Rapid Measurement of Cardiac Markers on Stratus CS

Biochemical markers of myocardial damage are of fundamental importance in ruling in and ruling out diagnostic strategies for acute coronary diseases, particularly when electrocardiographic findings do not allow a diagnosis. An increasing body of evidence has demonstrated the value of strategies based on specific (cardiac troponin I or T) and sensitive [myoglobin (Myo)] assays in the diagnosis (1), prognosis, and monitoring (2) of patients with acute coronary syndrome. The impact of laboratory information on the management of patients has led to the development and definition of recommendations designed to reduce the therapeutic turnaround time, thus improving the medical outcome for patients. Because of the delay often associated with transport-related problems, the concept of point-of-care testing has been introduced for the measurement of cardiac markers (3). Some requirements, which seem to be of value in assuring quality for devices and diagnostic systems designed for point-of-care testing of cardiac markers, require whole blood or heparinized specimens for effective therapeutic turnaround time and must guarantee quantitative, accurate, and precise results comparable to those provided by central laboratories. The aim of our study, therefore, was to evaluate the analytical performance of the Stratus CS (SCS; Dade Behring), a fluorometric enzyme immunoassay analyzer based on solid-phase radial partition immunoassay technology, for quantitative measurement of creatine kinase isoenzyme MB mass concentration (CK-MB), Myo, and troponin I (TnI) in whole blood samples collected using lithium heparinate as anticoagulant. The samples may be processed automatically by the analyzer, which provides for centrifugation on board before the analytical phase, or heparinized plasma may be dispensed directly into a sample cup for analysis. Briefly, after the first antibody is added to a glass fiber paper linked to the dendrimer, the sample and the second antibody are pipetted. Finally, enzyme activity is started by the substrate wash solution with the simultaneous washing of the unbound, labeled antibody. The antibodies, glass fiber paper, and substrate-wash solution are assembled into test packs, each containing all of the reagents for one test, which are ready and easy to use. Our evaluation, carried out for all three tests measurable simultaneously by the analyzer, addressed aspects that are of value in introducing the system in the emergency department. Moreover, because of the importance of TnI results in the management of patients with minimal myocardial damage, we evaluated the lower detection limit, linearity, and upper reference value for TnI. The study was carried out using plasma pools, commercial controls, and plasma samples; quality control was carried out using the trilevel Dade TRU-Liquid Cardiac Control. Plasma pools were prepared using samples from routine analysis, collected into lithium heparinate, centrifuged (700g for 10 min), divided into aliquots within 2 h after blood collection, and kept frozen at below −20 °C until use. Plasma samples for method comparison and TnI reference range evaluation were collected from acute coronary syndrome patients and from healthy subjects, respectively, all specimens being tested within 2 h after blood collection. The imprecision study (Table 1 ) was carried out with two controls (C1 and C3) and a plasma pool. Because the SCS processes one sample at a time, providing up to four results for each sample, within-run imprecision was calculated by consecutively testing controls and pool quadruplicate three times, and between-run imprecision was tested on 20 different days. Myo and CK-MB assays showed CVs within the target value, defined on the basis of biological variability [5.6% for Myo (4); 9.2% for CK-MB (5)]. We evaluated the stability of calibration curves by testing control samples C1 and C3 weekly for 10 weeks; the results, expressed as the CVs (for TnI, C1 = 3.4%, C3 = 3.8%; for CK-MB, C1 = 2.7%, C3 = 3.0%; for Myo, C1 = 3.6%, C3 = 3.0%) showed that the variability was comparable to the within-run imprecision without a significantly increased or decrease trend in values, allowing the same calibration curve to be maintained for at least 10 weeks (8 weeks according to the manufacturer). In fact, the 95% confidence intervals (CIs) of the differences between means obtained by comparing the first six and the second five determinations were not statistically significant: for TnI, −0.0482 to 0.0035 for C1, −0.28 to 1.62 for C3; for CK-MB, −0.184 to 0.769 for C1, −2.64 to 1.65 for C3; for Myo, −2.76 to 0.756 for C1, −27.2 to 15.7 for C3. To evaluate the stability of samples and reagents at room temperature, of practical importance in the emergency department, the experimental protocol was divided into two parts. In the first part, TnI, Myo, and CK-MB were assayed on samples (C1, C3, and plasma pool) at time zero (corresponding to the time when samples C1 and C3 were taken from the refrigerator, and the plasma pool from the freezer), and 3, 6, and 24 h later, using loaded samples at room temperature and reagents stored at 2–8 °C, as recommended by manufacturer. The results are listed in Table 1 as sample stability. In the second part of the study, the tests were measured within 24 h as described above, using reagents kept at room temperature (time zero corresponding to the time when the reagents and controls were taken from the refrigerator and plasma samples were taken from the freezer). The results, expressed as means, and the CVs are reported in Table 1 as reagent stability. In both the sample and reagent stability evaluations, no increasing or decreasing trends were found, and all CVs calculated were lower than those obtained in the imprecision study. In particular, with the exception of the C3 sample in the CK-MB assay, lower CVs were observed in the sample stability study than in the reagent stability study, thus underlining the stability of epitopes recognized by the antibodies used in these assays, a fundamental analytical tool to be verified, in particular for the TnI assay (6). Concerning the analytical aspects evaluated for the TnI assay only, the lower detection limit, calculated by diluting a plasma pool (mean TnI value, 0.11 μg/L) and by comparing the target and measured (mean value from two replicates) concentrations (y = 0.912x + 0.011; r = 0.962), was 0.02 μg/L (Fig. 1 A), whereas the functional sensitivity (TnI concentration showing a CV = 20% in a between-day imprecision profile) was 0.03 μg/L. From the imprecision profile, the lowest concentration that assured a CV of 10% was 0.06 μg/L. Experimental data from the imprecision study (Table 1 ) and the imprecision profile confirmed that this analyzer assures CVs of 10–14% at TnI concentrations of 0.06–0.08 μg/L. The linearity study, evaluated by a stepwise dilution series of the plasma pool with a mean TnI value of 55.96 μg/L (mean value obtained from different dilutions), demonstrated decreasing linearity for concentrations >20 μg/L (analytical range of the assay declared by manufacturer, 0–50 μg/L); moreover, the bias according to Bland-Altman analysis (−2.02) was not statistically significant (95% CI, −3.32 to −0.72; Fig. 1B ). Finally, we tested 85 plasma samples from healthy subjects (37 females and 48 males; ages, 19–75 years); the upper reference value, calculated as the 97.5th percentile (7), was 0.03 μg/L (99th percentile, 0.05 μg/L). Method comparisons were carried out with the main instruments located in the Central Laboratory: RxL Dimension (Dade Behring) for TnI and CK-MB and Stratus II (Dade Behring) for Myo. This choice was made because at the time of the study, the immunometric method for RxL Myo was not available. The results showed the following linear regressions: SCS Myo = 1.07(Stratus II Myo) − 0.37; r = 0.990; Sy|x = 3.85 (n = 60; range, 15.1–473.1 μg/L); SCS CK-MB = 0.954(RxL CK-MB) − 0.119; r = 0.996; Sy|x = 1.55 (n = 50; range, 0–132.1 μg/L); SCS TnI = 0.693(RxL TnI) + 0.946; r = 0.966; Sy|x = 1.56 (n = 74; range, 0–49.02 μg/L). The best results according to Bland-Altman statistical analysis were obtained for CK-MB (bias = 0.76; 95% CI, −0.0420 to 1.56), whereas TnI (bias = 2.57; 95% CI, 1.48 to 3.65) showed a wide dispersion of values, particularly at high concentrations. This was probably related to the decreasing linearity for concentrations >20 μg/L, as confirmed by the improvement found in comparison studies, when only samples with TnI value <16 μg/L (n = 50; range, 0–15.52 μg/L) were analyzed [SCS TnI = 0.956(RxL TnI) + 0.049; r = 0.981; Sy|x = 0.94; bias = −0.103; 95% CI, −0.39 to 0.18]. For Myo determinations, the statistically significant bias found (−8.122; 95% CI, −13.79 to −2.46) stresses the need for better standardization, even for methods from the same manufacturer. In conclusion, our findings confirm the overall good performance of the Stratus CS analyzer, thus indicating the analytical reliability of all methods used currently, as observed elsewhere (8). Furthermore, the system evaluated satisfies the requirements specified in IFCC (7) and National Academy of Clinical Biochemistry (9) recommendations for the use of biochemical markers in acute coronary syndrome, thus allowing the simultaneous measurement of early (Myo) and definitive (troponin) markers for the rapid ruling in or out of patients, and of CK-MB when of clinical importance, in relation to hospital policy. Finally, the verified stability of reagents, samples, and calibration curves, as well as the on-board centrifugation capability, system and quality-control lockout, and the low turnaround time (14 min plus 4 min for further results) enhance the practicability of the analyzer for use by either laboratory or non-laboratory personnel in the management of patients with chest pain, even in the emergency department. Imprecision and sample-reagent stability studies. C1 and C3, commercial control samples; pl, plasma pool. Imprecision and sample-reagent stability studies. C1 and C3, commercial control samples; pl, plasma pool. TnI analytical sensitivity evaluated as regression analysis (A), and linearity as Bland-Altman analysis (B).