- Email: [email protected]
Defibrillation probability and impedance change between shocks during resuscitation from out-of-hospital cardiac arrest
Resuscitation 80 (2009) 773–777
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Defibrillation probability and impedance change between shocks during resuscitation from out-of-hospital cardiac arrest夽 Robert G. Walker a,∗ , Rudolph W. Koster b , Charles Sun c,1 , George Moffat c,2 , Joseph Barger d , Pamela P. Dodson d , Fred W. Chapman a a
Physio-Control Inc. A Division of Medtronic, 11811 Willows Road NE, Redmond, WA 98052, USA Department of Cardiology, F3-239, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands British Columbia Ambulance Service, Victoria, BC, Canada d Contra Costa County EMS, 1340 Arnold Drive, Suite 126, Martinez, CA 94553, USA b c
a r t i c l e
i n f o
Article history: Received 3 October 2008 Received in revised form 21 March 2009 Accepted 1 April 2009 Keywords: Automated external defibrillator (AED) Cardiac arrest Defibrillation Transthoracic impedance
a b s t r a c t Objective: Technical data now gathered by automated external defibrillators (AEDs) allows closer evaluation of the behavior of defibrillation shocks administered during out-of-hospital cardiac arrest. We analyzed technical data from a large case series to evaluate the change in transthoracic impedance between shocks, and to assess the heterogeneity of the probability of successful defibrillation across the population. Methods: We analyzed a series of consecutive cases where AEDs delivered shocks to treat ventricular fibrillation (VF) during out-of-hospital cardiac arrest. Impedance measurements and VF termination efficacy were extracted from electronic records downloaded from biphasic AEDs deployed in three EMS systems. All patients received 200 J first shocks; second shocks were 200 J or 300 J, depending on local protocols. Results presented are median (25th, 75th percentiles). Results: Of 863 cases with defibrillation shocks, 467 contained multiple shocks because the first shock failed to terminate VF (n = 61) or VF recurred (n = 406). Defibrillation efficacy of subsequent shocks was significantly lower in patients that failed to defibrillate on first shock than in patients that did defibrillate on first shock (162/234 = 69% vs. 955/1027 = 93%; p < 0.0001). The failed VF terminations were distributed heterogeneously across the population; 5% of patients accounted for 71% of failed shocks. Shock impedance decreased by 1% [0%, 4%] and peak current increased by 1% [0%, 4%] between 200 J first and 200 J second shocks. Shock impedance decreased 4% [2%, 6%] and current increased 27% [25%, 29%] between 200 J first and 300 J second shocks. In all 499 pairs of same-energy consecutive shocks, impedance changed by less than 1% in 226 (45%), increased >1% in 124 (25%) and decreased >1% in 149 (30%). Conclusions: Impedance change between consecutive shocks is minimal and inconsistent. Therefore, to increase current of a subsequent shock requires an increase of the energy setting. Distribution of failed shocks is far from random. First shock defibrillation failure is often predictive of low efficacy for subsequent shocks. © 2009 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Cardiac arrest remains a leading cause of mortality. Of the hundreds of thousands of patients that suffer out-of-hospital cardiac arrest each year, many initially present with ventricular fibrillation
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2009.04.002. ∗ Corresponding author. Tel.: +1 425 867 4484; fax: +1 425 867 4462. E-mail address: [email protected] (R.G. Walker). University of British Columbia, and Island Medical Program, University of Victoria, 2127 Gourman Place, Victoria, BC V9B 6C5, Canada. 2 British Columbia Ambulance Service, PO Box 9600 Stn Prov Govt, Victoria, BC V8W 9P1, Canada. 1
0300-9572/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.resuscitation.2009.04.002
(VF) and many others experience VF during resuscitation attempts. Since VF must be terminated in these patients before spontaneous circulation can resume, defibrillation remains an important link in their chain of survival. Data on technical details of out-of-hospital defibrillation has, until recent years, been extremely difficult to gather and is correspondingly scarce. In the absence of such data, defibrillation practices have been based largely on presumptions informed by animal experiments and very limited sets of clinical data. However, modern defibrillators have made it feasible to gather technical data from large cohorts, allowing more careful evaluation of these presumptions. Until 2005, protocols for resuscitation from out-of-hospital cardiac arrest prescribed delivery of defibrillation countershocks in
774
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
“stacks” of up to three shocks prior to initiating or resuming CPR. Part of the rationale for delivering shocks in stacks had been an expected decrease in transthoracic impedance between consecutive shocks; a decrease in impedance caused by delivery of a first shock would result in higher current flow, and thus a higher probability of defibrillation success, for a subsequent shock.1 However, clinical evidence supporting this presumption has been limited to two small studies conducted with monophasic waveform shocks in settings other than out-of-hospital cardiac arrest.2,3 Recent guidelines have moved away from stacked shocks, but the fact that there was limited data supporting the original rationale for stacked shocks makes it desirable to determine whether the change in guidelines is further supported by the more extensive clinical data now available.4,5 Definitive evidence has been lacking over the years on whether, when one countershock fails to terminate VF, it is better to increase the dose or to keep the same dose for the next countershock. The rationale for repeating the same dose has been based, in part, on the expectation that impedance will decrease meaningfully from one shock to the next. Another part of the rationale is the presumption of “constant defibrillation probability”. This is the presumption that the same defibrillation probability obtained for a particular countershock dose across the general population will also be obtained in the subset of patients left in VF after a failed first shock or after failed first and second shocks.6 This presumption has not been evaluated. In the current investigation, we analyzed technical data collected by automated external defibrillators (AEDs) in a large consecutive case series of out-of-hospital cardiac arrests. We used these data to test two presumptions: the presumption that transthoracic impedance decreases between shocks, and the presumption of “constant defibrillation probability”.
impedance values and knowledge of the design of the energy delivery circuit of these AEDs. Two different types of impedance measurements were extracted from the AED records. The first type, “high-frequency impedance”, was an estimate made prior to the countershock by delivery of a very low intensity, high frequency (62.5 kHz) carrier signal across the thorax. This impedance estimation technique was first described by Geddes et al.,7 and has been shown to correlate well with impedance measurements made during shock discharge. The second type, “shock impedance”, was a direct measurement made during shock discharge. Due to non-linear behavior of the thorax, the shock impedance naturally decreases with increasing shock intensity.7 Therefore, we analyzed the change in shock impedance for the cases with two shocks at the same energy setting (200 J–200 J) separately from the cases with a higher energy setting for the second shock (200 J–300 J). On the other hand, the high-frequency impedance is measured before shock delivery and is not influenced by shock intensity. Therefore, data from all cases, regardless of second shock setting, could be combined for analysis of impedance change using this high-frequency impedance value. Descriptive statistics for impedance and current are reported as median [25th, 75th percentiles] unless otherwise indicated. Differences between proportions were tested with Chi-square statistics. 3. Results A total of 863 AED cases containing defibrillation shocks were available for analysis. VF was terminated by the initial 200 J shock in 802 (93%) cases. Due to failed VF termination with the first shock (n = 61) or refibrillation (n = 406), 467 cases contained at least two defibrillation shocks, and thus could be analyzed for impedance and defibrillation trends.
2. Methods 3.1. Impedance change between consecutive shocks A retrospective review was performed on electronic records downloaded from biphasic waveform AEDs (LIFEPAK 500, Medtronic Inc.) deployed with first-responding BLS teams in three EMS systems. Ethical approval for this retrospective review was obtained in each of the three participating systems. For each system, we analyzed a consecutive series of downloaded records where one or more shocks were delivered for treatment of ventricular fibrillation (VF) during out-of-hospital cardiac arrest. All AEDs were configured per local protocols to an energy sequence of either 200 J–200 J–360 J, or 200 J–300 J–360 J. Thus first shocks were always delivered at 200 J, and second shocks were delivered at either 200 J or 300 J depending on the configured energy protocol. In accordance with the then current guidelines, stacks of up to three shocks were delivered prior to initiating or resuming CPR. Electronic device records were reviewed to determine VF termination efficacy and extract impedance values for each delivered shock. Defibrillation efficacy was determined from manual review of ECG records, and was defined per commonly accepted convention as termination of VF for at least 5 s after shock delivery.1 Peak current was calculated for all shocks based on the recorded
Among the 467 cases with at least two shocks, the median shock impedance for the initial 200 J shock was 86 [73,103], and the median peak current was 17.4 A [14.8, 20.1]. In patients receiving second shocks at the same energy as the first shock (200 J), shock impedance decreased by a median of 1% and peak current thus increased by 1% between first and second shocks (Table 1). In patients receiving larger (300 J) second shocks, second-shock impedance decreased only 4% while second-shock peak current increased 27%, primarily as a consequence of the higher energy setting. Across all 467 cases with at least two shocks, the median change in high-frequency impedance between first and second shocks was 0% [−3%, 2%]. The median change between first and second-shock impedance was not influenced by whether the second shock was delivered immediately after the first as a part of a shock stack (0% [−3%, 2%]) or delivered later after an interval of CPR (0% [−3%, 2%]). Neither the change in shock impedance nor the change in high-frequency impedance was correlated to the impedance of the first shock (shock impedance R2 = 0.05; high-frequency impedance R2 = 0.04).
Table 1 Impedance and peak current change between first and second shocks.
Decrease in shock impedance % shock impedance decrease Increase in peak current % peak current increase Decrease in high-frequency impedance % high-frequency impedance decrease
Second shock at same energy (200 J, n = 109)
Second shock at higher energy (300 J, n = 358)
1 [0, 3] 1% [0%, 4%] 0.2 A [0.0, 0.7] 1% [0%, 4%]
3 [2, 5] 4% [2%, 6%] 4.7 A [3.7, 5.5] 27% [25%, 29%] 0 [−2, 2] 0% [−3%, 2%]
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
775
Table 2 VF termination probability for subsequent shocks stratified by first shock outcome, with 95%, two-sided confidence limits.
VF termination rate for second shock at 200 J VF termination rate for second shock at 300 J VF termination rate for all subsequent shocks
Patients in whom the first shock terminated VF (n = 406)
Patients in whom the first shock failed to terminate VF (n = 61)
92% (96/104) [85–97%] 95% (286/302) [92–97%] 93% (955/1027) [91–94%]
40% (2/5) [5–85%] 70% (39/56) [56–81%] 69% (162/234) [63–75%]
Table 3 Impedance change, peak current and defibrillation probability for the first shock at each energy level, among the 236 cases containing at least one shock at each of the three energy levels. Impedance change values are relative to the first shock at the prior energy level.
Change in high-frequency impedance Change in shock impedance Peak current VF termination rate
200 J
300 J
360 J
– – 17.3 A [14.9, 19.7] 82% (194/236)
0 [+1, −3] −3 [−2, −5] 21.8 A [18.8, 25.6] 86% (202/236)
0 [0, −1] −1 [0, −2] 24.0 A [20.2, 28.2] 90% (212/236)
Shock impedance changes were similar for third and subsequent shocks. Overall, 499 instances of same-energy consecutive shock pairs occurred within the first six delivered shocks. In 226 (45%) of these consecutive shock pairs, shock impedance changed by less than 1%. When shock impedance changed by more than 1%, it was nearly as likely to increase (n = 124) as to decrease (n = 149) (Fig. 1). 3.2. Defibrillation probability Subsequent shock defibrillation efficacy – the efficacy of all shocks after the first shock – was significantly lower in the 61 cases where the first shock failed to terminate VF (162 of 234 = 69%) than in the 802 cases where the first shock successfully terminated VF (955 of 1027 = 93%, p < 0.0001) (Table 2). Subsequent shock defibrillation efficacy in the 61 cases with a failed first shock was also significantly lower than first shock defibrillation efficacy in the overall population (162 of 234 = 69% vs. 802 of 863 = 93%, p < 0.0001). Overall, VF was terminated by 1919 of 2124 shocks (90%) delivered in the 863 cases. The 205 failed VF terminations were distributed heterogeneously across the population; some patients were more-difficult-to-defibrillate (Fig. 2). For example, 71% of all failed shocks can be accounted for by only 45 patients (5% of the total patient population). Among these 45 cases, the first shock VF termination rate was 36%. First shock impedance did not differ between this subset of 45 cases (88 ± 21 ) and the other 818 cases (88 ± 26 ). Only 12% of all patients, and 22% of those patients receiving multiple shocks, experienced a failed shock. As a consequence of the configured energy protocol and various combinations of recurrent and persistent VF, a total of 236 cases
Fig. 1. Shock impedance change for 499 same-energy consecutive shock pairs.
Fig. 2. Distribution of failed shocks across the population of 863 patients (solid black line). For example (dashed grey lines), 45 patients (5%) accounted for 145 of the 205 shocks (71%) that failed to terminate VF.
contained at least one shock at each of the three possible energy settings. In these 236 cases, impedance changed negligibly for the first shock delivered at each successive energy setting. However, defibrillation probability increased in parallel with the peak current of each energy dose (Table 3). 4. Discussion We used technical defibrillation data from a large cohort of prehospital cardiac arrests to evaluate two common presumptions about defibrillation during out-of-hospital cardiac arrest: (1) that impedance decreases between consecutive defibrillation shocks, and (2) that defibrillation probability is the same in the subset of patients with initially unsuccessful shocks as it is in the overall population. Both of these presumptions have been important components in the rationale for two defibrillation practices: delivering shocks in stacks, and, in defibrillators capable of delivering shocks of higher energy, delivering a second shock at the same energy when a first shock failed to terminate VF. Our observations refute both of these presumptions. In this large patient cohort treated with biphasic shocks, impedance change between consecutive shocks was minimal and inconsistent. Defibrillation probability was highly heterogeneous, and was substantially lower for subsequent shocks in patients with a failed first shock than in patients with a successful first shock.
776
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
4.1. Impedance change between consecutive shocks
4.2. Defibrillation probability
The observation of a decrease in transthoracic impedance with repeated countershocks was first reported in animal studies involving shock delivery during spontaneous circulation,8 or during serial episodes of brief VF followed by defibrillation to a perfusing rhythm.9 Subsequent clinical observations of the phenomenon that have provided the basis for relevant statements in resuscitation guidelines 1,4,5 are limited to a study of 28 patients undergoing elective cardioversion,3 and a study describing 10 patients undergoing emergency in-hospital defibrillation.2 Several aspects of these prior studies differ from our present investigation, including the use of monophasic shocks in all of these prior studies, varying impedance measurement techniques and, in some instances, use of hard paddles rather than adhesive pads. The difference that is likely to be the most salient, however, is that these earlier observations come from settings other than prolonged out-of-hospital cardiac arrest. Data from a more recent experimental study of prolonged VF are consistent with our present clinical findings. Niemann et al.10 observed no decrease in impedance between consecutive monophasic or biphasic shocks delivered after 5 min of unsupported VF in swine. They postulated that critical hypoperfusion of the skin and subcutaneous tissue during prolonged arrest and CPR precludes the hyperemia and edema beneath the shocking electrodes previously proposed as primary mechanisms of impedance decline with repeated monophasic shocks during normal perfusing rhythms.11 Also consistent with our findings are data from a clinical study of recurrent fibrillation during prehospital advanced life support care, which found that impedance in a 36-patient subset receiving first and second 200 J biphasic shocks did not change meaningfully between shocks.12 In another recent clinical study, Deakin et al.13 reported a small but statistically significant decrease in impedance between escalating biphasic shocks delivered during elective cardioversion. While the method of impedance measurement was not reported, the results are consistent with the modest decrease in shock impedance with increasing shock intensity first described by Geddes et al.7 , and also observed in our data. In our analysis of 863 AED records, impedance change between consecutive shocks was found to be negligible, particularly for consecutive same-energy shocks. For those patients that received a second 200 J shock following their first 200 J shock, the minimal impedance change resulted in a negligible median peak current increase of less than 200 mA. When the analysis was expanded to include all same-energy consecutive shocks, impedance remained essentially unchanged in close to half the instances, and was nearly as likely to exhibit a modest increase as a modest decrease in the other half of the instances (Fig. 1). Thus these data contradict the common presumption that decline in impedance is an operative mechanism by which shock current, and thus defibrillation probability, might increase for a subsequent shock when a previous shock at that same energy setting has failed to terminate VF. For those patients that received a 200 J shock followed by a 300 J shock, a moderate (27%) increase in second-shock peak current was observed. As physics would predict,7 the impedance decreased a small amount (4%) when dose was increased. The much larger increase in current was thus due predominantly to the higher capacitor voltage associated with the higher energy setting. If impedance did decrease meaningfully between shocks, shock current would indeed increase and move the defibrillation operating point up the probability-of-success curve. However, in this large patient cohort, the only mechanism by which the current of subsequent shocks increased appreciably was escalation of the shock energy dose.
As shown in Table 3, we observed an increase in defibrillation probability with increasing shock dose among the subset of patients who received shocks at each of the three energy levels. This increase in defibrillation probability with increasing dose is predicted by the well-known concept of the defibrillation probability-ofsuccess curve – a fundamental property of defibrillation14 – and is also consistent with the results of a recent randomized clinical trial.15 Although this fundamental dose–response property of defibrillation implies only that a higher dose has a relative higher defibrillation probability within a given patient or population, a common mis-extrapolation has been the notion that any given energy has a fixed, constant probability to defibrillate, even across different populations. In particular, it has been suggested that a shock dose observed to have a given VF termination probability in a large population maintains that same VF termination probability for second shocks delivered to the subset of patients left in VF after a failed first shock, and even for third shocks delivered to the subset of patients left in VF after failed first and second shocks.6 Inherent in this suggestion is the presumption that defibrillation probability is relatively homogenous throughout a population. Stated differently, this notion presumes that failure of a shock to terminate VF is predominantly a randomly distributed event – both longitudinally in a patient as well as across a population – and therefore that VF termination probability for second and subsequent shocks is mostly independent of the actual success or failure of the first shock. Contrary to the presumption of “constant defibrillation probability”, we found that VF termination failure was far from randomly distributed. A small subset of the patients accounted for most of the failed shocks, and in these patients, the first shock VF termination rate was substantially lower than it was in the overall population. Across all patients the probability of VF termination for subsequent shocks was significantly lower in cases with a failed first shock (69%) than in cases with a successful first shock (93%). Additionally, in direct contradiction to the common presumption,6 subsequent shock VF termination success in cases with a failed first shock was significantly lower than the first shock VF termination rate across the overall population. Thus our data indicate that, rather than representing a mostly random event, first shock defibrillation failure is often indicative of a more-difficult-to-defibrillate patient, and may serve to select out a subgroup of patients in which a given shock dose provides a significantly lower defibrillation probability than in the overall population. In our data, this subset of more-difficult-to-defibrillate patients was not distinguishable by impedance; further research is needed to determine the characteristics of these patients and the causes of defibrillation failure. 4.3. Limitations Our observations reflect data collected from a large patient cohort treated with one specific defibrillator model and shock waveform. Other defibrillators currently on the market, and used in prior studies of impedance behavior between consecutive shocks, feature a variety of different methods of measuring impedance, as well as a variety of shock waveforms. To date, ours is the only large clinical evaluation of impedance behavior between consecutive shocks specifically in the setting of out-of-hospital cardiac arrest, and our findings corroborate those from the only experimental evaluation of this behavior in a comparable cardiac arrest model.10 Given the present understanding of the underlying mechanisms,10,11,14 as well as the concordance of our clinical findings with experimental data using other monophasic and
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
biphasic shock waveforms,10 there is no basis to believe that the actual impedance behavior between consecutive shocks during cardiac arrest resuscitation would differ for any available defibrillator. However, considering the variety of methods of measuring impedance implemented in different devices, it would be important for any future studies of this phenomenon to specify the method(s) of impedance measurement, and if possible include a direct measure of impedance during shock discharge, to allow meaningful comparison to existing data. 5. Conclusions The findings of this large consecutive case series analysis directly contradict two common presumptions about defibrillation during out-of-hospital cardiac arrest, and support recent changes to resuscitation guidelines. Our finding that repeated shocks do not cause either a meaningful decrease in transthoracic impedance, or a meaningful impedance-mediated increase in shock current, adds support to the move away from stacked shocks made in the 2005 guidelines.4,5 Furthermore, the absence of any appreciable increase in shock current when a second shock at the same setting is administered supports the guidelines statement that if the first shock fails to terminate VF and the defibrillator is capable of delivering higher dose shocks, it is rational to increase the energy for subsequent shocks.5 Thus in terms of successful termination of VF, our findings indicate that the true benefit to be derived from the practices of stacking shocks and delivering second shocks at the same energy as failed first shocks are likely to be less than expected. This reduction in the VF termination yield that should reasonably be expected from such practices, weighed against the substantial cost of each defibrillation attempt in terms of CPR interruption time,16,17 supports the move away from stacked shocks and, for the many defibrillators that provide two or more shock doses for treatment of cardiac arrest, encourages ongoing re-evaluation of the practice of repeating sub-maximal dose shocks when VF is not terminated by the previous shock. Conflict of interest statement Robert G. Walker is a current employee of Physio-Control. Rudolph W. Koster was a paid medical advisor for Physio-Control until 2007, and has received research grants from Physio-Control. Fred W. Chapman is a current employee of Physio-Control. Authors Sun, Moffat, Barger, and Dodson have no conflicts to declare.
777
Acknowledgements We thank the EMTs, paramedics, and data collection staff at each of the sites for downloading the defibrillator data to enable our analysis. References 1. Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 4: the automated external defibrillator: key link in the chain of survival. The American Heart Association in Collaboration with the International Liaison Committee on Resuscitation. Circulation 2000;102:I-60–76. 2. Kerber RE, Grayzel J, Hoyt R, Marcus M, Kennedy J. Transthoracic resistance in human defibrillation. Influence of body weight, chest size, serial shocks, paddle size and paddle contact pressure. Circulation 1981;63:676–82. 3. Sirna SJ, Ferguson DW, Charbonnier F, Kerber RE. Factors affecting transthoracic impedance during electrical cardioversion. Am J Cardiol 1988;62:1048–52. 4. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2005;112:IV-1–203. 5. Deakin CD, Nolan JP. European Resuscitation Council guidelines for resuscitation 2005. Section 3. Electrical therapies: automated external defibrillators, defibrillation, cardioversion and pacing. Resuscitation 2005;67(Suppl. 1):S25–37. 6. Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 6: advanced cardiovascular life support: section 2: defibrillation. The American Heart Association in collaboration with the International Liaison Committee on Resuscitation. Circulation 2000;102:I-90–4. 7. Geddes LA, Tacker Jr WA, Schoenlein W, Minton M, Grubbs S, Wilcox P. The prediction of the impedance of the thorax to defibrillating current. Med Instrum 1976;10:159–62. 8. Dahl CF, Ewy GA, Ewy MD, Thomas ED. Transthoracic impedance to direct current discharge: effect of repeated countershocks. Med Instrum 1976;10:151–4. 9. Geddes LA, Tacker WA, Cabler P, Chapman R, Rivera R, Kidder H. The decrease in transthoracic impedance during successive ventricular defibrillation trials. Med Instrum 1975;9:179–80. 10. Niemann JT, Garner D, Lewis RJ. Transthoracic impedance does not decrease with rapidly repeated countershocks in a swine cardiac arrest model. Resuscitation 2003;56:91–5. 11. Sirna SJ, Kieso RA, Fox-Eastham KJ, Seabold J, Charbonnier F, Kerber RE. Mechanisms responsible for decline in transthoracic impedance after DC shocks. Am J Physiol 1989;257:H1180–3. 12. Koster RW, Walker RG, Chapman FW. Recurrent ventricular fibrillation during advanced life support care of patients with prehospital cardiac arrest. Resuscitation 2008;78:252–7. 13. Deakin CD, Ambler JJ, Shaw S. Changes in transthoracic impedance during sequential biphasic defibrillation. Resuscitation 2008;78:141–5. 14. Tacker WA. Fibrillation causes and criteria for defibrillation. In: Tacker WA, editor. Defibrillation of the heart: ICDs, AEDs, and manual. St. Louis, MO: Mosby; 1994. p. 1–14. 15. Stiell IG, Walker RG, Nesbitt LP, et al. BIPHASIC trial: a randomized comparison of fixed lower versus escalating higher energy levels for defibrillation in out-ofhospital cardiac arrest. Circulation 2007;115:1511–7. 16. Berg RA, Hilwig RW, Kern KB, Sanders AB, Xavier LC, Ewy GA. Automated external defibrillation versus manual defibrillation for prolonged ventricular fibrillation: lethal delays of chest compressions before and after countershocks. Ann Emerg Med 2003;42:458–67. 17. Kramer-Johansen J, Edelson DP, Abella BS, Becker LB, Wik L, Steen PA. Pauses in chest compression and inappropriate shocks: a comparison of manual and semi-automatic defibrillation attempts. Resuscitation 2007;73:212–20.
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical paper
Defibrillation probability and impedance change between shocks during resuscitation from out-of-hospital cardiac arrest夽 Robert G. Walker a,∗ , Rudolph W. Koster b , Charles Sun c,1 , George Moffat c,2 , Joseph Barger d , Pamela P. Dodson d , Fred W. Chapman a a
Physio-Control Inc. A Division of Medtronic, 11811 Willows Road NE, Redmond, WA 98052, USA Department of Cardiology, F3-239, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands British Columbia Ambulance Service, Victoria, BC, Canada d Contra Costa County EMS, 1340 Arnold Drive, Suite 126, Martinez, CA 94553, USA b c
a r t i c l e
i n f o
Article history: Received 3 October 2008 Received in revised form 21 March 2009 Accepted 1 April 2009 Keywords: Automated external defibrillator (AED) Cardiac arrest Defibrillation Transthoracic impedance
a b s t r a c t Objective: Technical data now gathered by automated external defibrillators (AEDs) allows closer evaluation of the behavior of defibrillation shocks administered during out-of-hospital cardiac arrest. We analyzed technical data from a large case series to evaluate the change in transthoracic impedance between shocks, and to assess the heterogeneity of the probability of successful defibrillation across the population. Methods: We analyzed a series of consecutive cases where AEDs delivered shocks to treat ventricular fibrillation (VF) during out-of-hospital cardiac arrest. Impedance measurements and VF termination efficacy were extracted from electronic records downloaded from biphasic AEDs deployed in three EMS systems. All patients received 200 J first shocks; second shocks were 200 J or 300 J, depending on local protocols. Results presented are median (25th, 75th percentiles). Results: Of 863 cases with defibrillation shocks, 467 contained multiple shocks because the first shock failed to terminate VF (n = 61) or VF recurred (n = 406). Defibrillation efficacy of subsequent shocks was significantly lower in patients that failed to defibrillate on first shock than in patients that did defibrillate on first shock (162/234 = 69% vs. 955/1027 = 93%; p < 0.0001). The failed VF terminations were distributed heterogeneously across the population; 5% of patients accounted for 71% of failed shocks. Shock impedance decreased by 1% [0%, 4%] and peak current increased by 1% [0%, 4%] between 200 J first and 200 J second shocks. Shock impedance decreased 4% [2%, 6%] and current increased 27% [25%, 29%] between 200 J first and 300 J second shocks. In all 499 pairs of same-energy consecutive shocks, impedance changed by less than 1% in 226 (45%), increased >1% in 124 (25%) and decreased >1% in 149 (30%). Conclusions: Impedance change between consecutive shocks is minimal and inconsistent. Therefore, to increase current of a subsequent shock requires an increase of the energy setting. Distribution of failed shocks is far from random. First shock defibrillation failure is often predictive of low efficacy for subsequent shocks. © 2009 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Cardiac arrest remains a leading cause of mortality. Of the hundreds of thousands of patients that suffer out-of-hospital cardiac arrest each year, many initially present with ventricular fibrillation
夽 A Spanish translated version of the summary of this article appears as Appendix in the final online version at doi:10.1016/j.resuscitation.2009.04.002. ∗ Corresponding author. Tel.: +1 425 867 4484; fax: +1 425 867 4462. E-mail address: [email protected] (R.G. Walker). University of British Columbia, and Island Medical Program, University of Victoria, 2127 Gourman Place, Victoria, BC V9B 6C5, Canada. 2 British Columbia Ambulance Service, PO Box 9600 Stn Prov Govt, Victoria, BC V8W 9P1, Canada. 1
0300-9572/$ – see front matter © 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.resuscitation.2009.04.002
(VF) and many others experience VF during resuscitation attempts. Since VF must be terminated in these patients before spontaneous circulation can resume, defibrillation remains an important link in their chain of survival. Data on technical details of out-of-hospital defibrillation has, until recent years, been extremely difficult to gather and is correspondingly scarce. In the absence of such data, defibrillation practices have been based largely on presumptions informed by animal experiments and very limited sets of clinical data. However, modern defibrillators have made it feasible to gather technical data from large cohorts, allowing more careful evaluation of these presumptions. Until 2005, protocols for resuscitation from out-of-hospital cardiac arrest prescribed delivery of defibrillation countershocks in
774
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
“stacks” of up to three shocks prior to initiating or resuming CPR. Part of the rationale for delivering shocks in stacks had been an expected decrease in transthoracic impedance between consecutive shocks; a decrease in impedance caused by delivery of a first shock would result in higher current flow, and thus a higher probability of defibrillation success, for a subsequent shock.1 However, clinical evidence supporting this presumption has been limited to two small studies conducted with monophasic waveform shocks in settings other than out-of-hospital cardiac arrest.2,3 Recent guidelines have moved away from stacked shocks, but the fact that there was limited data supporting the original rationale for stacked shocks makes it desirable to determine whether the change in guidelines is further supported by the more extensive clinical data now available.4,5 Definitive evidence has been lacking over the years on whether, when one countershock fails to terminate VF, it is better to increase the dose or to keep the same dose for the next countershock. The rationale for repeating the same dose has been based, in part, on the expectation that impedance will decrease meaningfully from one shock to the next. Another part of the rationale is the presumption of “constant defibrillation probability”. This is the presumption that the same defibrillation probability obtained for a particular countershock dose across the general population will also be obtained in the subset of patients left in VF after a failed first shock or after failed first and second shocks.6 This presumption has not been evaluated. In the current investigation, we analyzed technical data collected by automated external defibrillators (AEDs) in a large consecutive case series of out-of-hospital cardiac arrests. We used these data to test two presumptions: the presumption that transthoracic impedance decreases between shocks, and the presumption of “constant defibrillation probability”.
impedance values and knowledge of the design of the energy delivery circuit of these AEDs. Two different types of impedance measurements were extracted from the AED records. The first type, “high-frequency impedance”, was an estimate made prior to the countershock by delivery of a very low intensity, high frequency (62.5 kHz) carrier signal across the thorax. This impedance estimation technique was first described by Geddes et al.,7 and has been shown to correlate well with impedance measurements made during shock discharge. The second type, “shock impedance”, was a direct measurement made during shock discharge. Due to non-linear behavior of the thorax, the shock impedance naturally decreases with increasing shock intensity.7 Therefore, we analyzed the change in shock impedance for the cases with two shocks at the same energy setting (200 J–200 J) separately from the cases with a higher energy setting for the second shock (200 J–300 J). On the other hand, the high-frequency impedance is measured before shock delivery and is not influenced by shock intensity. Therefore, data from all cases, regardless of second shock setting, could be combined for analysis of impedance change using this high-frequency impedance value. Descriptive statistics for impedance and current are reported as median [25th, 75th percentiles] unless otherwise indicated. Differences between proportions were tested with Chi-square statistics. 3. Results A total of 863 AED cases containing defibrillation shocks were available for analysis. VF was terminated by the initial 200 J shock in 802 (93%) cases. Due to failed VF termination with the first shock (n = 61) or refibrillation (n = 406), 467 cases contained at least two defibrillation shocks, and thus could be analyzed for impedance and defibrillation trends.
2. Methods 3.1. Impedance change between consecutive shocks A retrospective review was performed on electronic records downloaded from biphasic waveform AEDs (LIFEPAK 500, Medtronic Inc.) deployed with first-responding BLS teams in three EMS systems. Ethical approval for this retrospective review was obtained in each of the three participating systems. For each system, we analyzed a consecutive series of downloaded records where one or more shocks were delivered for treatment of ventricular fibrillation (VF) during out-of-hospital cardiac arrest. All AEDs were configured per local protocols to an energy sequence of either 200 J–200 J–360 J, or 200 J–300 J–360 J. Thus first shocks were always delivered at 200 J, and second shocks were delivered at either 200 J or 300 J depending on the configured energy protocol. In accordance with the then current guidelines, stacks of up to three shocks were delivered prior to initiating or resuming CPR. Electronic device records were reviewed to determine VF termination efficacy and extract impedance values for each delivered shock. Defibrillation efficacy was determined from manual review of ECG records, and was defined per commonly accepted convention as termination of VF for at least 5 s after shock delivery.1 Peak current was calculated for all shocks based on the recorded
Among the 467 cases with at least two shocks, the median shock impedance for the initial 200 J shock was 86 [73,103], and the median peak current was 17.4 A [14.8, 20.1]. In patients receiving second shocks at the same energy as the first shock (200 J), shock impedance decreased by a median of 1% and peak current thus increased by 1% between first and second shocks (Table 1). In patients receiving larger (300 J) second shocks, second-shock impedance decreased only 4% while second-shock peak current increased 27%, primarily as a consequence of the higher energy setting. Across all 467 cases with at least two shocks, the median change in high-frequency impedance between first and second shocks was 0% [−3%, 2%]. The median change between first and second-shock impedance was not influenced by whether the second shock was delivered immediately after the first as a part of a shock stack (0% [−3%, 2%]) or delivered later after an interval of CPR (0% [−3%, 2%]). Neither the change in shock impedance nor the change in high-frequency impedance was correlated to the impedance of the first shock (shock impedance R2 = 0.05; high-frequency impedance R2 = 0.04).
Table 1 Impedance and peak current change between first and second shocks.
Decrease in shock impedance % shock impedance decrease Increase in peak current % peak current increase Decrease in high-frequency impedance % high-frequency impedance decrease
Second shock at same energy (200 J, n = 109)
Second shock at higher energy (300 J, n = 358)
1 [0, 3] 1% [0%, 4%] 0.2 A [0.0, 0.7] 1% [0%, 4%]
3 [2, 5] 4% [2%, 6%] 4.7 A [3.7, 5.5] 27% [25%, 29%] 0 [−2, 2] 0% [−3%, 2%]
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
775
Table 2 VF termination probability for subsequent shocks stratified by first shock outcome, with 95%, two-sided confidence limits.
VF termination rate for second shock at 200 J VF termination rate for second shock at 300 J VF termination rate for all subsequent shocks
Patients in whom the first shock terminated VF (n = 406)
Patients in whom the first shock failed to terminate VF (n = 61)
92% (96/104) [85–97%] 95% (286/302) [92–97%] 93% (955/1027) [91–94%]
40% (2/5) [5–85%] 70% (39/56) [56–81%] 69% (162/234) [63–75%]
Table 3 Impedance change, peak current and defibrillation probability for the first shock at each energy level, among the 236 cases containing at least one shock at each of the three energy levels. Impedance change values are relative to the first shock at the prior energy level.
Change in high-frequency impedance Change in shock impedance Peak current VF termination rate
200 J
300 J
360 J
– – 17.3 A [14.9, 19.7] 82% (194/236)
0 [+1, −3] −3 [−2, −5] 21.8 A [18.8, 25.6] 86% (202/236)
0 [0, −1] −1 [0, −2] 24.0 A [20.2, 28.2] 90% (212/236)
Shock impedance changes were similar for third and subsequent shocks. Overall, 499 instances of same-energy consecutive shock pairs occurred within the first six delivered shocks. In 226 (45%) of these consecutive shock pairs, shock impedance changed by less than 1%. When shock impedance changed by more than 1%, it was nearly as likely to increase (n = 124) as to decrease (n = 149) (Fig. 1). 3.2. Defibrillation probability Subsequent shock defibrillation efficacy – the efficacy of all shocks after the first shock – was significantly lower in the 61 cases where the first shock failed to terminate VF (162 of 234 = 69%) than in the 802 cases where the first shock successfully terminated VF (955 of 1027 = 93%, p < 0.0001) (Table 2). Subsequent shock defibrillation efficacy in the 61 cases with a failed first shock was also significantly lower than first shock defibrillation efficacy in the overall population (162 of 234 = 69% vs. 802 of 863 = 93%, p < 0.0001). Overall, VF was terminated by 1919 of 2124 shocks (90%) delivered in the 863 cases. The 205 failed VF terminations were distributed heterogeneously across the population; some patients were more-difficult-to-defibrillate (Fig. 2). For example, 71% of all failed shocks can be accounted for by only 45 patients (5% of the total patient population). Among these 45 cases, the first shock VF termination rate was 36%. First shock impedance did not differ between this subset of 45 cases (88 ± 21 ) and the other 818 cases (88 ± 26 ). Only 12% of all patients, and 22% of those patients receiving multiple shocks, experienced a failed shock. As a consequence of the configured energy protocol and various combinations of recurrent and persistent VF, a total of 236 cases
Fig. 1. Shock impedance change for 499 same-energy consecutive shock pairs.
Fig. 2. Distribution of failed shocks across the population of 863 patients (solid black line). For example (dashed grey lines), 45 patients (5%) accounted for 145 of the 205 shocks (71%) that failed to terminate VF.
contained at least one shock at each of the three possible energy settings. In these 236 cases, impedance changed negligibly for the first shock delivered at each successive energy setting. However, defibrillation probability increased in parallel with the peak current of each energy dose (Table 3). 4. Discussion We used technical defibrillation data from a large cohort of prehospital cardiac arrests to evaluate two common presumptions about defibrillation during out-of-hospital cardiac arrest: (1) that impedance decreases between consecutive defibrillation shocks, and (2) that defibrillation probability is the same in the subset of patients with initially unsuccessful shocks as it is in the overall population. Both of these presumptions have been important components in the rationale for two defibrillation practices: delivering shocks in stacks, and, in defibrillators capable of delivering shocks of higher energy, delivering a second shock at the same energy when a first shock failed to terminate VF. Our observations refute both of these presumptions. In this large patient cohort treated with biphasic shocks, impedance change between consecutive shocks was minimal and inconsistent. Defibrillation probability was highly heterogeneous, and was substantially lower for subsequent shocks in patients with a failed first shock than in patients with a successful first shock.
776
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
4.1. Impedance change between consecutive shocks
4.2. Defibrillation probability
The observation of a decrease in transthoracic impedance with repeated countershocks was first reported in animal studies involving shock delivery during spontaneous circulation,8 or during serial episodes of brief VF followed by defibrillation to a perfusing rhythm.9 Subsequent clinical observations of the phenomenon that have provided the basis for relevant statements in resuscitation guidelines 1,4,5 are limited to a study of 28 patients undergoing elective cardioversion,3 and a study describing 10 patients undergoing emergency in-hospital defibrillation.2 Several aspects of these prior studies differ from our present investigation, including the use of monophasic shocks in all of these prior studies, varying impedance measurement techniques and, in some instances, use of hard paddles rather than adhesive pads. The difference that is likely to be the most salient, however, is that these earlier observations come from settings other than prolonged out-of-hospital cardiac arrest. Data from a more recent experimental study of prolonged VF are consistent with our present clinical findings. Niemann et al.10 observed no decrease in impedance between consecutive monophasic or biphasic shocks delivered after 5 min of unsupported VF in swine. They postulated that critical hypoperfusion of the skin and subcutaneous tissue during prolonged arrest and CPR precludes the hyperemia and edema beneath the shocking electrodes previously proposed as primary mechanisms of impedance decline with repeated monophasic shocks during normal perfusing rhythms.11 Also consistent with our findings are data from a clinical study of recurrent fibrillation during prehospital advanced life support care, which found that impedance in a 36-patient subset receiving first and second 200 J biphasic shocks did not change meaningfully between shocks.12 In another recent clinical study, Deakin et al.13 reported a small but statistically significant decrease in impedance between escalating biphasic shocks delivered during elective cardioversion. While the method of impedance measurement was not reported, the results are consistent with the modest decrease in shock impedance with increasing shock intensity first described by Geddes et al.7 , and also observed in our data. In our analysis of 863 AED records, impedance change between consecutive shocks was found to be negligible, particularly for consecutive same-energy shocks. For those patients that received a second 200 J shock following their first 200 J shock, the minimal impedance change resulted in a negligible median peak current increase of less than 200 mA. When the analysis was expanded to include all same-energy consecutive shocks, impedance remained essentially unchanged in close to half the instances, and was nearly as likely to exhibit a modest increase as a modest decrease in the other half of the instances (Fig. 1). Thus these data contradict the common presumption that decline in impedance is an operative mechanism by which shock current, and thus defibrillation probability, might increase for a subsequent shock when a previous shock at that same energy setting has failed to terminate VF. For those patients that received a 200 J shock followed by a 300 J shock, a moderate (27%) increase in second-shock peak current was observed. As physics would predict,7 the impedance decreased a small amount (4%) when dose was increased. The much larger increase in current was thus due predominantly to the higher capacitor voltage associated with the higher energy setting. If impedance did decrease meaningfully between shocks, shock current would indeed increase and move the defibrillation operating point up the probability-of-success curve. However, in this large patient cohort, the only mechanism by which the current of subsequent shocks increased appreciably was escalation of the shock energy dose.
As shown in Table 3, we observed an increase in defibrillation probability with increasing shock dose among the subset of patients who received shocks at each of the three energy levels. This increase in defibrillation probability with increasing dose is predicted by the well-known concept of the defibrillation probability-ofsuccess curve – a fundamental property of defibrillation14 – and is also consistent with the results of a recent randomized clinical trial.15 Although this fundamental dose–response property of defibrillation implies only that a higher dose has a relative higher defibrillation probability within a given patient or population, a common mis-extrapolation has been the notion that any given energy has a fixed, constant probability to defibrillate, even across different populations. In particular, it has been suggested that a shock dose observed to have a given VF termination probability in a large population maintains that same VF termination probability for second shocks delivered to the subset of patients left in VF after a failed first shock, and even for third shocks delivered to the subset of patients left in VF after failed first and second shocks.6 Inherent in this suggestion is the presumption that defibrillation probability is relatively homogenous throughout a population. Stated differently, this notion presumes that failure of a shock to terminate VF is predominantly a randomly distributed event – both longitudinally in a patient as well as across a population – and therefore that VF termination probability for second and subsequent shocks is mostly independent of the actual success or failure of the first shock. Contrary to the presumption of “constant defibrillation probability”, we found that VF termination failure was far from randomly distributed. A small subset of the patients accounted for most of the failed shocks, and in these patients, the first shock VF termination rate was substantially lower than it was in the overall population. Across all patients the probability of VF termination for subsequent shocks was significantly lower in cases with a failed first shock (69%) than in cases with a successful first shock (93%). Additionally, in direct contradiction to the common presumption,6 subsequent shock VF termination success in cases with a failed first shock was significantly lower than the first shock VF termination rate across the overall population. Thus our data indicate that, rather than representing a mostly random event, first shock defibrillation failure is often indicative of a more-difficult-to-defibrillate patient, and may serve to select out a subgroup of patients in which a given shock dose provides a significantly lower defibrillation probability than in the overall population. In our data, this subset of more-difficult-to-defibrillate patients was not distinguishable by impedance; further research is needed to determine the characteristics of these patients and the causes of defibrillation failure. 4.3. Limitations Our observations reflect data collected from a large patient cohort treated with one specific defibrillator model and shock waveform. Other defibrillators currently on the market, and used in prior studies of impedance behavior between consecutive shocks, feature a variety of different methods of measuring impedance, as well as a variety of shock waveforms. To date, ours is the only large clinical evaluation of impedance behavior between consecutive shocks specifically in the setting of out-of-hospital cardiac arrest, and our findings corroborate those from the only experimental evaluation of this behavior in a comparable cardiac arrest model.10 Given the present understanding of the underlying mechanisms,10,11,14 as well as the concordance of our clinical findings with experimental data using other monophasic and
R.G. Walker et al. / Resuscitation 80 (2009) 773–777
biphasic shock waveforms,10 there is no basis to believe that the actual impedance behavior between consecutive shocks during cardiac arrest resuscitation would differ for any available defibrillator. However, considering the variety of methods of measuring impedance implemented in different devices, it would be important for any future studies of this phenomenon to specify the method(s) of impedance measurement, and if possible include a direct measure of impedance during shock discharge, to allow meaningful comparison to existing data. 5. Conclusions The findings of this large consecutive case series analysis directly contradict two common presumptions about defibrillation during out-of-hospital cardiac arrest, and support recent changes to resuscitation guidelines. Our finding that repeated shocks do not cause either a meaningful decrease in transthoracic impedance, or a meaningful impedance-mediated increase in shock current, adds support to the move away from stacked shocks made in the 2005 guidelines.4,5 Furthermore, the absence of any appreciable increase in shock current when a second shock at the same setting is administered supports the guidelines statement that if the first shock fails to terminate VF and the defibrillator is capable of delivering higher dose shocks, it is rational to increase the energy for subsequent shocks.5 Thus in terms of successful termination of VF, our findings indicate that the true benefit to be derived from the practices of stacking shocks and delivering second shocks at the same energy as failed first shocks are likely to be less than expected. This reduction in the VF termination yield that should reasonably be expected from such practices, weighed against the substantial cost of each defibrillation attempt in terms of CPR interruption time,16,17 supports the move away from stacked shocks and, for the many defibrillators that provide two or more shock doses for treatment of cardiac arrest, encourages ongoing re-evaluation of the practice of repeating sub-maximal dose shocks when VF is not terminated by the previous shock. Conflict of interest statement Robert G. Walker is a current employee of Physio-Control. Rudolph W. Koster was a paid medical advisor for Physio-Control until 2007, and has received research grants from Physio-Control. Fred W. Chapman is a current employee of Physio-Control. Authors Sun, Moffat, Barger, and Dodson have no conflicts to declare.
777
Acknowledgements We thank the EMTs, paramedics, and data collection staff at each of the sites for downloading the defibrillator data to enable our analysis. References 1. Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 4: the automated external defibrillator: key link in the chain of survival. The American Heart Association in Collaboration with the International Liaison Committee on Resuscitation. Circulation 2000;102:I-60–76. 2. Kerber RE, Grayzel J, Hoyt R, Marcus M, Kennedy J. Transthoracic resistance in human defibrillation. Influence of body weight, chest size, serial shocks, paddle size and paddle contact pressure. Circulation 1981;63:676–82. 3. Sirna SJ, Ferguson DW, Charbonnier F, Kerber RE. Factors affecting transthoracic impedance during electrical cardioversion. Am J Cardiol 1988;62:1048–52. 4. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation 2005;112:IV-1–203. 5. Deakin CD, Nolan JP. European Resuscitation Council guidelines for resuscitation 2005. Section 3. Electrical therapies: automated external defibrillators, defibrillation, cardioversion and pacing. Resuscitation 2005;67(Suppl. 1):S25–37. 6. Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 6: advanced cardiovascular life support: section 2: defibrillation. The American Heart Association in collaboration with the International Liaison Committee on Resuscitation. Circulation 2000;102:I-90–4. 7. Geddes LA, Tacker Jr WA, Schoenlein W, Minton M, Grubbs S, Wilcox P. The prediction of the impedance of the thorax to defibrillating current. Med Instrum 1976;10:159–62. 8. Dahl CF, Ewy GA, Ewy MD, Thomas ED. Transthoracic impedance to direct current discharge: effect of repeated countershocks. Med Instrum 1976;10:151–4. 9. Geddes LA, Tacker WA, Cabler P, Chapman R, Rivera R, Kidder H. The decrease in transthoracic impedance during successive ventricular defibrillation trials. Med Instrum 1975;9:179–80. 10. Niemann JT, Garner D, Lewis RJ. Transthoracic impedance does not decrease with rapidly repeated countershocks in a swine cardiac arrest model. Resuscitation 2003;56:91–5. 11. Sirna SJ, Kieso RA, Fox-Eastham KJ, Seabold J, Charbonnier F, Kerber RE. Mechanisms responsible for decline in transthoracic impedance after DC shocks. Am J Physiol 1989;257:H1180–3. 12. Koster RW, Walker RG, Chapman FW. Recurrent ventricular fibrillation during advanced life support care of patients with prehospital cardiac arrest. Resuscitation 2008;78:252–7. 13. Deakin CD, Ambler JJ, Shaw S. Changes in transthoracic impedance during sequential biphasic defibrillation. Resuscitation 2008;78:141–5. 14. Tacker WA. Fibrillation causes and criteria for defibrillation. In: Tacker WA, editor. Defibrillation of the heart: ICDs, AEDs, and manual. St. Louis, MO: Mosby; 1994. p. 1–14. 15. Stiell IG, Walker RG, Nesbitt LP, et al. BIPHASIC trial: a randomized comparison of fixed lower versus escalating higher energy levels for defibrillation in out-ofhospital cardiac arrest. Circulation 2007;115:1511–7. 16. Berg RA, Hilwig RW, Kern KB, Sanders AB, Xavier LC, Ewy GA. Automated external defibrillation versus manual defibrillation for prolonged ventricular fibrillation: lethal delays of chest compressions before and after countershocks. Ann Emerg Med 2003;42:458–67. 17. Kramer-Johansen J, Edelson DP, Abella BS, Becker LB, Wik L, Steen PA. Pauses in chest compression and inappropriate shocks: a comparison of manual and semi-automatic defibrillation attempts. Resuscitation 2007;73:212–20.