REF-1700 · Jones, Starrs & Hunter, 2026 · Menopause · Paper · PubMed

High-intensity, impact and strength training raised bone density a little in postmenopausal women but not significantly in premenopausal women

A meta-analysis of 8 RCTs (379 women: 6 trials, 267 postmenopausal; 2 trials, 112 premenopausal) found pooled effects of 0.47 at the femoral neck (n = 141) and 0.36 at the lumbar spine (n = 182) after menopause; the abstract gives no confidence intervals and does not report what its risk-of-bias assessment found

Main caveat: The authors': small effect sizes and methodological limitations; few trials, especially in premenopausal women.
moderate evidenceMeasured on non-athletes to trained athletesRead from the abstract only

Jones, Starrs and Hunter asked whether high-intensity, impact or strength training raises bone mineral density after menopause, and whether the answer differs before menopause.

They searched PubMed and the Cochrane trials register in July 2025 for randomized controlled trials published from March 2018 that compared one of these kinds of training with a control arm in premenopausal or postmenopausal women and measured bone density. They rated quality with the PEDro scale and risk of bias with RoB 2.

Eight trials with 379 women qualified, six in postmenopausal women (267) and two in premenopausal women (112). With the three training types pooled, training raised bone density at the femoral neck (effect 0.47, P = 0.014, 141 women) and the lumbar spine (0.36, P = 0.012, 182 women) in postmenopausal women; the effects in premenopausal women were not significant.

The authors call the effects small, point to methodological limitations, and ask whether stronger protocols would have worked in premenopausal women or whether bone responds differently in the two groups.

Participants

8 RCTs, 379 women: 6 trials with 267 postmenopausal women, 2 trials with 112 premenopausal women.
Pooled analyses (postmenopausal): femoral neck n = 141, lumbar spine n = 182.
Age, years since menopause, baseline bone status (normal, osteopenia, osteoporosis), training history and hormone therapy use: not in the abstract.

Protocol

Search: PubMed and CENTRAL on 17 July 2025; randomized trials published on or after 1 March 2018; terms on menopause, exercise and bone.
Inclusion: premenopausal or postmenopausal women; at least one arm of high-intensity, impact or strength training; a control arm; bone mineral density measured.
Appraisal: PEDro scale (quality) and RoB 2 (risk of bias); results not reported in the abstract.
Analysis: meta-analysis with intervention types pooled; the effects are reported with t statistics (t[3] = 5.22 femoral neck, t[4] = 4.36 lumbar spine). The effect-size metric, the training doses, durations and control conditions are not stated in the abstract.

Key Results

Site (training vs control)GroupEffectTestPn
Femoral neckpostmenopausal0.47t[3] = 5.220.014141
Lumbar spinepostmenopausal0.36t[4] = 4.360.012182
Femoral neck and lumbar spinepremenopausalnot significant (values not in the abstract)——112 in 2 trials
The authors describe the postmenopausal effects as small but significant.
Confidence intervals, heterogeneity, results by training type (high-intensity, impact or strength separately) and total-hip or other sites: not in the abstract.

Practical Application

After menopause, lifting and impact work moved bone density a little in the right direction. Pooled over the postmenopausal trials, high-intensity, impact and strength training raised femoral-neck (0.47) and lumbar-spine (0.36) density against controls, in 141 and 182 women; the authors call the effects small. This supports SYNTH-96's advice to add heavy lifting with impact, which until now rested on one trial (REF-1405) and a physiology principle (PHYS-37).
Before menopause the two trials found no significant effect. With 112 women in 2 trials this is a lack of evidence, not proof of no benefit; the authors leave open whether stronger protocols would have worked. Inference, not tested here: SYNTH-96's advice to start in the transition rests on the timing of bone loss (REF-1404), and this result neither supports nor refutes it.
The three kinds of training were pooled, so this review does not say whether impact, heavy lifting or high-intensity work does the job on its own; SYNTH-96's heavy-lifting-with-impact shape stays the tested template (REF-1405).
Bone density, not fractures, and not in athletes: inference, not tested here — for a cyclist or swimmer whose sport does not load bone (PHYS-37), lifting and impact outside the sport remain the lever, transferred from women who were not stated to be athletes.
This is not treatment advice: osteoporosis, a past fragility fracture or the choice of medication is a matter for her physician, and heavy loading with known osteoporosis starts with the physician's agreement and a qualified coach.

Limitations

The authors': small effect sizes and methodological limitations; few trials, especially in premenopausal women. The t statistics carry 3 and 4 degrees of freedom, which suggests that only some of the six postmenopausal trials entered each pooled effect (the abstract does not give the count).
Only trials published from March 2018 and two databases searched, so earlier trials (whether LIFTMOR, published online in 2017, is included is not stated) may be missing.
Training types were pooled; doses, durations, supervision and control conditions are not described in the abstract.
From the abstract alone, unknown: the effect-size metric, confidence intervals, heterogeneity, the PEDro and RoB 2 results, participants' age, bone status and training history, adverse events, registration and funding.
Bone density is a stand-in for fracture risk; no fracture outcome.
Written from the abstract and the PubMed record; full text not read.
How this entry was checked and changed
Why this level

risk of bias not read (abstract only): PEDro and RoB 2 were applied but the abstract reports only "methodological limitations", not what they found; small pooled samples (n = 141 and 182) and no confidence intervals in the abstract