Calcium and Vitamin D Barely Moved Fracture Risk in 69 Trials. What Researchers Point to Instead

The review searched the medical literature through February 2025 for randomized trials comparing calcium, vitamin D, or both against a placebo or no treatment, in adults who were not taking drug treatment for osteoporosis. The typical participant was older, with a median age of about 71 across trials, and the vast majority were living independently in the community rather than in care homes. Only about one trial in four focused on people at high risk of fractures or falls, which matters a lot for how the findings apply, and we'll return to it below.

The headline results for any fracture were consistent across all three approaches. According to the published meta-analysis in The BMJ, calcium alone produced a risk ratio of 0.91, vitamin D alone produced a risk ratio of 1.00, and the combination produced a risk ratio of 0.91. A risk ratio of 1.00 means no difference at all between groups, while 0.91 means roughly 9% fewer fractures in relative terms. The same pattern held for hip fractures and for the chance of having a fall.

A quick word on the last column, because it does a lot of work in this story. GRADE is the system reviewers use to rate how confident they are that an estimate is close to the truth. "High" certainty means further research is very unlikely to change the conclusion much. "Very low" means the estimate is shaky and could easily move. For vitamin D and for the combination, the fracture evidence carries the highest rating the system gives.

The authors also tested whether the results changed for particular groups, including by age, sex, a history of fractures or falls, and how much calcium people already got from food. They found no credible subgroup effects for any of the three strategies. In other words, they didn't find a hidden group of people in these trials for whom supplements suddenly worked much better.

Look closely at the combined calcium and vitamin D rows and you'll notice something that most headlines skipped. The confidence interval for any fracture (0.84 to 0.99) and for hip fracture (0.74 to 0.96) sits entirely below 1.00. In statistical terms, those results are significant. So why did the authors still describe the effect as "little to no"?

The answer lies in a decision they made before looking at the results. The team set a minimal important difference for each outcome, a threshold for how large an absolute benefit needs to be before it's worth caring about in real life. They set that bar at a 2% absolute reduction for any fracture, 0.7% for hip fracture, and 3% for the risk of falling. Translated into groups of people, the bar was 20 fewer fractures per 1,000 people, 7 fewer hip fractures per 1,000, and 30 fewer people falling per 1,000. The combined supplement effects came in under those lines.

It helps to picture it like a bathroom scale that reliably reads a quarter-pound lighter after a change. The difference is real and repeatable, but it isn't the kind of change most people would notice or plan around. The authors were candid that these thresholds are a judgment call: they note they did not involve patients in setting them. A reader who would value a smaller benefit might weigh the same numbers differently, which is exactly why seeing the thresholds matters.

The review's discussion also explains why this question has produced such mixed headlines over the years. Earlier meta-analyses of combined supplements did report significant effects, but the absolute benefits worked out to about 1% fewer fractures at any site and 0.3% fewer hip fractures, and much of that came from a single trial in a very high-risk population. Remove that one trial, and the benefit in the general population looks even smaller.

One more detail is worth mentioning plainly. Vitamin D on its own showed a hip fracture risk ratio of 1.13, with an interval running from 1.00 to 1.27. That sits right at the edge of statistical significance, in the unfavorable direction. The authors didn't treat it as evidence of harm, and neither should a single number in a large table, but it's one reason the review offers no support for vitamin D alone as a fracture strategy.

When the BMJ Group's release summarized the study, it described the authors as suggesting a shift in focus toward balance and resistance exercise, along with approaches that pair exercise with home hazard checks or education tailored to a person's risk. Inside the paper, the authors put it more carefully: apart from exercise and osteoporosis drug treatment, few interventions have moderate or high certainty evidence of consistently reducing fractures.

So what does that exercise evidence look like when you hold it to the same standard? The largest synthesis is a Cochrane review of exercise for preventing falls that pooled 108 randomized trials with 23,407 older adults living in the community across 25 countries. Its core findings carry the same top certainty rating as the best supplement evidence. Exercise reduced the rate of falls by 23%, about 195 fewer falls per 1,000 people each year, and reduced the number of people who fell at least once by 15%, about 72 fewer people falling per 1,000.

Balance training carries the strongest signal

Not all exercise performed equally, and this is where the details matter for anyone picturing a gym routine. Programs built around balance and functional movement, like controlled stepping, standing on one leg, and getting up and down from a chair, reduced the rate of falls by 24%, backed by 39 studies and high-certainty evidence. Programs that combined several types of exercise, most often balance work plus resistance training, probably reduced falls by about a third, with moderate certainty. Tai Chi may reduce falls by around a fifth, though with lower certainty.

The finding that surprises many readers is this one: for programs that mainly involved resistance training on its own, as well as dance or walking, the reviewers were uncertain of the effect on falls. The abridged version of the Cochrane review in the British Journal of Sports Medicine spells out that uncertainty. Strength work has plenty of other well-documented benefits, but for falls specifically, the evidence points most clearly to balance, often paired with strength, rather than strength alone.

If you're curious how your own balance compares with others your age, our breakdown of single-leg balance by age walks through the reference data and a simple self-test.

Here is the comparison at the heart of this analysis. It lines up the supplement evidence and the exercise evidence on the same four measures: how many trials, how many people, what the effect was, and how confident researchers are in it.

One important caveat before reading it: no trials have compared supplements and exercise head to head. These numbers come from separate bodies of research, with different participants, different follow-up periods, and slightly different outcome definitions. The BMJ review counted any fracture, for example, while the Cochrane review counted fractures caused by falls. The scorecard shows how strong each body of evidence is on its own terms. It isn't a race run on the same track.

A second line of fracture evidence supports the exercise column. A 2022 meta-analysis of 20 exercise trials, published in the Journal of Bone and Mineral Research, found about a third fewer overall fractures and major osteoporotic fractures in exercise groups. That analysis measured person-years of follow-up rather than individual participants, and it didn't report a GRADE certainty rating, so it strengthens the direction of the exercise case more than its precision.

What the numbers look like per 1,000 people

Relative percentages can feel abstract, so here's a translation. Falls have a solid US baseline: according to the US Preventive Services Task Force, 27.5% of adults 65 and older living in the community report at least one fall in a year. Applying each risk ratio to that same starting point gives a rough sense of scale. These are estimates built on a stated assumption, not measured results from any single trial.

Estimates assume a baseline of 275 people falling per 1,000 each year, taken from USPSTF data. Ranges reflect each study's 95% confidence interval.

For fractures, there's no single defensible baseline to apply, because fracture risk varies enormously by age, bone density, and history. A cleaner way to picture it: for every 100 fractures that would otherwise happen, the combined supplement estimate suggests about 9 fewer, and the exercise estimate for fall-related fractures suggests about 27 fewer. The first figure is high certainty but small. The second is larger but rests on low-certainty evidence from far fewer people.

It would be easy to read the scorecard as a knockout, and that wouldn't be fair to the evidence. The exercise research is strong for falls, but falls and fractures are different outcomes. Most falls don't break a bone, and the fracture evidence for exercise comes from 10 studies and about 4,000 people, rated low certainty. That's a promising signal, not a settled answer.

The USPSTF's own review of exercise adds another honest wrinkle. When the Task Force assessed fall-prevention interventions in 2024, its pooled estimate for exercise lowering the chance of falling at all was a risk ratio of 0.92, numerically identical to combined supplements in the BMJ review. What sets exercise apart there is that its confidence interval stays clear of 1.00 and that it also reduced how often people fell and how often falls caused injuries, with an injurious-fall ratio of 0.84. The Task Force gave exercise a grade B recommendation for adults 65 and older, based on 37 trials with more than 16,000 participants.

Two more practical details are worth keeping in view. First, the programs in these trials were usually structured and sustained, commonly two to three sessions a week for a year, built around gait, balance, and functional training. Second, supervision seems to matter: in the 2022 fracture analysis, predominantly supervised programs showed roughly twice the benefit of unsupervised ones, whose results were not statistically significant. The Cochrane reviewers also noted that side effects were rarely reported and usually minor, such as joint or muscle soreness, though one trial recorded a pelvic stress fracture. The lesson isn't that any movement equals prevention. It's that specific, well-designed programs are what the research actually tested.

For a sense of the lower-body strength these programs build, our guide to the 30-second chair stand explains the test and the reference ranges by age.

This is the most important section for anyone who takes calcium or vitamin D on a clinician's advice. The review's conclusions are about generally eligible adults, and the authors drew clear lines around them.

Several other limits shape how far the findings travel. Only 19 of the 69 trials focused on people at high risk of fractures or falls, and there was too little data on people living in residential care to draw firm conclusions for that group. The authors also note that many trials used low vitamin D doses and enrolled participants who weren't at high risk, often with normal vitamin D levels to begin with. That describes a large share of adults who take these supplements, but not everyone.

The review also excluded trials where exercise or dietary instructions were given to only one group. That design choice kept the supplement results clean, and it's also why the exercise comparison has to come from separate research.

The review lands in a US policy picture that's been moving in the same direction for several years. The Task Force's current final recommendation on this question dates from 2018. It recommended against low-dose daily vitamin D and calcium for preventing fractures in postmenopausal women living in the community and found the evidence insufficient for other groups and higher doses.

A December 2024 draft from the Task Force proposed going further, recommending against vitamin D, with or without calcium, to prevent fractures and against vitamin D to prevent falls in postmenopausal women and in men 60 and older. As of September 2026, that draft is still listed as in progress rather than final, so it's best understood as a proposal. Like the BMJ review, both the 2018 statement and the draft exclude people with a history of osteoporotic fracture, osteoporosis, or diagnosed vitamin D deficiency.

The broader nutrition question, how much calcium and vitamin D people need from their overall diet at different life stages, is separate from whether supplements prevent fractures. Our look at calcium and vitamin D needs by decade covers that side of the story.

Stepping back, the review doesn't say calcium and vitamin D are unimportant nutrients, and it doesn't say supplements are harmful. It says that, across a very large body of well-run trials in mostly healthy older adults, taking them as pills didn't reduce fractures or falls by an amount the researchers considered meaningful. The certainty behind that conclusion is about as strong as nutrition research gets.

The practice side of the ledger tells a richer, more hopeful story, with its own honest limits. Balance-focused exercise has high-certainty evidence for reducing falls, and structured programs that combine balance with strength appear to do the most. Whether that translates into fewer broken bones is more likely than not, but still less certain than many headlines suggest.

This is a pattern we see often in health research: as studies get larger and better designed, effects tend to shrink toward their true size. Our analysis of two decades of hormone therapy and Alzheimer's research traces the same arc in a different field.

For anyone weighing their own situation, especially people with a bone condition, a history of fractures, or a current prescription, the most useful next conversation is with a clinician who knows their full picture. What this review adds to that conversation is clarity: a clearer sense of what the evidence says about the pill, and what it says, and doesn't yet say, about the practice.