Skip to main content

Advertisement

Table 4 Predictive characteristics of admission biomarkers and their combinations for intensive care unit mortality

From: Evaluation of clinically available renal biomarkers in critically ill adults: a prospective multicenter observational study

Logistic regression model AUC-ROCa Cutoffb Sensitivity Specificity (+) LR (−) LR PPV NPV
Univariate models
 sCysC 0.727 (0.660–0.793) 1.12 mg/L 0.62 0.77 2.67 0.49 0.15 0.97
 uNAG 0.793 (0.743–0.842) 37.75 U/g Cre 0.82 0.71 2.83 0.26 0.16 0.98
 uACR 0.777 (0.721–0.832) 63.66 mg/g Cre 0.77 0.70 2.61 0.32 0.15 0.98
Multivariate models
 sCysC+ uNAG 0.811 (0.760–0.863)c,d 0.05e 0.80 0.75 3.17 0.26 0.17 0.98
 uNAG + uACR 0.809 (0.763–0.856)f 0.05e 0.88 0.70 2.97 0.17 0.16 0.99
 sCysC + uACR 0.756 (0.696–0.816) 0.06e 0.59 0.82 3.27 0.50 0.18 0.97
  1. Abbreviations: (+) LR Positive likelihood ratio, (−) LR Negative likelihood ratio, PPV Positive predictive value, NPV Negative predictive value, sCysC Serum cystatin C, uNAG Urinary N-acetyl-β-d-glucosaminidase, Cre Creatinine concentration, uACR Urinary albumin/creatinine ratio
  2. a Values are presented as AUC-ROC (95% CI). Among 1084 adult critically ill patients, 66 patients died in the intensive care unit
  3. b Ideal cutoff value according to Youden’s index
  4. c P < 0.05 vs. sCysC
  5. d P < 0.05 vs. sCysC + uACR
  6. e Cutoff points of the biomarker panels were the predicted probabilities generated from the multiple logistic regression model