Association between change in serum uric acid and rapid decline in kidney function in China

Study design and population

The detailed study design and methods of CHARLS have been detailed elsewhere²⁰. In summary, CHARLS is a nationally representative longitudinal study that focuses on the middle-aged and elderly population in China. It aims to collect comprehensive data on demographics, socioeconomic status, lifestyle habits, and health conditions. The initial survey encompassed a cohort of 17,708 individuals, all aged 45 years or older, hailing from 450 villages spread across 28 provinces. The data collection spanned from June 2011 to March 2012 and was executed through a multistage, stratified sampling technique, employing a probability proportional to the size of the population. Each participant was subjected to an in-depth assessment through personalized interviews using a standardized questionnaire. Following the baseline survey, participants have undergone biennial in-person follow-up interviews. During each follow-up phase, comprehensive physical examinations and systematic blood sample collections have been conducted at regular intervals. All procedures performed in this study were in accordance with the Declaration of Helsinki (revised in 2013) and were approved by the Ethics Committee of Peking University (IRB 00001052-11014).

We conducted a prospective, longitudinal analysis using data from the two waves of CHARLS for 2011, and 2015. A total of 5421 participants with complete measurements on baseline SUA, baseline and exit visit kidney outcomes were enrolled in this analysis (Supplemental Fig. 1).

Assessment of covariates

At baseline, information on age, gender, smoking and drinking status (no or yes), living residence (rural or urban), marital status, household income, and educational level were obtained from the questionnaires. Household income and educational level were classified as being at the mean income level or higher and secondary education or above, respectively. Marital status was classified into two groups: married and single (never married, separated, divorced, and widowed). Health-related factors included self-reported physician-diagnosed hypertension, diabetes, dyslipidemia, and heart disease. Body mass index (BMI) was calculated as weight in kilograms divided by height in square meters.

Laboratory assays

Venous blood specimens were procured by skilled personnel and dispatched to the nearby laboratory under refrigerated conditions at 4 °C. Upon arrival, the samples were promptly cryopreserved at − 20 °C. Within a fortnight, they were shipped to the Chinese Center for Disease Control and Prevention’s headquarters in Beijing, where they were maintained at an ultra-low temperature of − 80 °C until they were subjected to analysis at the Capital Medical University laboratory. The serum glucose and lipid profiles were quantified through an enzymatic colorimetric assay. The concentration of Hemoglobin A1c (HbA1c) was ascertained via a boronate affinity high-performance liquid chromatography technique. SUA levels were determined by employing a uric acid plus methodology. The presence of high-sensitivity C-reactive protein (CRP) was evaluated using an immunoturbidimetric method. Both Scr and cyc were measured using a rate-blanked, compensated Jaffe creatinine assay and a particle-enhanced turbimetric assay, respectively. It is noteworthy that in the CHARLS, blood sampling was conducted during the study visits of 2011 and 2015. Consequently, the assessment of Scr and cyc in our research was confined to the baseline year of 2011 and the final visit in 2015.

Definition of rapid eGFR decline

Consequently, the eGFR (mL/min/1.73 m²) was determined using serum levels of cystatin C and Scr, along with gender and age data. This calculation was executed utilizing the 2021 race-free Chronic Kidney Disease Epidemiology Collaboration equation (CKD-EPI)²¹:

Scr = standardized serum creatinine in mg/dL, κ = 0.7(females) or 0.9(males), α = − 0.219(females) or − 0.144(males), min(Scr/κ, 1) is the minimum of Scr/κ or 1.0, max(Scr/κ, 1) is the maximum of Scr/κ or 1.0, cys = standardized serum cystatin C in mg/L, Age (years).

A rapid decline in kidney function, defined as an annualized decrease in eGFRcr-cys of 5 mL/min per 1.73 m² or more, or an eGFR in 2011 is greater than 15 mL/min per 1.73 m² and less than 15 mL/min per 1.73 m² in 2015²².

Statistical analyses

The baseline characteristics of the participants were examined through descriptive statistical methods. Continuous variables, adhering to a normal distribution, were depicted as means accompanied by standard deviations (SD). In contrast, categorical variables were expressed as counts (and their corresponding percentages).

Comparisons among the four groups, based on the quartile of change in eGFR, were performed using one-way analysis of variance (ANOVA) with Tukey’s post hoc method for multiple comparisons. The risk factor for rapid eGFR decline was evaluated by univariable and multivariable logistic regression models.Multivariable logistic regression models included the following covariates: age, gender, smoking, drinking status, living residence, marital status, household income, BMI, hypertension, diabetes status, dyslipidemia, CRP, Hba1c, baseline eGFR, baseline SUA, and change in SUA. The ROC curve delineates a model’s performance by plotting the True Positive Rate against the False Positive Rate across different thresholds. The area under the curve (AUC) represents the model’s discriminatory power, with higher values indicating better classification accuracy. Due to gender difference in SUA distribution, we also stratified our multivariable regression analyses by gender. The correlations between changes in SUA and changes in eGFR during the follow-up period were investigated using Pearson’s correlation analysis, with stratification by gender.

The statistical methodologies employed in our study were predicated on a two-tailed testing paradigm, with the threshold for statistical significance established at a p-value below 0.05. Our computations were executed within the R Studio environment, specifically the 2023.12.0.369 iteration, an offering of the R Foundation for Statistical Computing, an esteemed institution headquartered in Vienna, Austria.