Study objective: To determine retrospectively the diagnostic accuracy of various ECG ST segment elevation criteria for the prehospital ECG diagnosis of acute myocardial infarction.
Design and setting: During a six-month period, paramedics acquired prehospital 12-lead ECGs on adult chest pain patients. Investigators interpreted ECGs independently, retrospectively, and blinded to patient outcome. ECGs were classified as meeting or not meeting the six ST segment elevation criteria regardless of ECG morphology if the criteria were present in two or more anatomically contiguous leads: 1 mm or more ST segment elevation; 2 mm or more ST segment elevation; 1 mm or more ST segment elevation in the limb leads or 2 mm or more ST segment elevation in the precordial leads; and the first three criteria with the simultaneous presence of reciprocal changes. ECGs that did not meet any ST segment elevation criteria were classified as normal, nonspecific ST/T wave changes, abnormal but not ischemic, and ischemic. Hospital charts were reviewed for final cardiac diagnosis.
Type of participant: Four hundred twenty-eight stable adult prehospital chest pain patients in whom paramedics acquired prehospital 12-lead ECGs.
Measurements and main results: Of the 428 cases, 123 (29%) met 1 mm or more ST segment elevation criteria. Sixty-three (51%) of these 123 patients did not have myocardial infarctions. ECG characteristics most frequently associated with these non-myocardial infarction ECGs were left bundle branch block (21%) and left ventricular hypertrophy (33%). The three criteria that required the presence of reciprocal changes had the highest positive predictive values (93% to 95%), with sensitivities ranging from 20% to 33%. The criteria of 1 mm or more ST segment elevation with the simultaneous presence of reciprocal changes had a positive predictive value of 94% and included 18 of the 21 (86%) myocardial infarction patients who had ST segment elevation and received thrombolytic therapy within five hours after hospital arrival. Of the 428 cases, 305 (71%) did not meet any ST segment elevation criteria and had a sensitivity of 81% and a negative predictive value of 49% for the absence of acute myocardial infarction.
Conclusion: Fifty-one percent of patients whose prehospital 12-lead ECG met 1 mm or more ST segment elevation criteria had non-myocardial infarction diagnoses. ST segment elevation alone lacks the positive predictive value necessary for reliable prehospital myocardial infarction diagnosis. Inclusion of reciprocal changes in prehospital ECG myocardial infarction criteria improved the positive predictive value to more than 90% and included a significant majority (62% to 86%) of acute myocardial infarction patients with ST segment elevation who received thrombolytic therapy within five hours after hospital arrival. ST segment elevation criteria that include reciprocal changes identify patients who stand to benefit most from early interventional strategies.