Reference

Study type

Characteristics

Determinant (I)

Controls (C)

Outcome measures and follow-up time

Results

Quality

assessment study

Chang, 2011

Design

Cross-sectional cohort study

N = 697

Dallas heart study

Aim of the study:

to determine in a multiethnic population-

based sample whether women with PCOS have greater measures

of subclinical atherosclerosis than women without PCOS

Inclusion criteria

-Women between the ages of 35 and 49

- PCOS according R’dam criteria

- Period 2000 - 2002

Exclusion criteria

- postmenopausal

women

Determinants

PCOS according R’dam criteria (exept for cycle interval > 45 days)

N = 144

Controls

Women with regular menses, total testosterone <2.78 nmol/l, no signs

of hirsutism and no polycystic ovarian morphology by magnetic

resonance imaging (MRI).

N =

Primary outcomes:

Subclinical atherosclerosis defined as coronary

artery calcium (CAC) by computed tomography and abdominal

aortic plaque by MRI.

Secundary outcomes:

Hirsutism (?), BMI,

Total Testosterone,  FAI, PCOM. Fasting Glucose and Insulin, HOMA-IR, Total cholesterol, LDL, HDL, VLDL, hs-CRP, Leptin, MCP-1. LV mass

Follow up:

Primary outcomes:

There was a greater prevalence of hypertension in women with PCOS (29.2% vs 18.8%, P = 0.03) but no significant difference in the prevalence of  impaired fasting glucose (IFG), diabetes or the metabolic syndrome than controls

In a multivariate logistic regression analysis adjusting for age, BMI, insulin and

hypertension, the odds ratio of aortic plaque for women with

PCOS compared with controls was not significant (0.63, 95% confidence

interval 0.36–1.11; unadjusted odds ratio, 0.66, 95% confidence

interval 0.39–1.13). Among women with PCOS, those with aortic plaque had significantly lower HDL [1.22 vs 1.33 mmol/l (47.0 vs 51.5 mg/dl),

P = 0.01] and higher fasting glucose [5.38 vs 5.00 mmol/l (97.0 vs

90.0 mg/dl), P = 0.01]. In the control groups, these risk factors

were not different in those with or without aortic plaque. In logistic regression analyses, significant predictors of aortic plaque in women with PCOS were glucose [OR 1.01 (1.00–102)], low HDL [OR 2.3 (1.0–5.4)] or the metabolic syndrome.[OR 2.4 (1.1–5.7)].

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

To assess the prevalence of subclinical atherosclerosis, electron beam computed tomography (EBCT) (n = 641) and aortic

MRI (n = 569) were performed at visit 3 to provide measurements of CAC and aortic plaque, respectively.

Selective loss to follow up:

No, they only included women who completed all three study visits

Identification confounders and correction in analysis:

Yes, adjustment for age, BMI, insulin and

Hypertension in the aortic plaque model.

Funding:

This work was supported by the Donald W. Reynolds Cardiovascular Clinical Research Center at Dallas, the American Heart Association (Fellow-to-Faculty Transition Award #0775033N to AYC) and the Burroughs-Wellcome Fund (Clinical Scientist Award in Translational Research grant #1005954 to RJA).

Coksuer, 2011

Design

Case- control

N = 64

Aim of the study:

To investigate the risk of cardiovascular disease by measuring carotid intima-media thickness (CIMT) and to evaluate the diameters and blood flow volume of vertebral arteries using Doppler ultrasound and the possible contribution of hyperandrogenemia that exists in these patients.

Inclusion criteria

- Caucasian women

- age 18 – 35 years

- nonsmokers

Exclusion criteria

pregnancy, hypothyroidism, hyperprolactinemia,

Cushings syndrome, nonclassical congenital

adrenal hyperplasia, and use of OAC

glucocorticoids, antiandrogens, ovulation induction agents, antidiabetic or antiobesity drugs or other hormonal

drugs within the previous 6 months. neoplastic, metabolic, hepatic, and cardiovascular

disorder or other concurrent medical illness were also

excluded from the study

Determinants

PCOS according R’dam criteria

N = 31

Controls

Healthy women

N = 33

Primary outcomes:

BMI, W/H ratio, Ferriman-Galwey score, SBP, DBP, heart rate, Blood flow in Vertebral arteries, CIMT, HOMA-IR, free and total T, Prl, E2, FSH, LH, Lipid profile, CRP

Secundary outcomes:

Follow up:

Primary outcomes:

A significant difference between the two groups in IMT on the carotid artery

and diameters of vertebral arteries were found (Table 2).

Right-left and total flow volumes of the vertebral arteries (VA) were significantly lower in the PCOS group than in the control group.  Compared to the control group, diameters of right and left VA decreased (P < 0.05).

Conclusion: The results of this study suggest that hyperandrogenemia in obese PCOS patients affects carotid and vertebro-basilar system arteries wall thickness.

In the regression analyses, significant correlation was detected between the IMT and testosterone levels (t = 3.35, P = 0.001).

Clearly defined groups:

Yes

Selection bias:

Possible, case-control

Method of assessing the outcome

appropriately:

All blood samples were obtained in the morning between 08.00–09.00 h after an overnight fast during the early follicular phase (second

to fourth day) of a spontaneous or progesterone induced menstrual cycle. Carotid and vertebral arteries ultrasound imaging were performed using Logic-5 Expert US system (General Electric Medical System, CA, USA) by the same experienced radiologist.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

Yes, regression analyses

Funding:

Not reported

De Groot, 2011

Design

systematic review and meta-analysis

N =

Cibula et al. 2000

Lundeand Tanbo et al. 2007

Shaw et al. 2008

Solomon et al. 2002

Wild et al. 2000b

Aim of the study:

To examine the risk of (non)fatal coronary

heart disease (CHD) or stroke in patients with PCOS and ovulatory women without PCOS

Inclusion criteria

Controlled studies comparing women with PCOS to women without PCOS were considered for eligibility. outcomes: incident CHD and/or stroke (non-fatal or fatal).

PCOS should have been defined according to the

WHO (WHO, 1973), NIH (Zawadski, 1992), Rotterdam (Rotterdam,

2004) or AE (Azziz et al., 2009) criteria and the control group should

be PCOS-free.

Exclusion criteria

Studies without a control group or that

did not distinguish women with or without PCOS

Determinants

PCOS according NIH, R’dam and AES-PCOS definitions

N =

Controls

N =

After identifying 1340 articles published before January 2010, 5 follow-up studies published between 2000 and 2008 were included.

Primary outcomes:

the pooled risk ratio for CHD or stroke or mortality related to CHD or stroke and BMI-adjusted pooled risk ratio

Secundary outcomes:

Follow up:

Primary outcomes:

The relative risk for cardiovascular events ranged from 0.92 to 4.24. Four of five studies showed individually a significant increased risk for women with PCOS. The pooled relative risk from a random effects model was 2.02 (95% confidence interval 1.47, 2.76), showing a 2-fold increased risk for CHD/stroke in women with PCOS. The I2 showed moderate heterogeneity (42%).

Two studies provided BMI-adjusted risk estimates (Wild et al., 2000b; Solomon et al., 2002; Fig. 3). The pooled relative risk adjusted for BMI was 1.55 (95% confidence interval 1.27, 1.89). In both analyses most weight came from the Solomon study (Solomon et al., 2002) that scored six out of nine points with the Newcastle–Ottawa Scale.

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

The definition of CHD and stroke was based on the criteria used by the various authors.

The studies showed heterogeneity in design, definitions and quality. The quality of the included studieswas assessed according to theNewcastle–

Ottawa Scale.

Selective loss to follow up:

Not reported in this meta-analysis

Identification confounders and correction in analysis:

Not in all studies.

Funding:

This study was supported by Leiden University Medical Center.

Newcastle-Ottawa Sclae:

One study scored seven out of nine points (Shaw et al., 2008) and one study

scored six points (Solomon et al., 2002), indicating high-quality. Three studies scored four points or less (Cibula et al., 2000; Wild

et al., 2000b; Lunde and Tanbo, 2007), indicating low quality.

Gateva, 2012a

Design

Retrospective case-control

N = 301

Aim of the study:

to investigate some classical cardiovascular risk factors in a crossectional study of Bulgarian women with PCOS and/or obesity.

Inclusion criteria

patients referred to the Clinic of Endocrinology in Alexandrovska University Hospital-Sofia for the last 21 years (starting year 1990):

- patients with PCOS or

obesity that had sufficient data in their medical charts to be

included in the analysis.

- premenopausal

Exclusion criteria

- age less than 18 or higher than 45 years

- severe hepatic, cardiovascular or endocrine disorders

(incl. diabetes), or other concurrent medical illnesses.

Determinants

PCOS according to the R’dam criteria

Group 2:

Lean PCOS

N = 94

Group 3:

Obese PCOS (BMI ≥ 30 kg/m2)

N = 82

Controls

Group 1:

Obese (BMI ≥ 30 kg/m2)., no PCOS

N = 125

Primary outcomes:

BMI, W/G ratio, TG, Total Cholesterol, HDL, LDL, HOMA_IR en Gluc/Ins ratio after OGTT, Sex steroid levels, FSH.LH, SHBG

Secundary outcomes:

Follow up:

Primary outcomes:

Both waist-to-stature ratio (WSR) (area under the curve 0.75, p = 0.002) and waist circumference (WC) (area under the curve 0.77, p = 0.001) but not waist-to-hip ratio (WHR) (area under the curve 0.62, p = 0.143) were shown to be good markers of increased cardiovascular risk, insulin resistance and dislipidemia in PCOS patients. The cut-off point for WSR of 0.50 is useful and the cut-off of 80 cm for WC is more appropriate than 88 cm in detecting cardiovascular risk in PCOS patients. Androgen levels and immunoreactive insulin during an oral glucose tolerance test had lower power for predicting increased CVD risk than WC and WSR. Dislipidemia (%) (18.2%) in Normal weight PCOS and  (44.4%) in Obese PCOS. Arterial hypertension (%)  0% in normal weight PCOS and 15.8% in Obese PCOS. A statically significant difference was found regarding the results from the OGTT (blood glucose on 0 and 60min and IRI on 0, 60, and 120 min), HOMA-index, glucose/insulin ratio, the rate of hypertension and dislipidemia,  systolic and diastolic blood pressure, HDL, VLDL, and TG (Table 2), the rate of liver steatosis and the levels of ALAT, GGT, and AP, and hematological

results (WBC and ESR) (Table 3) between the lean and obese PCOS women.

Clearly defined groups:

Yes

Selection bias:

Possible, after agematching of the three groups, 74 women with obesity were excluded.

Method of assessing the outcome

appropriately:

Yes, Dislipidemia was accepted if there

was such a diagnosis in the file and/or there was any treatment for dislipidemia and/or measured during the hospital stay TC > 5.2 mmol/L, and/or HDL <1.3 mmol/L and/or TG > 1.8 mmol/L. Arterial Hypertensionwas accepted if therewas such a

diagnosis in the file and/or there was antihypertensive treatment and/or measured during the hospital stay arterial blood pressure ≥140/90mmHg.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

Not reported

Huang 2010

Design

Case-control

N = 36

Aim of the study:

The aim of this study was to evaluate the characteristics of ECGs in patients with PCOS.

Inclusion criteria

- Taiwanese women

who visited the Reproductive Endocrinology Clinic,

Wan-Fang Hospital for irregular menstruation from April 2004 to June 2009.

- patients with PCOS met two of the three criteria (oligo-ovulation and/or anovulation, a polycystic ovary, and an androgen excess)

Exclusion criteria

hyperprolactinaemia,

hypogonadotropic hypogonadism, premature ovarian

failure, congenital adrenal hyperplasia, androgensecreting

tumor, Cushing’s syndrome, disorders of

the uterus and chromo-somal anomalies.

No sufficient clinical or biochemical records, menarche less than 3 years or at >40 years of age. Subjects without a sinus rhythm and

those who did not have at least 10 evaluable leads

Determinants

PCOS

N = 24

Controls

No PCOS

N = 12

Primary outcomes:

heart rate, PR interval, QRS

duration, Sokolow-Lyon voltage (SV1 þ RV5/6), Cornell voltage (RaVL þ SV3), QT interval and QTc interval were

measured in 12-lead ECGs.

Secundary outcomes:

Follow up:

Primary outcomes:

The QRS duration was wider in patients with PCOS than those without PCOS (91+8 vs. 81+10 ms, p50.05).

The heart rate, PR interval, Sokolow-Lyon voltage, product of the QRS duration times Cornell voltage combination, QT

interval, QTc interval, QT dispersion and QTc dispersion were similar between the two groups.

Clearly defined groups:

Definition PCOS not very clear.

Selection bias:

Possible, selection of controls not clear.

Method of assessing the outcome

appropriately:

Yes

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

No

Funding:

This study was supported by the Center of Excellence for Clinical Trials and Research of Wan Fang Hospital (DOH99-TD-B-111-003) and grants (98TMU-WFH- 05-4) from Taipei Medical University-Wan Fang Hospital, (NSC96-2628-B-038-012-MY3, NSC96- 2314-B-010-006, NSC97-2314-B-038-030-MY3) from the National Science Council, Taiwan.

Study limitations

The data should be interpreted with cautions due to

the potential limitations. We did not routinely

perform echocardiography or chest X-ray with the

measurement of cardiothoracic ratio in the study

patients. It would be better if we can show the

anatomical or structural differences between the

patients with and without PCOS. In addition,

the wider QRS duration in the patients with PCOS

may be caused by a high sympathetic tone. Future

study is warranted to evaluate the autonomic function

such as heart rate variability in the patients with PCOS.

Hudecova 2011

Design

Case-control

N = 167

Aim of the study:

To examine the prevalence of the metabolic syndrome (MetS) according to the scientific statement of

the American Heart Association and the US National Cholesterol Education Program/Adult Treatment Panel III in

middle-aged Swedish women previously diagnosed with polycystic ovary syndrome (PCOS) in comparison with

age-matched healthy controls.

Inclusion criteria

At the time of the follow-up investigation all women, independent

of their current PCOS status (resolved or persisting), were included. Healthy controls, born during the same month as the index patients, were

randomly selected from population registers.

Exclusion criteria

Determinants

PCOS according to the R’dam criteria (modified

N = 84

Controls

General population

N = 83

Primary outcomes:

W/H ratio, BMI, WC.

D- SBP, Heart rate, LDL,HDL, TG, Total cholesterol, HOMA-IR fasting blood glucose levels.

Secundary outcomes:

Follow up:

Primary outcomes:

At the follow-up, 31 women (36.9%) still fulfilled the diagnostic criteria for PCOS (persisting PCOS), whereas 30 women (35.7%) no longer did (resolved PCOS). In 23 women (27.4%) PCOS criteria could not be assessed because of postmenopausal status (n = 9, 10.7%), prior surgery (n=2, 2.4%), or ongoing use of hormonal contraception (n=12, 14.3%)

The prevalence of MetS in women with PCOS (mean _ SD age, 43.0 _ 5.8 years) was 23.8% and in controls was 8.0%, and it did not differ according to PCOS phenotype at the index assessment (polycystic ovaries [PCO], oligomenorrhea, and hirsutism: 10 [22.7%]; PCO and oligomenorrhea: 8 [22.2%]) or according to the persistence of PCOS features at follow-up (persisting PCOS: 25.8%; resolved PCOS: 16.7%).

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Yes

Selective loss to follow up:

Yes, described earlier

Identification confounders and correction in analysis:

Yes

Funding:

This study was supported by grants from the Family Planning Foundation, Uppsala, Sweden and the Selander Foundation, Uppsala, Sweden.

Limitations:

- no patients without the ultrasound criteria were included in the study, and very few had PCO and hyperandrogenism only.

- the variable endocrine status in participants

Iftikhar 2012

Design

Cohort study

N = 625

Aim of the study:

We conducted a

community-based retrospective cohort study in women

with PCOS in Olmsted County, Minnesota, and compared CV risk factors and incidence of CV events with those in

women without PCOS.

Inclusion criteria

patients aged 18 to 40 years, residing in Olmsted County, who were diagnosed

with PCOS between 1966 and 1988

Exclusion criteria

active thyroid disease, prolactin elevation, adrenal or ovarian tumours, or late-onset 21-hydroxylase deficiency (as shown by either a basal serum 17-hydroxyprogesterone >2.0 ng/ml or an elevated one-hour adrenocorticotropic hormone stimulation test).

Determinants

PCOS according to the R’dam criteria

N = 309

Controls

No PCOS, age matched

N = 343

Primary outcomes:

Incidence of CVD (MI, Stroke and coronary bypass surgery), Hypertension, Type II DM, BMI, Framingham risk score.

Secundary outcomes:

Follow up:

23.7 years

Primary outcomes:

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

By using the Rochester Epidemiology Project resources, a system that links and indexes the recordsof virtually all medical providers in Olmsted County, investigators can electronically identify and review records for all patients who received a particular diagnosis during

a defined time period. International Classification of Diseases 9 code 256.4, and Berkson code 027904. The keywords used were polycystic ovaries, Stein-Leventhal syndrome, and

sclerocystic ovaries.

Ten control subjects for each case were first identified using an established computerised matching algorithm,15 where age and calendar year during their clinic visit plus three years were matching factors. Medical records were abstracted for events through 2005. A follow-up survey was done to update their CV health status through 2007.

We incorporated gender and most recent complete

data on age, total cholesterol, HDL cholesterol, smoking status, systolic blood pressure, and use of antihypertensive medications into a risk score calculator.18 In the absence of records on systolic blood pressure readings on all patients, we used physician diagnosis of hypertension or use of

antihypertensive medications as evidence of hypertension.

Selective loss to follow up:

One patient was lost to follow-up.

Identification confounders and correction in analysis:

Assessed confounders included age;

BMI; smoking status; presence of T2DM, hypertension or hyperlipidaemia; and family history of CVD, hypertension or T2DM.

Funding:

This study was funded by Solvay

Pharmaceuticals through Women’s Health Fellowship (S.I.) funds and was made possible by the Rochester Epidemiology Project (grant AR30582 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases).

Karabulut 2012

Design

Case-control

N = 89

Aim of the study:

to compare body fat

distribution in PCOS with healthy controls and to investigate

the factors associated with carotid artery intima media thickness

(IMT) and insulin resistance.

Inclusion criteria

- September 2009

and September 2010. - Study population was most probably drawn

from the patients who attended the outpatient clinic of gynecology of Denizli State Hospital

- Control group: non-smoker patients with regular menses, without signs of hirsutismus, normal hormonal status and sonographic examination, who were matched for age and BMI.

Exclusion criteria

- type-2 diabetes, hyperprolactinemia,

hypogonadotrophic hypogonadism, thyroid disease, congenital

adrenal hyperplasia, androgen secreting tumors and Cushing’s

syndrome,

- smoking

Determinants

PCOS according R’dam criteria

N = 46

Controls

age

matched

N = 43

Primary outcomes:

Anthropometrical measurements, hormonal

levels, lipid and glucose profile were evaluated. Body fat thickness

in four regions and carotid IMT were measured

Secundary outcomes:

Follow up:

Primary outcomes:

Visceral and subcutaneous fat thickness and the mean carotid artery IMT were significantly higher in PCOS subjects (p < 0.01). In correlation analysis, age, body mass index (BMI) and waist hip ratio (WHR) showed correlation with

carotid artery IMT (r = 0,55, p < 0,001; r = 0.41, p < 0.008 and r = 0.34 p = 0.03, respectively), whereas visceral fat thickness presented a correlation with HOMA-IR index as a sign of insulin resistance.

Clearly defined groups:

Yes

Selection bias:

Possible, case-control study

Method of assessing the outcome

appropriately:

Yes

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

Not applicable

Funding:

Not reported.

Liang, 2012

Design

Case-control

N = 290

Aim of the study:

To evaluate the contribution to glucose intolerance and metabolic syndromeof

obesity combinedwith the diagnostic criteria of polycystic ovary syndrome

(PCOS).

Inclusion criteria

- Period 31 August 2010 – 31 august 2011

- Taipei University

-

Exclusion criteria

(i) women who had been diagnosed with hyperprolactinemia, hypogonadotropic hypogonadism,

premature ovarian failure, congenital adrenal hyperplasia, androgen-secreting tumor,Cushing’s syndrome, disorders of the uterus and chromosomal anomalies;

(ii) women who were less than three years past menarche or who were older than 45 years;

(iii) women who received hormones

ormedication for majormedical diseases (diabetes or

cardiovascular disease); and

(iv) women who had had ovarian cysts or ovarian tumors identified by ultra-sonographic

examination.

Determinants

PCOS according R’dam criteria

N = 220

Controls

 Women who did not present

with any of three PCOS criteria

N = 70

Primary outcomes:

obesity, hyperandrogenism, oligo-anovulation and

polycystic ovary morphology on impaired glucose tolerance and metabolic disturbances

Secundary outcomes:

Follow up:

Primary outcomes:

Logistic regression analyses demonstrated that obesity was the only parameter which had a significant association with impaired glucose tolerance (OR=8.0, 95% CI 3.9–16.6) andmetabolic syndrome (OR = 54.0, 95% CI 18.3–159.8).

Clearly defined groups:

Yes, however it’s not clear where control women were recruited.

Selection bias:

Possible, see clearly defined groups.

Method of assessing the outcome

appropriately:

Yes, The insulin sensitivity index was evaluated by the HOMAIR using the following formula: HOMA-IR = [fasting insulin (in μIU/mL) × fasting glucose (in mg/dL)]/405.World

Health Organization 2006 diagnostic criteria for diabetes were employed (fasting plasma glucose ≥7.0 mmol/L or two hour plasma glucose ≥11.1 mmol/L). Impaired glucose tolerance was defined as two hour glucose levels of 7.8–11.1 mmol/L in the 75 g oral glucose tolerance test. Inwomen with impaired glucose tolerance, the fasting plasma

glucose level should be <7 mmol/L.Metabolic syndrome was defined (2005 National Cholesterol

Education Program, Adult Treatment Panel III) as

the presence of at least three of the following criteria: abdominal obesity (waist circumference >80 cm in women);serumtriglycerides≥1.7 mmol/L; serumHDL<1.3 mmol/L;

systolic blood pressure ≥130 mmHg and/or diastolicblood pressure ≥85 mmHg; and fasting plasma glucose ≥7.0 mmol/L.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

Logistic regression analysis

Funding:

This work was supported by the National Science Council Grant NSC 98-2629-B-038-001-MY3 and Taipei Medical University,Wan FangHospital Grant 98TMU-WFH-05-03.

Meyer 2012

Design

systematic review and meta-analysis

N =

From the 36 eligible full-text studies, 8 studies were included in the systematic review and 19 studies were included in the metaanalysis

Aim of the study:

This systematic review and meta-analysis was conducted to determine whether carotid

intima-media thickness (CIMT), a marker of subclinical atherosclerosis, is higher in women with PCOS compared with women without

PCOS.

Inclusion criteria

Primary articles investigating CIMT among women with PCOS and without PCOS (controls) were included if they: (i) were a peer-reviewed primaryarticle, (ii) had a study population of women with PCOS [diagnostic criteria

for PCOS specified by the National Institutes of Health (NIH)](Zawadzki and Dunaif, 1992), the European Society of Human Reproduction

and Embryology/American Society of Reproductive Medicine(ESHRE/ASRM, 2004) and/or the Androgen Excess Society (AES) criteria(Azziz et al., 2006), and were compared with controls without PCOS, (iii)

reported a measure of CIMT (unadjusted or adjusted) and (iv) were published in the English language.

. The primary search was conducted in

Ovid MEDLINE through 19 November 2010

Exclusion criteria

Studies without a control group were

excluded

Determinants

PCOS

N = 1123

Controls

Controls

N = 923

Primary outcomes:

the mean difference in CIMT

between women with PCOS and women without PCOS

Secundary outcomes:

Follow up:

Primary outcomes:

Clearly defined groups:

Yes

Selection bias:

Assesment of risk of bias by authors

Method of assessing the outcome

appropriately:

Yes

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

Depends on the study

Funding:

This study was supported through the National Heart, Lung and Blood

Institute predoctoral fellowship (NHLBI 5T32HL083825-01A2, -02, -03)

and the postdoctoral fellowship (NHLBI T32 5T32HL007055-35).

Moran, 2010

Design

systematic review and meta-analysis

N = 35 studies

Aim of the study:

to assess and weight the evidence to provide a realistic estimate of differences in prevalence and incidence of IGT, DM2 and metabolic syndrome in women with and without PCOS and to assess the effect of the confounding factor adiposity on IGT, DM2 and metabolic syndrome prevalence in women with and without PCOS.

Inclusion criteria

PCOS defined by the National Institute of Health or Rotterdam

- Eachstudy had to have at least three of the five factors as defined by each criterion.

Exclusion criteria

- studies where abnormal glucose tolerance, DM2 or IGT were exclusion criteria

Determinants

PCOS

N =

Controls

No PCOS

N =

Primary outcomes:

incidence of IGT, DM2 or metabolic syndrome in women with and without PCOS.

Secundary outcomes:

the effects of obesity

Follow up:

Primary outcomes:

A total of 2192 studies were reviewed and 35 were selected for final analysis. Women with PCOS had increased prevalence of IGT (OR 2.48, 95% CI 1.63, 3.77; BMI-matched studies OR 2.54, 95% CI 1.44, 4.47), DM2 (OR 4.43, 95% CI 4.06, 4.82; BMI-matched studies OR 4.00, 95% CI 1.97, 8.10) and metabolic syndrome (OR 2.88, 95% CI 2.40, 3.45; BMI-matched studies OR 2.20, 95% CI

1.36, 3.56). One study assessed IGT/DM2 incidence and reported no significant differences in DM2 incidence (OR 2.07, 95% CI 0.68, 6.30). One study assessed conversion from normal glucose tolerance to IGT/DM2 (OR 2.4, 95% CI 0.7, 8.0). No studies reported metabolic syndrome incidence.

Clearly defined groups:

Yes

Selection bias:

Only 5 of the 35 included studies were found to have a high risk of detection bias; however, all were found to be at high risk of selection

bias, performance bias, attrition bias (incomplete outcome data) and reporting bias (selective outcome reporting).

Method of assessing the outcome

appropriately:

A literature search was conducted (MEDLINE, CINAHL, EMBASE, clinical trial registries and hand-searching) identifying studies reporting prevalence or incidence of IGT, DM2 or metabolic syndrome in women with and without PCOS. Data were presented as odds ratio (OR) [95% confidence interval (CI)] with fixed- and random-effects meta-analysis by Mantel–Haenszel methods. Quality testing was based on Newcastle–Ottawa Scaling and The Cochrane Collaboration’s risk of bias assessment tool. Literature searching,

data abstraction and quality appraisal were performed by two investigators.

Selective loss to follow up:

It was difficult to assess withdrawals and losses to follow-up.

Identification confounders and correction in analysis:

Not applicable

Funding:

Not reported

Morgan 2012

Design

Retrospective cohort study

N =

Aim of the study:

to determine the relative risk of type 2 diabetes, cancer, large-vessel disease (LVD), and all-cause mortality for women diagnosed with PCOS.

Inclusion criteria

- diagnosis of PCOS between 1990 and 2010

Exclusion criteria

- any diagnosis of diabetes, LVD, or cancer before the index date.

Determinants

PCOS

N = 21740

Controls

No PCOS

N = 86936

Primary outcomes:

Type 2 diabetes, LVD (first record of MI, stroke, angina, or central or peripheral revascularization), cancer, all-cause mortality

Secundary outcomes:

Follow up:

Median follow-up was 4.7 yr (interquartile range

 2.0–8.6 yr) in those with PCOS and 5.8 yr (2.7–9.6 yr) in the reference group.

Primary outcomes:

In the PCOS group, there were 713 cases of first incident diabetes during the follow-up period compared with 969 in the control group. The respective rates per 1000 patient-years (kpy) were 5.7 and 1.7, a relative-risk ratio of 3.12 (95% CI  3.10–3.6). In the CPHM, the HR was 3.015 (95% CI 2.733–3.327)

Clearly defined groups:

Yes, patients were matched to four controls from within their own primary-care practice. The matching criterion was age (±2 yr), with the same requirement of a 12-month wash-in period before the case index date and the exclusion of patients with any diagnosis of diabetes, LVD, or cancer before the case index date.

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

using data from the General Practice Research Database (GPRD), a longitudinal, anonymized research database derived from nearly 600 primarycare practices in the United Kingdom.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

Yes, BMI and age

Funding:

This work was supported by Centre for Endocrine and Diabetes Sciences, Cardiff University.

Schmidt 2011

Design

Cohort study

N = 93

Aim of the study:

To examine whether postmenopausal PCOS women differ from controls regarding cardiovascular risk factors, myocardial infarction (MI), stroke and mortality.

Inclusion criteria

- PCOS according R’dam criteria, aged 40–59 yr, all Caucasian

Exclusion criteria

- Congenital adrenal

hyperplasia and androgen secreting tumors

Determinants

PCOS according to R’dam criteria

N = 25

Controls

No PCOS

N = 68

Primary outcomes:

Blood pressure, glucose, insulin, triglycerides, total cholesterol, highand

low-density lipoprotein, apolipoprotein A1 and B, fibrinogen, and plasminogen activator

inhibitor antigen were studied. Incidences of MI, stroke, hypertension, diabetes, cancer, cause of

death, and age at death

Secundary outcomes:

Follow up:

21 year

Primary outcomes:

In 2008, seven of 32 PCOS women (22%) and 13 of 95 controls (14%) had diabetes (not significant).

In 2008, 22 of 32 (69%) of the PCOS women and 39 of 95 (41%) of the controls (P=0.008) had hypertension. In 1987, the corresponding rates were 39% (13 of 33) of the PCOS women and 11% (14 of 132) of the controls (P =0.001)

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

The population participated in a study in 1987 (9) and was reinvestigated in 2008. Hypertension was defined as use of antihypertensive agent and/or an ICD diagnosis of hypertension. Diabetes was defined as medical or dietary treatment for diabetes and/or diagnosis of type 2 or type 1 diabetes. MI and stroke (including one case of

subarachnoidal bleeding among the referents) were defined as being hospitalized due to the diagnosis according to the ICD. Only the first MI or stroke was counted. Fasting venous blood samples were taken between 0700 and 1000 h

Selective loss to follow up:

Yes, reported.

Identification confounders and correction in analysis:

No

Funding:

This work was supported by grants from the Health andMedical Care Committee of the Regional Executive Board, Region

Västra Götaland, the Hjalmar Svensson Research Foundation, the Go¨ teborg Medical Society, the Sahlgrenska University Hospital ALF agreement, the Swedish Heart Lung Foundation, and

the Swedish Research Council.

Wang 2011b

Design

Cohort study

N = 1127

Aim of the study:

To estimate whether women aged 19–32 who fulfilled National Institutes of Health

(NIH) criteria for polycystic ovary syndrome (PCOS) would be at a higher risk for subsequent development of incident diabetes, dyslipidemia, and hypertension, and to estimate whether

normal-weight women with PCOS would have the same degree of cardiovascular risk as

overweight women with PCOS.

Inclusion criteria

- Cardia study

- Age 18-30 years at baseling in 1985-1986

- Attended at the Year 15 examination

- ≥ 1 ovary

- Not pregnant

- PCOS at ages 20–32: if they reported oligomenorrhea between 20–30 years of age and either reported hirsutism between

20–30 years of age or fulfilled criteria for biochemical hyperandrogenism at Year 2.

Exclusion criteria

- FSH >40mIU/ml

- self-report of no menses within the last 12 months (Year 16)

- OAC

- Women who fulfilled criteria for PCOS at Year 16 only

- Prevalent cases of diabetes, hypertension or dyslipidemia

Determinants

PCOS according to NIH criteria

N = 53

Controls

No PCOS

N = 1074

Primary outcomes:

Hypertension, diabetes and dyslipidemia

Secundary outcomes:

Follow up:

18 years

Primary outcomes:

Clearly defined groups:

Yes

Selection bias:

Not expected

Method of assessing the outcome

appropriately:

Participants underwent a baseline exam and

follow-up exams at Years 2, 5, 7, 10, 15, and 20 with retention rates of 91, 88, 81, 79, 74,

and 72%, respectively. The CARDIA Women’s Study (CWS) is an ancillary study of

women who attended an additional examination at Year 16; CWS was designed to examine

the role of androgens and polycystic ovaries in the development of cardiovascular disease. The study sample included 1127 women present at both the Year 2 (ages 20–32) and Year 16 (ages 34–46) examinations.

Diabetes was defined as a fasting plasma glucose ≥ 126 mg/dL or use of diabetic medications. Dyslipidemia was defined as LDL > 160 mg/dL, HDL < 40 mg/ dL, triglycerides > 200 mg/dL, or use of cholesterol-lowering medications. Hypertension was defined as blood pressure ≥ 140/90 mmHg or the use of antihypertensive medications.

Selective loss to follow up:

Women were only included when present at both the Year 2 (ages 20-32) and Year 16 (ages 34-36) examinations.

Identification confounders and correction in analysis:

adjusted for age, race, BMI, education, and parity obtained at Year 2, and family history of diabetes obtained at Year 0.

Funding:

Supported by the National Heart, Lung and Blood Institute (N01-HC-95095, N01-HC-48047, N01-HC-48048, N-01-HC-48049, N01-HC-48050, N01-HC-05187, N01-HC-45205, N01-HC-45204, N01-HC-45134), as was the CARDIA Women’s Study (R01-HL-065611).

Wiltgen 2010

Design

Case-control

N =

Aim of the study:

to compare clinical, hormonal, and metabolic variables in women with classic polycystic

ovary syndrome (PCOS), in ovulatory women presenting hirsutism, normal androgen levels, and polycystic ovaries

(HþPCO), and in a group with isolated hirsutism (IH) presenting with normal ovaries and androgen levels.

Inclusion criteria

- outpatients aged 14–35 years and consulting for

hirsutism at the Gynecological Endocrinology Unit at Hospital de Clınicas de Porto Alegre, Brazil

- PCOS according NIH and R’dam criteria

- Controls: women in the same age range, with regular and proven ovulatory cycles (luteal-phase P >3.8 ng/mL)

- None of the subjects or controls

had received oral contraceptives, progestins, antiandrogens,

antihypertensive, insulin-sensitizing, or weight-control drugs for

at least 3 months before the study

Exclusion criteria

other hyperandrogenic

disorders (nonclassic congenital adrenal hyperplasia,

Cushing’s syndrome, androgen-secreting neoplasms), thyroid disorders,

or hyperprolactinemia

Determinants

PCOS according NIH and R’dam criteria

N = 195

Controls

No PCOS

N = 25

Primary outcomes:

Secundary outcomes:

Follow up:

Primary outcomes:

Clearly defined groups:

Yes, but not clear where PCOS- and control women were recruited

Selection bias:

Possible, see ‘clearly defined groups’

Method of assessing the outcome

appropriately:

The biochemical and ultrasound evaluation were made between days 2 and 10 of the menstrual cycle (and on days 21–24 for P levels) or on any

day if the patient was amenorrheic. After an overnight fast, blood samples were drawn from an antecubital vein for biochemical determinations

and oral glucose tolerance test. All samples were obtained between 8 AM and 10 AM.

Selective loss to follow up:

Not applicable

Identification confounders and correction in analysis:

Yes, BMI

Funding:

Not reported

Tabel 4.2.1: risico’s op het krijgen van een fataal en/of niet-fataal CV gebeurtenis bij vrouwen met PCOS (in de voorgeschiedenis) ten opzichte van vrouwen zonder PCOS

Studie

Definitie PCOS

N

Gem. FU (jr)

Uitkomst

Aantal gebeurtenissen

(N)

Gecorrigeerde risico’s (95% BI)

Gecorrigeerd voor

A

NA

A

NA

Lunde, 2007

WHOII, NIH, R’dam, AE

131

723

15-25

NF IHZ

2

4

RR 2.76 (0.51-14.91)

Geen

NF B

0

8

RR 0.32 (0.02-5.56)

Shaw, 2008

WHOII, NIH, R’dam, AE

104

286

6

F HVZ &

NF IHZ

22

32

HR 1.59 (1.19-.2.12)

Leeftijd, DM, BMI,

mate van slagaderverkalking

Wild, 2000

WHO II, AE, R’dam

319

1060

31

NF IHZ

15

42

OR 1.2 (0.5-2.6)

BMI

NF B

10

13

OR 3.4 (1.2-9.6)

Iftikhar 2012

R’dam

309

343

23.7

NF MI

15

16

HR 0.74  (0.32-1.72)

Lft, BMI, fertiliteits

behandeling,

postmenopauzale hormoon

behandeling,

familie-voorgeschiedenis

voor hypertensie

NF IAP

10

6

HR 1.32  (0.42-4.13)

NF B

6

7

HR 1.05 (0.28-3.92)

F HVZ

4

2

HR 5.67  (0.51-63.7)

Schmidt

2011

R’dam

32

95

19

F & NF IHZ

4

9

RR 1.32 (0.44-3.99)

Geen

F & NF B

6

10

RR 1.78 (0.70-4.51)

F & NF HVZ

10

20

RR 1.48 (0.78-2.83)

Morgan

Read code (UK; huisartsen

21 740

86 936

4.7/ 5.8

F & NF HVZ

43

162

HR 1.08 (0.77-1.51)

Lft, aantal contacten

met huisarts, BMI

Afkortingen: A = aangedaan door ziektebeeld; AE = Androgen-excess;

B = beroerte; BI = betrouwbaarheidsinterval; BMI = body mass index;  DM = diabetes mellitus; F = fataal; FU = follow-up;

HR = hazard ratio; HVZ = hart- en vaatziekten; IAP = instabiele angina pectoris; IHZ = ischemische hartziekten; lft = leeftijd;

MI = myocard infarct; NA = niet aangedaan door ziektebeeld; NF = niet fataal; NIH = National Institutes of Health; OR = odds ratio;

R’dam = Rotterdam; RR = relatief risico; UK = United Kingdom; WHO II = World Health Organization

Tabel 4.2.2: risico’s op het ontwikkelen van hypertensie in vrouwen met PCOS vergeleken met vrouwen zonder PCOS.

Studie

Definitie PCOS

N

Gem FU (jr)

Gem lft FU

Definitie HT

Aantal vrouwen met HT (N)

Gecorrigeerde risico (95% BI)

Gecorrigeerd voor

A

NA

A

NA

A

NA

Onderzoeken met patiënten series waarin vrouwen met PCOS zijn vergeleken met vrouwen zonder PCOS

Cibula

WHOII, NIH, R’dam, AE

28

752

NA

51.9

51.0

II/III/IV

14

290

RR 1.30 (0.89-1.90)

Geen

Gateva 2012a

R’dam

125

82

nvt

26.5

26.3

II/III/IV/VII

16

15

RR 0.70 (0.37-1.34)

NB: controle groep bestaat uit obese vrouwen.

Iftikhar, 2012

R’dam

309

343

23.7

44.4

48.8

I/II

80

73

RR 1.22 (0.92-1.61)

Geen

Liang, 2012

R’dam

220

70

NA

26.9

28.3

VI

68

20

RR 1.08 (0.71-1.65)

Geen

Lunde, 2007

WHOII, NIH, R’dam, AE

131

723

15-25

Nb

Nb

II

11

58

RR 1.05 (0.56-1.94)

Geen

Schmidt

2011

R’dam

32

95

19

70.4

70.7

II of V

22

39

RR 1.67 (1.20-2.34)

Geen

Shaw, 2008

WHOII, NIH, R’dam, AE

104

286

6

62.5

65.8

Nb

63

126

RR 1.38 (1.12-1.68)

Geen

Wang, 2011b

NIH

53

1074

18

38-50

II, III, IV

14

274

OR 1.8 (0.9-3.6)

Lft, ras, BMI baseline, educatie, pariteit, familieVG DM, BMI

RR 1.04 (0.65-1.64)

Geen

Wild, 2000

WHO II, AE, R’dam

319

1060

31

Nb

Nb

VII

72

201

OR 1.4 (0.9-2.0)

BMI

RR 1.19 (0.94-1.51)

Geen

Wiltgen, 2010

NIH/ R’dam

195

25

NA

22.3

29.7

VI

`78

3

RR 3.33 (1.14-9.77)

Geen

Afkortingen: A = aangedaan door ziektebeeld;  BI = betrouwbaarheidsinterval; BMI = body mass index;  DM = diabetes mellitus;  FU = follow-up; lft = leeftijd; N = aantal; NA = niet aangedaan door ziektebeeld; nb = niet bekend; OR = odds ratio; RR = relatief risico; VG = voorgeschiedenis;


Definitie hypertensie: I zelfrapportage; II medicatie gebruik; III systolische bloeddruk ≥140 mmHg; IV diastolische bloeddruk ≥90 mmHg; V  ICD code; VI  systolische bloeddruk ≥135 mmHg en/of  diastolische bloeddruk ≥85 mmHg; VII rapportage huisarts

Tabel 4.2.3: risico’s op het ontwikkelen van diabetes mellitus in vrouwen met PCOS vergeleken met vrouwen met zonder PCOS

Studie

Definitie PCOS

N

Gem FU (jr)

Gem lft FU

Definitie DM

Aantal vrouwen met DM (N)

Gecorrigeerde risico (95% BI)

Gecorrigeerd voor

A

NA

A

NA

A

NA

Meta-analyse

Moran, 2010

NIH, R’dam

441

63

Nb

Nb

Nb

III/ IV/VIII/ IX

19

63

OR 4.00 (1.97-8.10)

BMI

12 105

56 959

1063

1274

RR 3.93 (3.63-4.25)

Geen

Cross-sectionele onderzoeken

Gambineri, 2012

NIH

249

Nb

16.9

45.9

Nb

Nb

39.3%

5.8%

Nb

Nb

Hudecova, 2011

R’dam

84

87

13.9

43

43.7

II

7

0

-

Geen

Iftikhar 2012

R’dam

309

343

23.7

44.4

48.8

I, II, VIII

2

9

RR 0.25 (0.05-1.13)

Liang, 2012

R’dam

220

70

NA

26.9

28.3

III/IV

11

1

RR 3.50 (0.46-26.63)

Geen

Lunde, 2007

WHOII, NIH, R’dam, AE

131

723

15-25

Nb

Nb

II

1

nb

SIR 6.1 (2.2-13.0)

Morgan, 2012

Read code classification (UK- huisarts)

21 740

86 936

4.7/ 5.8

27.1

27.1

V

713

969

OR 3.02 (2.73-3.33)

Lft, aantal contacten met huisarts, BMI

RR 2.94 (2.67-3.24)

Geen

Schmidt

2011

R’dam

32

95

19

70.4

70.7

II of VIII

7

13

RR 1.60 (0.70-3.65)

Geen

Wang, 2011b

NIH

53

1074

18

38-50

II of IX

12

137

OR 2.6 (1.3-5.3)

Lft, ras, BMI baseline, educatie, pariteit, familieVG, DM, BMI

RR 1.77 (1.05-2.99)

Geen

Wild, 2000

WHO II, AE, R’dam

319

1060

31

Nb

Nb

V/VIII

22

32

OR 2.2 (0.9-5.2)

BMI

RR 2.28 (1.35-3.87)

Wiltgen, 2010

NIH/ R’dam

195

25

NA

22.3

29.7

VII

7

0

-

Geen

Afkortingen: A = aangedaan door ziektebeeld;  BI = betrouwbaarheidsinterval; BMI = body mass index;  DM = diabetes mellitus;  FU = follow-up; lft = leeftijd; N = aantal; NA = niet aangedaan door ziektebeeld; nb = niet bekend; OR = odds ratio; RR = relatief risico; SIR =  standardized incidence ratio; VG = voorgeschiedenis;

Definitie DM: I zelfrapportage; II medicatie gebruik; III nuchtere glucose ≥ 7.0 mmol/l; IV random glucose ≥ 11.1 mmol/l; V  ICD-9 codes; VI WHO; criteria; VII = onbekend; VIII medisch dossier; IX nuchtere glucose ≥ 9.0 mmol

Tabel 4.2.4.: risico’s op het ontwikkelen van het metabool syndroom in vrouwen met PCOS in vergelijking met vrouwen zonder PCOS

Studie

Definitie PCOS

N

Gem FU (jr)

Gem lft FU

Definitie MetS

Aantal vrouwen met MetS(N)

Gecorrigeerde risico (95% BI)

Gecorrigeerd voor

A

NA

A

NA

A

NA

Meta-analyses

Moran, 2010

NIH/ R’dam

273

276

Nb

NB

I - IV

61

37

OR 2.20 (1.36-3.56)

BMI

2256

4130

508

731

RR 1.27 (1.15-1.41)

Geen

Onderzoeken met patiënten series waarin vrouwen met PCOS zijn vergeleken met vrouwen zonder PCOS

Liang, 2012

R’dam

220

70

Nb

26.9

28.3

I

66

10

RR 2.10 (1.14-3.86)

Geen

Wiltgen, 2010

NIH/ R’dam

195

25

Nb

22.3

29.7

I

61

0

-

Geen

Hudecova, 2011

R’dam

84

87

13.9

43

43.7

I

22

7

RR 2.96 (1.30-6.76)

Geen

Rahmanpout, 2012

NIH

30

71

Nb

17.7

17.7

IV

10

8

RR 2.96 (1.30-6.76)

Geen

Afkortingen: A = aangedaan door ziektebeeld;  BI = betrouwbaarheidsinterval; BMI = body mass index;  FU = follow-up; lft = leeftijd; MetS = metabool syndroom; N = aantal; NA = niet aangedaan door ziektebeeld; nb = niet bekend; OR = odds ratio; RR = relatief risico

Definitie metabool syndroom: I Adult Treatment Panel III (ATPIII); II American Heart Association/National Heart Lung Blood Institute (AHA); III  World Health Organization (WHO); IV International Diabetes Federation (IDF)

Tabel 4.2.5.: risico’s op het ontwikkelen van dyslipidemie in PCOS in de voorgeschiedenis vergeleken met vrouwen  zonder PCOS in  de voorgeschiedenis

Studie

Definitie PCOS

N

Gem FU (jr)

Gem lft FU

Definitie dyslip

Aantal vrouwen met dyslipidemie (N)

Gecorrigeerde risico (95% BI)

Gecorrigeerd voor

A

NA

A

NA

A

NA

Onderzoeken met patiënten series waarin vrouwen met PCOS zijn vergeleken met vrouwen zonder PCOS

Hudecova, 2011

R’dam

84

87

13.9

43

43.7

IV

18

7

RR 2.66 (1.17-6.05)

Geen

Liang, 2012

R’dam

220

70

NA

26.9

28.3

IV

29

3

RR 3.08 (0.97-9.79)

Geen

Schmidt

2011

R’dam

25

68

19

70.4

70.7

II

6

18

RR 0.91 (0.41-2.02)

Geen

Shaw, 2008

WHOII, NIH, R’dam, AE

104

286

6

62.5

65.8

Nb

64

166

RR 1.06 (0.88-1.27)

Geen

Wang, 2011b

NIH

53

1074

18

38-50

II/V/VII/VIII

18

238

OR 2.0 (1.0-3.9)

Lft, ras, BMI baseline, educatie, pariteit, familievoorgeschiedenis DM, BMI

Wild, 2000

WHO II, AE, R’dam

319

1060

31

Nb

Nb

IX

Nb

Nb

OR 3.2 (1.7-6.0)

BMI

Wiltgen, 2010

NIH/ R’dam

195

25

NA

22.3

29.7

IV

45

4

RR 1.44 (0.57-3.67)

Geen

Afkortingen: A = aangedaan door ziektebeeld; BMI = body mass index;   NA = niet aangedaan door ziektebeeld; nb = niet bekend;

Definitie dyslipidemie: I zelfrapportage; II medicatie gebruik; III  cholesterol ≥5.0 mmol/l ; IV triglyceriden ≥1.7 mmol/l; V HDL ≤1.0 mmol/l; VI Totaal cholesterol ≥6.2 mmol/l; VII LDL >4.1 mmol/l; VIII  triglyceriden ≥2.3 mmol/l; IX  cholesterol ≥7.8 mmol/l