The clinical applicability of rate-responsive pacing (RRP) by means of activity sensing has been tested in 15 patients. The patients (ages 24-85) had sinus node dysfunction (2), atrial fibrillation (7), or sinus rhythm (6) combined with complete atrioventricular block. Exercise capacity was investigated on a bicycle ergometer and on a treadmill in a double-blind cross-over study design following one week each of fixed rate ventricular pacing (70 bpm) and rate-responsive pacing (60/125-150 bpm). The patients answered a questionnaire concerning subjective symptoms. A Holter ECG was recorded during 24 hours of all day activity on rate-responsive pacing. During exercise in the rate-responsive mode, heart rate increased more on the treadmill than on the bicycle. A majority of the patients (13 of 15) preferred rate-responsive pacing mainly due to less dyspnea and tiredness. Exercise capacity improved significantly both on bicycle (+7%; p less than 0.01) and on treadmill (+19%; p less than 0.01) during rate-responsive pacing. There were no complications during the follow-up period. In conclusion, the activity-sensing pacemaker is a valuable supplement to existing types of pacemakers. It should be used in patients in whom an atrial electrogram cannot be used for rate triggering.