High-Sensitivity Generation 5 Cardiac Troponin T Sex- and Age-Specific 99th Percentiles in the CALIPER Cohort of Healthy Children and Adolescents

To the Editor: International societies and expert groups recommend using sex-specific upper reference limits (URLs)1 for interpretation of high-sensitivity cardiac troponin (hs-cTnI/T) assays (1, 2). Despite advances in our understanding of cTn in adults, the effect of age and sex on cTn in pediatric populations is less understood. The Canadian Laboratory Initiative on Pediatric Reference Intervals (CALIPER) previously determined 97.5th and 99th percentiles for the Abbott hs-cTnI assay in a healthy pediatric cohort (3). These data represented the first attempt to characterize hs-cTnI in healthy children but did not include neonates nor were sex-specific differences evident. It is unclear, however, what cTnT URLs would be in this population because cTnT is a different protein and there are differences in analytical sensitivity and specificity between hs-cTnT and hs-cTnI assays (1–4). Establishing pediatric reference values for cTn is important because its measurement has several clinical indications in pediatrics, including monitoring children with postoperative myocarditis and congenital pediatric heart disease. Serum samples from 598 children and adolescents were selected from the CALIPER biobank. Samples from participants <1 year were obtained from apparently healthy infants at local maternity wards or outpatient clinics. Samples were measured with the Roche hs-cTnT (Gen 5 assay) STAT assay on the cobas 8000 e602 analyzer and reported in whole numbers per laboratory practice recommendations (2), with 3 ng/L used to designate undetectable concentrations (4). Any result <3 ng/L was reported as 3 ng/L, consistent with prior work (4). Data were visually inspected, and partitions were confirmed by the Harris and Boyd method, as previously described (3). Although hs-cTnT concentrations from 0 to <1 year were normally distributed, hs-cTnT concentrations for both sexes of age 1 to <19 years were highly skewed and could not be normalized via transformation (3). No outliers were identified by the Reed method (3). The URLs were calculated with the nonparametric method (>120 participants) or robust method (>40 and <120 participants) (3), with 90% CIs determined according to Clinical and Laboratory Standards Institute guidelines (Fig. 1). Percentage of detectable hs-cTnT results based on Roche package insert limit of blank (3 ng/L) and limit of detection (5 ng/L) and Abbott hs-cTnI limit of detection (1 ng/L) from the International Federation of Clinical Chemistry Committee on Clinical Applications of Cardiac Bio-Markers (http://www.ifcc.org/media/477441/high-sensitivity-cardiac-troponin-i-and-t-assay-analytical-characteristics-designated-by-manufacturer-v08232018.pdf). hs-cTnT concentrations were markedly increased from 0 to <6 months and subsequently decreased and narrowed at 1 year. The sex-specific 99th percentiles from 1 to <19 years were 14 ng/L for boys and 11 ng/L for girls. Fig. 1 reports the percentage of hs-cTnT results detectable (>3 ng/L and >5 ng/L). Although the sensitivity of the Roche hs-cTnT and Abbott hs-cTnI assays in CALIPER could not be directly compared because different samples were measured, the percent detectable (hs-cTnI >1 ng/L) was markedly higher in the hs-cTnI study (3) as illustrated in Fig. 1. The clinical value of using sex-specific 99th percentiles for hs-cTn has been recently debated (1, 2). Several groups suggest that owing to intrinsic sex-specific pathophysiological differences in adults, sex-specific URLs would improve the accuracy of hs-cTn result interpretation, especially in the acute coronary syndrome setting (1, 2, 4). However, the pathophysiological reasons for increased cTn in children are not the same as those in adults (3). Importantly, the data illustrate a time-dependent decrease in hs-cTnT over the first year of life, with age-specific URLs being important in this population to identify myocardial injury. The higher concentrations in this age group could be a partial result of transient hypoxia at parturition and/or fetal cTn expression in skeletal muscle; however, more studies are needed to further investigate this finding. After 1 year, hs-cTnT 99th percentiles (girls = 11 ng/L; boys = 14 ng/L) are similar to what has been reported for the AACC Universal Sample Bank (USB) 99th percentiles (females = 10 ng/L; males = 16 ng/L) but lower than the US 5th-generation cTnT package insert cutoffs (females = 14 ng/L; males = 22 ng/L) for which the population selection criteria were not stated (4). In this regard, the lower CALIPER 99th percentiles for hs-cTnT, especially for boys, may be preferred in the pediatric population to prevent misdiagnoses. An important study limitation to consider is that there were fewer than 300 individuals in each subgroup after partitioning (2). It has been reported that 300 samples are required to attain a 95% probability that 99% of results will not exceed the estimated URL (i.e., tolerance level) (5). Thus, owing to smaller sample sizes in this study, reported 99th percentiles should be interpreted cautiously. However, at study onset, we had nearly 600 individuals, which would have been sufficiently powered to derive the 99th percentile, if there was no further partitioning required, similar to hs-cTnI (3). Another limitation was the lack of imaging or other cardiac biomarkers to confirm participant cardiovascular health. However, the similarity of CALIPER and AACC USB 99th percentiles suggests that this population was, in fact, a healthy population in a cardiovascular sense. In summary, the current study provides novel and useful information regarding the effect of age and sex on hs-cTnT concentrations in healthy children and fills an important gap for laboratories offering hs-cTnT testing by providing age- and sex-specific URLs. upper reference limit high-sensitivity cardiac troponin I high-sensitivity cardiac troponin T cardiac troponin Canadian Laboratory Initiative on Pediatric Reference Intervals. The authors thank all participants, their families and CALIPER volunteers who made this study possible. The authors also thank the technical and managerial staff at Mount Sinai Hospital (Toronto, ON) for sample analysis.