Abstract
Objective To determine the association of physical activity (PA) and sedentary time (ST; leisure and total ST), commuting mode with hyperuricemia in a multiethnic Chinese population, and to analyze the difference between sexes.
Methods Baseline data were analyzed from 22,094 participants from the China Multi-Ethnic Cohort study in the Yunnan region, China. PA and sedentary behavior were assessed through questionnaires. Hyperuricemia was defined as serum urate > 7.0 mg/dL among men and > 6.0 mg/dL among women. A restricted cubic spline (RCS) was created to model the possible nonlinear relationship of PA and ST with hyperuricemia. Logistic regression was used to estimate the odds ratio (OR) and 95% CI.
Results Hyperuricemia prevalence in the observed population was 15.5% (men 25.5%, women 10.7%). Compared to those with light PA, participants with moderate-to-vigorous PA had lower odds of hyperuricemia (adjusted ORs were 0.85 [95% CI 0.77–0.94] and 0.88 [95% CI 0.79–0.97]). However, RCS showed a U-shaped nonlinear relationship between PA and hyperuricemia, and a linear relationship between hyperuricemia prevalence and increasing ST. Total ST ≥ 4 hours/day increased the risk of hyperuricemia in women but not in men. Mode of transportation revealed that sedentary behavior increased the risk of hyperuricemia, but there were inconsistent results based on sex.
Conclusion Moderate PA may be more beneficial in reducing the risk of hyperuricemia. Reducing ST may have a greater effect on preventing hyperuricemia in females than in males.
Hyperuricemia, characterized by the elevation of serum uric acid concentration, has been associated with many chronic metabolic diseases, such as cardiovascular disease, chronic kidney disease, and metabolic syndrome.1,2,3 Specifically, prolonged hyperuricemia leads to the formation of monosodium urate crystals that accumulate in joints and other tissues, causing gout.4 Previous epidemiological data found that approximately 21% of adults in the US had hyperuricemia,5 and the prevalence ranges from 13% to 25.8% in some Asian countries.6,7,8,9 In China, the prevalence of hyperuricemia was 13.3% between 2000 and 2014.6,7 Hyperuricemia has been steadily increasing in recent years, causing significant health and economic burdens, and is considered an important public health problem worldwide.10,11 However, to date, a precise pathological mechanism of hyperuricemia has not been fully elucidated.12 Thus, identifying its risk factors is urgently needed to provide appropriate early prevention strategies.
Lifestyle behaviors, such as dietary patterns and physical activity (PA), play an essential role in individual health. A report from Asian participants showed that physical exercise was able to reverse the 27% increase in mortality associated with high serum urate (SU) and that among individuals with high SU levels, being fully active extended life expectancy.13 The positive effect of lifestyle measures for the prevention or treatment of hyperuricemia has been well documented. Participation in regular PA and reduced sedentary time (ST) was highly recommended to prevent hyperuricemia, as displayed in a cohort study of 158,713 participants.14 A recent study in the rural Chinese population demonstrated that prolonged ST could significantly increase SU and hyperuricemia.15
In China, studies frequently reported that prevalence of hyperuricemia was higher in males than in females.6,15,16 Similar results have been observed in the US.17 Male sex is a significant risk factor for hyperuricemia and gout, with men up to 4 times more likely to be affected than women.18,19 There are many reasons for this difference, such as estrogen, genetics, and transcription factors.20,21 A previous study showed that a greater likelihood for developing hyperuricemia is associated with eating habits for males, and with lifestyle factors such as type of work, commuting method, and exercise for females.16 In our present study, we used data from the China Multi-Ethnic Cohort study (CMEC), Yunnan province, to evaluate the associations between sedentary behavior, PA, and hyperuricemia, and to examine the potential modifying effects of sex on these associations, as well as the possible joint effect of ST and PA on the risk of hyperuricemia.
METHODS
Study design and participants. This study was carried out using baseline data from the CMEC, an ongoing large-scale prospective cohort study that comprises 5 provinces/cities (Tibet, Sichuan, Yunnan, Guizhou, and Chongqing) in Southwest China, conducted between May 2018 and September 2019.22 Briefly, a multistage stratified cluster sampling method was used to obtain samples from community-based populations. First, minority settlements in the cities of Lijiang, Dali, and Chuxiong Yunnan province were selected as our study sites. Then, Yongsheng County, Heqing County, Yongren County, and Wuding County were selected as secondary sampling units from these prefectures. Last, 1–8 communities were randomly selected by the local Centers for Disease Control and Prevention. Community size, health conditions, migration status, and ethnic structure were considered. The plateau residents of Han nationality were investigated in Lijiang City, whereas the Bai and Yi nationality were gathered from Dali City and Chuxiong City, respectively.
A total of 23,143 individuals aged 30–79 years were recruited and interviewed. Participants were excluded if they were unable to provide a unique national identification card, had severe physical or mental diseases (eg, schizophrenia, bipolar disorder), or could not comply with the requirements of the study. We further excluded individuals who had incomplete questionnaires or had no SU results (n = 840), as well as those who reported implausible values for PA or sitting time (n = 179) and extreme values (> 24 h) for sleeping and sitting time (n = 30). Finally, 22,094 participants were included in this analysis (Figure 1). All participants provided written informed consent prior to data collection. The study protocol was approved by the Ethics Committee of Kunming Medical University (KMMU2020MEC078).
Study flow chart.
Measurements. An electronic questionnaire was used along with face-to-face interviews. Body measurements such as height, weight, and blood pressure (BP) were performed 3 times by licensed physicians in the communities according to the standard operating procedures of the CMEC. Height and weight were measured to the nearest 0.1 cm and 0.1 kg, respectively, while wearing light clothes and without socks and shoes. BP was measured using an electronic sphygmomanometer (Omron U30), the first taken after the subject had been sitting quietly for at least 5 minutes, and the second and third taken at 1-minute intervals. Participants were requested to maintain an empty stomach for at least 8 hours before blood was drawn. Clinical laboratory tests were performed by professional testing institutions. Study protocols have been published previously.22
Definition of hyperuricemia. Outcome ascertainment of hyperuricemia was defined as SU > 7.0 mg/dL in males and > 6.0 mg/dL in females, consistent with the definition of most epidemiological studies.
Assessment of PA and sedentary behavior. The questions on PA and sedentary behavior were adapted from validated questionnaires used in the China Kadoorie Biobank (CKB).23 Data collection was divided into agricultural and nonagricultural workers. Participants were investigated about their usual type and duration of activities related to work, commuting, household chores, sedentary and leisure time, and exercise during the past year. Activity types were classified as follows: heavy manual work, manual work, standing work, sedentary work, manual work in the farming season, semi-mechanized work in the farming season, fully mechanized work in the farming season, commuting mode (walking, bicycle, motorbike, private or public transportation [eg, bus, car, underground, ferry]), household activity, tai chi/qigong/leisure walking, jogging/aerobics, swimming, ball games (eg, basketball, badminton, table tennis), exercise with fitness equipment, and other exercises (eg, mountain walking, home exercise, jump rope).
PA was calculated by multiplying the metabolic equivalent (MET) value for a particular type of PA by hours spent on that activity per day and summing the MET hours for all activities. METs were based on the 2011 Compendium of Physical Activities.24 PA (MET h/day) included physical activity related to jobs, transportation, leisure time, and housework. Leisure ST activities were recorded, such as playing on a mobile phone or tablet, watching television, reading, playing cards or mahjong, and using a computer outside of work (quantified as hours/day). Total ST included leisure ST and work ST. Participants were asked to report 1 (regular) commuting mode in the past year, and response options included the following: walk or bicycle, motorbike or private car or bus, working at home or nearby, housework, or disabled to work.
Demographics. Besides age, sex, and ethnic group, other variables were defined as follows: (1) education level was divided according to junior high school; (2) occupation was classified agricultural and nonagricultural according to the PA questionnaire; and (3) economic status was classified according to annual household income.
Lifestyle variables. Smoking was defined as a total > 100 cigarettes smoked to date. Previous smokers were defined as having quit for > 6 months. Alcohol and tea consumption was categorized as never or ever. Fat intake was calculated according to the food frequency questionnaire in the Chinese Food Composition List (2nd edition)25; ≥ 75 g/day was defined as a high-fat diet. Sleep disorders were measured using a questionnaire from CKB by asking 3 questions: (1) Did you have difficulty in falling asleep (sleep onset latency ≥ 30 min), or wake up in the middle of the night at least 3 days a week; (2) Did you wake up too early in the morning and have difficulty falling asleep again at least 3 days a week; and (3) Did you have trouble being clear-minded during the day due to poor sleep at least 3 days a week? If they answered yes to any of the 3 questions, they were classified as having sleep disorders.26
Diseases. BMI (kg/m2) was classified as normal (18.5–23.9 kg/m2), overweight (24.0–27.9 kg/m2), and obese (≥ 28.0 kg/m2). Diabetes was based either on fasting plasma glucose ≥ 7.0 mmol/L, random blood glucose ≥ 11.1 mmol/L, 2-hour postprandial blood glucose ≥ 11.1 mmol/L, HbA1c ≥ 6.5%, or current use of blood glucose–lowering agents or insulin. Hypertension (HTN) was defined as diastolic BP ≥ 90 mmHg, systolic BP ≥ 140 mmHg, self-reported physician-diagnosed condition, or current use of antihypertensive medication. Diabetes and HTN were each categorized as yes or no.
Sensitivity analysis. Sensitivity analysis was performed by redefining hyperuricemia as SU > 6.8 mg/dL, in which monosodium urate crystals form in vitro at physiological temperatures and pH levels.
Statistical analysis. Individuals were categorized into hyperuricemia and nonhyperuricemia according to the cut-off points defined. Continuous variables were expressed as mean ± SD if normally distributed and as median and IQR otherwise, whereas categorical variables were expressed as frequency and percentage. The Kolmogorov-Smirnov test was used to determine whether data were normally distributed. The significance was determined by t test, Mann-Whitney U test, Kruskal-Wallis H test for continuous variables, and by chi-square test for categorical variables.
Restricted cubic spline (RCS) was used to model the nonlinear association between continuous exposures (leisure ST, total ST, and PA) and hyperuricemia. Three knots located at the 5th, 50th, and 95th percentiles were used in the RCS function, while the median of the lowest PA and ST were used as a reference value. PA (leisure ST and total ST only) and total ST (PA only), age, sex, ethnicity, education, occupation, income, smoking, alcohol and tea consumption, high-fat diet, sleep disorders, BMI, diabetes, and HTN were included in RCS models as covariates to control their potential confounding effects.
A binary logistic regression model was applied to estimate the odds ratio (OR) and 95% CI for hyperuricemia. Collinearity diagnostics were done with the variance inflation factor (VIF) test, with VIF < 10 and the tolerance statistic > 0.1. Leisure ST and total ST were shown as categorical variables (< 2, 2–3.9, 4–5.9, ≥ 6 h/day). PA was divided into tertiles based on total MET hours/day (light: < 21.5 MET h/day; moderate: 21.5–41.1 MET h/day; vigorous: > 41.1 MET h/day). The same covariates in RCS were included in the binary logistic regression model. Trend tests were conducted by treating the exposure variables categorized into tertiles (PA) or different levels (ST) and putting them into the models as continuous variables. Interaction tests between PA and ST were conducted simultaneously by adding the respective multiplicative terms in the models. Subgroup analysis was performed according to sex.
All reported P values were 2-tailed and α level of ≤ 0.05 was considered statistically significant. Statistical analyses were performed as indicated using SPSS version 20.0 (IBM Corp.) and graphics were produced using SAS version 9.4 (SAS Institute).
RESULTS
The basic characteristics of the study participants according to the hyperuricemia status are presented in Table 1. A total of 22,094 individuals were eligible for inclusion in the analysis. Hyperuricemia prevalence was 15.5% (3435/22,094; men 25.5% [1851/7245], women 10.7% [1584/14,849]). Participants with hyperuricemia tended to be older, male, of Han ethnicity, with higher education, nonfarmers, smokers, drinkers, have a high average annual income, and consume a high-fat diet (all P < 0.001). Moreover, they also had higher BMI, diabetes mellitus, and HTN (all P < 0.001). Participants with hyperuricemia were more likely to have light PA, highest ST levels, and more likely to work at home, nearby, engage in housework only, or were disabled workers (all P < 0.001).
Characteristics of the study participants with and without hyperuricemia.
We compared SU according to categories of sedentary behavior and PA (Table 2). Average SU levels were observed to decrease with increasing PA tertiles, whereas SU concentrations were higher with increasing ST (P < 0.001). We also observed that walking/riding a bicycle to work had lower SU levels relative to other commuting modes (P < 0.001). These results were similar in men and women (all P < 0.001).
Serum urate (mg/dL) level in the study participants according to sedentary behavior and physical activity.
We examined the possible nonlinear relationship between PA, ST, and hyperuricemia using RCS models in all participants (Figure 2). No significant nonlinear relationship was observed for ST with the risk of hyperuricemia (P for nonlinearity > 0.05), with a linear positive dose-response relationship between them (Figures 2A,B). A U-shaped relationship was shown between PA and hyperuricemia (P for nonlinearity = 0.002; Figure 2C).
Nonlinear relationship of (A) leisure ST, (B) total ST, (C) and PA with hyperuricemia among all participants. In the RCS functions, we used 3 knots (5th, 50th, and 95th percentiles) and set the median of the first level of ST and the median of the first tertiles of PA as the reference point for all participants. Model was adjusted for physical activity (leisure ST and total ST only) and total ST (PA only), age, sex, ethnicity, education, occupation, income, smoking, alcohol drinking, tea drinking, high-fat diet, sleep disorders, BMI, diabetes, hypertension. The odds ratio is shown in the solid red line, and 95% CIs in the shaded area. The green horizontal dashed lines represent reference line y = 1. The black vertical dashed line represents reference line ST = 1 h/day, PA = 11.86 MET h/day. ST: sedentary time; MET: metabolic equivalent; PA: physical activity.
Figure 3 shows the ORs and 95% CIs for sedentary behavior and PA with hyperuricemia for all participants. After adjusting for covariates, participants who spent ≥ 6 h/day of leisure ST were more likely to have hyperuricemia than those who spent < 2 h/day (OR 1.28, 95% CI 1.05–1.56). The ORs for hyperuricemia comparing total ST 2–3.9 h/day, 4–5.9 h/day, and ≥ 6 h/day with < 2 h/day were 1.01 (95% CI 0.92–1.12), 1.16 (95% CI 1.02–1.33), and 1.22 (95% CI 1.05–1.42), respectively (P for trend = 0.002). Additionally, there was a negative association between PA and hyperuricemia after adjusting for potential confounding factors (P for trend = 0.01). Compared with the lowest PA tertile, moderate PA and vigorous PA had decreased risk of hyperuricemia, with OR 0.85 (95% CI 0.77–0.94) and OR 0.88 (95% CI 0.79–0.97). Participants in the groups working at home or nearby, and those performing housework or disabled to work had higher risk of hyperuricemia than walking/riding a bicycle to work, with OR 1.20 (95% CI 1.03–1.39) and OR 1.45 (95% CI 1.27–1.65), respectively. Meanwhile, the interaction showed no statistical significance in PA and ST (P for interaction > 0.05).
Odds ratios and 95% CIs for sedentary time, physical activity, and commuting mode with hyperuricemia for all participants (n = 22,094). Model 1: adjusted for physical activity (ST only) and total ST level (PA only). Model 2: adjusted for Model 1 plus age, sex, ethnicity, education, occupation, income, smoking, alcohol and tea consumption, high-fat diet, sleep disorders, BMI, diabetes, and hypertension. Tests for trends were conducted by leisure ST, total ST, and PA as continuous variables and tests for interactions were conducted by adding the respective multiplicative terms in the models simultaneously. * Model 1 additionally adjusted for PA and total ST. MET: metabolic equivalent; PA: physical activity; ST: sedentary time.
Sex-stratified analysis is presented in Figure 4. Leisure ST had no increased risk of hyperuricemia in men and women after adjustment by covariates. A statistically significant positive association between total ST and hyperuricemia was only found in women (P for trend = 0.009; P for trend in men = 0.21). A statistically significant negative association was observed between hyperuricemia and PA in men (P for trend < 0.001) and moderate PA in women (OR 0.80, 95% CI 0.70–0.93); however, high PA in women had an OR 0.89 (95% CI 0.77–1.05). The relationship between commuting mode and hyperuricemia also presented inconsistent results based on sex. Compared with employees walking/riding a bicycle to work, women working at home or nearby, and doing housework or being disabled to work were more likely to have hyperuricemia (OR 1.34, 95% CI 1.08–1.65 and OR 1.60, 95% CI 1.33–1.92, respectively). Men who did housework or were disabled to work had an increased risk of hyperuricemia (OR 1.28, 95% CI 1.05–1.56). The interaction of ST and PA was also not statistically significant in men and women (P for interaction > 0.05).
Odds ratios and 95% CIs for sedentary time, physical activity, and commuting mode with hyperuricemia for males (n = 7245) and females (n = 14,849). Model 1: adjusted for physical activity (ST only) and total ST level (PA only). Model 2: adjusted for Model 1 plus age, sex, ethnicity, education, occupation, income, smoking, alcohol and tea consumption, high-fat diet, sleep disorders, BMI, diabetes, and hypertension. Tests for trends were conducted by leisure ST, total ST, and PA as continuous variables and tests for interactions were conducted by adding the respective multiplicative terms in the models simultaneously. * Model 1 additionally adjusted for PA and total ST. MET: metabolic equivalent; PA: physical activity; ST: sedentary time.
Notably, in the sensitivity analysis conducted after using a uniform definition of > 6.8 mg/dL for men and women, there were similar results as the primary analysis. The ORs of hyperuricemia for moderate and vigorous PA groups were 0.84 (95% CI 0.75–0.94) and 0.80 (95% CI 0.71–0.91), respectively. RCS has shown a U-shaped nonlinear relationship between PA and hyperuricemia, and a linear relationship between total ST and hyperuricemia. There was no association between ST and hyperuricemia in men, but ST ≥ 4 h/day was a hazard for women. Compared with employees walking/riding a bicycle to work, doing housework or being disabled to work was more likely to have hyperuricemia (OR 1.28, 95% CI 1.06–1.55 among men and OR 1.52, 95% CI 1.17–2.00 among women). The interaction of ST and PA was also not statistically significant (P for interaction > 0.05; data not shown).
DISCUSSION
In this multiethnic sample, the major findings are the significant positive association of ST and the significant negative association of PA with hyperuricemia. However, we did not observe the combined effect of ST with PA. In particular, we observed a U-shaped relationship between PA and hyperuricemia and a dose-response association of ST with hyperuricemia among all participants. Meanwhile, commuting mode was associated with hyperuricemia. However, sex stratification revealed a sex difference in sedentary behavior and hyperuricemia.
It is generally accepted that, apart from PA, sitting time is an independent factor affecting the health outcome.27 A cross-sectional study including 161,064 participants showed that those who spent ≥ 10 h/day were more likely to have hyperuricemia than those who spent < 5 h/day engaging in sedentary behavior.14 Cui et al reported that prolonged sitting was an independent risk factor for hyperuricemia.28 In our study, participants who spent total ST ≥ 4 h/day had a positive association of hyperuricemia, and the dose-response analyses demonstrated a linear relationship between the risk of hyperuricemia prevalence and ST. Unlike the total ST, there was no association between leisure ST ≥ 4 h/day and hyperuricemia, but there was a positive association for leisure ST ≥ 6 h/day; the reason could be that leisure ST usually involves intermittent sedentary behavior. Playing mahjong, watching TV, and mobile entertainment on cell phones are common leisure time behaviors in rural Chinese areas, and typically take the form of intermittent sitting. Previous studies showed that when comparing continuous sedentary behavior and periodic sedentary behavior, the related physiological indicators such as decreased blood glucose and increased insulin sensitivity, C-peptide, and triglyceride content improved in those with intermittent sedentary behavior.29,30,31,32,33 The commuting mode also further indicated that prolonged sitting and exercising less will increase the risk of hyperuricemia. Compared with commuting to work by walking or riding a bicycle, participants working at home or nearby and those performing housework or disabled to work were more likely to have hyperuricemia. Only 1 study focused on transportation type on the risk of hyperuricemia, and found that women were influenced by the commuting method.14 This may be related to transportation mode affecting obesity, and obesity affecting hyperuricemia. The exact mechanism underlying the effects of sedentary behavior remains unclear. Previous studies had shown that sedentary behavior and hyperuricemia were associated with obesity and insulin resistance.34 Increasing insulin levels in the blood promote the reabsorption of sodium in the renal tubules, thereby reducing uric acid clearance in the renal tubules and increasing uric acid levels in the blood, and possibly affecting the cellular processes responsible for metabolic abnormalities.35,36,37 Additionally, the pentose phosphate pathway activated during prolonged sitting was associated with the excitation of xanthine oxidase, increasing SU production.38
There was a U-shaped relationship between PA and hyperuricemia. Regular participation in PA can help increase muscle mass and decrease all-cause mortality in men, according to previous studies.39 However, in clinical trials, production of SU was increased after strenuous exercise, leading to a transiently higher level of SU.40 Dehydration because of high PA and excessive sweating can lead to elevated uric acid and precipitate gout in susceptible individuals.41 Nishida et al also found that SU production was related to the intensity of PA. Compared to heavy exercise, moderate regular exercise only produces a minimal elevation of SU, resulting in beneficial effects for the prevention of hyperuricemia.42
We did not find an interaction of ST and PA in our study. A recent study found that combined light PA and a higher ST (≥ 8 h/day) may significantly increase the risk of hyperuricemia.15 A cross-sectional study among Mexican Americans reported that the beneficial effect of PA on subclinical atherosclerosis was offset when ST exceeded 3 h/day.43 Moreover, positive health outcomes provided by regular exercise disappeared and chances of chronic illness are increased when the daily average ST was exceeded.44 The discrepant findings on the interaction of ST and PA on hyperuricemia may be partly explained by the differences in the sample size, participant characteristics, adjustment for confounders, and methods used to assess ST and PA.
On the other hand, in our present study, we observed different effects in men and women in the relationship of sedentary behavior with hyperuricemia. The association of total ST and commuting mode on hyperuricemia seemed to be stronger in women than in men. Conversely, a cross-sectional study in the Henan Rural Cohort found similar associations of ST with hyperuricemia in both males and females.15 The reasons for this sex difference are likely complex. Efforts to perform daily activities in different social roles are presented differently for women and men and can influence their behaviors in health maintenance. Banks et al found after adjusting confounding factors that sedentary behavior was associated with obesity in women, whereas there was no correlation in men.45 Similarly, in an international study including adults from 10 geographically and culturally diverse countries, ST was curvilinearly associated with BMI in women from the US sites included in the study, but not in men or adults from the other 9 countries.46 Suminski and colleagues found ST was positively associated with body fat percentage only in women who did not meet the PA guidelines; no such associations were identified in men who were insufficiently active.47 Results from a prospective cohort of 146,000 individuals showed that longer leisure time spent sitting was associated with a higher total cancer risk in women but was not associated with cancer risk in men.48 Our study showed a similar sex difference in hyperuricemia. It is plausible that the metabolic alterations due to sedentary behavior are not substantial enough to affect hyperuricemia risk in men or that there remains some residual confounding by obesity or other factors.
Our study had some limitations. First, a cross-sectional study could not establish a cause-and-effect relationship between PA, sedentary behavior, and hyperuricemia. Second, reporting bias could not be avoided because many factors were self-reported. Third, some residual confounding factors may have affected our estimations, especially since we lacked gout as a covariate in the analysis. Gout can cause activity limitation (particularly during gout flares), and it can affect physical function, including less PA and more sedentary behavior. Unfortunately, we lacked the ability to diagnose gout at baseline for participants of the large-scale cohort. Some published studies also did not use gout as a covariate when analyzing the relationship between hyperuricemia and PA,15,16,49 allowing our results to be cross-referenced. Considered together, the relationship between ST and PA on hyperuricemia in this study was clear. Despite these limitations, our findings may help develop strategies to prevent and manage hyperuricemia in China.
In summary, this study in a multiethnic Chinese population demonstrates a positive linear dose-response association between ST and the risk of hyperuricemia and a U-shaped nonlinear relationship between PA and hyperuricemia. These findings suggest that moderate PA and low ST may be less likely to cause hyperuricemia. Sex played a role in the association between ST and hyperuricemia, with a positive association of ST with hyperuricemia in women but not in men. We considered a stronger effect on hyperuricemia in women vs men in terms of reducing ST. Longitudinal studies are required to explore the interpretation of cause-and-effect relationships.
Footnotes
This study was supported by the National Key R&D Program of China (Grant no. 2017YFC0907302), National Natural Science Foundation of China (Grant no. 81860597), Young Talents Special Project of Ten Thousand Talents Program of Yunnan Province (Grant no. YNWR-QNBJ-2019-042), and Kunming Medical University Biological Resources Digital Development and Application Project (Grant no.202002AA100007).
R. Hong, J. Huang, and C. Xu are joint first authors.
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication January 26, 2022.
- Copyright © 2022 The Journal of Rheumatology
