In Reply: Dr Ferket and colleagues examined a sample of studies that are reviewed in our analyses and suggest that eligible studies did not aim to improve the FRS but to identify an independent predictor of cardiovascular events or all-cause mortality. We believe that these aims coincide when conventional risk factors were defined by the FRS (or what the authors alluded to as the FRS). This was the case for all studies included in our review. In addition, all eligible articles included 1 or more analyses comparing the predictive performance of the FRS against the predictive performance of the FRS plus some additional predictor for a prospectively assessed outcome. The aim of such analyses is to assess improved prediction over the FRS, regardless of whether the exact analysis or phrasing may differ across studies (“independent predictors,” “increased area under the curve,” “better prediction,” etc). Many of these studies may have had additional objectives as well and may have used additional baseline models other than just the FRS, such as the Dutch or PROCAM (Prospective Cardiovascular Münster Heart Study) risk score. Nevertheless, the probing of incremental prediction beyond FRS was always one of their analyses. Under this definition, the statement in the study by Stork et al1 that “adding information on CRP, IL-6, or plaque burden to any risk function improved the discriminatory power” constitutes an eligible claim for improved prediction since 1 of the 3 baseline risk functions evaluated in their study was the FRS.
Tzoulaki I, Liberopoulos G, Ioannidis JPA. Assessing Predictive Performance Beyond the Framingham Risk Score—Reply. JAMA. 2010;303(14):1368-1369. doi:10.1001/jama.2010.390