Study archive · page 11

In a meta-analysis of six randomised trials with 236 people with obstructive sleep apnoea, exercise programmes lasting more than 12 weeks reduced lighter stage N2 sleep and increased deep stage N3 sleep, while total sleep time, sleep efficiency and time to fall asleep did not change.

⚠What it shows: Encouraging but far from settled: pooled results from six trials in 236 people, and the deep sleep shift only reached significance in programmes longer than 12 weeks or once heart disease patients were set aside.Meta-analysis · J Sci Med Sport

When 63 people with insomnia were randomly given fabricated feedback about their previous night's sleep, those told their sleep efficiency had been poor reported more sleepiness and fatigue and less alert thinking by the evening than those told it had been good. Their scores on a basic reaction-time test and a measure of sleep-related attentional bias did not differ.

⚠What it shows: Suggestive rather than settled: one experiment in which 63 people with insomnia were randomly assigned a fake sleep score for a single day, so it says nothing about long-term tracker use or about people who sleep well. Everything that shifted was how participants rated themselves; their actual reaction-time performance barely moved.Study · J Sleep Res

In a pooled comparison of 15 studies covering 1,877 overweight or obese people with moderate-to-severe sleep apnoea, GLP-1 drugs cut the apnoea-hypopnoea index by 15.28 events per hour versus placebo, lowered BMI by 1.78 points and nudged up blood oxygen levels. The evidence was rated moderate certainty at best, and comparisons against exercise were mostly indirect.

⚠What it shows: A promising lead, not a verdict: results pooled from 15 studies with 1,877 people, none of the estimates reached high certainty, and the drug versus exercise comparisons rested largely on indirect evidence. The authors themselves call their ranking hypothesis-generating.Meta-analysis · Front Endocrinol (Lausanne)

In a 26-week trial of 646 adults with sleep apnoea who could not use CPAP, a pill combining aroxybutynin and atomoxetine cut breathing events by a model-estimated 44.1% from baseline against 17.6% on placebo, a difference of 4.0 events per hour. Fatigue scores showed no significant improvement, and 21.2% of those on the drug stopped because of side effects.

⚠What it shows: One large trial, and a serious one of its kind: 646 adults with mild to severe sleep apnoea who could not tolerate CPAP were assigned at random to the pill or a placebo for 26 weeks. Worth knowing that the drug's maker ran the trial and employs several of the authors, and that the average drop of 4.0 events an hour came with no significant change in fatigue.RCT · Am J Respir Crit Care Med

In pooled randomised trials, Pilates produced large improvements in pain and everyday function for women with chronic low back pain, and postmenopausal women benefited as much as younger women.

⚠What it shows: Consistent benefit, but treat the size with care: pooled effects this large often shrink as bigger trials arrive, and the comparison groups ranged from doing nothing to other treatments, so some of the gain may reflect structured exercise of any kind. The review also does not report how many women took part overall.Study · J Back Musculoskelet Rehabil

Across 46 randomised trials in 2,860 people with knee osteoarthritis, isokinetic strength training improved knee muscle strength, range of motion and knee function scores, and reduced pain and stiffness compared with control treatments.

⚠What it shows: Encouraging but shaky: results pooled from 46 trials in 2,860 people with knee osteoarthritis all pointed the same way, yet the review's own authors rated the underlying evidence as poor because the trials disagreed with each other and many gave imprecise results. Bigger, better trials are still needed before this counts as settled.Meta-analysis · J Orthop Surg Res

Pooled results from nine randomised trials found hip strengthening improved balance reach distances, self-reported ankle stability scores and hip abductor strength in people with chronic ankle instability. In subgroup analysis, the balance gains only appeared when hip work was added to ankle rehabilitation; hip strengthening on its own beat neither no training nor ankle rehab alone.

⚠What it shows: Suggestive rather than settled: results pooled from nine trials that the reviewers describe as small and low quality, so the size of any benefit is uncertain. Everyone taking part already had a repeatedly unstable ankle, so this says nothing about preventing a first sprain.Meta-analysis · Complement Ther Med

A systematic review of randomised placebo-controlled trials found testosterone therapy improves sexual function in postmenopausal women, with some studies also reporting higher bone mineral density and lower HDL cholesterol; long-term safety remains unclear.

⚠What it shows: The sexual function benefit rests on several trials where women were assigned at random to testosterone or placebo, so it deserves real weight, but this sweep of the studies is clear that long-term safety has not been established. The drop in protective HDL cholesterol seen in some studies is the other reason for caution.Systematic review · Clin Endocrinol (Oxf)

Pooled results from 4 randomised trials in 609 postmenopausal women found that isoflavone supplements combined with exercise training did not significantly increase bone mineral density at any site measured, and the combination added nothing over exercise alone or isoflavones alone.

⚠What it shows: A believable null built on thin ground: results pooled from just 4 trials in 609 postmenopausal women, and the authors note the trials were small with some method problems. It cannot rule out small effects that bigger trials might detect.Meta-analysis · J Chin Med Assoc

A meta-analysis of 23 randomised trials in 1880 postmenopausal women found phytoestrogen supplements made no difference to body weight, BMI, waist or hip circumference, fat mass or body fat percentage. In women with pre-existing metabolic disorders such as prediabetes, type 2 diabetes, prehypertension or high cholesterol, supplementation was associated with a 0.78 kg weight gain.

⚠What it shows: A fairly solid null: results pooled from 23 trials in 1880 postmenopausal women showed no change in weight, waist or body fat overall. The weight gain seen in women with metabolic disorders comes from smaller subgroup comparisons, so treat it as a warning flag rather than proven harm.Meta-analysis · Maturitas

A meta-analysis of 47 randomised trials found that group-based diet and physical activity programmes helped overweight and obese adults lose 3.44 kg more than controls at 12 months, with 2.56 kg of advantage still present at 24 months; programmes explicitly targeting weight loss, and men-only groups providing feedback and dietary goals, were associated with greater effectiveness.

⚠What it shows: About as solid as weight-loss evidence gets: results pooled from 47 trials in overweight and obese adults show a clinically meaningful loss versus controls. The programmes varied hugely in design, though, so exactly which ingredients make one group work better than another is still unclear.Meta-analysis · Appl Psychol Health Well Being

Across roughly 10 years of repeated drink questionnaires and hip scans in 9704 women aged 65 and over, tea drinking was associated with slightly higher total hip bone density (0.718 versus 0.715 g/cm2, a difference of 0.003), while coffee showed no overall association with hip or femoral neck density. Only above five cups of coffee a day did the analysis hint at lower density.

⚠What it shows: Treat it as a hint rather than a verdict: researchers followed 9704 women aged 65 and over for about 10 years, repeatedly asking what they drank and scanning their hips, so they watched existing habits rather than assigning drinks. The tea difference was 0.003 g/cm2, small enough to mean little for any one woman, and the coffee signal above five cups a day came from a curve-fitting analysis rather than the main comparison.Study · Nutrients

Pooled results from 28 trials in 1,574 adults with type 2 diabetes linked vitamin C supplementation to a 6.27 mmHg drop in systolic blood pressure and a 0.54% fall in HbA1c, but most trials were short and small, and the authors say it cannot yet be recommended as a therapy.

⚠What it shows: A lead, not a therapy: results pooled from 28 trials in 1,574 adults, but most ran under six months with fewer than 100 people each, and the blood sugar finding rests on very low certainty evidence.Meta-analysis · Diabetes Care

A systematic review of seven randomised trials in 192 sedentary or immobilised people found that five of the six studies measuring muscle size reported higher muscle volume and mass with omega-3 supplementation and no training at all, and both studies measuring muscle protein synthesis found a higher rate.

⚠What it shows: Early and small: a sweep of seven trials covering just 192 sedentary or immobilised people, so nobody knows whether the effect adds anything for people who actually train.Systematic review · Crit Rev Food Sci Nutr

A narrative review of 18 randomised trials of positive psychology programmes, including gratitude journaling, optimism training and mindfulness delivered in person, in hybrid form or digitally over 6 to 12 weeks, reported consistent improvements in cardiovascular risk factors, health behaviours and medication adherence. Whether those gains persist is still unclear.

⚠What it shows: Encouraging but unsettled: the authors gathered 18 trials in which people were assigned at random, plus single-arm studies and pooled analyses, then read across them rather than combining the numbers formally. They are open that nobody has established the right dose, that the populations studied were not diverse, and that longer-term outcomes beyond the usual 6 to 12 weeks remain uncertain.Review · Cardiol Clin

Pooled results from eight randomised trials in 471 healthy participants, 81% of them female, found high-intensity interval training produced no significant reduction in anxiety or depressive symptoms compared with a non-exercising control group. The reviewers rate the certainty of that evidence as very low.

⚠What it shows: A real null, but a thin one: results pooled from only eight trials with 471 healthy participants, and the reviewers themselves call the certainty very low. Everyone involved was already healthy, so this says nothing about people living with diagnosed depression or anxiety.Meta-analysis · Scand J Med Sci Sports

Across 25 randomised controlled trials, people taking GLP-1 receptor agonists such as semaglutide and liraglutide showed no higher rates of suicidal behaviour than controls, and the null held for suicidal ideation, attempts and completed suicides alike.

⚠What it shows: About as reassuring as trial evidence gets: results pooled from 25 trials where people were assigned at random showed no rise in any form of suicidal behaviour, on any of the drugs, and the trials did not disagree with each other. Suicidal behaviour is rare, though, so very small effects in specific groups cannot be fully ruled out.Meta-analysis · J Diabetes

People with so-called metabolically healthy obesity move more, sit less and are fitter than those with unhealthy obesity, yet pooled cohort data still show a 24-33% higher risk of death and cardiovascular disease compared with metabolically healthy normal-weight people. The one study to test cardiorespiratory fitness directly suggests fitness differences may largely explain that gap.

⚠What it shows: Strong as summaries go but not proof: it pools dozens of studies that followed people rather than testing anything on them, so it cannot show cause, and the extra death risk sat on the edge of significance. Only a single study examined whether fitness explains the difference.Systematic review · Prog Cardiovasc Dis

In a pooled comparison of 12 randomised trials with 2957 healthy adults, wearables paired with online coaching and social media support significantly reduced total sedentary time, outperforming control conditions, traditional non-wearable interventions and wearables used alone; no wearable strategy significantly reduced prolonged unbroken sitting.

⚠What it shows: Reasonably convincing for total sitting time: results pooled from 12 trials in 2957 healthy adults consistently ranked a wearable plus online support as the best strategy. Very little research looked at long unbroken stretches of sitting, so that part of the question remains genuinely open.Meta-analysis · BMJ Open

Digital tools such as apps, trackers and online coaching helped adults with diabetes take more daily steps and break up their sitting time more often, in a pooled analysis of 13 randomised trials with 980 participants. Programmes packing in more behaviour change techniques, or adding a face-to-face element, showed the clearest gains.

⚠What it shows: Encouraging rather than settled: results pooled from 13 trials with 980 adults, and the authors themselves call for more high-quality studies. Nearly all of the evidence covers type 2 diabetes, so it may not apply to type 1.Meta-analysis · Int J Behav Med

Pooled evidence from 8 long-term and 34 single-session studies in healthy adults suggests that adding blood flow restriction cuffs to heavy resistance training, meaning loads of at least 60% of maximal strength, does not further improve muscle strength or size compared with conventional heavy training, likely because the physiologic demands are comparable.

⚠What it shows: A believable null, but the authors themselves rate the certainty of the pooled evidence as very low: only 8 studies tracked actual strength and size adaptations, all in healthy adults. The single-session findings on effort and repetitions completed also disagreed with each other.Meta-analysis · J Strength Cond Res

Across seven randomised trials with 558 community-dwelling adults aged 60 and over with sarcopenic obesity, exercise alone or with dietary supplements increased grip strength by 1.30 kg, gait speed by 0.05 m/s and appendicular muscle mass by 0.40 kg, while cutting waist circumference by 1.40 cm and total fat mass by 1.77 kg.

⚠What it shows: A consistent signal with modest sizes: results pooled from seven trials in 558 older adults, and the gains, like 0.05 m/s of walking speed, are real but small. The authors still call for better-designed trials to work out which exercise or supplement mix does the work.Meta-analysis · Maturitas

Probiotic supplements reduced eczema severity in adults in a pooled analysis of nine trials with 402 participants, with benefits in both the short and long term, and improved long-term quality of life. A mixture of Lactobacillus salivarius LS01 and Bifidobacterium BR03 ranked as the most promising combination.

⚠What it shows: Early and small: results pooled from nine trials with 402 adults, and the short-term benefit only just crossed the line for significance. The best-strain ranking comes from statistical modelling, not head-to-head tests.Meta-analysis · J Dermatolog Treat

A meta-analysis of 30 randomised trials with 30,263 participants found vitamin D supplementation had no significant effect on preventing acute respiratory infections. Apparent benefits from daily and short-term dosing disappeared when only the highest quality trials were analysed, and publication bias was detected.

⚠What it shows: A convincing null: pooling 30 trials covering 30,263 people found no overall protection against respiratory infections, and the hints of benefit from daily dosing vanished in the best-run studies. Signs that unflattering results went unpublished weaken even those hints further.Meta-analysis · Nutrients