These cases you mention are all called ANECTDOTAL incidents. I would like to see a controlled, double-blind study in the NEJM or such with a large number of participants before I can decide what is "the truth". Otherwise they're not worth paying attention to.