European Heart Journal Advance Access originally published online on April 16, 2008
European Heart Journal 2008 29(10):1335-1336; doi:10.1093/eurheartj/ehn150
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
New cardiovascular risk determinants do exist and are clinically useful: reply
UV University Medical Center
Internal Medicine
PO Box 7057
Amsterdam 1007MB
The Netherlands
Tel: +31 20 44 44 307
Fax: +31 20 44 44 313
Email: y.smulders{at}vumc.nl
Dr Levinson remarks that the clinical value of biomarkers can only be defined in studies showing improved outcome after incorporation of these biomarkers in risk models. We agree that this formally provides the final piece of evidence for which is the optimum prevention strategy. However, the optimum strategy always starts of with improved risk stratification, in order to better identify high-risk patients who will have the largest absolute benefit. The effectivity of subsequent intervention strategies is usually not specific for particular risk phenotypes. For example, for the selection of patients who will have a large absolute risk reduction from the use of statins, there is no evidence that the selection of risk factors or biomarkers for risk prediction models has any impact on the relative risk reduction. The same is true for blood pressure reduction: it lowers risk in all types of patients, so the only real issue seems to identify the high-risk groups, regardless of which variables are used in the risk models.
Many of the biomarkers have a risk ratio/odds ratio value that is comparable with ‘conventional’ risk factors, and are highly prevalent in those patient groups who qualify for risk assessment (usually middle-aged to elderly people patients with at least one clearly elevated conventional risk factor, automatically elevating their risk to above the population average). Hence, such risk determinants are equally important candidates for incorporation into improved risk models as are the conventional risk factors. The worse performance may be a theoretical disadvantage if risk models are applied to low-risk populations, but less so if the focus is on those at intermediate or elevated risk, which is where the relevant reclassifications occur. The correctness of such reclassifications is the final piece of evidence to obtain.
The main purpose of our paper was to extend the toolbox of risk model makers to more than just discrimination. One may argue whether biomarkers have already successfully qualified, but the correct tools to judge this are essential, as also Dr Wang appears to acknowledge in his accompanying article.1
Reference
- Wang TJ. New cardiovascular risk factors exist, but are they clinically useful? Eur Heart J (2008) 29:441–444.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||