Vitamin D after sixty: what the large trials in older adults found, and why the guidelines disagree

The searches that bring people to this topic sound almost identical: "vitamin D for the elderly", "how long can you take vitamin D", "can you take it all year round". In other words, the person has already decided that a deficiency is worth correcting — we have a separate piece on that, vitamin D and longevity — and is asking something else: what changes after sixty, what exactly the large trials in older adults tested, why "more" turned out worse in several of them, and how two authoritative guidelines came to read the same blood test differently. This piece is about that.

Why it is a separate question after sixty

Skin makes vitamin D less well with age. In a study of human skin samples (donors aged 8 to 92), the concentration of the precursor, 7-dehydrocholesterol, fell with age, and on exposure to ultraviolet light the skin of 77- and 82-year-olds produced more than twofold less previtamin D3 than the skin of 8- and 18-year-olds. Add to that what no laboratory measures: an older person spends less time in the sun and exposes less skin.

Hence the plain arithmetic of the bodies that set intake standards. The US Institute of Medicine (2011) derived its recommended dietary allowance from conditions of minimal sun exposure: 600 IU a day from age 1 to 70 and 800 IU from age 71 — the amount at which a 25-hydroxyvitamin D level of at least 20 ng/mL (50 nmol/L) is reached in 97.5% of the population. The gap between those two figures is the price of age in that model.

The second feature of older age is not biochemical but practical: harm has a different meaning. A young person has little to lose from a spare tablet; for someone past seventy the main risk is a fall, and it is on falls that vitamin D behaved unexpectedly in several trials. We will come back to that.

Fractures: where it worked and where it did not

The oldest and most cited positive trial is French, from 1992: 3,270 women, mean age 84, living in nursing homes (a detail from the same authors' 1994 publication on the same cohort), received 800 IU of vitamin D3 together with 1.2 g of calcium or placebo for 18 months. Among those who completed the study, hip fractures were 43% fewer and all non-vertebral fractures 32% fewer; parathyroid hormone fell by 44%, and femoral bone density rose by 2.7% against a 4.6% decline on placebo. This is a population with deficiency (mean baseline 25(OH)D around 13–16 ng/mL), old, sedentary, with calcium in the same tablet — all four features matter.

Trials then moved to people living at home, and the result changed.

• RECORD (2005), 5,292 people aged 70 or older after a low-trauma fracture, 800 IU of vitamin D3 and/or 1,000 mg of calcium daily: none of the options reduced the rate of further fractures (hazard ratio for vitamin D 1.02).

• Women's Health Initiative (2006), 36,282 postmenopausal women, 400 IU plus 1,000 mg of calcium a day for seven years: hip bone density 1% higher, no statistically significant reduction in hip fracture (hazard ratio 0.88 with a confidence interval including one), more kidney stones (1.17).

• VITAL (2022), 25,871 people — men from 50 and women from 55, not selected for deficiency or low bone mass — 2,000 IU a day, median follow-up 5.3 years: 1,991 fractures, hazard ratio 0.98 for total fractures, 1.01 for hip fractures. No effect in any subgroup, including those with low baseline levels.

• D-Health (2023), 20,326 Australians aged 60–84, 60,000 IU once a month for up to five years: hazard ratio for total fractures 0.94 with an interval from 0.84 to 1.06 — statistically indistinguishable from null; the hazard ratio fell with longer follow-up, but the test of that interaction did not reach significance (p = 0.14).

The summaries say the same as the individual trials. The 2014 Cochrane review (53 trials, 91,791 participants): vitamin D alone, in the doses tested, is unlikely to prevent hip fracture (11 trials, 27,693 participants, risk ratio 1.12); vitamin D with calcium gives a small reduction (nine trials, 49,853 participants, 0.84) — and the authors translate that into absolute numbers at once: one hip fracture prevented per 1,000 community-dwelling older adults per year, nine per 1,000 among those living in institutions. A 2018 meta-analysis with trial sequential analysis (81 RCTs, 53,537 participants): no effect on total fractures (relative risk 1.00), hip fractures (1.11), falls (0.97) or bone density; the same result in trials of high doses.

The pattern in the populations reads plainly: the positive result came where people were deficient, old, living in an institution and given calcium in the same tablet; in people living at home and not selected for deficiency there was no effect. None of these trials compared "deficient" with "not deficient" head to head, so this is a description of whom the result was obtained in, not proof that correcting the level is enough. It is consistent with the conclusion of our first article — and that is as far as its strength goes.

Falls: why "more" turned out worse

Here lies the most unexpected part of the evidence, and it bears directly on the question "how should it be taken".

• Australia, 2010: 2,256 women aged 70 or older at high risk of fracture received 500,000 IU of vitamin D3 once a year (in autumn or winter) or placebo for 3–5 years. The vitamin D group had more falls (83.4 versus 72.7 per 100 person-years, rate ratio 1.15) and more fractures (4.9 versus 3.9 per 100 person-years, 1.26). The largest excess of falls came in the first three months after dosing (1.31) — when the blood level was at its peak, around 120 nmol/L.

• Zurich, 2016: 200 people aged 70 or older with a prior fall, three monthly regimens — 24,000 IU, 60,000 IU, or 24,000 IU plus calcifediol. The higher regimens raised 25(OH)D above 30 ng/mL more reliably — and did not improve lower-extremity function, while the incidence of falls over the year in those groups was 66.9% and 66.1% versus 47.9% on the low dose.

• STURDY (2021), 688 people aged 70 or older at elevated fall risk with 25(OH)D between 25 and 72.5 nmol/L; 200 IU a day against 1,000, 2,000 and 4,000 IU. In the dose-finding stage there were more falls on 2,000 and 4,000 IU than on 1,000; 1,000 IU was selected as the best dose and participants were moved to it; 1,000 IU against 200 did not prevent falls (hazard ratio 0.94), while for serious falls and falls with hospitalisation the risk was higher (1.87 and 2.48). The trial has caveats: the control group received 200 IU rather than placebo, and the dose-finding stage ended before its prespecified thresholds were reached.

• ViDA (2017), 5,108 New Zealanders aged 50–84, 100,000 IU a month (after an initial 200,000) for 2.5–4.2 years: the same number of falls (hazard ratio 0.99), non-vertebral fractures 6% versus 5% (1.19, not statistically significant).

Two different conclusions follow from this group of studies, and it matters not to fuse them. First: very large infrequent doses — annual, or large monthly — were accompanied by more falls in several trials in older adults, and that is not one chance finding. Second: a moderate monthly regimen in itself did not increase fractures in two large trials (D-Health, ViDA); the D-Health authors write explicitly that their data "do not support concerns" about a monthly bolus. So the problem is not the words "once a month" but the size of the dose and who receives it.

This is what the US Preventive Services Task Force recommendation (2018) rests on: for community-dwelling adults aged 65 or older without known osteoporosis or deficiency — do not prescribe vitamin D to prevent falls (grade D, meaning "harm or no benefit"). For fall prevention the same body recommends exercise — grade B for community-dwelling adults aged 65 or older at increased risk of falls; multifactorial programmes — grade C, individually. And it is what the new Endocrine Society guideline (2024) rests on: for people over 50 in whom vitamin D is indicated — daily administration rather than intermittent high doses.

Mortality, heart, cancer: what was tested and what came out

Observational studies linked low 25(OH)D to higher mortality for decades; why that is an association rather than an effect we covered in the first article. Here is what the trials found where vitamin D was given at random.

• VITAL (2019), 25,871 people (men from 50, women from 55), 2,000 IU a day, 5.3 years: invasive cancer — hazard ratio 0.96, major cardiovascular events — 0.97, death from any cause — 0.99. Death from cancer — 0.83 with an interval from 0.67 to 1.02, that is, on the boundary; it is a secondary end point and the authors did not claim it as a result. There was no excess hypercalcaemia.

• D-Health (2022), 21,315 people aged 60–84, 60,000 IU a month, 5 years: all-cause mortality 1.04 (interval 0.93–1.18), cardiovascular 0.96, cancer 1.15 (0.96–1.39). The 25(OH)D level in the vitamin D group was 115 nmol/L against 77 on placebo — the supplement worked as a supplement, and that did not carry over into mortality.

• D-Health, heart (2023): major cardiovascular events 6.0% versus 6.6%, hazard ratio 0.91 with an interval of 0.81–1.01. The authors phrase it cautiously: "might reduce", but the absolute difference is small (5.8 events per 1,000 over five years, 172 people treated per event avoided) and the interval is consistent with no effect. The signal was stronger in those taking cardiovascular drugs at baseline (0.84), but the interaction test did not reach significance.

• DO-HEALTH (2020), 2,157 people aged 70 or older without major comorbidities (mean age 74.9), 2,000 IU a day, omega-3 and strength training in a factorial design, 3 years: none of the interventions produced a statistically significant effect on any of the six end points — blood pressure, physical performance, a cognitive test, non-vertebral fractures, infections.

• ViDA (2017), 5,108 people aged 50–84, monthly 100,000 IU: cardiovascular disease 11.8% versus 11.5%, hazard ratio 1.02 — and the same in those who were deficient at baseline.

Summaries. A 2019 meta-analysis (52 RCTs, 75,454 participants): all-cause mortality — relative risk 0.98, cardiovascular — 0.98; cancer mortality — 0.84 (0.74–0.95), one of the few consistent signals in the field. The 2014 Cochrane review (56 trials, 95,286 participants, mostly women over 70; in 26 of the 45 trials with a known baseline level it was below 20 ng/mL): 12.5% versus 12.7% deaths, relative risk 0.97; split by form, the effect held only for vitamin D3 (0.94), not for D2 or the active metabolites. The systematic review underpinning the 2024 Endocrine Society guideline: no significant effect in healthy adults aged 19–74; in people over 75 — a very small reduction in mortality with high certainty. That line is what the guideline's main recommendation for older adults is built on.

Infections: where the dosing regimen shows most clearly

Respiratory infections are the one outcome for which the difference between daily and infrequent dosing has been shown not in a single trial but in a pooled analysis of individual data.

A 2017 meta-analysis of individual participant data from 10,933 people in 25 RCTs (aged 0 to 95): vitamin D reduced the proportion of people with at least one acute respiratory infection — odds ratio 0.88. But split by regimen: daily or weekly without a bolus — 0.81; with at least one large dose — 0.97, that is, no effect. Within the daily and weekly regimens, protection was strongest in those starting below 25 nmol/L (0.30) and markedly weaker in those starting higher (0.75). The 2021 update on aggregate data from 46 RCTs (75,541 participants) gave a more modest overall figure — 0.92 — and the same profile: an effect in daily regimens (0.78) and at daily dose equivalents of 400–1,000 IU (0.70); this time no subgroup by baseline level was found. These are data across all ages, not older adults specifically — worth remembering when reading them.

Why the guidelines disagree

The disagreement is not about the data but about which question counts as the main one.

The Institute of Medicine (2011) answered "how much does the population need": a level of 20 ng/mL (50 nmol/L) covers bone requirements in 97.5% of people, an allowance of 600–800 IU, a tolerable upper intake level of 4,000 IU a day; the committee also recorded that the prevalence of deficiency in North America had been overestimated, and that for some outcomes the association is U-shaped — risk at both low and high levels.

The Endocrine Society (2011) answered "how to treat a patient at risk" and defined deficiency as below 20 ng/mL, insufficiency as 21–29, and named a level above 30 ng/mL as the treatment goal. Hence a decade of two "normals" on one lab report.

The Endocrine Society (2024) revised its own position, and this is the most substantial change in the field in years. Three points. First: no trial was found in which routine 25(OH)D testing improved an outcome — not in the general population, not in obesity, not in people with dark skin — so the panel came out against routine testing in all the populations considered and declined to name a "target level". Second: for people aged 75 and older, empiric supplementation is suggested — without testing — because of a possible small reduction in mortality; for healthy adults under 75, supplementation above the dietary allowance to prevent disease is not suggested. Third: over 50 — daily, not intermittent high doses. The panel notes at once that the optimal dose for these groups remains unclear: doses in the trials varied widely and participants were often allowed to continue their own supplements.

The US Preventive Services Task Force (2018) on fractures in community-dwelling adults: for postmenopausal women, no benefit is established for 400 IU or less with 1,000 mg of calcium or less (a recommendation against); for higher doses, for men and for premenopausal women the evidence is insufficient; and — with adequate certainty — vitamin D combined with calcium increases the incidence of kidney stones. On falls — a recommendation against, described above.

The upshot: one guideline says "over 75, take it without testing", another says "for fracture prevention at home, not established", and both are right about their own questions. The first looks at mortality in the oldest group; the second at fractures in women living at home.

"How long" — what is actually known

The question "in courses or all year round" is better rephrased as "what happens to the level when intake stops" — for that there are measurements.

Vitamin D3 itself accumulates in fat tissue and leaves the body slowly — a half-life of about two months; its transport form 25(OH)D, the thing the test measures, about 15 days. In a winter study of 67 men in Omaha (41° north), the 25(OH)D level changed in proportion to dose — roughly 0.7 nmol/L for each additional microgram a day — and the authors calculated that summer stores from skin synthesis covered more than 80% of these men's winter requirement. In other words, "seasonality" is not a dosing schedule but a property of the system itself: the level rises over months and falls over months.

Two things follow. A short "course" does not create a new steady state — the plateau takes on the order of three months, visible both in the dose-response studies and in the Australian trial, where after the annual dose the level fell from 120 to 90 nmol/L within three months. And continuous multi-year intake of moderate doses did not surprise the trials on safety: five years of 2,000 IU in VITAL passed without excess hypercalcaemia, three years of 2,000 IU in DO-HEALTH with similar numbers of deaths in all groups. Toxicity — hypercalcaemia — in the published cases with a known dose is described with sustained intake of 40,000 IU a day or more and at 25(OH)D levels far above those produced by ordinary regimens; the Institute of Medicine's tolerable upper intake level for adults is 4,000 IU a day.

And one figure on why "one dose for everyone" does not work. In a trial of 163 postmenopausal women (aged 57–90, mean baseline 39 nmol/L), 800 IU a day raised 25(OH)D above 50 nmol/L in 97.5% of participants, the dose-response curve plateaued around 112 nmol/L, and women with obesity had a level 17.8 nmol/L lower at the same dose. Body mass is the second factor after age that stops someone else's prescription from transferring.

What the field does not have

There is no trial in which the level was measured, supplementation given according to the result, and outcomes compared with people who were not tested. That is exactly why the Endocrine Society could not recommend the test in 2024 — not because the test is bad, but because nobody has checked its benefit.

There is no agreed "target level" for disease prevention. The 20 and 30 ng/mL thresholds are thresholds for bone health and for defining deficiency, not targets to which raising a healthy person is proven to help.

There is no trial showing benefit from a level "above normal". In D-Health the level in the vitamin D group was 115 nmol/L against 77 — and mortality did not differ.

There are no data on harm from a moderate monthly regimen in older adults — but there are on harm from an annual dose and from a large monthly one. These two statements sit side by side without contradicting each other.

What to do with this at the appointment

1. Whether to test is an open question today, and different guidelines answer it differently. It is a conversation with a doctor, not something settled by an article.

2. If supplementation is prescribed — a daily regimen, not large infrequent doses: the one thing on which the guidelines and the trials in older adults agree.

3. Calcium is a separate decision with separate risks (kidney stones), not "what comes in the package".

4. For fall prevention the USPSTF guideline recommends exercise (in community-dwelling older adults at increased risk of falls), and not vitamin D.

5. Dose and duration are the doctor's decision, and it depends on age, body mass, baseline level and what the person already takes. The trials above have their doses, but they also have their populations.

Two other quantities that determine the risk of falls and fractures in older adults are measured without a laboratory — muscle mass and gait speed; what grip strength predicts we cover separately.

In the Lonevi record, a 25(OH)D level, if it has been measured, sits next to its date and laboratory — and that is the point here: one result answers "where am I now", two results a season apart answer "is this changing at all".

This material is informational and does not replace a consultation with a doctor. Decisions about taking any medication or supplement, about doses and duration, and about which tests you need are made together with your treating physician.

References

• MacLaughlin J., Holick M.F. (1985). Aging decreases the capacity of human skin to produce vitamin D3. Journal of Clinical Investigation, 76(4), 1536–1538. PMID 2997282.

• Ross A.C. et al. (2011). The 2011 report on dietary reference intakes for calcium and vitamin D from the Institute of Medicine: what clinicians need to know. Journal of Clinical Endocrinology & Metabolism, 96(1), 53–58. PMID 21118827.

• Holick M.F. et al. (2011). Evaluation, treatment, and prevention of vitamin D deficiency: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology & Metabolism, 96(7), 1911–1930. PMID 21646368.

• Demay M.B. et al. (2024). Vitamin D for the prevention of disease: an Endocrine Society clinical practice guideline. Journal of Clinical Endocrinology & Metabolism, 109(8), 1907–1947. PMID 38828931.

• Shah V.P. et al. (2024). A systematic review supporting the Endocrine Society clinical practice guidelines on vitamin D. Journal of Clinical Endocrinology & Metabolism, 109(8), 1961–1974. PMID 38828942.

• US Preventive Services Task Force, Grossman D.C. et al. (2018). Vitamin D, calcium, or combined supplementation for the primary prevention of fractures in community-dwelling adults: recommendation statement. JAMA, 319(15), 1592–1599. PMID 29677309.

• US Preventive Services Task Force, Grossman D.C. et al. (2018). Interventions to prevent falls in community-dwelling older adults: recommendation statement. JAMA, 319(16), 1696–1704. PMID 29710141.

• Chapuy M.C. et al. (1992). Vitamin D3 and calcium to prevent hip fractures in elderly women. New England Journal of Medicine, 327(23), 1637–1642. PMID 1331788.

• Chapuy M.C. et al. (1994). Effect of calcium and cholecalciferol treatment for three years on hip fractures in elderly women. BMJ, 308(6936), 1081–1082. PMID 8173430.

• Grant A.M. et al. (2005). Oral vitamin D3 and calcium for secondary prevention of low-trauma fractures in elderly people (RECORD): a randomised placebo-controlled trial. Lancet, 365(9471), 1621–1628. PMID 15885294.

• Jackson R.D. et al. (2006). Calcium plus vitamin D supplementation and the risk of fractures. New England Journal of Medicine, 354(7), 669–683. PMID 16481635.

• LeBoff M.S. et al. (2022). Supplemental vitamin D and incident fractures in midlife and older adults. New England Journal of Medicine, 387(4), 299–309. PMID 35939577.

• Waterhouse M. et al. (2023). The effect of monthly vitamin D supplementation on fractures: a tertiary outcome from the population-based, double-blind, randomised, placebo-controlled D-Health trial. Lancet Diabetes & Endocrinology, 11(5), 324–332. PMID 37011645.

• Avenell A. et al. (2014). Vitamin D and vitamin D analogues for preventing fractures in post-menopausal women and older men. Cochrane Database of Systematic Reviews, (4), CD000227. PMID 24729336.

• Bolland M.J., Grey A., Avenell A. (2018). Effects of vitamin D supplementation on musculoskeletal health: a systematic review, meta-analysis, and trial sequential analysis. Lancet Diabetes & Endocrinology, 6(11), 847–858. PMID 30293909.

• Sanders K.M. et al. (2010). Annual high-dose oral vitamin D and falls and fractures in older women: a randomized controlled trial. JAMA, 303(18), 1815–1822. PMID 20460620.

• Bischoff-Ferrari H.A. et al. (2016). Monthly high-dose vitamin D treatment for the prevention of functional decline: a randomized clinical trial. JAMA Internal Medicine, 176(2), 175–183. PMID 26747333.

• Appel L.J. et al. (2021). The effects of four doses of vitamin D supplements on falls in older adults (STURDY): a response-adaptive, randomized clinical trial. Annals of Internal Medicine, 174(2), 145–156. PMID 33284677.

• Khaw K.T. et al. (2017). Effect of monthly high-dose vitamin D supplementation on falls and non-vertebral fractures: secondary and post-hoc outcomes from the randomised, double-blind, placebo-controlled ViDA trial. Lancet Diabetes & Endocrinology, 5(6), 438–447. PMID 28461159.

• Scragg R. et al. (2017). Effect of monthly high-dose vitamin D supplementation on cardiovascular disease in the Vitamin D Assessment Study (ViDA): a randomized clinical trial. JAMA Cardiology, 2(6), 608–616. PMID 28384800.

• Manson J.E. et al. (2019). Vitamin D supplements and prevention of cancer and cardiovascular disease (VITAL). New England Journal of Medicine, 380(1), 33–44. PMID 30415629.

• Neale R.E. et al. (2022). The D-Health Trial: a randomised controlled trial of the effect of vitamin D on mortality. Lancet Diabetes & Endocrinology, 10(2), 120–128. PMID 35026158.

• Thompson B. et al. (2023). Vitamin D supplementation and major cardiovascular events: D-Health randomised controlled trial. BMJ, 381, e075230. PMID 37380191.

• Bischoff-Ferrari H.A. et al. (2020). Effect of vitamin D supplementation, omega-3 fatty acid supplementation, or a strength-training exercise program on clinical outcomes in older adults: the DO-HEALTH randomized clinical trial. JAMA, 324(18), 1855–1868. PMID 33170239.

• Zhang Y. et al. (2019). Association between vitamin D supplementation and mortality: systematic review and meta-analysis. BMJ, 366, l4673. PMID 31405892.

• Bjelakovic G. et al. (2014). Vitamin D supplementation for prevention of mortality in adults. Cochrane Database of Systematic Reviews, (1), CD007470. PMID 24414552.

• Martineau A.R. et al. (2017). Vitamin D supplementation to prevent acute respiratory tract infections: systematic review and meta-analysis of individual participant data. BMJ, 356, i6583. PMID 28202713.

• Jolliffe D.A. et al. (2021). Vitamin D supplementation to prevent acute respiratory infections: a systematic review and meta-analysis of aggregate data from randomised controlled trials. Lancet Diabetes & Endocrinology, 9(5), 276–292. PMID 33798465.

• Jones G. (2008). Pharmacokinetics of vitamin D toxicity. American Journal of Clinical Nutrition, 88(2), 582S–586S. PMID 18689406.

• Heaney R.P. et al. (2003). Human serum 25-hydroxycholecalciferol response to extended oral dosing with cholecalciferol. American Journal of Clinical Nutrition, 77(1), 204–210. PMID 12499343.

• Vieth R. (1999). Vitamin D supplementation, 25-hydroxyvitamin D concentrations, and safety. American Journal of Clinical Nutrition, 69(5), 842–856. PMID 10232622.

• Gallagher J.C. et al. (2012). Dose response to vitamin D supplementation in postmenopausal women: a randomized trial. Annals of Internal Medicine, 156(6), 425–437. PMID 22431675.