Introduction:
Type 2 diabetes (T2D) and mild cognitive impairment (MCI) are interrelated conditions that significantly impair quality of life. This study aimed to identify a feasible biomarker for assessing T2D-MCI risk and to evaluate a potential therapeutic strategy.

Material and methods:
We integrated data from the National Health and Nutrition Examination Survey (NHANES) with Mendelian randomization (MR) to investigate genetic causal relationships between T2D, MCI, and their shared biomarkers. Transcriptomic analysis identified T2D-associated genes. Clinical trials evaluated the short-term effects of modified fasting therapy (MFT) on glucose regulation and cognitive function. Cellular assays and patient samples validated key genes’ roles in biochemical markers and downstream pathways.

Results:
Among 6,356 T2D and 1,138 MCI subjects, vitamin D, high-density lipoprotein cholesterol (HDL-C), globulin, and creatinine were associated with both conditions. MR analysis showed that higher HDL-C levels reduced T2D risk (0.9059, 95% CI: 0.8666–0.9470) but increased MCI risk (OR = 1.0482, 95% CI: 1.0216–1.0755). Nuclear Factor I A (NFIA) was identified as a key HDL-C regulator. In a clinical cohort (17 T2D patients and 23 controls), MFT reduced body mass index fasting glucose, and HDL-C, increased NFIA expression, and improved Montreal Cognitive Assessment scores, especially in T2D-MCI patients. HDL-C rebounded at six months. In vitro, NFIA overexpression increased intracellular HDL-C and suppressed NF-κB signaling, while NFIA knockdown reduced APOA1 and APOE.

Conclusions:
HDL-C has divergent genetic effects on T2D and MCI. NFIA modulates HDL-C and NF-κB activity, supporting metabolic and cognitive improvements. Targeting NFIA through MFT may represent a promising strategy for T2D-MCI prevention and treatment.