Article TypeSpecial Correspondence To the Editor:

Lancet1 on the new atherosclerotic cardiovascular disease (ASCVD) risk equation released by the American College of Cardiology and American Heart Association.2 They showed that the new risk equa-tion overestimates ASCVD risk when applied in several cohort studies. One of these studies was the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study,3 which was used in the American College of Cardiology/American Heart Association guidelines to validate the ASCVD risk equation. As REGARDS investigators, we are writing to provide clarification on design com-ponents of REGARDS and other contemporary cardiovascular epi-demiology studies cited by Ridker and Cook that could contribute to lower-than-expected event numbers. We offer 4 possible explanations for the apparent overestimation of ASCVD risk by the new equation. In our opinion, these explanations suggest that the ASCVD risk equa-tion may be more valid than has been suggested by Ridker and Cook. First, the studies used to develop the ASCVD risk equation included “active surveillance ” to supplement self-report for identifying possi-ble events. For example, in the Atherosclerosis Risk in Communities (ARIC) Study (1 of the studies used in developing the ASCVD risk equation), investigators identified possible ASCVD events using 2 approaches: (1) study staff telephoned participants or their proxies annually to identify hospitalizations and deaths; and (2) they searched lists of International Classification of Diseases codes from area hos-pitals and state death files.4 For all death and hospital records that may have involved an ASCVD event, the ARIC study staff abstracted diagnostic information and obtained Minnesota coding for electro-cardiograms. Also, coronary events that occurred out of the hospital setting were documented by study staff who contacted physicians, coroners, and next of kin for information. However, the validation cohorts (REGARDS and the Multiethnic Study of Atherosclerosis [MESA]) and, to our knowledge, the studies evaluated by Ridker and Cook did not include similar surveillance procedures, which are costly and far less feasible since the implementation of the Healt