Abstract
Background:
Cardiopulmonary resuscitation (CPR) by healthcare responders is often suboptimal during out-of-hospital cardiac arrest (OHCA). Mechanical CPR devices have been promoted as a strategy to improve CPR quality. Whether their use in routine care is associated with improved OHCA survival at emergency medical service (EMS) agencies is unclear.
Methods:
Within the Cardiac Arrest Registry to Enhance Survival, we assessed agency-level rates of OHCA survival at EMS agencies with ≥2 years of OHCA data before and after their first documented use of a mechanical CPR device during 2013–2019 using an observational cohort study design. Temporal trends in favorable neurological survival (without severe neurological disability) and survival to discharge were first assessed in EMS agencies that did not initiate mechanical CPR device use (control agencies) using multivariable hierarchical logistic regression. Then, an interrupted time series analysis evaluated whether introduction of mechanical CPR devices at EMS agencies was associated with higher OHCA survival.
Results:
Among 51,994 patients with OHCA at 73 control agencies, there were no temporal trends in risk-adjusted rates of favorable neurological survival (annual rates ranged between 9.6–10.6%; Ptrend=0.23) or survival to discharge (annual rates ranged between 11.1–12.0%; Ptrend=0.31). At 49 EMS agencies that introduced mechanical CPR devices (14,110 OHCAs before and 17,804 OHCAs after), the mean EMS agency risk-adjusted rate of favorable neurological survival was 8.9%±2.2% before mechanical CPR device introduction and 8.3%±1.3% after, with no change in model intercept (adjusted OR, 0.94 [95% CI: 0.80, 1.11]; P=0.48) or slope after introduction of mechanical CPR devices (adjusted OR per year, 1.03 [95% CI: 0.96, 1.12]; P=0.41). Similarly, the mean EMS agency risk-adjusted rate of survival to discharge was 11.0%±2.2% before and 10.0%±1.0% after device introduction, with no significant change in model intercept or slope after device introduction.
Conclusions:
In a large U.S. registry of OHCA, EMS agency rates of favorable neurological survival and survival to discharge were not higher after the initiation of mechanical CPR device use.
Keywords: cardiac arrest, CPR, survival
INTRODUCTION
Although high-quality cardiopulmonary resuscitation (CPR) is a critical link in the chain of survival for out-of-hospital cardiac arrest (OHCA),1 CPR is often suboptimal in routine practice.2, 3 In recent years, mechanical CPR devices have been promoted as a way to mitigate variability in CPR quality. CPR is physically demanding and difficult to sustain for long periods; in contrast, mechanical CPR devices may offer advantages by providing consistent chest compressions with a standardized rate and depth with minimal interruptions and free up emergency medical responders to focus on identifying and treating the etiology of the OHCA, airway management, and defibrillation. In real-world practice, mechanical CPR device use has grown, increasing from 1.9% in 2010 to 8.0% in 2016 of all OHCAs in one emergency medical service (EMS) registry.4 and to 44.7% of EMS-treated OHCAs in 2022 within the Cardiac Arrest Registry to Enhance Survival (CARES) (personal communication with Dr. Bryan McNally, Executive Director of CARES).
Despite this growth in use, the evidence supporting mechanical CPR devices for cardiac arrest is mixed. In a 2018 Cochrane review of 7 randomized and quasi-randomized trials of mechanical CPR devices, only two studies (one for OHCA, one for in-hospital cardiac arrest) found mechanical CPR devices increased rates of survival to discharge, and no study found device use increased rates of survival with good neurological function.5 Notably, these 7 studies comprised a total of only 8067 OHCAs. Given the limited number of OHCAs in these trials, and as participating EMS agencies in the trials may have had higher-quality manual CPR for OHCA as compared to the majority of EMS agencies, there remains a need to better understand whether mechanical CPR device use is associated with improved survival, especially among a broader range of EMS agencies in routine practice.
Accordingly, in a national OHCA registry, we conducted an interrupted time series analysis to assess the association between mechanical CPR use and OHCA survival outcomes at EMS agencies initiating device use, given the growing uptake of this technology despite its cost and limited real-world data. We examined whether EMS agency rates of OHCA survival increased after mechanical CPR device introduction, evaluating both rates of favorable neurological survival and overall survival.
METHODS
Data for this study are available from the corresponding author on request and approval by the CARES registry. The study adhered to the Strengthening and Reporting of Observational Studies in Epidemiology reporting guidelines for observational studies and was approved by Saint Luke’s Hospital’s Institutional Review Board, which waived the requirement for informed consent as the study involved deidentified data.
Data Source
CARES is a prospective, multicenter registry of patients with an OHCA in the U.S., with a current catchment area of approximately 200 million residents, representing approximately 60 % of the U.S. population. Established by the Centers for Disease Control and Emory University, the design of the registry has been previously described.6, 7 Briefly, all patients with a non-traumatic OHCA for whom resuscitation is attempted are identified by EMS agencies. Standardized international Utstein definitions for reporting clinical variables and outcomes are used to ensure data uniformity.8 Each event is verified by a CARES analyst for completeness and accuracy.
Study Population
The adoption of mechanical CPR devices at EMS agencies markedly increased after the onset of the COVID-19 pandemic, leaving very few EMS agencies to serve as controls. Moreover, the onset of the pandemic was associated with a significant reduction in OHCA survival throughout the U.S.,9 making it challenging to disentangle the effect of mechanical CPR device introduction from the pandemic on OHCA survival outcomes. Therefore, we restricted our analyses of CARES data to the pre-pandemic period. Between January 1, 2013, and December 31, 2019, we identified 457,367 patients with a non-traumatic OHCA. We excluded 67,515 arrests occurring at a nursing home or healthcare facility, 50,261 arrests in which the OHCA was witnessed by 9-1-1 personnel (to be consistent with other OHCA studies), and 11,659 cases in children under 18 years of age to focus the analyses on adults with OHCA outside healthcare facilities Figure S1 in the Supplementary Appendix). We then excluded 66,657 OHCAs from 877 EMS agencies which had less than 4 years of participation in CARES, as we sought to compare OHCA survival at an agency with at least 2 years of data before and at least 2 years of data after mechanical CPR device use to generate more reliable agency-level survival rates, leaving 327,129 OHCAs at 568 EMS agencies with at least 4 years of CARES data. Of these, 398 agencies reported using a mechanical CPR device during the study period. Among the 398 agencies which used mechanical CPR devices, we excluded 310 agencies as they did not have at least 2 years of OHCA data in CARES either before or after introduction of mechanical CPR devices, 35 low-volume agencies with fewer than 20 OHCA cases annually, and 4 agencies with <1% mechanical CPR device use (as this suggested low dissemination of mechanical CPR devices at the agency), leaving 49 EMS agencies which adopted mechanical CPR devices in the study cohort. Among the 170 agencies which did not introduce mechanical CPR devices, we excluded 97 low-volume sites with fewer than 20 OHCA cases annually, leaving 73 control agencies in the study cohort. Our final study cohort comprised 83,908 OHCAs from 122 EMS agencies (49 which introduced mechanical CPR devices and 73 which did not).
Study Outcomes and Independent Variable
The primary outcome was favorable neurological survival, defined as survival to discharge with a Cerebral Performance Category score (range: 1–5) of 1 or 2, where 1 denotes no to mild neurological disability and 2 denotes moderate disability.10 The secondary outcome was survival to hospital discharge. We examined EMS agency rates of both outcomes before and after introduction of mechanical CPR devices at a given agency.
The independent variable was mechanical CPR device use. The earliest date an EMS agency reported the use of a mechanical CPR device for an OHCA in the CARES registry was defined as the agency’s date of introducing mechanical CPR device use.
Statistical Analysis
We determined, a priori, to conduct our study at the EMS agency-level rather than at the patient-level. Since mechanical CPR devices are only applied to patients who remain in cardiac arrest despite some period of manual CPR, defibrillation and epinephrine administration, a patient-level analysis comparing those who receive a mechanical CPR device vs. manual CPR will be biased against patients treated with mechanical CPR because patients who achieve return of spontaneous circulation with initial interventions (defibrillation, epinephrine) are typically assigned to the manual CPR group, as has been done in prior analyses. To overcome this limitation, the current study was performed at the EMS-agency level to determine whether introduction of mechanical CPR devices was associated with an improvement in OHCA survival outcomes,
Given the large sample size, baseline characteristics of patients before vs. after an EMS agency introduced mechanical CPR devices were compared using standardized differences, wherein a standardized mean difference of >0.10 indicated a clinically meaningful difference.11 Additionally, characteristics of patients at control agencies by calendar year were evaluated using a test for trend. At control agencies, baseline characteristics of patients by calendar year were assessed using p for trend.
We first assessed for temporal trends in OHCA survival at the 73 control agencies which did not introduce mechanical CPR devices during the study period. Among control agencies, we constructed a multivariable hierarchical logistic regression model with favorable neurological survival as the outcome, EMS agency as a random effect (to account for clustering of patient outcomes within site), and the following variables as fixed effects: patient demographics (age, race/ethnicity, sex), calendar year of arrest, etiology of arrest (presumed cardiac, respiratory, drug overdose, other), witnessed status of arrest, location of arrest (home, public building, industrial facility, street, recreational facility, transport center [e.g., airport]), whether bystander CPR was provided, and initial cardiac arrest rhythm (asystole, pulseless electrical activity, unknown non-shockable rhythm, ventricular fibrillation, pulseless ventricular tachycardia, and unknown shockable rhythm). We evaluated for temporal changes in rates of favorable neurological outcome using a p for trend.
Next, among the 49 EMS agencies which introduced mechanical CPR devices, we compared EMS agency-level rates of OHCA survival in the period after mechanical CPR device use began as compared to before their introduction. To accomplish this, we conducted an interrupted time series analysis, within a hierarchical logistic regression framework, to compare trends in favorable neurological survival before and after mechanical CPR device introduction at an EMS agency. Such an approach also allows us to examine trends in survival outcomes before introduction of mechanical CPR devices. This separate multivariable hierarchical logistic regression model also had EMS agency as a random effect, favorable neurological survival as the outcome, and adjusted for the same patient and cardiac arrest variables. In this analysis, trend lines (slope and intercept) were estimated for each EMS agency before and after mechanical CPR device introduction. Nonlinearity was assessed using restricted cubic splines but was nonsignificant, so straight-line trends were fitted. The slopes and intercepts were treated as correlated random effects within the hierarchical model using an unstructured covariance matrix. The resultant average slope and intercept from the model represented the overall population trend, and a joint statistical test was performed to assess whether the average intercept or slope for the outcome of favorable neurological survival differed after mechanical CPR device introduction, as compared with survival trends before their introduction. Specifically, either a more positive slope or higher intercept would indicate improved EMS rates of favorable neurological outcome after mechanical CPR device introduction beyond what would be expected based on pre-implementation trends. We then evaluated for an interaction between introduction of mechanical CPR devices at an agency and the frequency with which mechanical CPR devices were used at each agency (<20% vs. >20% and <45% vs. >45% of all OHCA cases), to assess whether the effect of mechanical CPR device use on agency-level rates of favorable neurological survival differed by the frequency with which mechanical CPR devices were used at an agency. Additionally, we evaluated whether our findings differed if 9-1-1 witnessed arrests were included in the study cohort.
We then repeated these analyses for the secondary outcomes of survival to discharge. Overall, patient data was nearly complete except for race and initial cardiac arrest rhythm. For race/ethnicity, we created a separate ‘unknown’ category for those with unknown or missing data. As initial cardiac arrest rhythm was missing in <0.05% of patients, imputation was not performed given its negligible rate.
All analyses were evaluated at a prespecified 2-sided significance level of 0.05 and were performed with SAS 9.4 (SAS Institute, Cary, NC).
RESULTS
At the 49 EMS agencies which introduced mechanical CPR devices, there were a total of 14,110 OHCAs before device introduction (mean pre-device time period of 3.2 ± standard deviation of 0.8 years) and 17,804 OHCAs after device introduction (mean post-device time period of 3.3 ± 0.8 years). Table 1 compares baseline characteristics of patients at these EMS agencies before and after mechanical CPR device introduction. There were no differences in age, sex, proportion of OHCAs which were witnessed, location of arrest, and first detected rhythm. However, patients in the period after mechanical CPR device introduction were more likely to be categorized as non-Hispanic White, owing to lower rates of unknown or missing race/ethnicity. Moreover, patients in the period after mechanical CPR device introduction were more likely to have a cardiac arrest etiology of drug overdose, since this etiology was introduced as an option for etiology of cardiac arrest in 2017.
Table 1. Patient Characteristics at EMS Agencies That Introduced Mechanical CPR Devices.
Characteristics of patients with OHCA in the time period before and after their EMS agency introduced mechanical CPR devices are shown.
| Before Mechanical CPR | After Mechanical CPR | Standardized | |
|---|---|---|---|
| N = 14,110 | N = 17,804 | Difference* | |
| Age, yrs | 0.02 | ||
| Mean ± standard deviation | 62.2 (17.1) | 61.8 (17.3) | |
| Median (interquartile range) | 63.0 (52.0, 75.0) | 63.0 (51.0, 74.0) | |
| Race and Ethnicity, n (%) | 0.42 | ||
| Non-Hispanic White | 6,328 (44.9) | 10,509 (59.0) | |
| Non-Hispanic Black | 2,324 (16.5) | 2,949 (16.6) | |
| Hispanic | 859 (6.1) | 1,402 (7.9) | |
| Non-Hispanic Asian | 165 (1.2) | 313 (1.8) | |
| Non-Hispanic Native American | 68 (0.5) | 95 (0.5) | |
| Unknown or missing | 4,366 (30.9) | 2,536 (14.2) | |
| Male Sex, n (%) | 9,048 (64.1) | 11,420 (64.1) | |
| Witnessed Arrest, n (%) | 6,032 (42.8) | 7,541 (42.4) | 0.01 |
| Bystander CPR, n (%) | 5,846 (41.4) | 7,244 (40.7) | 0.02 |
| Location Type, n (%) | 0.09 | ||
| Home Residence | 11,284 (80.0) | 14,493 (81.4) | |
| Public/Commercial Building | 1,524 (10.8) | 1,877 (10.4) | |
| Street/Hwy | 822 (5.8) | 906 (5.1 | |
| Place of Recreation | 215 (1.5) | 341 (1.9) | |
| Industrial Place | 84 (0.6) | 79 (0.4) | |
| Transport Center | 80 (0.6) | 64 (0.4) | |
| Other | 101 (0.7) | 44 (0.2) | |
| Cardiac Arrest Etiology, n (%) | 0.34 | ||
| Presumed Cardiac Etiology | 12,554 (89.0) | 14,980 (84.1) | |
| Respiratory/Asphyxia | 948 (6.7) | 1,302 (7.3) | |
| Drug Overdose | 43 (0.3) | 1,048 (5.9) | |
| Other | 565 (4.0) | 474 (2.7) | |
| First Detected Rhythm, n (%) | 0.04 | ||
| Non-Shockable | |||
| Asystole | 7,077 (50.2) | 9,231 (51.9) | |
| Pulseless Electrical Activity | 2,423 (17.2) | 2,907 (16.3) | |
| Unknown Unshockable Rhythm | 1,398 (9.9) | 1,920 (10.8) | |
| Shockable | |||
| Ventricular Fibrillation | 2,247 (15.9) | 2,630 (14.8) | |
| Ventricular Tachycardia | 165 (1.2) | 173 (1.0) | |
| Unknown Shockable Rhythm | 800 (5.7) | 936 (5.3) | |
| Missing | 0 | 7 |
A standardized difference of >0.10 denotes a significant difference
Table 2 summarizes baseline characteristics of 51,994 patients with OHCA at the 73 control agencies during the study period. Generally, there were no meaningful changes in patient and cardiac characteristics between 2013 and 2019 except for higher rates of drug overdose as a cardiac arrest etiology during later years.
Table 2. Patient Characteristics at EMS Agencies Without Mechanical CPR Device Use.
Patient characteristics at control agencies by calendar year are shown.
| 2013 | 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | |
|---|---|---|---|---|---|---|---|
| n = 5,625 | n = 6,638 | n = 7,359 | n = 7,515 | n = 8,393 | n = 8,217 | n = 8,247 | |
| Age, yrs | |||||||
| Mean ± standard deviation | 63.4 (16.9) | 63.08 (16.5) | 63.1 (16.7) | 62.4 (16.9) | 62.1 (17.3) | 62.5 (16.8) | 62.4 (17.0) |
| Median (IQR) | 64.0 (53.0, 76.0) | 64.0 (53.0, 75.0) | 64.0 (53.0, 76.0) | 63.0 (52.0, 75.0) | 63.0 (51.0, 75.0) | 64.0 (53.0, 75.0) | 64.0 (52.0, 75.0) |
| Race and Ethnicity, n (%) | |||||||
| Non-Hispanic White | 2,203 (39.2) | 2,571 (38.7) | 2,826 (38.4) | 2,842 (37.8) | 3,428 (40.8) | 3,299 (40.1) | 3,403 (41.3) |
| Non-Hispanic Black | 888 (15.8) | 1,015 (15.3) | 1,102 (15.0) | 1,076 (14.3) | 1,308 (15.6) | 1,220 (14.8) | 1,259 (15.3) |
| Hispanic | 141 (2.5) | 131 (2.0) | 140 (1.9) | 111 (1.5) | 140 (1.7) | 144 (1.8) | 187 (2.3) |
| Non-Hispanic Asian | 152 (2.7) | 192 (2.9) | 252 (3.4) | 286 (3.8) | 312 (3.7) | 326 (4.0) | 373 (4.5) |
| Non-Hispanic Native American | 144 (2.6) | 152 (2.3) | 145 (2.0) | 142 (1.9) | 154 (1.8) | 140 (1.8) | 198 (2.4) |
| Unknown or Missing | 2,097 (37.3) | 2,577 (38.8) | 2,894 (39.3) | 3,058 (40.7) | 3,051 (36.4) | 326 (37.6) | 2,827 (34.3) |
| Male Sex, n (%) | 3,610 (64.2) | 4,270 (64.3) | 4,721 (64.2) | 4,888 (65.0) | 5,447 (64.9) | 5,392 (65.6) | 5,389 (65.3) |
| Witnessed Arrest, n (%) | 2,661 (47.3) | 3,190 (48.1) | 3,443 (46.8) | 3,545 (47.2) | 4,078 (48.6) | 3,964 (48.2) | 4,034 (48.9) |
| Bystander CPR, n (%) | 2,378 (42.3) | 2,882 (43.4) | 3,192 (43.4) | 3,382 (45.0) | 3,528 (42.0) | 3,443 (41.9) | 3,643 (44.2) |
| Location Type, n (%) | |||||||
| Home Residence | 4,577 (81.4) | 5,434 (81.9) | 5,961 (81.0) | 6,102 (81.2) | 6,848 (81.6) | 6,619 (80.6) | 6,611 (80.2) |
| Public/Commercial Building | 435 (7.7) | 556 (8.4) | 595 (8.1) | 640 (8.5) | 733 (8.7) | 753 (9.2) | 771 (9.3) |
| Street/Hwy | 265 (4.7) | 337 (5.1) | 435 (5.9) | 409 (5.4) | 464 (5.5) | 488 (5.9) | 458 (5.6) |
| Place of Recreation | 147 (2.6) | 173 (2.6) | 210 (2.85) | 216 (2.9) | 205 (2.4) | 228 (2.8) | 238 (2.9) |
| Industrial Place | 34 (0.6) | 33 (0.5) | 44 (0.60) | 49 (0.7) | 40 (0.5) | 59 (0.7) | 43 (0.5) |
| Transport Center | 26 (0.5) | 27 (0.4) | 45 (0.61) | 33 (0.4) | 59 (0.7) | 44 (0.5) | 73 (0.9) |
| Other | 141 (2.5) | 78 (1.2) | 69 (0.94) | 66 (0.9) | 44 (0.5) | 26 (0.3) | 53 (0.6) |
| Cardiac Arrest Etiology, n (%) | |||||||
| Presumed Cardiac Etiology | 4,763 (84.7) | 5,590 (84.1) | 6,175 (83.9) | 6,277 (83.5) | 6,849 (81.6) | 6,831 (83.1) | 6,808 (82.6) |
| Respiratory/Asphyxia | 342 (6.1) | 423 (6.4) | 571 (7.8) | 550 (7.3) | 637 (7.6) | 574 (7.0) | 641 (7.8) |
| Drug Overdose | 0 (0.0) | 0 (0.0) | 0 (0.0) | 1 (0.0) | 605 (7.2) | 503 (6.1) | 538 (6.5) |
| Other | 520 (9.2) | 625 (9.4) | 613 (8.3) | 687 (9.1) | 302 (3.6) | 309 (3.8) | 260 (3.2) |
| First Detected Rhythm, n (%) | |||||||
| Non-Shockable | 2,239 (39.8) | 2,581 (38.9) | 3,100 (42.1) | 3,124 (41.6) | 3,763 (44.8) | 3,701 (45.04) | 3,615 (43.8) |
| Asystole | 883 (15.7) | 966 (14.6) | 1,076 (14.6) | 1,088 (14.5) | 1,367 (16.3) | 1,307 (15.9) | 1,426 (17.23) |
| PEA | 1,143 (20.3) | 1,492 (22.5) | 1,440 (19.6) | 1,548 (20.6) | 1,385 (16.5) | 1,287 (15.7) | 1,344 (16.3) |
| Unknown Unshockable | |||||||
| Shockable | 816 (14.5) | 963 (14.5) | 1,071 (14.6) | 1,071 (14.3) | 1,267 (15.1) | 1,282 (15.6) | 1,205 (14.6) |
| Ventricular Fibrillation | 46 (0.8) | 62 (0.9) | 91 (1.2) | 81 (1.1) | 87 (1.0) | 79 (1.0) | 87 (1.1) |
| Ventricular Tachycardia | 496 (8.8) | 574 (8.7) | 580 (7.89) | 601 (8.0) | 524 (6.2) | 560 (6.8) | 569 (6.9) |
| Unknown Shockable | 2 | 0 | 1 | 2 | 0 | 1 | 1 |
Abbreviations: IQR, interquartile range; PEA, pulseless electrical activity
Survival Outcomes
At EMS agencies that did not introduce mechanical CPR devices, annual rates of favorable neurological survival at the patient level were between 9.6% to 10.6% annually, with no trend in either direction during the 7-year period after risk adjustment (p for trend=0.23) (Table 3). Similarly, there were no significant trends in risk-adjusted rates of survival to discharge (p for trend=0.31). After accounting for patient and cardiac arrest characteristics, there were no temporal trends in either survival outcome at the EMS agency level (results not shown).
Table 3. Trends in Survival Outcomes at Control Agencies.
There were no temporal trends in rates of favorable neurological survival and survival to discharge at EMS agencies that did not introduce mechanical CPR devices.
| 2013 | 2014 | 2015 | 2016 | 2017 | 2018 | 2019 | P for trend | ||
|---|---|---|---|---|---|---|---|---|---|
| Outcomes | n = 5,625 | n = 6,638 | n = 7,359 | n = 7,515 | n = 8,393 | n = 8,217 | n = 8,247 | Unadjusted | Adjusted* |
| Favorable Neurological Survival | 555 (9.9%) | 647 (9.7%) | 719 (9.8%) | 794 (10.6%) | 853 (10.1%) | 850 (10.3%) | 789 (9.6%) | 0.80 | 0.23 |
| Survival to Discharge | 640 (11.4%) | 781 (11.8%) | 829 (11.3%) | 908 (12.1%) | 993 (11.8%) | 985 (12.0%) | 919 (11.1%) | 0.97 | 0.31 |
Models adjusted for EMS agency as random effect and the following as fixed effects: patient age, race/ethnicity, and sex; whether bystander CPR was provided; calendar year of arrest, etiology of arrest; witnessed status of arrest; location of arrest; and initial cardiac arrest rhythm.
At EMS agencies that introduced mechanical CPR devices, the median agency-level rate of mechanical CPR device use after their introduction was 16.4% (interquartile range: 3.8% to 56.5%) (Figure 1). Overall, 1,231 (8.7%) patients had favorable neurological survival in the time period before device introduction and 1466 (8.2%) in the time period after device introduction. At the EMS agency level, the mean risk-adjusted rate of favorable neurological survival was 8.9% ± 2.2% before mechanical CPR device introduction and 8.3% ± 1.3% after. When comparing risk-adjusted agency-level rates of favorable neurological survival, there was no change in model intercept or slope after introduction of mechanical CPR devices (Table 4). The model intercept had a nonsignificant decrease of 6% after device introduction (adjusted odds ratio [OR], 0.94 [95% CI: 0.80, 1.11]; P=0.48), and the change in slope was not significantly higher after device introduction (adjusted OR per year, 1.03 [95% CI: 0.96, 1.12]; P=0.41). Notably, there was no significant trend in the slope for favorable neurological survival prior to device introduction (adjusted OR per year, 0.96 [0.90, 1.03], P=0.25).
Figure 1. Variation in Use of Mechanical CPR Devices.
Among the 49 EMS agencies which started using mechanical CPR devices, there was wide variation in device use in the period after their introduction at the agency.
Table 4. Survival Outcomes After Introduction of Mechanical CPR Devices.
At EMS agencies which introduced mechanical CPR devices, there was neither a change in the model intercept or slope to suggest an improvement in rates of favorable neurological survival or survival to discharge for OHCA after introduction of mechanical CPR devices.
| Adjusted OR (95% CI) * | P value | |
|---|---|---|
| Favorable Neurological Survival | ||
| Annual survival slope before mechanical CPR introduction† | 0.96 (0.90, 1.03) | 0.25 |
| Annual survival slope after mechanical CPR introduction‡ | 1.00 (0.94, 1.05) | 0.89 |
| Change in survival slope after mechanical CPR introduction# | 1.03 (0.96, 1.12) | 0.41 |
| Change in model intercept after mechanical CPR introduction ** | 0.94 (0.80, 1.11) | 0.48 |
| Survival to Discharge | ||
| Annual survival slope before mechanical CPR introduction† | 0.96 (0.90, 1.01) | 0.13 |
| Annual survival slope after mechanical CPR introduction‡ | 1.01 (0.96, 1.06) | 0.81 |
| Change in survival slope after mechanical CPR introduction# | 1.05 (0.98, 1.13) | 0.17 |
| Change in model intercept after mechanical CPR introduction ** | 0.94 (0.80, 1.08) | 0.32 |
Models adjusted for EMS agency as random effect and the following as fixed effects: patient age, race/ethnicity, and sex; whether bystander CPR was provided; calendar year of arrest, etiology of arrest; witnessed status of arrest; location of arrest; and initial cardiac arrest rhythm.
Examines annual trend in survival before the EMS agency introduced mechanical CPR devices
Examines annual trend in survival after the EMS agency introduced mechanical CPR devices
Evaluates whether the slope of survival changed after introduction of mechanical CPR devices. A significant OR >1.0 would suggest improvement in survival rate after device introduction
Evaluates whether the model intercept for the survival outcome changed after introduction of mechanical CPR devices. A significant OR >1.0 would suggest instantaneous improvement in survival rate after device introduction.
Similarly, 1528 (10.8%) patients survived to discharge before mechanical device introduction and 1784 (10.0%) survived after device introduction. At the EMS agency level, the mean risk-adjusted rate of survival to discharge was 11.0% ± 2.2% before mechanical CPR device introduction and 10.0% ± 1.0% afterward. EMS agency rates of survival to discharge were not significantly higher (by model intercept or change in slope) in the time period after mechanical device introduction, and there was no survival trend prior to device introduction (see Table 4).
Because of variability in mechanical CPR device use across EMS agencies, we evaluated for an interaction between mechanical CPR device introduction and frequency of device use (<20% vs. >20% of all OHCA cases at an agency) for each survival outcome. Of the 49 agencies which introduced mechanical CPR devices, 22 (44.9%) used a mechanical CPR device in <20% of their OHCAs and 27 (55.1%) used a device in >20% of their OHCAs after device introduction. At EMS agencies with <20% mechanical CPR device use, rates of favorable neurological survival and survival to discharge were declining prior to device introduction and plateaued after device introduction. Therefore, there was a significant improvement in the slope of survival outcomes observed after device introduction when comparing survival slopes before and after device introduction (Table 5). At EMS agencies with >20% mechanical CPR device use, there was a trend for increasing rates of favorable neurological survival and survival to discharge prior to device introduction, which plateaued after device introduction. Therefore, there was a significant decrease in the slope of survival outcomes observed after device introduction when comparing survival slopes before and after device introduction. The interaction p-values were 0.001 and 0.002 for favorable neurological survival and survival to discharge, respectively. Results were similar when we repeated the interaction analyses at thresholds of <45% (bottom 2 tertiles) vs. >45% (top tertile) of mechanical CPR device use (Supplementary Appendix Table S1).
Table 5. Survival Outcomes, Stratified by Mechanical CPR Device Use Rate.
At EMS agencies with <20% mechanical CPR device use, survival was declining prior to device introduction and plateaued after device introduction. Therefore, there was a significant improvement in the slope of survival outcomes observed after device introduction. At EMS agencies with >20% mechanical CPR device use, there was a non-significant trend toward increased survival prior to device introduction, which plateaued after device introduction. Therefore, there was a significant decrease in the slope of survival outcomes observed after device introduction.
| Agencies with <20% mCPR use | Agencies with >20% mCPR use | ||||
|---|---|---|---|---|---|
| (n = 27) | (n = 22) | Interaction | |||
| aOR (95% CI) | P | aOR (95% CI) | P | P-value | |
| Favorable Neurological Survival | 0.001 | ||||
| Annual survival slope before mechanical CPR introduction† | 0.93 (0.86, 1.00) | 0.04 | 1.09 (0.97, 1.23) | 0.15 | |
| Annual survival slope after mechanical CPR introduction‡ | 1.03 (0.96, 1.10) | 0.44 | 0.93 (0.85, 1.02) | 0.13 | |
| Change in survival slope after mechanical CPR introduction# | 1.11 (1.01, 1.22) | 0.03 | 0.86 (0.74, 0.99) | 0.04 | |
| Change in model intercept after mechanical CPR introduction ** | 0.86 (0.70, 1.05) | 0.15 | 1.08 (0.81, 1.44) | 0.61 | |
| Survival to Discharge | 0.002 | ||||
| Annual survival slope before mechanical CPR introduction† | 0.92 (0.86, 0.98) | 0.01 | 1.08 (0.98, 1.20) | 0.12 | |
| Annual survival slope after mechanical CPR introduction‡ | 1.03 (0.97, 1.10) | 0.19 | 0.95 (0.88, 1.03) | 0.22 | |
| Change in survival slope after mechanical CPR introduction# | 1.12 (1.03, 1.22) | 0.006 | 0.88 (0.77, 1.00) | 0.04 | |
| Change in model intercept after mechanical CPR introduction ** | 0.87 (0.73, 1.05) | 0.16 | 0.99 (0.76, 1.28) | 0.92 |
Abbreviation: aOR, adjusted odds ratio
Models adjusted for EMS agency as random effect and the following as fixed effects: patient age, race/ethnicity, and sex; whether bystander CPR was provided; calendar year of arrest, etiology of arrest; witnessed status of arrest; location of arrest; and initial cardiac arrest rhythm.
Examines annual trend in survival before the EMS agency introduced mechanical CPR devices
Examines annual trend in survival after the EMS agency introduced mechanical CPR devices
Evaluates whether the slope of survival changed after introduction of mechanical CPR devices
Evaluates whether the model intercept for the survival outcome changed after introduction of mechanical CPR devices.
Finally, our main study findings were unchanged when we included 9-1-1 witnessed OHCAs in the study cohort. EMS agency rates of favorable neurological survival and survival to discharge were not higher (by model intercept or change in slope) in the time period after mechanical CPR device introduction (Supplementary Appendix Table S2).
DISCUSSION
In the largest registry for OHCA in the U.S., we found that a number of EMS agencies were already using mechanical CPR devices prior to the COVID-19 pandemic. Use of mechanical CPR devices varied across EMS agencies, with a median agency-level rate of 16.4% for all their OHCAs. However, introduction of mechanical CPR devices was not associated with higher EMS agency rates of favorable neurological survival or survival to discharge during a time period when there were no temporal changes in OHCA survival among agencies not using these devices.
Our study provides the first agency-level comparison of survival outcomes for OHCA before and after introduction of mechanical CPR devices and extends the existing literature on this technology. In a 2018 Cochrane review of 7 randomized and quasi-randomized trials, the authors found that only two studies (one for OHCA, one for in-hospital cardiac arrest) showed mechanical CPR devices increased rates of survival to discharge, and no study found device use increased rates of survival with good neurological function.5 No meta-analysis was done in this Cochrane review, given substantial heterogeneity. However, in this review, the largest randomized trial of mechanical CPR devices for OHCA (the PARAMEDIC trial12) found that patients randomized to treatment with mechanical CPR devices had no difference in 30-day survival compared with those receiving manual compressions.
Yet, uptake of mechanical CPR devices has increased substantially in routine practice.4 For this reason, there have been several observational studies that have evaluated OHCA survival among patients treated and not treated with mechanical CPR devices in real-world practice. Some studies have reported lower survival rates among patients treated with mechanical CPR devices as compared with manual CPR,13–16 while others have reported improved survival.17 However, a patient-level analysis comparing mechanical CPR with manual CPR in an observational study is prone to resuscitation time bias, as patients successfully treated with defibrillation or epinephrine during the early minutes of a resuscitation are typically assigned to the manual CPR group in prior studies and only patients with more prolonged resuscitations could be included in the mechanical CPR group. Data regarding the time of prehospital interventions including application of mechanical CPR device and for return of spontaneous circulation have not been available in prior studies, precluding a time-dependent propensity score analysis to address this survival bias.
To overcome this limitation in observational studies, we evaluated OHCA survival at the EMS agency level using an interrupted time series analysis. This approach assumes 1) no temporal trends in survival at agencies not using mechanical CPR devices and 2) no survival trends at agencies introducing device use in the time period before their introduction. We found there were no temporal trends in survival at EMS agencies that did not introduce mechanical CPR devices during the study period. Then, among agencies that introduced mechanical CPR devices, we compared risk-adjusted EMS rates of OHCA survival before vs. after their introduction and found neither an instantaneous change in agency-level rates of survival (i.e., model intercept) or in the trajectory of survival (i.e., change in slope), thus suggesting that introduction of mechanical CPR devices at was not associated with higher EMS rates of OHCA survival. Importantly, we found no temporal trend in survival outcomes in the years preceding their use that may have affected the survival slope after device introduction. Given the substantial costs to equip all EMS vehicles with a mechanical CPR device at an EMS agency, concerns about improper use of mechanical CPR devices and complications associated with their use (including traumatic visceral injuries to the thorax and abdomen),18, 19 our findings raise questions about their use for OHCA in routine care. Future research should identify where there may be patient settings in which mechanical CPR device use may improve OHCA survival (e.g., during transport of OHCA patients without return of circulation in refractory ventricular fibrillation).
We did find significant interactions between mechanical CPR device introduction and the frequency with which EMS agencies used the device for each survival outcome. For agencies with lower mechanical CPR device use (<20% of OHCAs), we found an improvement in the slope of survival after device introduction, but this was because survival rates were declining prior to device introduction and plateaued afterward, thus creating the positive slope change (significant adjusted OR >1). In contrast, for agencies with higher mechanical CPR device use (>20% of OHCAs), we found a decrease in the slope of survival after device introduction due to the fact that there was a trend toward improving survival outcomes annually prior to device introduction, which plateaued after device introduction, thus creating a negative slope change (significant adjusted OR <1). While these findings of changes in the slope of survival after device introduction may merely be due to temporal trends prior to device introduction, it is clear that the annual slope of survival after device introduction was flat in EMS agencies, regardless of the frequency with which a mechanical CPR devices was used.
During the COVID-19 pandemic, many more EMS agencies began using mechanical CPR devices, in part, because these devices minimized healthcare staff exposure to a potentially deadly pathogen, especially during the early years of the pandemic. We did not include the pandemic and post-pandemic years of 2020–2024 in our study for two reasons. First, there would only be 30 EMS agencies which would serve as controls as most EMS agencies had adopted mechanical CPR devices by 2024, and agencies which remained non-adopters of this technology may have substantial selection bias and therefore would not be robust controls to assess temporal trends. Additionally, as OHCA survival decreased markedly for all EMS agencies in the U.S. during the initial years of the pandemic,9 conducting a pre- vs. post- comparison of OHCA survival using data from this period would make it difficult to disentangle the effect of mechanical CPR on survival outcomes from that of the COVID-19 pandemic. This is because agencies introducing mechanical CPR devices during 2018 to March of 2020 would have a significant proportion of their follow-up period during the worst years of the COVID-19 pandemic (March 2020 to 2022). We would likely find that agencies which introduced mechanical CPR devices during this timeframe would have lower survival rates after introduction of mechanical CPR devices, and adjustment for temporal trends at control agencies may not be sufficient. In contrast, inclusion of agencies which introduced mechanical CPR devices during 2021–2022 would likely have shown a sharp decline in survival slope before mechanical CPR device introduction and a sharp rise in survival slope after mechanical device introduction, thus leading to an artefactual increase in change in survival slope after device introduction. Therefore, our analysis of the 2013–2019 period provides the best contemporary data to compare OHCA survival at EMS agencies before and after introduction of mechanical CPR devices.
Our study should be interpreted with the following limitations. First, there is the possibility that unmeasured patient-level confounding exists. For instance, detailed information on patient comorbidities is not systematically collected within CARES. However, unmeasured confounding is less an issue than for typical observational studies as we compared OHCA survival before and after introduction of mechanical CPR devices within the same agency, and so the case-mix of patients in a given community is unlikely to have changed substantially during the 7-year period. Second, as we did not have information on changes in training, staffing, transport practices, airway management strategies, post-arrest care, or other improvements in resuscitation at EMS agencies during the study period, there is the possibility of site-level confounding which we could not account for in our interrupted time series analysis. Third, we did not have information as to whether every EMS vehicle was equipped with a mechanical CPR device at a given agency, the type of training provided to ensure their proper use, and the time it took for a mechanical CPR device to be placed on a pulseless patient, to provide more granular information on agencies which introduced mechanical CPR devices. Fourth, our analyses were restricted to EMS agencies submitting data to CARES; thus, our findings may not pertain to EMS agencies not participating in this national registry.
Conclusion
In a large U.S. registry of OHCA, EMS agency rates of favorable neurological survival and survival to discharge were not higher at agencies which started using mechanical CPR devices.
Supplementary Material
CLINCAL PERSPECTIVE.
What is New?
In the U.S-based Cardiac Arrest Registry to Enhance Survival, an interrupted time series analysis assessed whether EMS agencies which started using mechanical CPR devices during 2013–2019 had higher rates of OHCA survival.
At 73 control agencies (no mechanical CPR device use) comprising 51,994 OHCAs, rates of OHCA survival were similar throughout the study period.
At 49 EMS agencies (31,914 total OHCAs) that began using mechanical CPR devices, device use was not associated with higher EMS agency rates of either favorable neurological survival or survival to discharge.
What Are the Clinical Implications?
Use of mechanical CPR devices in routine practice does not appear to increase rates of OHCA survival.
Given these study findings in real-world practice and lack of efficacy data in clinical trials, the settings in which mechanical CPR devices improve OHCA survival remain undefined.
ACKNOWLEDGMENTS:
Funding:
Drs. Chan and Girotra receive research funding from the National Heart, Lung, and Blood Institute (R01HL160734 and R01HL178638).
Dr. Chan and Dr. Girotra receives funding from the American Heart Association
Dr. Girotra is also supported by funding from the National Heart, Lung, and Blood Institute (R01HL166305).
Dr. McNally is the Executive Director of CARES, which receives funding from the CDC CARES Expansion and Modernization Grant as part of the Cardiovascular Advances in Research and Opportunities Legacy (CAROL) Act.
None of the above funders had any role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Dr. Chan had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Non-standard Abbreviations and Acronyms
- CARES
Cardiac Arrest Registry to Enhance Survival
- CPR
cardiopulmonary resuscitation
- EMS
emergency medical service
- OHCA
out-of-hospital cardiac arrest
- OR
odds ratio
Footnotes
Disclosures of Conflicts of Interest: none
Data Sharing Statement:
The present study analyzed data using the CARES registry. Data for this study are available from the corresponding author on request and approval by CARES.
REFERENCES
- 1.Bhanji F, Mancini ME, Sinz E, Rodgers DL, McNeil MA, Hoadley TA, Meeks RA, Hamilton MF, Meaney PA, Hunt EA, et al. Part 16: education, implementation, and teams: 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2010;122(18 Suppl 3):S920–933. [DOI] [PubMed] [Google Scholar]
- 2.Wik L, Kramer-Johansen J, Myklebust H, Sorebo H, Svensson L, Fellows B, Steen PA. Quality of cardiopulmonary resuscitation during out-of-hospital cardiac arrest. JAMA. 2005;293(3):299–304. [DOI] [PubMed] [Google Scholar]
- 3.Stiell IG, Brown SP, Christenson J, Cheskes S, Nichol G, Powell J, Bigham B, Morrison LJ, Larsen J, Hess E, et al. What is the role of chest compression depth during out-of-hospital cardiac arrest resuscitation? Crit Care Med. 2012;40(4):1192–1198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kahn PA, Dhruva SS, Rhee TG, Ross JS. Use of Mechanical Cardiopulmonary Resuscitation Devices for Out-of-Hospital Cardiac Arrest, 2010–2016. JAMA Netw Open. 2019;2(10):e1913298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wang PL, Brooks SC. Mechanical versus manual chest compressions for cardiac arrest. Cochrane database of systematic reviews (Online). 2018;8(8):CD007260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.McNally B, Stokes A, Crouch A, Kellermann AL. CARES: Cardiac Arrest Registry to Enhance Survival. Ann Emerg Med. 2009;54(5):674–683 e672. [DOI] [PubMed] [Google Scholar]
- 7.McNally B, Robb R, Mehta M, Vellano K, Valderrama AL, Yoon PW, Sasson C, Crouch A, Perez AB, Merritt R, Kellermann A. Out-of-hospital cardiac arrest surveillance --- Cardiac Arrest Registry to Enhance Survival (CARES), United States, October 1, 2005--December 31, 2010. MMWR Surveill Summ. 2011;60(8):1–19. [PubMed] [Google Scholar]
- 8.Jacobs I, Nadkarni V, Bahr J, Berg RA, Billi JE, Bossaert L, Cassan P, Coovadia A, D'Este K, Finn 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(21):3385–3397. [DOI] [PubMed] [Google Scholar]
- 9.Chan PS, Girotra S, Tang Y, Al-Araji R, Nallamothu BK, McNally B. Outcomes for Out-of-Hospital Cardiac Arrest in the United States During the Coronavirus Disease 2019 Pandemic. JAMA Cardiol. 2021;6(3):296–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jennett B, Bond M. Assessment of outcome after severe brain damage. Lancet. 1975;1(7905):480–484. [DOI] [PubMed] [Google Scholar]
- 11.Austin PC. Using the Standardized Difference to Compare the Prevalence of a Binary Variable Between Two Groups in Observational Research. Communications in Statistics – Simulation and Computation. 2009;38:1228–1234. [Google Scholar]
- 12.Perkins GD, Lall R, Quinn T, Deakin CD, Cooke MW, Horton J, Lamb SE, Slowther AM, Woollard M, Carson A, et al. Mechanical versus manual chest compression for out-of-hospital cardiac arrest (PARAMEDIC): a pragmatic, cluster randomised controlled trial. Lancet. 2015;385(9972):947–955. [DOI] [PubMed] [Google Scholar]
- 13.Gonzales L, Oyler BK, Hayes JL, Escott ME, Cabanas JG, Hinchey PR, Brown LH. Out-of-hospital cardiac arrest outcomes with “pit crew” resuscitation and scripted initiation of mechanical CPR. Am J Emerg Med. 2019;37(5):913–920. [DOI] [PubMed] [Google Scholar]
- 14.Beyer M, Mowbray F, Wanis N, Berger D, Brent C, Dunne RB, Shields T, Ball M, Miller J, Klausner HA. Comparing Outcomes of Mechanical and Manual CPR in Out-of-Hospital Cardiac Arrest: A Retrospective Cohort Study. Available at SSRN: 10.2139/ssrn.5188051. Accessed August 2, 2025. [DOI] [Google Scholar]
- 15.Buckler DG, Burke RV, Naim MY, MacPherson A, Bradley RN, Abella BS, Rossano JW. Association of Mechanical Cardiopulmonary Resuscitation Device Use With Cardiac Arrest Outcomes: A Population-Based Study Using the CARES Registry (Cardiac Arrest Registry to Enhance Survival). Circulation. 2016;134(25):2131–2133. [DOI] [PubMed] [Google Scholar]
- 16.Karasek J, Ostadal P, Klein F, Rechova A, Seiner J, Strycek M, Polasek R, Widimsky P. LUCAS II Device for Cardiopulmonary Resuscitation in a Nonselective Out-of-Hospital Cardiac Arrest Population Leads to Worse 30-Day Survival Rate Than Manual Chest Compressions. J Emerg Med. 2020;59(5):673–679. [DOI] [PubMed] [Google Scholar]
- 17.Chen YR, Liao CJ, Huang HC, Tsai CH, Su YS, Liu CH, Hsu CF, Tsai MJ. The Effect of Implementing Mechanical Cardiopulmonary Resuscitation Devices on Out-of-Hospital Cardiac Arrest Patients in an Urban City of Taiwan. Int J Environ Res Public Health. 2021;18(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Milling L, Mikkelsen S, Astrup BS. Characteristics of mechanical CPR-related injuries: A case series. J Forensic Leg Med. 2020;70:101918. [DOI] [PubMed] [Google Scholar]
- 19.Gaisendrees C, Gerfer S, Ivanov B, Sabashnikov A, Merkle J, Luehr M, Schlachtenberger G, Walter SG, Eghbalzadeh K, Kuhn E, et al. Outcomes after mechanical versus manual chest compressions in eCPR patients. Expert Rev Med Devices. 2021;18(10):1023–1028. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The present study analyzed data using the CARES registry. Data for this study are available from the corresponding author on request and approval by CARES.