Influence of obesity on atherogenic dyslipidemia in women with polycystic ovary syndrome

Influence of obesity on atherogenic dyslipidemia inwomen with polycystic ovary syndromeAntonio Hern�andez-Mijares*,†,‡,§,1, Celia Ba~nuls*,‡,§,1, Marcelino G�omez-Balaguer*, Marina Bergoglio*, Victor M.V�ıctor*,†,§,¶,** and Milagros Rocha*,†,§,¶

*Service of Endocrinology, University Hospital Dr. Peset, Valencia, Spain, †Foundation for the Promotion of Healthcare andBiomedical Research in the Valencian Community (FISABIO), Valencia, Spain, ‡Department of Medicine, Faculty of Medicine,University of Valencia, Valencia, Spain, §Institute of Health Research INCLIVA, Valencia, Spain, ¶CIBER CB06/04/0071Research Group, CIBER Hepatic and Digestive Diseases, University of Valencia, Valencia, Spain, **Department of Physiology,Faculty of Medicine, University of Valencia, Valencia, Spain

ABSTRACT

Background Obesity is known to underlie, at least partially, dyslipidemia in polycystic ovary syndrome (PCOS),but it is unclear whether PCOS status per se increases the risk of alterations of lipoprotein subfractions, whichdiffer in size and atherogenic potential. Our objective was to evaluate whether PCOS influences lipoproteinprofile and LDL and HDL subfractions and to study the impact of obesity on these parameters.

Materials and methods This was a case–control study conducted in an academic medical centre. The studypopulation consisted of 54 women of fertile age with PCOS and 60 controls adjusted for age and BMI.Biochemical lipid profile and LDL and HDL lipoprotein subfractions (measured using Lipoprint System).

Results Lean PCOS women exhibited lower HDL cholesterol and apolipoprotein AI levels than controls,although these differences were not associated with alterations of lipoprotein subfractions. All obese subjects,whether PCOS or controls, displayed lipid parameters typical of atherogenic dyslipidemia, although the formergroup had lower levels of large HDL, higher levels of small HDL subfractions and a higher percentage of VLDLthan the latter. These differences were associated with a greater prevalence of non-A LDL pattern (25.0%) inobese PCOS subjects than in obese controls (4.3%).

Conclusions PCOS does not constitute an additional risk factor for cardiovascular disease in lean women, butleads to a lipid profile characteristic of atherogenic dyslipidemia and an altered pattern of lipoprotein subfractionwhen associated with obesity.

Keywords Atherogenic dyslipidemia, HDL subfractions, LDL subfractions, obesity, polycystic ovary syndrome.

Eur J Clin Invest 2013; 43 (6): 549–556

Introduction

Polycystic ovary syndrome (PCOS) is a common endocrine and

metabolic disorder that occurs in 5–10% of premenopausal

women [1,2]. Symptoms vary among women and include

menstrual irregularities, infertility, hyperandrogenism and

hirsutism [2]. In recent years, several studies have related car-

diometabolic events with PCOS, including insulin resistance,

vascular dysfunction, hypertension and dyslipidemia [3,4]. In

particular, a spectrum of abnormal lipid and lipoprotein pro-

files is found in patients with PCOS, such as low HDL choles-

terol (HDLc), elevated triglyceride concentrations and, less

commonly, increased total and LDL cholesterol (LDLc) levels

[5]. The findings of research in this field, however, have been

highly variable and have depended on several factors, includ-

ing obesity and, consequently, degree of insulin resistance.

Intriguingly, reduced HDLc and increased triglycerides, in

addition to a higher proportion of small and dense LDL parti-

cles (sdLDL), lead to atherogenic dyslipidemia, which has

emerged as an important marker of the increased cardio-

vascular disease (CVD) risk observed in patients with obesity,

metabolic syndrome, insulin resistance and type 2 diabetes

mellitus [6]. In such cases, a simple quantitative measurement

of LDLc concentration may be misleading, because LDL parti-

cles are heterogeneous in terms of size, density and physical

properties.1These authors contributed equally to the work.

European Journal of Clinical Investigation Vol 43 549

DOI: 10.1111/eci.12080

ORIGINAL ARTICLE

In addition to this, it is also well known that HDLc does

not represent a sum of identical particles, but rather a col-

lection of discrete subfractions that differ in physicochemical

properties; namely size, density, composition and charge [7].

It seems that HDL atherogenicity increases as particle size

decreases, and that, in this population, small HDL are more

atherogenic than large HDL. These qualitative alterations in

lipoprotein metabolism could account for the increased risk

of cardiovascular disease in PCOS; however, this is a subject

that requires clarification, as studies of lipoprotein sub-

fractions in PCOS are scarce and have produced conflicting

results [8–14]. These discrepant results may be due to the

varying definitions of PCOS that have evolved over time,

different methods used to assess subfraction distribution, and

the heterogeneous baseline characteristics of the patients

evaluated.

Recently, it has been concluded that body mass index (BMI)

has a greater impact on insulin resistance than testosterone

levels in women with PCOS [15]. Although obesity, a highly

prevalent condition associated with PCOS, is known to

undermine insulin resistance, which in turn enhances CVD

risk, it is unclear what role certain adipose-related factors may

play. This information is important to fully understand the

metabolic effects of PCOS and to improve long-term evaluation

of cardiovascular risk.

In this context, the purpose of the present study was to

evaluate whether the diagnosis of PCOS per se has an influence

on LDL and HDL subfractions and to assess the impact of

obesity on lipoprotein profile in young women with PCOS

compared with age- and BMI-adjusted controls.

Subjects and methods

SubjectsFifty-four women of fertile age with PCOS and sixty healthy

women adjusted by age and BMI were recruited at the Out-

patient’s Department of the Endocrinology Service of the Dr.

Peset University Hospital. Subjects were included in the study

if they fulfilled all the Rotterdam criteria for diagnosis of PCOS

[16]. These criteria consisted of oligoovulation (cycles lasting

longer than 35 days or less than 26 days), elevated free test-

osterone levels (>0.5 ng/dL; the cut-off level for free testoster-

one level was the mean � 2SD according to normal levels in

controls), hirsutism (total Ferriman–Gallwey score >7) andpolycystic ovaries identified by transvaginal ultrasonography

[defined as the presence of at least 12 small (2–9 mm) follicles

in each ovary]. Control subjects had regular menses, levels of

testosterone lower than 0.9 ng/mL and no family history of

PCOS, diabetes or familial combined hyperlipidemia. To study

the influence of obesity on lipoprotein metabolism, PCOS and

control subjects were divided into two groups – lean or obese –

according to whether their BMI was lower than 25 kg/m2 or

higher than 30 kg/m2, respectively. Specifically, overweight

women were excluded.

Exclusion criteria were pregnancy or lactation, galactorrhoea

or any endocrine or systemic disease that could affect repro-

ductive physiology, organic, malignant, haematological, infec-

tious or inflammatory disease, diabetes mellitus, a history of

cardiovascular disease and the taking of lipid lowering or

antihypertensive drugs.

In all subjects, anthropometric, blood pressure and analytical

evaluations were performed, and weight (kg), height (m) and

waist (cm) were measured. Body mass index [BMI = weight

(kg)/height2 (m)] was then calculated.

Written informed consent was obtained from all subjects

prior to participation. The study was approved by the Ethics

Committee of Hospital Dr Peset and was performed in

accordance with the Helsinki Declaration.

Blood sampling/Laboratory methodsBlood was collected from the antecubital vein at

8.00–10.00 a.m. on the second/third day of the menstrual

cycle (follicular phase), after 12 h of fasting. Glucose levels

were measured using enzymatic techniques and a Dax-72

autoanalyzer (Bayer Diagnostic, Tarrytown, NY, USA). Insu-

lin was measured by an enzymatic luminescence technique.

Insulin resistance was calculated by homoeostasis model

assessment [HOMA-IR = (fasting insulin (lU/mL) 9 fasting

glucose (mmol/L)/22.5]. Follicle-stimulating hormone (FSH),

luteinising hormone (LH), total testosterone, androstendione,

dehydroepiandrosterone sulphate (DHEAS), estradiol, 17

a-hydroxyprogesterone (17a-OH-P) and sex hormone-binding

globulin (SHBG) were measured by specific chemilumines-

cence techniques in the Clinical Analysis Service of the

Hospital.

Free androgenic index (FAI) was calculated as total testos-

terone level divided by SHBG level – both in nanomols per litre

– and then multiplying by 100. High-sensitivity C-reactive

protein (hs-CRP) was quantified by a latex-enhanced immu-

nonephelometric assay (Behring Nephelometer II; Dade Beh-

ring, Inc., Newark, DE, USA) with an intra-assay coefficient of

variation of 8.7% and sensitivity of 0.01 mg/L. Levels higher

than 10 mg/L were specifically excluded to rule out the

influence of acute inflammation.

Measurement of serum lipids and lipoproteinsubclassesTotal cholesterol and triglycerides were measured by means

of enzymatic assays, and HDLc concentrations were recorded

with a Beckman LX-20 autoanalyzer (Beckman Coulter, La

Brea, CA, USA) using a direct method. The intraserial varia-

tion coefficient was <3.5% for all determinations. LDLc

550 ª 2013 Stichting European Society for Clinical Investigation Journal Foundation. Published by John Wiley & Sons Ltd

A. HERN�ANDEZ-MIJARES ET AL. www.ejci-online.com

concentration was calculated using the Friedewald method.

Non-HDLc concentration was calculated according to the

difference between total cholesterol and HDLc. Atherogenic

index of plasma (AIP) was obtained by calculating the loga-

rithm of the ratio of plasma concentration of triglycerides to

HDLc. Apolipoprotein AI (Apo AI) and B (Apo B) were

determined by immunonephelometry (Dade Behring BNII,

Marburg, Germany) with an intra-assay variation coefficient

of <5.5%.

LDL and HDL subfractions were separated using the

Quantimetrix Lipoprint system (Redondo Beach, CA) [17].

High-resolution polyacrylamide gel tubes specific for LDL or

HDL were used for electrophoresis. Twenty-five microlitres of

sample were mixed with 200 lL (for LDL) or 300 lL (for

HDL) of Lipoprint loading gel containing Sudan Black B dye

to stain the lipoproteins. This mixture was then placed on the

upper part of the polyacrylamide gel tube. After 30 min of

photopolymerisation at room temperature, the samples in

each gel tube were electrophoresed with 3 m�A for 55 or

60 min for LDL and HDL, respectively. The tubes were then

maintained in the dark for 1 h, after which densitometry was

performed at 610 nm. Raw data from the densitometer were

imported into a Microsoft excel spreadsheet, and subfractions

were identified and quantified using a computerised method

developed for the Quantimetrix Lipoprint system and NIH

image program version 1.62 (Bethesda, MD, USA) for research

use. The Liposure� (Quantimetrix Corporation, Redondo

Beach, CA, USA) was used for quality control. Very low-

density lipoprotein (VLDL), 7 LDL and 10 HDL subclasses

were quantified and further classified as large, intermediate or

small subfractions. The LDL electrophoretic profile allows

three patterns to be defined: pattern A/large and buoyant

LDL (cut-off size more than 268 �A); intermediate pattern (cut-

off size more than 265 and � 268 �A); and pattern B/sdLDL

(cut-off size � 265 �A).

Statistical analysisSPSS 15.0 software (SPSS Statistics Inc., Chicago, IL, USA) was

employed for statistical analysis. Continuous variables were

expressed as mean and standard deviation (SD) or as median

and 25th and 75th percentiles for parametric and nonpara-

metric data, respectively. Qualitative data were expressed as

percentages. Parametric data were compared with an inde-

pendent T-test or Mann–Whitney U when the dependent vari-

able was not normally distributed. Pearson’s correlation or

Spearman’s correlation coefficients were used to measure the

strength of the association between two variables of parametric

and nonparametric data, respectively. The Chi-square test was

used to compare proportions among groups of subjects. A

confidence interval of 95% was employed for all the tests and

differences were considered significant when P < 0.05.

Results

The clinical and metabolic characteristics of participants

according to BMI are shown in Table 1. No differences in age or

BMI were found between PCOS subjects and controls in lean or

obese groups. All PCOS subjects displayed higher levels of

insulin, HOMA-IR index and androgens (total testosterone,

androstendione, DHEAS) and lower levels of SHBG than their

respective controls. Lean women with PCOS showed higher

waist circumference and diastolic blood pressure (P < 0.05)

than their respective controls. As expected, obesity was asso-

ciated with higher waist circumference and blood pressure and

altered carbohydrate metabolism (glucose, insulin and HOMA-

IR index) in the whole obese population.

No differences in lipid parameters were observed between

lean PCOS subjects and controls, except for a slight reduction in

HDLc and Apo AI levels, which nonetheless were within the

range recommended by the ATPIII (Table 2). However, obesity

was related with a proatherogenic lipid profile consisting of

higher LDLc, non-HDLc, triglyceride and Apo B levels, higher

ApoB/Apo AI ratio, and lower HDLc and Apo AI levels

(P < 0.05). Furthermore, the highest levels of LDLc, non-HDLc,

triglycerides and AIP were detected in the obese PCOS subjects.

As shown in Table 2, lipoprotein subfractions did not differ

between lean PCOS subjects and controls, whereas obesity

aggravated lipoprotein subfractions with respect to women of

normal weight – both control and PCOS –, who showed higher

levels of VLDL, intermediate and small HDL and lower levels

of large HDL. Although both obese groups – PCOS and con-

trols – displayed an altered lipid profile associated with lower

LDL particle size, a trend was observed towards a greater

percentage of sdLDL among those with PCOS (data not

shown). In addition, a significantly greater number of obese

women with PCOS showed a non-A LDL pattern than did

obese controls (v2, P < 0.05; Fig. 1). In particular, 97.3% and

100% of lean controls and PCOS subjects, respectively, exhib-

ited an A LDL pattern. However, 4.3% of control obese women

showed an intermediate LDL pattern, while 10.0% and 15.0% of

obese women with PCOS showed an intermediate and B LDL

pattern, respectively (Fig. 1), in accordance with a concomitant

increase in the sdLDL, which was associated with a higher

VLDL percentage. In fact, LDL particle size negatively corre-

lated with percentage of VLDL in PCOS (r = �0.407; P = 0.002)

subjects but not in controls (r = �0.091, P = 0.488).

To test the hypothesis that insulin resistance and/or andro-

gens are predictors of the presence of a more atherogenic lipid

profile within this pathology, we analysed possible correlations

between lipid parameters and insulin, HOMA-IR or FAI in the

two different populations (control and PCOS subjects; Table 3).

Insulin and HOMA-IR correlated with triglyceride levels, AIP,


PCOS, OBESITY AND LIPOPROTEIN SUBFRACTIONS

HDL and large and small HDL percentages in both popula-

tions. LDL particle size was specifically associated with insulin

and HOMA-IR in controls, whereas non-HDLc and VLDL

correlated with insulin and HOMA-IR only in women with

PCOS. In the control group, FAI was not associated with any of

the parameters analysed, while conversely, FAI correlated with

all the parameters except LDL particle size in women with

PCOS. Thus, according to the results shown in Table 2, the

atherogenic dyslipidemia observed in obese PCOS subjects was

related with androgens.

Discussion

The results of the present study show that lean women with

PCOS exhibit an altered lipoprotein profile characterised by

low levels of HDLc and Apo AI. As expected, obesity was

associated with a more atherogenic lipoprotein profile in both

PCOS and control women, evident in the reduced HDLc levels

and increased LDLc, non-HDLc and triglyceride levels in said

subjects. Furthermore, obesity was related with an increase in

small HDL subfractions and VLDL percentage and smaller LDL

particle size. PCOS exacerbated the deleterious effects of obes-

ity on lipoprotein profile; higher LDLc, non-HDLc and tri-

glyceride levels, AIP and VLDL percentage and small HDL

subfractions associated with a higher prevalence of atherogenic

B LDL pattern were observed in our obese PCOS subjects.

While insulin resistance is not a diagnostic criterion for

PCOS, it is recognised that this common hyperandrogenic dis-

order among women is a multifaceted syndrome in which

insulin resistance and obesity play critical roles. The

Table 1 Clinical and metabolic characteristics of patients with PCOS and control women according to BMI

BMI < 25 kg/m2 BMI > 30 kg/m2

Control PCOS Control PCOS

Subjects (n) 37 34 23 20

Age (years) 25.2 � 3.6 23.3 � 4.2 26.9 � 5.4 25.4 � 7.5

BMI (kg/m2) 21.2 � 1.8 21.2 � 1.9 35.1 � 2.4† 35.4 � 4.6‡

Waist (cm) 73.2 � 6.2 76.6 � 7.0* 103.6 � 8.5† 104.7 � 12.8‡

Systolic BP (mmHg) 105 � 7 108 � 11 115 � 13† 122 � 16‡

Diastolic BP (mmHg) 65 � 7 70 � 9* 71 � 11† 78 � 14‡

Glucose (mg/dL) 80.2 � 9.2 81.3 � 9.1 90.6 � 9.2† 88.3 � 10.8‡

Insulin (lU/mL) 6.4 � 2.7 8.5 � 5.3* 11.7 � 6.3† 20.9 � 15.6*,‡

HOMA-IR index 1.27 � 0.55 1.75 � 1.07* 2.63 � 1.38† 4.78 � 4.44*,‡

Total testosterone (ng/mL) 0.47 � 0.20 0.74 � 0.47* 0.42 � 0.20 0.99 � 0.47*

FSH (mU/mL) 4.4 � 2.8 4.9 � 1.3 4.16 � 1.68 4.41 � 1.25

LH (mU/mL) 4.1 � 3.56 6.3 � 4.5* 6.41 � 5.71 4.33 � 2.61

Estradiol (pg/mL) 68.9 � 79.6 41.6 � 32.5 91.5 � 62.7 39.1 � 13.4*

Androstenedione (ng/mL) 2.50 � 1.03 3.85 � 1.78* 2.42 � 0.99 4.04 � 2.03*

DHEAS (lg/dL) 256.3 � 88.6 370.9 � 153.9* 189.4 � 93.2† 294.0 � 135.8*

17a-OH-P (ng/mL) 0.61 � 0.68 0.83 � 0.41 0.87 � 0.42 0.71 � 0.35

FAI 2.13 � 1.31 5.26 � 5.41* 2.90 � 2.26 13.02 � 12.08‡

SHBG (nM) 121.9 � 76.8 70.4 � 37.5* 88.7 � 66.1 42.1 � 38.7*,‡

hs-CRP (mg/L) 1.58 � 1.38 2.01 � 2.38 3.67 � 2.73† 4.51 � 3.01‡

Data are expressed as mean � SD.

BMI, body mass index; BP, blood pressure; FSH, follicle-stimulating hormone; LH, luteinising hormone; DHEAS, dehydroepiandrosterone sulphate; 17a-OH-P,

17 a -hydroxyprogesterone; FAI, free androgenic index; HOMA-IR, homoeostasis model assessment; PCOS, polycystic ovary syndrome; SHBH, sex hormone-

binding globulin; hs-CRP, high-sensitivity C-reactive protein.

*P < 0.05 when comparing controls and PCOS (lean or obese) using an unpaired Student’s t-test.†P < 0.05 when comparing lean and obese controls using an unpaired Student’s t-test.‡P < 0.05 when comparing lean and obese PCOS using an unpaired Student’s t-test.



pathophysiological process behind the impact of obesity on

lipid metabolism has been described in the literature, although

the effect of PCOS on lipid metabolism is debatable. Dyslipi-

demia is possibly the most common metabolic abnormality

produced in relation to PCOS. In the present study, lean

women with PCOS exhibited lower HDLc and Apo AI levels

than their respective controls, which could be representative of

an altered lipid profile, although the levels in question can be

considered to be within the range recommended by the ATPIII

[18]. Furthermore, no changes were observed in triglyceride or

LDLc levels in lean women with PCOS when compared with

their respective controls, in accordance with that reported

previously by some authors [14], although not by others [19,20].

Most previous studies about dyslipidemia in PCOS have been

carried out in overweight and obese women with PCOS due to

the direct association between PCOS and obesity. The BMI of

our population could have disguised alterations in these

parameters by PCOS, in which case obesity would be a con-

founding factor. In fact, most lipid parameters were found to

worsen as BMI increased, with differences being detected

between LDLc, non-HDLc, triglycerides and AIP levels of obese

Table 2 Lipoprotein profile in PCOS subjects and controls women according to BMI

BMI < 25 kg/m2 BMI > 30 kg/m2

Controls PCOS Controls PCOS

TC (mg/dL) 170.9 � 27.8 160.7 � 27.6 178.6 � 35.5 198.0 � 31.1‡

LDLc (mg/dL) 99.6 � 20.1 95.5 � 20.8 113.0 � 30.2† 131.5 � 20.9*,‡

HDLc (mg/dL) 60.8 � 10.5 53.5 � 10.6* 45.5 � 11.9† 38.8 � 10.8 ‡

Non-HDLc (mg/dL) 114.4 � 19.7 105.9 � 22.6 133.1 � 34.3† 159.2 � 29.5*,‡

Triglycerides (mg/dL) 57 (46, 73) 51 (44, 62) 85 (61, 156) † 118 (78, 157)*,‡

AIP 0.003 � 0.167 �0.028 � 0.241 0.302 � 0.299† 0.511 � 0.290*,‡

Apo B (mg/dL) 71.9 � 13.8 65.9 � 11.9 87.8 � 24.1† 98.4 � 23.4‡

Apo A-I (mg/dL) 156.7 � 30.6 141.6 � 24.5* 132.1 � 23.4† 128.8 � 18.9‡

Apo B/Apo A-I 0.494 � 0.122 0.471 � 0.104 0.675 � 0.180† 0.789 � 0.257‡

VLDL (%) 10.3 � 3.1 11.2 � 3.2 12.3 � 3.7† 16.5 � 4.8*,‡

Large HDL (%) 40.0 � 6.9 40.8 � 8.0 30.8 � 7.3† 25.1 � 7.6*,‡

Intermediate HDL (%) 46.5 � 3.8 47.4 � 5.4 51.4 � 4.4† 52.0 � 3.7‡

Small HDL (%) 13.5 � 5.4 11.8 � 3.6 17.9 � 6.7† 22.8 � 7.8*,‡

LDL particle size (�A) 272.8 � 1.6 273.5 � 1.6 271.8 � 1.9† 271.5 � 3.6‡

Data are expressed as mean � SD, except for triglycerides which is represented as median and IQ range. Values of serum triglycerides concentrations were

normalised using a log transformation.

BMI, body mass index; PCOS, polycystic ovary syndrome; TC, total cholesterol; LDLc, low-density lipoprotein cholesterol; HDLc, high-density lipoprotein

cholesterol; AIP, atherogenic index of plasma; Apo, apolipoprotein; VLDL, very low-density lipoprotein.

*P < 0.05 when comparing controls and PCOS (lean or obese) using an unpaired Student’s t-test.†P < 0.05 when comparing lean and obese controls using an unpaired Student’s t-test.‡P < 0.05 when comparing lean and obese PCOS using an unpaired Student’s t-test.

97·3 100 95·7

15·0

10·02·7 4·3

75·0

0%

50%

100%

Control PCOS Control PCOSBMI>30Kg/m2BMI<25Kg/m2

Per

cent

age

Pattern AIntermediatePattern B

Figure 1 Percentage of LDL patterns in controls andpolycystic ovary syndrome (PCOS) subjects according to theirbody mass index. *P < 0.05 when comparing percentages ofnon-A pattern between control and PCOS in obese women bymeans of the Chi-square test.



PCOS subjects and those of obese controls. In line with our

findings, Castelo-Branco et al. [21] concluded that obesity – but

not being overweight – is associated with dyslidipemia in

women with PCOS irrespective of Rotterdam phenotypes, as

they observed significantly higher levels of total cholesterol,

LDLc and triglycerides and significantly lower HDLc levels in

the serum of said patients. Moreover, previous studies confirm

that the prevalence and severity of dyslipidemia in PCOS

depend on the diagnostic criteria employed [20]. In fact, Rizzo

and collaborators reported that ovulatory PCOS showed milder

forms of atherogenic dyslipidemia than anovulatory PCOS,

which seemed to be related to the extent of insulin resistance

[11].

Low HDLc levels in women with PCOS are consistently

reported in the literature, although little is known about the

composition of HDL particles with conflicting results. In fact,

we have observed a differential response in HDL subfractions

depending on BMI. Lean women did not exhibit qualitative

changes in HDL subfractions, whereas obesity worsened HDL

parameters (increasing small HDL and decreasing large HDL

subfractions) in both PCOS subjects and controls, with these

parameters proving to be more atherogenic in the former

group. Previous studies in obese women with PCOS have

shown no qualitative changes in HDL subfractions [14],

although some authors have reported a significantly lower

percentage of large HDL and higher levels of small HDL [12],

which is consistent with our results and with the decrease in

HDL size reported by Sidhwani et al. [13].

In addition, despite the fact that triglyceride levels were

similar in both our lean groups, our data suggest that lipo-

protein metabolism is altered in patients with PCOS with

higher BMI. In fact, the present findings demonstrate a small

but significant increase in VLDL percentage in obese PCOS

subjects with respect to obese controls. In the presence of

increased plasma levels of VLDL and normal activity of cho-

lesterol ester transfer protein (CETP), VLDL triglycerides can be

exchanged for LDL- and HDL cholesterol. This exchange pro-

duces LDL particles enriched in triglycerides, which are rapidly

lipolysed by hepatic lipase, resulting in smaller, denser parti-

cles. Triglyceride-rich HDL particles are also smaller and can

undergo further modification, including hydrolysis of their

triglycerides by hepatic lipase and reduced cholesterol efflux

from cells, which, in turn, contributes to lower concentrations

of HDLc [22]. Numerous studies have demonstrated that the

predominance of sdLDL particles (even though they can carry

the same total cholesterol content) correlates with the progres-

sion of atherosclerosis and earlier and more severe cardiovas-

cular disease [23–25]. Our results provide further evidence that

only obese women, whether PCOS or not, are characterised by

an increase in sdLDL. Nevertheless, we have detected a 10.0%

prevalence of atherogenic B LDL and 15.0% of intermediate

pattern among PCOS subjects compared with 4.3% prevalence

of intermediate LDL profile in obese controls, which is in

accordance with previous reports [9,10,14]. In contrast, other

studies carried out in overweight-obese women with PCOS

Table 3 Correlation coefficient between insulin, HOMA-IRindex or FAI and lipid parameters in controls and PCOSsubjects

Controls PCOS

r P r P

Insulin

Non-HDLc 0.266 0.057 0.374 0.008

Triglycerides 0.595 <0.001 0.443 0.001

AIP 0.629 <0.001 0.607 <0.001

VLDL 0.182 0.175 0.505 <0.001

LDL size �0.385 0.003 �0.296 0.037

HDL �0.408 0.002 �0.613 <0.001

Large HDL �0.468 <0.001 �0.499 <0.001

Small HDL 0.338 0.010 0.589 <0.001

HOMA-IR

Non-HDLc 0.295 0.030 0.310 0.025

Triglycerides 0.567 <0.001 0.577 <0.001

AIP 0.629 <0.001 0.562 <0.001

VLDL 0.175 0.184 0.347 0.011

LDL size �0.308 0.018 �0.227 0.103

HDL �0.427 <0.001 �0.560 <0.001

Large HDL �0.493 <0.001 �0.299 0.030

Small HDL 0.374 0.004 0.405 0.003

FAI

Non-HDLc �0.133 0.344 0.345 0.011

Triglycerides �0.214 0.110 0.356 0.008

AIP �0.074 0.583 0.448 <0.001

VLDL 0.037 0.783 0.335 0.013

LDL size 0.030 0.821 �0.191 0.166

HDL �0.061 0.744 �0.384 0.004

Large HDL �0.028 0.833 �0.397 0.003

Small HDL �0.135 0.313 0.429 0.001

Correlation coefficients were estimated by Pearson’s correlation for all

parameters except for triglycerides, in which case Spearman’s correlation

was used.

AIP, atherogenic index of plasma; FAI, free androgenic index; HOMA-IR,

homoeostasis model assessment; PCOS, polycystic ovary syndrome.



have not identified any real quantitative change in LDL

lipoprotein profile [12,13]. These discrepancies may be due to

confounding variables (e.g. age, body size, race or method of

measuring LDL subparticles) that can affect LDL particle size.

Indeed, it should be said that most of the studies in question

did not divide groups according to BMI.

Because hyperandrogenism and lipid metabolism are closely

related, we evaluated the influence of testosterone on dyslipi-

demia. A significant correlation between FAI and most lipid

parameters – except for LDL particle size – was found in

women with PCOS, suggesting that hyperandrogenemia

accounts for the atherogenic lipid profile of these patients. In

fact, we have observed that testosterone has a predominantly

deleterious effect on lipid profile, which confirms findings of

other studies in women with PCOS [15,26]. Testosterone has

been implicated in lowering HDLc levels, an effect attributed

to the up-regulation of two genes involved in the catabolism of

HDL; namely scanveger receptor B1 (SR-B1) and hepatic lipase

[27]. Intriguingly, it has been suggested that the quality of LDL

particles, although subject to genetic factors, also depends on

insulin levels. In a previous study, a negative correlation was

reported between insulin levels and LDL particle size in over-

weight women with PCOS, thus implying a more LDL athero-

genic type in the setting of hyperinsulinemia [9], as confirmed

by the present findings.

To summarise, our study provides evidence that PCOS per se

does not constitute an additional risk factor for developing a

proatherogenic lipid profile, while PCOS associated with

obesity exacerbates the deleterious effect of obesity on athero-

genic dyslipidemia and lipoprotein subfractions. In the light of

this, analysis of lipid profile in these patients should include

lipoprotein subclasses if a more thorough evaluation of

cardiovascular risk is to be performed.

Acknowledgements

This work was supported by grants PI10/1195 and PS09/01025

from FIS and co-funded by the European Regional Develop-

ment Fund of the European Union (FEDER) and ACOMP/

2012/045 and ACOMP/2012/042 from the Regional Ministry of

Education of Valencian Community. M Rocha is a recipient of a

Miguel Servet contract (CP10/00360) from Carlos III Health

Institute. VM V�ıctor is a recipient of Valencian Regional Min-

istry of Health and Carlos III Health Institute contract (CES10/

030). We kindly thank B Normanly and I Soria-Cuenca for their

contribution to the present study.

Author contributions

The authors’ responsibilities were as follows: AH-M conducted

the study. MG, and MB provided supervision of the subjects in

the study. CB, and MR performed the laboratory analyses and

collected data. VMV assisted in the design of the experiment

and provided support throughout the course of the trial and

analysis. CB and MR performed statistical analyses, interpreted

the data and prepared the manuscript. AH-M, CB and MR were

responsible for its final content. All authors read and approved

the final version of the manuscript. None of the authors has any

personal or financial conflict of interest.

Address

Service of Endocrinology, University Hospital Dr. Peset, Avda

Gaspar Aguilar 90, Valencia 46017, Spain (A. Hern�andez-Mij-

ares, C. Ba~nuls, M. G�omez-Balaguer, M. Bergoglio, V. M. V�ıctor,

M. Rocha); Foundation for the Promotion of Healthcare and

Biomedical Research in the Valencian Community (FISABIO),

Avda Gaspar Aguilar 90, Valencia 46017, Spain (A. Hern�andez-

Mijares, V. M. V�ıctor, M. Rocha); Department of Medicine,

Faculty of Medicine, University of Valencia, Avda Blasco

Ib�a~nez 13, Valencia 46010, Spain (A. Hern�andez-Mijares, C.

Ba~nuls); Institute of Health Research INCLIVA, Avda Blasco

Ib�a~nez 17, Valencia 46010, Spain (A. Hern�andez-Mijares, C.

Ba~nuls, V. M. V�ıctor, M. Rocha); CIBER CB06/04/0071

Research Group, CIBER Hepatic and Digestive Diseases, Uni-

versity of Valencia, Avda Blasco Ib�a~nez 13, Valencia 46010,

Spain (V. M. V�ıctor, M. Rocha); Department of Physiology,

Faculty of Medicine, University of Valencia, Avda Blasco

Ib�a~nez 13, Valencia 46010, Spain (V. M. V�ıctor).

Correspondence to: Antonio Hern�andez-Mijares, Service of

Endocrinology, University Hospital Dr. Peset, Av. Gaspar Ag-

uilar 90, 46017 Valencia, Spain. Tel.: (034) 961622492; fax: (034)

961622492; e-mail: [email protected]

Received 23 November 2012; accepted 25 February 2013

References1 Diamanti-Kandarakis E, Kouli CR, Bergiele AT, Filandra FA,Tsianateli TC, Spina GG et al. A survey of the polycystic ovarysyndrome in the Greek island of Lesbos: hormonal and metabolicprofile. J Clin Endocrinol Metab 1999;84:4006–11.

2 Ehrmann DA. Polycystic ovary syndrome. N Engl J Med2005;352:1223–36.

3 Sukalich S, Guzick D. Cardiovascular health in women withpolycystic ovary syndrome. Semin Reprod Med 2003;21:309–15.

4 Toulis KA, Goulis DG, Mintziori G, Kintiraki E, Eukarpidis E,Mouratoglou SA et al. Meta-analysis of cardiovascular disease riskmarkers in women with polycystic ovary syndrome. Hum ReprodUpdate 2011;17:741–60.

5 Cussons AJ, Stuckey BG, Watts GF. Cardiovascular disease in thepolycystic ovary syndrome: new insights and perspectives.Atherosclerosis 2006;185:227–39.

6 Musunuru K. Atherogenic dyslipidemia: cardiovascular risk anddietary intervention. Lipids 2010;45:907–14.

7 Kontush A, Chapman MJ. Antiatherogenic small, dense HDL-guardian angel of the arterial wall? Nat Clin Pract Cardiovasc Med2006;3:144–53.



8 Dejager S, Pichard C, Giral P, Bruckert E, Federspield MC, Beucler Iet al. Smaller LDL particle size in women with polycystic ovarysyndrome compared to controls. Clin Endocrinol (Oxf)2001;54:455–62.

9 Berneis K, Rizzo M, Lazzarini V, Fruzzetti F, Carmina E.Atherogenic lipoprotein phenotype and low-density lipoproteinssize and subclasses in women with polycystic ovary syndrome.J Clin Endocrinol Metab 2007;92:186.

10 Doi SA, Abbas JM, Parkinson L, Chakraborty J, Akanji AO. LDLspecies heterogeneity in the atherogenic dyslipidemia of polycysticovary syndrome. Am J Clin Pathol 2008;129:802–10.

11 Rizzo M, Berneis K, Hersberger M, Pepe I, Di Fede G, Rini GB et al.Milder forms of atherogenic dyslipidemia in ovulatory versusanovulatory polycystic ovary syndrome phenotype. Hum Reprod2009;24:2286–92.

12 Legro RS, Blanche P, Krauss RM, Lobo RA. Alterations in low-density lipoprotein and high-density lipoprotein subclasses amongHispanic women with polycystic ovary syndrome: influence ofinsulin and genetic factors. Fertil Steril 1999;72:990–5.

13 Sidhwani S, Scoccia B, Sunghay S, Stephens-Archer CN, Mazzone T,Sam S. PCOS is associated with atherogenic changes in lipoproteinparticle number and size independent of body weight. ClinEndocrinol (Oxf) 2011, in press, doi: 10.1111/j.1365-2265.2011.04015.x.

14 Phelan N, O’Connor A, Kyaw-Tun T, Correia N, Boran G, RocheHM et al. Lipoprotein subclass patterns in women with polycysticovary syndrome (PCOS) compared with equally insulin-resistantwomen without PCOS. J Clin Endocrinol Metab 2010;95:3933–9.

15 Cupisti S, Giltay EJ, Gooren LJ, Kronawitter D, Oppelt PG,Beckmann MW et al. The impact of testosterone administration tofemale-to-male transsexuals on insulin resistance and lipidparameters compared with women with polycystic ovary syndrome.Fertil Steril 2010;94:2647–53.

16 The Rotterdam ESHRE/ASRM-Sponsored PCOS ConsensusWorkshop Group. Revised 2003 consensus on diagnostic criteria andlong-term health risks related to polycystic ovary syndrome. FertilSteril 2004;81:19–25.

17 Hoefner DM, Hodel SD, O’Brien JF, Branum EL, Sun D, Meissner Iet al. Development of a rapid, quantitative method for LDL

subfractionation with use of the Quantimetrix Lipoprint LDLSystem. Clin Chem 2001;47:266–74.

18 Expert Panel on Detection, Evaluation, and Treatment of High BloodCholesterol in Adults. Executive Summary of The Third Report ofThe National Cholesterol Education Program (NCEP) Expert Panelon Detection, Evaluation, And Treatment of High Blood CholesterolIn Adults (Adult Treatment Panel III). JAMA 2001;285:48–97.

19 Valkenburg O, Steegers-Theunissen RP, Smedts HP, Dallinga-ThieGM, Fauser BC, Westerveld EH et al. A more atherogenic serumlipoprotein profile is present in women with polycystic ovarysyndrome: a case-control study. J Clin Endocrinol Metab2008;93:470–6.

20 Wild RA, Rizzo M, Clifton S, Carmina E. Lipid levels in polycysticovary syndrome: systematic review and meta-analysis. Fertil Steril2011;95:1073–9.e1–11.

21 Castelo-Branco C, Steinvarcel F, Osorio A, Ros C, Balasch J.Atherogenic metabolic profile in PCOS patients: role of obesity andhyperandrogenism. Gynecol Endocrinol 2010;26:736–42.

22 Packard CJ. Triacylglycerol-rich lipoproteins and the generation ofsmall, dense low-density lipoprotein. Biochem Soc Trans2003;31:1066–9.

23 Koba S, Yokota Y, Hirano T, Ito Y, Ban Y, Tsunoda F et al. SmallLDL-cholesterol is superior to LDL-cholesterol for determiningsevere coronary atherosclerosis. J Atheroscler Thromb 2008;15:250–60.

24 Moon JY, Kwon HM, Kwon SW, Yoon SJ, Kim JS, Lee SJ et al.Lipoprotein(a) and LDL particle size are related to the severity ofcoronary artery disease. Cardiology 2007;108:282–9.

25 Packard CJ. Small dense low-density lipoprotein and its role as anindependent predictor of cardiovascular disease. Curr Opin Lipidol2006;17:412–7.

26 Fruzzetti F, Perini D, Lazzarini V, Parrini D, Genazzani AR.Adolescent girls with polycystic ovary syndrome showingdifferent phenotypes have a different metabolic profileassociated with increasing androgen levels. Fertil Steril2009;92:626–34.

27 Diamanti-Kandarakis E, Papavassiliou AG, Kandarakis SA,Chrousos GP. Pathophysiology and types of dyslipidemia in PCOS.Trends Endocrinol Metab 2007;18:280–5.



Documents

Influence of obesity on atherogenic dyslipidemia in women with polycystic ovary syndrome