In vitro evaluation of a sensor sensitive to acceleration forces included in a new rate modulated pacemaker

Scoring method for assessing rate adaptive pacemakers: application to two different activity sensors

To optimize programming of rate adaptive pacemakers (RAPs), we explored a new mathematical method to assess the performance of RAPs during daily-life tests, using customized Windows-based software. By stepwise discriminant analysis and linear regression, this method allows calculation of the acceleration and deceleration capacity of pacemakers and their general behavior during effort and recovery phases. Twenty-three patients (10 females and 13 males; 68 +/- 8 years) with chronic atrial fibrillation and a slow ventricular response were evaluated. They randomly received an accelerometer-controlled VVIR Dash Intermedics pacemaker (10 patients) or a vibration piezoelectric-controlled VVIR Sensolog III Siemens pacemaker (13 patients). All patients underwent the same test protocol: 6 minutes walking, 1.5 minutes climbing stairs, 1.5 minutes descending stairs, and 0.5 minutes sit-ups. By definition, the pacemaker responsiveness slope was programmed so that the heart rate response of paced patients during the walking test corresponded best to that of healthy controls. The slope was left unchanged for the other tests. We considered four scores: an acceleration score (EA score), an effort rate score (ER score), a deceleration score (RD score), and a recovery rate score (RR score). Scores ranged from -10 (hypochronotropic behavior of the pacemaker) to +10 (hyperchronotropic behavior), based on daily-life tests of 15 healthy controls (7 females and 8 males, 65 +/- 9 years). A score of 0 represented exact concordance with healthy controls. During stair descent, the Sensolog III produced excessive acceleration (EA score = +2.9 +/- 1.1) compared to: (1) stair climbing (EA score = -4.0 +/- 1.9; P = 0.01, with the same pacemakers); and (2) the Dash (+1.8 +/- 1.9; P = 0.04) and healthy controls (P = 0.02). The sit-up tests revealed a hypochronotropic response of both pacemakers compared to healthy controls, with a larger difference for the Sensolog III (EA score = -2.0 +/- 5.8; P = 0.04; RD score = -6.8 +/- 3.8' P = 0.02). We conclude that activity-driven pacemakers can accommodate brief activities, except for isovolumetric exercise such as sit-ups. During daily activities, accelerometer-driven pacemakers seem to provide a heart rate resoibse closer to that of healthy controls. Our new mathematical analysis is a simple and reproducible method for evaluating and quantifying the efficacy of any sensor-driven pacemaker.

Vibration, acceleration, gravitation, and movement: activity controlled rate adaptive pacing during treadmill exercise testing and daily life activities

Activity-based sensors for rate adaptive pacing have been available for several years and now include several different types: vibration; acceleration; gravitation; and movement. However, a systematic comparison evaluating the relative advantages and disadvantages of these various sensors has received little study. The purpose of the present study was to compare these sensor subtypes using treadmill testing and an outdoor test circuit, which simulated daily life activities and included both uphill and downhill walking. Pacemakers were strapped on the chest of healthy volunteers and connected to one channel of an ambulatory recording device, which also recorded the subject's intrinsic heart rate. The pacemakers were programmed using an initial treadmill test to standardize the rate responsive parameters for each device. Nine different pacemaker models were studied including 3 vibration-based (Elite, Synchrony, Metros), 4 acceleration-based (Relay, Excel, Ergos, Trilogy), 1 gravitational-based (Swing), and 1 movement-based (Sensorithm) device. All devices demonstrated a prompt rate response with casual walking on flat ground. The vibration-, gravitational-, and movement-based pacemakers showed a pronounced rate decline during more strenuous work, e.g., walking uphill. This phenomenon was absent in the accelerometer-based units. In particular, the vibration- and movement-based units showed a higher rate with walking downhill compared to uphill. An optimally tuned rate behavior on the treadmill usually did not provide an optimal rate behavior during daily activities and there was a tendency to overstimulation during low workload. The development of the two newest sensors (gravitational and movement) did not result in an improved performance of rate response behavior. Overall, the accelerometer-based pacemakers simulated or paralleled sinus rate behavior the most closely.

Stepwise analysis of the calibration procedure of an accelerometer-based pacemaker

The rate modulation of Relay 294-03 is individually adjusted during a 3-minute calibration test corresponding to the collection of moderate acceleration signals (MAL), which are related to a programmable moderate pacing rate (PR). Those rate/acceleration values form a calibration point in the flat part of the rate response curve (RRC). To help in a more clinically oriented analysis of this feature, the stimulator was first calibrated while strapped on a volunteer who remained seated (low MAL), walked (medium MAL), or jogged (high MAL). The device was then mechanically submitted to three calibrated to and fro movements corresponding to the sensitivity spectrum of the sensor. Each movement was repeated ten times to test the influence of the slope (1-10).

RESULTS

(1) One-way ANOVA significant effect of the MAL recorded during the calibration test on subsequent rate modulation (P = 0.0001); (2) The sensor amplifier gain was inversely related to the MAL: high MAL induced lower amplifier gain and lower PR for identical accelerations; (3) Lower amplifier gain allowed to discriminate highest acceleration magnitudes before reaching an overflow of the sensor; (4) The maximum achievable PR increased not only with the programmed slope, but also with decreasing MAL (P = 0.0055): a low MAL shifts to the left calibration point located on the part of the RRC and makes the last steep part of the RRC start earlier, thus leading to higher maximum achievable PR. In conclusion the calibration procedure is crucial not only in defining a moderate acceleration intensity but also in determining the sensor amplifier gain and the maximum achievable PR.