Background
High levels of serum uric acid (UA) are prevalent in the general population. In the National Health and Nutrition Examination Survey (NHANES) 2007–2008 UA levels higher than 339 μmol/L were found in 21.6 % of the women, and among men 21.2 % had UA levels higher than 416 μmol/L [
1]. Similar prevalence has been found in China [
2]. The incidence and prevalence of hyperuricemia is increasing, as reflected by the increase in the incidence and prevalence of gout since the 1960s [
3]. In the US, the prevalence of gout more than doubled between 1969 and 1985 [
4], may have increased further over the past two decades [
1], and has paralleled a significant increase in prevalence of hyperuricemia [
1].
The metabolic syndrome (MetS) is a constellation of interrelated risk factors that increases the risk of cardiovascular disease and type 2 diabetes [
5]. MetS is associated with more than two-fold risk of atherosclerotic cardiovascular disease and cardiovascular death [
6]. The prevalence of MetS is high in most populations, and in the NHANES 2003–2006 about 34 % of US adults ≥20 years of age fulfilled the MetS definition [
7]. One study estimated the worldwide prevalence of MetS to range from <10 % to as much as 84 %, depending on region, sex, age and ethnicity [
8]. The prevalence of MetS increased significantly between NHANES 1988–1994 and NHANES 1999–2006, and one of the main reasons for this was the increase in abdominal obesity [
9]. Overweight and obesity is an increasing global burden [
10] and the number of overweight and obese is projected to continue to grow into the future [
11].
UA has been reported to be a risk factor for cardiovascular disease or cardiovascular death in many studies [
12‐
14], but not all [
15]. Studies have suggested that hyperuricemia is associated with all of the components of MetS individually: elevated blood pressure [
16], obesity [
17], high triglycerides [
18], low HDL [
18] and elevated fasting glucose [
19]. Several cross-sectional studies have shown an association between UA and MetS [
20,
21], although, after multivariable adjustment, the association disappeared in one study [
22]. The role of UA as an independent predictor of the development of MetS has also been examined in several prospective studies. In one study no such association was found [
23]. A recent meta-analysis comprising 11 prospective studies concluded that there was an independent, linear dose–response relationship between increasing UA and the development of MetS [
24]. As the prevalence of hyperuricemia increases along with the prevalence of overweight and MetS, the causal association between the phenomena remains unsolved. The purposes of the present prospective cohort study were to examine the role of UA and change in UA as a predictor of the MetS and its components after 7 years, and to assess to what extent overweight modified the associations between UA and the metabolic components.
Discussion
In this large prospective study of subjects without diabetes from the general population, elevated UA at baseline was independently associated with increased risk of elevated blood pressure in the overweight individuals 7 years later. We found no association between UA and future elevated blood pressure in the normal-weight subjects. Moreover, UA at baseline predicted new-onset impaired fasting glucose in the overweight persons, but not in the normal-weight group. Baseline UA was a predictor of MetS in all subjects. Finally, a longitudinal increase in UA of 59 μmol/L over 7 years raised the odds of developing MetS by 28 %.
The association between UA and MetS is in accordance with previous prospective studies [
17,
30,
31]. Few studies have examined the association between longitudinal UA change and MetS. In a healthy Japanese cohort, no significant association was found between 1 mg/dL (59 μmol/L) UA increase and incident MetS [
32]. However, in the Japanese study, follow-up time was shorter than in our study, and the authors did not adjust for baseline UA. These methodological differences may in part explain the discrepancies between the results of our study and the study from Japan.
To the best of our knowledge, there are no other studies of this scale where the population is stratified into normal-weight/overweight before examining the association between UA and MetS and its components. A small study (
n = 69) from the United Arab Emirates examined the univariable relationship between a set of biomarkers, among them UA, and components of MetS in healthy, young females, stratified into normal-weight (BMI ≤ 25 kg/m
2), overweight (BMI > 25, < 30 kg/m
2), and obese (BMI ≥ 30 kg/m
2) [
33]. This study found statistically significant correlations between UA and the waist circumference and triglycerides components only, and the associations were confined to the obese group. The authors found no significant correlation between UA and blood pressure in the strata; this may be due to small sample size and a population of uniform age and sex. In our study, we did not find any statistically significant interaction between the BMI-cutoff of obesity (BMI < 30 kg/m
2 and BMI ≥ 30 kg/m
2) for neither MetS nor any of its components. This may be due to a small group of obese in our cohort.
The association between hypertension and UA was first noted in the 1870s and has been demonstrated in numerous publications. In a recent meta-analysis, UA increase was reported to be associated with a statistically significant elevation in incident hypertension [
16]. It has been claimed that an elevated serum UA is the independent risk factor for hypertension that is the most reproducible to date [
34]. A multitude of studies, in an effort to explain how hyperuricemia can lead to hypertension and cardiovascular disease, have proposed interlinked mechanisms such as endothelial dysfunction and reduction in endothelial nitric oxide (NO) levels [
35], oxidative stress [
36], activation of the renin-angiotensin-aldosterone-system (RAAS) [
37] and renal microvascular lesions [
38]. However, we found that UA was a predictor of elevated blood pressure in the overweight, but not in the normal-weight. Few studies have explored this phenomenon. The precursor of UA is xanthine, and the reaction from the latter to the former is catalyzed by the enzyme xanthine oxidoreductase (XOR), which can exist in two forms, xanthine dehydrogenase (XDH) or xanthine oxidase (XO) [
39]. The enzyme is mostly in its XDH form, but can be transformed into XO by proteolytic cleavage or oxidation. In the XO form, reactive oxygen species are a by-product of the reaction of xanthine to UA [
40]. Therefore, under certain circumstances, increased activity of XO, detected as elevated production of UA, will lead to increased oxidative stress, which, in turn, can be detrimental in the state of reduced antioxidant capacity that accumulated fat creates [
41]. Furthermore, UA can affect adipocytes by inducing upregulation of pro-inflammatory factors and downregulation of the insulin sensitizer and anti-inflammatory factor adiponectin [
42]. Adiponectin is negatively associated with BMI and body-fat [
43]. Since low levels of adiponectin is associated with the development of hypertension [
44] and insulin resistance [
45], it could be speculated that adiponectin is part of the link between UA and elevated blood pressure and insulin resistance, and explain why UA is associated with new onset elevated blood pressure and impaired fasting glucose in the overweight but not the normal-weight in our study. Furthermore, a study found increased angiotensinogen levels in the hypertensive overweight (BMI ≥ 25 kg/m
2), compared to the hypertensive normal-weight (BMI < 25 kg/m
2), in the presence of hyperuricemia [
46], and a rodent model demonstrated that UA-mediated upregulation of adipose RAAS caused insulin resistance [
47]. UA might also directly contribute to the development of insulin resistance in adipose tissue, possibly through redox modulation [
48]. These could also be mechanisms in which UA is associated with overweight-related elevated blood pressure and elevated fasting glucose.
Epidemiologically, UA is associated with insulin resistance [
49], and the development of insulin resistance is often preceded by hyperuricemia [
50]. MetS does not comprise a uniform group of subjects; Sperling et al. of The Cardiometabolic Think Tank present a subtype where insulin resistance is dominant [
51]. An association between hyperuricemia and insulin resistance could in part explain the development of MetS.
The present study has important strengths: the large size, solid attendance rate, long follow-up time, use of UA as a continuous variable, and the ability to correct for confounders such as eGFR, use of diuretics and all the traditional cardiovascular risk factors. However, a major shortcoming of our study is the lack of fasting blood samples. In particular, glucose and triglycerides, and thereby the definition of MetS, are affected by this. The incorporation of time since last meal and adjustment of the cut-offs in the definition of elevated fasting glucose and elevated triglycerides compensated in part, but not fully, for this limitation. In addition, only a single measurement of serum UA was done in each survey. Another shortcoming of this study may be the fact that our baseline data were collected 21–22 years ago, and 14–15 years have passed since follow-up. Both lifestyles and pharmacological treatment have changed in that time. However, if the effects of overweight on UA’s association with MetS can be reproduced in studies on newer data, our findings may be even more relevant as overweight and obesity is an even greater challenge in the world of today. That our study population comprised largely of healthy, middle-aged to elderly Caucasians can be viewed as both a weakness and a strength; the results may not be generalizable to dissimilar populations, but the homogeneity of our cohort may have prevented dilution of our findings due to important diversities in baseline properties.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
Study design: JVN, HMS and MDS. Data analyses: JVN, HMS and MDS. Writing the first draft: JVN, HMS and MDS. Data interpretation, discussion and preparation of the final manuscript: JVN, HMS, KY, TGJ, SNZ, BOE and MDS. All authors read and approved the final manuscript.