Classic cable theory was used to analyze the relation between the activation-recovery interval measured from unipolar electrograms and transmembrane action potential duration. Theoretic analysis demonstrated that the temporal derivative of the extracellular potential is proportional to a spatial weighting of the third temporal derivative of the transmembrane action potentials along a cable with uniform propagation in a homogeneous medium. Thus, the activation-recovery interval, measured as the interval between times of minimum derivative (V̇(min)) of the QRS and maximum derivative (V̇(max)) of the T wave, should be related to action potential duration, measured as the interval between times of V̇(max) of the upstroke and V̇(min) of the downstroke of the transmembrane action potential. This relation was examined experimentally in 12 anesthetized dogs. Unipolar electrograms and transmembrane action potentials were recorded simultaneously from sites within 2 mm of each other during control states, cardiac sympathetic nerve stimulation, localized epicardial warming, and graded reductions in myocardial perfusion. The heart was paced from several sites. There was close correlation between activation-recovery interval and action potential duration measurements taken during cardiac sympathetic nerve stimulation and local epicardial warming (r = 0.96 and 0.99 for cardiac sympathetic nerve stimulation and warming, respectively). In five animals in which coronary perfusion pressure was gradually lowered, the variables correlated closely over a range of values from 62 to 212 msec (r = 0.98). However, although the overall correlation was good and mean differences between activation-recovery interval and action potential duration were small, in individual cases there were differences up to 24 msec. This study provides new data that are consistent with the theoretic prediction that time of V̇(max) of the T wave is related to time of V̇(min) of the local action potential and that activation-recovery interval is directly related to action potential duration. In addition, the results provide evidence that the activation-recovery interval is a good estimate of the action potential duration independent of T waveform, drive site, or the presence of localized steep gradients of repolarization.
