Methods

Patients who sustained out-of-hospital cardiac arrest in the Greater Birmingham, Alabama area from January to August 2008 were studied. The subjects had been enrolled in the Resuscitation Outcomes Consortium-Cardiac Arrest Population-based Registry (ROC-Epistry). The Resuscitation Outcomes Consortium is a clinical research network consisting of 10 clinical centers in the United States and Canada organized to conduct randomized trials pertaining to trauma and out-of-hospital sudden cardiac arrest. The ROC-Epistry Registry was designed to assess how evidence-based resuscitation is implemented in community-based practice. The present observational study, based on existing data sources, met the requirements for minimal risk and waiver of informed consent. Local institutional review board approval was obtained before having access to de-identified data files. Outcomes data were not obtained.

All patients studied were resuscitated in the field using LifePak 12 and 500 defibrillators (Physio-Control, Redmond, Washington). These defibrillators can continuously record electrocardiographic and thoracic impedance waveforms. The rhythm strips recorded electronically were obtained from the defibrillation patches used. Proprietary software (CODE-STAT, version 7.0, Physio-Control) was used to analyze the electronic files, retrieved from the defibrillators. The software allows for analysis of both rhythm strips and thoracic impedance waveforms.

Large or chaotic artifacts produced by CCs can render rhythm strips obtained during CPR uninterpretable. Frequently, however, a stable, oscillating artifact is seen ( Figure 1 ). In these cases, we searched for evidence of ventricular capture caused by CCs. Periods of pulseless electrical activity or asystole, either as the initial rhythm or after successful defibrillation, were used for analysis.

Artifact produced by CCs, as initiated (asterisk) and interrupted (A,C) from patient in asystole. (B,D) Same tracings, with superimposition of thoracic impedance waveform from defibrillator pads (green), which can be used to analyze CC events. Red arrows generated by algorithm mark each CC event.

The external defibrillators measured and recorded impedance continuously by passing a safe, high-frequency (14 to 20 kHz), alternating current between the defibrillation pads and measuring the voltage generated. A computer algorithm identified CC events ( Figure 1 ) using a method that has been clinically validated. The software marks each CC event with a red arrow ( Figures 1 to 5 ) .

Rhythm strip showing evidence of ventricular capture (QRS complexes marked with asterisk). (A) Note, similarities in marked complexes. (B) Note, similar timing between impedance (green) and initiation of marked QRS complexes. Note, also, QRS seen after each of first 3 CCs but intermittently thereafter. Arrows generated by computer algorithm mark each CC event. Dashed circles point to stable relation between QRS complexes, CC artifact, and impedance waveform. (C) Nine QRS complexes considered ventricular capture from CCs are superimposed. Note, how QRS complexes and impedance waveform line up.

Example of evidence of ventricular capture. (A) Ventricular capture (QRS complexes marked with asterisk ) seen after some CC artifacts. Note, ventricular capture is intermittent but QRS complexes kept a stable relation with impedance waveform (B, green tracing). Arrows generated by algorithm mark each CC event.

Example of evidence of ventricular capture (QRS complexes marked with asterisk ), seen once for every 2 CCs until seventh capture and once for every 3 CCs thereafter. (A) Arrows generated by algorithm mark each CC event. (B) Note, stable relation among QRS complexes, CC artifact, and impedance waveforms ( green). (B) Gain of electrocardiographic signal 5 mm/mV.

(A) Example of VF arising shortly after CCs begun ( asterisk marks initiation of VF). Note, patient was in pulseless electrical activity and had idioventricular rhythm allowing third CC to be delivered during repolarization phase, after which VF ensued (possible “compression-on-T” phenomenon). (B) Example from another patient of VF arising during CCs.

Intermittent ventricular capture was defined as the presence of QRS complexes of similar morphology, that were intermittently associated with the CC cycles determined from thoracic impedance, and that exhibited a constant time relation with the CC artifact seen on the rhythm strip and with the impedance signal. Only cases of intermittent ventricular capture were considered, because of the difficulty in differentiating constant ventricular capture in a 1:1 ratio with the CC from a constant motion artifact or an intrinsic rhythm of coinciding rate. Two groups were defined according to the presence or absence of ventricular capture by CCs. Ventricular activations were defined as QRS complexes that were either intrinsic beats or considered from ventricular capture. Successful defibrillation was defined as the absence of VF 5 seconds after defibrillation.

In the 24 cases in which CCs produced only minor or stable motion artifacts, the number of ventricular activations during a 15-second period while CCs were being performed was compared to a 15-second period in which CCs were not (immediately before the initiation of CCs when possible). In the group of patients in whom ventricular capture was found, the first segment in which it was seen was used. In 2 cases, however, CCs were started within a few seconds of successful defibrillation. In these cases, the 15 seconds before and after the cessation of CCs were used. In the group of patients in whom ventricular capture was not found, the first segment in which CCs produced motion artifact and the 15 seconds preceding it were used.

The number of episodes of refibrillation (in cases in which the initial rhythm was VF) was analyzed. In the cases in which the initial rhythm was not VF, episodes of de novo VF were quantified, and for analysis purposes were categorized as refibrillation. A standard electrocardiographic gain and sweep speed (10 mm/mV and 25 mm/s, respectively) was used for all the figures shown, unless noted otherwise.

The Statistical Package for Social Sciences was used for statistical analysis. The Wilk-Shapiro test was used to assess for normality of the distribution of the data. The number of ventricular activations during a 15-second period was normally distributed. Continuous variables are expressed as the mean ± SD. The paired t test with equal variances not assumed was used to compare the number of ventricular activations before and during CCs in each group (with and without ventricular capture). An independent samples Student’s t test was used to compare the change in the number of ventricular activations before and during CCs in both groups. Statistical significance was defined as a 2-tailed p value <0.05.