>If you actually read the referenced studies from an evidence-based medicine standpoint they're pretty weak. They're mostly just observational with unreliable subject reported data. They haven't fully controlled for all confounding factors such as the healthy subject effect. And effect sizes are small.

The first citation (23) links to https://pubmed.ncbi.nlm.nih.gov/32827219/, which analyzed 15 randomized controlled trials. The effect size is 17% reduction in "combined cardiovascular events". Considering that "heart disease" is the top cause of death in the US (https://www.cdc.gov/injury/wisqars/pdf/leading_causes_of_dea...), that seems like a big effect size.