Susceptibility of automated external defibrillators to train overhead lines and metro third rails

Introduction: Immediate accessibility to automated external defibrillators (AED) is recommended for highly frequented public areas. In train terminals and metro stations electromagnetic interference (EMI) is present. In preparation for a public access defibrillation (PAD) programme in this environment possible effects on AED safety and accuracy were studied. Methods: In typical public transportation settings 11 different AED models were bench tested for their sensitivity and specificity of ECG analysis with shockable and nonshockable rhythms provided by an ECG simulator. The devices were exposed to the electromagnetic interference of a rail system operating with 15 kV alternating current (ac) with a frequency of 16 2/3 Hz and a subway system powered with 750 V direct current (dc). AED cables were setup parallel and perpendicular to the tracks, the tests were carried out at 3 m distance from the rails in an empty station and with incoming trains. Results: A total of 5280 tests were recorded, each device was tested a total of 480 times. Fifteen kilovolts 16 2/3 Hz ac interfered more than 750 V dc with the tachyarrhythmia detection systems (P<0.0001). An AED setup with electrode cables perpendicular to track and power line reduced interference (P<0.0001), while incoming trains had no significant effect on ECG analysis (P=0.19). Depending on the AED model, sensitivity ranged from 60 to 100% and specificity from 54 to 100%, representing a positive likelihood-ratio from 1.3 to 241 and a negative likelihood-ratio from 0.7 to 0.0. In the public transportation setting tested, four AED models were unsuitable for automated defibrillation as these devices demonstrated an unacceptable performance in respect of accuracy and safety. In the train setting two devices performed with an accuracy of 57 and 65%. One AED recommended shocks for sinus rhythm at normal frequency. In the metro setting one AED did not advise shocks for ventricular tachycardia. Conclusion: Shock advisory systems of some AED models are susceptible to electromagnetic interference, especially in terminals with 15 kV 16 2/3 Hz ac power supplies. Interference is minimized, if patient position is parallel and electrode cables are perpendicular to overhead line. The choice of AED model for train or metro stations depends on its lack of susceptibility to typical electromagnetic interference.