Study archive

Structured exercise programmes reduced university students' Pittsburgh Sleep Quality Index scores by 2.43 points versus non-exercise controls in a meta-analysis of seven trials with 701 participants, but very low evidence certainty and heavy variation between studies temper the finding.

⚠What it shows: Meta-analysis of only 7 RCTs (701 university students) with very high heterogeneity (I²=91.7%); the 95% prediction interval crossed zero and GRADE certainty was very low to low, so the effect may not generalise.Meta-analysis · Front Public Health

A meta-analysis of ten randomised trials in healthy adults found warm-water immersion thermotherapy increased total sleep time (SMD 0.63) and sleep efficiency (SMD 0.64), but did not significantly change sleep latency or any Pittsburgh Sleep Quality Index score.

⚠What it shows: Meta-analysis of ten RCTs in healthy adults; both significant effects were borderline, water temperature, timing and immersion protocols differed between trials, and subjective sleep quality did not improve on any PSQI component.Meta-analysis · Int J Biometeorol

A meta-analysis of 12 randomised trials in 179 team-sport athletes found early-afternoon napping produced a large improvement in endurance and a moderate gain in agility versus no nap, with naps longer than 60 minutes showing the biggest effects.

⚠What it shows: Meta-analysis of 12 RCTs but only 179 athletes in total; findings apply after presumed normal night sleep, and effects on cognitive and technical performance remain uncertain.Systematic review · Front Physiol

A systematic review of ten randomised trials in 420 shift workers found that eight reported significant improvements in sleep measures such as PSQI score, total sleep time, sleep efficiency or wake after sleep onset following structured exercise, and three reported better alertness, reaction time or short-term memory. Effect sizes were inconsistent and 80% of the trials had some concerns or high risk of bias.

⚠What it shows: Narrative systematic review of ten small RCTs, 420 adult shift workers (60% in healthcare); interventions and outcome measures were too heterogeneous to pool, 80% of trials had bias concerns, and adherence was hampered by fatigue and irregular schedules.Systematic review · Front Public Health

In a meta-analysis of 20 studies in healthy adults, a single hot-water immersion session raised heart rate by 28 bpm and lowered diastolic blood pressure by 5 mmHg, while repeated immersion reduced resting heart rate by 3 bpm. Only one study reported cardiorespiratory fitness data, and the authors call the overall cardiovascular benefits inconclusive.

⚠What it shows: Meta-analysis of 20 studies in healthy adults; only one study measured cardiorespiratory fitness, other cardiovascular markers showed no significant effects, and the authors describe the benefits as inconclusive.Meta-analysis · Physiol Rep

Across 15 randomised trials in people with prehypertension or hypertension, exercise significantly improved aerobic capacity (SMD 0.88; 95% CI 0.61-1.15), with multicomponent training showing the strongest effect.

⚠What it shows: Meta-analysis of 15 RCTs in adults with prehypertension or hypertension; protocols varied widely and the subgroup findings on programme length, session frequency, session duration, weekly minutes and supervision are labelled exploratory by the authors, not a prescription.Study · J Hypertens

In a dose-response network meta-analysis of 28 randomised trials (2084 adults with metabolic syndrome), aerobic exercise, combined aerobic-plus-resistance training, mind-body exercise and HIIT were all associated with lower systolic blood pressure, with HIIT at 690 metabolic equivalent minutes per week the most effective dose at -5.07 mmHg. Effects appeared larger in participants with hypertension and in programmes lasting 16 weeks or more.

⚠What it shows: Bayesian network meta-analysis of 28 trials in 2084 adults with metabolic syndrome (mean age 55, mean BMI 31.68); the best doses are modelled estimates with wide credible intervals, not head-to-head trial results, and the authors say the dose ranges still need validating.Meta-analysis · Maturitas

In a 6-week randomised trial of 45 physically active male university students, twice-weekly 4 × 4 min HIIT improved VO2max and time to exhaustion about as much as three weekly sessions, while once weekly produced only a small VO2max rise; completion rates were 100%, 93% and 80% for the 1-, 2- and 3-session groups.

⚠What it shows: One 6-week RCT, n=45 physically active male university students (15 per group); homogeneous sample and short intervention period, so the 'optimal frequency' conclusion needs testing in larger, more diverse groups.Study · Front Physiol

A systematic review found only four randomised trials of meal-timing changes in shift workers: nearly 11 hours of overnight fasting was linked to lower post-meal glucose and non-esterified fatty acid responses versus eating and snacking at night, while insulin resistance improved in just one trial and HDL-C changed in none.

⚠What it shows: Systematic review of only four RCTs in adult shift workers, with different fasting windows and short durations; the authors explicitly question the robustness and generalisability of these preliminary positive signals.Review · Nutr Health

Pooling 30 randomised controlled trials with 37 072 overweight or obese patients, GLP-1 receptor agonists lowered systolic blood pressure by 3.37 mmHg and diastolic by 1.05 mmHg compared with placebo. The effect was consistent across diabetes status, drug formulation, follow-up duration and route of administration, with the exception of exenatide subgroups.

⚠What it shows: Meta-analysis of 30 RCTs in 37 072 overweight or obese patients, but blood pressure was a secondary outcome in most of them and the absolute reductions are modest; these are prescription medicines whose use and monitoring sit with a clinician.Meta-analysis · J Hypertens

A meta-analysis of placebo-controlled randomised trials found GLP-1 receptor agonists modestly reduced LDL cholesterol (mean difference -2.93 mg/dL) and cut total cholesterol by about 7 mg/dL, while triglycerides, VLDL-C and HDL-C did not change significantly. Meta-regression found weight change did not explain the LDL or total cholesterol reductions.

⚠What it shows: Meta-analysis of placebo-controlled RCTs with extremely high heterogeneity (I2 above 99% in the meta-regressions); the absolute lipid changes are small, most trials were not designed with lipids as the primary endpoint, and these are prescription medicines managed clinically.Meta-analysis · Curr Med Res Opin

An umbrella review of 43 systematic reviews found resistance training in adults with type 2 diabetes was associated with HbA1c reductions of roughly 0.3-0.6 percentage points, systolic blood pressure drops of about 4-7 mmHg, small lipid improvements and large gains in muscle strength.

⚠What it shows: Umbrella review of 43 reviews with moderate GRADE certainty for key outcomes, but substantial heterogeneity, heavy overlap of primary trials across reviews, and frequent methodological flaws in the included syntheses.Systematic review · Metabolism

In a network meta-analysis of 235 randomised trials covering 20,980 untrained older adults, whey protein plus resistance training ranked as the most effective combination for muscle mass (SMD 1.29) and leg strength (SMD 1.16), while whey plus multicomponent training ranked highest for mobility tasks.

⚠What it shows: Network meta-analysis of 235 RCTs in untrained older adults, so most rankings rest on indirect comparisons; certainty was moderate for muscle mass and leg strength, effects were moderated by participant condition and protein dose, and findings may not transfer to already-trained people.Meta-analysis · Nutrients

A meta-analysis of 14 randomised trials (327 intervention and 284 control participants with metabolic syndrome) found berry consumption raised HDL cholesterol by 1.771 mg/dL and left almost everything else unchanged: LDL-C, total cholesterol, triglycerides, leptin, BMI, waist circumference, blood pressure, TNF-α, IL-6, CRP, adiponectin, HbA1c, HOMA-IR, glucose and insulin were all unaffected overall.

⚠What it shows: Pooled data from 14 small trials in people with metabolic syndrome, using different berries, doses and durations; the HDL rise is tiny and of unclear clinical meaning, and the longer-duration and larger-study signals come from subgroup analyses that are exploratory.Meta-analysis · Int J Food Sci Nutr

A systematic review of 15 randomised trials in 1161 pre- and postmenopausal women found animal proteins such as whey, meat and milk cut body fat more consistently than soy over 8-16 weeks — but mainly when combined with calorie restriction and exercise, and benefits of raising protein alone could not be confirmed.

⚠What it shows: Systematic review of 15 RCTs (1161 women), but effects were strongly confounded by co-interventions like energy restriction and exercise, and results across trials were heterogeneous.Systematic review · FASEB J

In a meta-analysis of 17 studies, plaza dancing — group aerobic dance performed to music in public spaces — significantly reduced body weight, BMI, body fat percentage, resting heart rate, systolic blood pressure, cholesterol and triglycerides in healthy women aged 45 and over, and improved vital capacity, with BMI benefits strongest in programmes of 12 weeks or less.

⚠What it shows: Meta-analysis of 17 studies in healthy middle-aged and older women, drawn heavily from Chinese databases; effect sizes were mostly small to moderate and the authors call for larger, long-term follow-up trials.Meta-analysis · Front Public Health

A meta-analysis of 17 randomised trials found that leisure-time physical activity alleviated depressive symptoms in menopausal women (SMD −1.23), with results similar regardless of exercise intensity, form, duration or supervision.

⚠What it shows: Meta-analysis of 17 RCTs with wide confidence intervals suggesting heterogeneity across studies; depression was measured with varied scales, and the authors call for further high-quality trials to confirm the findings.Systematic review · Front Psychiatry

A meta-analysis of 29 studies in 726 physically active premenopausal women found combined oral contraceptive use made little to no difference to heart rate, respiratory measures, IL-6, BMI or metabolic markers compared with non-users — though the certainty of the evidence was low.

⚠What it shows: Meta-analysis of 29 studies (726 active women), but certainty was rated low across all outcomes due to small samples, high risk of bias and varied methods — reassuring, not definitive.Systematic review · Front Physiol

Across 86 randomised trials (n=5,670), many exercise-induced gains — including body mass, waist circumference and maximal oxygen uptake — remained better than baseline even after training stopped, with residual benefits fading in an outcome-specific, time-dependent way while muscle mass, resting heart rate and LDL cholesterol benefits did not persist.

⚠What it shows: Meta-analysis of 86 RCTs (n=5,670); residual effects varied widely by outcome and detraining duration, and several markers — resting heart rate, LDL cholesterol, muscle mass and lean body mass — were no longer significantly improved at any follow-up.Systematic review · Front Physiol

A meta-analysis of 18 randomised controlled trials in 539 adults aged 60 and over found resistance training reduced C-reactive protein (effect size -0.72), TNF-α (-0.56) and IL-10 (-3.34), and tended to reduce IL-6 (-0.59). The CRP reduction held across every subgroup, including age, training method, intensity, weekly frequency and programme duration.

⚠What it shows: Pooled data from 18 RCTs and 539 adults aged 60 and over, using blood markers rather than disease outcomes; protocols were heterogeneous and the anti-inflammatory cytokine IL-10 also fell, which complicates the story.Meta-analysis · Int J Environ Res Public Health

A systematic review of four randomised trials in 148 frail adults aged 65 or over found elastic-band resistance training over 12-28 weeks improved functionality, handgrip and lower-limb strength, especially in progressive, longer programmes; adding nutritional education further reduced frailty.

⚠What it shows: Only four RCTs with 148 participants, and three of the four carried a high risk of bias; autonomy did not improve and long-term follow-up is lacking.Systematic review · Rehabilitacion (Madr)

In a pooled analysis of six randomised controlled trials in 330 older adults, exercise on its own significantly improved serum albumin levels, while combining exercise with nutritional supplementation showed no additional significant effect. Most included trials were judged to carry a significant risk of bias and overall certainty was rated low.

⚠What it shows: Meta-analysis of only six randomised trials, 330 older adults (mean age 72.1 years, 61% female); most trials at significant risk of bias and the pooled certainty of evidence was graded low, so the effect size is unreliable.Meta-analysis · Biomolecules

A network meta-analysis of 61 randomised trials found that mind-body exercise, acutherapy, cognitive training, non-invasive brain stimulation, meditation and physical exercise all improved global cognition in older adults versus passive controls, with mind-body exercise showing the largest effect (SMD 1.384).

⚠What it shows: Network meta-analysis of 61 RCTs in older adults with and without mild cognitive impairment; confidence intervals were wide, no active intervention significantly beat another, and intervention methods varied across trials.Meta-analysis · Sci Rep

A network meta-analysis of 33 studies in 1121 adolescents and young adults found a single session of mind-body exercise was the most effective type for reducing anxiety (SMD -0.36), aerobic exercise also produced significant reductions (SMD -0.16), while resistance and multicomponent exercise did not.

⚠What it shows: Network meta-analysis of 33 studies (1121 participants); effects were small-to-moderate, certainty was moderate for mind-body exercise, and more multi-arm head-to-head trials are needed.Meta-analysis · Scand J Med Sci Sports