I haven't researched it, but the two statistics could be reconciled if there are fewer smokers than non-smokers.