Evidence is consistent, with known limits.
Two big analyses of the trials found magnesium lowers blood pressure by about 2 to 3 points, mainly in people whose blood pressure is high. For older adults' night leg cramps, Cochrane found nothing. And the headline sleep result, 17 minutes, comes from two small trials.
Two big trials and a meta-analysis agree that testosterone therapy lifts sex drive in men with confirmed low testosterone; its effects on energy and mood are small. The largest trial found no rise in heart attacks or strokes, but more heart-rhythm problems, lung clots and kidney injury, and the largest trial to count fractures found 43% more of them.
Of 3,334 men, 277 had a normal total testosterone but a low free one, and more sexual and physical symptoms on average; 96 had the reverse and mostly did not. A smaller study disagrees. Plus the three ways a lab gets a free number, and the one the guideline says not to use.
Processed meat goes with an 18% higher colorectal cancer risk per 1.8 oz (50 g) a day. Two 2019 reviews measured a smaller cut, three servings a week, and found about 1 fewer case per 1,000 people over a lifetime. Both are true. Here is why, what IARC's Group 1 actually means, and why steak is a weaker case.
The famous PREDIMED trial was retracted from the New England Journal of Medicine and republished — and the nuts arm did just as well as the olive oil arm. Across 92,000 Americans over 28 years, swapping butter or margarine for olive oil went with lower death rates. Against other vegetable oils combined, no significant difference at all.
Same trial, same 59 people, same 40-gram dose: lactitol caused severe symptoms across the board and maltitol caused no laxative effect at all. Both are labelled sugar alcohols. Europe's regulator set a limit on erythritol for diarrhoea — not clots — and refused to drop the warning label. And the erythritol and xylitol heart scares came from one dataset.
Nineteen people ate 441 calories a day less without being told to, purely from moving protein from 15% to 30% of their meals — with the appetite hormones pushing the other way. But when 38 foods were ranked head to head, the most filling was a boiled potato, and water content predicted fullness better than protein did.
Soluble fibre supplements produced about 2.8 lb more weight loss on average across 22 trials — an average, not a prediction for any one person, and most of those trials paired fibre with a calorie-restricted diet. The Lancet's 25–29 g/day target comes from a review that excluded supplement trials entirely.
Ten adults did the same 14-day diet twice: once on 8.5 hours' sleep, once on 5.5. The scale showed the same loss both times. On short sleep, 55% less of it was fat and 60% more of it was muscle. A separate 80-person trial found sleeping longer cut intake by 270 calories a day, with no change in what people burned.
22 randomised trials say brisk walking takes off about 4.7 lb — and 5.7 lb of fat, because fat-free mass went up at the same time. The famous step target is not a finding: it is the name of a Japanese pedometer sold in 1965.