Does running protect your bones better than swimming?
Running clearly stimulates bone formation better than activities without impact loading, but a direct comparison with swimming is absent from the research. If you want to protect your bones, impact-loading activities such as running are in any case more beneficial than swimming or cycling alone.
In a cross-sectional study, older runners aged 50 to 72 had approximately 40% more bone mineral than comparable peers who did not run. This applied to both men and women. Running is a high-impact activity: every step exerts force on bones, and that stimulates the formation of new bone tissue. Mouse research shows how this works mechanically: loading activates bone-building cells via specific signalling pathways. Whether this mechanism works equally strongly in humans has not yet been confirmed clinically.
A direct comparison with swimming cannot be made on the basis of the available studies. The research contains no data on the effect of swimming on bone density. What is known is that activities without impact loading do not, or barely, stimulate bone. Road cycling is the clearest example: it actually lowers bone density, particularly in the hip, even in relatively young cyclists. Cyclists who add running or other impact-loading exercises to their training do improve their bone density. Swimming, like cycling, involves no impact loading, which from a biological perspective suggests it has a similarly limited effect on bones. But that is reasoning from a mechanism, not from direct swimming data.
One caveat regarding running deserves attention, specifically for women. If female runners eat too little relative to their training load, hormonal balance is disrupted. This lowers oestrogen levels and actually damages bone formation. Research shows that 16 to 47% of lean female athletes have disordered eating behaviour, with an associated risk of lower bone density. Running therefore only protects your bones if you eat enough.
The 40% claim about runners is based on a single cross-sectional study (PMID 3945033); causality has not been proven. The effect of swimming is entirely absent from the supplied sources; the comparison is partly reasoning from the mechanism of impact loading. Cycling data (PMID 35734792) is indirect evidence for the importance of impact loading.