Determination of pacing capture in implantable defibrillators: benefit of evoked response detection using RV coil to can vector.

Automatic detection of capture in ICDs would be useful for ensuring normal pacing function and lead integrity and may increase device longevity. Evoked response detection can be difficult due to postpace polarization. Polarization on the RV coil to can vector, however, should be absent when pacing with a true bipolar lead (pace tip to ring). Polarization on the RV coil to can vector should be low in an integrated bipolar lead due to the large surface area of the coil. Ventricular-paced responses were prospectively recorded in 20 patients during ICD implantation or replacement. Capture and loss of capture responses were noted during threshold searches with electrograms recorded between the RV coil and can. A detector was designed to discriminate between capture and noncapture-paced responses using data from the first 11 patients and validated on the remaining 9. The detector had a sensitivity of 99.9% (detected capture on capture beats), and had a specificity of 100% (detected no capture on noncapture beats) for all lead configurations. There was no measurable polarization with true bipolar leads. In integrated bipolar leads, maximum polarization ranged from 0.0 to 16.7 mV. In conclusion, paced evoked responses can be detected in ICDs using the RV coil to can vector using standard pacing waveforms. Special polarization reducing pacing waveforms are not required. These observations could be used to design ICDs with automatic pacing threshold detection.