Early, small, conflicting, or animal-only. Not actionable alone.
The review that pooled both found cortisol fell significantly and perceived stress did not move at all. Sleep is the best-supported claim, testosterone rises in men but not women, and 25 cases of liver injury have been published worldwide.
The trial behind the 2024 headlines was retracted in September 2025 — four days after a meta-analysis pooled it. Ten months on, that meta-analysis still carries no correction. What survives: a real but modest effect on post-meal blood sugar, and about four pounds in the best weight trial, most of it back within a month of stopping.
The 2023 study linking erythritol to heart attacks came with a real mechanism — more reactive platelets, more clotting in animals. But the cohorts measured erythritol in blood, not diet, and your body makes its own: plasma levels track metabolic disease and fall after weight loss. Marker or cause is genuinely unresolved.
Swapping sugary drinks for diet ones took off 2.3 lb across 17 randomised trials, and swapping for water did nothing. But the Framingham study found daily cola drinkers had 3.7% lower hip bone density — and diet cola looked the same, so switching does not buy you out of that one.
The WHO guideline covers 'all synthetic and naturally occurring' sweeteners and names stevia in the same list as aspartame — natural is written out as an exemption on purpose. And stevia has about a third the randomised trial evidence of aspartame, so the confidence people feel about it is not coming from a deeper evidence base.
The 2023 study behind 'sweetener damages DNA' is real, competently run and industry-free — and every experiment in it happened in a dish. Its own authors write that the health effects 'are not yet known'. That sentence did not make the coverage.
The WHO advises against non-sugar sweeteners — as a conditional recommendation, on evidence it grades low certainty, from a harm signal that is observational, in a document that names reverse causation itself. Meanwhile 17 randomised trials found swapping sugary drinks for sweetened ones took off 2.3 lb, while swapping for water did nothing.
13 trials, 606 people, and a real dose-response — but an I-squared of 94%, meaning the trials flatly disagree, and no figure you could act on. Meanwhile the famous green coffee claim of 17.7 lb 'with or without diet and exercise' came from a study whose weights were altered. It cost $3.5 million and was retracted.
The famous 30% metabolism boost works out to about 24 calories, and two later labs could not reproduce it at all. What two randomised trials do support is duller and more useful: a glass before meals, for a couple of pounds over twelve weeks.
You lose about a pound and a half in an hour and the researchers called it 'bodily fluids' in their own conclusion. The one trial that randomised people found sauna added no significant fat loss on top of exercise — but with 15 per arm, it couldn't have detected much.