Apolipoproteins A-I and B As Components of Metabolic Syndrome with Respect to Diabetes Status: A Factor Analysis

Abstract

Background:

The aim of the current study was to elucidate the clustering pattern of metabolic syndrome components along with apolipoproteins (Apo) A-I and B in diabetic and nondiabetic subjects.

Methods:

Factor analysis of conventional variables of metabolic syndrome [i.e., waist circumference, homeostasis model assessment of insulin resistance (HOMA-IR), triglycerides (TG), high-density lipoprotein-cholesterol (HDL-C), and systolic blood pressure (SBP)] with or without addition of Apo A-I and B was performed on 567 and 327 diabetic and nondiabetic subjects, respectively. Thereafter, analyses were repeated after substitution of TG and HDL-C by the TG-to-HDL-C ratio (TG/HDL-C).

Results:

Regarding conventional variables of metabolic syndrome, one or two underlying factors were identified, depending on whether lipid measures were entered as two distinct variables or as a composite measure. Apolipoproteins were consistent with a one-factor structure model of metabolic syndrome and did not change the loading pattern remarkably in nondiabetics. TG and HDL-C tended to cluster with Apo B and A-I, respectively, in different models.

Conclusion:

The current study confirms that addition of Apo A-I and B is consistent with the one-factor model of metabolic syndrome and does not modify the loading pattern remarkably in nondiabetic subjects.

Introduction

M etabolic syndrome clusters a range of cardiovascular and diabetes mellitus risk factors, including central obesity, insulin resistance, dyslipidemia, and hypertension.¹ Evaluation of apolipoproteins (Apo) A-I and B would help clinicians to better understand the atherogenic dyslipidemia that is a common feature of metabolic syndrome.² Apo A-I is the main structural protein of high-density lipoprotein cholesterol (HDL-C), which has atheroprotective effects.³ On the contrary, the structural subunit of atherogenic lipoproteins [i.e., very-low-density (VLDL-C), intermediate- (IDL-C), and low-density lipoprotein (LDL-C) cholesterols) is comprised of Apo B.^4,5 Because there is one Apo B per each of those aforementioned lipid particles, the total Apo B level reflects the atherogenic potential of lipoproteins.⁵

Due to strong intercorrelations that exist between metabolic syndrome components, it is difficult to establish independent associations with the help of multivariate statistical models.^6
–8 To circumvent this problem, some studies have used so-called factor analysis models, which analyze the interrelatedness of measured variables in terms of a smaller group of latent factors. In factor analysis, each extracted factor is interpreted as the manifestation of a distinct underlying pathophysiologic process.^8,9 Prior studies performing factor analysis on abnormalities related to metabolic syndrome have identified from one to seven factors.^8

–12 The discrepancies between outcomes of various studies could be a result of the heterogeneity in the number and nature of the variables chosen for analysis. It has been suggested that using two or more closely interrelated measures (e.g., triglycerides [TG] and HDL-C) interferes with factor analysis and leads to identification of higher numbers of factors than expected.⁹

Factor analysis of abnormalities related to metabolic syndrome is yet to be explored more, especially in diabetic subjects. In this respect, this study benefits from a stratified sample by diabetes status and for the first time in a Middle Eastern population evaluates clustering of metabolic syndrome components along with Apo A-I and B.

Methods

A total of 894 individuals (aged 30–70 years), who were consecutively visited at an outpatient clinic of Vali-Asr Hospital (a Tehran University–affiliated medical center) from June, 2008, to August, 2011, were enrolled in the study. Subjects with hepatic, renal, thyroid, or adrenal problems, along with those who were taking insulin or lipid-modifying agents, were not included in the study. Diabetic participants were under treatment with metformin, glibenclamide, or both simultaneously for controlling their hyperglycemia. All participants gave oral informed consent before study commencement. The study was conducted in accordance with the Helsinki declaration and was performed in line with the recommendations of local ethics review committee of Tehran University of Medical Sciences.

Anthropometric measures, including weight and height, were determined with the subjects in in light clothing and without shoes. Waist circumference was measured at the end of a normal expiration, at mid-distance between iliac crest and rib cage, and was rounded to the nearest 0.1 cm. The participants were asked to rest for at least 5 min before having their blood pressure checked two times with at least a 5-min interval. The average of these two measurements was used for the analyses. Venous blood samples were drawn after a 12-h overnight fast. Fasting plasma glucose was measured by the glucose oxidase test [intra- and interassay coefficients of variation (CV) less than 2.1 and 2.6, respectively]. TG and HDL-C were assayed by enzymatic techniques (Parsazmun, Karaj, Iran), and Apo A-I and B by the immunoturbidimetric method (Roche, Basel, Switzerland). The intra- and interassay CV were lower than 1.0 and 2.4 for Apo A-I and lower than 1.2 and 3.2 for Apo B, respectively. Insulin was determined by radioimmunoassay using an antibody with no cross-reactivity for pro-insulin and C-peptide (Immunotech, Prague, Czech Republic). The intra- and interassay CV were lower than 4.3 and 3.4, respectively. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as fasting insulin (U/L)×fasting plasma glucose (mg/dL)/405, as described by Matthews et al.¹³

Metabolic syndrome was defined according to either Adult Treatment Panel III (ATP III) or modified International Diabetes Federation (IDF) declarations. According to ATP III criteria, a person with metabolic syndrome must have three or more of the following conditions: Presence of the abdominal obesity (waist circumference ≥102 cm and ≥88 cm in men and women, respectively), elevated blood pressure [systolic blood pressure (SBP) ≥130 mmHg and/or diastolic blood pressure (DBP) ≥85 mmHg], low HDL-C (<40 mg/dL and <50 mg/dL in men and women, respectively), TG ≥150 mg/dL, and fasting plasma glucose ≥100 mg/dL (or diabetes).^1,14 According to the modified IDF criteria for use in an Iranian population, a person with metabolic syndrome must have abdominal obesity (waist circumference ≥90 cm in both men and women) plus any two or more of the following conditions: Elevated blood pressure (see above) or treatment of previously diagnosed hypertension, low HDL-C (see above) or on HDL-C therapy, TG ≥150 mg/dL or on TG therapy, and fasting plasma glucose ≥100 mg/dL (or diabetes).¹⁵

Data of diabetic and nondiabetic subjects were analyzed separately using SPSS software (version 16.0; SPSS Inc., Chicago, IL). Natural log transformations were used to improve the normality of skewed variables [i.e., TG, TG-to-HDL-C ratio (TG/HDL-C), and HOMA-IR] in the subsequent analyses, as appropriate. Principal characteristics of diabetic and nondiabetic subjects were compared by the Student t-test. Bivariate Pearson correlation coefficients between individual variables were determined. Exploratory factor analysis was performed using the principal component method of factor extraction, a technique that reduces original variables into fewer latent factors. Only factors with an eigenvalue (the amount of variance attributable to the factor) of greater than 1 were retained and transformed by the varimax rotation method to facilitate interpretation. Varimax rotation cannot be performed if factor analysis extracts just one factor. In line with many prior studies,^{10
–12,16
–18} variables with factor loadings of ≥|0.40| were considered to be significant constituents of that factor, because this ensures that the variable shares at least 15% of the variance with the factor.¹⁹ Factor scores, as the estimates of individual factors with the mean and standard deviation equal to 0 and 1, respectively, were determined by a regression method. Factor analysis was performed in two different models. In model 1, waist circumference, HOMA-IR, TG, HDL-C, and SBP were entered into factor analysis because two of these variables are related to lipid profile (i.e., TG and HDL-C). In model 2 we substituted the TG/HDL-C ratio for both TG and HDL-C. Later, factor analyses were repeated after addition of Apo A-I and B to the variables of interest in both models. Finally, using logistic regression analysis, the predictive abilities of extracted factor scores for ATP III- and IDF-defined metabolic syndrome were assessed.

Results

Principal characteristics of the study population are shown in Table 1. The proportions of subjects with diabetes mellitus and those without diabetes mellitus were 63.4% and 36.6%, respectively. Generally, all variables were significantly worse in diabetic subjects than in nondiabetics. The correlation matrix between the variables of interest is shown in Table 2. The correlation of waist circumference with all other variables reached significant levels in nondiabetic subjects. In both diabetic and nondiabetic subjects, the highest correlation coefficients were observed between Apo A-I and HDL-C.

Table 1.

Principal Characteristics of the Study Population

Variables	Nondiabetics (n=327)	Diabetics (n=567)
Age (years)	45.7±11.8	55.3±9.5^b
BMI (kg/m²)	28.5±5.0	29.7±4.5^a
Waist circumference (cm)	91.7±11.6	99.0±11.2^b
FPG (mg/dl)	95.2±13.3	153.9±49.4^b
HOMA-IR	2.60±1.99	3.55±2.02^b
TG (mg/dL)	161.4±99.0	182.9±104.9^a
HDL-C (mg/dL)	50.9±12.8	46.2±11.0^b
TG/HDL-C	3.64±3.10	4.34±3.01^a
Non-HDL-C (mg/dL)	151.3±38.9	141.6±42.7^a
LDL-C (mg/dL)	121.3±33.5	105.5±35.7^a
SBP (mmHg)	117.8±14.1	127.6±16.4^b
DBP (mmHg)	77.7±9.0	80.7±7.6^b
Apo A-I (mg/dL)	145.1±28.8	138.0±27.1^a
Apo B (mg/dL)	94.3±25.6	88.0±26.6^a

Variables are expressed as mean±SD.

P<0.01.

P<0.001.

BMI, body mass index; FPG, fasting plasma glucose; HOMA-IR, homeostasis model assessment of insulin resistance; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood pressure; Apo, apolipoprotein.

Table 2.

Pearson Correlation Coefficients Between Variables of Interest

	Waist circumference	HOMA-IR	TG	HDL-C	SBP	Apo A-I
Nondiabetics
HOMA-IR	0.511^a
TG	0.287^a	0.255^a
HDL-C	−0.197^b	−0.139	−0.298^a
SBP	0.378^a	0.072	0.150^c	−0.051
Apo A-I	−0.223^b	−0.105	−0.103	0.515^a	−0.097
Apo B	0.161^c	0.156^c	0.258^a	0.067	0.166^c	0.018
Diabetics
HOMA-IR	0.300^a
TG	0.125^c	0.224^a
HDL-C	−0.147^c	−0.186^a	−0.240^a
SBP	0.130^c	0.112	0.011	−0.033
Apo A-I	−0.070	−0.064	−0.034	0.383^a	−0.053
Apo B	0.002	0.077	0.215^a	0.032	0.075	0.044

P<0.001.

P<0.01.

P<0.05.

HOMA-IR, homeostasis model assessment of insulin resistance; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure; Apo, apolipoprotein.

Factor loading patterns of the correlation matrix are shown in Table 3. Considering waist circumference, HOMA-IR, TG, HDL-C, and SBP as components of metabolic syndrome, factor analysis identified two underlying factors, explaining 58.4% and 54.3% of the total variance in nondiabetic and diabetic subjects, respectively. Regardless of diabetes status, waist circumference, HOMA-IR, and SBP loaded on factor 1 and lipid measures loaded on factor 2. Likewise, in nondiabetics, after addition of apolipoproteins to the variables, analysis extracted two factors; the first one was associated with waist circumference, HOMA-IR, TG, SBP, and Apo B and the second one was mainly related to HDL-C and Apo A-I. By contrast, in diabetic subjects, addition of apolipoproteins to the variables of interest of model 1 led to a three-factor structure model, in which Apo B along with TG loaded on the newly identified factor 3.

Table 3.

Factor Loadings of Conventional Variables of Models 1 and 2 with or without Inclusion of Apolipoproteins on the Selected Factors

	Model 1					Model 2
	Conventional variables		Conventional variables+apolipoproteins			Conventional variables	Conventional variables+apolipoproteins
	Factor 1	Factor 2	Factor 1	Factor 2	Factor 3	Factor 1	Factor 1	Factor 2
Nondiabetics
Waist circumference	0.77^a	0.29	0.75^a	−0.23	—	0.82^a	0.80^a	—
HOMA-IR	0.51^*	0.41	0.64^a	−0.14	—	0.68^a	0.65^a	—
TG	0.33	0.65^a	0.61^a	−0.20	—	—	—	—
HDL-C	0.10	−0.85^a	−0.13	0.86^*	—	—	—	—
TG/HDL-C	—	—	—	—	—	0.58^a	0.61^a	—
SBP	0.79^a	−0.14	0.53^a	−0.02	—	0.57^a	0.51^a	—
Apo A-I	—	—	−0.05	0.84^a	—	—	−0.43^a	—
Apo B	—	—	0.56^a	0.18	—	—	0.41^a	—
% variance explained	35.1	23.3	27.9	22.8	—	45.3	34.9	—
Diabetics
Waist circumference	0.75^a	−0.10	0.78^a	−0.06	−0.08	0.70^a	0.74^a	−0.16
HOMA-IR	0.69^a	−0.26	0.77^a	−0.07	0.16	0.78^a	0.77^a	0.07
TG	0.28	−0.67^a	0.34	−0.17	0.68^*	—	—	—
HDL-C	0.01	0.78^a	−0.27	0.76^a	−0.09	—	—	—
TG/HDL-C	—	—	—	—	—	0.60^a	0.55^a	0.34
SBP	0.50^a	0.39	0.45^a	−0.03	0.04	0.42^a	0.47^a	0.10
Apo A-I	—	—	0.09	0.86^a	0.02	—	−0.27	−0.20
Apo B	—	—	−0.13	0.06	0.84^a	—	0.12	0.65^a
% variance explained	32.2	22.1	20.4	19.5	16.6	42.7	25.6	18.7

Factors with an eigenvalue ≥1 were selected for analysis. Unrotated factor loadings are shown for one-factor structure models.

Factor loadings ≥|0.40|.

Performing factor analysis on variables of interest of model 2 (i.e., waist circumference, HOMA-IR, TG/HDL-C ratio, and SBP), one underlying factor was identified in both nondiabetic and diabetic subjects (accounting for 45.3% and 42.7% of the total variance in nondiabetic and diabetic subjects, respectively). Addition of apolipoproteins did not interfere with the one-factor structure model of metabolic syndrome in nondiabetics. However, in diabetic subjects, Apo B loaded on the second identified factor and Apo-A-I was weakly associated with both extracted factors.

The first extracted factor of each studied model was directly associated with metabolic syndrome regardless of diabetes status (Table 4). Generally, loading of apolipoproteins on a factor was accompanied by slight improvement in the odds ratio of that factor for metabolic syndrome.

Table 4.

Odds Ratios (Confidence Intervals) of Calculated Factor Scores for ATP III- and IDF-Defined Metabolic Syndrome

	Model 1					Model 2
	Conventional variables		Conventional variables+apolipoproteins			Conventional variables	Conventional variables+apolipoproteins
	Factor 1	Factor 2	Factor 1	Factor 2	Factor 3	Factor 1	Factor 1	Factor 2
Nondiabetics
ATP III	5.0 (3.1–8.0)	1.6 (1.1–2.3)	7.6 (4.3–13.5)	0.56 (0.41–0.76)		6.0 (3.6–10.1)	6.3 (3.8–10.6)
IDF	6.7 (3.9–11.5)	1.9 (1.3–2.7)	10.3 (5.4–19.7)	0.61 (0.45–0.82)		8.5 (4.7–15.1)	9.2 (5.0–17.0)
Diabetics
ATP III	4.8 (3.0–7.8)	0.58 (0.42–0.80)	3.5 (2.2–5.7)	0.54 (0.39–0.75)	1.7 (1.1–2.5)	8.3 (4.2–16.5)	7.2 (3.7–13.7)	1.5 (1.0–2.2)
IDF	5.4 (3.2–9.0)	0.52 (0.38–0.72)	3.6 (2.0–6.7)	0.71 (0.52–0.95)	2.0 (1.3–3.1)	10.4 (6.1–17.8)	12.1 (6.5–21.9)	1.7 (1.1–2.6)

ATP III, Adult Treatment Panel III; IDF, International Diabetes Federation.

Discussion

Most studied variables were worse in diabetic subjects in proportion to nondiabetics. Apo A-I and Apo B were highly correlated with HDL-C and TG, respectively, in both diabetic and nondiabetic subjects. Performing analysis on the conventional variables of metabolic syndrome, one or two factors were extracted, depending on whether lipid measures were considered as two distinct variables (i.e., TG and HDL-C) or as a composite measure (i.e., TG/HDL-C ratio). Inclusion of apolipoproteins into factor analysis did not change the factor structure model in nondiabetic subjects remarkably. On the contrary, an extra factor was identified after addition of apolipoproteins to the variables of interest in diabetic subjects. Generally, loading of apolipoproteins led to improvement in the odds ratios of factor scores for metabolic syndrome in nondiabetic subjects.

Association of apolipoproteins with insulin resistance, metabolic syndrome, and its components has been topic of some prior studies.^20

–23 In this respect, it has been shown that low levels of Apo A-I associate with metabolic syndrome.²⁴ On the contrary, Sung and Hwang have shown that Apo B is positively related to metabolic syndrome and insulin resistance.²⁵ Moreover, Apo B compared to non-HDL-C is a better indicator of metabolic syndrome²⁶ and is also more strongly associated with cardiovascular risk factors (e.g., central adiposity, insulin resistance, thrombosis, and inflammation).²⁷ Furthermore, Apo B is a superior atherogenic risk factor than LDL-C.²⁸ In line with aforementioned findings, we showed that Apo A-I and Apo B are inversely and directly associated with cardiovascular risk factors, respectively, especially in nondiabetic subjects.

Studies performing factor analysis on metabolic syndrome components along with apolipoproteins are limited. Due to differences in the number and nature of selected components, in methods used for interpretation of analysis, and also in sample collection, it is difficult to perform formal direct comparisons between various studies. However, an overview reveals some common patterns. For instance, in previous studies, Apo B and Apo A-I tended to load on the same factors with TG and HDL-C, respectively, which is consistent with our findings.^29,30 Moreover, in another study performed on children, the Apo B/Apo A-I ratio and lipid measures were related to the second extracted factor.³¹ We found that loading patterns of metabolic syndrome components are not substantially modified by apolipoproteins in nondiabetics. Furthermore, in general agreement with our previous studies, the combination of two lipid measures of metabolic syndrome led to a one-factor structure model,^8,9 which was maintained in nondiabetics following addition of apolipoproteins to the analysis. It has been suggested that insulin resistance is the common pathophysiologic characteristic that weaves together components of metabolic syndrome.⁶ The current study confirms that the idea of a single underlying factor is plausible, and levels of apolipoproteins are mainly controlled by the pathophysiologic pathway that underlies development of metabolic syndrome, especially in nondiabetic subjects.

One of the strengths of our study is that by performing our analysis on diabetic and nondiabetic subjects separately, we tried to control the influence of diabetes status on clustering of metabolic syndrome components. On the contrary, the major limitation of our study is that we used a cross-sectional study design, which is also employed by many prior similar studies. Because cross-sectional analyses provide information at a single point in time, they cannot be used to infer a causal relationship between metabolic syndrome components.

In conclusion, our study reveals that it is reasonable to conceive a single underlying pathophysiologic pathway for development of metabolic syndrome. Identification of a one-factor structure model, regardless of apolipoproteins in nondiabetic subjects, implies that apolipoproteins are mediated by the common pathophysiologic pathway that clusters core components of metabolic syndrome. However, further longitudinal studies are needed to determine whether inclusion of apolipoproteins in the definition improves the predictive ability of metabolic syndrome for diabetes and coronary artery disease.

Footnotes

Author Disclosure Statement

The authors declare no conflicts of interests.

References

Grundy

, Hansen

, Smith

et al. Clinical management of metabolic syndrome: Report of the American Heart Association/National Heart, Lung, and Blood Institute/American Diabetes Association conference on scientific issues related to management. Circulation, 2004; 109:551–556.

Sniderman

, Faraj

. Apolipoprotein B, apolipoprotein A-I, insulin resistance and the metabolic syndrome. Curr Opin Lipidol, 2007; 18:633–637.

Walldius

, Jungner

. The apoB/apoA-I ratio: A strong, new risk factor for cardiovascular disease and a target for lipid-lowering therapy—a review of the evidence. J Intern Med, 2006; 259:493–519.

Marcovina

, Packard

. Measurement and meaning of apolipoprotein AI and apolipoprotein B plasma levels. J Intern Med, 2006; 259:437–446.

Davidson

. Apolipoprotein measurements: Is more widespread use clinically indicated? Clin Cardiol, 2009; 32:482–486.

Meigs

. Invited commentary: Insulin resistance syndrome? Syndrome X? Multiple metabolic syndrome? A syndrome at all? Factor analysis reveals patterns in the fabric of correlated metabolic risk factors. Am J Epidemiol, 2000; 152:908–911.

Hanson

, Imperatore

, Bennett

et al. Components of the “metabolic syndrome” and incidence of type 2 diabetes. Diabetes, 2002; 51:3120–3127.

Esteghamati

, Zandieh

, Khalilzadeh

et al. Clustering of leptin and physical activity with components of metabolic syndrome in Iranian population: An exploratory factor analysis. Endocrine, 2010; 38:206–213.

Esteghamati

, Zandieh

, Khalilzadeh

et al. Clustering of metabolic syndrome components in a Middle Eastern diabetic and non-diabetic population. Diabetol Metab Syndr, 2010; 2:36.

10.

Kraja

, Province

, Arnett

et al.

Do inflammation and procoagulation biomarkers contribute to the metabolic syndrome cluster?

Nutr Metab (Lond), 2007; 4:28.

11.

Lafortuna

, Adorni

, Agosti

et al. Factor analysis of metabolic syndrome components in obese women. Nutr Metab Cardiovasc Dis, 2008; 18:233–241.

12.

Sakkinen

, Wahl

, Cushman

et al. Clustering of procoagulation, inflammation, and fibrinolysis variables with metabolic factors in insulin resistance syndrome. Am J Epidemiol, 2000; 152:897–907.

13.

Matthews

, Hosker

, Rudenski

et al. Homeostasis model assessment: Insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia, 1985; 28:412–419.

14.

Genuth

, Alberti

, Bennett

et al. Follow-up report on the diagnosis of diabetes mellitus. Diabetes Care, 2003; 26:3160–3167.

15.

International Diabetes Federation. The IDF consensus worldwide definition of the metabolic syndrome [article online] 2006. www.idf.org/webdata/docs/IDF_Meta_def_final.pdf/. 2012 March 14.

16.

Hodge

, Boyko

, de Courten

et al. Leptin and other components of the Metabolic Syndrome in Mauritius—a factor analysis. Int J Obes Relat Metab Disord, 2001; 25:126–131.

17.

Leyva

, Godsland

, Ghatei

et al. Hyperleptinemia as a component of a metabolic syndrome of cardiovascular risk. Arterioscler Thromb Vasc Biol, 1998; 18:928–933.

18.

Pyorala

, Miettinen

, Halonen

et al. Insulin resistance syndrome predicts the risk of coronary heart disease and stroke in healthy middle-aged men: The 22-year follow-up results of the Helsinki Policemen Study. Arterioscler Thromb Vasc Biol, 2000; 20:538–544.

19.

Lawlor

, Ebrahim

, May

et al. (Mis)use of factor analysis in the study of insulin resistance syndrome. Am J Epidemiol, 2004; 159:1013–1018.

20.

Wallenfeldt

, Bokemark

, Wikstrand

et al. Apolipoprotein B/apolipoprotein A-I in relation to the metabolic syndrome and change in carotid artery intima-media thickness during 3 years in middle-aged men. Stroke, 2004; 35:2248–2252.

21.

Sierra-Johnson

, Romero-Corral

, Somers

et al. ApoB/apoA-I ratio: An independent predictor of insulin resistance in US non-diabetic subjects. Eur Heart J, 2007; 28:2637–2643.

22.

Sierra-Johnson

, Somers

, Kuniyoshi

et al. Comparison of apolipoprotein-B/apolipoprotein-AI in subjects with versus without the metabolic syndrome. Am J Cardiol, 2006; 98:1369–1373.

23.

Belfki

, Ben Ali

, Bougatef

et al. The apolipoprotein B/apolipoprotein A 1 ratio in relation to metabolic syndrome and its components in a sample of the Tunisian population. Exp Mol Pathol, 91:622–625.

24.

Mattsson

, Magnussen

, Ronnemaa

et al. Metabolic syndrome and carotid intima-media thickness in young adults: roles of apolipoprotein B, apolipoprotein A-I, C-reactive protein, and secretory phospholipase A2: The cardiovascular risk in young Finns study. Arterioscler Thromb Vasc Biol, 2010; 30:1861–1866.

25.

Sung

, Hwang

. Association between insulin resistance and apolipoprotein B in normoglycemic Koreans. Atherosclerosis, 2005; 180:161–169.

26.

Han

, Park

, Kim

et al. Apolipoprotein B is a better marker than non-HDL-cholesterol for the metabolic syndrome in Koreans. Atherosclerosis, 2008; 197:333–338.

27.

Sattar

, Williams

, Sniderman

et al. Comparison of the associations of apolipoprotein B and non-high-density lipoprotein cholesterol with other cardiovascular risk factors in patients with the metabolic syndrome in the Insulin Resistance Atherosclerosis Study. Circulation, 2004; 110:2687–2693.

28.

Williams

, Sniderman

, Sattar

et al.

Comparison of the associations of apolipoprotein B and low-density lipoprotein cholesterol with other cardiovascular risk factors in the Insulin Resistance Atherosclerosis Study (IRAS)

Circulation, 2003; 108:2312–2316.

29.

Laaksonen

, Niskanen

, Nyyssonen

et al. Dyslipidaemia as a predictor of hypertension in middle-aged men. Eur Heart J, 2008; 29:2561–2568.

30.

Shmulewitz

, Auerbach

, Lehner

et al. Epidemiology and factor analysis of obesity, type II diabetes, hypertension, and dyslipidemia (syndrome X) on the Island of Kosrae, Federated States of Micronesia. Hum Hered, 2001; 51:8–19.

31.

Retnakaran

, Zinman

, Connelly

et al. Nontraditional cardiovascular risk factors in pediatric metabolic syndrome. J Pediatr, 2006; 148:176–182.