Enhancing the diagnostic performance of troponins in the acute care setting

Abstract

Background: Current guidelines define cardiac troponin I (TnI) as an indicator of necrosis when the concentration exceeds the 99% upper limit of a healthy reference population, a reference value near the assay's lowest detectable level. We assessed the utility of a modified TnI cutoff point derived from a population at low risk for coronary artery disease (CAD) and evaluated its utility in determining acute myocardial infarction (MI).

Methods: A modified TnI cutoff point was derived by the receiver operating characteristic (ROC) curve from 737 consecutive patients who underwent serial TnI measurements for exclusion of MI. Creatinine kinase isoenzyme MB (CK-MB) evolutionary change was used to define MI. The new derived cutoff point was validated using another subset of 320 patients who were evaluated for MI.

Results: ROC-derived TnI cutoff point (A) was 0.65 μg/L, and its performance was compared to the recommended cutoff point ([B] 0.15 μg/L). Cutoff point A had greater specificity (94.5% vs. 86.9%, p < 0.001) but slightly lower sensitivity (96.5% vs. 100%, p < 0.01). Cutoff point A provided significantly greater positive predictive value (PPV) for MI (74.1% vs. 55.5%, p < 0.0001) and fewer false-positive errors, while preserving comparable negative predictive value (NPV) (98.9% vs. 100%).

Conclusion: A higher cutoff point derived from a reference population of patients at low risk for CAD may improve the TnI performance assay. The PPV for diagnosis of MI was significantly higher and false-positive values were fewer without affecting the NPV. The more reliable diagnosis of MI may have resulted, which, in turn, may have significant clinical and economic implications.