Transtelephone Pacemaker Monitoring

Transcription

1 Transtelephone Pacemaker Monitoring Five Years Later Seymour Furman, M.D., and Doris J. W. Escher, M.D. ABSTRACT Six hundred nineteen patients have been followed by remote monitoring of pacemaker function using ECG and rate or rate alone; 278 of 280 have had battery exhaustion or electronic failure demonstrated. Ten percent of exhausted pacemakers failed prior to the average longevity of the particular model, and 32% (89 of 280) exceeded 36 months longevity; of these, 13% (37 of 280) lasted more than 40 months and 4.6% (13 of 280) exceeded 50 months. The error rate is 0.7% (2 of 280). With pulse generator longevity increasing, monitoring is done less frequently during the first 2 years, then calls are made weekly after 24 months. T he durability of pacemaker systems has always been variable. Patient welfare has required observation following implantation to reduce the incidence of sudden and unpredicted failure of the implanted system as well as to detect early failure of some units and maximize the useful longevity of others. Pacemaker Failure The major reason for the necessity of follow-up over a prolonged period is that the time of failure is spread over a large portion of the total longevity of any pulse generator model. Elective removal of all generators during normal function, as when 10% of a series has failed, produces extremely short functional periods and accepts an undetected failure rate of 10%. A recent model of ventricular inhibited pacemaker (VVI)* had 10% failure at 15 months. Removal of the entire series at that point would have produced an average longevity of 13 months. When all the units were followed to exhaustion or electronic deterioration, half the series continued to function at 28 months, with an average longevity of 28.3 months and a standard deviation of 8.8 months (Fig. 1). A model contemporary with it had a 10% failure time of 16 months and would have had an average longevity of 15 months if all units had been removed at that time. Fifty percent had failed at 22 months, and the average longevity was 23 months with a standard deviation of 4.8 months [S]. Such wide standard deviations required that one group of patients be carefully monitored over 18 months (? 1 SD) and the other over 10 months (& 1 SD) in order to achieve a 66% detection rate for impending failure; these periods had to From the Cardiothoracic Service, Division of Surgery and the Cardiology Service, Division of Medicine, Montefiore Hospital and Medical Center, Bronx, N.Y. Supported in part by U.S. Public Health Service Grant no. HE Presented at the Eleventh Annual Meeting - of The Society of Thoracic Surgeons, - Montreal, Que., Canada, Jan , Address remint reauests to Dr. Furman. 111 E. 210th St., Bronx, N.Y *Terminology for pulse generators follows that suggested by Parsonnet, Furman, and Smyth [ THE ANNALS OF THORACIC SURGERY

2 Transtelephone Pacemaker Monitoring c % E - -..,, I 1 i s Number Of Months After Implant Number Of Failurem FIG. 1. This model, successful for its time, was implanted during Failure was approximately normal in distribution, with mean and median at 28 months and a standard deviation of 8.8 months. Ninety-five percent of failures occurred over a period of 35 months. Data on Cordis Model 143A7: total sample, 101; total calculated, 50; total dead or removed, 36; lost to follow-up, 5; remaining, 10 (average, 41.O mo); average, 28.3 mo; median, 28 mo; standard deviation, 8.8 mo. be doubled to detect 95% of failures. For neither group would elective replacement at 15 to 16 months have been consistent with patient safety or efficient use of the pulse generator. Were the spread, or standard deviation, of failure to narrow, monitoring could begin later and would be required for only a short period. A standard deviation of only 1 month for pulse generator failure would require monitoring for a 4-month period to detect 95% of all failures. Battery exhaustion has been and remains the single most common reason for pulse generator replacement, but electronic and packaging defects must also be considered. During , of 32 1 pulse generators (average longevity, 27 months) removed for some variety of failure, 123 of the failures were electronic in nature (average longevity, 19.1 months) and 198 were due to battery exhaustion (average longevity, months). A late postimplant rising threshold still occurs at a rate of about 3% each year. Electrode fracture remains a persistent event, causing 4 to 8% of the yearly secondary procedures [lo]. The need for monitoring will change if the average duration and reliability of pulse generators improve and if the initial failures occur later. The average useful functional period of generators removed for battery exhaustion during 1974 was 32 months (Fig. 2), during which 35% of the patients died. A reliable increase of the 10% pacemaker failure time from 16 to 32 months would substantially reduce the need for monitoring for a very large group and would shift the monitoring burden to failures occurring later. The mercury-zinc cell has established generator longevity to be between 18 and 36 months, with a sudden voltage decline and battery failure at end of life. An energy source that decays linearly and progressively can be monitored less frequently and removed when rate or another determinant reaches a point incom- VOL. 20, NO. 3, SEPTEMBER,

3 FURMAN AND ESCHER a.w 3a 37 M j; - 30! 20 I7 2 I6 rae c.- 18 pr, f 12 u mil 12 I a I011 I FIG. 2. Normal battery exhaustion is occurring later with improved pulse generator design. Pacemakers removed for exhaustion only, tabulated on a monthly basis, have increased in longevity from 21.3 months in 1971 to 31.8 months in patible with normal function. In the next few years, significant changes in time of onset as well as frequency of monitoring can be anticipated as new energy sources are used. A high incidence of electronic malfunction reduces the accuracy of failure detection, as comprehension of a pulse generator failure pattern is the basis of any follow-up system. Sudden changes in this pattern, or a high incidence of random failures, allow a high number of failures to go undetected until the monitoring system is adjusted to a-new schedule. From 1970 to 1973 the overall accuracy of failure prediction was about 80%. Emergencies during 1974 occurred because of electrode fracture (4%) and electronic failure (9.8%). Battery exhaustion is a process that usually falls within anticipated time limits and is almost always detected. Of the 19 pulse generator replacements done in 1974 for undetected failure, 16 (84%) were necessitated by failure of electronic components that occurred during the first year after implanting. Seven other impending electronic failures (3.7%) were detected by telephone and the generators replaced before failure occurred. This period previously had been one of light observation. The problems of failure detection were eventually resolved by early frequent telephone monitoring and by elective removal of a large number of suspect pulse generators (32 of 194, or 16.5% of the total for 1974). Methods of Monitoring The electronic clinic demonstrated that measurement of electronic function allows pulse generator failure to be predicted. One month, the shortest practical 328 THE ANNALS OF THORACIC SURGERY

4 Transtelephone PacemaRer Monitoring interval between patient visits, is not frequent enough to achieve the necessary accuracy for detecting battery depletion, which requires weekly data determination [2,7]. Electronic monitoring served as the mainstay of pacemaker follow-up for many years [l 1, 141 and is now complementary to the transtelephone system. Use of both telephonic and electronic follow-up has shown that visits at intervals of 4 months during the first year after implantation, every 3 months during the second year, and telephonic follow-up weekly thereafter are most efficient and productive, but the schedule must be modified as experience with specific pacemaker models dictates. If direct electronic follow-up is not practiced, telephone monitoring should be begun initially after implantation. The accuracy of telephone monitoring in detection of battery depletion is unmatched by that of direct electronic monitoring because of the ease and frequency of the transmissions [4]. The single area of failure has been the inability to detect impending lead fracture, which is, unfortunately, undetectable by most means. The purposes of clinic monitoring by electronic wave-form analysis are: 1. Evaluation of return of symptoms despite apparently normal pacing. This problem may require maximum operator and electronic skill to detect intermittent malfunction. 2. Adjustment and control of generators capable of variability in impulse duration (Medtronic and 5961 *). 3. Adjustment and optimum control of noninvasive programmable pacers (Cordis Omnicort). 4. Initial observation of pulse generators before telephone monitoring is required. 5. Evaluation of pulse generators suspected of having electronic defects that are not manifested by a change in rate or a clear-cut failure to pace or sense spontaneous cardiac activity. Electronic clinic monitoring continued on a monthly basis for prolonged periods is burdensome for staff and patient and should be replaced by telephone monitoring. Telephone monitoring alone, with rate determination and an electrocardiogram, can be used at frequent intervals following pacer implantation to ascertain electrode stability, then at longer intervals, and as the unit ages, at ever shorter intervals. The purposes of telephone monitoring are: 1. Careful pacemaker follow-up to achieve maximum safety and maximum longevity. 2. Careful and frequent observation of pacemakers in a series known to have a high incidence of electronic failure or of a unit that is no longer operating normally but for which clear-cut malfunction has not been demonstrated. *Medtronic, Inc., Minneapolis, Minn. Wordis Corp., Miami, Fla. VOL. 20, NO. 3, SEPTEMBER,

5 FURMAN AND ESCHER 3. Remote follow-up for patients who would otherwise receive little or no aftercare directed at the pacemaker. The capacity of the system to be used as a laboratory resource in which the primary physician receives adequate data about the patient to make a judgment, but does not relinquish responsibility for patient care, is especially important in the United States and has encouraged extension of pacemaker follow-up to several thousand patients from whom it otherwise might have been withheld [ The extension of monitoring to remote areas where travel is so great an impediment that follow-up would not be done. TRA NSTELEPHON IC MONITORING Rate Only. Rate-only monitoring was the initial approach. It is suitable for any variety of pacemaker and requires only that the unit have a magnetic switch to convert to asynchronous operation or that pacing be frequent enough that intervals between pacemaker stimuli can be determined. Battery failure is detected when the first cell is exhausted and pacer rate declines sharply [6, 131. Sensing function is difficult to determine except when tracking pulses can be used to distinguish between paced and sensed beats.* This method is still useful and is frequently employed. Rate and Digital Plethysmography. This early approach was adapted from the rate-only technique because of the occasional need to determine whether ventricular capture has occurred. The lack of positive evidence of capture is the major defect of the rate-only technique and is only partially solved by rate and digital plethysmography, as the incidence of false-positive and false-negative results may be substantial, though statistical data have never been made available [6, 121. Rate and Electrocardiogram. This provides rate determination and an ECG, usually but not necessarily lead I, to ascertain the adequacy of capture and pacer sensing. The ECG is superior to digital plethysmography since it is the universally recognized indicator of cardiac activity, and it is preferred when remote follow-up starts shortly after pacer implantation because failure to capture or sense ventricular activity is readily recognized. If follow-up is begun late, perhaps 1 to 2 years after pacer implantation, the ECG is less necessary since cessation of ventricular capture rarely occurs in the presence of a normal pacemaker rate and presumed normal output. Two general varieties of rate and ECG monitoring exist. Both have a carrier frequency modulated by the ECG, producing a wavering tone acoustically coupled to the telephone. In one,? a highly accurate ECG results from an 1,800 Hz carrier signal that transmits the pacer spike and ECG to the receiver. Both are demodulated and converted to a signal capable of operating an ECG and an interval timer that maintains both artifact amplitude and polarity. The other4 is a carrier-interrupt system with a 2,000 cycle carrier signal interrupted by the pacemaker impulse. The impulse is not transmitted, and a gap is left in the carrier signal where the pacer stimulus had occurred. An artifact is then added to the ECG at the receiver that distorts the low-frequency spectrum of the ECG. The *Starr-Edwards Cardiac Pacemaker System, Edwards Laboratories, Santa Ana, Calif. tmedalert Gorp., New York, N.Y. Spacer-Check, ESBIMedcor, Yardley, Pa. 330 THE ANNALS OF THORACIC SURGERY

6 Transtelephone Pacemaker Monitoring pacemaker spike as seen on the final ECG is a total artifact and bears only a time relation to the stimulus generated by the pacemaker. Neither amplitude nor polarity can be determined. The carrier-interrupt system is easier for the patient to work and allows an unskilled operator to derive useful data. The more discriminating system provides a superior ECG but requires an operator of greater skill and is most readily performed in a central laboratory facility. In each instance the transmission is performed from the patient s home. Rate and ECG ahd Electronic Analysis. Potentially the most sophisticated system for remote information transmission, it requires that the patient visit a peripheral facility with a relatively elaborate transmitter from which data are transmitted to a central station for analysis.* The need for the patient to travel to the transmitter makes the technique similar to a conventional electronic clinic operation, removing the ease of home transmission. A device of similar capability to transmit from the patient s home is under development [3]. ELECTROCARDIOGRAPHY IN TRANSTELEPHONE MONITORING The two varieties of ECG provide interpretation and artifacts which are similar to conventional electrocardiography, and both clearly indicate failure to capture, sense, or both. A few artifacts are unique to transtelephone monitoring. The most serious artifact is that caused by the use of a single-lead rhythm strip, usually lead I, in which the QRS or the pacemaker artifact may be isoelectric or for some reason insufficiently diagnostic. Another lead should be selected and transmission repeated. The diagnostic lead can be used thereafter. Telephone noise is frequently picked up from continuous interference and electrical transients. Usually both are detected readily. Patient reversal of the limb leads is readily detected and is of little consequence. Motion artifacts occur frequently. There can be troublesome 60 Hz interference with all systems. The stimulus rate and the ECG can usually be discriminated. Because the magnetic determination of rate requires a magnet to be held over the pacemaker, its movement (as in a patient with Parkinson s disease) can cause a variety of artifacts, including cessation of pacing and inhibition of the pacer [51. Respiratory movement is recognized, as on a conventional ECG, by a slow, rhythmic oscillation of the baseline. The degradation of the ECG signal on the carrier-interrupt monitor may cause difficulty with interpretation of the QRS complex. A wide variety of changes may be detected with ECG monitoring. These include the range of pacemaker malfunction: rate change and loss of pacing or sensing or both (Fig. 3). Relatively common intercurrent findings are multifocal premature ventricular contractions, tachycardia, atrial fibrillation, and other arrhythmias (Fig. 4). In more than 99.9% of transmissions the induction of a competitive, magnet-produced rhythm to detect pacemaker rate is without meaningful effect. Pacemaker-produced premature ventricular contractions occur but are not sustained, and only very rarely do multiple PVCs require cessation of use of the *Pacemaker Diagnostic Clinic of America, Gainesville, Fla. VOL. 20, NO. 3, SEPTEMBER,

7 FURMAN AND ESCHER FIG. 3, Normal pacing with intermittent unsensed complexes (carrier-interrupt transmission). magnet (see Fig. 4). No episode of ventricular fibrillation or sustained tachycardia, competitively induced, has occurred. Monitoring Technique The schedule of follow-up visits or telephone calls should be based on the best available knowledge of the characteristics of pacemaker operation and should be adhered to as closely as possible. Careful records will provide quick diagnostic FIG. 4. In addition to findings related to pacemaker function, a variety of other events can be documented: (A) Competitive pemature ventricular contractions and an asynchronous pacemaker. (B) Competition between a noncompetitive pacemaker in the magnetic asynchronous mode and spontaneous contractions poduces short, self-limited runs of ventricular tachycardia (the only such instance in this series). (C) Premature ventricular contractions, compensatory pause, and a pseudofusion beat. (0) Spontaneous multijocal ventricular tachycardia, pacemaker normal. 332 THE ANNALS OF THORACIC SURGERY

8 Transtelephone Pacemaker Monitm'ng information when failure occurs and must be readily available for comparison with earlier data (Fig. 5). Over half of the patients, even in large metropolitan areas with good medical facilities, do not receive adequate pacemaker follow-up [ 13. Pacemaker systems tend to fail during three distinct periods after implantation and should be monitored correspondingly. At present the schedule for telephone calls is every 2 months for the first 6 months after implantation and at shorter intervals until the eighteenth or twentyfourth month (depending on pacemaker model), when weekly calls are begun. Weekly calls are started earlier if the behavior of a specific pacemaker model or individual unit is suspicious. Early Postimplant (months I -3). Electrode malfunction occurs during the first 2 weeks and should be detected during postimplant observation. At Montefiore Hospital and Medical Center, 3.4% of transvenous pacemaker implants require revision during the early period following implantation. Surgical problems of wound healing cannot be detected remotely, and conventional observation by the surgeon is required. Random electronic failure occurs infrequently. Failure to sense or pace properly happens often during this period in the general population of pacemaker implants. A follow-up service* accepting patients from a variety of physicians reports that 10% of all pacemaker system failures occur during the first 3 months after implantation, representing problems of electrode placement. Middle Period (4-24 months). Three varieties of failure occur, though all are relatively uncommon with a modern, well-designed pacemaker system: (1) Battery exhaustion, which should not occur in more than 5% of modern pulse generators before the twenty-fourth month - some models appear to have reached that goal; (2) Lead fracture (4% in 1974); and (3) Electronic failure. The method of follow-up should be matched to experience with the specific pulse generator. Late Period (beyond24 months). Battery exhaustion is the typical failure mode. Lead fracture continues at a constant rate (4% per year). Electronic failures increase as pulse generator components drift out of specification despite satisfactory battery function. Results Our follow-up program has made only 2 errors. One occurred with rate-only monitoring when ventricular capture was erroneously assumed to be present because the pacemaker rate was constant. The other happened with rate and ECG monitoring in which battery voltage was down: the rate was stable and pacing was consistent during each transmission, but pacing had become intermittent and syncope had returned. The error rate of transtelephone monitoring is thus about 0.7% of the failure determinations made on patients followed to generator removal, 2 patients of 280 (Table). *Medalert Corp., New York, N.Y. VOL. 20, NO. 3, SEPTEMBER,

9 FURMAN AND ESCHER hhjz-4.- I c , - I ; A FIG. 5. Record keeping is as important as data acquisition. (A) Face ofjle card used for telephone transmissions. Of the 6 19 patients followed since 1969,280 have had failure detected before cessation of pacing and return of symptoms. Two hundred two patients are now actively followed. The average number of calls has been 35 per patient, 1 being the least and 132 the most for a pacemaker suspected of premature battery failure but which lasted 56 months. The average longevity of pacemakers removed was 34 months. Ten percent of exhausted pacemakers failed prior to the average life for the particular model, and 32% (89 of 280) exceeded 36 months' longevity. Of 334 THE ANNALS OF THORACIC SURGERY

10 Transtelephone Pacemaker Monitoring FIG. 5. (Continued.) (B) Back of file card. Rate stability is clearly visible, as is a rate change during the twenty-jifth month, the significance of which was unknown. End of pacemaker lfe was heralded by a beginning decline and then a sharp decline of 6.8 beats per minute in one week. The pulse generator was replaced during the thirty-jifth month of operation. (Retouched to hide patient s identity.) these, 13% (37 of 280) lasted longer than 40 months and 4.6% (13 of 280) exceeded 50 months. One pulse generator was removed 63 months after implantation. These results have been derived from a carefully controlled hospital-based patient population. Pacemaker follow-up in the general population is of equal interest. Data provided by a non-hospital-based follow-up system* that accepts patients without any preselection provides an informative comparison. During the period March, 1972, to September, 1974, a total of 606 patients were followed for some period, with an average enrollment of 182. Fifty-four pacemakers were electively replaced despite follow-up. Failure or impending failure was detected in 104 pacemaker systems. This was 57% of the average number followed, and 10 failures were discovered at the very first (enrollment) transmission, so that function during the preceding period of no pacing is unknown. Forty-three (41%) of the failures were due to battery exhaustion, 3 (3%) to electrode displacement, 6 (6%) to wire breaks, 8 (8%) to sensing failures, 13 *Medalert Gorp., New York, N.Y VOL. 20, NO. 3, SEPTEMBER,

11 FURMAN AND ESCHER RESULTS OF TRANSTELEPHONE MONITORING IN 619 PATIENTS Result Failure detected Errors Pacer electronic failure (undetected) Lead fracture - pacer normal Patient died - pacer normal Elective cessation of monitoring Patients actively followed No. of Patients (12.5%) to loss of ventricular capture with pacemaker rate normal, 13'(12.5%) to rate increases of more than 10 beats per minute in pulse generators that should have decreased in rate with battery depletion, and a variety to sudden cessation of pacer function and other failures, the nature of which was never determined. The failure rate requiring surgical intervention was 23% per year of pacemaker operation.* References 1. Bilitch, M., Cassady, E. E., and Lloyd, J. S. Physician Follow-up of Patients with Permanent Cardiac Pacemakers. In H. J. T. Thalen (Ed), Cardiac Pacing: Proceedings of the 4th International Symposium. Assen, Netherlands: van Gorcum, P Escher, D. J. W., and Furman, S. Oscilloscopic and recent other methods of implantable pacemaker follow-up. Ann Cardiol Angiol 20:503, Escher, D. J. W., Hochberg, H., King, E., Chudoba, P., and Kaufman, P. Pacemaker electronic follow-up: Complete unique system. Circulation 49,50 (Suppl 111): 94, Furman, S. Transtelephone observation of implanted cardiac pacemakers. Med Znstrum 7:196, Furman, S., Escher, D. J. W., and Parker, B. Failure of triggered pacemakers. Am Heart J 82:28, Furman, S., Escher, D. J. W., and Parker, B. The pacemaker follow-up clinic. Prog Cardiovasc DW 14:515, Furman, S., Parker, B., and Escher, D. J. W. Transtelephone pacemaker clinic.] Thorac Cardiovarc Surg 61:827, Morse, D. P., Tesler, U. F., and Lemole, G. M. The actual lifespan of pacemakers. Chest 64:454, Parsonnet, V., Furman, S., and Smyth, N. Implantable cardiac pacemakers: Status report and resource guideline. (Pacemaker study group.) Circulation 50:A21, Parsonnet, V., Gilbert, L., and Zucker, I. R. Natural history of pacemaker wires. J Thorac Cardiovasc Surg 65:315, Parsonnet, V., Meyers, G. H., Gilbert, L., and Zucker, I. R. Prediction of impending pacemaker failure in a pacemaker clinic. Am J Cardiol 25:311, Pennock, R. S., Dreifus, L. S., Morse, D. P., and Watanabe, Y. Cardiac pacemaker function. JAMA 222:1379, Ruben, S. Sealed zinc-mercuric oxide cells for implantable cardiac pacemakers. Ann NY Acad Sci 167:627, *A. Auerbach, personal communication. 336 THE ANNALS OF THORACIC SURGERY

12 Transtelephone Pacemaker Monitoring 14. Van Den Berg, J. W., Rodrigo, F. A., Thalen, H. J. T., and Koops, J, Photoanalysis of the condition of implanted pacemaker electrode circuits. Proc Kon Ned Akad Wet (Biol Med) 70:419, Discussion DR. DAVID C. MACGREGOR (Toronto, Ont., Canada): I would like to congratulate Dr. Furman on his pioneer work in transtelephone monitoring and to thank him for his assistance in establishing our own follow-up program at the Toronto General Hospital. We have recently reviewed our experience with 224 pulse generators replaced over the past 28 months on the basis of telephone evaluation alone. ECG and rate determination were used in every case, and only units which had been in place for 18 months or more were considered. Our results differ considerably from those of Dr. Furman. Seventy-eight (35%) of the pulse generators were replaced electively, 99 (44%) were predicted failures on the basis of rate change alone, and 47 (21%) were unpredicted failures in that a rate change was not observed prior to loss of capture or failure of sensing function or both. The mean longevity of all pulse generators was 29.7 months. Ofthe 47 unpredicted pulse generator failures, 25, or 11.2% ofthe total group, lost the ability to sense or pace, or both, without an accompanyingchange in rate, and as such would not be detected by rate-only monitoring. Four failures resulted from unpredictable electrode fractures. Eight patients had clinical symptoms, but there were no deaths between the recognition of pulse generator failure and the time of replacement. Although it is often stated that, at present, the stimulation rate of virtually all pacemakers varies as battery exhaustion occurs, numerous patients still have pacemakers that do not incorporate this feature. For this reason we recommend strict adherence to manufacturers replacement recommendations. We also believe that transtelephone monitoring should include an electrocardiogram to identify premature battery exhaustion in this group of patients as well as to detect random component failures, which continue to occur in all types of pacemakers. Furthermore, the electrocardiogram may identify pacemaker-related or other underlying cardiac arrhythmias that often require treatment. DR. ALBERT STARR (Portland, Ore.): We certainly owe a great deal to Dr. Furman for his pioneering work with pacemakers, especially in pointing out the need for an organized program of follow-up and also in demonstrating that the telephone system has certain advantages. We regard this kind of follow-up system as a supplement to standard cardiological care, but not as a substitute for it, and that is the important concept. Our own transtelephone system began its clinical use in 1970, and we are currently following 350 patients. The system has been made possible by the development of a proved, truly rate stable pulse generator of a ventricular tracking type. Rate is thus an exquisite indicator of battery voltage, and hence rate alone provides adequate pacemaker analysis. Competitive pacing is avoided during telephone transmission, since at both the high and low rates the pulse generator is always in the demand mode. The tracking signal in this pacemaker facilitates phone monitoring by marking the patient s own QRS complex; thus it can be considered a kind of simplified digitized electrocardiogram. We have organized this so that the clinician receives a pacergram as a laboratory service, not as a professional service. This reduces the cost enormously and avoids the practice of cardiology on the telephone. DR. FURMAN: I would like to thank the discussants for their very kind and searching comments on the technology involved. I must disagree with Dr. Starr concerning the electrocardiogram, which is the universally recognized means of determining cardiac action and should be used when at all possible. Many of the intercurrent findings we have made that were not related to pacer function were detected because of the electrocardiogram, and neither rate-only monitoring nor tracking would have yielded that information. VOL. 90, NO. 3, SEPTEMBER,

13 FURMAN AND ESCHER My second point is that we probably will see a noticeable decline in the loss of cardiac capture without a significant rate change, as that was representative of only one major manufacturer and only one major model. All the units now in use seem to have a rate change or some other clear-cut indication of battery depletion. In any event, even in that rate-stable model we were able to determine that a very small rate change was the indicator of battery depletion, and except for 1 patient we had very little trouble even in that single model. NOTICE FROM THE AMERICAN BOARD OF THORACIC SURGERY A limited number of copies of the 1974 Self-Assessment Examination for Thoracic Surgery (SATSE), together with an answer sheet, are available for $25 per copy from the office of the American Board of Thoracic Surgery, E. Seven Mile Rd., Detroit, Mich THE ANNALS OF THORACIC SURGERY