Real-Time Ventricular Fibrillation Amplitude-Spectral Area Analysis to Guide Timing of Shock Delivery Improves Defibrillation Efficacy During Cardiopulmonary Resuscitation in Swine

Abstract

Background: The ventricular fibrillation amplitude spectral area (AMSA) predicts whether an electrical shock could terminate ventricular fibrillation and prompt return of spontaneous circulation. We hypothesized that AMSA can guide more precise timing for effective shock delivery during cardiopulmonary resuscitation.

Methods and results: Three shock delivery protocols were compared in 12 pigs each after electrically induced ventricular fibrillation, with the duration of untreated ventricular fibrillation evenly stratified into 6, 9, and 12 minutes: AMSA-Driven (AD), guided by an AMSA algorithm; Guidelines-Driven (GD), according to cardiopulmonary resuscitation guidelines; and Guidelines-Driven/AMSA-Enabled (GDAE), as per GD but allowing earlier shocks upon exceeding an AMSA threshold. Shocks delivered using the AD, GD, and GDAE protocols were 21, 40, and 62, with GDAE delivering only 2 AMSA-enabled shocks. The corresponding 240-minute survival was 8/12, 6/12, and 2/12 (log-rank test, P=0.035) with AD exceeding GDAE (P=0.026). The time to first shock (seconds) was (median [Q1-Q3]) 272 (161-356), 124 (124-125), and 125 (124-125) (P<0.001) with AD exceeding GD and GDAE (P<0.05); the average coronary perfusion pressure before first shock (mm Hg) was 16 (9-30), 10 (6-12), and 3 (-1 to 9) (P=0.002) with AD exceeding GDAE (P<0.05); and AMSA preceding the first shock (mV·Hz, mean±SD) was 13.3±2.2, 9.0±1.6, and 8.6±2.0 (P<0.001) with AD exceeding GD and GDAE (P<0.001). The AD protocol delivered fewer unsuccessful shocks (ie, less shock burden) yielding less postresuscitation myocardial dysfunction and higher 240-minute survival.

Conclusions: The AD protocol improved the time precision for shock delivery, resulting in less shock burden and less postresuscitation myocardial dysfunction, potentially improving survival compared with time-fixed, guidelines-driven, shock delivery protocols.

Keywords: amplitude spectral area; animal model; defibrillation; resuscitation; sudden cardiac arrest; ventricular fibrillation; ventricular fibrillation waveform analysis; waveform analysis.

© 2017 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley.

Figures

Figure 1

AMSA‐Driven (AD) protocol featuring 1 representative experiment for each shock advisory criterion and the corresponding total number of shocks given guided by such criterion for the entire AD group. Shown in each graph are the AMSA at the time CPR started (AMSAt0) and the time threshold according to: 15−(0.0222·[tcpr−t0]); where t0 is the time at which CPR started and tcpr is the CPR time in seconds with each “step” representing the interval between ventilation pauses during which shock advise and subsequent shock delivery occurred (if advised). The AMSA threshold criterion (A) and AMSA delta criterion (B) are time independent based on reaching an AMSA value ≥15 mV·Hz (AMSA threshold) or having an AMSA increase ≥30% relative to the preceding AMSA value at any time during CPR (AMSA delta). The time threshold criterion (C) and the time maximum criterion (D) are independent of the AMSA level during CPR but dependent on CPR duration, with the time threshold setting a time to advise shock based on AMSAt0 exceeding 15−(0.0222·[tcpr−t0] in seconds) and the time maximum setting a time (360 s) by which a shock must be delivered. As shown by the representative experiments, the threshold criterion (A) and the delta criterion (B) were met before the time threshold criterion. The time threshold criterion was met after failing to meet AMSA threshold or AMSA delta criterion while having an AMSAt0 that enabled shock advice within 360 s (C). When none of the above criteria were met, the time maximum criterion advised shock at 360 s (D). Please refer to the Methods section for the approach developed to reset the algorithm after a shock is advised. AMSA indicates amplitude spectral area; CPR, cardiopulmonary resuscitation.

Figure 2

Swine model of electrically induced VF with resuscitation attempted by conventional CPR using a mechanical piston device and bag‐valve ventilation through an endotracheal tube with delivery of 100% oxygen. CPR indicates cardiopulmonary resuscitation; VF, ventricular fibrillation.

Figure 3

Kaplan–Meier survival curves analyzed from the time VF was induced in all 36 animals (A) and from the time ROSC was achieved in 27 animals (B), comparing the 3 defibrillation protocols: AMSA‐Driven (AD), Guidelines‐Driven (GD), and Guidelines‐Driven/AMSA‐Enabled (GDAE). The overall P value calculated using the log‐rank test is shown in the graph along with the pairwise comparison between AD and GDAE using the Holm–Sidak test. ROSC indicates return of spontaneous circulation; VF, ventricular fibrillation.

Figure 4

Shown are the initial 12 min of the resuscitation effort from the start of CPR (ie, min 0) depicting the main effects of the defibrillation protocols along with differences between animals that achieved ROSC and those that did not. Differences were analyzed by a linear mixed‐effect model treating time as continuous variable to assess the main effects of time (min) and either the defibrillation protocol (DP)—left—or the CPR outcome (ROSC)—right—and their interaction (DP or ROSC×min) and also treating time as a discrete variable to show for descriptive reasons differences at specific moments using the Sidak multiple comparisons pairwise method (left). AD indicates AMSA‐Driven; AMSA, amplitude spectral area; CPP, coronary perfusion pressure; CPR, cardiopulmonary resuscitation; GD, Guidelines‐Driven; GDAE, Guidelines‐Driven/AMSA‐Enabled; ROSC, return of spontaneous circulation. The number of animals receiving CPR at each time measurement is depicted graphically on top of the figure, indicating that only 10 animals remained receiving CPR after 12 min (ie, AD=3/12, GD=2/12, and GDAE=5/12). Arrows denote delivery of epinephrine (EPI) (1 mg into the right atrium) at 4 and 8 min of CPR. The absence of AMSA data in the GD and GDAE groups at 2, 4, 6, 8, 10, and 12 min corresponds to the pauses when electrical shocks were delivered, precluding calculation of AMSA. Data are shown as mean±SEM; “a” different from AD; *P≤0.05, †P≤0.01, and ‡P≤0.001.

Figure 5

First shock analysis showing (A) CPR duration before first shock; (B) Epinephrine given before first shock; (C) Average coronary perfusion pressure (CPP) before first shock; (D) AMSA before first shock; (E) Individual AMSA values showing the AMSA immediately before start of CPR and the corresponding AMSA before shock delivery; with white symbols denoting AMSA‐Driven (AD), black symbols Guidelines‐Driven (GD), and gray symbols Guidelines‐Driven/AMSA‐Enabled (GDAE) and with the duration of untreated VF shown by circles for 6 min, by squares for 9 min, and by diamonds for 12 min (the animal that received a shock after 12 min, had a pulseless electrical activity at the end of untreated VF, and reversed to VF after 7 min if CPR, time at which the AD algorithm was activated); and (F) the first shock outcome determined at the next CPR pause or 18 s later if CPR was not resumed. The CPR duration, epinephrine, and CPP data were analyzed by the Kruskal–Wallis on ranks test and AMSA by 1‐way ANOVA with corresponding tests for pairwise comparisons. The first shock outcome was analyzed by χ2 (*P=0.016). The group data are shown as box plots and means±SD with the P values for the overall test displayed in each graph and the pairwise comparisons denoted by symbols (*P<0.05, ‡P<0.001 vs AD). AMSA indicates amplitude spectral area; CPR, cardiopulmonary resuscitation; PEA, pulseless electrical activity; ROCA, return of cardiac activity; ROSC, return of spontaneous circulation; VF, ventricular fibrillation.

Figure 6

Postresuscitation effects on hemodynamic function. Baseline measurements were obtained 5 min before inducing ventricular fibrillation. Differences were analyzed by a linear mixed‐effect model treating time as continuous variable to assess the main effects of time (min) and defibrillation protocol (DP) and their interaction (DP×min) and treating each time point as a discrete moment for descriptive reasons using the Sidak multiple comparisons pairwise method to identify statistically significant differences among groups at the specific time point. AD indicates AMSA‐driven; AMSA, amplitude spectral area; DO2I, systemic oxygen delivery index; GD, guidelines‐driven; GDAE, guidelines‐driven/AMSA‐enabled; LVSWI, left ventricular stroke work index; MAP, mean aortic pressure; PETCO2, end‐tidal PCO2; VO2I, systemic oxygen consumption index. The data are shown as mean±SEM; “a” different from AD; *P≤0.05 and ‡P≤0.001.

Figure 7

Upper graphs depict the association between AMSA values measured during the pause before shock delivery and the immediate cardiac outcomes for the first shock (A) and for all shocks (B). Differences were analyzed by 1‐way ANOVA followed by the Holm‐Sidak pairwise comparisons test with the P values for the overall test shown in each graph. AMSA indicates amplitude spectral area; AUC, area under the curve; PEA, pulseless electrical activity; ROCA, return of cardiac activity; ROSC, return of spontaneous circulation; VFp, persistence of ventricular fibrillation. The data are shown as mean±SD; “a” different from VFp, “b” different from PEA; *P<0.05, ‡P<0.001. Lower graphs depict the areas under the receiver operator characteristic curves with their 95% confidence intervals shown in brackets for the predictive value of AMSA on the immediate outcome for the first shock outcome (C) and for all shocks (D). The outcome was defined as ROSC or no‐ROSC (ie, ROCA, PEA, or VFp).

Figure 8

A, Histograms showing the distribution of the shocks delivered per animal according to the defibrillation protocol (AD, AMSA‐Driven; GD, Guidelines‐Driven; GDAE, Guidelines‐Driven/AMSA‐Enabled). B, Shock burden (calculated as described in the text) analyzed by Kruskal–Wallis test on ranks followed by the Tukey test for pairwise comparisons, with the data shown in box plots. The P value for the overall test is shown in the graph (*P<0.05 vs AD). C, Linear regression between shock burden and survival time after ROSC, showing the corresponding 95% confidence interval for the regression analysis.1One GD data point is behind 3 AD data points (2 at 240 min and 1 at 237 min); 2Two GD data points are behind 3 AD data points; 3One GD and 1 GDAE data points are behind the 2 AD data points. D, Linear regression between shock burden and LVSWI at 240 min in 15 animals that survived the entire postresuscitation interval, showing the corresponding 95% confidence interval for the regression analysis. Hemodynamic data were lost because of technical issues in 1 AD animal that had zero shock burden and survived 240 min. 1Two superimposed GD data points. AMSA indicates amplitude spectral area; LVSWI, left ventricular stroke work index; ROSC, return of spontaneous circulation.