Nutrition and wellness, with the sources attached
Cohort studies tied the MIND diet to a lower risk of dementia. The one trial that tested it directly found no real difference after three years.
Sleep-lab studies show a high dose of alcohol shortens the time it takes to fall asleep, but even a low dose disrupts dream sleep, and the later hours of the night suffer.
Meditation programs move anxiety scores down against a wait-list, but the gap shrinks once the comparison group also gets attention, to about zero in social anxiety. Real, but smaller than the headlines.
In trials that split people at random into a low-carb group or a low-fat group, weight loss came out close either way, about a kilogram apart, and in named diet programs the gap was gone by twelve months.
The low FODMAP diet is built as a short test, not a permanent way of eating. Trials show it beats a normal diet on IBS symptoms, and the real answer comes from reintroducing foods afterward.
A review of seventeen trials found no consistent benefit from any alternative form of creatine, and only three tested one against plain monohydrate. The newer bottles are selling a story, not a result.
The commonly cited limit of about 25 grams of added sugar a day comes from a review of 73 earlier analyses covering 8,601 studies. What it actually rests on is more specific, and more limited, than the round number suggests.
Balance and functional exercise cut the rate of falls by 24% across 39 trials, and the biggest gains came from combining it with resistance work, done for months, not weeks.
The trials that exist mostly measure blood sugar, not sleep itself, and one controlled feeding study found that how much people ate during the day did not change how long or how well they slept.
None of the trials reviewed here measured a sugar high followed by a crash in alertness. What the trials do show is a steady link between habitual sugar intake and a higher risk of depression over years, not minutes.
Trials that randomly assigned adults to eat or skip breakfast found skippers lost slightly more weight, not less, even though they ate a bit more at lunch.
Why am I bloated after eating healthy? Trials point to what ferments in the gut, not what is wrong with it: a diet low in fermentable carbs was the only one tested more than once that beat a normal diet for bloating.
Taking iron every other day gets more of it into you than taking it daily, and splitting a dose across the day gains nothing. Pregnancy looks like the exception.
Feed people extra eggs and their cholesterol rises, the harmful kind and the protective kind together. The studies that followed people for decades cannot agree what that means.
Heavy training raised bone density at the lower spine about three times as much as light training. Walking on its own showed no significant effect at the spine, and at the hip only after six months.
Take insomnia therapy apart across 241 trials and the sleep hygiene talk contributes nothing. Restricting time in bed and reserving the bed for sleep do the work. Regularity has data but no trials.
One 200 mg capsule makes people a little quicker on an attention test within the hour. The calm it is sold for is a smaller, shakier finding, measured a capsule at a time.
Fasting and calorie counting came out about even. Only alternate-day fasting won, by 1.29 kg, and in the trials that ran 24 weeks or longer even that edge was gone.
One trial stands alone in testing fermented food as a whole diet. Over 17 weeks it raised the variety of gut bacteria and lowered inflammation, while the measure it was built around did not move.
Across 86 trials in 162,796 people, omega-3 capsules made little or no difference to deaths or heart events. People who eat fatty fish keep turning out healthier anyway.
Cutting saturated fat lowered heart attacks, strokes and related events by 17%. But what replaced the fat changed the size of that benefit, and so did the eater's starting risk.
Across 69 trials and 153,902 adults, calcium, vitamin D and the two combined had little to no effect on fracture risk. The K2 record is thinner than either.
Nearly 300,000 people were given antioxidant pills in trials. They did not live longer, the best-run studies hint they died slightly sooner, and beta-carotene raised lung cancer risk in smokers.
Cortisol falls and stress scores move a long way. The trials disagree wildly with each other, the certainty is rated low, and the liver injury reports arrived later.
The single number everyone quotes is not a constant, and it is not what the timing trials measured. Those trials point at early afternoon, not evening.
Lifting weights raised resting metabolism by about 96 calories a day. Green tea added roughly 38. Eating more often changed nothing at all.
Creatine gives a small memory benefit and leaves overall thinking unchanged. In healthy older adults, the group most sold on it, the evidence is two blinded trials.
A daily multivitamin shifts cancer risk by around one percentage point and does nothing for heart disease or death. The memory headline traces back to a single trial.
Whether a probiotic turns up in stool depends on the strain, except above 10 billion live bacteria a day, when every strain was recovered. The trials that found benefit did not need the bacteria to stay.
Measured energy expenditure falls by roughly 150 kcal a decade across adult life, and much of that comes from moving less, not only a slower engine. Nothing in the data happens at thirty.
Serve a lunch at half the calories per gram and slower to chew, and people eat 573 calories less than at the dense, quick version, without trying. Lighten only part of the plate and they make it back.
Exercise cut depression symptoms by a moderate amount, and walking did as much as anything else. In strength trials, how much people trained barely mattered. Who they trained with mattered more.
Fiber's benefit rests on studies covering seventeen million people and peaks between 25 and 29 grams a day. Probiotic evidence is a set of small, strain-specific trials.
Ask people to change how they eat and their depression scores improve a little; anxiety does not move at all. The bigger claims come from surveys, not from trials.
The band where risk fell furthest was 25 to 29 g a day, and what mattered was the total. Which kind of fiber, the soluble and insoluble split printed on the label, barely predicted anything.
The official number is a floor for avoiding deficiency, not a target. It works out to about 56 grams a day at 70 kg, and eating past it pays mainly if you lift or you are dieting.
Risk is lowest at about seven hours and climbs on both sides, with the long side climbing faster. In the lab, cutting sleep to four hours builds a deficit one recovery night does not clear.
Protein supplements on their own did nothing for muscle in adults over 50. Alongside lifting, the same protein added about half a kilo of arm and leg muscle and a firmer grip.
Four of eight trials in anxious people found magnesium helped. Many mixed it with other ingredients, and none recruited anyone on the strength of an anxiety diagnosis.
In poor sleepers, magnesium bisglycinate beat a dummy capsule by 1.6 points on an insomnia score over four weeks. Older adults with insomnia fell asleep about a quarter of an hour sooner.
Only one human trial here compared magnesium forms side by side. Citrate and an amino-acid chelate were absorbed better than oxide, which did no better than a dummy capsule. Glycinate was not in it.
Melatonin stops working better above about 4 mg a night, and the whole effect is minutes of sleep rather than hours. The hour you take it matters more than the number on the bottle.
Older muscle builds less from the same meal, so protein targets go up after 65 - about 70 to 84 grams a day at 70 kg. The vitamin gaps the reviews found are about what people eat, and they could not measure absorption.
Across 26 trials, omega-3 improved depression scores by a small margin, and only EPA-heavy formulas carried the effect. It is real, modest, and easy to overstate in either direction.
Plant protein trailed animal protein for muscle mass by a small margin, and only in adults under 60. Soy matched milk outright, and 26 grams of soy did the work of 19 grams of whey.
Inulin and galacto-oligosaccharides feed bacteria you already have, and about 2.4 grams a day tripled Bifidobacterium in one trial in children. The gas some people get is the same reaction.
Two uses hold up in the trials: diarrhea while you are on antibiotics, and the C. difficile infection that can follow. In children on antibiotics, the dose mattered more than the strain name on the front of the box.
Trade carbohydrate for protein at the same calories and you lose about 0.79 kg more in twelve weeks. Protein really does make people feel fuller. Its sturdier benefit is keeping weight off.
Eating protein evenly across the day raised muscle-building by a quarter over 24 hours. The two trials that ran for weeks and weighed the muscle found no significant difference. Total intake matters most.
Thirty trials raised people's intake of the main fat in seed oils and measured inflammation. Nothing moved. The argument worth having is about what you would eat instead.
Melatonin has real evidence behind a small effect. Herbal formulas have a mountain of studies and a credibility problem. Most of the shelf has nothing tested at all.
Ultra-processed means built from extracted substances and additives, not food that has merely been cooked or canned. Fed matched menus and left alone, people ate hundreds of calories more of it.
With age, eating too little B12 is rarely the problem. The step that frees B12 from food protein breaks: 38% of older adults with a worn-out stomach lining showed deficiency, against 15% of those with neither that nor an acid blocker.
How far a dose moves your blood level depends on what you weigh and where you start. Dose relative to body weight explained 34.5% of the variation between groups of people taking supplements.
Vitamin D absorbed 32% better when it was swallowed with a meal containing fat. Iron is the reverse problem, and the coffee rule runs one way: before the meal is fine, after is not.
Each calculator runs a different equation. In adults with obesity the best puts three in four within 10% of a measured resting rate, and the activity multiplier on top is barely validated.
