Table 2.
Comparing the AUC of CRP, NLR and their combination in predicting infection in the development and validation groups.
| Variable | CRP ≥ 40 mg/L | NLR ≥ 7.41 | Combined CRP and NLR | P value* |
|---|---|---|---|---|
| Development cohort (n = 620) | ||||
| AUC (95% CI) | 0.75 (0.71–0.80) | 0.68 (0.63–0.73) | 0.76 (0.72–0.81) | P1 = 0.57 |
| P | < 0.001 | < 0.001 | < 0.001 | P2 < 0.001 |
| Sensitivity (%) | 75 | 68 | 69 | |
| Specificity (%) | 76 | 68 | 84 | |
| + LR | 3.08 | 2.06 | 4.18 | |
| − LR | 0.33 | 0.5 | 0.37 | |
| Validation cohort (n = 154) | ||||
| AUC (95% CI) | 0.73 (0.63–0.83) | 0.62 (0.51–0.73) | 0.75 (0.66–0.84) | P1 = 0.62 |
| P | < 0.001 | 0.03 | < 0.001 | P2 = 0.01 |
| Sensitivity (%) | 82 | 66 | 72 | |
| Specificity (%) | 64 | 58 | 78 | |
| + LR | 2.26 | 1.58 | 3.24 | |
| − LR | 0.29 | 0.6 | 0.36 | |
*The receiver operating characteristic (ROC) curves in the development and validation groups were compared by DeLong’s method.
P1-P value of combined CRP and NLR vs CRP ≥ 40 mg/L; and.
P2-P value of combined CRP and NLR vs NLR ≥ 7.41.