- Relative to normal-weight peers, women with obesity had a 21% higher rate of dementia at 15 or more years of follow-up in this longitudinal study, but the association of BMI and dementia changed over time.
Why this matters
- Interventions for preventing and treating dementia are limited.
- In an editorial, May A. Beydoun, PhD, and Mika Kivimaki, FMedSci, write that the associations among weight status, related behaviors, and dementia are complex, and call for more long-term studies of the associations of these factors with different stages of dementia over time.
- At mean 18-year follow-up, 2.7% cohort had experienced hospital admission with a dementia diagnosis code.
- Compared with normal-weight peers (BMI, 20-24.9 kg/m2), women with obesity (BMI, ≥30 kg/m2) had dementia risk that was:
- Lower during years 5-9: rate ratio [RR], 0.79 (95% CI, 0.71-0.89);
- Similar during years 10-14: RR, 1.04 (95% CI, 0.99-1.11); and
- Higher during years ≥15: RR, 1.21 (95% CI, 1.16-1.26).
- Risk during years ≥15 driven by vascular dementia, dementia of unspecified type.
- Low BMI (2), low caloric intake (lowest quintile), and inactivity (exercise rarely/never or less than once weekly) all predicted dementia during the first decade but not clearly thereafter.
- Prospective population-based cohort study of 1,136,846 UK women with a mean age at baseline of 56 years (Million Women Study).
- Main outcome: dementia.
- Funding: UK Medical Research Council; Cancer Research UK.
- Baseline measures self-reported.
- Reliance on diagnosis codes.
- Milder dementia not captured.
- Residual and unmeasured confounding.