Background: The time required for patients with acute chest pain to reach a hospital emergency department varies, possibly lowering the diagnostic performance of a general cut-off value for high-sensitivity cardiac troponin T (hs-cTnT) in diagnosing acute myocardial infarction (AMI). The aim of this study was to determine the cut-off values as calculated by receiver operating characteristic (ROC) of hs-cTnT at different times from onset of symptoms to admission, and to evaluate their diagnostic performance.
Methods: Our study included 3096 patients with symptoms suggestive of AMI. These patients were classified according to time from onset of symptoms to admission. The diagnostic accuracy was quantified by the area under the ROC curve (AUC).
Results: Of the patients, 1082 (49.3%) were diagnosed as having AMI (317 were non-ST segment elevation myocardial infarction [NSTEMI]). The AUC for hs-cTnT to diagnose AMI was 0.881 at <3 h after symptom onset, 0.940 at 3-6 h after symptom onset, 0.966 at 6-12 h after symptom onset, and 0.974 at >12 h after symptom onset. The threshold as determined by ROC of hs-cTnT was 13.5 ng/L to diagnose AMI at ≤3 h after symptom onset with a sensitivity of 81.8% and a specificity of 80.1%, 17.8 ng/L at 3-6 h after symptom onset with a sensitivity of 94.6% and a specificity of 84.3%, 30.0 ng/L at 6-12 h after symptom onset with a sensitivity of 95.9% and a specificity of 85.5%, and 58.0 ng/L at >12 h after symptom onset with a sensitivity of 92.7% and a specificity of 93.3%. The same observations were performed for the diagnosis of NSTEMI.
Conclusions: The ROC-determined cut-off value of hs-cTnT for AMI or NSTEMI diagnosis gradually increased with time from onset of symptoms to presentation. Using a higher cut-off value by ROC for hs-cTnT will improve its accuracy in diagnosing AMI or NSTEMI patients in late presenters. The higher value will enable physicians to more quickly rule in patients compared to the 99th percentile cut-off, and can rule out patients safely.