a b s t r a c t

Background: In out of hospital cardiac arrest (OHCA), the prognostic influence of conversion to Shockable rhythms during resuscitation for initially Non-shockable rhythms remains unknown. This study aimed to assess the relationship between initial and subsequent shockable rhythm and post-arrest Survival and neurological outcomes after OHCA. Methodology: This was a retrospective analysis of all OHCA cases collected from the Pan-Asian Resuscitation Out- comes Study (PAROS) registry in 7 countries in Asia between 2009 and 2012. We included OHCA cases of presumed cardiac etiology, aged 18-years and above and resuscitation attempted by EMS. We performed multivariate logistic regression analyses to assess the relationship between initial and subsequent shockable rhythm and survival and neurological outcomes. 2-stage seemingly unrelated bivariate probit models were developed to jointly model the survival and neurological outcomes. We adjusted for the clustering effects of country variance in all models.

Results: 40,160 OHCA cases met the inclusion criteria. There were 5356 OHCA cases (13.3%) with Initial shockable rhythm and 33,974 (84.7%) with initial non-shockable rhythm. After adjustment of baseline and prehospital charac- teristics, OHCA with initial shockable rhythm (odds ratio/OR = 6.10, 95% confidence interval/CI = 5.06-7.34) and subsequent conversion to shockable rhythm (OR = 2.00,95%CI = 1.10-3.65) independently predicted better survival-to-hospital-discharge outcomes. Subsequent shockable rhythm conversion significantly improved survival-to-admission, discharge and post-arrest overall and cerebral performance outcomes in the multivariate lo- gistic regression and 2-stage analyses.

Conclusion: Initial shockable rhythm was the strongest predictor for survival. However, conversion to subsequent shockable rhythm significantly improved post-arrest survival and neurological outcomes. This study suggests the importance of early resuscitation efforts even for initially non-shockable rhythms which has prognostic implications and selection of subsequent post-resuscitation therapy.

(C) 2016

Abbreviations: OHCA, out-of hospital cardiac arrest; PAROS, Pan-Asian resuscitation outcomes study; EMS, emergency medical services; VF, ventricular fibrillation; PVT, pulseless ven- tricular tachycardia; PEA, pulseless electrical activity; VT, ventricular tachycardia; CPR, cardio-pulmonary resuscitation; TTM, Targeted temperature management; PCI, percutaneous cor- onary intervention; ECMO, extracorporeal membrane oxygenation; AED, automatic external defibrillators; BLS, basic life support; ED, emergency department; ROSC, return of spontaneous circulation; OPC, overall performance category; CPC, cerebral performance category; SUBPR, seemingly unrelated bivariate probit regression; OR, odds ratio; CI, confidence interval.

* Corresponding author at: Saw Swee Hock School of Public Health, National University of Singapore, 21 Lower Kent Ridge Road, 119077, Singapore.

E-mail address: [email protected] (W. Wah).

http://dx.doi.org/10.1016/j.ajem.2016.10.042

0735-6757/(C) 2016

1. Introduction
   1. Background

Out-of hospital cardiac arrest (OHCA) is a major cause of mortality in Asia and around the world. There were more than 300 000 OHCA each year with an overall survival-to-hospital discharge rate of 9.5% in US in 2013 [1]. Recent data from the Pan-Asian Resuscitation Outcomes Study (PAROS) showed lower survival-to-hospital discharge rates than the United States and Europe, possibly due to lower incidence of initial shockable rhythm, longer Response times and differences in Emergency Medical Services (EMS) systems in Asia [2].

Heart rhythms associated with cardiac arrest are divided into two groups: shockable rhythms (ventricular fibrillation/VF, pulseless ven- tricular tachycardia/PVT) and non-shockable rhythms (asystole, pulseless electrical activity/PEA). Traditional resuscitation efforts are fo- cused on Early defibrillation to patients with initial shockable rhythm [3], as this is known to be a significant predictor of survival outcomes after OHCA [4]. However, over the past few decades, the incidence of ini- tial shockable rhythm has declined and initial non-shockable rhythms have increased, possibly due to the modifications in risk factors and treatment of coronary heart disease with cardiovascular drugs that may shorten the duration of VF/VT [5,6]. For initially non-shockable rhythms, it has been suggested that early efforts should focus on unin- terrupted Cardio-pulmonary resuscitation without pauses for rhythm checks [7]. However, this is still controversial and the evidence is conflicting [8]. CPR is believed to improve myocardial tissue perfusion and excitability in assisting subsequent shockable rhythm conversion [9]. If the patient does not receive CPR or defibrillation within the first few minutes of arrest, an initially coarse VF changes into a fine VF of low amplitude and subsequently converts into an asystole within 10 to 12 min which would be more difficult to resuscitate [10].

Importance

Especially in Asia, only a minority of OHCA present with initial shockable rhythm [2] and a number of patients with initial non- shockable rhythm actually convert to a shockable rhythm after a period of resuscitation. If conversion to subsequent shockable rhythm is a strong predictor of subsequent outcomes, this finding might have an important implication for consideration in termination of resuscitation rules, clinical prognostication and selection of subsequent post- resuscitation therapies. However, the prognostic influence of conver- sion to shockable rhythms during resuscitation for initially non- shockable rhythms remains unknown in the Pan-Asian population.

Aims

Using retrospective data, this study aimed to assess the relationship between initial and subsequent shockable rhythm and post-arrest sur- vival and neurological outcomes after OHCA. It also hypothesized that subsequent shockable rhythm had better survival and neurological out- comes than persistent-non-shockable rhythm.

1. Methodology
   1. Study Setting

The PAROS registry, a prospective Asia pacific multinational registry, is a resuscitation clinical research network established in collaboration with EMS agencies and academic centers in Singapore, Japan, South Korea, Malaysia, Thailand, Taiwan and United Arab Emirates-Dubai [11]. Twelve sites from seven countries are involved in this study. The registry aims to improve Survival outcomes and EMS systems of partic- ipating countries by providing baseline information about OHCA preva- lence, management and outcomes, describe and compare regional

variations in EMS, systemic and structural interventions for OHCA using common data definitions and collection methods [11]. The major- ity of PAROS EMS systems are single tiered, public-funded and fire- based except for Thailand and Malaysia which have hospital-based EMS. Thailand EMS has nurses and physicians on ambulances whereas the other countries have emergency medical technicians. Ambulances are equipped to provide automatic external defibrillation (AED) and basic life support (BLS) treatment. However, each EMS has variations in dispatch system, response time, ambulance management and treat- ment capability.

All participating countries use a Standardized taxonomy and case re- cord form to collect common variables. Each participating country is in charge of its own data collection process. All data are entered through a secured shared internet electronic data capture (EDC) system based in the study coordination center in Singapore [11]. Japan, South Korea and Taiwan which have existing data registries export data via an auto- matic export field data entry process to the PAROS registry. Countries without a national data registry enter data through the EDC system. In- stitutional review board approval was gained from each country in ac- cordance with its own regulations and guidelines. The PAROS registry collects all OHCA, as confirmed by the absence of pulse, unresponsive- ness and apnea, of both presumed cardiac and non-cardiac etiology pre- senting to emergency department (ED) or brought in by EMS. The registry excludes OHCA patients who were immediately pronounced dead and for whom resuscitation was not attempted due to decapita- tion, Rigor mortis, and dependent lividity and “do not attempt resuscita- tion” orders. EMS providers in all PAROS sites have an obligation to transport all OHCA to hospitals (except OHCA with obvious death, rigor mortis, decapitation, dependent lividity), except for Malaysia which practices termination of resuscitation in the field.

Study Design, Selection of Participants

This study was a retrospective analysis of all OHCA cases collected from the PAROS registry in 7 countries in Asia between 2009 and 2012. We included OHCA cases of presumed cardiac etiology, aged 18 years and above and resuscitation attempted by EMS.

Outcomes, Methods and Measurements

The primary outcome was survival-to-hospital-discharge (discharged alive/remained in hospital at 30th day post-arrest). Second- ary outcomes were total return of spontaneous circulation (ROSC) (ROSC at scene or arrival at ED), survival-to-admission and good neuro- logical status with post-arrest overall performance category (OPC) and Cerebral Performance Category scales at 1 or 2. Survival outcomes were determined through the review of hospital records and neurolog- ical outcomes were determined through the combination of the review of hospital records and interview. The included study variables were age, gender, location type, medical comorbidity, arrest witnessed status, bystander CPR, initial arrest rhythm, prehospital defibrillation, prehospital airway, prehospital drug administration and EMS response time within 8 min [12]. In the entire eligible cohort, the outcome of OHCA were evaluated based on three groups stratified by initial/first rhythm recorded on automatic external defibrillator(AED) – initial shockable rhythm (VF/VT/Unspecified shockable rhythm), conversion to subsequent shockable rhythm (initial non-shockable rhythm PEA/ Asystole/Unspecified unshockable rhythm and defibrillated at prehospital or ED) and remained in non-shockable rhythm (initial non-shockable PEA/Asystole/Unspecified unshockable rhythm and not defibrillated). We performed sub-group analyses on OHCA with initial non-shockable rhythm including PEA, asystole and unspecified non- shockable rhythms respectively. Time to first shock (time from EMS CPR to first prehospital defibrillation) was assessed among patients with initial shockable rhythm and conversion to subsequent shockable rhythm.

Analysis

The analyses were performed using Stata. The association between initial rhythm and outcomes were assessed using chi-squared tests or Fisher’s exact test and independent samples t-test or Mann-Whitney- U test as appropriate. We adjusted for the clustering effects of country variance in all univariate and multivariate logistic regression models. 2-stage seemingly unrelated bivariate probit regression (SUBPR) model was developed to assess the influence of initial rhythm and sub- sequent conversion rhythm on survival-to-admission (first stage) and survival-to-discharge outcomes (second stage) [13]. In this SUBPR model, the Probability of survival-to-admission and survival-to- discharge outcomes were jointly modeled using maximum likelihood estimates with robust standard error estimates. A 2-stage equation was used when the second equation observed was conditional on the first outcome. For example, patients who did not survive on admission would not survive on discharge. The ‘seemingly unrelated’ model was used since different sets of covariates were used in the two equations. Prehospital characteristics such as age, gender, brought in by EMS, location of arrest, witnessed status, bystander CPR, pre-hospital air- way, pre-hospital drug administration, response time and initial arrest rhythm were adjusted in the model. Univariate significant fac- tors were adjusted in the multivariate models based on p value b 0.1. In this multivariate seemingly unrelated regression (SUR) bivariate probit model, we used backward step-wise regression procedure. The Wald Chi-square test which is reflected by the statistical

significance of p value was used to determine whether the final fitted model would be best estimated jointly in a recursive manner or not. In addition, we also assessed the statistical significance of the likelihood-ratio test of rho, which implies that the model is appropri- ately estimated with the SUBPR method. The final effect size of inde- pendent significant risk factors were quantified using the odds ratio (OR) estimate and its associated 95% confidence interval (CI). A two- sided p value of less than 0.05 was considered statistically significant in the multivariate model. The frequency of patients with total ROSC, survival-to-admission, survival-to-discharge, favorable post-arrest OPC and CPC were compared by the interval of EMS shock delivery (time from EMS CPR to first prehospital defibrillation) using the chi-squared test for trend.

1. Results

Fig. 1 shows the Selection process of included OHCA cases in this study. There were 66,780 OHCA cases between 2009 and 2012. This study included 39,330 cases after excluding those cases which were aged b 18 years, non-presumed cardiac etiology, resuscitation not attempted by EMS/private ambulance and missing initial rhythm data. Out of the entire eligible cohort, there were 33,974 cases (86.4%) who presented with initial non-shockable rhythm and 5356 cases (13.6%) with initial shockable rhythm. Among the initial non-shockable rhythm cohort, around 2691 (6.8%) cases converted to subsequent shockable

Fig. 1. Flow diagram of included OHCA cases.

Table 1

Characteristics of the entire eligible OHCA cases stratified by initial rhythm.

Covariates	Initial shockable rhythma n = 5356 (13.6%)	Initial non-shockable rhythmb n = 33,974 (86.4%)		p value
		Conversion to subsequent shockable rhythmc	Remain in non-shockable rhythmd
		n = 2691 (6.8%)	n = 31,283 (79.6%)
Age, median (IQR)	64 (54-74)	68 (57-78)	78 (68-86)	b 0.001?
Gender male	4296 (80.21)	1911 (71.01)	17,529 (56.03)	b 0.001?
Brought in by EMS	5345 (99.79)	2687 (99.85)	31,279 (99.99)	b 0.001?
Home residence	1456 (27.18)	1280 (47.57)	9726 (31.09)	b 0.001?
Arrest witnessed	3796 (70.87)	1454 (54.03)	11,192 (35.78)	b 0.001?
Bystander CPR	2404 (44.88)	965 (35.86)	11,744 (37.54)	b 0.001?
Prehospital advanced airway	2256 (42.12)	1129 (41.95)	12,752 (40.76)	0.001?
Prehospital drug administration	1165 (21.75)	571 (21.22)	3158 (10.09)	b 0.001?
Response time less than 8 min	3894 (72.7)	1824 (67.78)	22,561 (72.12)	b 0.001?
Prehospital or ED defibrillation	5198 (97.05)	2691 (100)	0 (0)	b 0.001?
Shock delivery time(minutes)e	2 (1-3)	8 (4-15)	–	N/A
ROSC at scene or ED	1493 (27.88)	266 (9.88)	1566 (5.01)	b 0.001?
Survival-to-admission	1790 (33.42)	493(18.32)	2498 (7.99)	b 0.001?
Survival-to-discharge	1374 (25.65)	222 (8.25)	896 (2.86)	b 0.001?
Post-arrest cerebral performance	938 (17.51)	119 (4.42)	398 (1.27)	b 0.001?
Post-arrest overall performance	718 (13.41)	100 (3.72)	351 (1.12)	b 0.001?

IQR; interquartile range.

* Statistically significant (p b 0.05).

^a VF/VT/Unspecified shockable rhythm.

^b PEA/Asystole/Unspecified non-shockable rhythm.

^c Initial PEA/Asystole/Unspecified unshockable rhythm and defibrillated at prehospital or ED.

^d Initial PEA/Asystole/Unspecified unshockable rhythm and not defibrillated.

^e Shock delivery time (time from EMS CPR to first prehospital defibrillation).

rhythm and 31,283 cases (79.6%) cases remained in non-shockable rhythm.

Table 1 displays the characteristics of the entire eligible OHCA cases stratified by initial rhythm. Those cases with initial shockable rhythm were more likely to be younger, male, had witnessed arrest, received bystander CPR, prehospital airway and drug, and had EMS response time of less than 8 min. Those patients who converted to subsequent shockable rhythm were more likely to arrest in a home residence. In the subsequent shockable rhythm conversion group, the frequency of patients who had total ROSC, survival-to-admission, discharge and good neurological outcomes significantly decreased with longer EMS shock delivery time as observed in the initial shockable Rhythm group (Fig. 2). The majority of patients with favorable survival and neurologi- cal outcomes received first prehospital shock within the first 20 min after initiation of EMS CPR (Fig. 2).

After adjustment of differences in prehospital characteristics, those cases with initial shockable rhythm and subsequent shockable rhythm conversion had better ROSC, survival-to-admission, survival-to- discharge, post-arrest overall and cerebral performance outcomes. After adjustment of baseline and prehospital characteristics, subsequent conversion to shockable rhythm significantly improved survival-to- admission (OR = 1.31 95% CI = 1.01-1.70), discharge (OR = 2, 95%

CI = 1.1-3.65) and post-arrest overall (OR = 5.12, 95% CI = 3.5-7.48) and cerebral performance outcomes (OR = 5.39, 95% CI = 4.32-6.73). OHCA with initial shockable rhythm (OR = 6.10, 95% CI = 5.06-7.34) and subsequent conversion to shockable rhythm (OR = 2.00, 95% CI = 1.10-3.65) independently predicted better survival-to-hospital- discharge outcomes (Table 2). In the subgroup of initial non-shockable rhythm, OHCA patients with subsequent rhythm conversion had signif- icantly better survival-to-admission, discharge, post-arrest overall and cerebral performance outcomes. Further sub-group analyses on PEA, asystole and unspecified non-shockable rhythm showed a consistent positive association with shockable rhythm conversion and favorable survival and neurological outcomes. In the 2-stage analysis of the entire cohort, subsequent conversion to shockable rhythm significantly influ- enced favorable survival-to-admission (OR = 1.18, 95% CI = 1.02- 1.36) and discharge (OR = 1.42, 95% CI = 1.03-1.95) (Table 3) and good cerebral (OR = 2.2, 95% CI = 1.97-2.46) and overall performance

(OR = 2.14, 95% CI = 1.92-2.38) (Table 4). Similarly in the initial non- shockable rhythm subgroup, subsequent rhythm conversion signifi- cantly improved survival-to-admission (OR = 1.20, 95% CI = 1.04- 1.37), discharge (OR = 1.41, 95% CI = 1.05-1.9), cerebral (OR = 2.23,

95% CI = 2.02-2.46) and overall performance outcomes (OR = 2.15, 95% CI = 1.96-2.36) (Tables 3, 4).

1. Discussion

In this study, a majority of patients (86.4%) with OHCA in Asia pre- sented with initial non-shockable rhythm but only a minority of those patients (6.8%) converted to subsequent shockable rhythm. This is con- sistent with other published studies [7,14,15]. There are global differ- ences in frequencies of shockable rhythm with lower incidence in Asia (4.1%-19.8%) than Europe (28-45%) and US (24%) which might be ex- plained by differences in EMS systems and etiology of OHCA [2,8,16].

Previous studies have shown that initial shockable rhythm has a bet- ter prognosis compared to non-shockable rhythm [5,16-20]. However, studies have been conflicting regarding patients with initially non- shockable rhythms who subsequently become shockable, with some showing higher survival [8,21-26] but others reporting worse survival [7,27-29]. For example, two OHCA and one in-hospital cardiac arrest studies reported poorer survival among patients who converted to sub- sequent VF/VT than those with initial shockable rhythm and persistent non-shockable rhythm [7,27,28].

We believe this large multinational population-based prospective cohort data shows clearly that survival and neurological outcomes after OHCA were most favorable in patients presenting with initial shockable rhythm and favorable in those with subsequent shockable rhythm conversion compared to those with persistent non-shockable rhythm. The 3 phase model of cardiac arrest (electrical, circulatory, met- abolic phase) in resuscitative physiology suggests the presence of shockable and non-shockable rhythm is a surrogate for ischemic dura- tion [30]. The majority of those patients with converted rhythm are as- sumed to be in metabolic phase which could have explained poorer outcomes than initial shockable rhythm.

In the subgroup analysis of initial non-shockable rhythm, subse- quent rhythm conversion had better survival and neurological

Fig. 2. Comparison of survival and neurological outcomes by time interval from EMS CPR to 1st prehospital defibrillation.

outcomes in patients with unknown non-shockable rhythm than those with asystole and PEA which could be explained by better prognostic ar- rest characteristics. One recent study reported better outcomes in the

initial asystole patients with rhythm conversions but no association was observed among those with initial PEA [31]. However, this study showed a consistent improvement in survival and neurological

Table 2

Multivariate analysis of the influence of initial and subsequent shockable rhythm on survival and neurological outcomes of OHCA. Adjusted odds ratio (95% confidence interval)^a

1.34 (1.06-1.69)?

	Reference category – remain in non-shockable rhythmc	The entire eligible cohort (n Initial shockable rhythmd	= 39,330) Conversion to subsequent shockable rhythme	Initial non-shockable rhythm subgroup (n = 33,974)b Conversion to subsequent shockable rhythme
	ROSC at scene or ED Survival-to-admission Survival-to-discharge Good post-arrest cerebral performance Good post-arrest overall performance	4.47 (3.31-6.03)? 2.73 (2.56-2.93)? 6.1 (5.06-7.34)? 11.35 (9.21-14)? 12.54 (9.15-17.17)?	1.59 (0.56-4.48) 1.31 (1.01-1.70)? 2 (1.1-3.65)? 5.12 (3.5-7.48)? 5.39 (4.32-6.73)?	1.57 (0.6-4.13) 1.97 (1.14-3.39)? 5.08 (4.17-6.2)?
	Sub-groups	Initial PEA subgroup (n = 5626)	Initial Asystole subgroup (n =	Initial Unspecified non-shockable rhythm
			19,429)	subgroup (n = 8919)
	Reference category – remain in non-shockable rhythmc	Conversion to subsequent shockable rhythme	Conversion to subsequent shockable rhythme	Conversion to subsequent shockable rhythme
	ROSC at scene or ED Survival-to-admission Survival-to-discharge Good post-arrest cerebral performance Good post-arrest overall performance	0.63 (0.29-1.39) 1.26 (0.80-1.97) 1.36 (0.72-2.55) 2.1 (1.18-3.73)? 2.29 (1.76-2.99)?	1.27 (0.47-3.42) 2.05 (1.08-3.91)? 2.25 (1.64-3.08)? 3.88 (1.85-8.13)? 2.75 (2.73-2.76)?	3.52 (2.52-4.9)? 3.37 (2.91-3.92)? 5.53 (3.82-8.01)?

4.95 (3.34-7.33)?

1.70 (1.56-1.85)?

5.87 (5.16-6.68)?

^a Adjusted for age, gender, brought in by EMS, location of arrest, witnessed status, bystander cardiopulmonary resuscitation, pre-hospital airway, pre-hospital drug administration, response time, initial arrest rhythm.

^b Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm).

^c Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm) and not defibrillated.

^d Initial shockable (VF/VT/Unspecified shockable) rhythm.

^e Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm) and defibrillated.

* Statistically significant (p b 0.05).

Table 3

Multivariate analysis of factors associated with survival-to-admission and survival-to- discharge outcomes of OHCA (2-stage seemingly unrelated bivariate probit regression model).

Table 4

Multivariate analysis of factors associated with good post-arrest overall and cerebral per- formance outcomes of OHCA (2-stage seemingly unrelated bivariate probit regression model).

Entire eligible cohort

Initial non-shockable rhythm sub-group^a

Entire eligible cohort

Initial non-shockable rhythm subgroup^a

1st stage – survival-to-admission OR (95% CI) OR (95% CI)

Age 0.998 (0.992-1.003) 0.999 (0.99-1.004)

1st stage – post arrest cerebral performance

OR (95% CI) OR (95% CI)

Brought in by EMS 1.4 (0.74-2.63) 1.01 (0.72-1.4)

Non-residential location 1.34 (1.27-1.42)? 1.37 (1.29-1.45)? Cardiac arrest rhythm

Age 0.98 (0.98-0.99)? 0.98 (0.98-0.99)?

Female 1.01 (0.95-1.07) 0.97 (0.95-0.99)?

Brought in by EMS 0.73 (0.64-0.84)?

Remain in non-shockable rhythm^b

Conversion to subsequent shockable rhythm^d

Reference Reference

1.18 (1.02-1.36)? 1.20 (1.04-1.37)?

Non-residential location 1.33 (1.27-1.4)? 1.34 (1.26-1.42)?

Arrest witnessed 1.78 (1.73-1.83)? 1.94 (1.85-2.04)?

Bystander CPR 0.97 (0.94-1) 0.85 (0.81-0.88)?

Prehospital airway 0.48 (0.46-0.5)? 0.54 (0.52-0.57)?

Initial shockable rhythm^c 1.83 (1.77-1.90)?

2nd stage – survival-to-discharge OR (95% CI)

Age 0.991 (0.986-0.996)?

OR (95% CI)

0.993 (0.989-0.997)?

Response time less than 8 min 1.07 (0.98-1.17)

Pre-hospital drug administration 0.97 (0.94-0.998)? Cardiac arrest rhythm

Remain in non-shockable rhythm^b Reference

Female 0.97 (0.93-1.01) 1.01 (0.97-1.05)

Initial shockable rhythm^c 3.5 (2.91-4.21)?

Brought in by EMS 1.65 (0.79-3.46)

Non-residential location 1.44 (1.41-1.48)?

0.93 (0.39-2.17)

1.48 (1.37-1.59)?

Conversion to shockable rhythm^d 2.2 (1.97-2.46)? 2.23 (2.02-2.46)?

Arrest witnessed 1.18 (1.1-1.27)? 1.23 (1.09-1.39)?

Bystander CPR 0.99 (0.91-1.06)

2nd stage – post arrest overall performance

OR (95% CI) OR (95% CI)

Prehospital airway 0.79 (0.67-0.92)? 0.74 (0.66-0.83)? Response time less than 8 min 1.07 (1.01-1.14)?

Cardiac arrest rhythm

Age 0.98 (0.978-0.985)? 0.98 (0.98-0.99)?

Female 0.98 (0.93-1.04) 0.98 (0.95-1.003)

Brought in by EMS 0.71 (0.63-0.82)?

Remain in non-shockable rhythm^b

Conversion to subsequent shockable rhythm^d

Reference Reference

1.42 (1.03-1.95)? 1.41 (1.05-1.9)?

Non-residential location 1.35 (1.29-1.41)? 1.37 (1.29-1.45)? Arrest witnessed 1.78(1.73-1.83)? 1.95(1.85-2.04)?

Bystander CPR 1 (0.98-1.03) 0.89 (0.86-0.92)?

Prehospital airway 0.48 (0.47-0.5)? 0.52 (0.49-0.55)?

Initial shockable rhythm^c 2.68 (2.31-3.12)?

Response time less than 8 min 1.07 (0.98-1.17)

OR; odds ratio, 95% CI; 95% confidence interval.

* Statistically significant (p b 0.05).

Cardiac arrest rhythm

Remain in non-shockable rhythm^b

Reference

*

_a Initial shockable rhythm^c 3.48 (2.87-4.2)

Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm).

^b Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm) and not defibrillated.

Conversion to subsequent shockable rhythm^d

2.14 (1.92-2.38)? 2.15 (1.96-2.36)?

^c Initial shockable (VF/VT/Unspecified shockable) rhythm.

^d Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm) and defibrillated.

outcomes in PEA, asystole and unspecified unshockable groups with subsequent shockable rhythm conversion. Therefore, our study sup- ports that some patients with initial non-shockable rhythm can benefit from defibrillation [8,21-26]. Recent studies explained that subsequent shock delivery was associated with improved post-arrest survival out- comes only when the shock delivery time was less than 20 min (from initiation of EMS CPR to 1st shock) supporting the need for rhythm checks in line with current guidelines [24,26,28,32]. This study showed more patients in the shockable rhythm conversion group achieved fa- vorable survival and neurological outcomes within the first 20 min after initiation of EMS CPR than after 20 min. One recent study also con- firmed better survival rates among those patients with conversion to subsequent shockable rhythm despite more pauses due to defibrillation attempts in those patients [8] and the benefit of early Rhythm analysis to provide shocks if indicated with a 2-minutes period of CPR rather than delayed rhythm analysis [3,33]. However, some studies have suggested that EMS should not interrupt CPR to do frequent rhythm analysis for initially non-shockable rhythms as this provides little benefit and ad- versely affects circulation to vital organs and in turn have detrimental effects on survival outcomes [34,35]. These studies suggested focusing less on defibrillation for these patients and to focus more on high quality CPR with minimal interruptions, ventilation, identification and treat- ment of Reversible causes and a need for new treatment strategies. However, our study suggests the importance of early resuscitation ef- forts even for initially non-shockable rhythms which has prognostic im- plications and selection of subsequent post-resuscitation therapy.

Termination of resuscitation/do not attempt Resuscitation guidelines usually recommend to have continued Resuscitative efforts for patients with shockable rhythm or subsequent shockable rhythm conversion, but for patients with terminal rhythm, PEA or asystole, termination of

OR; odds ratio, 95% CI; 95% confidence interval.

* Statistically significant (p b 0.05).

^a Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm).

^b Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm) and not defibrillated.

^c Initial shockable (VF/VT/Unspecified shockable) rhythm.

^d Initial non-shockable rhythm (PEA/Asystole/Unspecified unshockable rhythm) and defibrillated.

resuscitation is recommended [36]. Our study results suggest this is a reasonable approach to take, as patients who convert to a shockable rhythm have significantly better prognosis compared to those who re- main in non-shockable rhythms. This could suggest that patients with converted shockable rhythm were similar in cardiac etiology to those presenting with shockable rhythms and different from those who remained in non-shockable rhythms that were suspected to have irre- versible cardiac etiology other than coronary ischemia.

Initial shockable rhythm is often used as a criteria for selection of subsequent post-resuscitation therapies such as targeted temperature management [37], percutaneous coronary intervention (PCI) with ongoing CPR [3,38] and extra-corporeal membrane oxygenation (ECMO) [39]. One recent review mentioned that TTM was reported to improve survival outcomes for patients with both initially shockable and non-shockable rhythms compared to the traditional care before TTM era [40]. Additionally, Primary PCI after successful resuscitation was significantly associated with improved neurological and survival outcomes after discharge irrespective of initial rhythm [41]. Our results suggest that patients with initial non-shockable rhythms with subse- quent shockable rhythm conversion could be considered in future stud- ies which investigate the therapeutic benefit of these aggressive post- resuscitative interventions.

There are some limitations in this study. Firstly, there might be a risk of misclassification bias since cardiac rhythm is dynamic and initial rhythm recorded might not reflect the patient’s true initial rhythm be- fore monitoring was applied. Additionally, subsequent defibrillation

received was used as a surrogate measure for shockable rhythm conver- sion. Although Defibrillator recordings were downloaded for review, this review was not 100%. There might also be other confounders which would have affected the outcomes of OHCA which were not col- lected in this study, such as quality and process of resuscitation, in- hospital care, number of shocks given, time to shocks, etc. There might be a risk of under-reporting and selection bias since some patients could have arrived at the hospital via private transport rather than by EMS, who might not be enrolled into the PAROS registry. Therefore, in those countries which adopt termination of resuscitation in field or use private transport as the major Mode of transport, those OHCA cases reported might not be fully representative of OHCA in those pop- ulations. Lastly, there might be variations in patient, EMS and hospital factors including Resuscitation protocols which might have affected the outcomes. However, this study managed these risks with the usage of multi-national registry involving a relatively large number of cardiac arrest cases and usage of a standardized Utstein template for data collection and adjustment of country variance in all analysis models.

1. Conclusion

Initial shockable rhythm was the strongest predictor for survival. However, conversion to subsequent shockable rhythm significantly im- proved post-arrest survival and neurological outcomes. This study sug- gests the importance of early resuscitation efforts even for initially non- shockable rhythms which has prognostic implications and selection of subsequent post-resuscitation therapy.

Funding Sources

This study is supported by grants from the National Medical Re- search Council (Singapore), Ministry of Health, Singapore, Korea Cen- ters for Disease Control and Prevention. Funding agencies had no role in the research presented in the paper, and the researchers were fully independent in pursuing this research.

Disclosure

Authors have no conflict of interest.

Acknowledgments

We would like to acknowledge the contributions of Ms Susan Yap and Ms Shahidah Ahmad from the Department of Emergency Medicine, Singapore General Hospital and the Singapore Clinical Research Insti- tute for coordination of the study.

References

1. American Heart Association. Cardiac Arrest Statistics; 2015[Available from: http:// www.heart.org/HEARTORG/General/Cardiac-Arrest-Statistics_UCM_448311_Article. jsp, Accessed 18.09.2016].
2. Ong MEH, Do Shin S, De Souza NNA, Tanaka H, Nishiuchi T, Song KJ, et al. Outcomes for out-of-hospital cardiac arrests across 7 countries in Asia: the Pan Asian Resusci- tation Outcomes Study (PAROS). Resuscitation 2015;96:100-8.
3. European Resuscitation Council. European Resuscitation Council: summary of the main changes in the resuscitation guidelines. ERC GUIDELINES 2015; 2015[Available from: http://www.cprguidelines.eu/assets/downloads/ERC_summary_booklet_HRES. pdf, Accessed 18.09.2016].
4. Sasson C, Rogers MA, Dahl J, Kellermann AL. Predictors of survival from out-of- hospital cardiac arrest a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2010;3:63-81.
5. Cobb LA, Fahrenbruch CE, Olsufka M, Copass MK. Changing incidence of out-of- hospital ventricular fibrillation, 1980-2000. JAMA 2002;288:3008-13.
6. Herlitz J, Engdahl J, Svensson L, Young M, Angquist K-A, Holmberg S. Decrease in the occurrence of ventricular fibrillation as the initially observed arrhythmia after out- of-hospital cardiac arrest during 11 years in Sweden. Resuscitation 2004;60:283-90.
7. Hallstrom A, Rea TD, Mosesso Jr VN, Cobb LA, Anton AR, Van Ottingham L, et al. The relationship between shocks and survival in out-of-hospital cardiac arrest patients initially found in PEA or asystole. Resuscitation 2007;74:418-26 [Epub 2007/04/25].
8. Olasveengen TM, Samdal M, Steen PA, Wik L, Sunde K. Progressing from initial non- shockable rhythms to a shockable rhythm is associated with improved outcome after out-of-hospital cardiac arrest. Resuscitation 2009;80:24-9 [Epub 2008/12/17].
9. Leng CT, Berger RD, Calkins H, Lardo AC, Paradis NA, Halperin HR. Electrical induc- tion of ventricular fibrillation for resuscitation from postcountershock pulseless and asystolic cardiac arrests. Circulation 2001;104:723-8.
10. Anantharaman V. Resuscitation update. College of family physicians Singapore, 20. ;

2013.

1. Ong ME, Shin SD, Tanaka H, Ma MH, Khruekarnchana P, Hisamuddin N, et al. Pan- Asian Resuscitation Outcomes Study (PAROS): rationale, methodology, and imple- mentation. Acad Emerg Med 2011;18:890-7.
2. Stiell IG, Wells GA, Field IB, et al. Improved out-of-hospital cardiac arrest survival through the inexpensive optimization of an existing defibrillation program: OPALS study phase II. JAMA 1999;281:1175-81.
3. STATA. biprobit — bivariate probit regression. [Available from: http://www.stata. com/manuals13/rbiprobit.pdf, Accessed 18.09.2016].
4. Nichol G, Thomas E, Callaway CW, Hedges J, Powell JL, Aufderheide TP, et al. Regional var- iation in out-of-hospital cardiac arrest incidence and outcome. JAMA 2008;300:1423-31.
5. Goto Y, Maeda T, Goto YN. Effects of prehospital epinephrine during out-of-hospital cardiac arrest with initial non-shockable rhythm: an observational cohort study. Crit Care 2013;17:R188.
6. Mader TJ, Nathanson BH, Millay S, Coute RA, Clapp M, McNally B. Out-of-hospital Cardiac arrest outcomes stratified by rhythm analysis. Resuscitation 2012;83: 1358-62.
7. Rea TD, Eisenberg MS, Sinibaldi G, White RD. Incidence of EMS-treated out-of- hospital cardiac arrest in the United States. Resuscitation 2004;63:17-24.
8. Atwood C, Eisenberg MS, Herlitz J, Rea TD. Incidence of EMS-treated out-of-hospital cardiac arrest in Europe. Resuscitation 2005;67:75-80.
9. Nadkarni V, Larkin G, Peberdy M, Carey S, Kaye W, Mancini M, et al. National Regis- try of Cardiopulmonary Resuscitation Investigators: first documented rhythm and clinical outcome from in-hospital cardiac arrest among children and adults. JAMA 2006;295:50-7.
10. Kitamura T, Iwami T, Kawamura T, Nagao K, Tanaka H, Hiraide A. Implementation Working Group for the All-Japan Utstein Registry of the Fire and Disaster Manage- ment Agency. Nationwide public-access defibrillation in Japan. N Engl J Med 2010; 362:994-1004.
11. Herlitz J, Svensson L, Engdahl J, Silfverstolpe J. Characteristics and outcome in out-of- hospital cardiac arrest when patients are found in a non-shockable rhythm. Resusci- tation 2008;76:31-6 [Epub 2007/08/22].
12. Kajino K, Iwami T, Daya M, Nishiuchi T, Hayashi Y, Ikeuchi H, et al. Subsequent ven- tricular fibrillation and Survival in out-of-hospital cardiac arrests presenting with PEA or asystole. Resuscitation 2008;79:34-40 [Epub 2008/08/06].
13. Deakin CD, Morrison LJ, Morley PT, Callaway CW, Kerber RE, Kronick SL, et al. Part 8: advanced life support: 2010 international consensus on cardiopulmonary resuscita- tion and Emergency Cardiovascular Care Science with treatment recommendations. Resuscitation 2010;81:e93-174.
14. Kitamura N, Nakada T-a, Shinozaki K, Tahara Y, Sakurai A, Yonemoto N, et al. Subse- quent shock deliveries are associated with increased favorable neurological out- comes in cardiac arrest patients who had initially non-shockable rhythms. Crit Care 2015;19:1-8.
15. Fukuda T, Matsubara T, Doi K, Fukuda-Ohashi N, Yahagi N. Predictors of favorable and poor prognosis in unwitnessed out-of-hospital cardiac arrest with a non-shock- able initial rhythm. Int J Cardiol 2014;176:910-5.
16. Goto Y, Maeda T, Nakatsu-Goto Y. Prognostic implications of conversion from nonshockable to shockable rhythms in out-of-hospital cardiac arrest. Crit Care 2014;18:528.
17. Meaney PA, Nadkarni VM, Kern KB, Indik JH, Halperin HR, Berg RA. Rhythms and outcomes of adult in-hospital cardiac arrest. Crit Care Med 2010;38:101-8.
18. Thomas AJ, Newgard CD, Fu R, Zive DM, Daya MR. Survival in out-of-hospital cardiac arrests with initial asystole or pulseless electrical activity and subsequent shockable rhythms. Resuscitation 2013;84:1261-6 [Epub 2013/03/05].
19. Andrew E, Nehme Z, Lijovic M, Bernard S, Smith K. Outcomes following out-of- hospital cardiac arrest with an initial cardiac rhythm of asystole or pulseless electri- cal activity in Victoria, Australia. Resuscitation 2014;85:1633-9.
20. Gilmore CM, Rea TD, Becker LJ, Eisenberg MS. Three-phase model of cardiac arrest: time-dependent benefit of bystander cardiopulmonary resuscitation. Am J Cardiol 2006;98:497-9.
21. Zheng R, Luo S, Liao J, Liu Z, Xu J, Zhan H, et al. Conversion to shockable rhythms is associated with better outcomes in out-of-hospital cardiac arrest patients with ini- tial asystole but not in those with pulseless electrical activity. Resuscitation 2016.
22. Baskett PJ, Steen PA, Bossaert L. European Resuscitation Council guidelines for resus- citation 2005. Section 8. The ethics of resuscitation and end-of-life decisions. Resus- citation 2005;67(Suppl 1):S171-80 [Epub 2005/12/03].
23. Stiell IG, Nichol G, Leroux BG, Rea TD, Ornato JP, Powell J, et al. Early versus later rhythm analysis in patients with out-of-hospital cardiac arrest. N Engl J Med 2011; 365:787-97.
24. Steen S, Liao Q, Pierre L, Paskevicius A, Sjoberg T. The critical importance of minimal delay between chest compressions and subsequent defibrillation: a haemodynamic explanation. Resuscitation 2003;58:249-58.
25. Tang W, Snyder D, Wang J, Huang L, Chang Y-T, Sun S, et al. One-shock versus three- shock defibrillation protocol significantly improves outcome in a porcine model of prolonged ventricular fibrillation cardiac arrest. Circulation 2006;113: 2683-9.
26. Bailey ED, Wydro GC, Cone DC. Termination of resuscitation in the prehospital setting for adult patients suffering nonTraumatic cardiac arrest. National Association of EMS Physi- cians Standards and Clinical Practice Committee. Prehosp Emerg Care 2000;4:190-5.
27. Dumas F, Grimaldi D, Zuber B, Fichet J, Charpentier J, Pene F, et al. Is hypothermia after cardiac arrest effective in both shockable and nonshockable patients? Insights from a large registry. Circulation 2011;123:877-86.
28. Thomas JL, Bosson N, Kaji AH, Ji Y, Sung G, Shavelle DM, et al. Treatment and out- comes of ST segment elevation myocardial infarction and out-of-hospital cardiac ar- rest in a regionalized system of care based on presence or absence of initial shockable cardiac arrest rhythm. Am J Cardiol 2014;114:968-71.
29. Johnson NJ, Acker M, Hsu CH, Desai N, Vallabhajosyula P, Lazar S, et al. Extracorpo- real life support as rescue strategy for out-of-hospital and emergency department cardiac arrest. Resuscitation 2014;85:1527-32.
30. Grunau BE, Christenson J, Brooks SC. Targeted temperature management after out- of-hospital cardiac arrest: who, when, why, and how? Can Fam Physician 2015; 61:129-34.
31. Grasner JT, Meybohm P, Caliebe A, Bottiger BW, Wnent J, Messelken M, et al. postresuscitation care with mild therapeutic hypothermia and coronary interven- tion after out-of-hospital cardiopulmonary resuscitation: a prospective registry analysis. Crit Care 2011;15:R61 [Epub 2011/02/16].