 2 85 ± 14** 116 ± 21 Free androgen index 15.9 ± ± 0.6** 2.0 ± 0.6 In group A rosiglitazone alone until month 6 was followed by the addition of EE/CPE till month 12 whereas in group B EE/CPA alone until month 6 was followed by addition of rosiglitazone till month 12. The data were analysed within groups by the paired t-test with the Bonferroni correction for multiple comparisons. *P < 0.05 month 6 versus month 0; P < 0.01 month 12 versus month 6; **P < 0.01 month 6 versus month 0; P < 0.01 month 12 versus month 0. DHEA-S = dehydroepiandrosterone sulphate; SHBG = sex hormone-binding globulin. blood pressure, cigarette smoking and alcohol drinking habits or the use of current or occasional medications (data not shown). A significant decrease in the hirsutism score was obtained only following the administration of EE/CPA (Figure 3). However, there was no difference between the two groups in the reduction of the score at the end of the study (month 12). During the initial 4 months of diet before drug intervention, there were only nine spontaneous menses in the 17 subjects who completed the study. More frequent spontaneous uterine bleedings (n = 29 episodes) occurred during months 0 to 6 of rosiglitazone taken alone in all 10 subjects of group A. Regular uterine bleeding occurred during the pause whenever EE/CPA was used in the two groups. Safety profile There was no apparent association of the occasional occurrence of a list of usual side-effects or peculiar symptoms and the different phases of the study. Mild to moderate symptoms of fatigue and sugar craving disappeared in most instances during the preparatory diet phase. There were no changes from baseline in the serum levels of hepatic transaminases and creatine kinase. Rosiglitazone had no effect on fibrinogen and PAI-1 whereas a 50% decrease in this pro-coagulation factor by EE/ CPA did not reach statistical significance (data not shown). End results for combined groups Since the changes brought by each of the two drugs were similar in each group whatever their sequential order of administration, results of groups A and B were combined for overall efficacy analysis. As shown in Table V, paired comparisons between values at randomization and values at the end of the study indicate highly significant improvements in all metabolic and androgen parameters except for no changes in TC, LDL-C, Apo B, AUC for glucose and a significant increase in TG. Furthermore no significant changes were noticed in anthropometric and clinical or safety parameters. following rosiglitazone (group A, months 6 to 12). The estrogen progestin also reduced total testosterone, androstenedione and DHEA-S but to variable extents in the two treated groups. Clinical efficacy parameters Throughout the medication periods there were no changes in body weight, body mass index (BMI), waist and hip girths, Figure 3. Means ± SEM of hirsutism score at baseline, end of single (month 6) and end of combined s (month 12). *P < 0.05 month 6 versus month 0; P < 0.01 month 12 versus month 6; P < 0.01 month 12 versus month 0. 5

6 A.Lemay et al. Table V. Overall efficacy of assessed parameters (means ± SEM) of combined groups A and B Discussion The main objective of this study was to assess the effects of a thiazolidinedione type of insulin sensitizer with or without an anti-androgen estrogen progestin as a new approach to treat the multiple endocrine metabolic problems of overweight PCOS women with elevated insulin. The results indicate that in the 64% of subjects not responding to a diet low in refined sugars, rosiglitazone alone was effective in improving insulin, HOMA, QUICKI and insulin response to OGTT. On the other hand, EE/CPA was effective to decrease androgens, to suppress FAI, to improve the hirsutism score and to increase HDL-C. The changes induced by each drug were similar whether one drug was administered before or after the other in their sequential combinations. Although there were no changes in anthropometric measures, highly significant improvements on metabolic (except for increased TG) and androgen parameters were observed at the end of when results of groups A and B were combined. Noticeably, the complementary beneficial effects of both medications were also without notable side-effects or changes in safety parameters. The results obtained with rosiglitazone as a single agent compare with those previously reported with thiazolidinediones in obese PCOS women. Troglitazone, rosiglitazone and pioglitazone have been shown to be effective in improving IR in all reported studies but none has been demonstrated effective in reducing BMI as also observed in our study (Dunaif et al., 1996; Ehrmann et al., 1997; Hasegawa et al., 1999; Azziz et al., 2001; Ghazeeri et al., 2003; Romualdi et al., 2003; Shobokshi and Shaarawy, 2003; Belli et al., 2004; Brettenthaler et al., 2004; Guido et al., 2004b; Sepilian and Nagamani, 2005). Among these studies, a significant decrease of total testosterone was obtained in four reports (Dunaif et al., 1996; Ehrmann 6 Baseline Month 0 End of s Month 12 Glucose (mmol/l) 5.1 ± ± Insulin (pmol/l) 166 ± ± 11 < HOMA 5.93 ± ± 0.41 < QUICKI ± ± AUC glucose (mmol/l.h) 5.3 ± ± AUC insulin (pmol/l.h) 1942 ± ± Triglycerides (mmol/l) 1.28 ± ± Total cholesterol (mmol/l) 4.41 ± ± LDL-cholesterol (mmol/l) 2.63 ± ± HDL-cholesterol (mmol/l) 1.19 ± ± 0.09 < Apolipoprotein A (g/l) 1.28 ± ± 0.07 < Apolipoprotein B (g/l) 0.88 ± ± Testosterone (mmol/l) 2.14 ± ± Androstenedione (mmol/l) 11.8 ± ± 0.9 < DHEA-S (pmol/l) 6.99 ± ± SHBG (mmol/l) 16 ± ± 18 < Free androgen index 14.3 ± ± 0.4 < Hirsutism score 17.2 ± ± 0.4 < The data were analysed by paired t-test. HOMA = homeostasis model assessment; QUICKI = quantitative sensitivity check index; AUC = area under the curve; LDL = low-density lipoprotein; HDL = high-density lipoprotein; DHEA-S = dehydroepiandrosterone sulphate; SHBG = sex hormone-binding globulin. P et al., 1997; Hasegawa et al., 1999; Sepilian and Nagamani, 2005). A significant reduction in FAI was also observed in three other studies due to a significant increase in SHBG (Azziz et al., 2001; Shobokshi and Shaarawy, 2003; Brettenthaler et al., 2004). There is no obvious reason explaining the nonsignificant changes in testosterone, FAI and SHBG in our study as also observed in others (Ghazeeri et al., 2003; Romualdi et al., 2003; Belli et al., 2004; Guido et al., 2004a). Our results with rosiglitazone are also in agreement with previous data reporting no significant changes in lipid fractions, notably in TG and HDL-C (Ehrmann et al., 1997; Hasegawa et al., 1999; Romualdi et al., 2003; Brettenthaler et al., 2004). Overall, data in obese PCOS women indicate that the usual dosage of each thiazolidinedione used alone is effective in decreasing insulin but has various efficacies on testosterone and does not improve the lipid profile nor reduce weight and BMI. Our results with EE/CPA as a single medication are similar to those of other studies in obese PCOS subjects with IR reporting effective reduction in testosterone, androstenedione, DHEA-S and FAI (Morin-Papunen et al., 2000; Harborne et al., 2003; Sabuncu et al., 2003). The important increase in SHBG which has high-affinity binding for testosterone probably explains the modest decrease in total testosterone. This may suggest an incomplete suppression of ovarian testosterone production although the free testosterone might be decreased due to extensive binding by the high amount of SHBG. EE/CPA did not cause significant changes in IR but increased low HDL-C and already elevated TG as also observed in our study (Morin-Papunen et al., 2000; Harborne et al., 2003; Sabuncu et al., 2003). The increased TG levels at the completed is explained by the wellknown effect of estrogen progestin formulations on TG. Our results demonstrate that the sequential combinations of rosiglitazone and EE/CPA and EE/CPA and rosiglitazone achieved effects that were similar to those obtained when used as a single initial agent. Statistical analysis did not reveal any trend in the fluctuations of insulin or its response to OGTT after combinations of both drugs. These results would indicate no apparent interactions between the two medications. The end results of both groups combined showed highly significant changes in metabolic and androgen parameters at the end of the period. The significance of these end results would then be due not only to the increased number of subjects evaluated but also to 12 months exposure to at least one of the two medications in each group. Besides this paper, only a few studies have evaluated EE/ CPA in association with another intervention aiming at reducing insulin in PCOS with IR. No significant differences were found in metabolic and androgen parameters in lean PCOS subjects treated for 4 months with EE/CPA and metformin as compared to EE/CPA alone (Elter et al., 2002). A recent study on obese PCOS indicates that EE/CPA causes additional reductions of HOMA and testosterone when added for 2 months at the beginning of metformin (Mitkov et al., 2004). However, EE/CPA has been reported to interfere with the reduction in IR by the anti-obesity agent sibutramine which was also effective in reducing weight and improving lipids (Sabuncu et al., 2003). These previous data indicate variable efficacy of agents aiming at reducing IR in the presence of EE/CPA.

7 Rosiglitazone EE/CPA in obese PCOS with high insulin Indeed, the absence of effect of EE/CPA on IR and of interference with rosiglitazone in our study would be unexpected considering that various OC decrease insulin sensitivity in normal subjects (Godsland et al., 1990; Petersen et al., 1999). Recent studies further indicate that low dose OC reduce insulin sensitivity in non-obese PCOS women (BMI <25 kg/m 2 ) with normal insulin (Cibula et al., 2005). Moreover, it has been shown that low dose OC also further increase elevated insulin levels in non-obese adult PCOS women (Cagnacci et al., 2003) and in lean PCOS girls (Ibanez and de Zegher, 2003, 2004, 2004). Our data in obese PCOS women indicate that, unlike the findings in lean PCOS girls, a low dose OC containing EE/ CPA does not increase insulin. This observation is in agreement with the results of several other studies having evaluated insulin during EE/CPA administration to PCOS patients (Morin-Papunen et al., 2000; Elter et al., 2002; Harborne et al., 2003; Sabuncu et al., 2003). The lack of increase in IR by the EE/CPA formulation used in this study could be related to the direct anti-androgenic action of CPA. However, the data on improvement of insulin sensitivity by anti-androgens in PCOS are heterogeneous, being related to weight and levels of androgens (Diamanti-Kandarakis et al., 1995; Moghetti et al., 1996; Ibanez et al., 2003; Vrbikova et al., 2004). Finally, our results show that rosiglitazone is effective in reducing insulin to a similar extent in the presence or in the absenceof EE/CPA. This is in contrast to other studies reporting few or no effects of metformin in the presence of other low dose OC (Ibanez and de Zegher, 2004; Cibula et al., 2005). The differences in response to the insulin sensitizer could be related to the property of CPA that inhibits testosterone receptor and action. This is supported by results of of metformin combined with the anti-androgen receptor flutamide showing normalization of insulin, lipids and androgens in lean PCOS girls (Ibanez and de Zegher, 2003). Alternatively rosiglitazone could be more effective than metformin since metformin was reported as ineffective in reducing insulin in lean PCOS patients also taking EE/CPA (Elter et al., 2002). Rosiglitazone could also reduce visceral fat which plays a significant role in IR (Carey et al., 2002). However, there is as yet no study combining a thiazolidinedione with another OC. Since the results of this study indicate complementary benefits, further studies are required in obese PCOS women to better understand the endocrine and metabolic changes by with other OC with or without an anti-androgen, an insulin sensitizer or a combination of interventions including modifications of diet and physical activity. In summary, we demonstrate the benefits of combining a thiazolidinedione type of insulin sensitizer with an oral estrogen progestin formulation having a direct anti-androgenic action. In obese PCOS women with high insulin which is not corrected initially by a diet low in refined sugars for 4 months, a 4 mg daily dose of rosiglitazone improves IR while 35 µg of EE and 2 mg of CPA has mixed effects on lipids (favourable increase of HDL-C versus unfavourable elevation of TG) and overall beneficial effects on androgens, hirsutism and uterine bleeding. This combination of medications appears without noticeable changes in side-effects and safety parameters. The simultaneous reduction of excess ovarian androgen production, blockage of androgen action and sensitization to insulin constitutes a new approach for the of the multiple endocrine and metabolic anomalies of overweight PCOS subjects in which high insulin is not corrected by an appropriate diet. Combinations of other OC, anti-androgens and insulin sensitizers could also be of benefit. This new approach needs to be further investigated for an optimal of an overweight PCOS patient presenting with high insulin and requesting an OC. Acknowledgements The authors are indebted to Monique Longpré and Annie Levesque- Langelier for their professional nursing assistance and research experience and to Sylvie Dostie for typing the manuscript. Supported by a grant from the Canadian Institutes of Health Research to A.Lemay, S.Dodin and J.C.Forest. Drugs were graciously supplied by Berlex Canada Inc. and GlaxoSmithKline Montreal, Québec, Canada. References Azziz R, Ehrmann D, Legro RS, Whitcomb RW, Hanley R, Fereshetian AG, O Keefe M and Ghazzi MN (2001) Troglitazone improves ovulation and hirsutism in the polycystic ovary syndrome: a multicenter, double blind, placebo-controlled trial. J Clin Endocrinol Metab 86, Balen AH, Laven JS, Tan SL and Dewailly D (2003) Ultrasound assessment of the polycystic ovary: international consensus definitions. Hum Reprod Update 9, Belli SH, Graffigna MN, Oneto A, Otero P, Schurman L and Levalle OA (2004) Effect of rosiglitazone on insulin resistance, growth factors, and reproductive disturbances in women with polycystic ovary syndrome. Fertil Steril 81, Brettenthaler N, De Geyter C, Huber PR and Keller U (2004) Effect of the insulin sensitizer pioglitazone on insulin resistance, hyperandrogenism, and ovulatory dysfunction in women with polycystic ovary syndrome. J Clin Endocrinol Metab 89, Cagnacci A, Paoletti AM, Renzi A, Orru M, Pilloni M, Melis GB and Volpe A (2003) Glucose metabolism and insulin resistance in women with polycystic ovary syndrome during therapy with oral contraceptives containing cyproterone acetate or desogestrel. J Clin Endocrinol Metab 88, Camp HS, Li O, Wise SC, Hong YH, Frankowski CL, Shen X, Vanbogelen R and Leff T (2000) Differential activation of peroxisome proliferator-activated receptor-gamma by troglitazone and rosiglitazone. Diabetes 49, Canada Health Services and Promotion Branch (1988) Nutrient Values of Some Common Foods. Ottawa. Carey DG, Cowin GJ, Galloway GJ, Jones NP, Richards JC, Biswas N and Doddrell DM (2002) Effect of rosiglitazone on insulin sensitivity and body composition in type 2 diabetic patients [corrected]. Obes Res 10, Carmina E and Lobo RA (2004) Use of fasting blood to assess the prevalence of insulin resistance in women with polycystic ovary syndrome. Fertil Steril 82, Cibula D, Fanta M, Vrbikova J, Stanicka S, Dvorakova K, Hill M, Skrha J, Zivny J and Skrenkova J (2005) The effect of combination therapy with metformin and combined oral contraceptives (COC) versus COC alone on insulin sensitivity, hyperandrogenaemia, SHBG and lipids in PCOS patients. Hum Reprod 20, Diamanti-Kandarakis E, Mitrakou A, Hennes MM, Platanissiotis D, Kaklas N, Spina J, Georgiadou E, Hoffmann RG, Kissebah AH and Raptis S (1995) Insulin sensitivity and antiandrogenic therapy in women with polycystic ovary syndrome. Metabolism 44, Dunaif A, Scott D, Finegood D, Quintana B and Whitcomb R (1996) The insulin-sensitizing agent troglitazone improves metabolic and reproductive abnormalities in the polycystic ovary syndrome. J Clin Endocrinol Metab 81, Ehrmann DA, Schneider DJ, Sobel BE, Cavaghan MK, Imperial J, Rosenfield RL and Polonsky KS (1997) Troglitazone improves defects in insulin action, insulin secretion, ovarian steroidogenesis, and fibrinolysis in women with polycystic ovary syndrome. J Clin Endocrinol Metab 82, Elter K, Imir G and Durmusoglu F (2002) Clinical, endocrine and metabolic effects of metformin added to ethinyl estradiol cyproterone acetate in nonobese women with polycystic ovarian syndrome: a randomized controlled study. Hum Reprod 17,

8 A.Lemay et al. Escobar-Morreale HF, Lasuncion MA and Sancho J (2000) Treatment of hirsutism with ethinyl estradiol desogestrel contraceptive pills has beneficial effects on the lipid profile and improves insulin sensitivity. Fertil Steril 74, ESHRE ASRM (2004) Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod 19, Ferriman D and Gallwey JD (1961) Clinical assessment of body hair growth in women. J Clin Endocrinol Metab 21, Friedewald WT, Levy RI and Fredrickson DS (1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 18, Ghazeeri G, Kutteh WH, Bryer-Ash M, Haas D and Ke RW (2003) Effect of rosiglitazone on spontaneous and clomiphene citrate-induced ovulation in women with polycystic ovary syndrome. Fertil Steril 79, Gionet NJ and Godin G (1989) Self-reported exercise behavior of employees: a validity study. J Occup Med 31, Godsland IF, Crook D and Wynn V (1990) Low-dose oral contraceptives and carbohydrate metabolism. Am J Obstet Gynecol 163, Guido M, Romualdi D, Giuliani M, Suriano R, Selvaggi L, Apa R and Lanzone A (2004a) Drospirenone for the of hirsute women with polycystic ovary syndrome: a clinical, endocrinological, metabolic pilot study. J Clin Endocrinol Metab 89, Guido M, Romualdi D, Suriano R, Giuliani M, Costantini B, Apa R and Lanzone A (2004b) Effect of pioglitazone on the adrenal androgen response to corticotrophin in obese patients with polycystic ovary syndrome. Hum Reprod 19, Harborne L, Fleming R, Lyall H, Sattar N and Norman J (2003) Metformin or antiandrogen in the of hirsutism in polycystic ovary syndrome. J Clin Endocrinol Metab 88, Hasegawa I, Murakawa H, Suzuki M, Yamamoto Y, Kurabayashi T and Tanaka K (1999) Effect of troglitazone on endocrine and ovulatory performance in women with insulin resistance-related polycystic ovary syndrome. Fertil Steril 71, Hsueh WA and Law R (2003) The central role of fat and effect of peroxisome proliferator-activated receptor-gamma on progression of insulin resistance and cardiovascular disease. Am J Cardiol 92,3J 9J. Ibanez L and de Zegher F (2003) Low-dose combination of flutamide, metformin and an oral contraceptive for non-obese, young women with polycystic ovary syndrome. Hum Reprod 18, Ibanez L and de Zegher F (2004a) Ethinylestradiol drospirenone, flutamide metformin, or both for adolescents and women with hyperinsulinemic hyperandrogenism: opposite effects on adipocytokines and body adiposity. J Clin Endocrinol Metab 89, Ibanez L and De Zegher F (2004b) Flutamide metformin plus an oral contraceptive (OC) for young women with polycystic ovary syndrome: switch from third- to fourth-generation OC reduces body adiposity. Hum Reprod 19, Ibanez L, Ong K, Ferrer A, Amin R, Dunger D and de Zegher F (2003) Lowdose flutamide metformin therapy reverses insulin resistance and reduces fat mass in nonobese adolescents with ovarian hyperandrogenism. J Clin Endocrinol Metab 88, Korytkowski MT, Mokan M, Horwitz MJ and Berga SL (1995) Metabolic effects of oral contraceptives in women with polycystic ovary syndrome. J Clin Endocrinol Metab 80, Legro RS, Castracane VD and Kauffman RP (2004) Detecting insulin resistance in polycystic ovary syndrome: purposes and pitfalls. Obstet Gynecol Surv 59, Lemay A, Dodin S, Cedrin I and Turcot-Lemay L (1995) Phasic serum lipid excursions occur during cyclical oral conjugated oestrogens but not during transdermal oestradiol sequentially combined with oral medroxyprogesterone acetate. Clin Endocrinol (Oxf) 42, Lemay A, Dodin S, Turcot L, Dechene F and Forest JC (2001) Estrogen/progesterone replacement versus pravastatin and their sequential association in hypercholesterolemic postmenopausal women. Maturitas 40, Lemay A, Dodin S, Kadri N, Jacques H and Forest JC (2002) Flaxseed dietary supplement versus hormone replacement therapy in hypercholesterolemic menopausal women. Obstet Gynecol 100, Livingstone C and Collison M (2002) Sex steroids and insulin resistance. Clin Sci (Lond) 102, Lord JM, Flight IH and Norman RJ (2003) Insulin-sensitising drugs (metformin, troglitazone, rosiglitazone, pioglitazone, D-chiro-inositol) for polycystic ovary syndrome. Cochrane Database Syst Rev, CD Mitkov M, Pekhlivanov B and Terzieva D (2004) [Combined use of metformin and Diane35 in polycystic ovary syndrome]. Akush Ginekol (Sofiia) 43, Moghetti P, Tosi F, Castello R, Magnani CM, Negri C, Brun E, Furlani L, Caputo M and Muggeo M (1996) The insulin resistance in women with hyperandrogenism is partially reversed by antiandrogen : evidence that androgens impair insulin action in women. J Clin Endocrinol Metab 81, Morin-Papunen LC, Vauhkonen I, Koivunen RM, Ruokonen A, Martikainen HK and Tapanainen JS (2000) Endocrine and metabolic effects of metformin versus ethinyl estradiol cyproterone acetate in obese women with polycystic ovary syndrome: a randomized study. J Clin Endocrinol Metab 85, Morin-Papunen L, Vauhkonen I, Koivunen R, Ruokonen A, Martikainen H and Tapanainen JS (2003) Metformin versus ethinyl estradiol cyproterone acetate in the of nonobese women with polycystic ovary syndrome: a randomized study. J Clin Endocrinol Metab 88, Petersen KR, Christiansen E, Madsbad S, Skouby SO, Andersen LF and Jespersen J (1999) Metabolic and fibrinolytic response to changed insulin sensitivity in users of oral contraceptives. Contraception 60, Prelevic GM (1997) Insulin resistance in polycystic ovary syndrome. Curr Opin Obstet Gynecol 9, Prelevic GM, Wurzburger MI, Trpkovic D and Balint-Peric L (1990) Effects of a low-dose estrogen antiandrogen combination (Diane-35) on lipid and carbohydrate metabolism in patients with polycystic ovary syndrome. Gynecol Endocrinol 4, Romualdi D, Guido M, Ciampelli M, Giuliani M, Leoni F, Perri C and Lanzone A (2003) Selective effects of pioglitazone on insulin and androgen abnormalities in normo- and hyperinsulinaemic obese patients with polycystic ovary syndrome. Hum Reprod 18, Sabuncu T, Harma M, Nazligul Y and Kilic F (2003) Sibutramine has a positive effect on clinical and metabolic parameters in obese patients with polycystic ovary syndrome. Fertil Steril 80, Sepilian V and Nagamani M (2005) Effects of rosiglitazone in obese women with polycystic ovary syndrome and severe insulin resistance. J Clin Endocrinol Metab 90, Shobokshi A and Shaarawy M (2003) Correction of insulin resistance and hyperandrogenism in polycystic ovary syndrome by combined rosiglitazone and clomiphene citrate therapy. J Soc Gynecol Invest 10, Vrbikova J, Hill M, Dvorakova K, Stanicka S, Vondra K and Starka L (2004) Flutamide suppresses adrenal steroidogenesis but has no effect on insulin resistance and secretion and lipid levels in overweight women with polycystic ovary syndrome. Gynecol Obstet Invest 58, Submitted on May 16, 2005; resubmitted on August 20, 2005; accepted on August 23,