572 studies and 486,538 cancer cases say yes, with a clear dose-response. The protective effect of moderate drinking disappeared once studies stopped counting sick quitters as non-drinkers.
Fifteen randomised trials, 56,675 people. Cutting saturated fat cut cardiovascular events by 17% and did nothing to how long anyone lived. And 21 cohort studies found no link at all. All three are true.
Thirty-one percent of American adults take one. After 84 randomised trials the task force could not say whether it helps or harms. There is one small real signal, and one ingredient they tell you to stop.
Across 16 randomised trials, protein source did not change strength or absolute lean mass. The one real edge for animal protein was 0.90 lb, in under-50s, and one of its intervals touches zero.
Across 121 randomised trials the strength effect was large. The effect on everyday physical ability, from the same review, was small. Both numbers are real and you should have been given both.
Across 4,871 tests, submaximal equations put people in the right fitness third 51% of the time. Guessing gets 33%. Here is what a real test is, and when you actually need one.
Something real does happen: blood levels of the vitamins go up. What has not been shown is that it buys you anything. The biggest brand's own trial ran thirty people for four weeks.
A diet nobody markets beat 21 other lifestyle interventions. The best exercise was a wall sit. And the supplements sold for blood pressure came last, by their own numbers.
The US reference standard is 22,379 measured tests. Three things move your percentile before you train: which machine, which edition of the chart, and whether anyone measured you at all.
Fitness falls about 4.0 mL/kg/min a decade. Interval training adds about 5.5. The catch is that the device on your wrist misses by more than that.