Consideration of high-sensitivity troponin values below the 99th percentile at presentation: does it improve diagnostic accuracy?

Background: The introduction of high-sensitivity cardiac troponin (hs-cTn) assays allows the assessment of clinical decision values below the 99th percentile.
Methods: Final diagnosis and one-year mortality were adjudicated in a multicenter, prospective cohort of 1181 patients presenting with acute chest pain to the emergency department. Hs-cTnT (Roche) and cTnI-ultra (Siemens) were measured in a blinded fashion.
Results: At presentation hs-cTnT and cTnI-ultra were below the limit of blank (LOB) in 201 (17%) and 549 (47%) patients, below the 75th percentile in 379 (32%) and 623 (53%) patients, below the 95th percentile in 603 (51%) and 808 (68%), and below the 99th percentile in 748 (63%) and 913 (77%), respectively. Sensitivities for the diagnosis of AMI were 100.0% and 96.8% respectively for hs-cTnT and cTnI-ultra (LOB as cut-off value), 99.5% and 96.2% (75th percentile), 96.8% and 93.0% (95th percentile), and 94.1% and 88.1% (99th percentile). The proportion of patients correctly classified as having or not AMI increased from 32.9% (LOB as cut-off value) to 47.8% (75th percentile), 65.9% (95th percentile) and 77.3% (99th percentile) for hs-cTnT and from 61.2% to 67.3%, 81.9% and 89.3% respectively for cTnI-ultra. At 1 year, all-cause mortality was very low and similar for patients below all of these cut-off levels (between 0.7% and 1.5%, p=0.748 for all-groups comparison).
Conclusion: cTn should be considered as a continuous variable. Decision values below the 99th percentile (e.g. the 75th percentile) are associated with a very high NPV for the diagnosis of AMI, but have a lower accuracy than the 99th percentile.
(Copyright © 2013 Elsevier Ireland Ltd. All rights reserved.)