Multiparity and Risk of Metabolic Syndrome: Isfahan Healthy Heart Program

Abstract

Introduction:

More than one third of Iranian adult women have the metabolic syndrome. We conducted this study to assess the effect of parity on the prevalence of metabolic syndrome in analyses controlling for sociodemographic and reproductive variables as well as behavioral risk factors.

Methods:

We evaluated the relationship between number of children and metabolic syndrome in 6331 adult nonpregnant women >20 years of age. The data source for this study was Isfahan Healthy Heart Program (IHHP). Metabolic syndrome was defined according to Adult Treatment Panel III (ATP III).

Results:

Overall, 34.2% of women met the criteria for metabolic syndrome. The number of children borne in women with metabolic syndrome was significantly higher than others (5.2 ± 3.1 vs. 3.5 ± 2.6; p < 0.0001). In logistic regression analyses, the odds of metabolic syndrome increased 24% (95% confidence interval [CI], 22–26%) with each additional child, but after adjustment for sociodemographic, reproductive, and behavioral characteristics, the odds of metabolic syndrome was attenuated (odds ratio [OR], 1.03; 95% CI, 1.00–1.06). Further adjustment for body mass index (BMI) yielded similar results (OR, 1.02; 95% CI, 0.98–1.05).

Conclusion:

A combination of lifestyle risk factors and/or biological changes associated with childbearing may explain the positive association between parity and increased risk of metabolic syndrome.

Introduction

More than one third of Iranian adult women have the metabolic syndrome.1 ^, 2 This represents an emerging public health problem, as shown by trends in cardiovascular mortality in Iran.3 Understanding the role of reproductive history in the development of metabolic syndrome will help to alleviate metabolic syndrome-associated morbidity and mortality.

The role of multiparity in the development and pathogenesis of diabetes mellitus, obesity, and dislipidemia is controversial.4 ⁻ 9 These conflicting results may arise from unadjusted effects for age and socioeconomic status that influence childbearing patterns as well as disease risk. To our knowledge, no prior study in Eastern countries has evaluated the relationship between childbearing and metabolic syndrome as a combination of predefined cardiovascular risk factors. Considering the effect of genetic and lifestyle factors on the metabolic syndrome, we conducted this study to assess the effect of parity on the prevalence of metabolic syndrome in analyses controlling for sociodemographic, reproductive, and behavioral characteristics.

Materials and Methods

The data source for this study was the Isfahan Healthy Heart Program (IHHP), which was developed as a community-based interventional program with ongoing evaluation in three cities in Iran between 2001 and 2005. This study aimed to reduce the modifiable cardiovascular risk factors in the general population through 10 distinct interventional projects targeting diet, physical activity, and smoking status. The details of study protocol have been described earlier.10

In a two-stage sampling method, census blocks were randomly selected from each county and divided into clusters, each having approximately 1000 households. Then, 5–10 households within these clusters were randomly selected for enumeration. After enumeration, a total of 6331 adult nonpregnant women >20 years of age were randomly selected for the survey. Those with known metabolic or thyroid disease and breast-feeding mothers were not included. Written informed consent was obtained from all participants before inclusion.

Parity was defined as the number of live births reported. Current use of oral contraceptive pills (OCP) was defined based on reported use of “birth control pills for any reason.” Current use of postmenopausal hormone therapy (HT) was defined based on reported use of oral or transdermal female hormones. The duration of OCP administration and history of abortion were also recorded.

Data for reported mean family income were collected and analyzed in quartiles. Physical activity was evaluated using the Baecke Questionnaire of habitual physical activity.11 Current employment was defined as having or working at a job or business for at least 4 months during the last year including the last month. Smoking status was determined through the self-administered questionnaire according to World Health Organization (WHO) guidelines¹² and categorized as current, past, or never.

Components of the metabolic syndrome

Components of metabolic syndrome were evaluated based on the Adult Treatment Panel III (ATP III)-defined cutoff points for women (systolic blood pressure [SBP] 130 mmHg, diastolic blood pressure [DBP] 85 mmHg, triglyceride 150 mg/dL, fasting blood sugar [FBS] 110 mg/dL, high-density lipoprotein [HDL] 50 mg/dL, waist circumference 88 cm).13 Fasting levels of plasma glucose and HDL were assessed with the enzymatic–colorimetric method (ParsAzmoun, Tehran, Iran). All laboratory measurements were performed in the laboratory of Isfahan Cardiovascular Research Center with adherence to external quality control. Seated SBP and DBP were measured on the right arm after 5 minutes of rest. The average of two measurements was used for analysis. Waist circumference was measured at the midpoint between the lowest rib and the iliac crest. Participants with three or more of these five possible components were considered to have metabolic syndrome.

Statistical analysis

Data were presented as mean (standard deviation [SD]), percentage or odds ratio (OR; 95% confidence interval [CI]), as appropriate. Educational level was classified as illiterate, 1–5 years, 6–12 years, and >12 years of education. Body mass index (BMI) was categorized as <18.5 kg/m², 18.5–24.9 kg/m², 25–29.9 kg/m², 30–34.9 kg/m², or >35 kg/m² for further analysis. Number of children, sociodemographic, and reproductive and behavioral characteristics of participants were compared between those with and without metabolic syndrome using a two-sample independent t-test or chi-squared test, as appropriate. P values <0.05 were considered to be significant.

The relationship between metabolic syndrome and parity was assessed by logistic regression analysis adjusted for sociodemographic, reproductive, and behavioral covariates. Further adjustment for BMI was performed in a separate model. The relationship of metabolic syndrome with those variables that were significantly associated with metabolic syndrome in univariate analyses was assessed in additional logistic regression models. Because the prevalence of metabolic syndrome was different between urban and rural populations, the analyses have been provided in models stratified by residence. Statistical analysis was performed using Statistical Package for Social Sciences software version 10.0 (SPSS Inc., Chicago, IL).

Results

Among 6326 participants, 34.2% of women met the criteria for metabolic syndrome. Rates for the individual components of metabolic syndrome are shown in Table 1. The number of live births per woman ranged from 0 to 14, with a weighted mean of 4.1 live births. As shown in Table 1, the number of childbearing in women with metabolic syndrome was significantly higher than those without metabolic syndrome (5.2 ± 3.1 vs. 3.5 ± 2.6, P < 0.0001). The prevalence of metabolic syndrome in nulliparous women, mothers with 1–4, and >4 children was 15.2%, 26.2%, and 49.6%, respectively.

Table 1.

Sociodemographic, Behavioral, and Reproductive Characteristics in Participating Women

Variables	Participants with metabolic syndrome (n = 2163)	Participants without metabolic syndrome(n = 4163)	P value
Sociodemographic and behavioral variables
Age (years)	46.8 (14.4)	34.7 (12.8)	<0.0001
Education			<0.0001
Illiterate	39.2	19.9
1–5 years	35.8	32.8
6–12 years	22.2	41
>12 years	2.8	6.3
Residence			<0.0001
Rural	23.7	29.0
Urban	76.3	71.0
Family income			<0.0001
<Quartile 1 of population	30.8	26.1
>Quartile 3 of population	9.6	12.7
Quartile 1 < income < Quartile 3	59.6	61.2
Currently employed	4.2	6.3	<0.0001
Smoking			<0.0001
Current	2.3	1.2
Past	1.3	0.8
Never	96.4	98
Physically inactive	82.1	75.7	<0.0001
Body mass index (kg/m²)	29.4 (5.2)	25.2 (4.9)	<0.0001
≤18.5	0.6	4.6	<0.0001
18.6–24.9	16.7	48.2
25–29.9	41.7	32.4
30–34.9	30.2	11.9
≥35	10.8	2.9
Reproductive variables
Parity (number of live births)	5.2 (3.1)	3.5 (2.6)	<0.0001
Age at first live birth (years)	17.8 (3.2)	18.7 (3.2)	<0.0001
Current use of oral contraceptives	44.5	41.7	0.034
On oral contraceptive (year)	5.2 (4.8)	3.8 (3.6)	<0.0001
Abortion	0.74 (1.25)	0.45 (0.93)	<0.0001
On postmenopausal hormonal therapy	6.0	5.9	0.94
Components of metabolic syndrome
Impaired fasting glucosea	19.7	1.6	<0.0001
Elevated blood pressureb	59.5	10.5	<0.0001
Fasting high-density lipoprotein <50 mg/dL	79.3	48.3	<0.0001
Fasting triglyceride level >150 mg/dL	87.6	21.6	<0.0001
Waist circumference >88 cm	94.4	45.8	<0.0001

Data are presented as percentiles or mean (standard deviation).

^aFasting glucose ≥110 mg/dL or taking diabetes medication.

^bBlood pressure ≥130/85 mmHg or taking antihypertensive medication.

Participants with metabolic syndrome were significantly older with lower educational level, lower degrees of family income, and lower rate of employment. Residency in urban areas was associated with the development of metabolic syndrome. Expectedly, behavioral risk factors, including smoking and inactivity as well as BMI, were also associated with the development of metabolic syndrome. In terms of reproductive variables, current use of OCP, the duration of OCP administration, history of abortion, and lower age at first live birth were positively associated with the development of metabolic syndrome. Current use of postmenopausal hormonal therapy was not different in clients with and without metabolic syndrome (Table 1).

In logistic regression analyses, the odds of metabolic syndrome increased 24% (95% CI, 22–26%) with each additional child, but after adjustment for sociodemographic, reproductive, and behavioral characteristics (model B), the odds of metabolic syndrome were attenuated (OR, 1.03; 95% CI, 1.00–1.06). Because BMI is significantly associated with the components of metabolic syndrome, especially waist circumference, to prevent multi-co-linearity, its effect on the OR of parity was examined in a separate model including all above-mentioned covariates (model C). Further adjustment for BMI yielded similar results (OR, 1.02; 95% CI, 0.98–1.05).

Number of children was a significant predictor of increased waist circumference, impaired fasting glucose, high blood pressure, and triglyceride level (model A), but adjustment for the covariates (models B and C) attenuated the predictive value of parity for the development of each component of metabolic syndrome (Table 2).

Table 2.

Odds Ratio of Parity for Metabolic Syndrome and Components of Metabolic Syndrome

Model outcome	Model A	Model B	Model C
Metabolic syndrome	1.24 (1.22–1.26)	1.03 (1.00, 1.06)	1.02 (0.98, 1.05)
Fasting glucose ≥110	1.27 (1.24–1.31)	1.08 (1.03, 1.12)	1.04 (1.00, 1.08)
Blood pressure ≥130/85 mmHg	1.29 (1.26–1.32)	1.00 (0.97, 1.03)	1.00 (0.97, 1.03)
Fasting high-density lipoprotein <50 mg/dL	0.99 (0.98–1.01)	1.03 (1.00, 1.05)	1.02 (0.99, 1.04)
Fasting triglyceride level ≥l150 mg/dL	1.19 (1.17–1.22)	1.04 (1.01, 1.07)	1.03 (1.00, 1.05)
Waist circumference >88 cm	1.15 (1.13–1.17)	1.01 (0.98, 1.04)	1.00 (0.98, 1.02)

Model A, not adjusted; model B, adjusted for sociodemographic, reproductive, and behavioral characteristics; model C, adjusted for covariates of model B plus body mass index.

Data are presented as odds ratio (95% confidence interval).

Because the prevalence of metabolic syndrome was different between urban and rural populations, we performed the analyses in models stratified by residence (Table 3). After adjustment for multiple covariates, parity was not a significant predictor of metabolic syndrome either in rural or in urban population. Age, low educational level, low family income, and current use of oral contraceptive were significantly associated with an increased risk of metabolic syndrome. However, the OR of low income, physical inactivity, and OCP administration for the development of metabolic syndrome were significant only in urban population.

Table 3.

Relationship of Parity and Characteristics of Women with Metabolic Syndrome Stratified by Residence

Independent variable	Total OR (95% CI) n = 6331	Urban OR (95% CI)n = 4610	rural OR (95% CI)n = 1721
Paritya	1.02 (0.98–1.05)	1.01 (0.96,1.07)	1.07 (0.93,1.22)
Age, years	1.06 (1.05–1.07)	1.06 (1.05,1.07)	1.05 (1.04,1.06)
Education
Illiterate	1.15 (0.82–1.65)	1.35 (0.91,2.01)	2.68 (0.54,13.19)
1–5 years	1.51(1.08–2.11)	1.59 (1.12,2.27)	3.30 (0.68,15.87)
6–12 years	1.24 (0.89–1.72)	1.18 (0.84,1.67)	3.09 (0.64,14.87)
>12 years	1.00 (referent)	1.00 (referent)	1.00 (referent)
Family income
<Quartile 1 of population	1.21 (1.07–1.39)	1.29 (1.11,1.48)	1.03 (0.81,1.26)
>Quartile 3 of population	0.95 (0.78–1.16)	0.89 (0.72,1.09)	1.11 (0.76,1.59)
Quartile 1<income< quartile3	1.00 (referent)	1.00 (referent)	1.00 (referent)
Currently employed	1.02 (0.78–1.35)	1.04 (0.76,1.44)	0.95 (0.56,1.63)
Smoking
Current	1.41 (0.90–2.20)	1.24 (0.78,1.97)	2.33 (0.34,15.81)
Past	0.81 (0.45–1.46)	0.76 (0.40,1.43)	0.99 (0.17,5.76)
Never	1.00 (referent)	1.00 (referent)	1.00 (referent)
Physically inactive	1.05 (0.90–1.22)	1.17 (1.01,1.38)	0.72 (0.49,1.04)
Current use of oral contraceptives	1.22 (1.09–1.38)	1.32 (1.15,1.51)	1.01 (0.79,1.27)

^aAdjusted for sociodemographic, reproductive, and behavioral characteristics as well as body mass index.

Abbreviations: CI, confidence interval; OR, odds ratio.

Discussion

The key finding in this study is the association between the increasing number of children and risk of metabolic syndrome among women. The attenuation of risk after adjustment for multiple variables suggests that a combination of lifestyle risk factors and/or biological changes associated with childbearing may mediate the effects of parity on the development of metabolic syndrome.

Some prior studies have suggested a J-shaped association between number of children and coronary heart disease risk, especially diabetes mellitus, with an increased risk among nulliparous women and those with more than 4 children.14 ^, 15 A potential explanation for the proposed increased risk of cardiovascular risk in nulliparous women is the role of insulin resistance and β-cell dysfunction in conditions with hormonal imbalances, such as the presence of polycystic ovary syndrome.5 ^, 16 ^, 17 However, in our study, the lowest prevalence of metabolic syndrome was observed in nulliparous women, which is in line with the recent large-scale study of Cohen and his colleagues.18

In earlier studies, grand multiparity (5 or more births) was associated with increased risk for cardiovascular diseases.14 ^, 15 The cumulative adverse effects of parity on metabolic disorders may be explained in part by physiologic changes during normal pregnancy such as hormonal alterations secondary to fewer ovulatory cycles,19 insulin resistance,20 ^, 21 and increased glucocorticoid activity.22 These physiologic changes may influence several features of metabolic syndrome, including diabetes mellitus,21 dyslipidemia,8 ^, 21 and obesity6 ^, 7 ^, 22 that could result in increased cardiovascular risk. Moreover, lifestyle changes such as smoking cessation, which frequently occurs once women learn of their pregnancy,23 more sedentary life style,24 ⁻ 26 and socioeconomic factors26 ^, 27 result in weight gain during pregnancy, which is unlikely to be lost after childbearing.9

The potential mechanisms for the observed association between number of children and risk of metabolic syndrome include a complex combination of biological and lifestyle consequences of pregnancy that may not be truly differentiated. Because of reduced exposure to estrogen and relative insulin resistance during pregnancy, most investigators have focused on biological pathways for this association.28 ⁻ 30 However, in favor of the “lifestyle consequences” assumption, some prior studies have shown that with an increasing number of children the cardiovascular risk increases similarly in both sexes.14 ^, 31 The issue needs further investigation in large-scale controlled studies.

In the study of Cohen and colleagues,18 which is the only published survey with comparable objectives, after adjustment for BMI, the odds of parity for the risk of metabolic syndrome decreased considerably, suggesting the important mediating role of weight gain in the effects of parity on the development of metabolic syndrome. In our study, significant ORs of the components of the metabolic syndrome have been attenuated after adjustment for sociodemographic, reproductive, and behavioral characteristics of women, and further adjustment for BMI did not make a considerable variation. These findings may reasonably explain the association between parity and metabolic syndrome with the mediating role of sociodemographic, reproductive, and behavioral characteristics of women. We believe that the observed role of weight or weight changes during pregnancy may be explained well enough by other covariates.

In support of previous studies,15 ^, 18 our findings in separate logistic regression analyses demonstrate that age, low educational level, and low family income are significantly associated with an increased risk for the components metabolic syndrome. Current use of OCP especially in urban populations was also a potent predictor of metabolic syndrome (OR, 1.32; 95% CI, 1.15–1.51), which is in contrast with the data of Cohen and colleagues,18 who found the use of OCP protective against the development of metabolic syndrome (OR, 0.30; 95% CI, 0.4–0.64). Potential mechanism for the adverse effects of OCP may be the association of its use with higher low-density lipoprotein cholesterol (LDL-C)32 and lower HDL-C33 levels. Other behavioral characteristics, including physical activity and smoking, were also associated with the development of metabolic syndrome, which is in line with most previous surveys.34 ⁻ 36 Although the prevalence of metabolic syndrome was different between urban and rural populations, the stratified analyses show that habitation is not a significant mediator of the relationship between childbearing and metabolic syndrome risk.

Study limitations

This study includes a large representative sample of Iranian adult women; however, the cross-sectional nature of the study limits the identification of a causality effect rather than a simple accompaniment of evaluated factors. One limitation of this study is that we did not evaluate the lactation history in our clients, which has been reported to attenuate the development of metabolic syndrome.¹⁸ Other factors to be included in future studies might be the history of gestational diabetes, infertility due to hormonal disorders, such as polycystic ovary syndrome, and age of children.

In conclusion, this study demonstrates the association between the increasing number of children borne and the risk of metabolic syndrome among women, with the mediating role of a combination of pregnancy-related behavioral risk factors and/or biological changes. Conducting further studies to better identify the independent behavioral and biological predictors of metabolic syndrome would help in directing the public health policy toward establishment of effective risk reduction strategies.

Footnotes

Acknowledgments

The study has been approved by Isfahan Cardiovascular Research Center Research Council by grant number 81126 and was funded by grant no. 31309304 from the Iranian Budget and Programming Organization. WHO has designated this project as a model in the region; it is indexed as code No. 86 in the Canadian Heart Health Promotion Projects (http://www.med.mun.ca/g8HeartHealth/page/list_projects.htm/). The authors would like to thank the large team who conducted Isfahan Healthy Heart Program.

References

Sarrafzadegan

, Kelishadi

, Baghaei

et al. Metabolic syndrome: An emerging public health problem in Iranian Women: Isfahan Healthy Heart Program. Int J Cardiol, 2008; 131:90–6.

Fakhrzadeh

, Ebrahimpour

, Pourebrahim

, Heshmat

, Larijani

. Metabolic syndrome and its associated risk factors in healthy adults: A population-based study in Iran. Metab Syndr Relat Disord, 2006; 4:28–34.

Sarraf-Zadegan

, Boshtam

, Malekafzali

, Bashardoost

, Tabatabaei

, Rafiei

. Secular trends in cardiovascular mortality in Iran, with special reference to Isfahan. Acta Cardiol, 1999; 54:327–333.

Pyke

. Parity and the incidence of diabetes. Lancet, 1956; 270:818–820.

Hanley

, McKeown-Eyssen

, Harris

, Hegele

, Wolever

, Kwan

, Zinman

. Association of parity with risk of type 2 diabetes and related metabolic disorders. Diabetes Care, 2002; 25:690–695.

Gunderson

, Murtaugh

, Lewis

, Quesenberry

, West

, Sidney

. Excess gains in weight and waist circumference associated with childbearing: the Coronary Artery Risk Development in Young Adults Study (CARDIA). Int J Obes Relat Metab Disord, 2004; 28:525–535.

Weng

, Bastian

, Taylor

DH Jr

, Moser

, Ostbye

. Number of children associated with obesity in middle-aged women and men: Results from the health and retirement study. J Womens Health, 2004; 13:85.

Lewis

, Funkhouser

, Raczynski

, Sidney

, Bild

, Howard

. Adverse effect of pregnancy on high-density lipoprotein (HDL) cholesterol in young adult women. The CARDIA Study. . Coronary Artery Risk Development in Young Adults. Am J Epidemiol, 1996; 144:247.

Troisi

, Wolf

, Mason

, Klingler

, Colditz

. Relation of body fat distribution to reproductive factors in pre- and postmenopausal women. Obes Res, 1995; 3:143.

10.

Sarraf-Zadegan

, Sadri

, Malekafzali

, Baghaei

et al. Isfahan Healthy Heart Program: a comprehensive integrated community-based program for cardiovascular disease prevention and control. Design, methods and initial experience. Acta Cardiol, 2003; 58:309–320.

11.

Pereira

, Fitzgerald

, Gregg

. A collection of physical activity questionnaires for health-related research. Med Sci Sports Exerc, 1997; 29:517–518.

12.

World Health Organisation . Guidelines for controlling and monitoring the tobacco epidemic. Geneva: World Health Organisation, 1998.

13.

Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III), final report. Circulation, 2002; 106:3143.

14.

Lawlor

, Emberson

, Ebrahim

et al. Is the association between parity and coronary heart disease due to biological effects of pregnancy or adverse lifestyle risk factors associated with child-rearing? Findings from the British Women’s Heart and Health Study and the British Regional Heart Study. Circulation, 2003; 107:1260.

15.

Nicholson

, Asao

, Brancati

, Coresh

, Pankow

, Powe

. Parity and risk of type 2 diabetes: the Atherosclerosis Risk in Communities Study. Diabetes Care, 2006; 29:2349–2354.

16.

Charles

, Pettitt

, McCance

, Hanson

, Bennett

, Knowler

. Gravidity, obesity, and non-insulin dependent diabetes among Pima Indian women. Am J Med, 1994; 97:250–255.

17.

Van

der Spuy ZM DS

. The pathogenesis of infertility and early pregnancy loss in polycystic ovary syndrome. Best Pract Res Clin Obstet Gynaecol, 2004; 18:755–771.

18.

Cohen

, Pieper

, Brown

, Bastian

. Number of children and risk of metabolic syndrome in women. J Womens Health, 2006; 15:763–773.

19.

Rossner

. Pregnancy, weight cycling and weight gain in obesity. Int J Obes Relat Metab Disord, 1992; 16:145–147.

20.

Godsland

. The influence of female sex steroids on glucose metabolism and insulin action. J Intern Med, 1996; 738,(Suppl):1.

21.

Kritz-Silverstein

, Barrett-Connor

, Wingard

. The effect of parity on the later development of noninsulin-dependent diabetes mellitus or impaired glucose tolerance. N Engl J Med, 1989; 321:1214.

22.

Harris

, Ellison

, Richter

, de Wet

, Levin

. Are overweight women at increased risk of obesity following pregnancy?. Br J Nutr, 1998; 79:489.

23.

Rossner

, Ohlin

. Pregnancy as a risk factor for obesity: Lessons from the Stockholm Pregnancy and Weight Development Study. Obes Res, 1995; 3,(Suppl 2):267s–275s.

24.

Ohlin

, Rossner

. Trends in eating patterns, physical activity and socio-demographic factors in relation to postpartum body weight development. Br J Nutr, 1994; 71:457–470.

25.

Walker

. Predictors of weight gain at 6 and 18 months after childbirth: A pilot study. J Obstet Gynecol Neonatal Nurs, 1996; 25:39–48.

26.

Harris

, Ellison

, Clement

. Do the psychosocial and behavioral changes that accompany motherhood influence the impact of pregnancy on long-term weight gain?. J Psychosom Obstet Gynaecol, 1999; 20:65–79.

27.

Potter

, Hannum

, McFarlin

, Essex-Sorlie

, Campbell

, Trupin

. Does infant feeding method influence maternal postpartum weight loss. J Am Diet Assoc, 1991; 91:441–446.

28.

Ness

, Schotland

, Flegal

et al. Reproductive history and coronary heart disease risk in women. Epidemiol Rev, 1994; 16:298–314.

29.

Ness

, Harris

, Cobb

et al. Number of pregnancies and the subsequent risk of cardiovascular disease. N Engl J Med, 1993; 328:1528–1533.

30.

Dekker

, Schouten

. Number of pregnancies and risk of cardiovascular disease. N Engl J Med, 1993; 329:1893–1894.

31.

Hardy

, Lawlor

, Black

, Wadsworth

MEJ

, Kuh

. Number of children and coronary heart disease risk factors in men and women from a British birth cohort. BJOG, 2007; 114:721–730.

32.

Hubert

, Eaker

, Garrison

et al. Life-style correlates of risk factor change in young adults: an eight-year study of coronary heart disease risk factors in the Framingham offspring. Am J Epidemiol, 1987; 125:812–831.

33.

Greenlund

, Webber

, Srinivasan

et al. Associations of oral contraceptive use with serum lipids and lipoproteins in young women: the Bogalusa Heart Study. Ann Epidemiol, 1997; 7:561–567.

34.

Bianchi

, Rossi

, Muscari

, Magalotti

, Zoli

, Pianoro Study Group. Physical activity is negatively associated with the metabolic syndrome in the elderly. QJM, 2008; 101::713–721.

35.

Brandão

, Brandão

, de Magalhães

, Pozzan

. Management of metabolic syndrome in young population. Am J Ther, 2008; 15:356–361.

36.

Buckland

, Salas-Salvadó

, Roure

, Bulló

, Serra-Majem

. Sociodemographic risk factors associated with metabolic syndrome in a Mediterranean population. Public Health Nutr, 2008; 15:1–7.