https://www.neurology.org/doi/pdf/10.1212/WN9.0000000000000057
chatGPT:
Summary
The paper studies whether serum 25-hydroxyvitamin D [25(OH)D] measured in early midlife is associated with later brain PET markers of preclinical Alzheimer-type pathology in dementia-free adults. The cohort came from the Framingham Heart Study Generation 3, with vitamin D measured at examination cycle 1 in 2002–2005, when participants were on average about 39 years old, and amyloid/tau PET imaging performed roughly 16 years later. The final analytic sample included 435 participants with vitamin D data; 424 had amyloid-PET and 369 had tau-PET.
The main outcomes were:
| Outcome | PET marker | Main finding |
|---|---|---|
| Global tau burden | FTP-PET across 34 cortical regions | Higher midlife vitamin D associated with lower later tau |
| Composite early AD-region tau | Entorhinal, parahippocampal, fusiform, amygdala, inferior/middle temporal cortices | Higher vitamin D associated with lower tau |
| Amyloid burden | PiB-PET frontal/lateral temporal/parietal/retrosplenial composite | No association with vitamin D |
In the fully adjusted model, higher log-transformed 25(OH)D was associated with lower global tau-PET burden and lower composite tau-PET burden, after adjustment for age, sex, time to PET, camera type, depression, season, smoking, blood pressure, antihypertensive use, diabetes, cardiovascular disease, and BMI. The effect sizes were modest but statistically significant: global tau β ≈ −0.022, p = 0.010; composite tau β ≈ −0.023, p = 0.016. There was no association with amyloid-PET.
The authors also tested vitamin D as a clinical category using <30 ng/mL vs ≥30 ng/mL, but this categorical analysis was not significant in the fully adjusted model. This suggests the signal may be more continuous, or that the study lacked power to identify threshold effects. Only 37 participants had vitamin D below 20 ng/mL, limiting analysis of conventional deficiency thresholds.
Sensitivity analyses broadly supported the tau finding. Excluding people taking vitamin D supplements did not materially change the results, and adjustment for amyloid burden still left the vitamin D–tau association present. No significant interaction was found by APOE ε4 status or by sex.
The authors conclude that lower vitamin D in early midlife may be a modifiable risk marker for later tau pathology, but they appropriately state that the finding requires confirmation and clinical-trial testing before being interpreted as causal.
Novelty
The main novelty is that this is not simply another study linking low vitamin D to dementia risk or cognitive decline in older adults. It links early midlife circulating vitamin D to later objective PET evidence of tau deposition in adults who were still dementia-free.
The paper is novel in four specific ways:
-
Timing of exposure: vitamin D was measured at a mean age of about 39, much earlier than many prior dementia/vitamin D studies, which often measure vitamin D in older adults closer to the onset of cognitive decline.
-
Preclinical outcome: the study uses tau-PET and amyloid-PET, rather than clinical dementia diagnosis or cognitive testing alone. This allows it to examine Alzheimer-type pathology before overt dementia.
-
Tau-specific signal: the finding was present for tau but not amyloid. That is interesting because tau burden is often more closely tied to neurodegeneration and cognitive decline than amyloid burden, although amyloid may still be upstream in many Alzheimer models.
-
Long interval: the exposure-to-imaging gap was about 16 years, giving the study a genuinely prospective structure rather than a cross-sectional association between vitamin D and brain pathology.
The most scientifically interesting aspect is the dissociation between tau and amyloid. The authors suggest this could fit with early tau accumulation being detectable before widespread cortical amyloid in younger/preclinical cohorts. Another possibility is that vitamin D biology may influence tau phosphorylation, oxidative stress, neuroinflammation, kinase activity, or phosphatase regulation more directly than amyloid plaque accumulation.
Critique
This is a useful and well-designed observational study, but the causal claim should be treated cautiously.
The strongest feature is the prospective design: vitamin D was measured many years before PET imaging, reducing the chance that preclinical dementia caused low vitamin D through reduced outdoor activity, frailty, or altered diet. The Framingham cohort also provides careful longitudinal phenotyping and adjustment for several vascular and lifestyle-related confounders. Use of PET biomarkers is another major strength.
However, the study has several important limitations.
First, vitamin D was measured only once. A single early-midlife 25(OH)D value may not represent long-term vitamin D exposure. Vitamin D varies with season, supplement use, adiposity, diet, sun exposure, latitude, physical activity, and illness. The authors adjusted for season and BMI, but they lacked detailed lifetime data on outdoor activity, diet, and physical activity. This leaves substantial potential for residual confounding.
Second, the association may partly reflect healthy lifestyle confounding. Higher vitamin D can be a marker for more outdoor activity, better health, lower adiposity, better diet, higher socioeconomic status, or greater exercise. Some of these factors could plausibly reduce tau pathology or brain aging independently of vitamin D itself.
Third, the effect size is small and the categorical clinical threshold analysis was weaker. Continuous vitamin D was associated with tau, but the clinically intuitive comparison of ≥30 vs <30 ng/mL was not significant in fully adjusted models. That does not invalidate the result, but it makes the clinical interpretation less straightforward. It is not clear whether raising someone from, for example, 22 to 35 ng/mL would meaningfully reduce tau burden.
Fourth, there were relatively few people with marked deficiency. Only 9.3% had vitamin D below 20 ng/mL, and only 3 participants were below 12 ng/mL. This limits the ability to test whether true deficiency is the important risk state, rather than variation across the normal-to-high range.
Fifth, PET imaging was done in a selected subset of the Framingham cohort. Participants had to survive, remain engaged, undergo MRI, be free of major neurologic disease, and receive PET imaging. This creates possible selection bias, especially if vitamin D status is related to health, survival, or participation.
Sixth, the sample was predominantly Caucasian, limiting generalisability. This matters particularly for vitamin D because skin pigmentation, sun exposure, supplementation patterns, vitamin D binding protein, ancestry, and social/environmental factors can alter both measured 25(OH)D and its interpretation.
Seventh, the paper does not establish whether vitamin D itself is protective, or whether low vitamin D is simply a risk marker. Mechanistic plausibility exists: vitamin D may influence oxidative stress, neuroinflammation, GSK3β/CDK5 activity, PP2A methylation, and tau phosphorylation pathways. But observational human PET data alone cannot establish that supplementation would reduce tau deposition.
Bottom line
This is a valuable prospective biomarker study suggesting that higher vitamin D in early midlife predicts lower tau-PET burden about 16 years later, without a parallel association with amyloid-PET. Its novelty lies in connecting midlife vitamin D to preclinical tau pathology, rather than only to late-life cognition or dementia diagnosis.
The result is biologically plausible and potentially important, but it should be interpreted as hypothesis-generating, not as proof that vitamin D supplementation prevents Alzheimer-type pathology. The next useful step would be replication in larger and more diverse cohorts with repeated vitamin D measures, richer lifestyle confounder data, and ideally trials testing whether correction of deficiency in midlife changes downstream tau, cognition, or dementia risk.