A new rhythm library for testing automatic external defibrillators: performance of three devices.
A library of arrhythmias obtained from patients with cardiac arrest was developed. Such a data base will permit both in vitro testing of the rhythm analysis system of automatic external defibrillators before clinical field trials are conducted and comparison of devices. Defibrillators equipped with voice/electrocardiographic tape recorders and used in the prehospital defibrillation programs in Iowa and King County, Washington provided the rhythm source. From these recordings, segments of ventricular fibrillation with minimal artifact and a duration of greater than or equal to 6 s were selected. Segments of ventricular fibrillation (n = 102) were categorized by average peak amplitude as fine (1 to less than 3 mm), medium (3 to less than 7 mm), coarse (7 to less than 12 mm) and extra coarse (greater than or equal to 12 mm), and transcribed onto high fidelity videocassette tapes. Nonventricular fibrillation cardiac arrest rhythms (n = 144), which included wide complex idioventricular rhythms, ventricular and supraventricular tachycardias, asystole and artifact, were also transcribed. Automatic external defibrillators developed by three manufacturers reached a treat (shock) decision on 88 to 93% of the ventricular fibrillation rhythms and on 5 to 10% of the nonventricular fibrillation rhythms. The latter decisions were defined as false positive, though for many rapid nonventricular fibrillation rhythms, countershock may be the appropriate treatment response. There were no statistically significant differences among the three devices in the shock/no shock decisions. A variety of ventricular fibrillation arrhythmias and terminology to express the preclinical performance of automatic external defibrillators are defined. Three commercially available automatic external defibrillators appear to successfully identify ventricular fibrillation and nonventricular fibrillation rhythms.(ABSTRACT TRUNCATED AT 250 WORDS)