Table 1.
Patient demographics and clinical data at baseline.
Figure 1.
In-hospital or 28-day mortality percentage by CFD quartiles (A) and procalcitonin (PCT) quartiles (B) on hospital admission.
CFD P-value for trend: p<0.001, PCT P-value for trend: p = 0.109.
Figure 2.
ROC curve of CFD levels and APACHE II score as predictors of in-hospital and 28-day mortality (A).
CFD P-value for trend: p<0.001, PCT P-value for trend: p = 0.109. Score: CFD; AUC = 0.79 (0.69–0.68), P<0.05 compared to APACHE II, IDI = 0.14 (0.03–0.25). APACHE-II+CFD; AUC = 0.79 (0.70–0.89), P<0.05, IDI = 0.18 (0.09-0.28). APACHE-II; AUC = 0.68 (0.56–0.79), Reference. ROC curve of CFD levels, PCT levels and APACHE II scores as predictors of in-hospital and 28-day mortality (B). Score: CFD; AUC = 0.79 (0.68–0.90), P<0.05 compared to APACHE II, IDI = 0.16 (0.04–0.29). PCT; AUC = 0.63(0.50–0.77), P>0.05, IDI = −0.04 (0.01–0.10). APACHE-II+CFD; AUC = 0.80(0.68–0.90), P<0.05, IDI = 0.20(0.09–0.31). APACHE-II+PCT; AUC = 0.65(0.51–0.78), P>0.05, IDI = 0.01(0.01–0.02). APACHE-II+CFD+PCT; AUC = 0.80(0.69–0.91), P<0.05, IDI = 0.16(0.04–0.27). CFD+PCT; AUC = 0.79(0.68–0.90), P<0.05, IDI = 0.13(−0.01–0.26). APACHE-II; AUC = 0.64(0.50–0.78), Reference.
Table 2.
Logistic regression model of in-hospital and 28-day mortality in patients with severe sepsis.
Figure 3.
Locally weighted polynomial regression (LOESS) analysis for CFD based probability of in-hospital or 28-day mortality adjusted for APACHE II score on hospital admission.