Professional Documents
Culture Documents
Background: In patients with out-of-hospital cardiac arrest lation before receipt of a third 1-mg dose of epinephrine), sur-
(OHCA), care requirements can conict with the need to vival rate at hospital discharge among patients meeting these
promptly focus efforts on organ donation in patients who are criteria, performance of the criteria, and number of patients eli-
pronounced dead. gible for organ donation.
Objective: To evaluate objective criteria for identifying patients Results: In the Paris SDEC 1-year cohort, the survival rate among
with OHCA with no chance of survival during the rst minutes of the 772 patients with OHCA who met the objective criteria was
cardiopulmonary resuscitation to enable prompt orientation to- 0% (95% CI, 0.0% to 0.5%), with a specicity of 100% (CI, 97% to
ward organ donation. 100%) and a positive predictive value of 100% (CI, 99% to 100%).
These results were veried in the validation cohorts. Ninety-ve
Design: Retrospective assessment using OHCA data from 2 reg- (12%) patients in the Paris SDEC 1-year cohort may have been
istries and 1 trial. eligible for organ donation.
Setting: France (Paris Sudden Death Expertise Center [SDEC] Limitation: Several patients had unknown outcomes.
prospective cohort [2011 to 2014] and PRESENCE multicenter
cluster randomized trial [ClinicalTrials.gov: NCT01009606] [2009 Conclusion: Three objective criteria enable the early identica-
to 2011]) and the United States (King County, Washington, pro- tion of patients with OHCA with essentially no chance of survival
spective cohort [2006 to 2011]). and may help in decision making about the organ donation
process.
Patients: 1771 patients from the Paris SDEC 1-year cohort (2011
to 2012) and 5192 from the validation cohorts. Primary Funding Source: French Ministry of Health.
Measurements: Evaluation of 3 objective criteria (OHCA not Ann Intern Med. 2016;165:770-778. doi:10.7326/M16-0402 www.annals.org
witnessed by emergency medical services personnel, nonshock- For author afliations, see end of text.
able initial cardiac rhythm, and no return of spontaneous circu- This article was published at www.annals.org on 13 September 2016.
The cohort included patients with out-of-hospital cardiac arrest of presumed cardiac origin from 16 May 2011 to 15 May 2012. ACLS = advanced
cardiac life support; SDEC = Sudden Death Expertise Center.
* Out-of-hospital cardiac arrest not witnessed by emergency medical services personnel, initial nonshockable rhythm, and no return of spontaneous
circulation before receipt of a third 1-mg dose of epinephrine.
aortic double-balloon catheter insertion had to be less value (NPV) (the probability of survival when 1 crite-
than 150 minutes (warm ischemia). These eligibility cri- rion is not met) and their respective 95% CIs for the
teria were applied to patients identied as having no objective criteria. The specicity and PPV were identi-
chance of survival to evaluate their potential for organ ed as the key test characteristics at the study design
donation. stage. We described the amount of missing data for
End Points each variable of interest. Sensitivity analyses based on
The primary end point was the survival rate at hos- multiple imputation were used to account for missing
pital discharge among patients with OHCA who met data on outcomes and the objective criteria in the
the proposed objective criteria. For the PRESENCE co- SDEC cohorts. The multiple imputation procedure (mi
hort (22), the survival rate at day 28 instead of at hos- impute command) in Stata, version 14.1 (StataCorp),
pital discharge was considered because the latter infor- was used. We performed the multiple imputations by
mation was unavailable. The secondary end point was chained equations under a missing-at-random assump-
the number of patients eligible for organ donation ac- tion. The imputation models included the following
cording to the French protocol for kidney retrieval from variables: sex, age, location, whether OHCA was wit-
UDCD in the SDEC cohorts. nessed by a bystander, whether CPR was performed by
Statistical Analysis a non-EMS bystander, type of witness, presence of ini-
Data are reported as means and SDs or medians tial shockable rhythm, epinephrine dose, and outcome.
and interquartile ranges (IQRs) for continuous variables Ten data sets were created, with missing values re-
and as percentages for qualitative variables. The sur- placed by imputed values. The results of the analyses of
vival rate and 95% CI were determined for patients with the individual imputed data sets were combined using
OHCA who fullled the objective criteria. We used Rubin rules. All other statistical tests were done using
standard 2 2 diagnostic test result tables to report SAS, version 9.3 (SAS Institute).
prehospital or in-hospital mortality among patients We assumed that for patients with OHCA who
meeting the objective criteria. We estimated the sensi- meet the objective criteria, the expected survival to dis-
tivity (the probability that the criteria identify patients charge from the hospital is 0.001% (instead of 0% to
with no chance of survival [for patients who died]), allow sample size calculation). Therefore, 732 patients
specicity (the probability that the criteria recommend were required in order to detect a difference from a
continuing ACLS [for patients discharged alive]), posi- survival rate of 0.4%, with 90% power and a 1-sided
tive predictive value (PPV) (the probability of death type I error rate of 0.05 in a 1-sample test of
when the criteria are met), and negative predictive proportions.
772 Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 www.annals.org
Table 2. Performance of Each Objective Criterion for Predicting Death in Patients With OHCA, Individually and in Pairs, in the
Paris SDEC 1-y Cohort (n = 1771)
Criteria OHCA Not Witnessed by EMS Nonshockable Initial Cardiac No Sustainable ROSC Before Receipt
Personnel (n 1532), Rhythm (n 1364), n/N (%) of Third 1-mg Epinephrine Dose
n/N (%) (n 1097), n/N (%)
Survival rate when the 113/1532 (7) 23/1364 (2) 20/1097 (2)
criteria were met
Specicity 26/139 (19) 115/138 (83) 119/139 (86)
PPV 1419/1532 (93) 1341/1364 (98) 1077/1097 (98)
OHCA Not Witnessed by EMS OHCA Not Witnessed by EMS Nonshockable Initial Cardiac Rhythm
Personnel Plus Nonshockable Personnel Plus No Plus No Sustainable ROSC Before
Initial Cardiac Rhythm Sustainable ROSC Before Receipt of Third 1-mg Epinephrine
(n 1174), n/N (%) Receipt of Third 1-mg Dose (n 898), n/N (%)
Epinephrine Dose (n 950),
n/N (%)
Survival rate when the 15/1174 (1) 17/950 (2) 2/898 (0.2)
criteria were met
Specicity 123/138 (89) 122/139 (88) 137/139 (99.0)
PPV 1159/1174 (99) 933/950 (98) 896/898 (99.8)
EMS = emergency medical services; OHCA = out-of-hospital cardiac arrest; PPV = positive predictive value; ROSC = return of spontaneous
circulation; SDEC = Sudden Death Expertise Center.
www.annals.org Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 773
Table 3. Sensitivity, Specicity, and Positive and Negative Predictive Values of the Objective Criteria for Predicting Death in
Patients With OHCA*
Cohort Patients Who Met Patients Who Did Sensitivity Specicity Positive Negative Survival Rate
the Objective Not Meet >1 Objective (95% CI), (95% CI), Predictive Predictive When the
Criteria, n Criterion, n % % Value Value Objective
(95% CI), % (95% CI), Criteria Were
Died Survived Died Survived % Met (95% CI), %
(True- (False- (False- (True-
Positive) Positive) Negative) Negative)
Paris SDEC 1-y 772 0 860 139 47 (4550) 100 (97100) 100 (99100) 14 (1216) 0/772=0 (00.2)
cohort (n = 1771)
Paris SDEC validation 1569 1 1750 336 47 (4549) 99.7 (99.0100.0) 99.9 (99.7100.0) 16 (1418) 1/1570=0.1 (00.2)
cohort (n = 3656)
PRESENCE 285 0 185 16 61 (5665) 100 (79100) 100 (99100) 8 (513) 0/285=0 (00.8)
cohort (n = 486)
King County 173 0 640 237 21 (1923) 100 (98100) 100 (98100) 27 (2430) 0/173=0 (00.4)
cohort (n = 1050)
OHCA = out-of-hospital cardiac arrest; SDEC = Sudden Death Expertise Center.
* Criteria included OHCA not witnessed by emergency medical services personnel, nonshockable initial cardiac rhythm, and no return of sponta-
neous circulation before receipt of a third 1-mg epinephrine dose.
Probability that the criteria identify patients with no chance of survival (for patients who died).
Probability that the objective criteria recommend continuing advanced cardiac life support (for patients discharged alive).
Probability of death when the objective criteria are met.
Probability of survival when 1 objective criterion is not met.
Table 1 summarizes the characteristics of the 3656 3 days (IQR, 1 to 5 days). Of the 173 patients aged 18
(94%) patients with complete data. When the objective to 54 years who met the objective criteria, the median
criteria were met, 1 patient survived to hospital dis- epinephrine dose administered was 5 mg (IQR, 4 to 6
charge in a persistent vegetative state (Cerebral Perfor- mg). No patient who met the objective criteria survived
mance Category score of 4) (Appendix Figure 1). Table to hospital discharge (Appendix Figure 3, available at
3 shows the diagnostic test characteristics of the appli- www.annals.org).
cation of the objective criteria to this internal validation The sensitivity, specicity, PPV, and NPV of the ob-
cohort. jective criteria for predicting death in patients with
OHCA in these external validation cohorts are shown in
External Validation Cohorts: France and the
Table 3.
United States
In the PRESENCE cohort, we analyzed 486 patients Sensitivity Analysis
with OHCA of presumed cardiac origin who received For the Paris SDEC 1-year cohort, the survival rate
ACLS from EMS personnel. The mean age was 68 among patients meeting the 3 objective criteria, includ-
years, and 64% were male. Bystanders were present in ing the 135 patients with missing data, was 0% (CI,
75% of cases, and CPR was performed by non-EMS by- 0.0% to 0.2%). For the Paris SDEC validation cohort, the
standers for 102 (28%) patients. Initial shockable survival rate among patients meeting the criteria, in-
rhythm was present in 12% of cases, and the median cluding the 242 patients with missing data, was 0.06%
epinephrine dose administered was 8 mg (IQR, 5 to 10 (CI, 0.0% to 0.2%), which is consistent with the results of
mg). Eighty-six patients (18%) survived to hospital ad- the main analysis.
mission, and 16 (3%) were still alive at day 28. No pa-
tient who met the objective criteria survived to day 28 UDCD Candidates
(Appendix Figure 2, available at www.annals.org). Paris SDEC 1-Year Cohort
The King County cohort included 2669 patients Among the 772 patients with no chance of survival
with OHCA of presumed cardiac origin with ACLS at- who were screened for organ donation eligibility ac-
tempted by EMS personnel; 1050 patients were cording to the French clinical criteria, 95 (12%) might
younger than 55 years and were included in this analy- have had organs that were likely to be suitable for
sis. The mean age was 45 years, and 65% were men. transplantation (Table 4 and Figure 2). These patients
Cardiac arrest occurred at home in 692 cases (66%). Of could have been referred quickly to organ donation
the 1050 cases of OHCA, 114 (11%) were witnessed by institutions under mechanical ventilation and continu-
EMS personnel and 936 (89%) were witnessed by by- ous automated external cardiac massage.
standers before EMS arrival. Non-EMS bystanders per-
formed CPR for 609 (58%) patients. Initial shockable
rhythm was present in 364 (35%) cases. Overall, 468 Internal Validation Cohort: France
(45%) patients survived to hospital admission, and 239 Among the 1570 patients in the Paris SDEC valida-
(23%) were discharged alive after a median length of tion cohort with no chance of survival, 122 (8%) might
stay of 7 days (IQR, 5 to 12 days). During hospitaliza- have had organs that were likely to be suitable for
tion, 229 patients died after a median length of stay of transplantation (Table 4).
774 Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 www.annals.org
ACLS = advanced cardiac life support; CPR = cardiopulmonary resuscitation; SDEC = Sudden Death Expertise Center; UDCD = uncontrolled
donation after cardiac death.
* Out-of-hospital cardiac arrest not witnessed by emergency medical services personnel, initial nonshockable rhythm, and no return of spontaneous
circulation before receipt of a third 1-mg dose of epinephrine.
Proposed by the French national protocol for kidney retrieval from UDCD (24).
www.annals.org Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 775
futile resuscitative efforts among patients with no Our study has limitations. One patient in the Paris
chance of survival could lead to physical and psycho- SDEC 1-year cohort and 9 in the Paris SDEC validation
logical pain for patients and their relatives and could cohort had missing outcome data. However, the results
divert care and resources from patients who are more of the sensitivity analyses that included these patients
likely to survive (12). Unfortunately, the current require- were consistent with the results of the main analyses.
ment of 30 minutes of CPR before on-site patient extri- Second, the King County cohort was restricted to pa-
cation (imposed by the French national protocol for tients aged 18 to 54 years (the age range for UDCD),
kidney retrieval from UDCD) may partially explain the with no survivors to hospital discharge among those
need to discard potential grafts, which has been re- meeting the objective criteria. It is unlikely that older
ported in the literature (17). patients with OHCA who were not included in our study
In Europe, UDCD data show that organ transplan- would have better outcomes (42 44).
tation programs are benecial in terms of successful Our study also has several important strengths. The
transplantation, although the transport time targets set data were collected prospectively in the SDEC and
by UDCD protocols can be difcult to achieve (3134). PRESENCE cohorts and were based on Utstein deni-
In 2006, the Institute of Medicine projected that nation- tions (21). Our study design reects the experience of 3
wide implementation of UDCD protocols could gener- cohorts of patients with OHCA managed in 2 EMS sys-
ate an additional 22 000 donation opportunities annu- tems, which optimized the clinical relevance of our
ally in the United States, thus substantially reducing data. The internal validity of these data is robust with
wait times for transplantation (35). Our results and narrow CIs, and external validation supports their gen-
those of Roberts and colleagues in the United Kingdom eralizability. Inclusion of additional traditional criteria,
(36) support this idea by specifying the potential num- although associated with poor outcomes, would not
ber of patients with OHCA who are eligible for organ contribute more but may be taken into consideration.
donation. However, our results may be overestimates Finally, the implementation of these objective criteria
because some UDCD eligibility criteria may be difcult for patients with OHCA is an important topic in revi-
to verify on the scene. In real-world practice in France, sions of the international CPR guidelines.
after cannulation, potential donors are actively In conclusion, these data from 3 cohorts of patients
screened for infections and pertinent medical history with OHCA receiving advanced life support from para-
that might preclude donation, and approximately 30% medics or physicians suggest that there is essentially no
of them are refuted for medical reasons (37). In addi- chance of survival in patients whose OHCA is not wit-
tion, although our ndings allow improvement of early nessed by EMS personnel, who have a nonshockable
assessment for UDCD protocols, several ethical, legal, initial cardiac rhythm, and in whom ROSC does not oc-
and organizational issues raised by UDCD transplanta- cur before receipt of a third 1-mg dose of epinephrine.
tion must be carefully considered before local imple- Our results may help in decision making about UDCD.
mentation of these protocols (6, 38, 39). Most impor-
tant, UDCD should be considered only when available From Paris Cardiovascular Research Center, Hopital Necker-
therapeutic options are unsuccessful or are not clini- Enfants Malades, Hopital Cochin, Universite Paris Descartes-
cally indicated to avoid denying optimum treatment to Sorbonne Paris Cite, Hopital Saint-Louis, Universite Paris
patients who might recover (40). In addition, this emer- Diderot Paris VII, Hopital Europeen Georges Pompidou, and
gency procedure with specic time constraints requires Paris Fire Brigade, Paris, France; Hopital Avicenne, Universite
highly coordinated multidisciplinary teamwork to pre- Paris 13, Bobigny, France; and Public HealthSeattle & King
County and University of Washington, Seattle, Washington.
serve organ function.
All communities should optimize retrieval of tissue
and organ donations in patients pronounced dead in Acknowledgment: The authors thank the physicians, para-
medics, nurses, and ambulance attendants at each center for
out-of-hospital settings (6). Emergency medical ser-
their valuable cooperation in the study. A list of the Paris
vices agencies should develop protocols and imple-
SDEC Investigators is provided in the Appendix (available at
mentation plans with their regional organ and tissue
www.annals.org).
donation program to optimize donation after OHCA
(6). We believe that EMS personnel should implement
Financial Support: The PRESENCE trial was supported by the
our 3 objective criteria in their updated protocols for
Programme Hospitalier de Recherche Clinique 2008 of the
cardiac arrest. All patients who meet these criteria, French Ministry of Health and by the Research Delegation of
whose organs are reasonably likely to be suitable for the Assistance PubliqueHopitaux de Paris.
transplantation, and who would otherwise have termi-
nation of efforts should be considered candidates for Disclosures: Dr. Jabre reports a grant from the French Minis-
organ donation. Rapid referral to an organ donation try of Health during the conduct of the study. Pr. Jacob re-
institution under mechanical ventilation and continuous ports personal fees from AbbVie, Alere, BioPorto, and Frese-
automated external cardiac massage should be consid- nius outside the submitted work. Pr. Adnet reports a grant
ered for patients with no chance of survival. In the from the French Ministry of Health during the conduct of the
United States, however, initiation of organ preservation study. Authors not named here have disclosed no conicts of
without prior patient consent is not allowed; eligibility interest. Disclosures can also be viewed at www.acponline
for UDCD is restricted to previously registered organ .org/authors/icmje/ConictOfInterestForms.do?msNum=M16
donors (41). -0402.
776 Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 www.annals.org
www.annals.org Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 777
VITAL STATISTICS
More than 148 000 physicians and other health professionals receive
Annals, and millions of people access it through institutional libaries, the
Web, or mobile devices. The most recent (2015) Impact Factor for Annals
of Internal Medicine is 16.593the highest of any specialty journal in the
Thomson Reuters' General and Internal Medicine category. The Impact
Factor is a measurement of the frequency with which the average article
has been cited. Annals has risen in rank to 5th among 151 general medi-
cine journals. It is one of the most highly cited and inuential journals in
the world.
778 Annals of Internal Medicine Vol. 165 No. 11 6 December 2016 www.annals.org
The cohort included patients with out-of-hospital cardiac arrest of presumed cardiac origin from 16 May 2012 through 15 May 2014. ACLS =
advanced cardiac life support; SDEC = Sudden Death Expertise Center.
* Out-of-hospital cardiac arrest not witnessed by emergency medical services personnel, initial nonshockable rhythm, and no return of spontaneous
circulation before receipt of a third 1-mg dose of epinephrine.
One patient survived to hospital discharge in a persistent vegetative state (Cerebral Performance Category score of 4).
PRESENCE cohort
(n = 570)