Genetic and Environmental Influences on the Associations Between Uric Acid Levels and Metabolic Syndrome Over Time

Abstract

Background:

The longitudinal associations between serum uric acid (UA) levels and metabolic syndrome (MetS) and its components, as well as the shared genetic and environmental correlations between these traits, were evaluated.

Participants and Methods:

In a total of 1803 participants (675 men and 1128 women; 695 monozygotic twin individuals, 159 dizygotic twin individuals, and 949 non-twin family members; 44.3 ± 12.8 years old) and 321 monozygotic twin pairs with data on UA levels and MetS components at baseline and follow-up, mixed linear model, conditional logistic regression, and bivariate variance component analysis were conducted.

Results:

After 3.7 ± 1.4 years, the incident and persistent prevalence of MetS were 5.3% and 11.6%, respectively. UA was positively associated with the concurrent and future number of MetS criteria, blood pressure (BP), and triglyceride (TG) levels, whereas an inverse association was observed between UA and future high-density lipoprotein cholesterol (HDL-C) levels after adjusting for twin and household effects, demographics, health behaviors at baseline, and other confounders according to outcome variables. In the adjusted bivariate analysis, UA had genetic and environmental correlations with the concurrent and future number of MetS criteria, and had genetic correlations with concurrent BP and TG levels and future diastolic BP and HDL-C levels. In the adjusted co-twin control analysis, twins with a higher UA level were more likely to have concurrent MetS [odds ratio (95% confidence interval) 1.59 (1.00–2.53)], high blood glucose levels [1.84 (1.06–3.17)], future MetS [2.35 (1.19–4.64)], and high TG levels [1.52 (1.03–2.24)] than co-twins with a lower UA level.

Conclusion:

Genetic and environmental factors affect the concurrent and longitudinal associations between UA and MetS as well as some of its components.

Introduction

Ahigher level of serum uric acid (UA), which is the end product of purine nucleoside degradation and hepatic glycolysis, is considered a risk factor for metabolic syndrome (MetS).^1,2 In a meta-analysis of prospective cohort studies, a dose–response association between serum UA level and the risk for MetS was observed,¹ and the pooled hazard ratio of MetS was 1.55 for the highest versus the lowest category of serum UA level.² Since MetS is a constellation of different metabolic components, several complex pathophysiologic mechanisms, including several enzyme activities related to insulin resistance, vasoconstriction, sodium retention, and hepatic gluconeogenesis, have been considered to be associated with the condition.^3

–9 However, whether genetic regulations affect these relationships remains unclear. Genetic correlation coefficient, which is a measure of the amount of additive genetic variation shared between UA and each MetS component, can help identify pleiotropic genetic variants and therefore may affect treatment strategies.¹⁰ Although a cross-sectional study reported a genetic correlation between UA and MetS components,¹¹ shared genetic influence on the longitudinal relationships between these traits remains to be elucidated. In addition, nongenetic effects on these associations have not been studied in a co-twin control analysis of monozygotic twins, although cross-sectional relationships, including environmental correlations, have been examined.¹⁰ In the present study that included participants from the Healthy Twin Study, the cross-sectional and longitudinal associations between UA and MetS and its components, as well as the shared genetic and environmental correlations between both features and nongenetic effects on these associations via a co-twin control analysis, were examined.

Participants and Methods

Study population

The present study subjects were participants of the Healthy Twin Study, a multicenter cohort study established in South Korea in 2005 for adults (>30 years) with same-sex twins and their first-degree adult family members.^12,13 This study included 1803 individuals from 475 families (comprising 675 male and 1128 female participants, 695 monozygotic twin individuals, 159 dizygotic twin individuals, and 949 non-twin family members with a mean age of 44.3 ± 12.8 years) with available data on their MetS components and serum UA levels at baseline and follow-up over a period of 3.7 ± 1.4 years. Of the participants, 1940 visited twice from 2005 to 2012. The subjects excluded from the current study (N = 137) tended to be older (55.9 ± 12.2 years vs. 44.3 ± 12.8 years, P < 0.001) and have higher body mass index (25.3 ± 2.9 kg/m² vs. 23.6 ± 3.2 kg/m², P < 0.001). The present study was carried out in accordance with the Declaration of Helsinki. Informed consent along with conflict of interest disclosure was obtained from the study participants. The study protocol was approved by the Institutional Review Board of Samsung Medical Center (IRB file No. 2005-08-113) and Busan Paik Hospital (IRB file No. 05-037).

Measurements

Waist circumference (WC), blood pressure (BP), as well as serum concentrations of UA, glucose [fasting serum glucose (FSG)], triglycerides (TGs), and high-density lipoprotein cholesterol (HDL-C), were measured at baseline and during follow-up visits. Trained study assistants measured the WC of the participants at the midpoint between the lower margin of the rib cage and the iliac crest using a stretch-resistant tape. In addition, BP was manually measured using a standard mercury sphygmomanometer under standard conditions. In overnight fasting condition, the serum concentrations of UA (uricase enzymatic colorimetric method), FSG (hexokinase enzymatic assay), TG (enzymatic assay), and HDL-C (enzymatic or homogeneous assay) were measured using ADVIA 1650 (Siemens, Germany) or Hitachi 7600-210/7180 (Hitachi Ltd., Japan). The interassay coefficient of variation for these measurements was set below 7%, and the analyses were conducted in a central laboratory authorized by the Korean Association of Quality Assurance for Clinical Laboratory Examination.¹⁴

The following criteria for MetS were adapted from a harmonized definition¹⁵: WC ≥90 cm for men and ≥85 cm for women; BP ≥130/85 mmHg or a history of hypertension; FSG ≥5.6 mmol/L (100 mg/dL) or a history of diabetes mellitus; TG ≥1.7 mmol/L (150 mg/dL); and HDL-C < 1.03 mmol/L (40 mg/dL) in men or <1.29 mmol/L (50 mg/dL) in women. When the number of MetS criteria was three or more, MetS was considered.

Self-reported questionnaires were used to assess the participants' medical history at baseline and follow-up, educational attainment at baseline, and lifestyle habits (such as alcohol consumption, cigarette smoking, and physical activity) at baseline.

Statistical analyses

The characteristics of the participants, such as demographics, health behaviors, zygosity, and UA levels (at baseline and follow-up) among the four categories of MetS at baseline and follow-up (no MetS at both times; no MetS at baseline and MetS at follow-up; MetS at baseline and no MetS at follow-up; and MetS at both times), were compared using chi-squared test or post hoc test in analysis of variance (Scheffe's test). Mixed linear models were used to assess the cross-sectional and longitudinal relationships between UA and each MetS component (or the number of MetS criteria) as continuous variables after adjusting for random effects (twin and household effects) and fixed effects [age, sex, education, smoking status, alcohol use, physical activity, and WC (except for the analysis on the number of MetS criteria) at baseline]. In the analysis of longitudinal associations, the models were adjusted for changes in UA levels and WC (except for the analysis on MetS criteria at follow-up) and the same MetS component (or the number of MetS criteria) at baseline. To control genetic factors, a co-twin control analysis of the associations between UA and concurrent and future MetS as well as its individual criteria in 321 monozygotic twin pairs was conducted using conditional logistic regression after adjusting for health behaviors, education, and WC (except for the analysis on MetS) at baseline. In the co-twin control analysis of longitudinal relationships, we further adjusted for changes in UA levels and WC (except for the analysis on MetS) and the same MetS component (or the number of MetS criteria) at baseline. All statistical analyses were performed using IBM Statistical Package for the Social Sciences (SPSS) version 24.0 (IBM Corp., Armonk, NY).

A bivariate variance component analysis based on the maximum likelihood ratio and variance component decomposition was carried out to evaluate shared genetic and environmental influences on the relationships between UA and concurrent and follow-up MetS components and the number of MetS criteria. This analysis was adjusted for age and sex in cross-sectional correlations as well as age, sex, and the same MetS component at baseline (or the number of MetS criteria at baseline) in longitudinal correlations. These analyses were conducted using the Sequential Oligogenic Linkage Analysis Routines (version 6.6.2). The phenotypic correlations were partitioned into relationships that are explained by genetic (ρ _G) and environmental sharing (ρ _E) after adjusting for age and sex. A genetic or environmental association between the two phenotypes was considered when the values of ρ _G or ρ _E significantly deviate from 0 (P < 0.05), respectively. Genetic correlations indicated the extent to which the additive effects of the same set of genes influenced more than one phenotype or the common genetic factors that influenced both phenotypes through shared pathways.¹⁶ Environmental correlations indicated the environmental and nonadditive genetic effects.¹⁷

Results

Of the participants, 5.3% had recently developed MetS, 10.9% did not have MetS, and 11.6% had persistent MetS after 3.7 ± 1.4 years. The results of the comparison between the participants' UA levels, demographic factors, and health behaviors toward MetS at baseline and follow-up are summarized in Table 1. Compared with individuals without MetS during both visits, those with MetS at baseline or follow-up were more likely to have higher UA levels and were older. Sex, educational attainment, zygosity, and health behaviors were significantly different among the subgroups.

Table 1.

Comparison of Characteristics Between Subjects with Metabolic Syndrome Over Time (N = 1803)

	MetS at baseline/follow-up, mean ± SD or N (%)
	MetS(−)/MetS(−) (N = 1300)	MetS(−)/MetS(+) (N = 96)	MetS(+)/MetS(−) (N = 197)	MetS(+)/MetS(+) (N = 210)	P^*
Characteristics at baseline
UA (mg/dL)	4.5 ± 1.3	5.2 ± 1.4^†	5.6 ± 1.3^†	5.7 ± 1.4^†	<0.001
Age (years)	41.5 ± 11.8	48.6 ± 11.9^†	48.37 ± 12.3^†	51.4 ± 12.3^†	<0.001
Men (N = 675 vs. 1128)	390 (30.0)	37 (38.5)	131 (66.5)	117 (55.7)	<0.001
Zygosity					<0.001
Monozygotic twin (N = 695)	549 (42.2)	40 (41.7)	59 (29.9)	35 (20.2)
Dizygotic twin (N = 159)	124 (9.5)	4 (4.2)	19 (9.6)	12 (5.7)
No twin (N = 949)	627 (48.3)	52 (54.2)	119 (60.5)	151 (71.9)
Graduated high school	622 (47.9)	36 (37.5)	77 (39.5)	61 (29.0)	<0.001
Current smoker	210 (16.2)	21 (21.9)	55 (27.9)	57 (27.1)	0.001
Alcohol use (grams/week)	93 ± 162	105 ± 120	140 ± 191	168 ± 208^†	<0.001
Physical activity (MET-min/week)	6239 ± 9977	7433 ± 10579	8166 ± 12188	7032 ± 9521	0.065
Change over time
UA (mg/dL)	0.1 ± 0.8	0.3 ± 0.8	0.1 ± 0.9	0.1 ± 1.0	0.076
WC (cm)	1.2 ± 5.4	3.8 ± 7.3^†	−0.6 ± 5.2^†	1.6 ± 4.5	<0.001

MetS was adapted from a harmonized definition.

Analysis of variance test and post hoc analysis by Scheffe test [^† P < 0.05 compared to MetS(−) at baseline and follow-up] or chi-squared test.

MetS, metabolic syndrome; SD, standard deviation; UA, uric acid; WC, waist circumference.

After adjusting for age, sex, education, health behaviors, WC (except for the analysis on the number of MetS criteria), and household and twin effects, results showed that UA was positively associated with the concurrent number of MetS criteria, BP, and TG levels (P < 0.05). After adjusting for the same MetS component at baseline and changes in WC (except for the analysis on the future number of MetS criteria) and UA levels, results showed that UA level at baseline was positively associated with the number of MetS criteria, BP, and TG levels at follow-up, whereas an inverse association was observed between UA level and HDL-C levels at follow-up (P < 0.05). In these longitudinal analyses, a change in UA levels was positively associated with diastolic BP and TG levels at follow-up (P < 0.05). With respect to the concurrent bivariate analysis, UA had genetic and environmental correlations with systolic BP, TG levels, and the number of MetS criteria after adjusting for age and sex. With respect to the longitudinal bivariate analysis, UA had significant genetic and environmental correlations with the future number of MetS criteria and HDL-C levels after adjusting for age, sex, and the same MetS component (or the number of MetS criteria) at baseline. Changes in UA level had genetic correlations with FSG and TG levels at follow-up as well as environmental correlations with FSG and HDL-C levels at follow-up after adjusting for age, sex, and the same MetS component at baseline (Table 2).

Table 2.

Relationships of Uric Acid at Baseline and Change in Uric Acid Over Time with Metabolic Syndrome and Its Components at Baseline and Follow-Up

	Beta coefficient (95% CI) per 1 mg/dL change in UA		Cross-trait correlations with UA at baseline ^a		Cross-trait correlations with change in UA over time
	At baseline	Change over time	Additive genetic	Environmental	Additive genetic	Environmental
At baseline^b
Number of MetS criteria	0.12 (0.08 to 0.15)	—	0.27 ± 0.05^*	0.26 ± 0.05^*	—	—
Systolic BP	1.15 (0.55 to 1.76)	—	0.18 ± 0.05^*	0.11 ± 0.05^*	—	—
Diastolic BP	0.64 (0.22 to 1.06)	—	0.15 ± 0.05^*	0.08 ± 0.05	—	—
Glucose/	0.20 (−0.29 to 0.68)	—	0.07 ± 0.06	0.08 ± 0.05	—	—
TGs	7.13 (4.33 to 9.93)	—	0.21 ± 0.06^*	0.22 ± 0.04^*	—	—
HDL-C	−0.15 (−0.66 to 0.34)	—	−0.09 ± 0.05^*	−0.06 ± 0.04	—	—
At follow-up^c
Number of MetS criteria	0.09 (0.05 to 0.13)	0.05 (−0.01 to 0.09)	0.16 ± 0.07^*	0.15 ± 0.05^*	0.20 ± 0.11	0.05 ± 0.04
Systolic BP	0.60 (0.01 to 1.18)	0.01 (−0.76 to 0.76)	0.13 ± 0.07	0.07 ± 0.05	0.01 ± 0.12	−0.02 ± 0.04
Diastolic BP	0.65 (0.20 to 1.10)	0.61 (0.03 to 1.19)	0.17 ± 0.06^*	0.01 ± 0.04	0.05 ± 0.10	0.01 ± 0.04
Glucose/	−0.24 (−0.83 to 0.41)	−0.28 (−1.03.0.47)	0.02 ± 0.06	0.05 ± 0.04	0.24 ± 0.11^*	−0.11 ± 0.04^*
TGs	6.62 (3.80 to 9.43)	6.23 (2.58 to 9.87)	0.04 ± 0.09^*	0.21 ± 0.05^*	0.43 ± 0.14^*	−0.04 ± 0.04
HDL-C	−0.84 (−1.27 to −0.41)	0.30 (−0.27 to 0.86)	−0.12 ± 0.05^*	−0.13 ± 0.04^*	−0.12 ± 0.09	0.14 ± 0.04^*

MetS was adapted from a harmonized definition.

Values represent OR (95% CI), ρ g ± SE, or ρ e ± SE.

Bivariate analysis after adjusting for age and sex ^cadjusting for age, sex, and same MetS components [or number of MetS criteria] at baseline.

Mixed linear model after adjusting for random effects (twin effect and household effect) and fixed effects [age, sex, education, WC (except for number of MetS criteria), smoking, alcohol, and physical activity at baseline].

Add same component [or number of MetS criteria] at baseline and changes in WC over time to the fixed effects in the ^bModel.

P < 0.05.

BP, blood pressure; CI, confidence interval; HDL-C, high-density lipoprotein cholesterol; OR, odds ratio; SE, standard error; TG, triglyceride.

Table 3 presents the results of the co-twin control analysis of 321 monozygotic twin pairs. Twins with a higher UA level at baseline were more likely to have concurrent MetS [odds ratio (OR): 1.59 and 95% confidence interval (95% CI): 1.00–2.53] and high FSG (OR: 1.84 and 95% CI: 1.06–3.17) after adjusting for education, health behaviors, and WC (for the analysis on concurrent high FSG). Furthermore, twins with a higher UA level at baseline were more likely to have MetS (OR: 2.35 and 95% CI: 1.19–4.64) and high TGs at follow-up (OR: 1.52 and CI: 1.03–2.24) than co-twins with low UA at baseline after adjusting for education, health behaviors, WC (for the analysis on high TG levels), same MetS components (TG or the number of MetS criteria) at baseline, and changes in WC (for the analysis on high TG levels) and UA.

Table 3.

Association of Uric Acid with Concurrent and Follow-Up Metabolic Syndrome in Monozygotic Twin Pairs

	UA at baseline	Change in UA over time
At baseline^a
MetS	1.59 (1.00–2.53)
High BP	1.19 (0.77–1.86)
High glucose	1.84 (1.06–3.17)
High TGs	1.33 (0.79–2.24)
Low HDL-C	0.95 (0.64–1.40)
At follow-up^b
MetS	2.35 (1.19–4.64)	1.51 (0.84–2.72)
High BP	0.89 (0.48–1.65)	0.92 (0.55–1.51)
High glucose	1.18 (0.62–2.26)	1.11 (0.57–2.14)
High TGs	1.52 (1.03–2.24)	1.17 (0.74–1.84)
Low HDL-C	0.92 (0.60–1.40)	0.88 (0.61–1.28)

Co-twin control analysis in 321 pairs of monozygotic twins. MetS was adapted from a harmonized definition.

Values are OR (95% CI) using conditional logistic regression analysis after adjusting for ^asmoking status, alcohol use, physical activity, education, WC [except for analysis for MetS itself] at baseline.

Baseline characteristics, same MetS component [MetS status in the analysis for MetS itself] at baseline, and changes in UA and WC [except for analysis for MetS itself].

Discussion

Based on the results of the present study, a higher UA level at baseline was associated with a greater number of MetS criteria at baseline and follow-up, and the association between UA levels and the future number of MetS criteria was independent of the number of MetS criteria at baseline. These associations were regulated by shared genetic and environmental factors in the bivariate analysis. After adjusting for the genetic factors of monozygotic twin pairs, a higher UA level at baseline was associated with higher ORs for concurrent and future MetS. These findings from the co-twin control analysis support the effect of nongenetic mechanisms on the associations between UA and MetS at baseline and follow-up.

The dose–response association between UA level and the risk of developing MetS was observed in the meta-analysis of prospective studies. The risk for incident MetS increased by 30% per 1 mg/dL increase in UA.¹ Moreover, a higher UA level was associated with the risk for increased TG levels and high BP by 41% and 30%, respectively.² The results of the two meta-analyses are consistent with our findings. However, studies on the genetic and environmental correlations between UA and MetS components are limited. Our cross-sectional findings on the genetic influences on the association between UA and concurrent TG and HDL-C levels as well as the environmental regulations that affect the relationships between UA and systolic BP and TG levels are comparable with those of the family study by the National Heart, Lung, and Blood Institute (NHLBI).¹¹ To the best of our knowledge, no comparable studies on the genetic and environmental influences on the longitudinal associations between UA and MetS components are available. In our longitudinal bivariate analysis, the genetic correlation between UA and future HDL-C levels and the environmental association between UA and future TG levels remained significant after adjusting for the same MetS component at baseline. Furthermore, an increase in the UA level over a period of 3.7 ± 1.4 years was independently associated with diastolic BP and TG levels at follow-up after considering UA level, WC, and the same MetS component at baseline. The result of the bivariate genetic analysis suggests the possible effect of genetic regulation on the association between changes in UA levels and future TG and FSG levels. The associations between the changes in UA level and future MetS components were similar to those observed in the 7-year follow-up study about the associations between increased UA level and the development of MetS and some MetS criteria.¹⁸ However, the effects of genetic regulation on the association between changes in UA level and future MetS components have not been presented.

Complex biological mechanisms may affect the associations between UA and MetS and its components. Hyperuricemia induces the following mechanisms related to MetS components³: inhibition of endothelial nitric oxide synthase activity that increases insulin resistance and BP⁴; inhibition of adenosine monophosphate kinase (AMPK) and increase in adenosine monophosphate dehydrogenase (AMPD) activity that increases hepatic gluconeogenesis and serum glucose level⁵; inhibition of epithelial sodium channels on the distal nephrons with consequent renal sodium retention causing an increase in BP⁶; upregulation of fructokinase, inhibition of AMPK, and an increase in AMPD in the hepatocytes that induces hepatic steatosis^7,8; and increase in nicotinamide adenine dinucleotide phosphate-oxidase activity that causes intracellular UA and consequently intracellular oxidative stress and mitochondrial injury.⁹ Although a significant association was observed between UA and TG as well as the MetS components, the biological mechanisms that influence this association are not fully elucidated.³

The genetic regulation that influences the associations between UA and changes in UA levels as well as MetS and its components may be attributed to relational or mosaic pleiotropy. The relational pleiotropy refers to the phenomenon in which underlying genetic/familial factors directly affect the initiating traits and influence the other genetically correlated traits via pathways mediated by those traits. On the other hand, mosaic pleiotropy refers to a situation in which responsible genetic/familial factors directly act on each of the traits.^11,19

The evidence on pleiotropic pathway was also supported by the results of a linkage analysis. In the study, a genomic region on chromosome 2 was found as a pleiotropic locus that affects UA and other MetS phenotypes.²⁰ If both shared environmental and genetic factors were associated with UA and MetS and its components, a within-pair twin case/control study that controls the effect of genetics and early environment that are shared by twins could validate the associations. Based on the significant associations observed in the co-twin control analysis in monozygotic twin pairs, we can hypothesize that acquired factors, such as unmonitored dietary and lifestyle habits, may affect the associations between UA and concurrent and future MetS through physiological mechanisms.

This study has key strengths. First, our findings were based on population-based twin and family members and were more generalizable compared with those of hospital-based studies. In addition, our findings could support the evidence on the role of genetic and environmental factors on the longitudinal relationships between UA and MetS components. However, the present study also has limitations. We did not control the confounding factors, such as current medication regimen and eating pattern (consumption of sugar-filled and purine-containing food), that influence UA levels and MetS components. Although the study design was not intended for an interventional study, we still presented the results, clinical interpretation of the findings, and important lifestyle habits that can improve the health of the participants. This study results could help participants change their dietary and lifestyle habits and start medication therapy management.²¹ Furthermore, the associations in dizygotic twin pairs were not examined. This analysis would provide information on the differences in these relationships after adjusting for shared early environment.

Nevertheless, this study confirmed the significant associations between changes in UA levels and the concurrent or future MetS after controlling genetic factors, and the possible effects of the pleiotropy pathway on these relationships were also presented. In conclusion, UA might be a risk factor for the concurrent development of the components of MetS through genetic and environmental mechanisms. Thus, future studies must be conducted to better elucidate the genetic and environmental factors influencing the risk of developing metabolic and cardiovascular diseases.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Yuan

, Yu

, Li

, et al. Serum uric acid levels and risk of metabolic syndrome: A dose-response meta-analysis of prospective studies. J Clin Endocrinol Metab, 2015; 100:4198–4207.

Liu

, Que

, Zhou

, et al. Dose-response relationship of serum uric acid with metabolic syndrome and non-alcoholic fatty liver disease incidence: A meta-analysis of prospective studies. Sci Rep, 2015; 5:14325.

Sharaf El Din

UAA

, Salem

, Abdulazim

. Uric acid in the pathogenesis of metabolic, renal, and cardiovascular diseases: A review. J Adv Res, 2017; 8:537–548.

Roy

, Perreault

, Marette

. Insulin stimulation of glucose uptake in skeletal muscles and adipose tissues in vivo is NO dependent. Am J Physiol, 1998; 274:E692–E699.

Cicerchi

, Li

, Kratzer

, et al. Uric acid-dependent inhibition of AMP kinase induces hepatic glucose production in diabetes and starvation: Evolutionary implications of the uricase loss in hominids. FASEB J, 2014; 28:3339–3350.

, Huang

, Li

, et al. Hyperuricemia induces hypertension through activation of renal epithelial sodium channel (ENaC). Metabolism, 2016; 65:73–83.

Lanaspa

, Sanchez-Lozada

, Cicerchi

, et al. Uric acid stimulates fructokinase and accelerates fructose metabolism in the development of fatty liver. PLoS One, 2012; 7:e47948.

Lanaspa

, Cicerchi

, Garcia

, et al. Counteracting roles of AMP deaminase and AMP kinase in the development of fatty liver. PLoS One, 2012; 7:e48801.

Sautin

, Nakagawa

, Zharikov

, et al. Adverse effects of the classic antioxidant uric acid in adipocytes: NADPH oxidase-mediated oxidative/nitrosative stress. Am J Physiol Cell Physiol, 2007; 293:C584–C596.

10.

Povel

, Boer

, Feskens

. Shared genetic variance between the features of the metabolic syndrome: Heritability studies. Mol Genet Metab, 2011; 104:666–669.

11.

Tang

, Hong

, Province

, et al. Familial clustering for features of the metabolic syndrome: The National Heart, Lung, and Blood Institute (NHLBI) Family Heart Study. Diabetes Care, 2006; 29:631–636.

12.

Sung

, Cho

, Lee

, et al. Healthy Twin: A twin-family study of Korea—Protocols and current status. Twin Res Hum Genet, 2006; 9:844–848.

13.

Gombojav

, Song

, Lee

, et al. The Healthy Twin Study, Korea updates: Resources for omics and genome epidemiology studies. Twin Res Hum Genet, 2013; 16:241–245.

14.

Lee

, Song

, Sung

. Genetic and environmental associations between C-reactive protein and components of the metabolic syndrome. Metab Syndr Relat Disord, 2013; 11:136–142.

15.

Alberti

, Eckel

, Grundy

, et al. Harmonizing the metabolic syndrome: A joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation, 2009; 120:1640–1645.

16.

Almasy

, Blangero

. Multipoint quantitative-trait linkage analysis in general pedigrees. Am J Hum Genet, 1998; 62:1198–1211.

17.

MacCluer

, Scavini

, Shah

, et al. Heritability of measures of kidney disease among Zuni Indians: The Zuni Kidney Project. Am J Kidney Dis, 2010; 56:289–302.

18.

Norvik

, Storhaug

, Ytrehus

, et al. Overweight modifies the longitudinal association between uric acid and some components of the metabolic syndrome: The Tromso Study. BMC Cardiovasc Disord, 2016; 16:85.

19.

Rieger

, Michaelis

, Green

. Glossary of Genetics: Classical and Molecular. Berlin, Germany: Springer-Verlag, 1991.

20.

Tang

, Miller

, Rich

, et al. Linkage analysis of a composite factor for the multiple metabolic syndrome: The National Heart, Lung, and Blood Institute Family Heart Study. Diabetes, 2003; 52:2840–2847.

21.

Song

, Sung

, Lee

. Associations between adiposity and metabolic syndrome over time: The Healthy Twin Study. Metab Syndr Relat Disord, 2017; 15:124–129.