Study archive · page 3

Pooled results from eight randomised trials in 426 adults with heavy alcohol use or alcohol use disorder found that cognitive behavioural therapy for insomnia produced a large drop in insomnia severity, an estimated 5.51 points on the standard insomnia scale, with benefits still present at six months. Alcohol-related outcomes such as craving and drinking days showed no significant improvement.

⚠What it shows: Solid for the sleep result: results pooled from eight trials covering 426 adults, judged at low to moderate risk of bias, with the benefit holding at six months. The drinkers studied were mostly men, and alcohol outcomes were reported so inconsistently that the lack of drinking benefit is far less certain.Meta-analysis · Sleep Med Rev

In pooled results from 96 randomised trials with 18,419 participants, digital CBT reduced insomnia severity scores by 4.24 points versus control, the only one of six digital sleep therapies to reach a clinically meaningful improvement. Neurofeedback, digital brief behavioural therapy and circadian support apps showed no significant effects on any outcome.

⚠What it shows: The headline result rests on moderate-quality evidence pooled from 96 trials in 18,419 people, which is unusually strong for sleep tech. Most other comparisons were less certain, and only 2 of the trials were judged fully rigorous, so rankings between app types stay tentative.Meta-analysis · BMC Med

Pooled data from studies following women from pregnancy to one year postpartum found 36% followed a nonoptimal sleep quality trajectory and 22% a nonoptimal sleep duration trajectory. These high-risk sleep paths were associated with worse outcomes for both mothers and infants.

⚠What it shows: Solid for describing patterns: it pools studies that each followed at least 100 women with repeated sleep checks from pregnancy to a year after birth. It tracks how sleep unfolds rather than testing any fix, so it cannot say poor sleep caused the worse outcomes it was linked to.Meta-analysis · Sleep

Pooled results from 22 trials found exercise therapy improved global cognitive function and reduced depression in older adults with mild cognitive impairment. Aerobic exercise was the most effective type, while multicomponent and neuromotor exercise did not improve cognition, and cognitive gains grew with higher exercise frequency.

⚠What it shows: A reasonably strong signal: results pooled from 22 trials in older adults with mild cognitive impairment showed a meaningful cognitive benefit. The depression finding rests on just 7 of those trials, and the studies varied in the exercise types and doses they tested.Meta-analysis · Arch Gerontol Geriatr

In pooled data from 34 studies of 21,024 adults with type 2 diabetes, 43.82% screened positive for anxiety symptoms and 41.83% for depressive symptoms. The estimates varied widely between studies and reflect questionnaire results, not diagnosed disorders.

⚠What it shows: Treat the exact figures loosely: this pools 34 studies covering 21,024 people, but the studies disagreed with each other so much that the authors warn the averages are rough summaries rather than universal rates. These were symptom questionnaires, not clinical diagnoses.Review · J Diabetes Metab Disord

Pooled results from 12 randomised trials covering 757 adults with major depression found exercise training reduced depressive symptoms and raised circulating BDNF and kynurenine. Improvements in BDNF, kynurenine and interleukin-6 tracked with the drop in symptoms, hinting at how exercise works as an antidepressant.

⚠What it shows: The mood benefit rests on pooled trials, but only 12 of them with 757 people, and the biomarker link comes from correlations across those trials rather than a direct test. Treat the BDNF story as a promising lead, not a settled mechanism.Meta-analysis · J Affect Disord

Pooled results from 20 randomised trials in 1824 people with primary osteoporosis found exercise therapy improved lumbar spine and femoral neck bone density and eased pain more than conventional treatment. The authors grade the underlying evidence as low quality.

⚠What it shows: Take the direction seriously and the size loosely: results pooled from 20 trials in 1824 people favoured exercise for bone density and pain, but the authors rate the evidence as low quality and the trials disagreed with each other a lot.Meta-analysis · Orthop Surg

Across 27 randomised trials in 965 people with type 2 diabetes, combined aerobic and resistance training improved body fat, HbA1c, fasting glucose and blood lipids whether cardio or weights came first. Direct comparisons between the two orders were not significant, so neither sequence can be declared superior.

⚠What it shows: The case for combining cardio and weights is solid here, pooled from 27 trials in 965 people with type 2 diabetes; the ordering question is not. Head-to-head comparisons between the two sequences were uncertain and rated low certainty, so the rankings are hints rather than verdicts.Meta-analysis · Endocrine

Pooled results from ten randomised trials suggest curcumin supplementation reduces post-exercise muscle soreness and the muscle damage marker creatine kinase, with modest improvements in strength and joint range of motion.

⚠What it shows: Fairly credible as supplement evidence goes: results pooled from ten randomised trials, all pointing the same modest direction. The trials were small, though, and the authors say larger, higher-quality studies are still needed to confirm the effects and settle long-term safety.Meta-analysis · Phytother Res

Omega-3 supplements did not significantly reduce deaths or the re-narrowing of treated arteries in results pooled from 29 trials of 35,144 high-risk cardiovascular patients. Larger and longer trials tended to show smaller benefits, suggesting bias may partly explain earlier positive findings.

⚠What it shows: Take this seriously but not as final: results pooled from 29 trials in 35,144 people at high heart risk found no clear survival benefit, though some modest effect remains possible. The telling detail is that the bigger and longer the trial, the smaller the apparent benefit, which points to weaker studies inflating the early enthusiasm.Meta-analysis · BMC Cardiovasc Disord

Pooled results from 13 trials in 544 people with type 2 diabetes link time-restricted eating to a 0.46 percentage point drop in HbA1c and a smaller waist. The authors rate the evidence as fragile and hypothesis-generating, with no effect found on sleep, lipids or blood pressure.

⚠What it shows: Treat this as a hint rather than a verdict: the 13 trials were small, they disagreed sharply with each other, and the authors themselves call the pooled results hypothesis-generating. One headline result, a drop in body mass index, vanished entirely when a single trial was removed.Systematic review · Front Nutr

In pooled trial data from people with type 2 diabetes, intermittent fasting lowered HbA1c, fasting glucose and body weight over the short term (under 3 months) compared with other diets, but performed no better in the long term and the metabolic benefits disappeared once the fasting stopped.

⚠What it shows: A fair verdict on durability: results pooled from 12 trials with 966 people with type 2 diabetes, though the studies disagreed with each other a fair amount, so treat the exact effect sizes loosely. The long-term data showed no advantage over other diets, and the gains vanished after people stopped fasting.Meta-analysis · Nutr Res

Probiotic supplements had no significant effect on blood levels of TMAO, a metabolite regulated by gut bacteria, in results pooled from 8 trials with 270 adults. A possible larger reduction in people under 50 remains unconfirmed.

⚠What it shows: A clear null but from thin data: pooled results from 8 trials covering just 270 adults, so a small real effect could still hide in numbers that size. The suggestion that people under 50 respond differently comes from slicing the data into subgroups and needs its own test.Meta-analysis · Clin Nutr ESPEN

Pooled results from ten randomised trials in 486 adults with depression found probiotic supplementation lowered the inflammatory markers C-reactive protein and TNF-α, with a borderline drop in IL-1β. Two other markers, IL-6 and IL-10, did not change significantly.

⚠What it shows: Reasonably solid as far as it goes: results pooled from ten trials covering 486 adults with depression. The drops were in blood markers of inflammation rather than in mood itself, and the authors say larger, longer trials are still needed to confirm the effect and find the best formulations.Review · Prog Neuropsychopharmacol Biol Psychiatry

A pooled analysis of 53 cohort studies found that long-term exposure to fine particulate air pollution (PM2.5) is associated with increased risk of premature death, with significant effects seen even below 10 µg/m3. Studies with poorer exposure measurement produced smaller effect estimates, suggesting the true harm may be underestimated.

⚠What it shows: About as robust as pollution evidence gets: 53 studies of large populations followed over time, spanning North America, Europe and Asia. It still cannot prove cause and effect on its own, and the finding that cruder pollution measurement produced smaller estimates hints the real risk may be larger than reported.Meta-analysis · Environ Res

Across more than a dozen city studies, each 100 micrograms per cubic metre rise in airborne particle concentration was associated with a 6% increase in daily deaths. The link held across warm and cold climates and survived detailed adjustment for weather, leading the author to judge a causal association the most reasonable interpretation.

⚠What it shows: A remarkably consistent signal pooled from more than a dozen city studies, though all of them tracked populations rather than testing anything, so cause cannot be proven outright; the author argues the sameness of the effect across climates makes a causal link the most reasonable reading. It also dates from 1994, when particle levels and measurement methods differed from today's.Meta-analysis · Environ Res

Across five randomised trials, probiotic, prebiotic and synbiotic supplements showed no prophylactic effect against COVID-19: infection rates and antibody levels were no different from placebo over one to three months.

⚠What it shows: A consistent null across five trials spanning exposed staff, care home residents, healthy adults and household contacts, which makes it reasonably trustworthy. The trials only ran one to three months, so longer use and other infections were never tested.Systematic review · Clin Med Res

Across 37 randomised trials, vitamin supplements showed no consistent benefit for preventing or managing COVID-19 or long COVID. The exceptions were on death rates, where vitamin C showed a benefit and vitamin D did too, but only when given regularly.

⚠What it shows: The headline null is fairly firm: 37 trials where people were assigned vitamins or a comparison at random were pooled, and most outcomes showed no clear benefit from any vitamin. The two positive signals on deaths came from trials that varied widely in doses and patients, so even the authors treat them cautiously and call for better-designed studies.Meta-analysis · Nutrients

An umbrella review spanning 15 meta-analyses and 161 studies finds post-exercise cold-water immersion speeds recovery of power, jump performance and soreness, but does not aid strength recovery and impairs strength and muscle growth gains when used routinely.

⚠What it shows: The widest sweep of this evidence yet, covering 161 studies gathered through 15 earlier pooled reviews, and the pattern is consistent. The catch is that the authors rate the underlying evidence as low quality and inconsistent, so the exact sizes of these effects remain uncertain.Systematic review · Sports Med Open

In a pooled analysis of 156 studies covering 6222 healthy adults, men had larger type I and type II muscle fibres and a greater share of fast-twitch type II fibre, while women had a greater proportion of slow-twitch type I fibre. The differences largely held regardless of age, physical activity level and muscle group.

⚠What it shows: About as firm as this kind of evidence gets, pooling 156 studies of 6222 healthy adults. A biopsy samples one small piece of one muscle, and these are averages, so plenty of individual women will be more fast-twitch than plenty of individual men.Meta-analysis · Physiol Rep

Adults with atopic dermatitis had roughly double the odds of depression and anxiety in pooled studies, and suicidal ideation was over four times more likely among adults and adolescents with the condition. Children with atopic dermatitis also showed higher odds of depression.

⚠What it shows: A consistent and sizeable link, but it comes from pooling studies that followed people rather than testing anything, so it cannot prove the skin condition causes the distress. The included studies also used differing definitions of depression and anxiety and rarely measured how severe the eczema was.Meta-analysis · J Am Acad Dermatol

Pooled rankings from 34 randomised trials in overweight or obese adults with lower-limb osteoarthritis placed diet plus resistance exercise as the most effective approach for muscle mass, resistance exercise alone for muscle strength, and aerobic exercise alone for walking speed.

⚠What it shows: Strong as evidence goes because it pools 34 randomised trials, but the rankings come from comparing treatments across different studies rather than head to head, and the trials used a mix of diets and exercise programmes, so the exact ordering is less certain than the overall message.Meta-analysis · Nutrients

Pooled results from 11 trials in 623 postpartum women found that lumbopelvic stabilising exercises significantly reduced low back and pelvic pain and disability, though quality of life scores did not significantly improve.

⚠What it shows: A solid signal for pain and function: results pooled from 11 trials in 623 postpartum women, and the benefit on pain and disability was clear. The trials were mostly short, quality of life did not improve, and the authors say longer, higher-quality studies are still needed.Meta-analysis · Biol Res Nurs

Across six studies of adults aged 18 to 29 with problematic alcohol use, a single bout of exercise was associated with a moderate short-term reduction in alcohol craving, often alongside less anxiety and negative mood.

⚠What it shows: Promising but thin: only six studies qualified, they disagreed with each other moderately, and the authors themselves say sturdier, more clinically focused research is needed before this counts as anything like treatment.Meta-analysis · Eur J Sport Sci