Dependence of assessment of coronary artery reperfusion during acute myocardial infarction on angiographic criteria and interobserver variability.

The angiographic films of 240 patients with acute myocardial infarction were studied in a randomized trial of intravenous anisoylated plasminogen streptokinase activator complex (APSAC) versus intracoronary streptokinase therapies. The interobserver variability of grading coronary artery perfusion by the Thrombolysis in Myocardial Infarction Study Group (TIMI) criteria was measured as well as the effect of different definitions of reperfusion on the determination of reperfusion rate. There was good agreement in the reading of infarct artery flow grades between 2 blinded observers for each grade considered separately (k = 0.726 +/- 0.014) and for grades 0 or 1 (no perfusion) versus grades 2 or 3 (perfusion) (k = 0.905 +/- 0.011). Discordance between grades 0 or 1 versus 2 or 3 occurred in 74 (5%) of the 1,615 angiographic readings. Discrepancies of clinical significance which affected qualification for study entry, reperfusion or reocclusion status occurred in only 15 patients (6%). Grade 1 flow was found to have the most variable interpretation. Reperfusion rates for APSAC and streptokinase differed significantly when reperfusion was defined by 3 different criteria. The reperfusion rate ranged from 51 to 72% for APSAC and from 60 to 75% for streptokinase depending upon criteria selected. For comparison of the results of different thrombolytic studies, a standard semiquantitative system for grading infarct artery perfusion should be used, readings should be blinded and the criteria used for the definition of reperfusion should be clearly specified.