Study archive · page 31

Resistance training gives healthy adult women large gains in upper and lower body strength and a medium-sized gain in muscle mass, according to the first meta-analysis to look exclusively at women, pooling 24 studies.

⚠What it shows: Reliable as an overall picture: results pooled from 24 studies of healthy women training anywhere from 4 weeks to 12 months. Research fields tend to publish their successes, and correcting for that shrank the lower body strength effect from large to medium, so the exact sizes are uncertain.Meta-analysis · Sports Med

In obese older adults on calorie-restricted diets, adding resistance training three times a week prevented 93.5% of the diet's lean mass loss while fat loss stayed essentially the same, according to a meta-analysis of six trials.

⚠What it shows: A consistent picture built on a thin base: results pooled from just six trials in obese older adults, with programmes running 12 to 24 weeks. Whether the same protection holds for younger or leaner dieters is untested here.Meta-analysis · Nutrients

Adding aerobic training to a strength programme does not compromise gains in muscle size or maximal strength, according to a meta-analysis of 43 studies in healthy adults. Explosive strength gains may be blunted, particularly when cardio and lifting happen in the same session.

⚠What it shows: About as solid as training evidence gets: results pooled from 43 studies in healthy adults, and the conclusion held regardless of age, training background or cardio type. The one genuine cost was explosive power, which dipped mainly when cardio and lifting shared a session.Meta-analysis · Sports Med

Flywheel resistance training improved change of direction performance more than traditional resistance training in male athletes, in a meta-analysis of seven studies. Two or fewer sessions a week, and 12 or fewer sessions in total, produced the largest effects.

⚠What it shows: Promising but thin: pooled results from just seven studies, all in male athletes, so nobody knows yet whether it holds for women or everyday gym-goers. The studies also varied in sports and programmes, which makes the exact size of the flywheel advantage uncertain.Meta-analysis · Int J Environ Res Public Health

A meta-analysis of 13 randomised trials found that adding neuromuscular electrical stimulation to resistance training produced modestly greater gains in strength and muscle mass than lifting alone.

⚠What it shows: Worth taking seriously, though the edge is small: results pooled from 13 trials where people were assigned at random favoured the combination for both strength and size. The strength studies disagreed with each other quite a bit, so the true size of that benefit is uncertain.Meta-analysis · Eur J Appl Physiol

Pooling 162 randomised trials with 10,475 adults aged 18 to 81.6 years, exercise produced small to moderate improvements in bone mineral density at five of six skeletal sites, including the femoral neck (g 0.28) and lumbar spine (g 0.22). The authors stress that rankings of which exercise type wins, and the suggested dose peak near 1100 MET-min/week for lumbar spine, are exploratory rather than prescriptions.

⚠What it shows: The headline effect is solid, the fine detail is not: results pooled from 162 randomised trials in 10,475 adults, with confidence ranging from high at the femoral neck down to very low at the total hip. The comparisons between exercise types were imprecise, the studies disagreed with each other, and the dose figures shifted depending on where the model's edges were drawn.Review · Bone

Physical activity improved cognition in sedentary older adults across seven randomised trials totalling 321 people, with aerobic exercise and programmes longer than 12 weeks showing the clearest effects.

⚠What it shows: Promising but thin: results pooled from only seven small trials totalling 321 sedentary older adults, and the authors themselves call for more rigorous studies. The trials also disagreed with each other about how large the benefit really is.Meta-analysis · J Alzheimers Dis

Across three pooled analyses spanning 15 years of research, exercise was associated with reduced anxiety, but only aerobic forms showed the effect. Sessions of at least 21 minutes appeared necessary, and programmes needed to exceed 10 weeks to shift a person's underlying anxious disposition.

⚠What it shows: Robust in direction, fuzzy on dose: three sets of results pooled from 15 years of research agree that exercise lowers anxiety, but only the aerobic kind, and it worked no better than relaxation techniques. The true minimum session length is still unclear because other factors muddied that finding.Meta-analysis · Sports Med

Physical exercise had a positive and significant effect on anxiety in older adults in a meta-analysis of 17 studies published between 2005 and 2021, with consistent results across two different ways of pooling the data.

⚠What it shows: Encouraging, but the field is thin: results pooled from 17 studies of older people showed a clear drop in anxiety scores, yet the authors themselves call for higher quality research before the size of the benefit can be trusted.Meta-analysis · Comput Math Methods Med

Across six large national cohorts in Europe, North America, Asia and Africa covering 89,056 middle-aged and older adults, declines in cognitive function were consistently associated with worsening depression scores, with an overall effect of -0.094.

⚠What it shows: A strikingly consistent pattern, but researchers followed people rather than testing anything on them, so it cannot say whether fading cognition worsens mood or the reverse. The link held across 89,056 adults on four continents and across age, sex, drinking habits and marital status.Meta-analysis · J Affect Disord

Adults with long COVID symptoms had a peak oxygen uptake 4.9 mL/kg/min lower than recovered peers when tested more than 3 months after infection, in a meta-analysis of 9 studies. Deconditioning was common but not the whole story: abnormal oxygen extraction by the muscles, dysfunctional breathing and blunted heart-rate responses were also described.

⚠What it shows: Treat this as a credible signal rather than a settled number: results pooled from 9 studies covering 464 people with symptoms and 359 without, and the certainty was rated low. The underlying studies were small, prone to bias and defined long COVID differently from one another.Meta-analysis · JAMA Netw Open

Pooled results from 10 randomised trials covering 400 adults with overweight or obesity found exercise training improved flow-mediated dilation, a marker of blood vessel function; the benefit was clear for aerobic training while resistance training fell just short of statistical significance.

⚠What it shows: Encouraging but thin: results pooled from 10 trials covering just 400 adults, and the quality checks raised some concerns, so the authors themselves urge caution. Lifting fell just short of a clear effect, which leaves its role unresolved rather than disproven.Meta-analysis · Sci Rep

Pooling 17 trials in 576 healthy women, exercise produced modest improvements in blood lipids: total cholesterol fell by 5.77 mg/dL, triglycerides by 5.60 mg/dL, and HDL cholesterol rose by 4.49 mg/dL. Combined training was most effective for triglycerides and total cholesterol, while aerobic exercise significantly raised HDL.

⚠What it shows: A real but small effect: results pooled from 10 articles covering 17 trials with 576 healthy women in total. The trials tested different types and intensities of exercise, so this cannot tell you how much or how hard to train to get the same shift.Meta-analysis · J Physiol Biochem

Across six randomised trials of 1246 overweight or obese adults with metabolic syndrome, 6 to 12 months of diet and physical activity change reduced waist circumference by 3.12 cm on average and lowered the inflammatory markers C-reactive protein and IL-6, while raising adiponectin.

⚠What it shows: A consistent direction of travel: results pooled from six trials in 1246 adults with metabolic syndrome, all lasting 6 to 12 months, favoured lifestyle change on every marker measured. The authors still want higher-quality studies before the inflammation findings are treated as settled.Meta-analysis · Biol Res Nurs

In a 36-week mid-stage trial of 230 adults with obesity or overweight, the once-daily oral GLP-1 pill aleniglipron produced up to 11.3% more weight loss than placebo, with mostly mild to moderate gut side effects.

⚠What it shows: Promising but not final: one 36-week trial in 230 adults, run and part-authored by the drug's developer, so the result needs confirming in the larger late-stage trials that decide whether a medicine gets approved.Study · Nat Med

Strength training on its own reduced total cholesterol, triglycerides, LDL cholesterol and C-reactive protein in adults, while raising HDL cholesterol and adiponectin, in a meta-analysis of trials comparing lifting with no exercise.

⚠What it shows: Solid pooled evidence: results pooled from many adult trials comparing lifting alone with a group that did not train, and all six blood markers moved the right way. The shifts were small to moderate, and the search only runs to May 2017, so newer trials are missing.Meta-analysis · J Phys Act Health

Meditation-based programmes lasting 4 to 12 weeks improved self-reported sleep quality in adults aged 55 and over with mild cognitive impairment, with a moderate to large pooled effect, and were also linked to less anxiety, stress and depression.

⚠What it shows: Promising but not settled: results pooled from seven small studies in 434 older adults, some without random assignment, and the studies disagreed with each other on how big the benefit is. Sleep was mostly self-reported rather than measured objectively.Review · Geriatr Nurs

Pooled results from 10 randomised trials in 2134 adults with insomnia or poor sleep found cannabinoids reduced insomnia severity and improved sleep quality scores, increased total sleep time by 33.99 min, lifted sleep efficiency by 4.31% and modestly cut time to fall asleep by 12.0 min. Side effects, mostly mild dizziness and dry mouth, were roughly four times more frequent than with comparators, with no serious harms recorded.

⚠What it shows: Encouraging but not the last word: results pooled from 10 randomised trials in 2134 adults, comparing cannabinoids against placebo or melatonin. The trials were mostly short, so the right dose and whether the benefit lasts are both unresolved, and mild dizziness and dry mouth were clearly more common.Review · Sleep Med

Oral collagen supplements improved both pain and function in people with knee osteoarthritis across 11 randomised trials totalling 870 participants, with pain scores improving by 13.63 points more than placebo and function by 6.46 points, improvements the authors call statistically and clinically noteworthy.

⚠What it shows: Reasonably strong for a supplement: results pooled from 11 trials in 870 people with knee osteoarthritis all pointed the same way, favouring collagen over placebo. The trials disagreed with each other on how large the benefit is, so the exact size remains uncertain.Meta-analysis · Clin Exp Rheumatol

Across nine pooled trials in 627 people, heat therapy reduced short-term pain and improved grip strength in hand osteoarthritis. Combined approaches such as paraffin and mud treatments outperformed heated gloves alone.

⚠What it shows: Solid for the short term: results pooled from nine trials in 627 people back the pain relief and grip gains, but the trials delivered heat in different ways and only tracked people briefly. Nobody knows yet whether the benefit lasts.Meta-analysis · J Rehabil Med

Greater exposure to ultra-processed food was associated with higher risks across 32 health outcomes in an umbrella review of meta-analyses covering nearly ten million people. Evidence was rated convincing for cardiovascular disease related deaths (risk ratio 1.50, a 50% higher risk), type 2 diabetes, anxiety and common mental disorders.

⚠What it shows: As strong as diet-tracking evidence gets without a trial: findings pooled from dozens of studies following nearly ten million people, so the pattern is unlikely to be a fluke. It still cannot prove the food caused the harm, and most of the underlying evidence was rated low or very low quality.Systematic review · BMJ

Pooled results from 19 randomised trials in 768 healthy adults found that taking collagen peptides daily alongside several weeks of training produced small but significant gains in fat-free mass (SMD 0.48), tendon morphology (SMD 0.67), muscle architecture (SMD 0.39) and recovery of reactive strength 48 hours after muscle-damaging exercise (SMD 0.43). Maximal strength shifted least of all, at SMD 0.19.

⚠What it shows: Decent evidence for a supplement, with sensible limits: results pooled from 19 randomised trials in 768 healthy adults who supplemented for at least eight weeks while training. The authors rated their own confidence as moderate for body composition, low for most outcomes and very low for the tendon measurements, so the size of these effects could still move.Meta-analysis · Sports Med

Exercise reduced the risk of depression both during and after pregnancy in a meta-analysis of 32 trials with 4,189 participants, with yoga-style mind-body exercise working best before birth and structured aerobic exercise most effective for treating postnatal depression.

⚠What it shows: Strong enough to take seriously: results pooled from 32 trials in 4,189 people across pregnancy and the postnatal period. Effects looked smaller when exercise was compared against active alternatives like health education rather than against nothing at all.Meta-analysis · BMC Pregnancy Childbirth

Interactive digital tools for self-managing low back pain have a weak evidence base: across 6 completed randomised trials with 2,706 adults, only one showed a benefit over control, though none showed harm.

⚠What it shows: Take the app-store hype down a notch: a sweep of all the studies up to March 2016 found just six completed trials with 2,706 people, and only one favoured the digital tool. Participants were mostly white, well educated and middle aged, so nobody knows whether these tools help anyone else.Meta-analysis · J Med Internet Res