The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Original Article

Early Cardiopulmonary Resuscitation

in Out-of-Hospital Cardiac Arrest
Ingela HasselqvistAx, R.N., Gabriel Riva, M.D., Johan Herlitz, M.D., Ph.D.,
Mrten Rosenqvist, M.D., Ph.D., Jacob Hollenberg, M.D., Ph.D.,
Per Nordberg, M.D., Ph.D., Mattias Ringh, M.D., Ph.D., Martin Jonsson, B.Sc.,
Christer Axelsson, R.N., Ph.D., Jonny Lindqvist, M.Sc., Thomas Karlsson, B.Sc.,
and Leif Svensson, M.D., Ph.D.

A BS T R AC T
BACKGROUND

Three million people in Sweden are trained in cardiopulmonary resuscitation

(CPR). Whether this training increases the frequency of bystander CPR or the
survival rate among persons who have out-of-hospital cardiac arrests has been
questioned.
METHODS

We analyzed a total of 30,381 out-of-hospital cardiac arrests witnessed in Sweden

from January 1, 1990, through December 31, 2011, to determine whether CPR was
performed before the arrival of emergency medical services (EMS) and whether
early CPR was correlated with survival.
RESULTS

CPR was performed before the arrival of EMS in 15,512 cases (51.1%) and was not
performed before the arrival of EMS in 14,869 cases (48.9%). The 30-day survival
rate was 10.5% when CPR was performed before EMS arrival versus 4.0% when
CPR was not performed before EMS arrival (P<0.001). When adjustment was made
for a propensity score (which included the variables of age, sex, location of cardiac arrest, cause of cardiac arrest, initial cardiac rhythm, EMS response time,
time from collapse to call for EMS, and year of event), CPR before the arrival of
EMS was associated with an increased 30-day survival rate (odds ratio, 2.15; 95%
confidence interval, 1.88 to 2.45). When the time to defibrillation in patients who
were found to be in ventricular fibrillation was included in the propensity score,
the results were similar. The positive correlation between early CPR and survival
rate remained stable over the course of the study period. An association was also
observed between the time from collapse to the start of CPR and the 30-day survival rate.

From the Center for Resuscitation Science, Solna (I.H.-A., G.R., J. Hollenberg,
P.N., M. Ringh, M.J., L.S.), and the Department of Clinical Sciences, Section of
Cardiology, Danderyd Hospital, Danderyd
(M. Rosenqvist), Karolinska Institutet,
Stockholm, the Center for Pre-Hospital
Research in Western Sweden (J. Herlitz)
and the School of Health Sciences (C.A.),
University of Bors, Bors, and Sahlgrenska University Hospital (J. Herlitz), the
Institute of Internal Medicine, Department of Metabolism and Cardiovascular
Research, Sahlgrenska University Hospital (J.L.), and the Center for Applied Biostatistics, Sahlgrenska Academy at the
University of Gothenburg (T.K.), Gothenberg all in Sweden. Address reprint
requests to Dr. Herlitz at the Center for
Pre-Hospital Research, University of Bors,
SE-501 90 Bors, Sweden, or at johan
[email protected].
Ms. Hasselqvist-Ax and Dr. Riva contributed equally to this article.
N Engl J Med 2015;372:2307-15.
DOI: 10.1056/NEJMoa1405796
Copyright 2015 Massachusetts Medical Society.

CONCLUSIONS

CPR performed before EMS arrival was associated with a 30-day survival rate after
an out-of-hospital cardiac arrest that was more than twice as high as that associated with no CPR before EMS arrival. (Funded by the Laerdal Foundation for Acute
Medicine and others.)

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

2307

The

n e w e ng l a n d j o u r na l

ut-of-hospital cardiac arrest is a

major public health concern, given that
there are approximately 420,000 cases in
the United States and 275,000 cases in Europe
annually.1-3 Decreasing the time to treatment is
crucial for improving outcomes in cases of cardiac arrest.4,5 As stated in American and European guidelines, the most important response
measures that currently can be taken outside a
hospital setting are recognizing early that a cardiac arrest is occurring, placing an alarm call,
performing cardiopulmonary resuscitation (CPR),
and performing defibrillation.6,7 Globally, CPR is
taught to millions of people each year. In Sweden,
more than 3 million people (of a population of
9.7 million)8 have undergone CPR training during the past three decades (Claesson A, Swedish
Resuscitation Council: personal communication).
However, in recent years, the value of bystander
CPR has been debated in the medical community.9,10 A major source of concern is the lack of
a randomized clinical trial to show that bystander CPR is life-saving. Scientific documentation of the value of bystander CPR with regard to
survival after an out-of-hospital cardiac arrest
has been based on studies in animals and studies involving registry data, in which the preponderance of evidence has shown an association
between CPR and survival.11,12 The main effect of
CPR is probably indirect, in that it may prolong
the time window for defibrillation.13,14 The sooner defibrillation can be performed, the better the
chance of survival.15,16
In the current study, our primary aim was to
assess whether CPR initiated before the arrival
of emergency medical services (EMS) was associated with an increase in the 30-day survival
rate among persons who collapsed due to an
out-of-hospital cardiac arrest, with adjustment
for several confounders, including the age and
sex of the patient, the place and cause of the
cardiac arrest, the initial cardiac rhythm, the EMS
response time, the time from collapse to the call
for EMS, and the time from collapse to defibrillation. A secondary aim was to assess the association between the estimated time from collapse
to the start of CPR and the 30-day survival rate.

of

m e dic i n e

recorded in the Swedish Cardiac Arrest Registry

from January 1, 1990, through December 31,
2011. Nonwitnessed cases and cases witnessed
only by the EMS crew were excluded.
Swedish Cardiac Arrest Registry

The Swedish Cardiac Arrest Registry now includes

more than 90% of all persons who had an outof-hospital cardiac arrest in Sweden and in whom
CPR was attempted. It is a national quality registry funded by the Swedish Association of Local
Authorities and Regions. All EMS centers in
Sweden report data to the registry in accordance
with the Utstein style17; the reporting includes
completion of a standard form with detailed
descriptions of the circumstances, time delays,
and forms of intervention in cases of out-of-hospital cardiac arrest in which CPR was attempted.
The form is reviewed and completed by the head
physician of the ambulance organization in each
county. This registry has been thoroughly described elsewhere.18
Cases are included in the registry only if there
was no breathing and no sign of circulation in
the patient and if CPR, defibrillation, or both
were started. In recent years, there has been a
progressive increase in retrospective monitoring,
including cross-checking between the registry
and the local EMS registry.
Ethics approval for the use of data from the
Swedish Cardiac Arrest Registry was obtained
from the Regional Ethics Board, Gothenburg,
Sweden. The protocol for this registry study is
available with the full text of this article at
NEJM.org.
Dispatch and Ambulance Organization

In Sweden, there are approximately 850 ambulances serving 9.7 million citizens.19 There is a
two-tiered EMS system in Sweden for responses
to all medical emergencies: the first tier consists
of EMS units that can respond to an out-ofhospital cardiac arrest with a basic level of life
support, and units in the second tier are able to
provide an advanced level of life support. Data
on EMS units in both tiers are included in the
Swedish Cardiac Arrest Registry and in this
analysis. Sweden has 15 dispatch centers, all of
which are similar in their organization and
Me thods
emergency-call processing. The dispatcher uses a
Population
standard protocol with a specific questionnaire
We included all cases of EMS-treated and by- for the identified emergency. In cases of suspected
stander-witnessed out-of-hospital cardiac arrests cardiac arrest, an ambulance is dispatched, and

2308

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

Early Cardiopulmonary Resuscitation in Cardiac Arrest

dispatchers are instructed to offer the caller a

chance to perform telephone-assisted CPR (i.e.,
if the caller is not trained in CPR, instructions
on how to perform CPR are provided). This system was introduced in 1998.20 The guidelines for
CPR changed during the study period; major
alterations occurred in 2000, 2005, and 2010 in
response to changes in international guidelines.7,21,22 These changes included the elimination of the pulse check and more focus on chest
compressions, with more and deeper compressions used.
In 2006, a dual-dispatch system was implemented in Stockholm that engages firefighters
and police in addition to EMS. However, in most
parts of Sweden, this dual-dispatch system was
not implemented until 2011; we estimate that
approximately 5% of the cases that were included
in this study occurred after that system was
implemented.
Initiation of CPR

Data on the start of CPR included whether CPR

was started before the arrival of EMS (yes or no)
and the estimated time between the patients
collapse and the initiation of CPR. In addition,
patients who underwent CPR during the latter
part of the study (January 1, 2009, through December 31, 2011) were categorized in one of two
groups patients who did and those who did
not undergo telephone-assisted CPR.
Times of Events and Procedures

The times of events and procedures (day, hour,

and minute) were reported for the patients collapse, the call to the dispatch center, the dispatch of EMS, the arrival of EMS at the scene,
the start of CPR (either before or after the arrival
of EMS), and defibrillation (if the patient had
ventricular fibrillation or pulseless ventricular
tachycardia as the first recorded arrhythmia).
The time of the emergency call, the time of dispatch, and the time of the arrival of EMS were
recorded digitally and collected from the dispatch center, whereas the times of collapse, the
start of CPR, and defibrillation were based on
statements from witnesses, first responders, and
EMS clinicians.

30-day survivors. Cerebral function is described

with the use of the cerebral performance categories score, in which category 1 or 2 indicates
favorable cerebral function, category 3 or 4 indicates poor cerebral function, and category 5 indicates brain death (Table S1 in the Supplementary Appendix, available at NEJM.org).
Statistical Analysis

Univariate analyses were performed with the use

of Fishers exact test for dichotomous variables
and the MannWhitney U test for continuous
variables. Logistic regression was used for the
calculation of odds ratios with corresponding
confidence intervals and for analyses of interaction. A propensity analysis was performed to adjust for potential confounders. First, a propensity
score for receipt of CPR before EMS arrival was
calculated by means of multiple logistic regression. The variables included in the score were
age, sex, cause of cardiac arrest, place of cardiac
arrest, initial cardiac rhythm, year of cardiac arrest, time from collapse to the call for EMS, time
from the call for those services to the arrival of
EMS, and among patients with an initial rhythm
of ventricular tachycardia or ventricular fibrillation, time from collapse to defibrillation. Using
forward stepwise selection, we tested for interactions between the variables and included them
in the score if the P value was less than 0.20. The
score was then used as an adjustment factor in a
logistic-regression models.
All tests were two-sided, and a P value of less
than 0.05 for the primary objective and of less
than 0.01 for all other analyses was considered
to indicate statistical significance. SAS software,
version 9.3 (SAS Institute), was used for all statistical analyses.

R e sult s
Cardiac Arrest and CPR

Data on 61,781 patients who had an out-of-hospital cardiac arrest and underwent CPR were
reported to the registry during the period from
1990 through 2011. Among these incidents of
cardiac arrest and CPR, 7898 were witnessed by
EMS responders (12.8%), and 17,935 (29.0%) were
not witnessed. Information on whether the inCerebral Function at Discharge
cident was witnessed was missing in 4585 cases
Since November 2008, the Swedish Cardiac Arrest (7.4%); these cases were excluded from the
Registry has collected information on cerebral analysis.
function at hospital discharge in a subgroup of
Among the patients who had a bystandern engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

2309

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 1. Baseline Characteristics of the Patients.*

Variable
Median age (10th to 90th percentile) yr

CPR Started after

Arrival of EMS
(N=14,869)

CPR Started before

Arrival of EMS
(N=15,512)

P Value

74 (5486)

69 (4684)

<0.001

Female sex %

30.2

26.8

<0.001

Cardiac cause of cardiac arrest %

73.4

72.4

0.04

Collapse at home %

73.2

55.5

<0.001

VF or VT as initial ECG rhythm %

30.7

41.3

<0.001
<0.001

Median intervals (10th to 90th percentile) min

Collapse to call for EMS

4 (011)

3 (010)

Call for EMS to arrival of EMS

6 (315)

8 (320)

<0.001

Collapse to start of CPR

11 (523)

4 (017)

<0.001

Patients in VF or VT no.
Median time from collapse to defibrillation (10th
to 90th percentile) min

4194

5900

11 (621)

13 (724)

<0.001

* The percentages of patients with missing data for each variable were as follows: age, 3.3%; sex, 3.2%; cause of cardiac
arrest, 6.7%; place of collapse, 0.8%; initial electrocardiographic (ECG) rhythm, 7.9%; delay from collapse to call for
emergency medical services (EMS), 20.0%; delay from call for EMS to arrival of EMS, 6.4%; delay from collapse to start
of CPR, 21.2%; and delay from collapse to defibrillation, 5.3%. The proportions of missing data were similar between
the two groups. P value calculations included data only from those patients who did not have missing data. CPR denotes cardiopulmonary resuscitation, VF ventricular fibrillation, and VT ventricular tachycardia.

witnessed out-of-hospital cardiac arrest, 682

(2.2%) were not included because of a lack of
information on whether CPR had been given
before the arrival of EMS. Of the remaining
30,681 patients, 300 (1.0%) were not included
because information on 30-day survival was
missing. Among the 30,381 patients who had a
bystander-witnessed out-of-hospital cardiac arrest
and recorded data on both the start of CPR and
survival, 15,512 (51.1%) received CPR before the
arrival of EMS, and 14,869 (48.9%) did not (Fig.
S1 in the Supplementary Appendix).
Characteristics of the Patients

As compared with patients who did not undergo

CPR before EMS arrival, patients who underwent
CPR before EMS arrival were younger, and their
collapse was less likely to have occurred at home;
this group also included fewer women (Table1).
Patients in the group that underwent CPR before
EMS arrival were more often found to be in ventricular fibrillation, despite no greater frequency
in that group of cardiac causes as the assumed
underlying cause of cardiac arrest (thehigher
frequency of ventricular fibrillation was also

2310

found when adjustment was made for confounders [Table S2 in the Supplementary Appendix]).
The times from collapse to the call to EMS and
from collapse to the start of CPR were shorter in
the group that underwent CPR before the arrival
of EMS. However, the time from the call for EMS
until the arrival of EMS and the time from collapse to defibrillation if the patient had been
found to be in ventricular fibrillation were longer in this group. The prolonged EMS response
time in the group that received CPR before EMS
arrival was found regardless of whether the patient had initial arrhythmia and regardless of
the place in which the out-of-hospital cardiac
arrest occurred (Table S3 in the Supplementary
Appendix). Changes over time in the number of
persons trained in CPR, the performance of
early CPR, and survival are shown in Figure1.
Survival

The 30-day survival rate was 10.5% among patients who underwent CPR before EMS arrival,
as compared with 4.0% among those who did
not (P<0.001). There was a significantly higher
survival rate among patients who received CPR

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

Early Cardiopulmonary Resuscitation in Cardiac Arrest

Telephone-Assisted CPR and Survival

In a subgroup analysis (with data collected from

January 1, 2009, through December 31, 2011), a
total of 35% of patients who underwent CPR
before EMS arrival had CPR performed by a person who received telephone-assisted CPR instructions (Table S4 in the Supplementary Appendix).
These patients had a 30-day survival rate of
10.9%, as compared with a rate of 15.4% among
those who received early CPR performed by a
person who did not receive telephone-assisted
instructions (P<0.001).
Post Hoc Propensity Analysis

A post hoc propensity score was calculated that

took into account the year of the out-of-hospital
cardiac arrest and all the variables listed in Table1 (except the time from collapse to defibrillation) and relevant interactions among them.
After adjustment for the propensity score, the
difference in the 30-day survival rate between
the two groups (CPR before vs. after EMS arrival)
remained highly significant (odds ratio, 2.15; 95%
confidence interval [CI], 1.88 to 2.45; P<0.001)
(Table3). Because data were missing for several
variables, we also performed multiple imputation by means of the Markov chain Monte Carlo
method, which resulted in a corresponding odds
ratio of 2.05 (95% CI, 1.84 to 2.28). In an analy-

A
CPR before EMS Arrival (%)

75

T-CPR

65
55
45

2.75

No. of persons
trained

2.25

CPR before
EMS arrival

1.75
1.25

35

0.75

25

0.25

No. of Persons Trained (millions)

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

B
25

T-CPR

20

Survival Rate (%)

before EMS arrival in all the subgroups analyzed

(Fig.2). A significant interaction was found between receipt of CPR before EMS arrival and
both the sex of the patient and the place in
which the cardiac arrest occurred (P=0.002 and
P=0.001, respectively, for the interaction). The
increase in the survival rate among patients who
underwent CPR before EMS arrival was more
marked among men than among women and
more marked when the out-of-hospital cardiac
arrest occurred outside the patients home (e.g.,
in a public location) than when it occurred inside
the patients home.
Table2 shows the 30-day survival rate in relation to the time between the patients collapse
and the start of CPR. Overall, there was a significant association between the time to the start of
CPR and the survival rate: the survival rate decreased with an increase in the time to the start
of CPR. The association was found in all the
subgroups analyzed.

CPR before
EMS Arrival

15
10
5
0

No CPR before
EMS Arrival

1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010

Figure 1. Changes over Time in CPR Training, the Performance of Early

CPR, and Survival Rates.
Panel A shows the number of persons in Sweden who were trained in cardiopulmonary resuscitation (CPR) and the proportion of patients in whom
CPR was started before the arrival of emergency medical services (EMS).
Panel B shows the survival rate when CPR was given and when CPR was
not given before EMS arrival. In both panels, the vertical line (T-CPR) indicates the year in which telephone-assisted CPR was introduced in Sweden.

sis of the subgroup of patients who were found

to be in ventricular fibrillation in which the time
between collapse and defibrillation was included
in the propensity score, there was also a significant difference between the groups (odds ratio,
2.27; 99% CI, 1.84 to 2.81; P<0.001).
Cerebral Function among Survivors

In total, there were 474 patients who survived for

30 days after collapse and for whom information
on the cerebral performance categories score,
for which higher scores indicate greater disability, was available. At the time of discharge from
the hospital, 81% of these patients had a score
of category 1, 14% a score of category 2, 5% a
score of category 3, less than 1% a score of category 4, and less than 1% a score of category 5.

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

2311

The

Subgroup

n e w e ng l a n d j o u r na l

Survival Rate Survival Rate Patients with

No CPR before
CPR before
No CPR before
EMS Arrival
EMS Arrival
EMS Arrival
%

All patients
Age
72 yr
>72 yr
Sex
Female
Male
Cause of cardiac arrest
Cardiac
Noncardiac
Location of cardiac arrest
At home
Other location
Initial ECG rhythm
VF or VT
Asystole or PEA
Year of cardiac arrest
19901995
19962001
20022007
20082011

of

m e dic i n e

Patients with
CPR before
EMS Arrival

Odds Ratio (95% CI)

no.

4.0

10.5

14,869

15,512

2.80 (2.473.18)

5.6
2.9

12.7
7.9

6,405
8,011

9,043
5,929

2.44 (2.072.87)
2.84 (2.303.50)

4.1
4.1

8.3
11.5

4,343
10,036

4,053
11,085

2.14 (1.672.73)
3.02 (2.603.51)

4.2
3.4

11.5
8.5

10,205
3,694

10,452
3,993

2.94 (2.533.41)
2.62 (1.993.45)

3.1
6.7

5.9
16.3

10,783
3,949

8,544
6,855

1.97 (1.642.37)
2.72 (2.263.27)

9.4
1.5

20.1
3.2

4,194
9,487

5,900
8,394

2.43 (2.072.85)
2.12 (1.622.78)

3.8
3.0
4.6
5.5

9.7
6.9
10.7
13.4

3,892
4,697
3,562
2,562

2,629
3,563
3,923
5,278

2.75 (2.093.62)
2.38 (1.803.14)
2.46 (1.933.14)
2.64 (2.073.88)
0.5

1.0

1.5

2.0

2.5

3.0

3.5

Figure 2. Subgroup Analysis of Survival Rates.

ECG denotes electrocardiographic, PEA pulseless electrical activity, VF ventricular fibrillation, and VT ventricular tachycardia.

Discussion
In the current study, we found increased survival
rates among patients who had an out-of-hospital
cardiac arrest and underwent CPR before the
arrival of EMS. This finding is consistent with
those of other studies.23-25 We also found that if
bystander CPR was started before the arrival of
EMS, the emergency call was initiated more
rapidly, which suggests that bystanders with CPR
training are better than bystanders without such
training at recognizing that a cardiac arrest is
occurring and taking action. Despite these findings, the time from collapse to the arrival of
EMS, as well as the time from collapse to first
defibrillation, was longer in cases in which CPR
was given before EMS arrived. Thus, the survival
rate among patients who received CPR before
EMS arrived was increased despite the fact that
the time to defibrillation was prolonged.
There are several possible explanations for
the observation that the survival rate was increased even when the time to defibrillation was
prolonged; one explanation is that CPR maintains
2312

a certain degree of circulation, which may prevent ventricular fibrillation from deteriorating to
asystole before EMS arrives. A possible explanation for the longer EMS response times in the
group that received early CPR is that having a
longer period between the call to EMS and EMS
arrival may increase the likelihood that CPR is
started before EMS arrival (e.g., if neighbors call
in, someone able to perform CPR passes by, or
telephone-assisted CPR is provided). Finally, some
EMS systems in Sweden dispatch the fire department when there is no ambulance available,
which results in a longer time between the call
to EMS and EMS arrival.
Patients who underwent CPR before the arrival
of EMS were less likely to have collapsed at home;
this finding is consistent with previous findings
that are best explained by the observation that a
preponderance of in-home, out-of-hospital cardiac arrests are witnessed by elderly persons in the
home. Such persons are most often not educated
in CPR or are not capable of performing it.26,27
The increase in the survival rate if CPR was
given before the arrival of EMS was greater when

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

Early Cardiopulmonary Resuscitation in Cardiac Arrest

Table 2. The Rate of 30-Day Survival in Relation to the Time from Collapse to the Start of CPR.
No. of Patients
with Data

Subgroup

Survival Rate According to Time

from Collapse to Start of CPR*
03 min

48 min

914 min

>14 min

percent
All patients

23,931

15.6

8.7

4.0

0.9

72 yr

12,169

19.2

11.0

5.6

1.5

>72 yr

10,968

10.8

6.7

2.6

0.4

Age

Sex
Female

6,424

12.2

7.5

4.5

1.0

16,842

16.9

9.4

3.8

0.9

Yes

16,534

16.6

9.5

4.0

0.9

No

5,979

12.9

6.5

3.8

1.0

Male
Cardiac cause

Location of cardiac arrest

At home

15,179

9.2

6.0

3.3

0.7

8,579

21.6

12.9

5.7

1.7

Yes

8,213

26.4

15.6

7.9

2.7

No

13,941

4.6

3.1

1.6

0.4

19901995

5,068

16.3

9.6

3.5

0.7

19962001

6,397

11.9

6.7

3.1

0.5

20022007

5,605

15.2

9.9

4.8

0.8

20082011

6,666

17.0

8.9

4.9

2.0

Other location
VF or VT as initial ECG rhythm

Year of cardiac arrest

* P<0.001 in all subgroups for the association between the time from collapse to the start of CPR and 30-day survival.

the out-of-hospital cardiac arrest took place outside the patients home, possibly because in that
setting, bystanders are often younger and may be
more likely to be trained in CPR. Our finding that
CPR before the arrival of EMS was more likely to
increase the survival rate among men than
among women could be explained by findings in
one study that more women who have an out-ofhospital cardiac arrest have them at home.28
The importance of early CPR was further supported by the strong association we observed
between the time from collapse to the start of
CPR and the 30-day survival rate. This association was found among all patients and among
all subgroups that we evaluated.
There was an increase over time in the proportion of persons who received CPR before the
arrival of EMS when they had an out-of-hospital

cardiac arrest. We speculate that widespread

teaching of CPR to laypeople, the implementation of telephone-assisted CPR, and more frequent dispatch of firefighters and the police
may have contributed to the increase. Other
strategies in addition to population-based CPR
training are also important. In this issue of the
Journal, Ringh et al.29 report increases in the
rates of bystander-initiated CPR associated with
the use of a mobile-phone positioning system
for dispatching CPR-trained lay volunteers to
respond to nearby out-of-hospital cardiac arrests.
Our study has some limitations. First, all time
data were obtained from the Swedish Cardiac
Arrest Registry, and in this registry, the time of
collapse is estimated, as is the time of the start
of CPR, although other time data for exam-

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

2313

The

n e w e ng l a n d j o u r na l

of

m e dic i n e

Table 3. Odds of 30-Day Survival among Patients Who Underwent CPR before EMS Arrival.*
No. of Patients
with Data

Odds Ratio for

30-Day Survival

95% Confidence
Interval

99% Confidence
Interval

30,381

2.80

2.553.09

2.473.18

2.95

2.603.34

2.503.47

2.15

1.882.45

1.802.56

Unadjusted

2.74

2.453.06

2.373.18

Adjusted

1.92

1.702.16

1.642.24

1.92

1.482.50

1.362.71

1.26

0.951.68

0.871.84

2.76

2.503.06

2.423.16

Unadjusted

2.63

2.273.04

2.163.19

Adjusted

2.27

1.942.67

1.842.81

Group and Propensity Score Adjustment

All patients
Propensity score

19,153

Unadjusted
Adjusted
Propensity score excluding time from collapse
to call

Propensity score for patients with missing data

for time from collapse to call

22,928

3,775

Unadjusted
Adjusted
Patients with VF or VT as initial ECG rhythm
Propensity score

10,094
7,025

* Adjusted results were adjusted for a propensity score that included the following variables, except where indicated: age, sex, cause of cardiac arrest, place of cardiac arrest, initial cardiac rhythm, year of cardiac arrest, time from collapse to the call for EMS, and time from the call
for EMS to the arrival of EMS. Unadjusted results were not adjusted for the propensity score but are from an analysis that involves the same
group of patients as the adjusted results (i.e., the analysis included patients for whom the data used to calculate the propensity score were
not missing).
Odds ratios for 30-day survival are for the comparison of patients who underwent CPR before EMS arrival with those who did not.
The propensity score includes the variables included in the main propensity score plus the time from collapse to defibrillation.

ple, the timing of the call to EMS, the dispatch

of EMS, and the arrival of EMS are digitally
registered at the dispatch center and appear to
be accurate in most cases. Second, approximately
25% of all persons with an out-of-hospital cardiac arrest in whom CPR is started are not prospectively reported to the registry, although data
on these patients have been reported retrospectively since 2011. Thus, reporting has become
more complete over time. Third, the protocols
for EMS providers have changed over time as a
result of changes in guidelines every 5 years. The
EMS response time has concomitantly increased.
As time passes, the initial rhythm of ventricular
fibrillation deteriorates into asystole, and therefore the proportion of patients found to be in
ventricular fibrillation decreases.13 However, the
number of Swedes educated in CPR has increased, and the proportion of cases in which
CPR is started before EMS arrival has increased
changes that are associated with increased
survival rates in this study. Finally, the use of
2314

therapeutic hypothermia, coronary revascularization, and implantable cardioverterdefibrillators has been implemented in recent years,
which may confound our results. These limitations, however, are tempered by the strengths of
our study, including our large sample, the use of
bystander-witnessed cases to better characterize
delays, and our robust statistical analysis.
In conclusion, among patients who had an
out-of-hospital cardiac arrest, CPR performed
before the arrival of EMS was associated with a
rate of 30-day survival that was more than twice
as high as that associated with no CPR before
EMS arrival. The association with a good outcome was greater when the time to the initiation
of CPR was short.
Supported by grants from the Laerdal Foundation for Acute
Medicine in Norway, the Swedish HeartLung Foundation, and
the Swedish Association of Local Authorities and Regions.
No potential conflict of interest relevant to this article was
reported.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

Early Cardiopulmonary Resuscitation in Cardiac Arrest

References
1. Go AS, Mozaffarian D, Roger VL, et al.
Heart disease and stroke statistics
2014 update: a report from the American
Heart Association. Circulation 2014;129(3):
e28-e292.
2. Berdowski J, Berg RA, Tijssen JG,
Koster RW. Global incidences of out-ofhospital cardiac arrest and survival rates:
systematic review of 67 prospective studies. Resuscitation 2010;81:1479-87.
3. Atwood C, Eisenberg MS, Herlitz J,
Rea TD. Incidence of EMS-treated out-ofhospital cardiac arrest in Europe. Resuscitation 2005;67:75-80.
4. Weaver WD, Cobb LA, Hallstrom AP,
Fahrenbruch C, Copass MK, Ray R. Factors
influencing survival after out-of-hospital
cardiac arrest. J Am Coll Cardiol 1986;7:
752-7.
5. Cummins RO, Eisenberg MS, Hallstrom AP, Litwin PE. Survival of out-ofhospital cardiac arrest with early initiation of cardiopulmonary resuscitation.
Am J Emerg Med 1985;3:114-9.
6. Field JM, Hazinski MF, Sayre MR, et al.
Part 1: executive summary: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;
122:Suppl 3:S640-S656.
7. Nolan JP, Soar J, Zideman DA, et al.
European Resuscitation Council Guidelines for Resuscitation 2010 Section 1.
Executive summary. Resuscitation 2010;
81:1219-76.
8. Statistics Sweden. Population statistics, JanuaryJune 2014 (http://www
.scb
.se/be0101-en).
9. Bardy GH. A critics assessment of our
approach to cardiac arrest. N Engl J Med
2011;364:374-5.
10. American Heart Association. CPR in
sudden cardiac arrest debated, 2012 (http://
pages.nxtbook.com/tristar/aha/day2_2012/
offline/aha_day2_2012.pdf).
11. Hollenberg J, Bng A, Lindqvist J, et al.
Difference in survival after out-of-hospital cardiac arrest between the two largest
cities in Sweden: a matter of time? J Intern
Med 2005;257:247-54.

12. Kitamura T, Iwami T, Kawamura T,

et al. Nationwide improvements in survival from out-of-hospital cardiac arrest
in Japan. Circulation 2012;126:2834-43.
13. Waalewijn RA, Nijpels MA, Tijssen JG,
Koster RW. Prevention of deterioration of
ventricular fibrillation by basic life support during out-of-hospital cardiac arrest.
Resuscitation 2002;54:31-6.
14. Christenson J, Andrusiek D, EversonStewart S, et al. Chest compression fraction determines survival in patients with
out-of-hospital ventricular fibrillation. Circulation 2009;120:1241-7.
15. Valenzuela TD, Roe DJ, Nichol G,
Clark LL, Spaite DW, Hardman RG. Outcomes of rapid defibrillation by security
officers after cardiac arrest in casinos.
N Engl J Med 2000;343:1206-9.
16. Caffrey SL, Willoughby PJ, Pepe PE,
Becker LB. Public use of automated external defibrillators. N Engl J Med 2002;347:
1242-7.
17. Jacobs I, Nadkarni V, Bahr J, et al.
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
Councils of Southern Africa). Circulation
2004;110:3385-97.
18. Strmse A, Svensson L, Axelsson
AB, Gransson K, Todorova L, Herlitz J.
Validity of reported data in the Swedish
Cardiac Arrest Register in selected parts
in Sweden. Resuscitation 2013;84:952-6.
19. Central government activities in ambulance services (RiR 2012:20). Stockholm:
Swedish National Audit Office, 2012
(http://www.riksrevisionen.se/PageFiles/
16781/summary%20RiR%202012_20.pdf).
20. Swedish Society of Medicine. Swedish
index for emergency medical assistance.

Stavanger, Norway:Laerdal, 1997. (In

Swedish.)
21. Handley AJ, Monsieurs KG, Bossaert
LL. European Resuscitation Council guidelines 2000 for adult basic life support. Med
Clin (Barc) 2002;118:385-90. (In Spanish.)
22. Handley AJ, Koster R, Monsieurs K,
Perkins GD, Davies S, Bossaert L. European Resuscitation Council guidelines for
resuscitation 2005. Section 2. Adult basic
life support and use of automated external defibrillators. Resuscitation 2005;67:
Suppl 1:S7-S23.
23. Holmberg M, Holmberg S, Herlitz J.
Factors modifying the effect of bystander
cardiopulmonary resuscitation on survival
in out-of-hospital cardiac arrest patients
in Sweden. Eur Heart J 2001;22:511-9.
24. Nordberg P, Hollenberg J, Herlitz J,
Rosenqvist M, Svensson L. Aspects on the
increase in bystander CPR in Sweden and
its association with outcome. Resuscitation 2009;80:329-33.
25. Wissenberg M, Lippert FK, Folke F,
et al. Association of national initiatives to
improve cardiac arrest management with
rates of bystander intervention and patient
survival after out-of-hospital cardiac arrest. JAMA 2013;310:1377-84.
26. Vaillancourt C, Kasaboski A, Charette
M, et al. Barriers and facilitators to CPR
training and performing CPR in an older
population most likely to witness cardiac
arrest: a national survey. Resuscitation
2013;84:1747-52.
27. Kitamura T, Morita S, Kiyohara K,
et al. Trends in survival among elderly patients with out-of-hospital cardiac arrest:
a prospective, population-based observation from 1999 to 2011 in Osaka. Resuscitation 2014;85:1432-8.
28. Kim C, Fahrenbruch CE, Cobb LA,
Eisenberg MS. Out-of-hospital cardiac arrest in men and women. Circulation 2001;
104:2699-703.
29. Ringh M, Rosenqvist M, Hollenberg J,
et al. Mobile-phone dispatch of laypersons for CPR in out-of-hospital cardiac
arrest. N Engl J Med 2015;372:-.
Copyright 2015 Massachusetts Medical Society.

posting presentations from medical meetings online

Online posting of an audio or video recording of an oral presentation at

a medical meeting, with selected slides from the presentation, is not considered
prior publication. Authors should feel free to call or send e-mail to the
Journals Editorial Offices if there are any questions about this policy.

n engl j med 372;24nejm.org June 11, 2015

The New England Journal of Medicine

Downloaded from nejm.org on January 27, 2016. For personal use only. No other uses without permission.
Copyright 2015 Massachusetts Medical Society. All rights reserved.

2315