Read this article or watch the video before you start knocking down a dozen cups of Joe to prolong your life.

if one observes a benefit in a population associated with consuming a food or beverage, and the benefit is not mediated by the active ingredient in that food or beverage, the finding is likely due to unmeasured confounding.  In other words, I think coffee is in the same camp as red wine: the observed benefits are likely due more to the type of person who drinks it than what’s actually in the drink.