Using a high-sensitivity cardiac troponin 1 (hs-cTnI) assay to diagnose myocardial infarction (MI) may lead to overdiagnosis with resulting inappropriate therapies, researchers conclude from a new study.
A team of British investigators studied 20,000 consecutive inpatients and outpatients undergoing blood tests for any clinical reason at University Hospital Southampton and found that 1 in 20 had levels of troponin greater than the manufacturer’s recommended upper limit of normal (ULN).
Most patients, however, were being seen for noncardiac conditions and showed no clinical signs or symptoms of MI.
“The study showed that 39% of all patients from critical care units, 14% of all medical inpatients, and 6% of all patients from the ED [emergency department] had a troponin concentration greater than the level that indicates an MI,” lead researcher Nick Curzen, BM(Hons), PhD, professor of interventional cardiology/consultant cardiologist, University Hospital Southampton, United Kingdom, told theheart.org | Medscape Cardiology.
“The test is an excellent one for ruling out an MI in ED patients; however, these data suggest that it is not a good way of diagnosing acute MI in patients in the hospital, unless they have a classic MI presentation,” he said.
Overdiagnosis and Inappropriate Therapies
“The use of increasingly sensitive troponin assays for excluding or diagnosing acute MI has become universal” and the “gold standard biomarker,” the authors write.
Elevated cardiac troponin concentrations, especially in patients not presenting with a typical history of cardiac pain, may be caused by myocardial injury or type 2 MI, but “are not well recognized when the troponin test is requested or the result interpreted.”
Additionally, the authors continue, the 99th centile ULN is “generally applied as a binary ‘rule in’ or ‘rule out’ threshold for acute MI,” with an assumption that a concentration greater than the recommended threshold implies acute MI.
Misinterpretation of troponin results may therefore lead to “inappropriate management” and excessively aggressive therapy.
“I spent time increasingly frustrated that patients were being referred with ‘elevated’ hs-cTnI levels but in whom I felt the clinical suspicion of MI was low,” Curzen said.
“I therefore wanted to work out what was the actually distribution of the hs-cTnI test in a large population of hospital patients, regardless of underlying condition, in order to find out what the 99th centile was for the hospital group, compared to the manufacturer-supplied 99th centile ULN,” he said.
To investigate the question, the researchers conducted a prospective, observational study of 20,000 consecutive patients (≥18 years; median age [SD], 61 ; 52.9% female) at University Hospital Southampton between June 29 and August 24, 2017 who required a blood test for clinical reasons.
The Beckman Coulter Access AccuTnI+3 assay (Brea, California) is the test used in routine practice, the authors note.
Careful Interpretation Required
The 99th centile hs-cTnI concentration for the entire study population was 296 ng/L, with one in 20 patients (5.4%; n = 1080) having an hs-cTnI concentration above 40 ng/L, which is the manufacturer’s recommended ULN.
When the researchers excluded all patients diagnosed as having acute MI on discharge from the hospital or in whom an hs-cTnI test had been requested for clinical reasons, 18,171 patients remained.
In these patients, the 99th centile was 189 ng/L, with 4.6% of patients having a level above 40 ng/L.
The most common reasons for the test in the 1707 patients whose clinical team requested it were chest pain, arrhythmia, and suspected blackouts.
In the outpatients, the observed 99th centile was 65 ng/L, with hs-cTnI levels greater than the recommended ULN in 2%.
In the patients, the 99th centile was 563 ng/L, with hs-cTnI levels greater than the recommended ULN in 7.29%.
In patients whose blood was sampled in the ED, 27.2% had hs-cTnI concentrations requested by doctors in that department. The 99th centile for the remaining ED population (n = 3706), the 99th centile was 215 ng/L, with 6.07% having hs-cTnI greater than the recommended ULN.
In critical care patients, 39.02% had hs-cTnI levels greater than the recommended ULN.
When all patients diagnosed with MI or whose hs-cTnI levels greater than the recommended ULN was requested by the clinical team were excluded, 14.16% of all medical patients (excluding cardiac patients) had hs-cTnI levels greater than the recommended ULN.
In the “older people wards,” the acute surgical unit, and orthopedic wards, the percentage of patients whose hs-cTnI levels were greater than the recommended ULN were 20.8%, 4.62%, and 5.24%, respectively.
“In none of these patients was an acute MI suspected or diagnosed,” the authors emphasize.
An association was found between increasing age and distribution of troponin concentration.
Significant differences were also found in mean hs-cTnI levels in men, compared with women; a total of 6.6% vs 4.38% of women had had hs-cTnI levels greater than the recommended ULN. Moreover, there were significant differences in mean hs-cTnI levels (62 vs 31 ng/L; P = .021).
A multivariate analysis excluding those with a diagnosis of MI or clinical tests requested by the clinical team found advancing age, male sex, and decreasing estimated glomerular filtration rate to be independent predictors of having an hs-cTnI level greater than the recommended ULN.
“The notion of using a single binary value greater than the supplied ULN of any assay to diagnose whether a patient has an acute myocardial infarction is flawed,” the authors conclude.
“This study shows the need for medical staff to interpret troponin levels carefully in order to avoid misdiagnosis of an MI and inappropriate treatment,” Curzen said.
Curzen added that “the results can be used to stimulate debate about the way troponin measurements are requested and interpreted in the future, because it is certainly not quite right at the moment.”
Commenting on the study for theheart.org | Medscape Cardiology, Paul Collinson, MBBCHR, MD, Department of Clinical Blood Scientists, St. George’s University Hospitals, NHS Foundation Trust, London, called it “an interesting study, but definitely of the ‘may you live in interesting times’ variety.”
He expressed several concerns. “The assay used by the researchers was not, in fact, a high-sensitivity assay but a contemporary sensitive assay.”
In examining the 99th percentile of 65 ng/L, this is identical to a study that his group did looking at a randomly selected sample of general practitioner patients, he noted, “but when those patients with any underlying cardiac disease were excluded, the 99th percentile dropped to 40 ng/L.”
“What this clearly illustrates is that the 99th percentile of 40 is a sensible value for healthy individuals and any number above that indicates some degree of myocardial injury,” observed Collinson, who was not involved with the current study.
Furthermore, the researchers indeed found evidence of myocardial injury with elevated troponin in patients admitted to the hospital, and in the sickest of all patients — those in the intensive care unit — they found the largest number of elevated troponins, he pointed out.
“Troponin elevation in intensive care predicts an adverse outcome — detectable heart damage is always bad news,” he stated.
He agreed that the study “very nicely indicates that troponin measurements are grossly over-requested, wasting money and spreading diagnostic confusion,” but the “implication from the title and text is that perhaps we should use a much higher 99th percentile, which will mean missing patients with MI who can be treated and lives saved, as well as missing (even more) women with preventable coronary disease, who are currently underdiagnosed.”
He emphasized that the “advent of high-sensitivity troponin is, in fact, a great breakthrough, as it allows very early and very accurate rule-out of myocardial injury and acute MI— and, paradoxically, it reduces the number of analytical false-positive diagnoses around the diagnostic cutoff.”
Myocardial Injury vs MI
Also commenting on the study or theheart.org | Medscape Cardiology, James Januzzi Jr, MD, Hutter Professor of Medicine at Harvard Medical School, Cambridge, and staff cardiologist at Massachusetts General Hospital, Boston, said he has “seen some press suggesting that these results suggest troponin assays are somehow flawed.”
Rather than looking at these results as implying a flawed test, “I’d suggest the interpretation of the result of the study is flawed: troponin tests only identify myocardial injury and cannot tell us that an MI is present, as an MI is a clinical diagnosis that required clinical correlation to confirm its presence,” said Januzzi, who was not involved with the study.
“Stated another way, although myocardial injury is central to the diagnosis of MI, myocardial injury does not confirm its presence,” Januzzi said.
Christopher deFilippi, MD, vice chair of academic affairs, Inova Heart and Vascular Institute, Falls Church, Virginia, who also was not involved with the study, called it a “pragmatic article that will be helpful to clinicians.”
“Hospitalized patients not infrequently have cardiac troponin levels about the ULN because they have systemic disease or are receiving therapies that cause cardiac injury without ischemic; therefore, indications for testing, particularly with a high-sensitivity assay, need to be carefully considered,” he told theheart.org | Medscape Cardiology.
Moreover, the study “highlights the importance of the recognition of myocardial injury as separate from acute MI, as detailed in the Fourth Universal Definition of Myocardial Infarction.”
The first randomized controlled trial of high-sensitivity cardiac troponin I assay (ARCHITECTSTAT high-sensitive troponin I assay; Abbott Laboratories) in patients with suspected acute coronary syndrome showed the test increased detection of those with myocardial injury, but only a third of the patients reclassified had a diagnosis of MI and outcomes at 1 year did not improve.
Results of the High-STEACS trial were presented last year at the European Society of Cardiology (ESC) Congress 2018 and simultaneously published online in the Lancet.
Beckman Coulter ( Brea , California ) provided an unrestricted research grant for the study. Curzen reports receiving unrestricted research grants from Boston Scientific, Haemonetics, Heartflow, and Beckmann Coulter; speaker fees or consultancy fees from Haemonetics, Abbot Vascular, Heartflow, and Boston Scientific; and travel sponsorship from Biosensors, Abbot, Lilly/D-S, St. Jude Medical, and Medtronic. Januzzi is supported in part by the Hutter Family Professorship in Cardiology; has received grant support from Singulex, Abbott, and Prevencio; has received consulting income from Roche Diagnostics, Critical Diagnostics, Philips, Abbott, Prevencio, and Novartis; and participates in clinical end-point committees or data safety monitoring boards for Siemens, AbbVie, Pfizer, Amgen, Janssen, and Boehringer Ingelheim. Collinson reports no conflicting interests. deFilippi reports receiving research support from Roche Diagnostics and serving as a consultant for Abbot Diagnostics, Ortho Diagnostics, Roche Diagnostics, and Siemens.
BMJ. Published online March 13, 2019. Full text