Background--The upper limit of vulnerability (ULV) correlates with the defibrillation threshold and can be determined with 1 episode of ventricular fibrillation (VF). To automate the ULV in an implantable cardioverter-defibrillator (ICD), the most vulnerable intervals must be identified from an ICD electrogram rather than the latest-peaking surface T wave (T_peak). We hypothesized that the recovery time (T_R), defined as the maximum derivative (dV/dt) of the T wave of the shock electrogram, correlates with the most vulnerable intervals.

Methods and Results--We determined ULV, defibrillation threshold, and the most vulnerable intervals in 25 patients at ICD implantation. The ULV was the weakest T-wave shock that did not induce VF. The most vulnerable intervals were the ones associated with the strongest shocks that induced VF. Telemetered shock electrograms were stored on digital tape and differentiated offline to measure T_R. T_peak and T_R were highly correlated (T_peak-T_R=-2±11 ms; =0.80, P<0.001). At least 1 most vulnerable interval timed between -20 ms and +20 ms relative to T_peak in all patients and between -40 ms and +20 ms relative to T_R in 96% of patients.

Conclusions--The recovery time of shock electrograms provides accurate information about global repolarization. T_R closely approximates T_peak. The ULV method may be automated in an ICD by timing T-wave shocks relative to T_R.