Abstract

The predictive value of the measurement of changes in ST segment elevation was assessed as a non-invasive marker of coronary artery reperfusion after thrombolytic treatment. Forty five patients with acute myocardial infarction (23 anterior, 22 inferior) of less than six hours' duration were given thrombolytic treatment by either the intravenous (n = 28) or the intracoronary route (n = 17). A proportional value for the shift in ST segment, termed the fractional change, was calculated both from 12 lead electrocardiograms and from the Holter tape for each patient. Coronary artery patency in an initial group of 22 patients (training group) was associated with a fractional change value of greater than or equal to 0.5 (100% specific, 88% sensitive by Holter analysis; 100% specific, 94% sensitive by 12 lead electrocardiogram). This rule performed well when it was applied to a test group of 17 patients (100% specific, 93% sensitive by Holter analysis; and 67% specific, 93% sensitive by 12 lead electrocardiogram). Linear discriminant analysis was then used to determine which features gave the best separation of those in whom there was reperfusion and those in whom there was not. This gave 100% specificity and 100% sensitivity when applied to the training group for either the 12 lead electrocardiogram or Holter monitoring. When it was applied to the test group, the sensitivity was maintained at 100%, but the specificity dropped to 33% irrespective of whether the basis of the test was Holter monitoring or the 12 lead electrocardiogram. These results suggest that a fractional change of >/= 0.5 calculated from a single lead showing myocardial injury is a useful non-invasive marker of reperfusion. The technique can be applied to either 12 lead electrocardiograms or Holter monitoring. The use of a more complex classification increased the sensitivity of the test at the expense of its specificity.