Abstract
The Siemens Point-of-Care Testing (POC) Atellica® VTLi high-sensitivity troponin I (hsTnI) device has been previously validated. Verification independently provides evidence that an analytical procedure fulfils concordance with laboratory assays, imprecision, and hemolysis interference requirements.
OBJECTIVES: The Siemens Point-of-Care Testing (POC) Atellica® VTLi high-sensitivity troponin I (hsTnI) device has been previously validated. Verification independently provides evidence that an analytical procedure fulfils concordance with laboratory assays, imprecision, and hemolysis interference requirements. Five whole blood samples spanning the measuring interval were analysed 20 times in succession. Hemolysis interference was assessed at three troponin concentrations by spiking five hemolysate concentrations to plasma to achieve free hemoglobin concentrations 35-1,000 mg/dL. Concordance between whole blood (VTLi) and plasma on laboratory analysers (Beckman, Roche, Siemens) was assessed by Pearson correlation and kappa statistics at the (LOQ) and upper reference limit (URL). This was repeated for frozen plasma samples.
METHODS: Five whole blood samples spanning the measuring interval were analysed 20 times in succession. Hemolysis interference was assessed at three troponin concentrations by spiking five hemolysate concentrations to plasma to achieve free hemoglobin concentrations 35-1,000 mg/dL. Concordance between whole blood (VTLi) and plasma on laboratory analysers (Beckman, Roche, Siemens) was assessed by Pearson correlation and kappa statistics at the (LOQ) and upper reference limit (URL). This was repeated for frozen plasma samples. Coefficients of variation for whole blood were <10 % for whole blood troponin concentrations of 9.2 and 15.9 ng/L, thus below the URL. Hemolysis positively interfered; at 250 mg/dL affecting the low troponin sample (+3 ng/L; +60 %) and high troponin sample (+37 ng/L; +24 %). Correlation coefficients were 0.98, 0.90 and 0.97 between VTLi and Beckman, Roche and Siemens assays respectively. Corresponding kappa statistics were 0.80, 0.73 and 0.84 at the LOQ and 0.70, 0.44 and 0.67 at the URL.
RESULTS: Coefficients of variation for whole blood were <10 % for whole blood troponin concentrations of 9.2 and 15.9 ng/L, thus below the URL. Hemolysis positively interfered; at 250 mg/dL affecting the low troponin sample (+3 ng/L; +60 %) and high troponin sample (+37 ng/L; +24 %). Correlation coefficients were 0.98, 0.90 and 0.97 between VTLi and Beckman, Roche and Siemens assays respectively. Corresponding kappa statistics were 0.80, 0.73 and 0.84 at the LOQ and 0.70, 0.44 and 0.67 at the URL. Concordances between VTLi and laboratory assays were at least non-inferior to those between laboratory assays. Imprecision met manufacturer claims and was consistent with a high sensitivity assay. There is potential for hemolysis interference, highlighting the need for quality samples. The results support performance characteristics previously reported in validation studies, and the device offers acceptable performance for use within intended medical settings.
CONCLUSIONS: Concordances between VTLi and laboratory assays were at least non-inferior to those between laboratory assays. Imprecision met manufacturer claims and was consistent with a high sensitivity assay. There is potential for hemolysis interference, highlighting the need for quality samples. The results support performance characteristics previously reported in validation studies, and the device offers acceptable performance for use within intended medical settings.