Background. In the canine model, an upper limit of shock strength exis
ts that can induce ventricular fibrillation during the vulnerable peri
od of the cardiac cycle. This shock strength (the upper limit of vulne
rability) closely correlates with the defibrillation threshold and sup
ports the ''upper limit of vulnerability'' hypothesis of defibrillatio
n. It is not known whether an upper limit of vulnerability exists in h
umans or whether this limit correlates with the defibrillation thresho
ld. Methods and Results. In 13 patients undergoing implantable cardiov
erter-defibrillator implantation, the shock strengths associated with
a 50% probability of reaching the upper limit of vulnerability (ULV50)
and a 50% probability of reaching the defibrillation threshold (DFT50
) were determined by the up-down algorithm. The ULV50 was determined o
nly for the mid-upslope of the positive T waves and for the mid-downsl
ope of the negative T waves. No major complications occurred during su
rgery. An upper limit of vulnerability was demonstrated in each patien
t. The ULV50 was 300+/-138 V or 6.8+/-5.8 J, which was significantly l
ower than the DFT50 of 347+/-167 V (p=0.038) or 9.1+/-7.3 J (p=0.013).
The correlation between the ULV50 and the DFT50 was significant (r=0.
90, p<0.001 for voltage; r=0.93, p<0.001 for energy). Conclusions. An
upper limit of vulnerability is present in humans. There is a signific
ant correlation between the ULV50 and the DFT50, and the ULV50 is sign
ificantly lower than the DFT50.