Causal relationship and shared genetic pathways between diabetic kidney disease and cognitive impairment: a Mendelian randomization study

Abstract

Background: Diabetic kidney disease (DKD) may increase cognitive impairment (CI) risk, but causal evidence remains limited. We aimed to investigate this causality and elucidate underlying genetic mechanisms.

Methods: Bidirectional two-sample Mendelian randomization (MR) was performed using summary statistics from four DKD genome-wide association studies (GWAS) and a large-scale cognitive function GWAS. Sensitivity analyses (heterogeneity, pleiotropy, leave-one-out, and Steiger directionality tests) confirmed the robustness of the results. Multivariable MR was used to adjust for potential confounders. Genetic correlation was assessed via linkage disequilibrium score regression (LDSC). Shared loci and pathways were identified through colocalization and functional enrichment analyses.

Results: DKD significantly increased cognitive decline risk (inverse-variance weighted (IVW) OR range: 0.55-0.88; all p < .05), with consistent results across sensitivity analyses. Multivariable MR confirmed that the association was independent of confounders. Significant genetic correlations were observed (LDSC rg = 0.072-0.201). Colocalization identified six shared risk loci (posterior probability for H4 (PP.H4) > 0.90). Enriched pathways included ribosomal function, mitochondrial oxidative phosphorylation, and neurodegeneration.

Conclusions: This study provides evidence supporting a causal relationship and genetic correlation between DKD and CI, while also identifying shared genetic features and biological pathways that may contribute to their association.

Keywords: Diabetic kidney disease; Mendelian randomization; cognitive impairment; genetic correlation; genome-wide association study.

Graphical abstract

Figure 1.

Schematic overview of the study design and experimental workflow.

Figure 2.

The FUMA analysis revealed significant tissue enrichment of cognition-related GWAS signals, primarily localized in 14 distinct neural structures, including brain cerebellar hemisphere, brain cerebellum, brain cortex, brain frontal cortex BA9, brain anterior cingulate cortex BA24, brain nucleus accumbens basal ganglia, brain hypothalamus, brain hippocampus, brain amygdala, brain caudate basal ganglia, brain putamen basal ganglia, pituitary, brain substantia nigra, and brain spinal cord cervical c-1.

Figure 3.

Forest plot displaying the causal effect of diabetic kidney disease on cognitive function from inverse-variance weighted Mendelian randomization analysis. IVW: inverse-variance weighted; nSNP: the number of SNPs used in the MR analysis; OR: odds ratio; 95% CI: 95% confidence interval; pval: p value.

Figure 4.

Forest plot displaying reverse Mendelian randomization results testing cognitive function’s effect on diabetic kidney disease risk. IVW: inverse variance weighted; nSNP: the number of SNPs used in the analysis; OR: odds ratio; pval: p value; heterogeneity p: p value of IVW heterogeneity test; pleiotropy p: p value of pleiotropy test.

Figure 5.

Forest plot displaying multivariate Mendelian randomization results examining the direct effect of diabetic kidney disease on cognitive function after adjustment for Alzheimer’s disease, heart failure, hypertension, anxiety, depression, education, smoking, alcohol consumption, and income. OR: odds ratio; 95% CI: 95% confidence interval; pval: p value.

Figure 6.

Genetic colocalization analysis between diabetic kidney disease and cognitive function at six susceptibility loci. Each panel displays: X-axis: −log10(p value) from cognitive function GWAS; Y-axis: −log10(p value) from diabetic kidney disease GWAS at the (A) OR14J1, (B) SUOX, (C) RAB5B, (D) IKZF4, (E) RPS26, and (F) ERBB3 loci; dots represent individual genetic variants colored by their linkage disequilibrium (r²) with the lead SNP. Analyses were performed using Coloc software with posterior probability >.90 defining significant colocalization.