Respiration rate (RR) and minute ventilation (MV) provide important clinical information on the state of the patient. This study evaluated the accuracy of determining these using a pacemaker impedance sensor. In 20 patients who were previously implanted with a Guidant PULSAR MAX group of pacemakers, the telemetered impedance sensor waveform was recorded simultaneously with direct volume respiration waveforms as measured by a pneumatometer. Patients underwent 30 minutes of breathing tests while supine and standing, and a 10-minute ergonometer bicycle exercise test at a workload of 50 W. Breathing tests included regular and rapid-shallow breathing sequences. RR was determined by a computerized algorithm, from impedance and respiration signals. The mean RR by impedance was 21.3 +/- 7.7 breaths/min, by direct volume was 21.1 +/- 7.6 breaths/min, range 7-66, the mean difference of RR measured by the impedance sensor, as compared with the true measurement, being 0.2 +/- 2.1 breaths/min. During the entire exercise, the mean correlation coefficient between impedance (iMV) and direct measured MV was 0.96 +/- 0.03, slope 0.13 +/- 0.05 L/Omega and range 0.07-0.26 L/Omega. Bland-Altman limits of agreement were +/- 4.6 L/min for MV versus iMV with each patient calibrated separately. The correlation coefficient for iMV versus MV over the entire 10 minutes of exercise, including the initial 4 minutes of exercise, was 0.99. The transthoracic impedance sensor of an implanted pacemaker can accurately detect respiration parameters. There was a large variation between subjects in the iMV versus MV slope during a bicycle exercise test, whereas for each subject, the slope was stable during submaximal bicycle exercise.