In a correction posted on the Web site of The New England Journal of Medicine, where the study, called APPROVe, was published in February, 2005, the journal removes the conclusion that it takes 18 months for risk to emerge from the Abstract and the discussion of their paper.