Clin Chem Lab Med 2014; 52(11): 1657–1665
Magdalena Krintus*, Marek Kozinski, Pascal Boudry, Nuria Estañ Capell, Ursula Köller, Karl Lackner, Guillaume Lefèvre, Lieselotte Lennartz, Johannes Lotz, Antonio Mora Herranz, Mads Nybo, Mario Plebani, Maria B. Sandberg, Wolfgang Schratzberger, Jessie Shih, Øyvind Skadberg, Ahmed Taoufik Chargui, Martina Zaninotto and Grazyna Sypniewska
European multicenter analytical evaluation of the Abbott ARCHITECT STAT high sensitive troponin I immunoassay Abstract Background: International recommendations highlight the superior value of cardiac troponins (cTns) for early diagnosis of myocardial infarction along with analytical requirements of improved precision and detectability. In this multicenter study, we investigated the analytical performance of a new high sensitive cardiac troponin I (hscTnI) assay and its 99th percentile upper reference limit (URL). *Corresponding author: Magdalena Krintus, PhD, Department of Laboratory Medicine, Nicolaus Copernicus University, Collegium Medicum, 9 Sklodowskiej-Curie Street, 85-094 Bydgoszcz, Poland, Phone: +48 52 5853602, Fax: +48 52 5853603, E-mail: [email protected] Marek Kozinski: Department of Principles of Clinical Medicine, Nicolaus Copernicus University, Collegium Medicum in Bydgoszcz, Bydgoszcz, Poland Pascal Boudry: Department of Clinical Biology, CHR Mons-Hainaut, Mons, Belgium Nuria Estañ Capell and Antonio Mora Herranz: Department of Clinical Biochemistry, Dr. Peset University Hospital, Valencia, Spain Ursula Köller and Wolfgang Schratzberger: Institute for Laboratory Diagnostics, Vienna Hospital Association, Hospital Hietzing, Vienna, Austria Karl Lackner and Johannes Lotz: Laboratory Medicine, University Medical Center, Johannes Gutenberg University Mainz, Mainz, Germany Guillaume Lefèvre and Ahmed Taoufik Chargui: Department of Biochemistry and Hormonology, AP-HP, Hôpital Tenon, Paris, France Lieselotte Lennartz: Abbott Laboratories, Wiesbaden, Germany Mads Nybo and Maria B. Sandberg: Department of Clinical Biochemistry and Pharmacology, Odense University Hospital, Odense, Denmark Mario Plebani and Martina Zaninotto: Department of Laboratory Medicine, University Hospital, Padua, Italy Jessie Shih: Abbott Laboratories, Abbott Park, IL, USA Øyvind Skadberg: Laboratory of Clinical Biochemistry, Stavanger University Hospital, Stavanger, Norway Grazyna Sypniewska: Department of Laboratory Medicine, Nicolaus Copernicus University, Collegium Medicum in Bydgoszcz, Bydgoszcz, Poland
Methods: Laboratories from nine European countries evaluated the ARCHITECT STAT high sensitive troponin I (hs-TnI) immunoassay on the ARCHITECT i2000SR/i1000SR immunoanalyzers. Imprecision, limit of blank (LoB), limit of detection (LoD), limit of quantitation (LoQ) linearity of dilution, interferences, sample type, method comparisons, and 99th percentile URLs were evaluated in this study. Results: Total imprecision of 3.3%–8.9%, 2.0%–3.5% and 1.5%–5.2% was determined for the low, medium and high controls, respectively. The lowest cTnI concentration corresponding to a total CV of 10% was 5.6 ng/L. Common interferences, sample dilution and carryover did not affect the hs-cTnI results. Slight, but statistically significant, differences with sample type were found. Concordance between the investigated hs-cTnI assay and contemporary cTnI assay at 99th percentile cut-off was found to be 95%. TnI was detectable in 75% and 57% of the apparently healthy population using the lower (1.1 ng/L) and upper (1.9 ng/L) limit of the LoD range provided by the ARCHITECT STAT hs-TnI package insert, respectively. The 99th percentile values were gender dependent. Conclusions: The new ARCHITECT STAT hs-TnI assay with improved analytical features meets the criteria of high sensitive Tn test and will be a valuable diagnostic tool. Keywords: analytical evaluation; analytical performance; high sensitive troponin; 99th percentile upper reference limit. DOI 10.1515/cclm-2014-0107 Received January 29, 2014; accepted May 4, 2014; previously published online June 4, 2014
Introduction The Third Universal Definition of Myocardial Infarction highlights the importance of cardiac troponin (cTn) as the preferred biomarker for the detection of myocardial
Brought to you by | Linnaeus University - Växjö Authenticated Download Date | 10/14/14 5:37 AM
1658 Krintus et al.: hs-cTnI evaluation infarction (MI) with a cTn concentration above the 99th percentile upper reference limit (URL) as the cut-off for diagnosis of MI with a total CV ≤ 10% at this concentration [1, 2]. Evolving recommendations of scientific organizations have led to improvements in cTn assays to meet these requirements [1–4]. Only a limited number of contemporary sensitive cTn assays meet the 10% CV requirement and as a result, newer high sensitivity cardiac troponin (hs-cTn) assays have been developed to meet these criteria. Currently, there are no definitive requirements for what defines a high sensitivity assay versus a contemporary sensitive assay. A recommendation was recently published by the Task Force of International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) on Clinical Applications of Cardiac Biomarkers that suggests that highly sensitive assays have the ability to measure at least 50% and ideally over 95% of healthy individuals above the assay’s limit of detection (LoD) along with an imprecision (expressed as CV) of ≤ 10% at the 99th percentile URL [2, 5]. Further support for this recommendation comes from the Study Group on Biomarkers in Cardiology in their critical clinical concepts regarding analytical characteristics of high sensitive cTn assays [6]. In view of the lack of cTn standardization and harmonization, there is a substantial need for assay-dependent determination of the diagnostic threshold of hs-cTn, essential for the clinical decision making process according to current recommendations. In the context of the analytical requirements and the expectations of the clinicians, the purpose of our study was to validate analytical performance of the ARCHITECT STAT hs-TnI assay and to verify its 99th percentile URL.
Materials and methods The ARCHITECT STAT hs-TnI assay performed on the Abbott ARCHITECT i2000SR/i1000SR immunoanalyzers (Abbott Laboratories, Wiesbaden, Germany) was evaluated in nine clinical laboratories in the following European countries: Austria, Belgium, Denmark, France, Germany, Italy, Norway, Poland and Spain from January through August 2013.
antibody coated microparticles whose epitope binding specificity is directed against amino acids 24–40 on the TnI protein [7, 8]. In the second step, a specific mouse-human chimeric, monoclonal anti-troponin I acridinium-labeled antibody acts as a detection antibody. This chimeric antibody, which has an epitope binding specificity directed against amino acids 41–49 on the cTnI molecule was designed to minimize the susceptibility to interferences by heterophilic antibodies.
Analytical performance A 5-day imprecision study was conducted at eight sites, according to the National Committee for Clinical Laboratory Standards/Clinical and Laboratory Standards Institute (NCCLS/CLSI) guidelines [9]. Two ARCHITECT runs were performed on each day with a minimum of 2 h separating the runs. Three cTnI control levels (low, medium and high across the range 12–21,000 ng/L) were assayed during the 5-day evaluation in replicates of two on each run for a total of 20 replicates for each control. Mean, standard deviation (SD) and CVs for each site were calculated. For the precision profile lithium heparin panels were prepared and tested at one site (France) using a high-concentration stock of cTnI derived from pooled clinical samples spiked into a negative serum pool. Samples were tested over 21 days in independent assay runs. Results were pooled and a model fit was created to estimate concentrations at 20%, 15% and 10% CV. Limit of blank (LoB), LoD and limit of quantitation (LoQ) were verified according to the CLSI guidelines at eight sites [10]. LoB was estimated by measuring 20 replicates of zero calibrator and calculating the mean result and the SD. LoD was determined using 20 replicates of a sample with concentration equal to the claimed LoD and estimation of the proportion of results exceeding the LoB was calculated. For LoQ a minimum of 30 replicates of a sample with a concentration close to the claimed LoQ were run. Dilution linearity of the hs-cTnI immunoassay was evaluated at five sites according to the CLSI EP6-A [11]. Low and high TnI pools with undiluted values that ranged between 10–50,000 ng/L were used. The percent recovery was calculated. A mean recovery ( ± SD) of 100% ( ± 20%) was required to demonstrate acceptable linearity of dilution. Evaluation of potential interference by hemoglobin, bilirubin and lipids was performed based on guidance from the CLSI document EP7-A2 [12]. For all analytes the change in concentration was compared and the recovery was calculated. Interference was regarded acceptable if the cTnI recovery was within 100% ± 10%.
Carryover The cTnI carryover was assessed by analyzing high cTn concentration heparin plasma pools followed by low cTn concentration heparin plasma pools. The percentage of carryover was calculated.
Assay description and procedure This assay is a two-step, double-monoclonal chemiluminescent microparticle immunoassay (CMIA) for the detection of cTnI in human serum or plasma. According to the manufacturer’s insert, this assay uses a sample volume of 160 μL. In the first step, cTnI present in the sample binds to the specific mouse monoclonal anti-cTnI (capture)
Sample type comparison Sample type comparison was performed at two sites. At the first site, 31 matched samples obtained from randomly selected patients with cTnI concentrations above the LoD were collected into serum,
Brought to you by | Linnaeus University - Växjö Authenticated Download Date | 10/14/14 5:37 AM
Krintus et al.: hs-cTnI evaluation 1659 ethylenediaminetetraacetic acid (EDTA) and lithium heparin plasma tubes. Each specimen type was measured in duplicate. The mean results were compared and the significance of the differences was evaluated. At the second site, 26 frozen paired serum and lithium heparin plasma samples were evaluated in duplicate. Both sites performed linear regression and Bland-Altman analyses.
percentiles. Comparison between the groups was performed by using the Mann-Whitney U-test. Dilution linearity was determined by least-squares regression analysis. Statistical comparison between different sample types was assessed with the Wilcoxon rank-sum test for paired samples. Passing-Bablok and Deming regression analyses were performed for both the sample and method comparison. The 99th percentile values were determined by right-sided nonparametric statistics as described in the CLSI C28-A3c [13]. Ninety percent confidence intervals (CI) were provided for the 99th percentile values, if possible. We excluded the outlier observations from the analyzed population based on the method described by Reed et al. [14]. Statistical analysis was performed using Analyse-it 2.2 for Microsoft Excel (Analyse-it Software, Leeds, UK) and MedCalc 12.7.0 (MedCalc Software, Ostend, Belgium).
Method comparison Analytical correlation of the ARCHITECT STAT hs-TnI to the contemporary ARCHITECT STAT TnI assay was evaluated using 2598 samples that spanned the dynamic range of the hs-TnI as well as the contemporary TnI assay. Correlation, slope, and y-intercept according to Passing-Bablok and Deming regression analyses were calculated. A concordance analysis was performed in order to assess the degree of agreement between both assays.
Results
99th percentile URL
Performance characteristics
For estimation of the 99th percentile URL, a cohort of 1769 apparently healthy individuals and blood donors recruited in seven countries. 993 women and 776 men, aged between 18 and 91 years, were selected. Individuals were either blood donors (2 countries, n = 409) or recruited for a reference range study (5 countries, n = 1360). Prior to the initiation of participant enrollment, the study was approved by local Ethics Committees/Investigational Review Boards, as required. All study participants completed questionnaires detailing their medical histories and medications in order to confirm their good health before the blood draw. This healthy reference population was defined as individuals who were free of cardiac diseases (no history of cardiac disease, cardiac treatment/medication or cardiac intervention). Additional exclusion criteria included pregnancy, current infection and any chronic inflammatory disease.
Results from the 5-day precision evaluation are summarized in Table 1. Total imprecision ranged between 3.3% and 8.9%, 2.0% and 3.5%, and 1.5% and 5.2% for the low, medium and high controls, respectively. Total interlaboratory CV at all sites was lower than 10% for samples ranging from 18.4 to 15,912.3 ng/L, supporting the manufacturer’s claim of “total imprecision of ≤ 10% CV with controls or panels across the range of 10–50,000 pg/mL (ng/L)”. The lowest cTnI concentration corresponding to a total CV of 10% was 5.6 ng/L (Figure 1); a total CV of 20%, 15% and 5% was at 2.2, 3.6 and 9 ng/L, respectively. The LoB and LoD results at all sites verified the manufacturer’s claim of LoB range of 0.7–1.3 ng/L and LoD range of 1.1–1.9 ng/L. The manufacturer’s LoQ was verified at ≤ 10 ng/L with a 10% CV (Table 1).
Statistical analysis Categorical data were summarized with number and percentages. Data were expressed as mean ± SD and median with 25th–75th
Table 1 Evaluation of precision and LoQ in eight sites following the CLSI EP5-A protocol [9]. Site
1 2 3 4 5 6 7 8
Country
Belgium Denmark France Germany Italy Norway Poland Spain
Instrument
i2000SR i2000SR -1 i2000SR -2 i2000SR -3 i2000SR -1 i2000SR -2 i2000SR i2000SR i2000SR i2000SR i1000SR
LoQ ng/L
7.7 4.6
5.8 4.6 7.8 7.2 4.8 8.1
%CV at LoQ
10.6 8.4
8.2 6.8 7.1 6.7 8.0 7.8
Low control ng/L
% CV
19.0 18.4 19.7 20.4 18.9 19.9 19.5 20.6 20.1 18.8 21.5
3.3 3.3 3.5 8.9 3.4 4.4 3.9 4.3 3.9 5.3 4.8
Medium control ng/L
189.2 190.2 193.1 201.2 186.3 186.0 194.4 193.7 193.8 187.5 198.7
CLSI, Clinical and Laboratory Standards Institute; CV, coefficient of variation; LoQ, limit of quantitation.
Brought to you by | Linnaeus University - Växjö Authenticated Download Date | 10/14/14 5:37 AM
% CV
2.5 2.4 3.5 3.1 2.4 2.0 3.0 2.5 2.8 3.2 3.0
High control ng/L
14,032.0 14,593.4 14,742.2 14,779.6 13,994.2 14,602.1 14,935.9 15,690.2 14,610.6 15,117.8 15,912.3
% CV
2.6 1.8 2.2 2.4 2.5 1.5 2.9 1.5 1.6 5.2 2.1
1660 Krintus et al.: hs-cTnI evaluation
Sample type comparison
14
12
CV, %
10
8
6
4 0
5
10
15
20
25
30
35
40
45
hs-Tnl, ng/L
Figure 1 Imprecision profile for a hs-cTnI assay on the Abbott ARCHITECT immunoanalyzer (Abbott Laboratories, Wiesbaden, Germany). hs-cTnI, high sensitivity cardiac troponin I.
Dilution linearity was evaluated at five sites using 26 plasma samples. Slopes of observed versus expected values ranged from 0.904 to 0.981 and the correlation coefficient did not differ significantly from 1.0. Overall recovery was 90.6%. The overall correlation was 0.988 over the range of 9.60–20,097.10 ng/L with a slope of 0.971 and intercept of 175.29. No interferences with hs-cTnI results were seen by elevated levels of bilirubin, hemoglobin and triglycerides. A change in hs-cTnI concentration of 500 mg/dL. The mean recovery of hs-cTnI ranged from 96% to 113%. The amount of plasma carryover was determined to be 0.006%.
The Passing-Bablok and Bland-Altman results for the first experiment carried out on 31 matched serum, EDTA plasma and lithium heparin plasma samples are summarized in Table 2. Results for serum and EDTA plasma did not show a significant difference using the Wilcoxon ranksum test (p = 0.32), but statistically significant differences were found between lithium heparin and EDTA plasma, and between lithium heparin plasma and serum (p
Magdalena Krintus*, Marek Kozinski, Pascal Boudry, Nuria Estañ Capell, Ursula Köller, Karl Lackner, Guillaume Lefèvre, Lieselotte Lennartz, Johannes Lotz, Antonio Mora Herranz, Mads Nybo, Mario Plebani, Maria B. Sandberg, Wolfgang Schratzberger, Jessie Shih, Øyvind Skadberg, Ahmed Taoufik Chargui, Martina Zaninotto and Grazyna Sypniewska
European multicenter analytical evaluation of the Abbott ARCHITECT STAT high sensitive troponin I immunoassay Abstract Background: International recommendations highlight the superior value of cardiac troponins (cTns) for early diagnosis of myocardial infarction along with analytical requirements of improved precision and detectability. In this multicenter study, we investigated the analytical performance of a new high sensitive cardiac troponin I (hscTnI) assay and its 99th percentile upper reference limit (URL). *Corresponding author: Magdalena Krintus, PhD, Department of Laboratory Medicine, Nicolaus Copernicus University, Collegium Medicum, 9 Sklodowskiej-Curie Street, 85-094 Bydgoszcz, Poland, Phone: +48 52 5853602, Fax: +48 52 5853603, E-mail: [email protected] Marek Kozinski: Department of Principles of Clinical Medicine, Nicolaus Copernicus University, Collegium Medicum in Bydgoszcz, Bydgoszcz, Poland Pascal Boudry: Department of Clinical Biology, CHR Mons-Hainaut, Mons, Belgium Nuria Estañ Capell and Antonio Mora Herranz: Department of Clinical Biochemistry, Dr. Peset University Hospital, Valencia, Spain Ursula Köller and Wolfgang Schratzberger: Institute for Laboratory Diagnostics, Vienna Hospital Association, Hospital Hietzing, Vienna, Austria Karl Lackner and Johannes Lotz: Laboratory Medicine, University Medical Center, Johannes Gutenberg University Mainz, Mainz, Germany Guillaume Lefèvre and Ahmed Taoufik Chargui: Department of Biochemistry and Hormonology, AP-HP, Hôpital Tenon, Paris, France Lieselotte Lennartz: Abbott Laboratories, Wiesbaden, Germany Mads Nybo and Maria B. Sandberg: Department of Clinical Biochemistry and Pharmacology, Odense University Hospital, Odense, Denmark Mario Plebani and Martina Zaninotto: Department of Laboratory Medicine, University Hospital, Padua, Italy Jessie Shih: Abbott Laboratories, Abbott Park, IL, USA Øyvind Skadberg: Laboratory of Clinical Biochemistry, Stavanger University Hospital, Stavanger, Norway Grazyna Sypniewska: Department of Laboratory Medicine, Nicolaus Copernicus University, Collegium Medicum in Bydgoszcz, Bydgoszcz, Poland
Methods: Laboratories from nine European countries evaluated the ARCHITECT STAT high sensitive troponin I (hs-TnI) immunoassay on the ARCHITECT i2000SR/i1000SR immunoanalyzers. Imprecision, limit of blank (LoB), limit of detection (LoD), limit of quantitation (LoQ) linearity of dilution, interferences, sample type, method comparisons, and 99th percentile URLs were evaluated in this study. Results: Total imprecision of 3.3%–8.9%, 2.0%–3.5% and 1.5%–5.2% was determined for the low, medium and high controls, respectively. The lowest cTnI concentration corresponding to a total CV of 10% was 5.6 ng/L. Common interferences, sample dilution and carryover did not affect the hs-cTnI results. Slight, but statistically significant, differences with sample type were found. Concordance between the investigated hs-cTnI assay and contemporary cTnI assay at 99th percentile cut-off was found to be 95%. TnI was detectable in 75% and 57% of the apparently healthy population using the lower (1.1 ng/L) and upper (1.9 ng/L) limit of the LoD range provided by the ARCHITECT STAT hs-TnI package insert, respectively. The 99th percentile values were gender dependent. Conclusions: The new ARCHITECT STAT hs-TnI assay with improved analytical features meets the criteria of high sensitive Tn test and will be a valuable diagnostic tool. Keywords: analytical evaluation; analytical performance; high sensitive troponin; 99th percentile upper reference limit. DOI 10.1515/cclm-2014-0107 Received January 29, 2014; accepted May 4, 2014; previously published online June 4, 2014
Introduction The Third Universal Definition of Myocardial Infarction highlights the importance of cardiac troponin (cTn) as the preferred biomarker for the detection of myocardial
Brought to you by | Linnaeus University - Växjö Authenticated Download Date | 10/14/14 5:37 AM
1658 Krintus et al.: hs-cTnI evaluation infarction (MI) with a cTn concentration above the 99th percentile upper reference limit (URL) as the cut-off for diagnosis of MI with a total CV ≤ 10% at this concentration [1, 2]. Evolving recommendations of scientific organizations have led to improvements in cTn assays to meet these requirements [1–4]. Only a limited number of contemporary sensitive cTn assays meet the 10% CV requirement and as a result, newer high sensitivity cardiac troponin (hs-cTn) assays have been developed to meet these criteria. Currently, there are no definitive requirements for what defines a high sensitivity assay versus a contemporary sensitive assay. A recommendation was recently published by the Task Force of International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) on Clinical Applications of Cardiac Biomarkers that suggests that highly sensitive assays have the ability to measure at least 50% and ideally over 95% of healthy individuals above the assay’s limit of detection (LoD) along with an imprecision (expressed as CV) of ≤ 10% at the 99th percentile URL [2, 5]. Further support for this recommendation comes from the Study Group on Biomarkers in Cardiology in their critical clinical concepts regarding analytical characteristics of high sensitive cTn assays [6]. In view of the lack of cTn standardization and harmonization, there is a substantial need for assay-dependent determination of the diagnostic threshold of hs-cTn, essential for the clinical decision making process according to current recommendations. In the context of the analytical requirements and the expectations of the clinicians, the purpose of our study was to validate analytical performance of the ARCHITECT STAT hs-TnI assay and to verify its 99th percentile URL.
Materials and methods The ARCHITECT STAT hs-TnI assay performed on the Abbott ARCHITECT i2000SR/i1000SR immunoanalyzers (Abbott Laboratories, Wiesbaden, Germany) was evaluated in nine clinical laboratories in the following European countries: Austria, Belgium, Denmark, France, Germany, Italy, Norway, Poland and Spain from January through August 2013.
antibody coated microparticles whose epitope binding specificity is directed against amino acids 24–40 on the TnI protein [7, 8]. In the second step, a specific mouse-human chimeric, monoclonal anti-troponin I acridinium-labeled antibody acts as a detection antibody. This chimeric antibody, which has an epitope binding specificity directed against amino acids 41–49 on the cTnI molecule was designed to minimize the susceptibility to interferences by heterophilic antibodies.
Analytical performance A 5-day imprecision study was conducted at eight sites, according to the National Committee for Clinical Laboratory Standards/Clinical and Laboratory Standards Institute (NCCLS/CLSI) guidelines [9]. Two ARCHITECT runs were performed on each day with a minimum of 2 h separating the runs. Three cTnI control levels (low, medium and high across the range 12–21,000 ng/L) were assayed during the 5-day evaluation in replicates of two on each run for a total of 20 replicates for each control. Mean, standard deviation (SD) and CVs for each site were calculated. For the precision profile lithium heparin panels were prepared and tested at one site (France) using a high-concentration stock of cTnI derived from pooled clinical samples spiked into a negative serum pool. Samples were tested over 21 days in independent assay runs. Results were pooled and a model fit was created to estimate concentrations at 20%, 15% and 10% CV. Limit of blank (LoB), LoD and limit of quantitation (LoQ) were verified according to the CLSI guidelines at eight sites [10]. LoB was estimated by measuring 20 replicates of zero calibrator and calculating the mean result and the SD. LoD was determined using 20 replicates of a sample with concentration equal to the claimed LoD and estimation of the proportion of results exceeding the LoB was calculated. For LoQ a minimum of 30 replicates of a sample with a concentration close to the claimed LoQ were run. Dilution linearity of the hs-cTnI immunoassay was evaluated at five sites according to the CLSI EP6-A [11]. Low and high TnI pools with undiluted values that ranged between 10–50,000 ng/L were used. The percent recovery was calculated. A mean recovery ( ± SD) of 100% ( ± 20%) was required to demonstrate acceptable linearity of dilution. Evaluation of potential interference by hemoglobin, bilirubin and lipids was performed based on guidance from the CLSI document EP7-A2 [12]. For all analytes the change in concentration was compared and the recovery was calculated. Interference was regarded acceptable if the cTnI recovery was within 100% ± 10%.
Carryover The cTnI carryover was assessed by analyzing high cTn concentration heparin plasma pools followed by low cTn concentration heparin plasma pools. The percentage of carryover was calculated.
Assay description and procedure This assay is a two-step, double-monoclonal chemiluminescent microparticle immunoassay (CMIA) for the detection of cTnI in human serum or plasma. According to the manufacturer’s insert, this assay uses a sample volume of 160 μL. In the first step, cTnI present in the sample binds to the specific mouse monoclonal anti-cTnI (capture)
Sample type comparison Sample type comparison was performed at two sites. At the first site, 31 matched samples obtained from randomly selected patients with cTnI concentrations above the LoD were collected into serum,
Brought to you by | Linnaeus University - Växjö Authenticated Download Date | 10/14/14 5:37 AM
Krintus et al.: hs-cTnI evaluation 1659 ethylenediaminetetraacetic acid (EDTA) and lithium heparin plasma tubes. Each specimen type was measured in duplicate. The mean results were compared and the significance of the differences was evaluated. At the second site, 26 frozen paired serum and lithium heparin plasma samples were evaluated in duplicate. Both sites performed linear regression and Bland-Altman analyses.
percentiles. Comparison between the groups was performed by using the Mann-Whitney U-test. Dilution linearity was determined by least-squares regression analysis. Statistical comparison between different sample types was assessed with the Wilcoxon rank-sum test for paired samples. Passing-Bablok and Deming regression analyses were performed for both the sample and method comparison. The 99th percentile values were determined by right-sided nonparametric statistics as described in the CLSI C28-A3c [13]. Ninety percent confidence intervals (CI) were provided for the 99th percentile values, if possible. We excluded the outlier observations from the analyzed population based on the method described by Reed et al. [14]. Statistical analysis was performed using Analyse-it 2.2 for Microsoft Excel (Analyse-it Software, Leeds, UK) and MedCalc 12.7.0 (MedCalc Software, Ostend, Belgium).
Method comparison Analytical correlation of the ARCHITECT STAT hs-TnI to the contemporary ARCHITECT STAT TnI assay was evaluated using 2598 samples that spanned the dynamic range of the hs-TnI as well as the contemporary TnI assay. Correlation, slope, and y-intercept according to Passing-Bablok and Deming regression analyses were calculated. A concordance analysis was performed in order to assess the degree of agreement between both assays.
Results
99th percentile URL
Performance characteristics
For estimation of the 99th percentile URL, a cohort of 1769 apparently healthy individuals and blood donors recruited in seven countries. 993 women and 776 men, aged between 18 and 91 years, were selected. Individuals were either blood donors (2 countries, n = 409) or recruited for a reference range study (5 countries, n = 1360). Prior to the initiation of participant enrollment, the study was approved by local Ethics Committees/Investigational Review Boards, as required. All study participants completed questionnaires detailing their medical histories and medications in order to confirm their good health before the blood draw. This healthy reference population was defined as individuals who were free of cardiac diseases (no history of cardiac disease, cardiac treatment/medication or cardiac intervention). Additional exclusion criteria included pregnancy, current infection and any chronic inflammatory disease.
Results from the 5-day precision evaluation are summarized in Table 1. Total imprecision ranged between 3.3% and 8.9%, 2.0% and 3.5%, and 1.5% and 5.2% for the low, medium and high controls, respectively. Total interlaboratory CV at all sites was lower than 10% for samples ranging from 18.4 to 15,912.3 ng/L, supporting the manufacturer’s claim of “total imprecision of ≤ 10% CV with controls or panels across the range of 10–50,000 pg/mL (ng/L)”. The lowest cTnI concentration corresponding to a total CV of 10% was 5.6 ng/L (Figure 1); a total CV of 20%, 15% and 5% was at 2.2, 3.6 and 9 ng/L, respectively. The LoB and LoD results at all sites verified the manufacturer’s claim of LoB range of 0.7–1.3 ng/L and LoD range of 1.1–1.9 ng/L. The manufacturer’s LoQ was verified at ≤ 10 ng/L with a 10% CV (Table 1).
Statistical analysis Categorical data were summarized with number and percentages. Data were expressed as mean ± SD and median with 25th–75th
Table 1 Evaluation of precision and LoQ in eight sites following the CLSI EP5-A protocol [9]. Site
1 2 3 4 5 6 7 8
Country
Belgium Denmark France Germany Italy Norway Poland Spain
Instrument
i2000SR i2000SR -1 i2000SR -2 i2000SR -3 i2000SR -1 i2000SR -2 i2000SR i2000SR i2000SR i2000SR i1000SR
LoQ ng/L
7.7 4.6
5.8 4.6 7.8 7.2 4.8 8.1
%CV at LoQ
10.6 8.4
8.2 6.8 7.1 6.7 8.0 7.8
Low control ng/L
% CV
19.0 18.4 19.7 20.4 18.9 19.9 19.5 20.6 20.1 18.8 21.5
3.3 3.3 3.5 8.9 3.4 4.4 3.9 4.3 3.9 5.3 4.8
Medium control ng/L
189.2 190.2 193.1 201.2 186.3 186.0 194.4 193.7 193.8 187.5 198.7
CLSI, Clinical and Laboratory Standards Institute; CV, coefficient of variation; LoQ, limit of quantitation.
Brought to you by | Linnaeus University - Växjö Authenticated Download Date | 10/14/14 5:37 AM
% CV
2.5 2.4 3.5 3.1 2.4 2.0 3.0 2.5 2.8 3.2 3.0
High control ng/L
14,032.0 14,593.4 14,742.2 14,779.6 13,994.2 14,602.1 14,935.9 15,690.2 14,610.6 15,117.8 15,912.3
% CV
2.6 1.8 2.2 2.4 2.5 1.5 2.9 1.5 1.6 5.2 2.1
1660 Krintus et al.: hs-cTnI evaluation
Sample type comparison
14
12
CV, %
10
8
6
4 0
5
10
15
20
25
30
35
40
45
hs-Tnl, ng/L
Figure 1 Imprecision profile for a hs-cTnI assay on the Abbott ARCHITECT immunoanalyzer (Abbott Laboratories, Wiesbaden, Germany). hs-cTnI, high sensitivity cardiac troponin I.
Dilution linearity was evaluated at five sites using 26 plasma samples. Slopes of observed versus expected values ranged from 0.904 to 0.981 and the correlation coefficient did not differ significantly from 1.0. Overall recovery was 90.6%. The overall correlation was 0.988 over the range of 9.60–20,097.10 ng/L with a slope of 0.971 and intercept of 175.29. No interferences with hs-cTnI results were seen by elevated levels of bilirubin, hemoglobin and triglycerides. A change in hs-cTnI concentration of 500 mg/dL. The mean recovery of hs-cTnI ranged from 96% to 113%. The amount of plasma carryover was determined to be 0.006%.
The Passing-Bablok and Bland-Altman results for the first experiment carried out on 31 matched serum, EDTA plasma and lithium heparin plasma samples are summarized in Table 2. Results for serum and EDTA plasma did not show a significant difference using the Wilcoxon ranksum test (p = 0.32), but statistically significant differences were found between lithium heparin and EDTA plasma, and between lithium heparin plasma and serum (p
