Bone is not scaffolding. It is living tissue that is dismantled and rebuilt continuously, and what decides the balance is not mainly what you feed it — it is what you ask it to carry.
1. Why astronauts are the useful experiment
In microgravity, astronauts lose bone from the hip and spine at roughly 1 to 1.5% per month. A post-menopausal woman on Earth typically loses in the region of 1 to 2% per year. So spaceflight compresses about a decade of bone loss into a year, in healthy people in their forties, under continuous measurement and with diet controlled to the gram.
And they were fed calcium, and it did not help. Nutritional supply was never the limiting factor: the skeleton was being dismantled because nothing was asking it to stay. That is the finding that transfers to Earth, and it explains an otherwise puzzling body of evidence.
2. The supply-side failure — what calcium supplements do and do not do
| Claim | What the evidence shows |
|---|---|
| Calcium supplements prevent fractures | They do not, on the best available synthesis. A 2017 meta-analysis in JAMA pooling 33 randomised trials and 51,145 community-dwelling adults over 50 found no association between calcium, vitamin D, or the two combined and the incidence of hip fracture, vertebral fracture, non-vertebral fracture, or fracture of any kind |
| They increase bone density, so they must help | They do raise density slightly. Density is a surrogate, and this is one of the places where a small change in it does not deliver the outcome — the same problem set out in how to read a trial |
| They are harmless anyway | Probably not entirely. Work led by Bolland and Reid reported an excess of myocardial infarction among older women taking calcium supplements. That finding is contested — other analyses have not reproduced it and the mechanism is debated — but it is enough that a supplement with no fracture benefit should not be taken on the assumption that it costs nothing |
| Dietary calcium is different | Yes, and this is the practical conclusion. Calcium from food is not associated with the same cardiovascular signal, arrives more slowly and in smaller boluses, and comes with everything else in the food. Meeting the requirement through diet is the sensible default |
Three groups where supplementation is a different question and this section does not apply: people who cannot meet the requirement from diet, people in residential care — where the original trials that showed benefit were done, in an institutionalised, deficient, sun-deprived population — and anyone starting a bone-protecting drug, since those drugs are generally trialled and given alongside calcium and vitamin D replacement. If a clinician has advised it for a specific reason, that reason is the answer, not this page.
Vitamin D, separately
The case for vitamin D is about deficiency rather than supplementation on top of sufficiency. UK guidance is that adults consider 10 µg (400 IU) daily through autumn and winter, and year-round for people with little sun exposure, darker skin, or who are housebound. Doses around 800 IU are commonly used where deficiency is likely or bone health is the specific concern. Very large intermittent doses have performed badly in trials, including increasing falls in some, so more is not better. See the sun section, which makes the related point that vitamin D looks more like a marker of sun exposure than the mechanism behind its benefits.
3. What actually worked in space, and the threshold that mattered
Astronauts always exercised. The question was what kind.
| Countermeasure | Result |
|---|---|
| Cardiovascular exercise — treadmill, cycle | Did not prevent bone loss |
| The first resistance device, limited to around 300 lb | Did not prevent bone loss either. Exercise was happening; the loads were not high enough |
| ARED — the Advanced Resistive Exercise Device, up to around 600 lb | Bone was preserved in a way it had not been before. Heavy squats and deadlifts, essentially |
Why this is the most useful thing in the chapter
- Same people, same food, same calcium, same duration. The only variable that changed was the magnitude of the load — and the outcome changed with it.
- The dose–response has a threshold, and it is high. "Exercise" did not work. "Weight-bearing exercise" did not work. Loads approaching what a person can barely move worked. That is a much more specific instruction than the one usually given.
- Bone responds to strain magnitude and rate, not to time spent. Walking more does not fix this, which is worth saying plainly because walking more is the advice people usually receive.
- It also explains the calcium result rather than merely contradicting it. Supply was never limiting. The signal was.
4. Translating it to Earth
The best test of whether that transfers is LIFTMOR, which did the thing everyone assumed was unsafe: heavy lifting in post-menopausal women who already had low bone mass.
| LIFTMOR | |
|---|---|
| Who | Post-menopausal women with low bone mass, T-score below −1.0, mean age 66 |
| What | Twice weekly, 30 minutes, supervised. Five sets of five repetitions at above 85% of one-rep maximum — deadlift, squat, overhead press — plus impact loading |
| Against | A low-intensity home exercise programme of the same duration |
| Result | Bone density improved at the spine and hip in the training group while it declined in the control group — femoral neck +0.3% against −1.9%. Note where the benefit sits: the training group barely gained, and the control group lost. In a population whose bone is going one way, holding position is the win |
| Safety | The intervention was well tolerated in a group conventionally told to avoid exactly this |
For everyone else, the same principle scales down without ceremony: progressive resistance training that gets genuinely heavy over time, at whatever your starting point happens to be, is the intervention. The word doing the work is progressive — a load your skeleton has adapted to has stopped being a signal.
The very low-effort version
For people who will not do any of the above, there is a smaller and genuinely interesting finding. Brief daily hopping — on one leg, a minute or two — has been shown to increase hip bone density in the hopping leg while the other leg, in the same person, did not. The design is what makes it convincing: each participant was their own control, so diet, hormones, vitamin D and everything else were identical between the two legs.
The effect is modest and local, and it is not a substitute for loading the spine, which cannot be hopped. But it establishes the principle cleanly: the bone that gets the signal responds, and the bone that does not, does not.
5. Where loading is not enough
- Loading does most for cortical bone — the dense outer shell — and less for trabecular bone, the honeycomb interior of the hip and vertebrae. That is unfortunate, because the trabecular compartment is where fragility fractures start.
- Which is why exercise is a foundation rather than an alternative in anyone who already has osteoporosis or has broken a bone from a standing-height fall. It is not a matter of trying hard enough.
- Established osteoporosis is a treatable condition and it is under-treated, particularly in men and particularly after a first fracture. A wrist or vertebral fracture in an older person is the warning, and it is routinely missed as one.
- The assessment in the UK is a fracture risk calculation — usually FRAX or QFracture — with a DEXA scan where the risk is intermediate or where it would change the decision. Ask about it after a fragility fracture rather than waiting to be offered it.
6. The drugs, and a risk that is routinely misrepresented
Bisphosphonates — alendronate weekly, or a yearly zoledronate infusion — reduce fractures substantially, cost very little, and are the usual first choice. Denosumab and the bone-building agents are alternatives with their own rules.
Osteonecrosis of the jaw, in proportion
- It is real, and at osteoporosis doses it is rare. Reported incidence for oral bisphosphonates in osteoporosis ranges from roughly 1 per 100,000 patient-years at the low end to a few hundred per 100,000 in some national cohorts, with the higher figures coming from longer treatment durations and populations with more dental extraction. The accurate statement is a range, not a single number — the denominators differ by how cases were ascertained. In tens of thousands of patient-years within the randomised osteoporosis trials, it did not appear as an adverse event at all.
- The impression most people have comes from a different setting. The condition was described in cancer patients receiving intravenous bisphosphonates at roughly ten times the dose, far more frequently, and often alongside chemotherapy and steroids. Applying that risk to a weekly alendronate tablet overstates it by orders of magnitude.
- It is not a reason to decline treatment, and treating it as one causes harm. The fracture being prevented is common; hip fracture in particular carries a mortality in the region of 20 to 30% within a year, and a large proportion of survivors never return to independent living. Weighed against roughly one in a hundred thousand, the arithmetic is not close.
- The dental advice is right even though the risk framing is wrong. The clearest risk factor is tooth extraction during treatment, which raises it markedly. So: have outstanding dental work done before starting where that is practical, keep up routine dental care, and tell your dentist you take it. That is worth doing precisely because it removes most of the remaining risk.
- Atypical femoral fracture is the other rare long-term concern, and is the main reason a treatment break — a "drug holiday" — is considered after about five years of oral treatment. New, unexplained thigh or groin pain on long-term treatment should be reported rather than ignored.
The pattern this chapter keeps running into. A rare, vivid, memorable harm is weighed against a common, statistical, invisible one, and the vivid one wins. It happens with statins and muscle pain, with anticoagulants and bleeding, and here with a jaw. Knowing the actual denominators is most of the defence against it.