In Reply As pointed out by Drs Bittencourt and Fernandes, the Duke score has recently been shown to overestimate the pretest likelihood of obstructive CHD, and improved risk models are now available.1 However, when CE-MARC 2 commenced, we used the most up-to-date clinical guidelines for the investigation of stable chest pain, and these incorporated the Duke score. Although improved models now exist, as this was a pragmatic trial rather than a diagnostic accuracy study, we cannot evaluate the effect of the newer scores.
Greenwood JP, Brown JM, Berry C, for the CE-MARC 2 Investigators. Alternatives in the Evaluation of Suspected Coronary Heart Disease—Reply. JAMA. 2017;317(2):212-213. doi:10.1001/jama.2016.18335