Association of atherogenic index of plasma with kidney dysfunction in diabetic individuals: findings from two national population-based studies

Background Extensive evidence suggests that dyslipidemia is associated with endothelial dysfunction, oxidative stress, and inflammation, all of which can contribute to kidney dysfunction. The atherogenic index of plasma (AIP) is a novel marker of lipid metabolism disorder, but its role in kidney dysfunction in diabetic individuals remains controversial. This study aims to clarify the association of AIP with kidney dysfunction in diabetic individuals. Methods This cross-sectional study analyzed a representative sample of participants aged 20 years and older from the United States (n = 2,386, NHANES 2007-2018) and Korea (n = 698, KNHANES 2012). Weighted multivariate logistic regression analyses and smoothed curve fitting were conducted to investigate the relationship between logarithmically transformed AIP (lgAIP) and multiple kidney dysfunction, including albuminuria and low estimated glomerular filtration rate (eGFR) in diabetic individuals. Additionally, we conducted interaction analyses and subgroup analyses to assess whether this relationship remained consistent across different populations. We utilized receiver operating characteristic (ROC) curves to assess and compare the diagnostic performance of AIP and other lipid indices for kidney dysfunction. Results In both databases, higher lgAIP was significantly associated with the occurrence of albuminuria in diabetic individuals (NHANES: OR = 7.69, 95%CI: 2.90-20.40; KNHANES: OR = 6.00, 95%CI: 1.05-34.36) in the fully adjusted model. However, the OR (95% CI) for the association between lgAIP and low-eGFR was 1.22 (0.33, 4.53) in the NHANES database and 2.50 (0.16, 38.62) in the KNHANES database, indicating no statistically significant association. Subgroup analysis revealed that the association between lgAIP and albuminuria in diabetic individuals was influenced by age and BMI stratification in the NHANES database, and by BMI stratification in the KNHANES database (p for interaction < 0.05). Compared to other lipid indicators, AIP appears to be more precise and discriminatory in predicting albuminuria in diabetic individuals. Conclusion Our findings highlight a strong association between lgAIP and albuminuria in diabetic individuals. Future research should explore the mechanisms that underlying this relationship.