Exercise capacity and general well-being are improved by appropriately programmed rate responsive pacemakers when compared to fixed rate units. Ten patients had activity sensing DDDR units implanted for combined AV block and sinus node incompetence. Ten patients had Sensolog activity sensing VVIR units implanted for complete heart block. The effects of over and under programming of rate response in both dual and single chamber activity sensor rate adaptive pacemakers has been assessed subjectively by visual analog scales and specific activity questionnaires and objectively by graded treadmill testing and the performance of standardized daily activities. Patients were randomly programmed to absent rate response (VVI in the Sensolog group), hyporesponsive (DDD in the dual chamber group), appropriate response (VVIR, DDDR according to Manufacturer's instructions) and over responsive (VVIR+, DDDR+) in a double-blind crossover design. Thirty percent of patients demanded early crossover from VVI, 30% from DDDR+ and 50% from VVIR+. Perception of Exercise Capability was similar to objective exercise treadmill times which were shorter in VVI than in VVIR or VVIR+ (P less than 0.05) or control subjects (P less than 0.001). There was no difference between any dual chamber mode or control subjects. General well-being was poorest in DDDR+ and VVIR+ modes despite objective improvement in exercise capacity. Symptoms were least in VVIR and DDDR and all but one patient chose appropriate programming as their overall preferred mode. Thus, even inaccurate rate response programming results in similar and improved exercise capacity compared to absent rate response but overprogramming is unacceptable to most patients, confirming that appropriate programming and sensor specificity is critical in rate responsive pacing.