Study archive · page 6

A systematic review of 92 human studies found that while sleep disturbance may be linked to changes in blood pressure, stress hormones, blood sugar and other biomarkers, the certainty of the evidence is very low, and no threshold of disturbance at which harm begins could be identified.

⚠What it shows: The headline here is uncertainty itself: a sweep of 92 human studies on disturbed sleep found possible links to blood pressure, hormones and blood sugar, but the studies disagreed with each other and many were at risk of bias. Nobody can yet say how much disturbance, or for how long, starts to matter.Systematic review · J Sleep Res

A sweep of 17 randomised trials found acetazolamide, a drug best known for altitude sickness, generally improved overnight oxygen levels and reduced breathing disruptions in sleep apnoea. In obstructive sleep apnoea at low altitude, reported reductions in breathing-event rates ranged from roughly 42 to 69.4%, though effects varied widely by context.

⚠What it shows: Encouraging but not yet a treatment you can count on: the 17 trials disagreed with each other so much that their results could not be combined into one number, and the size of the benefit swung widely between types of apnoea and settings. The reviewers themselves say large, long-term trials are needed to work out dosing and who actually responds.Review · Int Orthod

In pooled driving tests in healthy adults, most sleep medications at approved doses left next-morning driving no worse than placebo. Drugs that did impair it included zolpidem 10 mg, triazolam 0.5 mg, ramelteon 8 mg, daridorexant 100 mg and mirtazapine, while most medications beat zopiclone.

⚠What it shows: Treat this as a useful map, not a verdict: the authors rated their own confidence in the evidence as low or very low. It pools 22 trials of next-morning driving tests, almost all in healthy volunteers rather than people with insomnia.Meta-analysis · Eur Neuropsychopharmacol

Pooled results from nine randomised trials in 1239 adults found that sleep restriction therapy on its own reduced insomnia severity, shortened the time to fall asleep and improved sleep efficiency, sleep quality and time awake during the night, while cutting total sleep time as the treatment intends. At follow-up the gains in insomnia severity, sleep efficiency and night-time waking held, but faster sleep onset did not.

⚠What it shows: About as solid as insomnia evidence gets in the short term: results pooled from nine trials in 1239 adults all pointed the same way. The wrinkle is durability, since the benefit for falling asleep faster had faded by longer follow-up and total sleep time drops at first by design.Meta-analysis · J Sleep Res

A meta-analysis of 13 randomised trials found that back braces and other lumbar orthoses, worn for two weeks or longer, significantly reduced low back pain and its associated disability.

⚠What it shows: Encouraging but light on detail: results pooled from 13 trials of people with low back pain, and the paper reports that pain and disability improved without saying by how much. Braces were often used alongside other therapies, so the brace alone may not deserve all the credit.Meta-analysis · J Back Musculoskelet Rehabil

Exercise therapy for tendinopathy works without the tendon visibly changing: a review of 20 studies in 625 patients found strong evidence refuting structural change as the explanation for improvement after eccentric training. Only heavy-slow resistance training showed moderate evidence of structural change.

⚠What it shows: Strong on what it rules out, silent on what the real mechanism is: a sweep of 20 studies covering 625 tendinopathy patients, though the studies were too varied to pool their numbers. The searches ran to 2012, so newer imaging work could refine the picture.Systematic review · Br J Sports Med

Pooled results from 28 randomised trials covering 2,494 people with fibromyalgia found aerobic exercise modestly reduced pain, fatigue, depressed mood and quality-of-life limitations and improved physical fitness. Pain fell with both land-based and water-based exercise of slight to moderate intensity, two or three times per week.

⚠What it shows: About as good as this evidence gets: results pooled from 28 randomised trials in 2,494 people with fibromyalgia, and the improvements held up when the authors tested for bias, though each one was modest in size. Few trials described their exercise programmes in detail, so which format works best is still unresolved.Meta-analysis · Arthritis Res Ther

In pooled data from 11 cohort studies of 778,929 people, total dairy intake was not linked to cancer mortality, but whole milk in men was tied to higher prostate cancer death risk, rising 43% per daily serving.

⚠What it shows: Reassuring on dairy overall, but treat the milk finding with care: these studies followed people rather than testing diets on them, so heavy milk drinkers may differ in other ways. The prostate signal came from a subgroup and only just reached statistical significance.Meta-analysis · Nutr J

Pooled results from 11 trials in 602 adults with overweight or obesity found that adding time-restricted eating to exercise produced modest extra losses of 1.64 kg body weight and 1.30 kg fat mass compared with exercise plus a control diet, without reducing fat-free mass. Insulin resistance also improved, while cholesterol, blood pressure, fasting glucose and HbA1c did not differ between groups.

⚠What it shows: Real but modest: results pooled from 11 trials in 602 adults with overweight or obesity, and the extra losses were measured in single kilograms and centimetres. Most trials were short, so nobody knows whether the edge holds over the long run.Study · Nutr Rev

A multilevel meta-analysis of 34 studies covering 692 female participants found creatine supplementation produced a small overall benefit (Hedges g 0.23), with small effects on performance (g 0.30) and body composition (g 0.24), though the certainty of the evidence was rated low to very low.

⚠What it shows: Encouraging direction, shaky confidence: results pooled from 34 studies covering 692 female participants point to small benefits, but the reviewers rated the certainty low or very low throughout. Protocols and outcomes varied so much that no firm sport-specific conclusion is possible yet.Systematic review · Front Nutr

Pooled results from four randomised trials suggest chilli pepper (Capsicum annuum) supplements lowered total cholesterol by about 4.98 mg/dL in adults with metabolic syndrome and related conditions, while triglycerides, HDL cholesterol and LDL cholesterol did not change.

⚠What it shows: Take it lightly: only four trials could be pooled, and the single change found was a small dip in total cholesterol while triglycerides, HDL and LDL all stayed put. Larger trials in people with genuinely raised blood fats are needed before this means much.Meta-analysis · Phytother Res

In pooled data from 19 randomised trials covering 38,160 adults with type 2 diabetes, semaglutide was associated with 19% higher odds of developing any neoplasm and 28% higher odds of benign growths. The link with malignant tumours was not statistically significant.

⚠What it shows: Solid as drug evidence goes: results pooled from 19 trials in 38,160 adults with type 2 diabetes, and the extra risk showed up mainly in benign growths, with the apparent link to cancer small enough that chance cannot be ruled out. Trials ran at most 156 weeks, so risk over decades remains unknown.Systematic review · Front Pharmacol

Pooled results from 144 randomised studies in 2216 adults suggest the best way to break up sitting depends on the goal: frequent activity bouts did most for blood glucose, longer or more intense bouts were needed to lower triglycerides, and moderate to vigorous intensity improved blood vessel function. Activity breaks also lowered insulin and both systolic and diastolic blood pressure in these short-term experiments.

⚠What it shows: Convincing for the hours after you move, unproven for the years ahead: results pooled from 144 short lab experiments covering 2216 adults, so whether these acute improvements add up to lasting health benefits was not tested here.Meta-analysis · Eur J Prev Cardiol

A 5-day fast in 38 healthy adults cut body mass index mostly through fat loss, with improvements partially persisting 12 weeks later; a model built on baseline gut microbiome and clinical measures predicted who kept the weight off, and it held up in three independent fasting cohorts.

⚠What it shows: Early but intriguing: one small trial in 38 healthy adults, so the prediction idea needs bigger groups before anyone plans a fast around their gut bacteria. It helps that the model held up in three separate fasting cohorts, including people fasting far longer.RCT · Genome Med

Pooling seven randomised placebo-controlled trials in healthy adults, probiotic supplementation produced a small but statistically significant improvement in symptoms of depression, anxiety and perceived stress compared with placebo.

⚠What it shows: Decent evidence for a small effect: results pooled from seven trials in healthy volunteers, all of which passed a quality check. Everyone studied was healthy, so this says nothing about treating diagnosed depression or anxiety.Meta-analysis · J Altern Complement Med

Pooled results from 65 randomised trials found that digitally delivered cognitive behavioural programmes modestly reduced alcohol and drug consumption and improved some psychosocial outcomes compared with minimal treatment, and cut consumption when added to usual care. They were no more effective than therapist-delivered therapy or other active interventions.

⚠What it shows: Solid but modest: results pooled from 65 trials in adults using alcohol or other drugs, and the average benefit was small, with the studies often disagreeing with each other. The apps matched, rather than beat, therapy delivered by a human.Meta-analysis · JMIR Ment Health

Pooled data from 32 studies of 1629 healthy adults show that acute doses of alcohol produce small to medium impairments in working memory, with the deficits most pronounced on tasks requiring active mental manipulation. The size of the effect varied with dose, task type, the control condition used and the sex mix of the sample.

⚠What it shows: About as solid as lab evidence gets: results pooled from 32 studies covering 1629 healthy adults, and the average study quality was good. These were one-off measured doses in controlled settings, so real-world drinking situations may play out differently.Meta-analysis · Psychopharmacology (Berl)

Pooled results from twelve randomised trials show that adding blood flow restriction cuffs to low or moderate intensity walking produced small to moderate extra gains in cardiovascular fitness, leg strength, muscle size and functional performance compared with identical walking without cuffs.

⚠What it shows: Promising but shaky: results pooled from twelve trials, and cuffed walking beat plain walking on every outcome measured. The authors rate the underlying studies as low to very low quality with a high risk of bias, so the size of these gains deserves caution.Review · J Sports Sci

A meta-analysis of 13 randomised trials in 214 physically active people found cold water immersion relieved delayed onset muscle soreness better than body cryotherapy within 24 hours of strenuous exercise, while cryotherapy showed only a small and fragile edge in jump performance recovery at 24 hours. By 48 hours neither method held a significant advantage.

⚠What it shows: Fairly solid for the soreness result: it pools results from 13 trials of 214 active people, and those trials agreed with each other on soreness. They disagreed on jump performance, so cryotherapy's small edge there deserves little weight.Meta-analysis · Medicine (Baltimore)

Across 22 randomised trials, adding protein supplementation to exercise training improved frailty status, lean mass, leg strength and walking speed in frail older adults. Protein paired with resistance training gave the clearest lean mass gains.

⚠What it shows: Solid for direction, modest in size: results pooled from 22 trials where frail older people were assigned at random to protein plus exercise or a comparison group. Study quality averaged 6.7 out of 10 and the programmes varied, so the best formula is still unsettled.Meta-analysis · Nutrients

A review of 15 randomised trials in 251 people maps how acute exercise moves testosterone: lifting raises it more and for longer than cardio, after which levels typically recover within 1 h, while extended high-intensity exercise can suppress testosterone during recovery for up to 72 h as cortisol works against it.

⚠What it shows: Useful as a map rather than a verdict: the 15 trials covered just 251 people between them and were reviewed precisely because earlier results disagreed. Every change described is short-lived, so nothing here shows exercise raising resting testosterone for good.Systematic review · PeerJ

A pooled analysis found no association between fine particulate air pollution (PM2.5) and ovarian cancer, and only a weak elevation in endometrial cancer risk. Across 145 studies, research on environmental chemicals and these cancers remains sparse.

⚠What it shows: Reassuring but thin: the ovarian result pools just four studies and the endometrial one five, so the picture could shift as more research lands. The authors themselves call the wider literature on these cancers sparse.Review · Environ Res

Pooled results from nine randomised trials in 571 people suggest Tai Chi lowers the inflammatory marker TNF-alpha, and that IL-6 and CRP also fall when the amount of practice is high. Low amounts of Tai Chi showed little effect on IL-6 and CRP.

⚠What it shows: An early signal, not a verdict: results pooled from nine trials totalling 571 people, which the authors rate as relatively low quality with findings that disagreed with each other. The suggestion it could help in COVID-19 goes well beyond what these trials tested.Meta-analysis · Ann Palliat Med

In a meta-analysis of 12 studies of adults, step counts from ActiGraph research monitors varied widely with device model, wear location and filter settings: one model recorded 123% of another's steps on the hip, wrist wear captured more steps than hip wear, and against a reference ankle device counts ranged from 73% to 178% depending on setup.

⚠What it shows: Solid on the main point but about research devices: results pooled from 12 studies of adults wearing ActiGraph monitors, a brand used in research rather than sold on the high street. The exact percentages apply to these devices only, so your own tracker's error will differ.Meta-analysis · Med Sci Sports Exerc