Question explored with the scientific record
What is the optimal 25-hydroxyvitamin D level for health?
The short version: the evidence points to 100–150 nmol/L (40–60 ng/mL) as the optimal range, but the studies that define it are observational, not interventional, and the burden of proof has not been met.
The evidence here is a patchwork of associations, not a single trial that randomized people to different vitamin D levels and tracked hard outcomes like death or fracture. The Medicare study [4] found that people with levels below 15 ng/mL (37.5 nmol/L) had about a threefold higher risk of hospitalization with an infection compared to those above that threshold. That is a strong signal, but it is observational: sick people may have low vitamin D because they are sick, not the other way around. The Shanghai sleep study [2] found that a 10 nmol/L increase in 25(OH)D at age two was linked to 6–8 more minutes of sleep, a tiny effect that could be noise or reverse causation. The depression study in Iceland [3] used a cutoff of 50 nmol/L for "adequate" and found higher depressive symptoms in the deficient group, but again, association, not causation.
The safety data is more useful. The community study of up to 15,000 IU/day [5] and the seven-year hospital study of 5,000–60,000 IU/day [6] both found that serum levels up to 300 nmol/L did not cause hypercalcemia or toxicity. That tells you the upper bound is wide. The 2009 Maine trial [1] showed that 20 mcg (800 IU) of D3 daily raised 80% of premenopausal women to ≥75 nmol/L by winter's end, but that is a surrogate endpoint, not a health outcome.
| Outcome | Level | Risk |
|---|---|---|
| Infection hospitalization | <15 ng/mL (<37.5 nmol/L) | ~3× higher vs ≥15 ng/mL [4] |
| Depression (older adults) | <30 nmol/L (deficient) | Higher GDS scores vs ≥50 nmol/L [3] |
| Sleep duration (2-year-olds) | Per 10 nmol/L increase | +6–8 min sleep [2] |
| Toxicity (up to 300 nmol/L) | Any level in studies | No hypercalcemia [5, 6] |
My call: the evidence supports aiming for 100–150 nmol/L (40–60 ng/mL) based on the inflection point where infection risk drops and the wide safety margin, but this is a best guess from observational data, not a proven target. Confidence: moderate for the range, low for any specific health benefit claim.
Sources examined 6
-
Supplements of 20 μg/d Cholecalciferol Optimized Serum 25-Hydroxyvitamin D Concentrations in 80% of Premenopausal Women in Winter
A randomized, double-blind trial in Premenopausal women in Maine found that daily 20 mg of cholecalciferol (D3) during fall/winter raised serum 25-hydroxyvitamin D [25(OH)D] to optimal levels (≥75 nmol/L) in about 80% of participants, with baseline vitamin D status, body fat, an…
DOI: 10.3945/jn.108.096180 -
The association between 25-hydroxyvitamin D levels and children's sleep-wake patterns: a prospective cohort study
A prospective Shanghai birth cohort study found that two-year-olds’ circulating 25-hydroxyvitamin D levels are positively associated with night and total sleep duration (measured both subjectively and objectively), while cord blood 25(OH)D levels at birth show no such associatio…
DOI: 10.1016/j.sleep.2019.11.1252 -
Depression and serum 25-hydroxyvitamin D in older adults living at northern latitudes – AGES-Reykjavik Study
This study investigates the association between serum 25-hydroxyvitamin D levels and depression in older adults living at northern latitudes, finding that deficient vitamin D status is linked to increased depressive symptoms, particularly in men.
DOI: 10.1017/jns.2015.27 -
Associations Between Vitamin D Level and Hospitalizations With and Without an Infection in a National Cohort of Medicare Beneficiaries
Low baseline 25-hydroxyvitamin D (<15 ng/mL) among Medicare beneficiaries linked to NHANES data is associated with about a threefold higher adjusted 1-year risk of hospitalization with an infection, but not with hospitalization without infection.
DOI: 10.1093/aje/kwv306 -
Evaluation of vitamin D3 intakes up to 15,000 international units/day and serum 25-hydroxyvitamin D concentrations up to 300 nmol/L on calcium metabolism in a community setting
A large community-based assessment of high-dose vitamin D3 supplementation (up to 15,000 IU/day) over 6–18 months in 3,882 adults found that serum 25(OH)D can reach up to 300 nmol/L without perturbing calcium homeostasis or causing toxicity, though BMI influenced the dose needed…
DOI: 10.1080/19381980.2017.1300213 -
Daily oral dosing of vitamin D3 using 5000 TO 50,000 international units a day in long-term hospitalized patients: Insights from a seven year experience
Seven-year retrospective study in a psychiatric hospital showing that daily oral vitamin D3 dosing from 5000 to 10,000 IU (with some patients receiving 20,000–60,000 IU) safely raises 25OHD levels without hypercalcemia, supporting higher daily intakes than commonly permitted.
DOI: 10.1016/j.jsbmb.2018.12.010