Premium
Lipoprotein metabolism indicators improve cardiovascular risk prediction
Author(s) -
Graaf Albert A,
Schalkwijk Daniel B,
Tsivtsivadze Evgeni,
Parnell Laurence D,
Werffvan der Vat Bianca J.C.,
Ommen Ben,
Greef Jan,
Ordovas Jose M
Publication year - 2013
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.27.1_supplement.874.19
Subject(s) - framingham risk score , medicine , receiver operating characteristic , very low density lipoprotein , multivariate statistics , lipoprotein , disease , cholesterol , statistics , mathematics
We investigated whether lipoprotein metabolism indicators derived from a computational model (van Schalkwijk et al J Clin Bioinf 2011) could improve cardiovascular (CVD) risk prediction for 1981 subjects (145 cases, 1836 controls) from the Framingham Heart Study offspring cohort. We applied a statistical learning algorithm using a support vector machine to select conventional risk factors and lipoprotein metabolism indicators that contributed to predicting risk for general CVD. Improvement was quantified by the change in the Area‐Under‐the‐ROC‐Curve (ΔAUC) and by risk reclassification, using low, middle, and high risk categories (Net Reclassification Improvement, NRI), and a category‐independent method (Integrated Discrimination Improvement, IDI). Two calculated VLDL lipoprotein metabolism indicators improved CVD risk prediction. Added to a multivariate model with the best performing Framingham risk markers, these markers significantly improved CVD prediction (ΔAUC=0·0177 p=0·0055), risk reclassification (NRI=0·090 p=0·014) and the category‐independent method (IDI=0·051 p<0·0001). The novel risk markers could reclassify 25% (p<0·0001) of all subjects, who were inappropriately classified in the ‘intermediate risk’ category by Framingham markers, to the low risk category. Conclusion two calculated VLDL metabolism indicators significantly improved CVD risk prediction.