Validation of a termination of resuscitation algorithm in out-of- hospital cardiac arrest patients with a non-shockable initial rhythm receiving advance life support MD, PhD Markus Skrifvars Liverpool Hospital ICU, SSWAHS
Validation of a termination of
resuscitation algorithm in out-of-
hospital cardiac arrest patients with
a non-shockable initial rhythm
receiving advance life support
MD, PhD Markus Skrifvars
Liverpool Hospital ICU, SSWAHS
Study Group
Markus Skrifvars, MD, PhD
Taneli Vayrynen MD
Markku Kuisma MD, PhD
Michael Parr MD, PhD
Maaret Castren MD, PhD
Johan Silfverstolpe MD, PhD
Leif Svensson MD, PhD
Johan Herlitz MD, PhD
Background
Survival after out-of-hospital cardiac arrest (OHCA) with a non-shockable rhythm is poor [1]
In order to identify patients in whom chance of survival is minimal termination of resuscitation (TOR) algorithms have been developed [2]
Most of these algorithms have been developed for emergency medical service (EMS) systems capable of basic life support [3]
[1] Iwami et al. Circulation 2009, [2] Bailey et al. Prehospital Emergency Care 2000, [3] Morrison et al. NEJM 2006
Background
Recently a TOR algorithm developed for EMS systems capable of advance life support (ALS) was been developed and validated [4]
1. Not witnessed by EMS personnel
2. No shock was delivered
3. No return of spontaneous circulation (ROSC) at any point during
the resuscitation
4. No bystander cardiopulmonary resuscitation performed
5. Arrest not witnessed by bystander
If none of the above criteria are present resuscitation can
be terminated
[4] Morrison et al. Resuscitation 2009, [5] Sasson et al. JAMA 2008
Survival from asystole and
PEA in Helsinki, FinlandHelsinki; Finland is a medium sized urban city with a population of 500,000
Prospective Utstein Style data collection of OHCAs since 1994
In two previous papers we examined survival to hospital discharge in patients resuscitated from asystole and PEA [4,5]
Based on these findings we develope a TOR algorithm for patients with OHCA and PEA or
asystole as the initial rhythm[5] Vayrynen et al. Acta Anaesth Scand 2008, [6] Vayrynen et al. Resuscitation 2008
Apply TOR in asystole if…
Unwitnessed arrest
Delay to arrival of
ambulance exceeds
10 minutes
No ROSC within
20 minutes from start
of ALS
Delay to arrival of ambulance exceeds
15 minutes
No ROSC within
20 minutes from start of ALS
In unwitnessed PEA if no ROSC within
10 minutes of ALS
Apply TOR in pulseless electrical
activity if…
Purpose
To retrospectively validate this Helsinki TOR algorithm and the algorithm by Morrison et al. in a
large OHCA database with survival with good neurological status at 30 days as the end-point
Study material
The Swedish Cardiac Arrest Registry
includes prospectively collected OHCAs
occurring in Sweden since 1990
Approximately 70% of all OHCA are
included annually and presently includes
about 44.000 patients
Data is collected prospectively and
entered into the database centrally
Outcome data is obtained from hospital
charts
Methods
We identified patients with asystole and
PEA as the initial rhythm
Patients with drowning, trauma, drug
overdose or accidental hypothermia were
excluded
Sensitivity, specificity, positive and
negative predictive values including 95%
confidence intervals were calculated for
both TOR algorithms
Results Asystole17.707
patients with Asystole
7.423 Unwitnessed Survival 0.4%
8.237Witnessed
Survival 1.4%
2.901 patients withdelay to ambulance
> 10 minSurvival 0.3%
5.336 patients with delay to ambulance
<10 minSurvival 2.0%
1171 patients with ROSC
Survival 8.2%
518 patients with ROSC delay < 20 min
from start of CPRSurvival 13.7%
653 patients with ROSC delay > 20 min
from start of CPRSurvival 4.4%
Conversion to VF
and
bystander
initiated life
support
Results PEA3.104 patients
with PEA
790 patients withUnwitnessed Survival 0.6%
14 patients with ROSC within 10 min from
start of CPRSurvival 7%
2.050 patients with Witnessed
Survival 1.5%
40 patients with ROSC exceeding 10 min from
start of CPRSurvival 0.2%
1.493 patients with delay to ambulance
less than 15 minSurvival 1.9%
460 patients with delay to ambulance exceeding 15 min
Survival 0.2%
209 patients with ROSC
Survival 8.6%
163 patients with ROSC delay < 20 min
Survival 11%
46 patients with ROSC > 20 min
Survival 0%
Performance of the Helsinki
TORAlive at 1 month Died before 1 month
Discontinue CPR 45 11,074
Continue CPR 162 9,631
Sensitivity 54% (53-55)
Specificity 78% (72-84)
Negative predictive value 1.7
Positive predictive value 99.6
Survival when algorithm recommends terminating 0.4%
Proportion of continued resuscitations 54%
95% Confidence intervals in parenthesis
Performance of TOR by
Morrison et al.
Alive at 1 month Died before 1 month
Discontinue CPR 5 5,465
Continue CPR 202 15,033
Sensitivity 26.7 (26.1-27.2)
Specificity 97.4 (94.4-99.0)
Negative predictive value 1.5
Positive predictive value 99.9
Survival when algorithm recommends terminating 0.1%
Proportion of continued resuscitations 74%
95% Confidence intervals in parenthesis
Discussion
The present TOR algorithm including special time
limits depending on the corresponding rhythm
worked well in patients with PEA (survival 0.1%)
In patients with asystole the survival of patients in
whom the TOR recommended discontinuing CPR
was low (0.4%) but some survivors where seen; The majority of these received bystander BLS or had conversion into
VF during the course of CPR
This study implies that survival of some patients is due to
other factors than just time delays
Problemce with rhythm interpretation?
Pseudo PEA or asystole?
Discussion
The TOR algorithm by Morrison et al. worked
very well in this large OHCA database with
only patients with non-shockable rhythms
(survival 0.1%)
However, the amount of continued
resuscitations was also higher (74% vs. 54%)
which will translate into more patients needing
intensive care
ConclusionTHe Morrison et al. algorithm seems to be
more applicable in practise and can be
recommended for use
The Helsinki TOR seems clinically justified
with limitations Correct identification of the initial rhythm
Conversion of asystole into VF?
Accuracy of measured time delays?
Bystander initiated basic life support?
Which one to use will depend on the EMS
system management protocols