12.6 more years of life without dementia if zero midlife vascular risk factors (helped by Vitamin D etc. )

Midlife Vascular Risk Burden and Dementia-Free Survival Years: The Atherosclerosis Risk in Communities Neurocognitive Study

Neurology® Open Access, September 2026, https://doi.org/10.1212/WN9.0000000000000152

Jiaqi Hu, Josef Coresh, Jason R. Smith, A. Richey Sharrett, Rebecca F. Gottesman, Pamela L. Lutsey, Thomas H. Mosley, Elizabeth Selvin, Michael Fang, Jordan Weiss

Summary by Claude - August, 2026

Prospective cohort, peer-reviewed. Funded by NIH/NHLBI and NIDDK; authors report no relevant disclosures. Open access (CC BY-NC-ND).

Why this study is on VitaminDWiki

This study measured no vitamin D. It is here because all four conditions it deals with — dementia, diabetes, hypertension, and smoking — have vitamin D page clusters on this site, linked at the bottom of this page. The strength of those four links is not equal, and it is worth being plain about it:

  • the diabetes evidence is the strongest, with randomized trials of prediabetes progression;
  • hypertension has been mostly null in unselected adults but better in those who were deficient to begin with; dementia has a large observational signal and no trial with dementia as an endpoint;
  • smoking the well-established direction is that smoking lowers vitamin D, with vitamin D possibly reducing the damage rather than the habit.

What this paper adds is the price tag — how many years of cognitive life those three risk factors cost when present at age 56. The vitamin D connection is ours; the arithmetic is theirs.

What it found

Three things measured once, at an average age of 56: diabetes, high blood pressure, and current smoking. Score 0 to 3. Then 12,409 people followed a median of 26 years — 3,008 developed dementia, 5,238 died without it.

Years lived alive and free of dementia between ages 55 and 95:

  • 0 risk factors — 30.1 years (95% CI 29.9–30.4)
  • 1 risk factor — 26.3 years (26.1–26.6)
  • 2 risk factors — 21.0 years (20.5–21.4)
  • 3 risk factors — 17.5 years (16.3–18.7)

A spread of 12.6 years, and a clean dose-response that holds in every subgroup: women, men, Black, White, APOE-ε4 carriers and non-carriers. Starting the clock later shrinks the gap in absolute terms but not the ratio — counted from age 75, it is 12.4 remaining dementia-free years for the zero-risk group against 6.0 for the three-risk group. From age 85, 5.9 against 3.0.

What a person would change

Only three things were measured, so only three are on offer here:

  • Don't smoke. Current smoking was the exposure; former and never smokers were pooled together as the reference, so quitting moved a person into the low-risk group.
  • Keep blood pressure under 140/90, treated or untreated — medication use counted the same as an elevated reading.
  • Avoid or control diabetes — defined as HbA1c ≥ 6.5%, medication use, or a physician's diagnosis.

The covariates in the paper's Table 3 add two more independent signals: ideal or intermediate physical activity ran about 15% lower dementia hazard than poor activity, and BMI ≥ 40 ran about 30% higher. Carrying two APOE-ε4 alleles tripled the dementia hazard (HR 3.31) but did not raise death without dementia at all — genetics loaded the dementia side of the scale only.

The trap in this paper

Total dementia by age 95 was higher in the healthy group (42%) than in the 3-risk-factor group (23%). This is not a benefit of poor vascular health. The high-risk group had 5.6x the hazard of dying without dementia; by 85 only 7% of them were still alive and dementia-free, against 35% of the healthy group. They never lived long enough to be diagnosed. In cause-specific models — which censor at death and adjust for sex, race, education, activity, BMI, stroke and APOE — they still carried 2.7x the hazard of developing dementia. Good vascular health buys extra years, and dementia becomes possible in those extra years. In the authors' own covariate-adjusted estimates (eTable 5), the reversal nearly vanishes: 37.1% vs 33.6%.

What this does not show / limitations

No vitamin D of any kind was measured — no 25(OH)D, no supplementation, no dose. Observational, so no causation. Risk factors were measured once in midlife and never re-checked, so this cannot tell you whether fixing hypertension at 60 recovers lost years — only that having it at 56 predicted fewer of them. The 3-risk-factor group is small (N = 176, 38 dementia cases), so those estimates are the shakiest in the paper. Only three exposures were modeled, chosen because ARIC measured them consistently — nothing on diet, sleep, hearing, or any nutrient. Dementia ascertainment likely differed between those who kept attending in-person visits and those who did not.

PDF

📄 Attached Etables PDF

📄 Attached efigures.pdf

📄 Attached emethods.pdf

Related in VitaminDWiki

Diabetes - Vitamin D and Magnesium

Hypertension - Vitamin D and Magnesium

Smoking

Dementia

Omega-3, etc.

Alzheimers

Study was also reviewed by