Department of Cardiology, Imperial College Hospitals National Health Service Trust, London, United Kingdom; National Heart and Lung Institute, Imperial College London, London, United Kingdom. Electronic address: [Email]
OBJECTIVE : This study sought to test specialized processing of laser Doppler signals for discriminating ventricular fibrillation (VF) from common causes of inappropriate therapies. BACKGROUND : Inappropriate implantable cardioverter-defibrillator (ICD) therapies remain a clinically important problem associated with morbidity and mortality. Tissue perfusion biomarkers, implemented to assist automated diagnosis of VF, sometimes mistake artifacts and random noise for perfusion, which could lead to shocks being inappropriately withheld. METHODS : The study tested a novel processing algorithm that combines electrogram data and laser Doppler perfusion monitoring as a method for assessing circulatory status. Fifty patients undergoing VF induction during ICD implantation were recruited. Noninvasive laser Doppler and continuous electrograms were recorded during both sinus rhythm and VF. Two additional scenarios that might have led to inappropriate shocks were simulated for each patient: ventricular lead fracture and T-wave oversensing. The laser Doppler was analyzed using 3 methods for reducing noise: 1) running mean; 2) oscillatory height; and 3) a novel quantification of electromechanical coupling which gates laser Doppler relative to electrograms. In addition, the algorithm was tested during exercise-induced sinus tachycardia. RESULTS : Only the electromechanical coupling algorithm found a clear perfusion cut off between sinus rhythm and VF (sensitivity and specificity of 100%). Sensitivity and specificity remained at 100% during simulated lead fracture and electrogram oversensing. (Area under the curve running mean: 0.91; oscillatory height: 0.86; electromechanical coupling: 1.00). Sinus tachycardia did not cause false positive results. CONCLUSIONS : Quantifying the coupling between electrical and perfusion signals increases reliability of discrimination between VF and artifacts that ICDs may interpret as VF. Incorporating such methods into future ICDs may safely permit reductions of inappropriate shocks.