What troponin concentration should I flag as an abnormal result?
A: The Global Task Force for the Universal Definition of Myocardial Infarction defined a cardiac troponin (cTn) > 99th percentile—derived from a normal, healthy reference population—as the decision threshold for the diagnosis of myocardial infarction (MI). The 10% coefficient of variation (CV) is defined as the optimal total imprecision of cTn assays at the 99th percentile, as most contemporary cTn assays lack the ability to deliver a 10% CV at or below the 99th percentile, or the ability to measure cardiac troponin below the 99th percentile. However, assays with a CV < 20% are considered clinically acceptable and will not cause false-positive acute MI diagnoses, provided serial samples are used to interpret acute changes (Clin Chem 2005;51:2198–200). Today, defining a high-sensitivity cTn assay requires an imprecision of < 10% CV at the 99th percentile. Therefore, abnormal should only be defined at the 99th percentile as long as the CV is < 20%.
What are the clinical ramifications for not using the 99th percentile?
Using the 99th percentile to define abnormal cTn optimizes the clinical diagnostic sensitivity of cTn assays and current data supports cTn assays with imprecision up to a 20% CV for both diagnosis and/or risk stratification use. While precision remains an important parameter, it is not the only metric that defines utility of cTn assays. Furthermore, we now know that a sensitive cTn assay with slightly more imprecision will correctly identify more patients at risk than an insensitive one with excellent imprecision (Clin Chem 2010;56:941–3). While one could achieve the 10% CV metric simply by increasing the assay threshold to define abnormal results, the consequence would be a decrease in clinical sensitivity. Ultimately, using the 10% CV identifies fewer patients with disease and increases the time after myocardial injury when a first positive result would be identified, delaying diagnosis of a true MI and impeding appropriate treatment.
What is the 99th percentile and how is it established?
Traditional reference intervals for clinical laboratory assays are most commonly established with an approach consistent with Clinical and Laboratory Standards Institute (CLSI) EP28-A3c guideline. This guideline recommends a minimum of 120 reference subjects for nonparametric statistics to estimate limits of a central 95% reference interval with limits at the 2.5th and 97.5th percentiles. However, determining the 99th percentile for cTn differs from the CLSI recommendations. Both the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) Task Force and the European Society of Cardiology recommend a total of 300 normal subjects to properly determine the 99th percentile.
Furthermore, when evaluating sex and age differences by decade, a minimum of 300 subjects per group are needed, for example 300 males and 300 females. Subjects included in a cTn reference interval study should be deemed healthy via use of health questionnaires along with biomarkers such as NT-proBNP and eGFR to screen for subclinical disease and/or other co-morbidities.
In summary, while expert opinion leaders and guidelines will continue to advocate for more precise and sensitive cTn assays, we should continue to follow guideline recommendations and use the 99th percentile to define abnormal cTn results, placing emphasis on serial acute changes to aid in diagnosing MI from chronic myocardial disease.
Amy Saenger, PhD, DABCC, FACB, is the senior cardiac medical and scientific affairs manager at Roche Diagnostics and associate professor of laboratory medicine and pathology at Mayo Clinic.