Totally fair question. This is the first thing I could chase down. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7044619/

My guess would be that the sample size here is small and therefore it's not a completely iron clad study. Maybe there are bigger studies out there, I don't know. This one would be easy to gainsay and/or maybe the publication doesn't think its readers respond to scientific research.

I guess this is like the early days of research showing smoking was bad. Long way to go until everyone is convinced.