Researchers initially screened 69,152 adults who had undergone at least two routine health examinations between 2014 and 2025. Eligibility required a baseline estimated glomerular filtration rate (eGFR) of at least 60 mL/min/1.73 m² and at least two follow-up examinations separated by intervals of three months or more. People with baseline eGFR below 60, evidence of acute infection, specified kidney-related diagnoses or a follow-up interval shorter than three months were excluded, leaving 48,304 participants. CKD was defined as two consecutive eGFR measurements below 60 mL/min/1.73 m², at least three months apart.

The investigators calculated 12 indices from anthropometric, glucose and lipid measurements: TyG, triglyceride-to-high-density lipoprotein cholesterol ratio, metabolic score for insulin resistance (METS-IR), TyG-BMI, TyG-WC, TyG combined with waist-to-height ratio, metabolic score for visceral fat, lipid accumulation product, Chinese visceral adiposity index, atherogenic index of plasma, cholesterol–high-density lipoprotein–glucose index (CHG) and remnant cholesterol. Their pairwise Spearman correlations ranged from 0.36 to 1.00, indicating substantial overlap among the measures, so the findings do not establish them as independent biomarkers.

Over a median follow-up of 4.74 years, 1,092 participants, or 2.26% of the cohort, developed CKD, corresponding to 4.56 cases per 1,000 person-years. In fully adjusted Cox models, every index showed a graded association with incident CKD across quartiles, with all tests for trend below 0.001. Comparing the highest with the lowest quartile, the adjusted hazard ratios were 2.48 (95% confidence interval 1.89–3.26) for TyG-WC, 2.41 (1.89–3.08) for METS-IR and 2.39 (1.79–3.18) for the Chinese visceral adiposity index. The TyG-WC association was consistent across prespecified subgroups, with no significant interaction tests.

Restricted cubic spline analyses indicated that the relationship between TyG-WC and incident CKD was nonlinear, with an inflection point at 811.58. Below this value, each standard-deviation increase in TyG-WC was associated with a higher CKD hazard (HR 1.71, 95% CI 1.47–1.98); above it, the association remained statistically significant but was attenuated (HR 1.10, 95% CI 1.07–1.13). The investigators described 811.58 as a statistical threshold rather than a validated biological cutoff, and noted that reverse causation cannot be excluded.

Among participants who developed CKD, 54 experienced rapid kidney function decline and 17 reached a clinical surrogate endpoint during a median 2.8 years of follow-up. Rapid decline was defined as an annual eGFR decrease of at least 5 mL/min/1.73 m², while surrogate endpoints comprised a 40% eGFR decline or a doubling of serum creatinine from the value recorded at CKD diagnosis. CHG was associated with rapid decline (odds ratio 1.56 per standard-deviation increase, 95% CI 1.23–1.96), whereas TyG-BMI was associated with surrogate endpoints (HR 2.55, 95% CI 1.38–4.71).

TyG-WC also had the highest measured discrimination for CKD at one year, with a time-dependent area under the curve of 0.817 (95% CI 0.776–0.851), although differences among the 12 indices narrowed over later follow-up. Adding TyG-WC to a baseline model improved three-year risk reclassification, with continuous net reclassification improvement of 0.192 (95% CI 0.072–0.318; p=0.012), while the integrated discrimination improvement was small at 0.001. Decision-curve analyses suggested potential clinical utility, and bootstrap internal validation produced similar optimism-corrected estimates.

Seven sensitivity analyses broadly supported the primary results. The TyG-WC association remained after excluding intervals longer than two years between eGFR measurements, applying a stricter CKD definition, excluding medicated participants, adding further covariates and excluding participants with a BMI of at least 28 kg/m², although some estimates weakened after the BMI restriction. An exploratory mediation analysis estimated that absolute monocyte count accounted for 4.76% of the TyG-WC association with CKD (95% CI 3.5–7.8%); the observational analysis cannot establish a causal mediation pathway.

Important limitations include the single-center, health-examination-based population, which may limit generalizability, and reliance on creatinine-based eGFR without systematic cystatin C or urinary albumin-to-creatinine measurements; urinary albumin data were unavailable for more than 99% of participants. The small number of progression events limited statistical power and covariate adjustment, mortality data were unavailable for competing-risk analyses, and progression was assessed through rapid eGFR decline and surrogate endpoints rather than detailed CKD stages. The observational design also leaves possible bidirectional relationships between metabolic measures and kidney function unresolved.

Subject of Research: Associations between surrogate insulin resistance indices and CKD incidence and progression in a Chinese health-examination cohort

Article Title: Stage-specific associations of nutrition-related surrogate insulin resistance indices across the chronic kidney disease continuum: evidence from a Chinese cohort

Article References: Fang, S., Li, Y., He, Y., Jiang, J., Zhu, Y., Chen, H., & Leng, S. (2026). Stage-specific associations of nutrition-related surrogate insulin resistance indices across the chronic kidney disease continuum: evidence from a Chinese cohort. Frontiers in Nutrition, 13, Article 1943087. https://doi.org/10.3389/fnut.2026.1943087

Image Credits: AI Generated

DOI: 10.3389/fnut.2026.1943087

Keywords: chronic kidney disease, insulin resistance, surrogate indices, waist circumference, triglyceride-glucose index, renal progression, retrospective cohort, risk prediction, Chinese health cohort, Stage-specific, associations, nutrition-related