Vitamin D Injection and IV Therapy: Mechanism, Evidence and Why Daily Beats the Megadose

In short: no cholecalciferol product carries a marketing authorisation for intravenous administration anywhere, so an intravenous vitamin D drip is not a licensed medicine by that route. The only injectable vitamin D licensed in the United Kingdom is intramuscular ergocalciferol 300,000 IU, indicated for malabsorption caused by gastrointestinal, liver or biliary disease, and its label requires serum and urinary calcium to be monitored — initially weekly. The intravenous vitamin D products that do exist are activated analogues for renal medicine, not nutrition. On outcomes, the literature separates by schedule rather than by dose: an annual 500,000 IU oral dose increased falls by about 15% and fractures by about 26%, monthly 60,000 IU increased falls, and Germany's federal risk assessment institute concluded in September 2025 that infrequent large doses carry health risks daily doses do not. The benefit signals are all daily — a 22% reduction in incident autoimmune disease at 2,000 IU per day, reduced telomere attrition in the same trial's sub-study, and a 12% cancer-mortality reduction that appeared only in the daily-dosing subgroup. For fractures the question now looks closed: pooling 69 trials and 153,902 adults in May 2026, vitamin D alone gave a risk ratio of 1.00 at high certainty.

This Journal has repeatedly asked whether a route earns its claim. Vitamin D asks a narrower question, and the answer is unusually clean: the route determines the schedule, and the schedule is what the trials keep separating.

What is actually in the syringe

Vitamin D3, cholecalciferol, is a fat-soluble secosteroid prohormone — a steroid nucleus with one ring opened. Ergocalciferol, vitamin D2, is the plant- and fungus-derived analogue. Neither dissolves in water, which is the single physical fact that governs everything that follows about route.

An injectable vitamin D preparation is therefore an oil. The United Kingdom's licensed presentation is ergocalciferol 300,000 IU in 1 mL of oil, labelled for intramuscular use only, with the further instruction that plastic syringes should not be used. Compounded cholecalciferol injections supplied by compounding pharmacies follow the same logic, dissolved in a vegetable oil and specified for deep intramuscular administration.

That is the whole formulation problem in one line. An oil depot can be placed in muscle. It cannot be put into a vein, and it cannot be put into a drip bag.

Two hydroxylations and a cofactor nobody measures

Cholecalciferol is inactive as administered. The liver hydroxylates it at carbon 25 to form 25-hydroxyvitamin D — the circulating storage form, and the molecule every laboratory report means when it says "vitamin D level". The kidney then hydroxylates that at carbon 1 to calcitriol, the hormone that binds the vitamin D receptor. A third enzyme, 24-hydroxylase, runs the deactivation arm.

All three of those enzymes are magnesium-dependent, as is vitamin D binding protein. A 2018 review in the Journal of the American Osteopathic Association put it plainly: every enzyme that metabolises vitamin D appears to require magnesium as a cofactor in the liver and the kidney. That is a real clinical consideration and a rarely measured one — the relationship between magnesium status and vitamin D metabolism runs in the direction of magnesium enabling vitamin D, not the reverse.

It also explains why the number on the report is a poor guide to the physiology. 25-hydroxyvitamin D is what is measured; calcitriol is what acts; and the conversion between them is regulated, cofactor-dependent and not proportional to the dose that went in.

There is no intravenous vitamin D

Three distinct products get conflated under the phrase "IV vitamin D", and separating them settles most of the question.

The first is intramuscular ergocalciferol. Its UK Summary of Product Characteristics restricts the indication to patients with gastrointestinal, liver or biliary disease associated with malabsorption of vitamin D, producing hypophosphataemia, rickets or osteomalacia. Dosing is a single 300,000 IU ampoule every three to six months. Hypercalcaemia, evidence of vitamin D toxicity and decreased renal function are contraindications, and the label instructs that serum and urinary calcium, phosphate and blood urea nitrogen be monitored at regular intervals, initially weekly.

Read that as a regulator's description rather than a manufacturer's: this is an injection for people who cannot absorb, requiring weekly bloods at the start. It is not a convenience format.

The second group is genuinely intravenous — and it is not cholecalciferol. Calcitriol injection at 1 microgram per millilitre is licensed for the management of hypocalcaemia in patients undergoing chronic renal dialysis. Paricalcitol injection at 5 micrograms per millilitre is licensed for the prevention and treatment of secondary hyperparathyroidism in chronic kidney disease. Both are activated analogues that bypass the hydroxylation steps entirely, which is precisely why they are reserved for kidneys that cannot perform them, and why they carry a calcium and parathyroid-hormone monitoring burden.

The third is the wellness drip. Whatever else is in that bag, it is not a licensed intravenous vitamin D, because no such product exists. The same distinction the series drew for B12 and folate applies here in a harder form: there, the intravenous route existed and was the wrong one; here, it does not exist at all.

What the depot does to the curve

The intuition behind the injection is that it delivers more. It does not.

A controlled pharmacokinetic study published in PLOS ONE in 2017 ran two winter cohorts in Berlin. One took escalating oral cholecalciferol — 2,000 IU daily for four weeks, then 4,000, then 8,000. The other received a single intramuscular injection of 100,000 IU. At day 28 the two routes were statistically indistinguishable. By day 84 the oral arm had reached 159.7 nmol/L and the injected arm 75.2 nmol/L. The injection produced a flatter, slower, more sustained curve, exactly as a depot should — and a lower one.

Scaling up does not change the shape. An open-label Australian study gave 50 vitamin D-deficient adults a single annual intramuscular injection of 600,000 IU. Mean 25-hydroxyvitamin D rose from 32 nmol/L to 114 nmol/L at four months and settled at 73 nmol/L by twelve. Parathyroid hormone fell by 30%. It also produced mild hypercalcaemia in 4% of participants and raised urinary calcium excretion in 20% at twelve months, in a study with no control arm.

The asymmetry that matters is not the peak. It is reversibility. An oral regimen can be stopped tomorrow. An intramuscular depot of 600,000 IU continues to release for months, and no decision taken afterwards can retrieve it.

Where the harm signals are

Two randomised trials define the boundary, and both tested the schedule an injection imposes.

In 2010 JAMA published a placebo-controlled trial of a single annual oral dose of 500,000 IU cholecalciferol in community-dwelling women aged 70 and over. The high-dose group had roughly 15% more falls and 26% more fractures than placebo. The intervention was intended to prevent exactly what it increased.

In 2016 JAMA Internal Medicine randomised 200 adults aged 70 and over with a prior fall to monthly vitamin D at 24,000 IU, 60,000 IU, or 24,000 IU with calcifediol. The higher monthly doses produced no benefit on functional decline and more falls than the lower one. Sixty per cent of participants fell during the twelve months.

Germany's Federal Institute for Risk Assessment examined this pattern directly in Opinion 031/2025, published on 3 September 2025, and concluded that high single doses taken at intervals of days or weeks are associated with health risks that daily dosing is not. Its mechanistic reading is worth noting: bolus administration produces large swings in circulating parent cholecalciferol and a prolonged elevation of the inactive 24,25-dihydroxy metabolite. The institute recommended a maximum of 20 micrograms per day in supplements and flagged that pushing 25-hydroxyvitamin D towards 100 nmol/L in people already above 70 nmol/L is where the association with adverse outcomes begins.

Where the benefit signals are

The mirror image is equally consistent, and it is the part sceptics tend to skip.

The VITAL trial randomised 25,871 adults to 2,000 IU of vitamin D3 daily or placebo. Its primary endpoints were null — cancer incidence with a hazard ratio of 0.96, major cardiovascular events 0.97, all-cause mortality 0.99. But a pre-specified analysis published in the BMJ in 2022 found 22% fewer incident autoimmune diseases on vitamin D, with a hazard ratio of 0.78 and a confidence interval that just cleared unity.

That finding comes with an instructive caveat. A follow-up published in Arthritis & Rheumatology in 2024 tracked the cohort for two years after the trial ended and found the effect gone, with a hazard ratio of 0.98 at seven years. Whatever vitamin D was doing required continued daily exposure. It was not stored.

A VITAL sub-study published in the American Journal of Clinical Nutrition in 2025 measured leukocyte telomere length in 1,054 participants at baseline, year two and year four. Daily vitamin D3 reduced telomere attrition by 0.14 kilobase pairs relative to placebo; marine omega-3 did not. This is a cellular surrogate rather than a clinical outcome, and telomere dynamics have a long history of over-interpretation — but it is a randomised result from a 2,000 IU daily dose.

Most pointed of all, an individual-patient-data meta-analysis in Ageing Research Reviews in 2023 pooled around 105,000 participants and found a non-significant 6% reduction in cancer mortality overall. Restricted to daily-dosing trials, the reduction was 12%. In bolus trials there was nothing. The schedule did not modify the effect; it was the effect.

The fracture question, settled at high certainty

The oldest claim for vitamin D is bone, and it is the claim that has aged worst.

In May 2026 the BMJ published a systematic review and meta-analysis of 69 randomised trials and 153,902 adults. For vitamin D alone, the risk ratio for any fracture was 1.00 with a confidence interval of 0.95 to 1.06 — 36 trials, 92,045 participants, graded high certainty. Calcium alone gave 0.91 at moderate certainty. Calcium and vitamin D together gave 0.91 at high certainty, statistically significant but below the reviewers' pre-specified threshold for clinical importance. Across all three comparisons the effect on falls was little to none.

The authors are explicit that these findings do not extend to people with specific bone disorders or to those on drug treatment for osteoporosis. That boundary matters, and it is the opposite of the boundary usually drawn in marketing, which extrapolates a treatment population's needs onto a healthy one.

The large primary trials point the same way. VITAL was null on cancer and cardiovascular events; the D-Health trial gave 21,315 Australians 60,000 IU monthly for five years and found an all-cause mortality hazard ratio of 1.04, with cancer mortality at 1.15 — not significant, but not in the hoped-for direction either, and delivered on precisely the monthly bolus schedule the meta-analytic evidence identifies as inert.

The excess you cannot take back

The European Food Safety Authority sets the tolerable upper intake level for adults at 100 micrograms, or 4,000 IU, per day, from all sources including supplements, and the same figure applies in pregnancy and lactation. NHS advice uses the same ceiling, against a UK reference intake of 10 micrograms daily.

Put a 300,000 IU ampoule on that scale. It contains roughly two years of the UK daily reference intake, administered at once into a tissue that will release it on its own schedule.

Documented toxicity is uncommon but not theoretical. A 2018 review in Nutrients collected 13 published cases in which serum 25-hydroxyvitamin D ran from 150 to 1,220 ng/mL and serum calcium from 11.1 to 23.1 mg/dL, with manufacturing errors and overdosing by patients or prescribers as the recurring causes. A 2021 letter in the Indian Journal of Anaesthesia made the operational point that 600,000-unit intramuscular injections remain in use alongside oral products of 1,000, 2,000 and 60,000 IU intended for daily, weekly or monthly use — a set of strengths spanning nearly three orders of magnitude, which is a dosing-error architecture rather than a dosing schedule.

Vitamin D is also fat-soluble and stored, which is why the fat-soluble vitamins as a class — including vitamin A — are the ones with meaningful upper limits. A water-soluble excess is largely excreted. An oil-depot excess is not.

The Gulf and the United Kingdom are not the same problem

None of the above means deficiency is imaginary, and in the Gulf it plainly is not.

A systematic review published in Frontiers in Nutrition in July 2025 pooled 35 studies and 28,260 subjects in the United Arab Emirates. The pooled mean 25-hydroxyvitamin D concentration in adults was 17.63 ng/mL, with a confidence interval of 14.28 to 20.99 — a population mean sitting below the 20 ng/mL deficiency threshold, in one of the sunniest places on earth. Indoor living, sun avoidance, covering clothing and dietary intake are the usual explanations, and the practical consequence is that repletion is a genuine clinical question across the region rather than a marketing construct.

The United Kingdom's position is different in kind. SACN and Public Health England advise 10 micrograms daily for everyone over one year of age, with particular attention to autumn and winter when UV-B is insufficient for cutaneous synthesis at British latitudes. That is a population dietary measure, not a diagnostic pathway.

The 2024 Endocrine Society clinical practice guideline draws the line between them. It recommends against routine 25-hydroxyvitamin D testing in the general population, noting that no clinical trial evidence was found to support it and that the panel could not establish blood-level thresholds for adequacy or for disease prevention. It advises against empiric supplementation above the dietary reference intakes in healthy adults under 75, while recommending it for children and adolescents, adults over 75 to lower mortality risk, pregnancy, and high-risk prediabetes. Real deficiency deserves treatment; a screening test attached to a product does not follow from it.

Vitamin K2 after DANCODE

The pairing of vitamin D with vitamin K2 is usually asserted rather than evidenced, so the recent randomised data deserve accurate reporting in both directions.

The same Danish group ran two trials of menaquinone-7 at 720 micrograms daily plus vitamin D at 25 micrograms daily against placebo for 24 months. The first, in JACC: Advances in 2023, randomised 389 men and found no significant reduction in coronary artery calcium progression overall. The DANCODE trial, published in Circulation on 28 August 2026, randomised 398 participants with severe calcification — a median baseline score of 903 Agatston units — and found progression of 196 units against 248 on placebo, with no increase in non-calcified plaque volume.

That is a real randomised result, and it should be read for exactly what it is: a slowing of an imaging surrogate in people who already have severe calcification, not a demonstrated reduction in cardiovascular events, and not a general case for adding K2 to a supplement. Note also the vitamin D dose that produced it — 25 micrograms, 1,000 IU, daily.

Evidence graded by outcome, schedule and route

Sorted by what was actually given and how often, the field is more coherent than its reputation suggests.

How to read a vitamin D offer

Three questions do most of the work.

First, what is the schedule? This is the variable the trials keep separating, and it is the one marketing suppresses in favour of the total dose. A regimen that delivers its year in a single administration is the regimen that generated the falls and fracture signals; a small daily dose is the regimen that generated every benefit signal the field has. Convenience is a real clinical value where adherence fails — it is simply not a pharmacological advantage here.

Second, what is licensed for the route being proposed? For intramuscular ergocalciferol the answer is malabsorption, with weekly calcium monitoring at the start. For intravenous vitamin D there is no answer, because there is no intravenous cholecalciferol product; the activated analogues that are given intravenously belong to renal medicine and carry their own monitoring requirements.

Third, is anything being measured that will change a decision? A 25-hydroxyvitamin D result in a healthy adult under 75 is a number the Endocrine Society explicitly declines to attach a threshold to. In a patient with malabsorption, chronic kidney disease or suspected osteomalacia it is a different test entirely, asked for a different reason.

Vitamin D is not the failure in this story. It is among the best-studied molecules in nutrition, with more large randomised evidence behind it than most prescription medicines, and a real deficiency problem across the Gulf that deserves a serious answer. What fails is the translation — from a genuine population deficiency to a megadose injection, from a stored fat-soluble vitamin to an intravenous bag it cannot enter. Water-soluble micronutrient profiles such as the MyerSence MD concept exist in the IVIXIR series precisely because solubility decides the route, not the other way round. Letting the molecule's chemistry set the format, and the randomised schedule set the dose, is the standard EFBA applies — and vitamin D is the clearest case in the series of a molecule whose evidence is strong and whose delivery claims are not.