Data-driven discovery of midlife cardiometabolic profile associated with incident early-onset and late-onset dementia.

Summary

In 289,494 UK Biobank participants followed for a median 14.1 years, five cardiometabolic clusters were identified. A high liver enzyme pattern conferred the largest risk for early-onset dementia (HR 2.58) and also increased late-onset dementia risk, while an inflammation pattern increased late-onset dementia risk (HR 1.39). Associations were independent of APOE ε4 status.

Key Findings

• Five data-driven cardiometabolic clusters were identified from 12 markers; the high liver enzymes pattern (cluster 3) showed the strongest association with early-onset dementia (HR 2.58; 95% CI 1.61–4.14).
• Cluster 3 also increased late-onset dementia risk (HR 1.36; 95% CI 1.09–1.71); an inflammation pattern (cluster 4) increased late-onset dementia risk (HR 1.39; 95% CI 1.13–1.72).
• No significant interactions with APOE ε4 genotype were detected, suggesting risk operates independently of this genetic factor.

Clinical Implications

Routine assessment of liver enzymes and systemic inflammation profiles in midlife could refine dementia risk stratification and motivate early interventions (e.g., metabolic/liver health optimization, anti-inflammatory strategies), irrespective of APOE ε4 genotype.

Limitations

• Observational design cannot establish causality; residual and unmeasured confounding possible.
• Generalizability beyond the UK Biobank population may be limited; biomarker cutpoints and cluster reproducibility need external validation.

Future Directions

Validate clusters across diverse populations, test interventions targeting liver health and systemic inflammation in high-risk clusters, and integrate imaging/omics to elucidate mechanisms.