Errm, not to mess with anyone, but how many of those dead users have been using the stuff for decades?
I bet that graph counts eighty-year-old people who die of smoking.
I doubt many ex users have been taking it for very long for a sustained period.
It stands to reason that people who have used something for thousands of times daily run a larger risk of getting killed by it.
Compared to people who only take something at the occasional party before they grow out of it.
That graph tells me exactly nothing.