The aim of this study was to compare, both subjectively and objectively, four modern rate-responsive pacing modes in a double-blind crossover design. Twenty-two patients, aged 18 to 81 years, had an activity-sensing dual chamber universal rate-responsive (DDDR) pacemaker implanted for treatment of high grade atrioventricular block and chronotropic incompetence. They were randomly programmed to VVIR (ventricular demand rate-responsive), DDIR (dual chamber demand rate-responsive), DDD (dual chamber universal) or DDDR (dual chamber universal rate-responsive) mode and assessed after 4 weeks of out-of-hospital activity. Five patients, all with VVIR pacing, requested early reprogramming. The DDDR mode was preferred by 59% of patients; the VVIR mode was the least acceptable mode in 73%. Perceived "general well-being," exercise capacity, functional status and symptoms were significantly worse in the VVIR than in dual rate-responsive modes. Exercise treadmill time was longer in DDDR mode (p less than 0.01), but similar in all other modes. During standardized daily activities, heart rate in VVIR and DDIR modes underresponded to mental stress. All rate-augmented modes overresponded to staircase descent, whereas the DDD mode significantly underresponded to staircase ascent. Echocardiography revealed no difference in chamber dimensions, left ventricular fractional shortening or pulmonary artery pressure in any mode. Cardiac output was greater at rest in the dual modes than in the VVIR mode (p = 0.006) but was similar at 120 beats/min. Beat to beat variability of cardiac output was greatest in VVIR mode (p less than 0.0001), with DDIR showing greater variability than DDD or DDDR modes (p less than 0.05). Mitral regurgitation estimated by Doppler color flow imaging was similar in all modes, but tricuspid regurgitation was significantly greater in VVIR than in dual modes (p less than 0.03). Subjects who preferred the DDDR mode and those who found the VVIR mode least acceptable had significantly greater increases in stroke volume when paced in the DDD mode than in the ventricular-inhibited (VVI) mode at rest (22%) when compared with subjects who preferred other modes (2%, p = 0.03). No other objective variable was predictive of subjective benefit from any rate-responsive pacing mode. Thus, dual sensor rate-responsive pacing (DDDR) is superior objectively and subjectively to single sensor (VVIR, DDIR and DDD) pacing and subjective benefit from dual chamber rate-augmented pacing is predictable echocardiographically.