Reference range, risk threshold, optimization claim. Three different numbers.
Almost every argument about a blood test is really an argument about which of these three a number is. This page is generated from the same records the panel is built from, so it cannot describe a threshold this site no longer publishes.
A lab number is not one kind of thing
This site publishes 38 markers across seven systems. Every threshold on every one of them is labeled with what KIND of number it is, because the four kinds carry very different amounts of evidence and are quoted interchangeably almost everywhere else.
Guideline threshold
A number a professional body has published as a decision point. It carries the weight of that body's review process, and it is a judgment about a population rather than a measurement of one person.
On this site five of 38 markers. Example: ApoBApolipoprotein B, a protein carried once on each cholesterol particle that can enter the artery wall, so it counts those particles directly., standard range < 130 mg/dL.
Lower-risk association
A concentration that tracks with lower risk in observationalResearch that watches what people already do rather than assigning a treatment. It can show that two things travel together, not that one causes the other. data. It describes who was healthier, not what happens if you move your number, and the two are not the same claim.
On this site nine of 38 markers. Example: HDL-C, standard range > 40 (M) / > 50 (F).
Practice convention
A number in wide use that no trial established. It may be sensible and it may be right. It is a convention, and this site says so rather than dressing it as a finding.
On this site seventeen of 38 markers. Example: Fasting Insulin, standard range < 25 mIU/L.
Not established
No target is published here, because the evidence does not support one. This is the honest state for more markers than the longevity industry admits.
On this site seven of 38 markers. Example: SII, standard range No clinical reference range.
A target is a category, not a measurement
A standard laboratory range says where most people in a reference population fall. A risk threshold says where risk begins to rise in the data somebody analyzed. An optimization target says where somebody believes you would be better off. Only the first is a property of the assay, and the third is the one longevity marketing sells.
That is why every marker here carries its standard range and any EBL reference range as separately sourced fields, and why seven of the 38 publish no optimal range at all.
Worked example: Vitamin D (25-OH)
Standard range ≥ 20 ng/mL (National Academies). Optimal range Not established.
VITAL (n=25,871) randomized 2000 IU/day for the prevention of cancer and cardiovascular disease. Neither primary endpointThe one outcome a trial names in advance as its main test. Results on other outcomes carry less weight. was reduced: invasive cancer of any type and major cardiovascular events were both null. Death from cancer, a secondary endpointAn outcome a trial also measures but did not choose as its main test., was lower. No randomized trial establishes an optimal concentration for longevity or disease prevention, so no optimal range is published here. The reference figure above is the National Academies sufficiency cutpoint; the 30 ng/mL target that circulates widely comes from the 2011 Endocrine Society guideline, which its own 2024 successor replaced while declining to set concentration thresholds for generally healthy people. How common low status actually is, and at which threshold, is set out under Nutrient Gaps.
Two careful analyses of the same data can disagree
When they do, this site shows both rather than picking the one that reads better. Disagreement is information about how much confidence the number can carry.
Uric Acid and coronary heart disease
Conventional analysis gave an odds ratio of 1.18 per standard deviation of urate and multivariable analysis gave 1.10, but MR-Egger, which corrects for pleiotropy at the cost of statistical power, gave 1.05 with a confidence intervalThe range the true effect probably falls in. A wide interval means the estimate is uncertain. from 0.92 to 1.20.
Population: 58 studies, 198,598 individuals; predominantly European ancestry
Evidence against, held on the Vitamin D (25-OH) record
311 healthy adults aged 55 to 70 randomized to 400, 4000 or 10 000 IU/day of vitamin D3 for three years. Achieved 25(OH)D was 132.2 nmol/L (53 ng/mL) in the 4000 IU arm and 144.4 nmol/L (58 ng/mL) in the 10 000 IU arm, both inside the 50 to 80 ng/mL window this site previously described as optimal. Radial volumetric bone density was lower in both higher-dose arms against 400 IU: -3.9 mg HA/cm3 (95% CI -6.5 to -1.3) at 4000 IU and -7.5 (-10.1 to -5.0) at 10 000 IU. Bone strength did not differ significantly (radius p=0.06, tibia p=0.12).
Randomized, double-blindNeither the participants nor the people assessing them know who received the real treatment until the trial ends., three-year dose-comparison trial. · PMID 31454046
Evidence against, held on the Free T3 record
Higher free T3 within the normal range associated with lower hip BMD and 20% increased nonvertebral fracture risk in healthy postmenopausal women (OPUS, 6-year prospective).
Prospective cohort. Observational, not randomized. · PMID 20410228
A citation beside a number is not proof the number came from it
two markers on this panel carry a written caveat about their own citation, kept in public rather than quietly corrected. Both are worth reading before trusting any threshold anywhere, including on this site.
Fasting Glucose
EBL previously cited PMIDPubMed ID: the permanent number the US National Library of Medicine gives a published paper. Each one links to that paper's record on PubMed. 15161807 here attributed to 'DECODE 2003, Lancet, n=25,364'. That PMID is 'Use and abuse of HOMA modeling' (Diabetes Care 2004) and the attribution did not exist. Withdrawn July 2026 and replaced with the two papers that actually address the claim. The page previously asserted a 'continuous dose-response'; the largest analysis finds the association non-linear in non-diabetics, so that wording was corrected rather than re-cited.
Free T3
PMID 20410228 is cited as a SAFETY SIGNAL, not as support for a target. Murphy et al. (JCEM 2010) is a 6-year prospective study in healthy postmenopausal women associating higher Free T3 within the reference range with lower hip BMD and 20% higher nonvertebral fracture risk. It is observational and cannot establish cause. It is cited because it is the only directly relevant finding located and it points away from the target this page used to publish.
Move the threshold and the problem changes size
The same blood, measured the same way, produces very different headlines depending on which cutoff is applied. This is the single most useful thing to understand about deficiency statistics.
Vitamin D (25-OH): Serum 25-hydroxyvitamin D
5.0% 95% CI 4.1 to 6.2, below 30 nmol/L (12 ng/mL) (at risk of deficiency), National Academies of Sciences (formerly Institute of Medicine), Dietary Reference Intakes
NHANES 2011-2014, United States, aged 1 year and older, n=16,180 · PMID 31076739
18.3% 95% CI 16.2 to 20.6, 30 to 49 nmol/L (12 to 19.6 ng/mL) (at risk of inadequacy), National Academies of Sciences (formerly Institute of Medicine), Dietary Reference Intakes
NHANES 2011-2014, United States, aged 1 year and older, n=16,180 · PMID 31076739
A prevalence figure describes how common a blood level is. It is not trial evidence that correcting that level changes an outcome, so it cannot carry a letter. The grade for supplementing vitamin D sits on the vitamin D3 record and rests on its own trials.
Vitamin B12: Serum vitamin B12, or plasma methylmalonic acid as a functional marker
3% to 26% range across commonly used cutoffs, not a confidence interval, varies by cutoff chosen for serum B12 (low B12 status by serum), no single body; the paper compares cutoffs in common clinical use
NHANES 1999-2004, United States adults aged over 19, n=12,612 · PMID 21677051
2% to 6% range across commonly used cutoffs, not a confidence interval, varies by cutoff chosen for methylmalonic acid (low B12 status by functional marker), no single body; the paper compares cutoffs in common clinical use
NHANES 1999-2004, United States adults aged over 19, n=12,612 · PMID 21677051
The prevalence range below measures how many people fall under a chosen line, not whether raising them above it changes anything. No letter follows from it.
RBC Magnesium: Serum magnesium is the common test. It is a poor index of total body magnesium.
approximately 50% not stated as an interval in the source, usual intake below the Estimated Average Requirement (dietary intake gap, not a blood measurement), Estimated Average Requirement, National Academies Dietary Reference Intakes; characterization of magnesium as a shortfall nutrient from the 2015 Dietary Guidelines Advisory Committee
dietary intake assessment cited in review, not a blood-based survey estimate reported 2016, citing the 2015 Dietary Guidelines Advisory Committee, United States, general population, varies by age group · PMID 28140318
Magnesium does not merely lack an answer, it lacks a measurement. Without a status assay that tracks body stores, biochemical prevalence cannot be estimated at all, and no trial can enroll people on the basis of a status this test cannot establish.
Why there is no single cutoff
There is no single agreed cutoff, and the most-quoted one has been withdrawn. The National Academies treat 50 nmol/L (20 ng/mL) as sufficient for most people. The Endocrine Society's 2011 guideline used 75 nmol/L (30 ng/mL), and that is the number still circulating in most clinical and consumer material. Its 2024 guideline replaced the 2011 one and declined to set concentration thresholds defining sufficiency, insufficiency or deficiency for generally healthy people, on the grounds that trials have not established outcome-specific thresholds. The prevalence figures above are computed against the National Academies cutpoints. Against the withdrawn 2011 target a far larger share of the population falls below the line, without a single blood level having changed.
What this page is, and is not
It is a description of how the numbers on this site are classified and sourced, assembled from the records themselves so it cannot fall out of step with them. It sets no target, endorses no range, and is not an assessment of any individual result. What any number means for one person depends on their baseline, their medications, the assay their laboratory ran, and clinical context this site has no access to.
Evidence Based Longevity, Inc. is owner-operated and edited by Arnold Sandlow, D.C., a California-licensed chiropractor. Evidence Based Longevity publishes in English only · Machine translations are not produced or reviewed by EBL for accuracy