Rate responsive pacing using transthoracic impedance minute ventilation sensors: A multicenter study on calibration stability

Previous studies showed that transthoracic impedance minute ventilation (IMV), as measured by a pacemaker sensor, is closely correlated to actual minute ventilation (V-E) determined by standard methods. The aim of this study was to analyze the changes in the calibration between IMV and V-E at rest and during exercise over time. Fifteen patients (age 60 +/- 13 years) with Medtronic Kappa 400 pacemakers completed a baseline visit followed by two visits separated by 1 (L/min) was monitored at rest in the supine and sit-month and 1 week, respectively. In each patient,sitting positions and during graded bicycle ergometer exercise using a standard cardiopulmonary metabolic gas analysis system with simultaneous recording of IMV (Omega/min) using DR-180 extended telemetry monitors. Calibration at rest was defined as the ratio of IMV to V,, calculated from 1-minute average values in the supine and sitting positions. Calibration during bicycle exercise was defined as intercept (IMV value V-E = 10 L/min-typical V-E, value at beginning of exercise), and slope of the IMV/V-E regression line. The at V calibration of IMV showed individual variability over time. The magnitude (absolute value) of observed fractional changes in calibration at 1 month was 0.23 +/- 0.20 (rest-supine), 0.20 +/- 0.15 (rest-sitting), 0.18 +/- 0.19 (exercise-intercept), 0.28 +/- 0.35 (exercise-slope), and 0.18 +/- 0.15, 0.15 +/- 0.09, 0.28 +/- 0.39, and 0.27 +/- 0.15, respectively, at 1 week. The magnitude of change at 1 month was not statistically different from the magnitude of change at 1 week. In conclusion, the calibration of IMV, as measured by a pacemaker sensor, versus actual V-E may demonstrate variability. However, this study also suggests that the observed changes are not cumulative over time. These results have implications for patient monitoring applications using these sensors and for development of future pacemaker rate response algorithms.