Association between serum copper levels and prevalence of hyperuricemia: a cross-sectional study

Hyperuricemia has been recognized as a worldwide public health concern. This study was conducted to examine the association between serum copper (Cu) concentration and the prevalence of hyperuricemia in a middle-aged and elderly population. Serum Cu concentration was measured by Roche modular P800 using the PAESA method. Serum uric acid (UA) concentration was detected by a Beckman Coulter AU 5800. Presence of hyperuricemia was defined as serum UA ≥ 416 μmol/L for men and ≥360 μmol/L for women. The association between serum Cu concentration and the prevalence of hyperuricemia was evaluated by logistic regression. The prevalence of hyperuricemia was 17.6% (n = 6,212) in the present study. Relative to the lowest quintile, the age- and sex-adjusted odds ratios for hyperuricemia were 1.38 (95% CI: 1.12 to 1.70), 1.34 (95% CI: 1.07 to 1.66), and 1.53 (95% CI: 1.23 to 1.91) in the third, fourth, and fifth serum Cu concentration quintiles (P for trend < 0.001). Similar results were found both in men and women subgroups. None of the findings were materially altered after adjustment for additional potential confounders. In conclusion, in this population-based cross-sectional study, serum Cu concentration was positively associated with the prevalence of hyperuricemia.

Assessment of exposures. The serum Cu concentration was measured by Roche modular P800 using the PAESA method. The inter-assay coefficients of variation were 6.15% (9.8 μmol/L) and 4.325% (16.5 μmol/L), and the intra-assay coefficients of variation were 5.17% (16.72 μmol/L) and 5.37% (8.86 μmol/L) for serum Cu respectively. Fast blood glucose was tested using glucose oxidase enzyme method. Blood pressure was checked on an electronic sphygmomanometer. Diabetes was diagnosed as fasting glucose level ≥7.0 mmol/L or if the subject was receiving drug treatment to control blood glucose. Hypertension was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg, or if the subject was taking antihypertensive treatment. The BMI of each subject was calculated as weight in kilograms divided by height in meters squared. The information of sports activities, smoking and drinking status was extracted from questionnaires answered by subjects. Sports activities included average frequency and average duration, and current status of smoking and drinking was evaluated in binary (yes or no).

Statistical analysis.
The quantitative data were presented in form of mean ± standard deviation, and the qualitative data were presented in form of percentage. The serum Cu concentration was divided into 5 categories in terms of the quintile distribution of the sample: ≤13.60, 13.61-15.40, 15.41-17.00, 17.01-19.00 and ≥19.01 μmol/L. The one-way classification ANOVA and the Kruskal-Wallis H test was performed to evaluate continuous data for normal distribution and abnormal distribution, and the χ 2 test was employed to assess the differences between qualitative data. The quintile with the lowest value of serum Cu concentration was regarded as reference, age-and sex-adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were reported to indicate the association between serum Cu and the prevalence of hyperuricemia in each quintile, respectively (Model 1). Moreover, two multivariable models (Model 2 and 3) were also used for logistic regression analyses in the whole, male and female population. The covariate adjusted model 2 included age, sex, BMI, smoking, drinking status, and we additionally adjusted for education, activity level, hypertension, diabetes in subsequent model 3. Median variables of serum Cu concentration in first to the highest quintile were used to test the linear trends by logistic regression. To increase statistical power, we additionally modelled continuous serum Cu concentrations, estimating the effect on hyperuricemia prevalence per 10 μmol/L increase. A scatter plot was employed to test linear relationship between serum Cu and UA after multivariable analyses.
The collected data were conducted by SPSS 17.0 and STATA 12.0; a P value of ≤0.05 was considered as the criterion for statistical significance.
Ethical approval and informed consent. The study protocol had been approved by the Ethics Committee of Xiangya Hospital, Central South University (reference number: 201312459). Written informed consents had been acquired from all included participants prior to the performance of research.

Results
A total of 6,212 subjects (3,700 male and 2,512 female) were recruited in the current analysis. The overall prevalence of hyperuricemia in this study was 17.6%. Table 1 shows basic characteristics of the whole population and subjects according to their hyperuricemia status. Significant differences were observed between the hyperuricemia and non-hyperuricemia population in terms of sex, smoking, drinking status, education, BMI, occupation, hypertension and diabetes.
The results of demographic characteristics and confounding factors are shown in Table 2, in accordance with the quintiles of serum Cu concentration. There were significant differences across all quartiles of serum Cu in terms of age, sex, smoking status, drinking status, education, activity level, BMI, occupation, hypertension and diabetes.
The outcomes of multivariable adjusted connections between serum Cu concentration and the prevalence of hyperuricemia are listed in Table 3. Comparing with the lowest quintile, age-and gender-adjusted ORs (Model 1) of the total population demonstrated significant increased prevalence of hyperuricemia in the third [ Table 3). In addition, serum Cu was significantly associated with UA (P < 0.001) after adjustment for age, sex, BMI, smoking, drinking, education, activity, occupation, hypertension and diabetes (Fig. 1).

Discussion
The present cross-sectional study showed a positive association between serum Cu concentrations and the prevalence of hyperuricemia, and the findings remained consistent after adjustment for confounders.
Few studies have examined the association between Cu and serum UA, and the findings are still inconclusive. By conducted a population-based study with 1197 subjects aged between 45 to 64 years, Bo et al. 17 noted that UA decreased significantly from the lowest to the highest tertile of Cu intake; however, in a randomly identified subgroup of male subjects (n = 231) from this study, Bo et al. did not find any significant association between serum  www.nature.com/scientificreports www.nature.com/scientificreports/ Cu and serum UA. Unlike Bo et al., the present study indicated the prevalence of hyperuricemia was positively associated with serum Cu concentration, which may be attribute to our large sample size. Fields et al. 19 found that rats with Cu-deficient diet exhibited high levels of plasma UA; while Derouiche et al. 20 reported that Cu supplementation caused an augmentation of UA in rats with or without diabetes. However, the above animal studies only provided evidence for the relationship between dietary Cu and serum UA.
Under certain conditions, Cu plays its role as an activator of Xanthine oxidase (XO). XO is a key enzyme in purine metabolism, which can oxidize hypoxanthine from nucleic acid metabolites into xanthine, and xanthine into UA. Though several studies have documented the inhibition effect of Cu on XO [21][22][23] , in the later study conducted by Hadizadeh et al. 24 , Cu was proved to be both a reversible inhibitor and an activator of XO, a progressive reduction of catalytic efficiency was observed when the Cu 2+ concentration raised from 5 to 700 μM.
The biological mechanisms linking serum Cu to hyperuricemia are unclear. It has been reported that UA may function either as an anti-oxidant (mainly in plasma) or pro-oxidant (mainly in cells) 25 . A high UA concentration may contribute to the antioxidant behavior of UA 16,26 , while both UA and hyperuricemia related pathologies, such as cardiovascular disease, hypertension and obesity, are associated with oxidative stress 8,9,12,[27][28][29] . Therefore, UA may play a very complex role through the anti-oxidant and pro-oxidant effects in hyperuricemia. Cu 2+ is able to oxidize UA directly 15 , Cu 2+ /LDL ratio may be an influence factor on pro-and anti-oxidant behaviors of UA 30 . The Cu-reduction ability of UA explains not only its pro-oxidant behavior but also, in part, its anti-oxidant activity 16 . Moreover, UA is a potent mediator of inflammation. A high concentration of UA has been demonstrated to mediate inflammation by stimulating inflammatory cytokines 31,32 and the nuclear factor kappa B (NF-κB) pathway 33,34 . On the other hand, serum Cu level was identified to be positively correlated with inflammation activity markers 17,35,36 , and Cu was proved to activate NF-κB 37 and enhance the production of pro-inflammatory cytokines [38][39][40] . Based on the analysis above, the serum Cu may promote the effect of UA through the inflammatory mechanism.
The present study is characterized with several strengths. First, it is the first research work performed on a large sample (6,212 subjects) that directly relates serum Cu to hyperuricemia. The findings of this study may provide new insights into the mechanism and treatment of hyperuricemia. Second, adjustments for some potential confounding factors, such as education, activity level, occupation, hypertension and diabetes, increased reliability of the results. It should be noted that our study had several potential limitations. First, based on the cross-sectional data provided in this study, we cannot draw a conclusion that reflect the causal correlations. Further intervention trials and prospective longitudinal studies are therefore expected to establish a causal relationship between serum Cu and hyperuricemia. Second, the total serum Cu concentration was detected in this study. While approximately 90-95% of the total amount of Cu in blood serum is strongly protein-bound, mostly with α2-globulin (ceruloplasmin) 13 , the rest of Cu remains non-bounded (non-ceruloplasmin Cu). Ogihara et al. 41 observed an elevation of non-ceruloplasmin as well as a reduction of UA in three of four untreated patients with Wilson's disease, but the sample size was too small (only 4 patients) and all subjects were in a Wilson's disease condition. A further study focusing on the relationship between ceruloplasmin and non-ceruloplasmin Cu with serum UA is therefore suggested.

conclusion
In this population-based cross-sectional study, serum Cu concentration was positively associated with the prevalence of hyperuricemia. Further studies are required to estimate the exact mechanisms of the association between Cu and hyperuricemia pathogenesis.