Resistance Training for Bone Health: What the Trials Show

Resistance training helps the hip that matters most. It does not finish the job.

If your goal is fracture prevention, the evidence does support progressive resistance training. It just does not support the lazy version of the story. In a 2021 meta-analysis by Ponzano et al., resistance training increased femoral neck bone mineral density by 0.02 g/cm² (95% CI 0.01 to 0.03) across 5 randomized trials with 521 participants. That is the narrow bridge of bone just below the ball of the hip joint, and it is where hip fractures often turn into life-changing events.

The catch is just as important. In the same meta-analysis of 53 randomized controlled trials, resistance training alone did not significantly improve lumbar spine bone density or total hip bone density. Lumbar spine BMD changed by 0.02 g/cm² (95% CI -0.01 to 0.05; 4 studies), which was not statistically significant. Total hip BMD changed by 0.00 g/cm² (95% CI 0.00 to 0.01; 4 studies), which was effectively no meaningful difference. That makes the effect anatomically selective, not general.

That selectivity makes sense on mechanical grounds. Hip-dominant lifts load the proximal femur hard. The femoral neck appears to respond. But the trials included in the meta-analysis did not directly test why the response showed up there and not elsewhere, so any explanation about impact thresholds or spinal loading remains inference, not a measured finding.

The functional gains are real, even if the bone gains are modest

Bone density gets the headline, but the best-supported benefits from resistance training in fracture-risk populations are broader. In the Ponzano review, quality of life improved with a standardized mean difference of 0.32 (95% CI 0.22 to 0.42) across 20 studies. Pain improved with an SMD of -0.26 (95% CI -0.37 to -0.16) across 17 studies. Timed Up and Go performance improved by 0.89 seconds (95% CI -1.01 to -0.78) across 13 studies. Adverse events were rare.

Those findings matter because fracture risk is not just a DXA printout. A person who moves better, hurts less, and reports better function may still benefit from training even if the bone density change is small. Still, small means small. The pain effect was pooled across heterogeneous studies using different measures, and the trials did not report how many people achieved a clearly noticeable pain reduction. The Timed Up and Go improvement was statistically significant, but the evidence provided here notes that it falls below commonly cited thresholds for functional independence in community-dwelling older adults.

So the honest summary is plain: resistance training helps. It just does not perform miracles, and the effect sizes do not let you pretend otherwise.

The uncomfortable result: falls did not improve, and may have worsened

The most clinically awkward finding in Ponzano et al. is the falls signal. The relative risk for falling was 1.23 (95% CI 1.00 to 1.51) across 5 studies. The fall incidence rate ratio was 1.05 (95% CI 0.91 to 1.21) across 7 studies. Neither estimate proves harm with high confidence, and the first confidence interval sits right on the statistical boundary. But neither estimate supports the claim that resistance training alone reduces falls.

This is the part many summaries glide past. They should not. You can build stronger muscle and still fail at the exact skill that keeps bones intact: not hitting the floor.

The current evidence does not tell us why this happened. Several explanations are plausible. Early training fatigue could temporarily raise fall risk. Higher-functioning participants may simply spend more time moving around, which gives them more chances to fall. Or resistance training may improve force production in controlled lifts without adequately training reactive balance, perturbation recovery, or proprioception in messy real life. Those explanations fit the pattern, but the trials summarized here did not experimentally verify them.

What seems to fix the problem: add balance and impact work

The best available correction comes from multicomponent exercise programs. In a 2021 meta-analysis by Rodrigues et al., interventions that combined resistance training with impact and balance work performed better on falls-related outcomes than single-modality approaches in people at fracture risk. That does not mean every circuit class with a foam pad qualifies. It means modality design matters.

Other trials in the evidence base point in the same direction. A 2021 secondary analysis by Daly et al. found that greater training frequency and volume within an 18-month multicomponent program were associated with better hip bone outcomes in middle-aged and older men. The analysis supports a dose-response pattern, but not a minimum effective dose. We know more work tended to do more. We do not know the floor below which the program stops being osteogenic.

A 2024 randomized trial by Blay et al. adds an important comparison. In postmenopausal women, a 12-month program that combined resistance and impact exercise produced relative benefits to hip bone structure that were comparable to risedronate on some structural parameters. That does not prove exercise replaces bisphosphonates in every patient, and the endpoint was hip structure rather than fracture events. But it does show that well-designed exercise is not a decorative add-on.

A 2023 secondary analysis by Talevski et al. found that an 18-month community-based program combining resistance, balance, impact, and mobility work improved patient-reported physical function and quality of life in older adults at fracture risk. That is the sort of result clinicians and patients can actually use.

Frailty changes the context, not the conclusion

The broader frailty literature supports exercise as a real intervention, not a wellness bumper sticker. A 2024 umbrella review by Ueshima et al. concluded that non-pharmacological interventions, especially exercise-based approaches, help prevent frailty progression in older adults. A 2025 meta-analysis by Yuan et al. extended that case to older adults with cognitive frailty, finding benefits for cognition and several physical-function outcomes, though the supplied summary notes that quality-of-life effects were not clearly significant there.

That matters because the people at highest fracture risk are often the least likely to enroll in polished gym-based resistance programs. Unfortunately, they are also the people least represented in much of this evidence. The major trials and the Ponzano meta-analysis mostly enrolled relatively healthy, community-dwelling adults. The results may not carry over cleanly to frail, institutionalized, or very old adults.

What the biology says, and what it still does not prove

A 2025 randomized trial by Bloch-Ibenfeldt et al. found that heavy resistance training increased short-term bone formation markers in well-functioning older adults. That is useful mechanistic support. Older bone still responds to loading signals. But biomarkers are not fractures, and they are not even DXA changes. They show that the pathway turns on, not that the final clinical outcome follows.

Program details also matter inside the resistance-training bucket itself. A 2024 randomized trial by Abdul-Al et al. compared hypertrophy training with contrast training in healthy women aged 60 to 70. In the abstract provided, contrast training increased whole-body, lumbar spine, total hip, and femoral neck BMD over 12 months, while hypertrophy training increased whole-body and lumbar spine BMD but left total hip and femoral neck unchanged. That study suggests not all resistance prescriptions are interchangeable, although it does not settle the best protocol for higher-risk populations.

What this evidence can and cannot promise

The biggest limitation is simple. No study in this evidence set showed that resistance training, alone or in combination, reduced actual fracture incidence. The main bone outcomes were surrogate measures such as DXA-derived BMD or structural estimates. Those markers matter, and lower BMD is associated with higher fracture risk, but the chain from exercise-induced BMD gain to fewer fractures was not directly tested here.

The femoral neck effect also sits near the technical precision limits of DXA. A group-level mean increase of 0.02 g/cm² can be statistically real while still being hard to interpret in any single patient. That does not make the result fake. It does mean readers should resist the fantasy that a tiny change on a scan automatically means their hip is now crash-proof.

The stakes are high enough that even modest gains deserve attention. In a 2022 cohort study by Campenfeldt et al., patients younger than 70 who had sustained a femoral neck fracture still showed sarcopenia, lower bone density, and severe functional impairment 10 years later compared with controls. Hip fractures are not a bad weekend. They are an excellent way to lose a decade.

The evidence supports resistance training as a necessary part of fracture prevention, not a complete one. If the program builds strength but ignores balance and impact capacity, it may improve the scan while leaving the fall problem unsolved.

The practical verdict is narrow but useful. Use progressive, hip-loading resistance exercise. Add explicit balance training. Add impact work if the person can do it safely. Expect better function and a small, site-specific bone gain at the femoral neck. Do not claim that lifting alone prevents falls, and do not claim that these trials proved fracture reduction. The data are better than "exercise is good," and less convenient than the gym-bro version of the story. That is usually where the truth lives.