Comparison of various simple insulin sensitivity and -cell function indices in lean hyperandrogenemic and normoandrogenemic young hirsute women

Transcription

1 FERTILITY AND STERILITY VOL. 80, NO. 1, JULY 2003 Copyright 2003 American Society for Reproductive Medicine Published by Elsevier Inc. Printed on acid-free paper in U.S.A. Comparison of various simple insulin sensitivity and -cell function indices in lean hyperandrogenemic and normoandrogenemic young hirsute women Yuksel Altuntas, M.D., a Muammer Bilir, M.D., b Bulent Ozturk, M.D., a and Sadi Gundogdu, M.D. b Internal Medicine Clinic, Sisli Etfal Education and Research Hospital, and Cerrahpasa Medical Faculty, Istanbul University, Istanbul, Turkey Received July 24, 2002; revised and accepted February 4, Reprint requests: Yuksel Altuntas, M.D., Haznedar mah, İlkyuva cad, Meriç sokak 26/5, Güngören, Istanbul, Turkey (FAX: ; yukselaltuntas@yahoo.com). a Department of Endocrinology and Metabolism, Internal Medicine Clinic, Sisli Etfal Education and Research Hospital. b Department of Endocrinology and Metabolism, Cerrahpasa Medical Faculty, Istanbul University /03/$30.00 doi: /s (03) Objective: To assess and compare various simple insulin sensitivity and -cell function indices in lean, hirsute, young women. Design: Prospective study. Setting: Departments of endocrinology and metabolism at a university and a state hospital. Patient(s): Seventy-one hirsute young women were classified as hyperandrogenemic or normoandrogenemic. Main Outcome Measure(s): Insulin sensitivity and -cell function indices derived from a single sample and an oral glucose tolerance test (OGTT). Result(s): Lean hyperandrogenemic hirsute women have insulin resistance and increased -cell function. The most sensitive indices of insulin resistance were total and 1-hour and 2-hour post-challenge insulin levels during OGTT. When a cut-off value of 3.2 or greater for homeostasis model assessment of insulin resistance (HOMA-IR) was accepted, 46% of hyperandrogenemic women and 30% of normoandrogenemic women were insulin resistant. Fasting insulin level was best correlated with the fasting insulin resistance index, HOMA-IR, and Quicky index. The HOMA-IR was best correlated with fasting insulin level and the hepatic insulin sensitivity index (ISI HOMA ). Conclusion(s): Insulin levels based on OGTT are the most useful index of insulin resistance and -cell function index in hirsute women. The HOMA-IR may be a proposed global test for insulin resistance; it correlated well with both OGTT-derived insulin resistance and -cell function indices and with global insulin resistance indices derived from a single sample (such as ISI HOMA, Quicky index, FIRI 1, fasting Belfiore index, and glucose/insulin ratio). (Fertil Steril 2003;80: by American Society for Reproductive Medicine.) Key Words: Insulin resistance, -cell function, normoandrogenemia, hyperandrogenemia, hirsutism Hirsutism is the presence of excess hair growth in women. It results from an increased blood androgen level or increased skin sensitivity to a normal androgen level (1). Because skin is relatively insensitive to androgen, however, the latter condition is called cryptic hyperandrogenemia (hyperandrogenism without skin manifestations). The most common causes of hirsutism are the polycystic ovary syndrome (PCOS) or idiopathic hirsutism. Women with PCOS may have a normal ratio of LH to FSH (30%), normal regular menses, and normal ovarian morphology on ultrasonography (20% to 40%) (2, 3). In addition, PCOS is common in otherwise apparently healthy women (4). Recently, idiopathic hirsutism was defined as existence of hirsutism in presence of regular menses and normal ovulatory function and normal androgen levels (3). However, a history of regular menses is not sufficient to exclude ovulatory dysfunction, since up to 40% of eumenorrheic hirsute women are anovulatory (5). Therefore, some hirsute women with PCOS are often misdiagnosed as having idiopathic hirsutism. Idiopathic hirsutism is relatively uncommon, affecting approximately less than 20% of all hirsute women (5, 6). 133

2 Although androstenodione is increased in both conditions, the activity of aromatase, which converts androstenodione to estrone, is decreased in idiopathic hirsutism (7). Therefore, increased estrone levels, which are responsible for anovulatory cycles in PCOS, are not seen in idiopathic hirsutism. In contrast to PCOS, ovulation and menstrual cycle may be normal in idiopathic hirsutism. The primary defect in PCOS is reported to be insulin resistance (8 11). Various studies indicate that nonobese patients with PCOS have substantial insulin resistance (11 13). In addition, increased -cell function was found even in nonobese women with PCOS. Insulin hypersecretion is thus probably linked with the pathogenesis of PCOS (14), but hyperandrogenemia itself is also reported to cause insulin resistance (15, 16). Insulin resistance affects 10% to 25% of the general population. Two common disorders frequently associated with insulin resistance are PCOS and type 2 diabetes mellitus, which affect 4% to 6% and 2% to 6% of reproductive-aged women, respectively. Overall, about 50% to 70% of women with PCOS and 80% to 100% of patients with type 2 diabetes mellitus have variable degrees of insulin resistance. The risk for type 2 diabetes mellitus among patients with PCOS is 5- to 10-fold higher than in healthy women (17). A subgroup of women with this syndrome has metabolic PCOS, which can be considered to be a prediabetic state. There is an urgent need to define the more subtle features in young lean women with PCOS, in whom the metabolic syndrome is yet to emerge. Such knowledge would allow prediction of future health risks (18). Simple methods are usually used to measure insulin sensitivity, but none of these various methods alone are ideal for assessing the action of insulin. Because women with PCOS can be considered to be in a prediabetic state, they should be investigated for both insulin sensitivity and -cell function. We assessed and compared various indices of simple insulin sensitivity and -cell function in lean hyperandrogenemic and normoandrogenemic hirsute young women. Because some patients with PCOS have normal androgenemia (19, 20) and 50% to 70% of patients with PCOS and 20% of nonobese healthy women have insulin resistance, we classified patients as hyperandrogenemic or normoandrogenemic instead of as having PCOS or idiopathic hirsutism. The fact that androgenemia is closely related to insulin resistance supports this classification. MATERIALS AND METHODS Patients We enrolled 71 lean (body mass index 25 kg/m 2 ) hirsute women and 35 healthy controls. The study group and control group were similar in terms of age and body mass index (P.05) (Table 1). Women in the study group were classified as hyperandrogenemic (free T level 3.8 pg/ml; n 37) or normoandrogenemic (free T level 3.8 pg/ml; n 34). Women with a homeostasis model assessment of insulin resistance (HOMA-IR) value greater than 3.2 were considered insulin resistant (22). The hyperandrogenemic and normoandrogenemic hirsute women were further divided into two subgroups according to insulin resistance (Tables 2 and 3). Those with a HOMA-IR value 3.2 were considered insulin resistant (20 hyperandrogenemic women and 12 normoandrogenemic women). Those with a HOMA-IR 3.2 were considered non insulin resistant (17 hyperandrogenemic women and 22 normoandrogenemic women). The diagnosis of hirsutism was evaluated by using the Ferriman Gallwey score ( 7). The diagnosis of PCOS was established by an elevated free T level, an elevated ratio of LH to FSH ( 2), ovarian ultrasonographic appearance, chronic anovulation (P level 5 ng/ml), and chronic oligomenorrhea ( 6 menses/y) or amenorrhea. Hyperandrogenemic women were considered to have PCOS according to the above criteria. Normoandrogenemic hirsutic women were considered to have unspecified hirsutism regardless of whether they had been classified as having PCOS or idiopathic hirsutism because normoandrogenemic PCOS and idiopathic hirsutism have similar features (normal menstrual cycles, normal ratio of LH to FSH ratio, or normal appearance of ovarian morphology on ultrasonography). Women with the Cushing syndrome, congenital adrenal hyperplasia, androgen-secreting ovary and adrenal tumor, adenoma of hypophysis, acromegaly, virilism, and hirsutism due to drugs were excluded from the study. All patients were euthyroid. Therapy with all drugs that could affect hormonal and carbohydrate metabolism, including oral contraceptives, antiandrogens, and metformin, were discontinued at least 3 months before the study. All women had normal glucose tolerance according to the criteria of the National Diabetes Group (22). No participant had functional deficiency of thyroid, kidney, or liver, and none were hyperprolactinemic. No participant engaged in a regular exercise or dietary program or had lost a substantial amount of weight recently. All patients gave informed consent to participate. The study was conducted according to the Declaration of Helsinki and guidelines for Good Clinical Practice, and it was approved by the local ethics committee. Testing and Evaluation Body mass index was obtained by dividing body weight in kilograms by the square of height in meters. A 300-g carbohydrate diet was given for 3 days before the oral glucose tolerance test (OGTT). The OGTT was performed with 75 g of glucose after 12 hours of fasting. Blood samples were obtained at fasting and at 1, 2, and 3 hours after testing to measure insulin 134 Altuntas et al. Insulin sensitivity indices in hirsute women Vol. 80, No. 1, July 2003

3 TABLE 1 General features of the study sample. Variable Hyperandrogenemic women Normoandrogenemic women Healthy controls P values HA vs. NA NA vs. control HA vs. control No. of participants Age (y) NS NS NS BMI (kg/m 2 ) NS NS NS FSH (miu/ml) NS NS NS LH (miu/ml) NS NS NS Free T (pg/ml) NS.0001 DHEAS ( g/ml) NS.001 PRL (ng/ml) NS NS NS Fasting glucose (mg/dl) NS NS NS 1-h glucose (mg/dl) NS NS NS 2-h glucose (mg/dl) NS NS NS 3-h glucose (mg/dl) NS NS NS Total glucose (mg/dl) NS NS NS Glucose AUC (mg/ml/min) NS NS NS Fasting insulin ( U/mL) NS NS.05 1-h insulin ( U/mL) NS.05 2-h insulin ( U/mL) NS.05 3-h insulin ( U/mL) NS.05 Total insulin ( U/mL) NS.05 Insulin AUC ( U/mL/min) NS NS NS Glucose AUC/insulin AUC NS NS NS HOMA-IR ( U/mg) NS NS.0001 Insulinogenic index 120min NS.05 (UI/mg) Fasting glucose/insulin ratio NS NS NS (U/mg) Quicky index ( U 1 /mg 1 ) NS NS NS FIRI NS NS FIRI 1 (mmol 1 /mui 1 /L) NS NS NS Raynaud index ( UI 1 /L) NS NS NS Fasting Belfiore index NS NS NS (pmol 1 mmol 1 l) HOMA -cell index (%) NS.0001 ISI HOMA ( U 1 /mg 1 ) NS NS NS Note: Data are means ( SE). AUC area under the curve; BMI body mass index; FIRI fasting insulin resistance index; HA hyperandrogenemic; HOMA homeostasis model assessment; ISI hepatic insulin sensitivity index; NA normoandrogenemic; NS not significant. Altuntas. Insulin sensitivity indicies in hirsute women. Fertil Steril and glucose levels. Blood glucose levels were studied in whole blood by using enzymatic calorimetric methods based on commercial devices, and plasma insulin levels were studied by recombinant immunoassay with double antibodies (DPC, Los Angeles, CA). Hormonal measurement was done during the first week of the menstrual cycle. We studied LH and FSH by immunoradiomorphometric assay (Byk-Sangtec Diagnostica GmbH, Dietzenbach, Germany), free T by recombinant immunoassay (DPC), and DHEAS by recombinant immunoassay (Pantex, Santa Monica, CA). -Cell Function and Insulin Sensitivity Indices To estimate -cell function and insulin sensitivity, we used various indices derived from fasting or OGTT (0, 60, 120 minutes) measurements of glucose and insulin (Table 4). Indices of -cell function, such as HOMA -cell index and insulinogenic index 120min, were estimated from OGTT-based glucose and insulin levels. Insulin sensitivity according to such criteria as HOMA- IR, hepatic insulin sensitivity (ISI HOMA ), the Quicky index, reciprocal fasting insulin resistance index, fasting Belfiore index, and fasting glucose/insulin ratio, were calculated by using mathematical estimation from a single fasting sample of glucose and insulin levels. The Raynaud index was calculated by using the mathematical estimation in a single fasting sample of insulin levels. The OGTT-derived insulin sensitivity indices were estimated from glucose and insulin levels 1, 2, and 3 hours after challenge. FERTILITY & STERILITY 135

4 TABLE 2 Classification of hyperandrogenemic patients, according to insulin resistance. Variable HA and IR (HOMA-IR 3.2) HA and no IR (HOMA-IR 3.2) No. of participants Age (y) BMI (kg/m 2 ) FSH (miu/ml) LH (miu/ml) Free T (pg/ml) a a a DHEAS ( g/ml) PRL (ng/ml) Fasting glucose (mg/dl) h glucose (mg/dl) h glucose (mg/dl) h glucose (mg/dl) Total glucose (mg/dl) Glucose AUC (mg/ml/min) Fasting insulin ( U/mL) b b 1-h insulin ( U/mL) b b 2-h insulin ( U/mL) b b 3-h insulin ( U/mL) Total insulin ( U/mL) a a Insulin AUC ( U/mL/min) Glucose AUC/insulin AUC HOMA-IR ( U/mg) b b Insulinogenic index 120min (UI/mg) Fasting glucose/insulin ratio a a (U/mg) Quicky index ( U 1 /mg 1 ) b b FIRI FIRI 1 (mmol 1 /mui 1 /L) Raynaud index ( UI 1 /L) a a Fasting Belfiore index a a (pmol 1 1/mmol 1 1) HOMA -cell index (%) b b ISI HOMA ( U 1 /mg 1 ) a a Note: Data are means ( SE). AUC area under the curve; BMI body mass index; HA hyperandrogenemia; FIRI fasting insulin resistance index; HOMA homeostasis model assessment; IR insulin resistance; ISI hepatic insulin sensitivity index. a P.05. b P.001. Altuntas. Insulin sensitivity indicies in hirsute women. Fertil Steril HOMA-IR The cut-off value for HOMA-IR was accepted as 3.2 or greater, according to the criteria of Marques-Vidal et al. (21). Fasting Glucose/Insulin Ratio The fasting glucose/insulin ratio was obtained by dividing the fasting glucose level (mg/dl) by the fasting insulin level ( U/mL). The cut-off point of glucose/insulin ratio for insulin resistance was accepted as 4.5, according to the criteria of Legro et al. (26). Area Under the Curve of Glucose and Insulin The integrated area under the curve (AUC) analysis for glucose and insulin were determined according to the formula of Tai et al. (32). The product of the glucose area under the plasma glucose curve and insulin area under the plasma insulin curve has been used as an index of insulin resistance (33, 34). Fasting Insulin Level The cut-off value of fasting insulin level for insulin resistance was accepted as 13 U/mL, according to the criteria of Ludvig et al. (35). 1-Hour (Peak) Insulin Levels The highest insulin levels achieved during the first hours of OGTT were considered to be the peak insulin level. Peak 136 Altuntas et al. Insulin sensitivity indices in hirsute women Vol. 80, No. 1, July 2003

5 TABLE 3 Classification of normoandrogenemic patients, according to insulin resistance. Variable NA and IR (HOMA-IR 3.2) NA and no IR (HOMA-IR 3.2) No. of participants Age (y) BMI (kg/m 2 ) FSH (miu/ml) LH (miu/ml) Free T (pg/ml) DHEAS ( g/ml) a a PRL (ng/ml) Fasting glucose (mg/dl) h glucose (mg/dl) h glucose (mg/dl) h glucose (mg/dl) Total glucose (mg/dl) Glucose AUC (mg/ml/min) Fasting insulin ( U/mL) a a 1-h insulin ( U/mL) h insulin ( U/mL) h insulin ( U/mL) Total insulin ( U/mL) Insulin AUC ( U/mL/min) Glucose AUC/insulin AUC HOMA-IR ( U/mg) Insulinogenic index 120min (UI/mg) Fasting glucose/insulin ratio a a (U/mg) Quicky index ( U 1 /mg 1 ) a a FIRI FIRI 1 (mmol 1 /mui 1 /L) Raynaud index ( UI 1 /L) Fasting Belfiore index (pmol 1 1/mmol 1 1) HOMA -cell index (%) a a ISI HOMA ( U 1 /mg 1 ) Note: Data are means ( SE). AUC area under the curve; BMI body mass index; FIRI fasting insulin resistance; HOMA homeostasis model assessment; ISI hepatic insulin sensitivity index. Altuntas. Insulin sensitivity indicies in hirsute women. Fertil Steril insulin levels greater than 150 U/mL at OGTT were accepted as insulin resistance (36). Total Insulin Response During OGTT The sum of insulin levels obtained at 0, 1, 2, and 3 hours after OGTT were considered the total insulin levels (37). Total insulin levels on OGTT greater than 155 U/mL were accepted as insulin resistance (36). Statistical Analysis Data analysis was performed by using SPSS software, version 11.0 (SPSS, Inc., Chicago, IL). All results are the mean ( SD). The nonparametric Mann Whitney U test was used to determine differences between groups. Spearman s correlation test was used for correlation analysis. Analysis of variance was used for multiple comparisons. P.05 was considered statistically significant. RESULTS Table 1 shows general characteristics of hyperandrogenemic and normoandrogenemic patients. -Cell Function Indices The insulinogenic index 120min and HOMA -cell index were used as -cell function indices. Insulinogenic index 120min was significantly higher in normoandrogenemic women than hyperandrogenemic women (P.01). The HOMA -cell index was significantly higher in hyperandrogenemic women than normoandrogenemic women (P.05). The total insulin level was positively correlated with insuli- FERTILITY & STERILITY 137

6 TABLE 4 Formulas and references of indices of -cell function and insulin sensitivity derived from fasting and OGTT measurements of glucose and insulin. Index Formula Reference -Cell function HOMA -cell index (%) 20 insulin 0 min/(glucose 0 min - 3.5) 23 Insulinogenic index 120min (UI/mg) AUC insulin/ AUC glucose 24 Insulin sensitivity Fasting insulin 1 ( U/mL) 1/insulin 0 min 25 Fasting glucose to insulin ratio (U/mg) glucose 0 min/insulin 0 min 26 Raynaud index ( UI 1 /L) 40/insulin 0 min 27 Fasting Belfiore index (pmol 1 1/mmol 1 1) 2/[(insulin 0 min glucose 0 min) 1] 28 FIRI 1 (mmol 1 /mui 1 /L) FIRI 1 (glucose 0 min insulin 0 min)/25 FIRI 1 1/FIRI 29 Insulin 120 min 1 ( U/mL) 1/insulin 120 min 29 Glucose AUC-insulin AUC OGTT ratio AUC glucose/auc insulin 23 (mg/ U) Quicky index ( U 1 /mg 1 ) 1/log insulin 0 min log glucose 0 min 31 ISI HOMA ( U 1 /mg 1 ) ( )/fasting insulin fasting glucose 27 HOMA-IR ( U/mg) (fasting insulin fasting glucose)/ Note: AUC area under the curve; FIRI fasting insulin resistance index; HOMA homeostasis model assessment; IR insulin resistance; ISI insulin sensitivity index; OGTT oral glucose tolerance test. Altuntas. Insulin sensitivity indicies in hirsute women. Fertil Steril nogenic index 120min (r.355; P.003) but was not correlated with the HOMA index. The 1-hour insulin level was positively correlated with insulinogenic index 120min (r.288; P.016). Insulin Sensitivity Table 5 shows results of evaluation of hyperandrogenemic and normoandrogenemic women according to various hyperinsulinemia criteria. Fasting Insulin When a cut-off value of 13 U/mL or greater was used for insulin resistance, 55% of patients were insulin resistant (Table 5). Fasting insulin levels correlated positively with 1-hour and 2-hour insulin levels (r.352 and r.358, respectively; P.002), total insulin levels during OGTT (r.472; P.001), HOMA -cell index (r.893, P.0001), and fasting insulin resistance index (FIRI) (r.981; P.0001). In contrast, fasting glucose levels correlated negatively with fasting glucose/insulin ratio (r.387; P.001), Quicky index (r.700; P.0001), FIRI 1 (r.328; P.006), fasting insulin 1 (r.387; P.001), Raynaud index (r 387; P.001), fasting Belfiore index (r.336; P.005), and ISI HOMA (r.328; P.006). The fasting insulin level correlated best with FIRI (r.98; P 0.001), HOMA-IR (r.89; P.001), and Quicky index (r.7; P.001). HOMA-IR Patients were considered insulin resistant on the basis of a HOMA-IR 3.2 (21). The HOMA-IR correlated positively with fasting (r.97; P.0001), 1-hour (r.38; P.001), TABLE 5 Evaluation of hyperandrogenemic and normoandrogenemic women, according to various hyperinsulinemia criteria. Criterion Hyperandrogenemic (n 37) Normoandrogenemic (n 34) P value HOMA-IR (45.9) 10 (29.41) NS Fasting insulin level 13 U/L 21 (56.7) 18 (52.9) NS 1-h (peak) insulin 150 U/L 15 (40.5) 14 (41.1) NS Total insulin level 155 U/mL 25 (67.5) 15 (44.1).046 Fasting glucose/insulin ratio (59.5) 20 (58.8) NS Note: Data are the number (percentage) of patients. HOMA homeostasis model assessment; IR insulin resistance; NS not significant. Altuntas. Insulin sensitivity indicies in hirsute women. Fertil Steril Altuntas et al. Insulin sensitivity indices in hirsute women Vol. 80, No. 1, July 2003

7 2-hour (r.6; P.0001), and total insulin levels (r.6; P.0001); insulin AUC (r.5, P.0001); HOMA -cell index (r.9; P.0001); FIRI (r 1; P.0001); and fasting insulin 1 (r.3; P.003). It correlated negatively with the fasting glucose/insulin ratio (r.4; P.001), Raynaud index (r.3; P.003), Belfiore index (r.31; P.01), FIRI 1 (r.304; P.01), Quicky index (r.413; P.0001), and ISI HOMA (r.641; P.0001). 1-Hour Post-Challenge Insulin Level Twelve percent of patients had 1-hour post challenge insulin levels greater than 150 U/mL. The 1-hour postchallenge insulin level correlated positively with HOMA-IR (r.38; P.0001), 3-hour post-challenge insulin level (r.340; P.004), total insulin level (r.865; P.0001), insulin AUC (r.822; P.0001), insulinogenic index 120min (r.288; P.016), and FIRI (r.319; P.008). It correlated negatively with fasting glucose/insulin ratio (r.244; P.042) and the Quicky index (r.386; P.001). The 1-hour post-challenge insulin level correlated best with the total insulin level. 2-Hour Post-Challenge Insulin Level The 2-hour insulin level correlated positively with HOMA-IR (r.324; P.05), 3-hour post-challenge insulin level (r.313; P.05), total insulin level (r.781; P.001), AUC insulin (r.796; P.001), and FIRI (r.349; P.05), and negatively with fasting glucose/insulin ratio (r.236; P.05) and the Quicky index (r.348; P.05). 3-Hour Post-Challenge Insulin Level The 3-hour insulin level correlated positively with the 1-hour post-challenge insulin level (r.34; P.05), total insulin level (r.584; P.0001), and insulin AUC (r.445; P.0001). Total Insulin Level Fifty-six percent of patients had a total insulin level greater than 155 U/mL. The total insulin level correlated positively with 1-hour (r.219; P.046) and 2-hour post-challenge glucose levels (r.237; P.048); glucose AUC (r.239; P.047); fasting insulin level (r.472; P.0001); 1-hour (r.865; P.0001), 2-hour (r.781; P.0001), and 3-hour post-challenge insulin levels (r.584; P.0001); insulin AUC (r.952; P 0001); insulinogenic index 120min (r.355; P.003); and FIRI (r.439; P.0001). Total insulin levels correlated negatively with fasting glucose/insulin ratio (r.335; P.005), the Quicky index (r.465; P.0001), FIRI 1 (r.293; P.015), fasting insulin level 1 (r.307; P.010), the Raynaud index (r.307; P.010), fasting Belfiore index (r.298; P.013), and ISI HOMA (r.293; P.015). Hepatic Insulin Sensitivity According to Glycemia and Insulinemia Hepatic insulin sensitivity reflecting basal hepatic glucose production had a statistically significant negative correlation with fasting glycemia (r.249; P.039), fasting insulinemia (r.328, P.006), total insulin level (r.293; P.015), and insulinemia AUC (r.281; P.019). Mean hepatic insulin sensitivity was higher in hyperandrogenemic women than normoandrogenemic women but was not statistically significant. The ISI HOMA correlated well with HOMA-IR (r.314; P.001), the Quicky index (r.870; P.0001), reciprocal fasting insulin resistance index (r 1; P.0001), fasting Belfiore index (r 1; P.0001), and fasting glucose/insulin ratio (r.885; P.0001), which were estimated from a single measurement of glucose and insulin. Relation Between Androgen and Insulin Levels Free T levels were significantly correlated with HOMA-IR (r.494; P.0001), fasting insulin level (r.476; P.0001), 3-hour post-challenge insulin level (r.237; P.046), and total insulin level (r.315; P.007). The free T level did not correlate with other insulin resistance variables. The DHEAS level correlated significantly only with the free testosterone level (r.408; P.0001). There were no correlations among DHEAS level and insulin resistance variables. DISCUSSION Simple methods are usually used to measure insulin sensitivity because they are easy to perform, quick, and inexpensive. However, none of these methods alone is ideal for assessing the action of insulin. We compared OGTT-derived post-challenge insulin resistance indices with simple insulin resistance indices derived from a single sample to evaluate insulin sensitivity and -cell function in young, lean, hyperandrogenemic and normoandrogenemic hirsute women. A principal reason to use OGTT is to identify persons with impaired glucose tolerance, because such persons are at increased risk for diabetes. In our study, only OGTT-derived insulin resistance indices (1-hour, 2-hour, and 3-hour postchallenge level and total insulin level) were significantly higher in hyperandrogenemic hirsute women than in normoandrogenemic hirsute women. In addition, OGTT-derived insulin resistance indices correlated best with various other insulin resistance indices based on fasting glucose and insulin levels derived from a single sample. The HOMA-IR correlated with OGTT-derived insulin resistance indices but not with the 3-hour post-challenge insulin level. Hyperandrogenemic women and normoandrogenemic women did not differ statistically in insulin resistance as assessed by HOMA-IR. However, 54% of hyperan- FERTILITY & STERILITY 139

8 drogenemic women and 32% of normoandrogenemic women were insulin resistant when a cut-off HOMA-IR value for insulin resistance of 3.2 or greater was used. The lower rate of insulin resistance suggests idiopathic hirsutism rather than PCOS in normoandrogenemic women. Hyperinsulinemia in women with idiopathic hirsutism as well as in nonobese women with PCOS has been reported (38). Hyperandrogenemic and normoandrogenemic women were subdivided further as having insulin resistance (HOMA-IR 3.2) or not having insulin resistance (HOMA- IR 3.2) (22). Fasting, 1-hour post-challenge, 2-hour postchallenge, and total insulin levels; Quicky index; reciprocal fasting insulin resistance index; Raynaud index; fasting Belfiore index; ISI HOMA ; and fasting glucose/insulin ratio were significantly higher in hyperandrogenemic women with insulin resistance than hyperandrogenemic women without insulin resistance. Only fasting insulin level, fasting glucose/ insulin ratio, and Quicky index were significantly higher in normoandrogenemic women with insulin resistance than normoandrogenemic without resistance. In the past two decades, the fasting insulin level has been used as a surrogate index of insulin sensitivity in several epidemiologic studies (39, 40). However, fasting insulin cannot explain more than 30% to 40% of the variance in glucose clamp determined insulin sensitivity (41). Therefore, it has not been recommended as a routine screening test in determining insulin sensitivity. In our study, mean fasting insulin levels were higher in hyperandrogenemic women than in normoandrogenemic women. When a cut-off level for insulin resistance of 13 U/mL or greater was used (39), 55% of all patients were insulin resistant. It has been shown in vivo and in vitro that hyperinsulinemia decreases the levels of adrenal androgen DHEAS. However, when insulin resistance develops, the effect of insulin on decreasing DHEAS levels diminishes; as a result, DHEAS levels remain the same, which in some cases of PCOS, results in increased levels of adrenal androgens (42). On this basis, it was reported that no relationship exists between insulin resistance and adrenal hyperandrogenism (43). Our study confirms this idea. The free T level had a significant positive correlation with fasting insulin, 3-hour post-challenge insulin, and total insulin levels. No correlation existed between the free T level and other insulin resistance parameters. Hyperandrogenism and insulin resistance are independent features of PCOS, and obesity exacerbates the clinical expression of hyperandrogenism (44). In addition, hyperandrogenism in PCOS may contribute to the associated insulin resistance in women with hyperandrogenism and is partially reversed by antiandrogen treatment (43). Studies in which insulin levels have been decreased by a variety of modalities indicate that hyperinsulinemia augments androgen production in PCOS. Furthermore, in many women with PCOS, lowering insulin level ameliorates but does not abolish hyperandrogenism (45). Despite having normal glucose tolerance and fasting normoinsulinemia, nonobese women with PCOS have a hyperinsulinemic response to glucose loading (46). Various studies indicate that nonobese patients with PCOS have a significant degree of insulin resistance (11 13). In addition, increased -cell function has been found in nonobese women with PCOS (14). In our study, the insulinogenic index 120min, an indicator of -cell function, was increased; this finding is similar to previous reports (14). We used insulinogenic and HOMA indices to evaluate -cell function. Only the insulinogenic index 120min, which is calculated from basal and OGTT-derived insulin and glucose levels, was correlated with total insulin and 1-hour postchallenge insulin levels. In contrast, the HOMA -cell index, which is estimated from a single sample of glucose and insulin level, was not correlated with total insulin and 1-hour post-challenge insulin levels. In our study, OGTT revealed no glucose intolerance in hyperandrogenic and normoandrogenic women, and the glycemic response did not differ between these groups. In addition, no difference was found in glycemic responses between hyperandrogenemic hyperinsulinemic and hyperandrogenemic normoinsulinemic women. The high insulin levels despite normal glucose tolerance observed in nonobese hyperandrogenic women correlated with findings of other studies (9). It has been suggested that a basal hepatic glucose overproduction defect may be a manifestation of the global insulin resistance in persons at risk for type 2 diabetes mellitus (47). The ISI HOMA, which reflects basal hepatic glucose production, was negatively correlated with fasting glycemia (P.05) in our study. No difference was observed between normoandrogenemic and hyperandrogenemic groups with regard to hepatic insulin sensitivity (P.05). Similar results were reported in obese women with PCOS (9). Therefore, hepatic insulin sensitivity should not be neglected in evaluation of insulin resistance. The ISI HOMA was strongly correlated with HOMA-IR, Quicky index, reciprocal fasting insulin resistance index, fasting Belfiore index, and fasting glucose/insulin ratio, which are estimated from a single measurement of glucose and insulin. In the healthy population, 20% of nonobese people have insulin resistance (36). In our study, total insulinemia at OGTT was found to be the most significant criterion in evaluating insulin resistance in nonobese hyperandrogenemic women. We previously reported that the insulinemia/ glycemia ratio is a good measure of insulin sensitivity in obese premenapousal women (47, 48). 140 Altuntas et al. Insulin sensitivity indices in hirsute women Vol. 80, No. 1, July 2003

9 When the fasting glycemia/insulinemia ratio (glucose/ insulin ratio) was used as a criterion for indicating insulin resistance, 59% of both hyperandrogenemic and normoandrogenemic women had insulin resistance. Therefore, the glucose/insulin ratio seems to be a less sensitive indicator in demonstration of hyperinsulinemia or insulin resistance compared with HOMA-IR, because hyperandrogenemia was correlated with hyperinsulinemia. In conclusion, nonobese hyperandrogenemic young hirsute women have insulin resistance and increased -cell function. Therefore, insulin resistance should be investigated in all hirsute women regardless of whether they have PCOS or idiopathic hirsutism. Even if they are lean and young, hirsute women with a family history of diabetes should be investigated to determine their risk of becoming diabetic. We suggest that OGTT-based glucose and insulin levels are the most useful measures of insulin sensitivity and -cell function in hirsute women. The OGTT also identifies persons with impaired glucose tolerance who are at increased risk for diabetes. The evaluation of insulin resistance formulas, which are estimated from measurement of glucose and insulin, should be used in hirsute women. We propose that HOMA-IR be used as a global insulin resistance test. It correlated well with OGTT-derived insulin resistance and -cell function indices and with insulin resistance indices derived from a single sample, such as the ISI HOMA, Quicky index, reciprocal fasting insulin resistance index, fasting Belfiore index, and fasting glucose/insulin ratio. References 1. Rittmaster RS. Hirsutism. Lancet 1997;349: Carmina E, Lobo RA. Polycystic ovaries in hirsute women with normal menses. Am J Med 2001;111: Azziz R, Carmina E, Sawaya ME. Idiopathic hirsutism. Endocr Rev 2000;21: Polson DW, Adams J, Wadsworth J, Franks S. Polycystic ovaries. A common finding in normal women. Lancet 1988;16: Azziz R, Waggoner WT, Ochoa T, Knochenhauer ES, Boots LR. Idiopathic hirsutism: an uncommon cause of hirsutism in Alabama. Fertil Steril 1998;70: Delahunt JW. Hirsutism. Drugs 1993;45: McKenna TD. Pathogenesis and treatment of polycystic ovary syndrome. N Engl J Med 1988;318: Burghen GA, Givens JR, Kitabchi AK. Correlation of hyperandrogenism with hyperinsulinism in polycystic ovarian disease. J Clin Endocrinol Metab 1980;50: Dunaif A, Segal KR, Futterweit W, Dobriansky A. Profound peripheral insulin resistance independent of obesity, in polycystic ovarian syndrome. Diabetes 1989;38: Jialal I, Naiker P, Reddi K, Moodley J, Joubert SM. Evidence for insulin resistance in non-obese patients with polycystic ovarian disease. J Clin Endocrinol Metab 1987;64: Legro RS, Bentley-Lewis R, Driscoll D, Wang SC, Dunaif A. Insulin resistance in the sisters of women with polycystic ovary syndrome: association with hyperandrogenemia rather than menstrual irregularity. J Clin Endocrinol Metab 2002;87: Chang RJ, Nakamura RM, Judd HI, Kaplan SA. Insulin resistance in non-obese patients with polycystic ovarian disease. J Clin Endocrinol Metab 1983;57: Toprak S, Yonem A, Cakir B, Guler S, Azal O, Ozata M, et al. Insulin resistance in non-obese patients with polycystic ovary syndrome. Horm Res 2001;55: Vrbikova J, Bendlova B, Hill M, Vankova M, Vondra K, Starka L. Insulin sensitivity and beta-cell function in women with polycystic ovary syndrome. Diabetes Care 2002;25: Smith S, Ravnikar VA, Barbieri RL. Androgen and insulin response to an oral glucose challenge in hyperandrogenic women. Fertil Steril 1987;48: Pugeat M, Ducluzeau PH, Mallion-Donadieu M. Association of insulin resistance with hyperandrogenia in women. Horm Res 2000;54: Ovalle F, Azziz R. Insulin resistance, polycystic ovary syndrome, and type 2 diabetes mellitus. Fertil Steril 2002;77: Conway GS. Hyperinsulinaemia and polycystic ovary syndrome. Hum Fertil (Camb) 2000;3: Conway GS, Honour JW, Jacobs HS. Heterogeneity of the polycystic ovary syndrome: clinical, endocrine and ultrasound features in 556 patients. Clin Endocrinol 1989;30: Yıldız BO, Gedik O. Insulin resistance in polycystic ovary syndrome: hyperandrogenemia versus normoandrogenemia. Eur J Obstet Gynecol Reprod Biol 2001;100: Marques-Vidal P, Mazoyer E, Bongard V, Gourdy P, Ruidavets JB, Drouet L, et al. Prevalence of insulin resistance syndrome in southwestern France and its relationship with inflammatory and hemostatic markers. Diabetes Care 2002;25: Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care 2001:24(Suppl 1):S5 S Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and -cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28: Drivsholm T, Hansen T, Urhammer SA Palacious RT, Volund A, Borch-Johnson K, et al. Assesment of insulin-sensitivity and beta-cell function from an oral glucose tolerance test [abstract]. Diabetologia 1999;42(Suppl 1):A Hermans MP, Levy JC, Morris RJ, Turner RC. Comparison of insulin sensitivity tests across a range of glucose tolerance from normal to diabetes. Diabetologia 1999;42: Legro RS, Finegood D, Dunaif A. A fasting glucose to insulin ratio is useful measure of insulin sensitivity in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1998;83: Raynaud E, Perez-Martin A, Brun JF, Benhaddad AA, Mercier J. Revised concepts for the estimation of insulin sensitivity from a single sample. Diabetes Care 1999;22: Belfiore F, Iannello S, Volpicelli G. Insulin sensitivity indices calculated from basal and OGTT-induced insulin, glucose, and FFA levels. Mol Genet Metab 1998;63: Duncan MH, Singh BM, Wise PH, Carter G, Alaghband-Zadeh J. A simple measure of insulin resistance. Lancet 1995;346: Stumvoll M, Mitrakou A, Pimenta W, Jenssen T, Yki-Jarvinen H, Van Haeften T, et al. Use of the oral glucose tolerance test to assess insulin release and insulin sensitivity. Diabetes Care 2000;23: Katz A, Nambi SS, Mather K, Baron AD, Follmann DA, Sullivan G, et al. Quantitative insulin sensitivity check index: a simple, accurate method for assessing insulin sensitivity in humans. J Clin Endocrinol Metab 2000;85: Tai MM. A mathematical model for the determination of total area under glucose tolerance and other metabolic curves. Diabetes Care 1994;17: Levine R, Haft R. Carbohydrate homeostasis. N Engl J Med 1970;283: Myllynen P, Koivisto V, Nikkila E. Glucose intolerance and insulin resistance accompany immobilization. Acta Med Scand 1987;222: Ludvig B, Nolan JJ, Baloga J, Sacks D, Olefsky J. Effect of obesity on insulin resistance in normal subject and patients with NIDDM. Diabetes 1995;44: Barbieri RL, Hornstein MD. Hyperinsulinaemia and ovarian hyperandrogenism: cause and effect. Endocrinol Metab Clin North Am 1988; 7: Haffner SM, Valdez RA, Hazuda HH, Mitchell BD, Morales PA, Stern MP. Prospective analysis of the insulin resistance syndrome (syndrome X). Diabetes 1992;41: Paoletti AM, Cagnacci A, Orru M, Ajossa S, Guerriero S, Melis GB. Treatment with flutamide improves hyperinsulinemia in women with idiopathic hirsutism. Fertil Steril 1999;72: Bonora E, Kiechl S, Willeit J, Oberhollenzer F, Egger G, Targher G, et al. Prevalence of insulin resistance in metabolic disorders. Diabetes 1998;47: Laakso M. How good a marker is insulin level for insulin resistance? Am J Epidemiol 1993;137: Falcone T, Finegood D, Fantos IG, Morris D. Androgen response to endogenous insulin secretion during the frequently sampled intravenous glucose tolerance test in normal and hyperandrogenic women. J Clin Endocrinol Metab 1990;71: Rittmaster RS, Deshwall N, Lehman L. The role of adrenal hyperandrogenism, insulin resistance, and obesity in the pathogenesis of polycystic ovarian syndrome. J Clin Endocrinol Metab 1993;76: FERTILITY & STERILITY 141

10 43. Dunaif A. Insulin resistance and the polycystic ovary syndrome: mechanism and implications for pathogenesis. Endocr Rev 1997;18: Moghetti P, Tosi F, Castello R, Magnani CM, Negri C, Brun E, et al. The insulin resistance in women with hyperandrogenism is partially reversed by androgen treatment: evidence that androgens impair insulin action in women. J Clin Endocrinol Metab 1996;81: Chang RJ, Nakamura RM, Judd HI, Kaplan SA. Insulin resistance in non-obese patients with polycystic ovarian disease. J Clin Endocrinol Metab 1983;57: Osei K. Increased basal glucose production and utilization in nondiabetic first-degree relatives of patients with NIDDM. Diabetes 1990;39: Altuntas Y, Korugan U, Hekim N. Comparison of insulin tolerance test with various simple and safe methods in the prediction of insulin sensitivity. Turk J Endocrinol Metab 1998;3: Hekim N, Korugan U, Altuntas Y. Comparison of different metabolic and anthropometric parameters in reflection of insulin sensitivity in premenopausal women [abstract]. Diabetologia 1996;39(Suppl 1):A Altuntas et al. Insulin sensitivity indices in hirsute women Vol. 80, No. 1, July 2003