Rapid Rule-Out of Myocardial Infarction After 30 Minutes as an Alternative to 1 Hour: The RACING-MI Cohort Study.

ABSTRACT

STUDY OBJECTIVE: The aim of this study was to investigate whether myocardial infarction can be safely ruled in or out after 30 minutes as an alternative to 1 hour. METHODS: This was a prospective, single-center clinical study enrolling patients admitted to the emergency department. Patients with chest pain suggestive of myocardial infarction were eligible for inclusion. There was no walk-in to the emergency department, and patients with highly elevated out-of-hospital troponin were transferred directly to an invasive heart center. High-sensitivity troponin I was measured at admission (0 hour), 30 minutes, 1 hour, and 3 hours. Diagnostic performance was assessed using the sensitivity and negative predictive value (primary endpoints) as measures of ability to rule out myocardial infarction. Specificity and positive predictive value of myocardial infarction were used as measures for the ability to rule in myocardial infarction (secondary endpoints). RESULTS: In total, 1,003 patients qualified for analysis. Median age was 64 (interquartile range 52 to 74) years, and 42% were women. Myocardial infarction was confirmed in 9% of patients. In the validation cohort (n=503), the 0-h/30-min algorithm assigned 242 (48%) patients to rule out, 54 (11%) to rule in, and 207 (41%) to the observational zone. This resulted in a sensitivity of 100% (92.0% to 100%), negative predictive value of 100% (95% confidence interval 98.5% to 100%), specificity of 96.7% (94.7% to 98.2%), and positive predictive value of 72.2% (58.4% to 83.5%). In comparison, the 0-h/1-h algorithm performed with a sensitivity of 100% (92.0% to 100%), negative predictive value of 100% (98.5% to 100%), specificity of 97.2% (95.2% to 98.5%), and positive predictive value of 75.5% (61.7% to 86.2%). CONCLUSION: The accelerated 0-h/30-min algorithm allowed for safe rule-out of myocardial infarction 30 minutes after admission. The rule-in ability of the 0-h/30-min algorithm was comparable to that of the 0-h/1h algorithm.