While observational studies have previously pointed to an association, being able to identify causation has been difficult due to overlapping risk factors and confounding variables.