Analytical Evaluation of 5 High-Sensitivity Cardiac Troponin Assays at New Lower Sex-Specific 99th Percentile Upper Reference Limits

To the Editor: High-sensitivity cardiac troponin (hs-cTn) assays are recommended for diagnosing myocardial infarction, and current guidelines endorse sex-specific 99th percentile upper reference limits (URLs) because women generally have lower troponin concentrations (1). The DANSPOT trial recently established sex-specific URLs from a healthy Danish population (2, 3). Because analytical imprecision increases at lower concentrations, reduced female URLs may affect assay performance. Consequently, we evaluated whether 5 hs-cTn assays used in Danish clinical practice meet high-sensitivity criteria by (a) demonstrating acceptable imprecision, defined as a coefficient of variation ≤10% at these newly established female 99th percentile URLs (Supplemental Table 1), and (b) measuring cardiac troponin values above the limit of detection (LoD) in ≥50% of healthy men and women. Sex-specific 99th percentile URLs were derived from samples from 2287 healthy Danish blood donors according to IFCC 2018 recommendations. Details of the DANSPOT biobank and URL derivation have been published previously (2, 3). Five hs-cTn assays used in Danish clinical practice were evaluated: Siemens Atellica hs-cTnI assay (Siemens Atellica IM Analyzer), Siemens Dimension Vista hs-cTnI assay (Dimension Vista 1500 analyzer), Roche Diagnostics Elecsys 2010 hs-cTnT assay (Roche Cobas e801 analyzer), Abbott Alinity i STAT hs-cTnI assay (Abbott Alinity I analyzer), and QuidelOrtho VITROS hs-TnI assay (QuidelOrtho VITROS 5600). Repeatability and within-laboratory precision at the female 99th percentile URLs were assessed using plasma samples with troponin concentrations near the assay-specific female URLs. Plasma samples were thawed and pooled to create 40 vials containing cardiac troponin I or cardiac troponin T concentrations at or below the newly established female 99th percentile URLs. For 2 of the assays, new patient samples were used either to spike or create the pool. These pooled samples were then refrozen at −20°C. The Alinity assay was newly implemented, and thus complications with the samples run on the Alinity assay resulted in fewer than 40 vials; therefore, another pool was created with new samples to get enough vials. Plasma samples were measured according to the CLSI EP05-A3 20 × 2 × 2 protocol. However, strict adherence was challenging for most assays (VITROS, Roche, Vista, and Alinity) due to routine assay operations, resulting in unbalanced data sets. The DANSPOT reference population data were analyzed to determine the proportion of healthy individuals with measurable troponin concentrations above the LoD using manufacturer-reported LoDs. Results were evaluated using estimated glomerular filtration rate cutoffs of both ≥60 and ≥90 mL/min/1.73 m2 because the original DANSPOT URLs were derived using the more restrictive cutoff. Statistical analyses were performed in R version 4.4.2. Repeatability and within-laboratory precision were estimated using variance component analysis according to CLSI EP05-A3 recommendations. Outlier detection for normally distributed data sets (Roche, Atellica, and VITROS) was performed using the 2-sided Grubbs test for single outliers at a significance level of 99%, showing no outliers. For the nonnormal data sets, outliers were detected using quartiles and 1.5 times the interquartile range, showing 5 outliers for Vista and 14 for Alinity. Outliers were subsequently removed. ANOVA was applied to balanced data sets, whereas restricted maximum likelihood estimation was used for unbalanced data sets. Analyses were performed using the VCA package. Four of 5 assays achieved a CV ≤ 10% at the female 99th percentile URL or a lower concentration, with Abbott Alinity being the only assay exceeding this threshold. Within-laboratory precision ranged from 2.4% to 10.5%, with Siemens Atellica demonstrating the lowest imprecision and Abbott Alinity the highest. Repeatability ranged from 2.1% to 7.3% (Table 1A). Analytical performance and measurable concentrations above the LoD. Panel A shows the repeatability and within-laboratory precision using the CLSI EP05-A3 20 × 2 × 2 protocol for this study and manufacturers. Panel B shows measurable cardiac troponin concentrations above the LoD at different eGFR cutoffs. Abbreviations: n, number of replicates; LoD, limit of detection; NA, not available; eGFR, estimated glomerular filtration rate. For Siemens Atellica and VITROS, the reporting limit exceeded the manufacturer-reported LoD, preventing assessment of measurable concentrations above the LoD. Among the remaining assays, measurable concentrations above the LoD ranged from 37.8% to 93.6% in women, 66.4% to 95.6% in men, and 50.9% to 94.5% overall (Table 1B). Siemens Vista demonstrated the highest proportion of measurable concentrations, whereas Abbott Alinity showed the lowest. According to the insert, the Abbott Alinity LoD ranges from 0.7 to 1.6 ng/L, resulting in different proportions of measurable cTn. Using an estimated glomerular filtration rate cutoff of ≥60 mL/min/1.73 m2 instead of ≥90 mL/min/1.73 m2 increased the proportion of healthy individuals with measurable concentrations above the LoD for all assays. Four of 5 evaluated assays fulfilled the hs-cTn imprecision criterion of CV ≤10% at the newly established Danish female 99th percentile URLs. In contrast, among the assays for which assessment was possible, only Siemens Vista detected troponin concentrations above the LoD in at least 50% of healthy women. These findings suggest that assay performance was more consistently maintained for the imprecision criterion than for the LoD-based criterion. For Abbott Alinity, the selected LoD substantially influenced whether the assay fulfilled the LoD-based hs-cTn criterion. Similarly, the proportion of measurable concentrations above the LoD was affected by the estimated glomerular filtration rate threshold used to define the healthy reference population, consistent with previous reports showing increased troponin concentrations in individuals with mildly impaired renal function (4). Our findings are consistent with previous analytical evaluations of hs-cTn assays (5), and the present study extends existing evidence by evaluating assay performance at female-specific URLs derived from a Danish reference population while providing additional data on Abbott Alinity, Roche Cobas e801, and VITROS 5600. Several limitations should be acknowledged. First, LoD values were obtained from manufacturers and were not independently verified. Second, technical issues during the Abbott Alinity evaluation necessitated the use of 2 pools and resulted in more outliers than anticipated, which may have influenced the imprecision estimates. Owing to limited sample material, it was not feasible to repeat the CLSI protocol, which would have been preferred. In addition, differences in sample size may have contributed to variation in the estimated imprecision, as larger data sets generally yield more robust estimates. Finally, a full multisite reproducibility evaluation was not performed, and intersite variability may therefore not be fully captured. In conclusion, 4 of 5 evaluated assays fulfilled the recommended CV ≤10% imprecision criterion at the newly established Danish female-specific 99th percentile URLs. However, fulfillment of the LoD-based criterion was more variable, with only Siemens Vista detecting troponin concentrations above the LoD in at least 50% of healthy women. Supplemental material is available at The Journal of Applied Laboratory Medicine online. Data are available from the corresponding author upon reasonable request. Nonstandard Abbreviations: hs-cTn, high-sensitivity cardiac troponin; LoD, limit of detection; URL, upper reference limit. Author Contributions: The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list. Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Research Funding: The study is funded in part by grants from Helsefonden (20-B-0431), Mauritzen La Fountaine Foundation, The Danish Heart Foundation (22-R156-A10257-22206), and the Independent Research Fund Denmark (2096-00157B). The study has also received support from Roche Diagnostics in terms of reagents for the Roche hs-TnT assay. Disclosures: M. Dahl has consultancies with Takeda and Grifols.