Objective:

To evaluate the screening value of the triglyceride-glucose index (TyG) combined with the systemic immune-inflammation index (SII) for diabetic kidney disease (DKD) in hospitalized patients with type 2 diabetes mellitus (T2DM).

Key Findings:

• TyG and lgSII were significantly higher in the DKD group compared to the non-DKD group (P<0.001), indicating a strong association.
• The final streamlined model achieved an AUC of 0.850, indicating good predictive performance, which is crucial for clinical applicability.
• Optimal cut-off values were TyG >10.03 and lgSII >2.76, with sensitivity of 74.3% and specificity of 85.0%, providing actionable thresholds for clinicians.
• The high-risk group (both indicators above cut-offs) had a DKD prevalence of 96.8%, underscoring the model's potential for identifying at-risk patients.
• Bootstrap validation yielded a mean AUC of 0.855, indicating robust model stability, which is essential for clinical trust.

Interpretation:

The combination of TyG and SII demonstrates favorable screening efficacy and model stability for identifying DKD among hospitalized T2DM patients.

Limitations:

• The study requires external validation before clinical implementation, and potential biases related to the retrospective design should be acknowledged.

Conclusion:

The model provides preliminary evidence for a low-cost, readily available tool for rapid risk stratification, but further research is necessary to confirm its efficacy in diverse populations.