hs-cTnI Thresholds Reframed Admission Decisions in Non-AMI Chest Pain
Most emergency department chest pain visits do not end in acute myocardial infarction, yet disposition after rule-out still varies widely from one hospital to the next. In this analysis of high-sensitivity cardiac troponin I-detectable, non-AMI chest pain visits, a lower troponin-informed admission strategy preserved safety while trimming projected hospital use and cost. The central finding is not subtle: adding sub-99th percentile biomarker information to clinical risk stratification changed who was admitted, not who was missed.
Across 29 U.S. emergency departments, investigators examined 1,481 non-AMI chest pain visits from the prospective HIGH-US trial. Every patient had a detectable baseline hs-cTnI measured with the Atellica IM TnIH assay, which matters because detectable values below the traditional 99th percentile are often treated as a gray zone rather than as actionable biological information. The cohort was then used to compare observed standard-of-care admission patterns with modeled pathways that combined hs-cTnI thresholds and risk scores. The outcome of interest was not only diagnostic performance for 30-day death or myocardial infarction, but also the downstream cost consequences of different admission rules.
One sentence captures the clinical tension in this space.
Only 1.0% of the cohort, 15 of 1,481 patients, experienced 30-day death or MI.
That low event rate explains why emergency physicians have long struggled with balance: admitted too broadly, chest pain pathways consume beds, testing, and procedural capacity; discharged too aggressively, they risk rare but consequential missed ischemic events. Standard-of-care practice admitted 59% of patients and achieved 93% sensitivity with 41% specificity. In other words, the usual approach found nearly all patients who later died or had MI, but it did so at the price of many admissions among patients who did not need them.
The modeled strategy using hs-cTnI ≥ 5 ng/L plus a non-low risk score shifted that balance. Projected admissions fell to 41%, specificity rose to 60%, and sensitivity remained 93%. The practical meaning is straightforward: the troponin threshold helped identify a larger group of low-risk patients who could avoid hospitalization without reducing the model’s ability to capture the small number of 30-day hard events. In a chest pain unit, this type of change directly affects throughput, bed availability, and the burden of unnecessary downstream testing.
Economics Followed the Diagnostic Signal
The cost analysis sharpened the clinical argument. Under Scenario 1, which included noninvasive diagnostic testing only, estimated per-patient costs were $1,025 with the hs-cTnI-informed pathway versus $1,139 with standard care. Under Scenario 2, which also included inpatient and invasive procedures, costs were $2,672 versus $3,536. Those differences are large enough to matter in high-volume emergency settings, where even modest per-patient savings can scale quickly across thousands of chest pain evaluations.
The economic logic is rooted in pathophysiology and test interpretation. Detectable hs-cTnI below the traditional 99th percentile is not equivalent to normal biology; it often reflects subclinical myocardial injury, chronic structural heart disease, or vulnerability that may not justify admission on its own but does carry prognostic weight. Combining that signal with a clinical risk score helps separate patients whose biomarker elevation is likely to reflect benign chronic injury from those whose overall profile suggests a higher chance of near-term decompensation. That is why the model can reduce unnecessary admissions without eroding sensitivity for death or MI.
The subgroup findings pointed in the same direction. Higher hs-cTnI thresholds produced additional economic benefit while preserving sensitivity. Even without a granular list of every subgroup structure in the abstract, the direction of effect is clinically important: as the threshold rises within the detectable range, fewer patients are routed into costly inpatient pathways, yet the safety profile does not collapse. That pattern suggests the assay is doing more than merely classifying “positive” versus “negative” troponin values; it is helping define a gradient of risk that better matches real-world chest pain decision-making.
Why Detectable Troponin Below the 99th Percentile Still Matters
Chest pain triage has long been shaped by the fear of missing AMI, but the more common problem is overshooting the need for admission in patients who are not having an infarction. High-sensitivity assays made that tension more visible by detecting tiny amounts of myocardial injury that older assays would have missed. When those values are interpreted in isolation, they can trigger defensive admissions. When they are paired with a clinical score, they become part of a more calibrated decision framework.
This study is valuable because it evaluates the question from two angles at once: patient safety and resource stewardship. Maintaining 93% sensitivity while improving specificity from 41% to 60% is not a cosmetic adjustment; it changes the number of people who are exposed to the costs and risks of hospitalization, serial testing, and invasive evaluation. In emergency medicine, where chest pain remains one of the most common complaints, that kind of improvement could reshape workflow if validated prospectively.
Still, the findings remain modeled rather than prospectively implemented in routine care. The analysis relied on observed standard-of-care admissions and simulated pathways built around troponin thresholds and risk scores. Real-world performance could shift once clinicians use the strategy in practice, particularly if local patient mix, assay calibration, or risk-score selection differs from the trial setting. The study does not resolve how best to operationalize the threshold across diverse emergency departments, nor does it establish a single universal cutoff for all institutions or patient subgroups.
Even so, the signal is coherent: detectable hs-cTnI below the 99th percentile can inform a more efficient admission strategy when interpreted alongside clinical risk. In a population where only 1.0% experienced 30-day death or MI, the model preserved safety while reducing projected admissions and lowering estimated costs. The next step is not more speculation about whether troponin matters; it is prospective validation of how to use it without creating new blind spots at the bedside.