Use of an Automated, Load-Distributing Band Chest Compression Device for Out-of-Hospital Cardiac Arrest Resuscitation JAMA. 2006;295:

Transcription

1 ORIGINAL CONTRIBUTION Use of an Automated, Load-Distributing Band Chest Compression Device for Out-of-Hospital Cardiac Arrest Resuscitation Marcus Eng Hock Ong, MD, MPH Joseph P. Ornato, MD David P. Edwards, MBA, EMT-P Harinder S. Dhindsa, MD, MPH Al M. Best, PhD Caesar S. Ines, MD, MS Scott Hickey, MD Bryan Clark, DO Dean C. Williams, MD Robert G. Powell, MD Jerry L. Overton, MPA Mary Ann Peberdy, MD APPROXIMATELY 400 TO individuals die every year from out-of-hospital cardiac arrest (OHCA), 1 representing approximately one third of all cardiovascular deaths 2 in the United States. Only 1% to 8% of individuals with OHCA survive to hospital discharge. 3-6 Patients who have ventricular fibrillation for less than 3 to 4 minutes (the electrical phase of cardiac arrest) 7 fare relatively well if rescuers arrive quickly and provide prompt defibrillation However, once ventricular fibrillation has been present longer, the myocardium becomes depleted of adenosine triphosphate and defibrillation usually results in conversion to asystole or a pulseless electrical rhythm. 7 Several studies suggest that a brief period of cardiopulmonary resuscitation See also pp 2620 and Context Only 1% to 8% of adults with out-of-hospital cardiac arrest survive to hospital discharge. Objective To compare resuscitation outcomes before and after an urban emergency medical services (EMS) system switched from manual cardiopulmonary resuscitation (CPR) to load-distributing band (LDB) CPR. Design, Setting, and Patients A phased, observational cohort evaluation with intention-to-treat analysis of 783 adults with out-of-hospital, nontraumatic cardiac arrest. A total of 499 patients were included in the manual CPR phase (January 1, 2001, to March 31, 2003) and 284 patients in the LDB-CPR phase (December 20, 2003, to March 31, 2005); of these patients, the LDB device was applied in 210 patients. Intervention Urban EMS system change from manual CPR to LDB-CPR. Main Outcome Measures Return of spontaneous circulation (ROSC), with secondary outcome measures of survival to hospital admission and hospital discharge, and neurological outcome at discharge. Results Patients in the manual CPR and LDB-CPR phases were comparable except for a faster response time interval (mean difference, 26 seconds) and more EMS-witnessed arrests (18.7% vs 12.6%) with LDB. Rates for ROSC and survival were increased with LDB-CPR compared with manual CPR (for ROSC, 34.5%; 95% confidence interval [CI], 29.2%-40.3% vs 20.2%; 95% CI, 16.9%-24.0%; adjusted odds ratio [OR], 1.94; 95% CI, ; for survival to hospital admission, 20.9%; 95% CI, 16.6%-26.1% vs 11.1%; 95% CI, 8.6%-14.2%; adjusted OR, 1.88; 95% CI, ; and for survival to hospital discharge, 9.7%; 95% CI, 6.7%-13.8% vs 2.9%; 95% CI, 1.7%-4.8%; adjusted OR, 2.27; 95% CI, ). In secondary analysis of the 210 patients in whom the LDB device was applied, 38 patients (18.1%) survived to hospital admission (95% CI, 13.4%-23.9%) and 12 patients (5.7%) survived to hospital discharge (95% CI, 3.0%- 9.3%). Among patients in the manual CPR and LDB-CPR groups who survived to hospital discharge, there was no significant difference between groups in Cerebral Performance Category (P=.36) or Overall Performance Category (P=.40). The number needed to treat for the adjusted outcome survival to discharge was 15 (95% CI, 9-33). Conclusion Compared with resuscitation using manual CPR, a resuscitation strategy using LDB-CPR on EMS ambulances is associated with improved survival to hospital discharge in adults with out-of-hospital nontraumatic cardiac arrest. JAMA. 2006;295: Author Affiliations: Departments of Epidemiology and Community Health (Dr Ong) and Biostatistics (Dr Best), Virginia Commonwealth University, and Department of Emergency Medicine (Drs Ong, Dhindsa, Ines, and Hickey), Department of Internal Medicine, Division of Cardiology (Dr Peberdy), and Department of Emergency Medicine and the Virginia Commonwealth University Reanimation, Engineering, and Shock Center (Drs Ornato and Peberdy), Virginia Commonwealth University Health System; The Richmond Ambulance Authority (Messrs Edwards and Overton); Department of Emergency Medicine, Chippenham & Johnston-Willis Hospital (Dr Clark); Department of Emergency Medicine, Richmond Community Hospital (Dr Williams); and Department of Emergency Medicine, St Mary s Hospital (Dr Powell), Richmond. Corresponding Author: Joseph P. Ornato, MD, Department of Emergency Medicine, Virginia Commonwealth University Medical Center, 1250 E Marshall St, 2nd Floor, Richmond, VA (ornato@aol.com) American Medical Association. All rights reserved. (Reprinted) JAMA, June 14, 2006 Vol 295, No

2 (CPR) before defibrillation can increase intracellular adenosine triphosphate levels and improve survival Attaining a coronary perfusion pressure of more than 15 mm Hg is one of the best predictors of return of spontaneous circulation (ROSC) in animals and humans. 22,23 Manual chest compression provides only approximately one third of the normal blood supply to the brain and 10% to 20% of the normal blood flow to the heart. 24 The use of a load-distributing band (LDB) device for chest compressions has been shown to achieve intrathoracic pressures higher than achievable safely during manual chest compression. The device improves coronary and systemic perfusion pressures and flows compared with those that can be achieved with manual CPR in animal models and in a small number of terminally ill patients. 15,25 In addition, in 1 study, 26 an LDB device was associated with improved ROSC compared with manual chest compression when used by paramedic fire captains in a large, urban emergency medical services (EMS) system. The goal of our study was to compare survival outcomes in patients with OHCA treated before and after the LDB device was used on urban EMS ambulances. METHODS EMS System Characteristics The Richmond Department of Fire and EMS provides first-response assistance on life or death emergency calls using a fire apparatus based at 20 fire stations. The trucks are staffed by emergency medical technicians who can perform manual CPR and defibrillate by using automated external defibrillators. The Richmond Ambulance Authority provides emergency advanced life support ambulance service for the city of Richmond, Virginia (population, ; service area, 62.5 square miles), and the ambulances are deployed using an Advanced System Status Management strategy. An average of 11 (range, 8-19) advanced life support ambulances is in service at any given time. The Richmond Ambulance Authority used a mechanic chest compression device (Michigan Instruments Thumper, Grand Rapids, Mich) as its standard CPR technique in the 1990s as a labor-saving device, but switched to manual CPR in 1998 when the aging compression devices became difficult to maintain. The Richmond Ambulance Authority used the LDB device (AutoPulse, ZOLL Circulation, Sunnyvale, Calif) as its new standard method for providing chest compression beginning in May 2003 (shortly after the US Food and Drug Administration [FDA] granted approval of the device as a labor saving device to free up paramedics to perform other tasks during resuscitation). Study Design The change in the standard method of performing chest compression during CPR provided an opportunity to conduct a phased, nonrandomized, observational study of the clinical outcomes of patients treated before and after the transition from manual CPR to LDB-CPR using the noninvasive device that received FDA clearance (K011046) on October 24, The LDB device generates artificial circulation by compressing the chest with an 8-inch-wide by 12.5-inchlong load-distributing band. The band is positioned directly over the sternum with the upper edge just below the patient s arms. The band is tightened, then relaxed, around the chest to provide a rhythmical squeezing effect that reduces the anteroposterior dimension of the chest by 20% measured at the sternum. The device adjusts automatically to the size and shape of each patient and is constructed around a backboard that contains a motorized rotating shaft under microprocessor control. The device delivers 80 chest compressions per minute. The compression depth is measured from the amount of belt spooled by the device. Cardiac arrests occurring between January 1, 2001, and March 31, 2003, received only manual CPR (manual CPR phase). Between April 1, 2003, to December 19, 2003, three LDB devices were used on a product evaluation basis on first-responder fire trucks. During this product evaluation period, LDB devices were applied on very few cases. It was decided that adding the devices to ambulances would provide better coverage. On December 20, 2003, 11 additional LDB devices were used on ambulances and a field supervisor unit to provide complete coverage for the city. Every attempt was made to apply LDB as early as possible during the resuscitation and use it as the standard method for delivering chest compressions. The LDB phase was between December 20, 2003, and March 31, Richmond uses a shock-first strategy only on first-responder or paramedic-witnessed cardiac arrest cases; in all other cases, first responders or paramedics are trained to provide 90 seconds of CPR before defibrillation (CPRfirst strategy). These practices were continued before and after introduction of the LDB device into the Richmond EMS system, with the initial 90 seconds of CPR being performed manually until the LDB device could be set up and applied to the patient. No significant OHCA treatment protocol changes were made between January 1, 2001, and March 31, 2005, and there were no other significant changes in the EMS system configuration or capabilities during this time (eg, no change in the use of automated external defibrillators by first responders). The number of public access defibrillation sites has been increasing slowly in Richmond for the last 10 years (now totaling 20 sites), but events from such sites account for less than 1% of all cardiac arrests. Study Population and End Points The study population included all adult patients with nontraumatic OHCA, defined as patients with absence of pulse, unresponsiveness, and apnea who received either CPR, defibrillation, or both, in Richmond during the study period. We excluded patients pronounced dead in the out-of-hospital set JAMA, June 14, 2006 Vol 295, No. 22 (Reprinted) 2006 American Medical Association. All rights reserved.

3 ting by paramedics without attempting resuscitation, OHCA caused by obvious major trauma, patients younger than 18 years, prisoners, mentally disabled, or pregnant women. The primary outcome measure was ROSC. Secondary outcomes included survival to hospital admission and hospital discharge, and neurological/ functional status at discharge. Variable and outcome definitions followed the Utstein recommendations Return of spontaneous circulation was defined as the presence of any palpable pulse, in the absence of chest compression, and detectable by manual palpation. Survival to admission was defined as patient admission to the hospital without ongoing CPR or other artificial circulatory support. Survival to hospital discharge was defined as a patient surviving the primary event and being discharged from the hospital alive. Neurological and functional status was assessed at discharge using standard Glasgow-Pittsburgh outcome measures. The Cerebral Performance Category score evaluates only cerebral performance capabilities, and the Overall Performance Category score reflects both cerebral and noncerebral status. 31 Institutional Review Board Review Data were collected from EMS patient care report forms, emergency department, and in-hospital patient records to determine clinical outcomes using a waiver of consent from each receiving hospital institutional review board. Other data sources included public accessible death certificate information. When necessary, we obtained consent and conducted an interview of the patient or their proxy to determine their current neurological (functional) outcome status. The study was reviewed and approved by the Virginia Commonwealth University Health System Committee for the Protection of Human Subjects as well as the institutional review boards of community hospitals receiving patients with OHCA in Richmond (Chippenham & Johnston Willis Hospitals, Bon Secours Hospitals, which included St Mary s, Richmond Community, and Richmond Memorial Hospitals). The institutional review boards waived the requirement for patient/ family consent and authorized a coinvestigator on the staff of each hospital to do a retrospective chart review to determine survival and the Cerebral Performance Category and Overall Performance Category at discharge. The institutional review boards preferred to waive the consent requirement to look at the chart to eliminate calling families of patients who died before hospital discharge. There was no significant difference in the method used or the completeness of data collection between phases of the study. Survival outcomes to hospital discharge were obtained on 97.6% of patients (manual CPR phase, 97.4%; and LDB-CPR phase, 97.8%; P=.81 by Fisher test). Quality Assurance Measures Our study included the following elements for quality assurance: development of standard protocols to perform all data collection and follow-up activities, use of standardized forms, uniform criteria for patient identification, standardized data processing, editing of incoming data, regular communications between the study investigators to resolve questions, and internal monitoring of data collection. Additional steps to ensure data quality included range checks and verification built into the data entry system, and a sequence of logic checking and examination of variables. Because this was a retrospective analysis, the only reliable time intervals available from the EMS system were the first-response fire units and EMS ambulance response time intervals (from call receipt at the 911 emergency medical dispatch center until arrival of the respective units at the call s location). For the last 15 calls during the LDB-CPR phase, paramedics were asked to call into medical dispatch on their portable radios when they arrived at the patient s side and when the LDB device was attached and started. The median time interval from EMS ambulance arrival at location to crew at patient side was 3.0 minutes and the median time interval from crew at patient side to LDB device attached and started was 3.6 minutes. Sample Size Before our study, the Richmond Ambulance Authority quality improvement program tracked only ROSC data (but not survival to hospital discharge data). Six months of ROSC data before (21% ROSC) and after (37% ROSC) the product evaluation period was used to estimate the sample size. To detect a 16% improvement in ROSC between LDB-CPR and manual CPR (37% vs 21%), with a 2-sided test size of 5% and a power of 90% would require 180 patients with cardiac arrest in the LDB-CPR group. To overcome loss to follow-up and nonparticipation, we sampled 150% of the required number of cardiac arrests for the intervention group. Hence, we estimated that 720 patient charts would be required to be reviewed, at an allocation ratio of 1:2 (270 in the LDB phase and 540 in the manual phase) to allow for loss to follow-up. Statistical Analysis Analysis was conducted by using JMP version 5.1 (SAS Institute Inc, Cary, NC). Frequency tables and descriptive statistics for all covariates were calculated. Univariate comparisons using t tests, 2 tests, or Fisher tests were conducted to identify differences in distribution of covariates between phases. Those comparisons with P.20 were included for consideration in the final logistic regression models. As prespecified, all statistical analyses were performed on an intention-to-treat basis, including all patients in the manual CPR phase with cardiac arrest and manual CPR, and all patients in the LDB-CPR phase with cardiac arrest, whether the LDB device was used or not applied. Associations between treatment groups and all end points were analyzed by using the 2 test with odds ratios (ORs) presented where applicable. Logistic regression was used to 2006 American Medical Association. All rights reserved. (Reprinted) JAMA, June 14, 2006 Vol 295, No

4 adjust for relevant covariates and adjusted ORs and 95% confidence intervals (CIs) are given for all end points. In the secondary analysis, the outcomes for patients in the LDB-CPR phase who actually received the device were estimated. RESULTS The Richmond EMS system responded to 2766 persons having cardiac arrest aged 18 years or older between January 1, 2001, and March 31, 2005 (FIGURE 1), which included 1475 persons in the manual CPR phase, 819 in the LDB-CPR phase, and 472 in the product evaluation phase. Of the 1475 persons, resuscitation was not attempted in 818 and 158 were found to Figure 1. Modified Utstein Reporting Template for Data Elements* 818 Resuscitation Not Attempted Manual CPR Phase 1475 Individuals With Absence of Signs of Circulation and/or Considered for Resuscitation (Age 18 y) 499 Patients Included in Manual CPR Phase have noncardiac etiology, yielding 499 cases that met study inclusion criteria during the manual CPR phase. Between April 1, 2003, and December 18, 2003 (the product evaluation period), the LDB device was applied to 160 patients with cardiac arrest (15.9% of cases). All cases (n=472) occurring during this product evaluation period were excluded from analysis. Between December 20, 2003, and March 31, 2005, of the 819 individuals with cardiac arrest, resuscitation was not attempted in 438 and 97 had a noncardiac etiology, yielding 284 cases eligible for and included in the analysis for the LDB- CPR phase regardless of whether the LDB device was applied (intention-totreat analysis). LDB-CPR Phase 819 Individuals With Absence of Signs of Circulation and/or Considered for Resuscitation (Age 18 y) 657 Resuscitation Attempted 381 Resuscitation Attempted 158 Noncardiac Etiology 97 Noncardiac Etiology 499 Presumed Cardiac Etiology 284 Presumed Cardiac Etiology 284 Patients Included in LDB- CPR Phase 210 Patients Had LDB Device Applied 74 Patients Did Not Have LDB Device Applied 50 LDB Device Not Indicated 22 Cease Resuscitation Request or Order 20 ROSC Established Shortly After Initial Defibrillation 8 Cardiac Arrest en Route to Hospital 14 LDB Device Not Available 4 Mechanical Failure of LDB Device 4 Inability to Fit LDB Device 2 Reason Missing 438 Resuscitation Not Attempted CPR indicates cardiopulmonary resuscitation; LDB, load-distributing band; ROSC, return of spontaneous circulation. *All cases (n=472) that occurred during the product evaluation period (April 1, 2003, to December 18, 2003) were excluded from the analysis. Pronounced dead on scene or do not resuscitate status. TABLE 1 shows the characteristics of patients treated during the manual CPR and LDB-CPR phases. There were no significant differences in age, sex, arrest location, bystander CPR, initial rhythm, whether the patients were defibrillated, and whether the cardiac arrest was bystander witnessed. There was a slightly faster ambulance response time interval (mean difference, 26 seconds) during the LDB-CPR phase (P =.03). There were fewer EMSwitnessed arrests during the manual CPR phase vs the LDB-CPR phase (12.6% vs 18.7%, P=.03). To adjust for these group differences, EMSwitnessed and ambulance response time were incorporated into the logistic regression models described below. Beginning in February 2004, one receiving hospital (Virginia Commonwealth University Health System) instituted a postresuscitation, mild hypothermia protocol for patients with OHCA who remained comatose after resuscitation. Ten patients received hypothermia during the LDB-CPR phase. We controlled for the effect of hypothermia treatment postresuscitation in the logistic regression model for survival to hospital discharge. The LDB device was applied to 210 (73.9%) of patients during the LDB- CPR phase. Reasons for nonapplication included not indicated when it arrived at the patient s side (69%), machine not available (19%), mechanical failure during attempted operation (6%), or inability to fit the device on an oversized or undersized patient (6%). Reasons the paramedics believed that the device application was not indicated were family requested cessation of resuscitation or presence of a valid do not resuscitate order (44%), establishment of ROSC shortly after initial defibrillation with or without a brief period of antecedent manual CPR (40%), or cardiac arrest occurring en route to a hospital with insufficient time for paramedics to apply the device before emergency department arrival (16%). TABLE 2 compares clinical outcomes in the manual CPR vs LDB-CPR phases using intention-to-treat analysis. The 2632 JAMA, June 14, 2006 Vol 295, No. 22 (Reprinted) 2006 American Medical Association. All rights reserved.

5 rates of ROSC and survival to hospital admission were all significantly higher in the LDB-CPR phase, after adjustment for differences in EMS response time intervals and EMS witnessed. The rate of survival to hospital discharge was significantly higher in the LDB-CPR phase, after adjustment for differences in EMS response time intervals, EMS witnessed, and postresuscitation hypothermia. The OR for survival with postresuscitation hypothermia compared with no hypothermia was 5.38 (95% CI, ). Of the 58 patients who survived to hospital admission, 38 had the device applied and 20 did not. Similarly, 12 of the 27 patients who survived to hospital discharge during the LDB-CPR phase had the device applied and 15 did not. Thus, secondary analysis showed that among 210 patients in the LDB-CPR phase who actually had the device applied, survival to hospital admission was 18.1% (95% CI, 13.4%-23.9%) and survival to hospital discharge was 5.7% (95% CI, 3.0%- 9.3%). TABLE 3 shows the Cerebral Performance Categories and Overall Performance Categories for patients in the manual CPR vs LDB-CPR phases. Overall, neurological status of survivors was not significantly different between the 2 phases. Based on the adjusted rates of survival to hospital discharge (Table 2), the absolute risk reduction was 6.8% or (95% CI, ), giving a number needed to treat of 15 (95% CI, 9-33). The relationship between ambulance response time interval and survival to hospital discharge by treatment phases only for patients whose cardiac arrest was not witnessed by firstresponse fire units or paramedics was also estimated. First-responder and EMS-witnessed cardiac arrests were excluded from the calculation, because they would essentially have a response time interval of zero. All of the benefits of LDB-CPR vs manual CPR phases occurred when paramedic ambulances arrived on location in less than 8 minutes from the time of 911 call receipt. For patients with ambulance response time intervals of less than 8 minutes, survival to hospital discharge was observed in 6 of 323 patients (1.9%; 95% CI, 0.9%-4.0%) for manual CPR phase and 15 of 185 patients (8.1%; 95% CI, 5.0%-13.0%) for LDB-CPR phase; whereas for ambulance response time intervals of 8 minutes or more, survival to hospital discharge was observed in 3 of 103 patients (2.9%; 95% CI, 1.0%-8.2%) for manual CPR phase and 1 of 37 patients (2.7%; 95% CI, 0.5%-13.8%) for LDB-CPR phase. Table 1. Patient Characteristics in the Manual CPR vs LDB-CPR Phases* Characteristics Manual CPR (n = 499) LDB-CPR (n = 284) P Value Age, mean (SD) 68.1 (15.6) 67.3 (16.2).49 Male 269 (53.9) 161 (56.7).45 Arrest location Residence 407 (81.6) 215 (78.2) Other 92 (18.4) 60 (21.8).26 Bystander witnessed 172 (34.5) 94 (33.5).77 EMS witnessed 63 (12.6) 53 (18.7).03 Bystander CPR 158 (31.7) 85 (30.5).73 Initial rhythm Ventricular fibrillation 99 (20.4) 63 (23.3) Ventricular tachycardia 3 (0.62) 2 (0.74) Asystole 265 (54.5) 140 (49.3).80 Pulseless electrical activity 119 (24.5) 65 (24.1) First-response fire unit time interval, 274 (93) 279 (90).70 mean (SD), s First-responder fire unit CPR performed 245 (49.1) 138 (49.8).85 before EMS arrival Defibrillated by firefighter AED 51 (10.2) 23 (8.27).38 EMS ambulance response time interval, 6.5 (2.8) 6.1 (2.4).03 mean (SD), min Hypothermia-induced postresuscitation 0 (0) 10 (3.50).001 LDB device applied 0 (0) 210 (73.9) NA Abbreviations: AED, automated external defibrillator; CPR, cardiopulmonary resuscitation; EMS, emergency medical services; LDB, load-distributing band; NA, not applicable. *Data are presented as number (percentage) unless otherwise specified. The EMS ambulance response time interval is from call receipt at the 911 emergency medical dispatch center until arrival of the respective units at the call s location. Some of the characteristics do not sum to the total number of patients in each phase due to missing data. t Test was used to compare mean values and 2 or Fisher test was used to compare percentages. Table 2. Comparison of Outcomes in the Manual CPR and LDB-CPR Phases* Manual CPR LDB-CPR OR (95% CI) of Patients % (95% CI) of Patients % (95% CI) Unadjusted Adjusted Return of spontaneous circulation 101/ ( ) 96/ ( ) 2.08 ( ) 1.94 ( ) Survival to hospital admission 54/ ( ) 58/ ( ) 2.11 ( ) 1.88 ( ) Survival to hospital discharge 14/ ( ) 27/ ( ) 3.23 ( ) 2.27 ( ) Abbreviations: CI, confidence interval; CPR, cardiopulmonary resuscitation; EMS, emergency medical services; LDB, load-distributing band; OR, odds ratio. *Both crude and adjusted ORs are presented in the logistic regression models. For the LDB-CPR phase, the total number of patients is not 284 due to missing data. Adjusted for differences in response time intervals and percentage of EMS witnessed. Adjusted for differences in response time intervals, percentage of EMS witnessed, and whether postresuscitation hypothermia was used. For the unadjusted and adjusted ORs and 95% CIs, a weighted logistic regression was performed American Medical Association. All rights reserved. (Reprinted) JAMA, June 14, 2006 Vol 295, No

6 TABLE 4 shows the outcomes grouped by clinically relevant subsets. Although several of the cells are too small to show statistically significant differences between phases, many of the larger subsets demonstrate improved LDB survival outcomes. FIGURE 2 displays the ROSC and survival to hospital discharge rates in 3-month intervals. Athough there is considerable overlap of the 95% CIs due to the small cell sizes, the figure provides a visual sense of the effect of duration of LDB use on the outcome measures. Table 3. Cerebral Performance Category and Overall Performance Category of Patients in the Manual CPR vs LDB-CPR Phases* Performance Categories Manual CPR (n = 101) No. of Patients (%) LDB-CPR (n = 96) P Value Cerebral Performance Category 1 5 (5.6) 13 (15.1) 2 3 (3.4) 3 (3.5) 3 2 (2.3) 2 (2.3) (3.4) 3 (3.5) 5 76 (85.4) 65 (75.6) Overall Performance Category 1 2 (2.3) 4 (4.7) 2 4 (4.5) 10 (11.6) 3 4 (4.5) 4 (4.7) (3.4) 3 (3.5) 5 76 (85.4) 65 (75.6) Abbreviations: CPR, cardiopulmonary resuscitation; LDB, load-distributing band. *Percentages based on patients with return of spontaneous circulation; numbers in performance categories do not sum to total number of patients in each phase due to missing data. Cerebral Performance Category scores describe good (1-2) and poor (3-5) outcomes. A score of 1 indicates conscious and alert with normal function or only slight disability; 2, conscious and alert with moderate disability; 3, conscious with severe disability; 4, comatose or persistent vegetative state; and 5, brain dead or death from other causes. An Overall Performance Category score of 1 indicates good overall performance: healthy, alert, and capable of leading a normal life; 2, moderate overall disability: conscious, performs independent activities of daily life or able to work part-time in sheltered environment; 3, severe overall disability: conscious, dependent on others for daily support; 4, coma/vegetative state: not conscious, unaware of surroundings, no cognition; and 5, brain death: certified brain dead or dead by traditional criteria. 2 Test was used to compare percentages. COMMENT In our study, OHCA clinical outcomes were improved following the introduction of LDB into an urban EMS system. The benefit was relatively robust across a range of patient subsets, especially for those patients with ventricular fibrillation initially, bystander witnessed events, and recipients of bystander CPR. A variety of techniques can produce circumferential chest compression during resuscitation. The first, an airfilled pneumatic vest, improved blood flow and survival in animals and improved coronary perfusion pressure in humans compared with manual chest compression. 15,32 It was energy inefficient and could not function as a practical, portable resuscitation device. The LDB device, cleared through the 510(k) FDA regulatory pathway as a laborsaving chest compression tool during resuscitation, has an energy efficient electric motor that compresses the chest with a microprocessor-controlled LDB. The LDB device generates systemic pressures and flows that are better than manual chest compression in animal models 25 and humans. 33 It produced a mean coronary perfusion pressure of 21 mm Hg compared with 14 mm Hg with Table 4. Outcomes by Clinically Relevant Subsets ROSC Hospital Admission Hospital Discharge of Patients (%) of Patients (%) of Patients (%) Subset Initial rhythm Ventricular fibrillation Ventricular tachycardia Pulseless electrical activity Manual CPR LDB-CPR OR (95% CI) Manual CPR LDB-CPR OR (95% CI) Manual CPR LDB-CPR OR (95% CI) 27/99 (27.3) 34/61 (55.7) 3.36 ( ) 16/93 (17.2) 19/59 (32.2) 2.29 ( ) 5/93 (5.4) 12/60 (20.0) 4.40 ( ) 1/3 (33.3) 2/2 (100) ND 1/3 (33.3) 2/2 (100) ND 0/3 (0) 2/2 (100) ND 47/119 (39.5) 28/64 (43.8) 1.19 ( ) 26/114 (22.8) 16/64 (25.0) 1.13 ( ) 7/115 (6.1) 4/64 (6.3) 1.03 ( ) Asystole 26/265 (9.8) 27/140 (19.3) 2.20 ( ) 11/262 (4.2) 16/138 (11.6) 2.99 ( ) 2/262 (0.8) 4/138 (2.9) 3.88 ( ) Witness to collapse Layperson 49/172 (28.5) 42/90 (46.7) 2.20 ( ) 26/166 (15.7) 29/92 (31.5) 2.48 ( ) 6/166 (3.6) 15/92 (16.3) 5.19 ( ) Emergency medical services 25/63 (39.7) 27/52 (51.9) 1.64 ( ) 15/59 (25.4) 19/52 (36.5) 1.69 ( ) 5/60 (8.3) 8/52 (15.4) 2.00 ( ) None 27/264 (10.2) 27/135 (20.0) 2.19 ( ) 13/260 (5.0) 9/132 (6.8) 1.39 ( ) 3/260 (1.2) 3/133 (2.3) 1.98 ( ) Bystander CPR Yes 26/158 (16.5) 21/84 (25.0) 1.69 ( ) 8/151 (5.3) 12/84 (14.3) 2.98 ( ) 1/151 (0.7) 8/84 (9.5) ( ) No 75/341 (22.0) 72/191 (37.7) 2.15 ( ) 46/334 (13.8) 43/188 (22.9) 1.86 ( ) 13/335 (3.9) 16/189 (8.5) 2.29 ( ) Abbreviations: CI, confidence interval; CPR, cardiopulmonary resuscitation; LDB, load-distributing band; ND, not determined; OR, odds ratio; ROSC, return of spontaneous circulation JAMA, June 14, 2006 Vol 295, No. 22 (Reprinted) 2006 American Medical Association. All rights reserved.

7 manual chest compression in a porcine animal model. 25 The LDB-CPR compressions produced 36% of normal coronary blood flow compared with 13% by manual CPR without pharmacological vasopressor support. With epinephrine, the LDB device generated normal heart and brain flow levels. In a pilot clinical study, 16 sequential patients who were terminally ill received alternating periods of manual CPR vs LDB-CPR for 90 seconds each after 10 minutes of failed standard advanced life support. 33 The LDB-CPR compression produced a significantly higher coronary perfusion pressure (mean [SD], 20 [12] vs 15 [11] mm Hg; P=.02) and peak aortic pressure (mean [SD], 153 [28] vs 115 [42] mm Hg; P.001) compared with values achieved with manual chest compression. Before our analysis, only 1 relatively small, published clinical study compared patient outcomes in those patients treated with the LDB device vs manual chest compression. The study by Casner et al 26 showed how the San Francisco fire department used the device on a few paramedic captain vehicles and responded to adults with cardiac arrest in their system relatively late (mean [SD] response time interval, 15 [5] minutes). Sixty-nine LDB-CPR cases were matched to 93 manual CPR cases. Patients treated with LDB-CPR experienced a higher rate of ROSC than patients treated with manual CPR (39% vs 29%, P=.003). The magnitude of the ROSC with LDB-CPR vs manual CPR was similar to that noted in our study. The majority of the survival to hospital discharge benefit observed in our Figure 2. Return of Spontaneous Circulation (ROSC) and Survival to Hospital Discharge for Manual CPR and LDB-CPR Phases Stratified by 3-Month Intervals A ROSC Manual CPR Phase LDB-CPR Phase ROSC, % ROSC, % Overall Overall of Patients 10/47 11/47 11/42 17/64 7/57 14/66 5/43 11/59 15/74 101/499 No./Total of Patients 2/11 29/62 18/35 14/47 13/52 20/71 96/278 B Survival to Hospital Discharge Manual CPR Phase LDB-CPR Phase Survival, % Survival, % Overall Overall of Patients 0/39 2/46 1/41 4/63 1/57 1/65 2/43 2/58 1/74 14/486 of Patients 1/11 6/62 6/34 3/47 5/56 6/68 27/278 CPR indicates cardiopulmonary resuscitation; LDB, load-distributing band. Error bars indicate 95% confidence intervals. For year 2001 data in the survival to hospital discharge manual CPR phase, the 95% confidence interval around zero survivors is shown American Medical Association. All rights reserved. (Reprinted) JAMA, June 14, 2006 Vol 295, No

8 LDB-CPR cases was due to improved outcomes among patients with initial ventricular fibrillation. Although patients with initial asystole had a higher rate of ROSC and a trend toward better survival to hospital discharge with LDB-CPR compared with manual CPR, the latter difference was not statistically significant, possibly due to the relatively small sample size in these subsets. No survival difference was noted in those patients with pulseless electrical activity initially, perhaps reflecting the wide range of underlying pathological triggers for cardiac arrest that may not be countered effectively by improved blood flow in such patients. Although rates of bystander CPR were similar between groups, the combined effect of LDB plus bystander CPR conferred a 16-fold increase odds of survival to hospital discharge. This observation, if confirmed in future studies, suggests that there may be an opportunity for improving survival by using a community strategy of aggressive bystander CPR training and prearrival telephone CPR instruction followed promptly by an optimized form of CPR. The main limitation of our study is that it was not a randomized controlled trial. Despite this, we believe that the results are valid because patients in the 2 phases were comparable in all respects except for a slightly faster EMS response time interval and more frequent EMS-witnessed events during the LDB- CPR phase. The improved outcomes observed during the LDB-CPR phase persisted even after adjusting for these differences using logistic regression. Similarly, although one of the receiving hospitals instituted a postresuscitation, mild hypothermia protocol during the LDB-CPR phase, it cannot explain the disparity in outcomes between the 2 groups. Although we were able to show the survival benefit of hypothermia in these patients, they represented less than 4% of patients in the LDB-CPR phase. This is not a large enough number of cases to sway the results, corroborated by our observation that logistic regression modeling continued to show improved survival to hospital discharge during the LDB- CPR phase vs the manual CPR phase, after adjusting for the effect of postresuscitation hypothermia. In addition, hypothermia induced after hospital admission would not have had any effect on the primary outcome of ROSC or the secondary outcome of survival to hospital admission. However, the final logistic regression models have relatively modest R 2 values, suggesting that these models may not fully account for the true sources of variation in the reported outcomes. The LDB device was used on all EMS ambulances in our system, on our supervisor vehicle, and several firstresponder fire units in a rapidly responding EMS system. Although we could not quantitate how long it took our paramedics to use the device (other than a crude estimate from a small sample of paramedic actions at the end of the LDB-CPR phase), every attempt was made to attach and start it as quickly as possible after arrival at the patient s side. All of the benefits of LDB-CPR vs manual CPR occurred when paramedic ambulances arrived on location in less than 8 minutes from the time of 911 call receipt. Our rapid LDB device protocol could, at least in part, explain the clinical benefit persisting to hospital discharge noted in our EMS system compared with a previous study in which a small number of devices were placed on paramedic fire captain vehicles that arrived relatively late in the resuscitation. 26 These hypothesisgenerating observations suggest that the protocol and timing of LDB application may be critical. Several recent studies have shown that both in-hospital and out-ofhospital CPR quality can be poor at times with long pauses in compressions. This raises the theoretical possibility that, if paramedics during the manual CPR phase were not performing the highest quality CPR (suggested by the relatively low survival rate), shifting to the use of the LDB device could be improving survival by simply improving the quality of CPR. If true, our observations could be confirmation that LDB-CPR is better than suboptimal manual CPR, but it may not be better than high-quality manual CPR. Richmond, like most EMS systems, did not have quality of CPR monitoring technology in place at the time these data were collected. Future studies should include such instrumentation, now that it is becoming available. Finally, despite our best efforts, we were unable to obtain outcomes for a small proportion ( 3%) of patients. This was due to unknown identity of patients with cardiac arrest at the time of arrest and subsequent difficulties in matching identities with hospital records. Records were most complete for the EMS phase of the study and the outcome ROSC. Thus, the amount of missing data are actually higher among potential survivors, at least to admission, suggesting that our results represent a slight underestimation of survival rates. CONCLUSION Our results suggest that a resuscitation strategy using the LDB-CPR on rapidly responding EMS ambulances is associated with improved outcomes, including survival to hospital discharge, in adults with OHCA. These results suggest that the LDB device may be a useful addition to current cardiac arrest treatment options, especially when used early for patients with cardiac arrest who do not respond immediately to a brief period of manual CPR, defibrillation, or both. However, further research (a large, adequately powered, prospective randomized clinical trial that blinds the rescuers to the intervention until they decide to initiate resuscitation) is needed to further define the value of LDB in resuscitation. Author Contributions: Dr Ong had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ong, Ornato, Dhindsa, Peberdy. Acquisition of data: Ong, Edwards, Ines, Hickey, Clark, Williams, Powell, Overton. Analysis and interpretation of data: Ong, Best. Drafting of the manuscript: Ong, Ornato, Best. Critical revision of the manuscript for important intellectual content: Ong, Edwards, Dhindsa, Best, Ines, Hickey, Clark, Williams, Powell, Overton, Peberdy JAMA, June 14, 2006 Vol 295, No. 22 (Reprinted) 2006 American Medical Association. All rights reserved.

9 Statistical analysis: Ong, Best. Administrative, technical, or material support: Ong, Edwards, Dhindsa, Ines, Hickey, Clark, Williams, Powell, Overton. Study supervision: Ornato, Peberdy. Financial Disclosures: Dr Ornato is a Science Advisor to ZOLL Circulation (Sunnyvale, Calif ), the manufacturer of the Autopulse device used in the study. Dr Ornato reported receiving reimbursement for travel expenses to Science Advisory board meetings approximately twice yearly and a small honorarium amounting to less than $2000 per year. He reported no other financial benefits (stock, stock options) from this relationship. Because of this relationship, Dr Ornato did not have access to data acquisition, entry, or analysis during this study. No other authors reported financial disclosures. Funding/Support: This study was not a sponsored project. The 3 LDB devices used during the evaluation phase were provided free by ZOLL Circulation. Eight additional devices were loaned to the EMS system by the manufacturer for 12 months when ambulance deployment occurred to provide feedback to the manufacturer on the device s design/durability on ambulances. The EMS system subsequently purchased all of the devices along with 7 additional units. Acknowledgment: We are grateful for the voluntary contributions of Thomas Franck, MD, MPH, Department of Epidemiology and Community Health, Virginia Commonwealth University; Andrew J. Anderson, Department of Emergency Medicine, Richmond Community Hospital; Patti Aldridge, RN, Department of Emergency Medicine, Retreat Hospital; and Lorie Liptak, Chris Schaeffer, Richard Pertgen, and Derek Andresen, all from the Richmond Ambulance Authority. REFERENCES 1. Zheng ZJ, Croft JB, Giles WH, Mensah GA. Sudden cardiac death in the United States, 1989 to Circulation. 2001;104: Becker LB, Smith DW, Rhodes KV. Incidence of cardiac arrest: a neglected factor in evaluating survival rates. Ann Emerg Med. 1993;22: Eisenberg MS, Horwood BT, Cummins RO, Reynolds-Haertle R, Hearne TR. Cardiac arrest and resuscitation: a tale of 29 cities. Ann Emerg Med. 1990;19: Ornato JP, McBurnie MA, Nichol G, et al. The Public Access Defibrillation (PAD) trial: study design and rationale. Resuscitation. 2003;56: Lombardi G, Gallagher J, Gennis P. Outcome of outof-hospital cardiac arrest in New York City: the Pre- Hospital Arrest Survival Evaluation (PHASE) study. JAMA. 1994;271: Becker LB, Ostrander MP, Barrett J, Kondos GT. Outcome of CPR in a large metropolitan area: where are the survivors? Ann Emerg Med. 1991;20: Weisfeldt ML, Becker LB. Resuscitation after cardiac arrest: a 3-phase time-sensitive model. JAMA. 2002;288: White RD, Hankins DG, Bugliosi TF. Seven years experience with early defibrillation by police and paramedics in an emergency medical services system. Resuscitation. 1998;39: Valenzuela TD, Roe DJ, Nichol G, Clark LL, Spaite DW, Hardman RG. Outcomes of rapid defibrillation by security officers after cardiac arrest in casinos. N Engl J Med. 2000;343: Caffrey SL, Willoughby PJ, Pepe PE, Becker LB. Public use of automated external defibrillators. N Engl J Med. 2002;347: Hallstrom AP, Ornato JP, Weisfeldt M, et al. Publicaccess defibrillation and survival after out-ofhospital cardiac arrest. N Engl J Med. 2004;351: Cobb LA, Fahrenbruch CE, Walsh TR, et al. Influence of cardiopulmonary resuscitation prior to defibrillation in patients with out-of-hospital ventricular fibrillation. JAMA. 1999;281: Eftestol T, Wik L, Sunde K, Steen PA. Effects of cardiopulmonary resuscitation on predictors of ventricular fibrillation defibrillation success during out-of-hospital cardiac arrest. Circulation. 2004;110: Wik L, Hansen TB, Fylling F, et al. Delaying defibrillation to give basic cardiopulmonary resuscitation to patients with out-of-hospital ventricular fibrillation: a randomized trial. JAMA. 2003;289: Halperin HR, Guerci AD, Chandra N, et al. Vest inflation without simultaneous ventilation during cardiac arrest in dogs: improved survival from prolonged cardiopulmonary resuscitation. Circulation. 1986;74: Ralston SH, Voorhees WD, Babbs CF. Intrapulmonary epinephrine during prolonged cardiopulmonary resuscitation: improved regional blood flow and resuscitation in dogs. Ann Emerg Med. 1984;13: Michael JR, Guerci AD, Koehler RC, et al. Mechanisms by which epinephrine augments cerebral and myocardial perfusion during cardiopulmonary resuscitation in dogs. Circulation. 1984;69: Kern KB, Ewy GA, Voorhees WD, Babbs CF, Tacker WA. Myocardial perfusion pressure: a predictor of 24- hour survival during prolonged cardiac arrest in dogs. Resuscitation. 1988;16: Sanders AB, Ewy GA, Taft TV. Prognostic and therapeutic importance of the aortic diastolic pressure in resuscitation from cardiac arrest. Crit Care Med. 1984;12: Sanders AB, Ogle M, Ewy GA. Coronary perfusion pressure during cardiopulmonary resuscitation. Am J Emerg Med. 1985;3: Wolfe JA, Maier GW, Newton JR Jr, et al. Physiologic determinants of coronary blood flow during external cardiac massage. J Thorac Cardiovasc Surg. 1988; 95: Paradis NA, Martin GB, Rivers EP, et al. Coronary perfusion pressure and the return of spontaneous circulation in human cardiopulmonary resuscitation. JAMA. 1990;263: McDonald JL. Coronary perfusion pressure during CPR in human beings [abstract]. Ann Emerg Med. 1983;12: Kern KB. Coronary perfusion pressure during cardio-pulmonary resuscitation. Baillieres Clin Anaesthesiol. 2000;14: Halperin H, Paradis N, Ornato J. Improved hemodynamics with a novel chest compression device during a porcine model of cardiac arrest [abstract]. Circulation. 2002;106(19 suppl 2): Casner M, Andersen D, Isaacs SM. The impact of a new CPR assist device on rate of return of spontaneous circulation in out-of-hospital cardiac arrest. Prehosp Emerg Care. 2005;9: 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. Resuscitation. 2004;63: 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. Circulation. 2004;110: Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein Style. Circulation. 1991;84: Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein Style. Ann Emerg Med. 1991;20: The Brain Resuscitation Clinical Trial II Study Group. A randomized clinical trial of calcium entry blocker administration to comatose survivors of cardiac arrest: design, methods, and patient characteristics. Control Clin Trials. 1991;12: Halperin HR, Tsitlik JE, Gelfand M, et al. A preliminary study of cardiopulmonary resuscitation by circumferential compression of the chest with use of a pneumatic vest. N Engl J Med. 1993;329: Timerman S, Cardoso LF, Ramires JA, Halperin H. Improved hemodynamic performance with a novel chest compression device during treatment of in-hospital cardiac arrest. Resuscitation. 2004;61: Wik L, Kramer-Johansen J, Myklebust H, et al. Quality of cardiopulmonary resuscitation during outof-hospital cardiac arrest. JAMA. 2005;293: Abella BS, Alvarado JP, Myklebust H, et al. Quality of cardiopulmonary resuscitation during inhospital cardiac arrest. JAMA. 2005;293: Abella BS, Sandbo N, Vassilatos P, et al. Chest compression rates during cardiopulmonary resuscitation are suboptimal: a prospective study during in-hospital cardiac arrest. Circulation. 2005;111: American Medical Association. All rights reserved. (Reprinted) JAMA, June 14, 2006 Vol 295, No