The study analyzed 8,358 hospital participants, including 1,800 with CKD, and 9,182 NHANES participants, including 2,751 with CKD. The hospital data came from routinely collected records at the Department of Cardiology of the First Affiliated Hospital of Dali University in Yunnan Province, covering March 2022 through June 2025. The NHANES component used survey periods from 1999–2000 through the 2017–March 2020 pre-pandemic release and incorporated survey weights, strata, and primary sampling units. Neither component was a prospective cohort, intervention, review, or animal study.

The primary exposure was the NLR–TyG index, or NTI, calculated as the product of the neutrophil-to-lymphocyte ratio and the triglyceride–glucose index. The neutrophil-to-lymphocyte ratio was calculated from absolute neutrophil and lymphocyte counts, while the triglyceride–glucose index was calculated as ln[triglycerides in mg/dL × fasting glucose in mg/dL/2]. The product was prespecified as a simple summary of concurrent inflammatory and metabolic burden. The investigators cautioned that its multiplicative form did not itself demonstrate biological or statistical interaction between its components.

Compared with the lowest cohort-specific NTI quartile, the highest quartile was associated with higher adjusted odds of CKD in the hospital population (OR, 2.85; 95% CI, 2.38–3.40) and in NHANES (OR, 1.93; 95% CI, 1.56–2.38; both p for trend<0.001). Restricted cubic splines supported an overall association in both populations. The fully adjusted relationship was nonlinear in the hospital cohort, where the increase in odds became more pronounced at higher NTI values, whereas no statistically significant departure from linearity was detected in NHANES after full adjustment (p for nonlinearity=0.075).

The models were adjusted for age, sex, race or ethnicity, education, marital status, smoking, drinking, body mass index, systolic blood pressure, diabetes, and heart failure. Restricting the analysis to participants with known non-use of lipid-lowering therapy did not remove the association. In that sensitivity analysis, each 1-unit increase in NTI was associated with adjusted odds ratios of 1.012 (95% CI, 1.009–1.015) in the hospital cohort and 1.026 (95% CI, 1.016–1.036) in NHANES. The highest-versus-lowest quartile comparisons were 2.79 (95% CI, 2.28–3.42) and 2.05 (95% CI, 1.58–2.65), respectively.

Subgroup analyses generally preserved the direction of the association across age, sex, race or ethnicity, smoking, drinking, diabetes, and body mass index categories. Formal interaction patterns, however, differed between populations. The hospital analysis indicated heterogeneity by sex and body mass index, while NHANES indicated heterogeneity by age and race or ethnicity. Because these patterns were not reproduced across both cohorts, the researchers characterized the subgroup findings as exploratory and hypothesis-generating rather than established biological effect modification.

Adding NTI to a basic clinical model produced statistically detectable but modest improvements in discrimination. The C statistic increased from 0.769 to 0.778 in the hospital population, a change of 0.009 (95% CI, 0.007–0.014), and from 0.772 to 0.778 in NHANES, a change of 0.005 (95% CI, 0.003–0.008). A model containing NLR and TyG as separate terms produced nearly identical discrimination, with a C statistic of 0.779 in both cohorts. Formal NLR-by-TyG interaction tests were not significant in either population, so the findings did not show that multiplying the components created multiplicative synergy or that the composite was superior to modeling them separately.

Several limitations constrain the interpretation. The observational, cross-sectional design and retrospective hospital component could not establish whether higher NTI preceded CKD, leaving temporality, reverse causation, residual confounding, and selection bias unresolved. CKD ascertainment also differed: NHANES defined it using estimated glomerular filtration rate or urine albumin-to-creatinine ratio, whereas the hospital data lacked albumin measurements and used estimated filtration rate or recorded diagnoses. Single-occasion measurements of inflammatory and metabolic markers may not represent long-term status, and acute illness could have influenced hospital measurements. The inpatient and community populations also had different clinical profiles, so their absolute NTI distributions and effect estimates should not be assumed to transfer directly between settings.

Overall, the study found that higher NTI was independently associated with prevalent CKD in two distinct hypertensive populations and offered a compact representation of inflammatory–metabolic burden with modest incremental discrimination. These cross-sectional findings do not establish causality or prospective predictive utility. The investigators stated that longitudinal validation with repeated biomarker measurements, harmonized CKD ascertainment, and external validation would be needed before NTI could be considered for routine clinical risk assessment.

Subject of Research: Cross-sectional association between the NLR–TyG index and prevalent chronic kidney disease in hypertensive adults

Article Title: NLR–TyG index and prevalent chronic kidney disease in hypertension: a cross-population study

Article References: Li, H., Zhao, C., Zhao, M., & Wu, X. (2026). NLR–TyG index and prevalent chronic kidney disease in hypertension: a cross-population study. Frontiers in Nutrition, 13, Article 1917577. https://doi.org/10.3389/fnut.2026.1917577

Image Credits: AI Generated

DOI: 10.3389/fnut.2026.1917577

Keywords: chronic kidney disease, hypertension, NLR–TyG index, neutrophil-to-lymphocyte ratio, triglyceride–glucose index, NHANES, inflammation, metabolic burden, cross-sectional study, index, prevalent, chronic