If you’ve been following vitamin D research for a while, the chances are that you’ve grown increasingly confused, frustrated, or cynical—or maybe some combination of these.
Starting in the 1990s, observational studies have reported that relatively high levels of blood vitamin D (25-hydroxyvitamin D, or 25(OH)D) are associated with greatly reduced risk of multiple diseases, including atherosclerotic cardiovascular disease (ASCVD), diabetes, various cancers, obesity, psychiatric diseases, age-related cognitive decline and dementia,1 and all-cause mortality.2,3,4,5 Sometimes the lowest-risk group had 25(OH)D levels well in excess of 30 ng/mL6,7—the level the Endocrine Society once considered “sufficient.”8
And yet for decades now, we have seen trial after trial test the ability of vitamin D supplements to prevent these diseases—or to reduce adverse outcomes from them—only to go nowhere. For example, in the VITAL trial, 25,871 participants (including 5106 African Americans, who have higher rates of deficiency) in their 50s and older were randomized to receive 2,000 IU per day of vitamin D3 or placebo daily. After a median of 5.3 years of followup, there were no differences between the groups in their risk of major cardiovascular events (MACE—heart attacks, strokes, or deaths from cardiovascular disease), or of each of these individually, or of death from any cancer or other cancer outcomes.9 And VITAL was just one of many similar failed trials.10–14 Even pooled together into meta-analyses, there was no hint of a benefit for cardiovascular outcomes,15,16 which are the main subject of this newsletter.
However, there was a clear problem with these trials. The point of supplementing vitamin D is to correct a vitamin D deficiency or “insufficiency”8 (or, more tendentiously, to achieve “optimal” levels) so that there is enough of the hormone to carry out its biological functions. But none of these trials were actually designed to test the idea that achieving a biologically appropriate level of 25(OH)D would protect people against ASCVD or MACE.
Instead, they were simplistic tests of the effect of vitamin D supplementation per se. Participants were taken from the general population rather than being recruited on the basis of having low vitamin D levels, and all participants randomized to vitamin D were given the same dose, regardless of what their levels were. Indeed, 55.1% of the subjects in VITAL had 25(OH)D levels equal to or greater than 30 ng/mL before supplementation began.9 And while the data in the other trials of vitamin D and cardiovascular outcomes are in different ways problematic, the same was true of about half of subjects in FIND,13 about a third of the subjects in ViDA,11,17 and half or more of people in D-Health,14 although 40.7% of the subjects in DO-HEALTH were frankly deficient (25(OH)D <20 ng/mL) and the number who were “sufficient” is not clear.12
As a review of these trials in Nature Reviews Endocrinology put it, these and similar trials showed “that vitamin D supplementation of vitamin D-replete adults … does not prevent cancer, cardiovascular events, falls or progression to type 2 diabetes mellitus [emphasis ours].”18 In other words: If it ain’t broke, don’t be surprised that you can’t fix it.
TARGET-D: treat to target
What critics (including us) have long called for instead is a ‘treat-to-target’ trial, in which investigators recruit people with low vitamin D levels and randomize them to either placebo or a titrated vitamin D dosing protocol.19,20,21 People would have their 25(OH)D levels tested periodically, and the investigators would adjust their vitamin D dose until they reach a targeted level of 25(OH)D. The investigators would then follow up to see if getting people’s vitamin D levels to target can reduce their risk of adverse outcomes such as MACE.
Now we have the first, preliminary attempt at such a trial. In the TARGET-D trial, 630 people were recruited within a month of suffering a heart attack and randomly assigned to either usual care or titrated vitamin D supplementation, with a target 25(OH)D of at least 40 ng/mL.22 (This target came from the authors’ unpublished analysis of a previous cohort study from their health system.22 They also cite four papers as supporting a 40 ng/mL target,8.23,24,25 though none of them actually do so*). The participants’ median age was 63, and they were overwhelmingly (78.1%) men.
At baseline, 87.0% of subjects had 25(OH)D levels of 40 ng/mL or under, with a median of 25 ng/mL (interquartile range: 18-33).22 The few subjects whose 25(OH)D was already 40 ng/mL or greater stayed pat and returned in a year for followup testing. The rest were given one of four starter doses of vitamin D3, depending on their baseline 25(OH)D level, and then re-tested in 3 months. If at that point their 25(OH)D had reached 40 ng/mL, great: They kept that dose and returned in a year for followup. Otherwise, their dose was doubled, and they were tested again in 3 months, repeating until they reached at least the 40 ng/mL target.22
A Rorschach result
After an average follow-up of 4.2±2.0 years, there was no statistically significant difference in MACE between the two groups, with 15.7% of the vitamin D group suffering MACE, versus 18.4% of subjects in the usual care group (hazard ratio (HR) 0.85, 95% CI 0.58–1.24, p = 0.40).22
However, that wasn’t necessarily the end of it. For starters, the trial was not able to fully enroll as a result of the COVID-19 pandemic; this left it underpowered relative to the investigators’ original study plan,22 leaving us to wonder if the nonsignificant 0.85 HR might have firmed up if they had fully enrolled the trial.
Moreover, when the investigators broke down MACE into its component outcomes, there seemed to be a protective effect specifically against new heart attacks: 3.8% of subjects in the vitamin D group suffered one, versus 7.9% of those in the usual care group (HR 0.48, 95% CI 0.24-0.96, p = 0.03).22 If that were a real effect, slashing heart attacks in half would be an exciting result, even without a reduction in MACE overall. Indeed, the previous D-Health trial also reported a reduction in heart attacks in its vitamin D group (HR 0.81, 95% CI 0.67–0.98), despite no effect on MACE.26
This isolated reduction in heart attacks in the vitamin D group relative to the usual-care group was enough to trigger the headline “Heart attack risk halved in adults with heart disease taking tailored vitamin D doses,” and for the lead author to say that she and her colleagues “encourage people with heart disease to discuss vitamin D blood testing and targeted dosing with their health care professionals to meet their individual needs.”27
But there are a few flies in that ointment. For one thing, the VITAL trial had heart attacks as a prespecified secondary outcome, and found no effect (HR 0.96, 95% CI 0.78-1.19).9 The same was true of three other major trials with a cardiovascular outcome.11,12,13 Of course, TARGET-D was a different kind of trial from its predecessors, but we mustn’t seize on the seeming confirmation on heart attacks in D-Health when multiple other trials with similar designs failed to find an effect.
For another, while the TARGET-D investigators reported the reduction in heart attacks as statistically significant, they did so while ignoring the “multiple comparisons problem.” When you measure many variables in a study, the odds of stumbling across positive-looking findings by sheer chance increase with each additional variable. There are statistical methods to account for this problem, but the investigators didn’t deploy them.
In this case, separately testing MACE, heart attack, death, heart failure hospitalization, and stroke without tightening the statistical pass-fail threshold might conceivably have pushed heart attacks across the line when a more rigorous test would have found it compatible with chance. In fact, the paper lists at least a dozen further secondary outcomes and four subgroup variables, none corrected for multiplicity; the secondary MI result is the one positive-looking outcome, and the one the investigators choose to highlight, but it’s one out of many shots on goal. And in fact, although not significant, others of these outcomes relevant to ASCVD pointed toward harm: nonsignificant nominal increases in the risk of heart failure hospitalization (4.2% vs 3.5%, HR 1.21, 95% CI 0.54-2.69) and stroke (1.6% vs 0.9%, HR 1.69. 95% CI 0.41-7.09).22
A third issue: The headline effect on heart attacks was in the “intention-to-treat” (ITT) analysis, meaning that it included everyone in the vitamin D and control groups. However, despite the elaborate test-and-dose-adjust protocol, 21% of the people in the vitamin D group never actually reached the 40 ng/mL 25(OH)D target. If the reduced risk of heart attacks in the vitamin D group was truly the result of the biological effect of raising a person’s level of circulating vitamin D to 40 ng/mL, as per the trial’s hypothesis, then vitamin D supplementation should have been at least as effective at cutting heart attack risk among people in the vitamin D group who actually achieved the 25(OH)D target (after adjusting for demographics and risk factors—the “per protocol” analysis) as it was in the vitamin D group as a whole.
There’s a limitation to this reasoning, since the trial can only randomize people to get the vitamin D intervention or not: achieving the target is not itself randomized, so restricting the analysis to those patients introduces potential selection bias. Indeed, the per-protocol analysis should if anything have exaggerated the true biological effect. That’s because people who tightly adhere to a medication regimen tend to differ from those who don’t in other ways, such as in general health-consciousness, vaccination, education, and adherence to their other therapies like statins and blood pressure meds.28–30
But instead, the difference in the risk of heart attacks got weaker and was no longer statistically significant in this achievers-only analysis (HR 0.69, 95% CI 0.33-1.44; log-rank p = 0.15).22
This weaker result is only a limited knock on the main (intention-to-treat) analysis result, for a few reasons. For one thing, the per-protocol estimates for MACE (HR 0.74) and for death (HR 0.62) nominally moved in the “right” direction in the per-protocol, though neither was statistically significant.22 Also, the achievers are a subset of the overall sample of vitamin D participants, so the two estimates aren’t independent: the 10 heart attacks in the vitamin D arm of the per-protocol analysis are a subset of the 12 in the all-comers vitamin D analysis, and the confidence intervals overlap almost completely. And the total number of participants in the per-protocol analysis was necessarily lower than in the all-comers intention-to-treat analysis, which reduces statistical power. But certainly the per-protocol analysis gives no sign that the heart-attack signal tracks with reaching the target, as was the hypothesis driving TARGET-D.
Our research team spends hundreds of hours each month vetting studies and distilling dense literature to deliver evidence-informed insights on health and longevity. If you find value in our work, consider becoming a premium member and supporting our mission.














Of performance, punctiliousness, and placebos
But there’s a more fundamental problem built into TARGET-D. Recall that there was no placebo group in the study: The control group was simply assigned to usual care.22 That opens up concerns about causality, because the lack of blinding, the involvement in elaborate medical procedures, and the effect of believing that one is receiving an intervention might have improved the supplemented participants’ outcomes for reasons other than their higher 25(OH)D levels.
First, there is performance bias. When doctors and patients know who is getting an intervention and who is not, the doctors might treat the participants differently depending on which group they are assigned to, such as by investing more into the care of people getting the tested treatment or unconsciously behaving more optimistically in their presence (having emotionally “written off” people not getting the intervention). Reciprocally, people who know they are in the active treatment group may be more optimistic and take their other health screenings and medications more seriously.
Indeed, some level of performance bias was built into the study design itself, irrespective of the mindsets and choices of individual doctors and patients: While people in the vitamin D group were coming back frequently for testing and dose adjustments, giving them more opportunities to interact with clinicians, people in the usual care group had one baseline round of testing done at the beginning and then went about their business. As the study design says, “No other contact is planned with the standard of care patients. Follow-up will be done by the querying of electronic records”.31
Another possible source of bias comes from the elaborate procedure necessitated by the trial’s treat-to-target design. Although the systematic evidence is mixed,32 many studies find that all placebos are not created equal: The more elaborate, dramatic, or involved an intervention is, the larger the placebo effect it generates.33,34,35 In this context, it is reasonable to suspect that prescribing a person a personalized dose of vitamin D, determined through repeated rounds of testing and adjustment, would be more likely to make patients psychologically invest in the treatment and change their health behaviors in ways that could impact the outcome measures than the one-dose-for-all strategy used in most clinical trials.
The potential bias created by the lack of placebos and the testing procedure could have been neutralized by having a placebo group instead of a usual-care group as controls, and having them undergo the same repeat testing as the vitamin D group followed by sham dose-adjustments. This was done (for instance) in the TRAVERSE trial of the cardiovascular safety of testosterone replacement therapy,36 but it was not the protocol in TARGET-D—and instead, there was no placebo of any kind.
Could these various sources of bias have had an influence on a hard outcome like heart attacks? Plausibly. People who are consistent in taking other chronic disease medications, especially statins, are more likely to also take their bisphosphonates for osteoporosis;30 to get vaccinated and go for their cancer screenings;29 and to have lower rates of adverse outcomes unrelated to their LDL levels, including notably food poisoning.28 And a meta-analysis found good placebo adherence in clinical trials is associated with lower cardiovascular mortality (odds ratio (OR) 0.68, 95% CI 0.60–0.77),37 suggesting again that something about being highly engaged with one’s health itself marks a person who will have better outcomes, independent of the therapeutic value of the treatments themselves.
Similarly, people in the vitamin D group of TARGET-D might have become more conscientious about taking their lipid-lowering, blood pressure, and antidiabetic medications after going through the extra medical engagement required of them as part of the trial. If so, then that extra medication adherence—rather than the vitamin D capsules or achieved 25(OH)D levels—could have reduced their risk of a heart attack.
The possibility of such an increase in medication adherence in the vitamin D arm of TARGET-D could have been probed if the investigators had re-tested subjects’ LDL-C, blood pressure, blood glucose, and other druggable causal risk factors at the end of the trial. If these risk factors were reduced in the vitamin D group but not the controls, it would tend to support this performance bias hypothesis. Unfortunately, these risk factors were only measured at baseline.22
Father Mendel would like a word
TARGET-D is in some ways the best trial we’ve had to test whether vitamin D supplementation might reduce the risk of cardiovascular disease endpoints—but as we’ve seen, it still has important methodological limitations that limit our confidence in the secondary outcome of heart attack reduction. What other evidence do we have?
The observational studies linking vitamin D to greatly reduced risk of ASCVD, diabetes, multiple cancers, all-cause mortality, and more are highly confounded.2,3,4,5 If you have high vitamin D levels, you’re more likely to be spending time outside in the sun—possibly socializing, possibly exercising, and not sitting alone, indoors, in front of a screen all day. You’re more likely to have a healthy body composition, since adipose tissue sequesters vitamin D, which leads to low vitamin D levels in people with too much of it.
You’re also less likely to be Black38,39 or American Indian,38 which ancestries typically have lower circulating vitamin D levels due to more pigmented skin blocking out UV light. People with these ancestries also have increased incidence of many poor health outcomes relative to Whites, including increased risk of some cancers40 and shorter life expectancy.41 But it’s clear that there are many complex reasons why health outcomes vary by race, including socioeconomic status, population genetics, and discrimination in healthcare,40,41 and there is some evidence that people of African ancestry may have evolved means of preserving vitamin D biological activity at a lower level of 25(OH)D,42–44 suggesting that Black people’s poor health outcomes are confounded with rather than caused by their low 25(OH)D.
Inflammation is also associated with lower 25(OH)D levels, especially if that inflammation is linked to a disease45—but it’s not simply that low vitamin D status broadly causes inflammation, since while vitamin D supplements lower inflammatory markers in people with abnormal glucose homeostasis,46 they do not do so in asthmatics46 or in overweight or obese people.46,47 Indeed, 25(OH)D levels fall acutely after a variety of inflammatory insults, leading some to propose that it is itself a negative acute-phase reactant48—i.e., that its fall is part of the rapid systemic response to insults like infections and heart attacks that works to minimize tissue damage and enhance repair processes.
So we have methodological problems with the clinical trials of vitamin D supplementation for ASCVD outcomes, and we have very challenging confounding in the observational evidence. But we have another way of getting at causality on this question: Mendelian randomization (MR). MR is a powerful genetic tool for uncovering natural experiments that approximate a randomized controlled trial, if its core assumptions hold. In MR, researchers look to see whether genetic variants that reliably influence an exposure of interest (for instance, a person’s 25(OH)D) are also associated with an outcome of interest (in this case, ASCVD—see our Premium Article that covers Mendelian randomization in depth).
Because genetic variants are randomly assigned to us by the accidents of our conception—by the chance shuffling of our parents’ DNA—MR sidesteps much of the confounding by lifestyle and environment that plagues conventional observational studies. For instance, the genetic variants that predispose toward higher circulating vitamin D levels are independent of whether you get out for a run in the sunshine every day. Conversely, a person’s genotype can’t be changed by disease, so MR is less vulnerable to “reverse causation” than conventional observational studies—that is, to cases where a disease might lower a person’s 25(OH)D and be misinterpreted as low 25(OH)D causing the disease.
MR isn’t a perfect substitute for a randomized trial: it depends on important assumptions, including notably that the variants affect the outcome only through the exposure of interest. In the case of circulating vitamin D, this could trip us up: Some MR studies have relied on variants involved in regulating 25(OH)D that some studies have found to also be related to plasma LDL-cholesterol (LDL-C)49,50 or other atherogenic lipoproteins51 (although others have not found this52); when such studies find associations between genetically-predicted 25(OH)D levels and ASCVD, they may therefore be confounded. There are methods to check for undue influence from such “pleiotropic” effects, which some but not all MR studies have used.49,
Another potential source of pleiotropic confounding: Some genetic variants that affect vitamin D levels do so through their effect on skin pigmentation, which could confound genetically-determined vitamin D levels with the complex association between race and health outcomes; two such variants were excluded from an MR study of vitamin D and skin cancer for this reason.53 However, the variants that affect skin pigmentation are either absent from the large data sets used in most of the MR studies on 25(OH)D and cardiovascular outcomes,54-57 and/or don’t appear to meaningfully affect vitamin D levels.58
Also, some more recent MR studies that reported an increased risk of cardiovascular diseases in people with low genetically-predicted 25(OH)D were retracted after problems with their statistical methods were uncovered,59,60 while others have remained in print despite statistical weaknesses,55,61,62 some of them with “findings” reported in the abstracts that are not consistent with the actual findings in the paper63 or headline analyses that don’t match the data tables and find discrepant results across measures of ASCVD.55
But here’s the punchline: When one of the two retracted papers was re-analyzed and republished with statistical methods that did not incorrectly “assume” that the genetic influence on 25(OH)D is constant across the population (in fact, it’s stronger at higher 25(OH)D levels and weaker at lower ones), the corrected analysis found no such association.64 Another paper that initially reported a similar association 65 found a null result when its investigators re-analyzed the data in response to the same error being pointed out.66
Similarly, other studies with statistical methods that test for and attempt to correct pleiotropic effects and single-variant outliers also found no likely effect of genetically-predicted 25(OH)D on risk of cardiovascular outcomes,56,57,67 as have multiple other MR studies.68–71
Our best evidence from the natural experiments of Mendelian randomization, then, is that there is no detectable causal effect of circulating vitamin D levels on risk of ASCVD and related cardiovascular outcomes—which is consistent with most of the trials, and with most of the outcomes in TARGET-D.
What about frank deficiency?
Even with the best MR studies finding no associations, there remain caveats. For instance, the variants that are known to influence 25(OH)D levels account for less than 10% of the variance across the population,54,72 so MR could miss effects that require larger differences in exposure. But you would expect that even relatively modest effects would shift some people from having 25(OH)D levels that were too low to provide any putative cardiovascular protection to levels that were sufficient to do so.
A related limitation is that the high-quality Mendelian randomization studies primarily assess lifelong, small, mostly linear effects: None of them could strongly rule out an effect of correcting frank deficiency (which is usually operationally defined as a 25(OH)D below 20 ng/mL8,73—though even this threshold has become contentious, with some arguing that even lower levels of 25(OH)D may be required to rule someone deficient24,74,75).
Whichever threshold one adopts, the MR studies are poorly positioned to test it, though they fall short in different ways. One of the better MR analyses was methodologically unable to look specifically at deficiency.57 Another looked at deficiency and found no effect, but was badly underpowered for this purpose, because the database of genetic variants on which it relied had too few cases of deficiency.56 A third paper didn’t look at deficiency per se, but performed stratified analyses to look at people with the lowest levels of genetically-predicted 25(OH)D.64 It still found null associations with multiple cardiovascular outcomes; however, the instrument is again underpowered to look at very low levels of predicted 25(OH)D, because the genetic effect is weaker at low levels of 25(OH)D.64 The same limitation applies to another stratified analysis that similarly found no effect of even the lowest levels of genetically-predicted vitamin D levels on all-cause, cardiovascular, or cancer mortality.66
However, while we have been using the MR studies to complement the results of TARGET-D and other vitamin D trials, in the case of deficiency we can supplement the weaknesses of the MRs with trial data. While critics are right to point out that VITAL and the other vitamin D cardiovascular trials recruited many people with relatively high levels of 25(OH)D, several of these trials did perform subgroup analyses looking specifically at people who began the trial with a 25(OH)D below 20 ng/mL. Even in deficient participants, there was no effect of vitamin D supplementation on MACE or other cardiovascular outcomes in VITAL (HR 1.09, 95% CI 0.68–1.76)9 or ViDA (HR 1.00, 95% CI 0.74-1.35).11
There may not have been enough deficient subjects in either of these trials individually to detect a real effect on MACE if there were one. However, the Endocrine Society pooled together all the deficient subjects across the 3 trials where baseline 25(OH)D levels were directly measured9,11,26 into a meta-analysis, which found no effect of vitamin D supplementation on heart attacks (relative risk (RR) 0.93, 95% CI 0.38 to 2.29), total cardiovascular events (RR 1.02, 95% CI 0.87 to 1.19), cancer (RR 0.91, 95% CI 0.70 to 1.19), or all-cause mortality (RR 1.11, 95% CI 0.85 to 1.46).15
All of these analyses have limitations, but together, the fact that the quality MR studies found no association between genetically-predicted 25(OH)D and cardiovascular outcomes and the lack of effect of vitamin D supplementation in deficient participants—the people most likely to benefit if a benefit did exist—argue against putting much faith in the single positive outcome in TARGET-D.
An unsatisfying conclusion
TARGET-D failed to find an effect of moderately aggressive treat-to-target vitamin D supplementation on its primary outcome (MACE),22 and while it found a significant effect on heart attacks, there are statistical and methodological reasons to suspect that this finding is overstated and possibly artifactual. The previous vitamin D trials, although lacking TARGET-D’s treat-to-target design, also failed to find an effect on their primary cardiovascular outcomes,9-14 even when restricted to people who began their trials with vitamin D deficiency15 or in a trial where close to half the participants (40.7%) were deficient.12
Together, the higher-quality MR analyses,56,57,64,66,67 each with its own strengths and weaknesses, jointly find no detectable linear lifelong effect of 25(OH)D on ASCVD, even studies using genetically-predicted vitamin D deficiency56 or in people with the lowest levels of genetically-predicted 25(OH)D.64,66
They can’t give us a definitive answer either, because of the limitations we discussed, but it may be the most information that can be extracted from MR.
It’s clear what the ideal trial to answer this question (and the similar questions about cancer and other chronic diseases) would be. Recruit a well-powered cohort of people who are both at elevated cardiovascular risk and vitamin D deficient, and randomize them to either treat-to-target vitamin D supplementation or placebo treatment with a sham test-and-titrate procedure. Then follow them up for as long as it takes to reliably see a difference in MACE, and find some way to keep people in the placebo arm from independently accessing vitamin D supplements or spending more time in the sun.
That’s a very tall ask. As the VITAL investigators wrote, “In observational studies, the 25(OH)D levels associated with lowest risks tend to be above … 20-25 ng/mL for cardiovascular disease. … [I]t remains possible that a trial among persons with extremely low vitamin D (i.e., well below the 20 ng/mL level recommended for bone health) would show stronger risk reductions. However, maintaining participants in a vitamin D deficient state and circumventing real-world clinical care for 5 years would be neither ethical nor feasible.”9
And if the main question that still needs asking is indeed whether correcting frank vitamin D deficiency reduces the risk of ASCVD or MACE, then such a trial would be of scientific but not practical value, since correcting vitamin D deficiency is already endorsed for avoiding demonstrable, vitamin D-specific harm to the body’s function in the form of impaired bone mineralization. Each of us can decide whether it’s worth getting tested for the Endocrine Society’s old deficiency (20 ng/mL) or biochemical insufficiency (30 ng/mL) cutpoints8,23–25 and performing our own treat-to-target protocols, though the Society itself has withdrawn both and only advocates testing for people with clinical reasons to suspect deficiency, such as low blood calcium levels.76,77 But based on available evidence, there may be no need to think or act differently about vitamin D out of concern for cardiovascular disease than there is for its established role in skeletal health.
That’s not the conclusion that inspired these several trials or sparked the excitement of many scientists and health-conscious people, but it may be the best practical answer we’re going to get.
* One cited paper found that vitamin D deficiency (20 ng/mL), but not vitamin D insufficiency (30 ng/mL), was associated with increased all-cause mortality;23 some cited papers support cutoffs of 208,23–25and/or 308 ng/mL 25(OH)D, and in one case mention cutpoints as low as 10 ng/mL;24 one mentions, but does not endorse or give evidence in support of, “experts or grassroots organizations [that] recommend that all adults should aim for serum levels of 25OHD of ≥125 nmol/l [50 ng/mL], as found in African individuals living in conditions similar to our Palaeolithic ancestors.”24 None give evidence in support of or endorse a 40 ng/mL target.
For a list of all previous weekly emails, click here.
References
1. Balion C, Griffith LE, Strifler L, et al. Vitamin D, cognition, and dementia: a systematic review and meta-analysis. Neurology. 2012;79(13):1397-1405.
2. Dalamaga M, Emfietzoglou R, Petropoulou D, et al. Vitamin D and health outcomes: State-of-the-art review of triangulated evidence and ongoing controversies. Curr Nutr Rep. 2026;15(1):26.
3. Autier P, Mullie P, Macacu A, et al. Effect of vitamin D supplementation on non-skeletal disorders: a systematic review of meta-analyses and randomised trials. Lancet Diabetes Endocrinol. 2017;5(12):986-1004.
4. Autier P, Boniol M, Pizot C, Mullie P. Vitamin D status and ill health: a systematic review. Lancet Diabetes Endocrinol. 2014;2(1):76-89.
5. Theodoratou E, Tzoulaki I, Zgaga L, Ioannidis JPA. Vitamin D and multiple health outcomes: umbrella review of systematic reviews and meta-analyses of observational studies and randomised trials. BMJ. 2014;348(apr01 2):g2035.
6. McDonnell SL, Baggerly CA, French CB, et al. Breast cancer risk markedly lower with serum 25-hydroxyvitamin D concentrations ≥60 vs <20 ng/ml (150 vs 50 nmol/L): Pooled analysis of two randomized trials and a prospective cohort. PLoS One. 2018;13(6):e0199265.
7. Garland CF, Gorham ED, Mohr SB, et al. Vitamin D and prevention of breast cancer: pooled analysis. J Steroid Biochem Mol Biol. 2007;103(3-5):708-711.
8. 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):1911-1930.
9. Manson JE, Cook NR, Lee IM, et al. Vitamin D supplements and prevention of cancer and cardiovascular disease. N Engl J Med. 2019;380(1):33-44.
10. Scragg R. The Vitamin D Assessment (ViDA) study – Design and main findings. J Steroid Biochem Mol Biol. 2020;198(105562):105562.
11. Scragg R, Stewart AW, Waayer D, et al. Effect of monthly high-dose vitamin D supplementation on cardiovascular disease in the vitamin D Assessment Study : A randomized clinical trial: A randomized clinical trial. JAMA Cardiol. 2017;2(6):608-616.
12. Gaengler S, Sadlon A, De Godoi Rezende Costa Molino C, et al. Effects of vitamin D, omega-3 and a simple strength exercise programme in cardiovascular disease prevention: The DO-HEALTH randomized controlled trial. J Nutr Health Aging. 2024;28(2):100037.
13. Virtanen JK, Nurmi T, Aro A, et al. Vitamin D supplementation and prevention of cardiovascular disease and cancer in the Finnish Vitamin D Trial: a randomized controlled trial. Am J Clin Nutr. 2022;115(5):1300-1310.
14. Neale RE, Baxter C, Romero BD, et al. The D-Health Trial: a randomised controlled trial of the effect of vitamin D on mortality. Lancet Diabetes Endocrinol. 2022;10(2):120-128.
15. 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):1961-1974.
16. Pei YY, Zhang Y, Peng XC, Liu ZR, Xu P, Fang F. Association of vitamin D supplementation with cardiovascular events: A systematic review and meta-analysis. Nutrients. 2022;14(15):3158.
17. Scragg R. Clinical trials of vitamin D supplementation and cardiovascular disease: A synthesis of the evidence. J Steroid Biochem Mol Biol. 2025;250(106733):106733.
18. Bouillon R, Manousaki D, Rosen C, Trajanoska K, Rivadeneira F, Richards JB. The health effects of vitamin D supplementation: evidence from human studies. Nat Rev Endocrinol. 2022;18(2):96-110.
19. Rejnmark L, Bislev LS, Cashman KD, et al. Non-skeletal health effects of vitamin D supplementation: A systematic review on findings from meta-analyses summarizing trial data. PLoS One. 2017;12(7):e0180512.
20. Amrein K, Scherkl M, Hoffmann M, et al. Vitamin D deficiency 2.0: an update on the current status worldwide. Eur J Clin Nutr. 2020;74(11):1498-1513.
21. Wimalawansa SJ. Enhancing the design of nutrient clinical trials for disease prevention-A focus on vitamin D: A systematic review. Nutr Rev. 2025;83(7):e1740-e1781.
22. 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;(ehag611). doi:10.1093/eurheartj/ehag611
23. Wang TY, Wang HW, Jiang MY. Prevalence of vitamin D deficiency and associated risk of all-cause and cause-specific mortality among middle-aged and older adults in the United States. Front Nutr. 2023;10:1163737.
24. Bouillon R. Comparative analysis of nutritional guidelines for vitamin D. Nat Rev Endocrinol. 2017;13(8):466-479.
25. Draft Evidence Review: Vitamin D, Calcium, or Combined Supplementation for the Primary Prevention of Falls and Fractures in Community-Dwelling Adults: Preventive Medication. December 17, 2024. Accessed November 5, 2026. https://www.uspreventiveservicestaskforce.org/uspstf/document/draft-evidence-review/vitamin-d-calcium-combined-supplementation-primary-prevention-falls-fractures-communitydwelling-adults
26. Thompson B, Waterhouse M, English DR, et al. Vitamin D supplementation and major cardiovascular events: D-Health randomised controlled trial. BMJ. 2023;381:e075230.
27. Heart attack risk halved in adults with heart disease taking tailored vitamin D doses. American Heart Association. Accessed November 9, 2026. https://newsroom.heart.org/news/heart-attack-risk-halved-in-adults-with-heart-disease-taking-tailored-vitamin-d-doses
28. Patrick AR, Shrank WH, Glynn RJ, et al. The association between statin use and outcomes potentially attributable to an unhealthy lifestyle in older adults. Value Health. 2011;14(4):513-520.
29. Brookhart MA, Patrick AR, Dormuth C, et al. Adherence to lipid-lowering therapy and the use of preventive health services: an investigation of the healthy user effect. Am J Epidemiol. 2007;166(3):348-354.
30. Curtis JR, Xi J, Westfall AO, et al. Improving the prediction of medication compliance: the example of bisphosphonates for osteoporosis: The example of bisphosphonates for osteoporosis. Med Care. 2009;47(3):334-341.
31. A Trial Evaluating Vitamin D Normalization on Major Adverse Cardiovascular-Related Events Among Myocardial Infarction Patients (The TARGET-D Study). Intermountain Health Care, Inc.; 2016. https://clinicaltrials.gov/study/NCT02996721
32. Fässler M, Meissner K, Kleijnen J, Hróbjartsson A, Linde K. A systematic review found no consistent difference in effect between more and less intensive placebo interventions. J Clin Epidemiol. 2015;68(4):442-451.
33. de Craen AJ, Tijssen JG, de Gans J, Kleijnen J. Placebo effect in the acute treatment of migraine: subcutaneous placebos are better than oral placebos. J Neurol. 2000;247(3):183-188.
34. Bannuru RR, McAlindon TE, Sullivan MC, Wong JB, Kent DM, Schmid CH. Effectiveness and implications of alternative placebo treatments: A systematic review and network meta-analysis of osteoarthritis trials: A systematic review and network meta-analysis of osteoarthritis trials. Ann Intern Med. 2015;163(5):365-372.
35. Meissner K, Linde K. Are blue pills better than green? How treatment features modulate placebo effects. Int Rev Neurobiol. 2018;139:357-378.
36. Lincoff AM, Bhasin S, Flevaris P, et al. Cardiovascular safety of testosterone-replacement therapy. N Engl J Med. 2023;389(2):107-117.
37. Yue Z, Cai C, Ai-Fang Y, Feng-Min T, Li C, Bin W. The effect of placebo adherence on reducing cardiovascular mortality: a meta-analysis. Clin Res Cardiol. 2014;103(3):229-235.
38. Ritterhouse LL, Lu R, Shah HB, et al. Vitamin d deficiency in a multiethnic healthy control cohort and altered immune response in vitamin D deficient European-American healthy controls. PLoS One. 2014;9(4):e94500.
39. Ginde AA, Liu MC, Camargo CA Jr. Demographic differences and trends of vitamin D insufficiency in the US population, 1988-2004. Arch Intern Med. 2009;169(6):626-632.
40. Zavala VA, Bracci PM, Carethers JM, et al. Cancer health disparities in racial/ethnic minorities in the United States. Br J Cancer. 2021;124(2):315-332.
41. GBD US Health Disparities Collaborators. Life expectancy by county, race, and ethnicity in the USA, 2000-19: a systematic analysis of health disparities. Lancet. 2022;400(10345):25-38.
42. Winder MB, Mason DL, Rangaswami J, Asif A, Vachharajani TJ, Mathew RO. Racial differences in the relationship between high-normal 25-hydroxy vitamin d and parathyroid hormone levels in early stage chronic kidney disease. J Bras Nefrol. 2021;43(1):34-40.
43. Thoreson CK, Chung ST, Ricks M, et al. Biochemical and clinical deficiency is uncommon in African immigrants despite a high prevalence of low vitamin D: the Africans in America study. Osteoporos Int. 2015;26(11):2607-2615.
44. Wright NC, Chen L, Niu J, et al. Defining physiologically “normal” vitamin D in African Americans. Osteoporos Int. 2012;23(9):2283-2291.
45. Kruit A, Zanen P. The association between vitamin D and C-reactive protein levels in patients with inflammatory and non-inflammatory diseases. Clin Biochem. 2016;49(7-8):534-537.
46. Hsu CY, Hussein HY, Al-Akkam KA, et al. The effects of vitamin D supplementation on inflammatory biomarkers: an umbrella study of meta-analysis on randomized controlled trials. Inflammopharmacology. 2026;34(7):4297-4310.
47. Zuk A, Fitzpatrick T, Rosella LC. Effect of vitamin D3 supplementation on inflammatory markers and glycemic measures among overweight or obese adults: A systematic review of randomized Controlled Trials. PLoS One. 2016;11(4):e0154215.
48. Silva MC, Furlanetto TW. Does serum 25-hydroxyvitamin D decrease during acute-phase response? A systematic review. Nutr Res. 2015;35(2):91-96.
49. Burgess S, Gill D. Genetic evidence for vitamin D and cardiovascular disease: choice of variants is critical. Eur Heart J. 2022;43(18):1740-1742.
50. Willer CJ, Schmidt EM, Sengupta S, et al. Discovery and refinement of loci associated with lipid levels. Nat Genet. 2013;45(11):1274-1283.
51. Ooi EM, Afzal S, Nordestgaard BG. Elevated remnant cholesterol in 25-hydroxyvitamin D deficiency in the general population: Mendelian randomization study: Mendelian randomization study. Circ Cardiovasc Genet. 2014;7(5):650-658.
52. Berry DJ, Vimaleswaran KS, Whittaker JC, Hingorani AD, Hyppönen E. Evaluation of genetic markers as instruments for Mendelian randomization studies on vitamin D. PLoS One. 2012;7(5):e37465.
53. Liyanage UE, Law MH, Melanoma Meta-analysis Consortium, Barrett JH, Iles MM, MacGregor S. Is there a causal relationship between vitamin D and melanoma risk? A Mendelian randomization study. Br J Dermatol. 2020;182(1):97-103.
54. Jiang X, O’Reilly PF, Aschard H, et al. Genome-wide association study in 79,366 European-ancestry individuals informs the genetic architecture of 25-hydroxyvitamin D levels. Nat Commun. 2018;9(1):260.
55. Xu B, Liu X, Liu X, Chen J, Cheng T, Chen L. Investigation of the causal effect and molecular signatures of serum hydroxyvitamin D on atherosclerosis: a Mendelian randomization and transcriptomic approach. Korean J Physiol Pharmacol. 2026;30(4):299-312.
56. Meena D, Dib MJ, Huang J, et al. Associations of genetically predicted vitamin D status and deficiency with the risk of carotid artery plaque: a Mendelian randomization study. Sci Rep. 2024;14(1):14743.
57. Jiang X, Ge T, Chen CY. The causal role of circulating vitamin D concentrations in human complex traits and diseases: a large-scale Mendelian randomization study. Sci Rep. 2021;11(1):184.
58. Wang X, Hivert V, Groot S, et al. Cross-ancestry analyses identify new genetic loci associated with 25-hydroxyvitamin D. PLoS Genet. 2023;19(11):e1011033.
59. 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):837-846.
60. Zhou A, Selvanayagam JB, Hyppönen E. Non-linear Mendelian randomization analyses support a role for vitamin D deficiency in cardiovascular disease risk. Eur Heart J. 2022;43(18):1731-1739.
61. Xue Y, Cao S, Zhang L, Chen J, Ning A. Investigating the mechanisms linking vitamin D to coronary artery disease: A mediating proteomics Mendelian randomisation study. Medicine (Baltimore). 2026;105(29):e49798.
62. Li X, Wang W, Liu H, Xia R, Lin Y, Pan G. Causal relationship between serum 25-hydroxyvitamin D and coronary heart disease: A bi-directional two-sample Mendelian randomization study. Nutr Metab Cardiovasc Dis. 2025;35(6):104020.
63. Li J, Song Z, Hu Y, et al. Non-linear association between serum 25-hydroxyvitamin D levels and cardiovascular outcomes: Evidence from NHANES and mendelian randomization analysis. Nutr Metab Cardiovasc Dis. 2025;35(10):104155.
64. 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. 2024;12(1):e2-e11.
65. 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):1552-1559.
66. Hyppönen E, Sutherland JP, Zhou A. Vitamin D deficiency increases mortality risk in the UK biobank. Ann Intern Med. 2024;177(12):1743-1746.
67. Huang T, Afzal S, Yu C, et al. Vitamin D and cause-specific vascular disease and mortality: a Mendelian randomisation study involving 99,012 Chinese and 106,911 European adults. BMC Med. 2019;17(1):160.
68. Brøndum-Jacobsen P, Benn M, Jensen GB, Nordestgaard BG. 25-hydroxyvitamin d levels and risk of ischemic heart disease, myocardial infarction, and early death: population-based study and meta-analyses of 18 and 17 studies: Population-based study and meta-analyses of 18 and 17 studies. Arterioscler Thromb Vasc Biol. 2012;32(11):2794-2802.
69. Afzal S, Brøndum-Jacobsen P, Bojesen SE, Nordestgaard BG. Genetically low vitamin D concentrations and increased mortality: Mendelian randomisation analysis in three large cohorts. BMJ. 2014;349(v18 2):g6330.
70. Brøndum-Jacobsen P, Benn M, Afzal S, Nordestgaard BG. No evidence that genetically reduced 25-hydroxyvitamin D is associated with increased risk of ischaemic heart disease or myocardial infarction: a Mendelian randomization study. Int J Epidemiol. 2015;44(2):651-661.
71. Manousaki D, Mokry LE, Ross S, Goltzman D, Richards JB. Mendelian randomization studies do not support a role for vitamin D in coronary artery disease. Circ Cardiovasc Genet. 2016;9(4):349-356.
72. Liu H, Shen X, Yu T, et al. A putative causality of vitamin D in common diseases: A mendelian randomization study. Front Nutr. 2022;9:938356.
73. Institute of Medicine, Food and Nutrition Board, Committee to Review Dietary Reference Intakes for Vitamin D and Calcium. Dietary Reference Intakes for Calcium and Vitamin D. (Del Valle HB, Yaktine AL, Taylor CL, Ross AC, eds.). National Academies Press; 2011.
74. Giustina A, Bouillon R, Binkley N, et al. Controversies in vitamin D: A statement from the Third International Conference. JBMR Plus. 2020;4(12):e10417.
75. 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):1146-1152.
76. 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):1907-1947.
77. 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):1948-1954.




