Quick Answer: Bone responds to load that is high in magnitude, applied quickly, and unusual in direction - which is why cycling and swimming, despite being outstanding cardiovascular exercise, do close to nothing for bone density, and why steady-state running does far less than most runners assume. If cardio is your whole programme, the highest-value additions are two short resistance sessions a week plus brief daily impact. This guide covers what each cardio type does, what to add, how much, and how to tell if it is working.

You run four times a week, you are fitter than most people you know, and your bone density scan came back worse than you expected.

That is not a contradiction. Cardiovascular fitness and skeletal loading are two different adaptations, and only one of them is guaranteed by a running habit.

Why does bone respond to some exercise and not others?

Bone adapts to strain, and specifically to strain that is large, fast, and unfamiliar. This is the mechanostat idea: bone tissue has a threshold, and loading below that threshold maintains what you have while loading above it triggers new bone formation.

Three features determine whether a given activity crosses the threshold.

Magnitude. How much force actually passes through the bone. Ground reaction forces in walking are roughly one times body weight; running is around two to three; jumping and landing can reach four or more. Lifting heavy loads produces high force through the loaded bones without the impact.

Rate. How quickly the force arrives matters as much as how large it is. A fast landing produces a much stronger osteogenic signal than the same peak force applied slowly, which is why hopping is more effective per repetition than slow grinding work.

Novelty and direction. Bone habituates. The same load in the same direction repeated thousands of times becomes background noise within a few weeks. Loading from unusual directions - lateral, rotational, multi-directional - keeps producing a signal when straight-ahead repetition has stopped.

This last point explains something that surprises people: a distance runner accumulates enormous numbers of loading cycles in a single direction, and after the initial adaptation, most of those cycles are contributing very little. Meanwhile, roughly forty short hops from a standing position - taking under a minute - keeps producing a response because the loading is high-rate and the body never fully habituates to it.

The practical version of all this: bone does not care how tired you are or how many calories you burned. It cares about peak force and how fast it arrived.

What does running actually do for your bones?

More than cycling and swimming, less than most runners believe, and very unevenly across the skeleton.

The evidence on distance runners is genuinely mixed. Studies of runners generally find higher bone density in the legs and hips than in sedentary controls - the loading is real and it does produce adaptation in the loaded bones. But the same studies frequently find spine density that is no better, and sometimes worse, than in non-runners.

There are two reasons for that pattern.

The first is habituation. Running loads the legs in a repetitive, predictable, vertical way. The initial adaptation is real, but after some months of consistent running the stimulus stops being novel, and continuing to add mileage adds cardiovascular fitness without adding much skeletal signal.

The second is energy availability, and it is the more serious issue. Distance running burns a lot of energy, and endurance athletes are a high-risk group for not eating enough to cover it. Chronic low energy availability suppresses reproductive hormones, and low oestrogen directly accelerates bone loss. A runner who is under-fuelling can be losing bone despite loading it - this is the well-documented pattern in the female athlete triad and its broader successor concept, relative energy deficiency in sport.

So running is not neutral for bone. It is genuinely useful for the hips and legs, close to useless for the spine and arms, and potentially harmful if it comes with an energy deficit.

What improves it: adding hills and sprints, both of which raise peak force and loading rate. And eating enough, which is not optional.

Why cycling and swimming are different in kind, not degree

These two are the important cases, because they feel like serious training and produce almost no skeletal benefit.

Cycling is non-weight-bearing. Your body weight is supported by the saddle and the bars, and the force through your legs is transmitted into pedals rather than into the ground. Studies of competitive road cyclists have repeatedly found lower bone density than in age-matched controls who do no sport at all - not merely no benefit, but a deficit.

The explanation combines several factors: the absence of impact loading, high energy expenditure that can create the same low-availability problem as running, long hours in a seated position, and calcium loss through sweat over long rides. None of that is an argument against cycling, which is excellent exercise. It is an argument that a cyclist needs a separate plan for their skeleton.

Swimming is the same logic taken further. Water supports essentially all of your body weight, so ground reaction force is near zero. Swimmers typically show bone density similar to or slightly below sedentary controls. The muscular pull on bone provides some stimulus, but it is far below the threshold that triggers meaningful bone formation.

Elliptical machines and rowing sit in a middle category - some loading, low impact, modest effect.

The honest framing: if your cardio is cycling or swimming, treat your bone programme as entirely separate. Nothing in those sessions is doing that job, and no amount of additional volume will change that.

The two resistance sessions that do the most

If you add one thing, add this. Resistance training loads bone through muscular pull, and heavy loading of the hip and spine is the intervention with the best evidence in this whole area.

The key characteristic is load, not duration. Programmes that produce measurable bone density improvements in women generally use progressively heavier resistance in compound movements - not light weights for high repetitions. The well-known clinical trials in postmenopausal women used heavy loading in a small number of exercises, twice a week, and produced improvements in spine and hip density rather than the maintenance most people expect.

A workable structure, twice weekly:

• A squat pattern - back squat, goblet squat, or leg press if barbells are not available. Loads the hip and spine directly.
• A hinge pattern - deadlift, Romanian deadlift, or hip thrust. This is the best available spine and hip loader.
• An overhead press. Loads the spine axially and the arms, which running never does.
• A row or pull-down. Balances the press and loads the upper back.
• A carry - a heavy suitcase carry for thirty seconds each side. Loads the spine asymmetrically, which is exactly the unusual direction bone responds to.

Three to five sets in the range of five to eight repetitions, with weight that is genuinely challenging by the last rep. Progress the load every few weeks. Forty minutes, twice a week.

The common mistake among runners moving into the gym is choosing light weights for high reps because it feels more familiar and less intimidating. That builds some endurance and very little bone. The load is the active ingredient.

The daily impact that costs one minute

Short, sharp, frequent impact is remarkably effective and almost free in terms of time and recovery.

The research on hopping and jumping interventions is encouraging: small daily doses of high-impact loading have produced measurable hip density improvements in premenopausal women over programmes of several months. The doses involved are small - in the range of ten to fifty hops a day, split into a couple of short bouts.

Why it works so well relative to its size comes back to rate and habituation. A hop delivers a high loading rate, and separating loading into short bouts through the day appears to produce a better response than the same number of repetitions done all at once, because bone cells become less responsive within a single session and re-sensitise with rest.

A practical version:

• Morning: 10 to 20 two-footed hops, landing softly, knees tracking over toes.
• Evening: another 10 to 20, or a set of drop-and-land from a low step.
• Rest a few hours between bouts. The separation is part of the effect.

Build up rather than starting at fifty. Tendons and joints adapt more slowly than the enthusiasm for a new habit, and shin or knee soreness in week one is the usual reason people stop.

This is also the piece most likely to be inappropriate for some readers. Anyone with existing osteoporosis, a history of fragility fracture, or significant joint problems should get individual advice before adding impact - the alternatives in the next section apply instead. This is general information, not medical advice.

What to do if impact is off the table

Plenty of people cannot hop, whether because of joint pain, existing low bone density, pelvic floor issues, or a recent injury. There is still a great deal available.

Heavy resistance training remains the priority. It does not require impact and it has the strongest evidence base of anything discussed here. Loading the hip and spine through squats, hinges, and carries produces the signal without the landing.

Increase load rather than adding impact. For someone avoiding jumping, progressively heavier resistance work over months is the main lever, and it is a large one.

Add stair climbing. Stepping produces higher forces than level walking without a landing impact, and it is easy to accumulate.

Use vibration cautiously. Whole-body vibration platforms have some supporting evidence for bone, though results are inconsistent and it is not a substitute for loading.

Train balance seriously. This one is often overlooked. The clinical outcome that matters is fracture, and fractures require falls. Balance and lower-body strength reduce fall risk substantially, and in older adults that pathway may protect more effectively than a small density change.

Pelvic floor considerations matter here too. If impact causes leaking, that is common, it is treatable, and a pelvic health physiotherapist can usually get someone back to jumping within a few months. Avoiding impact permanently because of it is not the only option.

How do you know whether any of it is working?

Slowly, and mostly not by feel. This is the frustrating part of bone training and the reason people abandon it.

Bone remodels on a cycle measured in months. A meaningful change on a scan generally takes one to two years, and typical improvements from an excellent programme are in the low single-digit percentages. That is genuinely worthwhile - a few percent of density is a meaningful shift in fracture risk - but it will never feel like progress the way a faster 5K does.

What you can track in the meantime:

Strength numbers. If your squat and deadlift loads are climbing steadily, the skeletal stimulus is increasing. This is the best available proxy and it updates weekly rather than annually.

Scan timing. A DXA scan repeated sooner than about two years usually cannot distinguish real change from measurement variation. If you have a baseline, plan the follow-up accordingly rather than rescanning out of impatience.

The context factors. Adequate energy intake, protein through the day, vitamin D status, and for many women the hormonal picture around perimenopause matter as much as the training. Bone loss accelerates sharply in the years around menopause, and no exercise programme fully offsets that on its own - it is worth a proper conversation with a doctor rather than a training-only approach.

The framing that helps most: this is not a training block with a finish line. It is a small permanent addition to a week that already contains cardio you enjoy. Two resistance sessions and a minute of hopping a day, sustained for years, is what the evidence actually supports. This is general information, not medical advice.

TL;DR:

• Bone responds to load that is large, fast, and unfamiliar - not to duration or calorie burn.
• Running helps hips and legs modestly, does little for the spine, and can harm bone if energy intake is low.
• Cycling and swimming provide essentially no skeletal stimulus; cyclists often show density below sedentary controls.
• Two heavy resistance sessions a week - squat, hinge, press, row, carry - is the single highest-value addition.
• Ten to fifty hops a day, split into short bouts, produces measurable hip benefit for about a minute of work.
• If impact is not possible, load heavier and train balance - fractures need falls, and falls are preventable.
• Change takes one to two years on a scan. Track strength numbers in the meantime.

The cardio is not the problem, and nothing here suggests doing less of it. It just does not cover your skeleton, and the thing that does takes about ninety minutes a week. This is general information, not medical advice.