Endogenous sex hormones and risk factors for atherosclerosis in healthy Greek postmenopausal women

Transcription

1 European Journal of Endocrinology (2006) ISSN CLINICAL STUDY Endogenous sex hormones and risk factors for atherosclerosis in healthy Greek postmenopausal women Irene Lambrinoudaki 1, George Christodoulakos 1, Demetrios Rizos 2, Emmanuel Economou 2, John Argeitis 1, Sofia Vlachou 1, Maria Creatsa 1, Evangelia Kouskouni 2 and Dimitrios Botsis 1 1 2nd Department of Obstetrics and Gynecology, and 2 Hormonal and Biochemical Laboratory, University of Athens, Aretaieio Hospital, Athens, Greece (Correspondence should be addressed to I Lambrinoudaki, 27 Themistokleous Street, Dionysos, GR-14578, Athens, Greece; ilambrinoudaki@hotmail.com) Abstract Objective: To assess the association between endogenous sex hormones and risk factors for atherosclerosis in healthy postmenopausal women. Design: Cross-sectional study in a university menopause clinic. Methods: Serum sex hormones and lipid lipoprotein profile, arterial pressure, homocysteine and insulin resistance, measured by the homeostasis model assessment of insulin resistance (HOMA-IR), were assessed in 598 healthy postmenopausal women not on hormone therapy. Results: Compared with women in the lowest testosterone quartile (Q), women in the highest testosterone quartile had higher total cholesterol (Q1: 225.2G41.3 vs Q4: 246.2G38.4 mg/dl, P!0.01), low-density lipoprotein (LDL)-cholesterol (Q1: 146.9G37.2 vs Q4: 171.8G35.3 mg/dl, P!0.001), atherogenic index of plasma (AIP) (Q1: 0.224G0.238 vs Q4: 0.087G0.254, P!0.01), apolipoprotein B (ApoB) (Q1: 100.7G23.1 vs Q4: 113.9G23.8 mg/dl, P!0.001) and higher highdensity lipoprotein (HDL)-cholesterol (Q1: 60.7G14.5 vs Q4: 52.9G13.0 mg/dl, P!0.01). Accordingly, women in the highest free androgen index (FAI) quartile had higher AIP (Q1: 0.232G0.254 vs Q4: 0.078G0.243, P!0.001) and ApoB (Q1: 102.4G25.5 vs Q4: 114.2G 25.8 mg/dl, P!0.01) and lower HDL-cholesterol (Q1: 62.0G15.7 vs Q4: 51.9G11.6 mg/dl, P!0.001) and apolipoprotein A (Q1: 159.6G25.6 vs Q4: 147.9G24.1 mg/dl, P!0.01) compared with women in the lowest FAI quartile. These differences remained significant after adjustment for age, body mass index (BMI), insulin resistance and social habits. The free estrogen index (FEI) exhibited similar associations to the FAI. HOMA-IR showed an independent positive association with total testosterone (Q1: 2.00G1.36 vs Q4: 2.66G1.60, P!0.01), FAI (Q1: 1.70G1.12 vs Q4: 3.04G1.66, P!0.001) and FEI (Q1: 1.70G0.91 vs Q4: 3.08G1.77, P!0.001). Conclusions: Increased androgenicity in healthy postmenopausal women is associated with an unfavorable cardiovascular risk profile. High endogenous estradiol is related to a pro-atherogenic lipid profile, an association which may, in part, be mediated by insulin resistance. European Journal of Endocrinology Introduction Cardiovascular disease (CVD) is a leading cause of death of women worldwide. While premenopausal women have a lower incidence of CVD compared with men of the same age, the incidence of the disease in women rises steeply after the age of 50 (1). The association between menopause and CVD was indicated in the Framingham cohort, where postmenopausal women had a 2 6 times greater incidence of CVD compared with premenopausal women in the same age range (2). Furthermore, premature menopause caused by bilateral oophorectomy in younger women is an established risk factor for ischemic heart disease (3). Menopause is associated with a pro-atherogenic lipid profile characterized principally by lower high-density lipoprotein-cholesterol (HDL-cholesterol), higher low-density lipoproteincholesterol (LDL-cholesterol) and triglyceride levels (4), central adiposity (5), increased diastolic pressure (6) and increased insulin resistance (7). Sex steroid deficiency is considered a crucial factor responsible for the menopause-associated changes in the CVD risk factor profile, and consequently for the increase in CVD risk (8). Although the menopausal ovary does not secrete estrogens, it continues to serve as a source of androgens. To variable degrees, ovarian D 4 andostendione and testosterone are aromatized peripherally, mainly in the adipose tissue, to estrone and estradiol respectively, thus determining the postmenopausal endogenous estrogen milieu (9). In contrast to the established cardiovascular benefit of premenopausal sex steroids, little is known q 2006 Society of the European Journal of Endocrinology DOI: /eje Online version via

2 908 I Lambrinoudaki and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 about the effect of postmenopausal estrogens of endogenous origin on CVD risk. Concerning the association of endogenous estrogens with the lipid profile in postmenopausal women, studies have yielded conflicting results. Some authors have reported positive correlations between serum estrone or estradiol with HDL-cholesterol and triglycerides and inverse associations with total cholesterol and LDL-cholesterol (10 14), while others have reported higher total cholesterol (15) as well as lower triglycerides in women with higher endogenous estrogen levels (16). Finally, some investigators have found no significant association between estrogen and lipid levels (17 19). Moreover, although menopausal transition increases insulin resistance (20) and arterial pressure (21), no data exist as to whether fluctuations in circulating estradiol within the menopausal range have any impact on these parameters. The same discrepancy holds for homocysteine metabolism. It still remains to be clarified whether variation of endogenous estrogen levels may causatively inflict changes in homocysteine metabolism or whether this association is confounded by age or adiposity (22). The matter is further complicated when studying the effect of exogenous estrogens on CVD risk factors and incidence in postmenopausal women. Although ample evidence exists that exogenous estrogens improve lipid lipoprotein profile, insulin resistance, homocysteine metabolism, and vasodilation and that estrogens have anti-inflammatory properties at the vascular wall (21), the results of recent randomized controlled studies have revealed an increased risk of ischemic cardiac events in postmenopausal women treated with estrogen progestin therapy (23). Beyond estrogens, androgen status may also play an important role in determining the cardiovascular risk in postmenopausal women. The association of endogenous androgens with CVD risk factors and the incidence of cardiovascular events are even more perplexing. Androgen excess in premenopausal women, as is the case in polycystic ovary syndrome (PCOS), has been associated with increased triglycerides and small, dense LDL particles, as well as reduced HDL-cholesterol (24, 25). Furthermore, hyperandrogenemic women demonstrate increased insulin resistance and incidence of CVD (26, 27). In contrast to the established negative impact of premenopausal hyperandrogenicity on the cardiovascular system, little is known on the effect of postmenopausal endogenous androgens on CVD risk. Increased testosterone levels in healthy postmenopausal women have been related to a risk factor profile which resembles that of hyperandrogenemic PCOS women (28), although these findings have not been corroborated by others (29). On the contrary, several recent studies have demonstrated that postmenopausal women with reduced circulating testosterone have an increased incidence of CVD, independent of metabolic factors (30, 31). Similarly, postmenopausal women with testosterone levels in the highest tertile demonstrated significantly lower intimal-medial thickness of carotid wall compared with women in the lower testosterone tertiles (32). Given the discrepancy in data on the association of endogenous sex steroids with cardiovascular risk factors in postmenopausal women and due to the possibility that the knowledge of this association may aid the clinician in the evaluation of postmenopausal women who are candidates for estrogen or androgen therapy, we undertook the present study in order to evaluate the association of endogenous sex steroids with the lipid lipoprotein profile, insulin resistance, homocysteine levels and arterial pressure in healthy postmenopausal women of Greek origin. Subjects and methods Subjects Subjects were recruited from the Menopause Clinic of the 2nd Department of Obstetrics and Gynecology, University of Athens, Aretaieion Hospital. This study was part of a survey aiming to evaluate risk factors for menopause-associated diseases in community-dwelling Greek women on their first counseling visit to our Menopause Clinic. For this purpose 1278 women were recruited during the time period from November 1998 to June Data reported in this study are based on 598 women for whom hormone measurements were available. Participants were at least 1 year postmenopausal and were not current users of hormone therapy. Ex-users were not included in the study unless they had been off therapy for at least 6 months. All women had a gynecological and biochemical evaluation which included: bimanual examination, PAP smear and transvaginal sonography, breast examination and mammography, thyroid liver renal function as well as blood coagulation tests and bone densitometry. Criteria for inclusion in the study were a sonographically assessed endometrial thickness %5 mm and the absence of gynecological malignancy, arterial hypertension, ischemic heart disease, thromboembolism, diabetes mellitus and non-treated thyroid dysfunction. Women on lipid-lowering or antihypertensive medication were not included in the study. Finally, women with abnormally high testosterone or dehydroepiandrosterone sulfate (DHEAS), as well as women with triglyceride levels above 400 mg/dl, were also excluded. All subjects signed an informed consent, and Institutional Review Board approval was obtained by the Ethics Committee of Aretaieion Hospital. Protocol Participating women were evaluated in a crosssectional design. A detailed medical history was

3 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 recorded for every subject. Blood pressure, weight and height were measured in the morning, in light clothing, and body mass index (BMI) was computed. Subsequently, fasting blood samples were drawn at 0900 h for the determination of serum lipids, lipoproteins, homocysteine and hormones. Samples were centrifuged immediately and serum was stored at K80 8C until assayed. Biochemical assays Serum total cholesterol, HDL-cholesterol and triglycerides were assessed enzymatically by an autoanalyzer (COBAS-MIRA, Roche Diagnostics Limited, Lewes, East Sussex, UK). LDL-cholesterol was estimated as described by the Friedewald equation (LDL-cholesterolZtotal cholesterol triglycerides/5 HDL-cholesterol). Apolipoprotein A1 (ApoA), apolipoprotein B (ApoB) and lipoprotein a (Lp(a)) were determined by an immunoturbimetric assay (ABX Diagnostics BP Montpellier, France). The atherogenic index of plasma (AIP) was computed by the following equation: log [triglycerides (mmol/l)/hdl-cholesterol (mmol/l)] (34). Serum homocysteine levels were measured by the Abbott Imx Homocysteine kit (Abbott Laboratories, Abbott Park, IL, USA). The total coefficient of variation (CV) (%) and the sensitivity of the kit were 4.3% and 0.5 mmol/l respectively. Hormone assays Follicle-stimulating hormone (FSH) was measured with the Microparticle Enzyme Immunoassay kit: FSH, Abbott Axsym measured on the AxsymR analyzer (Abbott Laboratories). The total CV ranged from 5.3% to 8.5%. Estradiol was measured with a commercial Enzyme Immunoassay kit: DSK (Diagnostic Systems Laboratories Inc., Webster, TX, USA). The total CV ranged from 4.3% to 6.1% and the analytical sensitivity was 8 pg/ml. D 4 -Androstendione was measured with an IBL Androstendione ELISA kit (IBL GmbH, Hamburg, Germany). The inter-assay CV ranged from 6.5% to 8.1%. Total testosterone and DHEA-SO4 were measured with DPC Total Testosterone and DHEA-SO4 Immulite analyzer kits (Diagnostic Products Corporation, Los Angeles, CA, USA). The total CV ranged from 8.0% to 16.0% and from 8.1% to 15% respectively. Sex hormone-binding globulin (SHBG) concentrations were measured with a DPC Immulite SHBG chemiluminescent enzyme immunometric assay kit (Diagnostic Products Corporation). The total CV ranged from 4.1% to 9.2%. The free estrogen index (FEI) was calculated using total estradiol and SHBG values by the following equation: FEIZestradiol (pg/ml)! 0.367/SHBG (nmol/l). The free androgen index (FAI) was calculated using total testosterone and SHBG values by the following equation: FAIZtestosterone (ng/ml)!3.47!100/shbg (nmol/l). Insulin was Postmenopausal sex hormones and cardiovascular risk factors 909 measured by an Abbott Insulin kit on an IMx analyzer (Abbott Laboratories). The total CV ranged from 4.4% to 6.0%. Insulin resistance was calculated by homeostasis model assessment of insulin resistance (HOMA-IR), based on fasting insulin and glucose levels, using the following equation: HOMA-IRZinsulin (mu/ml)!glucose (mmol/l)/22.5. Statistical analysis Statistical analysis was performed by use of SPSS version 8.0 (Statistical Package for the Social Sciences, Chicago, IL, USA). Distributions of continuous variables were tested for normality by use of the Kolmogorov Smirnov test. Skewed variables were logarithmically transformed. Means of lipids, lipoproteins, atherogenic index of plasma (AIP), HOMA, systolic and diastolic arterial pressure and serum homocysteine were compared across quartiles of serum hormones by ANOVA and were tested for linear trends. Adjustments for possible confounding factors were performed by multiple regression analysis. Statistical significance was set at the 0.05 level. Results The baseline characteristics of the 598 women participating in the study are presented in Table 1. Triglycerides, Lp(a), FSH, sex hormones and insulin had a skewed distribution and thus were logarithmically transformed for further statistical analysis. Tables 2 and 3 present the mean levels of lipids, AIP, lipoproteins, systolic and diastolic arterial pressure, HOMA-IR and homocysteine according to hormone quartiles. Triglycerides, AIP, ApoB, blood pressure and HOMA-IR decreased while HDL and ApoA increased from the lower to the highest FSH quartile. Women in quartile 1 had mean AIP levels of compared with in women in the highest quartile (P for linear trend Z0.0001). Accordingly, women in the first quartile had mean HOMA-IR levels of 2.74 compared with 1.82 in women in the 4th quartile (P for trend Z0.0001). In contrast to exogenously administered estrogen, the FEI associated positively with AIP, ApoB, blood pressure and HOMA-IR. Women in the lowest FEI quartile had mean AIP levels of compared with in women in the highest FEI quartile (P for linear trend Z0.0001). Accordingly, mean HOMA-IR levels were 1.70 in women in the lowest FEI quartile compared with 3.08 in women in the highest FEI quartile (P for linear trend Z0.0001). Testosterone showed a strong positive association with total and LDL-cholesterol (Fig. 1), triglycerides, AIP, ApoB and HOMA-IR, and a negative association with HDLcholesterol and ApoA (Tables 2 and 3). Mean AIP was K0.224 in the lowest testosterone quartile compared with K0.087 in the highest quartile (P for linear trend Z0.001). Accordingly, mean HOMA-IR levels increased from 2.00 in the lowest to 2.66 in the highest

4 910 I Lambrinoudaki and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 Table 1 Baseline demographic characteristics of the 598 healthy postmenopausal women who participated in the study. Continuous variables Mean S.D. Min Max Age (years) YSM BMI (kg/m 2 ) Systolic AP Diastolic AP Cholesterol (mg/dl) Triglycerides (mg/dl) HDL (mg/dl) LDL (mg/dl) Lp(a) (mg/dl) ApoA1 (mg/dl) ApoB (mg/dl) FSH (miu/ml) Estradiol (pg/ml) Testosterone (ng/ml) D 4 A (ng/dl) DHEAS (ng/dl) SHBG (nmol/l) HOMA-IR Homocysteine (mmol/l) Categorical variables Smoking (current, O10 22 cigarettes/day) (%) Daily alcohol intake (%) 0.8 Physical exercise (O 1 h/week) (%) 18.3 YSM: years since menopause, BMI: body mass index; AP: arterial pressure; D 4 A: D 4 -Androstendione; DHEAS: dehydroepindrosterone sulfate; SHBG: sex hormone-binding globulin; HOMA-IR: homeostasis model assessment of insulin resistance. testosterone quartile (P for linear trendz0.01). The FAI exhibited the same associations as total testosterone (Figs 1and 2). Mean AIP increasedfrom K0.232 in the lowest FAI quartile to K0.078 in the highest FAI quartile (P for linear trendz0.0001). Likewise, mean HOMA-IR increased from 1.70 in the lowest quartile to 3.04 in the highest (P for linear trendz0.0001). SHBG exhibited strong negative associations with triglycerides, AIP, ApoB, systolic and diastolic arterial pressure and HOMA-IR (Fig. 2). There was a significant decrease in AIP (K0.104 in quartile 1 vs K0.248 in quartile 4, P for linear trendz0.0001), in triglycerides (106.3 mg/dl in quartile 1 vs 82.6 mg/dl in quartile 4, P for linear trendz0.001), in ApoB (111.4 mg/dl in quartile 1 vs mg/dl in quartile 4, P for linear trendz0.02), in systolic arterial pressure (AP) (122.6 mmhg in quartile 1 vs mmhg in quartile 4, P for linear trendz0.005), in diastolic AP (79.5 mmhg in quartile 1 vs 70.5 mmhg in quartile 4, P for linear trendz0.001) and in HOMA-IR (3.11 in quartile 1 vs 1.54 in quartile 4, P for linear trendz ) by moving from the lower to the higher SHBG quartile. In contrast, there was an increase in HDL (53.4 mg/dl in quartile 1 vs 61.5 mg/dl in quartile 4, P for linear trendz0.0001) and ApoA (151.7 mg/dl in quartile 1 vs mg/dl in quartile 4, P for linear trendz0.002) across increasing SHBG quartiles. D 4 A and DHEAS did not associate with lipids, lipoproteins, arterial pressure, homocysteine or HOMA-IR. The results of the multivariate analysis on the relation between hormones and lipids, lipoproteins, arterial pressure, HOMA-IR and homocysteine are presented in Table 4. Besides hormones, the model included age, BMI, current cigarette smoking, alcohol consumption and physical exercise as independent variables. After adjustments for the above-mentioned covariates, FSH showed a significant negative association with triglycerides, AIP, diastolic arterial pressure and HOMA-IR and a positive association with HDLcholesterol and ApoA. Furthermore, while estradiol showed a negative association only with Lp(a), FEI associated marginally positively with triglycerides, significantly positively with AIP and HOMA-IR, and negatively with HDL and ApoA. Both testosterone and FAI showed a strong positive association with total and LDL-cholesterol, triglycerides, AIP, ApoB and HOMA-IR and a negative association with HDL-cholesterol, Lp(a) (testosterone) and ApoA (FAI) (Table 4). Contrariwise, SHBG associated negatively with triglycerides, AIP, ApoB, systolic and diastolic arterial pressure and HOMA-IR and positively with HDL-cholesterol and ApoA. D 4 A associated positively with LDL-cholesterol and AIP and negatively with HDLcholesterol, but these associations were weaker compared with those of testosterone. DHEAS did not associate with lipids, lipoproteins, arterial pressure, HOMA-IR or homocysteine. Discussion The results of our study suggest that healthy postmenopausal women with a higher index of androgenicity, indicated by higher total testosterone, higher FAI and lower SHBG, exhibit a pro-atherogenic risk profile, characterized by lower HDL-cholesterol and higher triglycerides, AIP, insulin resistance and diastolic blood pressure. Our findings are in concordance with previous reports examining the association between serum androgens and lipid profile in postmenopausal women. Haffner et al. (35), in a population study of 253 community-dwelling postmenopausal women, concluded that total testosterone correlated positively with total cholesterol, while free testosterone related inversely to HDL-cholesterol. Furthermore, Shelley et al. reported a positive independent association of the FAI with LDL-cholesterol, but not with HDL-cholesterol or triglycerides (10). It has to be borne in mind, however, that Shelley et al. evaluated menopausal transition, and their population, therefore, comprised both pre- and postmenopausal women. In a small sample of Japanese postmenopausal women, free testosterone predicted total and LDL-cholesterol, HDL-cholesterol

5 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 Postmenopausal sex hormones and cardiovascular risk factors 911 Table 2 Mean (S.D.) levels of serum lipids, the atherogenic index of plasma (AIP) and lipoprotein (a) according to hormone quartiles (Q) in 598 healthy postmenopausal women. TC (mg/dl) Triglycerides (mg/dl) LDL (mg/dl) HDL (mg/dl) AIP Lipoprotein (a) (mg/dl) FSH (miu/ml) Q1 ( ) (44.5) (52.6) (40.3) 55.9 (12.8) K0.125 (0.262) 16.6 (16.4) Q2 ( ) (37.8) (49.8) (34.6) 57.9 (12.5) K0.157 (0.247) 18.8 (21.5) Q3 ( ) (38.8)) 92.9 (40.9) (37.7) 58.2 (13.4) K0.173 (0.230) 23.1 (24.5) Q4 ( ) (43.1) 85.2 (38.3)*** (39.6) 62.8 (14.3)*** (0.230)*** 15.8 (15.7) Estradiol (pg/ml) Q1 ( ) (40.4) 94.7 (46.4) (38.4) 60.5 (14.6) K0.188 (0.254) 23.9 (25.5) Q2 ( ) (42.7) 97.9 (46.7) (39.1) 59.8 (13.0) K0.179 (0.240) 16.1 (17.1) Q3 ( ) (43.0) (53.5) (38.6) 55.4 (12.5) K0.135 (0.251) 15.9 (17.6) Q4 ( ) (38.9) 92.1 (39.8) (35.6)* 57.5 (13.3)* K0.189 (0.243) 17.7 (17.7)* Testosterone (ng/ml) Q1 ( ) (41.3) 86.5 (43.1) (37.2) 60.7 (14.5) K0.224 (0.238) 19.3 (21.4) Q2 ( ) (37.5) 93.6 (43.2) (34.6) 58.4 (12.3) K0.186 (0.223) 18.1 (23.5) Q3 ( ) (45.0) 91.1 (36.7) (40.9) 59.2 (13.9) K0.195 (0.227) 13.0 (14.7) Q4 ( ) (38.4)** (56.4)** (35.3)*** 52.9 (13.0)** K0.087 (0.254)** 16.6 (20.8) D 4 A (ng/dl) Q1 ( ) (45.2) 95.6 (46.5) (41.2) 59.6 (14.1) K0.199 (0.234) 15.3 (20.9) Q2 ( ) (37.7) 92.8 (40.0) (36.2) 57.1 (11.9) K0.176 (0.240) 14.0 (16.2) Q3 ( ) (40.3) 93.3 (45.8) (38.8) 55.2 (13.2) K0.158 (0.253) 13.2 (14.1) Q4 ( ) (41.9) 93.1 (37.8) (34.1) 56.8 (14.1) K0.169 (0.215) 18.2 (19.8) DHEAS (ng/dl) Q1 ( ) (40.1) (57.5) (35.4) 58.9 (12.1) K0.171 (0.258) 19.9 (24.0) Q2 ( ) (39.1) 93.7 (39.3) (38.4) 59.4 (13.3) K0.180 (0.238) 17.7 (18.4) Q3 ( ) (37.6) 98.9 (47.4) (36.1) 56.4 (13.0) K0.150 (0.251) 12.3 (13.8) Q4 ( ) (42.0) 95.6 (42.1) (37.3) 57.6 (13.3) K0.165 (0.235) 21.3 (20.0) SHBG (nmol/l) Q1 ( ) (41.1) (50.6) (36.9) 53.4 (12.2) K0.104 (0.229) 18.6 (21.1) Q2 ( ) (45.4) 94.8 (44.3) (43.5) 58.2 (13.6) K0.177 (0.255) 21.0 (25.0) Q3 ( ) (39.9) 92.9 (45.8) (36.3) 59.2 (13.6) K0.192 (0.238) 14.1 (16.5) Q4 ( ) (36.6) 82.6 (27.1)** (34.5) 62.5 (14.6)*** K0.248 (0.213)*** 16.0 (18.5) FAI Q1 ( ) (42.4) 86.5 (44.4) (37.7) 62.0 (15.7) K0.232 (0.254) 16.0 (18.1) Q2 ( ) (33.8) 88.7 (31.5) (31.4) 60.7 (11.6) K0.214 (0.203) 19.8 (24.8) Q3 ( ) (44.5) 92.3 (42.0) (44.1) 58.1 (13.5) K0.189 (0.230) 12.7 (12.9) Q4 ( ) (41.9)* (53.8)** (36.2)** 51.9 (11.6)*** K0.078 (0.247)*** 16.9 (21.9) FEI Q1 ( ) (38.0) 86.6 (31.2) (36.4) 64.0 (14.5) K0.237 (0.213) 20.8 (27.0) Q2 ( ) (40.4) 87.6 (43.2) (36.0) 60.6 (14.5) K0.224 (0.238) 16.4 (19.7) Q3 ( ) (42.1) (53.4) (40.1) 54.5 (11.2) K0.123 (0.233) 17.3 (19.4) Q4 ( ) (42.0) 95.5 (39.7) (38.4) 55.5 (13.0)*** K0.153 (0.252)*** 15.9 (14.8) *P!0.05, **P!0.01, ***P!0.001 (P for linear trend, ANOVA). D 4 A: D 4 -Androstendione; DHEAS: dehydroepindrosterone sulfate; SHBG: sex hormonebinding globulin. and triglycerides independently of insulin resistance and BMI (18). Similarly, bioavailable testosterone has been reported as an independent predictor of VLDL-TG (36). Very recently, a study conducted on a selected sample from the Atherosclerosis Risk in Communities (ARIC) population (15) revealed that the FAI was positively associated with higher LDL-cholesterol levels independently of demographic and metabolic factors and health behaviors. In contrast to total and free testosterone, SHBG, in our study as well as in other studies, associated with a favorable lipid profile. SHBG is secreted by the liver and binds testosterone, dihydrotestosterone and estradiol with high affinity, thus regulating their free concentrations (37). Haffner et al. (38) have hypothesized that SHBG may reflect intracellular bioavailable testosterone better than total testosterone, with reduced SHBG expressing greater androgenicity. Insulin is also considered an important inhibitor of SHBG production (37). SHBG has been shown consistently to relate positively to HDL-cholesterol (15, 35, 39) and negatively to triglycerides, total and LDL-cholesterol (15), as well as to ApoB (18). Contrariwise, two small studies have not found an association between endogenous testosterone or SHBG and lipid profile in postmenopausal women (19, 39). Our study is the first to demonstrate a positive association between the AIP and indexes of androgenicity in postmenopausal women. The AIP, defined as the logarithm of the ratio TG:HDL-cholesterol, has recently been proposed as a marker of the atherogenic potential of plasma. The significance of the AIP as a marker is based on the fact that (a) it is found increased in cohorts at high risk for coronary artery disease, (b) it is

6 912 I Lambrinoudaki and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 Table 3 Mean levels of serum apolipoproteins, systolic and diastolic adrenal pressure (AP), HOMA-IR and homocysteine (Hcy) according to hormone quartiles in 598 healthy postmenopausal women. ApoA1 (mg/dl) ApoB (mg/dl) Systolic AP Diastolic AP HOMA-IR Hcy (mg/dl) FSH (miu/ml) Q1 ( ) (28.9) (25.9) (9.5) 79.8 (8.1) 2.74 (1.84) 11.2 (2.78) Q2 ( ) (27.8) (25.2) (10.1) 75.2 (7.7) 2.33 (1.33) 10.7 (2.28) Q3 ( ) (25.3) (25.1) (10.1) 75.3 (10.1) 2.20 (1.28) 11.0 (2.52) Q4 ( ) (27.8)** (25.3)* (7.2)* 73.1 (8.8)* 1.82 (1.00)*** 11.5 (2.94) Estradiol (pg/ml) Q1 ( ) (31.6) (25.9) (8.85) 79.4 (8.27) 2.05 (1.22) 10.1 (2.30) Q2 ( ) (25.8) (24.8) (9.92) 74.7 (9.60) 2.24 (1.49) 11.4 (2.80) Q3 ( ) (25.9) (27.7) (10.6) 73.8 (8.75) 2.22 (1.04) 11.0 (2.07) Q4 ( ) (26.8) (24.0) (8.13) 76.9 (8.45) 2.51 (1.67)* (2.85) Testosterone (ng/ml) Q1 ( ) (25.6) (23.1) (11.2) 75.0 (11.0) 2.00 (1.36) 11.6 (2.72) Q2 ( ) (28.0) (20.8) (10.0) 78.1 (6.67) 2.06 (1.00) 11.4 (2.69) Q3 ( ) (23.6) (28.7) (7.61) 75.0 (7.40) 2.14 (1.09) 11.2 (2.37) Q4 ( ) (22.7)* (23.8)*** (5.77) 72.5 (10.6) 2.66 (1.60)** 10.8 (2.48) D 4 A (ng/dl) Q1 ( ) (24.0) (25.4) (6.19) 74.4 (8.73) 2.21 (1.33) 11.2 (2.68) Q2 ( ) (24.6) (23.8) (9.35) 74.4 (10.9) 2.32 (1.66) 11.2 (2.44) Q3 ( ) (23.3) (25.9) (9.81) 74.4 (8.92) 2.18 (1.49) 10.9 (2.42) Q4 ( ) (28.0) (26.0) (11.7) 76.7 (9.63) 2.49 (1.35) 11.8 (3.01) DHEAS (ng/dl) Q1 ( ) (27.7) (26.8) (10.8) 73.7 (10.0) 2.15 (1.62) 11.0 (2.50) Q2 ( ) (29.6) (24.1) (9.5) 76.2 (10.2) 2.23 (1.55) 11.3 (2.86) Q3 ( ) (27.0) (26.1) (9.67) 73.3 (9.07) 2.43 (1.42) 11.0 (2.13) Q4 ( ) (27.5) (26.5) (9.63) 77.7 (7.52) 2.25 (1.20) 11.6 (2.83) SHBG (nmol/l) Q1 ( ) (25.3) (24.8) (8.89) 79.5 (8.05) 3.11 (1.76) 10.7 (2.73) Q2 ( ) (22.9) (24.6) (7.99) 77.3 (8.20) 2.25 (1.22) 11.2 (2.40) Q3 ( ) (27.8) (23.7) (6.50) 74.2 (9.10) 1.97 (0.85) 11.1 (2.35) Q4 ( ) (24.9)** (26.0)* (11.4)** 70.5 (9.74)** 1.54 (0.75)*** 11.3 (2.95) FAI Q1 ( ) (25.6) (25.5) (11.1) 71.9 (9.87) 1.70 (1.12) 11.6 (2.91) Q2 ( ) (30.3) (23.2) (8.79) 77.0 (8.60) 1.83 (0.88) 11.4 (2.43) Q3 ( ) (19.2) (24.9) (7.99) 77.3 (9.61) 2.23 (1.08) 10.7 (1.92) Q4 ( ) (24.1)** (25.8)** (8.12) 76.7 (8.17) 3.04 (1.66)*** 11.1 (2.76) FEI Q1 ( ) (25.7) (24.5) (10.7) 71.5 (9.88) 1.70 (0.906) (2.88) Q2 ( ) (24.7) 99.9 (23.5) (9.0) 75.3 (10.6) 1.96 (0.990) (2.76) Q3 ( ) (20.8) (24.9) (8.5) 74.1 (8.61) 2.17 (1.17) (1.98) Q4 ( ) (27.4)*** (26.1)* (8.8)* 79.2 (7.02)** 3.08 (1.77)*** (2.85) *P!0.05, **P!0.01, ***P!0.001 (P for linear trend, ANOVA). D 4 A: D 4 -Androstendione; DHEAS: dehydroepindrosterone sulfate; SHBG: sex hormonebinding globulin; HOMA-IR: Homeostasis Model Assessment of Insulin Resistance. positively correlated with the fractional esterification rate of HDL-cholesterol, which is perhaps the most dependable marker for the atherogenic capacity of the lipid lipoprotein profile, and (c) it is inversely correlated to LDL-cholesterol particle size (40, 41). In our study, higher total testosterone and FAI levels as well as lower SHBG were associated with increased insulin resistance. Free testosterone has been reported to associate positively with fasting plasma glucose and insulin levels (42 44), as well as with the insulin/ glucose ratio (45). In the more recent evaluation of the Rancho Bernando cohort, Oh et al. (42) showed that, after adjusting for age, adiposity and systolic blood pressure, elevated free testosterone predicted longitudinally higher insulin levels and HOMA-IR as well as incident type 2 diabetes mellitus. Furthermore, Golden et al., studying the ARIC population (46), concluded that hyperandrogenism in postmenopausal women most strongly associated with the glucose and insulin component of the metabolic syndrome. SHBG showed an independent negative association with diastolic blood pressure in our study. Earlier reports have indicated that androgen levels are higher in women with essential hypertension (47). Haffner et al. (35) reported a positive association between endogenous total testosterone and systolic and diastolic blood pressure in postmenopausal women. Furthermore, hypertensive postmenopausal women are reported to have higher mean testosterone and free-to-total testosterone ratio compared with normotensive women (48). The same group demonstrated a significant association between free testosterone and blood pressure in healthy and diabetic postmenopausal women (49). Finally, Golden et al. reported that healthy postmenopausal women

7 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 Postmenopausal sex hormones and cardiovascular risk factors 913 r = 0.46, P = r = 0.19, P = r = 0.23, P = r = 0.44, P = Figure 1 Linear regression analysis of LDL-cholesterol to lntransformed serum testosterone (top) and of the atherogenic index of plasma (AIP) to the ln-transformed free androgen index (FAI) (bottom). Figure 2 Linear regression analysis of ln-transformed Homeostasis Model Assessment of Insulin Resistance (HOMA-IR) to lntransformed serum sex hormone-binding globulin (SHBG) (top) and ln-transformed free androgen index (FAI) (bottom). of the ARIC population with an FAI in the highest quartile were nine times more likely to present with the metabolic syndrome compared with women in the lowest FAI quartile (46). In keeping with this finding, results from the SWHAN study, based on 2961 perimenopausal women, indicated that SHBG was associated strongly and inversely with the metabolic syndrome (50). From all the above information it is apparent that increased androgenicity in postmenopausal women is associated with an increased risk for CVD. Contrary to this assumption, however, are the results of other studies, suggesting that postmenopausal women with low testosterone levels have increased carotid atherosclerosis (32) and increased incidence of coronary events (30, 31). A possible explanation for this discrepancy is that optimal cardiovascular function requires a certain range of circulating testosterone. Lower levels of testosterone may compromise endothelial function and blood viscosity (46), while higher levels may be associated with a constellation of atherogenic risk factors, namely increased visceral adiposity, increased blood pressure, insulin resistance and a pro-atherogenic lipid profile, as demonstrated in our study. It should be remembered, however, that testosterone in postmenopausal women is a precursor for estradiol and it may actually be estrogen deficiency that hides behind the association of low testosterone and increased CVD incidence. In keeping with the effect of higher androgens, higher estradiol levels associated with higher triglycerides, AIP, ApoB and lower HDL-cholesterol and ApoA. In keeping with our results, Mudali et al. (15) recently reported a strong positive association between estrone and total cholesterol and triglycerides among postmenopausal women with significant carotid atherosclerosis. Contrariwise, Ossewaarde et al. (16) demonstrated a significant

8 914 I Lambrinoudaki and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 Table 4 Linear regression analysis evaluating the association between serum lipids, lipoproteins, the atherogenic index of plasma, arterial pressure, HOMA-IR and serum hormones in 598 healthy postmenopausal women. The model is adjusted for age, BMI, cigarette smoking, alcohol intake and physical exercise. Ln-transformed hormones b coefficient P value Ln-transformed hormones b coefficient P value Total cholesterol Apolipoprotein A FSH K FSH Estradiol Estradiol K Testosterone Testosterone K D 4 A D 4 A DHEAS DHEAS SHBG K SHBG FAI FAI K FEI FEI K Triglycerides Apolipoprotein B FSH K FSH K Estradiol K Estradiol K Testosterone Testosterone D 4 A D 4 A DHEAS DHEAS SHBG K SHBG K FAI FAI FEI FEI LDL cholesterol Systolic AP FSH K FSH K Estradiol Estradiol K Testosterone Testosterone K D 4 A D 4 A DHEAS DHEAS SHBG K SHBG K FAI FAI FEI FEI HDL cholesterol Diastolic AP FSH FSH K Estradiol K Estradiol K Testosterone K Testosterone K D 4 A K D 4 A DHEAS K DHEAS SHBG SHBG K FAI K FAI FEI K FEI AIP HOMA-IR FSH K FSH K Estradiol K Estradiol Testosterone Testosterone D 4 A D 4 A DHEAS DHEAS K SHBG K SHBG K FAI FAI FEI FEI Lp(a) Homocysteine FSH K FSH K Estradiol K Estradiol Testosterone K Testosterone K D 4 A D 4 A DHEAS DHEAS SHBG K SHBG K FAI K FAI K FEI K FEI D 4 A: D 4 -Androstendione; DHEAS: dehydroepindrosterone sulfate; SHBG: sex hormone-binding globulin; HOMA-IR: homeostasis model assessment of insulin resistance; AP: arterial pressure. positive association between plasma estrone and HDLcholesterol, and a negative association with triglycerides and very low-density lipoprotein-cholesterol in a small sample of healthy postmenopausal women. Finally, other investigators found no significant association between endogenous estrogens and lipid levels or other markers of atherosclerosis (10, 17 19, 39, 51). Variability in the health status of the study population,

9 EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) 154 as well as in the estrogen measured, is a possible reason for the discordant results. Many studies employed estrone as a marker of endogenous estrogen levels, which may be more representative in the postmenopausal population compared with estradiol, which is the principal premenopausal sex steroid. In our sample, 6% of the women had estradiol levels below 8 pg/ml, which is the analytical sensitivity of the assay employed, while 51% of the sample had estradiol values lying between 8 pg/ml and 20 pg/ml. Unfortunately, an ultra-sensitive estradiol assay with sensitivity below 5 pg/ml was not available. Furthermore, metabolic parameters that were not accounted for in the previous studies may have confounded the associations of lipids and lipoproteins with endogenous estrogens. In our study, adjustment for HOMA-IR removed many of the associations between atherogenic risk factors and endogenous estradiol, thus indicating that these associations are not causative but are probably mediated through metabolic factors insulin resistance being one of them. Certain limitations should be kept in mind when interpreting our data. First, our study was conducted on a clinic-based sample, a fact implying that our population may have been more motivated and health concerned compared with the community-dwelling women. Secondly, as pointed out above, we were not able to measure estrone, the predominant circulating estrogen in postmenopausal women. Thirdly, in being cross-sectional in design, our study does not permit the inference of causal associations between sex steroids and cardiovascular risk factors. The present study has several strengths. We were able simultaneously to examine the association of multiple cardiovascular risk factors such as lipids, lipoproteins, insulin resistance, blood pressure and homocysteine with both androgen and estrogen status in a large sample, comprising healthy postmenopausal women. As is the case with premenopausal women with androgen excess, endogenous androgenicity in postmenopausal women seems to relate to an unfavorable cardiovascular risk profile. Since androgen replacement therapy is increasingly being used for the treatment of sexual dysfunction and decreased well-being in postmenopausal women, its impact on lipid and metabolic parameters related to cardiovascular health should be studied thoroughly. References 1 Matthews KA, Meilahn E, Kuller LH, Kelsey SF, Caggiula AW & Wing RR. Menopause and risk factors for coronary heart disease. The New England Journal of Medicine Kannel WB, Hjortland MC, McNamara PM & Gordon T. Menopause and risk of cardiovascular disease. The Framingham study. Annals of Internal Medicine Postmenopausal sex hormones and cardiovascular risk factors Barrett-Connor E & Bush TL. Estrogen and coronary heart disease in women. The Journal of the American Medical Association Warren PM & Halpert S. Hormone replacement therapy: controversies, pros and cons. Best Practice and Research Clinical Endocrinology and Metabolism Avgoulea A, Mastorakos G, Lambrinoudaki I, Christodoulakos G & Creatsas G. Role of postmenopausal hormone replacement therapy on body fat gain and leptin levels. Gynecological Endocrinology Reckelhoff JF. Basic research into the mechanisms responsible for postmenopausal hypertension. International Journal of Clinical Practice (Suppl) Wu SI, Chou P & Tsai ST. The impact of years since menopause on the development of impaired glucose tolerance. Journal of Clinical Epidemiology Liu Y, Ding J, Bush TL, Longenecker JC, Nieto FJ, Golden SH & Szklo M. Relative androgen excess and increased cardiovascular risk after menopause: a hypothesized relation. American Journal of Epidemiology Schiff I. Menopause. In Principles and Practice of Endocrinology and Metabolism. 2nd edn, pp Ed. KL Becker. Philadelphia, PA: JB Lippincott, Shelley JM, Green A, Smith AM, Dudley E, Dennerstein L, Hopper J & Burger H. Relationship of endogenous sex hormones to lipids and blood pressure in mid-aged women. Annals of Epidemiology Kuller LH, Gutai JP, Meilahn E, Matthews KA & Plantinga P. Relationship of endogenous sex steroid hormones to lipids and apoproteins in postmenopausal women. Arteriosclerosis Svendsen OL, Hassager C & Christiansen C. Relationships and independence of body composition, sex hormones, fat distribution and other cardiovascular risk factors in overweight postmenopausal women. International Journal of Obesity and Related Metabolic Disorders Soler JT, Folsom AR, Kaye SA & Prineas RJ. Associations of abdominal adiposity, fasting insulin, sex hormone binding globulin, and estrone with lipids and lipoproteins in postmenopausal women. Atherosclerosis Pasquali R, Vicennati V, Bertazzo D, Casimirri F, Pascal G, Tortelli O & Labate AM. Determinants of sex hormone-binding globulin blood concentrations in premenopausal and postmenopausal women with different estrogen status, Virgilio-Menopause-Health Group. Metabolism Mudali S, Dobs AS, Ding J, Jingzhong D, Cauley JA, Moyses S & Golden SH. Endogenous postmenopausal hormones and serum lipids: the Atherosclerosis Risk in Communities Study. Journal of Clinical Endocrinology and Metabolism Ossewaarde ME, Bots ML, van der Schouw YT, Thijssen JHH, Westerveld T, de Jong FH & Grobbee DE. Plasma and urinary sex hormones are differently related to lipids in healthy postmenopausal women. Maturitas Cauley JA, Gutai JP, Glynn NW, Paternostro-Bayles M, Cottington E & Kuller LH. Serum estrone concentrations and coronary artery disease in postmenopausal women. Arteriosclerosis and Thrombosis Kumagai S, Kai Y & Sasaki H. Relationship between insulin resistance, sex hormones and sex hormone-binding globulin in the serum lipid and lipoprotein profiles of Japanese postmenopausal women. Journal of Atherosclerosis and Thrombosis Cauley JA, Gutai JP, Kuller LH & Powell JG. The relation of endogenous sex steroid hormone concentrations to serum lipid and lipopoprotein levels in postmenopausal women. American Journal of Epidemiology Christodoulakos G, Lambrinoudaki I, Panoulis C, Sioulas V, Rizos D, Caramalis G, Botsis D & Creatsas G. Serum androgen levels and insulin resistance in postmenopausal women: association with hormone therapy, tibolone and raloxifene. Maturitas

10 916 I Lambrinoudaki and others EUROPEAN JOURNAL OF ENDOCRINOLOGY (2006) Dubey RK, Imthurn B, Barton M & Jackson EK. Vascular consequences of menopause and hormone therapy: importance of timing of treatment and type of estrogen. Cardiovascular Research Christodoulakos GE, Panoulis CP, Rizos D, Moustakarias T, Phocas I & Creatsas G. Homocysteine and folate levels in postmenopausal women. Maturitas The Writing Group for the Women s Health Initiative investigators. Risks and benefits of estrogen plus progestin in healthy postmenopausal women. The Journal of the American Medical Association Rajkhowa M, Neary RH, Kumpatla P, Game FL, Jones PW, Obhrai MS & Clayton RN. Altered composition of high density lipoproteins in women with the polycystic ovary syndrome. Journal of Clinical Endocrinology and Metabolism Pirwany IR, Fleming R, Greer IA, Packard CJ & Sattar N. Lipids and lipoprotein subfractions in women with PCOS: relationship to metabolic and endocrine parameters. Clinical Endocrinology Rexrode KM, Manson JE, Lee IM, Ridker PM, Sluss PM, Cook NR & Buring JE. Sex hormone levels and risk of cardiovascular events in postmenopausal women. Circulation Reinecke H, Bogdanski J, Woltering A, Breithardt G, Assmann G, Kerber S & von Eckardstein A. Relation of serum levels of sex hormone binding globulin to coronary heart disease in postmenopausal women. American Journal of Cardiology Wild S, Pierpoint T, McKeigue P & Jacobs H. Cardiovascular disease in women with polycystic ovary syndrome at long-term follow-up: a retrospective cohort study. Clinical Endocrinology Bernini GP, Sgro M, Moretti A, Argenio GF, Barlascini CO, Cristofani R & Salvetti A. Endogenous androgens and carotid intimal-medial thickness in women. Journal of Clinical Endocrinology and Metabolism Kaczmarek A, Reczuch K, Majda J, Banasiak W & Ponikowski P. The association of lower testosterone level with coronary artery disease in postmenopausal women. International Journal of Cardiology Rako S. Testosterone deficiency: a key factor in the increased cardiovascular risk to women following hysterectomy or with natural aging? Journal of Women s Health Bernini GP, Moretti A, Sgro M, Argenio GF, Barlascini CO, Cristofani R & Salvetti A. Influence of endogenous androgens on carotid wall in postmenopausal women. Menopause Hoffmann GE, Hiefinger R & Weiss L. Five methods for measuring low-density lipoprotein cholesterol concentration in serum compared. Clinical Chemistry Tan MH, Johns D & Glazer BN. Pioglitazone reduces atherogenic index of plasma in patients with type 2 diabetes. Clinical Chemistry Haffner SM, Newcomb PA, Marcus PM, Klein BE & Klein R. Relation of sex hormones and dehydroepiandrosterone sulfate (DHEA-SO4) to cardiovascular risk factors in postmenopausal women. American Journal of Epidemiology Van Beek AP, de Ruijter-Heijstek FC, Jansen H, Erkelens DW & de Bruin TW. Sex steroids and plasma lipoprotein levels in healthy women: The importance of androgens in the estrogen-deficient state. Metabolism Rosner W. Plasma steroid-binding proteins. Endocrinology and Metabolism Clinics of North America Haffner SM, Katz MS & Dunn JF. Increased upper body and overall adiposity is associated with decreased sex hormone binding globulin in postmenopausal women. International Journal of Obsesity Noyan V, Yucel A & Sagsoz N. The association of androgenic sex steroids with serum lipid levels in postmenopausal women. Acta Obstetricia et Gynecologica Scandinavica Dobiasova M & Frohlich J. The plasma parameter log (TG/HDL-C) as an atherogenic index: correlation with lipoprotein particle size and esterification rate in ApoB-lipoprotein-depleted plasma (FER HDL ). Clinical Biochemistry Cleeman J. Executive summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Journal of the American Medical Association Oh JY, Barrett-Connor E, Wedick NM & Wingard DL. Endogenous sex hormones and the development of type 2 diabetes in older men and women: the Rancho Bernardo Study. Diabetes Care Goodman-Gruen D & Barrett-Connor E. Sex differences in the association of endogenous sex hormone levels and glucose tolerance status in older men and women. Diabetes Care Khaw KT & Barrett-Connor E. Fasting plasma glucose levels and endogenous androgens in non-diabetic postmenopausal women Clinical Science Maturana MA & Spritzer PM. Association between hyperinsulinemia and endogenous androgen levels in peri- and postmenopausal women. Metabolism Golden S, Ding J, Szklo M, Schmidt MI, Duncan BB & Dobs A. Glucose and insulin components of the metabolic syndrome are associated with hyperandrogenism in postmenopausal women. The Atherosclerosis Risk in Communities Study. American Journal of Epidemiology Hughes GS, Mathur RS & Margolius HS. Sex steroid hormones are altered in essential hypertension. Journal of Hypertension Phillips GB, Pinkernell BH & Jing TY. Relationship between sex hormones and coronary artery disease in postmenopausal women. Arteriosclerosis, Thrombosis, and Vascular Biology Phillips GB, Tuck CH, Jing TY, Boden-Albala B, Lin IF, Dahodwala N & Sacco RL. Association of hyperandrogenemia and hyperestrogenemia with type 2 diabetes in Hispanic postmenopausal women. Diabetes Care Santoro N, Torrens J, Crawford S, Allsworth JE, Finkelstein JS, Gold EB, Korenman S, Lasley WL, Luborsky JL, McConnell D, Sowers MF & Weiss G. Correlates of circulating androgens in mid-life women: the study of women s health across the nation. Journal of Clinical Endocrinology and Metabolism Golden SH, Maguire A, Ding J, Crouse JR, Cauley JA, Zacur H & Szklo M. Endogenous postmenopausal hormones and carotid atherosclerosis: a case-control study of the Atherosclerosis Risk in Communities Cohort. American Journal of Epidemiology Received 10 January 2006 Accepted 20 March 2006