Obesity differentially affects serum levels of androstenedione and testosterone in polycystic ovary syndrome

Transcription

1 Obesity differentially affects serum levels of androstenedione and testosterone in polycystic ovary syndrome Carlos Moran, M.D., M.Sc., a,b Jose L. Renteria, M.D., M.Sc., a Segundo Moran, M.D., M.Sc., a Joaquin Herrera, B.Sc., a Sandra Gonzalez, B.Sc., a and Jose A. Bermudez, M.D. a a Health Research Council and b Medical Unit of High Specialty in Gynecology and Obstetrics No. 4, Mexican Institute of Social Security, Mexico City, Mexico Objective: To assess androstenedione (A) and T levels in obese and nonobese patients with polycystic ovary syndrome (PCOS) after GnRH and oral glucose tolerance tests (OGTT). Design: Cross sectional study. Setting: Clinical research center. Patient(s): Thirty patients with PCOS, of whom 15 were obese and 15 were nonobese, and 7 women without PCOS were included in the study. Intervention(s): The GnRH test and OGTT were performed in all subjects. Main Outcome Measure(s): Basal and stimulated levels of LH, FSH, insulin, A, and total T were measured. Areas under the curve (AUCs) and AUC change after stimulation for these hormones were calculated. Result(s): The basal T levels were significantly higher in obese than in nonobese patients with PCOS. In contrast, the basal levels of A were similar in obese and nonobese patients with PCOS. The T AUC after GnRH was significantly greater in obese than in nonobese patients with PCOS but was not significantly different after OGTT. The A AUC after GnRH and OGTT was significantly greater in nonobese than in obese patients with PCOS. However, there were no significant differences in T AUC and A AUC changes after GnRH and OGTT. Conclusion(s): A different pattern in the levels of T and A with respect to obesity in PCOS was observed, suggesting a shift in ovarian enzymatic function. (Fertil Steril Ò 28;9: Ó28 by American Society for Reproductive Medicine.) Key Words: Hyperandrogenism, polycystic ovary syndrome, obesity, testosterone, androstenedione Polycystic ovary syndrome (PCOS) affects 4% 8% of women of reproductive age (1) and is found in 5% 8% patients with hyperandrogenism (2, 3). Polycystic ovary syndrome is characterized by two of three diagnostic criteria: oligoovulation or anovulation, clinical and/or biochemical hyperandrogenism, and the presence of polycystic ovaries, after excluding other etiologies (4). Approximately 8% of patients with PCOS are either overweight or obese (3, 5). Received February 11, 27; revised and accepted September 7, 27. Supported in part by the following research grants: M from Consejo Nacional de Ciencia y Tecnologıa (CONACYT), Mexico City, and 25/1/I/92 and 26/1A/I/19 from the Mexican Institute of Social Security (Mexico City, Mexico). Presented in part at the 88 th Annual Meeting of The Endocrine Society, Boston, Massachusetts, June 24 27, 26. Reprint requests: Carlos Moran, M.D., M.Sc., Health Research Council, Mexican Institute of Social Security, 222 Bassett Avenue, , EI Paso, Texas 7991 (FAX: ; cemoranv@ hotmail.com). In comparison to androgen levels in normal women, androgen hypersecretion is found in patients with PCOS (6, 7). Augmented LH levels (8) and hyperinsulinemia (9) also are found. Luteinizing hormone stimulates ovarian production of androgens by theca cells (1). In normal conditions, ovarian androgens are converted to estrogens by means of aromatase in the granulosa cells (11), but in PCOS, this conversion is reduced (12, 13), and T increases as a result of overactive steroidogenesis (13). Obesity alone promotes a greater conversion of androstenedione (A) to estrone (14). Insulin resistance (IR) has been found in 5% 75% of patients with PCOS and causes compensatory hyperinsulinemia (5, 15, 16); IR is of a higher magnitude (15, 17) and is more frequently detected (17) among obese than nonobese patients with PCOS. It has been observed that insulin and/or LH are associated with greater production of A and T after incubation of ovarian stroma and theca cells that are obtained from normal and hyperandrogenic women (18, 19). It also has been observed that insulin with LH synergistically stimulates the production of A by tumor cells of ovarian origin (2). Insulin infusion has been shown to induce an increase of A in normal women and in those with hyperandrogenism (21). It appears that hyperandrogenism observed in PCOS is caused by stimulation by excess LH that is produced by gonadotropic dysfunction and hyperinsulinemia resulting from IR (22, 23). It has been reported that insulin affects gonadal steroid production independently of changes in gonadotropin secretion in PCOS (24). However, we have observed elsewhere that obese patients with IR have a higher dissociation of the LH-FSH ratio after GnRH test (17). 231 Fertility and Sterility â Vol. 9, No. 6, December /8/$34. Copyright ª28 American Society for Reproductive Medicine, Published by Elsevier Inc. doi:1.116/j.fertnstert

2 Androgen production in obese and nonobese patients with PCOS is controversial. Some investigators have reported that basal production of T (22, 25 28) or of A (22, 26, 28) is similar when comparing obese and nonobese patients with PCOS. Other researchers have reported that serum basal concentrations of T (29, 3) are higher in obese patients with PCOS, compared with in nonobese patients with PCOS. In contrast, other investigators have reported lower A concentrations in obese than nonobese patients with PCOS (25). The discrepancy between these studies may be a result of the different diagnostic criteria used to classify obesity and PCOS (22, 25 3). We hypothesized that obese and nonobese patients with PCOS present different patterns of A and T serum levels. To test this hypothesis, the concentrations of these androgens in serum were determined in obese and nonobese patients with PCOS, both basally and when stimulated by GnRH and oral glucose tolerance tests (OGTT). MATERIALS AND METHODS Subjects The research was reviewed and accepted by the institutional research board of the Mexican Institute of Social Security. All patients signed an informed consent form before their inclusion in the protocol. Thirty women diagnosed with PCOS and seven women with regular menstrual cycles and no hyperandrogenism were included in the study. The age range for all 37 patients was 2 35 years. They were classified into three groups: obese patients with PCOS (n ¼ 15), nonobese patients with PCOS (n ¼ 15), and control women (n ¼ 7). The sample size was calculated by the difference of means method, with data from a report elsewhere (28). Polycystic ovary syndrome was defined by oligo- or anovulation, clinical hyperandrogenism and/or hyperandrogenemia, and polycystic ovaries on ultrasound; other disorders such as Cushing syndrome, hypothyroidism, hyperprolactinemia, androgen-secreting neoplasms, and nonclassic adrenal hyperplasia had been excluded in research elsewhere (4). Additional criteria for noninclusion were the presence of diabetes mellitus, pregnancy, and the use of medicines that affect androgen production or carbohydrate metabolism (hormonal or hypoglycemic agents) during the 2 months before the study. Patients excluded later from the study were those who did not accept the stimulation tests, who declined to continue the study, and who were diagnosed with diabetes mellitus on OGTT. Obesity was defined as a body mass index (BMI) of R27 (31). Body mass index was calculated as the weight in kilograms divided by the height in meters squared. Waist and hip circumference were measured to evaluate body fat distribution by using the waist-to-hip ratio (32). Oligoanovulation was considered to be present when menstrual cycles were found to be >35 days or <26 days, or there was amenorrhea, which was considered to be present when menstruation was absent for >9 days (3). Patients were evaluated for hirsutism by using a modified Ferriman-Gallwey scale (33, 34), but its magnitude was not recorded. Protocol Pelvic ultrasound was determined by a 5-MHz transvaginal transducer (General Electric, Milwaukee, WI), on day 6 8 of the cycle in oligomenorrheic patients and control women and on any day in amenorrheic patients. The criteria for determining polycystic ovaries was the presence of R12 follicles in each ovary, each measuring <1 mm in diameter, and/ or increased ovarian volume (>1 cm 3 ), taking into account findings of reports published elsewhere (4, 35). Gonadotropin-releasing hormone and OGTT stimulation tests were performed when patients were in a fasting state, beginning at 8 AM, on 2 consecutive days chosen at random (17). These tests were performed during the early follicular phase (3 5 d) of a spontaneous menstrual cycle or on any given day in patients presenting amenorrhea. Samples were taken at 15,, 3, 6, 9, 12, and 18 minutes and were processed by centrifuge to determine the glucose concentration immediately. Samples for hormonal analysis were stored at 2 C until processed. Gonadotropin-releasing hormone (Relisorm-L, 1 mg IV; Serono, Italy) was administered immediately after the basal sample was taken. The OGTT was performed with 75 g of glucose (Sigma Aldrich Company, Ltd.; Irvine, Ayrshire, United Kingdom), and patients consumed a diet containing 25 g of carbohydrate daily, for 3 days before the test. Hormone and Glucose Analysis Levels of LH and FSH were determined by RIA by using commercial kits (RIA-agnost, France), and insulin was determined by available RIA kits (Cis Bio International, Gif-sur- Yvette, France). Androstenedione and total T were measured by using specific RIA techniques after thin-layer chromatography using silica gel plates (Merck, Darmstadt, Germany), as reported elsewhere (36). Serum glucose was measured by using the glucose-oxidase technique (Stanbio Laboratory Inc., Boerne, TX). All hormones were measured in duplicate, and the intraassay coefficients of variation were <5%. Analysis of Data Nonparametric measures such as median and range were used. The areas under the curve (AUCs) were calculated for hormonal values after stimulation. Also, the AUC of net change was calculated by subtracting the AUC corresponding to the median of the basal values (for each test time) from the AUC, after stimulation with GnRH and OGTT. The basal levels, the AUC, and the AUC change of all groups were compared by using the Kruskal-Wallis analysis of variance, and that for each pair of groups, by using the Mann-Whitney U test. P<.5 was considered to be statistically significant. Fertility and Sterility â 2311

3 FIGURE 1 FIGURE 1 CONTINUED Insulin ( U/mL) T (ng/ml) FSH (IU/L) LH (IU/L) a, b a a, b a b b a, c a, b b, c Box-and-whiskers plots of basal levels of insulin, LH, FSH, total T, and A in obese and nonobese patients with PCOS and in control women. The line within each box represents the median. Upper and lower boundaries of each box indicate 75 th and 25 th percentiles, respectively. The whiskers (above and below boxes) show the upper and lower adjacent value respectively. Superscript letters indicate significant differences (P<.5), within each plot, between pairs labeled with the same letter. The insulin levels were significantly greater in obese than in nonobese patients with PCOS and control women. The values of LH were significantly greater in both the obese and nonobese patients with PCOS than in the control group. The levels of FSH and A did not show any change between groups. The levels of T were significantly greater in the obese patients with PCOS, compared with in nonobese patients with PCOS and control women. Also, the levels of T were significantly greater in nonobese patients with PCOS than in control women. with PCOS (24 [2 28] y), and control women (27 [22 32] y). As expected by design, the BMI of obese patients with PCOS (31.1 [ ] kg/m 2 ) was significantly (P<.5) higher than the BMI of nonobese patients with PCOS (21.5 [ ] kg/m 2 ) and also was higher than the BMI of control women (23.8 [ ] kg/m 2 ). The waist-to-hip ratio of obese patients with PCOS (.88 [ ] cm) was significantly (P<.5) higher than that of nonobese patients with PCOS (.78 [.68.89] cm) and also was higher than the waist-to-hip ratio of control women (.77 [.74.79] cm). There were no significant differences in the median values of BMI and waist-to-hip ratio between nonobese patients with PCOS and control women. A (ng/ml) Obese PCOS Nonobese PCOS Controls RESULTS Clinical and Anthropometric Measures There were no differences among the median [range] ages of obese patients with PCOS (24 [2 3] y), nonobese patients Basal Hormonal Values Figure 1 shows the basal serum values of insulin, LH, FSH, A, and T in obese and nonobese patients with PCOS and in the control group. The basal levels of insulin were significantly higher (P<.5) in the obese than in the nonobese patients with PCOS and control women. Basal insulin levels of nonobese patients with PCOS were not significantly different from those of control women. The values of LH were similar between obese and nonobese patients with PCOS but were significantly higher in both (P<.5) than in the control group. Basal levels of FSH were not different between groups. Basal T levels were significantly higher (P<.5) in the obese patients with PCOS, compared with in nonobese patients with PCOS and control women. Also, basal T levels were significantly higher (P<.5) in nonobese patients with PCOS than in control women. The basal A levels were similar in all groups Moran et al. Androgen levels in PCOS and obesity Vol. 9, No. 6, December 28

4 TABLE 1 Areas under curve and AUC change of the hormonal values after the stimulation with GnRH in obese and nonobese patients with PCOS and controls. Parameter Obese PCOS Nonobese PCOS Controls (n [ 7) LH AUC (IU/L min) 31.3 ( ) a 34.3 ( ) b 11.4 ( ) a,b LH AUC change (IU/L min) 2.6 ( ) a 27.3 ( ) b 6.4 ( ) a,b FSH AUC (IU/L min) 1.2 ( ) 9.6 ( ) 1. ( ) FSH AUC change (IU/L min) 3.8 (.7 9.2) 2.8 ( ) 3. ( ) T AUC (ng/ml min) 1.14 ( ) a,b.93 ( ) a,c.62 (.54.65) b,c T AUC change (ng/ml min).3 (.55.34).6 (.33.39).18 (.35.19) A AUC (ng/ml min).9 ( ) a,b.99 ( ) a,c.81 ( ) b,c A AUC change (ng/ml min).4 (.39.23).1 (.19.17).16 (.27.12) Note: Data are median (range). a,b,c Superscript letters indicate a statistically significant difference (P<.5) between pairs. Areas Under the Curve of Hormones After GnRH Stimulation The AUC and AUC change of LH, FSH, T, and A after GnRH are shown in Table 1. There were no significant differences in values of LH AUC and LH AUC change after GnRH between obese and nonobese PCOS patient groups. The LH AUC and LH AUC change after GnRH were significantly greater (P<.5) in both obese and nonobese patients with PCOS when compared with the case of the control group (Fig. 2). The FSH AUC and FSH AUC change response to GnRH showed no significant differences between groups. The T AUC after GnRH was significantly greater (P<.5) in obese than in nonobese patients with PCOS and control women (Fig. 2). Also, the T AUC was found to be significantly greater (P<.5) in nonobese patients with PCOS than in control women. In contrast, the A AUC after GnRH was significantly greater in nonobese than obese patients with PCOS and was significantly greater (P<.5) in obese and nonobese patients with PCOS than in control women (Fig. 2). However, there were no significant differences in T AUC and A AUC changes after GnRH. Areas Under the Curve of Hormones After OGTT The AUCs and AUC change data for insulin, T, and A after OGTT are shown in Table 2. The insulin AUC and insulin AUC change after OGTT were significantly greater (P<.5) in obese patients with PCOS than in both nonobese patients with PCOS and control women (Fig. 2). Also, the insulin AUC and insulin AUC change were significantly greater (P<.5) in nonobese patients with PCOS than in control women. The T AUC with OGTT was greater, but not significantly so, in obese than in nonobese patients with PCOS and control women (Fig. 2). In contrast, the A AUC after OGTTwas significantly greater (P<.5) in nonobese patients with PCOS when compared with both obese patients with PCOS and control women (Fig. 2). There were no significant differences in the T AUC and A AUC changes after OGTT. Relationship of T and A Production Table 3 shows the basal and stimulated T-A ratios. The basal T-A ratio was significantly greater in obese than in nonobese patients with PCOS and control women. Also, it was significantly greater (P<.5) in nonobese patients with PCOS than in control women. After stimulation with GnRH, there were no statistically significant differences in the ratio of T AUC - A AUC between obese and nonobese patients with PCOS, but the ratio was significantly greater (P<.5) in obese patients with PCOS, compared with control women. With respect to OGTT, there were no statistically significant differences in the T AUC -A AUC ratio among the groups. DISCUSSION This study showed that basal and GnRH-stimulated T levels are significantly higher in obese than in nonobese patients with PCOS. Although basal A levels are similar in obese and nonobese patients with PCOS, the GnRH- and OGTTstimulated levels of A are higher in nonobese than in obese patients with PCOS. The fact that no differences are detected in the net change of T and A after GnRH and OGTT indicates that stimulated values are mainly affected by basal levels. We can speculate that the stimulation of LH and insulin for androgen production in PCOS is not acute but chronic. The results of this study agree with those of investigators who have reported basal serum T concentrations to be greater in obese than nonobese patients with PCOS (29, 3). However, other researchers have found that T was higher, but not statistically significantly so, in obese than nonobese patients with PCOS (22, 26 28), revealing perhaps a problem in the Fertility and Sterility â 2313

5 FIGURE 2 Areas under the curve of hormonal responses to the GnRH test (left panel) and OGTT (right panel). The LH AUC in response to GnRH was higher (P<.5) in obese and nonobese patients with PCOS than in control women. The insulin AUC after OGTT was significantly greater (P<.5) in obese patients with PCOS than in nonobese patients with PCOS and control women. Also, insulin AUC was significantly greater (P<.5) in nonobese patients with PCOS than in control women. The T AUC after the stimulation with GnRH was significantly higher (P<.5) in obese than in nonobese patients with PCOS but was not significantly different after OGTT. The A AUC after GnRH and OGTT was significantly higher (P<.5) in nonobese than in obese patients with PCOS and control women. Also, the A AUC after GnRH was significantly greater (P<.5) in obese patients with PCOS than in control women. Data are medians at each time point, with groups indicated as follows: squares, obese patients with PCOS; circles, nonobese patients with PCOS; and triangles, the control group. GnRH (1 µg IV) Oral glucose (75 g) LH (IU/L) Insulin ( U/mL) T (ng/ml).8 T (ng/ml) A (ng/ml).9 A (ng/ml) Time (min) Time (min) sample size and potency of the statistical evaluation. The synergic effect of LH and insulin on the stroma and theca cells (18, 19) could explain the greater basal concentration and T response to GnRH stimulation in obese patients with PCOS. In some studies published elsewhere (22, 26), as in ours, it was found that basal A was higher, but not statistically significantly so, in nonobese compared with obese patients with PCOS. By sampling every 1 minutes over several hours, 2314 Moran et al. Androgen levels in PCOS and obesity Vol. 9, No. 6, December 28

6 TABLE 2 Areas under curve and AUC change of the hormonal values after OGTT in obese and nonobese patients with PCOS and controls. Parameter Obese PCOS Nonobese PCOS Controls (n [ 7) Insulin AUC (mu/ml min) 22. ( ) a,b 12. ( ) a,c 43.3 ( ) b,c Insulin AUC change (mu/ml min) ( ) a,b 76. ( ) a,c 26.3 ( ) b,c T AUC (ng/ml min) ( ).84 ( ).8 (.51.99) T AUC change (ng/ml min).1 ( ).4 (.52.71).7 (.3.3) A AUC (ng/ml min).95 ( ) a 5 ( ) a,b.86 ( ) b A AUC change (ng/ml min).4 (.38.16).7 (.36.42).4 (.7.1) Note: Data are median (range). a,b,c Superscript letters indicate a statistically significant difference (P<.5) between pairs. higher A concentrations were found in nonobese than obese patients with PCOS (25). However, levels of A as stimulated by GnRH and OGTT are greater in nonobese than in obese patients with PCOS. These findings may be explained, at least in part, by an increased rate of aromatization from A to estrone in the adipose tissue of obese women, as has been reported elsewhere (14, 25). Another hypothesis is a more active conversion from A to T by 17b-hydroxysteroid dehydrogenase (17b-HSD) in obese PCOS women. The activity of 17b-HSD has been suggested to be increased in some patients with hyperandrogenism (13, 37). Results of the present study suggest that it is higher in obese than nonobese patients with PCOS. It has been mentioned that in some cases of hyperandrogenism, hyperinsulinemia may contribute to the stimulation of stromal 17b-HSD (37). This study and others (17, 25) have found that obese patients with PCOS have greater hyperinsulinemia than do nonobese patients with PCOS. However, we have to take into account that in PCOS, most of the T is secreted from the ovary, and approximately only one third is derived by peripheral conversion from A (38, 39). The correlation of basal T and A levels with insulin has been reported in obese patients with PCOS, with T being greater than with A (9). However, the synergic effect of LH and insulin has been shown to increase production of A but not T. This has been the case in vitro (2) as well as in clinical studies of obese patients and of those with hyperandrogenism and acanthosis nigricans (21). The basal T-A ratio is significantly greater in obese than nonobese patients with PCOS. The essential difference between the two groups with PCOS is the amount of adipose tissue present. Therefore, the basal difference in the T-A ratio may be due mainly to the action of aromatase of the adipose tissue, which converts more A to estrone (14). Perhaps a more active conversion from A to T by 17b-HSD also is present, because there is a higher level of hyperinsulinemia in obese than in nonobese PCOS women (37). The basal T-A ratio is significantly greater in nonobese patients with PCOS than in control women. The main difference between these groups is the presence of PCOS and not being overweight. For this reason, the main difference in A and T is TABLE 3 Relation of basal values and areas under curve of T and A after the stimulation with GnRH and OGTT in obese and nonobese patients with PCOS and controls. Ratio Obese PCOS Nonobese PCOS Controls (n [ 7) Basal T-A 9 ( ) a,b.91 ( ) a,c.73 ( ) b,c T AUC -A AUC after GnRH 1.1 ( ) a.95 ( ).69 (.45.86) a T AUC -A AUC after OGTT 5 ( ).87 ( ).94 ( ) Note: Data are median (range). a,b,c Superscript letters indicate a statistically significant difference (P<.5) between pairs. Fertility and Sterility â 2315

7 probably, at least in part, a result of increased action of 17b- HSD, which has the role of converting A to T. The action of this enzyme has been reported to be increased in granulosaluteal and stromal cells of hyperandrogenic women in vitro (37) and in hormonal studies in patients with functional ovarian hyperandrogenism (13). The T-A ratio is significantly greater in obese patients with PCOS than in control women, basally and after GnRH stimulation. The differences between the two groups are the presence of both PCOS and obesity and therefore could be a result of the combined action of aromatase and 17b-HSD. Several human 17b-HSD isoforms have been described, but only four appear to participate in the biosynthesis steps of gonadal steroid hormones in human beings (11, 4, 41). The present study suggests that one of these isoenzymes may have greater activity in the conversion of A to T in the ovary of patients with PCOS and may have greater activity still in those with obesity. Limitations of this study are the lack of androgen clearance measures in the obese and nonobese patients with PCOS and the lack of differentiation between ovarian and adrenal production of androgens. However, those aspects were beyond the objective of the study. Finally, it is possible that more time is required to see the complete stimulatory effect of GnRH and OGTT on T and A production. In conclusion, this study reveals a different pattern in A and T production, one modified by obesity in PCOS. These findings help understanding of the androgen values that are found in patients with PCOS. More research is required to clarify the action of isoforms of the 17b-HSD enzyme in patients with PCOS. REFERENCES 1. Azziz R, Woods KS, Reyna R, Key TJ, Knochenhauer ES, Yildiz BO. The prevalence and features of the polycystic ovary syndrome in an unselected population. J Clin Endocrinol Metab 24;89: Moran C, Tapia MC, Hernandez E, Vazquez G, Garcia-Hernandez E, Bermudez JA. Etiological review of hirsutism in 25 patients. Arch Med Res 1994;25: Azziz R, Sanchez LA, Knochenhauer ES, Moran C, Lazenby J, Stephens KC, et al. Androgen excess in women: experience with over 1 consecutive patients. J Clin Endocrinol Metab 24;89: The Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. Revised 23 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril 24;81: Romaguera J, Moran C, Diaz-Montes TP, Hines GA, Cruz RI, Azziz R. Prevalence of 21-hydroxylase-deficient nonclassic adrenal hyperplasia and insulin resistance among hirsute women from Puerto Rico. Fertil Steril 2;74: Lamb EJ, Dignam WJ, Pion RJ, Simmer HH. Plasma androgens in women. I. Normal and non-hirsute females oophorectomized and adrenalectomized patients. Acta Endocrinol (Copenh) 1964;45: Dignam WJ, Pion RJ, Lamb EJ, Simmer HH. Plasma androgens in women. II. Patients with polycystic ovaries and hirsutism. Acta Endocrinol (Copenh) 1964;45: Yen SSC, Vela P, Rankin J. Inappropriate secretion of follicle-stimulating hormone and luteinizing hormone in polycystic ovarian disease. J Clin Endocrinol 197;3: Burghen GA, Givens JR, Kitabchi AE. Correlation of hyperandrogenism with hyperinsulinism in polycystic ovarian disease. J Clin Endocrinol Metab 198;5: Gilling-Smith C, Willis DS, Beard RW, Franks S. Hypersecretion of androstenedione by isolated thecal cells from polycystic ovaries. J Clin Endocrinol Metab 1994;79: Payne AH, Hales DB. Overview of steroidogenic enzymes in the pathway from cholesterol to active steroid hormones. Endocr Rev 24;25: Erickson GF, Hsueh AJW, Quigley ME, Rebar RW, Yen SSC. Functional studies of aromatase activity in human granulosa cells from normal and polycystic ovaries. J Clin Endocrinol Metab 1979;49: Rosenfield RL, Barnes RB, Ehrmann DA. Studies of the nature of 17- hydroxyprogesterone hyperresponsiveness to gonadotropin-releasing hormone agonist challenge in functional ovarian hyperandrogenism. J Clin Endocrinol Metab 1994;79: Edman CD, MacDonald PC. Effect of obesity on conversion of plasma androstenedione to estrone in ovulatory and anovulatory young women. Am J Obstet Gynecol 1978;13: Dunaif A, Segal KR, Futterweit W, Dobrjansky A. Profound peripheral insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 1989;38: Carmina E, Koyama T, Chang L, Stanczyk FZ, Lobo RA. Does ethnicity influence the prevalence of adrenal hyperandrogenism and insulin resistance in polycystic ovary syndrome? Am J Obstet Gynecol 1992;167: Moran C, Garcia-Hernandez E, Barahona E, Gonzalez S, Bermudez JA. Relationship between insulin resistance and gonadotropin dissociation in obese and nonobese women with polycystic ovary syndrome. Fertil Steril 23;8: Barbieri RL, Makris A, Ryan KJ. Insulin stimulates androgen accumulation in incubations of human ovarian stroma and theca. Obstet Gynecol 1984;64(Suppl 3):73S Barbieri RL, Makris A, Randall RW, Daniels G, Kistner RW, Ryan KJ. Insulin stimulates androgen accumulation in incubations of ovarian stroma obtained from women with hyperandrogenism. J Clin Endocrinol Metab 1986;62: Nagamani M, Stuart CA, Van Dinh T. Steroid biosynthesis in the Sertoli- Leydig cell tumor: effects of insulin and luteinizing hormone. Am J Obstet Gynecol 1989;161: Stuart CA, Prince MJ, Peters EJ, Meyer WJ III. Hyperinsulinemia and hyperandrogenemia: in vivo androgen response to insulin infusion. Obstet Gynecol 1987;69: Dale PO, Tanbo T, Vaaler S, Abyholm T. Body weight, hyperinsulinemia, and gonadotropin levels in the polycystic ovarian syndrome: evidence of two distinct populations. Fertil Steril 1992;58: Fulghesu AM, Cucinelli F, Pavone V, Murgia F, Guido M, Caruso A, et al. Changes in luteinizing hormone and insulin secretion in polycystic ovarian syndrome. Hum Reprod 1999;14: Dunaif A, Graf M. Insulin administration alters gonadal steroid metabolism independent of changes in gonadotropin secretion in insulinresistant women with the polycystic ovary syndrome. J Clin Invest 1989;83: Dunaif A, Mandeli J, Fluhr H, Dobrjansky A. The impact of obesity and chronic hyperinsulinemia on gonadotropin release and gonadal steroid secretion in the polycystic ovary syndrome. J Clin Endocrinol Metab 1988;66: Kiddy DS, Sharp PS, White DM, Scanlon MF, Mason HD, Bray CS, et al. Differences in clinical and endocrine features between obese and non-obese subjects with polycystic ovary syndrome: an analysis of 263 consecutive cases. Clin Endocrinol 199;32: Singh KB, Mahajan DK, Wortsman J. Effect of obesity on the clinical and hormonal characteristics of the polycystic ovary syndrome. J Reprod Med 1994;39: Dos Reis RM, Foss MC, Dias de Moura M, Ferriani RA, Silva de Sa MF. Insulin secretion in obese and non-obese women with polycystic ovary syndrome and its relationship with hyperandrogenism. Gynecol Endocrinol 1995;9: Holte J, Bergh T, Berne C, Lithell H. Serum lipoprotein lipid profile in women with the polycystic ovary syndrome: relation to anthropometric, 2316 Moran et al. Androgen levels in PCOS and obesity Vol. 9, No. 6, December 28

8 endocrine and metabolic variables. Clin Endocrinol 1994;41: Acien P, Quereda F, Matallin P, Villarroya E, Lopez-Fernandez JA, Acien M, et al. Insulin, androgens, and obesity in women with and without polycystic ovary syndrome: a heterogeneous group of disorders. Fertil Steril 1999;72: National Institutes of Health Consensus Development Panel. Health implications of obesity. Ann Intern Med 1985;13: Moran C, Hernandez E, Ruiz JE, Fonseca ME, Bermudez JA, Zarate A. Upper body obesity and hyperinsulinemia are associated with anovulation. Gynecol Obstet Invest 1999;47: Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocr 1961;21: Hatch R, Rosenfield RL, Kim MH, Tredway D. Hirsutism: implications, etiology, and management. Am J Obstet Gynecol 1981;14: Pache TD, Wladimiroff JW, Hop WCJ, Fauser BCJM. How to discriminate between normal and polycystic ovaries: transvaginal US study. Radiology 1992;183: Lee-Benitez D, Leon C, Cervantes C, Parra A, Gonzalez-Diddi M, Aznar R, et al. Plasma concentration of androgens during the ovulatory menstrual cycle. Arch Invest Med (Mex) 1974;5: Barbieri RL. Human ovarian 17-ketosteroid oxidoreductase: unique characteristics of the granulosa-luteal cell and stromal enzyme. Am J Obstet Gynecol 1992;166: Bardin CW, Lipsett MB. Testosterone and androstenedione blood production rates in normal women and women with idiopathic hirsutism or polycystic ovaries. J Clin Invest 1967;46: Horton R, Neisler J. Plasma androgens in patients with the polycystic ovary syndrome. J Clin Endocr 1968;28: Barbieri RL, Gao X. Presence of 17 ß-hydroxysteroid dehydrogenase type 3 messenger ribonucleic acid transcript in an ovarian Sertoli-Leydig cell tumor. Fertil Steril 1997;68: Nelson VL, Qin KN, Rosenfield RL, Wood JR, Penning TM, Legro RS, et al. The biochemical basis for increased T production in theca cells propagated form patients with polycystic ovary syndrome. J Clin Endocrinol Metab 21;86: Fertility and Sterility â 2317