There's a difference between statistical inference and necessary and sufficient conditions. Smoking is linked to cancer via the former, the latter has yet to be demonstrated.
Exactly the same for your Teflon boogieman, condemned by, in this case scant, statistical inference while no mechanism...