|
Sensitivity and specificity for the prediction of various complications in multiple-trauma patients |
|||||
| Parameter (cutoff value) |
Predicted risk |
||||
|
|
|||||
| Outcome |
Infection |
Sepsis/severe sepsis/septic shock |
SOFA score ≥12 within 21 days |
ICU stay >7 days |
|
|
|
|||||
| AUC (odds ratio, 95% CI) (P < 0.002), chi-square test |
Sensitivity/specificity (cutoff value) |
AUC (odds ratio, 95% CI) (P < 0.001) |
Sensitivity/specificity (cutoff value) |
AUC (odds ratio, 95% CI) (P < 0.002) |
|
|
|
|||||
| Initial procalcitonin (≥1 ng/ml) |
0.732 (-, not increased) |
93%/41% (0.62 ng/ml) |
0.805 (6.1, 2.5–15.7) |
90%/48% (0.53 ng/ml) |
0.798 (7.8, 2.8–21.9) |
| 80%/59% (1.31 ng/ml) |
80%/64% (0.97 ng/ml) |
||||
| 73%/60% (1.37 ng/ml) |
70%/74% (1.36 ng/ml) |
||||
| Initial C-reactive protein (>100 mg/l) |
0.564 (-, not increased) |
87%/37% (105 mg/l) |
0.674 (-, not increased) |
90%/30% (76 mg/l) |
0.639 (-, not increased) |
| 80%/41% (111 mg/l) |
80%/64% (104 mg/l) |
||||
| 73%/45% (122 mg/l) |
70%/52% (121 mg/l) |
||||
| Injury Severity Score (≥20) |
0.869 (1.3, 1.1–1.6) |
100%/61% (26.0) |
0.760 (10.6, 3.3–34.1) |
90%/54% (20.0) |
0.721 (7.0, 2.2–22.4) |
| 80%/77% (31.5) |
88%/54% (21.5) |
||||
| 73%/80% (34.0) |
73%/70% (26.0) |
||||
| Initial SOFA score (≥12) |
0.864 (11.6, 3.3–40.6) |
87%/68% (10.5) |
0.805 (6.0, 2.0–18.4) |
95%/42% (7.5) |
0.788 (4.9, 1.7–13.9) |
| 73%/76% (11.5) |
82%/70% (9.5) |
||||
| 67%/85% (12.5) |
65%/78% (10.5) |
||||
| APACHE II score (≥20) |
0.783 (13.1, 3.5–48.9) |
93%/39% (13.5) |
0.685 (-, not increased) |
90%/40% (12.5) |
0.678 (-, not increased) |
| 87%/48% (14.5) |
82%/46% (13.5) |
||||
| 73%/61% (15.5) |
57%/66% (15.5) |
||||
|
APACHE, Acute Physiology and Chronic Health Evaluation; AUC, area under the curve of the receiver operating characteristic; CI, confidence interval; SOFA, Sepsis-related Organ Failure Assessment. AUCs were not significantly different among parameters, when compared with the statistical method of Hanley and McNeil (z < 1.88) for outcome [16]. | |||||
Meisner et al. Critical Care 2006 10:R1 doi:10.1186/cc3910 |
|||||