Demographic, regional and temporal trends of hyperuricemia epidemics in mainland China from 2000 to 2019: a systematic review and meta-analysis

ABSTRACT Background Hyperuricemia (HUA) is becoming a global public health problem and associated with multiple diseases. Objective We conducted a systematic review to synthesize the pooled prevalence of HUA in mainland China and delineate its demographic, regional, and temporal trends over the past two decades. Methods Systematic literature searches of PubMed, SCOPUS, Web of Science, the China National Knowledge Infrastructure (CNKI), and the Wanfang digital database were conducted to screen studies published from 1 January 2000 to 31 August 2019, reporting the prevalence of HUA in mainland China. The search strings were (‘hyperuricemia’ OR ‘hyperuricaemia’ OR ‘uric acid’) AND (‘prevalence’ OR ‘epidemiology’) AND ‘China’. Article quality was appraised quantitatively from 11 items. Before formal meta-analysis, age-standardized prevalence was transformed. The random-effects model was applied to synthesize the pooled prevalence due to its high heterogeneity. Then we stratified the prevalence estimates by age, gender, area, nationality, and publication year for subgroup analysis. Results Totally 177 eligible studies with a whole population of 2,277,712 were included in the present meta-analysis. The pooled prevalence in mainland China was estimated at 16.4% (95% CI: 15.3%~17.6%). In studies with the onset age at 20 ~ 29 years old, males had a twice times higher HUA prevalence than females (21.5% vs. 8.9%). The prevalence of HUA was 13.7% (11.8%~15.7%) in people aged 15~ years old, 16.5% in 30~ (14.8%~18.4%), 17.9% in 40~ (16.4%~19.5%), 19.4% in 50~ (17.8%~21.0%), 20.5% in 60~ (18.8%~22.3%), and 24.9% in over 70 (22.9%~27.1%). Stratified by regions, southern (25.5%) and southwestern (21.2%) China shared the highest prevalence, and the lowest prevalence was observed in the northwest (12.6%). From 2001 to 2017, the prevalence estimates of HUA steadily climbed from 8.5% to 18.4% with minor fluctuations. Multiple regression revealed that HUA prevalence was positively correlated to the larger sample size, area, advanced onset age, and published year. Conclusions The last two decades witnessed the rapid growth prevalence of HUA in China. Early screening and personalized health education for HUA need to be given enough attention.


Background
Hyperuricemia (HUA), characterized as a metabolic disease of purine disorders, is becoming a global public health problem. Traditional recognition of HUA was limited to its triggering of gout and related disability-adjusted of life years [1]. Recently, more studies have proved that HUA was also associated with cardiovascular diseases, renal dysfunction, and cancers [2][3][4]. For each increase of 1 mg/dL in serum uric acid (SUA) level, the pooled relative risk of stroke and coronary heart disease mortality was 1.10 and 1.13, respectively [5]. With socioeconomic development, gout incidence has more than doubled worldwide in the past two decades [6]. Meanwhile, the prevalence of HUA also keeps growing in many developed countries. In Italy, HUA prevalence increased from 8.5% in 2005 to 11.9% in 2009 [7]. Another longitudinal study in Australia revealed a significant increase in SUA both in men and women from 1959 to 1980 [8].
China is the largest developing country. 1.4 billion Chinese people are spreading in 34 regions with the varied socioeconomic, cultural, and behavioral background. Since the geographic characteristics are widely divergent, few surveys on HUA were conducted in a cross-regional scope [9,10]. Moreover, the pathophysiologic of HUA is complex. Increased uric acid production and/or decreased renal excretion capacity could result in HUA. Specific dietary patterns and alcohol use contributed to the diverse prevalence of HUA in different regions. Hence, understanding the epidemics of HUA and its social factors is essential to estimate the disease burden in China. With this information, targeted and personalized anti-hyperuricemia healthcare can be implemented.
To this end, we carried out a systematic review to synthesize the pooled prevalence of HUA in mainland China and to delineate the demographic, regional, and temporal distribution over the past two decades. ). Then we use the filter panel to select studies with full text. For SCOPUS database, the search strings were (TITLE-ABS ('uric acid' OR 'Hyperuricemia' OR 'hyperuricaemia') AND TITLE-ABS ('prevalence' OR 'epidemiology') AND TITLE-ABS ('China') AND LANGUAGE (English)) AND DOCTYPE (ar) AND PUBYEAR > 1999 AND PUBYEAR < 2020. For Web of Science database, the search strings were (TI = (uric acid) OR TI = (hyperuricemia) OR TI = (hyperuricaemia) OR AK = (uric acid) OR AK = (hyperuricemia) OR AK = (hyperuricaemia)) AND (TI = (prevalence) OR AB = (prevalence) OR TI = (epidemiology) OR AB = (epidemiology)) AND (TI = (China) OR AB = (China) OR TS = (China)) AND (LANGUAGE = (English) AND (Indexes = SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH,  ESCI,  CCR-EXPANDED)  AND  (IC  Timespan = 2000-2019). Similarly, Chinese-language literature was searched, from CNKI and Wanfang database, by using the medical subject heading terms and/or keywords including 'hyperuricemia', 'uric acid', 'prevalence', and 'epidemiology'. In addition, we also searched the related references for the screened literature if necessary.

Inclusion and exclusion criteria
The criteria of literature screening were as follows: (1) original study conducted in mainland China; (2) community-based cross-sectional study design; (3) the published time was during 1 January 2000 and 31 August 2019; (4) the onset age of survey was ≥15 years old, including both males and females; (5) age-specific prevalence of HUA was reported; (6) sample size ≥ 1000; (7) with detailed information on the participants' demographic, HUA diagnostic criteria and sampling process. We further excluded studies that were hospital-based, combined with other diseases, animal experiments, duplicated, case-control or intervention studies, conference abstracts or reviews, diagnostic criteria lacking or not classified by gender, and not available for age standardization.

Study selection
Literature was imported into Endnote X9 software. Firstly, the title and abstract were screened independently by two researchers (Y.L. and Z.Y.S.). Articles that did not met the above eligible criteria of (1)(2)(3) were excluded from this process. Then we downloaded the full text and reviewed them for compliance with the remaining criteria. Any disagreements were referred to the third researcher (X.Y.Z.) for judgment to avoid selection bias.

Data extraction and data items
We extracted the necessary data items for meta-analysis. It included first author, published year, survey time, age range, survey region, HUA diagnostic criteria (male/female), sample size (male/female), cases (male/female), age-adjusted prevalence (male/ female), sampling method, sample source, and study design, etc.

Quality appraisal
The quality assessment was performed simultaneously according to the 'cross-sectional or prevalence study quality' checklist [11]. It was designed by the U.S. Agency for Healthcare Research and Quality in 2004, initially to evaluate the researches on celiac disease. This checklist had 11 items, including data source, inclusion criteria, survey time, community-based population, avoiding selective bias, measurement quality assessment, explaining the reasons for exclusion, potential confounding control, dealing with missing data, participant response rate, follow-up conditions. Each qualified item was awarded one point out of eleven. More detailed descriptions of this checklist could be seen in Supplementary Table 1.

HUA diagnosis and classification
HUA was defined as a fasting serum uric acid level >420 μmol/L in males and >360 μmol/L in females measured on two separate days after a regular purine diet [12]. Each diagnostic criterion was listed in Supplementary Table 2. Due to the geographic and economic disparities, we classified the 31 mainland provinces, municipalities and autonomous regions in mainland China as the eastern, northern, central, southern, southwestern, northwestern, and northeastern regions of China. According to the onset age, the study population was divided into five groups: 15~, 20~, 30~, 40~, and 50~. Other subgroups were analyzed in gender (male vs. female), nationality (Han vs.

Statistical analysis
The fifth and sixth China National Population Census of 2000 and 2010 provided the overall population data by 5-year age group and gender [13,14]. We applied these data to calculate the standardized prevalence of HUA in each literature, for reducing the demographic differences in studies. Meta-analysis was performed in 'meta' packages of R program 3.6.0 (R core team). Firstly, the Freeman-Tukey double arcsine transformation was utilized to convert the prevalence rates into a quantity. It made them satisfy a normal distribution and suitable for either the fixed-or random-effect model. The heterogeneity of the studies was evaluated by the Chi-squared-based Q test and I 2 test. The I 2 values of 25%, 50%, and 75% referred to low, moderate, and high heterogeneity, respectively. The pooled prevalence estimates were calculated in a fixed-effect model (if I 2 < 25% or p > 0.05 in Q test) or a random-effect model (if I 2 > 25% or p < 0.05 in Q test). Subgroup analysis was applied to delineate the demographic, regional, and temporal trends of HUA prevalence. In addition, meta-regression analysis were conducted to identify the major contributors to high heterogeneity. The potential publication bias was assessed by using the Begg's test and Egger's test. The significant level of α was set at 0.05, and all tests were two-tailed.  into five groups: 15~(N = 51, n = 1,099,992), 20~ (N = 96, n = 1,041,320), 30~(N = 11, n = 46,758), 40~ (N = 9, n = 33,762), 50~(N = 10, n = 55,880).

Basic characteristics and quality appraisal of enrolled studies
The detailed characteristics of 177 studies were shown in Supplementary Tables 2-4 and Supplementary Text 1. All studies were cross-sectional, conducted in community population (n = 51), physical examination population (n = 85), occupational population (n = 40), and college students (n = 1). Multistage randomized sampling was applied in nearly one-quarter of studies (n = 44). The response rate of 166 was over 90.0%. As for quality appraisal, the average score of eligible studies was 7.81 ± 1.07, with a maximum of 10 and a minimum of 6. Of them, fifty-seven studies scored 8 (32.2%), thirty-four studies scored 9 (19.2%), and twelve studies scored 10 (6.8%).

Analysis of heterogeneity and publication bias
The overall heterogeneity between studies was noted as a significantly high level (I 2 = 99.8%, p < 0.001); even in subgroup analysis, the heterogeneity only decreased slightly. Hence, we performed the metaregression to identify the major contributors to high heterogeneity ( Table 2). Univariate analysis revealed that the HUA prevalence increased 16.4% and 11.3% in participants from the southern and southeastern as compared with those from northwestern China. The higher the article quality score gained, the lower the HUA prevalence was (β = −0.022, p = 0.002). In multiple regression, such a positive correlation was also found in studies with larger sample size, advanced onset age (50~), and published year since 2006. Egger's linear regression test and Begg's test showed that no significant publication bias was in both overall and subgroup analysis (Table 1).

Discussion
It is the first study attempting to synthesize the pooled standardized prevalence estimations of HUA among Chinese populations. Through a systematic literature search, we enrolled 177 articles (2,277,712 participants) to delineate the epidemiological distribution of HUA in mainland China over the past 20 years. The age-standardized prevalence was transformed before meta-analysis to minimize the differences of demographic characteristics between studies. In studies with the onset age at 20 ~ 29 years old, the HUA prevalence in males and females were 21.5% and 8.9%, confirming HUA a considerable disease burden in China. Moreover, the prevalence also kept increasing from 13.7% in 15 ~ 29 years old to 24.9% in the over 70 years old. It may be explained by the physiological reduction of the renal glomerular filtration rate with aging. SUA is the end product of purine metabolism, generated by nucleic acids and other purine compounds produced by catabolism of systematic nuclear proteins and food intake. Over two-thirds of SUA is excreted through the kidney, filtrated 100% in glomerulus, retaken 90% in proximal tubule, and the rest discharged [15]. The glomerular filtration rate has been proved to physiologically decline 0.75 ~ 1 ml/min with healthy aging in humans after 40 years old [16]. However, this trend of gender differs moderately. In males, the pooled prevalence was high along with the all adult age, peaked in the groups of 40 ~ 49 and 70 ~ . Unlike hypertension or type 2 diabetes, HUA is also commonly encountered in the young and middle-aged, which could be explained by the food impact on SUA level [17,18]. About 20% of SUA daily generated is from food intake, while the barbecue and hot pot with purine-rich food are getting popular in the young [19]. Moreover, males tend to indulge in alcohol and sugared beverages than females, and these drinks have been proved to interfere with uric acid metabolism and increase the risk of HUA [20,21]. This trend indicates that health education and screening should be conducted in young males. While, in females, the prevalence of HUA maintained relatively low from 5.3% to 8.1%, but increased sharply from 50 years old, till similar to that of males in 70 ~ . Such a trend is associated with menopause.
Estrogen inhibits the biosynthesis and promotes the excretion of SUA, and the estrogen in females before menopause is 2 ~ 3 times of males of the same age [22]. Elevated body mass index, insulin resistance, and alcohol ingestion may also be the factors contributing to the increased prevalence of HUA in postmenopausal females [23]. Thus, in females after menopause, the awareness of HUA should also be enhanced.
The distribution of HUA also showed geographical differences. The prevalence was higher in southern China (25.5%), southwestern China (21.2%), and northeastern China (15.9%), which may be attributed to regional eating habit, either is rich in purine content or interferences with the metabolism of purine. In southern, seafood is among the daily diets, while the hot pot is the most popular in southwestern. Alcohol is most consumed in northeast China than in the rest regions. In our study, HUA's extremum prevalence was observed in Guangdong (42.2%) and Xinjiang (1.8%). Minority localities and limited sample size in border areas of China may contribute to the extremely low prevalence of Xinjiang, indicating a need for studies with a larger sample size and gene polymorphisms in these areas. The relatively poor awareness rate, treatment of HUA in these areas also should be paid attention to. Health education    should be expanded to cover these populations, and regional specified dietary instruction should be focused on high-risk groups.
It was observed that, in the past two decades, the prevalence of HUA in mainland China doubled with a steady increase rate in both males and females. Augmented life pressure, shifted diet patterns and habits, reduced physical exercise are all potential factors contributing to this phenomenon [19,[24][25][26][27][28]. Given the huge population base in China, this doubled prevalence has tremendous implications for public health. Intensified interventions must be taken now to deal with this situation, which was neglected before. Apart from active treatment, prevention measures such as health education should be taken immediately at the community level, which can profoundly impact the whole population.
This meta-analysis estimated the age-standardized prevalence of HUA and analyzed the demographic, regional, and temporal distribution in mainland China throughout the past 20 years. Still, there are some limitations needed to be clarified. Firstly, high heterogeneity was noted in subgroup analysis, due to the vast geographic area and publication time span. We performed meta-regression to identify the major contributors to high heterogeneity. Surveys focusing on the southern and southeastern regions, larger sample size, advanced onset age, and published year were associated with higher HUA prevalence. Secondly, data of other factors may influence heterogeneity and partly contribute to the HUA prevalence (e.g., dietary patterns, alcohol consumption, stressful life, sedentary lifestyle, etc.) were not complete in the reviewed studies. Further large-scale studies should take these aspects into account. Thirdly, no longitudinal study was found in literature screening, and the included studies are almost cross-sectional design, which can only provide the prevalence but not infer the causal sequence.
To conclude, the last 20 years witnessed a double prevalence of HUA in mainland China with steady growth. Young and middle-aged males, residents in south and southwest China need to be given enough attention to with early and personalized education for HUA, and the elderly also should be screened and treated in time.