Skip to main content
Fig. 4 | Annals of Intensive Care

Fig. 4

From: Nucleated red blood cells as predictors of mortality in patients with acute respiratory distress syndrome (ARDS): an observational study

Fig. 4

a Receiver operator curve (ROC) for the assessment of predictive validity of NRBC measurements. The value that maximized the vertical distance between ROC and diagonal line (representing the highest sum of sensitivity and specificity) was used for calculation of a cutoff value, as described elsewhere [17]. ROC AUC: 0.71; 95% CI 0.66–0.75; p < 0.0001. b Probability of survival depicted as Kaplan–Meier curves of ARDS patients grouped by NRBC cutoff level of 220 NRBC/µl from ROC analysis. Log-rank test: p < 0.001. c Receiver operator curves (ROC) for the assessment of predictive validity of NRBCs in comparison to severity of illness scores APACHE, SAPS, SOFA and TISS at ICU admission. NRBC AUC = 0.713 (95% confidence interval = 0.665–0.758); APACHE AUC = 0.626 (95% confidence interval = 0.576–0.675); SAPS AUC = 0.618 (95% confidence interval = 0.567–0.667); SOFA AUC = 0.60 (95% confidence interval = 0.549–0.650); TISS AUC = 0.605 (95% confidence interval = 0.554–0.654). AUC of the NRBC ROC curve was significantly different from those of the other parameters: pairwise comparison: NRBC vs. APACHE: p = 0.018; NRBC vs. SOFA: p < 0.001; NRBC vs. SAPS: p = 0.009; NRBC vs. TISS: p = 0.003). NRBC nucleated red blood cells; SAPS Simplified Acute Physiology Score; APACHE Acute Physiology And Chronic Health Evaluation; SOFA Sequential Organ Failure Assessment, TISS Therapeutic Intervention Scoring System

Back to article page