Table 5.
Relative Contribution of Plasma Inflammation and Endothelial Dysfunction Markers in Model Fit and Discrimination by Race/Ethnicity
| Model | White
|
Black
|
Hispanic/Latina
|
Asian/Pacific Islander
|
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| −2 LnL | BIC | AUC | −2 LnL | BIC | AUC | −2 LnL | BIC | AUC | −2 LnL | BIC | AUC | |
| Clinical risk factorsa+fasting glucose level (reference model) | 367.59 | 375.16 | 0.93 | 363.69 | 370.69 | 0.88 | 148.47 | 154.59 | 0.90 | 61.96 | 67.64 | 0.92 |
|
|
||||||||||||
| Variable LR χ2 Statistic | BIC | AUC | Variable LR χ2 Statistic | BIC | AUC | Variable LR χ2 Statistic | BIC | AUC | Variable LR χ2 Statistic | BIC | AUC | |
| +WBC count | 2.61 | 375.01 | 0.93 | 0.75 | 397.29 | 0.88 | 3.54 | 173.22 | 0.90 | 1.82 | 104.90 | 0.92 |
| +TNF-R2 level | 1.44 | 376.17 | 0.93 | 0.01 | 398.02 | 0.88 | 4.15b | 172.61 | 0.91 | 1.19 | 105.53 | 0.92 |
| +IL-6 level | 1.27 | 376.35 | 0.93 | 6.21b | 391.82 | 0.88 | 0.74 | 176.02 | 0.90 | 0.00 | 106.72 | 0.92 |
| +hsCRP level | 6.62b | 371.00 | 0.93 | 1.66 | 396.37 | 0.88 | 1.65 | 175.11 | 0.90 | 1.86 | 104.86 | 0.92 |
| +WBC count and TNF-R2, IL-6, and hsCRP levels | 9.88b | 390.44 | 0.93 | 6.49 | 412.55 | 0.88 | 9.25 | 185.87 | 0.91 | 4.99 | 118.77 | 0.92 |
| +E-selectin level | 2.48 | 376.77 | 0.93 | 0.18 | 397.85 | 0.88 | 1.89 | 174.87 | 0.90 | 0.51 | 106.21 | 0.91 |
| +ICAM-1 level | 1.51 | 377.82 | 0.93 | 5.46b | 392.57 | 0.88 | 7.47b | 169.29 | 0.90 | 0.40 | 106.32 | 0.92 |
| +VCAM-1 level | 0.56 | 378.72 | 0.93 | 0.20 | 397.83 | 0.88 | 5.87b | 170.89 | 0.90 | 0.61 | 106.11 | 0.92 |
| +E-selectin, ICAM-1, and VCAM-1 levels | 3.13 | 391.28 | 0.93 | 5.56 | 406.47 | 0.88 | 10.25b | 178.75 | 0.90 | 0.91 | 117.17 | 0.92 |
Abbreviations; AUC, area under the receiver operating characteristic curve; BIC, Bayesian information criteria; LnL, log likelihood; LR, likelihood ratio; WBC, white blood cell. Other abbreviations: See Table 1.
Described in Table 4.
P<.05 in deviance test (likelihood ratio test) compared with the reference model.