A prospective study of an algorithm using cardiac troponin I and myoglobin as adjuncts in the diagnosis of acute myocardial infarction and intermediate coronary syndromes in a veteran's hospital.
BACKGROUND: Accurate and cost-effective evaluation of acute chest pain has been problematic for years. The high prevalence of missed myocardial infarctions (MI) has led to conservative triage behavior on the part of physicians, leading to expensive admissions to coronary care units. New algorithms are sorely needed for more rapid and accurate triage of patients with chest pain to appropriate treatment settings. HYPOTHESIS: We sought to test an algorithm for rapid diagnosis of MI and acute coronary syndromes using cardiac troponin I (cTnI) and myoglobin as adjuncts to creatine kinase (CK)-MB. We hypothesized our algorithm would be both sensitive and specific at early time points, and would allow safe stratification of patients not ruling in by conventional CK-MB criteria. METHODS: This was a 6-month prospective study of 505 consecutive patients who presented with chest pain at a university-affiliated veteran's hospital. The percentage of MIs at various time points was identified using combinations of markers. Safety outcomes were assessed by follow-up of patients discharged home. Cost savings analysis was assessed by surveying the physicians as to whether the use of the algorithm affected their disposition of patients. Forty-nine patients ruled in for MI. Using the combination of cTnI, 2-h doubling of myoglobin, and CK-MB, 37 (76%) ruled in at the time of presentation, 43 (88%) at 2 h, and 100% by 6 h. RESULTS: Cardiac troponin I plus a 2-h myoglobin was as accurate as the combination of all three markers and performed better than CK-MB in detecting patients presenting late and as a predictor for complications when CK-MB was normal. Of the 456 patients with normal markers after 6 h, only 140 were sent to the coronary care unit (CCU), and 176 were sent home. A 3-month follow-up showed minimal adverse events. One-half of physicians completing a survey stated the use of markers changed their disposition of patients, leading to an estimated 6-month cost savings of a half-million dollars. CONCLUSIONS: We developed an algorithm using troponin I and myoglobin as adjuncts to usual CK-MB levels that allowed for rapid and accurate assessment of patients with acute MI. It also afforded physicians important input into their decision making as to how best to triage patients presenting with chest pain. Their comfort in sending home certain subgroups of patients who otherwise would have been admitted to the CCU was rewarded with a good short-term prognosis and a large cost savings to the hospital.