We have recently described the electrophysiological basis of a new algorithm for the QT (TX) sensing rate responsive pacemaker. By using the new software program running on the standard programmer it has been possible to simulate the new algorithm in ten patients with complete heart block (seven patients had implanted TX units and three were paced with an external TX pacemaker) during routine exercise testing. In this way a single-blind, intra-patient comparison of the pattern of pacing rote change using both the existing and new algorithms was possible. In nine out of the ten cases the time taken to increase the pacing rate from 70 to 80 bpm was reduced significantly when the new algorithm was used (P = 0.037). Additionally, the correlation between the atrial and ventricular rates in those patients with normal sinus node function (seven patients) was determined. In all cases we have observed a significantly improved correlation between the atrial and ventricular paced rates during exercise with the new algorithm (P < 0.001).