Study archive · page 24

A pooled analysis across resistance training studies found that muscle growth improved as sets were taken closer to failure, while strength gains were similar across a wide range of repetitions in reserve.

⚠What it shows: Take it as a strong steer rather than a settled rule: how close to failure lifters actually trained had to be estimated from each study's write-up, and the authors themselves call the analysis exploratory. The trend for muscle growth was consistent, but the exact best stopping point is still unclear.Meta-analysis · Sports Med

In pooled results from controlled trials in 93 highly trained middle- and long-distance runners, adding strength and plyometric training 2-3 times per week for 8-12 weeks produced a large improvement in running economy.

⚠What it shows: A strong and consistent pooled result, but built on only five controlled trials covering 93 runners, so the single best programme design is not pinned down. Every participant was a high-level competitive runner, so the size of the benefit for everyday runners is untested here.Meta-analysis · J Strength Cond Res

Across 18 randomised trials in 1180 adults with type 2 diabetes and overweight or obesity, resistance training alone improved waist size, fasting glucose, fasting insulin, glycated haemoglobin, HDL cholesterol and triglycerides compared with standard care without exercise.

⚠What it shows: As strong as exercise evidence gets: results pooled from 18 trials in 1180 adults with type 2 diabetes and excess weight, all assigned at random. The certainty of the underlying evidence still ranged from very low to moderate, so the exact size of each benefit is uncertain.Meta-analysis · Br J Sports Med

Pooled data from 29 resistance training studies shows men and women gain muscle at similar relative rates when following the same programme. Absolute size gains slightly favoured men, mainly in the upper body, and training experience did not change the pattern.

⚠What it shows: About as solid as this question gets for now: results pooled from 29 studies of healthy adults aged 18 to 45 doing identical programmes. Most participants were not highly trained, so whether the pattern holds in seasoned lifters is still an open question.Meta-analysis · PeerJ

In a pooled analysis of 12 randomised trials with 11,758 participants, tirzepatide reduced body weight, BMI and waist circumference more than GLP-1 receptor agonists, placebo or insulin, though gastrointestinal side effects were more common.

⚠What it shows: Strong pooled evidence: 12 trials with 11,758 people, all assigned at random, and the results pointed the same way against every comparator. The clear trade-off is tolerability, with nausea and other gut side effects markedly more common than placebo.Meta-analysis · Front Public Health

Across 50 randomised trials lasting at least a year, GLP-1 receptor agonists were not linked to any overall increase in cancer risk; a signal for higher thyroid cancer risk, and a possible reduction in uterine cancer in obesity trials, both warrant further study.

⚠What it shows: Broadly reassuring: results pooled from 50 drug trials that each ran for at least a year found no overall cancer difference. The thyroid signal deserves dedicated study rather than alarm, and several authors report fees from the companies that make these drugs.Meta-analysis · Diabetes Obes Metab

Across 28 trials in 302 healthy adults, single fasted workouts were associated with larger rises in glucose and insulin and larger drops in free fatty acids than fed workouts, with no difference in triglycerides or the body's fuel-use ratio. The pooled results give no support to the idea that fasted sessions are better for glucose or fat metabolism.

⚠What it shows: Worth taking seriously but deliberately narrow: results pooled from 28 trials in 302 healthy adults, all measuring single workouts lasting 36 to 150 minutes. It says nothing about whether weeks of habitual fasted training change fat loss, fitness or long-term metabolic health.Meta-analysis · Clin Nutr ESPEN

In a pooled comparison of 28 randomised trials in 1,620 adults with overweight or obesity, aerobic exercise ranked best for reducing body weight and BMI, while HIIT ranked best for waist circumference, body fat, triglycerides, fasting glucose and VO2 max. No training mode significantly improved total cholesterol or blood pressure.

⚠What it shows: A useful ranking rather than a final verdict: results pooled from 28 trials in 1,620 adults with overweight or obesity, comparing four training styles side by side. Rankings like this lean partly on indirect comparisons between trials that never faced each other, so the gaps between modes are less certain than they look.Meta-analysis · Front Endocrinol (Lausanne)

Across 13 randomised trials in 1,470 perimenopausal and postmenopausal women, exercise of many forms, from aerobic work and yoga to Pilates, resistance training, stretching and relaxation, significantly improved sleep quality and psychological well-being.

⚠What it shows: Broadly trustworthy on direction, less certain on size: a sweep of all the trials on this covered 13 trials and 1,470 women, but the results were not pooled into a single number, so how big the benefit is remains unclear. Trial quality also varied across the included studies.Systematic review · Menopause

Pooled results from five small trials in 263 women suggest swimming may improve lumbar spine bone mineral density in postmenopausal women, but only when swimming time was between 3 and 6 hours; below 3 hours there was no effect.

⚠What it shows: Treat this as a lead, not a verdict: results pooled from just five trials with 263 women in total, and the studies disagreed with each other on the spine result. The authors themselves say the benefit needs further confirmation.Meta-analysis · Biomed Res Int

Pooled randomised trials show GLP-1 receptor agonists reduce body weight, BMI, waist measures and insulin resistance in women with PCOS more than metformin or placebo, but hormone levels did not change, HDL cholesterol fell, and nausea, vomiting and dizziness increased.

⚠What it shows: Encouraging but incomplete: results pooled from randomised trials in women with PCOS show clear weight and insulin wins, yet the authors themselves call for longer studies. Side effects were common, HDL cholesterol actually fell, and hormone levels did not budge.Meta-analysis · Sci Rep

Resistance training produced small to moderate gains in bone mineral density at the lumbar spine, femoral neck and total hip in postmenopausal women, across 17 controlled trials. Unexpectedly, programmes with fewer than 2 sessions a week produced bigger gains at the spine and hip than more frequent training, and free weights outperformed machines at the hip.

⚠What it shows: Solid for the headline benefit, shaky for the training advice: results pooled from 17 trials show real but modest bone gains in postmenopausal women. The frequency and free-weight findings come from slicing the data into subgroups, and the authors themselves say those slices cannot support firm recommendations.Meta-analysis · Osteoporos Int

Fasting for 14 hours a day may be as effective as 16 hours for weight loss, according to a sweep of 24 randomised trials of time-restricted feeding. The pattern produced moderate weight loss in people with overweight or obesity, and losses exceeded 5% of starting body weight when combined with calorie restriction.

⚠What it shows: Encouraging but not settled: a sweep of 24 trials found consistent short-term benefits in people carrying extra weight, but the trials were mostly brief and the authors themselves call for longer, better designed studies.Systematic review · Nutrients

Across 167 trials with 11,998 adults, intermittent fasting patterns produced weight loss comparable to ordinary calorie restriction of the same size, and fasting regimens showed weight rebound after 12 weeks while continuous restriction did not. What mattered was the depth of the energy cut, not the meal timing.

⚠What it shows: Strong evidence: results pooled from 167 trials in 11,998 adults with overweight, obesity or metabolic problems. Certainty ranged from low to high across the comparisons, and the rebound finding comes from a subgroup analysis, so it is the least firm piece.Meta-analysis · Nutr Rev

Eating earlier in the day under early time-restricted eating cut body weight by 1.84 kg, fat mass by 1.10 kg and visceral fat area by 9.76 cm2 versus unrestricted eating across 13 pooled trials, without a significant loss of fat-free mass. A 16:08 schedule, sixteen hours fasting to eight hours eating, looked best at preserving lean tissue.

⚠What it shows: Fairly convincing on fat loss: results pooled from 13 trials in 859 people pointed the same way. The muscle-sparing conclusion is shakier, since the change in lean mass only just missed statistical significance and the trials were short.Meta-analysis · Diabetes Metab Syndr

In postmenopausal women, whey protein supplementation boosted biceps curl strength and lower limb lean mass only when paired with resistance training. Taken without training, it showed no significant benefit for strength or lean mass, and was linked to lower dietary protein intake.

⚠What it shows: Strong enough to act on the headline point: results pooled from 10 trials in women aged 55 and over, and the benefit appeared only alongside resistance training. Whey moved neither fat mass nor body weight in either group, so this is about muscle, not weight loss.Meta-analysis · Nutrients

Pooled results from 28 randomised trials in 830 adults suggest coenzyme Q10 supplementation lowers blood markers of exercise-induced muscle damage and oxidative stress, with each extra 100 mg per day linked to bigger reductions.

⚠What it shows: Worth noting but not settled: results pooled from 28 trials in 830 adults show blood markers moving, which is not the same as feeling less sore or performing better. Most of the trials were run in Asia, so how far the findings travel is unclear.Meta-analysis · Clin Nutr ESPEN

Across 18 randomised trials, yoga reduced chronic low back pain versus no exercise for up to 6 to 7 months and improved disability for up to 12 months, but it was no better than physical therapy exercise on either outcome, and quality of life did not improve.

⚠What it shows: Believable but unglamorous: results pooled from 18 trials, with the strength of the evidence ranging from very low to moderate. The pain benefit had faded by the 12-month mark, and yoga showed no edge over ordinary exercise at any point.Meta-analysis · PLoS One

Runners retrained to land more softly had roughly two-thirds lower risk of knee injury in one trial of 320 people. Meanwhile footwear choices, graduated running programmes, multicomponent exercise and injury-prevention education showed no influence on knee injury risk across 17 other prevention trials.

⚠What it shows: Promising but fragile: the standout result comes from a single trial of 320 runners, rated low certainty within a wider sweep of 30 trials in new and recreational runners. It needs repeating before it counts as anything like settled.Meta-analysis · Br J Sports Med

Sleep deprivation cuts endurance performance by a moderate amount, according to pooled results from 31 studies of 478 people, and efforts lasting more than 30 minutes suffer most. One night of total sleep loss and one or more nights of partial loss compromised performance to a similar degree.

⚠What it shows: Consistent enough to take seriously: results pooled from 31 controlled studies with 478 healthy people all pointed the same way, though most studies had unclear risk of bias. Whether elite and professional endurance athletes are hit the same way is still unknown.Meta-analysis · Eur J Sport Sci

Older walkers increasingly brace the ankle by tensing the shin and calf muscles together early in each step, yet produce weaker push-off. In 107 healthy adults aged 26 to 86, ageing was linked to this stability-first strategy that sacrifices propulsion and efficiency.

⚠What it shows: A snapshot comparison rather than a study that followed people as they aged: 107 healthy adults aged 26 to 86 were each measured walking once, so it shows how gait differs with age, not why it changes or whether training can reverse it.Study · Gait Posture

Adults with high cardiorespiratory fitness had a 53% lower risk of dying from any cause than those with low fitness, across pooled evidence from 199 cohort studies and over 20.9 million observations. Every 1-MET higher level of fitness was linked to an 11% to 17% lower risk of all-cause death and an 18% lower risk of heart failure.

⚠What it shows: About as convincing as follow-the-people evidence gets: pooled results from 199 studies tracking over 20.9 million observations all point the same way. Because researchers measured fitness and then watched what happened, it shows a powerful link rather than proof that raising your fitness directly causes a longer life.Meta-analysis · Br J Sports Med

Exercise may moderately reduce symptoms of depression in adults, and in head-to-head trials it performed about as well as psychological therapy or antidepressant medication, according to a Cochrane review of 73 randomised trials in at least 4985 people.

⚠What it shows: Big on quantity, weaker on quality: results pooled from 73 trials in at least 4985 adults show a moderate benefit, but most trials had design flaws and the effect shrank in the most rigorously run ones. The comparisons with therapy and medication rest on a handful of small trials, so call it a rough draw rather than a verdict.Meta-analysis · Cochrane Database Syst Rev

Smartphone apps for self-monitoring blood pressure behaviours lowered systolic blood pressure by an average of just 1.64 mmHg across 15 randomised trials in 7415 people treated for hypertension. Apps that added tailored advice achieved a 2.92 mmHg drop, against 0.72 mmHg without it.

⚠What it shows: Believable but modest: results pooled from 15 trials where people were assigned at random, covering 7415 people on blood pressure treatment. The average drop is small, and the bigger effect from tailored advice comes from comparing groups of studies with each other rather than from a direct head-to-head test.Meta-analysis · JMIR Mhealth Uhealth