Background

High serum uric acid (SUA) is a well-known risk factor for gout and kidney stones [1] and an increasing number of studies links SUA as a risk factor for hypertension, cardiovascular disease, chronic kidney disease, insulin resistance, diabetes and cancer [2,3,4,5,6]. On the other hand, low levels of SUA have been recently related to Alzheimer and Parkinson diseases, chronic obstructive pulmonary disease [7] and even cardiovascular mortality [8,9,10]. Moreover, there is data evidencing the relation of SUA to fasting plasma glucose levels [11] as well as reduction in kidney function [12] in an U-shape curve fashion, indicating that both lower and higher values are associated with the outcome.

Many previous studies which have established SUA reference ranges in different populations were performed more than 40 years ago, when nutritional and clinical scenarios were different from nowadays [13,14,15,16,17,18,19,20], while SUA levels have been increasing in the general population in the last three decades at least [21, 22]. One of the most important limitations of previous studies is the use of the phosphotungistic method for measuring SUA [13, 14, 23, 24]. The uricase method is currently recognized as the gold standard for SUA levels determination [25], which was not available until the 1970s.

Several studies did not use validated international recommendations [26] for collection, transportation, processing and analysis of blood samples that are important for achieving good standards in laboratorial measures. In addition, factors known to modulate SUA levels such as alcohol consumption, chronic kidney disease and use of medications (e.g. diuretics) were most not excluded from study samples [23, 27]. More recent studies have also established SUA reference ranges for general population, however, without strict control for SUA determinant factors [28,29,30,31].

Production of reviewed data about distribution and SUA reference ranges is relevant to identify groups who may have higher risk to related diseases [32], and there is no universal reference range for SUA to date, as determined for fasting plasma glucose [33] or cholesterol [34], for example. Furthermore, laboratories around the world often create their own reference values based on local populations in pilot studies or use values indicated by laboratory kits. Prevalence of hyperuricemia might also be higher than previous reports due to the higher obesity prevalence in population [35]. Therefore, the aim of this study was to establish reference ranges for SUA for men and women in a large sample following rigorous laboratory quality standards and excluding factors which may directly interfere with SUA levels, as well as to establish the prevalence of hyperuricemia.

Methods

Study design

The Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) is a cohort study of 15,105 civil servants from five universities (Federal Universities of Bahia, Espírito Santo, Minas Gerais, and Rio Grande do Sul and the University of São Paulo) and one research institute (Oswaldo Cruz Foundation) located in different regions of Brazil [36,37,38]. All active or retired employees from the 6 institutions aged 35–74 years were eligible for the cohort. Exclusion criteria were current or recent pregnancy, intention to quit working at the institution in the near future, severe cognitive or communication impairment, and, if retired, residence outside of a study center’s corresponding metropolitan area. Sample size estimation was based on the main study outcomes—diabetes and myocardial infarction [36]. The first examination was carried out from 2008 through 2010 (baseline data). The ELSA-Brasil protocol was approved at each of the six study centers by the local Institutional Review Board addressing research in human participants. All participants provided signed informed consent.

This study is a cross-sectional analysis from ELSA-Brasil baseline data and sample size was composed of all participants with valid SUA levels at baseline, which is not sufficient to represent the entire population in Brazil, although it is very large and multicentric.

Each participant was interviewed at the workplace and again during a visit in the research center. Trained staff conducted interviews and examination following strict quality control procedures [38] and only data related to this study will be described in this section.

Definitions: all sample, reference sample, reference range and hyperuricemia

All participants from the ELSA-Brasil cohort included in the analyses we called all sample, and we established a reference sample after exclusion of participants with known factors that directly affect SUA levels: use of aspirin, thiazides, urate-lowering drugs, estrogen replacement therapy, excessive alcohol intake (> 140 g/week for women and > 210 g/week for men) [39] and estimated glomerular filtration rate < 60 ml/min/1.73 m2.

The reference range was defined as the central values between the 2.5th and 97.5th percentiles of SUA distribution in the reference sample for men and women, as established by Clinical and Laboratory Standards Institute (CLSI) [40]. Hyperuricemia was defined for SUA higher than or equal 7 mg/100 ml [41] for all sample for men and women.

We determined SUA according to the sociodemographic characteristics, habits, menopause and its distribution according to body mass index (BMI) and age in all and reference sample for men and women.

SUA and other variables

Procedures for collection of biological samples in ELSA-Brasil were standardized to assure uniformity at all investigation centres and followed the recommendations of the Brazilian Society of Clinical Pathology/Laboratory Medicine for blood collection [42]. Blood samples were drawn after 12-h fasting. Up to 30 min after the end of blood collection, all tubes were centrifuged under refrigeration for 15 min. Blood samples were stored in cryotubes at – 80 °C to the date of transport and processing, at most one month. Analysis of samples was centered at the laboratory of the University Hospital at University of São Paulo. SUA was measured by uricase method (automated colorimetric enzymatic) [43] by the ADVIA 1200 Siemens® equipment [36, 42]. This method is the recommended by Clinical and Laboratory Standards Institute [26] and the American College of Rheumatology and European League Against Rheumatism guideline [25].

Age, sex, education level (categorized as < 9, 9 to 11 and > 11 years of formal education), monthly family income (≤ US$1245, US$1246–3319, and ≥ US$3320), ethnicity according to the Brazilian census classification (black, brown or mixed, white, other—Asian, indigenous) [36], smoking status (never, past or current), excessive alcohol intake (> 140 g/week for women and > 210 g/week for men) [39] and menopause were self-reported in questionnaires at baseline. All participants were asked about their use of prescribed and nonprescribed drugs in the previous 2 weeks. BMI was calculated as weight divided by height in meters squared (kg/m2) [36, 37]. Physical activity was classified as poor, intermediate, or ideal according to the American Heart Association ideal cardiovascular health score [44]. Glomerular filtration rate was calculated using Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) [37, 45] and chronic kidney disease (CKD) was defined as a glomerular filtration rate < 60 ml/min [46].

Statistical analysis

Continuous variables are expressed as mean and standard deviation or median and interquartile range (IQR) and compared using analysis of variance (ANOVA) or Mann–Whitney U-test as appropriate after assessing normality assumptions. Categorical variables are expressed as proportions and compared using the chi-square or Fisher exact test. We calculated 95% confidence intervals (95% CI) for the frequency of hyperuricemia, whenever applicable. The Komolgoroff–Smirnov test was used for normality assessment. Significance level was set at 0.05.

Results

From the 15,105 participants from the ELSA-Brasil cohort, 5 individuals (< 0.1%) were excluded because they did not have a valid SUA level determination at study baseline. The all sample represents the 15,100 participants included in the analyses and the reference sample comprises 10,340 individuals after exclusion of participants with known factors that directly affect SUA levels: 873 (5.8%) individuals using aspirin, 1797 (11.9%) using thiazides, 90 (0.6%) using urate-lowering drugs, 539 (3.6%) under estrogen replacement therapy, 900 (6.0%) reporting excessive alcohol intake, 528 (3.5%) with estimated glomerular filtration rate < 60 ml/min/1.73 m2 and 33 (0.2%) with missing information in any of these criteria.

The all sample (N = 15,100) had a median age of 51 [IQR 45–58] years, mean BMI 27.0 (SD 4.7) Kg/m2 and consisted of 8216 (54.4%) women (Table 1). Median SUA was 6.3 [IQR 5.5–7.3] mg/100 ml for men and 4.7 [IQR 4.0–5.5] mg/100 ml for women. The reference sample (N = 10,340) had a median age of 50 [IQR 44–56] years, mean BMI 26.5 (SD 4.5) Kg/m2 and consisted of 5716 women (55.3%) (Table 1). Median SUA was 6.1 [IQR 5.3–7.0] mg/100 ml for men and 4.5 [IQR 3.9–5.3] mg/100 ml for women. Distribution of SUA were not normal, shifted to the right, for all and reference sample (Fig. 1).

Table 1 Sociodemographic and habits characteristics, menopause and prevalence of hyperuricemia in reference and all sample for men and women
Full size table
Fig. 1
figure 1

Distribution of serum uric acid by sex in reference (left) and all sample (right) for men and women

Full size image

Reference range for SUA (P2.5 to P97.5) was defined for reference sample: 4.0 to 9.2 mg/100 ml for men and 2.8 to 6.9 mg/100 ml for women (Tables 2, 3). Considering only individuals with BMI < 25 kg/m2, the SUA reference ranges would be 3.7 to 8.5 mg/100 ml for men and 2.7 to 6.2 mg/100 ml for women (Table 2). Prevalence of hyperuricemia for all sample was 31.9% (95% CI: 30.8–33.0%) in men and 4.8% (95% CI: 4.3–5.3%) in women and in reference sample was 25.3% (95% CI: 24.1–26.6%) in men and 2.2% (95% CI: 1.9–2.7%) in women (Table 1).

Table 2 Distribution of serum uric acid according to BMI in reference and all sample for men and women
Full size table
Table 3 Distribution of SUA levels according to age strata in reference and all sample for men and women
Full size table

Sex determined an important difference in the median values of SUA: 1.6 mg/100 ml higher in men in both reference and all sample. Higher BMI determined higher SUA levels for both men and women (p for trend < 0.001 for both sexes) in both reference and all sample (Table 2). For men there was no difference in SUA of reference sample among different age groups (categorized by 5 years) in reference sample (p for trend = 0.088), but age determined slightly higher SUA (p for trend = 0.038) in all sample. Older age determined higher SUA levels in reference and all sample for women (p for trend < 0.001 for both samples) (Table 3).

SUA levels according to other sociodemographic and clinical characteristics for reference and all sample are shown in Table 4. There was no difference in SUA levels according to ethnicity in men, but withe women had lower SUA levels (p = 0.01 in reference sample and p < 0.001 in all sample). SUA is slightly higher in individuals with lower educational level in women in both reference and all sample (p for trend < 0.001). Monthly mean family income influences the SUA levels only for women in all sample. Former smokers have higher SUA than current and never smokers in reference (p = 0.009 for men and p < 0.001 for women) and in all sample (p < 0.001 for men and woman). Menopause is associated with higher SUA levels for reference (4.7 [IQR 4.0–5.4] vs. 4.3 [IQR 3.8–5.3] mg/100 ml, p < 0.001) and all sample (4.9 [IQR 4.1–5.8] vs. 4.4 [IQR 3.8–5.1] mg/100 ml, p < 0.001). More frequent physical activity is associated with lower SUA in reference (p for trend = 0.01 for men and < 0.001 for women) and all sample (p for trend < 0.001 for men and women).

Table 4 Serum uric acid according to sociodemographic characteristics, habits and menopause in reference and all sample for men and women
Full size table

Discussion

This study determined the SUA reference ranges of a large and multicentric sample in a population defined as reference with exclusion of known conditions which affect SUA levels, using uricase method and following rigorous laboratory quality standards. It also determined the distribution of SUA regarding sex, age and BMI, the influence of sociodemographic characteristics, smoking and menopause on SUA along with the prevalence of hyperuricemia.

SUA reference range values in the present study are higher than those found previously. In a German study of the 1970s [23], it was found SUA levels from 3.4 to 7.0 mg/100 ml for men and 2.4 to 5.7 mg/100 ml for women, however, authors did not use the uricase method to determine the SUA levels. In an Indian study using uricase method to measure SUA [27], the reference range in 1470 subjects was slightly lower than those found in our study: 3.5 to 8.7 mg/100 ml in men and 2.5 to 6.9 mg/100 ml in women. Other studies included only participants with BMI between 18–25 kg/m2 and found 3.5–8.2 and 2.7–6.9 mg/100 ml for men, and 2.7–6.5 and 2.1–5.9 mg/100 ml for women [28, 30]. These results may reflect different populations studied, and exclusion criteria used.

Hyperuricemia is classically defined as SUA ≥ 7 mg/100 ml, considering the risk for incident gout [41, 47]. We found a high prevalence of hyperuricemia in men, even considering our reference sample. Nevertheless, the current gout classification criteria highlights SUA ranges and not a cut-off for hyperuricemia, indicating progressive disease risk for higher SUA values [25]. We also showed the prevalence of hyperuricemia considering the 6 mg/100 ml cut-off, considering the recent literature valuing it to the cardiovascular risk [3]. The association between SUA and hypertension, cardiovascular diseases, metabolic syndrome, diabetes, heart failure and CKD seems to be correlated with SUA levels either. Thus, the updated SUA values distribution in population receives special importance for quantifying the individuals at risk for these conditions [32].

SUA levels distribution (sex-specific) is not Gaussian. Other studies [14, 24, 27] also suggested a non-normal distribution, although they did not inform whether SUA levels distribution is parametric or not after applying a statistical test.

Sex was a major determinant for the SUA levels: in this study, the difference in median values for men and women was 1.6 mg/100 ml, higher in men. Other authors [13,14,15,16,17,18,19, 24, 27] found narrower SUA differences between sexes (0.8 to 1.6 mg/dl). In fact, our study’s (all sample) SUA levels are approximately 2.0 mg/100 ml higher in men and 1.5 mg/100 ml in women when compared with those past studies [13, 14, 16, 18, 20, 21, 23, 48]. Most studies of SUA levels distribution in large populations date from the 1960s and 1970s [13, 14, 16]. Since then, there has been a progressive SUA levels increasing [21, 24, 27], probably associated with current higher BMI values [35] in population, as well as changes in diet and alcohol consumption increase [49, 50]. Thus, SUA distribution curves needed to be updated to correctly reflect the nutritional and clinical profile of individuals.

SUA levels were directly related to BMI as observed by other authors [14, 21, 27]. There was a difference of 1.4 mg/100 ml in men and 1.1 mg/dl in women between the median SUA values in the highest and lowest strata of BMI (< 25 vs. > 35 kg/m2). Regarding age, we found similar SUA levels across age strata in men. However, older women showed slightly higher SUA levels. These results were similar to previous studies [13, 14, 21, 27] and menopause may be associated to this increase as suggested by Stockl et al. [51]. In fact, the SUA levels variations related to age and menopause are narrow and may have no clinical impact.

There was no relationship between SUA and self-reported ethnicity in our study as in another Brazilian population-based study [52] (except for women, with no clinical relevance values), which might be explained by the frequent miscegenation. However, we can find description in the literature of higher SUA levels and gout prevalence in Maoris, a specific population from New Zealand [53], and in African Americans [54,55,56], both related to the greater frequency of acquired risk factors, including obesity, physical inactivity, hypertension, diabetes, CKD, high seafood intake, elevated blood levels of lead, and use of antihypertensive medications [54].

There was a relationship of lower SUA with higher family income only for women in all sample. A former retrospective study has also found lower SUA levels in higher income groups [57]. Regarding smoking, other authors [48, 58] found that men who smoke have lower SUA levels, in spite of a study showing the opposite [59]. Frequent physical activity was associated with lower SUA levels, as also reported previously by Chen and Kawamoto [60, 61]. Fragala et al. [62] found this association only for males.

This study has some limitations. First, it is not confirmatory in determining the association between SUA and variables such as BMI and age due to its cross-sectional design. Second, the original cohort was not designed to establish the SUA reference range and prevalence of hyperuricemia and therefore may not represent the entire Brazilian population, although the sample is large and multicentric. Finally, we could not demonstrate the levels of SUA for individuals outside ELSA-Brasil age range (35 to 74 years).

Conclusions

Our manuscript proposes SUA reference ranges for men and women. We found high prevalence of hyperuricemia in men, even in a reference sample, after exclusion of conditions which influence SUA levels. Sex and BMI were major determinants for SUA. Updated SUA reference ranges and prevalence of hyperuricemia are higher nowadays and might be used to guide laboratories and screen diseases related to SUA.