Association between different insulin resistance indices and all-cause mortality in patients with diabetic kidney disease: a prospective cohort study.

Aim: Previous research has shown a strong association between insulin resistance (IR) and both the onset and advancement of diabetic kidney disease (DKD). This research focuses on examining the relationship between IR and all-cause mortality in individuals with DKD. Methods: This study utilized data obtained from the National Health and Nutrition Examination Survey (NHANES), spanning the years 2001 to 2018. Insulin resistance was assessed using reliable indicators (HOMA-IR, TyG, TyG-BMI, and METS-IR). The relationship between IR indices and survival outcomes was evaluated through weighted multivariate Cox regression, Kaplan-Meier survival analysis, and restricted cubic spline (RCS) modeling. To examine non-linear associations, the log-likelihood ratio test was employed, with piecewise regression models used to establish confidence intervals and identify threshold values. Diagnostic precision and efficacy were gauged using Receiver Operating Characteristic (ROC) curves, Area Under the Curve (AUC) evaluations, and calibration plots. Moreover, to verify the consistency of our results, stratified analyses and interaction tests were conducted across variables including age, gender, Body Mass Index (BMI), hypertension, and cardiovascular status. Results: This research involved a group of 1,588 individuals diagnosed with DKD. Over a median observation period of 74 months, 630 participants passed away. Using weighted multivariate Cox regression along with restricted cubic spline modeling, we identified non-linear associations between the four insulin resistance indices and all-cause mortality. An analysis of threshold effects pinpointed essential turning points for each IR index in this research: 1.14 for HOMA-IR, 9.18 for TyG, 207.9 for TyG-BMI, and 35.85 for METS-IR. It was noted that levels below these thresholds inversely correlated with all-cause mortality. In contrast, values above these points showed a significantly positive correlation, suggesting heightened mortality risks. The accuracy of these four IR metrics as indicators of all-cause mortality was confirmed through ROC and calibration curve analyses. Conclusion: In patients with DKD, an L-shaped association is noted between HOMA-IR and all-cause mortality, while TyG, TyG-BMI, and METS-IR exhibit U-shaped relationships. All four IR indices show good predictive performance.