Resuscitation 84 (2013) Contents lists available at ScienceDirect. Resuscitation. journal homepage:

Transcription

1 Resuscitation 84 (2013) Contents lists available at ScienceDirect Resuscitation journal homepage: Clinical paper Post-resuscitation care and outcomes of out-of-hospital cardiac arrest: A nationwide propensity score-matching analysis Joo Yeong Kim a, Sang Do Shin a,, Young Sun Ro b, Kyoung Jun Song c, Eui Jung Lee d, Chang Bae Park a, Seung Sik Hwang e, For the CardioVascular Disease Surveillance (CAVAS) investigators a Department of Emergency Medicine, Seoul National University College of Medicine, Seoul, South Korea b Department of Preventive Medicine, Graduate School of Public Health, Seoul National University, Seoul, South Korea c Department of Emergency Medicine, Seoul National University Boramae Medical Center, Seoul, South Korea d Department of Emergency Medicine, Korea University College of Medicine, Seoul, South Korea e Department of Social Medicine, Inha University School of Medicine, Incheon, South Korea article info abstract Article history: Received 10 December 2012 Received in revised form 4 February 2013 Accepted 16 February 2013 Keywords: Cardiac arrest Hypothermia Intensive care Outcome Objective: This study aimed to determine whether active post-resuscitation care (APRC) was associated with improved out-of-hospital cardiac arrest (OHCA) outcomes on a nationwide level. Methods and results: We used a national OHCA cohort database consisting of hospital and ambulance data. We included all survivors of OHCA, excluding patients with non-cardiac etiology, younger than 15 years, and with unknown outcomes, from (2008 to 2010). The APRC was defined when the OHCA patients received mild therapeutic hypothermia (MTH) or active cardiac care (ACC), such as intravenous thrombolysis, percutaneous coronary intervention, coronary artery bypass surgery, and pacemaker/implantable cardioverter defibrillator insertion, as well as routine intensive care; patients receiving conservative postresuscitation care (CPRC) served as the other group. The primary and secondary outcomes were survival to discharge and a good neurological outcome (cerebral performance category [CPC] 1 2), respectively. We extracted propensity-matched samples to control for selection bias. A multivariable logistic regression analysis was used to compare the APRC and CPRC groups adjusting for potential risks to calculate the adjusted odds ratios (ORs) and 95% confidence intervals (95% CIs). Of total 64,155 patients, 4557 survived to admission and were included in the final analysis. Out of these patients, 1599 (35.1%) cases survived to discharge, and 499 (11.0%) cases were discharged with good neurological recoveries. Overall, 695 cases (15.3%) received any APRC, including MTH (n = 377, 8.3%) and ACC (370, 8.1%). The outcomes was better in the APRC group than in the CPRC group for survival to discharge (58.7% vs. 30.8%, p < 0.001) and good neurological outcome (27.2% vs. 8.0%, p < 0.001), respectively. In the total cohort, the adjusted ORs of the APRC group compared to those the CPRC group were 2.15 (95% CI ) for survival to discharge and 2.54 (95% CI ) for a good neurological outcome. In the propensity score-matched cohort, the adjusted ORs for survival to discharge and good neurological outcome of APRC were significantly favorable. Conclusions: Active post-resuscitation care resulted in significantly improved outcomes in adult OHCA patients with a presumed cardiac etiology in a nationwide, retrospective, observational study Elsevier Ireland Ltd. All rights reserved. 1. Introduction A Spanish translated version of the summary of this article appears as Appendix in the final online version at Corresponding author at: Department of Emergency Medicine, Seoul National University College of Medicine, 101 Daehak-Ro, Jongno-Gu, Seoul , South Korea. Tel.: ; fax: addresses: blj01he@gmail.com (J.Y. Kim), shinsangdo@medimail.co.kr, sdshin@snu.ac.kr (S.D. Shin), Ro.youngsun@gmail.com (Y.S. Ro), skciva@gmail.com (K.J. Song), ironlyj@gmail.com (E.J. Lee), ipanema2@medimail.co.kr (C.B. Park), cyberdoc@inha.ac.kr (S.S. Hwang). Out-of-hospital cardiac arrest (OHCA) is a major cause of unexpected deaths worldwide. 1 The outcomes of OHCA have been improving thanks to the evidence-based cardiopulmonary resuscitation guidelines followed by numerous clinical and communities efforts in the chain of survival, 2,3 while most other outcomes have remained low. 1 Post-resuscitation care procedures include more active postresuscitation care (APRC) procedures, such as mild therapeutic hypothermia (MTH), active cardiac care (ACC) including percutaneous coronary intervention (PCI), coronary artery bypass /$ see front matter 2013 Elsevier Ireland Ltd. All rights reserved.

2 J.Y. Kim et al. / Resuscitation 84 (2013) surgery (CABG), and the insertion of an implantable cardioverterdefibrillator (ICD)/pacemaker (PM), as well as the usual conservative post-resuscitation care (CPRC) procedures, including stabilization of vital signs using vasopressors, ventilation support, general supportive care, glucose control and antiepileptic management. 4 Of these procedures, APRC procedures are challenging options in many hospital settings because they require more multidisciplinary approaches and more advanced skills, as well as much greater resources than are required for the usual intensive conservative care for critically ill patients. In 2003, the American Heart Association issued a guideline recommending that MTH be used to treat cardiac arrest with shockable rhythm, 5 and the AHA subsequently incorporated the procedure into its 2005 treatment guidelines, 6 such that it finally became the 5th link in the chain of survival that is emphasized in post-cardiac arrest care. 4 Active cardiac care, such as with PCI/CABG and ICD/PM, has also been proven to result in better outcomes in OHCA survivors in advanced hospital settings. 7,8 However, a larger proportion of hospitals still prefer to provide only CPRC rather than APRC 9 11 due to unclear indications when selecting treatment options, cultural barriers, lack of interdisciplinary cooperation, lack of resources and knowledge, and cost issues. 12,13 This applies to the current care systems, except for the urban academic medical centers, in Korea. APRC might not be considered as a single treatment option but rather as a multidisciplinary treatment care bundle, which can provide a higher standard care accompanying the procedure itself. There have been few studies on how many patients have been treated, which patients have been treated, and what size of effects has been found with APRC, particularly on a nationwide level. Post-resuscitation care options are too dynamic and complex, according to patients statuses by time period, to implement each protocol or treatment option easily and step by step in clinical practice. 4,14 Hospitals cannot provide treatment options such as MTH or ACC unless there is a dedicated multidisciplinary active post-resuscitation care team with a great amount of experience, skill and resources. A combined procedure (PCI with MTH), as well as each individual procedure, for OHCA was reported to be safe and effective in small case series 15,16 or in a population series. 17 Many combinations of treatment options, such as MTH and CABG, MTH with ICD, and PCI followed by PM or ICD, can be considered according to the patient s status. However, combinations can only be chosen by clinicians when they are available. For example, although the clues are very ambiguous for diagnosing ST elevation in MI in a clinical setting, the clinician can choose PCI if he or she has access to additional cardiologic intervention teams in real time. This accessibility can be a proxy measure of organized bundles of post-resuscitation care. Therefore, APRC can be regarded as a proxy of accessibility to a particular comprehensive post-resuscitation care level to a larger extent than can each segmented procedure provided to patients in most other studies. The objective of this study was to determine the effects of active post-resuscitation care procedures, including MTH and ACC (PCI/CABG, and ICD/PM), on the outcomes of OHCA in a nationwide, population-based OHCA survivor cohort. 2. Methods 2.1. Study setting The Korean emergency medical services (EMS), which is a single-tiered, government-based system, provides a basic-tointermediate service level of ambulance services from the sixteen provincial headquarters of the national fire department and supports a population of approximately 50 million. 18,19 Ambulance crews can administer CPR at the scene and during transport with automatic external defibrillation and advanced airway management under direct medical control. Advanced cardiac life support is limited in most areas of the country, but it is available in emergency departments. The public access defibrillation program began in 2009 but was not widely accepted until recently. The national government designates emergency departments (EDs) according to the resources and functional requirements of three levels; level 1 (n = 20) and level 2 (n = 110) EDs, which have more resources and better facilities for emergency care, must be staffed by emergency physicians 24 h a day and 365 days a year, whereas level 3 EDs (n = 330) can be staffed by general physicians. 18 Level 4 is non-ed facility. The EMS Act indicates that the structural and functional status of all EDs should be evaluated every year by an audit committee. This annual mandatory evaluation program includes the EDs resource fulfillment of requirements and functional performance, such as door to needle or balloon time for ST segment elevation myocardial infarction, whereas postresuscitation care is not evaluated for survivors after the OHCA. Every level 1 ED has 20 emergency intensive care unit (EICU) beds for emergency patients that are independently attended by the emergency department, to which survivors from OHCA are admitted and in which they are managed. Level 2 EDs can use ICU beds only when they are available after competing with other departments. These ICUs are usually operated by other specialty departments. Cardiopulmonary resuscitation guidelines from international societies are accepted by the academic societies and are recommended in clinical practice For example, the 2005 and 2010 AHA guidelines are accepted by national organizations Data source This research was part of a population-based retrospective study, the Cardiovascular Disease Surveillance (CAVAS) project, which has been conducted since 2006 in collaboration with and supported by the National Emergency Management Agency (NEMA) and the Centers for Disease Control and Prevention (CDC) of the government of the Republic of Korea This project has constructed a nationwide cohort for EMS-assessed OHCA, which was drawn from an EMS run sheet for demographic and Utstein information. This cohort study was followed by a hospital medical record review for hospital care and post-resuscitation care, and this review continued through the current year. All information from the EMS run sheets (patient records) is input by EMS providers immediately after the transport of OHCA patients and is stored in the NEMA electronic server. Medical record review experts from the Korean CDC visited all of the hospitals and captured the information for the outcomes and for postresuscitation care using a structured survey form. All of the items are defined in the medical record review guidelines developed by the project quality management committee (QMC), including definitions, inclusion and exclusion criteria, examples, and warnings. All reviewers are trained prior to joining the project by the project QMC, which consists of emergency medicine physicians, epidemiologists, statistical experts, cardiologists, and medical record review experts. When the reviewers cannot determine the information characteristics for a case, they can contact an emergency medicine physician from the quality management committee for decision-making, for the acquisition of more knowledge and for supervision Selection of study participants The data were extracted between January 2008 and December All adults with OHCA with a cardiac etiology were included.

3 1070 J.Y. Kim et al. / Resuscitation 84 (2013) Patients were excluded who had a non-cardiac etiology, who were younger than 15 years old, or who were without available outcomes. Patients were identified as having an arrest of cardiac etiology by medical record review. We presumed non-cardiac etiology if the medical record described the cause of death definitely as trauma, drowning, poisoning, burn, asphyxia, and hanging. Without definite non-cardiac etiology, we presumed it cardiac etiology. Patients were also excluded if they had an alert mental status at arrival at the ED through EMS resuscitation and successful defibrillation. Of the included patients, those cases without an alert mental state that survived to be admitted to the hospital intensive care unit were analyzed for this study. Allocation of the patients to each group was also done by medical record review. If a patient received either MTH or ACC, that OHCA victim was defined as an APRC group. On the contrary, if a patient received none of APRC, that OHCA victim was defined as a CPRC group Main exposure: conservative post-resuscitation care vs. active post-resuscitation care Conservative post-resuscitation care (CPRC) was defined as admission to an intensive care unit, which provided ventilator care, vasopressor administration, sedatives, neuromuscular blockers, anticonvulsive agents, antipyretics, and antibiotics if indicated. Active post-resuscitation care (APRC) was operatively defined as follows: (1) mild therapeutic hypothermia (MTH), which was defined as a case receiving mild therapeutic hypothermia using methods such as external cooling (water, fanning, or ice padding), internal cooling (gastric lavage, bladder cooling, or intravascular cooling using a catheter) or mixed cooling; although the target temperature should be between 32 and 34 C, and the minimum duration of hypothermia should be at least 12 h, we regarded the case as part of the hypothermia-treated group if the hospital attempted the procedure regardless of either the duration or the withdrawal of the procedure due to death during induced hypothermia; (2) active cardiac care (ACC), which included one of following procedures: (a) a case receiving reperfusion therapy for suspected myocardial infarction, regardless of intravenous (IV) thrombolysis or PCI, as soon as possible after survival to admission; if the procedure was stopped in the angiography room during the procedure after the insertion of the catheter due to arrest, the case was defined as part of the reperfusion therapy group; (b) a case receiving CABG surgery; and (c) a case receiving a temporary or permanent PM or ICD. MTH should be performed as soon as possible and be finished within after 48 h after survival to admission, whereas the other procedures might be performed at the ED as soon as possible or during the hospital stay. The APRC group was regarded as the patients who received APRC, if indicated Outcome measurements The primary outcome was survival to discharge, which was defined as an OHCA victim being admitted to a hospital and then finally being discharged to home or transferred to another facility such as a rehabilitation center. Survival to discharge was defined when survival to discharge was described in a discharge abstract or in other documentation. The secondary outcome was good neurological status, which was defined as a cerebral performance categories (CPC) score of 1 (good recovery) or 2 (moderate disability). The CPC score was categorized by medical recorder reviewers on the basis of discharge summary abstracts or documentation in medical records. Discharge summary abstracts or documentation was drafted by inpatient care doctors, and the summaries were usually reviewed and revised for greater validity by the hospital medical record review team because the summaries were utilized for national health insurance claim data after discharge. However, there was no standard recording frame for the CPC scale among the participating hospitals. Therefore, the categories were decided on and coded by medical record reviewers of the Korean CDC who were supervised by the project QMC via on- and off-line consultation for each case in which the CPC category was controversial Additional covariates Additional variables obtained from the hospital records included sex, age, witnessed arrest (yes, no, and unknown), place of the event (public vs. private vs. unknown), bystander CPR (yes and no), pre-hospital ROSC (yes and no), pre-hospital defibrillation (yes and no), initial ECG (ventricular fibrillation/ventricular tachycardia, pulseless electrical activity, asystole, and unknown), elapsed time interval from call to ambulance arrival at the scene, time interval from call to ambulance arrival at the ED, and level of ED (levels 1 4). We defined data as unknown when the information was not available from medical records regardless of site (e.g., original ECG sheet). For example witnessed was regarded as unknown when there was no record about who witnessed the event. The use of an electronic database prevented skipped entries Main analysis A multivariable logistic regression analysis was used for the total cohort to compare survival to discharge between the APRC and CPRC groups and to calculate the adjusted ORs and 95% confidence intervals (95% CIs). The main exposure (intervention in this study) was whether patients received APRC or CPRC. We used the Utstein template risk factors listed in the previous literature, which included sex, age, witness status, pre-hospital defibrillation, place, bystander CPR, elapsed time interval from call to ambulance arrival at the scene, time interval from call to ambulance arrival at the ED, and initial ECG rhythm. 22 We added the following covariates: level of ED (levels 1 4) adjusted for ED capacity and metropolitan vs. non-metropolitan status by a population of one million to adjust for geographical variation in community performance and resource factors. We tested interaction between the exposure variable and potential covariates. First, we performed the chunk test and calculated 2 log L for the full model with potential interaction terms and for a restricted model without any interaction terms to determine whether there was a significant difference between the two models or not. Consequently, we performed the backward elimination process for each potential interaction term by order of significance level. Finally, we obtained a final model considering the interaction between the exposure variable and other covariates. We assessed the confounding effects of covariates. We tested the associations and calculated the adjusted ORs with 95% CIs in the final model and compared all adjusted ORs with 95% CIs by dropping each covariate (one, two, three, etc., to all) from the model. First, the adjusted OR was placed at 10% plus or minus the original adjusted OR, and we input the covariates into the model. If the adjusted OR was within the 10% limits, we calculated the difference between the upper and lower confidence intervals and compared the differences between the original and other models. We chose for the final analysis the model that showed the smallest difference between confidence intervals. We matched the propensity score for APRC vs. CPRC to determine the effects of APRC vs. CPRC on the outcomes of patients under equivalent distribution of potential covariates. The APRC procedures appeared likely to be selected for potentially indicated patients with suspected myocardial infarction (reperfusion therapy or CABG) and suspected fatal arrhythmia (PM/ICD insertion). Although we already excluded patients with an alert mental state, APRC might be selected by clinicians when the patient was likely

4 J.Y. Kim et al. / Resuscitation 84 (2013) Fig. 1. Study population and outcomes. EMS; emergency medical service; OHCA: out-of-hospital cardiac arrest good neurological outcome: cerebral performance category class 1 or 2. to survive. To remove this selection bias, propensity-matched samples were extracted for analysis. We calculated propensity scores to a maximum of 10 decimal places. Patients receiving APRC (the exposure group) were matched to the closest non-exposure group (CPRC) member in each model whose propensity score differed by less than There was no overlapping of non-exposure cases. The number of propensity-matched samples was dependent on the proportion of APRC group cases among all of the eligible patients. The adjusted odds ratios (ORs) with 95% confidence intervals (95% CIs) for the exposure variable were calculated for the total cohort and for the propensity-matched cohort. The goodness of fit of the multivariable logistic regression model was tested for calibration using the Hosmer Lemeshow test with a chi-square analysis. The analysis was performed using SAS software (SAS Institute, Cary, S.C., USA), version 9.1 for Windows, and we referred to a SAS Macro to extract the propensity-matched cohort (see the page at p pdf) Subgroup analysis We determined the effects of MTH, ACC, and coronary artery reperfusion (CAR) on the outcomes of the total cohort. MTH included all methods (external evaporation, surface cooling material, intravenous device cooling, and internal cavity cooling) provided to the patients. ACC also included all cardiac procedures, such as CAR therapy (IV or PCI), CABG, and PM/ICD insertion. Finally, we compared outcomes between patients who received CAR therapy (CAR group) vs. did not receive any CAR (non-car). The adjusted odds ratios (ORs) with 95% confidence intervals (95% CIs) for the each procedure group were calculated, adjusting for MTH for ACC, ACC for MTT, CAR for MTH and CABG/PM/ICD as well as for the covariates in the models in the main analysis. The goodness of fit of the multivariable logistic regression model was tested for calibration using the Hosmer Lemeshow test with a chi-square analysis. 3. Results 3.1. Demographic findings Among the 64,155 EMS-assessed OHCA patients with available outcomes, 44,794 adult patients with presumed cardiac etiologies were enrolled after excluding patients with non-cardiac etiologies (n = 17,970) and pediatric (n = 1391) patients. Of these enrollees, 4557 patients survived to admission and were used for the final analysis. There were no patients with a normal alert mental state among the 496 patients (10.9%) receiving ROSC before arriving at the ED. Of these survivors, 1599 (35.1%) cases survived to discharge, and 499 (11.0%) cases were discharged with good neurological recovery (Fig. 1). Overall, 695 cases (15.3%) were treated with APRC, and 3862 (84.7) cases were treated with CPRC. Members of the APRC group were significantly younger than those of the CPRC group (66.5% vs. 56.6% between 15 and 64 years old). Bystander CPR (15.7% vs. 6.3%), ROSC by EMS (19.6% vs. 9.3%), and initial shockable rhythm (26.2% vs. 10.8%) were also significantly higher in the APRC group than in the CPRC group. Members of the APRC group had statistically higher survival to discharge rates (58.7% vs. 30.8%) and better neurological outcomes (27.2% vs. 8.0%) than did those of the CPRC group (Table 1). MTH was performed in patients from the metropolitan group, the bystander CPR group, and the level 1 ED group. Of the initial ECGs, MTH was performed more often in the PEA group than in the VF/VT group. Coronary reperfusion therapy was performed more often in non-metropolitan regions, but the remaining factors were otherwise similar to those for MTH (Table 2). APRC treatment consisted of MTH (n = 377, 8.3%; 207 cases of external cooling using water, fan, and cold blanket; 94 cases of external cooling using surface cooling materials; 47 cases of intravenous cooling; 118 cases of internal cooling using Levin tube or urine catheter) and ACC (coronary reperfusion therapy (n = 370, 8.1%; 44 cases of IV thrombolysis), 267 cases of PCI, 18 cases of CABG, and 61 cases of PM/ICD insertion) (Table 2). The proportion

5 1072 J.Y. Kim et al. / Resuscitation 84 (2013) Table 1 Demographic findings of the conservative intensive care and active post-resuscitation care groups for 4557 patients who survived to admission. Variables Total Post-resuscitation care p-value CPRC APRC Total, N (%) Sex, N (%) <0.001 Male Female Age group, N (%) <0.001 Adult (15 64 years) Elderly (older than 65 years) Metropolitan, N (%) No Yes Place, N (%) <0.001 Public Private Unknown Witnessed status, N (%) <0.001 No Yes Unknown Bystander CPR, N (%) <0.001 No Yes Initial ECG, N (%) <0.001 VF/VT PEA Asystole Unknown Interval, mean (SD), minutes Call to arrival at scene Call to arrival at ED Pre-hospital defibrillation, N (%) < No Yes ROSC by EMS, N (%) < No Yes ED level, N (%) <0.001 Level Level Level Level Outcomes, N (%) Survival to discharge <0.001 Good neurological status <0.001 CPCR: conservative post-resuscitation care. APRC: active post-resuscitation care. CPR: cardiopulmonary resuscitation; ECG: electrocardiography; VF: ventricular fibrillation; VT: ventricular tachycardia: PEA; pulseless electrical activity: ROSC: return of spontaneous circulation: EMS emergency medical services: ED: emergency department. of any APRC increased according to year: 10.6% in 2008, 16.1% in 2009, and 17.9% in Propensity score-matched cohort We extracted the propensity score-matched cohort. There were 678 patients in each of the two groups (APRC and CPRC). Both groups showed a similar distribution of demographic features and risk factors. The survival to discharge rate was significantly higher in the APRC group (58.4%) than in the CPRC group (44.5%) (p-value < 0.001). The good neurological outcome rate was also significantly higher in the APRC group (26.4%) than in the CPRC group (15.8%) (p-value < 0.001) (Table 3) Multivariable analysis for the total cohort and the propensity score-matched cohort In the total cohort, the adjusted ORs of the APRC group vs. the CPRC group were 2.15 (95% CI ) for survival to discharge (H L chi-square = 2.1; p-value = 0.977) and 2.54 (95% CI ) for good neurological outcomes (H L chi-square = 7.8; p-value = 0.451) (Table 4). In the propensity score-matched cohort, the adjusted ORs of the APRC group vs. the CPRC group were 1.95 (95% CI ) for survival to discharge (H L chi-square = 12.2; p-value = 0.145) and 2.29 (95% CI ) for good neurological outcomes (H L chisquare = 3.8; p-value = 0.873) (Table 4) Subgroup analysis The adjusted ORs of the MTH group vs. the non-mth group were 1.84 (95% CI ) for survival to discharge (H L chisquare = 3.2; p-value = 0.922) and 1.29 (95% CI ) for good neurological outcomes (H L chi-square = 4.8; p-value = 0.778). The adjusted ORs of the ACC group vs. the non-acc group were 2.35 (95% CI ) for survival to discharge (H L chi-square = 3.2; p-value = 0.922) and 3.73 (95% CI ) for good neurological outcomes (H L chi-square = 4.8; p-value = 0.778). The adjusted ORs of the CAR group vs. the non-car group were 2.10 (95% CI ) for survival to discharge (H L chi-square = 5.6;

6 Table 2 Distribution of active post-resuscitation care by potential risk factors. Covariates Total Mild therapeutic hypothermia Active cardiac care Any MTT procedure EXT-EVA EXT-SUR IVD INT-CAV Any ACC procedure IV-THR PCI CBGA PM/ICD N N % % % % % N % % % % % Year, total Sex, N (%) Male Female Age group, N (%) Adult (15 to 64 years) Elderly (more than 65 years) Metropolitan, N (%) No Yes Place, N (%) Public Private Unknown Witnessed status, N (%) No Yes Unknown Bystander CPR, N (%) No Yes ECG VF/pulseless VT PEA Asystole Unknown Interval, mean (SD), minutes Call to arrival to scene 6.2(2.5) 6.3(4.7) Call to arrival to ED 20.1(8.5) 20.4(10.8) Prehospital defibrillation, N (%) No Yes ROSC by EMS No Yes ED level Level Level Level Level Outcomes Survival to discharge Good neurological status MTT: mild therapeutic hypothermia; EXT-EVA: external cooling using cold water, fan, and cold blanket; EXT-SUR: external cooling attaching surface materials; IVD: intravenous device cooling; INT-CAV: internal cavity cooling via Levin tube or urinary catheter. ACC: active cardiac care; IV-THR: intravenous thrombolysis; PCI: primary coronary artery intervention; CABG: coronary artery bypass graft; PM/ICD: pacemaker/implantable cardioverter/defibrillator. CPR: cardiopulmonary resuscitation; ECG: electrocardiography; VF; ventricular fibrillation; VT: ventricular tachycardia; PEA: pulseless electrical activity; ROSC: return of spontaneous circulation; EMS emergency medical services; ED: emergency department. J.Y. Kim et al. / Resuscitation 84 (2013)

7 1074 J.Y. Kim et al. / Resuscitation 84 (2013) Table 3 Distribution of potential risk factors in the propensity score-matched cohort. Variables Total Post-resuscitation care P-value CPRC APRC Total, N (%) Sex, N (%) Male Female Age, mean (SD), year Metropolitan, N (%) No Yes Place, N (%) Public Private Unknown Witnessed status, N (%) No Yes Unknown Bystander CPR, N (%) No Yes Initial ECG, N (%) VF/pulseless VT PEA Asystole Unknown Interval, mean (SD), minutes Call to arrival to scene Call to arrival to ED Prehospital defibrillation, N (%) No Yes ROSC by EMS, N (%) No Yes ED level, N (%) Level Level Level Level Outcomes, N (%) Survival to discharge <0.001 Good neurological status <0.001 CPCR: conservative post-resuscitation care. APRC: active post-resuscitation care. CPR: cardiopulmonary resuscitation; ECG: electrocardiography; VF: ventricular fibrillation; VT: ventricular tachycardia; PEA: pulseless electrical activity; ROSC: return of spontaneous circulation; EMS emergency medical services; ED: emergency department. p-value = 0.693) and 2.98 (95% CI ) for good neurological outcomes (H L chi-square = 5.9; p-value = 0.695) (Table 5). 4. Discussion This study aimed to determine the association between APRC and improved outcomes at the nationwide level. We found the APRC group showed a significantly higher survival to discharge rate (adjusted OR = 2.17) and a good neurological discharge rate in the total cohort (Adjusted OR = 2.54). Significant associations were found in the propensity score-matched cohort. Previous studies showed a wide variation in the proportion of cases receiving MTH in OHCA. A recent survey study from Italy reported that MTH was received for post-resuscitation care in 16% of participating hospitals (n = 404) and used for both patients with shockable and non-shockable rhythm. 23 A survey on the use of hypothermia carried out in United Kingdom and sixty-seven (28.4%) ICUs have answered that they cooled patients after cardiac arrest among 246 ICUs. 24 A total of 11.3% of the 203 participating hospitals provided MTH in another study. 9 During 2004 and 2005, patients with OHCA were treated with MTH in 19 out of 20 intensive care units in Finland, and the proportion of patients treated with MTH increased from 4% in 2002 to 28% in Our data showed a low proportion of induced MTH (8.3%) over the course of three years among 315 participating hospitals (Table 2). The trend was favorable for the implementation of hypothermia (5.6% in % in 2010). Coronary reperfusion therapy was the most important procedure for removing the presumed cause of the collapse and for supporting the hemodynamic status during the post-resuscitation care. However, PCI was provided to patients with OHCA in only 5.9% of cases in this study (Table 2). Favorable effects of coronary intervention on the outcomes were reported in case series and studies in a specifically selected patient group with a suspected cause of MI 25 and with primary rhythms of VF/VT 26 and in the entire patient group without a specific category. 17,27 These studies showed better outcomes in patients who received a PCI, regardless of MTH or other treatment options. The proportion of patients who received a PCI was much lower than that observed in the above studies. More MTH and ACC were provided in the highest level of ED (21.8% and 13.5%) than in the lower levels of ED (7.1% and 8.8% in level 2 and both less than 2% in levels 3 and 4). No interactions between APRC and level of ED were observed in the models, indicating that the effects were consistent regardless of the level of

8 J.Y. Kim et al. / Resuscitation 84 (2013) Table 4 Multivariable logistic regression analysis for outcomes in the total cohort and the propensity score-matched cohort. Cohort Outcome Total Survival Unadjusted OR Adjusted OR a HL test b N N % OR 95% CI OR 95% CI Chi square P-value Total cohort Survival to discharge CPRC APRC Good neurological outcome CPRC APRC Propensity score-matched cohort Survival to discharge CPRC APRC , Good neurological outcome CPRC APRC CPRC: conservative intensive care group. APRC: active post-resuscitation care group. OR: odds ratio; 95% CI: 95% confidence interval. a Adjusted for age, sex, metropolitan vs. non-metropolitan, witness (yes, no, unknown), place (public vs. private vs. unknown), bystander cardiopulmonary resuscitation (yes, no), prehospital defibrillation (yes, no), return of spontaneous circulation at emergency medical service stage (yes, no), elapsed time interval from call to ambulance arrival at the scene (minutes), time interval from call to ambulance arrival at the emergency department (minutes), initial ECG, and emergency department level (levels 1 4). All interaction terms were tested and removed in the final model. Confounding effects of all variables were assessed. b The Hosmer Lemeshow test was used for the goodness of fit of the model. The chi-square and p-value tests showed good fitness. ED. A cardiac resuscitation center program has been encouraged to provide mandatory evaluations of coronary artery and reperfusion treatment if needed Our study supports the association between active coronary reperfusion therapy and improved outcomes. In a subgroup analysis, MTH resulted in a significantly higher survival to discharge rate but not a good neurological recovery rate (adjusted OR = 1.29 (95% CI )), while ACC showed significantly improved outcomes in both rates. Original studies have shown the positive effects of treatment with MTH on better neurological outcomes in well-designed, randomized trials. 28,29 One study reported that MTH was associated with favorable outcomes in OHCA with non-shockable rhythms, having excluded mortality cases within the first 24 h after ROSC. 30 In this study, we found that more than 80% of patients who received MTH were patients with non-shockable rhythms (13.5% in PEA, 43.0% in asystole, and 23.9% in unknown rhythm). Many clinicians have performed therapeutic hypothermia for non-shockable rhythms, although the evidence for success with such treatment is not high. 4 MTH might not be associated with good neurological outcomes due to higher proportions of non-shockable rhythms. Table 5 The effect of mild therapeutic hypothermia and active cardiac care on the outcomes of adult OHCA with presumed cardiac etiology (N = 4557). Outcome by procedure Total Outcome, yes Unadjusted OR Adjusted OR a HL test for adjusted model b N N % OR 95% CI OR 95% CI Chi square P-value Mild therapeutic hypothermia (MTH) Survival to discharge Non-MTH MTH Good neurological outcome Non-MTH MTH Active cardiac care (ACC) Survival to discharge Non-ACC ACC Good neurological outcome Non-ACC ACC Coronary artery reperfusion (CAR) Survival to discharge Non-CAR CAR Good neurological outcome Non-CAR CAR OR: odds ratio, 95% CI: 95% confidence interval. a Adjusted for age, sex, metropolitan vs. non-metropolitan, witness (yes, no, unknown), place (public vs. private vs. unknown), bystander cardiopulmonary resuscitation (yes, no), prehospital defibrillation (yes, no), return of spontaneous circulation at emergency medical service stage (yes, no), elapsed time interval from call to ambulance arrival at the scene (minutes), time interval from call to ambulance arrival at the emergency department (minutes), initial ECG, and emergency department level (levels 1 4). The MTH or ACC was added to be adjusted for each final counterpart model (ACC for analysis of MTH vs. non-mth group and MTH for analysis of ACC vs. non-acc group). All interaction terms were tested and removed in the final model. Confounding effects of all variables were assessed. The analysis on CAR vs. non-car was analyzed adjusted for coronary artery bypass graft and pacemaker insertion as well as all variables used in ACC model. b The Hosmer Lemeshow test (HL test) was used for the goodness of fit of the model. The chi-square and p-value tests showed good fitness.

9 1076 J.Y. Kim et al. / Resuscitation 84 (2013) ACC, including CABG and PM/ICD, are not always tried immediately after arriving ED. This could raise the problem of selection bias that those who die early will be excluded from the APRC group. In this study, we only enrolled those who survived to admission, which in other words, excluded those who did not achieve ROSC and died extremely early. The coronary artery reperfusion intervention excluding other cardiac intervention (CABG or PM/ICD insertion) to avoid selection bias was significantly associated with better outcomes in subgroup analysis. 5. Limitations This study has several limitations to the interpretation of its results and the generalization of these results to other settings. First, this study was not a controlled trial but was a retrospective, observational study. APRC might be selected for patients with a higher likelihood of survival, although we tried to match the risk using propensity scores. Second, the study setting was different from the paramedic level of EMS systems such as those in North American or European communities. A major part of ROSC was achieved in the ED rather than by EMS in this study. Third, no standardized treatment protocol for APRC (MTH and ACC) exists throughout the entire country, although many hospitals accept scientific guidelines for the use of MTH and ACC. Those hospitals might apply different enrollment criteria, protocols and methods, but we categorized all MTH treatments as the same protocol; we only restricted the definition of this treatment to its use as soon as possible and to its being finished within 48 h after ROSC. However, it is well known that the time required to initiate hypothermia treatment can lead to very different outcomes. Additionally, external cooling and internal cooling were regarded as the same method. These crude definitions might have biased the outcomes. Fourth, some important factors had a wide range of unknown information due to the use of retrospective medical record reviews. For example, the inclusion of excluded cases with unknown outcomes might have led to different results. Fifth, initial ECGs were usually obtained at the ED rather than by the EMS provider. Even fewer EMS providers obtained ECGs at the scene and transported the patients to the ED while providing CPR. Thus, the first ECG was generally obtained at the ED by a physician. This time difference might be responsible for the very low proportion of patients with shockable rhythms, compared with patients from other countries. This type of systemic difference should limit the generalization of the study results to other countries. Finally, hospital characteristics related to patient volume have their own effects on the outcomes. 18 We tested for an association between hospital volume for post-resuscitation care and outcomes. No association was found, so we removed the variable in the final models. Other hospital variables, including human resources, ICU capacities, and number of hospital beds, were not adjusted for in the models. Unexplained effects due to these factors may have been hidden in this study. 6. Conclusion In this nationwide, observational cohort study, OHCA victims who received any APRC including either MTH or ACC treatment options showed significantly better outcomes than those who only received CPRC. These findings might support the systemic inclusion of the fifth link in the chain of survival to improve the outcomes of OHCA. Prospective studies are needed to strengthen the outcome of APRC. Conflict of interest statement No conflicts of interest for all authors are in this study. Previous presentation This paper was presented at the National Association of Emergency Medical Services Physician annual congress in January 2011 and at the American Heart Association Resuscitation Science Symposium in Los Angeles in Acknowledgement This study was supported by the National Emergency Management Agency of Korea and the Korean Centers for Disease Control and Prevention. The study was funded by the Korean Centers for Disease Control and Prevention ( ). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at References 1. Berdowski J, Berg RA, Tijssen JG, Koster RW. Global incidences of out-of-hospital cardiac arrest and survival rates: systematic review of 67 prospective studies. Resuscitation 2010;81: Perkins GD, Brace SJ, Smythe M, Ong G, Gates S. Out-of-hospital cardiac arrest: recent advances in resuscitation and effects on outcome. Heart 2012;98: Fugate JE, Brinjikji W, Mandrekar JN, et al. Post-cardiac arrest mortality is declining: a study of the us national inpatient sample Circulation 2012;126: Peberdy MA, Callaway CW, Neumar RW, et al. Part 9: post-cardiac arrest care: 2010 american heart association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122:S Nolan JP, Morley PT, Hoek TL, Hickey RW. Therapeutic hypothermia after cardiac arrest. An advisory statement by the advancement life support task force of the international liaison committee on resuscitation. Resuscitation 2003;57: ECC Committee, Subcommittees and Task Forces of the American Heart Association american heart association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2005;112:IV Noc M. Urgent coronary angiography and percutaneous coronary intervention as a part of postresuscitation management. Crit Care Med 2008;36:S Kern KB, Rahman O. Emergent percutaneous coronary intervention for resuscitated victims of out-of-hospital cardiac arrest. Catheter Cardiovasc Interv 2010;75: Walters EL, Morawski K, Dorotta I, et al. Implementation of a post-cardiac arrest care bundle including therapeutic hypothermia and hemodynamic optimization in comatose patients with return of spontaneous circulation after out-of-hospital cardiac arrest: a feasibility study. Shock 2011;35: Wang HE, Thomas JJ, James D, et al. Post-cardiac arrest therapeutic hypothermia: overcoming the barrier of workplace culture and other implementation lessons. Jt Comm J Qual Patient Saf 2011;37: Oksanen T, Pettila V, Hynynen M, Varpula T. Therapeutic hypothermia after cardiac arrest: implementation and outcome in finnish intensive care units. Acta Anaesthesiol Scand 2007;51: Nichol G, Aufderheide TP, Eigel B, et al. Regional systems of care for out-ofhospital cardiac arrest: a policy statement from the american heart association. Circulation 2010;121: Bobrow BJ, Kern KB. Regionalization of postcardiac arrest care. Curr Opin Crit Care 2009;15: Sunde K, Pytte M, Jacobsen D, et al. Implementation of a standardised treatment protocol for post resuscitation care after out-of-hospital cardiac arrest. Resuscitation 2007;73: Wolfrum S, Pierau C, Radke PW, Schunkert H, Kurowski V. Mild therapeutic hypothermia in patients after out-of-hospital cardiac arrest due to acute stsegment elevation myocardial infarction undergoing immediate percutaneous coronary intervention. Crit Care Med 2008;36: Batista LM, Lima FO, Januzzi Jr JL, Donahue V, Snydeman C, Greer DM. Feasibility and safety of combined percutaneous coronary intervention and therapeutic hypothermia following cardiac arrest. Resuscitation 2010;81: Dumas F, White L, Stubbs BA, Cariou A, Rea TD. Long-term prognosis following resuscitation from out of hospital cardiac arrest: role of percutaneous coronary intervention and therapeutic hypothermia. J Am Coll Cardiol 2012;60: Shin SD, Suh GJ, Ahn KO, Song KJ. Cardiopulmonary resuscitation outcome of out-of-hospital cardiac arrest in low-volume versus high-volume emergency

10 J.Y. Kim et al. / Resuscitation 84 (2013) departments: an observational study and propensity score matching analysis. Resuscitation 2011;82: Shin SD, Ahn KO, Song KJ, Park CB, Lee EJ. Out-of-hospital airway management and cardiac arrest outcomes: a propensity score matched analysis. Resuscitation 2012;83: Park CB, Shin SD, Suh GJ, et al. Pediatric out-of-hospital cardiac arrest in korea: a nationwide population-based study. Resuscitation 2010;81: Ahn KO, Shin SD, Suh GJ, et al. Epidemiology and outcomes from non-traumatic out-of-hospital cardiac arrest in korea: a nationwide observational study. Resuscitation 2010;81: 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: Bianchin A, Pellizzato N, Martano L, Castioni CA. Therapeutic hypothermia in italian intensive care units: a national survey. Minerva Anestesiol 2009;75: Laver SR, Padkin A, Atalla A, Nolan JP. Therapeutic hypothermia after cardiac arrest: a survey of practice in intensive care units in the united kingdom. Anaesthesia 2006;61: Garot P, Lefevre T, Eltchaninoff H, et al. Six-month outcome of emergency percutaneous coronary intervention in resuscitated patients after cardiac arrest complicating st-elevation myocardial infarction. Circulation 2007;115: Cronier P, Vignon P, Bouferrache K, et al. Impact of routine percutaneous coronary intervention after out-of-hospital cardiac arrest due to ventricular fibrillation. Crit Care 2011;15:R Grasner JT, Meybohm P, Caliebe A, et al. Postresuscitation care with mild therapeutic hypothermia and coronary intervention after out-of-hospital cardiopulmonary resuscitation: a prospective registry analysis. Crit Care 2011;15:R Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346: Group HaCAS. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346: Testori C, Sterz F, Behringer W, et al. Mild therapeutic hypothermia is associated with favourable outcome in patients after cardiac arrest with non-shockable rhythms. Resuscitation 2011;82: