- Email: [email protected]
Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort
Atherosclerosis xxx (2015) 1e7
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort Dhananjay Yadav a, d, Eun Soo Lee a, Hong Min Kim a, Eunhee Choi b, Eun Young Lee c, Jung Soo Lim a, Song Vogue Ahn d, e, Sang Baek Koh d, e, Choon Hee Chung a, b, * a
Department of Internal Medicine, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea Institute of Lifestyle Medicine, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea Department of Internal Medicine, Soonchunhyang University Cheonan Hospital, Cheonan, Republic of Korea d Department of Preventive Medicine, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea e Institute of Genomic Cohort, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 November 2014 Received in revised form 30 March 2015 Accepted 21 April 2015 Available online xxx
Objective: Recent studies have demonstrated an association between serum uric acid (SUA) levels and metabolic syndrome (MetS). However, paucity of available data regarding the cause and effect relationship between SUA and MetS in healthy adults is still a big challenge which remains to be studied. Therefore, we investigated whether SUA predicts new onset of MetS in a population-based cohort study. Methods: The study included 1590 adults (661 men and 929 women) aged 40e70 years without MetS at baseline (2005e2008) and subjects were prospectively followed for 2.6 years. To evaluate the relationship between SUA and MetS, we divided the aforementioned subjects into quintiles (SUA-I to SUA-V) from the lowest to the highest values of SUA. SUA was measured by the enzymatic colorimetric method. We used category-free net reclassification improvement (NRI) and integrated discrimination improvement (IDI) to characterize the performance of predicted model. Results: During a mean of 2.6 years of follow-up, 261(16.4%) adults developed MetS. MetS variables were significantly related to the baseline SUA level. Waist circumference (WC), blood pressure (BP), and serum triglyceride (TG) were significantly higher in the highest quintile of SUA compared to the lowest SUA quintile in men and women. After adjustment for age, total cholesterol and low-density lipoprotein cholesterol (LDL-C) in men and women, subjects in the fifth quintiles of SUA showed significantly higher ORs for incident MetS. The association between hyperuricemia and new onset of MetS were consistently stronger in women than men. Additionally, among women, we found an improvement in the area under the ROC curve in the models that added SUA to core components of MetS. Conclusion: Our study suggests that SUA is significantly correlated with future risk of WC, BP, TG and may predicted as a risk factor for developing MetS. SUA may have a clinical role in predicting new-onset metabolic syndrome among women. Large prospective study is needed to reveal the clinical significance of SUA in metabolic disease. © 2015 Published by Elsevier Ireland Ltd.
Keywords: Serum uric acid Metabolic syndrome Hyperuricemia Korean adults
1. Introduction Metabolic syndrome (MetS) is characterized by the grouping of cardiovascular risk factors, including high blood pressure (BP), abdominal obesity, dyslipidemia, and increased glucose
* Corresponding author. Department of Internal Medicine, Yonsei University Wonju College of Medicine, 162 Ilsan-Dong, Wonju, Kangwon-Do 220-701, Republic of Korea. E-mail address: [email protected] (C.H. Chung).
concentration [1]. Subjects with MetS are linked with an increased risk of developing type 2 diabetes and cardiovascular disease (CVD) [2,3]. The prevalence of MetS in Korea has been immensely increased from 24.9% in 1998 to 31.3% in 2007 due to rapid lifestyle changes [4]. Apart from the dietary habit and variables of MetS, other factors have also played a significant role in higher prevalence and incidence of developing MetS. Several epidemiological studies have established the link between hyperuricemia and the development of MetS in various populations which may relate to the burden of
http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797 0021-9150/© 2015 Published by Elsevier Ireland Ltd.
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
2
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
diabetes and CVD [5e9]. Furthermore, studies evaluating a sexrelated association between SUA and development of MetS have also been highly intensified [5,10,11]. Recently, the association between serum uric acid (SUA) and MetS was monitored not only in hyperuricemia, but also in the normal range of SUA [12,13]. Nevertheless, very limited information is available about its ability as a predictor of MetS and its components and additional clinical utility in epidemiological perspective. Hence, we designed a follow-up study to prospectively examine the association between SUA and the incidence of MetS and its components as well as the predictive ability of SUA in identifying people who will develop new onset of MetS stratified by sex in Korean rural cohort. 2. Methods 2.1. Study participants The Korean Genome and Epidemiology Study on Atherosclerosis Risk of Rural Areas in the Korean General Population (KoGESARIRANG), a population-based prospective cohort study designed to estimate the prevalence, incidence, and risk factors for chronic degenerative disorders such as diabetes, hypertension, osteoporosis, and cardiovascular disease [14e16]. The adults aged between 40 and 70 years residing in the rural areas of Wonju and Pyeongchang in South Korea participated in the study. The baseline study was executed from November 2005 to January 2008, comprised of 5178 adults (2127 men and 3051women). 3862 (74.6%) subjects were attended during the first follow-up survey (April 2008eJanuary 2011). We then excluded 1020 subjects with MetS at baseline, 1218 subjects without baseline SUA measurement, 20 subjects with a history of cardiovascular disease at baseline and 14 subjects without complete data. In total, 1590 participants were included in the present analysis (661 men and 929 women) (Fig. 1). The study protocol was approved by the institutional review board of Yonsei University College of Medicine. Informed consent was obtained from each participant in the study. 2.2. Data collection and measurements At study entry, participants completed the medical history and
lifestyle questionnaire and went through a standardized health examination according to the optimum procedures. For anthropometrical measurements, body weight and height were measured, with the participants barefooted and wearing light indoor clothing. Waist circumference (WC) was measured using a tape (SECA-200; SECA, Hamburg, Germany) in a horizontal plane midway between the inferior margin of the ribs and the superior border of the iliac crest. BP was measured with a standard mercury sphygmomanometer (Baumanometer, Copiague, NY) twice on the right arm of the participants. The mean of the two BP readings was used for the data analyses. Baseline information on current smoking status and alcohol intake was collected by self-reported questionnaire. Subjects who answered “yes” to the question “Do you perform physical exercise regularly enough to make you sweat?” were designated as the regular exercise group. Venous blood samples were collected from all participants after fasting for >12 h or overnight. Fasting blood glucose (FBG) was measured by a glucose oxidaseebased assay. Serum concentrations of high-density lipoprotein cholesterol (HDL-C) and triglyceride (TG) were determined by enzymatic methods (Advia 1650; Siemens, Tarrytown, NY). Low-density lipoprotein (LDL)-cholesterol was calculated using Friedewald's formula: LDLcholesterol ¼ Total cholesterol e (HDL-C þ TG/5). SUA was measured by the enzymatic colorimetric method. In this method, SUA is oxidized by uricase to allantoin and hydrogen peroxide based on the uricase-peroxidase method. In the following reactions, the oxidative condensation of N-ethyl-N-(2-hydroxy-3sulfopropyl)-3-methylaniline and 4-aminophenazone produces a red chromogen in the presence of peroxidase and hydrogen peroxide [17]. An automated analyzer was used to measure the intensity of the color produced in the sample and the values were reported in mg/dL. 2.3. Definition of MetS MetS was defined according to the criteria of the harmonized definition [18] for MetS, which was the end-point of this study at the follow-up visit. This definitions includes the presence of at least three out of five metabolic abnormalities: 1) abdominal obesity, defined as a waist circumference 90 cm for men or 85 cm for
Fig. 1. Descrpition of the study population.
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
women (following Korean specific cutoffs for abdominal obesity defined by the Korean Society of Obesity) [19]; 2) hypertriglyceridemia, defined as a serum TG concentration 150 mg/dL (1.69 mmol/L); 3) high blood pressure, defined as a systolic blood pressure (SBP) 130 mmHg, a diastolic blood pressure (DBP) 85 mmHg, or treatment with antihypertensive agents; 4) low serum HDL cholesterol, defined as a serum HDL cholesterol concentration < 40 mg/dL (1.04 mmol/L) for men or < 50 mg/dL (1.29 mmol/L) for women; and 5) high fasting glucose defined as a fasting serum glucose 100 mg/dL or previously diagnosed type 2 diabetes. 2.4. Quintile of SUA stratified by gender To evaluate the association between the individual components of MetS and SUA levels, study subjects were divided into quintiles according to their SUA levels (SUA-I to SUA-V, first to fifth quintile of SUA). The quintile measured the dispersion in the optimum level of SUA to predict the future estimated risk of each component of MetS, stratified by men and women. The lowest and highest quintile of SUA in men and women were 0.7e4.5, 6.3e7.0 and 0.6e3.3, 4.8e6.0 mg/dL. 2.5. Statistical analysis We analyzed all of the results separately for men and women. Study data are expressed as means ± standard deviation (SD) or frequency (%). Baseline characteristics between the groups with and without incident MetS were assessed by two-sample t-test or chi-square test. Participants were characterized into five groups accordingly to the components of MetS disorder (presence of MetS variables such as 0, 1, 2, 3, 4). Group specific mean ± SD for SUA was calculated. One-way ANOVA was used for comparing the differences between the quintiles. For post hoc comparison, the Scheffe test was applied. In this study, the primary aim was to measure MetS incidence and then to estimate the prevalence of each component of MetS and MetS itself at follow-up with respect to the different quintiles of SUA level at baseline values. This was quantified by odds ratios (ORs) using binary logistic regression. The results were also evaluated after adjusting for several variables such as age, total cholesterol and LDL-cholesterol. We calculated the added discrimination value contributed by SUA level to predict the new onset of MetS, more than the information provided by each component of MetS. We estimated the areas under the receiver operating characteristic (ROC) curve in models which included waist circumference, triglyceride, HDL-cholesterol, blood pressure, glucose level with and without SUA. Additionally, we calculated NRI and IDI to measures the incremental prognostic value of SUA as a new marker when added in the traditional existing model. All analyses were performed using SAS, version 9.2 (SAS Institute, Cary, NC). A P value <0.05 was considered statistically significant. 3. Results After a follow-up period of 2.6 years, 113 men (17.09%) and 148 women (15.93%) (Fig. 1) developed MetS. The baseline characteristics of subjects (n ¼ 1590) stratified by gender in relation to the development of MetS and concentration of SUA in relation to number of metabolic parameters are shown in Table 1. Baseline WC, BMI, TG, LDL-C, FBG, and SBP were significantly higher and HDLcholesterol was significantly lower in men and women who developed MetS compared with non-MetS (Table 1). Women were significantly older in MetS group compared with non-MetS while no changes were observed in men. Alcohol drinking and smoking
3
status were significantly different between the MetS group and the non-MetS group in men, but not in women. Table 1 also shows the relationship between numbers of MetS disorders used to diagnosis MetS and increasing concentration of SUA. The mean concentration of SUA was significantly higher with increasing MetS components in both men and women. The mean concentration of SUA was 6.1 mg/dl and 4.6 mg/dl in men and women with 4 MetS components. To further evaluate the relationship between individual components of MetS and SUA level, the baseline data were categorized into quintiles from the lowest to the highest range of SUA (SUA-1 to SUA-V) are shown in Table 2. The highest quintile of SUA was significantly associated with elevated WC, BMI, TC, TG, LDL-C, and SBP in men and women. However, HDL-C and FBG were not found to be significant in different quintiles of SUA. A correlation matrix Online Supplementary Appendix 1 showed the linear relationships between SUA at baseline and each MetS component except for HDL-C. WC, BP, FBG, and TG were positively correlated with baseline SUA levels. A negative correlation was seen with HDL-C. Table 3 presents the prevalence as ORs (95% CI) for each component of MetS and MetS itself at follow-up in the different SUA quintiles in men. Compared with the first quintile of SUA, participants in the fifth quintile of SUA had significantly higher ORs for WC, BP, TG and 2.02 times higher ORs (95% CI 1.11e3.68; P < 0.05) for developing MetS. The ORs was slightly attenuated after adjustment for age, TC, and LDL-C for developing MetS in men. Associations of SUA quintiles for individual MetS components at follow-up in women are represented in Table 4. We observed a trend in increments of WC and TG variables in SUA from the lowest to the highest quintiles with ORs [3.0(1.9e4.73), 3.33(1.84e6.02) P < 0.05]. The ORs of WC and TG in women were significantly increased from the third to fifth quintiles of SUA and thus could be an effective variable for predicting the development of MetS. Greater magnitude of SUA association with the development of MetS was shown in women as compared to men. Compared to the women of SUA first quintile, participants in the SUA fifth quintile had significantly higher ORs for having abnormal WC, BP, TG and 2.2 times higher ORs for developing MetS, except for FG and HDL-C. Fruits contain high levels of fructose, so we included the baseline data on the consumption of different fruits (yes or no) in the studied subjects stratified by gender presented in online Supplementary Appendix 2. We assessed the baseline SUA levels to predict the new onset of MetS beyond the information given by baseline range of individual components of MetS (Table 5). The area under the ROC curve to predict future development of MetS by employing WC, TG, HDLCholesterol, blood pressure and blood glucose was 0.725 (95% CI 0.676e0.774) in men and 0.735 (0.694e0.776) in women. After, SUA were added to the model, the subsequent areas under the curve were 0.73 (0.68e0.78) and 0.75 (0.711e0.79) respectively. The comparison in areas under the ROC curve was represented by P values of 0.442 and 0.014 in men and women, respectively. The category free NRI was 0.10 (95% CI: 0.09e0.30, P ¼ 0.30), and the IDI was 0.006 (95% CI: 0.002e0.014, P ¼ 0.12) for men. For women, the category free NRI was 0.24 (95% CI: 0.06e0.41, P ¼ 0.007) and the IDI was 0.004 (95% CI: 0.002e0.01, P ¼ 0.20). Thus, addition of SUA to the model correctly reclassified 9% more cases in men and 23% more cases in women respectively. 4. Discussion This prospective follow-up cohort study in a Korean adult population showed that a higher level of SUA was a positive predictor for MetS incidence both in men and women, independent of age, total cholesterol, and low-density lipoprotein. Additionally, our
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
4
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
Table 1 Baseline characteristics of study subjects stratified for the absence and presence of MetS by gender and evaluation of SUA in relation to the number of metabolic syndrome components at follow-up. Men
Number of subjects Age (years) WC (cm) BMI (kg/m2) Total cholesterol (mg/dl) TG (mg/dl) HDL-C (mg/dl) LDL-C (mg/dl) FBG (mg/dl) SBP (mm Hg) DBP (mm Hg) SUA (mg/dl) Regular exercise (%) Alcohol drinking (%) Current smoking (%) Number of MetS components 0 1 2 3 4
Women
MetS
Non-MetS
p-value
113 56.02 ± 7.8 87.1 ± 6.4 24.7 ± 2.3 200.1 ± 35.6 151.4 ± 86.8 45.8 ± 9 120.6 ± 32.2 97.3 ± 17.5 129.6 ± 17.7 83.7 ± 10.8 5.9 ± 1.33 24.7 81.4 76.7 SUA 5.397 ± 1.144 5.554 ± 1.245 5.603 ± 1.108 5.862 ± 1.279 6.125 ± 1.488
578 56.6 ± 8.1 82.7 ± 6.7 23.01 ± 2.6 195.4 ± 34.4 122.8 ± 70.5 50.1 ± 11.7 113 ± 30.4 92.9 ± 16.7 124.9 ± 15.8 82.3 ± 10.7 5.5 ± 1.1 22.06 70.9 67.2
0.433 <0.001 <0.001 0.188 0.0013 <0.001 0.0176 0.0111 0.0054 0.221 0.0011 0.528 0.023 0.046 <0.0001
MetS
Non-MetS
p-value
148 55.0 ± 7.7 82.7 ± 7.2 25.4 ± 2.8 207.5 ± 36.3 127.5 ± 64.8 47.5 ± 10.5 125.9 ± 29.1 91.8 ± 20.1 128.4 ± 18.3 82.2 ± 11.2 4.31 ± 0.94 22.9 29.2 4.05 SUA 3.772 ± 0.836 4.06 ± 0.884 4.223 ± 0.933 4.22 ± 0.898 4.63 ± 1.036
781 52.5 ± 8.01 77.6 ± 7.4 23.5 ± 2.8 199 ± 37.5 99.5 ± 48.8 51.4 ± 10.6 117.3 ± 31.4 87.7 ± 9.6 120.2 ± 16.2 78.6 ± 11.6 4.04 ± 0.90 29.09 26.32 1.93
0.001 <0.001 <0.001 0.0115 <0.001 <0.001 0.0020 0.015 <0.001 0.001 0.0012 0.129 0.460 0.128 <0.0001
Data are expressed as mean ± standard deviation. MetS, Metabolic syndrome; WC, Waist circumference; BMI, Body mass index, HDL-C, High density lipoprotein cholesterol; LDL-C, Low density lipoprotein cholesterol; VLDL, Very low density lipoprotein; TG, Triglyceride, FBG, Fasting blood glucose; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; SUA, Serum uric acid.
Table 2 Comparison of demographic and biochemical variables at baseline in different quintiles of serum uric acid in men and women. Men
UA-I
UA-II
(n ¼ 130) Age WC (cm) BMI (kg/m2) Total cholesterol (mg/dl) TG (mg/dl) HDL-C (mg/dl) LDL-C (mg/dl) FBG (mg/dl) SBP (mmHg) DBP (mmHg) Uric acid (mg/dl) Women
Age Waist circumference (cm) Body mass index (kg/m2) Total cholesterol (mg/dl) TG (mg/dl) HDL-C (mg/dl) LDL-C (mg/dl) FBG (mg/dl) SBP (mmHg) DBP (mmHg) Uric acid (mg/dl)
57.5 81.2 22.5 190.4 117.7 50.1 108.5 95.6 125.2 81.6 3.9
± ± ± ± ± ± ± ± ± ± ±
8.2 6.3 2.5 33.7 67.3 11.2 29.8 24.4 15.9 11.2 0.57
UA-III
(n ¼ 110) 56.08 82.5 22.9 190.7 108.8 50.2 110.7 92.9 124 82.4 4.86
± ± ± ± ± ± ± ± ± ± ±
7.9 5.69 1.8 30 56.2 11.3 26.7 16.2 14.2 8.4 0.177***
UA-IV
(n ¼ 133) 57.3 83.3 23.3 197.8 126.8 49.6 115.6 93.9 129.6 84.1 5.45
± ± ± ± ± ± ± ± ± ± ±
7.8 6.85 2.61 34.0 79.3 11.9 31.0 17.0 18.1 12.8 0.16***
UA-V
(n ¼ 121) 55.1 83.4 23.3 194.6 128.2 48.4 114.4 90.5 122.6 82.4 6.03
± ± ± ± ± ± ± ± ± ± ±
7.8 6.9 2.8 37.7 76.9 12.8 30.2 10.6 15.7 10.6 0.18***
UA-1
UA-II
UA-III
UA-IV
(n ¼ 194)
(n ¼ 191)
(n ¼ 181)
(n ¼ 186)
53.06 76.8 23.1 194.3 97.7 50.6 113.2 88.6 119.8 78.2 2.9
± ± ± ± ± ± ± ± ± ± ±
8.2 7.8 2.8 36.0 47.1 10.0 30.4 15.9 15.6 10.3 0.40
52.3 77.06 23.4 195.4 96.0 51.3 114.6 89.03 122.1 80.1 3.6
± ± ± ± ± ± ± ± ± ± ±
8.2 7.6 2.7 33.7 40.9 10.7 27.9 11.5 18.2 12.7 0.13***
51.7 79.0 24.2 199.4 103.4 50.9 116.7 88.0 120.3 78.1 4.0
± ± ± ± ± ± ± ± ± ± ±
8.09 7.2 2.7** 36.0 54.9 11.2 29.0 9.5 16.6 11.7 0.10***
52.8 79.2 24.1 201.6 108.9 50.7 120.1 87.2 119.9 78.3 4.5
(n ¼ 167) 56.5 85.8 24.1 203.8 148 48.7 119.9 94.6 126.2 82.2 7.1
± ± ± ± ± ± ± ± ± ± ±
8.3 7.2*** 2.7*** 35.1* 79.6* 10.1 33.7* 13.2 16.1 10.2 0.74***
P value (#) 0.105 <0.001 <0.001 0.004 0.001 0.638 0.0189 0.142 0.0086 0.442 <0.001
UA-V
± ± ± ± ± ± ± ± ± ± ±
7.6 7.7* 2.9* 33.7 61.1 11.3 27.9 11.4 15.3 11.6 0.16***
(n ¼ 177) 54.5 80.3 24.5 211.7 114.7 50.4 129.3 89.1 125.3 81.3 5.4
± ± ± ± ± ± ± ± ± ± ±
7.5 7.1*** 2.9*** 44.7*** 56.0** 10.09 37.6*** 10.3 17.9** 11.5 0.56***
P value (#) 0.0135 0.001 <0.001 <0.001 0.002 0.945 <0.001 0.542 0.0081 0.026 <0.001
Data are expressed as mean ± standard deviation. WC, Waist circumference; BMI, Body mass index, HDL-C, High density lipoprotein cholesterol; LDL-C, Low density lipoprotein cholesterol; TG, Triglyceride; FBG, Fasting blood glucose; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; SUA, Serum uric acid. ***P < 0.001 compared with UA-I, **P < 0.01 compared with UA-I, *P < 0.05 compared with UA-I, # Difference between 5 groups (UA-I-UA-V).
analysis reported that in women, SUA increases the predictive ability for identification of future MetS subjects beyond the information given by the MetS components and may serve as a prognostic tool in women. SUA is an end product of purine metabolism in human and higher primates. The uric acid is formed in the liver by the breakdown of nucleic acid and protein. It has already been known and
evaluated that SUA is a potential risk factor for the development of MetS [5,11,20,21]. However, the strength of association was quite different in relation to sex, age, and ethnicity. The possible biological mechanism related to hyperuricemia and the future development of MetS has been studied in both the cell lines and animal model. A previous study reported that an increase in SUA has an effective role in the development of insulin resistance through the
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
5
Table 3 Odds ratio and 95% CI of each component of MetS at follow-up according to varying concentration of SUA in men.
Men WC (cm) Prevalence (%) OR (95%CI) BP Prevalence (%) OR (95%CI) HDL-C Prevalence (%) OR (95%CI) TG Prevalence (%) OR (95%CI) FBG Prevalence (%) OR (95%CI) MetS Incidence (%) OR (95%CI) Adjusted OR d (95%CI)
SUA-I
SUA-II
SUA-III
SUA-IV
SUA-V
(n ¼ 130)
(n ¼ 110)
(n ¼ 133)
(n ¼ 121)
(n ¼ 167)
15.38% 1 (Ref)
21.8% 1.53(0.79e2.96)
21.05% 1.46(0.77e2.76)
18.18% 1.22(0.629e2.37)
33.5% 2.77*(1.56e4.92)
31.5% 1 (Ref)
30.9% 0.971(0.561e1.68)
37.59% 1.30(0.785e2.17)
25.62% 0.748(0.431e1.29)
43.71% 1.68*(1.043e2.724)
29.2% 1 (Ref)
23.46% 0.749(0.420e1.33)
26.32% 0.865(0.504e1.48)
29.75% 1.02(0.596e1.76)
23.95% 0.763(0.453e1.28)
25.38% 1 (Ref)
21.82% 0.82(0.450e1.49)
29.32% 1.22(0.708e2.1)
26.45% 1.05(0.601e1.85)
36.53% 1.69*(1.021e2.80)
25.38% 1 (Ref)
23.64% 0.910(0.504e1.64)
25.56% 1.0(0.58e1.75)
15.70% 0.548(0.292e1.027)
28.14% 1.15(0.685e1.935)
14.62% 1 (Ref) 1 (Ref)
12.73% 0.852(0.406e1.79) 0.804(0.381e1.69)
18.05% 1.28(0.667e2.48) 1.22(0.630e2.37)
10.74% 0.703(0.331e1.494) 0.630(0.294e1.35)
25.75% 2.02*(1.11e3.68) 1.89*(1.03e3.47)
SUA, Serum uric acid; WC, Waist circumference; BP, Blood pressure; HDL-C, High density lipoprotein, TG, Triglyceride; FBG, Fasting blood glucose; MetS, Metabolic syndrome. * P < 0.05 compared with reference gp (SUA-I). d Adjusted for age, total cholesterol and low density lipoprotein cholesterol.
Table 4 Odds ratio and 95% CI of each component of MetS at follow-up according to varying concentration of SUA in women.
Women WC (cm) Prevalence (%) OR (95%CI) BP Prevalence (%) OR (95%CI) HDL-C Prevalence (%) OR (95%CI) TG Prevalence (%) OR (95%CI) FG Prevalence (%) OR (95%CI) MetS Incidence (%) OR (95%CI) Adjusted OR d (95%Cl)
SUA-I
SUA-II
SUA-III
SUA-IV
SUA-V
(n ¼ 194)
(n ¼ 191)
(n ¼ 181)
(n ¼ 186)
(n ¼ 177)
21.13% 1 (Ref)
22.51% 1.08(0.668e1.75)
33.15% 1.85*(1.164e2.94)
37.63% 2.25*(1.42e3.54)
44.63% 3.0*(1.9e4.73)
31.44% 1 (Ref)
26.18% 0.773(0.497e1.20)
26.52% 0.787(0.503e1.23)
29.57% 0.915(0.591e1.41)
41.81% 1.56*(1.02e2.39)
48.45% 1 (Ref)
51.31% 1.21(0.752e1.67)
54.70% 1.28(0.856e1.92)
54.84% 1.292(0.863e1.93)
53.67% 1.23 (0.819e1.85)
9.28% 1 (Ref)
15.7% 1.82(0.978e3.39)
16.57% 1.94*(1.04e3.62)
20.97% 2.59*(1.42e4.72)
25.42% 3.33*(1.84e6.02)
12.37% 1 (Ref)
10.47% 0.828(0.441e1.55)
9.94% 0.782(0.409e1.49)
11.29% 0.902(0.483e1.68)
14.69% 1.22(0.672e2.215)
12.37% 1 (Ref) 1 (Ref)
11.52% 0.922(0.498e1.70) 0.935(0.503e1.74)
15.47% 1.29(0.720e2.33) 1.33(0.739e2.42)
17.2% 1.47(0.830e2.609) 1.44(0.813e2.58)
23.73% 2.20*(1.27e3.820) 1.93* (1.105e3.400)
SUA, Serum uric acid; WC, Waist circumference; BP, Blood pressure; HDL-C, High density lipoprotein; TG, Triglyceride; FG, Fasting glucose; MetS, Metabolic syndrome.*P < 0.05 compared with the reference group. d Adjusted for age, total cholesterol and low density lipoprotein cholesterol.
inhibition of endothelial nitric oxide synthase [22]. Furthermore, Cook et al. observed the features of MetS in mice lacking endothelial nitric oxide synthase [23]. In addition, SUA stimulates inflammatory and oxidative changes in adipocytes, vascular smooth muscle cells, and human mononuclear cells [24,25]. The most significant proof of SUA as a predictor of MetS was furnished by the experimental studies of Nakagawa et al. in 2006 [26] in fructose-fed rat model (animal model of MetS) that showed a reduction in SUA level by allopurinol (inhibitor of xanthine oxidase) and reverses the feature of MetS. Despite overwhelming evidence of the association between SUA and insulin resistance (root cause of MetS), there are only a few data on the predictive value of SUA for new onset of MetS. However,
those studies did not have concerns about the information related to the additional predictive ability of SUA over and above that provided by different components of MetS at baseline. Our study reported a graded increase in the incidence of MetS among women with the increasing levels of SUA quintiles, whereas this was not noticeable in men. Sex-related variations of SUA in diagnosing MetS have also been reported. Also, several reports were underlined a stronger association between SUA level and MetS in females than in males [10e12]. The mechanisms underlying this association are not still clear. However, few studies have been shown that SUA level is independently associated with serum leptin level and advocated that leptin could be a risk factor responsible for hyperuricemia in metabolic syndrome [27]. Leptin directly impairs uric acid
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
6
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
Table 5 Comparison of area under the ROC curve of each variable of metabolic syndrome and additional SUA model. Each components AUC (95% CI)
SUA additional model AUC (95% CI)
p-value
Men High waist circumference High triglyceride Low HDL cholesterol High blood pressure High glucose 5 components
0.599 0.548 0.522 0.533 0.555 0.725
(0.554e0.644) (0.501e0.595) (0.483e0.561) (0.484e0.583) (0.512e0.598) (0.676e0.774)
0.647 0.591 0.584 0.594 0.613 0.73
(0.589e0.706) (0.531e0.651) (0.524e0.644) (0.535e0.653) (0.553e0.672) (0.68e0.78)
0.031 0.1027 0.0421 0.0658 0.0246 0.4422
Women High waist circumference High triglyceride Low HDL cholesterol High blood pressure High glucose 5 components
0.583 0.551 0.538 0.577 0.528 0.735
(0.544e0.623) (0.518e0.585) (0.494e0.582) (0.533e0.62) (0.504e0.553) (0.694e0.776)
0.649 0.606 0.584 0.615 0.617 0.75
(0.6e0.697) (0.554e0.658) (0.534e0.634) (0.567e0.664) (0.566e0.667) (0.711e0.79)
0.0004 0.0111 0.0583 0.0147 0.0002 0.0144
AUC, area under the ROC curve; 95% CI, 95% confidence interval, P value is the comparison between each components of metabolic syndrome and additional SUA level.
excretion in the kidney [28]. The sexual differences in the SUA result for incident MetS in women could also be explained by hormonal effect. Postmenopausal women have higher uric acid level due to the effect of estrogens, which is uricosuric in nature and it favors uric acid excretion [29]. Baseline data of serum leptin concentration (men & women) and menopausal status in women according to the different quintiles of serum uric acid (SUA) are presented in online Supplementary Appendix 3. The ORs for development of MetS were 2.02 fold higher in men and 2.2 fold higher in women in the highest quintile of SUA levels compared with the lowest reference quintile. The results are consistent with the finding by Rye et al. 2007 [30]. Interestingly, large epidemiological studies have observed a positive relationship between elevated SUA and the growing prevalence of BP, FG, TG and WC in adolescent and adults [7e9,31e33]. In the present study, we also observed a significant association between higher SUA quintile and elevated WC, BP and TG (Tables 3 and 4). In women, we found an improvement in the area under the ROC curve in the model that added SUA to MetS components while the improvement was not seen in men. We calculated the category free NRI and the IDI to measure the predictive ability of SUA as a new biomarker after adding SUA to model with or without incident MetS. Pencina et al. suggested that NRI and IDI are more sensitive than the area under the ROC curve for the determination of improvement in the predictive value [34]. In our study, the addition of SUA significantly improved the NRI in women. The IDI measuring the overall sensitivity did not improve in both genders. However, this data specify that SUA may have a clinical role in the screening for metabolic risk in women but needed to clarify with large sample size stratified by gender. There are some limitations to the present study. Our studied subjects were restricted to middle age (40e70 years) and elderly Koreans residing in a rural area with proportionately high baseline prevalence and the high incidence of MetS [35]. The study did not have information regarding some important determinants of SUA such as dietary factors, including intake of purine-rich foods, fructose consumption which could contribute to hyperuricemia. We may have some limit to generalize our findings as our results were based on a sample of relatively healthy middle-aged Koreans. The follow-up period of our cohort was 2.6 years and about 25% of the sample did not complete the follow-up visit and thus we could not assess the presence of cardiometabolic abnormalities in a large sample. Finally, our study was based on a single determination of SUA that may lead to random measurement error. In conclusion, we found that SUA was significantly correlated with future risk of WC, BP, TG and it can be used in predicting the
development of MetS. The present study provides additional evidence that SUA has an important role in predicting new-onset of MetS among women. Further large prospective follow-up studies are needed to corroborate these findings. Authors contributions The study was conceived and designed by DY, CHC. Data collection in cohort study was carried out by SVA and SBK. Statistical analysis was done by EHC. Interpretation of data analysis was done by DY, CHC, and EYL. ESL, HMK, JSL, EYL helped in editing the manuscript. DY wrote the manuscript and final copy of article was approved by all authors. CHC supervised the whole project. Acknowledgments This research was supported in part by a fund (2005-E71013-00, 2006-E71002-00, 2007-E71013-00, 2008-E71004-00, 2009E71006-00, 2010-E71003-00) for research by the Korea Centers for Disease Control and Prevention. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atherosclerosis.2015.04.797. Conflict of interest There is no conflict of interest between the authors for this study. References [1] R.H. Eckel, S.M. Grundy, P.Z. Zimmet, The metabolic syndrome, Lancet 365 (2005) 1415e1428. [2] S. Mottillo, K.B. Filion, J. Genest, et al., The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis, J. Am. Coll. Cardiol. 56 (2010) 1113e1132. [3] N. Sattar, A. McConnachie, A.G. Shaper, et al., Can metabolic syndrome usefully predict cardiovascular disease and diabetes? outcome data from two prospective studies, Lancet 371 (2008) 1927e1935. [4] S. Lim, H. Shin, J.H. Song, et al., Increasing prevalence of metabolic syndrome in Korea: the Korean National Health and Nutrition Examination Survey for 1998-2007, Diabetes Care 34 (2011) 1323e1328. [5] X. Sui, T.S. Church, R.A. Meriwether, et al., Uric acid and the development of metabolic syndrome in women and men, Metabolism 57 (2008) 845e852. [6] P.W. Liu, T.Y. Chang, J.D. Chen, Serum uric acid and metabolic syndrome in Taiwanese adults, Metabolism 59 (2010) 802e807. [7] V. Lohsoonthorn, B. Dhanamun, M.A. Williams, Prevalence of hyperuricemia and its relationship with metabolic syndrome in Thai adults receiving annual
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7 health exams, Arch. Med. Res. 37 (2006) 883e889. [8] A. Onat, H. Uyarel, G. Hergenc, et al., Serum uric acid is a determinant of metabolic syndrome in a population-based study, Am. J. Hypertens. 19 (2006) 1055e1062. [9] H.K. Choi, E.S. Ford, Prevalence of the metabolic syndrome in individuals with hyperuricemia, Am. J. Med. 120 (2007) 442e447. [10] W.K. Chiou, M.H. Wang, D.H. Huang, et al., The relationship between serum uric acid level and metabolic syndrome: differences by sex and age in Taiwanese, J. Epidemiol. 20 (2010) 219e224. [11] T. Yang, C.H. Chu, C.H. Bai, et al., Uric acid level as a risk marker for metabolic syndrome: a Chinese cohort study, Atherosclerosis 220 (2012) 525e531. [12] M.L. Zhang, Y.X. Gao, X. Wang, et al., Serum uric acid and appropriate cutoff value for prediction of metabolic syndrome among Chinese adults, J. Clin. Biochem. Nutr. 52 (2013) 38e42. [13] L.K. Niskanen, D.E. Laaksonen, K. Nyyssonen, et al., Uric acid level as a risk factor for cardiovascular and all-cause mortality in middle-aged men: a prospective cohort study, Arch. Intern Med. 164 (2004) 1546e1551. [14] S.B. Koh, J.K. Park, J.H. Yoon, et al., Preliminary report: a serious link between adiponectin levels and metabolic syndrome in a Korean nondiabetic population, Metabolism 59 (2010) 333e337. [15] S.B. Koh, J. Yoon, J.Y. Kim, et al., Relationships between serum adiponectin with metabolic syndrome and components of metabolic syndrome in nondiabetic Koreans: ARIRANG study, Yonsei Med. J. 52 (2011) 234e241. [16] J.Y. Kim, S.V. Ahn, J.H. Yoon, et al., Prospective study of serum adiponectin and incident metabolic syndrome: the ARIRANG study, Diabetes Care 36 (2013) 1547e1553. [17] K. Tsukimori, T. Yoshitomi, S. Morokuma, et al., Serum uric acid levels correlate with plasma hydrogen peroxide and protein carbonyl levels in preeclampsia, Am. J. Hypertens. 21 (2008) 1343e1346. [18] K.G. Alberti, R.H. Eckel, S.M. 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 120 (2009) 1640e1645. [19] S.Y. Lee, H.S. Park, D.J. Kim, et al., Appropriate waist circumference cutoff points for central obesity in Korean adults, Diabetes Res. Clin. Pract. 75 (2007) 72e80. [20] J.P. Goncalves, A. Oliveira, M. Severo, et al., Cross-sectional and longitudinal associations between serum uric acid and metabolic syndrome, Endocrine 41 (2012) 450e457. [21] S. Hara, H. Tsuji, Y. Ohmoto, et al., High serum uric acid level and low urine pH as predictors of metabolic syndrome: a retrospective cohort study in a
7
Japanese urban population, Metabolism 61 (2012) 281e288. [22] T. Nakagawa, D.H. Kang, D. Feig, et al., Unearthing uric acid: an ancient factor with recently found significance in renal and cardiovascular disease, Kidney Int. 69 (2006) 1722e1725. [23] S. Cook, O. Hugli, M. Egli, et al., Clustering of cardiovascular risk factors mimicking the human metabolic syndrome X in eNOS null mice, Swiss Med. Wkly. 133 (2003) 360e363. [24] W. Baldwin, S. McRae, G. Marek, et al., Hyperuricemia as a mediator of the proinflammatory endocrine imbalance in the adipose tissue in a murine model of the metabolic syndrome, Diabetes 60 (2011) 1258e1269. [25] S. Furukawa, T. Fujita, M. Shimabukuro, et al., Increased oxidative stress in obesity and its impact on metabolic syndrome, J. Clin. Invest 114 (2004) 1752e1761. [26] T. Nakagawa, H. Hu, S. Zharikov, et al., A causal role for uric acid in fructoseinduced metabolic syndrome, Am. J. Physiol. Ren. Physiol. 290 (2006) F625eF631. [27] A. Bedir, M. Topbas, F. Tanyeri, et al., Leptin might be a regulator of serum uric acid concentrations in humans, Jpn. Heart J. 44 (2003) 527e536. [28] B. Fruehwald-Schultes, A. Peters, W. Kern, et al., Serum leptin is associated with serum uric acid concentrations in humans, Metabolism 48 (1999) 677e680. [29] A. Nicholls, M.L. Snaith, J.T. Scott, Effect of oestrogen therapy on plasma and urinary levels of uric acid, Br. Med. J. 1 (1973) 449e451. [30] S. Ryu, J. Song, B.Y. Choi, et al., Incidence and risk factors for metabolic syndrome in Korean male workers, ages 30 to 39, Ann. Epidemiol. 17 (2007) 245e252. [31] J.Y. Wang, Y.L. Chen, C.H. Hsu, et al., Predictive value of serum uric acid levels for the diagnosis of metabolic syndrome in adolescents, J. Pediatr. 161 (2012) 753e756 e2. [32] Y. Taniguchi, T. Hayashi, K. Tsumura, et al., Serum uric acid and the risk for hypertension and type 2 diabetes in Japanese men: the Osaka Health Survey, J. Hypertens. 19 (2001) 1209e1215. [33] S. Ryu, Y. Chang, Y. Zhang, et al., A cohort study of hyperuricemia in middleaged South Korean men, Am. J. Epidemiol. 175 (2012) 133e143. [34] M.J. Pencina, R.B. D'Agostino Sr., R.B. D'Agostino Jr., et al., Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond, Stat. Med. 27 (2008) 157e172 discussion 207e12. [35] Kyounghoon Lee Sbk, Myungsang Shin, Minsoo Ahn, Jang-Young Kim, ByungSu Yoo, Seung-Hwan Lee, Junghan Yoon, Tae Yong Lee, So Yeon Ryu, Jong Ku Park, Kyung-Hoon Choe, Prevalence and sex-related characteristics of metabolic syndrome in a Korean rural cohort, Soc. Lipidol. Atheroscler. J. 18 (2008) 226e238.
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort Dhananjay Yadav a, d, Eun Soo Lee a, Hong Min Kim a, Eunhee Choi b, Eun Young Lee c, Jung Soo Lim a, Song Vogue Ahn d, e, Sang Baek Koh d, e, Choon Hee Chung a, b, * a
Department of Internal Medicine, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea Institute of Lifestyle Medicine, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea Department of Internal Medicine, Soonchunhyang University Cheonan Hospital, Cheonan, Republic of Korea d Department of Preventive Medicine, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea e Institute of Genomic Cohort, Yonsei University Wonju College of Medicine, Wonju, Republic of Korea b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 10 November 2014 Received in revised form 30 March 2015 Accepted 21 April 2015 Available online xxx
Objective: Recent studies have demonstrated an association between serum uric acid (SUA) levels and metabolic syndrome (MetS). However, paucity of available data regarding the cause and effect relationship between SUA and MetS in healthy adults is still a big challenge which remains to be studied. Therefore, we investigated whether SUA predicts new onset of MetS in a population-based cohort study. Methods: The study included 1590 adults (661 men and 929 women) aged 40e70 years without MetS at baseline (2005e2008) and subjects were prospectively followed for 2.6 years. To evaluate the relationship between SUA and MetS, we divided the aforementioned subjects into quintiles (SUA-I to SUA-V) from the lowest to the highest values of SUA. SUA was measured by the enzymatic colorimetric method. We used category-free net reclassification improvement (NRI) and integrated discrimination improvement (IDI) to characterize the performance of predicted model. Results: During a mean of 2.6 years of follow-up, 261(16.4%) adults developed MetS. MetS variables were significantly related to the baseline SUA level. Waist circumference (WC), blood pressure (BP), and serum triglyceride (TG) were significantly higher in the highest quintile of SUA compared to the lowest SUA quintile in men and women. After adjustment for age, total cholesterol and low-density lipoprotein cholesterol (LDL-C) in men and women, subjects in the fifth quintiles of SUA showed significantly higher ORs for incident MetS. The association between hyperuricemia and new onset of MetS were consistently stronger in women than men. Additionally, among women, we found an improvement in the area under the ROC curve in the models that added SUA to core components of MetS. Conclusion: Our study suggests that SUA is significantly correlated with future risk of WC, BP, TG and may predicted as a risk factor for developing MetS. SUA may have a clinical role in predicting new-onset metabolic syndrome among women. Large prospective study is needed to reveal the clinical significance of SUA in metabolic disease. © 2015 Published by Elsevier Ireland Ltd.
Keywords: Serum uric acid Metabolic syndrome Hyperuricemia Korean adults
1. Introduction Metabolic syndrome (MetS) is characterized by the grouping of cardiovascular risk factors, including high blood pressure (BP), abdominal obesity, dyslipidemia, and increased glucose
* Corresponding author. Department of Internal Medicine, Yonsei University Wonju College of Medicine, 162 Ilsan-Dong, Wonju, Kangwon-Do 220-701, Republic of Korea. E-mail address: [email protected] (C.H. Chung).
concentration [1]. Subjects with MetS are linked with an increased risk of developing type 2 diabetes and cardiovascular disease (CVD) [2,3]. The prevalence of MetS in Korea has been immensely increased from 24.9% in 1998 to 31.3% in 2007 due to rapid lifestyle changes [4]. Apart from the dietary habit and variables of MetS, other factors have also played a significant role in higher prevalence and incidence of developing MetS. Several epidemiological studies have established the link between hyperuricemia and the development of MetS in various populations which may relate to the burden of
http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797 0021-9150/© 2015 Published by Elsevier Ireland Ltd.
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
2
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
diabetes and CVD [5e9]. Furthermore, studies evaluating a sexrelated association between SUA and development of MetS have also been highly intensified [5,10,11]. Recently, the association between serum uric acid (SUA) and MetS was monitored not only in hyperuricemia, but also in the normal range of SUA [12,13]. Nevertheless, very limited information is available about its ability as a predictor of MetS and its components and additional clinical utility in epidemiological perspective. Hence, we designed a follow-up study to prospectively examine the association between SUA and the incidence of MetS and its components as well as the predictive ability of SUA in identifying people who will develop new onset of MetS stratified by sex in Korean rural cohort. 2. Methods 2.1. Study participants The Korean Genome and Epidemiology Study on Atherosclerosis Risk of Rural Areas in the Korean General Population (KoGESARIRANG), a population-based prospective cohort study designed to estimate the prevalence, incidence, and risk factors for chronic degenerative disorders such as diabetes, hypertension, osteoporosis, and cardiovascular disease [14e16]. The adults aged between 40 and 70 years residing in the rural areas of Wonju and Pyeongchang in South Korea participated in the study. The baseline study was executed from November 2005 to January 2008, comprised of 5178 adults (2127 men and 3051women). 3862 (74.6%) subjects were attended during the first follow-up survey (April 2008eJanuary 2011). We then excluded 1020 subjects with MetS at baseline, 1218 subjects without baseline SUA measurement, 20 subjects with a history of cardiovascular disease at baseline and 14 subjects without complete data. In total, 1590 participants were included in the present analysis (661 men and 929 women) (Fig. 1). The study protocol was approved by the institutional review board of Yonsei University College of Medicine. Informed consent was obtained from each participant in the study. 2.2. Data collection and measurements At study entry, participants completed the medical history and
lifestyle questionnaire and went through a standardized health examination according to the optimum procedures. For anthropometrical measurements, body weight and height were measured, with the participants barefooted and wearing light indoor clothing. Waist circumference (WC) was measured using a tape (SECA-200; SECA, Hamburg, Germany) in a horizontal plane midway between the inferior margin of the ribs and the superior border of the iliac crest. BP was measured with a standard mercury sphygmomanometer (Baumanometer, Copiague, NY) twice on the right arm of the participants. The mean of the two BP readings was used for the data analyses. Baseline information on current smoking status and alcohol intake was collected by self-reported questionnaire. Subjects who answered “yes” to the question “Do you perform physical exercise regularly enough to make you sweat?” were designated as the regular exercise group. Venous blood samples were collected from all participants after fasting for >12 h or overnight. Fasting blood glucose (FBG) was measured by a glucose oxidaseebased assay. Serum concentrations of high-density lipoprotein cholesterol (HDL-C) and triglyceride (TG) were determined by enzymatic methods (Advia 1650; Siemens, Tarrytown, NY). Low-density lipoprotein (LDL)-cholesterol was calculated using Friedewald's formula: LDLcholesterol ¼ Total cholesterol e (HDL-C þ TG/5). SUA was measured by the enzymatic colorimetric method. In this method, SUA is oxidized by uricase to allantoin and hydrogen peroxide based on the uricase-peroxidase method. In the following reactions, the oxidative condensation of N-ethyl-N-(2-hydroxy-3sulfopropyl)-3-methylaniline and 4-aminophenazone produces a red chromogen in the presence of peroxidase and hydrogen peroxide [17]. An automated analyzer was used to measure the intensity of the color produced in the sample and the values were reported in mg/dL. 2.3. Definition of MetS MetS was defined according to the criteria of the harmonized definition [18] for MetS, which was the end-point of this study at the follow-up visit. This definitions includes the presence of at least three out of five metabolic abnormalities: 1) abdominal obesity, defined as a waist circumference 90 cm for men or 85 cm for
Fig. 1. Descrpition of the study population.
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
women (following Korean specific cutoffs for abdominal obesity defined by the Korean Society of Obesity) [19]; 2) hypertriglyceridemia, defined as a serum TG concentration 150 mg/dL (1.69 mmol/L); 3) high blood pressure, defined as a systolic blood pressure (SBP) 130 mmHg, a diastolic blood pressure (DBP) 85 mmHg, or treatment with antihypertensive agents; 4) low serum HDL cholesterol, defined as a serum HDL cholesterol concentration < 40 mg/dL (1.04 mmol/L) for men or < 50 mg/dL (1.29 mmol/L) for women; and 5) high fasting glucose defined as a fasting serum glucose 100 mg/dL or previously diagnosed type 2 diabetes. 2.4. Quintile of SUA stratified by gender To evaluate the association between the individual components of MetS and SUA levels, study subjects were divided into quintiles according to their SUA levels (SUA-I to SUA-V, first to fifth quintile of SUA). The quintile measured the dispersion in the optimum level of SUA to predict the future estimated risk of each component of MetS, stratified by men and women. The lowest and highest quintile of SUA in men and women were 0.7e4.5, 6.3e7.0 and 0.6e3.3, 4.8e6.0 mg/dL. 2.5. Statistical analysis We analyzed all of the results separately for men and women. Study data are expressed as means ± standard deviation (SD) or frequency (%). Baseline characteristics between the groups with and without incident MetS were assessed by two-sample t-test or chi-square test. Participants were characterized into five groups accordingly to the components of MetS disorder (presence of MetS variables such as 0, 1, 2, 3, 4). Group specific mean ± SD for SUA was calculated. One-way ANOVA was used for comparing the differences between the quintiles. For post hoc comparison, the Scheffe test was applied. In this study, the primary aim was to measure MetS incidence and then to estimate the prevalence of each component of MetS and MetS itself at follow-up with respect to the different quintiles of SUA level at baseline values. This was quantified by odds ratios (ORs) using binary logistic regression. The results were also evaluated after adjusting for several variables such as age, total cholesterol and LDL-cholesterol. We calculated the added discrimination value contributed by SUA level to predict the new onset of MetS, more than the information provided by each component of MetS. We estimated the areas under the receiver operating characteristic (ROC) curve in models which included waist circumference, triglyceride, HDL-cholesterol, blood pressure, glucose level with and without SUA. Additionally, we calculated NRI and IDI to measures the incremental prognostic value of SUA as a new marker when added in the traditional existing model. All analyses were performed using SAS, version 9.2 (SAS Institute, Cary, NC). A P value <0.05 was considered statistically significant. 3. Results After a follow-up period of 2.6 years, 113 men (17.09%) and 148 women (15.93%) (Fig. 1) developed MetS. The baseline characteristics of subjects (n ¼ 1590) stratified by gender in relation to the development of MetS and concentration of SUA in relation to number of metabolic parameters are shown in Table 1. Baseline WC, BMI, TG, LDL-C, FBG, and SBP were significantly higher and HDLcholesterol was significantly lower in men and women who developed MetS compared with non-MetS (Table 1). Women were significantly older in MetS group compared with non-MetS while no changes were observed in men. Alcohol drinking and smoking
3
status were significantly different between the MetS group and the non-MetS group in men, but not in women. Table 1 also shows the relationship between numbers of MetS disorders used to diagnosis MetS and increasing concentration of SUA. The mean concentration of SUA was significantly higher with increasing MetS components in both men and women. The mean concentration of SUA was 6.1 mg/dl and 4.6 mg/dl in men and women with 4 MetS components. To further evaluate the relationship between individual components of MetS and SUA level, the baseline data were categorized into quintiles from the lowest to the highest range of SUA (SUA-1 to SUA-V) are shown in Table 2. The highest quintile of SUA was significantly associated with elevated WC, BMI, TC, TG, LDL-C, and SBP in men and women. However, HDL-C and FBG were not found to be significant in different quintiles of SUA. A correlation matrix Online Supplementary Appendix 1 showed the linear relationships between SUA at baseline and each MetS component except for HDL-C. WC, BP, FBG, and TG were positively correlated with baseline SUA levels. A negative correlation was seen with HDL-C. Table 3 presents the prevalence as ORs (95% CI) for each component of MetS and MetS itself at follow-up in the different SUA quintiles in men. Compared with the first quintile of SUA, participants in the fifth quintile of SUA had significantly higher ORs for WC, BP, TG and 2.02 times higher ORs (95% CI 1.11e3.68; P < 0.05) for developing MetS. The ORs was slightly attenuated after adjustment for age, TC, and LDL-C for developing MetS in men. Associations of SUA quintiles for individual MetS components at follow-up in women are represented in Table 4. We observed a trend in increments of WC and TG variables in SUA from the lowest to the highest quintiles with ORs [3.0(1.9e4.73), 3.33(1.84e6.02) P < 0.05]. The ORs of WC and TG in women were significantly increased from the third to fifth quintiles of SUA and thus could be an effective variable for predicting the development of MetS. Greater magnitude of SUA association with the development of MetS was shown in women as compared to men. Compared to the women of SUA first quintile, participants in the SUA fifth quintile had significantly higher ORs for having abnormal WC, BP, TG and 2.2 times higher ORs for developing MetS, except for FG and HDL-C. Fruits contain high levels of fructose, so we included the baseline data on the consumption of different fruits (yes or no) in the studied subjects stratified by gender presented in online Supplementary Appendix 2. We assessed the baseline SUA levels to predict the new onset of MetS beyond the information given by baseline range of individual components of MetS (Table 5). The area under the ROC curve to predict future development of MetS by employing WC, TG, HDLCholesterol, blood pressure and blood glucose was 0.725 (95% CI 0.676e0.774) in men and 0.735 (0.694e0.776) in women. After, SUA were added to the model, the subsequent areas under the curve were 0.73 (0.68e0.78) and 0.75 (0.711e0.79) respectively. The comparison in areas under the ROC curve was represented by P values of 0.442 and 0.014 in men and women, respectively. The category free NRI was 0.10 (95% CI: 0.09e0.30, P ¼ 0.30), and the IDI was 0.006 (95% CI: 0.002e0.014, P ¼ 0.12) for men. For women, the category free NRI was 0.24 (95% CI: 0.06e0.41, P ¼ 0.007) and the IDI was 0.004 (95% CI: 0.002e0.01, P ¼ 0.20). Thus, addition of SUA to the model correctly reclassified 9% more cases in men and 23% more cases in women respectively. 4. Discussion This prospective follow-up cohort study in a Korean adult population showed that a higher level of SUA was a positive predictor for MetS incidence both in men and women, independent of age, total cholesterol, and low-density lipoprotein. Additionally, our
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
4
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
Table 1 Baseline characteristics of study subjects stratified for the absence and presence of MetS by gender and evaluation of SUA in relation to the number of metabolic syndrome components at follow-up. Men
Number of subjects Age (years) WC (cm) BMI (kg/m2) Total cholesterol (mg/dl) TG (mg/dl) HDL-C (mg/dl) LDL-C (mg/dl) FBG (mg/dl) SBP (mm Hg) DBP (mm Hg) SUA (mg/dl) Regular exercise (%) Alcohol drinking (%) Current smoking (%) Number of MetS components 0 1 2 3 4
Women
MetS
Non-MetS
p-value
113 56.02 ± 7.8 87.1 ± 6.4 24.7 ± 2.3 200.1 ± 35.6 151.4 ± 86.8 45.8 ± 9 120.6 ± 32.2 97.3 ± 17.5 129.6 ± 17.7 83.7 ± 10.8 5.9 ± 1.33 24.7 81.4 76.7 SUA 5.397 ± 1.144 5.554 ± 1.245 5.603 ± 1.108 5.862 ± 1.279 6.125 ± 1.488
578 56.6 ± 8.1 82.7 ± 6.7 23.01 ± 2.6 195.4 ± 34.4 122.8 ± 70.5 50.1 ± 11.7 113 ± 30.4 92.9 ± 16.7 124.9 ± 15.8 82.3 ± 10.7 5.5 ± 1.1 22.06 70.9 67.2
0.433 <0.001 <0.001 0.188 0.0013 <0.001 0.0176 0.0111 0.0054 0.221 0.0011 0.528 0.023 0.046 <0.0001
MetS
Non-MetS
p-value
148 55.0 ± 7.7 82.7 ± 7.2 25.4 ± 2.8 207.5 ± 36.3 127.5 ± 64.8 47.5 ± 10.5 125.9 ± 29.1 91.8 ± 20.1 128.4 ± 18.3 82.2 ± 11.2 4.31 ± 0.94 22.9 29.2 4.05 SUA 3.772 ± 0.836 4.06 ± 0.884 4.223 ± 0.933 4.22 ± 0.898 4.63 ± 1.036
781 52.5 ± 8.01 77.6 ± 7.4 23.5 ± 2.8 199 ± 37.5 99.5 ± 48.8 51.4 ± 10.6 117.3 ± 31.4 87.7 ± 9.6 120.2 ± 16.2 78.6 ± 11.6 4.04 ± 0.90 29.09 26.32 1.93
0.001 <0.001 <0.001 0.0115 <0.001 <0.001 0.0020 0.015 <0.001 0.001 0.0012 0.129 0.460 0.128 <0.0001
Data are expressed as mean ± standard deviation. MetS, Metabolic syndrome; WC, Waist circumference; BMI, Body mass index, HDL-C, High density lipoprotein cholesterol; LDL-C, Low density lipoprotein cholesterol; VLDL, Very low density lipoprotein; TG, Triglyceride, FBG, Fasting blood glucose; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; SUA, Serum uric acid.
Table 2 Comparison of demographic and biochemical variables at baseline in different quintiles of serum uric acid in men and women. Men
UA-I
UA-II
(n ¼ 130) Age WC (cm) BMI (kg/m2) Total cholesterol (mg/dl) TG (mg/dl) HDL-C (mg/dl) LDL-C (mg/dl) FBG (mg/dl) SBP (mmHg) DBP (mmHg) Uric acid (mg/dl) Women
Age Waist circumference (cm) Body mass index (kg/m2) Total cholesterol (mg/dl) TG (mg/dl) HDL-C (mg/dl) LDL-C (mg/dl) FBG (mg/dl) SBP (mmHg) DBP (mmHg) Uric acid (mg/dl)
57.5 81.2 22.5 190.4 117.7 50.1 108.5 95.6 125.2 81.6 3.9
± ± ± ± ± ± ± ± ± ± ±
8.2 6.3 2.5 33.7 67.3 11.2 29.8 24.4 15.9 11.2 0.57
UA-III
(n ¼ 110) 56.08 82.5 22.9 190.7 108.8 50.2 110.7 92.9 124 82.4 4.86
± ± ± ± ± ± ± ± ± ± ±
7.9 5.69 1.8 30 56.2 11.3 26.7 16.2 14.2 8.4 0.177***
UA-IV
(n ¼ 133) 57.3 83.3 23.3 197.8 126.8 49.6 115.6 93.9 129.6 84.1 5.45
± ± ± ± ± ± ± ± ± ± ±
7.8 6.85 2.61 34.0 79.3 11.9 31.0 17.0 18.1 12.8 0.16***
UA-V
(n ¼ 121) 55.1 83.4 23.3 194.6 128.2 48.4 114.4 90.5 122.6 82.4 6.03
± ± ± ± ± ± ± ± ± ± ±
7.8 6.9 2.8 37.7 76.9 12.8 30.2 10.6 15.7 10.6 0.18***
UA-1
UA-II
UA-III
UA-IV
(n ¼ 194)
(n ¼ 191)
(n ¼ 181)
(n ¼ 186)
53.06 76.8 23.1 194.3 97.7 50.6 113.2 88.6 119.8 78.2 2.9
± ± ± ± ± ± ± ± ± ± ±
8.2 7.8 2.8 36.0 47.1 10.0 30.4 15.9 15.6 10.3 0.40
52.3 77.06 23.4 195.4 96.0 51.3 114.6 89.03 122.1 80.1 3.6
± ± ± ± ± ± ± ± ± ± ±
8.2 7.6 2.7 33.7 40.9 10.7 27.9 11.5 18.2 12.7 0.13***
51.7 79.0 24.2 199.4 103.4 50.9 116.7 88.0 120.3 78.1 4.0
± ± ± ± ± ± ± ± ± ± ±
8.09 7.2 2.7** 36.0 54.9 11.2 29.0 9.5 16.6 11.7 0.10***
52.8 79.2 24.1 201.6 108.9 50.7 120.1 87.2 119.9 78.3 4.5
(n ¼ 167) 56.5 85.8 24.1 203.8 148 48.7 119.9 94.6 126.2 82.2 7.1
± ± ± ± ± ± ± ± ± ± ±
8.3 7.2*** 2.7*** 35.1* 79.6* 10.1 33.7* 13.2 16.1 10.2 0.74***
P value (#) 0.105 <0.001 <0.001 0.004 0.001 0.638 0.0189 0.142 0.0086 0.442 <0.001
UA-V
± ± ± ± ± ± ± ± ± ± ±
7.6 7.7* 2.9* 33.7 61.1 11.3 27.9 11.4 15.3 11.6 0.16***
(n ¼ 177) 54.5 80.3 24.5 211.7 114.7 50.4 129.3 89.1 125.3 81.3 5.4
± ± ± ± ± ± ± ± ± ± ±
7.5 7.1*** 2.9*** 44.7*** 56.0** 10.09 37.6*** 10.3 17.9** 11.5 0.56***
P value (#) 0.0135 0.001 <0.001 <0.001 0.002 0.945 <0.001 0.542 0.0081 0.026 <0.001
Data are expressed as mean ± standard deviation. WC, Waist circumference; BMI, Body mass index, HDL-C, High density lipoprotein cholesterol; LDL-C, Low density lipoprotein cholesterol; TG, Triglyceride; FBG, Fasting blood glucose; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; SUA, Serum uric acid. ***P < 0.001 compared with UA-I, **P < 0.01 compared with UA-I, *P < 0.05 compared with UA-I, # Difference between 5 groups (UA-I-UA-V).
analysis reported that in women, SUA increases the predictive ability for identification of future MetS subjects beyond the information given by the MetS components and may serve as a prognostic tool in women. SUA is an end product of purine metabolism in human and higher primates. The uric acid is formed in the liver by the breakdown of nucleic acid and protein. It has already been known and
evaluated that SUA is a potential risk factor for the development of MetS [5,11,20,21]. However, the strength of association was quite different in relation to sex, age, and ethnicity. The possible biological mechanism related to hyperuricemia and the future development of MetS has been studied in both the cell lines and animal model. A previous study reported that an increase in SUA has an effective role in the development of insulin resistance through the
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
5
Table 3 Odds ratio and 95% CI of each component of MetS at follow-up according to varying concentration of SUA in men.
Men WC (cm) Prevalence (%) OR (95%CI) BP Prevalence (%) OR (95%CI) HDL-C Prevalence (%) OR (95%CI) TG Prevalence (%) OR (95%CI) FBG Prevalence (%) OR (95%CI) MetS Incidence (%) OR (95%CI) Adjusted OR d (95%CI)
SUA-I
SUA-II
SUA-III
SUA-IV
SUA-V
(n ¼ 130)
(n ¼ 110)
(n ¼ 133)
(n ¼ 121)
(n ¼ 167)
15.38% 1 (Ref)
21.8% 1.53(0.79e2.96)
21.05% 1.46(0.77e2.76)
18.18% 1.22(0.629e2.37)
33.5% 2.77*(1.56e4.92)
31.5% 1 (Ref)
30.9% 0.971(0.561e1.68)
37.59% 1.30(0.785e2.17)
25.62% 0.748(0.431e1.29)
43.71% 1.68*(1.043e2.724)
29.2% 1 (Ref)
23.46% 0.749(0.420e1.33)
26.32% 0.865(0.504e1.48)
29.75% 1.02(0.596e1.76)
23.95% 0.763(0.453e1.28)
25.38% 1 (Ref)
21.82% 0.82(0.450e1.49)
29.32% 1.22(0.708e2.1)
26.45% 1.05(0.601e1.85)
36.53% 1.69*(1.021e2.80)
25.38% 1 (Ref)
23.64% 0.910(0.504e1.64)
25.56% 1.0(0.58e1.75)
15.70% 0.548(0.292e1.027)
28.14% 1.15(0.685e1.935)
14.62% 1 (Ref) 1 (Ref)
12.73% 0.852(0.406e1.79) 0.804(0.381e1.69)
18.05% 1.28(0.667e2.48) 1.22(0.630e2.37)
10.74% 0.703(0.331e1.494) 0.630(0.294e1.35)
25.75% 2.02*(1.11e3.68) 1.89*(1.03e3.47)
SUA, Serum uric acid; WC, Waist circumference; BP, Blood pressure; HDL-C, High density lipoprotein, TG, Triglyceride; FBG, Fasting blood glucose; MetS, Metabolic syndrome. * P < 0.05 compared with reference gp (SUA-I). d Adjusted for age, total cholesterol and low density lipoprotein cholesterol.
Table 4 Odds ratio and 95% CI of each component of MetS at follow-up according to varying concentration of SUA in women.
Women WC (cm) Prevalence (%) OR (95%CI) BP Prevalence (%) OR (95%CI) HDL-C Prevalence (%) OR (95%CI) TG Prevalence (%) OR (95%CI) FG Prevalence (%) OR (95%CI) MetS Incidence (%) OR (95%CI) Adjusted OR d (95%Cl)
SUA-I
SUA-II
SUA-III
SUA-IV
SUA-V
(n ¼ 194)
(n ¼ 191)
(n ¼ 181)
(n ¼ 186)
(n ¼ 177)
21.13% 1 (Ref)
22.51% 1.08(0.668e1.75)
33.15% 1.85*(1.164e2.94)
37.63% 2.25*(1.42e3.54)
44.63% 3.0*(1.9e4.73)
31.44% 1 (Ref)
26.18% 0.773(0.497e1.20)
26.52% 0.787(0.503e1.23)
29.57% 0.915(0.591e1.41)
41.81% 1.56*(1.02e2.39)
48.45% 1 (Ref)
51.31% 1.21(0.752e1.67)
54.70% 1.28(0.856e1.92)
54.84% 1.292(0.863e1.93)
53.67% 1.23 (0.819e1.85)
9.28% 1 (Ref)
15.7% 1.82(0.978e3.39)
16.57% 1.94*(1.04e3.62)
20.97% 2.59*(1.42e4.72)
25.42% 3.33*(1.84e6.02)
12.37% 1 (Ref)
10.47% 0.828(0.441e1.55)
9.94% 0.782(0.409e1.49)
11.29% 0.902(0.483e1.68)
14.69% 1.22(0.672e2.215)
12.37% 1 (Ref) 1 (Ref)
11.52% 0.922(0.498e1.70) 0.935(0.503e1.74)
15.47% 1.29(0.720e2.33) 1.33(0.739e2.42)
17.2% 1.47(0.830e2.609) 1.44(0.813e2.58)
23.73% 2.20*(1.27e3.820) 1.93* (1.105e3.400)
SUA, Serum uric acid; WC, Waist circumference; BP, Blood pressure; HDL-C, High density lipoprotein; TG, Triglyceride; FG, Fasting glucose; MetS, Metabolic syndrome.*P < 0.05 compared with the reference group. d Adjusted for age, total cholesterol and low density lipoprotein cholesterol.
inhibition of endothelial nitric oxide synthase [22]. Furthermore, Cook et al. observed the features of MetS in mice lacking endothelial nitric oxide synthase [23]. In addition, SUA stimulates inflammatory and oxidative changes in adipocytes, vascular smooth muscle cells, and human mononuclear cells [24,25]. The most significant proof of SUA as a predictor of MetS was furnished by the experimental studies of Nakagawa et al. in 2006 [26] in fructose-fed rat model (animal model of MetS) that showed a reduction in SUA level by allopurinol (inhibitor of xanthine oxidase) and reverses the feature of MetS. Despite overwhelming evidence of the association between SUA and insulin resistance (root cause of MetS), there are only a few data on the predictive value of SUA for new onset of MetS. However,
those studies did not have concerns about the information related to the additional predictive ability of SUA over and above that provided by different components of MetS at baseline. Our study reported a graded increase in the incidence of MetS among women with the increasing levels of SUA quintiles, whereas this was not noticeable in men. Sex-related variations of SUA in diagnosing MetS have also been reported. Also, several reports were underlined a stronger association between SUA level and MetS in females than in males [10e12]. The mechanisms underlying this association are not still clear. However, few studies have been shown that SUA level is independently associated with serum leptin level and advocated that leptin could be a risk factor responsible for hyperuricemia in metabolic syndrome [27]. Leptin directly impairs uric acid
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
6
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7
Table 5 Comparison of area under the ROC curve of each variable of metabolic syndrome and additional SUA model. Each components AUC (95% CI)
SUA additional model AUC (95% CI)
p-value
Men High waist circumference High triglyceride Low HDL cholesterol High blood pressure High glucose 5 components
0.599 0.548 0.522 0.533 0.555 0.725
(0.554e0.644) (0.501e0.595) (0.483e0.561) (0.484e0.583) (0.512e0.598) (0.676e0.774)
0.647 0.591 0.584 0.594 0.613 0.73
(0.589e0.706) (0.531e0.651) (0.524e0.644) (0.535e0.653) (0.553e0.672) (0.68e0.78)
0.031 0.1027 0.0421 0.0658 0.0246 0.4422
Women High waist circumference High triglyceride Low HDL cholesterol High blood pressure High glucose 5 components
0.583 0.551 0.538 0.577 0.528 0.735
(0.544e0.623) (0.518e0.585) (0.494e0.582) (0.533e0.62) (0.504e0.553) (0.694e0.776)
0.649 0.606 0.584 0.615 0.617 0.75
(0.6e0.697) (0.554e0.658) (0.534e0.634) (0.567e0.664) (0.566e0.667) (0.711e0.79)
0.0004 0.0111 0.0583 0.0147 0.0002 0.0144
AUC, area under the ROC curve; 95% CI, 95% confidence interval, P value is the comparison between each components of metabolic syndrome and additional SUA level.
excretion in the kidney [28]. The sexual differences in the SUA result for incident MetS in women could also be explained by hormonal effect. Postmenopausal women have higher uric acid level due to the effect of estrogens, which is uricosuric in nature and it favors uric acid excretion [29]. Baseline data of serum leptin concentration (men & women) and menopausal status in women according to the different quintiles of serum uric acid (SUA) are presented in online Supplementary Appendix 3. The ORs for development of MetS were 2.02 fold higher in men and 2.2 fold higher in women in the highest quintile of SUA levels compared with the lowest reference quintile. The results are consistent with the finding by Rye et al. 2007 [30]. Interestingly, large epidemiological studies have observed a positive relationship between elevated SUA and the growing prevalence of BP, FG, TG and WC in adolescent and adults [7e9,31e33]. In the present study, we also observed a significant association between higher SUA quintile and elevated WC, BP and TG (Tables 3 and 4). In women, we found an improvement in the area under the ROC curve in the model that added SUA to MetS components while the improvement was not seen in men. We calculated the category free NRI and the IDI to measure the predictive ability of SUA as a new biomarker after adding SUA to model with or without incident MetS. Pencina et al. suggested that NRI and IDI are more sensitive than the area under the ROC curve for the determination of improvement in the predictive value [34]. In our study, the addition of SUA significantly improved the NRI in women. The IDI measuring the overall sensitivity did not improve in both genders. However, this data specify that SUA may have a clinical role in the screening for metabolic risk in women but needed to clarify with large sample size stratified by gender. There are some limitations to the present study. Our studied subjects were restricted to middle age (40e70 years) and elderly Koreans residing in a rural area with proportionately high baseline prevalence and the high incidence of MetS [35]. The study did not have information regarding some important determinants of SUA such as dietary factors, including intake of purine-rich foods, fructose consumption which could contribute to hyperuricemia. We may have some limit to generalize our findings as our results were based on a sample of relatively healthy middle-aged Koreans. The follow-up period of our cohort was 2.6 years and about 25% of the sample did not complete the follow-up visit and thus we could not assess the presence of cardiometabolic abnormalities in a large sample. Finally, our study was based on a single determination of SUA that may lead to random measurement error. In conclusion, we found that SUA was significantly correlated with future risk of WC, BP, TG and it can be used in predicting the
development of MetS. The present study provides additional evidence that SUA has an important role in predicting new-onset of MetS among women. Further large prospective follow-up studies are needed to corroborate these findings. Authors contributions The study was conceived and designed by DY, CHC. Data collection in cohort study was carried out by SVA and SBK. Statistical analysis was done by EHC. Interpretation of data analysis was done by DY, CHC, and EYL. ESL, HMK, JSL, EYL helped in editing the manuscript. DY wrote the manuscript and final copy of article was approved by all authors. CHC supervised the whole project. Acknowledgments This research was supported in part by a fund (2005-E71013-00, 2006-E71002-00, 2007-E71013-00, 2008-E71004-00, 2009E71006-00, 2010-E71003-00) for research by the Korea Centers for Disease Control and Prevention. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atherosclerosis.2015.04.797. Conflict of interest There is no conflict of interest between the authors for this study. References [1] R.H. Eckel, S.M. Grundy, P.Z. Zimmet, The metabolic syndrome, Lancet 365 (2005) 1415e1428. [2] S. Mottillo, K.B. Filion, J. Genest, et al., The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis, J. Am. Coll. Cardiol. 56 (2010) 1113e1132. [3] N. Sattar, A. McConnachie, A.G. Shaper, et al., Can metabolic syndrome usefully predict cardiovascular disease and diabetes? outcome data from two prospective studies, Lancet 371 (2008) 1927e1935. [4] S. Lim, H. Shin, J.H. Song, et al., Increasing prevalence of metabolic syndrome in Korea: the Korean National Health and Nutrition Examination Survey for 1998-2007, Diabetes Care 34 (2011) 1323e1328. [5] X. Sui, T.S. Church, R.A. Meriwether, et al., Uric acid and the development of metabolic syndrome in women and men, Metabolism 57 (2008) 845e852. [6] P.W. Liu, T.Y. Chang, J.D. Chen, Serum uric acid and metabolic syndrome in Taiwanese adults, Metabolism 59 (2010) 802e807. [7] V. Lohsoonthorn, B. Dhanamun, M.A. Williams, Prevalence of hyperuricemia and its relationship with metabolic syndrome in Thai adults receiving annual
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797
D. Yadav et al. / Atherosclerosis xxx (2015) 1e7 health exams, Arch. Med. Res. 37 (2006) 883e889. [8] A. Onat, H. Uyarel, G. Hergenc, et al., Serum uric acid is a determinant of metabolic syndrome in a population-based study, Am. J. Hypertens. 19 (2006) 1055e1062. [9] H.K. Choi, E.S. Ford, Prevalence of the metabolic syndrome in individuals with hyperuricemia, Am. J. Med. 120 (2007) 442e447. [10] W.K. Chiou, M.H. Wang, D.H. Huang, et al., The relationship between serum uric acid level and metabolic syndrome: differences by sex and age in Taiwanese, J. Epidemiol. 20 (2010) 219e224. [11] T. Yang, C.H. Chu, C.H. Bai, et al., Uric acid level as a risk marker for metabolic syndrome: a Chinese cohort study, Atherosclerosis 220 (2012) 525e531. [12] M.L. Zhang, Y.X. Gao, X. Wang, et al., Serum uric acid and appropriate cutoff value for prediction of metabolic syndrome among Chinese adults, J. Clin. Biochem. Nutr. 52 (2013) 38e42. [13] L.K. Niskanen, D.E. Laaksonen, K. Nyyssonen, et al., Uric acid level as a risk factor for cardiovascular and all-cause mortality in middle-aged men: a prospective cohort study, Arch. Intern Med. 164 (2004) 1546e1551. [14] S.B. Koh, J.K. Park, J.H. Yoon, et al., Preliminary report: a serious link between adiponectin levels and metabolic syndrome in a Korean nondiabetic population, Metabolism 59 (2010) 333e337. [15] S.B. Koh, J. Yoon, J.Y. Kim, et al., Relationships between serum adiponectin with metabolic syndrome and components of metabolic syndrome in nondiabetic Koreans: ARIRANG study, Yonsei Med. J. 52 (2011) 234e241. [16] J.Y. Kim, S.V. Ahn, J.H. Yoon, et al., Prospective study of serum adiponectin and incident metabolic syndrome: the ARIRANG study, Diabetes Care 36 (2013) 1547e1553. [17] K. Tsukimori, T. Yoshitomi, S. Morokuma, et al., Serum uric acid levels correlate with plasma hydrogen peroxide and protein carbonyl levels in preeclampsia, Am. J. Hypertens. 21 (2008) 1343e1346. [18] K.G. Alberti, R.H. Eckel, S.M. 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 120 (2009) 1640e1645. [19] S.Y. Lee, H.S. Park, D.J. Kim, et al., Appropriate waist circumference cutoff points for central obesity in Korean adults, Diabetes Res. Clin. Pract. 75 (2007) 72e80. [20] J.P. Goncalves, A. Oliveira, M. Severo, et al., Cross-sectional and longitudinal associations between serum uric acid and metabolic syndrome, Endocrine 41 (2012) 450e457. [21] S. Hara, H. Tsuji, Y. Ohmoto, et al., High serum uric acid level and low urine pH as predictors of metabolic syndrome: a retrospective cohort study in a
7
Japanese urban population, Metabolism 61 (2012) 281e288. [22] T. Nakagawa, D.H. Kang, D. Feig, et al., Unearthing uric acid: an ancient factor with recently found significance in renal and cardiovascular disease, Kidney Int. 69 (2006) 1722e1725. [23] S. Cook, O. Hugli, M. Egli, et al., Clustering of cardiovascular risk factors mimicking the human metabolic syndrome X in eNOS null mice, Swiss Med. Wkly. 133 (2003) 360e363. [24] W. Baldwin, S. McRae, G. Marek, et al., Hyperuricemia as a mediator of the proinflammatory endocrine imbalance in the adipose tissue in a murine model of the metabolic syndrome, Diabetes 60 (2011) 1258e1269. [25] S. Furukawa, T. Fujita, M. Shimabukuro, et al., Increased oxidative stress in obesity and its impact on metabolic syndrome, J. Clin. Invest 114 (2004) 1752e1761. [26] T. Nakagawa, H. Hu, S. Zharikov, et al., A causal role for uric acid in fructoseinduced metabolic syndrome, Am. J. Physiol. Ren. Physiol. 290 (2006) F625eF631. [27] A. Bedir, M. Topbas, F. Tanyeri, et al., Leptin might be a regulator of serum uric acid concentrations in humans, Jpn. Heart J. 44 (2003) 527e536. [28] B. Fruehwald-Schultes, A. Peters, W. Kern, et al., Serum leptin is associated with serum uric acid concentrations in humans, Metabolism 48 (1999) 677e680. [29] A. Nicholls, M.L. Snaith, J.T. Scott, Effect of oestrogen therapy on plasma and urinary levels of uric acid, Br. Med. J. 1 (1973) 449e451. [30] S. Ryu, J. Song, B.Y. Choi, et al., Incidence and risk factors for metabolic syndrome in Korean male workers, ages 30 to 39, Ann. Epidemiol. 17 (2007) 245e252. [31] J.Y. Wang, Y.L. Chen, C.H. Hsu, et al., Predictive value of serum uric acid levels for the diagnosis of metabolic syndrome in adolescents, J. Pediatr. 161 (2012) 753e756 e2. [32] Y. Taniguchi, T. Hayashi, K. Tsumura, et al., Serum uric acid and the risk for hypertension and type 2 diabetes in Japanese men: the Osaka Health Survey, J. Hypertens. 19 (2001) 1209e1215. [33] S. Ryu, Y. Chang, Y. Zhang, et al., A cohort study of hyperuricemia in middleaged South Korean men, Am. J. Epidemiol. 175 (2012) 133e143. [34] M.J. Pencina, R.B. D'Agostino Sr., R.B. D'Agostino Jr., et al., Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond, Stat. Med. 27 (2008) 157e172 discussion 207e12. [35] Kyounghoon Lee Sbk, Myungsang Shin, Minsoo Ahn, Jang-Young Kim, ByungSu Yoo, Seung-Hwan Lee, Junghan Yoon, Tae Yong Lee, So Yeon Ryu, Jong Ku Park, Kyung-Hoon Choe, Prevalence and sex-related characteristics of metabolic syndrome in a Korean rural cohort, Soc. Lipidol. Atheroscler. J. 18 (2008) 226e238.
Please cite this article in press as: D. Yadav, et al., Prospective study of serum uric acid levels and incident metabolic syndrome in a Korean rural cohort, Atherosclerosis (2015), http://dx.doi.org/10.1016/j.atherosclerosis.2015.04.797