Pituitary desensitization to gonadotropinreleasing hormone increases abdominal adiposity in hyperandrogenic anovulatory women

Transcription

1 FERTILITY AND STERILITY VOL. 70, NO. 1, JULY 1998 Copyright 1998 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A. Pituitary desensitization to gonadotropinreleasing hormone increases abdominal adiposity in hyperandrogenic anovulatory women Received November 11, 1997; revised and accepted January 13, Supported by grants from the Mayo Foundation, Rochester, Minnesota, MO1-RR00585 from the General Clinical Research Centers, Division of Research Resources; and P51-RR00167 from the National Institutes of Health, Bethesda, Maryland; and by TAP Pharmaceuticals, Chicago, Illinois. Presented at the 53rd Annual Meeting of the American Society for Reproductive Medicine, Cincinnati, Ohio, October 18 22, Reprint requests and present address: Daniel A. Dumesic, M.D., Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, Mayo Clinic, Rochester, Minnesota (FAX: ; * Department of Obstetrics and Gynecology, Mayo Clinic. Wisconsin Regional Primate Research Center, University of Wisconsin. Department of Obstetrics and Gynecology, University of Wisconsin. Endocrinology- Reproductive Physiology Training Program, University of Wisconsin. Department of Radiology, Mayo Clinic. Department of Endocrinology, Mayo Clinic /98/$19.00 PII S (98) Daniel A. Dumesic, M.D., David H. Abbott, Ph.D., Joel R. Eisner, B.S., Rebekah R. Herrmann,* Judd E. Reed, Timothy J. Welch, M.D., and Michael D. Jensen, M.D. Mayo Clinic, Rochester, Minnesota, and University of Wisconsin, Madison, Wisconsin Objective: To determine whether hyperandrogenism in anovulatory women affects body fat distribution. Design: Prospective nonrandomized study. Setting: An academic research environment. Patient(s): Ten hyperandrogenic anovulatory patients and 10 healthy women matched by body mass index. Intervention(s): Regional body fat analysis was performed before and after 3 months of GnRH analogue (GnRH-a) therapy. Main Outcome Measure(s): Body fat distribution was measured by waist-to-hip circumference ratio, single-slice computed tomography imaging (L2-3 interspace), and total body dual-energy x-ray absorptiometry. Result(s): Weight, body mass index, waist-to-hip circumference ratio, total body and leg fat mass, and subcutaneous adipose area were unaffected by the presence of hyperandrogenism or the use of GnRH-a therapy. Basal abdominal fat mass, abdomen-to-leg fat mass ratio, visceral adipose area, and total visceral adipose volume were comparable in both study groups. The abdominal fat mass increased in both groups during GnRH-a therapy, whereas the abdomen-to-leg fat mass ratio rose significantly only in the hyperandrogenic patients. During GnRH-a therapy, the hyperandrogenic patients demonstrated a significant increase in visceral adipose area compared with the healthy women so that total visceral adipose volume increased significantly in the former but not the latter. Conclusion(s): Three months of GnRH-a administration preferentially increased abdominal fat, as measured by singleslice computed tomography imaging and total body dual-energy x-ray absorptiometry, in hyperandrogenic anovulatory women. (Fertil Steril 1998;70: by American Society for Reproductive Medicine.) Key Words: Hyperandrogenic anovulation, androgens, body fat distribution In men, androgens stimulate abdominal lipolysis, a process in which triglyceride is hydrolyzed to fatty acids and glycerol (1). A reciprocal relation exists between serum testosterone concentrations and the amount of intraabdominal (visceral) adipose (2). Moreover, administration of testosterone to men enhances catecholamine-induced lipolysis, suppresses triglyceride uptake, and increases triglyceride turnover in SC abdominal adipose (3, 4). Administration of testosterone to men also decreases triglyceride uptake in visceral adipose and reduces visceral adipose stores, as determined by computed tomography (CT) imaging (5, 6). The decrease in visceral adipose induced by testosterone is unaccompanied by changes in body mass index (BMI) or waist-to-hip circumference ratio measurements (6). Although sex differences in adipose metabolism undoubtedly exist, abdominal adiposity also is a common feature of women with hyperandrogenic anovulation (7). Therefore, an intriguing question is whether the underlying androgenic environment in these women affects their body fat distribution. Because circulating testosterone levels in hyperandrogenic anovulatory women correlate positively with catecholamine-induced lipolysis and negatively with triglyceride uptake in SC abdominal adipose, ovarian hyperandrogenism in these women appears to counteract abdominal adi- 94

2 posity through enhanced abdominal lipolysis, decreased abdominal triglyceride uptake, or both (7). However, suppression of androgen action in these women by oral contraceptives or flutamide does not alter their waist-to-hip circumference ratio (7, 8). Because anthropometric measurements may not detect radiographic changes in visceral adipose mass (9), the present study examined whether the administration of a GnRH analogue (GnRH-a) for 3 months to women with hyperandrogenic anovulation would increase abdominal adiposity as measured by single-slice CT imaging of the L2-3 vertebral interspace and total body dual-energy x-ray absorptiometry. The results demonstrate that 3 months of GnRH-a administration causes a preferential increase in abdominal fat and total visceral adipose volume in hyperandrogenic anovulatory patients but not in healthy women. MATERIALS AND METHODS Study Subjects Informed consent was obtained from 20 white volunteers, consisting of 10 hyperandrogenic anovulatory patients and 10 healthy women. Body fat distribution in all participants was determined by waist-to-hip circumference ratio, singleslice CT imaging of the L2-3 vertebral interspace, and total body dual-energy x-ray absorptiometry measurement at baseline and again after 3 months of GnRH-a therapy with leuprolide acetate depot (TAP Pharmaceuticals, Chicago, IL) at a dosage of 7.5 mg monthly IM. This dosage of GnRH-a was chosen because it previously has been shown to induce complete pituitary desensitization to GnRH in obese hyperandrogenic anovulatory women (10). All participants ranged in age from years and were in good health. They had no history of thyroid disease or medication use known to affect reproductive function for at least 3 months before the study. Subjects were instructed to maintain their regular patterns of exercise and dietary intake. Hyperandrogenic anovulation was diagnosed by the finding of hirsutism or biochemical hyperandrogenism and chronic anovulation when specific ovarian, adrenal, and pituitary disorders were excluded. Biochemical hyperandrogenism was defined as an elevation in serum testosterone, non sex hormone-binding globulin-bound testosterone, or dihydrotestosterone that was 2 SD above the mean value for the normal female population, excluding hyperprolactinemia and late-onset 21-hydroxylase deficiency, as shown by either a basal serum 17-hydroxyprogesterone level of 2 ng/ml or a normal 1-hour ACTH stimulation test result (11). Chronic anovulation was defined as amenorrhea of 3 months duration (n 7 patients) or oligomenorrhea (e.g., intermenstrual intervals of 35 days) with adequately timed serum progesterone levels of 3 ng/ml (n 3 patients). All the healthy women had regular menstrual cycles every days with ovulation confirmed by luteal phase progesterone levels over a 2-month interval. They had no history of hirsutism or galactorrhea. Each healthy woman was individually matched to a single hyperandrogenic patient according to a BMI within 5 kg/m 2. Because the primary intention was to recruit healthy women who were closely matched in BMI to hyperandrogenic patients, we accepted a group of healthy women that was older in age ( years) than the group of hyperandrogenic patients ( years) (P 0.05 by paired t-test). The study was approved by the Mayo Clinic Institutional Review Board, and written consent was obtained from each participant. Radiographic Imaging Computed tomography was performed with the use of an Imatron C-150 scanner (Imatron Inc., San Francisco, CA) as previously described (12). Scanning was performed at 130 kv of exposure with a 0.4-second scan time and a 6-mm thickness. A single CT image was obtained at the L2-3 vertebral interspace and then transferred to a Sun workstation (Sun Microsystems, Inc., Mountainview, CA) for analysis. Each CT image was examined with image processing software to assess the Hounsfield units most representative of adipose. The lower limit of adipose density was Hounsfield units, and the upper limit was 27 2 Hounsfield units. To assess the area of visceral adipose, each abdominal CT image was edited by erasing the image exterior to the innermost abdominal wall muscles with a mousedriven cursor. For each participant, the visceral (edited set of image files) adipose area was subtracted from the total abdominal (first set of image files) adipose area to predict the SC abdominal adipose area. The visceral adipose area was divided by the total abdominal adipose area to determine the fraction of abdominal adipose at the L2-3 interspace that was located within the abdominal cavity (i.e., the percentage of intra-abdominal adipose). Total body dual-energy x-ray absorptiometry (Lunar Radiation Corp., Madison, WI) was used to obtain regional tissue composition analysis (software version 3.4); the 20- minute scan time was chosen (12). Each participant was positioned with her arms sufficiently separated from her trunk to allow the creation of a region of interest that included only truncal contents. The abdominal region was defined as that area between the dome of the diaphragm (cephalad limit) and the top of the greater trochanter (caudal limit). The anatomic landmark of the superior edge of the greater trochanter was chosen because it is readily recognized on the total body dual-energy x-ray absorptiometry scan image and only minute amounts of visceral fat are present caudal to this level ( 1% 2% of visceral fat). The leg region was defined as that area below the top of the greater trochanter (cephalad limit). Total body dual-energy x-ray absorptiometry images were saved to determine the fat-free body mass and the FERTILITY & STERILITY 95

3 amounts of total body, abdomen, and leg fat. Total body dual-energy x-ray absorptiometry images also were used to calculate the ratio of abdomen-to-leg fat mass and the percentage of total body fat. The percentage of intra-abdominal adipose as determined by the L2-3 interspace CT slice was multiplied by the amount of total abdominal fat as measured by total body dual-energy x-ray absorptiometry to predict the total visceral adipose volume, as previously described (12). Because total body dual-energy x-ray absorptiometry measures fat mass and CT provides volume data, total body dual-energy x-ray absorptiometry fat mass values were converted to adipose volume using the average density of triglyceride (0.900 g/ml). Anthropometric Analysis The waist-to-hip circumference ratio was calculated from the following circumferences measured in the supine position: waist (i.e., widest girth at the level of the umbilicus); hip (i.e., largest girth over the greater trochanters). Eight hyperandrogenic patients and seven healthy women had upper body obesity, as defined by a waist-to-hip circumference ratio of 0.85 (13). Hormone Assays Blood sampling initially was performed during the midfollicular phase (days 5 10 of the menstrual cycle) in the healthy women and during a period of documented anovulation, as determined by serum progesterone levels, in the hyperandrogenic patients. Blood sampling was repeated after 3 months of GnRH-a therapy. Blood samples were used to measure LH, FSH, testosterone, unbound testosterone, dihydrotestosterone, androstenedione, DHEAS, estrone, E 2, and unbound E 2. Radioimmunoassay measurement of serum testosterone, unbound testosterone, dihydrotestosterone, E 2, unbound E 2, and estrone was performed by Corning Nichols Institute (San Juan Capistrano, CA). The intra-assay coefficients of variation (CVs) were as follows: testosterone, 9.0%; unbound testosterone, 5.0%; dihydrotestosterone, 8.0%; E 2, 9.4%; unbound E 2, 4.0%; and estrone, 9.2%. The interassay CVs were as follows: testosterone, 13.0%; unbound testosterone, 7.8%; dihydrotestosterone, 12.0%; E 2, 13.4%; unbound E 2, 5.6%; and estrone, 14.0%. Luteinizing hormone, FSH, and androstenedione were measured by RIA in the Immunochemical Core Laboratory of the Mayo General Clinical Research Center. The intraassay CVs were as follows: LH, 4.3%; FSH, 3.6%; and androstenedione, 4.3%. The interassay CVs were as follows: LH, 5.4%; FSH, 5.1%; and androstenedione, 6.0%. Dehydroepiandrosterone was measured by RIA in the Endocrine Laboratory of the Mayo Clinic. The intra-assay and interassay CVs for DHEAS were 4.7% and 8.3%, respectively. Statistical Analysis Hormonal, radiographic, and anthropometric variables were compared by two-way analysis of variance (ANOVA) with repeated measures using the study group (healthy versus hyperandrogenic women) and GnRH-a treatment (before versus after) as factors to determine the independent effects of these variables and their possible interactions. When significant statistical interactions were present by ANOVA, post hoc univariate analysis was performed on the variables (Systat, Macintosh Version 5.2; Evanston, IL). Changes in visceral adipose area (L2-3 vertebral interspace) and total visceral adipose volume during GnRH-a treatment were compared in the hyperandrogenic patients and the healthy women with the use of a paired t-test. Log and arc sin transformation of the data were performed to achieve homogeneity of variance and to increase linearity (14). All data are expressed as antilog of the transformed mean and 95% confidence intervals. RESULTS Hormone Measurements There were statistically significant interactions between the study group and GnRH-a treatment for serum testosterone (P 0.01), unbound testosterone (P 0.01), and dihydrotestosterone (P 0.05) levels. Basal serum testosterone, unbound testosterone, and dihydrotestosterone levels were significantly greater in the hyperandrogenic patients than in the healthy women, who consistently had normal serum androgen values (study group effects: testosterone, P 0.01; unbound testosterone, P 0.01; dihydrotestosterone, P 0.025) (Table 1). Three months of GnRH-a administration to both study groups significantly lowered serum testosterone, unbound testosterone, and dihydrotestosterone concentrations to comparable levels, with a greater decline in androgen secretion seen in the hyperandrogenic patients than in the healthy women. Serum androstenedione concentrations tended to be higher in the hyperandrogenic patients than in the healthy women (study group effect: P 0.07). Because there was no statistically significant interaction between the study group and GnRH-a treatment (P 0.06), the data for serum androstenedione concentrations for both study groups were combined for GnRH-a treatment analysis. Serum androstenedione levels decreased significantly in both study groups during GnRH-a treatment (GnRH-a treatment effect: P 0.025). In contrast, serum DHEAS levels were unaffected by the study group or GnRH-a treatment (study group GnRH-a treatment interaction: P 0.9). There were no statistically significant interactions between the study group and GnRH-a treatment for serum E 2 (P 0.7), unbound E 2 (P 0.7), or estrone (P 0.1) levels. Basal serum E 2, unbound E 2, and estrone levels were similar in both study groups and decreased significantly in all 96 Dumesic et al. Body fat and hyperandrogenic anovulation Vol. 70, No. 1, July 1998

4 TABLE 1 Serum hormone levels in hyperandrogenic patients and healthy women before and after 3 months of GnRH-a therapy. Hyperandrogenic patients Healthy women Hormone Before therapy After therapy Before therapy After therapy Testosterone (ng/dl) 80.7 ( )* 21.9 ( ) 28.3 ( ) 16.1 ( ) Unbound testosterone (pg/ml) 15.7 ( )* 4.0 ( ) 4.6 ( ) 3.0 ( ) Dihydrotestosterone (ng/dl) 18.1 ( ) 8.2 ( ) 8.3 ( ) 5.1 ( ) Androstenedione (ng/ml) 3.0 ( ) 1.7 ( ) 1.5 ( ) 1.4 ( ) DHEAS ( g/ml) 1.8 ( ) 1.7 ( ) 1.5 ( ) 1.4 ( ) Estrone (pg/ml) 85.3 ( ) 25.8 ( ) 47.5 ( ) 26.8 ( ) E 2 (pg/ml) 69.8 ( ) 22.2 ( ) 62.4 ( ) 22.7 ( ) Unbound E 2 (pg/ml) 1.4 ( ) 0.4 ( ) 1.2 ( ) 0.5 ( ) LH (IU/L) 7.0 ( ) 0.5 ( ) 4.6 ( ) 0.3 ( ) FSH (IU/L) 4.3 ( ) 2.5 ( ) 4.4 ( ) 3.1 ( ) LH/FSH 1.6 ( ) 0.2 ( ) 1.0 ( ) 0.1 ( ) Note: All values represent antilog of the transformed mean (95% CI). * P versus healthy women before GnRH-a therapy, by ANOVA. P versus healthy women before GnRH-a therapy, by ANOVA. P 0.01 versus the same women before GnRH-a therapy, by ANOVA. P before versus after GnRH-a therapy (all women combined), by ANOVA. P 0.01 before versus after GnRH-a therapy (all women combined), by ANOVA. P 0.05 healthy versus hyperandrogenic women (before and after GnRH-a therapy combined), by ANOVA. women combined during GnRH-a treatment (GnRH-a treatment effects: E 2, P 0.01; unbound E 2, P 0.01; estrone, P 0.01). Serum LH and FSH levels were similar in both study groups (study group effects: LH, P 0.1; FSH, P 0.5). Both serum gonadotropin concentrations decreased significantly in all women combined during GnRH-a treatment (GnRH-a treatment effects: LH, P 0.01; FSH, P 0.01 [study group GnRH-a treatment interactions: LH, P 0.9; FSH, P 0.4]). Serum LH-to-FSH ratios were significantly higher in the hyperandrogenic patients than in the healthy women (study group effect, P 0.05) and decreased significantly in both groups during GnRH-a treatment (GnRH-a treatment effect: P 0.01; study group GnRH-a treatment interaction: P 0.5). Body Fat Measurements There were no statistically significant interactions between the study group and GnRH-a treatment for weight (P 0.9), BMI (P 0.7), waist-to-hip circumference ratio (P 0.7), or total body fat mass as measured by total body dual-energy x-ray absorptiometry (P 0.3); none of these parameters were affected by the study group (weight, P 0.5; BMI, P 1.0; waist-to-hip circumference ratio, P 0.7; total body fat mass, P 0.8). Basal weight, BMI, waist-to-hip circumference ratio, and total body fat mass as measured by total body dual-energy x-ray absorptiometry were comparable in the healthy women and the hyperandrogenic patients and were unchanged in all women combined during GnRH-a treatment (GnRH-a treatment effects: weight, P 0.4; BMI, P 0.5; waist-to-hip circumference ratio, P 0.1; total body fat mass, P 0.2) (Table 2). In contrast, there were statistically significant interactions between the study group and GnRH-a treatment for the basal percentage of total body fat and fat-free body mass as measured by total body dual-energy x-ray absorptiometry TABLE 2 Weight, body mass index, waist-to-hip circumference ratio, and total body fat mass as measured by total body dual-energy x-ray absorptiometry in hyperandrogenic patients and healthy women before and after 3 months of GnRH-a therapy. Hyperandrogenic patients Healthy women Parameter Before therapy After therapy Before therapy After therapy Weight (kg) 94.8 ( ) 95.5 ( ) 90.0 ( ) 90.6 ( ) Body mass index (kg/m 2 ) 33.7 ( ) 33.8 ( ) 33.5 ( ) 33.8 ( ) Waist-to-hip circumference ratio 0.91 ( ) 0.90 ( ) 0.90 ( ) 0.88 ( ) Total fat mass (kg) 35.7 ( ) 37.3 ( ) 34.8 ( ) 34.9 ( ) Note: All values represent antilog of the transformed mean (95% CI). FERTILITY & STERILITY 97

5 FIGURE 1 Regional body fat in hyperandrogenic patients and healthy women before and after 3 months of GnRH-a therapy. Values are antilog of the transformed mean and 95% CI. * P 0.05 versus all the women before GnRH-a therapy; P 0.05; P versus the same women before GnRH-a therapy; by ANOVA. The solid bar represents the hyperandrogenic patients; the open bar represents the healthy women. (P 0.05) (P 0.025). Although the basal percentage of total body fat and fat-free body as measured by total body dualenergy x-ray absorptiometry mass were not significantly different in the healthy women (percentage of total body fat, 46% [range, 44% 49%]; fat-free body mass, 44.3 kg [range, kg]) and in the hyperandrogenic patients (percentage of total body fat, 45% [range, 43% 47%], P 0.7; fat-free body mass, 48.5 kg [range, kg], P 0.06 versus healthy women), the hyperandrogenic patients who received GnHR-a therapy had an increase in the basal percentage of total body fat (47% [range, 45% 50%]; P 0.05 versus before GnRH-a therapy) without a change in fat-free body mass (46.7 kg [range, kg], P 0.1 versus before GnRH-a therapy), whereas the healthy women who received GnRH-a therapy did not (percentage of total body fat, 46% [range, 44% 49%], P 0.6; fat-free body mass, 46.0 kg [range, kg], P 0.1 versus before GnRH-a therapy). Abdominal fat mass as measured by total body dualenergy x-ray absorptiometry was similar in both groups of women (study group GnRH-a treatment interaction: P 0.2) (Fig. 1). Both study groups demonstrated a statistically significant increase in abdominal fat mass during GnRH-a treatment (P 0.05). Although the 2 groups had similar ratios of abdomen-to-leg fat mass before GnRH-a treatment, the hyperandrogenic patients showed a statistically significant increase in the ratio of abdomen-to-leg fat mass during GnRH-a therapy (GnRH-a treatment effect: P 0.025, study group GnRH-a treatment interaction: P 0.01), whereas the healthy women did not (GnRH-a treatment effect: P 0.5). Leg fat mass was comparable in both study groups and did not change during the 3 months of GnRH-a therapy (study group GnRH-a treatment interaction: P 0.8). The amount of SC adipose at the L2-3 vertebral interspace was unaffected by study group GnRH-a treatment (study group GnRH-a treatment interaction: P 0.9) (Fig. 1). The basal percentage of intra-abdominal adipose was comparable in the hyperandrogenic patients (26% [range, 23% 27%]) and in the healthy women (31% [range, 29% 34%]) (P 0.3) and was not significantly altered in either study group by GnRH-a treatment (hyperandrogenic patients: 27% [range, 25% 30%], P 0.3; healthy women: 28% [range, 26% 31%], P 0.07 versus before GnRH-a treatment). Nevertheless, there was a statistically significant interaction between the study group and GnRH-a treatment 98 Dumesic et al. Body fat and hyperandrogenic anovulation Vol. 70, No. 1, July 1998

6 FIGURE 2 Changes in [1] visceral adipose area at the L2-3 vertebral interspace and [2] total visceral adipose volume in hyperandrogenic patients and healthy women after 3 months of GnRH-a treatment. Values are antilog of the transformed mean and 95% CI. The visceral adipose area and total visceral adipose volume increased significantly with GnRH-a treatment in the hyperandrogenic patients compared with the healthy women (paired t-test). The solid bar represents the hyperandrogenic patients; the open bar represents the healthy women. for visceral adipose area at the L2-3 vertebral interspace (study group GnRH-a treatment interaction: P 0.05). Although it was similar in the 2 study groups before GnRH-a treatment, the visceral adipose area at the L2-3 vertebral interspace increased only in the hyperandrogenic patients after GnRH-a treatment; they demonstrated a statistically significant increase in visceral adipose area compared with the healthy similarly-treated women (P 0.05) (Figs. 1 and 2). Visceral adipose area at the L2-3 vertebral interspace rose in the hyperandrogenic patients by 21.1% (95% confidence interval, 1% 44%) but decreased in the healthy women by 4.6% (95% confidence interval, 14.4% 5.4%) (Fig. 3). Consequently, the basal total visceral adipose volume was comparable in both groups of women and increased during GnRH-a treatment in the hyperandrogenic patients (GnRH-a treatment effect: P 0.05) (Figs. 1 and 2) but not in the healthy women (GnRH-a treatment effect: P 0.1, study group GnRH-a treatment interaction: P 0.025). DISCUSSION Body fat distribution in humans is governed by a complicated balance between the mobilization and storage of regional adipose. As an essential component of regional adipose metabolism, lipolysis refers to an enzyme-mediated process in which triglyceride is hydrolyzed to fatty acids and glycerol (1). Triglyceride storage is increased by lipoprotein lipase, a capillary-bound enzyme that hydrolyzes circulating triglyceride before the uptake of fatty acids into adipocytes. Conversely, triglyceride breakdown from adipose is promoted by intracellular hormone-sensitive lipase, which in turn is stimulated by -adrenergic hormones and inhibited by insulin and 2 -adrenergic hormones (1, 15). These enzymes are regulated by sex steroids so that body fat distribution normally exhibits a sexual dimorphism. Gluteofemoral adiposity typically is present in women, whereas abdominal adiposity, a risk factor for cardiovascular disease and non insulin-dependent diabetes mellitus, commonly is observed in men (13). In men, abdominal adipose is a target tissue for androgen action. Elderly men with declining androgen production exhibit an increase in abdominal adipose mass (15). In middle-aged men, low plasma free testosterone concentrations accompany high visceral fat mass (2), whereas testosterone administration diminishes visceral adipose triglyceride uptake and decreases visceral adipose mass (5, 6). Testosterone therapy in such men also inhibits lipoprotein lipase activity and stimulates catecholamine-induced lipolysis in SC abdominal adipose, increasing triglyceride turnover (3, 4). Taken together, these studies in men show that androgens regulate the amount of abdominal adipose. The present study strongly suggests that androgens also regulate abdominal adipose stores in women with hyperandrogenic anovulation, an endocrinopathy characterized by abdominal obesity, ovarian hyperandrogenism, and insulin resistance (13). Three months of GnRH-a administration caused a preferential increase in abdominal fat in the hyperandrogenic patients but not in the healthy women, as evidenced by an increased ratio of abdomen-to-leg fat mass. Consistent with this finding, circulating bioactive testosterone levels in hyperandrogenic women correlate negatively with lipoprotein lipase activity and positively with catecholamine-stimulated lipolysis in SC abdominal adipocytes (7). Further, androgen excess in such women probably stimulates visceral lipolysis because total visceral adipose volume increased when ovarian hyperandrogenism was suppressed in this study. The latter observation is consistent FERTILITY & STERILITY 99

7 FIGURE 3 Visceral adipose area at the L2-3 vertebral interspace in a single healthy woman (BMI, 34.9 kg/m 2 ) and a matched hyperandrogenic patient (BMI, 31.8 kg/m 2 ) before and after 3 months of GnRH-a treatment. The visceral adipose area decreased by 9% (from to cm 2 ) with GnRH-a treatment in the healthy woman but increased by 52% (from 72.4 to cm 2 ) in the hyperandrogenic patient. with the finding that testosterone stimulates lipolysis in rat adipose precursor cells by increasing -adrenoceptor binding and postreceptor signal transduction (16). Hyperandrogenic women who received GnRH-a therapy for 3 months experienced an increase in total visceral adipose volume without a concomitant change in the waist-tohip circumference ratio. However, there are several factors responsible for the inability of the waist-to-hip circumference ratio to predict changes in the amount of visceral adipose. The waist-to-hip circumference ratio is influenced by SC abdominal and visceral adipose, bone structure, and muscle volume (9, 15). It also is affected by abdominal muscle tone, respiratory movement, and patient position. Although it is a simple and inexpensive method of quantifying adipose distribution in large population studies, the waist-to-hip circumference ratio provides only a rough estimate of regional adipose stores. Sophisticated radiographic imaging techniques, highly accurate for body fat analysis, are necessary for detailed investigation of adipose metabolism and its regulation by sex steroids. Single-slice CT imaging failed to confirm previous anthropometric findings that SC abdominal adipose mass is greater in hyperandrogenic patients than in healthy women (17). Further, single-slice CT imaging showed a slight increase in SC abdominal adipose area in GnRH-a treated hyperandrogenic women, with an elevation in total visceral adipose volume. This parallel increase in SC abdominal area and total visceral adipose volume observed in GnRH-a treated hyperandrogenic patients contrasts with the ability of testosterone administration in men to increase lipid uptake into SC abdominal adipose while decreasing lipid uptake into visceral adipose (5). Although single-slice CT imaging may not be as accurate in quantifying SC abdominal adipose as serial CT imaging (which scans multiple abdominal sections and then sums the volume of adipose within each section), we used single-slice CT imaging in this study because it reduced cost, scanning time, and radiation exposure (12). Further, single-slice CT imaging of the midabdomen in obese humans has been shown to predict SC abdominal adipose volume, as measured by serial CT imaging (18). A key question is why sex differences exist in the relation between circulating androgen levels and abdominal adiposity. Unlike men, in whom serum testosterone levels are negatively correlated with abdominal adiposity (2), androgen excess in anovulatory women accompanies increased abdominal obesity and insulin resistance (13). The insulin resistance in obese hyperandrogenic anovulatory women appears to be a primary defect because inhibition of androgen action by oral contraceptives, anti-androgens, or GnRH-a therapy does not completely reverse insulin resistance (7, 8), whereas improvement of hyperinsulinemia with an insulinsensitizing agent lowers serum androgen levels (19). 100 Dumesic et al. Body fat and hyperandrogenic anovulation Vol. 70, No. 1, July 1998

8 In hyperandrogenic anovulatory women, therefore, hyperinsulinemia due to insulin resistance or some other factor may predispose to abdominal obesity, with ovarian hyperandrogenism counteracting this type of adiposity through enhanced abdominal lipolysis, decreased abdominal triglyceride uptake, or both (20). Such a pathophysiologic adaptation to hormonal, nutritional, or other factors agrees with the findings of diminished 2 -adrenoceptor density and reduced cyclic adenosine monophosphate dependent hormone-sensitive lipase activation in adipocytes from hyperandrogenic anovulatory women (21). The present study was not designed to determine whether androgens and estrogens differentially regulate body fat distribution in women. Unlike androgens, estrogens influence gluteofemoral adiposity, as evidenced by preferential gluteofemoral fat deposition with the pubertal onset of ovarian function. Lipoprotein lipase activity in reproductive-aged women is greater in gluteofemoral adipose than abdominal adipose and increases further with pregnancy (7, 22). Later, lipoprotein lipase activity of gluteofemoral adipose decreases with menopause but returns to premenopausal levels with hormone replacement therapy (23, 24). Gluteofemoral adipose lipoprotein lipase activity also is progesterone-dependent and increases when progesterone is applied percutaneously to the gluteofemoral region of premenopausal women (25). Although GnRH-a induced decreases in estrogen and progesterone secretion may favor abdominal adipose deposition by diminishing gluteofemoral adipose lipoprotein lipase activity, basal estrone and E 2 secretion was similar in the hyperandrogenic patients and the healthy women in our study and decreased to the same degree in both study groups during GnRH-a treatment. Nevertheless, the ratio of abdomen-to-leg adipose mass increased with GnRH-a treatment in the hyperandrogenic women but not in the healthy women, despite their reduced estrogen and progesterone secretion. Further, gluteofemoral lipoprotein lipase activity in hyperandrogenic anovulatory women has been shown to be unaffected by the administration of progestin-containing oral contraceptives (7). Experimental protocols that eliminate ovarian hyperandrogenism while maintaining the circulating estrogenic milieu are necessary to investigate the differential regulation of body fat distribution by androgens and estrogens in hyperandrogenic anovulatory women. In conclusion, 3 months of GnRH-a administration caused a preferential increase in abdominal fat and total visceral adipose volume in hyperandrogenic anovulatory patients but not in healthy women, as measured by single-slice CT imaging of the L2-3 vertebral interspace and total body dual-energy x-ray absorptiometry. These findings strongly suggest that androgen excess in anovulatory women enhances abdominal lipolysis, decreases abdominal triglyceride uptake, or both. References 1. Coppack SW, Jensen MD, Miles JM. In vivo regulation of lipolysis in humans. J Lipid Res 1994;35: Seidell JC, Bjorntorp P, Sjostrom L, Kvist H, Sannerstedt R. Visceral fat accumulation in men is positively associated with insulin, glucose, and C-peptide levels, but negatively with testosterone levels. Metabolism 1990;39: Rebuffe-Scrive M, Marin P, Bjorntorp P. Effect of testosterone on abdominal adipose tissue in men. Int J Obes 1991;15: Marin P, Oden B, Bjorntorp P. Assimilation and mobilization of triglycerides in subcutaneous abdominal and femoral adipose tissue in vivo in men: effects of androgens. J Clin Endocrinol Metab 1995;80: Marin P, Lonn L, Andersson B, Oden B, Olbe L, Bengtsson BA, et al. Assimilation of triglycerides in subcutaneous and intraabdominal adipose tissues in vivo in men: effects of testosterone. J Clin Endocrinol Metab 1996;81: Marin P, Holmang S, Gustafsson C, Jonsson L, Kvist H, Elander A. Androgen treatment of abdominally obese men. Obes Res 1993;1: Rebuffe-Scrive M, Cullberg G, Lundberg PA, Lindstedt G, Bjorntorp P. Anthropometric variables and metabolism in polycystic ovary disease. Horm Metab Res 1989;21: Diamanti-Kandarakis E, Mitrakou A, Hennes MMI, Platanissiotis D, Kaklas N, Spina J, et al. Insulin sensitivity and antiandrogen therapy in women with polycystic ovary syndrome. Metabolism 1995;44: Van der Kooy K, Leenen R, Seidell JC, Deurenberg P, Droop A, Bakker CJG. Waist-hip ratio is a poor predictor of changes in visceral fat. Am J Clin Nutr 1993;57: Dumesic DA, Matteri RL. Obesity affects circulating estradiol levels of premenopausal hirsute women receiving leuprolide acetate depot. Int J Fertil 1993;38: Dewailly D, Vantyghem-Haudiquet MC, Sainsard C, Buvat J, Cappoen JP, Ardaens K, et al. Clinical and biological phenotypes in late-onset 21-hydroxylase deficiency. J Clin Endocrinol Metab 1986;63: Jensen MD, Kanaley JA, Reed JE, Sheedy PF. Measurement of abdominal and visceral fat with computed tomography and dual energy x-ray absorptiometry. Am J Clin Nutr 1995;61: Wild RA. Hyperandrogenism: implications for cardiovascular, endometrial, and breast disease. In: Adashi EY, Rock JA, Rosenwaks Z, editors. Reproductive endocrinology, surgery, and technology. Philadelphia: Lippincott-Raven Publishers, 1996: Sokal RR, Rohlf FJ. Biometry, the principles and practice of statistics in biological research. 3rd ed. New York: WH Freeman and Co., Abate N, Garg A. Heterogeneity in adipose tissue metabolism: causes, implications and management of regional adiposity. Prog Lipid Res 1995;34: Xu X, De Pergola G, Bjorntorp P. The effects of androgens on the regulation of lipolysis in adipose precursor cells. Endocrinology 1990; 126: Holte J, Bergh T, Berne C, Berglund L, Lithell H. Enhanced early insulin response to glucose in relation to insulin resistance in women with polycystic ovary syndrome and normal insulin tolerance. J Clin Endocrinol Metab 1994;78: Tokunaga K, Matsuzawa Y, Ishikawa K, Tarui S. A novel technique for the determination of body fat by computed tomography. Int J Obes 1983;7: Dunaif A, Scott D, Finegood D, Quintana B, Whitcomb R. The insulin-sensitizing agent troglitazone improves metabolic and reproductive abnormalities in the polycystic ovary syndrome. J Clin Endocrinol Metab 1996;81: Reynisdottir S, Ellerfeldt K, Wahrenberg H, Lithell H, Arner P. Multiple lipolysis defects in the insulin resistance (metabolic) syndrome. J Clin Invest 1994;93: Ek I, Arner P, Bergqvist A, Carlstrom K, Wahrenberg H. Impaired adipocyte lipolysis in nonobese women with the polycystic ovary syndrome: a possible link to insulin resistance. J Clin Endocrinol Metab 1997;82: Rebuffe-Scrive M, Enk L, Crona N, Lonnroth P, Abrahamsson L, Smith U, et al. Fat cell metabolism in different regions in women. Effects of menstrual cycle, pregnancy and lactation. J Clin Invest 1985;75: Rebuffe-Scrive M, Eldh J, Hafstrom L-O, Bjorntorp P. Metabolism of mammary, abdominal and femoral adipocytes in women before and after menopause. Metabolism 1986;35: Rebuffe-Scrive M, Lonnroth P, Marin P, Wesslau C, Bjorntorp P, Smith U. Regional adipose tissue metabolism in men and postmenopausal women. Int J Obes 1987;11: Rebuffe-Scrive M, Basdevant A, Guy-Grand B. Effect of local application of progesterone on human adipose tissue lipoprotein lipase. Horm Metab Res 1983;15:566. FERTILITY & STERILITY 101