Study archive · page 13

A review of 40 randomised trials found acupressure, reflexology and auricular acupressure beat routine care or no treatment for insomnia, but acupressure fared only marginally better than a sham version and auricular acupressure gave equivocal results. The authors say the evidence does not allow a clear conclusion on benefit.

⚠What it shows: Treat the positive headlines with caution: across 40 trials, only 10 used a convincing fake version and every single study had at least one area at high risk of bias. When a therapy wins against no treatment but not against a sham, expectation may be doing much of the work.Meta-analysis · Sleep Med

Pooling 63 studies of healthy, non-clinical populations, sleep duration showed only a weak positive correlation with resilience (r = 0.11) while sleep quality showed a somewhat stronger one (r = 0.27); the correlation was slightly weaker in prospective studies (r = 0.18) and the authors found high publication bias for the sleep quality work.

⚠What it shows: Worth some weight: results pooled from 63 studies in healthy people rather than one small trial. Most measured sleep and resilience at the same moment, so neither is shown to cause the other, and the authors flagged that positive results were probably over-represented.Meta-analysis · Sleep Med Rev

A review of six randomised trials with 244 participants found most reported improvements in sleep onset latency, total sleep time and sleep efficiency with Yoga Nidra, a guided meditative practice, compared with controls including cognitive behavioural therapy for insomnia, progressive muscle relaxation and music-based relaxation. Most trials carried moderate to high risk of bias and adverse events were poorly reported.

⚠What it shows: Promising rather than proven: a sweep of the literature turned up only six trials with 244 people between them, and most were rated at moderate to high risk of bias because participants knew what they were getting and results were reported incompletely. Protocols and outcome measures also varied a lot, and safety data was barely reported.Systematic review · J Integr Complement Med

A review of 21 studies found sexual function was affected similarly in women and men with obstructive sleep apnoea, but effective interventions were reported only for men. Continuous positive airway pressure improved erectile and orgasmic function as well as excessive daytime sleepiness, although sildenafil was superior to CPAP for erectile dysfunction.

⚠What it shows: Treat as an evidence gap rather than a verdict: a sweep of 21 studies, of which only 6 included women and none of those tested a treatment by assigning people at random. The authors say firm generalisations cannot be made, so the women's side of this rests on very thin ground.Systematic review · Int J Clin Pract

In pooled results from five trials covering 323 adults, aerobic exercise produced no significant change in C-reactive protein, a reduction of about 3%, while body weight fell by 3.4 kg, body fat by 1.4% and maximum oxygen consumption rose by about 12%.

⚠What it shows: Solid on the fitness side, thin on the inflammation side: results pooled from five trials of 323 adults, which is a small base from which to rule anything out. Trials only had to last 4 weeks, so a longer or larger programme might still shift the inflammation marker.Meta-analysis · Metabolism

In pooled results from 17 studies covering 829,098 workers, active commuting was associated with 9% lower all-cause mortality and 15% lower cardiovascular mortality; cycling commuters showed 21% lower all-cause and 33% lower cardiovascular mortality, while low levels of active commuting were linked to no reduction in any type of mortality.

⚠What it shows: Reasonably strong, but not proof: results pooled from 17 studies following 829,098 workers who chose their own commute, so cyclists may differ from drivers in other ways. Cancer deaths were not lower with walking or cycling once the fullest adjustments were applied.Meta-analysis · Sports Med

Pooling 22 randomised trials of 552 men and women aged 60 and over, resistance training raised VO2max by an average of 1.8 ml/kg/min compared with controls, and again showed no evidence of true differences between individuals in how much they responded.

⚠What it shows: Decent evidence for a small effect: results pooled from 22 trials in 552 people aged 60 and over. The trials disagreed considerably with each other and the plausible range of outcomes ran from a slight decline to a clear gain, so improvement is not guaranteed for any one person.Meta-analysis · Sci Prog

Pooled results from 47 controlled studies found that low-volume high-intensity interval training (no more than 500 metabolic equivalent minutes per week) produced no significant difference in total body fat mass, body fat percentage or lean body mass compared with either no exercise or moderate-intensity continuous training. It did improve cardiorespiratory fitness more than both (p < 0.001 versus no exercise; p = 0.017 versus continuous training).

⚠What it shows: Solid ground: results pooled from 47 controlled studies in normal weight, overweight and obese adults, so the fitness gain and the flat body composition are both reasonably firm. The trials ran from 4 weeks upwards and included non-randomised ones, so longer or better-controlled programmes could still read differently.Meta-analysis · Sports Med

Pooling nine studies with 640 participants, eating fish raised blood 25-hydroxyvitamin D by an average of 4.4 nmol/L; fatty fish drove almost all of that at 6.8 nmol/L, while lean fish showed no clear effect. The authors conclude that recommended fish intakes cannot optimise vitamin D status.

⚠What it shows: Trustworthy but modest: results pooled from randomised trials in 640 people, and the trials largely agreed with each other on fatty fish. The longer studies suggested a bigger rise, so short trials may understate what months of regular fish eating would do.Meta-analysis · Am J Clin Nutr

Pooling six trials with 788 adults, diet and diet plus exercise both lowered total cholesterol, LDL cholesterol and triglycerides, while aerobic exercise alone only lowered triglycerides. No approach produced a significant change in HDL cholesterol.

⚠What it shows: Decent evidence, narrow base: results pooled from six trials in 788 men and women, each of which compared diet, exercise, both and a control group directly, which is rare and useful. Six trials is still few, and the trials ran from 4 weeks upwards, so this says little about years of either habit.Meta-analysis · Clin Nutr

In a meta-analysis of 32 randomised trials, weight loss reduced bone mineral density at the hip by about 0.008 g/cm2 and at the lumbar spine by about 0.018 g/cm2 but did not change total body BMD; calorie restriction, alone or combined with exercise, lowered hip bone density, whereas weight loss achieved through exercise training without dietary restriction was associated with increased hip BMD.

⚠What it shows: Solid on the direction of travel: results pooled from 32 trials in adults, with hip losses emerging after more than 4 months and spine losses mainly in calorie restriction lasting over a year. The trials differed in diets, exercise and participants, so the size of the effect for any one person remains uncertain.Meta-analysis · Osteoporos Int

Across 10 trials, berberine supplementation was associated with small reductions in body mass index of 0.29 kg/m2 and in waist circumference of 2.75 cm, but no significant change in body weight.

⚠What it shows: Real but tiny: results pooled from 10 trials, and the changes are small enough that most people would not notice them on a tape measure over a normal supplement run. Body weight itself did not shift, and earlier individual trials had disagreed, which is what prompted the pooling in the first place.Meta-analysis · Complement Ther Clin Pract

Pooled results from 13 trials in 838 people found capsinoid supplements had no significant effect on body weight, BMI, waist circumference, fat mass, fat-free mass, visceral fat or body fat percentage. A modest weight reduction appeared only in overweight participants in trials lasting 12 weeks or longer.

⚠What it shows: Strong enough to doubt the fat-burner marketing: results pooled from 13 trials in 838 people found nothing across eight separate body measures. The hint of weight loss in overweight people on longer courses came from a smaller slice of the data, which is a weaker kind of evidence, and the authors say it needs confirming.Meta-analysis · Clin Nutr ESPEN

A meta-analysis of 11 randomised trials in 618 people found omega-3 supplementation had no effect on insulin sensitivity compared with placebo. A small significant improvement appeared only in the subgroup of trials using one particular measurement method.

⚠What it shows: A fairly firm null for this specific question: results pooled from 11 trials in 618 people showed no change in insulin sensitivity versus placebo. One slice of trials using a single measurement method did show a small benefit, and the search only runs to October 2010, so later trials are not counted here.Meta-analysis · Clin Nutr

A systematic review of 247 studies with pooled results from 31 studies of new Alzheimer's cases linked low educational attainment (risk ratio 1.99), high homocysteine (1.93) and current or ever smoking (1.37) to higher risk, while higher physical activity and n-3 fatty acids appeared protective. The share of cases attributable to the factor was highest for physical activity (31.9%) and smoking (31.09%), whereas caffeine had the least consistent findings of all the factors reviewed.

⚠What it shows: Broad and fairly consistent, but built on studies that followed people rather than testing changes on them: 247 studies reviewed and 31 pooled for new Alzheimer's cases. People who exercise more or study longer differ in many other ways, so the share-of-cases figures are best read as ceilings rather than promises.Meta-analysis · BMC Public Health

Across 40 randomised trials covering 1720 people with depression, 18.1% dropped out of exercise programmes, and dropout was lower in the exercise groups than in control conditions. Supervision by physiotherapists or exercise physiologists predicted fewer dropouts, while more severe baseline symptoms predicted more.

⚠What it shows: Solid ground for the tolerability claim: results pooled from 40 trials in 1720 people with depression. It measures who kept turning up in trials rather than how well exercise treats depression, and people who volunteer for a trial may already be keener than average.Meta-analysis · J Affect Disord

In pooled data from 12 randomised trials of 1113 people with diabetes or prediabetes, adding probiotics to standard care raised gut Lactobacillus and Bifidobacterium levels and modestly reduced fasting blood glucose. Courses of three months or less showed the clearest improvements in blood sugar markers.

⚠What it shows: Reasonably solid on direction, less so on size: results pooled from 12 trials in 1113 people with diabetes or prediabetes, but the trials disagreed with each other a great deal on the microbe changes. The drop in fasting glucose was modest, and everyone here was on standard diabetes care, so this says nothing about healthy people.Meta-analysis · Br J Nutr

Pooled data from 22 randomised trials covering 1068 healthy people found probiotic supplements produced no significant change in gut microbiota diversity on any of four standard indices. The result held regardless of probiotic family or how long people supplemented.

⚠What it shows: A convincing null for the healthy shopper: results pooled from 22 trials in 1068 healthy people found no shift on any of four diversity measures, and the finding held across probiotic types, study quality and course length. Diversity is only one way of judging gut health, so this does not rule out other effects.Meta-analysis · BMC Med

Pooling trials of workplace physical activity programmes, employees in the intervention groups reported less general musculoskeletal pain (standardised difference -0.40) and less neck and shoulder pain (-0.37) than comparison groups, while the few studies of low back pain and of arm, elbow, wrist, hand or finger pain showed no clear evidence of benefit.

⚠What it shows: Reasonably solid for the neck and shoulders: results pooled from workplace trials published between 1990 and March 2013, each comparing employees given an exercise programme with employees who were not. Pain was reported by the employees themselves, and there were too few trials on the low back, arms and hands to say anything either way.Meta-analysis · Workplace Health Saf

Pooled data from 29 studies found stabilisation or core stability exercises produced statistically significant but clinically insignificant long-term improvements in non-specific low back pain and disability compared with any alternative treatment or control, with mean differences of -6.39 for pain and -3.92 for disability on a 0 to 100 scale. Compared with other forms of active exercise, there was no statistical or clinical difference.

⚠What it shows: Strong evidence and unusually consistent: results pooled from 29 studies covering more than 2,000 people for pain and for disability, with the studies largely agreeing and rated of high quality at long-term follow-up. The authors judge further research unlikely to change the conclusion, though it does not tell you which exercise any individual will stick with.Meta-analysis · BMC Musculoskelet Disord

Pooling 15 randomised trials of 587 men and women aged 60 and over with overweight or obesity, resistance training of at least eight weeks improved body weight, percent body fat, fat mass, BMI and lean mass, yet the spread of results showed no true differences between individuals in how much they responded.

⚠What it shows: Reasonably solid: results pooled from 15 trials in 587 people aged 60 and over who were carrying extra weight, all training for at least eight weeks. It does not prove individual variation is absent, only that the trials' own numbers give no evidence for it once random variation is accounted for.Meta-analysis · Sci Prog

Pooled results from nine randomised trials found cold water immersion slightly better than passive recovery for muscle soreness, both immediately after exercise and days later. The pooled data suggested a dose-response pattern, with the best results at water temperatures of 11 to 15 °C and immersion times of 11 to 15 minutes.

⚠What it shows: Reasonably trustworthy but modest: results pooled from nine trials where people were assigned at random, all comparing cold water immersion with simply resting. The benefit for soreness was small, and with so few studies the ideal temperature and time are best treated as a signal rather than a settled recipe.Meta-analysis · Sports Med

In a randomised trial of adults with obesity, prediabetes and fatty liver, a matched ~10% weight loss on a very-low-carbohydrate ketogenic diet raised liver insulin sensitivity two to three times more than Mediterranean or very-low-fat diets, and produced the biggest falls in liver fat, glycated haemoglobin and round-the-clock glucose and insulin. LDL cholesterol and apolipoprotein B did not differ between the diets.

⚠What it shows: Worth taking seriously but not settled: one trial where adults with obesity plus prediabetes and fatty liver were assigned a diet at random, with weight loss deliberately matched at about 10% across groups. It cannot say whether the liver advantage lasts long term or applies to people without these metabolic problems.Study · Cell Metab

Pooled results from 11 randomised trials in athletes (215 participants, mean age 26.09 years, 91% men) found heat acclimation improved time-trial performance and lowered maximal heart rate by 7 beats per minute, but produced no significant change in VO2max, exercise heart rate, core temperature, mean skin temperature, plasma volume or blood lactate. Changes in thermal comfort and perceived exertion were judged too small to be meaningful.

⚠What it shows: Reasonable evidence but thin on the ground: results pooled from 11 trials with 215 athletes between them, and 91% of those athletes were men, so it says little about women. The performance gain was modest and most of the physiological markers people expect heat training to shift did not budge.Meta-analysis · J Sports Sci Med