BACKGROUND

Insulin resistance (IR) is central to metabolic syndrome and contributes to the development of type 2 diabetes mellitus (T2DM) as well as mild cognitive impairment (MCI). Several low-cost surrogate markers have been proposed to assess IR, such as the triglyceride-glucose (TyG) index, TyG-BMI, TG/HDL-C, metabolic score for insulin resistance (METS-IR), and single-point insulin sensitivity estimator (SPISE). This study aimed to develop nomogram models integrating these indices with clinical data to predict MCI in patients with T2DM.

METHODS

A total of 600 patients diagnosed with T2DM were recruited. Demographic, clinical, and biochemical parameters were documented, and cognitive performance was assessed using the Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE). Logistic regression analyses identified predictors of MCI, and receiver operating characteristic (ROC) curves evaluated their predictive accuracy. Two nomogram models were constructed: Model 1 included METS-IR, age, sex, education level, and hypertension; Model 2 substituted SPISE for METS-IR, retaining other clinical variables.

RESULTS

All IR surrogate indices were significantly associated with MCI and reduced MMSE scores (P < 0.001). METS-IR and SPISE exhibited higher predictive accuracy (AUC: METS-IR = 0.809, SPISE = 0.805) compared to TyG, TyG-BMI, and TG/HDL-C, particularly among female participants. Nomogram models demonstrated robust predictive performance (AUC: Model 1 = 0.846; Model 2 = 0.838).

CONCLUSIONS

Nomogram models incorporating METS-IR or SPISE alongside key clinical parameters effectively predicted the risk of MCI among patients with T2DM. These indices notably outperformed other surrogate markers, highlighting their clinical value for early assessment of cognitive risk.