Table 3.

Area under receiver operating characteristic curve (AUC) parameters (mean, 95% Confidence Interval (CI)) of analyzed predictive models over 1000 bootstraps.

Risk factors	Mean AUC	95% CI
Results based on dichotomization of the CIMT data using 90th percentile threshold (0.75 mm, CIMT > 90th percentile was considered high)
Traditional risk factors (Model 1)	0.757	[0.696, 0.811]
Traditional risk factors + ceramides (Model 2)	0.762	[0.697, 0.818]
Traditional risk factors + CERT (Model 3)	0.760	[0.696, 0.819]
Age + sex + Ceramides (Model 4)	0.726	[0.661, 0.784]
Age + sex + CERT (Model 5)	0.720	[0.660, 0.778]
Results based on dichotomization of the CIMT data using 75th percentile threshold (0.68 mm, CIMT > 75th percentile was considered high)
Traditional risk factors (Model 1)	0.722	[0.675, 0.766]
Traditional risk factors + ceramides (Model 2)	0.729	[0.688, 0.771]
Traditional risk factors + CERT (Model 3)	0.734	[0.690, 0.776]
Age + sex + Ceramides (Model 4)	0.710	[0.666, 0.754]
Age + sex + CERT (Model 5)	0.705	[0.663, 0.749]

Abbreviations: CERT, coronary event risk test.