Activity-Based Pacing: Comparison of a Device Using an Accelerometer Versus a Piezoelectric Crystal DAVID W. BACHARACH,* TIMOTHY S. HILDEN,* JAY O. MILLERHAGEN,** BARBARA L. WESTRUM/* and JOHN M. KELLY* From the *Human Performance Laboratory. Department PERSS, St. Cloud State University, St. Cloud, and **Cardiac Pacemakers, Inc., St. Paul, Minnesota
BACHARACH, D.W., ET AL.: Activity-Based Pacing: A Comparison of a Device Using an Accelerometer Versus a Piezoelectric Crystal. The EXCEL'" VR, an accelevometer-based pacemaker (AC), and the Legend'", a pacemaker utilizing a piezoelectric crystal (PZ), were compared under ergometric conditions and during stair climbing to assess the appropriateness oftheirrate responses. The pacemakers, programmed to the manufacturers' nominai settings in order to compare different technologicaliy based sensors under identicaJ conditions, were strapped over subjects' left mid-pectoraJ region. Placement of the devices was randomized to control for positional effects. Ten healthy subjects (55-72 yearsj completed a graded exercise treadmiiJ test to 80% of maximum predicted heart rate (HH). An additionai group often subjects (50-66 years) completed exercise protocols involving bicycle ergometry and stair climbing. Throughout all tests, pacemaker pulse rates and subjects'intrinsic HR were monitored continuously. For the treadmill exercise, the average correlations between the AC and PZ pacemakers' pulse rate and HR for the group as a whole were r - 0.92 and r = 0.82, respectively. Individual subject comparisons were also made between each pacemaker rate and intrinsic HR. The mean difference from intrinsic rate was 11 ppm for the AC pacemaker and 24 ppm for the PZ pacemaker. In addition, the PZ pacemaker's maximal pulse rate was significantly lower (105 ± 9.6 ppm) than the other two rates (AC 137 ± 6 ppm; intrinsic HR 129 ± 2 beats/min). Throughout the bicycle ergometry testing, the intrinsic HR was higher than the AC and PZ pacing rates. However, the AC's rate was significantly higher than the PZ's rate. When subjects ascended stairs, the intrinsic HR and AC rate were closely correlated, but the PZ rate was significantly Jower. When subjects descended stairs, neither pacemaker's rate matched intrinsic HR. These results indicate that the AC pacemaker more closely matches intrinsic HR in healthy subjects during ergometric activity and stair climbing than does the PZ pacemaker. Although the three activities evaluated in this study may be lacking representation of other common daily tasks, these results do lend support for considering the use of an AC pacemaker. These data also suggest the need for further research using activity-based rate-responsive pacemakers after parameter optimization for each subject to address issues of sensitivity to and specificity of human movement. (PACE, Vol. 15, February 1992} adaptive-rate pacemakers, activity-controlled pacemakers
Introduction Many adaptive-rate pacemaker systems use a Supported by a grant from Cardiac Pacemakers, Inc. Address for reprints: Dr. David W. Bacharach. Human Performance Laboratory Halenbeck Hall Dept. PERSS St Cloud State University, St. Cloud, MN 56301. Fax: (612) 255-2099. Received April 8, 1991; revision September 3, 1991: revision November 7,1991; accepted November 19,1991. 188
piezoelectric Crystal (PZ) bonded tO the inside of the pacemaker can to sense body movements.^'^ j h i s mechanical sensor bases its rate response On . , , , , . , , r , counting a threshold signal generated trom the inVernal crystal.-' Although these pacemaker systems aie quick to respond, relatively easy to use, and
February 1992
PACE, Vol. 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
have unquestioned support for adaptive-rate pacing,^ some limitations do exist.^•*~^ The performance of these devices varies from patient to patient, and acceptable performance often requires fine tuning of parameter settings. Even with careful attention to parameter settings, the rates achieved are not always proportional to the intensity of exercise.^'^'^ In addition, inappropriate pacing can sometimes result from vibrations not associated with physical activity. This may occur because some pacemakers are most sensitive to frequencies in the range of 10-50 Hz,® whereas many physical movements only generate frequencies up to 6 Hz.^'^°-^^ Additionally, a PZ mounted to the canister is dependent upon distortion of the canister. With this in mind, the purpose of this study was to compare the performance of a typical pacemaker with a PZ bonded to the canister (Legend™, Medtronic, Inc., Minneapolis, MN, USA) and a pacemaker that uses an integrated circuit silicon accelerometer as a sensor for rate modulation (EXCEL'" VR, Cardiac Pacemakers, Inc., St. Paul, MN, USA). The accelerometer bases its rate response on the change in resistance generated by a silicon mass, completely contained within the pulse generator's circuitry, as the mass moves in concert with body motions and independent of the canister itself. The authors recognize that studies of this nature, by design, have some limitations. These limitations include externally mounted pacemakers that may respond differently than implanted pacemakers; the sample group studied may not closely represent the patient population who would receive these pacemakers; and that laboratory exercise tests may not simulate daily activity of pacemaker patients. With respect to these limitations, Mianulli et al.^^ and Benditt et al.^^ found strap-on rate-adaptive pacemakers are valid representations of implanted rate-adaptive pacemaker response when standardized techniques are used. When compared at rest and at peak, Mianulli et al.^^ found the mean difference in pacing rates were 6 pulses per minute (ppm) and 4 ppm, respectively. Although no study to date has reported valid data derived specifically from the PZ pacemaker at nominal settings used in this study, it was assumed that a strap-on system would provide valid
PACE, Vol. 15
data for comparison between devices using the same strap-on system. Subjects who volunteered for this study were age matched (65 ± 10 years) to typical pacemaker patients. We also included two standard lahoratory/clinical tests [treadmill and bicycle ergometry) along with a daily task (stair climbing) at two different rates, in an attempt to include an activity that might provide a different response to exercise than standard clinical tests. By incorporating these procedures, the most crucial limitations of this study were considered.
Methods The performance of the AC pacemaker and the PZ pacemaker was evaluated during graded treadmill exercise, bicycle ergometry, and stair walking at two different rates of ascent and descent. Both devices have parameter adjustments to account for speed of rate response, amount of response, and the rate of return to the lower rate limit. Since parameter adjustments to optimize adaptive rate pacemaker responses are not standardized, nominal values provided a method of comparing different technologically hased sensors under identical conditions. The nominal values for response time, threshold, and recovery period were considered to be similar across pacemakers. Upper rate limit and lower rate limit were programmed at 170 ppm and 60 ppm, respectively, for both devices. Treadmill Testing Ten asymptomatic, fit subjects (seven male, three female) from a cardiac rehabilitation/adult fitness program were evaluated during graded exercise treadmill testing using the Chronotropic Assessment Exercise Protocol (CAEP).^^ The mean age was 63 years, with a range of 55-72 years. The pacemakers (AC and PZ) were securely strapped to the subjects' left chest wall in the midpectoral region using a specialized chest harness. The order of placement within the chest harness was randomized to control for any possible positional effects. Subjects wore soft-soled tennis shoes when walking on the treadmill. They were instructed not to use the handrail for support during the test. Subjects rested 2-5 minutes prior to exercise to ensure pacemaker rates were within 5
February 1992
189
BACHARACH. ET AL.
ppm of their lower rate limit. They then performed a CAEP protocol to 80% of maximal predicted heart rate (HR) (220-age), followed by a recovery period of 5-10 minutes. Intrinsic HR and pacing rates from each pulse generator were averaged and recorded every 5 seconds by a multi-channel recording system (Labview'" II, National Instruments, Austin, TX, USA). Bicycle Ergometry Another group of ten fit subjects (ages 50-66) performed a two-part progressive exercise protocol on a bicycle ergometer. Subjects were fitted with a UNIQ™ {Computer Instrument Corp., Hempstead, NY, USA) CIC heart watch electrode chest strap and the two pacemakers. The pacemakers were held flat against the chest using Elasticon™ (Johnson and Johnson Co., Buffalo. NY. USA), once again alternating placement to control for positional effects. The three leads—AC pacmaker, PZ pacemaker, and intrinsic HR—were interfaced to separate digital displays for recording HR and pacing rates every 5 seconds. Part one of the bicycle test was intended to determine whether increased pedal frequency at a constant power output would influence HR and/ or pacing rates. Subjects sat quietly on the bike while the HR, pacing rates, and blood pressure were recorded. The test was initiated at a pedal frequency of 50 revolutions per minute [rpm] and constant power output of 25 watts (W). After a short period (— 1 min), exercise was terminated, and a recovery period was initiated to allow all rates to return to baseline values. Stage 2 was then performed at 60 rpm, 25 W, followed by a recovery period. Stage 3 was performed at 70 rpm, 25 W, followed by recovery. Part two of tbe bicycle test was a standard progressive protocol with workrates increasing 25 W every minute until 80% of predicted maximum HR was achieved. Subjects were asked to maintain 50 rpm to allow comparison of results with common clinical tests that use 50 rpm as a pedal frequency.^^ HR and pacemaker rates were displayed throughout the test and recorded every 5 seconds during the initial rest and subsequent exercise and recovery periods. Each test was terminated when the HR and pacing rates were below 100 beats/min
190
and within 10% of their lower rate limits, respectively. Stair Climbing The same group of subjects performing on the bicycle ergometer also completed a protocol involving stair climbing at a normal and fast pace. Wearing the same monitoring equipment, subjects first walked up and down stairs at 80 steps per minute (normal pace) using a metronome to keep the proper cadence. In the second phase, subjects performed the same protocol, except at 100 steps per minute. A rest period between ascent and descent and after each phase of the test allowed pacing rates to return to lower rate limits and HR to return to within 10% of baseline.
Results Treadmill Testing All subjects achieved 80% of maximum predicted HR. No complications were noted. Descriptive data, along with maximal HR and pulse rate (PR) results for the treadmill testing, are presented in Table I. The differences between mean maximal HR and mean maximal pacing rates were 8 ppm for the AC pacemaker and 24 ppm for the PZ pacemaker. These differences were tested using a one-way analysis of variance (ANOVA) and found to be statistically significant (F[2,18] = 9.61, P < 0.001). A post hoc analysis (Scheffe F-test, P < 0.05) identified significant differences between the maximal intrinsic HR (129 ± 8 beats/min) and the PZ pacing rate (105 ± 30 ppm) and between the AC pacing rate (137 ± 18 ppm) and the PZ pulse rate. No significant difference was found between maximal HR and the AC pacing rate. Individual comparisons of pulse generator rates versus subject's intrinsic rate are listed in Table II. The average individual correlation was r = 0.92 and r = 0.82, with differences between mean maximal HR and mean maximal pacing rates of 11 ppm and 24 ppm for the AC and the PZ, respectively. Correlations were made between each pulse generator's rate and the subjects' HR for the group as a whole (Fig. 1). The correlation between the AC pacing rate and the subject's HR was r ^ 0.80. The correlation between the PZ pacing rate and
February 1992
PACE, Vol, 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
Table 1 Descriptive Data for All Treadmill Subjects
Subject
Sex
Age
80% Max Predicted Heart Rate
Max Intrinsic Heart Rate
Max AC Pulse Rate
Max PZ Pulse Rate
1 2 3 4 5 6 7 8 9 10 Mean SD
M M M M M F M F F M
72 55 62 62 63 67 62 55 68 64 63 5
118 132 126 126 126 122 126 132 122 125 126 4
123 142 116 134 122 130 131 135 125 131 129 8
124 142 111 157 132 135 157 157 142 110 137 18
112 161 77 140
78 66 98 100 129
89 105 30
AC = accelerometer-based pacemaker; PZ = piezoelectric-based pacemaker.
the subject's HR was r = 0.27 under identical conditions. The ppm variations calculated by the root mean square (RMS), which is the absolute mean difference for each data point from the line of best fit, were 11 ppm for the AC device and 26 ppm for the PZ device. These pooled r values represent the range of responses that were observed with each
pacemaker. Individual correlations between intrinsic HR and the pacemaker rates indicate that responses from either pacemaker would be acceptable; however, the greater absolute mean difference for the PZ pacemaker suggests that the accelerometer driven pacemaker may provide better specificity to exercise.
y-.922x* 17.309. r2 - ,641
y - .45x * 44,055, r2 - .072 180
S
100
60
60
80
100 120 Subject HR
MO
160
40 40
180
60
80
100 120 140 160 180 Subject HR
Figure 1. Response of the accelerometer and piezoelectric devices versus the subject's intrinsic heart rate (HR) during treadmill exercise.
PACE, Vol. 15
February 1992
191
BACHARACH, ET AL.
Table M. Comparisons of Pulse Generator Rates Versus Intrinsic Rates During Treadmiii Exercise
Subject
1 2 3 4 5 6 7
8 9 10 Mean
AC Pacemaker
PZ Pacemaker
r
Mean x-y
r
Mean x-y
0.93 0.92 0.94 0.96 0.88 0,96 0.96 0.95 0.89 0,76 0.92
21 1 9 16 4 20 13 17 5
0.91 0.82 0.26 0.72 0.91 0,9 0.92 0,95 0.86
0.9
5 42 19 28 28 32 22 20 8 40
11
0.82
24
1
lute mean difference = 11 ppm), while tbe correlation between the PZ rate response and subjects' HR was r = 0.65 (absolute mean difference = 14 ppm). Bicycle Ergometry
During the progressive bicycle test, all subjects achieved 80% of the predicted maximum HR. Descriptive data and maximal HR and PR data for subjects taking part in the bicycle ergometry and stair climbing tests are shown in Table III. Mean values (± standard error) are shown in Figures 3 and 4 for the two phases of bicycle ergometry testing. A two-way repeated measure ANOVA was used to determine global differences for the progressive bicycle test. Figures 3 and 4 show that intrinsic HR was consistently higher than the pacing rate of either pacemaker throughout the bicycle testing protocol, at constant power output (Fig. 3) and during incremental testing (Fig.
Mean x-y data represents the deviation of the pacemaker rate from the intrinsic rate for each subject. AC = accelerometerbased pacemaker; PZ = piezoelectric-based pacemaker.
4).
Recovery data, being a function of constant decay from achieved peak values, reveal similar trends for each of the pacemakers (Fig. 2). For the group as a whole, the correlation between the AC rate response and subjects' HR was r = 0.87 (abso-
In addition, the AC pacemaker has a significantly higher pacing rate than the PZ pacemaker for all workrates during each phase of the cycling task. While cycling at 50, 60, and 70 rpm showed an increase in rates for all measures, the increase was only significant for the AC pacemaker, as shown in Figure 3.
y - .746x * 12.472, r2 -.425
y - 1.208X- 20.659. r2 • .762
60
BO
100 120 140 Subject HR
160
40 40
160
60
80
100 120 140 160 180 Subject HR
Figure 2. Recovery o/ the acceJerometer and piezoeJectric devices versus t\\e subject's intrinsic heart rate (HR) during treadmill exercise.
192
February 1992
PACE, Vol. 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
Table III. Descriptive Data for all Bicycle Ergometer and Stair-Climbing Subjects Sex
Age
80% Max Predicted Heart Rate
Max Intrinsic Heart Rate
Max AC Pulse Rate
Max PZ Pulse Rate
1 2 3 4 5 6 7 8 9 10 Mean SD
M M M M M F M F
55 55 61 52 61 59 50 60 50 51 56 4
132 132 127 134 127 127 136 128 136 136 131 4
135 132 128 140 128 148 134 140 144 137 137 6
83 78 83 99 95 78 99 99 78 73 87 10
60 65 60 85 68 92 78 85 60 80 71 12
Ll-
Subject
M
AC = accelerometer-based pacemaker; PZ = piezoelectric-based pacemaker.
Percent changes in HR and pacing rates were also calculated. Increases from rest to exercise at 25 W for intrinsic HR were 14%, 21%, and 23% at 50, 60, and 70 rpm, respectively. Increases in pacing rate for the AC pacemaker were 10%, 17%, and 28% at 50, 60, and 70 rpm, respectively. The PZ pacemaker, however, only increased 0.5%, 3%,
100
•t
I
* indicates a signllicani dirfersnc* between HR andfor PRt. t Indicates a signiiicani dllfeiance between rpm Irialt wllhin ea
E] SOipm O BOipm Q 70rpm
7S W RZ
12S W
150 W
B«eBV«i>
Workload
group.
Figure 3. Hate response to changes in rpm during bicycling at 25 watts. HR ^ heart rate; PR = pulse rate.
PACE, Vol. 15
100 W
Figure 4. Rate response fmean ± SE) during incremental bicycling. HR = heart rate; PR = pulse rate; AC = accelerometer-based pacemaker; PZ = piezoelectricbased pacemaker.
February 1992
193
BACHARACH, ET AL.
7y
¥
8C spm Acs.
U
BO>pm Desc.
m 10D spm Ac5. u 100 spm Dose.
Figure 5. Rate response (mean ± SEJacsending (AcsJ and descending fDescJ stmvs at 80 and 100 steps per minute. HR ^ heart rate; PR = puJse rate; AC ^ accelerometer-based pacemaker; PZ = piezoeJectric-based pacemaker.
balween Ihs vatious rate
and 4% at 50, 60, and 70 rpm, respectively. Similar results were seen during the incremental bicycle test. Mean percent increases from resting values to peak exercise were 75% for intrinsic HR, 41% for AC pacing rate, and 8% for PZ pacing rate. Stair Climbing and Descent
Mean values [ ± standard error) for the stair climbing protocol are shown in Figure 5, A oneway ANOVA (P < 0.05) was used to compare maximal HR and rates of the two pacemakers. When ascending stairs at 80 and 100 steps per minute, no significant differences between HR and the AC rate were identified; however, the PZ rate was significantly lower than the other two rates during each ascent. During descent, only the AC at 80 steps per minute was significantly different from the HR. The AC rate changes paralleled the HR rate changes to ascent and descent, respectively. However, the PZ device exhibited an increase on descent that is paradoxical when compared to HR for both step frequencies.
Discussion Recognizing the intent of this study was to evaluate the way in which rate modulation is accomplished by an AC sensor verses a PZ sensor bonded to the canister, the results indicate that at the manufacturer's nominal settings, the AC pacemaker provided a more proportional response to incremental treadmill activity than did the PZ
194
pacemaker. Considerable variations occurred between the intrinsic HR and the PZ paced rate for six of the ten subjects (Table III). These findings indicate that nominal values for the PZ pacemaker may not be appropriate for some subjects and that additional fine tuning would be required to attempt to achieve an acceptable rate response. The PZ pacemaker's inappropriately high rates near the onset of exercise, as shown in Figure 2, have been reported previously^^ and provide evidence against the notion that the activity threshold setting was too high. In addition, some subjects demonstrated blunted rate response, also indicated in Figure 2, which have been commonly observed by clinicians. The relationship of the AC pacemaker to intrinsic HR suggests that its nominal values may provide a better chronotropic response to treadmill exercise in a majority of subjects. During cycling, both pacemakers had difficulty keeping up with intrinsic HR. Little upper body motion is generated during cycling, which may explain the PZ pacemaker's muted response at nominal settings. Considering activity-based pacemakers may not be specific to all activities, the rate increase observed in the AC pacemaker during cycling would be advantageous provided rate specificity has not been compromised. A closer look at Figure 3 does reveal increasing rates with higher pedaling speeds for the two pacemakers as well as intrinsic HR. The AC pacemaker rate increased similarly to increases in HR; however, the standard error for the AC pacemaker was so small that these differences were statisti-
February 1992
PACE, Vol. 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
cally significant, while the changes in intrinsic HR as a function of rpm were not statistically significant. As indicated, percent increases for HR and the AC pacemaker were similar, while the PZ pacemaker showed virtually no change. This indicates the AC pacemaker responded more appropriately. The mean resting HR was 76 ± 4 beats/min as compared to mean pacing rates of 66 ± 0.3 ppm (AC) and 63 ± 0.5 ppm (PZ), making percent increases in rates biased toward the activity-based pacemakers. The PZ pacemaker's paradoxical rate increase during descent of stairs was noted. However, results of the AC pacemaker for ascending and descending stairs at 80 and 100 steps per minute are in the same direction, and similar to previously reported data.**'^ Adjustments in parameter settings of the AC and PZ pacemakers can improve rate modulation during exercise for those subjects demonstrating insufficient activity or an inordinate amount of disturbance to generate the expected response. It has been suggested that activity-based rate responsive pacemakers such as the PZ pacemaker provide a less than satisfactory rate response because of the frequency range to which they are most sensitive. Alt et al.^ stated, "A maximum sensitivity in the area of 10 Hz has disadvantageous consequences to an activity-controlled system, since interference influences have their amplitude peak in this frequency range. It is therefore impossible to discriminate between exercise induced signals and signals which should not lead to a frequency response." (p. 1676) It is the opinion of these authors that there may be several reasons for the AC pacemaker's greater sensitivity and specificity to body motion. In observing the characteristics of the two pacemakers, the PZ device, with an optimal frequency range from 10-50 Hz, may be susceptible to nonspecific pacing responses or extraneous pressure applied to the canister since the piezoelectric crystal is dependent upon distortion of the crystal bonded to the pacemaker canister. The AC device, with an optimal frequency range of 1-8 Hz, has the sensor mounted within the pacemaker yet independent of the canister. Therefore, the AC device's specificity to activity should not be compromised with the improved proportionality and sensitivity. The scope of this study
PACE, Vol. 15
did not, however, incorporate additional tests to address the issue of specificity and it is as yet undetermined as to whether the results achieved here are attributable to differences in frequency ranges of the two pacemakers. Certainly, these data support rationale for further investigations using an AC pacemaker and a PZ pacemaker optimized for each individual while executing a variety of daily tasks attempting to answer the specificity question for human movement. A review of the recovery data demonstrates the importance of a correct response parameter setting. At nominal settings, the AC pacemaker provided an appropriate rate during recovery; however, the variability of the results achieved by the PZ pacemaker make it difficult to evaluate its recovery data. The recovery algorithms in each device are related to time rather than to amount of activity measured by tbe sensor. Recovery is simply dependent on the difference between the present rate, the lower rate limit, and the programmed recovery parameter. If an incorrect response setting causes an inappropriate present rate, the return to the lower rate limit will also be altered.
Conclusions As suggested by Sulke et al.,^ accurate response programming is essential for optimal performance of variable-rate pacemakers. The AC and PZ pacemakers can provide rate modulation for individual subjects at nominal settings; however, this study showed the AC pacemaker's rate is more closely associated to intrinsic HR during treadmill walking, bicycle ergometry, and ascending stairs than a PZ pacemaker's rate. The AC pacemaker set at nominal values is better correlated to an individual's intrinsic HR response (absolute mean difference = 11 ppm) than was the PZ pacemaker's rate when set at matching nominal values (absolute mean difference = 26 ppm] during treadmill exercise. Although neither pacemaker performed well enough to closely match intrinsic HR during cycling, the AC pacemaker appears to be more sensitive to increases in workrate than the PZ pacemaker during progressive bicycle ergometry. Finally, the AC pacemaker can perform better than the PZ pacemaker when ascending stairs at 80 or 100 steps per minute; however, both pacemakers
February 1992
195
BACHARACH, ET AL.
were higher than intrinsic HR while descending stairs at 80 or 100 steps per minute. Overail, these data favor the use of an AC pacemaker to provide a more appropriate rate response than a PZ paceReferences
maker in which the piezoelectric crystal is bonded to the canister to incremental treadmill, bicycle exercise, and stair climbing when response parameter settings are standardized.
1. Humen DP, Kostuk TWJ, Klein GJ. Activity sensing, rate-responsive pacing: Improvement in myocardial performance with exercise. PACE 1985; 8:52-59. 2. Lindemans FW, Rankin JR, Murtasugh R, et al. Clinical experience with an activity sensing pacemaker. PACE 1986; 9:978-986. 3. Furman S, Hayes DL, Holmes DR. A Practice of Cardiac Pacing. 2nd ed. Mount Kisco, New York, Futura Publishing Company, Inc., 1989, pp. 414-428. 4. Miura DS. A new concept in pacemaker therapy: Rate responsive pacing. Cardiol Rev Rep 1987; 8:40-44. 5. Suike N, Dritsa A, Chambers J, et al. Is accurate rate response programming necessary? PACE 1990; 3:1031-1044, 6. Sulke N, Pipilis A, Henderson RA, et al. Comparison of normal sinus node with seven types of rate responsive pacemakers during everyday activity. Br Heart J 1990; 4:25-31. 7. Lau C-P, Mehta D, Toff WD, et al. Limitations of rate-response of an activity-sensing rate-responsive pacemaker to different forms of activity. PACE 1988; 11:141-150. 8. Alt E, Matula M, Theres H, et al. The basis for activity controlled rate variable cardiac pacemakers: An analysis of mechanical forces on the human body induced by exercise and environment. PACE 1989; 12:1667-1680.
Lau C-P, Butrous GS, Ward DE, et al, Comparison of exercise performance of six rate-adaptive right ventricular cardiac pacemakers. Am J Cardiol 1989; 63:833-838. Bunge T, Thompson D. Sensing internal and external body activities. In F Perez Gomez (ed.): Cardiac Pacing: Electrophysiology: Tachyarrythmias. Madrid, Spain, Editorial Grouz, 1985, pp, 786-791, Lau C-P, Stott JRR, Toff WD, et al. Selective vibration sensing: A new concept for activitysensing rate-responsive pacing, PACE 1988; 11:1299-1309. MianuUi M, Benditt DC, Markowitz T. et al. A comparison of strap-on versus implanted activitybased rate-responsive pacemakers: Are strap-on studies valid? (abstract) PACE 1991; 14:742. Benditt DG, MianuUi M, Fetter J, et al. An officebased exercise protocol for predicting chronotropic response of activity-triggered rate-variable pacemakers. Am ] Cardiol 1989; 64:27-32. Wilkoff BL, Corey J, Blackburn G. A mathematical model of the cardiac chronotropic response to exercise. J Electrophysiol 1989; 3:176-180. Astrand PO, Rodahl K. Textbook of Work Physiology: Physiological Bases of Exercise. New York, NY, McGraw-Hill Book Company, 1977, pp. 369-371.
196
10.
11.
12.
13.
14.
15.
16.
February 1992
Mond H, Line 1, Hunt D. A third generation activity pacemaker: Is the rate response superior? (abstract) PAGE 1990; 13:514.
PACE, Vol. 15
BACHARACH, D.W., ET AL.: Activity-Based Pacing: A Comparison of a Device Using an Accelerometer Versus a Piezoelectric Crystal. The EXCEL'" VR, an accelevometer-based pacemaker (AC), and the Legend'", a pacemaker utilizing a piezoelectric crystal (PZ), were compared under ergometric conditions and during stair climbing to assess the appropriateness oftheirrate responses. The pacemakers, programmed to the manufacturers' nominai settings in order to compare different technologicaliy based sensors under identicaJ conditions, were strapped over subjects' left mid-pectoraJ region. Placement of the devices was randomized to control for positional effects. Ten healthy subjects (55-72 yearsj completed a graded exercise treadmiiJ test to 80% of maximum predicted heart rate (HH). An additionai group often subjects (50-66 years) completed exercise protocols involving bicycle ergometry and stair climbing. Throughout all tests, pacemaker pulse rates and subjects'intrinsic HR were monitored continuously. For the treadmill exercise, the average correlations between the AC and PZ pacemakers' pulse rate and HR for the group as a whole were r - 0.92 and r = 0.82, respectively. Individual subject comparisons were also made between each pacemaker rate and intrinsic HR. The mean difference from intrinsic rate was 11 ppm for the AC pacemaker and 24 ppm for the PZ pacemaker. In addition, the PZ pacemaker's maximal pulse rate was significantly lower (105 ± 9.6 ppm) than the other two rates (AC 137 ± 6 ppm; intrinsic HR 129 ± 2 beats/min). Throughout the bicycle ergometry testing, the intrinsic HR was higher than the AC and PZ pacing rates. However, the AC's rate was significantly higher than the PZ's rate. When subjects ascended stairs, the intrinsic HR and AC rate were closely correlated, but the PZ rate was significantly Jower. When subjects descended stairs, neither pacemaker's rate matched intrinsic HR. These results indicate that the AC pacemaker more closely matches intrinsic HR in healthy subjects during ergometric activity and stair climbing than does the PZ pacemaker. Although the three activities evaluated in this study may be lacking representation of other common daily tasks, these results do lend support for considering the use of an AC pacemaker. These data also suggest the need for further research using activity-based rate-responsive pacemakers after parameter optimization for each subject to address issues of sensitivity to and specificity of human movement. (PACE, Vol. 15, February 1992} adaptive-rate pacemakers, activity-controlled pacemakers
Introduction Many adaptive-rate pacemaker systems use a Supported by a grant from Cardiac Pacemakers, Inc. Address for reprints: Dr. David W. Bacharach. Human Performance Laboratory Halenbeck Hall Dept. PERSS St Cloud State University, St. Cloud, MN 56301. Fax: (612) 255-2099. Received April 8, 1991; revision September 3, 1991: revision November 7,1991; accepted November 19,1991. 188
piezoelectric Crystal (PZ) bonded tO the inside of the pacemaker can to sense body movements.^'^ j h i s mechanical sensor bases its rate response On . , , , , . , , r , counting a threshold signal generated trom the inVernal crystal.-' Although these pacemaker systems aie quick to respond, relatively easy to use, and
February 1992
PACE, Vol. 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
have unquestioned support for adaptive-rate pacing,^ some limitations do exist.^•*~^ The performance of these devices varies from patient to patient, and acceptable performance often requires fine tuning of parameter settings. Even with careful attention to parameter settings, the rates achieved are not always proportional to the intensity of exercise.^'^'^ In addition, inappropriate pacing can sometimes result from vibrations not associated with physical activity. This may occur because some pacemakers are most sensitive to frequencies in the range of 10-50 Hz,® whereas many physical movements only generate frequencies up to 6 Hz.^'^°-^^ Additionally, a PZ mounted to the canister is dependent upon distortion of the canister. With this in mind, the purpose of this study was to compare the performance of a typical pacemaker with a PZ bonded to the canister (Legend™, Medtronic, Inc., Minneapolis, MN, USA) and a pacemaker that uses an integrated circuit silicon accelerometer as a sensor for rate modulation (EXCEL'" VR, Cardiac Pacemakers, Inc., St. Paul, MN, USA). The accelerometer bases its rate response on the change in resistance generated by a silicon mass, completely contained within the pulse generator's circuitry, as the mass moves in concert with body motions and independent of the canister itself. The authors recognize that studies of this nature, by design, have some limitations. These limitations include externally mounted pacemakers that may respond differently than implanted pacemakers; the sample group studied may not closely represent the patient population who would receive these pacemakers; and that laboratory exercise tests may not simulate daily activity of pacemaker patients. With respect to these limitations, Mianulli et al.^^ and Benditt et al.^^ found strap-on rate-adaptive pacemakers are valid representations of implanted rate-adaptive pacemaker response when standardized techniques are used. When compared at rest and at peak, Mianulli et al.^^ found the mean difference in pacing rates were 6 pulses per minute (ppm) and 4 ppm, respectively. Although no study to date has reported valid data derived specifically from the PZ pacemaker at nominal settings used in this study, it was assumed that a strap-on system would provide valid
PACE, Vol. 15
data for comparison between devices using the same strap-on system. Subjects who volunteered for this study were age matched (65 ± 10 years) to typical pacemaker patients. We also included two standard lahoratory/clinical tests [treadmill and bicycle ergometry) along with a daily task (stair climbing) at two different rates, in an attempt to include an activity that might provide a different response to exercise than standard clinical tests. By incorporating these procedures, the most crucial limitations of this study were considered.
Methods The performance of the AC pacemaker and the PZ pacemaker was evaluated during graded treadmill exercise, bicycle ergometry, and stair walking at two different rates of ascent and descent. Both devices have parameter adjustments to account for speed of rate response, amount of response, and the rate of return to the lower rate limit. Since parameter adjustments to optimize adaptive rate pacemaker responses are not standardized, nominal values provided a method of comparing different technologically hased sensors under identical conditions. The nominal values for response time, threshold, and recovery period were considered to be similar across pacemakers. Upper rate limit and lower rate limit were programmed at 170 ppm and 60 ppm, respectively, for both devices. Treadmill Testing Ten asymptomatic, fit subjects (seven male, three female) from a cardiac rehabilitation/adult fitness program were evaluated during graded exercise treadmill testing using the Chronotropic Assessment Exercise Protocol (CAEP).^^ The mean age was 63 years, with a range of 55-72 years. The pacemakers (AC and PZ) were securely strapped to the subjects' left chest wall in the midpectoral region using a specialized chest harness. The order of placement within the chest harness was randomized to control for any possible positional effects. Subjects wore soft-soled tennis shoes when walking on the treadmill. They were instructed not to use the handrail for support during the test. Subjects rested 2-5 minutes prior to exercise to ensure pacemaker rates were within 5
February 1992
189
BACHARACH. ET AL.
ppm of their lower rate limit. They then performed a CAEP protocol to 80% of maximal predicted heart rate (HR) (220-age), followed by a recovery period of 5-10 minutes. Intrinsic HR and pacing rates from each pulse generator were averaged and recorded every 5 seconds by a multi-channel recording system (Labview'" II, National Instruments, Austin, TX, USA). Bicycle Ergometry Another group of ten fit subjects (ages 50-66) performed a two-part progressive exercise protocol on a bicycle ergometer. Subjects were fitted with a UNIQ™ {Computer Instrument Corp., Hempstead, NY, USA) CIC heart watch electrode chest strap and the two pacemakers. The pacemakers were held flat against the chest using Elasticon™ (Johnson and Johnson Co., Buffalo. NY. USA), once again alternating placement to control for positional effects. The three leads—AC pacmaker, PZ pacemaker, and intrinsic HR—were interfaced to separate digital displays for recording HR and pacing rates every 5 seconds. Part one of the bicycle test was intended to determine whether increased pedal frequency at a constant power output would influence HR and/ or pacing rates. Subjects sat quietly on the bike while the HR, pacing rates, and blood pressure were recorded. The test was initiated at a pedal frequency of 50 revolutions per minute [rpm] and constant power output of 25 watts (W). After a short period (— 1 min), exercise was terminated, and a recovery period was initiated to allow all rates to return to baseline values. Stage 2 was then performed at 60 rpm, 25 W, followed by a recovery period. Stage 3 was performed at 70 rpm, 25 W, followed by recovery. Part two of tbe bicycle test was a standard progressive protocol with workrates increasing 25 W every minute until 80% of predicted maximum HR was achieved. Subjects were asked to maintain 50 rpm to allow comparison of results with common clinical tests that use 50 rpm as a pedal frequency.^^ HR and pacemaker rates were displayed throughout the test and recorded every 5 seconds during the initial rest and subsequent exercise and recovery periods. Each test was terminated when the HR and pacing rates were below 100 beats/min
190
and within 10% of their lower rate limits, respectively. Stair Climbing The same group of subjects performing on the bicycle ergometer also completed a protocol involving stair climbing at a normal and fast pace. Wearing the same monitoring equipment, subjects first walked up and down stairs at 80 steps per minute (normal pace) using a metronome to keep the proper cadence. In the second phase, subjects performed the same protocol, except at 100 steps per minute. A rest period between ascent and descent and after each phase of the test allowed pacing rates to return to lower rate limits and HR to return to within 10% of baseline.
Results Treadmill Testing All subjects achieved 80% of maximum predicted HR. No complications were noted. Descriptive data, along with maximal HR and pulse rate (PR) results for the treadmill testing, are presented in Table I. The differences between mean maximal HR and mean maximal pacing rates were 8 ppm for the AC pacemaker and 24 ppm for the PZ pacemaker. These differences were tested using a one-way analysis of variance (ANOVA) and found to be statistically significant (F[2,18] = 9.61, P < 0.001). A post hoc analysis (Scheffe F-test, P < 0.05) identified significant differences between the maximal intrinsic HR (129 ± 8 beats/min) and the PZ pacing rate (105 ± 30 ppm) and between the AC pacing rate (137 ± 18 ppm) and the PZ pulse rate. No significant difference was found between maximal HR and the AC pacing rate. Individual comparisons of pulse generator rates versus subject's intrinsic rate are listed in Table II. The average individual correlation was r = 0.92 and r = 0.82, with differences between mean maximal HR and mean maximal pacing rates of 11 ppm and 24 ppm for the AC and the PZ, respectively. Correlations were made between each pulse generator's rate and the subjects' HR for the group as a whole (Fig. 1). The correlation between the AC pacing rate and the subject's HR was r ^ 0.80. The correlation between the PZ pacing rate and
February 1992
PACE, Vol, 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
Table 1 Descriptive Data for All Treadmill Subjects
Subject
Sex
Age
80% Max Predicted Heart Rate
Max Intrinsic Heart Rate
Max AC Pulse Rate
Max PZ Pulse Rate
1 2 3 4 5 6 7 8 9 10 Mean SD
M M M M M F M F F M
72 55 62 62 63 67 62 55 68 64 63 5
118 132 126 126 126 122 126 132 122 125 126 4
123 142 116 134 122 130 131 135 125 131 129 8
124 142 111 157 132 135 157 157 142 110 137 18
112 161 77 140
78 66 98 100 129
89 105 30
AC = accelerometer-based pacemaker; PZ = piezoelectric-based pacemaker.
the subject's HR was r = 0.27 under identical conditions. The ppm variations calculated by the root mean square (RMS), which is the absolute mean difference for each data point from the line of best fit, were 11 ppm for the AC device and 26 ppm for the PZ device. These pooled r values represent the range of responses that were observed with each
pacemaker. Individual correlations between intrinsic HR and the pacemaker rates indicate that responses from either pacemaker would be acceptable; however, the greater absolute mean difference for the PZ pacemaker suggests that the accelerometer driven pacemaker may provide better specificity to exercise.
y-.922x* 17.309. r2 - ,641
y - .45x * 44,055, r2 - .072 180
S
100
60
60
80
100 120 Subject HR
MO
160
40 40
180
60
80
100 120 140 160 180 Subject HR
Figure 1. Response of the accelerometer and piezoelectric devices versus the subject's intrinsic heart rate (HR) during treadmill exercise.
PACE, Vol. 15
February 1992
191
BACHARACH, ET AL.
Table M. Comparisons of Pulse Generator Rates Versus Intrinsic Rates During Treadmiii Exercise
Subject
1 2 3 4 5 6 7
8 9 10 Mean
AC Pacemaker
PZ Pacemaker
r
Mean x-y
r
Mean x-y
0.93 0.92 0.94 0.96 0.88 0,96 0.96 0.95 0.89 0,76 0.92
21 1 9 16 4 20 13 17 5
0.91 0.82 0.26 0.72 0.91 0,9 0.92 0,95 0.86
0.9
5 42 19 28 28 32 22 20 8 40
11
0.82
24
1
lute mean difference = 11 ppm), while tbe correlation between the PZ rate response and subjects' HR was r = 0.65 (absolute mean difference = 14 ppm). Bicycle Ergometry
During the progressive bicycle test, all subjects achieved 80% of the predicted maximum HR. Descriptive data and maximal HR and PR data for subjects taking part in the bicycle ergometry and stair climbing tests are shown in Table III. Mean values (± standard error) are shown in Figures 3 and 4 for the two phases of bicycle ergometry testing. A two-way repeated measure ANOVA was used to determine global differences for the progressive bicycle test. Figures 3 and 4 show that intrinsic HR was consistently higher than the pacing rate of either pacemaker throughout the bicycle testing protocol, at constant power output (Fig. 3) and during incremental testing (Fig.
Mean x-y data represents the deviation of the pacemaker rate from the intrinsic rate for each subject. AC = accelerometerbased pacemaker; PZ = piezoelectric-based pacemaker.
4).
Recovery data, being a function of constant decay from achieved peak values, reveal similar trends for each of the pacemakers (Fig. 2). For the group as a whole, the correlation between the AC rate response and subjects' HR was r = 0.87 (abso-
In addition, the AC pacemaker has a significantly higher pacing rate than the PZ pacemaker for all workrates during each phase of the cycling task. While cycling at 50, 60, and 70 rpm showed an increase in rates for all measures, the increase was only significant for the AC pacemaker, as shown in Figure 3.
y - .746x * 12.472, r2 -.425
y - 1.208X- 20.659. r2 • .762
60
BO
100 120 140 Subject HR
160
40 40
160
60
80
100 120 140 160 180 Subject HR
Figure 2. Recovery o/ the acceJerometer and piezoeJectric devices versus t\\e subject's intrinsic heart rate (HR) during treadmill exercise.
192
February 1992
PACE, Vol. 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
Table III. Descriptive Data for all Bicycle Ergometer and Stair-Climbing Subjects Sex
Age
80% Max Predicted Heart Rate
Max Intrinsic Heart Rate
Max AC Pulse Rate
Max PZ Pulse Rate
1 2 3 4 5 6 7 8 9 10 Mean SD
M M M M M F M F
55 55 61 52 61 59 50 60 50 51 56 4
132 132 127 134 127 127 136 128 136 136 131 4
135 132 128 140 128 148 134 140 144 137 137 6
83 78 83 99 95 78 99 99 78 73 87 10
60 65 60 85 68 92 78 85 60 80 71 12
Ll-
Subject
M
AC = accelerometer-based pacemaker; PZ = piezoelectric-based pacemaker.
Percent changes in HR and pacing rates were also calculated. Increases from rest to exercise at 25 W for intrinsic HR were 14%, 21%, and 23% at 50, 60, and 70 rpm, respectively. Increases in pacing rate for the AC pacemaker were 10%, 17%, and 28% at 50, 60, and 70 rpm, respectively. The PZ pacemaker, however, only increased 0.5%, 3%,
100
•t
I
* indicates a signllicani dirfersnc* between HR andfor PRt. t Indicates a signiiicani dllfeiance between rpm Irialt wllhin ea
E] SOipm O BOipm Q 70rpm
7S W RZ
12S W
150 W
B«eBV«i>
Workload
group.
Figure 3. Hate response to changes in rpm during bicycling at 25 watts. HR ^ heart rate; PR = pulse rate.
PACE, Vol. 15
100 W
Figure 4. Rate response fmean ± SE) during incremental bicycling. HR = heart rate; PR = pulse rate; AC = accelerometer-based pacemaker; PZ = piezoelectricbased pacemaker.
February 1992
193
BACHARACH, ET AL.
7y
¥
8C spm Acs.
U
BO>pm Desc.
m 10D spm Ac5. u 100 spm Dose.
Figure 5. Rate response (mean ± SEJacsending (AcsJ and descending fDescJ stmvs at 80 and 100 steps per minute. HR ^ heart rate; PR = puJse rate; AC ^ accelerometer-based pacemaker; PZ = piezoeJectric-based pacemaker.
balween Ihs vatious rate
and 4% at 50, 60, and 70 rpm, respectively. Similar results were seen during the incremental bicycle test. Mean percent increases from resting values to peak exercise were 75% for intrinsic HR, 41% for AC pacing rate, and 8% for PZ pacing rate. Stair Climbing and Descent
Mean values [ ± standard error) for the stair climbing protocol are shown in Figure 5, A oneway ANOVA (P < 0.05) was used to compare maximal HR and rates of the two pacemakers. When ascending stairs at 80 and 100 steps per minute, no significant differences between HR and the AC rate were identified; however, the PZ rate was significantly lower than the other two rates during each ascent. During descent, only the AC at 80 steps per minute was significantly different from the HR. The AC rate changes paralleled the HR rate changes to ascent and descent, respectively. However, the PZ device exhibited an increase on descent that is paradoxical when compared to HR for both step frequencies.
Discussion Recognizing the intent of this study was to evaluate the way in which rate modulation is accomplished by an AC sensor verses a PZ sensor bonded to the canister, the results indicate that at the manufacturer's nominal settings, the AC pacemaker provided a more proportional response to incremental treadmill activity than did the PZ
194
pacemaker. Considerable variations occurred between the intrinsic HR and the PZ paced rate for six of the ten subjects (Table III). These findings indicate that nominal values for the PZ pacemaker may not be appropriate for some subjects and that additional fine tuning would be required to attempt to achieve an acceptable rate response. The PZ pacemaker's inappropriately high rates near the onset of exercise, as shown in Figure 2, have been reported previously^^ and provide evidence against the notion that the activity threshold setting was too high. In addition, some subjects demonstrated blunted rate response, also indicated in Figure 2, which have been commonly observed by clinicians. The relationship of the AC pacemaker to intrinsic HR suggests that its nominal values may provide a better chronotropic response to treadmill exercise in a majority of subjects. During cycling, both pacemakers had difficulty keeping up with intrinsic HR. Little upper body motion is generated during cycling, which may explain the PZ pacemaker's muted response at nominal settings. Considering activity-based pacemakers may not be specific to all activities, the rate increase observed in the AC pacemaker during cycling would be advantageous provided rate specificity has not been compromised. A closer look at Figure 3 does reveal increasing rates with higher pedaling speeds for the two pacemakers as well as intrinsic HR. The AC pacemaker rate increased similarly to increases in HR; however, the standard error for the AC pacemaker was so small that these differences were statisti-
February 1992
PACE, Vol. 15
COMPARISON OF ACTIVITY-BASED PACEMAKERS
cally significant, while the changes in intrinsic HR as a function of rpm were not statistically significant. As indicated, percent increases for HR and the AC pacemaker were similar, while the PZ pacemaker showed virtually no change. This indicates the AC pacemaker responded more appropriately. The mean resting HR was 76 ± 4 beats/min as compared to mean pacing rates of 66 ± 0.3 ppm (AC) and 63 ± 0.5 ppm (PZ), making percent increases in rates biased toward the activity-based pacemakers. The PZ pacemaker's paradoxical rate increase during descent of stairs was noted. However, results of the AC pacemaker for ascending and descending stairs at 80 and 100 steps per minute are in the same direction, and similar to previously reported data.**'^ Adjustments in parameter settings of the AC and PZ pacemakers can improve rate modulation during exercise for those subjects demonstrating insufficient activity or an inordinate amount of disturbance to generate the expected response. It has been suggested that activity-based rate responsive pacemakers such as the PZ pacemaker provide a less than satisfactory rate response because of the frequency range to which they are most sensitive. Alt et al.^ stated, "A maximum sensitivity in the area of 10 Hz has disadvantageous consequences to an activity-controlled system, since interference influences have their amplitude peak in this frequency range. It is therefore impossible to discriminate between exercise induced signals and signals which should not lead to a frequency response." (p. 1676) It is the opinion of these authors that there may be several reasons for the AC pacemaker's greater sensitivity and specificity to body motion. In observing the characteristics of the two pacemakers, the PZ device, with an optimal frequency range from 10-50 Hz, may be susceptible to nonspecific pacing responses or extraneous pressure applied to the canister since the piezoelectric crystal is dependent upon distortion of the crystal bonded to the pacemaker canister. The AC device, with an optimal frequency range of 1-8 Hz, has the sensor mounted within the pacemaker yet independent of the canister. Therefore, the AC device's specificity to activity should not be compromised with the improved proportionality and sensitivity. The scope of this study
PACE, Vol. 15
did not, however, incorporate additional tests to address the issue of specificity and it is as yet undetermined as to whether the results achieved here are attributable to differences in frequency ranges of the two pacemakers. Certainly, these data support rationale for further investigations using an AC pacemaker and a PZ pacemaker optimized for each individual while executing a variety of daily tasks attempting to answer the specificity question for human movement. A review of the recovery data demonstrates the importance of a correct response parameter setting. At nominal settings, the AC pacemaker provided an appropriate rate during recovery; however, the variability of the results achieved by the PZ pacemaker make it difficult to evaluate its recovery data. The recovery algorithms in each device are related to time rather than to amount of activity measured by tbe sensor. Recovery is simply dependent on the difference between the present rate, the lower rate limit, and the programmed recovery parameter. If an incorrect response setting causes an inappropriate present rate, the return to the lower rate limit will also be altered.
Conclusions As suggested by Sulke et al.,^ accurate response programming is essential for optimal performance of variable-rate pacemakers. The AC and PZ pacemakers can provide rate modulation for individual subjects at nominal settings; however, this study showed the AC pacemaker's rate is more closely associated to intrinsic HR during treadmill walking, bicycle ergometry, and ascending stairs than a PZ pacemaker's rate. The AC pacemaker set at nominal values is better correlated to an individual's intrinsic HR response (absolute mean difference = 11 ppm) than was the PZ pacemaker's rate when set at matching nominal values (absolute mean difference = 26 ppm] during treadmill exercise. Although neither pacemaker performed well enough to closely match intrinsic HR during cycling, the AC pacemaker appears to be more sensitive to increases in workrate than the PZ pacemaker during progressive bicycle ergometry. Finally, the AC pacemaker can perform better than the PZ pacemaker when ascending stairs at 80 or 100 steps per minute; however, both pacemakers
February 1992
195
BACHARACH, ET AL.
were higher than intrinsic HR while descending stairs at 80 or 100 steps per minute. Overail, these data favor the use of an AC pacemaker to provide a more appropriate rate response than a PZ paceReferences
maker in which the piezoelectric crystal is bonded to the canister to incremental treadmill, bicycle exercise, and stair climbing when response parameter settings are standardized.
1. Humen DP, Kostuk TWJ, Klein GJ. Activity sensing, rate-responsive pacing: Improvement in myocardial performance with exercise. PACE 1985; 8:52-59. 2. Lindemans FW, Rankin JR, Murtasugh R, et al. Clinical experience with an activity sensing pacemaker. PACE 1986; 9:978-986. 3. Furman S, Hayes DL, Holmes DR. A Practice of Cardiac Pacing. 2nd ed. Mount Kisco, New York, Futura Publishing Company, Inc., 1989, pp. 414-428. 4. Miura DS. A new concept in pacemaker therapy: Rate responsive pacing. Cardiol Rev Rep 1987; 8:40-44. 5. Suike N, Dritsa A, Chambers J, et al. Is accurate rate response programming necessary? PACE 1990; 3:1031-1044, 6. Sulke N, Pipilis A, Henderson RA, et al. Comparison of normal sinus node with seven types of rate responsive pacemakers during everyday activity. Br Heart J 1990; 4:25-31. 7. Lau C-P, Mehta D, Toff WD, et al. Limitations of rate-response of an activity-sensing rate-responsive pacemaker to different forms of activity. PACE 1988; 11:141-150. 8. Alt E, Matula M, Theres H, et al. The basis for activity controlled rate variable cardiac pacemakers: An analysis of mechanical forces on the human body induced by exercise and environment. PACE 1989; 12:1667-1680.
Lau C-P, Butrous GS, Ward DE, et al, Comparison of exercise performance of six rate-adaptive right ventricular cardiac pacemakers. Am J Cardiol 1989; 63:833-838. Bunge T, Thompson D. Sensing internal and external body activities. In F Perez Gomez (ed.): Cardiac Pacing: Electrophysiology: Tachyarrythmias. Madrid, Spain, Editorial Grouz, 1985, pp, 786-791, Lau C-P, Stott JRR, Toff WD, et al. Selective vibration sensing: A new concept for activitysensing rate-responsive pacing, PACE 1988; 11:1299-1309. MianuUi M, Benditt DC, Markowitz T. et al. A comparison of strap-on versus implanted activitybased rate-responsive pacemakers: Are strap-on studies valid? (abstract) PACE 1991; 14:742. Benditt DG, MianuUi M, Fetter J, et al. An officebased exercise protocol for predicting chronotropic response of activity-triggered rate-variable pacemakers. Am ] Cardiol 1989; 64:27-32. Wilkoff BL, Corey J, Blackburn G. A mathematical model of the cardiac chronotropic response to exercise. J Electrophysiol 1989; 3:176-180. Astrand PO, Rodahl K. Textbook of Work Physiology: Physiological Bases of Exercise. New York, NY, McGraw-Hill Book Company, 1977, pp. 369-371.
196
10.
11.
12.
13.
14.
15.
16.
February 1992
Mond H, Line 1, Hunt D. A third generation activity pacemaker: Is the rate response superior? (abstract) PAGE 1990; 13:514.
PACE, Vol. 15
