The baseline clinical data included in this analysis were sourced from individuals who underwent health examinations at the Health Promotion Center of Sir Run Run Shaw Hospital, Zhejiang University, in Hangzhou, China, from January 2017 to December 2019. A total of 169,964 individuals were initially included after excluding patients with malignancies, cerebral hemorrhage, cerebral infarction, heart disease, liver dysfunction, end-stage renal disease, and autoimmune diseases. Subsequently, participants with missing baseline clinical data were excluded, leaving a final sample of 97,645 individuals.

The primary outcome measure in this study was the presence or absence of kidney stones in the subjects. Throughout the study, renal ultrasonography (UTUS) was conducted by trained radiologists using the same model of ultrasound machines from the United States, equipped with 3.0–5.0 MHz frequency transducers. Detailed records were maintained for stone size, quantity, location, degree of renal hydronephrosis, and other urinary tract abnormalities.

The main exposure factor of interest was the Chinese visceral adiposity index (CVAI), which was utilized as the primary variable. CVAI was calculated based on gender-specific mathematical models as follows: for females, CVAI=-187.32 + 1.71×age + 4.23×BMI + 1.12×WC + 39.76×lgTG-11.66×HDL-C, and for males, CVAI=-267.93 + 0.68×age + 0.03×BMI + 4.00×WC + 22.00×lgTG-16.32×HDL-C. The calculated CVAI values are presented in Table 1. CVAI primarily reflects the visceral fat content within the body, with higher CVAI values indicating greater visceral fat content and a higher predicted incidence of cardiovascular disease.

The medical history was systematically collected by well-trained general practitioners at Zhejiang University Affiliated Sir Run Run Shaw Hospital. It included comprehensive information such as chief complaints, current medical conditions, past medical history, personal history, family history, and physical examinations. Alcohol consumption was categorized into current drinkers (those who consumed alcohol daily for more than 6 months) and non-current drinkers. Smoking status was divided into current smokers (those who smoked daily for more than 6 months) and non-current smokers.

Measurements of body weight, height, blood pressure (BP), and waist circumference (WC) were taken by trained nurses. The Body Mass Index (BMI) was calculated as the ratio of weight (in kg) to the square of height (in m^2). Various laboratory tests were conducted, including assessments of total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), blood urea nitrogen (BUN), serum creatinine (CR), and serum uric acid (UA). Urinalysis was performed at the hospital laboratory and included parameters such as urine pH, specific gravity, red blood cells, white blood cells, protein, bacteria, and more.

Continuous variables are reported as mean ± standard deviation (SD), while categorical variables are presented as proportions. Analytical comparisons utilized t-tests for continuous variables and chi-square tests for categorical variables.

The association between CVAI and kidney stones was explored through logistic regression models. Model 1 remained unadjusted, while Model 2 incorporated adjustments for sex, smoking, alcohol consumption, hypertension, and diabetes. Model 3 extended these adjustments to include systolic blood pressure, diastolic blood pressure, serum uric acid (UA), glycosylated hemoglobin(HbA1c), blood urea nitrogen (BUN), serum creatinine (CR), urine specific gravity, uric pH, and urine protein, building upon Model 2.

To comprehensively investigate the relationship between CVAI and kidney stones, multivariable logistic regression was performed. CVAI was considered both as a continuous variable and categorized into four quartiles. Trends were assessed by treating CVAI quartiles as continuous variables. Additionally, we proceeded to examine if there existed a non-linear relationship between CVAI and the likelihood of kidney stones through the utilization of a generalized additive model (GAM) and curve fitting. If such a relationship was identified, we applied a two-piecewise linear regression model to determine the threshold effect of CVAI on kidney stones, based on the smoothing plot. We employed a recursive technique to automatically ascertain the inflection point, leveraging the maximum model likelihood. Subgroup analyses were conducted using hierarchical logistic regression models, encompassing all potential confounding factors outlined in the baseline table.

The statistical analyses were carried out using R version 4.0.3 software (http://www.R-project.org/) and relevant packages, including “mgcv”, “visreg”, and “broom”. A two-sided p-value < 0.05 was considered statistically significant, ensuring a robust evaluation of the associations observed in the study.