Myocardial Injury Thresholds for 4 High-Sensitivity Troponin Assays in a Population-Based Sample of US Children and Adolescents

Abstract

Background: Myocardial injury is an important pediatric diagnosis. Establishing normative data from a representative pediatric sample is vital to provide accurate upper reference limits (URLs) for defining myocardial injury using high-sensitivity cardiac troponin. Methods: Among participants 1 to 18 years of age in the 1999–2004 National Health and Nutrition Examination Survey, we measured high-sensitivity troponin T using one assay (Roche) and high-sensitivity troponin I using 3 assays (Abbott, Siemens, and Ortho). In a strictly defined healthy subgroup, we estimated 97.5th and 99th percentile URLs for each assay using the recommended nonparametric method. Results: Of 5695 pediatric participants, 4029 met criteria for the healthy subgroup (50% males; mean age 12.6 years). Our 99th percentile URL estimates for all 4 high-sensitivity troponin assays among children and adolescents were lower than the manufacturer-reported URLs (derived from adults). The 99th percentile URLs (95% CI) were 15 ng/L (95% CI, 12–17) for high-sensitivity troponin T, 16 ng/L (95% CI, 12–19) for high-sensitivity troponin I with the Abbott assay, 38 ng/L (95% CI, 25–46) for high-sensitivity troponin I with the Siemens assay, and 7 ng/L (95% CI, 5, 12) for high-sensitivity troponin I with the Ortho assay. The 95% CIs for age-, sex-, and race and ethnicity-specific 99th percentile URLs overlapped. However, the 97.5th percentile URL for each assay was measured with superior statistical precision (ie, tighter 95% CIs) and demonstrated differences by sex. For male compared with female children and adolescents, 97.5th percentile URLs were 11 ng/L (95% CI, 10–12) versus 6 ng/L (95% CI, 6–7) for high-sensitivity troponin T, 9 ng/L (95% CI, 7–10) versus 5 ng/L (95% CI, 4–6) for high-sensitivity troponin I with the Abbott assay, 21 ng/L (95% CI, 18–25) versus 11 ng/L (95% CI, 9–13) for high-sensitivity troponin I with the Siemens assay, and 4 ng/L (95% CI, 3–5) versus 2 ng/L (95% CI, 1–3) for high-sensitivity troponin I with the Ortho assay. In contrast to the 99th percentiles, the point estimates of 97.5th percentile pediatric URLs for high-sensitivity troponin were also much more stable to differences in the analytic approaches taken to estimate URLs. Conclusions: Because myocardial infarction is rare in children and adolescents, the use of statistically more precise and reliable sex-specific 97.5th percentile high-sensitivity troponin URLs might be considered to define pediatric myocardial injury.