Table 2. Predictive value of Cox regression models for each inflammation-related biomarker assessed by AUC for incident coronary events.

Biomarker	AUC1a	ΔAUC1 (95% CI)a	AUC2b	ΔAUC2 (95% CI)b
None	0.807	—	0.845	—
hsCRP	0.809	0.002 (0.001–0.008)	0.845	0.000 (−0.001–0.003)
IL-6	0.810	0.003 (0.001–0.008)	0.846	0.001 (−0.000–0.004)
IL-18	0.808	0.001 (0.001–0.005)	0.845	0.000 (−0.001–0.003)
TGF-β1	0.808	0.001 (0.001–0.005)	0.845	0.000 (−0.000–0.003)
MIF	0.809	0.002 (0.001–0.008)	0.845	0.000 (−0.000–0.003)
MCP-1	0.808	0.001 (0.001–0.005)	0.845	0.000 (−0.001–0.002)
IL-8	0.808	0.001 (0.001–0.003)	0.845	0.000 (−0.000–0.002)
IP-10	0.808	0.001 (0.000–0.005)	0.845	0.000 (−0.001–0.005)
RANTES	0.808	0.001 (0.001–0.006)	0.845	0.000 (−0.000–0.003)
Adiponectin	0.808	0.001 (0.000–0.007)	0.845	0.000 (−0.000–0.005)
Leptin	0.811	0.004 (0.001–0.010)	0.845	0.000 (−0.000–0.004)
sE-selectin	0.817	0.010 (0.004–0.018)	0.847	0.002 (−0.000–0.009)
sICAM-1	0.817	0.010 (0.005–0.021)	0.848	0.003 (0.000–0.011)
With all 13 biomarkers	0.825	0.018 (0.013–0.038)	0.851	0.006 (0.003–0.021)
With IL-6,sICAM-1 c	0.819	0.012 (0.006–0.024)	0.849	0.004 (0.001–0.012)

Bold print denotes statistical significance for ΔAUC (P<0.05). “−0.000” denotes values between −0.0005 and 0.0000.

^aAdjusted for age, sex and survey (model 1).

^bAdjusted for age, sex, survey, BMI, systolic blood pressure, ratio of total cholesterol/HDL cholesterol, smoking, alcohol, physical activity, parental myocardial infarction and prevalent diabetes (model 2).

^cWith biomarkers that were significantly associated with incident coronary events in multivariable-adjusted models (IL-6, sICAM-1).

ΔAUC denotes the differences between the model with the respective inflammation-related biomarker and the model without any inflammation-related biomarker. ΔAUC for the difference between the model adjusted for age, sex, survey and cardiometabolic risk factors (model c) and the basic model adjusted for age, sex and survey (model a) was 0.038 [95% CI 0.026–0.055].