Polycystic Ovary Syndrome

POLYCYSTIC OVARY SYNDROME

Received A

6, 2008;

H.J.T. has

nothing

on the C

Solvay P

This work

Cardiov

tralia an

Health

L.J.M. r

Presented

church,

crine So

2008.

Reprint re

Women

Locked

toria 316

monash

FC

184

Endothelial function and insulin resistance inpolycystic ovary syndrome: the effects of medicaltherapyHelena J. Teede, Ph.D.,a,b Caroline Meyer, Ph.D.,b Samantha K. Hutchison, M.B., B.S.,a

Sophia Zoungas, Ph.D.,a,b Barry P. McGrath, M.D.,c and Lisa J. Moran, Ph.D.a

a The Jean Hailes Foundation for Women’s Health, Monash Institute of Health Services Research, Monash University,

Melbourne; b Diabetes Unit, Southern Health, Melbourne; and c Department of Vascular Sciences and Medicine, Centre for

Vascular Health, Monash University, Clayton, Victoria, Australia

Objective: To assess the interaction between insulin resistance and endothelial function and the optimal treatmentstrategy addressing cardiovascular risk in polycystic ovary syndrome.Design: Randomized controlled trial.Setting: Controlled clinical study.Patient(s): Overweight age- and body mass index–matched women with polycystic ovary syndrome.Intervention(s): Six months metformin (1 g two times per day, n ¼ 36) or oral contraceptive pill (OCP) (35 mgethinyl E2–2 mg cytoproterone acetate, n ¼ 30).Main Outcome Measure(s): Fasting and oral glucose tolerance test glucose and insulin levels, endothelial function(flow-mediated dilation, asymmetric dimethylarginine, plasminogen activator inhibitor-1, von Willebrand factor),inflammatory markers (high-sensitivity C-reactive protein), lipids, and hyperandrogenism.Result(s): The OCP increased levels of glucose and insulin on oral glucose tolerance test, high-sensitivity C-re-active protein, triglycerides, and sex-hormone binding globulin and decreased levels of low-density lipoproteincholesterol and T. Metformin decreased levels of fasting insulin, oral glucose tolerance test insulin, high-densitylipoprotein cholesterol, and high-sensitivity C-reactive protein. Flow-mediated dilation increased only with met-formin (þ2.2% � 4.8%), whereas asymmetric dimethylarginine decreased equivalently for OCP and metformin(�0.3 � 0.1 vs. �0.1 � 0.1 mmol/L). Greater decreases in plasminogen activator inhibitor-1 occurred for theOCP than for metformin (�1.8 � 1.6 vs. �0.7 � 1.7 U/mL).Conclusion(s): In polycystic ovary syndrome, metformin improves insulin resistance, inflammatory markers, andendothelial function. The OCP worsens insulin resistance and glucose homeostasis, inflammatory markers, and tri-glycerides and has neutral or positive endothelial effects. The effect of the OCP on cardiovascular risk in polycysticovary syndrome is unclear. (Fertil Steril� 2010;93:184–91. �2010 by American Society for Reproductive Medi-cine.)

Key Words: Polycystic ovary syndrome, insulin resistance, endothelial function, oral contraceptive pill, metformin

pril 3, 2008; revised September 5, 2008; accepted September

published online November 18, 2008.

nothing to disclose. C.M. has nothing to disclose. S.K.H. has

to disclose. S.Z. has nothing to disclose. B.P.M. has served

ardiovascular Advisory Boards of Boehringer Ingelheim and

harmaceuticals. L.J.M. has nothing to disclose.

was an investigator-initiated trial supported by a competitive

ascular Lipid (CVL) Research Grant sponsored by Pfizer Aus-

d through internal department funds. H.J.T. received a National

and Medical Research Council Career Development Award.

eceived the Jean Hailes Fellowship.

at the Endocrine Society of Australia National Meeting, Christ-

New Zealand, September 2–5, 2007, and the American Endo-

ciety Annual Meeting, San Francisco, California, June 15–18,

quests: Lisa J. Moran, Ph.D., The Jean Hailes Foundation for

’s Health, Monash Institute of Health Services Research,

bag 29, Monash Medical Centre, 246 Clayton Rd., Clayton, Vic-

8, Australia (FAX: 61-03-9594-7554; E-mail: lisa.moran@med.

.edu.au).

ertility and Sterility� Vol. 93, No. 1, January 2010opyright ª2010 American Society for Reproductive Medicine, P

Polycystic ovary syndrome (PCOS) affects approximately6.6% of women of reproductive age and presents withreproductive manifestations including anovulation, infertil-ity, and hyperandrogenism. It is recognized as a metaboliccondition with insulin resistance (IR) a key pathophysiologiccomponent contributing to hyperandrogenism and the repro-ductive and metabolic features of PCOS (1). Overweight orobesity affects 50% to 80% of women with PCOS and exac-erbates IR and the reproductive and metabolic features ofPCOS (2). Polycystic ovary syndrome also is associatedwith increased cardiovascular risk factors including the met-abolic syndrome, impaired glucose tolerance, type 2 diabetesmellitus, endothelial dysfunction, and subclinical atheroscle-rosis (3). Addressing these complications is thus advisableand offers a potential to prevent future metabolic disease inthis population.

0015-0282/10/$36.00ublished by Elsevier Inc. doi:10.1016/j.fertnstert.2008.09.034

mailto:[email protected]

mailto:[email protected]

Endothelial dysfunction is an early stage in the develop-ment of cardiovascular disease. It can be assessed noninva-sively with techniques including flow-mediated dilationand by measurement of circulating markers produced bythe endothelium such as plasminogen activator inhibitor-1(PAI-1; a prothrombotic factor that inhibits fibrinolysis),asymmetric dimethylarginine (ADMA; an endogenous com-petitive inhibitor of nitric oxide [NO] synthase), and vonWillebrand factor (vWF; a carrier for the coagulation proteinfactor VIII) (4, 5). Reduced flow-mediated dilation (6) andincreased vWF (7), PAI-1 (8), and ADMA (9) independentlypredict cardiovascular disease. To date, flow-mediateddilation, PAI-1, ADMA, and vWF have been reported to beboth elevated (10–13) and similar (14) in PCOS comparedwith controls. Inflammatory markers, including high-sensitivity C-reactive protein (hsCRP), are also predictiveof cardiovascular disease (15) and are reportedly similar(16) or increased (17) in PCOS. Where endothelial dysfunc-tion is reported in PCOS, it has been associated variably withIR, hyperandrogenism, and other factors such as lipids orinflammatory markers such as hsCRP protein (10, 17, 18).

Insulin resistance is proposed to be associated with endo-thelial dysfunction through mechanisms including insulinstimulating NO synthase gene expression (19) and endothe-lin-1 expression through activation of the mitogen-activatedprotein kinase pathway (20). In PCOS, lifestyle interventionis first-line treatment with the oral contraceptive pill (OCP)and metformin commonly used in addition. In PCOS, metfor-min improves reproductive features and IR, and, although en-dothelial function has not been well studied, flow-mediateddilation (18, 21), PAI-1 (22), and ADMA (13) also havebeen shown to improve with changes related to changes inIR in some studies (21, 22) but not others (13, 18). TheOCP also improves reproductive features; however, it ad-versely affects glucose, IR, and some cardiovascular risk fac-tors in PCOS (23). Conversely, the OCP improves othercardiovascular risk factors such as low-density lipoproteincholesterol (LDL-C) and, in studies to date in PCOS, alsoimproves ADMA despite increased IR (12).

Although the effect of metformin or the OCP on endothe-lial function therefore has been assessed previously, there isas yet no comparison of their relative effect on endothelialfunction in PCOS. This study aimed to examine compara-tively the effect of pharmacologic interventions that differen-tially affect IR (metformin vs. the OCP) on cardiometabolicfeatures of PCOS including markers of endothelial function,inflammation, lipids, and hyperandrogenism.

MATERIALS AND METHODS

Subjects and Recruitment

This study comprises all subjects from a larger interventionstudy for whom additional blood samples were availablefor measurement of markers of endothelial function (23).Overweight women (body mass index [BMI] >25 kg/m2)with PCOS (n ¼ 66) were recruited from community adver-tisements. Polycystic ovary syndrome was diagnosed on the

Fertility and Sterility�

basis of National Institutes of Health criteria (24). Exclusioncriteria were type 2 diabetes mellitus (World Health Organi-zation criteria), smoking, and pregnancy. For 3 months beforeclinical measurements, participants were required to ceaseoral contraceptives, endocrine hormonal treatment, or insu-lin-sensitizing agents and received standard lifestyle advice(National Heart Foundation of Australia recommendations).This study was approved by the Southern Health ResearchAdvisory and Ethics Committee (Institutional ReviewBoard). All participants gave informed written consent.

Intervention

Subjects were allocated to one of two open-label groups onthe basis of computer-generated random numbers: [1] met-formin 1 g two times per day with doses titrated up overa 4-week period starting at 500 mg two times per day (n ¼36) and [2] high-dose OCP (35 mg ethinyl E2–2 mg cyproter-one acetate) (n ¼ 30) for 6 months. All subjects were fre-quency matched for age and weight to ensure equaldistribution of these potential confounding factors. All clini-cal measurements occurred at baseline and 6 months aftercommencing medications. The high-dose OCP was selectedas a commonly prescribed OCP in PCOS in both Australiaand Europe. Of the 66 subjects, 34 were seen with menstrualirregularity, hyperandrogenism, and polycystic ovaries on ul-trasound examination (n ¼ 16 metformin, n ¼ 18 OCP), and32 were seen with menstrual irregularity and hyperandrogen-ism (n ¼ 20 metformin, n ¼ 12 OCP).

CLINICAL AND BIOCHEMICAL MEASUREMENTS

Body mass index was calculated by weight divided bysquared height, and waist circumference was measured atthe umbilicus. Venous blood samples were collected afteran overnight fast for glucose, insulin, T, sex-hormone bind-ing globulin (SHBG), free fatty acids (FFA), total cholesterol,high-density lipoprotein cholesterol (HDL-C), LDL-C,triglycerides (23), hsCRP (16), vWF, and PAI-1 (25) aspreviously described. Asymmetric dimethylarginine wasmeasured with use of a commercially available chromogenicimmunoenzymatic assay (DLD Diagnostika GmbH, Ham-burg, Germany). The free androgen index (FAI) was calcu-lated as FAI ¼ (T/SHBG) � 100. A 120-minute 75-g oralglucose tolerance test (OGTT) was performed with assess-ment of glucose and insulin at 0, 30, 60 and 120, minutes.The homeostatic model assessment (HOMA-IR) as a surro-gate measure of IR was calculated as (Fasting serum insulin� Fasting plasma glucose)/22.5 (26). Glucose and insulinareas under the curve (AUC) during the OGTT were calcu-lated geometrically with use of the trapezoidal rule (27). Be-cause of the erratic anovulatory menstrual cycles, data werenot collected at specific cycle stages.

Endothelial Function

Brachial artery diameter was measured from B-mode ultra-sound images captured on a Diasonics DRF-400 machine

185

(Diasonics, Milpitas, CA) with use of a 10-MHz transducer,and an electrocardiogram trace was simultaneously recorded.Longitudinal scanning identified the clearest image of thebrachial artery above the elbow, with continuous scanningheld for 30 seconds before and 4 minutes after ischemia, in-duced via a pneumatic tourniquet inflated around the upperarm to 40 mm Hg above systolic pressure for 4 minutes. Ves-sel diameter was measured during systole and diastole andaveraged over five cardiac cycles. Flow-mediated dilationwas determined as the percentage change from baseline to60 seconds after ischemia, the point of maximal dilation(28). All flow-mediated dilation measurements wereperformed in a climate-controlled environment at constanttemperature (22�C). Published internal repeatability datademonstrate the accuracy and repeatability of these parame-ters (28).

Statistics

All data are presented as mean � SEM. Fasting insulin,HOMA-IR, FAI, hsCRP, and triglycerides were not normallydistributed and were log-transformed, and data were pre-sented as median (range). Results are presented for the 56subjects who completed the intervention (n ¼ 30 metformin,n ¼ 26 OCP) except for lipids, fasting insulin, fasting glu-cose, and HOMA-IR (n¼ 54); T, SHBG, FAI, OGTT insulin,and OGTT glucose (n ¼ 53); FFA, vWF, PAI-1, and ADMA(n ¼ 52); and hsCRP (n ¼ 46) because of incomplete data.Subjects with an isolated hsCRP >10 mg/L were excludedfrom the hsCRP analysis. Intention to treat analysis (baselinevalue carried forward) was conducted for the main outcomesof IR (OGTT insulin) and flow-mediated dilation. Two-tailedstatistical analysis was performed with use of SPSS for Win-dows 14.0 software (SPSS Inc., Chicago, IL) with statisticalsignificance set at a level of P%.05. Baseline data were

FIGURE 1

Changes in AUC OGTT insulin (A) and AUC OGTT glucosData are presented as mean � SEM. Data were assessedthe within-subject factor and intervention as the betweentreatment (P< .001) with decrease in AUC insulin for metf**Significant differential effect of treatment (P¼ .011) with

Teede. Endothelial function and PCOS. Fertil Steril 2010.

186 Teede et al. Endothelial function and PCOS

assessed with use of one-way analysis of variance (ANOVA)with intervention as between-subject factor, and comparisonsbetween time points were assessed with use of repeated-measures ANOVA with intervention as between-subject fac-tor. Relationships between variables were examined with useof bivariate and partial correlations with age and the changein weight or BMI over the intervention as covariates for allanalysis.

RESULTS

Subjects, Weight, and Body Composition

Data are reported for 56 women who completed the study(n ¼ 30 metformin, n ¼ 26 OCP) (age 33.5 � 6.7 years,weight 97.1 � 21.1 kg, BMI 36.1 � 7.2 kg/m2, BMI range25.49–55.73 kg/m2). Subject weights did not change duringthe 3 months before clinical measurements (98.6 � 4.1 vs.99.8 � 3.9 kg, P¼.163) with no differences in the before-study change in weight between the metformin and theOCP groups (P¼.499). There were no differences in baselinecharacteristics for subjects randomly assigned to either group(Table 1), and no differences in baseline characteristics forsubjects who completed or did not complete the study (datanot shown). There were no changes in weight or BMI overthe study for either group. Increases in waist circumference(1.5 � 5.1 cm, P¼.034) occurred for all subjects (n ¼ 56)with no difference between groups. Data have been reportedpreviously on study withdrawals. There were 10 withdrawals(n ¼ 6 metformin group; n ¼ 4 OCP group); all withdrawalswere for personal reasons with the exception of one woman inthe OCP group withdrawing because of having mood swings(23).

There was a time-by-treatment effect for SHBG (P<.001)and FAI (P<.001). The OCP group demonstrated an increase

e (B) after metformin or OCP treatment for 6 months.with use of repeated-measures ANOVA with time as

-subject factor. *Significant differential effect oformin (P¼ .003) and increase for OCP (P< .001).increase in AUC glucose for OCP (P< .001).

Vol. 93, No. 1, January 2010

in SHBG and decrease in FAI levels compared with no signif-icant change for the metformin group (Table 1) and higherSHBG and lower FAI levels after treatment. Data havebeen reported previously for the effects of metformin andthe OCP on menstrual regularity (23).

Insulin and Glucose Homeostasis

There was a time-by-treatment effect for fasting insulin(P¼.007), HOMA-IR (P¼.007), and OGTT insulin(P<.001) such that the metformin group demonstrateddecreases in fasting insulin (7.4 (�48.6, 46.7) mU/L,P¼.021), HOMA-IR (1.6 [�10.5, 17.1], P¼.013), andOGTT insulin (5,454.3 � 8,632.7 mU/L per 120 minutes,P¼.003), and the OCP group demonstrated no significantchanges in fasting insulin and HOMA-IR and an increasein OGTT insulin (4,397.7 � 5,242.7 mU/L per 120 minutes,P<.001) (Fig. 1). After treatment, the metformin group dem-onstrated lower fasting insulin (P¼.041), HOMA-IR(P¼.049), and OGTT insulin (P¼.015). On intention to treatanalysis for OGTT insulin, no differences in main effectswere reported (data not shown).

There were no changes in fasting glucose and no differen-tial effect of treatment on fasting glucose over the study (Ta-ble 1). There was a time-by-treatment effect for OGTTglucose (P¼.011) with no change in OGTT glucose for themetformin group but an increase in OGTT glucose with theOCP (117.7 � 133.5 mmol/L per 120 minutes, P<.001)(Fig. 1).

Lipid, Free Fatty Acids, and Inflammatory Markers (hsCRP)

There were no changes in total cholesterol over the study(Table 1). Increases in FFA (85.6 � 238.5 mmol/L, P¼.015)occurred for all subjects. There was a time-by-treatment effectfor hsCRP (P¼.001), HDL-C (P¼.002), LDL-C (P¼.042), andtriglycerides (P¼.015). High-sensitivity C-reactive protein andHDL-C decreased for the metformin group (P¼.028 andP¼.002, respectively), and hsCRP increased with the OCP(P¼.043). Low-density lipoprotein cholesterol decreased forthe OCP group (P¼.006) and triglycerides decreased for theOCP (P¼.006) with no change with metformin. Low-densitylipoprotein cholesterol was significantly lower and HDL-Cwas significantly higher in the OCP group after treatment(P¼.001).

Markers of Endothelial Function

There was a time-by-intervention effect for flow-mediateddilation (P¼.037), PAI-1 (P¼.017), and vWF (P¼.038). Inthe metformin group, vWF did not change, flow-mediated di-lation increased (P¼.017), and PAI-1 (P¼.004) and ADMA(P<.001) decreased. In the OCP group, vWF and flow-medi-ated dilation did not change significantly, and PAI-1(P¼.004) and ADMA (P<.001) decreased. A greater de-crease in PAI-1 occurred with the OCP (P¼.004) thanwith metformin (P¼.044). The OCP displayed lower PAI-1after treatment (P<.001). Although vWF did not change


significantly within either group, a differential effect of treat-ment was noted, and the posttreatment level was higher withthe OCP than with metformin (P¼.012) (Fig. 2). Equivalentreductions in ADMA (0.2 � 0.1 mmol/L, P<.001) occurredover the study with metformin (0.1 � 0.1 mmol/L) and theOCP (0.3 � 0.1 mmol/L) with no differential effect of treat-ment (Fig. 2). On intention to treat analysis for flow-mediateddilation, no differences in main effects were reported (datanot shown).

Correlations

There was no correlation between the change in flow-medi-ated dilation with intervention and the change in IR,ADMA, PAI-1, and vWF. The change in ADMA correlatedwith the change in LDL-C (R ¼ 0.312, P¼.026), the changein vWF correlated with the change in SHBG (R ¼ 0.432,P¼.002) and the change in FAI (R ¼ �0.381, P¼.007), andthe change in PAI-1 correlated with the change in SHBG(R ¼ �0.341, P¼.015) and the change in FAI (R ¼ 0.377,P¼.008). These relationships remained on adjustment forage, weight, and BMI.

DISCUSSION

We report for the first time the comparative study of medicalinterventions that differentially alter IR on endothelial func-tion (flow-mediated dilation and circulating markers) inoverweight women with PCOS. Specifically, metformin im-proved IR, inflammation (hsCRP), and circulating (ADMAand PAI-1) and vascular ultrasound markers of endothelialfunction (flow-mediated dilation). The OCP increased IRand hsCRP, yet still improved some circulating markers ofendothelial function (ADMA and PAI-1). This suggests anindirect relationship between IR and endothelial functionand provides novel insights into the relationships betweenIR and endothelial function.

Insulin resistance has a central pathophysiologic role con-tributing to reproductive and metabolic features of PCOS(1). Confirmation that metformin improves (21) and theOCP worsens (29) IR is consistent with the findings of themajority of studies in PCOS and non-PCOS populations. Fur-thermore, we confirm previous reports of reductions in hsCRPwith metformin (30) and increases in hsCRP with the OCP(31). The underlying mechanisms for this effect include thefirst-pass metabolic effect of the OCP on increased productionof coagulation proteins and hsCRP (32). Given the associationof IR with endothelial dysfunction, the metabolic syndrome,impaired glucose tolerance and type 2 diabetes mellitus andthe association of inflammatory markers with the pathogenesisof cardiovascular disease (10), the differential effect of theOCP and metformin on IR and inflammatory markers mayhave clinical implications for their long-term use in PCOS.

In the general population IR increases endothelial dysfunc-tion and reduces NO bioavailability through hyperglycemia,oxidative stress, dyslipidemia, and inflammation (5). In obser-vational PCOS studies, we and other groups have

187

FIGURE 2

Changes in endothelial function: flow-mediated dilation (FMD) (A), vWF (B), PAI-1 (C), and ADMA (D) aftermetformin or OCP treatment for 6 months. Data are presented as mean� SEM. Data were assessed with use ofrepeated-measures ANOVA with time as the within-subject factor and intervention as the between-subjectfactor. *Significant differential effect of treatment (P¼ .037) with increase in flow-mediated dilation for metformin(P¼ .017). **Significant differential effect of treatment (P¼ .038) for vWF with significantly higher vWF levels for theOCP group after treatment (P¼ .012). ySignificant differential effect of treatment (P¼ .017) with greater decreasein PAI-1 for OCP (P< .001) and decrease in PAI-1 for metformin (P¼ .044). yySignificant differential effect ofmetformin or OCP treatment (P< .001) for ADMA for all subjects.


demonstrated endothelial dysfunction related to IR (16–18),androgens (14, 16, 18), lipids (10), or inflammation (17). In in-terventional PCOS studies, metformin improves endothelialdysfunction (18, 21) with the exception of one small short-term study in a population that potentially had less IR (Rotter-dam criteria diagnosed PCOS) (33). However, the mechanismsby which metformin improves flow-mediated dilation are notyet clear. In this current study and in other work (18), changesin IR and flow-mediated dilation were not correlated. Improvedendothelial function with metformin may be caused by a varietyof factors independent of IR including metformin-induced lipidchanges (21) and direct vessel wall effects via increased NOsynthesis or activity in vivo (34).

Consistent with impaired flow-mediated dilation, endothe-lial dysfunction in PCOS has been supported by higher levels


of circulating endothelial markers (ADMA, PAI-1) (11–13)and inflammatory markers (hsCRP) (17). Previous studies re-ported reductions in ADMA with studies that only assessedmetformin (13) or OCP (12) in isolation. In this first compar-ative study assessing a diverse range of cardiometabolic fac-tors, we have shown similar reductions in ADMA with theOCP and metformin that occurred independent of changesin IR. Insulin resistance potentially regulates ADMA throughmechanisms including oxidative stress, dyslipidemia, anddirect effects of IR or hyperglycemia. In addition to this,the formation of ADMA by protein arginine N-methyltrans-

ferases is regulated by LDL-C (35). In the current study

LDL-C level fell with the OCP, and this change correlated

with the change in ADMA. Estrogen also up-regulates dime-

thylarginine dimethylaminohydrolyase activity, which


increases ADMA breakdown (36), and estrogen thereforemay decrease ADMA by directly regulating its synthesisand breakdown independent of IR. This is consistent withobserved reductions in ADMA after estrogen therapy inpostmenopausal women (37). Given the demonstrated im-provement in ADMA with the OCP and metformin despitedifferential effects on IR, the regulation of ADMA in IRstates including PCOS is likely to be complex and not directlyrelated to IR.

Plasminogen activator inhibitor-1 is a prothrombotic factorproduced by the endothelium that inhibits fibrinolysis and reg-ulates vascular smooth muscle proliferation. In vitro, insulinup-regulates PA1-1 gene transcription (38) and stimulates he-patic (39) and endothelial PAI-1 production (40). We reporthere that despite OCP-induced increases in IR, PAI-1 decreasedmore with the OCP than metformin. This challenges the theoryof a direct effect of IR on PAI-1. In non-PCOS populations, weand other groups have shown that estrogen decreases PAI-1 (41,42), which occurs via reductions in NO production (43) or sup-pression of PAI-1 gene expression (44). However, the negativeeffect of the OCP on IR does not appear to negate its estrogen-induced beneficial effects on PAI-1. Neither the OCP nor met-formin intervention significantly changed vWF, although the

TABLE 1Subject characteristics at baseline and study end.

Metformin (n [ 30)

Beforetreatment

Aftertreatment

Weight (kg) 98.02 � 4.0 96.4 � 3.7BMI (kg/m2) 36.5 � 1.3 35.9 � 1.2Waist circumference (cm) 109.1 � 2.7 110.0 � 2.9a

Fasting glucose (mmol/L) 4.7 � 0.1 4.5 � 0.1Fasting insulin (mU/L) 16.4 (4.4–74.0) 10.5 (5.2–97.HOMA-IR 3.3 (0.9–15.7) 2.0 (1.0–32.8Total cholesterol (mmol/L) 5.3 � 0.2 5.1 � 0.2LDL-C (mmol/L) 3.4 � 0.2 3.3 � 0.2HDL-C (mmol/L) 1.2 � 0.1 1.1 � 0.1a

Triglycerides (mmol/L) 1.3 (0.6–2.9) 1.2 (0.6–4.7)Free fatty acids (mmol/L) 448.5 � 32.6 535.6 � 50.7hsCRP (mg/L) 3.4 (0.9–14.0) 2.9 (0.2–12.9T (nmol/L) 2.4 � 0.1 2.3 � 0.2SHBG (nmol/L) 31.0 � 3.0 38.4 � 6.4FAI 8.0 (2.1–27.1) 7.3 (0.9–37.8

Note: Data are presented as mean � SEM except for fasting ipresented as median (range). Baseline data were assessedtween-subject factors, and intervention data were assessedwithin-subject factor and intervention as the between-subje

a P< .05 for within-group change over study intervention.b P< .05 for difference between OCP and metformin group at



vWF levels after treatment did differ between groups. Giventhe association of flow-mediated dilation, PAI-1, and ADMAwith metabolic syndrome and cardiovascular disease, theseimprovements in endothelial cardiometabolic function arelikely to be clinically beneficial.

Estrogen consistently has been reported to have both indi-rect favorable effects on the endothelium mediated by NO,lipids, and inflammatory and adhesion markers and direct ef-fects on the vessel wall smooth muscle, intima, and endothe-lium (32). The beneficial effects of estrogen therefore maycounteract the negative effects of impaired IR and glucose ho-meostasis on endothelial function. However, the OCP in-creases thrombosis, and we have previously showna worsening of arterial stiffness after higher-dose OCP usein PCOS potentially negating the beneficial effects of theOCP on circulating markers of endothelial function (23).This discrepancy also may be due to arterial stiffness as mea-sured by pulse wave velocity representing a different stage ofcardiovascular risk (arteriosclerosis and arterial stiffness inthe media) to flow-mediated dilation (atherosclerosis and en-dothelial function of the intima) (45). Furthermore, althoughIR may directly or indirectly adversely affect the endothe-lium, IR encompasses a number of components including

OCP (n [ 26)

Beforetreatment

Aftertreatment

P value changeover study,metforminvs. OCP

95.8 � 4.1 96.6 � 4.1 .08535.8 � 1.5 36.1 � 1.5 .130

104.2 � 3.0 106.3 � 2.8a .3904.5 � 0.1 4.4 � 0.1 .815

1)a 15.6 (2.9–79.7) 19.2 (4.6–46.0)b .007)a 3.1 (0.5–17.0) 3.9 (0.9–9.0)b .007

5.1 � 0.2 5.0 � 0.2 .8113.2 � 0.2 2.8 � 0.2a,b .0421.4 � 0.1 1.5 � 0.1b .002

1.2 (0.5–2.4) 1.6 (0.4–3.4)a .015a 453.78 � 34.7 537.6 � 38.3a .332

)a 3.5 (0.3–9.7) 4.7 (0.5–15.6)a .0012.1 � 0.1 1.7 � 0.2b .354

33.4 � 3.2 148.5 � 14.6a,b < .001) 6.5 (2.6–28.8) 1.2 (0.2–3.9)a,b < .001

nsulin, HOMA-IR, triglycerides, hsCRP, and FAI, which arewith use of one-way ANOVA with intervention as the be-with use of repeated-measures ANOVA with time as the

ct factor.

study beginning or end.

189

dyslipidemia, hypertension, inflammation, endothelial dys-function, and altered cytokines. This may explain why thecurrently noted opposite effects on IR do not translatedirectly to the circulating markers of endothelial function.Additionally, flow-mediated dilation may be considered asan indicator of total endothelial function, representing thecombined effects of endothelial-derived factors. Despite ourobserved positive changes in ADMA and PAI-1 with theOCP, their relationship to flow-mediated dilation and globalendothelial function is still not clear. Also, although surrogatemarkers predict cardiovascular disease outcomes, they stillremain substitute, indirect indicators of cardiovascular dis-ease. Despite the positive effects of the OCP on some circu-lating markers of endothelial function, there was a trendtoward a negative effect of the OCP on flow-mediated dila-tion. The net cardiovascular effects of the OCP, although ofsignificant clinical interest, remain unclear.

Potential limitations of the current study warrant discus-sion. We did not measure renal dysfunction, which canhave an impact on ADMA (46), although participants werehealthy young women with no history of renal disease.Although the accurate measurement of IR is challenging,the HOMA-IR score and OGTT AUC insulin both havebeen shown to be comparable to results of clamp studies(47, 48). Finally, endothelial function also has been shownto vary across the menstrual cycle (49), and we were unableto standardize data collection to a specific cycle stage be-cause of erratic cycle lengths.

In women with PCOS, metformin improves a range of riskfactors for type 2 diabetes mellitus and cardiovascular dis-ease including IR, inflammatory markers, flow-mediated di-lation, and circulating markers of endothelial function.However, despite negative effects of the OCP on IR, glucosehomeostasis, inflammatory markers, and triglycerides, theOCP still improved circulating markers of endothelial func-tion potentially because of beneficial effects of estrogencounteracting the negative effects of impaired IR on endothe-lial function. With these contradictory effects, the overall ef-fect of the OCP on cardiovascular risk is not clear, but it ispossible that the different components of the OCP mayhave opposing effects that may balance the risk out and neu-tralize it overall. Given the increased cardiometabolic riskpresent in women with PCOS, further research is requiredto determine the optimal pharmaceutical therapy. This in-cludes expansion of these findings to metformin or OCP ther-apy in lean women with PCOS and the effects of combinationmetformin and OCP therapy on insulin resistance, glucosehomeostasis, and cardiovascular risk factors including endo-thelial function in PCOS.

Acknowledgments: Douglas Pharmaceutical Australia provided the metfor-

min. Pathology was completed at Southern Health pathology laboratories.

REFERENCES1. Teede H, Hutchison SK, Zoungas S. The management of insulin resis-

tance in polycystic ovary syndrome. Trends Endocrinol Metab

2007;18:273–9.


2. Norman RJ, Masters SC, Hague W, Beng C, Pannall P, Wang JX. Meta-

bolic approaches to the subclassification of polycystic ovary syndrome.

Fertil Steril 1995;63:329–35.

3. Lakhani K, Prelevic GM, Seifalian AM, Atiomo WU, Hardiman P. Poly-

cystic ovary syndrome, diabetes and cardiovascular disease: risks and

risk factors. J Obstet Gynaecol 2004;24:613–21.

4. Deanfield J, Halcox JP, Rabelink TJ. Endothelial function and

dysfunction: testing and clinical relevance. Circulation 2007;115:

1285–95.

5. Jansson P-A. Endothelial dysfunction in insulin resistance and type 2

diabetes. J Intern Med 2007;262:173–83.

6. Gokce N, Keaney JF Jr, Hunter LM, Watkins MT, Menzoian JO, Vita JA.

Risk stratification for postoperative cardiovascular events via noninva-

sive assessment of endothelial function: a prospective study. Circulation

2002;105:1567–72.

7. Morange PE, Simon C, Alessi MC, Luc G, Arveiler D, Ferrieres J, et al.

Endothelial cell markers and the risk of coronary heart disease: the Pro-

spective Epidemiological Study of Myocardial Infarction (PRIME)

study. Circulation 2004;109:1343–8.

8. Thogersen AM, Jansson JH, Boman K, Nilsson TK, Weinehall L,

Huhtasaari F, et al. High plasminogen activator inhibitor and tissue

plasminogen activator levels in plasma precede a first acute myocardial

infarction in both men and women: evidence for the fibrinolytic system

as an independent primary risk factor. Circulation 1998;98:2241–7.

9. Valkonen VP, Paiva H, Salonen JT, Lakka TA, Lehtimaki T, Laakso J,

et al. Risk of acute coronary events and serum concentration of asymmet-

rical dimethylarginine [see comment]. Lancet 2001;358:2127–8.

10. Meyer C, McGrath BP, Cameron J, Kotsopoulos D, Teede HJ. Vascular

dysfunction and metabolic parameters in polycystic ovary syndrome.

J Clin Endocrinol Metab 2005;90:4630–5.

11. Carmassi F, De Negri F, Fioriti R, De Giorgi A, Giannarelli C,

Fruzzetti F, et al. Insulin resistance causes impaired vasodilation and hy-

pofibrinolysis in young women with polycystic ovary syndrome. Thromb

Res 2005;116:207–14.

12. Charitidou C, Farmakiotis D, Zournatzi V, Pidonia I, Pegiou T,

Karamanis N, et al. The administration of estrogens, combined with

anti-androgens, has beneficial effects on the hormonal features and

asymmetric dimethyl-arginine levels, in women with the polycystic

ovary syndrome. Atherosclerosis 2008;196:958–65.

13. Heutling D, Schulz H, Nickel I, Kleinstein J, Kaltwasser P,

Westphal SA, et al. Asymmetric dimethylarginine, inflammatory

and metabolic parameters in women with polycystic ovary syndrome

before and after metformin treatment. J Clin Endocrinol Metab

2008;93:82–90.

14. Kelly CJ, Speirs A, Gould GW, Petrie JR, Lyall H, Connell JM. Altered

vascular function in young women with polycystic ovary syndrome.


15. Pai JK, Pischon T, Ma J, Manson JE, Hankinson SE, Joshipura K, et al.

Inflammatory markers and the risk of coronary heart disease in men and

women. N Engl J Med 2004;351:2599–610.

16. Meyer C, McGrath BP, Teede HJ. Overweight women with polycystic

ovary syndrome have evidence of subclinical cardiovascular disease.


17. Tarkun I, Arslan BC, Canturk Z, Turemen E, Sahin T, Duman C. Endo-

thelial dysfunction in young women with polycystic ovary syndrome: re-

lationship with insulin resistance and low-grade chronic inflammation.


18. Diamanti-Kandarakis E, Alexandraki K, Protogerou A, Piperi C,

Papamichael C, Aessopos A, et al. Metformin administration improves

endothelial function in women with polycystic ovary syndrome. Eur J

Endocrinol 2005;152:749–56.

19. Kuboki K, Jiang ZY, Takahara N, Ha SW, Igarashi M, Yamauchi T, et al.

Regulation of endothelial constitutive nitric oxide synthase gene expres-

sion in endothelial cells and in vivo : a specific vascular action of insulin.

Circulation 2000;101:676–81.

20. Cardillo C, Nambi SS, Kilcoyne CM, Choucair WK, Katz A, Quon MJ,

et al. Insulin stimulates both endothelin and nitric oxide activity in the

human forearm. Circulation 1999;100:820–5.


21. Orio F Jr, Palomba S, Cascella T, De Simone B, Manguso F, Savastano S,

et al. Improvement in endothelial structure and function after metformin

treatment in young normal-weight women with polycystic ovary syn-

drome: results of a 6-month study. J Clin Endocrinol Metab 2005;90:

6072–6.22. Velazquez EM, Mendoza SG, Wang P, Glueck CJ. Metformin therapy is

associated with a decrease in plasma plasminogen activator inhibitor-1,

lipoprotein(a), and immunoreactive insulin levels in patients with the

polycystic ovary syndrome. Metabolism 1997;46:454–7.

23. Meyer C, McGrath BP, Teede HJ. Effects of medical therapy on insulin

resistance and the cardiovascular system in polycystic ovary syndrome.

Diabetes Care 2007;30:471–8.

24. Zawdaki J, Dunaif A. Diagnostic criteria for polycystic ovary syndrome:

towards a rational approach. In: Dunaif A, Givens J, Haseltine F, Mar-

rian G, eds. Polycystic ovary syndrome current issues in endocrinology

and metabolism. Vol. 4. Boston: Blackwell Scientific, 1992:377–84.

25. Teede HJ, Dalais FS, Kotsopoulos D, McGrath BP, Malan E, Gan TE,

et al. Dietary soy containing phytoestrogens does not activate the hemo-

static system in postmenopausal women. J Clin Endocrinol Metab

2005;90:1936–41.

26. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF,

Turner RC. Homeostasis model assessment: insulin resistance and

beta-cell function from fasting plasma glucose and insulin concentra-

tions in man. Diabetologia 1985;28:412–9.

27. Wolever TM, Jenkins DJ, Jenkins AL, Josse RG. The glycemic

index: methodology and clinical implications. Am J Clin Nutr

1991;54:846–54.

28. Liang YL, Teede H, Kotsopoulos D, Shiel L, Cameron JD, Dart AM, et al.

Non-invasive measurements of arterial structure and function: repeatabil-

ity, interrelationships and trial sample size. Clin Sci 1998;95:669–79.

29. Morin-Papunen LC, Vauhkonen I, Koivunen RM, Ruokonen A,

Martikainen HK, Tapanainen JS. Endocrine and metabolic effects of

metformin versus ethinyl estradiol–cyproterone acetate in obese women

with polycystic ovary syndrome: a randomized study. J Clin Endocrinol

Metab 2000;85:3161–8.

30. Diamanti-Kandarakis E, Paterakis T, Alexandraki K, Piperi C,

Aessopos A, Katsikis I, et al. Indices of low-grade chronic inflammation

in polycystic ovary syndrome and the beneficial effect of metformin.

Hum Reprod 2006;21:1426–31.

31. van Rooijen M, Hansson LO, Frostegard J, Silveira A, Hamsten A,

Bremme K. Treatment with combined oral contraceptives induces

a rise in serum C-reactive protein in the absence of a general inflamma-

tory response. J Thromb Haemost 2006;4:77–82.

32. Teede HJ. Sex hormones and the cardiovascular system: effects on arte-

rial function in women. Clin Exp Pharmacol Physiol 2007;34:672–6.

33. Lowenstein L, Damti A, Pillar G, Shott S, Blumenfeld Z. Evaluation of

endothelial function in women with polycystic ovary syndrome. Eur J

Obstet Gynecol Reprod Biol 2007;134:208–12.

34. Davis BJ, Xie Z, Viollet B, Zou MH. Activation of the AMP-activated

kinase by antidiabetes drug metformin stimulates nitric oxide synthesis

in vivo by promoting the association of heat shock protein 90 and endo-

thelial nitric oxide synthase. Diabetes 2006;55:496–505.

35. Boger RH, Sydow K, Borlak J, Thum T, Lenzen H, Schubert B, et al.

LDL cholesterol upregulates synthesis of asymmetrical dimethylarginine

in human endothelial cells: involvement of S-adenosylmethionine–de-

pendent methyltransferases. Circ Res 2000;87:99–105.


36. Holden DP, Cartwright JE, Nussey SS, Whitley GS. Estrogen stimu-

lates dimethylarginine dimethylaminohydrolase activity and the me-

tabolism of asymmetric dimethylarginine. Circulation 2003;108:

1575–80.

37. Verhoeven MO, Hemelaar M, van der Mooren MJ, Kenemans P,

Teerlink T. Oral, more than transdermal, oestrogen therapy lowers asym-

metric dimethylarginine in healthy postmenopausal women: a random-

ized, placebo-controlled study. J Intern Med 2006;259:199–208.

38. Banfi C, Eriksson P, Giandomenico G, Mussoni L, Sironi L, Hamsten A,

et al. Transcriptional regulation of plasminogen activator inhibitor type 1

gene by insulin: insights into the signaling pathway. Diabetes 2001;50:

1522–30.

39. Kooistra T, Bosma PJ, Tons HA, van den Berg AP, Meyer P, Princen HM.

Plasminogen activator inhibitor 1: biosynthesis and mRNA level are in-

creased by insulin in cultured human hepatocytes. Thromb Haemost

1989;62:723–8.

40. Schneider DJ, Nordt TK, Sobel BE. Stimulation by proinsulin of expres-

sion of plasminogen activator inhibitor type-I in endothelial cells. Diabe-

tes 1992;41:890–5.

41. Teede HJ, McGrath BP, Turner A, Majewski H. Effects of oral combined

hormone replacement therapy on platelet aggregation in postmenopausal

women. Clin Sci 2001;100:207–13.

42. Meijers JC, Middeldorp S, Tekelenburg W, van den Ende AE, Tans G,

Prins MH, et al. Increased fibrinolytic activity during use of oral contra-

ceptives is counteracted by an enhanced factor XI–independent down

regulation of fibrinolysis: a randomized cross-over study of two low-

dose oral contraceptives [see comment]. Thromb Haemost 2000;84:

9–14.

43. Brown NJ, Abbas A, Byrne D, Schoenhard JA, Vaughan DE. Compara-

tive effects of estrogen and angiotensin-converting enzyme inhibition on

plasminogen activator inhibitor-1 in healthy postmenopausal women.

Circulation 2002;105:304–9.

44. Smith LH, Coats SR, Qin H, Petrie MS, Covington JW, Su M, et al. Dif-

ferential and opposing regulation of PAI-1 promoter activity by estrogen

receptor alpha and estrogen receptor beta in endothelial cells [see com-

ment]. Circ Res 2004;95:269–75.

45. Ter Avest E, Stalenhoef AF, de Graaf J. What is the role of non-invasive

measurements of atherosclerosis in individual cardiovascular risk predic-

tion? Clin Sci 2007;112:507–16.

46. Fliser D, Kronenberg F, Kielstein JT, Morath C, Bode-Boger SM,

Haller H, et al. Asymmetric dimethylarginine and progression of chronic

kidney disease: the mild to moderate kidney disease study. J Am Soc

Nephrol 2005;16:2456–61.

47. Ciampelli M, Leoni F, Cucinelli F, Mancuso S, Panunzi S, De Gaetano A,

et al. Assessment of insulin sensitivity from measurements in the fasting

state and during an oral glucose tolerance test in polycystic ovary syn-

drome and menopausal patients. J Clin Endocrinol Metab 2005;90:

1398–406.

48. Mather KJ, Hunt AE, Steinberg HO, Paradisi G, Hook G, Katz A, et al.

Repeatability characteristics of simple indices of insulin resistance: im-

plications for research applications. J Clin Endocrinol Metab 2001;86:

5457–64.

49. Williams MR, Westerman RA, Kingwell BA, Paige J, Blombery PA,

Sudhir K, et al. Variations in endothelial function and arterial compli-

ance during the menstrual cycle. J Clin Endocrinol Metab 2001;86:

5389–95.

191

Documents

Polycystic Ovary Syndrome