A-031 Evaluation of the Analytical and Clinical Performance of the NT-proBNPII Assay on the Atellica CI Analyzer

Background Measurements of NT-proBNP are used to aid diagnosis of heart failure (HF) in patients with clinical suspicion of new onset or worsening HF. This study evaluated the analytical and clinical performance of the Atellica IM NT-proBNPII (PBNPII) Assay on the Atellica CI Analyzer in emergency department (ED) and outpatient (OP) populations. Methods Analytical Performance: Limits of blank, detection, and quantitation (LoB, LoD, and LoQ) samples were tested on three reagent lots. Linearity studies tested N=18 levels of mixed high and low samples (N=7/level). Precision studies were performed using human serum samples and quality control samples. Repeatability and within-lab precision samples were tested in duplicate/run, 2 runs/day for 20 days (N=80). Reproducibility samples were assayed in three replicates/run, 2 runs/day for 5 days using three analyzers (N=90 measurements/sample). Method comparison (MC) studies were performed on Atellica IM and Atellica CI analyzers using singlicate native human serum and contrived samples (N=145). Clinical Performance: Subjects were tested for the ED (N=3126) and OP (N=1029) studies. The ED study primary objective of the clinical testing was to verify performance of the cut-offs as established in the ICON-RELOADED study1: exclusion of HF (300 pg/mL, all ages) and inclusion of HF by age (<50 y.o.= 450 pg/mL, 50-75 y.o.= 900 pg/mL, >75 y.o.= 1800 pg/mL). Primary statistical endpoints of this evaluation were the likelihood ratios (LR) for both cutoffs and indeterminate results between cut-offs. The OP study primary objective was to verify use of a single cut-off value of 125 pg/mL to aid in the diagnosis of new onset HF. Primary statistical endpoints were sensitivity, specificity, and negative/positive predictive values. Results LoB, LoD, and LoQ results were 13, 20, and 35 pg/mL, respectively. Linear range was 35-35,000 pg/mL. Coefficients of variation (CV) for repeatability were 0.8-2.6% and within lab, 1.3-3.5% at 106-32,846 pg/mL. Reproducibility CVs were lowest, 1.9% at 450 pg/mL and highest, 4.0% at 35,681 pg/mL. MC Passing-Bablok regression yielded, y=1.01x+ 5 pg/mL at 56-31,345 pg/mL with Pearson correlation coefficient, r=0.999. The LR+ for the ED study compares to the ICON Reloaded study with LR+ of 3.60 (ICON: 5.99) for diagnosis of HF, 0.55 (ICON: 0.96), indeterminate, and 0.07 (ICON: 0.09), exclusion of HF. Using a single cut-off value, 125 pg/mL, in the OP population resulted in AUROC curve of 0.837 (95% CI: 0.805, 0.870). Sensitivity of 85.9%, specificity, 63.7%, NPV, 95.4%, and PPV, 34.0% were achieved for PBNPII on the Atellica CI for identification of new onset HF. Conclusion The Atellica IM PBNPII on the Atellica CI Analyzer demonstrated acceptable analytical performance. MC showed equivalent results between Atellica CI and Atellica IM Analyzers. The clinical performance meets ICON Reloaded criteria for diagnosing or exclusion of HF, and supports use of a single cut-off of 125 pg/mL to identify new onset of HF.