25-hydroxyvitamin D (the vitamin D test)

In brief

What the test actually measures

When a doctor orders a vitamin D level, the laboratory does not measure vitamin D. Vitamin D itself, the molecule made in skin or swallowed in a capsule, has to be changed twice before it becomes the hormone ^1,34. The first change happens in the liver, where an enzyme attaches a hydroxyl group, an oxygen atom bonded to a hydrogen, at the molecule's twenty-fifth carbon ³⁵. The product is 25-hydroxyvitamin D, and it was identified as a biologically active metabolite of vitamin D3 in 1968 ¹. The liver enzyme chiefly responsible, CYP2R1, was pinned down only in 2004, by finding a patient who had inherited two broken copies of the gene, which left him with low 25-hydroxyvitamin D and the classic signs of deficiency ³⁵. Sold as a drug, the same molecule is called calcifediol, and in a Barcelona hospital series it was the supplement most often found in patients with very high results ³⁰.

The result on the report is a sum. Most supplements contain one of two forms, vitamin D2 (ergocalciferol) or vitamin D3 (cholecalciferol) ¹⁸, and each is hydroxylated in turn, so the blood carries 25-hydroxyvitamin D2 and 25-hydroxyvitamin D3. Total 25-hydroxyvitamin D, the figure guidelines use, is the two added together ^2,3. The two halves do not behave identically: in a tracer study, in which volunteers swallowed a small labeled dose and its disappearance from the blood was timed over 33 days, in 36 healthy men from The Gambia and the United Kingdom, 25-hydroxyvitamin D2 had a half-life of 13.9 days against 15.1 days for D3 ⁴, and in a 2021 comparison of 32 routine assays most of them underestimated the D2 form ³⁶. A patient taking prescription vitamin D2 can therefore look worse supplied on some machines than they are ³⁶.

The second change happens mainly in the kidney, which adds another hydroxyl at carbon 1 to make 1,25-dihydroxyvitamin D ³⁵. This is the hormone, isolated in 1971 as the metabolite that acts on the intestine ³⁴, and it is what the vitamin D receptor in cells actually binds ³⁷. It circulates at a far lower concentration: 25-hydroxyvitamin D runs at roughly 25 to 200 nmol/L ⁵, while laboratories report the hormone in picograms per milliliter, with low defined in one large kidney study as under 22 pg/mL ³⁸.

Neither form floats free. In normal serum about 88 percent of 25-hydroxyvitamin D is bound to vitamin D binding protein, a carrier protein made by the liver, and most of the rest to albumin ⁶. In mice engineered to lack the binding protein, 25-hydroxyvitamin D cleared from the blood much faster, evidence in animals that the carrier is what turns 25-hydroxyvitamin D into a long-lived reserve ³⁹. A 2025 commentary proposed that muscle adds a second reservoir, taking up the binding protein with its cargo and releasing 25-hydroxyvitamin D back into the blood as the protein is broken down, and argued that this cycling accounts for the long half-life; it is the authors' interpretation in a commentary, not a new measurement ⁴⁰. The test measures that reserve, bound and unbound together ⁴¹.

• Total 25-hydroxyvitamin D by automated immunoassayAn antibody or binding protein competes for 25-hydroxyvitamin D on a high-throughput analyzer, usually without first extracting the sample with solvent ⁸.The routine method in many clinical laboratories since automated analyzers arrived ⁸Only 50 percent of 32 ligand-binding assays, the immunoassays and protein-binding assays that capture the molecule with an antibody or binding protein, met the 5 percent bias goal in the 2021 standardization study ³⁶; in 1064 paired outpatient samples, among the pairs that disagreed, one immunoassay called deficiency 118 times more often than mass spectrometry ⁷⁴.Most underestimate 25-hydroxyvitamin D2 and several may cross-react with 24,25-dihydroxyvitamin D ³⁶; errors move inversely with binding-protein concentration, which is high in pregnancy and low in intensive care ⁶²; high-dose biotin, a B vitamin tested as a treatment for multiple sclerosis, can distort kits that use streptavidin, a protein that grips biotin, to hold the test together ^71,72; and a myeloma antibody has faked intoxication ³².
• Total 25-hydroxyvitamin D by LC-MS/MSLiquid chromatography separates the molecules and tandem mass spectrometry weighs them against an isotope-labeled internal standard ⁵⁷.Reference laboratories, research, infants, and the reference methods of NIST, Ghent University and the CDC ^8,31Only 53 percent of 14 laboratory-built methods met the 5 percent bias goal, and four were off by 12 to 21 percent ².Methods that do not separate the C3 epimer read high, most of all in infants ^2,31; like every total assay, it cannot see the free fraction or the hormone ⁴¹.
• 1,25-dihydroxyvitamin DThe hormone itself, measured in picograms per milliliter, a unit a thousand times smaller than the nanograms used for 25-hydroxyvitamin D ³⁸.Hypercalcemia with low parathyroid hormone, granulomatous disease and lymphoma, CYP24A1 defects, hereditary rickets, and chronic kidney disease ^38,46,53Its half-life of about 15 hours ⁵ means a single value reflects the last day of regulation, not the body's stores.It stays normal in deficiency until 25-hydroxyvitamin D is down to about 10 nmol/L, so it cannot diagnose deficiency ⁴³.
• 24,25-dihydroxyvitamin D and the metabolite ratioThe first breakdown product of 25-hydroxyvitamin D, measured beside it by LC-MS/MS and often expressed as a ratio ⁵⁰.Suspected CYP24A1 deficiency, in which the 25-hydroxyvitamin D3 to 24,25-dihydroxyvitamin D3 ratio reached 80 to 100 ⁵⁰; research on functional vitamin D status ¹¹⁷In a six-laboratory comparison, mean bias against NIST's candidate reference method ranged from −1.6 to 36 percent ¹²⁰.No agreed clinical cutoffs; in 204 older adults the ratio's correlation with parathyroid hormone was modest, rho −0.292 ¹¹⁷.
• Free or bioavailable 25-hydroxyvitamin DThe unbound fraction, either calculated from total 25-hydroxyvitamin D, binding protein and albumin, or measured directly ⁴¹.Research only ⁴¹Calculated values explained only 13 percent of the variation in directly measured ones ²⁸.Calculations inherit the bias of the binding-protein assay, the source of the 2013 race claim ⁶⁶; direct methods are complex or prone to interference ⁴¹.

Why the storage form, and not the hormone

The choice of what to measure follows from two clocks. 25-hydroxyvitamin D3 has a half-life of about 15 days; the hormone, about 15 hours ⁵. With that half-life, 25-hydroxyvitamin D reflects most of what has come in over the previous month or two, from sunlight, food and pills ^4,5. A molecule that turns over within a day reports only what the body has decided to do since yesterday ⁵. If the question is how much vitamin D a person has been getting, the slow pool is the one to sample.

The second reason is that the hormone is not allowed to wander. Its production in the kidney is under tight control. In rats on a low-calcium diet, removing the thyroid and parathyroid glands cut production of 1,25-dihydroxyvitamin D to negligible levels within 40 hours, and parathyroid extract restored it, an effect that changes in calcium or phosphorus could not explain ⁴². Inorganic phosphate also controls the step ⁴⁴. In 2004 a third controller was added: a single injection of FGF23, fibroblast growth factor 23, a hormone released by bone, lowered 1,25-dihydroxyvitamin D in rodents within 3 hours by reducing the messenger RNA, the working copy of a gene that a cell reads to build a protein, for the activating enzyme, 1-alpha-hydroxylase, and raising it for the inactivating one, 24-hydroxylase ⁴⁵. All of this is animal evidence, but the human data described next fit the same pattern ⁴³.

That control is exactly why the hormone is a poor marker of supply. When 25-hydroxyvitamin D runs short, the parathyroid glands secrete more parathyroid hormone, which drives the kidney to make more of the hormone from what precursor remains ^42,48. This is secondary hyperparathyroidism, and in the elderly it costs bone: high turnover, bone loss, mineralization defects and hip fractures ⁴⁸. The hormone itself is defended to the last. In the records of 3661 Adelaide patients, among the 319 whose 25-hydroxyvitamin D was 40 nmol/L or less, 1,25-dihydroxyvitamin D and calcium absorption did not fall until 25-hydroxyvitamin D reached about 10 nmol/L ⁴³. A doctor who measured the hormone in a deficient patient would usually find it normal and be falsely reassured.

The precursor, by contrast, is not held constant, and the body's response to it can be seen. In 1569 French adults sampled through a winter, parathyroid hormone sat on a plateau of 36 pg/mL as long as 25-hydroxyvitamin D stayed above 78 nmol/L (31 ng/mL), and climbed below it ⁴⁹. That relationship, a rising parathyroid hormone as 25-hydroxyvitamin D falls, is the physiological basis of the whole test. It is also where the arguments begin, because where exactly the curve bends, and whether it bends at the same place in everyone, has never been settled ¹³.

When the hormone is measured on purpose

There are situations in which 25-hydroxyvitamin D is the wrong question and the hormone is the right one, and they share a feature: the trouble lies not in supply but in how the hormone is made, broken down or received ⁴⁶. The most common is hypercalcemia that the parathyroid glands are not causing. Hypercalcemia affects up to 4 percent of the population, and besides cancer, primary hyperparathyroidism and excess vitamin D, its causes include production of 1,25-dihydroxyvitamin D outside the kidney and failure to break it down ⁴⁶.

Production outside the kidney was first shown in 1981, in a patient with sarcoidosis whose kidneys had been removed and who still became hypercalcemic, evidence that something else was making the hormone ⁴⁷. That something is the granuloma. In sarcoidosis, tuberculosis and some lymphomas, macrophages or tumor cells express 1-alpha-hydroxylase and make 1,25-dihydroxyvitamin D without the kidney's feedback controls ⁴⁶. In such patients the excess lies in the hormone rather than in the store, so only the second test explains the calcium ^15,46.

Failure to break the hormone down came into focus in 2011. Idiopathic infantile hypercalcemia, so named because its cause was unknown, a disease of severe hypercalcemia, failure to thrive, vomiting, dehydration and calcium deposits in the kidney, turned out in six affected children to be caused by recessive loss-of-function mutations in CYP24A1, the gene for 24-hydroxylase, faults that disable the enzyme and in that study caused disease when both copies of the gene carried one; CYP24A1 mutations were also found in a second group of infants who had become hypercalcemic after single large preventive doses of vitamin D, known as bolus prophylaxis ³³. The authors connected this to the epidemic of infantile hypercalcemia in Britain in the 1950s, when milk was heavily fortified, an inference rather than a tested claim ³³. Patients with two faulty copies, and sometimes only one, have high hormone, suppressed parathyroid hormone, excess calcium in the urine, kidney stones and calcification of the kidney, and first-time calcium stone formers show signs of the same impaired breakdown ⁴⁶. Here the most informative measurement is a third metabolite: the ratio of 25-hydroxyvitamin D3 to 24,25-dihydroxyvitamin D3 stayed below 25 in vitamin D-sufficient people but reached 80 to 100 in patients with idiopathic infantile hypercalcemia ⁵⁰. The defect can surface in adulthood. In a 2026 case report, a 32-year-old woman who had had no hypercalcemia in earlier pregnancies became hypercalcemic while pregnant and taking vitamin D and calcium supplements; very high hormone and a high ratio of 25-hydroxyvitamin D to 24,25-dihydroxyvitamin D pointed to the diagnosis, and she carried two copies of the same CYP24A1 variant ⁵¹. The authors attributed the timing to the normal rise of the activating enzyme in pregnancy, with the supplements as an added trigger ⁵¹.

The inherited rickets are the clearest case for measuring both forms, because each breaks a different link in the chain and leaves a different fingerprint, set out in the table below ^{35,37,52,53,54}. The most recently described, published in 2018, was found in two unrelated patients with early-onset rickets and low levels of both 25-hydroxyvitamin D and the hormone; each carried the same new mutation in the enzyme CYP3A4, which made it oxidize 1,25-dihydroxyvitamin D ten times faster than the normal enzyme in laboratory tests ⁵³. The disease is not a shortage of vitamin D but its accelerated destruction ⁵³.

Chronic kidney disease is the last setting, and the one where the hormone falls first. In 1814 outpatients at 153 American centers, 71 percent of them primary care practices, low 1,25-dihydroxyvitamin D was present in 13 percent of those with an estimated glomerular filtration rate (eGFR), the standard measure of kidney function, above 80 mL/min, and in more than 60 percent of those below 30 mL/min, while 25-hydroxyvitamin D3 did not differ across the range of kidney function ³⁸. A falling hormone with a steady precursor is the signature of the kidney's own failing capacity, probably helped by hormones such as FGF23, which suppresses the activating step in rodents ⁴⁵.

From a binding protein to a mass spectrometer: the whole history

The test is almost as old as the molecule's discovery. Three years after 25-hydroxycholecalciferol was identified ¹, two groups published competitive binding assays for it: Belsey, DeLuca and Potts in September 1971 ⁵⁵, and Haddad and Chyu, with a protein-binding radioassay, in December ⁵⁶. The principle is a contest: a protein that binds 25-hydroxyvitamin D is mixed with a fixed amount of radioactively labeled 25-hydroxyvitamin D and with the patient's extracted serum, and the more unlabeled molecule the patient's sample contains, the less radioactivity the protein holds. Within a year the method was answering clinical questions. Normal London adults and schoolchildren averaged 16 ng/mL, epileptic patients on high-dose anticonvulsants averaged 5 ng/mL, and two further epileptic patients with anticonvulsant osteomalacia healed quickly on small doses of 25-hydroxycholecalciferol ²⁶.

Binding assays could not distinguish 25-hydroxyvitamin D2 from D3 and suffered interference from other hydroxylated metabolites ⁵⁷. In 1985 Hollis and Napoli described a radioimmunoassay built on a rabbit antibody that reacted equally with the D2 and D3 metabolites it was tested against, though not with the hormone or the parent vitamins, and needed only a microliter of plasma per tube ⁵⁸. In the early 1990s they introduced an iodine-labeled tracer, which the DEQAS history dates to 1992 ⁸ and which their method paper, describing what it called the first radioimmunoassay for a vitamin D metabolite with a radioiodinated tracer, reported in March 1993 ⁵⁹; the kit DiaSorin marketed with it, followed by other commercial radioimmunoassays, put the test within reach of ordinary hospital laboratories ⁸. The next step removed radioactivity altogether: fully automated immunoassays that skipped solvent extraction made 25-hydroxyvitamin D a routine button on any chemistry analyzer, and as their limitations became apparent more laboratories turned to liquid chromatography-tandem mass spectrometry (LC-MS/MS) ⁸.

The scale of the problem was shown bluntly in 2004. Postmenopausal women measured in one laboratory averaged 46 ng/mL and similar women in another averaged 21 ng/mL; 17 percent fell below 32 ng/mL in the first and 90 percent in the second ²⁰. Split samples from ten healthy adults sent to six laboratories gave means from 17.1 to 35.6 ng/mL, and when 20 ng/mL of pure 25-hydroxyvitamin D was added, the laboratories recovered between 7.7 and 18.0 of it ²⁰. Whether a person was called deficient depended on where the blood was sent.

Mass spectrometry entered routine use around the same time. A laboratory in northwest London reported in 2005 a method that measured 25-hydroxyvitamin D2 and D3 separately, with a stable-isotope internal standard, a known amount of the same molecule built with heavier atoms that lets the instrument correct for losses, and had run more than 6000 clinical samples in nine months ⁵⁷. Within a year the new method showed a weakness of its own: the Mayo Clinic found unexpectedly high results in infants on its new mass-spectrometry method, traced them to C3 epimers, forms of 25-hydroxyvitamin D in which one group at carbon 3 points the other way and which its mass spectrometry was counting as the real thing, and found them in 22.7 percent of 172 infants; an extracted radioimmunoassay, by contrast, gave accurate results ³¹.

The response was to build a ruler. The Vitamin D Standardization Program (VDSP) was set up in November 2010 by the Office of Dietary Supplements of the National Institutes of Health (NIH) with the Centers for Disease Control and Prevention (CDC), the National Institute of Standards and Technology (NIST) and Ghent University, with the goal of making every assay agree with reference methods to within 5 percent bias ¹². NIST certified a serum reference material for vitamin D metabolites, SRM 972, in 2011, and found that separating the epimer was essential to assigning its values ⁹; replaced it with SRM 972a, made mostly of unmodified human serum ¹⁰, and added SRM 2973 at a higher level near 100 nmol/L ¹¹. The Vitamin D External Quality Assessment Scheme (DEQAS), which has circulated shared samples since 1989, switched in 2013 from comparing laboratories with one another to scoring them against the NIST reference method ⁸.

How accurate the number is

Agreement has improved but not arrived. When DEQAS began in 1989, the spread between laboratories, expressed as a coefficient of variation, ran from 29.3 percent at 42.7 nmol/L to 53.7 percent at 20.0 nmol/L; by January 2017, with 918 participants in 56 countries using 30 methods, it was 10.3 to 15.3 percent ⁸. A 15 percent spread at 30 nmol/L, close to the threshold for rickets risk, still means the same blood can be classified differently in different places ^3,8.

Automated immunoassays carry specific weaknesses. Because most skip solvent extraction ⁸, the carrier protein stays in the reaction, and accuracy can depend on how much of it there is. In samples from 51 healthy people, 52 pregnant women, 50 hemodialysis patients and 50 intensive care patients, most of six routine assays deviated significantly from isotope-dilution mass spectrometry, and in four of the five automated assays the error moved inversely with the binding-protein concentration, which is high in pregnancy and low in critical illness ⁶². The most comprehensive comparison of such assays, published in 2021, sent 50 single-donor sera with reference values to 17 laboratories running 32 ligand-binding assays, the family that includes immunoassays and the older protein-binding assays, in which an antibody or binding protein captures the molecule: only half met the 5 percent bias goal, four of the 13 distinct assays were consistently within it and four consistently outside it wherever they were run, most underestimated 25-hydroxyvitamin D2, and five may have been fooled by 24,25-dihydroxyvitamin D3, the first breakdown product, a problem called cross-reaction ³⁶. Employees of the manufacturers whose assays were tested were among the authors ³⁶.

Mass spectrometry is not automatically right. In the companion study, 14 laboratories running their own LC-MS/MS methods did only slightly better: 53 percent met the goal, four were off by 12 to 21 percent, and failure to separate the 3-epimer accounted for much of the bias ². The epimer is not only an infant problem. NIST's high-level human serum reference material contains 98.4 nmol/L of 25-hydroxyvitamin D3 and 5.23 nmol/L of the 3-epimer, about 5 percent ¹¹, though in infants it can be most of the total ³¹.

Some errors come from the patient's other pills. Many automated immunoassays use a streptavidin-biotin capture step, in which streptavidin, a protein that grips the B vitamin biotin, holds the test together, so biotin in the patient's blood can interfere ^71,72. In 23 plasma samples from people taking high-dose biotin, at 31.7 to 1160 micrograms per liter, most hormone results on such assays, in a panel that included 25-hydroxyvitamin D, were abnormal and returned to normal once the biotin was removed ⁷¹. Two of that study's authors worked for MedDay Pharmaceuticals, which was developing high-dose biotin for progressive multiple sclerosis, a benefit the study stated as established ^71,72. The company's own phase 3 trial, SPI2, the large final stage of testing, found that 300 mg a day did not improve disability or walking in 642 patients (odds ratio for improvement 1.35, 95 percent confidence interval 0.81 to 2.26), and confirmed that the drug produced inaccurate results on laboratory tests using biotin-tagged antibodies ⁷².

Three 2026 comparisons show that certification has not erased the differences. In 400 outpatient samples spread evenly across the clinical range, three widely used automated immunoassays all differed significantly from a mass-spectrometry method traceable to the Vitamin D Standardization-Certification Program (VDSCP), the CDC certification program that grew out of the standardization effort ^12,73: one consistently read high, one flagged too many people as deficient at the 75 nmol/L cutoff, and one agreed best, with kappa, a measure of agreement beyond chance, of 0.818 to 0.895; a laboratory's DEQAS performance did not consistently reflect how well it agreed on patient samples ⁷³. In 1064 paired outpatient samples in Milan, an automated mass-spectrometry analyzer, Thermo Scientific's Cascadion, read proportionally higher than a chemiluminescence immunoassay, one read out by emitted light: a regression slope of 1.33 and an average difference of 35.5 percent mean its results ran about a third higher ⁷⁴. The two ranked patients almost identically (intraclass correlation 0.928), but sorted them into six clinical categories with an unweighted kappa of only 0.32, or 0.72 when near-misses between neighboring categories count as partial agreement, and 0.56 with only two categories; among the pairs that disagreed, the immunoassay labeled patients deficient 118 times more often ⁷⁴. In its first preliminary evaluation, Roche's fully automated Cobas i601 mass-spectrometry system read 25-hydroxyvitamin D 16 percent lower than the same Madrid hospital's immunoassay at an intermediate decision level ⁷⁵; a second evaluation of the same instrument, in German commercial laboratories, found strong agreement with their existing mass-spectrometry method and bias against target values within 6.0 percent across the hormones tested ⁷⁶. A laboratory that changes method can reclassify a substantial share of its patients overnight ⁷⁴.

Units, thresholds, and what each was built on

The first source of confusion is arithmetic. American laboratories usually report nanograms per milliliter and most others nanomoles per liter, and the factor between them is about 2.5: the Institute of Medicine's 20 ng/mL is 50 nmol/L ⁷, and a parathyroid plateau at 78 nmol/L is 31 ng/mL ⁴⁹. A patient told they are at 50 in one country and 20 in another has been given the same number.

The Institute of Medicine (IOM) committee of 2011 set out to define how much vitamin D a population needs, not to diagnose individuals ⁷. It concluded that only bone health had evidence consistent with cause and effect, that the evidence for cancer, cardiovascular disease, diabetes and autoimmune disease was inconsistent and insufficient, and that intakes of 600 international units (IU) a day for ages 1 to 70 and 800 IU from age 71, corresponding to a serum level of at least 20 ng/mL, would meet the needs of at least 97.5 percent of the population ⁷. The 20 was therefore a level that almost everyone's requirement falls below, not a line below which a person is deficient. The committee warned that the prevalence of inadequacy in North America had been overestimated, noted U-shaped associations with risk at both low and high levels for some outcomes, and called for laboratory ranges to be reassessed ⁷.

Months later the Endocrine Society's guideline took the clinical view, judging deficiency very common in all age groups and suggesting testing of people at risk with a reliable assay ⁷⁷. It defined sufficiency as at least 30 ng/mL (75 nmol/L) and insufficiency as above 20 but below 30 ng/mL ¹³. Evidence that pointed toward a higher figure was in print: a 1998 letter from Holick's group titled 'Redefining vitamin D insufficiency' ⁶⁰; a 2006 review of studies on bone density, lower-limb function, dental health, falls, fractures and colorectal cancer, which concluded that the most advantageous levels began at 75 nmol/L (30 ng/mL) and were best at 90 to 100 nmol/L ⁶¹; in French adults parathyroid hormone began to climb below 31 ng/mL ⁴⁹; calcium absorption was 65 percent higher at a mean of 86.5 than at 50.2 nmol/L in a small study comparing two experiments a year apart ⁷⁸; and in 675 autopsy bone biopsies from northern Germany no one above 75 nmol/L had pathological accumulation of unmineralized bone, though the authors could find no level that inevitably produced it ⁷⁹. Evidence pointing the other way was also in print: in 492 women aged 20 to 80, calcium absorption efficiency showed no relation to 25-hydroxyvitamin D at all ⁸⁰. In 2012 members of the IOM committee published a rebuttal arguing that the Society's target levels, its definition of deficiency and its idea of who is at risk were not well supported ⁸¹.

The two figures coexisted for thirteen years ^7,13. A 2017 international conference in Pisa tried to narrow the ground: below 12 ng/mL (30 nmol/L) the risk of rickets and osteomalacia rises, 20 to 50 ng/mL appears safe and sufficient for the skeleton in the general population, and the variation between assays had contributed to the chaos ³.

In 2024 the Endocrine Society reversed itself. A new guideline panel concluded that the trial evidence does not permit any 25-hydroxyvitamin D threshold that predicts meaningful benefit from supplementation, rated its certainty very low, and announced that the Society no longer endorses its 2011 definitions of sufficiency and insufficiency ¹³. The companion guideline advised against routine testing in every population it considered ¹⁸. Its recommendations do not apply to people with established indications for vitamin D treatment or testing ¹⁸.

The reversal did not end the argument. A 2024 consensus from the sixth international conference on Controversies in Vitamin D said that optimal levels remain debated, blamed the lack of assay standardization for much of the difficulty in pooling studies, and attributed the negative trials largely to the enrollment of people who were not deficient ⁶⁸. In 2025 Holick, first author of the 2011 guideline ⁷⁷, published a critique titled 'The Debatable Clinical Utility of the 2024 Vitamin D Guideline' ⁶⁹. Other countries draw the line lower. The United Kingdom's Scientific Advisory Committee on Nutrition defined deficiency in 2016 as below 25 nmol/L (10 ng/mL), and in 2026 a multidisciplinary group called ACT NOW Vitamin D argued, in a position statement that is expert opinion rather than new data, for raising it to 50 nmol/L ¹⁴.

The testing boom, and the turn against screening

Between 2000 and 2010, among enrollees in Medicare Part B, the part of the US program for older Americans that pays for outpatient care and laboratory tests, the number of vitamin D tests reimbursed per 10,000 enrollees rose at least 83-fold: vitamin D was among the tests, alongside drug-of-abuse screens, whose volume in 2010 divided by that in the earliest year, usually 2000, ranged from 83 to 1510 ¹⁹. The 2024 Endocrine Society guideline traced the growth in testing, and in supplement use, to the associations between low levels and a variety of common disorders, associations it described as not clearly causal ¹⁸. The overuse persisted. A 2025 analysis of 2019 claims covering 232 million Americans with Medicare or commercial insurance found that 7 to 51 percent of four common tests, 25-hydroxyvitamin D among them, were repeated more often than recommended; the abstract does not give vitamin D's own share, and two of the three authors work for Diaceutics, a laboratory-data company ⁸².

Pushback came first from payers and specialty societies. In December 2010 Ontario stopped paying for vitamin D screening in people without specific indications, and screening fell by 92.7 percent; the Choosing Wisely recommendations against low-value vitamin D tests, released in the United States in February 2013 and in Canada in October 2014, were followed by much smaller reductions of 4.5 to 14.0 percent ⁶³. The authors concluded that recommendations alone may be insufficient without policy change ⁶³. A safety-net system, a public network that treats patients regardless of their ability to pay, of 11 hospitals and 70 clinics in New York cut orders by 44 percent for inpatients and 46 percent for outpatients after it required an indication with every order and added an alert against repeating the test within three months ⁸³.

The evidence reviews then asked the question that should have come first: does screening healthy people help? The review for the US Preventive Services Task Force found no study that had tested screening itself, and across the 46 studies it included, treating community-dwelling adults with low levels did not change mortality, fractures, falls, diabetes, cardiovascular disease, cancer or depression ²¹. The Task Force issued an I statement, meaning the evidence is insufficient to weigh benefits against harms, and noted that no single serum level defines deficiency ¹⁷. It was the Task Force's second such verdict; the first, published in January 2015, covered the same group, community-dwelling, nonpregnant adults seen in primary care ^16,17. The Endocrine Society's 2024 review of 151 studies likewise found no trial of screening ⁸⁴, and its guideline advised against routine testing in the general population and specifically in people with obesity or dark complexion ¹⁸.

The habit has not broken. At one Barcelona hospital, annual 25-hydroxyvitamin D testing rose a further 54.1 percent between 2020 and 2025, and the whole distribution of results shifted upward as the tail of very high values grew ³⁰. The authors called the pattern overdiagnosis followed by oversupplementation ³⁰.

When it reads low

Many low readings follow geography, season and skin ^22,23,49. Skin makes vitamin D when ultraviolet B light converts 7-dehydrocholesterol, a cholesterol relative in the skin, into previtamin D3, and latitude and season change how much of that band reaches the ground ²². Human skin samples exposed to cloudless winter sunlight in Boston, at 42.2 degrees north, made no previtamin D3 from November through February; in Edmonton, at 52 degrees north, the dead season ran from October through March, while at 34 and 18 degrees north sunlight still worked in midwinter ²². Those were laboratory measurements on skin and on the precursor, not on people, but population data match them: among 1569 French adults sampled between November and April, levels fell from south to north (correlation with latitude r = −0.79), and 14 percent were at or below 12 ng/mL ⁴⁹. The seasonal swing shows within the same people: across 268,877 health examinations of 69,131 Korean adults between 2014 and 2025, 25-hydroxyvitamin D peaked in August, while homocysteine, an amino acid in the blood with its own seasonal rhythm, peaked in early November, a rhythm that 25-hydroxyvitamin D did not explain ⁸⁵. Sunscreen, long suspected, seems to matter less than expected in real life: it blocked vitamin D production under artificial ultraviolet light, but two randomized field trials found no effect of daily use and most of 69 observational studies found no reduction, though no trial has tested the high sun protection factor (SPF) products now recommended ⁸⁶.

Skin pigment reduces how much vitamin D a given dose of light makes. In a 1982 experiment, one standard dose of ultraviolet radiation raised serum vitamin D up to 60-fold in two lightly pigmented volunteers but did not change it significantly in three heavily pigmented volunteers, one of whom needed six times the dose to reach the same level ²³. Five people is a small study.

Body size lowers the number, and the explanation has changed. In 2000, after whole-body ultraviolet exposure, obese adults' blood vitamin D3 rose 57 percent less than lean adults' although their skin held as much precursor, and the authors concluded that vitamin D was being deposited in fat, out of reach ⁸⁷. In 2012 an analysis of 686 unsupplemented adults found that a simple dilution model, the same vitamin spread through a larger body, removed the obesity-related variation entirely when body weight was used rather than body mass index (BMI), and found no evidence of sequestration ²⁴. A genetic study supports the direction of cause. In up to 42,024 adults from 21 cohorts, each 1 kg/m² higher BMI went with 1.15 percent lower 25-hydroxyvitamin D; a 10 percent higher genetically predicted BMI lowered it by 4.2 percent, while gene variants that lower vitamin D did not raise BMI ⁸⁸. The practical consequence is the same, dosing has to be adjusted for body size ²⁴, but the picture of fat as a vault that locks the vitamin away has given way to one of fat as a larger volume that dilutes it.

The gut, the liver and certain drugs can each lower it. After a single 50,000 IU dose of vitamin D2, serum vitamin D rose above 50 ng/mL within 12 hours in seven healthy people but stayed below 10 ng/mL in five of seven patients with fat malabsorption; two patients with severe inflammatory bowel disease, unexpectedly, absorbed normally ²⁵. Epileptic patients on high-dose anticonvulsants averaged 5 ng/mL against 16 ng/mL in healthy Londoners in 1972, which the authors attributed to drug-induced liver enzymes breaking vitamin D down ²⁶; a rare gain-of-function mutation in CYP3A4 does the same thing to the hormone and its precursor ⁵³. In cirrhosis, total 25-hydroxyvitamin D and its carriers fall together ²⁸, and the liver's own 25-hydroxylase can fail outright in the rare patient with a CYP2R1 mutation ³⁵. Chronic kidney disease, by contrast, lowered the hormone but not 25-hydroxyvitamin D3 in a large outpatient cohort ³⁸. Heavy loss of protein in the urine is different. In 106 adults with nephrotic syndrome, total and calculated bioavailable 25-hydroxyvitamin D were both lower than in 40 healthy controls, a fall the authors attributed to the binding protein leaking into the urine, and hip bone density tracked the bioavailable level (r = 0.358) but not the total (r = 0.174) ⁸⁹. It was a cross-sectional study, and the bioavailable levels were calculated rather than measured ⁸⁹.

Illness itself lowers the reading. In 33 patients having a knee replaced, 25-hydroxyvitamin D fell by about 40 percent within two days as the inflammatory marker C-reactive protein soared, had not recovered by day five, and was still 20 percent lower three months later ²⁷. The authors concluded that 25-hydroxyvitamin D is unlikely to be a reliable measure of status in anyone with a significant inflammatory response ²⁷. That finding matters far beyond surgery: some of the low levels seen in sick people are likely a consequence of illness rather than its cause ^27,90.

Magnesium may modify the number in either direction. In 180 participants of a randomized trial, magnesium supplementation raised 25-hydroxyvitamin D3 when baseline levels were near 30 ng/mL and lowered it when they were between 30 and 50 ⁹¹. That is a single interaction analysis within one trial, two of whose authors were from the Center for Magnesium Education and Research, and it measured vitamin D metabolites, not health outcomes ⁹¹.

Does low always mean deficient? The binding protein and race

A total measured in the blood is not necessarily what tissues can use ⁴¹. Most 25-hydroxyvitamin D is bound to its carrier ⁶, and the free-hormone hypothesis holds that the unbound fraction, or the fraction loosely held by albumin, is what matters ⁴¹. The idea has animal support: mice with no binding protein at all had low total vitamin D metabolites yet were otherwise normal on a replete diet, although they developed secondary hyperparathyroidism quickly when vitamin D was withdrawn ³⁹.

In 2013 the idea was applied to a long-noticed pattern: low total levels are common among Black Americans ⁶⁴. Among 2085 Baltimore adults, Black participants had much lower total 25-hydroxyvitamin D than white participants, 15.6 against 25.8 ng/mL, yet higher bone density, 1.05 against 0.94 g per square centimeter ⁶⁴. They also had about half the binding protein, 168 against 337 micrograms per milliliter, and common genetic variants of the protein explained 79.4 percent of its variation; among homozygous participants, those carrying two identical copies of the gene at both variant sites, estimated bioavailable 25-hydroxyvitamin D came out nearly identical, 2.9 and 3.1 ng/mL ⁶⁴. If that were right, the standard test overstated deficiency in Black Americans; the authors attributed the similar bioavailable levels to racial differences in common binding-protein variants ⁶⁴.

The binding-protein numbers did not survive. In 2015 a mass-spectrometry assay that could also identify the protein's genetic variants found isoform-specific errors in a widely used monoclonal immunoassay, and measured similar concentrations in white and Black participants, 262 against 266 micrograms per milliliter ⁶⁵. In 2016, in 1057 men from the United States, the United Kingdom and The Gambia, the monoclonal assay again read the protein about 50 percent lower in men of African ancestry while polyclonal antibody assays and proteomic methods, which identify proteins by mass spectrometry, showed no difference; directly measured free 25-hydroxyvitamin D tracked total 25-hydroxyvitamin D closely (r = 0.84) and was lower in African American than in white American men ⁶⁶. The authors concluded that the earlier racial difference in binding protein was most likely assay bias and that total 25-hydroxyvitamin D remains a useful measure of status in the general population ⁶⁶. The higher bone density measured in Black participants stands ⁶⁴, but the reassuring explanation for their low total levels came from an antibody's blind spot ⁶⁶.

The free-hormone idea itself has not been discarded; it has been narrowed to the settings where the carrier is truly abnormal, such as cirrhosis and nephrotic syndrome ^28,41,89. In 155 people, patients with cirrhosis had the lowest total 25-hydroxyvitamin D, binding protein and albumin but the highest directly measured free 25-hydroxyvitamin D, pregnant women had the most binding protein but ordinary free levels, and calculated free values explained only 13 percent of the variation in measured ones ²⁸. Measured free levels correlated with parathyroid hormone; calculated ones did not ²⁸. A 2025 review concluded that calculated free levels inherit the population and assay differences of the binding-protein measurement, that direct methods remain complicated or prone to interference, that mass spectrometry is promising, and that the clinical role of free 25-hydroxyvitamin D is still undetermined ⁴¹.

When it reads high

A high reading usually means supplements, typically extremely high doses taken by mistake or without supervision ¹⁵, so the question with a very high number is what was taken, how much and how often. Vitamin D toxicity is hypercalcemia, and its symptoms are confusion, apathy, recurrent vomiting, abdominal pain, heavy urination, thirst and dehydration; a 25-hydroxyvitamin D above 150 ng/mL (375 nmol/L) is the hallmark of toxicity from overdose ¹⁵. It was once assumed that the hormone itself caused the hypercalcemia of vitamin D poisoning. Animal experiments overturned that: in intoxicated animals 25-hydroxyvitamin D rose as high as 2.5 micromoles per liter while the hormone did not ⁵. The leading theories now are that 25-hydroxyvitamin D at such concentrations binds the vitamin D receptor despite its weaker grip ^5,46, or crowds the hormone off its carrier and raises its free level; occasional anecdotal reports from intoxicated people appear to support the second, and that is the extent of the human evidence the review cites ⁵. One reviewer judged that toxicity probably requires more than 750 nmol/L, about 300 ng/mL, but kept 250 nmol/L, about 100 ng/mL, as a prudent upper limit ⁵.

The case reports show how high the number can go. In 13 published reports, patients presented with 25-hydroxyvitamin D of 150 to 1220 ng/mL and serum calcium of 11.1 to 23.1 mg/dL, from manufacturing errors, overdosing by patients or prescribers, or both ²⁹. A 2025 report described three adults in one household who developed severe hypercalcemia and kidney injury within three days of each other after cooking with unpackaged oil later found to contain vitamin D at 1286 micrograms per milliliter; one needed dialysis and all recovered ⁹². In Kashmir, where very large intramuscular injections of 600,000 IU have been common, vitamin D toxicity accounted for eight of 40 hospital admissions for hypercalcemia in the first half of 2025, with 25-hydroxyvitamin D averaging 123.5 ng/mL, below the classic hallmark ⁹³. Eight patients in a retrospective series from one center cannot move the 150 ng/mL line, but they are a reminder that it is a marker drawn from case reports, not a cutoff tested in studies designed to find one ^15,29. Single cases show how the doses arise. A 17-year-old in Los Angeles who drank homemade milkshakes laced with vitamin D3 daily for about a year reached 594 ng/mL with a calcium of 15.5 mg/dL ⁹⁴, and an infant in Chennai given 6000 IU a day for 45 days instead of the recommended 400, after a mismatch between prescription and dispensing, had a level reported only as above 120 ng/mL, with high calcium ⁹⁵.

The rise is now visible in hospital data. At a Barcelona tertiary hospital, among 185,604 adult results between 2020 and 2025, results above 100 ng/mL more than doubled, from 31 (0.14 percent) to 108 (0.31 percent); calcium was available for 324 of 393 such results, and 14 (4.3 percent) showed hypercalcemia ³⁰. Calcifediol, 25-hydroxyvitamin D itself taken as a drug, was documented in 91.0 percent of the patients with hypervitaminosis whose supplementation was recorded and in 12 of the 14 with hypercalcemia, often on short-interval dosing ³⁰.

Below the toxic range, a high number is not obviously better. In a planned secondary analysis of a four-year trial in 2303 postmenopausal women, using standardized values, levels up to 60 ng/mL carried no extra risk of falls, while women at 60 ng/mL or more had about twice the odds of two or more falls (odds ratio 1.99), though they made up only about 2 percent of observations ⁹⁶. The Pisa conference placed the safe and sufficient range for the skeleton at 20 to 50 ng/mL ³.

Some high readings are false or reflect unusual biology. In infants, C3 epimers can make up most of the measured total on methods that do not separate them ³¹; high-dose biotin can distort streptavidin-based immunoassays ⁷¹; some automated assays consistently read above mass spectrometry ⁷³; and in a 76-year-old man with severe hypercalcemia who took only 400 IU a day, routine testing reported more than 250 nmol/L (100 ng/mL) while mass spectrometry gave 92 nmol/L. He turned out to have multiple myeloma, and the false result disappeared as treatment sharply lowered the abnormal antibody the cancer made ³². At the other end of the biology, people with faulty CYP24A1 cannot break down vitamin D metabolites and can become hypercalcemic on doses most people tolerate, including recommended doses and single large preventive doses in infants ^15,33.

What 25-hydroxyvitamin D causes: the genetic evidence

Low 25-hydroxyvitamin D is associated with a long list of common diseases, and the trials of supplements have mostly been negative ^18,21. Mendelian randomization offers a third way to ask whether the level itself causes anything. Because the gene variants that raise or lower 25-hydroxyvitamin D are dealt out at conception, they should not be tangled up with diet, weight, illness or wealth, and if people who inherit lower levels also have more of a disease, low vitamin D is more likely a cause ⁹⁷. The method has its own assumptions, chiefly that the variants affect the disease only through vitamin D; a variant that also acts by other routes, called pleiotropy, can mislead, and that possibility can never be fully excluded ⁹⁸.

A 2020 genetic study of the test, in 417,580 people of European ancestry, found 143 places in the genome that influence 25-hydroxyvitamin D, implicating genes for lipid transport, skin properties, and sulfonation and glucuronidation, chemical tags that prepare molecules for excretion ⁹⁰. Its Mendelian randomization analyses found no robust evidence that 25-hydroxyvitamin D causes any of the candidate traits it tested, among them body mass index and psychiatric disorders, but found that many of those traits change 25-hydroxyvitamin D ⁹⁰. For the traits it examined, it suggests that much of the association runs backward, from the condition to the number ⁹⁰. Observational studies keep showing why that matters. In 934 hip-fracture patients in Padua enrolled from 2023 to 2025, deficiency's association with death within a year (hazard ratio 1.65) shrank to 1.31, with a 95 percent confidence interval of 0.88 to 1.96, after adjustment for age, sex and illness burden ⁹⁹. In 35,887 American adults surveyed between 2007 and 2023, higher levels went with more osteoarthritis, which the authors attributed to reverse causation, such as people with the disease taking supplements, rather than to harm ¹⁰⁰.

Multiple sclerosis is the exception. Using four variants in 14,498 people with the disease and 24,091 controls, each genetically determined fall of one standard deviation, a standard step in a measure's spread, in 25-hydroxyvitamin D measured on a logarithmic scale doubled the odds of multiple sclerosis (odds ratio 2.0, 95 percent confidence interval 1.7 to 2.5), and the result held at 1.7 after removing variants that might act through other pathways, though such pleiotropy can never be fully excluded ⁹⁸. In 2025 a French trial gave 316 untreated patients with a first multiple sclerosis-like episode and 25-hydroxyvitamin D below 100 nmol/L either 100,000 IU of vitamin D3 every two weeks or placebo for two years: disease activity, a combination of relapses and new lesions on magnetic resonance imaging (MRI), occurred in 60.3 against 74.1 percent (hazard ratio 0.66, 95 percent confidence interval 0.50 to 0.87), but clinical relapses did not differ significantly, 17.9 against 21.8 percent (hazard ratio 0.69, 95 percent confidence interval 0.42 to 1.16) ¹⁰¹. The genetic evidence concerns who develops multiple sclerosis; the trial concerns what happens after a first episode, and its effect was on imaging, with none of its 10 clinical secondary outcomes significant ¹⁰¹. Both favor vitamin D, but they answer different questions, and neither shows an effect on disability over decades ^98,101.

The most dramatic genetic claim came from a refinement called nonlinear Mendelian randomization, which splits the population into strata and asks whether the effect is concentrated among people with low levels ^102,103. In 2021 a consortium analyzing 386,406 people reported that, among those below 25 nmol/L, each 10 nmol/L higher genetically predicted level cut the odds of death by 31 percent (odds ratio 0.69), with benefit up to about 40 nmol/L ¹⁰². In 2022 a UK Biobank analysis of 307,601 people found an L-shaped curve, with the odds of death 25 percent higher at 25 than at 50 nmol/L ¹⁰⁴. Both papers appeared to show what the trials had not: that correcting true deficiency saves lives ^102,104.

Then the method itself was examined. The residual method used in the 2021 paper divides people by the part of their 25-hydroxyvitamin D that their genes do not predict, and it assumes that the genetic score, the combined effect of the variants that raise 25-hydroxyvitamin D, shifts the level by the same amount in everyone ^97,103. In 2023 a reanalysis found that the score's effect on 25-hydroxyvitamin D differed about five-fold between the lowest and highest tenths of the population, and that the protective effect in people with low levels vanished with an alternative way of forming the strata, the doubly ranked method ¹⁰³. The journal placed an expression of concern on the 2021 paper in July 2023 ¹⁰⁵, and in December retracted and republished it ¹⁰⁶. The republished version found no association of genetically predicted 25-hydroxyvitamin D with coronary disease, stroke or death at any level, and concluded that substantial reductions in mortality from low-dose supplementation are unlikely even in people with low status ¹⁰⁷. In 2024 a Bristol group using negative-control outcomes, diseases that vitamin D should not affect, found clearly biased estimates from both nonlinear methods in some situations ⁹⁷. Writing in 2024, Stender, Gellert-Kristensen and Davey Smith pointed to these vitamin D findings, then being retracted or corrected, as the example of new methods applied before they had been properly tested ¹⁰⁸. The UK Biobank paper received a correction in 2025, the content of which the PubMed record does not summarize ¹⁰⁹. A 2026 method that stratifies by recruitment center rather than by modeled exposure found no evidence of a causal effect of vitamin D on coronary artery disease across 20 UK Biobank centers, though the centers differed only modestly in vitamin D levels ¹¹⁰.

The genetic literature has since been flooded. Public summary statistics make two-sample Mendelian randomization quick to run, and the same 2024 letter advised journals to reject papers that report nothing else ¹⁰⁸. In 2025 and 2026 such studies reported protective causal effects of 25-hydroxyvitamin D on hypothyroidism ¹¹¹ and on laryngeal cancer, the latter with an odds ratio of 0.9993, an effect its authors called slight ¹¹². A 2026 study that genotyped its own participants, 2061 community-dwelling adults and 200 patients with vertebral fractures in Shanghai, found no effect on bone density or vertebral fractures, though it reported one on a marker of bone formation in older participants ¹¹³.

Testing and treating to a number: the trials

If the test has a purpose beyond diagnosing rickets, it should identify people who benefit from treatment. The trials that selected people by their level, or treated to a target, are the direct test of that idea. The US Preventive Services Task Force's review found that, among community-dwelling adults with low levels, treatment did not change mortality, fractures, falls, diabetes, cardiovascular disease, cancer or depression ²¹. The usual defense of supplementation is that the big trials failed because they enrolled people who were not deficient ⁶⁸. That has been tested only indirectly. In VITAL, a trial of 25,871 American adults not selected for deficiency, 2000 IU a day did not reduce total fractures (hazard ratio 0.98, 95 percent confidence interval 0.89 to 1.08), and the effect did not differ by baseline 25-hydroxyvitamin D ¹¹⁴. Because its participants were not chosen for low levels, that comparison is a weaker test than a trial confined to deficient people would be ¹¹⁴.

CORONAVIT tested the strategy at population scale during the COVID-19 pandemic ¹¹⁵. Of 3100 British adults offered a postal finger-prick test, 2958 accepted and 2674, 86.3 percent, were below 75 nmol/L and were sent six months of vitamin D at 800 or 3200 IU a day; compared with 3100 people offered nothing, acute respiratory infections occurred in 5.7 and 5.0 percent against 4.6 percent, odds ratios of 1.26 and 1.09, and COVID-19 was not reduced either ¹¹⁵. Test-and-treat did not work for the outcome it was designed to prevent.

TARGET-D, published in September 2026, tested a more aggressive version in people who had had a heart attack, in a pragmatic trial, one run within routine care, based at Intermountain Medical Center in Utah: each patient's dose of vitamin D3 was adjusted by a set algorithm to reach and hold a target range of 40 to 80 ng/mL ⁷⁰. Of 630 patients, with a median baseline of 25 ng/mL and 87 percent at or below 40, those randomized to the target had major adverse cardiovascular events in 15.7 percent against 18.4 percent with usual care, a difference that was not significant (hazard ratio 0.85, 95 percent confidence interval 0.58 to 1.24; P = .40, meaning that a difference this size could easily arise by chance) ⁷⁰. Among the secondary end points, repeat heart attack occurred in 3.8 against 7.9 percent (hazard ratio 0.48, P = .03), while death, heart failure admission and stroke did not differ ⁷⁰. A primary result that misses and a secondary one that hits is a reason for another trial, not a conclusion, and the control group received usual care rather than placebo ⁷⁰.

The best positive evidence for treating to a level comes from prediabetes, and it is weaker than it looks. In an individual-participant meta-analysis of three trials, which pooled the raw records of every participant rather than each trial's summary results, vitamin D reduced new diabetes by 15 percent (hazard ratio 0.85), and among people given cholecalciferol, those who maintained at least 125 nmol/L had a 76 percent lower risk than those at 50 to 74 nmol/L (hazard ratio 0.24) ¹¹⁶. The second comparison is between people who reached different levels, not between randomized groups, so whatever made some people reach higher levels may also have protected them. One of the authors was a Takeda Pharmaceuticals employee, and one of the three trials tested eldecalcitol, a vitamin D analog ¹¹⁶.

In the disease where the genetic evidence for causation is strongest ⁹⁸, the test was used only to select participants: the French multiple sclerosis trial enrolled people below 100 nmol/L and gave everyone in the treatment arm the same dose ¹⁰¹.

What the test misses

It misses the hormone. A normal 25-hydroxyvitamin D says nothing about whether the kidney is making 1,25-dihydroxyvitamin D, and a low one says little about whether it has stopped: the hormone was held normal in deficient patients until 25-hydroxyvitamin D fell to about 10 nmol/L ⁴³, and in chronic kidney disease the hormone fell while the precursor did not ³⁸.

It misses what tissues make for themselves. Granulomas and lymphomas make the hormone locally ^46,47, and neither the amount nor the effect shows up in 25-hydroxyvitamin D.

It misses the body's response. Parathyroid hormone is the body's own readout of whether the supply is adequate ⁴⁹, and the relationship varies between people: at every level of parathyroid hormone, Black participants in the Baltimore study had lower total 25-hydroxyvitamin D than white participants ⁶⁴. Calcium absorption, the job the hormone was first identified doing ³⁴, was unrelated to 25-hydroxyvitamin D in 492 women ⁸⁰.

It misses breakdown. Below 25 nmol/L of 25-hydroxyvitamin D the body nearly stops making 24,25-dihydroxyvitamin D ⁵⁰, and the ratio of the two, the vitamin D metabolite ratio, has been proposed as a marker of functional deficiency. In 204 Belgian older adults the ratio correlated more strongly with parathyroid hormone than either metabolite alone, but in adjusted models both a low ratio and a low 25-hydroxyvitamin D were risk factors for death ¹¹⁷. A 2026 study of healthy Japanese adults proposed tentative reference intervals on a fully automated analyzer, 3.0 to 11.8 percent in men and 2.5 to 10.4 percent in women, with several authors employed by Shimadzu, an instrument maker ¹¹⁸.

It misses the free fraction ⁴¹, it is pulled down by acute illness ²⁷, and magnesium supplementation can shift it up or down depending on the starting level ⁹¹. And it does not, on any evidence yet published, identify healthy people who will benefit from treatment ^13,21.

The 2025 to 2026 literature, verified

The trials. In April 2025 the D-Lay MS trial reported that 100,000 IU of vitamin D3 every two weeks reduced disease activity after a first multiple sclerosis-like episode, hazard ratio 0.66, with the effect seen on MRI rather than in relapses ¹⁰¹. In September 2026 TARGET-D reported that treating heart attack patients to a 25-hydroxyvitamin D target range of 40 to 80 ng/mL did not reduce major cardiovascular events, hazard ratio 0.85, with a secondary reduction in repeat heart attack, hazard ratio 0.48 ⁷⁰.

The laboratory. Two 2026 comparisons found clinically important disagreement between widely used methods: three automated immunoassays differed from mass spectrometry in 400 samples, with external quality scores failing to predict patient-sample agreement ⁷³, and in 1064 paired outpatient samples an immunoassay called deficiency 118 times more often than mass spectrometry among the pairs that disagreed ⁷⁴. An early evaluation of Roche's new fully automated mass-spectrometry analyzer, the Cobas i601, read 25-hydroxyvitamin D 16 percent below the hospital's immunoassay at an intermediate decision level ⁷⁵, while a second found close agreement with conventional mass spectrometry ⁷⁶. In 128 children under three, measured by mass spectrometry in dried blood spots, subtracting the C3 epimer changed the deficiency classification in 17.2 percent, and in 54.5 percent of those between 18 and 22 ng/mL; one author is from Masdiag, a Warsaw company ¹¹⁹. A 2026 case report traced an apparent vitamin D intoxication above 250 nmol/L to interference from a myeloma protein; mass spectrometry gave 92 nmol/L ³². A 2025 claims analysis found that a large share of four common tests, vitamin D among them, were repeated more often than recommended ⁸². A 2025 review set out why free 25-hydroxyvitamin D is still a research measurement ⁴¹, and a 2026 study proposed reference intervals for the metabolite ratio ¹¹⁸.

Toxicity. A 2026 Barcelona audit found testing up 54.1 percent and results above 100 ng/mL more than doubled between 2020 and 2025, with calcifediol behind most of them ³⁰. A 2025 case series traced a family outbreak of hypercalcemia to contaminated cooking oil ⁹², and a 2026 Kashmir series attributed eight of 40 hypercalcemia admissions to vitamin D ⁹³. Single 2025 cases included an adolescent at 594 ng/mL from homemade fortified milkshakes ⁹⁴ and an infant poisoned by a dispensing error ⁹⁵; a 2026 case unmasked CYP24A1 deficiency in pregnancy ⁵¹, and another showed a myeloma protein faking intoxication ³².

Causation. A correction to the 2022 UK Biobank nonlinear analysis was published in June 2025 ¹⁰⁹, and a 2026 statistical method using recruitment centers rather than modeled exposure found no evidence of a causal effect of vitamin D on coronary artery disease ¹¹⁰. Two 2026 observational studies showed confounding and reverse causation at work: an association of deficiency with death after hip fracture largely disappeared after adjustment for illness ⁹⁹, and higher levels went with more osteoarthritis, probably because patients take supplements after diagnosis ¹⁰⁰. Nothing we found published in 2025 or 2026 restored the 2011 thresholds or the case for screening healthy adults. The argument has not stopped: the first author of the 2011 guideline published a critique of the 2024 guideline in 2025 ⁶⁹, and a UK group in 2026 called for raising that country's deficiency line from 25 to 50 nmol/L ¹⁴. Meanwhile two-sample Mendelian randomization papers claiming causal effects continued to appear ^111,112, against a 2024 call to stop publishing them without supporting evidence ¹⁰⁸.

What was believed, and is not

That a normal level of the active hormone means a normal vitamin D supply. The hormone is defended by the parathyroid glands and stays normal until deficiency is severe ⁴³, which is why the precursor, not the hormone, is the test of supply ^3,43.

That the hormone causes vitamin D poisoning. Early assumptions that 1,25-dihydroxyvitamin D drove the hypercalcemia were replaced once animal intoxication showed that the hormone does not rise ⁵.

That a vitamin D level is a vitamin D level. In 2004 the laboratory determined the diagnosis ²⁰, in 2021 about half the assays still missed the accuracy goal ^2,36, and in 2026 a change of method could still reclassify a substantial share of patients ⁷⁴.

That 30 ng/mL is the floor of sufficiency. It was the Endocrine Society's definition from 2011 until 2024, when the Society withdrew it because no trial evidence identifies a threshold that predicts benefit ¹³.

That low total 25-hydroxyvitamin D in Black Americans is offset by low binding protein. The binding-protein difference was an artifact of a monoclonal assay, and directly measured free 25-hydroxyvitamin D is lower too ^65,66.

That fat locks vitamin D away. Dilution in a larger body explains the lower levels of obesity without any sequestration ²⁴, replacing the earlier interpretation ⁸⁷.

That genetics proved deficiency kills. The nonlinear Mendelian randomization result was retracted and republished as null ¹⁰⁷, after the method's central assumption was shown to fail for vitamin D ^97,103.

That everyone should be screened. No trial has tested screening, and both the Task Force and the Endocrine Society declined to endorse it ^17,18.

That much of the population is vitamin D deficient, as a fact rather than a definition. The figures depended on the cutoff. A 2006 review placed the most advantageous levels at 75 nmol/L (30 ng/mL) and above ⁶¹, the 2011 Endocrine Society guideline judged deficiency very common in all age groups ⁷⁷, the Institute of Medicine committee concluded that the prevalence of inadequacy had been overestimated ⁷, in 2016 three authors of that committee's summary asked in print whether there was really a pandemic ⁶⁷, and in 2024 the Endocrine Society withdrew its 30 ng/mL definition of sufficiency ¹³.

That a low level during an acute illness means deficiency. After knee replacement, 25-hydroxyvitamin D fell about 40 percent in two days as inflammation rose, and the authors judged the test unreliable in anyone with a significant inflammatory response ²⁷.

What is still unknown

Where deficiency begins for an individual. The trials do not identify any threshold that predicts benefit ¹³, and the Pisa consensus that levels below 12 ng/mL carry rickets risk rests on bone, not on the many other outcomes the test is ordered for ³.

Whether a better marker exists. The metabolite ratio ¹¹⁷, directly measured free 25-hydroxyvitamin D ⁴¹ and the newer metabolites all lack standardized methods and adequate outcome data, and even 24,25-dihydroxyvitamin D was measured with mean biases of up to 36 percent among expert laboratories ¹²⁰.

Whether treating to a target helps anyone in particular. TARGET-D's reduction in repeat heart attack was a secondary end point ⁷⁰, the prediabetes benefit at high achieved levels was not a randomized comparison ¹¹⁶, and the multiple sclerosis effect was on imaging rather than relapses ¹⁰¹.

Whether the negative trials missed a benefit confined to truly deficient people. The 2024 consensus argues that they did ⁶⁸; VITAL found no effect modification by baseline level ¹¹⁴, and test-and-treat in CORONAVIT did not prevent infections ¹¹⁵, but none of the trials described here enrolled only people below 12 ng/mL, the level the Pisa conference tied to rickets risk ^3,70,101,115.

Whether today's high-SPF sunscreens lower levels, since none has been tested in a field trial ⁸⁶; how much of the low levels seen in sick people is a consequence of illness ^27,90; and whether automated mass spectrometry will finally end the disagreement between laboratories or simply move everyone's numbers ^74,75,76.

Where it connects

Sources

120 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Blunt JW, DeLuca HF, Schnoes HK. 25-hydroxycholecalciferol. A biologically active metabolite of vitamin D3.Biochemistry · 1968 · 7(10):3317-22doi:10.1021/bi00850a001 · PMID 4300699

   The 1968 identification of 25-hydroxycholecalciferol, the molecule the test measures, as a biologically active metabolite of vitamin D3. PubMed holds no abstract, so it is cited only for the discovery and date its title records.

2. 2
   Wise SA, Camara JE, Burdette CQ, et al. Interlaboratory comparison of 25-hydroxyvitamin D assays: Vitamin D Standardization Program (VDSP) Intercomparison Study 2 - Part 1 liquid chromatography - tandem mass spectrometry (LC-MS/MS) assays - impact of 3-epi-25-hydroxyvitamin D3 on assay performance.Anal Bioanal Chem · 2021 · 414(1):333-349doi:10.1007/s00216-021-03576-1 · PMID 34432104

   Fifty single-donor sera with reference values sent to 14 laboratories using 14 LC-MS/MS methods: 53 percent met the 5 percent bias goal, four were off by 12 to 21 percent, and failure to separate the 3-epimer drove much of the bias. Co-authors include employees of Chromsystems, Waters, LabCorp and Endoceutics. PubMed dates it 2021; the issue is 2022.

3. 3
   Sempos CT, Heijboer AC, Bikle DD, et al. Vitamin D assays and the definition of hypovitaminosis D: results from the First International Conference on Controversies in Vitamin D.Br J Clin Pharmacol · 2018 · 84(10):2194-2207doi:10.1111/bcp.13652 · PMID 29851137

   Consensus of the 2017 Pisa conference: total 25-hydroxyvitamin D (D2 plus D3) remains the measure of status; assay variation has produced chaos; below 12 ng/mL (30 nmol/L) rickets and osteomalacia risk rises; 20 to 50 ng/mL appears safe and sufficient for skeletal health. Expert opinion.

4. 4
   Jones KS, Assar S, Harnpanich D, et al. 25(OH)D2 half-life is shorter than 25(OH)D3 half-life and is influenced by DBP concentration and genotype.J Clin Endocrinol Metab · 2014 · 99(9):3373-81doi:10.1210/jc.2014-1714 · PMID 24885631

   Deuterated tracers in 36 healthy men from The Gambia and the UK: half-life 15.1 days for 25-hydroxyvitamin D3 and 13.9 days for D2, longer with more binding protein. Small, men only.

5. 5
   Jones G. Pharmacokinetics of vitamin D toxicity.Am J Clin Nutr · 2008 · 88(2):582S-586Sdoi:10.1093/ajcn/88.2.582S · PMID 18689406

   Review: vitamin D half-life about 2 months, 25-hydroxyvitamin D3 about 15 days at 25 to 200 nmol/L, the hormone about 15 hours; animal intoxication raises 25-hydroxyvitamin D to 2.5 micromol/L without raising the hormone; toxicity probably needs more than 750 nmol/L, with 250 nmol/L kept as a prudent upper limit. Much of the toxicity mechanism comes from animal work and anecdote.

6. 6
   Bikle DD, Gee E, Halloran B, Kowalski MA, Ryzen E, Haddad JG. Assessment of the free fraction of 25-hydroxyvitamin D in serum and its regulation by albumin and the vitamin D-binding protein.J Clin Endocrinol Metab · 1986 · 63(4):954-9doi:10.1210/jcem-63-4-954 · PMID 3745408

   88 percent of 25-hydroxyvitamin D bound to binding protein in normal serum and 83 percent in cirrhosis, the rest mainly to albumin; changes in the carriers changed total and free fractions but not the free concentration. Small numbers of subjects.

7. 7
   Ross AC, Manson JE, Abrams SA, et al. The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know.J Clin Endocrinol Metab · 2011 · 96(1):53-8doi:10.1210/jc.2010-2704 · PMID 21118827

   The Institute of Medicine committee's summary: requirements based on bone health; extraskeletal evidence inconsistent; Recommended Dietary Allowances of 600 IU (ages 1 to 70) and 800 IU (71 and older) correspond to at least 20 ng/mL (50 nmol/L), covering 97.5 percent of the population; U-shaped associations for some outcomes; prevalence of inadequacy overestimated. PubMed dates it November 2010.

8. 8
   Carter GD, Berry J, Durazo-Arvizu R, et al. Hydroxyvitamin D assays: An historical perspective from DEQAS.J Steroid Biochem Mol Biol · 2017 · 177:30-35doi:10.1016/j.jsbmb.2017.07.018 · PMID 28734989

   History told by the Vitamin D External Quality Assessment Scheme's own advisory panel, so an insider's account: first distribution in June 1989 to 13 UK laboratories with interlaboratory coefficients of variation of 29.3 to 53.7 percent; 918 participants in 56 countries with coefficients of variation of 10.3 to 15.3 percent in January 2017; the 1992 iodinated tracer used in a DiaSorin kit; automated non-extraction assays; accuracy-based assessment against the NIST method from 2013. PubMed dates the record 2017; volume 177 is the 2018 print issue.

9. 9
   Phinney KW, Bedner M, Tai SS, et al. Development and certification of a standard reference material for vitamin D metabolites in human serum.Anal Chem · 2012 · 84(2):956-62doi:10.1021/ac202047n · PMID 22141317

   NIST SRM 972: four serum pools with certified values for 25-hydroxyvitamin D2, D3 and the 3-epimer, assigned by three isotope-dilution mass spectrometry approaches at NIST and the CDC; separating the epimer proved essential. PubMed dates the record December 2011.

10. 10
    Phinney KW, Tai SS, Bedner M, et al. Development of an Improved Standard Reference Material for Vitamin D Metabolites in Human Serum.Anal Chem · 2017 · 89(9):4907-4913doi:10.1021/acs.analchem.6b05168 · PMID 28375002

    SRM 972a, the replacement for SRM 972, with three of its four levels made of unmodified human serum and all value-assignment methods resolving the epimer. Includes a Health Canada collaboration; no commercial authors listed.

11. 11
    Tai SS, Nelson MA, Bedner M, et al. Development of Standard Reference Material (SRM) 2973 Vitamin D Metabolites in Frozen Human Serum (High Level).J AOAC Int · 2017 · 100(5):1294-1303doi:10.5740/jaoacint.17-0182 · PMID 28917261

    SRM 2973, a serum near 100 nmol/L with a certified 25-hydroxyvitamin D3 of 98.4 nmol/L and a 3-epimer reference value of 5.23 nmol/L, plus 24,25-dihydroxyvitamin D3 values for SRM 972a and 2973.

12. 12
    Sempos CT, Vesper HW, Phinney KW, Thienpont LM, Coates PM. Vitamin D status as an international issue: national surveys and the problem of standardization.Scand J Clin Lab Invest Suppl · 2012 · 243:32-40doi:10.3109/00365513.2012.681935 · PMID 22536760

    The founding description of the Vitamin D Standardization Program (the Office of Dietary Supplements of the National Institutes of Health with the Centers for Disease Control and Prevention, NIST and Ghent University, November 2010): performance goals of 10 percent or less imprecision and 5 percent or less bias, master equations to re-standardize old survey data, and a planned certification program at the Centers for Disease Control and Prevention. A program description by its organizers.

13. 13
    McCartney CR, McDonnell ME, Corrigan MD, Lash RW. Vitamin D Insufficiency and Epistemic Humility: An Endocrine Society Guideline Communication.J Clin Endocrinol Metab · 2024 · 109(8):1948-1954doi:10.1210/clinem/dgae322 · PMID 38828961

    The Endocrine Society no longer endorses its 2011 definitions of sufficiency (at least 30 ng/mL) and insufficiency (20 to 30 ng/mL), because trial evidence does not identify thresholds that predict benefit; certainty very low. Two authors are Endocrine Society staff.

14. 14
    Christie S, Uday S, Hill TR, et al. Vitamin D in the UK: an urgent call to redefine the threshold for deficiency.Front Endocrinol (Lausanne) · 2026 · 17:1870408doi:10.3389/fendo.2026.1870408 · PMID 42453499

    Position statement from the ACT NOW Vitamin D collective: the UK definition of deficiency, below 25 nmol/L, set by the Scientific Advisory Committee on Nutrition in 2016, may underestimate population risk; it argues for raising the threshold to below 50 nmol/L. Opinion drawing on national survey data; one author is from a nutrition communications consultancy.

15. 15
    Marcinowska-Suchowierska E, Kupisz-Urbańska M, Łukaszkiewicz J, Płudowski P, Jones G. Vitamin D Toxicity-A Clinical Perspective.Front Endocrinol (Lausanne) · 2018 · 9:550doi:10.3389/fendo.2018.00550 · PMID 30294301

    Review: symptoms of toxicity (confusion, apathy, vomiting, abdominal pain, polyuria, polydipsia, dehydration); 25-hydroxyvitamin D above 150 ng/mL (375 nmol/L) as the hallmark of overdose; endogenous forms from granulomas, lymphomas and CYP24A1 loss.

16. 16
    LeFevre ML; U.S. Preventive Services Task Force. Screening for vitamin D deficiency in adults: U.S. Preventive Services Task Force recommendation statement.Ann Intern Med · 2015 · 162(2):133-40doi:10.7326/M14-2450 · PMID 25419853

    The Task Force's first I statement, published 20 January 2015: evidence insufficient to assess the benefits and harms of screening community-dwelling, nonpregnant adults seen in primary care who have no signs of deficiency.

17. 17
    US Preventive Services Task Force; Krist AH, Davidson KW, Mangione CM, et al. Screening for Vitamin D Deficiency in Adults: US Preventive Services Task Force Recommendation Statement.JAMA · 2021 · 325(14):1436-1442doi:10.1001/jama.2021.3069 · PMID 33847711

    An I statement: evidence insufficient to assess benefits and harms of screening asymptomatic adults; no single serum level defines deficiency. Updates the Task Force's 2014 recommendation, published in January 2015, for the same community-dwelling, nonpregnant adults.

18. 18
    Demay MB, Pittas AG, Bikle DD, et al. Vitamin D for the Prevention of Disease: An Endocrine Society Clinical Practice Guideline.J Clin Endocrinol Metab · 2024 · 109(8):1907-1947doi:10.1210/clinem/dgae290 · PMID 38828931

    The 2024 guideline: empiric vitamin D for ages 1 to 18, adults over 75, pregnancy and high-risk prediabetes; against routine 25-hydroxyvitamin D testing in all populations considered, including people with obesity or dark complexion, because no trial supports screening and no target level is defined.

19. 19
    Shahangian S, Alspach TD, Astles JR, Yesupriya A, Dettwyler WK. Trends in laboratory test volumes for Medicare Part B reimbursements, 2000-2010.Arch Pathol Lab Med · 2014 · 138(2):189-203doi:10.5858/arpa.2013-0149-OA · PMID 23738761

    Centers for Disease Control and Prevention analysis of Medicare Part B claims per 10,000 enrollees: vitamin D was among the tests whose volume ratios from 2000 to 2010 ranged from 83 to 1510. The abstract does not give vitamin D's own figure, so only the lower bound is used. PubMed dates the record 2013.

20. 20
    Binkley N, Krueger D, Cowgill CS, et al. Assay variation confounds the diagnosis of hypovitaminosis D: a call for standardization.J Clin Endocrinol Metab · 2004 · 89(7):3152-7doi:10.1210/jc.2003-031979 · PMID 15240586

    Two laboratories gave mean values of 46 and 21 ng/mL for clinically similar postmenopausal women (17 against 90 percent below 32 ng/mL), and split samples from 10 adults sent to six laboratories gave means from 17.1 to 35.6 ng/mL. Small samples; it shows how large the disagreement was, not which laboratory was right.

21. 21
    Kahwati LC, LeBlanc E, Weber RP, et al. Screening for Vitamin D Deficiency in Adults: Updated Evidence Report and Systematic Review for the US Preventive Services Task Force.JAMA · 2021 · 325(14):1443-1463doi:10.1001/jama.2020.26498 · PMID 33847712

    46 studies (16,205 people): no study tested screening itself; among community-dwelling adults with low levels, treatment did not change mortality, fractures, falls, diabetes, cardiovascular disease, cancer or depression.

22. 22
    Webb AR, Kline L, Holick MF. Influence of season and latitude on the cutaneous synthesis of vitamin D3: exposure to winter sunlight in Boston and Edmonton will not promote vitamin D3 synthesis in human skin.J Clin Endocrinol Metab · 1988 · 67(2):373-8doi:10.1210/jcem-67-2-373 · PMID 2839537

    Human skin samples and 7-dehydrocholesterol exposed to sunlight made no previtamin D3 in Boston (42.2 degrees N) from November to February or in Edmonton (52 degrees N) from October to March, but did in midwinter at 34 and 18 degrees N. An ex vivo model, not measurements in people.

23. 23
    Clemens TL, Adams JS, Henderson SL, Holick MF. Increased skin pigment reduces the capacity of skin to synthesise vitamin D3.Lancet · 1982 · 1(8263):74-6doi:10.1016/s0140-6736(82)90214-8 · PMID 6119494

    One standard ultraviolet dose raised serum vitamin D up to 60-fold in two lightly pigmented volunteers but not significantly in three heavily pigmented ones; one needed six times the dose. Five people.

24. 24
    Drincic AT, Armas LA, Van Diest EE, Heaney RP. Volumetric dilution, rather than sequestration best explains the low vitamin D status of obesity.Obesity (Silver Spring) · 2012 · 20(7):1444-8doi:10.1038/oby.2011.404 · PMID 22262154

    In 686 unsupplemented adults a hyperbolic fit of 25-hydroxyvitamin D against total body weight removed the obesity-related variation; no evidence of sequestration. Cross-sectional modeling.

25. 25
    Lo CW, Paris PW, Clemens TL, Nolan J, Holick MF. Vitamin D absorption in healthy subjects and in patients with intestinal malabsorption syndromes.Am J Clin Nutr · 1985 · 42(4):644-9doi:10.1093/ajcn/42.4.644 · PMID 4050723

    After 50,000 IU of D2, serum vitamin D rose above 50 ng/mL by 12 hours in seven healthy people but stayed under 10 ng/mL in five of seven with fat malabsorption; two with severe inflammatory bowel disease absorbed normally. Fourteen people.

26. 26
    Stamp TC, Round JM, Rowe DJ, Haddad JG. Plasma levels and therapeutic effect of 25-hydroxycholecalciferol in epileptic patients taking anticonvulsant drugs.Br Med J · 1972 · 4(5831):9-12doi:10.1136/bmj.4.5831.9 · PMID 4342760

    Early clinical use of the protein-binding assay: mean 16 ng/mL in normal London adults and schoolchildren against 5 ng/mL in epileptic patients on high-dose anticonvulsants, and rapid healing of two patients with osteomalacia on small doses of 25-hydroxycholecalciferol. Small observational study; the liver enzyme-induction explanation is the authors' interpretation, not something the study tested directly.

27. 27
    Reid D, Toole BJ, Knox S, et al. The relation between acute changes in the systemic inflammatory response and plasma 25-hydroxyvitamin D concentrations after elective knee arthroplasty.Am J Clin Nutr · 2011 · 93(5):1006-11doi:10.3945/ajcn.110.008490 · PMID 21411617

    In 33 patients, 25-hydroxyvitamin D fell about 40 percent by day 2 after knee replacement as C-reactive protein rose, and was still 20 percent lower at 3 months. Small, single procedure; shows a low level can be a consequence of illness.

28. 28
    Schwartz JB, Lai J, Lizaola B, et al. A comparison of measured and calculated free 25(OH) vitamin D levels in clinical populations.J Clin Endocrinol Metab · 2014 · 99(5):1631-7doi:10.1210/jc.2013-3874 · PMID 24483159

    In 155 people, cirrhosis gave the lowest total 25-hydroxyvitamin D but the highest directly measured free level; pregnancy raised binding protein without changing free levels; calculated free values explained only 13 percent of measured ones. The free assay was an immunoassay; small groups.

29. 29
    Galior K, Grebe S, Singh R. Development of Vitamin D Toxicity from Overcorrection of Vitamin D Deficiency: A Review of Case Reports.Nutrients · 2018 · 10(8):953doi:10.3390/nu10080953 · PMID 30042334

    Thirteen case reports: 25-hydroxyvitamin D 150 to 1220 ng/mL and calcium 11.1 to 23.1 mg/dL, from manufacturing errors and overdosing by patients or prescribers. Case reports select severe cases.

30. 30
    Fabregat-Bolufer AB, Martín-Alonso C, Escolà-Rodríguez A, Morales-Ruiz M. Overdiagnosed and oversupplemented: the iatrogenic rise of vitamin D toxicity.Eur J Endocrinol · 2026 · 194(6):K21-K25doi:10.1093/ejendo/lvag088 · PMID 42145024

    185,604 adult results at a Barcelona hospital, 2020 to 2025, on a certified immunoassay: testing up 54.1 percent, median 25.3 to 28.6 ng/mL, results above 100 ng/mL from 31 (0.14 percent) to 108 (0.31 percent), hypercalcemia in 14 of 324 with calcium measured, calcifediol in 91.0 percent of those with dosing data. Retrospective, one center.

31. 31
    Singh RJ, Taylor RL, Reddy GS, Grebe SK. C-3 epimers can account for a significant proportion of total circulating 25-hydroxyvitamin D in infants, complicating accurate measurement and interpretation of vitamin D status.J Clin Endocrinol Metab · 2006 · 91(8):3055-61doi:10.1210/jc.2006-0710 · PMID 16720650

    Mayo Clinic laboratory study: C3 epimers in 39 of 172 infants (22.7 percent), making up 8.7 to 61.1 percent of the total, and none in older children or adults with or without liver disease by their method. An extracted radioimmunoassay gave accurate results. Later reference materials found small amounts of the epimer in pooled human serum too.

32. 32
    Jainandunsing S, van Velsen EFS, Verschuure P, Willemsen D. Apparent vitamin D intoxication due to paraprotein interference.JCEM Case Rep · 2026 · 4(10):luag297doi:10.1210/jcemcr/luag297 · PMID 42780527

    A 76-year-old man with severe hypercalcemia on 400 IU a day: 25-hydroxyvitamin D above 250 nmol/L on routine testing, reproduced in an independent reference laboratory, but 92 nmol/L by LC-MS/MS; the cause was an immunoglobulin A kappa multiple myeloma, and the false result normalized as treatment lowered the paraprotein. A single case.

33. 33
    Schlingmann KP, Kaufmann M, Weber S, et al. Mutations in CYP24A1 and idiopathic infantile hypercalcemia.N Engl J Med · 2011 · 365(5):410-21doi:10.1056/NEJMoa1103864 · PMID 21675912

    Recessive loss-of-function CYP24A1 mutations in six children with idiopathic infantile hypercalcemia and in a second cohort who became hypercalcemic after single large preventive doses of vitamin D (bolus prophylaxis); the authors link this to Britain's 1950s epidemic during heavy milk fortification, an inference rather than a tested claim.

34. 34
    Holick MF, Schnoes HK, DeLuca HF, Suda T, Cousins RJ. Isolation and identification of 1,25-dihydroxycholecalciferol. A metabolite of vitamin D active in intestine.Biochemistry · 1971 · 10(14):2799-804doi:10.1021/bi00790a023 · PMID 4326883

    The isolation of the hormone, 1,25-dihydroxyvitamin D, as the metabolite active in intestine. PubMed holds no abstract; cited only for the discovery and date in its title.

35. 35
    Cheng JB, Levine MA, Bell NH, Mangelsdorf DJ, Russell DW. Genetic evidence that the human CYP2R1 enzyme is a key vitamin D 25-hydroxylase.Proc Natl Acad Sci U S A · 2004 · 101(20):7711-5doi:10.1073/pnas.0402490101 · PMID 15128933

    One patient with low 25-hydroxyvitamin D and classic deficiency symptoms, homozygous for a CYP2R1 mutation that eliminated 25-hydroxylase activity. A single case, but it identified the liver enzyme the test depends on.

36. 36
    Wise SA, Camara JE, Burdette CQ, et al. Interlaboratory comparison of 25-hydroxyvitamin D assays: Vitamin D Standardization Program (VDSP) Intercomparison Study 2 - Part 2 ligand binding assays - impact of 25-hydroxyvitamin D2 and 24R,25-dihydroxyvitamin D3 on assay performance.Anal Bioanal Chem · 2021 · 414(1):351-366doi:10.1007/s00216-021-03577-0 · PMID 34435207

    The same 50 sera in 17 laboratories running 32 ligand-binding assays: 50 percent met the 5 percent bias goal; of 13 distinct assays, 4 were consistently within it and 4 consistently outside wherever they were run; most underestimated 25-hydroxyvitamin D2 and five may cross-react with 24R,25-dihydroxyvitamin D3. Employees of Abbott, bioMérieux, Bio-Rad, IDS, Siemens and Awareness Technology, whose assays were tested, are co-authors.

37. 37
    Hughes MR, Malloy PJ, Kieback DG, et al. Point mutations in the human vitamin D receptor gene associated with hypocalcemic rickets.Science · 1988 · 242(4886):1702-5doi:10.1126/science.2849209 · PMID 2849209

    Two families with hereditary vitamin D-resistant rickets carried different single mutations in the receptor's DNA-binding zinc fingers; the receptor bound the hormone normally but bound DNA poorly.

38. 38
    Levin A, Bakris GL, Molitch M, et al. Prevalence of abnormal serum vitamin D, PTH, calcium, and phosphorus in patients with chronic kidney disease: results of the study to evaluate early kidney disease.Kidney Int · 2007 · 71(1):31-8doi:10.1038/sj.ki.5002009 · PMID 17091124

    In 1814 outpatients at 153 centers, low 1,25-dihydroxyvitamin D (under 22 pg/mL) appeared in 13 percent at eGFR above 80 and in more than 60 percent below 30, while 25-hydroxyvitamin D3 did not differ across eGFR deciles. Cross-sectional; the funding is listed only as non-government in PubMed. PubMed dates it November 2006.

39. 39
    Safadi FF, Thornton P, Magiera H, et al. Osteopathy and resistance to vitamin D toxicity in mice null for vitamin D binding protein.J Clin Invest · 1999 · 103(2):239-51doi:10.1172/JCI5244 · PMID 9916136

    Mice without binding protein had low total vitamin D metabolites but were otherwise normal on a replete diet, developed secondary hyperparathyroidism quickly when deprived, and resisted vitamin D toxicity. Animal evidence only.

40. 40
    Fraser DR, Mason RS. Commentary: Cellular functions of vitamin D-binding protein.Comp Biochem Physiol A Mol Integr Physiol · 2025 · 305:111848doi:10.1016/j.cbpa.2025.111848 · PMID 40147814

    A commentary proposing that skeletal muscle cells take up empty binding protein through the megalin and cubilin receptors, hold 25-hydroxyvitamin D on it, and release it back to the blood as the protein is broken down, and that this cycling accounts for the long half-life. An authors' synthesis, not a new measurement.

41. 41
    Alexandridou A, Stokes CS, Volmer DA. Measurement of Serum Free Vitamin D Concentrations: Importance, Challenges, and the Emerging Role of Mass Spectrometry.Clin Chem · 2025 · 71(2):254-265doi:10.1093/clinchem/hvae202 · PMID 39661472

    2025 review: calculated free 25-hydroxyvitamin D depends on binding-protein assays that differ by population and method; direct methods are complex or interference-prone; LC-MS/MS is promising; the clinical role of free levels remains undetermined. A narrative review.

42. 42
    Garabedian M, Holick MF, Deluca HF, Boyle IT. Control of 25-hydroxycholecalciferol metabolism by parathyroid glands.Proc Natl Acad Sci U S A · 1972 · 69(7):1673-6doi:10.1073/pnas.69.7.1673 · PMID 4340153

    Rat experiment: removing the thyroid and parathyroid glands cut production of 1,25-dihydroxycholecalciferol to negligible levels within 40 hours, and parathyroid extract restored it, independently of serum calcium or phosphorus. Animal evidence.

43. 43
    Need AG, O'Loughlin PD, Morris HA, Coates PS, Horowitz M, Nordin BE. Vitamin D metabolites and calcium absorption in severe vitamin D deficiency.J Bone Miner Res · 2008 · 23(11):1859-63doi:10.1359/jbmr.080607 · PMID 18597633

    Records of 3661 patients, 319 with 25-hydroxyvitamin D of 40 nmol/L or less: the hormone and calcium absorption fell, and parathyroid hormone rose, only in those at 10 nmol/L or below. Cross-sectional clinic data from one Adelaide center.

44. 44
    Tanaka Y, Deluca HF. The control of 25-hydroxyvitamin D metabolism by inorganic phosphorus.Arch Biochem Biophys · 1973 · 154(2):566-74doi:10.1016/0003-9861(73)90010-6 · PMID 4691503

    Rat study (MeSH) of phosphate control of the kidney step. PubMed holds no abstract; cited only for what its title states.

45. 45
    Shimada T, Hasegawa H, Yamazaki Y, et al. FGF-23 is a potent regulator of vitamin D metabolism and phosphate homeostasis.J Bone Miner Res · 2004 · 19(3):429-35doi:10.1359/JBMR.0301264 · PMID 15040831

    In rats and mice, one injection of FGF23 lowered 1,25-dihydroxyvitamin D within 3 hours, reaching its lowest at 9 hours, by reducing kidney 1-alpha-hydroxylase and increasing 24-hydroxylase messenger RNA, without any rise in parathyroid hormone. Animal evidence; the authors worked in the pharmaceutical laboratories of Kirin Brewery. PubMed dates it December 2003.

46. 46
    Tebben PJ, Singh RJ, Kumar R. Vitamin D-Mediated Hypercalcemia: Mechanisms, Diagnosis, and Treatment.Endocr Rev · 2016 · 37(5):521-547doi:10.1210/er.2016-1070 · PMID 27588937

    Mayo Clinic review: hypercalcemia in up to 4 percent of the population; vitamin D intoxication acting through very high 25-hydroxyvitamin D; 1-alpha-hydroxylase made outside the kidney in granulomas and lymphomas; CYP24A1 loss with high hormone, suppressed parathyroid hormone, kidney stones and calcification of the kidney.

47. 47
    Barbour GL, Coburn JW, Slatopolsky E, Norman AW, Horst RL. Hypercalcemia in an anephric patient with sarcoidosis: evidence for extrarenal generation of 1,25-dihydroxyvitamin D.N Engl J Med · 1981 · 305(8):440-3doi:10.1056/NEJM198108203050807 · PMID 6894783

    A single case report. PubMed holds no abstract; cited only for what its title states, the first evidence that the hormone can be made outside the kidney.

48. 48
    Lips P. Vitamin D deficiency and secondary hyperparathyroidism in the elderly: consequences for bone loss and fractures and therapeutic implications.Endocr Rev · 2001 · 22(4):477-501doi:10.1210/edrv.22.4.0437 · PMID 11493580

    Review: deficiency causes secondary hyperparathyroidism, high bone turnover, bone loss, mineralization defects and fractures; myopathy and falls less certain; strict diagnostic criteria hampered by assay differences.

49. 49
    Chapuy MC, Preziosi P, Maamer M, et al. Prevalence of vitamin D insufficiency in an adult normal population.Osteoporos Int · 1997 · 7(5):439-43doi:10.1007/s001980050030 · PMID 9425501

    1569 adults from 20 French cities in winter: levels fell with latitude (r = −0.79), 14 percent were at or below 30 nmol/L (12 ng/mL), and parathyroid hormone sat on a plateau of 36 pg/mL above 78 nmol/L (31 ng/mL) and rose below it. Cross-sectional, one radioimmunoassay method; later studies did not agree on a single plateau.

50. 50
    Kaufmann M, Gallagher JC, Peacock M, et al. Clinical utility of simultaneous quantitation of 25-hydroxyvitamin D and 24,25-dihydroxyvitamin D by LC-MS/MS involving derivatization with DMEQ-TAD.J Clin Endocrinol Metab · 2014 · 99(7):2567-74doi:10.1210/jc.2013-4388 · PMID 24670084

    More than 1000 samples: little 24,25-dihydroxyvitamin D3 was made below 25 nmol/L of 25-hydroxyvitamin D, and the 25-hydroxyvitamin D3 to 24,25 ratio stayed under 25 in vitamin D-sufficient subjects but reached 80 to 100 in idiopathic infantile hypercalcemia. One co-author worked for Waters Corporation, a mass-spectrometry manufacturer.

51. 51
    Trost SU, Burmeister L. CYP24A1 Deficiency Presenting With Hypercalcemia and Hypervitaminosis D in Pregnancy: A Case Report.Cureus · 2026 · 18(1):e100639doi:10.7759/cureus.100639 · PMID 41635392

    A 32-year-old pregnant woman on vitamin D and calcium supplements with parathyroid-independent hypercalcemia, very high calcitriol and a high ratio of 25-hydroxyvitamin D to 24,25-dihydroxyvitamin D; homozygous for a CYP24A1 variant; no hypercalcemia in earlier pregnancies. A single case in an open-access journal.

52. 52
    Kitanaka S, Takeyama K, Murayama A, et al. Inactivating mutations in the 25-hydroxyvitamin D3 1alpha-hydroxylase gene in patients with pseudovitamin D-deficiency rickets.N Engl J Med · 1998 · 338(10):653-61doi:10.1056/NEJM199803053381004 · PMID 9486994

    Cloning of the human 1-alpha-hydroxylase gene (CYP27B1) on chromosome 12q13.3 and four homozygous mutations, each abolishing enzyme activity, in four unrelated patients with pseudovitamin D-deficiency rickets.

53. 53
    Roizen JD, Li D, O'Lear L, et al. CYP3A4 mutation causes vitamin D-dependent rickets type 3.J Clin Invest · 2018 · 128(5):1913-1918doi:10.1172/JCI98680 · PMID 29461981

    Two unrelated patients with early rickets, low 25-hydroxyvitamin D and low hormone, and the same de novo CYP3A4 gain-of-function mutation that oxidized the hormone 10 times faster than normal CYP3A4 in vitro. Two patients; the in vitro activity explains the clinical picture.

54. 54
    Levine MA. Diagnosis and Management of Vitamin D Dependent Rickets.Front Pediatr · 2020 · 8:315doi:10.3389/fped.2020.00315 · PMID 32596195

    Review, read in full in PubMed Central: Prader and colleagues described pseudodeficiency rickets in 1961; in type 1A, 25-hydroxyvitamin D is normal or high and the hormone low or undetectable; in type 1B both are low; in type 2A, 25-hydroxyvitamin D is normal or high and the hormone high, 50 to 1,000 pg/mL against 30 to 100 normally in children; the kidney's activating enzyme is induced by parathyroid hormone and suppressed by FGF23. The author declares no conflicts of interest.

55. 55
    Belsey R, Deluca HF, Potts JT. Competitive binding assay for vitamin D and 25-OH vitamin D.J Clin Endocrinol Metab · 1971 · 33(3):554-7doi:10.1210/jcem-33-3-554 · PMID 4328344

    Published in September 1971, three months before Haddad and Chyu. PubMed holds no abstract; cited only for the method and date its title and record give.

56. 56
    Haddad JG, Chyu KJ. Competitive protein-binding radioassay for 25-hydroxycholecalciferol.J Clin Endocrinol Metab · 1971 · 33(6):992-5doi:10.1210/jcem-33-6-992 · PMID 4332615

    One of the first competitive protein-binding assays for 25-hydroxyvitamin D, published in December 1971, three months after Belsey, DeLuca and Potts. PubMed holds no abstract; cited only for the method and year its title records.

57. 57
    Maunsell Z, Wright DJ, Rainbow SJ. Routine isotope-dilution liquid chromatography-tandem mass spectrometry assay for simultaneous measurement of the 25-hydroxy metabolites of vitamins D2 and D3.Clin Chem · 2005 · 51(9):1683-90doi:10.1373/clinchem.2005.052936 · PMID 16020493

    An early routine LC-MS/MS method measuring the D2 and D3 forms separately: between-run coefficients of variation of 5.1 to 9.5 percent, more than 6000 clinical samples in 9 months, r = 0.91 against a commercial radioimmunoassay. The abstract does not address the C3 epimer.

58. 58
    Hollis BW, Napoli JL. Improved radioimmunoassay for vitamin D and its use in assessing vitamin D status.Clin Chem · 1985 · 31(11):1815-9PMID 4053351

    The radioimmunoassay: a rabbit antibody that recognized the 25-hydroxylated D2 and D3 metabolites equally but not 1,25-dihydroxyvitamin D or the parent vitamins, a detection limit of 3.0 pg per tube and 1 microliter of plasma per tube. A methods paper, not outcome evidence. The PubMed record lists no DOI.

59. 59
    Hollis BW, Kamerud JQ, Selvaag SR, Lorenz JD, Napoli JL. Determination of vitamin D status by radioimmunoassay with an 125I-labeled tracer.Clin Chem · 1993 · 39(3):529-33PMID 8448871

    Described by its authors as the first radioimmunoassay for a vitamin D metabolite with a radioiodinated tracer: a goat antibody that reacted equally with the D2 and D3 metabolites tested but not with the hormone, 50 microliters of serum, a detection limit of 2.8 micrograms per liter. Published March 1993; the DEQAS history dates the tracer to 1992. The PubMed record lists no DOI.

60. 60
    Malabanan A, Veronikis IE, Holick MF. Redefining vitamin D insufficiency.Lancet · 1998 · 351(9105):805-6doi:10.1016/s0140-6736(05)78933-9 · PMID 9519960

    A letter from Holick's group. PubMed holds no abstract; cited only for its title and date.

61. 61
    Bischoff-Ferrari HA, Giovannucci E, Willett WC, Dietrich T, Dawson-Hughes B. Estimation of optimal serum concentrations of 25-hydroxyvitamin D for multiple health outcomes.Am J Clin Nutr · 2006 · 84(1):18-28doi:10.1093/ajcn/84.1.18 · PMID 16825677

    Review of studies that looked for thresholds for bone density, lower-extremity function, dental health, falls, fractures and colorectal cancer: the most advantageous levels began at 75 nmol/L (30 ng/mL) and were best at 90 to 100 nmol/L (36 to 40 ng/mL). A dosage figure in the abstract was later corrected. A review of threshold studies, not a trial.

62. 62
    Heijboer AC, Blankenstein MA, Kema IP, Buijs MM. Accuracy of 6 routine 25-hydroxyvitamin D assays: influence of vitamin D binding protein concentration.Clin Chem · 2012 · 58(3):543-8doi:10.1373/clinchem.2011.176545 · PMID 22247500

    Five automated immunoassays and a radioimmunoassay against isotope-dilution LC-MS/MS in 51 healthy people, 52 pregnant women, 50 hemodialysis and 50 intensive care patients; most deviated significantly, and in 4 of the 5 automated assays the error moved inversely with binding-protein concentration. Assays have since been revised, so the specific numbers date.

63. 63
    Henderson J, Bouck Z, Holleman R, et al. Comparison of Payment Changes and Choosing Wisely Recommendations for Use of Low-Value Laboratory Tests in the United States and Canada.JAMA Intern Med · 2020 · 180(4):524-531doi:10.1001/jamainternmed.2019.7143 · PMID 32040158

    Interrupted time series in 54 million people: Ontario's December 2010 end of payment for low-value vitamin D screening cut it by 92.7 percent, while the February 2013 US and October 2014 Canadian Choosing Wisely recommendations were followed by reductions of 4.5 to 14.0 percent. Observational.

64. 64
    Powe CE, Evans MK, Wenger J, et al. Vitamin D-binding protein and vitamin D status of black Americans and white Americans.N Engl J Med · 2013 · 369(21):1991-2000doi:10.1056/NEJMoa1306357 · PMID 24256378

    In 2085 Baltimore adults, total 25-hydroxyvitamin D (15.6 against 25.8 ng/mL) and binding protein (168 against 337 micrograms/mL) were lower in Black than white participants, bone density higher, and estimated bioavailable levels, calculated only in participants homozygous for the two common binding-protein variants, similar. The binding-protein values came from a monoclonal immunoassay later shown to under-read the variant common in people of African ancestry, so the central finding has not held.

65. 65
    Henderson CM, Lutsey PL, Misialek JR, et al. Measurement by a Novel LC-MS/MS Methodology Reveals Similar Serum Concentrations of Vitamin D-Binding Protein in Blacks and Whites.Clin Chem · 2016 · 62(1):179-87doi:10.1373/clinchem.2015.244541 · PMID 26453697

    A validated mass-spectrometry assay for binding protein found isoform-specific error in the widely used monoclonal immunoassay and similar concentrations in white and Black participants, 262 against 266 micrograms/mL. PubMed dates it October 2015.

66. 66
    Nielson CM, Jones KS, Chun RF, et al. Free 25-Hydroxyvitamin D: Impact of Vitamin D Binding Protein Assays on Racial-Genotypic Associations.J Clin Endocrinol Metab · 2016 · 101(5):2226-34doi:10.1210/jc.2016-1104 · PMID 27007693

    In 1057 men from the US, UK and The Gambia, the monoclonal assay read binding protein about 50 percent lower in African-ancestry men while polyclonal and proteomic methods did not; directly measured free 25-hydroxyvitamin D tracked total (r = 0.84) and was lower in African American than white US men. Cross-sectional; men only.

67. 67
    Manson JE, Brannon PM, Rosen CJ, Taylor CL. Vitamin D Deficiency - Is There Really a Pandemic?N Engl J Med · 2016 · 375(19):1817-1820doi:10.1056/NEJMp1608005 · PMID 27959647

    A perspective whose first three authors also wrote the Institute of Medicine committee's 2011 summary; the fourth is from the Office of Dietary Supplements of the National Institutes of Health. PubMed holds no abstract; cited only for its title, authors and date.

68. 68
    Giustina A, Bilezikian JP, Adler RA, et al. Consensus Statement on Vitamin D Status Assessment and Supplementation: Whys, Whens, and Hows.Endocr Rev · 2024 · 45(5):625-654doi:10.1210/endrev/bnae009 · PMID 38676447

    Consensus of the sixth international conference 'Controversies in Vitamin D': optimal 25-hydroxyvitamin D levels remain debated; lack of assay standardization hinders the pooling of studies; the initial negative extraskeletal trials are likely due to the enrollment of vitamin D-replete people. Expert opinion; the possible benefits it lists come from post hoc analyses.

69. 69
    Holick MF. The Debatable Clinical Utility of the 2024 Vitamin D Guideline: Bridging the Gap Between Current Guidelines, Practical Clinical Recommendations, and Utilization of Emerging Evidence in Vitamin D Disease Prevention.Endocr Pract · 2025 · 31(3):399-402doi:10.1016/j.eprac.2025.02.002 · PMID 39938796

    A critique of the 2024 Endocrine Society guideline by the first author of the 2011 guideline. PubMed holds no abstract; cited only for its existence, author and title.

70. 70
    May HT, Le VT, Anderson JL, et al. Targeted vitamin D supplementation and major adverse cardiovascular events after myocardial infarction: the TARGET-D trial.Eur Heart J · 2026 · published online 7 September 2026doi:10.1093/eurheartj/ehag611 · PMID 42702500

    Pragmatic trial in 630 heart attack patients (median baseline 25 ng/mL) at Intermountain Medical Center, Utah, randomized to usual care or vitamin D3 adjusted by algorithm to a 25-hydroxyvitamin D target of 40 to 80 ng/mL: major adverse cardiovascular events 15.7 against 18.4 percent (hazard ratio 0.85, 95 percent confidence interval 0.58 to 1.24, P = .40); repeat heart attack, a secondary end point, 3.8 against 7.9 percent (hazard ratio 0.48, P = .03). Usual-care control rather than placebo; the secondary finding is hypothesis-generating.

71. 71
    Piketty ML, Prie D, Sedel F, et al. High-dose biotin therapy leading to false biochemical endocrine profiles: validation of a simple method to overcome biotin interference.Clin Chem Lab Med · 2017 · 55(6):817-825doi:10.1515/cclm-2016-1183 · PMID 28222020

    In 23 plasmas from people on high-dose biotin (31.7 to 1160 micrograms/L), most hormone results on streptavidin-biotin Roche assays, in a panel that included 25-hydroxyvitamin D, were abnormal and normalized once biotin was removed; the abstract does not give the direction of the 25-hydroxyvitamin D error. Two authors were employees of MedDay Pharmaceuticals, which was developing high-dose biotin for progressive multiple sclerosis, a benefit the paper stated as established and MedDay's later SPI2 phase 3 trial did not confirm.

72. 72
    Cree BAC, Cutter G, Wolinsky JS, et al. Safety and efficacy of MD1003 (high-dose biotin) in patients with progressive multiple sclerosis (SPI2): a randomised, double-blind, placebo-controlled, phase 3 trial.Lancet Neurol · 2020 · 19(12):988-997doi:10.1016/S1474-4422(20)30347-1 · PMID 33222767

    642 patients with progressive multiple sclerosis randomized to biotin 100 mg three times daily or placebo: confirmed improvement in 12 against 9 percent (odds ratio 1.35, 95 percent confidence interval 0.81 to 2.26), not significant; the drug produced inaccurate results on tests using biotinylated antibodies despite mitigation strategies. Funded by MedDay Pharmaceuticals; two authors were MedDay employees.

73. 73
    Racz S, Jozsa L, Bhattoa-Buzas AP, et al. Comprehensive Comparison of Three Automated 25-Hydroxyvitamin D Immunoassays with a VDSCP Traceable LC-MS/MS Method Across Clinically Relevant Decision Thresholds.Nutrients · 2026 · 18(18)doi:10.3390/nu18182986 · PMID 42796968

    400 outpatient sera on DiaSorin, Beckman Coulter and Roche immunoassays against an LC-MS/MS method traceable to the Vitamin D Standardization-Certification Program: Roche overestimated, Beckman flagged more false deficiency at 75 nmol/L, DiaSorin agreed best (kappa 0.818 to 0.895); external quality-scheme performance did not consistently reflect agreement in patient samples. One laboratory.

74. 74
    Corbetta C, Corsaro L, Caradonna E, et al. Comparison of LC-MS/MS and CLIA Methods for Vitamin D Measurement in an Unselected Outpatient Cohort: Statistical Evaluation and Impact on Clinical Classification.Diagnostics (Basel) · 2026 · 16(16)doi:10.3390/diagnostics16162514 · PMID 42650916

    1064 paired samples on a Siemens chemiluminescence immunoassay and an automated Thermo Scientific Cascadion LC-MS/MS, both VDSP participants: LC-MS/MS read proportionally higher (slope 1.33, average difference 35.5 percent); intraclass correlation 0.928; kappa 0.32 unweighted and 0.72 weighted on six categories and 0.56 on two; in discordant pairs the immunoassay classified deficiency 118-fold more often. The authors conclude that automated LC-MS/MS can substantially reclassify outpatients toward sufficiency. From a private Milan diagnostic company; one author is from Bracco, an imaging company.

75. 75
    Serres-Gómez M, González-Casaus ML, Serrano Olmedo MG, et al. Preliminary analytical evaluation of the fully automated Cobas i601 LC-MS/MS system for steroid hormones and vitamin D in a clinical laboratory.Clin Chim Acta · 2026 · 593:121312doi:10.1016/j.cca.2026.121312 · PMID 42674091

    First preliminary evaluation of the Roche Cobas i601 fully automated LC-MS/MS system: good precision, but 25-hydroxyvitamin D read 16 percent lower than the laboratory's immunoassay at an intermediate decision level. Preliminary, one hospital.

76. 76
    Eckel N, Stemmerich K, Lutz D, et al. First performance verification of automated cobasi601 LC-MS/MS in a high-throughput clinical laboratory using 25-hydroxyvitamin D, 5α-dihydrotestosterone, and 17α-hydroxyprogesterone.Clin Chem Lab Med · 2026 · 64(11):2461-2473doi:10.1515/cclm-2026-0421 · PMID 42330081

    Verification in German commercial laboratories (Bioscientia and Labor Augsburg): imprecision within 4.2 percent and bias against target values within 6.0 percent across the analytes, and strong agreement with the laboratory's conventional LC-MS/MS for 25-hydroxyvitamin D. No manufacturer authors are listed; one laboratory group.

77. 77
    Holick MF, Binkley NC, Bischoff-Ferrari HA, et al. Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline.J Clin Endocrinol Metab · 2011 · 96(7):1911-30doi:10.1210/jc.2011-0385 · PMID 21646368

    The 2011 guideline: vitamin D deficiency judged very common, 25-hydroxyvitamin D by a reliable assay suggested as the test for people at risk, no screening of those not at risk; the task force declared no corporate funding. Its 30 ng/mL sufficiency definition is not in the abstract and is cited here from the 2024 communication that withdrew it.

78. 78
    Heaney RP, Dowell MS, Hale CA, Bendich A. Calcium absorption varies within the reference range for serum 25-hydroxyvitamin D.J Am Coll Nutr · 2003 · 22(2):142-6doi:10.1080/07315724.2003.10719287 · PMID 12672710

    Two spring experiments a year apart using marketed calcium supplements: absorption 65 percent higher at a mean 86.5 than at 50.2 nmol/L. Small pharmacokinetic study comparing different experiments rather than randomized groups; a larger study (Aloia 2010) did not confirm the relationship.

79. 79
    Priemel M, von Domarus C, Klatte TO, et al. Bone mineralization defects and vitamin D deficiency: histomorphometric analysis of iliac crest bone biopsies and circulating 25-hydroxyvitamin D in 675 patients.J Bone Miner Res · 2010 · 25(2):305-12doi:10.1359/jbmr.090728 · PMID 19594303

    Autopsy bone biopsies from 675 northern Germans: no pathological osteoid accumulation in anyone above 75 nmol/L, though no single level inevitably caused defects; the authors concluded 30 ng/mL should be the minimum. Post-mortem blood levels and a single population.

80. 80
    Aloia JF, Chen DG, Yeh JK, Chen H. Serum vitamin D metabolites and intestinal calcium absorption efficiency in women.Am J Clin Nutr · 2010 · 92(4):835-40doi:10.3945/ajcn.2010.29553 · PMID 20660223

    In 492 Black and white women aged 20 to 80, there was no relation between 25-hydroxyvitamin D and calcium absorption efficiency; menopause, calcium intake, estradiol and the hormone were the determinants. Cross-sectional; 25-hydroxyvitamin D by radioreceptor assay.

81. 81
    Rosen CJ, Abrams SA, Aloia JF, et al. IOM committee members respond to Endocrine Society vitamin D guideline.J Clin Endocrinol Metab · 2012 · 97(4):1146-52doi:10.1210/jc.2011-2218 · PMID 22442278

    Members of the IOM committee argued that the Endocrine Society's target levels, its definition of deficiency and its idea of who is at risk were not well supported. A commentary, not new data.

82. 82
    Smart D, Schreier J, Singh IR. Inappropriate Laboratory Testing: Significant Waste Quantified by a Large-Scale Year-Long Study of Medicare and Commercial Payer Reimbursement.Arch Pathol Lab Med · 2025 · 149(3):253-261doi:10.5858/arpa.2023-0486-OA · PMID 38839058

    2019 claims for 64 million Medicare and 168 million commercially insured people: 7 to 51 percent of four tests (25-hydroxyvitamin D, prostate-specific antigen, lipid panel and hemoglobin A1c) exceeded recommended frequencies, at a conservative excess cost above $350 million. The abstract does not give vitamin D's own share; two of the three authors work for Diaceutics, a laboratory-data company.

83. 83
    Cho HJ, Mestari N, Israilov S, et al. Reducing 25-Hydroxyvitamin D Testing in a Large, Urban Safety Net System.J Gen Intern Med · 2023 · 38(10):2326-2332doi:10.1007/s11606-023-08201-8 · PMID 37131103

    A mandatory indication prompt and an alert against repeat testing within 3 months in 11 New York hospitals and 70 clinics cut inpatient and outpatient orders by 44 and 46 percent. Quality-improvement study without a control group.

84. 84
    Shah VP, Nayfeh T, Alsawaf Y, et al. A Systematic Review Supporting the Endocrine Society Clinical Practice Guidelines on Vitamin D.J Clin Endocrinol Metab · 2024 · 109(8):1961-1974doi:10.1210/clinem/dgae312 · PMID 38828942

    151 studies from 37,007 citations; no trials of the benefits or harms of screening with 25-hydroxyvitamin D were found.

85. 85
    Kim SH, Chae JH, Chae CH. Season of sampling and the interpretation of serum homocysteine results: a within-person analysis of 268,877 health-screening examinations.Clin Chem Lab Med · 2026 · published online 23 September 2026doi:10.1515/cclm-2026-1304 · PMID 42770710

    268,877 examinations of 69,131 adults at one Korean center, 2014 to 2025: within-person homocysteine peaked in early November and 25-hydroxyvitamin D in August; adjusting for 25-hydroxyvitamin D did not reduce the homocysteine rhythm. Observational; one center.

86. 86
    Neale RE, Khan SR, Lucas RM, Waterhouse M, Whiteman DC, Olsen CM. The effect of sunscreen on vitamin D: a review.Br J Dermatol · 2019 · 181(5):907-915doi:10.1111/bjd.17980 · PMID 30945275

    Systematic review of 4 experimental studies, 3 field trials and 69 observational studies: sunscreen blocked vitamin D production under artificial ultraviolet light, but randomized field trials (SPF about 16) found no effect and observational studies mostly none; no trials of today's high-SPF products.

87. 87
    Wortsman J, Matsuoka LY, Chen TC, Lu Z, Holick MF. Decreased bioavailability of vitamin D in obesity.Am J Clin Nutr · 2000 · 72(3):690-3doi:10.1093/ajcn/72.3.690 · PMID 10966885

    After whole-body irradiation the rise in vitamin D3 was 57 percent lower in obese than lean white adults despite the same skin precursor, and body mass index correlated inversely with peak levels after 50,000 IU of D2. The authors proposed deposition in fat; a later analysis argued for dilution instead.

88. 88
    Vimaleswaran KS, Berry DJ, Lu C, et al. Causal relationship between obesity and vitamin D status: bi-directional Mendelian randomization analysis of multiple cohorts.PLoS Med · 2013 · 10(2):e1001383doi:10.1371/journal.pmed.1001383 · PMID 23393431

    21 adult cohorts, up to 42,024 people: each 1 kg/m² higher body mass index went with 1.15 percent lower 25-hydroxyvitamin D; a 10 percent higher genetically predicted body mass index lowered it by 4.2 percent, while vitamin D-lowering variants did not change body mass index. Cohorts from Europe and North America; MeSH lists white participants.

89. 89
    Aggarwal A, Yadav AK, Ramachandran R, et al. Bioavailable vitamin D levels are reduced and correlate with bone mineral density and markers of mineral metabolism in adults with nephrotic syndrome.Nephrology (Carlton) · 2016 · 21(6):483-9doi:10.1111/nep.12638 · PMID 26428957

    106 adults with idiopathic nephrotic syndrome and 40 healthy controls in Chandigarh: total and calculated bioavailable 25-hydroxyvitamin D both lower in patients, a fall the authors attribute to urinary loss of binding protein; hip bone density correlated with the bioavailable level (r = 0.358) but not the total (r = 0.174). Cross-sectional; bioavailable levels calculated, not measured.

90. 90
    Revez JA, Lin T, Qiao Z, et al. Genome-wide association study identifies 143 loci associated with 25 hydroxyvitamin D concentration.Nat Commun · 2020 · 11(1):1647doi:10.1038/s41467-020-15421-7 · PMID 32242144

    417,580 Europeans: 143 loci, and Mendelian randomization finding no robust causal effect of 25-hydroxyvitamin D on the phenotypes tested but many phenotypes affecting 25-hydroxyvitamin D. European ancestry only.

91. 91
    Dai Q, Zhu X, Manson JE, et al. Magnesium status and supplementation influence vitamin D status and metabolism: results from a randomized trial.Am J Clin Nutr · 2018 · 108(6):1249-1258doi:10.1093/ajcn/nqy274 · PMID 30541089

    Ancillary study of 180 people in a randomized trial: magnesium raised 25-hydroxyvitamin D3 when baseline levels were near 30 ng/mL and lowered it when they were 30 to 50. Interaction analysis in one trial; two co-authors are from the Center for Magnesium Education and Research.

92. 92
    Jha V, Rouniyar R, Rao SD, Dua A, Bhadada SK. An Outbreak of Vitamin D Toxicity in a Family: Report of 3 Cases From Unregulated Fortification.JCEM Case Rep · 2025 · 3(12):luaf230doi:10.1210/jcemcr/luaf230 · PMID 41179273

    Three adults in one household with severe hypercalcemia, kidney injury and very high 25-hydroxyvitamin D after using unpackaged cooking oil that LC-MS/MS found contaminated with vitamin D at 1286 micrograms/mL; one needed dialysis, all recovered. Three cases.

93. 93
    Sheikh JA, Baigh Z, Dawar BMO. Etiological Spectrum of Hypercalcemia in a North Indian Tertiary Care Center: Emerging Role of Vitamin D Toxicity.Cureus · 2026 · 18(5):e108257doi:10.7759/cureus.108257 · PMID 42245902

    Forty adults hospitalized with hypercalcemia in Kashmir in the first half of 2025: vitamin D toxicity accounted for 8 (20 percent), with 25-hydroxyvitamin D 123.5 ± 25 ng/mL; prior regional reports implicate 600,000 IU intramuscular injections. Small retrospective series in an open-access journal with light peer review.

94. 94
    Gordon L, Zipursky R, Cheung CC, Sasidharan Pillai S. Vitamin D Intoxication in an Adolescent due to Over-Supplementation: A Case Report and Review of Literature.AACE Endocrinol Diabetes · 2025 · 13(1):93-97doi:10.1016/j.aed.2025.10.008 · PMID 41641290

    A 17-year-old seen at Children's Hospital Los Angeles who drank homemade milkshakes supplemented with vitamin D3 daily for about a year: calcium 15.5 mg/dL, 25-hydroxyvitamin D 594 ng/mL, suppressed parathyroid hormone; recovered with fluids and a low-calcium diet. A single case.

95. 95
    Sudarsan A, Pasupathy U, Bharathi U. Prescription-dispensing mismatch leading to vitamin D intoxication in an infant.BMJ Case Rep · 2025 · 18(11)doi:10.1136/bcr-2025-267253 · PMID 41276311

    A term infant in Chennai given 6000 IU a day for 45 days instead of the recommended 400, a cumulative 320,000 IU: 25-hydroxyvitamin D above 120 ng/mL, high blood and urine calcium, no kidney calcification; improved within 48 hours of stopping. A single case.

96. 96
    Binkley N, Sempos CT, Borchardt G, et al. Association of standardized serum 25-hydroxyvitamin D with falls in post-menopausal women.Osteoporos Int · 2025 · 36(2):255-264doi:10.1007/s00198-024-07310-2 · PMID 39585370

    Secondary analysis of a 4-year trial in 2303 women with retrospectively standardized values: no excess falls up to 60 ng/mL; at 60 ng/mL or more, odds of two or more falls about doubled (odds ratio 1.99), but that group was about 2 percent of observations. PubMed dates it November 2024.

97. 97
    Hamilton FW, Hughes DA, Spiller W, Tilling K, Davey Smith G. Non-linear Mendelian randomization: detection of biases using negative controls with a focus on BMI, Vitamin D and LDL cholesterol.Eur J Epidemiol · 2024 · 39(5):451-465doi:10.1007/s10654-024-01113-9 · PMID 38789826

    Negative-control outcomes, which vitamin D should not affect, produced clearly biased estimates with both nonlinear methods in some situations, and the doubly ranked method failed to reproduce the trial-established shape for LDL (low-density lipoprotein) cholesterol. Recommends caution and sanity checks before further use.

98. 98
    Mokry LE, Ross S, Ahmad OS, et al. Vitamin D and Risk of Multiple Sclerosis: A Mendelian Randomization Study.PLoS Med · 2015 · 12(8):e1001866doi:10.1371/journal.pmed.1001866 · PMID 26305103

    Four variants for 25-hydroxyvitamin D tested in 14,498 people with multiple sclerosis and 24,091 controls: each genetically determined one-standard-deviation decrease in log 25-hydroxyvitamin D doubled the odds of multiple sclerosis (odds ratio 2.0, 95 percent confidence interval 1.7 to 2.5), 1.7 after excluding possibly pleiotropic variants. Residual pleiotropy cannot be excluded; European ancestry.

99. 99
    Arcidiacono GP, Torres MO, Ceolin C, et al. Prevalence, predictors, and clinical impact of vitamin D deficiency on 12-month outcomes in hip fracture patients.J Endocrinol Invest · 2026 · 49(10):2913-2924doi:10.1007/s40618-026-02952-x · PMID 42364060

    934 hip-fracture patients in Padua, 2023 to 2025: deficiency (below 50 nmol/L) in 53.6 percent; its association with death within 12 months (hazard ratio 1.65) fell to 1.31 (95 percent confidence interval 0.88 to 1.96) after adjustment for age, sex and comorbidity. Prospective observational cohort.

100. 100
     Khalil I, Hrebat M, Laouadi N, et al. Association between serum 25-hydroxyvitamin D and osteoarthritis prevalence in US adults: a cross-sectional study using NHANES 2007-2023.Ann Med Surg (Lond) · 2026 · 88(10):6349-6358doi:10.1097/MS9.0000000000005552 · PMID 42866781

     35,887 US adults in the National Health and Nutrition Examination Survey, measured by standardized LC-MS/MS: each 10 nmol/L higher 25-hydroxyvitamin D went with 5.3 percent higher odds of osteoarthritis, which the authors attributed to reverse causation or to patients taking supplements after diagnosis rather than to harm. Cross-sectional.

101. 101
     Thouvenot E, Laplaud D, Lebrun-Frenay C, et al. High-Dose Vitamin D in Clinically Isolated Syndrome Typical of Multiple Sclerosis: The D-Lay MS Randomized Clinical Trial.JAMA · 2025 · 333(16):1413-1422doi:10.1001/jama.2025.1604 · PMID 40063041

     316 untreated patients with a first episode typical of multiple sclerosis and 25-hydroxyvitamin D under 100 nmol/L; 100,000 IU every 2 weeks for 24 months cut disease activity from 74.1 to 60.3 percent (hazard ratio 0.66, 95 percent confidence interval 0.50 to 0.87), driven by lesions on magnetic resonance imaging, while relapses did not differ significantly (hazard ratio 0.69, 95 percent confidence interval 0.42 to 1.16); none of the 10 clinical secondary outcomes differed significantly.

102. 102
     Emerging Risk Factors Collaboration/EPIC-CVD/Vitamin D Studies Collaboration. Estimating dose-response relationships for vitamin D with coronary heart disease, stroke, and all-cause mortality: observational and Mendelian randomisation analyses.Lancet Diabetes Endocrinol · 2021 · 9(12):837-846doi:10.1016/S2213-8587(21)00263-1 · PMID 34717822

     RETRACTED. The original version reported, among people under 25 nmol/L, an odds ratio of 0.69 for death per 10 nmol/L higher genetically predicted level, from a stratified (nonlinear) Mendelian randomization. After a July 2023 expression of concern it was retracted and replaced in December 2023 by a version whose stratified analyses were null at every level.

103. 103
     Burgess S. Violation of the Constant Genetic Effect Assumption Can Result in Biased Estimates for Non-Linear Mendelian Randomization.Hum Hered · 2023 · 88(1):79-90doi:10.1159/000531659 · PMID 37651993

     Reanalysis showing that the genetic score's effect on 25-hydroxyvitamin D differed five-fold between the lowest and highest deciles, so the residual method's protective effect in people with low levels disappeared with the doubly ranked method. A methods paper; both methods can be biased when the assumption fails.

104. 104
     Sutherland JP, Zhou A, Hyppönen E. Vitamin D Deficiency Increases Mortality Risk in the UK Biobank : A Nonlinear Mendelian Randomization Study.Ann Intern Med · 2022 · 175(11):1552-1559doi:10.7326/M21-3324 · PMID 36279545

     307,601 white UK Biobank participants, 18,700 deaths: an L-shaped genetic association with mortality, steep up to 50 nmol/L, and an odds ratio of 1.25 at 25 against 50 nmol/L. Its nonlinear Mendelian randomization belongs to the family of methods later shown to give biased answers for vitamin D when the genetic effect varies with level; a correction was published in 2025.

105. 105
     The Lancet Diabetes & Endocrinology. Expression of Concern-Estimating dose-response relationships for vitamin D with coronary heart disease, stroke, and all-cause mortality: observational and Mendelian randomisation analyses.Lancet Diabetes Endocrinol · 2023 · 11(9):634doi:10.1016/S2213-8587(23)00198-5 · PMID 37454668

     The journal's July 2023 expression of concern on the 2021 paper. No abstract in PubMed; cited for its existence and date.

106. 106
     The Editors of The Lancet Diabetes & Endocrinology. Retraction and republication-Estimating dose-response relationships for vitamin D with coronary heart disease, stroke, and all-cause mortality: observational and Mendelian randomisation analyses.Lancet Diabetes Endocrinol · 2023 · 12(1):8doi:10.1016/S2213-8587(23)00364-9 · PMID 38048795

     The December 2023 retraction-and-republication notice. No abstract in PubMed; cited for the action and its date. PubMed dates it 2023; the issue is January 2024.

107. 107
     Emerging Risk Factors Collaboration/EPIC-CVD/Vitamin D Studies Collaboration. Estimating dose-response relationships for vitamin D with coronary heart disease, stroke, and all-cause mortality: observational and Mendelian randomisation analyses.Lancet Diabetes Endocrinol · 2023 · 12(1):e2-e11doi:10.1016/S2213-8587(23)00287-5 · PMID 38048800

     The republished version: 386,406 people of European ancestry; no association of genetically predicted 25-hydroxyvitamin D with coronary disease (odds ratio 0.98), stroke (1.01) or death (0.99) per 10 nmol/L, and null stratified results at all levels; substantial benefit from low-dose supplementation called unlikely even in people with low status.

108. 108
     Stender S, Gellert-Kristensen H, Smith GD. Reclaiming mendelian randomization from the deluge of papers and misleading findings.Lipids Health Dis · 2024 · 23(1):286doi:10.1186/s12944-024-02284-w · PMID 39244551

     A letter: public genome-wide summary statistics have produced a flood of low-value two-sample Mendelian randomization papers, which editors are advised to reject when they report nothing else; new nonlinear methods applied without proper testing gave misleading results, exemplified by vitamin D findings being retracted or corrected. Opinion.

109. 109
     Annals of Internal Medicine. Correction: Vitamin D Deficiency Increases Mortality Risk in the UK Biobank.Ann Intern Med · 2025 · 178(8):1215-1216doi:10.7326/ANNALS-25-02471 · PMID 40523290

     Published erratum to the 2022 paper. PubMed holds no abstract, so what was corrected is not summarized here; cited only for the fact and date of the correction.

110. 110
     Burgess S, Woolf BAR, Mason AM. Context-Stratified Mendelian Randomization: Exploiting Regional Exposure Variation to Explore Causal Effect Heterogeneity and Nonlinearity.Stat Med · 2026 · 45(20-22):e70727doi:10.1002/sim.70727 · PMID 42680961

     A new stratification by recruitment center rather than by modeled exposure; applied to 20 UK Biobank centers it found no causal effect of vitamin D on coronary artery disease and no heterogeneity. Its power depends on how much the exposure differs between centers.

111. 111
     Akbarzadeh M, Fateh ST, Moeinafshar A, et al. Genetically Predicted 25-Hydroxyvitamin D Levels on Hypothyroidism: A Two-Sample Mendelian Randomization.Biomed Res Int · 2026 · 2026(1):e8086558doi:10.1155/bmri/8086558 · PMID 42253156

     Two-sample Mendelian randomization on European-ancestry summary statistics: higher genetically predicted 25-hydroxyvitamin D associated with lower risk of hypothyroidism (beta −0.197) but not of Hashimoto thyroiditis; the authors read it as support for supplementation. Summary-data analysis only.

112. 112
     Wang D, Zhang L, Cui Y, et al. Causal relationships between 25-hydroxyvitamin D levels and laryngeal cancer risk: A two-sample Mendelian randomization study.Food Nutr Res · 2025 · 69doi:10.29219/fnr.v69.11489 · PMID 41323137

     Two-sample Mendelian randomization on published summary data: odds ratio 0.9993 (95 percent confidence interval 0.9987 to 0.9999) for laryngeal cancer, which the authors called a slight protective effect; not significant after adjustment for alcohol alone. Summary-data analysis only.

113. 113
     Chen L, Zhao J, Zhang Z, Gao C. Circulating serum 25-hydroxyvitamin D levels and bone mineral density and osteoporotic vertebral fractures: A Mendelian randomization study.J Transl Int Med · 2026 · 14(4):610-619doi:10.1515/jtim-2026-0058 · PMID 42657082

     15 variants genotyped in 2061 community-dwelling adults and 200 outpatients with vertebral fractures, studied at a Shanghai hospital: no association of genetically predicted 25-hydroxyvitamin D with bone density, bone turnover markers or vertebral fractures overall; in those aged 55 to 92, low levels were linked to a higher bone-formation marker (odds ratio 1.38). Small for a Mendelian randomization study.

114. 114
     LeBoff MS, Chou SH, Ratliff KA, et al. Supplemental Vitamin D and Incident Fractures in Midlife and Older Adults.N Engl J Med · 2022 · 387(4):299-309doi:10.1056/NEJMoa2202106 · PMID 35939577

     VITAL ancillary study: 25,871 US adults not selected for deficiency; 2000 IU of vitamin D a day did not reduce total fractures (hazard ratio 0.98, 95 percent confidence interval 0.89 to 1.08), nonvertebral or hip fractures over a median 5.3 years, and the effect did not vary with baseline 25-hydroxyvitamin D. Funded by the National Institute of Arthritis and Musculoskeletal and Skin Diseases.

115. 115
     Jolliffe DA, Holt H, Greenig M, et al. Effect of a test-and-treat approach to vitamin D supplementation on risk of all cause acute respiratory tract infection and covid-19: phase 3 randomised controlled trial (CORONAVIT).BMJ · 2022 · 378:e071230doi:10.1136/bmj-2022-071230 · PMID 36215226

     6200 UK adults; of 3100 offered a finger-prick test, 2958 accepted and 2674 (86.3 percent) were below 75 nmol/L and received 800 or 3200 IU a day; respiratory infections were not reduced (odds ratios 1.26 and 1.09). Open label, 6 months.

116. 116
     Pittas AG, Kawahara T, Jorde R, et al. Vitamin D and Risk for Type 2 Diabetes in People With Prediabetes : A Systematic Review and Meta-analysis of Individual Participant Data From 3 Randomized Clinical Trials.Ann Intern Med · 2023 · 176(3):355-363doi:10.7326/M22-3018 · PMID 36745886

     Three trials: vitamin D cut new diabetes by 15 percent (hazard ratio 0.85); among those given vitamin D, maintaining at least 125 nmol/L against 50 to 74 was associated with a hazard ratio of 0.24, a post-randomization comparison by achieved level that is not protected by randomization. One author is a Takeda Pharmaceuticals employee; one trial tested eldecalcitol.

117. 117
     Ladang A, Gendebien AS, Kovacs S, et al. Investigation of the Vitamin D Metabolite Ratio (VMR) as a Marker of Functional Vitamin D Deficiency: Findings from the SarcoPhAge Cohort.Nutrients · 2024 · 16(19)doi:10.3390/nu16193224 · PMID 39408192

     204 Belgian older adults: the metabolite ratio correlated more strongly with parathyroid hormone (rho −0.292) than either metabolite alone, but in Cox models both low ratio and low 25-hydroxyvitamin D predicted death. Small cohort.

118. 118
     Miyamoto H, Kisugi R, Tajima A, et al. Sex-Based Differences in the Vitamin D Metabolite Ratio and Its Associated Factors Among Healthy Japanese Adults: A Cross-Sectional Study.Biomedicines · 2026 · 14(9)doi:10.3390/biomedicines14091993 · PMID 42792735

     Tentative reference intervals on a fully automated LC-MS/MS: 24,25-dihydroxyvitamin D 0.3 to 3.0 ng/mL in men and 0.2 to 2.3 in women; metabolite ratio 3.0 to 11.8 and 2.5 to 10.4 percent. Several co-authors are employees of Shimadzu, the instrument's maker.

119. 119
     Julia T, Krzysztof G, Zofia GP, et al. C3-Epimer Inclusion Influences Vitamin D Deficiency Classification Near Diagnostic Thresholds in Early Childhood.Nutrients · 2026 · 18(8)doi:10.3390/nu18081188 · PMID 42075001

     128 children aged 0 to 36 months in Bydgoszcz, Poland, measured by LC-MS/MS in dried blood spots: subtracting the C3 epimer changed deficiency classification (below 20 ng/mL) in 22 (17.2 percent), and in 54.5 percent of those at 18 to 22 ng/mL against 4.2 percent outside that range. Cross-sectional; one author is from Masdiag, a Warsaw company. PubMed lists the authors with given and family names inverted; the first author is Julia Tworowska.

120. 120
     Wise SA, Tai SS, Nelson MA, et al. Interlaboratory Comparison for the Determination of 24,25-Dihydroxyvitamin D3 in Human Serum Using Liquid Chromatography with Tandem Mass Spectrometry.J AOAC Int · 2017 · 100(5):1308-1317doi:10.5740/jaoacint.17-0183 · PMID 28741469

     Six VDSP laboratories measuring 24,25-dihydroxyvitamin D3 against NIST's candidate reference method: one at −1.6 percent mean bias, three at 10 to 15 percent, one at 36 percent; one laboratory found a calculation error. Shows that the newer metabolite is not yet harmonized.