Vous êtes sur la page 1sur 8

Resuscitation Science

Waveform AnalysisGuided Treatment Versus a Standard


Shock-First Protocol for the Treatment of Out-of-Hospital
Cardiac Arrest Presenting in Ventricular Fibrillation
Results of an International Randomized, Controlled Trial
John P. Freese, MD; Dawn B. Jorgenson, PhD; Ping-Yu Liu, PhD; Jennifer Innes;
Luis Matallana, EMT-P; Krishnakant Nammi, MS; Rachael T. Donohoe, PhD; Mark Whitbread;
Robert A. Silverman, MD, MS; David J. Prezant, MD
BackgroundVentricular fibrillation (VF) waveform properties have been shown to predict defibrillation success and
outcomes among patients treated with immediate defibrillation. We postulated that a waveform analysis algorithm
could be used to identify VF unlikely to respond to immediate defibrillation, allowing selective initial treatment with
cardiopulmonary resuscitation in an effort to improve overall survival.
Methods and ResultsIn a multicenter, double-blind, randomized study, out-of-hospital cardiac arrest patients in 2 urban
emergency medical services systems were treated with automated external defibrillators using either a VF waveform
analysis algorithm or the standard shock-first protocol. The VF waveform analysis used a predefined threshold value
below which return of spontaneous circulation (ROSC) was unlikely with immediate defibrillation, allowing selective
treatment with a 2-minute interval of cardiopulmonary resuscitation before initial defibrillation. The primary end point
was survival to hospital discharge. Secondary end points included ROSC, sustained ROSC, and survival to hospital
admission. Of 6738 patients enrolled, 987 patients with VF of primary cardiac origin were included in the primary
analysis. No immediate or long-term survival benefit was noted for either treatment algorithm (ROSC, 42.5% versus
41.2%, P=0.70; sustained ROSC, 32.4% versus 33.4%, P=0.79; survival to admission, 34.1% versus 36.4%, P=0.46;
survival to hospital discharge, 15.6% versus 17.2%, P=0.55, respectively).
ConclusionsUse of a waveform analysis algorithm to guide the initial treatment of out-of-hospital cardiac arrest patients
presenting in VF did not improve overall survival compared with a standard shock-first protocol. Further study is
recommended to examine the role of waveform analysis for the guided management of VF.
Clinical Trial RegistrationURL: http://www.clinicaltrials.gov. Unique identifier: NCT00535106.
(Circulation. 2013;128:995-1002.)
Key Words:atrial fibrillation cardiopulmonary resuscitation defibrillation, electric resuscitation

suggested a 3-phase model for VF that included an electrical,


a circulatory, and a metabolic phase, with immediate defibrillation being the optimal treatment for only the first phase in
which the interval from the onset of VF to the time of defibrillation is of short duration. At about the same time, initial
studies were published that described the use of delayed defibrillation or a period of cardiopulmonary resuscitation (CPR)
before the initial defibrillatory shock as a means to improve
survival for VF of longer duration.9,10
This concept of delayed defibrillation was included in
the 2005 resuscitation guidelines as a treatment option to
be considered in adults with out-of hospital ventricular

ore than 50 years have passed since the first description


of successful transthoracic defibrillation for the treatment of ventricular fibrillation (VF), and it has been >40 years
since defibrillators were first introduced into the prehospital
setting in the United States and United Kingdom.14 However,
despite decades of experience and >500000 cardiac arrests
annually in the United States and Europe, the optimal strategy
for the initial management of VF has yet to be defined.5,6

Clinical Perspective on p 1002


Immediate defibrillation had long been considered the
standard treatment for VF.7 In 2002, Weisfeldt and Becker8

Received March 22, 2012; accepted May 7, 2013.


From the Office of Medical Affairs, Fire Department of New York, Brooklyn, NY (J.P.F., L.M., D.J.P.); Philips Healthcare, Seattle, WA (D.B.J., K.N.);
Fred Hutchinson Cancer Research Center, Seattle, WA (P.-Y.L.); London Ambulance Service, London, UK (J.I., R.T.D., M.W.); Department of Emergency
Medicine, Long Island Jewish Medical Center, New Hyde Park, NY (R.A.S.); and Pulmonary Medicine Division, Department of Medicine, Montefiore
Medical Center and Albert Einstein College of Medicine, Bronx, NY (D.J.P.).
Correspondence to John P. Freese, MD, Chief Medical Director, Fire Department of New York, 9 MetroTech Center, 4W-1 Brooklyn, NY 11201. E-mail
freesej@fdny.nyc.gov or jfreesemd@hotmail.com
2013 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org

DOI: 10.1161/CIRCULATIONAHA.113.003273

995

996CirculationAugust 27, 2013


fibrillationand EMS [emergency medical services] response
(call to arrival) intervals >4 to 5 minutes.11 Most recently,
however, randomized, controlled trials and meta-analyses designed to assess the impact of delayed defibrillation found no benefit and led to recent consensus was that
there is inconsistent evidence to support or refute a delay
in defibrillation.1217
Various properties of the VF waveform have been shown
to be strongly predictive of response to defibrillation.1821
Correlations between VF waveform properties and survival
have also been shown.22,23 For this reason, the utility of waveform analysis for the guided management of VF has been suggested.2426 Conceptually, using waveform analysis to identify
those patients in the electrical phase would allow them to
receive immediate defibrillation, whereas those patients in the
circulatory phase could selectively receive delayed defibrillation. Thus, the known benefit of early defibrillation would
be retained for those patients most likely to benefit, and an
alternative treatment (CPR) could be provided to the remaining patients in an attempt to improve outcomes in this group
with traditionally poor survival.
This prospective, international study was designed to assess
the efficacy of guided initial resuscitative management using
a waveform analysis algorithm compared with a standard
shock-first protocol for the management of out-of-hospital
cardiac arrest presenting in VF.

Methods
Study Setting and Design
This multicenter, double-blind, randomized study was conducted by
the primary EMS systems in New York, NY, and London, UK. The
parallel-group trial was intended to demonstrate the superiority of a
waveform analysis algorithm compared with a standard shock-first
protocol for the initial treatment of out-of-hospital cardiac arrests
presenting in VF.
Study inclusion required that the patients arrest was of cardiac
origin as described by the rescuers in accordance with the Utstein
style,13 that the initial defibrillator used during the resuscitation was
a study device, and that the subject presented in VF as determined by
the initial rhythm assessment of the automated external defibrillator
(AED). Pediatric patients (age <18 years) and those for whom resuscitative care was terminated as a result of a do not resuscitate order
were excluded, as were patients for whom data from the study device
were not able to be obtained or for whom the initial AED analysis
was incomplete.
The waveform analysis algorithm used in this study was a Food
and Drug Administrationcleared product and did not require an
investigational device exemption. In New York, the study was
approved by the Institutional Review Board of the New York City
Department of Health and Mental Hygiene and included a waiver of
informed consent process consistent with existing requirements. In
London, the study was approved by the Lewisham Local Research
Ethics Committee.
The study was overseen by an independent data safety monitoring board comprising experienced resuscitation scientists chosen for
their extensive work in the field of out-of-hospital resuscitation and
without financial conflict of interest pertaining to defibrillator manufacturers. This board undertook a planned interim efficacy analysis
when primary outcomes were known for 75% of the anticipated sample size, also allowing an adaptive sample size adjustment. Following
the predefined interim analysis plan, this board did not recommend
sample size adjustment and instead recommended study termination for futility. At the time of that recommendation, the originally
planned sample size requirements had been met.

Treatment Protocol
Patient care was provided by certified first responders/firefighters
in New York and by emergency medical technicians or paramedics
in both cities. All patient care was provided according to standardized prehospital protocols. These did not vary between the sites with
respect to initial resuscitation management and were consistent with
2005 international resuscitation guidelines.
After the arrival of certified first responders or emergency medical
technicians/paramedics, CPR was initiated only until an AED was
applied. All AED CPR intervals were set to 2 minutes. The study
device (FR2+, Philips Healthcare, Seattle, WA) used an impedancecompensating biphasic truncated exponential waveform and fixed
150-J energies. No other interventions (advanced airway management, vascular access, pharmaceutical administration) were performed during this initial treatment phase.
AEDs were randomized to either a standard shock-first protocol or VF waveform analysis algorithm. The AED recommended an immediate defibrillatory shock for all VF in the
standard shock-first arm. In the waveform analysis arm, the
initial rhythm analysis used a proprietary algorithm (Philips
Healthcare) to analyze the VF waveform and to assign a resulting numeric VF score. 27 This VF score was then compared
with a predefined threshold value with a sensitivity >80% and
a specificity >60% with respect to the probability of achieving return of spontaneous circulation (ROSC) after immediate
defibrillation.24 Patients whose waveform analysis score met or
exceeded that threshold value were recommended for immediate defibrillation. Those whose VF waveform analysis score
was below the threshold value were recommended for a 2-minute CPR interval before reanalysis. This waveform analysis was
incorporated into the rhythm assessment of the AED and did
not result in any delay in treatment recommendation. All subsequent rhythm analyses in the VF waveform analysis group and
all rhythm analyses in the shock-first protocol used an identical
VF identification and treatment algorithm. Prehospital providers
were blinded to device randomization, which was reversed at
specified intervals.
Because treatment differed only for those patients whose VF scores
were below the predefined threshold, the primary comparison group
was made up of those patients in the 2 arms whose initial VF scores
were below that threshold.

Data Collection
Patient data were collected according to the updated Utstein template.13 Computerized dispatch records, prehospital care forms
completed by EMS personnel, and phone interviews with the
EMS providers were used to collect prehospital and demographic
data. Hospital data were collected by each agency directly from
the receiving hospitals. Validation of outcomes using available
national registries was also performed. Individuals responsible
for clinical data acquisition were blinded to all ECG data until the
study was complete.
AED data were transmitted from the study device to a removable
data card, which was retrieved after each use. ECG data were then
transferred to the study sponsor for analysis. VF scores generated by
the AED were validated for all patients in the waveform analysis arm,
and initial VF scores were calculated for all patients in the shock-first
arm post hoc. For those patients in the waveform analysis arm who
were recommended to a 2-minute CPR interval before reanalysis,
VF scores were also calculated after the CPR interval. Individuals
responsible for ECG data analysis were blinded to all clinical data
until the study was complete.

Outcomes
The primary outcome for this study was survival to hospital discharge.
Secondary outcomes included ROSC, sustained ROSC defined as
ROSC maintained until hospital arrival, and survival to hospital
admission, each defined in accordance with the Utstein template.13

Freese et al SmartCPR Trial: VF Waveform Analysis for OOHCA 997

Results

Sample Size and Statistical Analysis


The primary end point analysis of the study was designed to assess
survival to hospital discharge for patients whose VF presented below
the predefined threshold. Comparison was made between those
patients recommended to CPR before defibrillation as a result of the
waveform analysis algorithm and those treated with immediate defibrillation. Predefined additional analyses included a primary outcome
analysis for all VF patients and an analysis of both the overall study
population and the primary comparison group for all secondary outcomes. As described below, an a priori decision was made to perform
a subgroup analysis of patients in the waveform analysis arm whose
VF scores improved after the 2-minute CPR interval.
Before the study, we postulated that the waveform analysis algorithm would result in improved survival to hospital discharge. On the
basis of the known VF survival rates in both study sites and the recognition that patients with lower waveform analysis scores are less
likely to survive than the population of VF arrests as a whole, we estimated a baseline 5% survival rate for the primary comparison group
and hypothesized an absolute improvement in survival to hospital
discharge of 10%.
With the use of a 2-sided value of 0.05 and a power of 90%,
the sample size requirement for primary comparison group in this
study was calculated to be 400 patients. After adjustment for anticipated exclusion criteria, a 10% exclusion rate was projected, and
the sample size for the primary comparison group was adjusted to
450. With an estimated 50% of waveform analysis scores presenting
below the predefined threshold, this yielded a total study population
of 900 patients that would be required from the combination of the
2 study sites.
The primary end point analysis was conducted with a Fisher exact
test, and other differences within the primary comparison group were
assessed with multivariate logistic regression analyses. Odds ratios
and 95% confidence intervals were calculated with SAS statistical
software (SAS Institute Inc, Cary, NC). A value P<0.05 was considered significant.

Study Population
Enrollment began in London on May 5, 2006, and in New
York City on May 13, 2007. Both sites concluded enrollment on June 30, 2009. A total of 6738 patients were
enrolled during the study (Figure1). Of these, 5614 were
excluded on the basis of initial rhythm assessment: 4966
had an initial rhythm that was not VF and 648 were missing
AED ECG data.
Of the 1124 patients presenting in VF, 137 additional
patients were excluded from the study as a result of meeting 1
predefined exclusion criteria: 79 patients with arrests of noncardiac origin, 56 patients for whom the initial AED analysis
was interrupted or aborted, 8 pediatric patients, 8 patients for
whom a nonstudy AED was initially used, and 5 patients for
whom a do not resuscitate order was presented on scene.
The characteristics and incident details for the 987 patients
comprising the overall study population are shown in Table1.
There were no significant differences between the 2 arms
with respect to age, sex, or ethnicity. Neither arm varied
significantly with respect to incident characteristics such as
location, bystander witnessed status, bystander CPR before
EMS arrival, EMS response time, number of defibrillatory
shocks, use of endotracheal intubation, or administration of
either epinephrine or antidysrhythmics. There were significant
differences between the 2 sites with respect to ethnic diversity, frequency of bystander CPR, and EMS response interval. Bystander CPR occurred more often in London (49.94%
versus 24.43%; P<0.0001), and EMS response times were

Figure 1. Study overview. The primary


comparison group comprised patients
with ventricular fibrillation (VF) below a
predefined threshold who were treated
with either immediate defibrillation or 2
minutes of cardiopulmonary resuscitation
(CPR) before defibrillation as determined
by the automated external defibrillator
(AED) randomization. DNR indicates do
not resuscitate; and Rx, initial treatment.

998CirculationAugust 27, 2013


Table 1. Baseline Patient and Incident Characteristics for the
Overall Study Population
Shock First
(n=500)

Table 2. Initial VF Waveform Analysis Scores


Greater Than
Median Maximum Minimum the Threshold
Score* Score*
Score* Fraction, %

Waveform Analysis
(n=487)
All arrests (n=984)

0.098

1.000

0.019

45.1

352 (72)

Male sex (n=739)

0.096

1.000

0.019

44.4

Race, n (%)

Female sex (n=285)

0.102

0.562

0.024

47.2

Asian

22 (4)

19 (4)

Age 80 y (n=822)

0.098

1.000

0.020

43.2

Black

68 (14)

76 (16)

Age <80 y (n=163)

0.096

0.376

0.019

45.5

White

286 (57)

257 (53)

38 (8)

48 (10)

Asian (n=40)

0.092

0.276

0.032

42.5

Black (n=143)

0.104

1.000

0.026

49.7

Age, median (range), y


Male sex, n (%)

Hispanic

65 (1897)
388 (78)

64 (1897)

Race

Mixed

0 (0)

2 (0)

Other

21 (4)

17 (3)

White (n=543)

0.097

0.562

0.019

44.2

Unknown

65 (13)

68 (14)

Hispanic (n=86)

0.099

0.388

0.027

46.5

Location of arrest, n (%)

Other (n=40)

0.084

0.346

0.026

35.0

Residence

276 (55)

Unknown (n=132)

0.100

0.487

0.030

47.0

Business

40 (8)

37 (8)

124 (25)

111 (23)

60 (12)

42 (9)

395 (79)

Public location
Other
Bystander witnessed, n (%)

297 (61)

EMS-witnessed arrests (n=52)

0.144

1.000

0.026

65.4

Bystander witnessed
arrests (n=780)

0.100

0.562

0.019

44.2

388 (80)

Unwitnessed arrests (n=178)

0.083

0.388

0.023

34.8

0.487

0.021

47.2

1.000

0.019

41.6

Bystander CPR, n (%)

183 (37)

161 (33)

Response time 5 min (n=614) 0.100

Response time, median (range), min

4.30 (027)

4.28 (020)

Response time >5 min (n=368) 0.094

2 (116)

2 (018)

Shocks, median (range), n


Endotracheal intubation, n (%)

367 (73)

371 (76)

Epinephrine administered, n (%)

374 (75)

366 (75)

Antidysrhythmic administered, n (%)

148 (30)

127 (26)

CPR indicates cardiopulmonary resuscitation.

significantly shorter in New York (4 minutes 3 seconds versus


5 minutes 41 seconds; P<0.0001).

Waveform Analysis
Table2 describes the normalized initial waveform analysis scores for the overall patient population and for specific
subgroups characterized by patient or incident characteristics.
Female patients and those whose arrests were witnessed by
EMS providers were more likely to have an initial waveform
analysis score above the predefined threshold, whereas those
with unwitnessed arrests and patients for whom the EMS
response time exceeded 5 minutes were more likely to have an
initial waveform analysis score below threshold.

Outcome
Of the 987 patients with out-of-hospital VF cardiac arrests
who made up the study population, 8 (0.8%) were lost to
follow-up. No survival differences were noted between the 2
arms for the overall study population or for the primary comparison group (Table3).
For the overall study population, no differences were noted
between the 2 arms for any outcome end point. ROSC (42.5%
versus 41.2%; P=0.70), sustained ROSC (32.4% versus 33.4%;
P=0.79), survival to hospital admission (34.1% versus 36.4%;
P=0.46), and survival to hospital discharge (15.6% versus
17.2%; P=0.55) were similar between the waveform analysis
algorithm and the standard shock-first protocol, respectively.

EMS indicates emergency medical services; and VF, ventricular fibrillation.


*All scores were normalized.
Percentage of patients presenting above the predefined threshold VF
waveform analysis score.

For the primary comparison group, those patients with an


initial VF score below the predefined threshold, there was no
difference between the shock-first arm and waveform analysis
arm for the primary end point of survival to hospital discharge
(7.25% versus 7.19%; P=1.00), and no difference was found
for any secondary survival end point (ROSC, 29.39% versus
26.98%, P=0.57; sustained ROSC, 21.76% versus 20.14%,
P=0.83; survival to admission, 21.76% versus 22.66%,
P=0.84).
Multivariate logistic regression analysis (Figure2) for the
primary comparison group demonstrated that only younger
age, female sex, and initial waveform analysis score were
positively associated with the primary end point of survival
to hospital discharge. A positive correlation was also noted
for all secondary survival end points for female sex and
Table 3. Outcome Measures for All Study Patients and the
Primary Analysis Group
All VF, n (%)

ROSC

VF Less Than the


Threshold, n (%)

Shock
First
(n=500)

Waveform
Analysis
(n=487)

Shock
First
(n=278)

Waveform
Analysis
(n=262)

206 (41)

206 (43)

75 (27)

77 (29)

Sustained ROSC

167 (33)

158 (32)

56 (20)

57 (22)

Survival to admission

181 (36)

166 (34)

62 (23)

58 (22)

Survival to discharge

83 (17)

74 (16)

20 (7)

19 (7)

ROSC indicates return of spontaneous circulation; and VF, ventricular


fibrillation.

Freese et al SmartCPR Trial: VF Waveform Analysis for OOHCA 999


Odds Ratio

95% CI

ROSC

1.081

(0.864-1.353)

0.4944

SROSC

1.066

(0.838-1.358)

0.6022

Admission

1.183

(0.909-11.539)

0.2122

Discharge

0.67

(0.472-0.952)

0.0256

ROSC

0.352

(0.169-0.732)

0.0052

SROSC

0.387

(0.179-0.835)

0.0155

Admission

0.253

(0.109-0.588)

0.0014

Discharge

0.229

(0.074-0.714)

0.111

ROSC

1.439

(1.130-1.832)

0.0032

SROSC

1.473

(1.133-1.914)

0.0038

Admission

1.837

(1.367-2.468)

<0.0001

Discharge

1.608

(1.045-2.475)

0.0307

ROSC

3.427

(1.726-6.805)

0.0004

SROSC

3.772

(1.761-8.078)

0.0006

Admission

6.035

(2.610-13.957)

<0.0001

Discharge

2.57

(0.809-8.167)

0.1095

Age

Gender (male)

Initial Waveform Analysis Score

Improved Waveform Analysis Score with CPR

Note: For graphical representation, the odds ratios were truncated at 10.0.

Figure 2. Multivariate logistic regression analysis for survival correlation (odds ratios and 95% confidence intervals).

initial waveform score and, among those in the waveform


analysis arm, an increase in VF score after the 2-minute CPR
interval.
Further subgroup analyses were performed for the waveform analysis arm, comparing those patients for whom the VF
score increased after the CPR interval with those for whom
the score declined. ECG data were available for 204 of 262

patients, and a total of 105 of these patients (51.5%) experienced an increase in VF score before the first defibrillatory
shock (Table4). Those whose VF scores had increased after
the CPR interval had higher survival rates for all secondary
survival end points (ROSC, 41.90% versus 19.19%, P<0.001;
sustained ROSC, 33.65% versus 13.13%, P<0.001; survival to
admission, 36.89% versus 11.11%, P<0.001), although there

Table 4. Secondary Analysis Among Patients in the Waveform Analysis Arm Whose VF Scores
Decreased or Increased After the Initial CPR Interval
Waveform Analysis With
Decreased VF Score (n=99)

Waveform Analysis With


Increased VF Score (n=105)

Age, median (range), y

63 (1897)

65 (2896)

0.20

Male sex, n (%)

77 (78)

75 (71)

0.33

White race, n (%)

52 (53)

58 (55)

0.78

EMS witnessed, n (%)


Bystander witnessed, n (%)

7 (8)

5 (5)

0.56

73 (74)

80 (76)

0.87

Bystander CPR, n (%)

23 (23)

Response time, median (range), min

4.4 (215)

Initial VF score (normalized)


ROSC achieved at any point, n (%)

0.069 (0.0280.104)
19 (19)

32 (30)

0.27

4.78 (120)

0.30

0.069 (0.0190.104)
44 (42)

0.41
<0.001

SROSC

13 (13)

35 (33)

<0.001

Hospital admission

11 (11)

38 (36)

<0.001

5 (6)

11 (10)

0.13

Live discharge from hospital

CPR indicates cardiopulmonary resuscitation; EMS, emergency medical services; ROSC, return of spontaneous circulation;
SROSC, sustained return of spontaneous circulation; and VF, ventricular fibrillation. Note: for graphical representation, the odds ratios
were truncated at 10.0.

1000CirculationAugust 27, 2013


was no difference in survival to hospital discharge (11.65%
versus 5.05%; P=0.13). In patients in whom the increase in
VF score exceeded the predefined threshold value, compared
with those whose score did not achieve the threshold, there
was a significant increase in all survival end points (ROSC,
61.29% versus 25.43%, P<0.001; sustained ROSC, 53.33%
versus 18.50%, P<0.001; survival to admission, 68.97% versus 16.76%, P<0.001; survival to discharge, 20.69% versus
6.36%, P=0.02).

Discussion
In this study that used waveform analysis algorithm to guide
the initial treatment of out-of-hospital cardiac arrest patients
presenting in VF, there were no improvements in overall survival compared with a standard shock-first protocol. Despite
decades of out-of-hospital resuscitation experience involving
the use of transthoracic defibrillation, the optimal initial resuscitation strategy for VF remains undefined. Recent advances
in the understanding of the pathophysiologic basis and electrophysiology of VF have led to the suggested 3-phase model
for VF and testing of alternative treatment strategies such as
CPR before initial defibrillation, particularly when the interval
from VF onset to initial treatment is prolonged.810 However,
more recent studies designed to validate such treatment strategies have failed to replicate the results of the earlier reports
that had demonstrated benefit from a CPR-first strategy.1217
In 2011, the Resuscitation Outcomes Consortium published
the results of the largest study to date examining the question
of whether CPR before defibrillation, applied as a standard
strategy for all VF, would improve outcomes. Enrolling a total
of 2432 patients with VF arrests, the study found no difference
in survival between patients receiving immediate defibrillation and those receiving CPR before defibrillation.17 Although
our study also sought to apply CPR before initial defibrillation attempts, it differed in that the use of the VF waveform
analysis algorithm allowed the selective application of this
therapy only for those patients whose analysis suggested that
immediate defibrillation was not likely to result in ROSC. The
predefined threshold value used in this study demonstrated
the ability to distinguish between patients who are likely to
respond to immediate defibrillation and those who are not.
This would suggest that waveform analysis would be useful
for recommending alternative treatment strategies such as
CPR before defibrillation for those with poorer-quality wave
forms in an attempt to improve their probability of survival.
However we found no difference between guided treatment
for VF using a waveform analysis algorithm and a universal
shock-first protocol with respect to survival to hospital discharge or any intermediate survival end point.
In addition to supporting prior work in demonstrating that
the initial waveform analysis score for VF is strongly correlated with all survival end points, our study found a correlation between survival and later VF waveform analysis scores,
which may explain some of the disparate results in the literature on the use of delayed defibrillation. Among patients who
received CPR before defibrillation, only half demonstrated an
increase in VF score after the CPR interval. However, this is
a potentially important finding given that CPR is expected to
result in an improvement in VF waveform properties.2832

It is important to consider a number of factors that could be


responsible for the varied changes in waveform analysis score
after the CPR interval. One possibility is that any meaningful
improvement in waveform quality is a function of the quality
of CPR administered to patients with poor waveform and that,
even though defibrillation was delayed for 2 minutes, better
CPR in some patients led to better waveform and therefore
better response to defibrillation. Because we did not determine
CPR quality in this study, we cannot confirm this hypothesis.
Other possibilities remain, and although we found no difference in available patient or incident characteristics between
those whose scores improved with CPR and those whose
scores did not improve, other factors such as the patients cardiac history, prior myocardial infarctions, or degree of active
myocardial ischemia at the onset of VF may have prevented
an improvement in VF quality during the CPR interval.3336 It
is also possible that the improvement in VF score was prognostic but not a direct result of the CPR interval itself, and as
the main study result indicates, there is no survival benefit of
administering 2 minutes of CPR before attempting defibrillation. Because our study was limited in that it was not designed
to provide such comparative analyses for the subgroup with
this postrandomization measure or to analyze survival differences between this population and any control group,
conclusions related to improvement in VF score cannot be
specifically supported, and further study is recommended to
address these questions.
This study has several additional limitations. Although
the protocol for resuscitation and postresuscitation prehospital care was standardized at each study site, there was no
standardization of postresuscitation inpatient care at either
site, including the lack of routine use of therapeutic hypothermia at either site during the study period. Second, our
study used only 1 waveform analysis parameter, and it is
possible that other waveform characteristics may have
yielded different results.
Additionally, there was no ability to measure CPR performance during the study, allowing the possibility that the lack
of any benefit in the waveform analysis group was the result
of poor CPR performance rather than a failure of the analysis.
Given the known variability in the quality of CPR that is provided in the prehospital setting, including the frequent failure
to deliver effective or any compressions, high-quality CPR
may not have been provided.37,38 Without quality compressions, the likelihood of successful defibrillation declines.39,40
Finally, the inability to evaluate CPR performance limits the
ability to further interpret the subgroup analysis that demonstrated improved outcomes among those patients treated with
CPR before defibrillation for whom an improvement in VF
score was noted after the 2-minute CPR interval. Without
such a measurement, it is impossible to discern whether this
improvement was a result of and directly related to highquality CPR performance or merely a prognostic value that
was unrelated to CPR delivery.

Conclusions
VF waveform analysis to selectively guide the initial management of out-of-hospital cardiac arrest patients presenting in VF did not result in improved overall survival. We

Freese et al SmartCPR Trial: VF Waveform Analysis for OOHCA 1001


demonstrated the prognostic value of waveform analysis
scores as a determinant of resuscitation outcomes, and we
further identified a subgroup of patients for whom guided
management via waveform analysis was associated with
improved waveform characteristics and for whom additional
study appears warranted to understand the factors associated
with this finding. Further studies are needed to objectively
measure the CPR being performed, to assess its relation to
changes in the VF characteristics, and to examine the use of
other strategies to optimize the initial resuscitation strategies
for patients with low-quality VF.

Acknowledgments
We thank the following individuals for their time, effort, and assistance in completing this study: Emily Mydynski, Meera Shenoy,
Chris Choe, Stacy Gehman, James Russell, Bradley Kaufman, Doug
Isaacs, Andrew Werner, Gilbert Caicedo, Dulce McCorvey, Freda
Scott, James Braun, Savahanna Lien, Paul Barbara, Kevin Munjal,
Fionna Moore, and, most important, the men and women of the Fire
Department of New York and the London Ambulance Service.

Sources of Funding
The Fire Department of New York and London Ambulance Service
received grant support for the study from Philips Healthcare.

Disclosures
None.

References
1. Zoll PM, Linenthal AJ, Gibson W, Paul MH, Norman LR. Termination of
ventricular fibrillation in man by externally applied electric countershock.
N Engl J Med. 1956;254:727732.
2. Nagel EL, Hirschman JC, Mayer PW, Frank D. Telemetry of physiologic data: an aid to fire rescue in a metropolitan area. South Med J.
1968;61:598601.
3. Grace WJ, Chadbourn JA. The mobile coronary care unit. Dis Chest.
1969;55:452455.
4. Cobb LA, Conn RD, Samson WE, Philbin JE. Early experiences in the
management of sudden death with a mobile intensive/coronary care unit.
Circulation. 1970;42:S144.
5. Berdowski J, Berg RA, Tijssen JG, Koster RW. Global incidences of outof-hospital cardiac arrest and survival rates: systematic review of 67 prospective studies. Resuscitation. 2010;81:14791487.
6. Rea TD, Eisenberg MS, Sinibaldi G, White RD. Incidence of EMStreated out-of-hospital cardiac arrest in the United States. Resuscitation.
2004;63:1724.
7. Guidelines 2000 for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation. 2000;102:I1I403.
8. Weisfeldt ML, Becker LB. Resuscitation after cardiac arrest: a 3-phase
time-sensitive model. JAMA. 2002;288:30353038.
9. Cobb LA, Fahrenbruch CE, Walsh TR, Copass MK, Olsufka M, Breskin
M, Hallstrom AP. Influence of cardiopulmonary resuscitation prior to defibrillation in patients with out-of-hospital ventricular fibrillation. JAMA.
1999;281:11821188.
10. Wik L, Hansen TB, Fylling F, Steen T, Vaagenes P, Auestad BH, Steen
PA. Delaying defibrillation to give basic cardiopulmonary resuscitation
to patients with out-of-hospital ventricular fibrillation: a randomized trial.
JAMA. 2003;289:13891395.
11. International Liaison Committee on Resuscitation. International consensus on CPR and ECC science with treatment recommendations, part 3:
defibrillation. Circulation. 2005;112:III-17III-24.
12. Hazinski MF, Nolan JP, Billi JE, Bttiger BW, Bossaert L, de Caen AR,
Deakin CD, Drajer S, Eigel B, Hickey RW, Jacobs I, Kleinman ME,
Kloeck W, Koster RW, Lim SH, Mancini ME, Montgomery WH, Morley
PT, Morrison LJ, Nadkarni VM, OConnor RE, Okada K, Perlman
JM, Sayre MR, Shuster M, Soar J, Sunde K, Travers AH, Wyllie J,
Zideman D. Part 1: executive summary: 2010 International Consensus
on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care

Science With Treatment Recommendations. Circulation. 2010;122(suppl


2):S250S275.
13. Jacobs V, Nadkarni J, Bahr RA, Berg JE, Billi L Bossaert, Cassan P,
Coovadia A, DEste K, Finn J, Halperin H, Handley A, Herlitz J, Hickey
R, Idris A, Kloeck W, Larkin GL, Mancini ME, Mason P, Mears G,
Monsiurs K, Montgomery W, Morley P, Nichol G, Nolan J, Okada
K, Perlman J, Shuster M, Steen PA, Stertz F, Tibballs J, Timerman
S, Triutt T, Zideman D. Cardiac arrest and cardiopulmonary resuscitation outcome reports: update and simplification of the Utstein
templates for resuscitation registries: a statement for healthcare professionals from a task force of the International Liaison Committee on
Resuscitation (American Heart Association, European Resuscitation
Council, Australian Resuscitation Council, New Zealand Resuscitation
Council, Heart and Stroke Foundation of Canada, InterAmerican Heart
Foundation, Resuscitation Council of Southern Africa). Resuscitation.
2004;63:233249.
14. Jacobs IG, Finn JC, Oxer HF, Jelinek GA. CPR before defibrillation in
out-of-hospital cardiac arrest: a randomized trial. Emerg Med Australas.
2005;17:3945.
15. Meier P, Baker P, Jost D, Jacobs I, Henzi B, Knapp G, Sasson C. Chest compressions before defibrillation for out-of-hospital cardiac arrest: a metaanalysis of randomized controlled clinical trials. BMC Med. 2010;8:52.
16. Simpson PM, Goodger MS, Bendall JC. Delayed versus immediate defibrillation for out-of-hospital cardiac arrest due to ventricular fibrillation:
a systematic review and meta-analysis of randomised controlled trials.
Resuscitation. 2010;81:925931.
17. Stiell IG, Nichol G, Leroux BG, Rea TD, Ornato JP, Powell J, Christenson J,
Callaway CW, Kudenchuk PJ, Aufderheide TP, Idris AH, Daya MR,
Wang HE, Morrison LJ, Davis D, Andrusiek D, Stephens S, Cheskes S,
Schmicker RH, Fowler R, Vaillancourt C, Hostler D, Zive D, Pirrallo RG,
Vilke GM, Sopko G, Weisfeldt M; ROC Investigators. Early versus later
rhythm analysis in patients with out-of-hospital cardiac arrest. N Engl J
Med. 2011;365:787797.
18. Lin LY, Lo MT, Ko PC, Lin C, Chiang WC, Liu YB, Hu K, Lin JL,
Chen WJ, Ma MH. Detrended fluctuation analysis predicts successful
defibrillation for out-of-hospital ventricular fibrillation cardiac arrest.
Resuscitation. 2010;81:297301.
19. Lightfoot CB, Nremt-P, Callaway CW, Hsieh M, Fertig KC, Sherman LD,
Menegazzi JJ. Dynamic nature of electrocardiographic waveform predicts
rescue shock outcome in porcine ventricular fibrillation. Ann Emerg Med.
2003;42:230241.
20. Strohmenger HU, Lindner KH, Brown CG. Analysis of the ventricular
fibrillation ECG signal amplitude and frequency parameters as predictors
of countershock success in humans. Chest. 1997;111:584589.
21. Marn-Pernat A, Weil MH, Tang W, Pernat A, Bisera J. Optimizing timing
of ventricular defibrillation. Crit Care Med. 2001;29:23602365.
22. Goto Y, Suzuki I, Inaba H. Frequency of ventricular fibrillation as predictor of one-year survival from out-of-hospital cardiac arrests. Am J Cardiol.
2003;92:457459.
23. Indik JH, Allen D, Shanmugasundaram M, Zuercher M, Hilwig RW, Berg RA,
Kern KB. Predictors of resuscitation in a swine model of ischemic and
nonischemic ventricular fibrillation cardiac arrest: superiority of amplitude
spectral area and slope to predict a return of spontaneous circulation when
resuscitation efforts are prolonged. Crit Care Med. 2010;38:23522357.
24. Snyder DE, White RD, Jorgenson DB. Waveform analysis of ven
tricular fibrillation to predict defibrillation. Curr Opin Crit Care.
2005;11:192199.
25. Chapman FW, Rea TD. A new piece in the unfinished ventricular fibrillation analysis puzzle. Crit Care Med. 2008;36:22102212.
26. Ristagno G, Gullo A. Is ventricular fibrillation waveform analysis suitable
for optimizing time of ventricular defibrillation? Yes it is. Crit Care Med.
2007;35:18041805.
27. Snyder DE, White RD, Jorgenson DB. Outcome prediction for guidance
of initial resuscitation protocol: shock first or CPR first. Resuscitation.
2007;72:4551.
28. Young C, Bisera J, Gehman S, Snyder D, Tang W, Weil MH. Amplitude
spectrum area: measuring the probability of successful defibrillation as
applied to human data. Crit Care Med. 2004;32(suppl):S356S358.
29. Noc M, Weil MH, Gazmuri RJ, Sun S, Biscera J, Tang W. Ventricular
fibrillation voltage as a monitor of the effectiveness of cardiopulmonary
resuscitation. J Lab Clin Med. 1994;124:421426.
30. Berg RA, Hilwig RW, Ewy GA, Kern KB. Precountershock cardiopulmonary resuscitation improves initial response to defibrillation from prolonged ventricular fibrillation: a randomized, controlled swine study. Crit
Care Med. 2004;32:13521357.

1002CirculationAugust 27, 2013


31. Menegazzi JJ, Wang HE, Lightfoot CB, Fertig KC, Chengelis NL, Sherman LD,
Callaway CW. Immediate defibrillation versus interventions first in a swine
model of prolonged ventricular fibrillation. Resuscitation. 2003;59:261270.
32. Eftestl T, Wik L, Sunde K, Steen PA. Effects of cardiopulmonary resuscitation on predictors of ventricular fibrillation defibrillation success during
out-of-hospital cardiac arrest. Circulation. 2004;110:1015.
33. Kern KB, Garewal HS, Sanders AB, Janas W, Nelson J, Sloan D, Tacker WA,
Ewy GA. Depletion of myocardial adenosine triphosphate during prolonged untreated ventricular fibrillation: effect on defibrillation success.
Resuscitation. 1990;20:221229.
34. Maldonado FA, Weil MH, Tang W, Bisera J, Gazmuri RJ, Johnson B,
DAlessio A. Myocardial hypercarbic acidosis reduces cardiac resuscitability. Anesthesiology. 1993;78:343352.
35. Indik JH, Donnerstein RL, Berg RA, Hilwig RW, Berg MD, Kern KB.
Ventricular fibrillation frequency characteristics are altered in acute myocardial infarction. Crit Care Med. 2007;35:11331138.

36. Ristagno G, Li Y, Tang W, Sun S, Weil MH. Comparison between ischemic and electrically induced ventricular fibrillation. Crit Care Med.
2006;34:S432S437.
37. Wik L, Kramer-Johansen J, Myklebust H, Sreb H, Svensson L, Fellows B,
Steen PA. Quality of cardiopulmonary resuscitation during out-of-hospital cardiac arrest. JAMA. 2005;293:299304.
38. Olasveengen TM, Wik L, Kramer-Johansen J, Sunde K, Pytte M, Steen PA.
Is CPR quality improving? A retrospective study of out-of-hospital cardiac
arrest. Resuscitation. 2007;75:260266.
39. Edelson DP, Abella BS, Kramer-Johansen J, Wik L, Myklebust H, Barry AM,
Merchant RM, Hoek TL, Steen PA, Becker LB. Effects of compression
depth and pre-shock pauses predict defibrillation failure during cardiac
arrest. Resuscitation. 2006;71:137145.
40. Eftestl T, Sunde K, Steen PA. Effects of interrupting precordial compressions on the calculated probability of defibrillation success during out-ofhospital cardiac arrest. Circulation. 2002;105:22702273.

Clinical Perspective
In the 3 most recent iterations of resuscitation guidelines, recommendations for ventricular fibrillation (VF) have varied,
including immediate shock, delayed defibrillation, and most recently the recognition of inconsistent evidence for the most
appropriate initial resuscitation strategy for VF. In this study, we sought to implement an automatic external defibrillator
based VF waveform algorithm for the initial treatment of VF among out-of-hospital cardiac arrest patients compared with
a standard shock-first protocol. Overall survival between the 2 study arms was equivalent; thus, this study does not provide
definitive evidence to clarify the inconsistency concerning the initial treatment of VF. Among patients with low-quality VF
who were given 2 minutes of cardiopulmonary resuscitation before initial defibrillation, despite published data suggesting
that an improvement in VF score is to be expected, there was wide variation among those patients with regard to the resulting VF score, ranging from significant improvement to marked decline. Although these changes in VF score positively correlated to outcome, the study design did not allow the identification of causative factors or comparison with the shock-first
arm. These results suggest a potential for the future use of such technologies to guide prognostic or treatment decisions,
emphasize the need to address the quality of CPR during any resuscitation, and yield a number of additional questions for
future research.