If we had even 1 honest meta-analysis for every 10 trials like this, well, we'd know more than we do now. Small trials with inconsistent/poor designs are the bane of medical science.
That said, I drink dark roast so this study must be true.
> coffee intake was associated with reduced risk of oral, pharynx, liver, colon, prostate, endometrial cancer and melanoma and increased lung cancer risk.
(Heh. Smokers drink nearly twice as much coffee as non-smokers. Smoking increases caffeine metabolism. I think this explains the lung cancer risk)
To your parenthetical comment: May be unrelated, but totally anecdotal, smokers tend to drink a very light roast that is easily drinkable like a folgers or maxwell house, maybe a coffee shop light roast, like a Dunkin. I don't think smokers are downing cups of Starbucks dark blend and if they are god help their stomachs.
This is probably true in the US, but in Europe I bet you would find that smokers drink espresso (which is usually made with dark roast coffee). Even for "regular strength" black coffee, people tend to drink "americano", which is espresso diluted with water. In the UK, in particular, I'm astounded at the volume of dark roast coffee people drink in the morning.
Any statement like "in Europe" is a fundamental disservice to how diverse Europe is in terms of taste. For example, I don't think I've ever met anyone in my life that likes to drink americano. I know a lot of people that smoke and drink coffee though.
Medium roasts have the most. With roast, both water evaporation and caffeine pyrolysis increase. Medium hits the maximum. I prefer very light roasts for the flavor.
The real variable isn’t roast its the brew concentration and to a lesser degree acidity, and most coffee I see served “American” style is pretty diluted so that you can see the cup through it.
That said, I drink dark roast so this study must be true.