Linear association between atherogenic index of plasma and diabetic kidney disease: evidence from a large clinical cohort and NHANES
Background Diabetic kidney disease (DKD) is a prevalent microvascular complication of type 2 diabetes mellitus (T2DM) and contributes substantially to end-stage renal disease. The atherogenic index of plasma (AIP) has exhibited biomarker utility. However, its relationship with DKD remains uncertain. Methods This cross-sectional investigation included 10,112 T2DM subjects from the National Metabolic Management Center of Yuhuan Second People’s Hospital between September 2017 and February 2025 and 5,573 individuals in NHANES cohort (1999–2020). AIP values were calculated and categorized into quartiles. Multivariable logistic regression, restricted cubic spline, and stratified subgroup analyses were applied for assessing the link between AIP and DKD. The area under the curve (AUC), net reclassification improvement (NRI) and integrated discrimination improvement (IDI) was used to assess the discriminative ability of AIP. Results After adjustment for confounding factors, each one-unit AIP rise resulted in a 72% higher likelihood of DKD (OR: 1.72, 95% CI: 1.51–1.96, p < 0.001). Relative to the lowest AIP quartile, the highest quartile showed a 54% higher odds DKD (OR: 1.54, 95% CI: 1.36–1.75, p < 0.001), with a clear dose–response effect (P for trend <0.001). Spline modeling indicated a positive linear link between AIP and DKD, which was apparent in all examined subgroups (all P for interaction >0.05) and externally validated in NHANES. Incorporating AIP into the baseline model modestly improved discrimination, increasing the area under the curve from 68.42 to 68.76%, with a continuous NRI of 0.118 (95% CI: 0.079–0.157, p < 0.001) and an IDI of 0.006 (95% CI: 0.004–0.007, p < 0.001). Conclusion AIP shows an independent positive association with DKD in individuals with T2DM, demonstrating a linear dose–response relationship, but the incremental predictive value is limited. The association between AIP and DKD still needs to be further validated by larger-scale prospective studies.