Skip to content

Vitamin pill shocker: “A complex web of vested interests promote calcium and vitamin D for osteoporosis, despite lack of evidence”

Paul Alper points us to this scary news article by Susan Perry:

Calcium and vitamin D supplements have been shown repeatedly to have no beneficial effect on preventing or treating osteoporosis . . . In fact, the evidence has not only demonstrated that calcium and vitamin D supplements do not reduce the risk of bone fractures, it has also found that they may cause harm in some cases. This harm includes a greater risk of developing kidney stones, gastrointestinal symptoms that require hospitalization, heart attacks, stroke and even (paradoxically) hip fractures.

OK, fair enough. Lots of ideas seem good until you try them out and gather lots of real-world data; calcium supplements could be one such idea.

So what’s the problem?

Perry continues:

Yet doctors continue to recommend the supplements to their patients. And people keep taking them. Indeed, studies have revealed that more than half of older Americans (70 percent of older women) take either prescription or over-the-counter calcium and/or vitamin D supplements — mostly because they’ve been told the products will strengthen their bones.


Why has the evidence about the ineffectiveness of these supplements in preventing broken bones and about their potential for harm — evidence that has been around since 2002 — failed to dampen people’s belief in them?

I guess doctors are as clueless as the rest of us. It’s hard to keep up with the news.

But maybe there’s more going on. Perry points to an article in the British Medical Journal by Andrew Grey and Mark Bolland, who write:

Industry gains scientific credibility, which protects or enhances the sales of its products, and indirect marketing through advocacy groups. Advocacy organisations and specialist societies gain funds to support their existence. Academics gain by maintenance of their status and by obtaining access to research funds and career enhancing publications and presentations.

Perry continues:

– Here are just a few of the reasons Grey and Bolland cite for why medical specialty societies, advocacy groups and academia continue to discount the science and promote supplements for preventing and treating osteoporosis:

Calcium and vitamin D supplements are enormously profitable — and not just for the companies that make and sell them.

– Health-advocacy groups, such as the National Osteoporosis Foundation (NOF) in the United States and the International Osteoporosis Foundation (IOF) in Europe, rely heavily on money from commercial sponsors.

– Some academic researchers also rely heavily on the support of the nutrition and supplement industry, although they have often failed to disclose that financial connection.

This all seems like bad news, and it’s good for this particular rock to be picked up so we can see the worms crawling underneath.

I have one question

In her news article, Perry contrasts calcium and vitamin D supplements with “other therapies once recommended for the prevention and treatment of osteoporosis — estrogen, calcitonin and fluoride [which] were essentially abandoned once they were shown to be ineffectual or harmful.”

So here’s my question. If there’s money to be made in calcium and vitamin D, why isn’t there money to be made in estrogen, calcitonin, and fluoride. What happened so there isn’t a web of vested interest for these other discredited treatments?

I posed this question to Alper and he responded as follows:

Susan Perry is the expert on this subject. My guess is that estrogen, calcitonin and flouride are far harder and more costly to obtain compared to nipping down to the nearest pharmacy for calcium and vitamin D. Perry appears to be the author of a book on menopause where you can see her comments on estrogen, calcitonin and fluoride.

The “web of vested interest” is more tenuous when there is an expensive intermediary such as a doctor’s visit. Besides, Americans are enamored with pill taking and fish oil swallowing. I have observed friends drinking a large glass of orange juice followed by a vitamin C pill.

I’m reminded of my Linus Pauling theory.

Alper posed my question to Perry who told us some more:

Estrogen has fallen out of favor for the treatment of osteoporosis because of the findings from the Women’s Health Initiative regarding serious adverse risks. That’s not to say that some sectors within the medical community do not continue to try to revive the use of the drug for osteoporosis prevention, but most women have wisely resisted that recommendation. As for calcitonin, the FDA ruled a couple of years ago that it shouldn’t be marketed for the treatment of osteoporosis, again because of research showing links to potentially serious adverse effects, so that would explain it going out of favor. Doctors would be unwilling to recommend a product to their patients that the FDA has explicitly said shouldn’t be used for that purpose. I suspect that fluoride supplements were found to be ineffective, although that is an older story (and one I’d have to research for the details). Also, fluoride is very out of favor with people who practice alternative medicine, so that may also have contributed to it being abandoned.

Estrogen is certainly more expensive and more difficult to obtain than vitamin D and calcium supplements. I’m not sure about calcitonin and fluoride. But I think the reasons I cite above may be stronger explanations for why estrogen, calcitonin and fluoride have been (essentially) abandoned for the treatment of osteoporosis. I don’t know for sure, though.

“I don’t know for sure” . . . that’s not something you’ll be hearing from David Brooks any time soon! I guess we can forget Susan Perry ever getting a column in the New York Times.


  1. Rahul says:

    So what’s the canonical meta-analysis on this? Is it absolutely clear that Calcium etc. do more harm than good?

    • John Thacker says:

      There’s a variety of conditions that they may help. I don’t think that the science is necessarily so clear (see e.g. here: or the Mayo Clinic on Vitamin D:

      There’s pretty clear benefits for people who have deficiencies, little benefits for people who have adequate levels, and fighting around what is considered an adequate level. They are also, on the whole, cheap (4-5 cents a pill) and paid out of pocket, so while it adds up to a large revenue because more people take them than necessary, hundred of pills can be cheaper than running a blood test to verify levels. It doesn’t take much for universal vitamin D to be cheaper for the health system than universal vitamin D blood tests followed by supplements only for those who need them.

      Doctors are going to response pretty differently to something when it’s “eh, won’t hurt and it’s cheap” to estrogen therapy, where the downside risks are pretty established now.

      • I think this is more or less right when it comes to Vit D. It seems to have lots of other benefits, including reduced risks of some cancers, and reduced risk of various inflammatory diseases (it’s used by the body to modulate down inflammation). It’s incredibly cheap when purchased in reasonable quantities that you’d use for long term supplementation, and a large fraction of the population have marginal blood levels.

        Calcium on the other hand seems to be mostly flimflam.

        • Also, there are multiple forms of “vitamin D”. One form, I think “calcidiol” or hydroxyvitamin D has a long half life (tens of hours) in the body, and is converted into “calcitriol” by hydroxylation, which then has a very short half life (tens of minutes?). Supplementation seeks to increase the circulation of this long-life pre-form, but because of the way equilibrium chemistry works, supplements push the reaction towards increased calcitriol and therefore increased elimination rate as well. So it’s not like you can just give a one-time dose of vitamin D and hope to get the levels up and “fix things” (though some people do try this). Daily supplementation makes sense for vitamin D. It’s a hormone you normally produce daily when exposed to sunlight, but lots of people avoid sun exposure and wind up with lowish vit D levels.

          (note: I am not a biochemist/chemist, some of what I wrote above may be inaccurate (particularly the naming), but I believe the basic story is correct, you can easily google and check this all out, so please do if you’re interested)

      • Rahul says:


        Thanks for elaborating a more nuanced view.

        The original post & the quotes made it sound as if doctors were ignorant & peddling quack medicine on the lobbying of the pharma companies.

    • gwern says:

      I don’t think that’s clear at all and the vitriol of the paper linked in OP is a little surprising to me. If you look through the meta-analyses of the RCTs on vitamin D’s effect on all-cause mortality, they all seem to turn in plausible point values of ~RR=0.95. I’ve been playing with the Gompertz curve, and an RR of 0.95 gives you about half a year of life, which at the usual health-care economics valuations like $50k/year, is $25k or considerably more than the lifetime cost of $0.5k-$1k or so (some vitamin D-3 pills will cost you something like $10-20 per year, assuming you don’t let yourself be ripped off, and there’s not much reason to start before age 30 so if you live to your 80s, that’s 50 years of $10 costs). Giving the elderly vitamin D supplements is doing the right thing perhaps for the wrong reason.

  2. Steve B says:

    I submit that the continued use of calcium and vitamin D supplements is due most to the desire to ‘do something’ (“Illusion of control” to use the behavioral decision-making term.) And a lag after decades of authorities saying it was a good thing. The companies profiting from calcium and vitamin D are too diverse and small to be driving the use.

  3. Jon says:

    A challenge in these issues is that these conditions evolve over decades. It has long been public that any benefits are most likely from avoiding defficiency well before osteoporosis becomes a risk.

    There is also a bias that people would rather try something than be resigned to spending their final years crippled.

  4. Martha says:

    One complexity that seems not to be getting adequate attention in this discussion is that osteoporosis (low bone density) isn’t a very strong predictor of fracture. I like the display at the very bottom of for illustrating this.

    So what do I (a smallish woman in her early seventies with lowish bone density and whose mother had both cervical and hip fractures in her early eighties) do? My doctor put me on a bisphosphonate when I was in my early sixties, but agreed that stopping it was wise when the gastric side effects were bad. He then put me on nasally administered calcitonin, but took me off of it a few years ago when research showed evidence of possible carcinogenic effects. (I didn’t object, even though the study was on men, not women, because I doubted that the positive effect, if any, would be very great.)

    He continues to tell me to take a certain amount of calcium and vitamin D. But a couple of years ago, I read that excesses of them may increase heart attack risk. So I roughly calculated my total intake of each from diet and try to take supplements only to the level that combined intake (from diet and supplements) is below what is considered risky.

    But mainly I focus on preventing serious falls — not by avoiding anything and everything, but by some avoidance (namely, hire someone if a job requires going above the second step on a stepladder) and mainly targeted exercises: exercises to improve balance (with the attitude that the goal is not to avoid fumbling so much as to be able to recover from a fumble), exercises to keep the ligaments and muscles around joints functional, and exercises to avoid kyphosis (“getting in touch with my traps and lats”).

    So that’s my version of trying to take the various bits of information I am aware of and put them together in a common-sense sort of way.

    I’d be interested in hearing what others do. (Bill?)

    • Paul Alper says:

      Martha: The link you mentioned,
      was really informative even if I have no idea who “Bill” is. The other link you allude to,
      is also interesting reading.
      I have heard from my wife’s physician that T score and Z score are machine dependent and operator dependent. If this is true, then year-to-year comparisons are useless, especially if you change clinics or new machines are installed.
      Further, I am under the impression that the T score definitions have (arbitrarily?) changed since the 1994 date such that more women are now in the category of osteopenia and osteoporosis. Have the “goal posts” moved?

      • Martha says:

        Bill is someone who sometimes follows this blog and who had an osteoporotic fracture a few years ago. So “Bill?” was an invitation for him to comment.

        “I have heard from my wife’s physician that T score and Z score are machine dependent and operator dependent. If this is true, then year-to-year comparisons are useless …”and “I am under the impression that …” I don’t recall hearing either of these, but would not be surprised if both are true.

        Also, I recall reading few years ago that a medical quality group in one of the Canadian provinces (BC, if I remember correctly) published a study (can’t locate the reference at the moment) asserting that a combination of five (if I recall correctly) other factors predicted fracture risk better than bone mineral density.

        • Peter says:

          Paul, you are right: T score definitions have changed since 1994. The old definition was a T score lower than -2.5. The new one is T score lower than -2.0. That change led to an 85% increase in the diagnosis of osteoporosis in the US (6,781,000 more diagnoses). This info comes from ch.2 of H. Gilbert Welch: Overdiagnosed: making people sick in the pursuit of health. 2011

          If you are not familiar with Welch’s work (he is a professor at Dartmouth Medical School), I recommend the video at the cited link (under “Author Updates”) about his latest book Less Medicine, More Health, 2015.
          I have read Overdiagnosed, Less Medicine, More Health, and Know Your Chances and found them to beexcellent books.
          Nortin Hadler (a medicine professor from UNC, Chapel Hill) is also a very interesting writer:

          • Rahul says:

            The mere shifting of a threshold shouldn’t be proof of over-diagnosis, right? The crucial question is what makes for a “right” threshold?

            Did we start overdiagnosing or were we underdiagnosing before? I guess an axiomatic threshold of osteoporosis is impossible unless we have some consideration of risk or limit of useful clinical intervention?

            In a certain sense, as medicine advances, and as society gets more prosperous we might expect more broad definitions of illness?

            • Martha says:

              As I alluded to (but perhaps did not articulate well enough) above, it is not clear that osteoporosis/low bone mineral density is an “illness”. It is just one of many factors that can influence the *risk* of fractures. For example, in a clinical trial for an intervention aimed at reducing fractures, bone density would not rationally be the primary outcome (number of fractures would). Unfortunately, bone density is sometimes taken as a proxy for fracture risk (since some interventions would affect bone density sooner than any effect on fracture rate would occur) — but evidence suggests that it is a poor proxy. And there might be effective interventions for reducing fracture incidence that would not lower bone density.

              • Martha says:

                (See also Peter’s comment below re surrogate outcomes — sorry, I didn’t see it before responding.)

          • Paul Alper says:

            I echo Peter’s recommendation of Gilbert Welch. Below is a fascinating link of Welch explaining his ideas to an audience of lay people and physicians. It is about 99 minutes long, entertaining and well worth the effort. The graphs are compelling.


        • Bill Jefferys says:

          That would be me. Martha knows me well, and Paul visited our place in Vermont once…we went for a nice walk.

          I suffered a fractured vertebra due to a fall about 12 or 13 years ago, and a bone density scan revealed a significant amount of osteoporosis. Probably I should have been more aware of this possibility since my dad also had this condition (which led to a fractured hip at one point).

          I was treated first with an injectable drug, Forteo, which encourages the activity of the osteoblasts (which build bone). We stopped that after two years because of a general recommendation for that drug not to use it for more than two years. The bone density scans showed considerable improvement, which moved me from an osteoporotic condition to an osteopenic condition (according to the readings). At this point I switched to a bisphosphonate, which acts by retarding the action of the osteoclasts (that destroy bone…the maintenance of bone health is a constant interaction between the clasts and the blasts since bone is not a static tissue but is constantly being renewed by the actions of these two cells). I used bisphosphonates for about 5 years but again it’s not recommended that they be used indefinitely and stopped taking this drug. I’ve not used prescription drugs for this condition since.

          All this time I’ve taken daily vitamin D/Calcium supplements as recommended. They are cheap and I guess they can’t hurt in the amounts that I take them. Biennial scans show a stable situation as regards bone density. And no further fractures. Like Martha, I do try to avoid situations that could increase fracture risk.

    • Anoneuoid says:

      I like this chart:

      Is there a name for this, or a package that does it in R?

    • Peter says:

      Martha wrote: “One complexity that seems not to be getting adequate attention in this discussion is that osteoporosis (low bone density) isn’t a very strong predictor of fracture.”

      This is a very important point. In medical jargon low bone density is a surrogate outcome. Surrogate outcomes are biomarkers (like T score or HbA1c in the case of diabetes) or intermediate outcomes that substitutes for patient-important outcomes, that is, outcomes that patients notice and care about such as survival, functional capacity, symptoms and health-related quality of life. Use of surrogate outcomes in drug evaluation does not provide sufficient clarity for understanding the actual benefits and harms for patients taking the drugs. Simply put: Surrogate outcomes are false idols. And we all know better than to pray to false idols.

      See John Yudkin et al.: The idolatry of the surrogate. BMJ 2011; 343 doi:
      Surrogate Outcomes in Clinical Trials: A Cautionary Tale. JAMA INTERN MED/VOL 173 (NO. 8), APR 22, 2013
      (one can find ungated copies of both articles online)

      • This is particularly the case for things like statins. Suppose they really do lower LDL Cholesterol. So what? Naturally occurring low LDL cholesterol is associated with lower heart disease risk… but is statin suppressed LDL Cholesterol also associated with lower heart disease risk? Maybe it is, and maybe it isn’t, but whatever the case, it’s critically important to keep our eye on the actual prize (lower risk of heart disease).

        LDL cholesterol is a perfect example of an intermediate or surrogate outcome, because it’s totally utterly meaningless by itself. People can’t “feel” cholesterol. High LDL cholesterol by itself doesn’t reduce quality of life. It’s only the things that it may cause (arteriosclerosis etc) that matter.

      • Paul Alper says:

        Here is a non-medical example of inappropriate surrogate outcomes/criteria: a dashboard light goes on in your car. One solution is to cover the light with duct tape and thus, problem solved.

        • Don’t use duct tape! Jeez, Duct tape has a rubber-based adhesive and lasts only a few months, leaving a difficult to clean smear after it falls off. The obvious choice here is Electrical tape because it’s opaque black and will block that light really well


      • Jon says:

        The problem is real endpoints are much harder to measure and track. Bone density can be measured at specific points in time. You can put someone on a drug, measure bone density a year later and then be done. One may have to follow people for years to observe bone fractures, as well as long term changes in bone density.

        Furthermore, bone fractures are determined by factors other than bone integrity, i.e. balance and risk taking.

        Surrogates can be useful if there is a plausible mechanism.

  5. Madman says:

    The profits from the sale of calcium and Vitamin D supplements are very marginal and very diverse. That is, calcium and D supplements are quite inexpensive and any vitamin company can sell these. So, the idea that somehow vitamin companies conspired to push the sale of these supplements to boost profits just doesn’t pass the reality check.

  6. Gospace says:

    My mother-in-law has obvious signs of osteoporosis. The hunched back, stooped neck. And my wife has 3 sisters, 5 years older, 3 and 5 years younger. When her 3 year younger sister broke a bone simply exiting a vehicle, she was diagnosed with osteoporosis. Her other two sisters went in for bone density checks, Both diagnosed with osteoporosis. She went in- above, but not by much, the density cutoff for an osteoporosis diagnosis. She’s been taking calcium citrate and vitamin D3 supplements, as well as a multi-vitamin since her mid- 20’s, when we got married. She’s now taking Fosomax as prescribed. Her sisters have never taken any supplements. Anecdotal evidence? Of course it is; I’m just a layman, not qualified to note such things. But, if they all had the same family doctor, and he noted it, it would be published in a medical journal proving the value of early supplementation in families with a history of osteoporosis.

Leave a Reply