Preliminary Evaluation of the Vitros ECi Cardiac Troponin I Assay

The recently published standards of practice for the use of cardiac marker testing for ruling in and ruling out acute myocardial infarction (AMI) recommend the implementation of cardiac troponin I (cTnI) or T (cTnT) as appropriate markers (1). The purpose of this preliminary study was to analytically and clinically evaluate the Ortho-Clinical Diagnostics (Rochester, NY) Vitros Troponin I immunodiagnostic assay. The Vitros ECi system uses reagents containing biotinylated monoclonal anti-cTnI antibody and goat polyclonal anti-cTnI antibody labeled with horseradish peroxidase. After an 8-min incubation for reagents and sample in a well precoated with streptavidin, the well is washed and a signal reagent containing horseradish peroxidase substrates is added to detect peroxidase bound to the well by utilizing an enhanced chemiluminescence reaction (2). The imprecision (as the CV, %) within the same calibration performed according to NCCLS guidelines (3) showed the following results: between-run (n = 80) means over 31 days on 20 occasions were 0.347 μg/L (10%), 0.768 μg/L (5.9%), 4.10 μg/L (4.3%), and 15.5 μg/L (3.2%). The analytical sensitivity, determined using 2 SD after 20 replicates of the zero calibration material were measured, was 0.018 μg/L. The lowest concentration giving rise to a between-assay CV ≤10% was 0.351 μg/L. Deming relationships, obtained from 36 fresh plasma samples across the linear range (0–100 μg/L) of the Vitros cTnI assay compared with the Dade-Behring RxL cTnI assay, showed the following: Vitros cTnI = 0.556 Dade cTnI + 0.039; r = 0.972. The 95% confidence intervals for slope and intercept were 0.509–0.602 and −0.394 to 0.472, respectively. The difference plot (Exstat Ver. 9; DDU Software) shown in Fig. 1 A demonstrates a proportional, increasing negative bias across the range of concentrations tested for by the ECi cTnI assay compared with the Dade RxL cTnI assay. The bias is likely a result of lack of standardization between cTnI assays (4). Deming regression analysis for the Vitros cTnI assay for fresh plasma samples (n = 50) compared with the same frozen, thawed plasma samples showed no significant differences (frozen cTnI = 0.965 fresh cTnI − 0.005; r = 0.997) over a cTnI range of 0–11.3 μg/L. Deming regression analysis between paired fresh serum and plasma (heparinized) samples showed a bias for plasma as follows: plasma Vitros cTnI = 0.760 serum Vitros cTnI − 0.201; r = 0.998 (n = 17). Based on 200 apparently healthy male and female blood donors, the upper reference 97.5th percentile was 0.1 μg/L in serum and 0.08 μg/L in plasma samples. Fig. 1B shows a comparison of ROC curves for the ECi and RxL assays determined from 74 chest pain patients admitted to rule in or rule out AMI. A modified WHO criterion, using a cTnI clinical decision cutpoint at 0.8 μg/L (Dade Behring RxL assay) for the biochemical criterion (predetermined at Hennepin County Medical Center), was used (5). Peak cTnI values, from the three to four plasma samples obtained serially within 24 h of presentation to the hospital, were used to construct the ROC curves. Thirty percent (22 of 74) of patients were diagnosed with AMI. The ECi cTnI cutoff of 0.40 μg/L demonstrated a sensitivity of 100% and a specificity of 79%. In comparison, the RxL cTnI cutoff of 0.55 μg/L demonstrated a sensitivity of 100% and a specificity of 77%. There were no statistical differences in the areas under the ROC curves between assays (0.9038 for ECi; 0.9008 for RxL). Patients with clinical diagnoses involving myocardial damage, including acute coronary syndromes (unstable angina), congestive heart failure, and cardiac surgery, but excluding AMI, were responsible for the 77% and 79% specificities of both cTnI assays at 100% sensitivities. Outcome studies involving risk assessment, however, were not part of this study design. In summary, the Vitros ECi Troponin I immunodiagnostic assay demonstrated acceptable analytical performance. The impression around the upper reference cutpoints was clinically acceptable (<10%). The 97.5th percentile was substantially different (four- to fivefold higher) than the analytical sensitivity, allowing for acceptable clinical specificity. Furthermore, there was excellent clinical concordance between the Vitros ECi and Dade RxL cTnI assays. The cTnI values obtained with the ECi were approximately one-half of the those obtained with the RxL cTnI, probably because of the lack of antigen and antibody standardization between assays (4). Thus, the Vitros ECi Troponin I immunodiagnostic assay appears to be an acceptable alternative for monitoring plasma or serum cTnI for ruling in and ruling out AMI. As for risk stratification in acute coronary syndrome patients (6), additional clinical evaluation studies will be necessary to validate it applicability in clinical practice. Difference plot (A) and ROC curves (B) for the comparison of the RxL and ECi cTnI assays. (B), concentration cutpoints (μg/L) are shown on each curve. The ECi is indicated by the solid line, closed circles, and dark numbers; the RxL is indicated by the dashed line, open circles, and lighter numbers. This work was supported in part by Ortho Clinical Diagnostics, Raratan, NJ.