Do calcium supplements cause heart attacks?

Calcium is good for us, right? Milk products are great sources of calcium, and we’re told to emphasize milk products in our diets. Don’t (or can’t) eat enough dairy? Calcium supplements are very popular, especially among women seeking to minimize their risk of osteoporosis. Osteoporosis prevention and treatment guidelines recommend calcium and vitamin D as an important measure in preserving bone density and reducing the risk of fractures. For those who don’t like dairy products, even products like orange juice and Vitamin Water are fortified with calcium. The general perception seemed to be that calcium consumption was a good thing – the more, the better. Until recently.

In a pattern similar to that I described with folic acid, there’s new safety signals from trials with calcium supplements that are raising concerns. Two studies published in the past two years suggest that calcium supplements are associated with an significantly increased risk of heart attacks. Could the risks of calcium supplements outweigh any benefits they offer?

Why Calcium? Osteoporosis

Osteoporosis is a progressive bone condition of reduced bone mass and deterioration of bone tissue, and a correlating increase in fracture risk. 80% of those diagnosed are women. Hips and spines are the most common fracture locations, but they can appear in any bone, and osteoporosis makes fractures more likely. In postmenopausal women over the age of 50, the lifetime risk of a vertebral fracture is about one in three, and one in five for a hip fracture. Because they are so common, hip and vertebral fractures cause considerable aggregate and individual morbidity and mortality. So prevention and treatment are major health issues.

The initial strategy to preventing and managing osteoporosis is ensuring adequate calcium and vitamin D dietary intake, as both influence bone density. Calcium intake influences overall calcium balance: adequate vitamin D and calcium ensure calcium balance is positive. This occurs at about 1000mg per day in premenopausal women, and 1500mg per day in postmenopausal women not taking estrogen. The North American Menopause Society’s (NAMS) 2006 osteoporosis guidelines recommends [PDF] adequate calcium and vitamin D for all postmenopausal women, regardless of osteoporosis risk factors. The guidelines note that requirements increase with age owing to reduced absorption, and recommending adequate intake (preferably via diet) as the preferred sources. The 2010 Canadian guidelines [PDF] are similar, recommending 1200mg of calcium (diet and supplements) and vitamin D for all individuals over the age of 50. The Institute of Medicine recently updated its calcium and vitamin D guidelines (pdf) as well. It concluded with the caution that the consumption of levels beyond those recommended have not been shown to offer additional health benefits, and may in fact be linked to other health problems.

The effectiveness of calcium and vitamin D for the prevention and treatment of osteoporosis has been studied in both observational and prospective clinical trials. Wile there are data to demonstrate that calcium and vitamin D can prevent bone loss, the data on fracture prevention are much less convincing, with some trials showing no effect. Beyond density effects, calcium is also associated with generally positive effects on muscle strength, balance, and the risk of falls. So for most men and women with (or at risk of) osteoporosis, calcium and vitamin D are standard treatments. Given dietary intake in those at greater risk of osteoporosis may be below recommended levels, supplements are often used to meet recommended amounts.

The Safety Signals

Prior studies of calcium supplements have pointed to a possible relationship between calcium supplementation and cardiovascular events. Bolland et al specifically examined the relationship of calcium with the risk of heart attacks and cardiovascular events in a 2010 BMJ meta-analysis. It included all RCTs of calcium supplements (≥500 mg/day), with a study size of 100 or more participants, an average age over 40, and a duration of more than one year. Trials that included vitamin D as an intervention were excluded. 15 trials were identified: some with patient-level data, and some with trial level data. Analyses of both sets of data identified a significant increase in heart attacks in those randomized to calcium supplements. The trial-level analysis show a hazard ratio (pdf) of 1.27 with a 95% confidence interval of 1.01 to 1.59 (p=0.038). The patient level analysis revealed a similar hazard ratio for myocardial infarction of 1.31 (95% confidence interval 1.02 to 1.67, p=0.035). Overall, the analysis suggests that calcium supplements increase the relative risk of myocardial infarction by about 30%. Reassuringly, there were no statistically significant increases in the risk of stroke, death, or the composite endpoint of MI+stroke+death in either analysis. Based on the patient-level data, the authors estimated that treating 69 people with calcium for five years will cause one additional heart attack. The authors suggested that in light of calcium’s unimpressive efficacy against fractures, that calcium’s role in osteoporosis prevention and treatment should be reevaluated.

Time to stop the calcium? As noted above, the data to support the use of calcium supplements alone to prevent fractures are, on balance, unimpressive. And there are possible models for how calcium could be causing these harms: vascular calcification is a potential (though not proven) consequence that might be more likely in the elderly patients. However, given calcification can take years, and harms appear shortly after dosing starts, it could be a due to effects on carotid plaque thickness, leading to aortic calcificiation, and subsequent cardiovascular events. (Reid describes potential mechanisms for these harms in a2010 paper in Clinical Endocrinology.)

What happened after this paper was released? There were criticisms of the endpoints, and the fact the composite endpoint was not significant. Concerns were also raised that the trials included were not designed with cardiovascular endpoints – a valid criticism. And many pointed to the fact the studies excluded vitamin D, contrary to treatment guidelines and common use. Now the same group has done a new analysis, incorporating vitamin D. Bolland and associates followed up their calcium-only therapy with a study of calcium + vitamin D. They used the Women’s Health Initiative (WHI) dataset to answer the vitamin D question, added in some other studies, and redid their meta-analysis.

The WHI was a massive 15-year trial of over 161,000 women that sought to answer a number of questions about women’s health. The most well known components were the hormone therapy trials which changed our understanding of the risks and benefits of hormone treatments. The calcium and vitamin D study was a component of the WHI which randomized 36,282 postmenopausal women aged 50-79 into two groups. One group received 1,000 mg of calcium carbonate and 400 UI of vitamin D once daily, the other, placebos. Interesting in the design was that 54% of women were already taking calcium, and 47% were already taking vitamin D, and they were allowed to continue with their therapy, even after randomization. This meant that actual calcium and vitamin D doses women consumed varied from zero to substantially more than the intervention dose. The clinical question the study sought to answer was to understand the effects on fracture risk and the prevention of colorectal cancer — and the results were disappointing: no effects on colorectal cancer, and insignificant effects on fractures (though in a subgroup analysis of compliant patients, significant reductions in hip fractures were noted.)

Bolland sought to analyze the WHI data for cardiovascular effects, and then add these data into the previous meta-analysis. In the over 16,000 women not taking their own calcium and vitamin D, there was a significant increase (hazard ratio 1.22) in myocardial infarction noted in the group randomized to calcium and vitamin D (p=0.04, 95% CI 1.00 to 1.50). Similarly, significant effects were also noted in other composite endpoints. In contrast, women taking their own calcium and vitamin D didn’t show any changes in their cardiovascular risk when randomized to calcium and vitamin D. In addition, no relationship was found between calcium dose and risk of cardiovascular events.

The authors then pooled their own WHI analysis with two other studies of calcium and vitamin D where trial-level data for cardiovascular events were available: In total, over 20,000 participants could be studied. In this pooled analysis, calcium and vitamin D were associated with a significant increases in myocardial infarction (relative risk 1.21), stroke (RR 1.20) and a composite endpoint of both (RR 1.16).

Finally the authors combined the trial level data from their calcium-only meta-analysis with their trial level calcium plus vitamin D data:, resulting in a pool of over 28,000 patients across nine trials. In this analysis, there was risk increase of 1.24 (95% confidence interval 1.07-1.45, P=0.004) for myocardial infarction and 1.15 for the combined endpoint (1.03-1.27, P=0.009).

Difficult to interpret? Yep. The lack of effect of “personal” use of calcium on endpoints, and the lack of dose response, means this isn’t case closed for the clinical question. But the persistent and significant correlation between randomization to calcium, with or without vitamin D, and myocardial infarction, does concern me. There are a number of additional criticisms outlined in the editorial that accompanied the Bolland WHI analysis, and the keen reader is referred there for more.


Is it possible that calcium supplements can be causing harms that could outweigh their benefits? Yes, but the evidence isn’t clear enough to give an definitive answer. These data need to be factored into individual evaluations of diet as well as risk factors for cardiovascular disease and osteoporosis. I’d like to see these findings validated by other groups, as both meta-analyses came from the same group of researchers. The meta-analysis can be a very useful tool, but it’s not without its own limitations, as is often pointed out by the contributors to this blog. Interestingly, a 2010 meta-analysis, from a different group of authors, and using a different methodology, has come to a different evaluation of calcium. So the question remains an open one. More data may help refine our estimates of number needed to treat, and number needed to harm, to inform treatment decisions. And it should help guide advice for younger, premenopausal women, as well as men. So until more data emerges, my tentative recommendations to consumers are as follows:

  • Calcium supplementation has been associated with increased risks of cardiovascular events like heart attacks. Until there is more evidence to confirm or refute this association, it’s prudent to be cautious when taking calcium supplements.
  • No harms have been shown from calcium consumption via dietary sources. Efforts should be made to first meet dietary requirements through food products, before considering supplements.
  • Routine supplementation, in the absence of a dietary deficiency, is not necessary or advisable.
  • Calcium supplements may still be advisable for those with low dietary intakes, or those at risk of or being treated for osteoporosis. The risk-benefit assessment for calcium supplements needs to consider risk factors for both osteoporosis and for cardiovascular disease.
  • Vitamin D supplements are advisable for most people, and are recommended for the prevention and treatment of osteoporosis. The suggested doses of calcium and vitamin D may vary based on diet, medical conditions, and other considerations. Sources for target doses could include the IOM or recent osteoporosis guidelines (Canada) (USA).


The emerging safety data on calcium may yet become another cautionary tale about the unexpected and undesirable outcomes of targeted supplements. Until more evidence emerges, the safety of calcium supplements will continue to be questioned and debated. But that’s science-based practice: Data can be conflicting, messy, and difficult to interpret. There is always the possibility of unintended consequences when we make therapeutic decisions, and only by rigorously evaluating what we’re doing can we continue to improve the way we prevent and treat disease.

Bolland MJ, Avenell A, Baron JA, Grey A, MacLennan GS, Gamble GD, & Reid IR (2010). Effect of calcium supplements on risk of myocardial infarction and cardiovascular events: meta-analysis. BMJ (Clinical research ed.), 341 PMID: 20671013

Bolland, M., Grey, A., Avenell, A., Gamble, G., & Reid, I. (2011). Calcium supplements with or without vitamin D and risk of cardiovascular events: reanalysis of the Women’s Health Initiative limited access dataset and meta-analysis BMJ, 342 (apr19 1) DOI: 10.1136/bmj.d2040

Posted in: Herbs & Supplements, Science and Medicine

Leave a Comment (19) ↓

19 thoughts on “Do calcium supplements cause heart attacks?

  1. majkinetor says:

    This is probably result of missing Vitamin K2 in the supplementation scheme (probably also the reason why dietary calcium didn’t show any harm, dairy products like butter are among highest sources of this vitamin).

    Vitamin D should probably be higher also, due to the age of participants and/or disease state which probably translates to sedentary home life.

    Also hypothesis that calcium alone can reduce osteporosis is suspicious or even wrong according to evidence by numerous epidemiological studies.

  2. Robin says:

    Thanks for this. It seems, one by one (vitamin C, vitamin E, folate, and now calcium) the benefit of supplements is of less-than-previously-thought significance and they’re best, and most safely, acquired through food.

    Wile there are data to demonstrate that calcium and vitamin D can prevent bone loss, the data on fracture prevention are much less convincing, with some trials showing no effect.

    though in a subgroup analysis of compliant patients, significant reductions in hip fractures were noted

    I remember when the WHI released their data on low fat diets, and the point was made that compliance and attenuation and a reliance on self-reportage are frustrating and serious weaknesses of these large, multi-year trials. But, that study did turn up a lot of data on the harmfulness of trans-fats, so, who knows?

    Further, regarding cardiovascular risk, from BMJ:

    To complicate the interpretation further, a recent meta-analysis from the group showed that the increased risk of cardiovascular events with calcium supplements was present only in subjects with baseline dietary calcium intake above the median

    So, it’s possible that cardiovascular risk is not from the supplements themselves but overall total daily intake?

    Confusing, to say the least.

  3. Dpeabody says:

    “No harms have been shown from calcium consumption via dietary sources. Efforts should be made to first meet dietary requirements through food products, before considering supplements.”

    Is this because people do not usually attain the same level of calcium intake via diet as they do from supplements. If not, what would be your best guess at some of the complicating factors that could lead to a difference in results?

    Personally I love milk and drink around 6 glasses a day. Not due to any of the reported health benefits, I just love the taste.

  4. majkinetor says:

    “Thanks for this. It seems, one by one (vitamin C, vitamin E, folate, and now calcium) ”

    It doesn’t seem that way to me. Even contrary. Its more and more recognised that RDI levels are not adequate [for most people]. There are different forms of supplements and different qualities. For instance, Vitamin C is best in synthetic form (and liposome technology due to short half life, glucose competition, low absorption due to gut environment, and potential insulin resistance). Vitamin E is the best in natural form, and synthetic form might even be toxic. Also, you must understand that most of the trials out there used low or minuscule amounts of vitamins (especially if we observe the fact that majority is done on sick people and vitamin levels are impaired in those folks). For instance, latest hypothesis that is being evaluated for C is “Dynamic flow” where C is taken every 3 hours or so due

    Vitamin K2 directs storage of calcium where it should be and prevents its accumulation where it shouldn’t be – soft tissues. Overcalcification and impairment of calcium homeostasis is one of the markers of aging.

    You probably shouldn’t drink that much milk, I would check calcification or arteries just to be sure. Fermented milk products (or goat milk) is IMO better choice. Since you already drink that amount of milk, you should probably supplement Magnesium (in the form of Citrate) or eat leafy vegetables.

  5. JPZ says:

    Thank you for a very nice review that pointed out the strengths and weaknesses of many of the study results and conclusions. Given the inconsistency of the findings and the heavy reliance on meta-analysis, I suppose I would favor a “wait and see” approach, but your recommendations are largely prudent. I would add two points about special subgroups: 1) young women (9-18 yrs) as a group have quite low calcium intakes and 2) lactose intolerance can drive low dietary intakes of calcium. Patients in these categories should receive special attention regarding calcium intake. Also, one preventive approach not addressed in many therapeutic calcium studies is optimizing peak bone mass for woman by the time they reach their 20’s. This may be a more effective intervention period for prudent calcium + vitamin D supplementation.

    Over-supplementation in the absence of science-based need (e.g. benefits of high dose vitamin A in retinitis pigmentosa) is an unfortunate result of “more is better” thinking. Once again, the U-shaped benefit curve brings us back to the middle ground for calcium.

  6. LovleAnjel says:


    Why is fermented milk better? Does it contain less calcium?

  7. cervantes says:

    Off topic, but I always knew Marc Siegel was a flaming, cardboard — well, you know. His publicist just sent me this:

    Contact: Katy Reddin, 212-999-5585,

    THE INNER PULSE: Unlocking the Secret Code of Sickness and Health

    – Marc Siegel, MD Available For Immediate Interview

    Many doctors view the seemingly inexplicable tragedies and extraordinary recoveries that happen in hospitals every day simply as aberrations from the medical norm. In this remarkable new book, doctor and leading media commentator Marc Siegel explores how the mind affects illness, both positively and negatively. In his just released book, THE INNER PULSE: Unlocking the Secret Code of Sickness and Health, Fox News senior Medical Correspondent Dr. Marc Siegel, drawing on his decades of experience treating patients, explores the sometimes miraculous effects that mind and spirit can have on disease and healing. The inner pulse is the essence that links the soul to the brain and body, a marker that predicts whether a person’s life force is fading or strengthening. THE INNER PULSE shows readers how they may tap into this pulse and in, some cases, even influence it.


    • Explores how your inner pulse can alert you to what is going on in your body

    • Offers a new perspective on the positive and negative effects of the mind on illness and healing

    • Includes dramatic case stories of Dr. Siegel’s work with his own patients—those who have healed and those who have not—as well as his own story

    Exploring the uncanny world where a patient’s expectation and sometimes even an outcome can be driven by personal intuition, THE INNER PULSE will give you a deeper understanding of how the mind relates to disease and how the mind and the body, working in sync, can heal.


    About the author:

    Marc Siegel, MD, (New York, NY) is a practicing internist, professor of medicine at NYU Langone Medical Center, and Medical Director of Doctor Radio at NYU and Sirius/XM Satellite Radio. He is a senior Medical Correspondent for Fox News. He is the author of several other books, including False Alarm.

  8. mark says:

    @ Scott

    Great post.

    One of the problems assessing adverse effects in clinical trials is that the trial is only powered to detect the treatment benefit and is almost always massively underpowered to detect adverse effects. Probably the only reason the cardiovascular risks of calcium were seen is that risks are about the same (or even greater) than the benefits on fracture reductions. In practical terms, it means that potential risks of an agent can almost always be dismissed as not statistically significant, even when large meta-analyses are carried out (eg referring to a 10-15% increased risk of stroke with calcium that was not significant as reassuring).

    Given that the issue of harm has been raised with calcium, it is difficult to see any future trials of calcium being carried out (for ethical, funding, and recruitment reasons). The BMJ editorialists suggest waiting for more data, but that seems unlikely, so we probably have to make decisions on these current data, however imperfect.

  9. JPZ says:


    Is suspicion of harm the same as evidence of harm? I thought this was SBM, and we could accept the limitations of meta-analysis. If you have to blend a n=36,282 poorly controlled WHI calcium study (but much better HRT study) with other studies to confirm what the same group already found in a previous meta-analysis, then you should question the science. Apply the same standards that you would apply to some study you don’t like.

    I think Scott has been prudent to advise caution, but for you to consider further studies unethical is akin to CTM practitioners claiming traditional use is enough and further studies are too “Western.” I apologize if that comparison seems heavy handed considering the forum, but I think you are over-reacting.

  10. mark says:


    ‘Is suspicion of harm the same as evidence of harm?’

    No, but should you have a lower threshold for suspecting harms than you would for attributing benefits? esp significant harms like heart attacks.

    Could you elaborate on what is wrong with the science here? They previously pooled data from trials that individually didn’t have power to address an issue and showed a consistent increase in heart attack with calcium across the trials. Now they have confirmed the same risk in another trial, and added that data to their previous analyses.

    “For you to consider further studies unethical…”

    I didn’t say they were unethical, but that there were ethical issues. Would you be able to convince an ethics committee to do a RCT of calcium when the primary outcome is one of harm? (ie research has shown a 25% increase in risk of heart attack with calcium, we want to confirm this in our trial) I wouldn’t think you could, nor convince someone to fund such a trial, nor convince people to enter such a trial.


  11. gippgig says:

    Instead of calcium supplementation, has anyone looked for correlations between calcium level in the blood and osteoporosis/cardiovascular events/everything else?

  12. majkinetor says:

    2 LovleAnjelon

    Some dairy products will contain less calcium then milk, simply because part of it is lost in whey. How much is lost depends on the method used to ferment milk.

    The other important reason is that there is more and more evidence that milk might not be that good for subset of humans (lactose intolerance, casein insensitivity and other alergies, etc…). Heavy industrialization of milk is another problem – BGH, UHT, homogenization, fat removal (hence removal of lipid soluable vitamins like A and D) and fortification with inorganic calcium and other vitamins/minerals with suspicious quality.

    Goat milk is less industrialized and has better content for human consumption then cows milk. For instance, Goat’s milk contains trace amounts of an allergenic casein protein, alpha-S1, found in cow’s milk and is more similar to human milk.

  13. JPZ says:

    @ Mark

    Scott covered the main points about the science, but I can highlight my own concerns. The Reid group is the only one to find this MI result in a meta-analysis, and the Sesso group did not replicate it in their meta-analysis. The WHI study did not find the MI result with n=36,282, but Reid did find it by dropping out almost half of the subjects. The re-analysis is a secondary one that can no longer claim randomization nor stratification and may have introduced biases. I am not saying that the Reid lab has come to an incorrect conclusion, but I am saying that the evidence is weak without other data.

    Clinical trials can be conducted with a primary efficacy endpoint and a primary safety endpoint. The next major study of calcium and some beneficial outcome can specifically address what the study’s power is to detect an increased risk of MI (and perhaps use more specific tests). The FDA put black box warnings about (then) unproven MI concerns on HRT well before WHI was run, and the subjects would have been advised of this in the consent process. Somehow they recruited over 161,000 women.

  14. Artour says:

    It is an established medical fact that heart patients have heavy breathing at rest, when in stable conditions, with minute ventilation rates of about 12-16 L/min, instead of 6 L/min (the medical norm):

    Chronic hyperventilation reduces oxygen levels in the heart muscle tissue, and tissue hypoxia in the heart causes angina pain. Hence, there is nothing difficult with heart disease. What about calcium effects?

    Several studies found that coronary artery spasms and heart attacks are readily provoked by the hyperventilation provocation test.

    The key mechanism of coronary spasms and tissue hypoxia (leading to heart attacks) involves vasoconstriction of arteries and arterioles due to effects of hypocapnia (low CO2 in the arterial blood). CO2 influences the tone of smooth muscles through calcium ions, and this can help to find the cause of heart attacks due to Ca supplementation.

    When I teach breathing retraining, the students start to slow down their breathing and Ca-Mg-Zn supplements have them to restore normal breathing parameters faster. Hence, with breathing retraining it is safe to use calcium. How much? I apply 3-day test in order to find nutritional deficiencies where changes in breathing is the guiding factor to decide if one is deficient and later to use the test to find the optimum Ca daily dose:

  15. mark says:


    You say the Sasso meta-analysis got different results to the ones from Reid. Sasso found a relative risk of cardiovascular events in 3 trials of 1.14 (CI 0.92 to 1.41), concluding ‘calcium supplements seem to have minimal cardiovascular effects’. In the calcium only meta-analysis by Reid, the risk for the composite endpoint was 1.18 (1.00 to 1.39) in the patient level analysis in 5 trials, and 1.12 (0.97-1.30) in the trial level analysis in 11 trials, which Scott described as reassuringly not statistically significant. In the calcium/vitamin D analysis by Reid, the result was 1.15 (1.03 to 1.27) in 13 trials. The estimates of risk don’t change materially, but the precision of the estimate increases as more trials are included. As I said, clinical trials are almost always underpowered to assess risk, so potential adverse effects can almost always be dismissed as not statistically signficiant, even in large meta-analyses.

    I re-read Scott’s post and don’t see concerns about the science. He says it’s complicated but that is not the same thing. They did a subgroup analysis which you refer to as dropping out half the subjects to find an effect, which seems a stretch. Sure you disagree with the interpretation of the results, but does that really mean it is bad science?

    The WHI trials started about 1992, when it was thought HRT prevented cardiovascular disease and lots of other things. I don’t think there was a black box warning well before the trials started as you say.

  16. JPZ says:


    I don’t think the Reid group is doing bad science. I think they are doing meta-analyses, and, like all meta-analyses, the data should be interpreted with caution keeping in mind the limitations of the technique. The fact that the Sasso group did not replicate their finding just adds more caution to the interpretation – it does not invalidate Reid’s results at all. I have not seen enough data to form an opinion about the results other than “wait and see.”

    Also, thanks for pointing out the black box warning slip. My bad. I was thinking about the endometrial cancer boxed warning back in 1975.

  17. mark says:

    Fair enough.

    Scott’s post reminded me of the vioxx and rosiglitasone MI stories. It struck me that whenever the possibility of adverse effects are raised, often the first defence/criticism is that the results aren’t statistically significant or have used meta-analysis and either way can’t be relied on. But unless trials are designed differently to current trials with adequate power to detect adverse effects (ie are enormous and consequently very expensive) we don’t have any other options. So, we’re stuck with less than ideal data to argue about the meaning and clinical relevance of.

    1. Scott Gavura says:

      I agree that new prospective data are unlikely to emerge for a host of reasons, including ethical issues. Reid made the following comment:

      Based on our own trial data, a placebo-controlled trial of 13,500 women over 5 years would have 80% power to detect a 20% difference in the composite end-point of myocardial infarction, stroke or sudden death. Such a study may well address the remaining clinical uncertainties, but would be both unethical and impractical, because the primary hypothesis is one of harm.

      A prospective trial with that endpoint and of that size is unlikely. So we’re faced with working with what we’ve got. I could have done a much deeper dive into both the problems with meta-analyses and subgroup analysis, as well as the study power issue Mark identifies, but that would have pushed the post another 1000 words or more, without changing my bottom line: Given the efficacy data are so unimpressive, the safety signals, while not definitive, push the risk-benefit assessment into the questionable zone for many people.

      I appreciate the feedback and comments on the post. Thanks.

Comments are closed.