Europace (2001) 3, doi: /eupc , available online at on

Transcription

1 Europace (21) 3, doi:1.153/eupc , available online at on ICDs Clinical evaluation of defibrillation efficacy with a new single-capacitor biphasic in patients undergoing implantation of an implantable cardioverter defibrillator J. Brugada 1, B. Herse 2, B. Sandsted 3, U. Michel 4, B. D. Schubert 4 and S. J. Hahn 5 1 Arrhythmia Unit, Cardiovascular Institute, Hospital Clinic, University of Barcelona, Spain; 2 University Clinic, Goettingen, Germany; 3 University Clinic, Gothenburg, Sweden; 4 Guidant Research, Brussels, Belgium; 5 Guidant Tachyarrhythmia Research, St Paul, U.S.A. Aims Improvements in the size and shape of implantable cardioverter defibrillators (ICDs) might be obtained by using one capacitor instead of the series connection of two capacitors traditionally used in ICDs. The aim of this study was to determine whether a biphasic delivered from a single 336 μf capacitor had the same defibrillation efficacy as a standard biphasic. Methods and Results Randomized, paired defibrillation threshold testing was acutely performed in 54 patients undergoing ICD implantation. A standard 14 μf 8% tilt biphasic (two 28 μf capacitors connected in series) was compared with an experimental biphasic delivered from a single 336 μf capacitor at either 6% tilt (33 patients) or 8% tilt (21 patients). All s had a 6/4 phase1/phase2 duration ratio. Compared with the standard, the 6% tilt experimental had a lower delivered energy ( vs joules, P< 2), lower peak voltage ( vs V, P< 1), and a slightly longer pulse duration ( vs ms, P< 1). Conversely, the 8% tilt experimental had a higher delivered energy ( vs joules, P< 1), a lower peak voltage ( vs V, P< 1) and a much longer pulse duration ( vs ms,P< 1). Conclusion Waveforms delivered from a large capacitance are feasible but require a lower tilt. This technique may allow smaller, thinner ICDs without jeopardizing defibrillation success. (Europace 21; 3: ) 21 The European Society of Cardiology Key Words: Human, defibrillation, pulse-, capacitor, pulse- tilt. Introduction Over the past several years, research and technological advances have allowed important improvements in Manuscript submitted 13 November 2, revised 18 June 21, and accepted 2 July 21. This study was supported in part by Guidant Research, St Paul, Minnesota, U.S.A. and Brussels, Belgium. Correspondence and/or reprint requests: Dr Josep Brugada, Director of the Arrhythmia Unit, Cardiovascular Institute, Hospital Clinic, c/ Villarroel 17, 836 Barcelona, Spain. jepbrugada@grn.es implantable cardioverter defibrillators (ICD). New and highly effective algorithms for detection, treatment and follow-up have been developed [1,2,3] along with significant reductions in the size of the device [4,5]. However, current device sizes and shapes are still not ideal and further research and technological advances are needed in this area. The size and shape of an implantable pulse generator is partially dependent on the physical size of the capacitors that are used to generate the defibrillating shock since they consume approximately 1/3 of the total volume of the ICD. The energy stored on a capacitor is proportional to its capacitance and to the square of its voltage (E= 1 2 CV2 ) but its size is usually primarily /1/ $35./ 21 The European Society of Cardiology

2 Single-capacitor defibrillation 279 dictated by its capacitance. Thus, lower capacitance, higher voltage capacitors would appear to be desirable, but traditional capacitor technologies suitable for ICDs are limited to maximum voltages between 35 and 39 Volts. Thus, ICDs have used a combination of capacitors connected in series to provide adequate voltage (7 8 V) for effective defibrillation. Defibrillating s from ICDs are commonly referred to as single-capacitor s, even though two or more physical capacitors are actually used. The term single capacitance would be more correct. Recently there has been a great deal of interest in improving the biphasic by searching for more optimum capacitance values. The majority of this research has focused on trying to reduce the energy requirements for defibrillation by shortening the pulse duration. This is most easily accomplished by reducing capacitance but at the expense of increasing the voltage and current requirements [6 11]. However, present and evolving capacitor technologies suggest a search for s that defibrillate with lower voltages and currents may be a more effective strategy for improving device size and shape. Lengthening the pulse duration, and reducing the tilt of the defibrillation are two ways that defibrillation voltages can be reduced [12 17]. One means to lengthen pulse duration is to increase the capacitance value used for defibrillation. Block et al. [18] showed a single 45 μf capacitor reduced voltage requirements by 45% with no change in energy requirements. However, 45 μf has a longer than is desired (>2 ms) and may provide more energy than necessary for future ICDs. In this study we investigated the use of an intermediate capacitance value (336 μf) and determined the importance of tilt on the successful application of this larger capacitance. Specifically, we compared the defibrillation efficacy of a standard 14 μf, 8% tilt biphasic to a new experimental biphasic generated from a single 336 μf capacitor at 6% tilt (1st series) and at 8% tilt (2nd series) in patients undergoing implantation of an implantable cardioverter-defibrillator. Methods Patients Fifty-four patients undergoing implantation of a new transvenous cardioverter-defibrillator system were included in the study (33 patients in the 1st series and 21 patients in the 2nd series). Written informed consent was obtained for all patients under a protocol approved by the Ethics Committee of each participating institution. Test procedure Under general anaesthesia, each patient underwent routine transvenous implantation of a single endocardial Volts Volts Volts ms ms (a) (b) (c) 14 µf, 8% tilt standard 6% tilt 1st series 8% tilt 2nd series ms Figure 1 Waveforms tested in this study. Each is shown with a peak voltage appropriate for a 1 Joule shock (stored energy). (a) Standard 14 μf, 8% tilt biphasic tested in all patients, (b) 6% tilt, 336 μf experimental tested in the first series of 33 patients, (c) 336 μf experimental tested in the second series of 21 patients. defibrillation lead (Endotak Model 95 or 1, Guidant Corporation, St Paul, MN, U.S.A.) using fluoroscopic guidance. A hot can emulator (Model 6967, Guidant Corporation) was placed in a subcutaneous pocket in the left pectoral region. Defibrillation efficacy was tested using a right ventricle (leading edge of 1st phase cathodic) to superior vena cava and hot can (leading edge of 1st phase anodic) lead configuration (RV SVC+Can). Two specifically modified external research defibrillators were used for delivery of the two biphasic s (ERD, Guidant Corporation). The standard 14 μf had an 8% overall tilt and a 6/4 phase1/phase2 duration ratio (Fig. 1a). In the 1st series

3 28 J. Brugada et al. of patients, the 336 μf experimental had an overall tilt of 6% with a 6/4 phase duration ratio (Fig. 1b). In the 2nd series of patients the 336 μf experimental had an overall tilt of 8% with a 6/4 phase duration ratio (Fig. 1c). In each patient, defibrillation efficacy using the standard was tested in parallel with the experimental using a randomized, interleaved, step-down defibrillation threshold protocol. Fibrillation was induced and allowed to continue for 1 s before application of the test shock. Only one test shock during each fibrillation episode was used for data analysis. Upon failure of any test shock, rescue shocks were delivered using a standard external cardioverter defibrillator. All testing began with 15 Joules. Testing continued by stepping down to successively lower energy levels (1, 5, 12, 1, 8, 5 J) until one failure was observed with each device or the lowest energy level of 5 Joules was attained. At least 5 min was allowed between each fibrillation episode. The defibrillation threshold (DFT), was defined as either the lowest energy that succeeded or 5 Joules. Upon completion of the defibrillation test protocol the hot can emulator was removed and the ICD implant was completed in a normal way. Data analysis and statistics All data were expressed as mean standard deviation. When quoting pairs of means, the data for the standard were always given first. Within each series of patients, paired Student s t-tests were used to compare the stored energy, delivered energy, peak voltage, peak current, system impedance and pulse duration at the defibrillation threshold between the standard and the experimental. Differences were considered statistically significant at the P< 5 level. Only patients who completed defibrillation threshold testing with both s were included in the analysis. Statistical analysis between 1st and 2nd series of patients was performed using a two-sample t-test (unpaired) on the corresponding data assuming equal variance in the two samples. Sample size was determined prospectively by assuming an α= 5 and a power (1 β) of 8% to detect a 2 Joule difference in energy requirements between the standard and experimental s. Assuming an expected mean energy of 9 5 Joules for the standard yielded a sample size of 41 patients for each series of patients. Results Patient demographics Fifty-six patients were enrolled in this study. The actual mean and standard deviation of the standard turned out to be significantly lower than assumed in the Table 1 Patient characteristics in 1st and 2nd series 1st series 2nd series n Male/female 27/6 2/1 Age (years) LVEF (%) Weight (kg) CAD 46% 48% Primary VF 37% 19% Amiodarone 18% 19% Sotalol 15% % CAD=coronary artery disease; LVEF=left ventricular ejection fraction; VF=ventricular fibrillation. sample size calculations so the first series of patients was stopped after 33 patients. These 33 patients had an actual power of 98% to resolve a 2 Joule difference in energy requirements. The second series of patients was stopped after 21 patients since a statistically significant result had been reached. The 21 patients in the second series had an actual power of 89% to resolve a 2 Joule difference in energy requirements. The clinical characteristics of each series of patients are shown in Table 1. No complications were observed during defibrillation threshold testing in any of the patients in the 1st or 2nd series. All patients were successfully implanted with an ICD system. 1st series of patients (standard vs 6% tilt experimental s) In the first series of 33 patients the single capacitor 336 μf experimental had a 6% tilt. The mean impedance of the lead system was 43 5 Ω with both standard and experimental s. At that impedance the experimental had a % longer pulse duration than the standard ( vs ms, P< 1). Mean stored energy for defibrillation was similar between the standard and the experimental s (8 3 3 vs Joules, P= 71). However, the mean delivered energy requirements were 15% lower with the experimental ( vs Joules, P= 27). Peak voltages were 35% lower with the experimental ( vs Volts, P< 1) and peak currents were 36% lower (8 1 6 vs Amps, P< 1). 2nd series of patients (standard vs 8% tilt experimental) In the second series of 21 patients the 336 μf single capacitor experimental had an 8% tilt. Mean impedance of the lead system was 46 5 Ω with both s. At this impedance, the experimental now had a much longer pulse duration than

4 Single-capacitor defibrillation 281 Series 1 33 patients Series 2 21 patients ± ± 2 8 P = ± ± 3 5 P = 1 Energy (J) µf, 6% tilt, 14 µf, Figure 2 Comparison of mean delivered energy defibrillation thresholds (DFTs) for the standard ( 14 μf; ) and experimental (336 μf) s (). The experimental had a 6% tilt in the first series of 33 patients, and an 8% tilt in the second series of 21 patients. the standard ( vs ms, P< 1). The experimental with 8% tilt required 38% more stored energy ( vs9 3 3 Joules, P= 2) and 44% more delivered energy ( vs Joules, P= 1) than the standard. While mean energy requirements increased, peak voltage (32 51 vs Volts, P< 1) and peak current ( vs Amps, P< 1) requirements were 23% lower using the experimental. second group of patients, it is likely that the true difference between 6% tilt and 8% tilt s was underestimated. To obtain a lesser assessment of the true differences between s the 336 μf experimental data was normalized by the control data within each patient. An unpaired t-test of the normalized data showed the 8% tilt had higher stored energy (41%, P< 1), higher delivered energy (62%, P< 1), and higher peak voltage (12%, P= 1). Cross-comparison of data from series 1 and 2 Unpaired comparisons between data from the two series of patients showed the standard had a 7% higher lead system impedance (P= 13), a 7% longer pulse duration and a 15% lower peak current (P= 12) in the second series of patients compared with the first series. The standard also showed a trend towards lower stored energy (P= 66), delivered energy (P= 58) (Fig. 2) and peak voltage (P= 8) in the second series of patients. Turning to the 336 μf experimental s, a 6% tilt was tested in series 1 and a 8% tilt was tested in series 2. The second series of patients had a 7% higher lead system impedance (P= 13) but a 92% longer (P< 1) (Fig. 3), due to the increase in tilt. Delivered energy at DFT was 35% higher with the 8% tilt (P= 7) (Fig. 2). Stored energy and peak voltage also trended higher but were not statistically significant. However, since the control s mean defibrillation requirements were lower in the Discussion In this study we tested patients undergoing defibrillator implantation to determine whether the use of a shock generated from single 336 μf capacitor would result in the same defibrillating efficacy as a standard 14 μf. Our data showed that a 6% tilt biphasic from a single 336 μf capacitor provided the same defibrillation efficacy as the standard with the additional benefit of significantly reduced peak voltage and current. However, an 8% tilt was significantly less effective than the 6% tilt when 336 μf was used. Biphasic strength-duration relationships: impact on device design It is well known that the shape of a defibrillating has a significant impact on defibrillation strength requirements. Strength-duration relationships for defibrillation with fixed tilt monophasic s

5 282 J. Brugada et al. 3 Series 1 33 patients 11 ± ± 1 4 P < 1 Series 2 21 patients 11 ± 1 7 ± 2 5 P <.1 Pulse duration (ms) µf, 6% tilt, 14 µf, Figure 3 Comparison of mean pulse durations at DFT for the standard ( 14 μf) and experimental (336 μf) s. The experimental had a 6% tilt in the first series of 33 patients, and an 8% tilt in the second series of 21 patients. Abbreviations as in Figure 2. are similar in shape to those known for pacing [12,14]. Defibrillation energies increase monotonically between 2 and 2 ms while voltage and current decrease to a stable level beyond 1 ms [14]. Therefore, it should be possible to reduce defibrillation energy by reducing capacitance to obtain a shorter duration, but at the expense of higher voltage and current [1,11]. However, several clinical studies showed that the combination of biphasic s with smaller capacitance, and the resulting shorter duration, failed to reduce energy despite a significant increase in voltage and current [6 9]. Thus, device size reduction by using smaller capacitance may not be as attractive as once believed, although it still appears to be feasible. Unlike monophasic s, strength-duration relationships for fixed-tilt biphasic s determined in animal models [13] showed that energy requirements are relatively constant between 3 and 1 ms and only increased at durations greater than 2 ms. Peak voltage and peak current decreased continuously over the entire range of pulse durations. Figure 4 illustrates defibrillation strength-duration relationships for both monophasic and biphasic s using data collected from a swine study [13]. Energy for a monophasic almost doubled when pulse duration was increased from 3 to 1 ms, while energy for a biphasic was unchanged over this same range of pulse durations. The biphasic energy requirement was only slightly higher at 2 ms and did not reach a statistically different value until 3 ms. The voltage-duration relationships show that voltage was reduced as pulse durations increased for both monophasic and biphasic s, but that biphasic s yielded a larger reduction in voltage at longer durations. This difference between monophasic and biphasic strength duration relationships explains the lack of energy reduction seen in the human biphasic studies with smaller capacitances. The biphasic strength-duration data also suggests that a device using a single, larger capacitance value might be feasible. Previous studies in humans [18] and in swine [19] showed that a single 45 5 μf capacitor and 6% tilt could be used for effective defibrillation. Peak voltages and currents were reduced by almost 45%, without any change in delivered energy compared with a standard 14 μf, 8% tilt biphasic. While these studies showed that defibrillation with a single larger capacitor was a feasible method to reduce device size, the animal studies [13] (Fig. 4) suggest 45 5 μf capacitance produces a longer than desired (>2 ms) and may store more energy than necessary for future ICDs ( 34 Joules). The present study used a single capacitor with an intermediate capacitance value and investigated the importance of tilt. Importance of tilt These new human data are the first to confirm the importance of tilt when using a larger capacitance. The results agree with the work of several other groups studying the effect of tilt with smaller capacitances. Swartz et al. [15] and Natale et al. [16] both showed lower tilts reduce energy and voltage requirements for defibrillation in humans. Data from a swine model also showed that tilts in the range of 5 6% performed best for a 336 μf [2]. The efficacy of the 8% tilt experimental was less than the

6 Single-capacitor defibrillation 283 Energy (J) Voltage (V) Capacitance for 8% tilt 78 µf 155 µf Monophasic Pulse duration (ms) Monophasic Biphasic Biphasic 311 µf 3 Capacitance for 8% tilt 78 µf 155 µf 311 µf Pulse duration (ms) 6% tilt despite the fact that more energy was delivered to the heart with the higher tilt. This may be due to spontaneous re-induction of ventricular fibrillation by long pulse durations with low trailing edge voltages [12]. While recent studies suggest large tilts (near 1%) may still be effective for certain types of s with small time constants and short pulse durations [21], the re-fibrillation phenomenon is probably still relevant to the larger time constant, longer pulse durations of the present study. These data as well as the cited references support the conclusion that delivered energy is not the most important determinant of defibrillation success, and in fact delivering too much energy, via a large tilt is actually counter-productive. Thus, optimization of tilt, as suggested by Swartz [15], Natale [16] and others [22 24], may lead to further improvements of ICDs, even though lower tilt s deliver less energy. Finally, optimization of tilt may be of particular importance in the case of s delivered from large capacitance since traditional 8% tilt s become much too long. 466 µf 466 µf Figure 4 Energy duration (a) and voltage duration (b) relationships for defibrillation with a 6% tilt monophasic and an 6/4 phase ratio, biphasic from a swine study [13]. The approximate capacitance required to produce an 8% tilt with a given pulse duration is also shown. Energy requirements for the monophasic almost double between 3 and 1 ms while there was no change with the biphasic. Only a 3 ms duration ( 466 μf) had a statistically significant increase in energy with the biphasic. Voltages decrease with increasing pulse duration for both monophasic and biphasic s, but biphasics have a larger voltage reduction that persist out to at least 2 ms pulse durations. 3 Clinical implications The size and shape of ICDs are dependent on the capacitors used to generate the defibrillating shocks since they consume approximately 1/3 of the total volume in an ICD. Traditional ICD capacitor technologies have maximum voltages between 35 and 39 Volts. Thus, to achieve the voltages required for defibrillation, two or more capacitors must be connected in series. However, connecting two capacitors in series reduces their effective capacitance by half, increases ICD size, and limits devices to a rather rectangular shape since that is the most efficient way to fit two capacitors. The rationale for the present study arises from the recognition that a somewhat larger single capacitor consumes less volume and is easier to package efficiently in a more patient-friendly shape than a two-capacitor system. For example, a single 336 μf aluminum electrolytic capacitor is approximately 33% smaller ( 7 cc) than two μf capacitors (1 μf when in series). The single capacitor would also allow a more rounded or elliptical shape for the ICD, rather than the usual rectangular shape that is most efficient for packaging two capacitors. Charge times should not be affected by changes in capacitance when used with the constant power charge circuits found in many modern ICDs. New developments finally made possible by single capacitor techniques must not compromise defibrillation safety. The data obtained in this study showed that future devices should be able to use the larger capacitance, single capacitor technology without jeopardizing defibrillation safety. As an example, one possible device might use a single 336 μf, 39 Volt capacitor. Such a device would store approximately 26 Joules and could be 7 8 cc, or more, smaller than a similar two capacitor device storing 31 Joules. In our study all patients were successfully defibrillated using a stored energy of 15 J or less. Applying the traditional safety margin of 1 Joules, all 54 patients could have received such a 26 Joule device. The success of commercially available devices with slightly lower energy outputs confirms that such a device should be successful for the vast majority of ICD patients []. Alternatively, the results of this and the other studies cited show that a wide range of capacitances with appropriately selected peak voltages could be employed to make an ICD with specifically designed size and energy specifications. Limitations While the results of this study are clear and an appropriate number of patients were tested to support statistically significant conclusions, the study was performed on a limited number of patients. Overall, both groups of patients tended to have a somewhat higher ejection fraction than a typical ICD patient population and no patients had DFTs greater than 15 Joules. The first series had a higher percentage of patients with primary ventricular fibrillation than might be typical and the first

7 284 J. Brugada et al. group had more patients on Sotalol than the second group (15% vs %). None of these factors should have affected the paired defibrillation results, but may have had some effect on the unpaired comparisons between the two series of patients. While no studies have shown a link between patient characteristics and DFTs, patients with lower ejection fractions and higher DFTs exist. Sotalol may also have some effect on DFTs. Thus, caution should be used in generalizing these results to a large ICD population. While the 336 μf, 6% tilt should be expected to perform at least as well as a standard ICD in all types of patients, a few patients may have higher defibrillation requirements that might not always allow implantation of a lower energy device. Conclusions Defibrillation using a single 336 μf capacitor with a 6% tilt required similar amounts of stored energy but significantly reduced delivered energy, peak voltage and peak current compared with the standard 14 μf that uses two capacitors in series. A single 336 μf capacitor and 6% tilt could be used to improve significantly the size and shape of future implantable defibrillators. References [1] Higgins SL, Lee RS, Kramer RL. Stability: an ICD detection criterion for discriminating atrial fibrillation from ventricular tachycardia. J Cardiovasc Electrophysiol 1995; 6: [2] Neuzner J, Pitschner HF, Schlepper M. Programmable VT detection enhancements in implantable cardioventer defibrillator therapy. Pacing Clin Electrophysiol 1995; 18: [3] Schaumann A, Von zur Muhlen F, Gonska, et al. Enhanced detection criteria in implantable cardioverted defibrillators to avoid inappropriate therapy. Am J Cardiol 1996; 78: [4] Klein H, Auricchio A, Huvelle E, et al. Initial experience with a new down-sized implantable cardioverter defibrillator. Am J Cardiol 1996; 78: [5] KenKnight BH, Jones BR, Thomas AC, et al. Technological advances in implantable cardioverter defibrillators before the year 2 and beyond. Am J Cardiol 1996; 78: [6] Poole JE, Kudenchuck PJ, Dolack GL, et al. A prospective randomized comparison in humans of 9 μf and 12 μf biphasic pulse defibrillation using a unipolar defibrillation system. J Cardiovasc Electrophysiol 1995; 6: [7] Bardy GH, Poole JE, Kudenchick PJ, et al. A prospective randomized comparison in humans of biphasic s 6 μf and 12 μf capacitance pulses using a unipolar defibrillation system. Circulation 1995; 91: [8] Block M, Hammel D, Böcker D, et al. Internal defibrillation with smaller capacitors: a prospective randomized cross-over comparison of defibrillation efficacy obtained with 9 μf and 1 μf capacitors in humans. J Cardiovasc Electrophysiol 1995; 6: [9] Swerdlow CD, Kass RM, Chen PS, et al. Effect of capacitor size and pathway resistance on defibrillation threshold for implantable defibrillators. Circulation 1994; 9: [1] Leonelli FM, Kroll MW, Brewer JE. Defibrillation thresholds are lower with smaller storage capacitors. Pacing Clin Electrophysiol 1995; 18: [11] Rist K, Tchou PJ, Mowrey K, et al. Smaller capacitors improve the biphasic. J Cardiovasc Electrophysiol 1994; 5: [12] Irnich W. The fundamental law of electrostimulation and its application to defibrillation. Pacing Clin Electrophysiol 199; 13: [13] Hahn SJ, Heil JE, Lin Y, et al. Defibrillation strengthduration relationships for fixed-tilt biphasic s. Circulation 1994; 9: I 228 (Abstract). [14] Wessale JL, Bourland JD, Tacker WA, et al. Bipolar catheter defibrillation in dogs using trapezoidal s of various tilts. J Electrocardiol 198; 13: [15] Swartz JF, Fletcher RD, Karasik PE. Optimization of biphasic s for nonthoracotomy defibrillation. Circulation 1993; 88: [16] Natale A, Sra J, Krum D, et al. Relative efficacy of different tilts with biphasic defibrillation in humans. Pacing Clin Electrophysiol 1996; 19: [17] Chapman PD, Wetherbee JN, Vetter JW, et al. Strengthduration curves of fixed pulse width variable tilt truncated exponential s for nonthoracotomy internal defibrillation in dogs. Pacing Clin Electrophysiol 1988; 11: [18] Block M, Hammel D, Bocker D, et al. Biphasic Defibrillation using a single capacitor with large capacitance: reduction of peak voltages and ICD device size. Pacing Clin Electrophysiol 1996; 19: [19] Hahn SJ, Heil JE, Lang DJ. Large capacitor defibrillation reduces peak voltage without increasing energies. Pacing Clin Electrophysiol 1995; 16: [2] Hahn SJ, Heil JE, Lang DJ. Defining the optimal tilt of a 336 μf biphasic defibrillation. Pacing Clin Electrophysiol 1997; 2: 1167 (Abstract). [21] Yamanouchi Y, Fishler MG, Mowrey KA, Wilkoff BL, Mazgalev TN, Tchou PJ. New approach to biphasic s for internal defibrillation: fully discharging capacitors. J Cardiovasc Electrophysiol 2; 11(8): [22] Walcott GP, Walker RG, Cates AW, et al. Choosing the optimal monophasic and biphasic s for ventricular defibrillation. J Cardiovasc Electrophysiol 1995; 6: [23] Kroll MW. A minimal mode of the single capacitor biphasic defibrillation. Pacing Clin Electrophysiol 1994; 17: [24] Hahn SJ, Heil JE, Lin Y, et al. Optimization of 9 μf biphasic s for ICD s using a theoretical model and central composite design of experiments. Pacing Clin Electrophysiol 1996; 19: 655 (Abstract). [] Neuzner J. Safety margins: lessons learned from the Low Energy Endotak Trial (LEET). Am J Cardiol 1996; 78: