High-Sensitivity Troponin Results Influenced by Age

With high-sensitivity cardiac troponin assays, the upper reference limits (URLs) used to pinpoint damage to the heart differ depending not only on the sex of the individual being tested but also their age, new research shows.

Myocardial injury is defined by the 2018 Fourth Universal Definition of MI (UDMI) as an abnormal cardiac troponin concentration that exceeds the benchmark of the 99th percentile URL.

“However, the URLs in current clinical use are typically those reported by manufacturers and were derived from convenience samples using inconsistent approaches, with heterogeneous or incomplete phenotypic information necessary to define a truly ‘healthy’ reference sample,” John W. McEvoy, MBBCh, PhD (University of Galway, Ireland, and Johns Hopkins University, Baltimore, MD), and colleagues write in the Journal of the American College of Cardiology. The existing URLs, they add, also are based on populations that skew younger in comparison to the patients who typically seek cardiovascular care.

To get a clearer look at variations with four high-sensitivity cardiac troponin assays on the market, the investigators turned to the 1999-2004 National Health and Nutrition Examination Survey (NHANES) to obtain a diverse, representative sample of healthy US adults.

“What emerged from our analyses was that there are significant differences by age, which we think are important to acknowledge,” McEvoy told TCTMD, adding that the UDMI currently endorses sex-specific thresholds to improve diagnostic accuracy. “What I believe our data have clarified and now [should] motivate a serious conversation about is the use of age-specific cut points for high-sensitivity troponin in the diagnosis of myocardial injury.”

Much like age-adjusted D-dimer thresholds for pulmonary embolism, high-sensitivity troponin assays may soon take this variable into account, he predicted. The idea “has been broached before, but the data haven’t been strong enough to allow the kind of confidence you need to implement [age cutoffs] clinically.”

McEvoy hopes that their results will “create a compelling case for reconsideration of this in the forthcoming Universal Definition of MI” now in the works. “At present,” he said, “we are potentially overdetecting myocardial injury in older adults, because we’re using 99th percentiles that don’t reflect this age phenomenon.”

Sex and Age, but Not Race/Ethnicity

Within the database of 12,545 NHANES participants, 2,746 (22%) qualified as healthy based on criteria including body mass index, use of medications to control CV risk factors, comorbidities, kidney function, and NT-proBNP level, among others. This subset, of whom half were men, had a mean age of 37 years.

For high-sensitivity cardiac troponin T (hs-cTnT), tested with a Roche assay, the 99th percentile URL in the healthy NHANES participants matched that reported by the manufacturer, at 19 ng/L. For high-sensitivity cardiac troponin I (hs-cTnI), the 99th percentile URL was consistently lower in this healthy subgroup than had been reported by the tests’ manufacturers (13 vs 28 ng/L with Abbott, 5 vs 11 ng/L with Ortho, and 37 vs 46.5 ng/L with Siemens).

These “results suggest that cohorts previously used by manufacturers to derive 99th percentile URLs may not have fully excluded people with underlying health conditions,” McEvoy et al note.

At present, we are potentially overdetecting myocardial injury in older adults. John W. McEvoy

As has been found before, there were sex differences: the 99th percentile URL values with each of the four assays in the NHANES subgroup were higher for men versus women (P < 0.001).

Additionally, individuals younger than 40 years had significantly lower 99th percentile URLs than did adults 60 years and older—though none of the tests’ manufacturers currently report URLs tailored to age. The Siemens hs-cTnI assay, for instance, had 99th percentile URLs of 33, 43, and 51 ng/dL for the healthy NHANES participants aged 18-39, 40-59, and ≥ 60 years, respectively.

The NHANES subgroup contained 405 adults in the eldest age bracket, all of whom qualified as healthy based on strict criteria. But “the URLs in current use (often derived by manufacturers using samples that included mostly young adults) may not apply well to older persons, even though these are the individuals who most commonly present for clinical care and are evaluated for myocardial injury,” the authors explain.

No differences were found based on race/ethnicity.

Edging Towards Real-World Use

McEvoy pointed out that, although their research was done independently, exploration of URL variations through the NHANES data set is something the assay “manufacturers are really keen to see as well.” In a related study, published online last month in Circulation, he and his colleagues looked at the same question but focused on children and adolescents in NHANES.

As a next step, “I’d love to see our age-specific 99th percentiles be studied in patients who are presenting for medical care with chest pain to see how our cut points perform in that context,” said McEvoy.

Cian P. McCarthy, MBBCh, BAO (Massachusetts General Hospital, Boston), and colleagues, writing in an accompanying editorial, agree that the NHANES results “have implications on how we define and interpret normal and abnormal troponin values and highlight the importance of using a standardized process for deriving the URL for these widely used assays.”

They, like McEvoy did in speaking with TCTMD, emphasize that acute myocardial “injury—defined by an hs-cTn rise and/or fall around the 99th percentile [URL]—is only one component of the Universal Definition of MI, and does not alone define it.”

As to the question of whether both age- and sex-specific thresholds should be adopted, the editorialists note that “such cutoffs may further complicate MI diagnostic criteria.” Still, they say, “this is superseded by the benefits of improved diagnostic accuracy in younger and female patients (a critical health equity step) while reducing MI overdiagnosis in the elderly, with the resultant harms that might follow, adverse psychosocial patient impact, and unnecessary healthcare expenditure from cascade testing.”