Abstract
Background
Out-of-hospital cardiac arrest (OHCA) is a serious condition. The volume-outcome relationship and various post–cardiac arrest care elements are believed to be associated with improved neurological outcomes. Although previous studies have investigated the volume–outcome relationship, adjusting for post-cardiac arrest care, intra-class correlation for each institution, and other covariates may have been insufficient.
Objective
To investigate the volume-outcome relationships and favorable neurological outcomes among OHCA cases in each institution.
Methods
We conducted a prospective observational study of adult patients with non-traumatic OHCA using the OHCA registry in Japan. The primary outcome was 30-day favorable neurological outcomes, and the secondary outcome was 30-day survival. We set the cutoff values to trisect the number of patients as equally as possible and classified institutions into high-, middle-, and low-volume. Generalized estimating equations (GEE) were performed to adjust for covariates and within-hospital clustering.
1 Department of Emergency Medicine, Teikyo University School of Medicine, 2-11-1 Kaga, Itabashi-ku, Tokyo 173-8606, Japan
Results
Among the 9909 registry patients, 7857 were included. These patients were transported to either low- (2679), middle- (2657), or high- (2521) volume institutions. The median number of eligible patients per institution in 19 months of study periods was 82 (range, 1–207), 252 (range, 210–353), and 463 (range, 390–701), respectively. After multivariable GEE using the low-volume institution as a reference, no significant difference in odds ratios and 95% confidence intervals were noted for 30-day favorable neurological outcomes for middle volume [1.22 (0.69–2.17)] and high volume [0.80 (0.47–1.37)] institutions. Moreover, there was no significant difference for 30-day survival for middle volume [1.02 (0.51–2.02)] and high volume [1.09 (0.53–2.23)] institutions.
Conclusion
The patient volume of each institution was not associated with 30-day favorable neurological outcomes. Although this result needs to be evaluated more comprehensively, there may be no need to set strict requirements for the type of institution when selecting a destination for OHCA cases.
Highlights
- •
We report the volume–outcome relationships and neurological outcome in OHCA cases.
- •
We adjusted intra-class correlation and post-hospital arrival covariates in GEE.
- •
There was no association between OHCA cases and favorable neurological outcomes.
- •
There was no association between OHCA cases and survival.
- •
We could decide where to transport regardless of the volume of the institution.
1
Introduction
Out-of-hospital cardiac arrest (OHCA) is an extremely serious condition. The American Heart Association reported that the overall survival to hospital discharge rate for patients with OHCA was approximately 10%, which remains low even in recent years [ ]. Various factors that are presumed to be associated with improved prognosis and neurological outcomes have been studied, including post-cardiac arrest care such as venoarterial extracorporeal membrane oxygenation (V-A ECMO), coronary angiography (CAG), intra-aortic balloon pumping (IABP), and targeted temperature measurement (TTM) [ , ]. The 2010 International Liaison Committee on Resuscitation (ILCOR) guidelines for cardiopulmonary resuscitation (CPR) were significantly updated in 2015, implementing the Grading of Recommendations Assessment, Development and Evaluation system [ ]. This revision included performing chest compressions at a depth of 5–6 cm, at least 100 compressions per minute (not to exceed 120 compressions per minute), and informing the importance of recoil during chest compression, the need for rapid extracorporeal cardiopulmonary resuscitation implementation, and the control of body temperature for at least 24 h at a target body temperature of 32 °C to 36 °C [ ]. Therefore, CPR and post-arrest care practices may have changed worldwide since this revision.
It has been hypothesized that as the number of transports at each institution increases, the treatment experience increases, and the overall quality of treatment, including post-cardiac arrest care may improve [ ]. This concept is referred to as the volume-outcome relationship and is well known regarding the number of various procedures or surgeries and better outcomes [ ]. Several studies have investigated the volume-outcome relationship between OHCA cases and favorable neurological outcomes [ , ]. These study results differ, which may be related to insufficient adjustment for posthospital arrival covariates, such as post-arrest care, or adjustment for hierarchy among institutions, referred to as within-hospital clustering. In addition, few studies have included patient data collected after the publication of the 2015 of ILCOR CPR guidelines [ ]. A study that did use patient data collected after the publication of these guidelines was not adjusted for post-hospital arrival covariates or within-hospital clustering [ ]. Therefore, the volume-outcome relationship and favorable neurological outcomes in OHCA patients has not been fully evaluated. If a volume-outcome relationship is shown, it may be possible to consolidate OHCA patients to specific institutions and enhance these institutions’ experience of clinical practice for OHCA patients. In addition, it may also influence the neurological prognosis of the OHCA patients.
This study was conducted prospectively to collect the necessary covariates and aimed to investigate the volume-outcome relationships between OHCA cases in each institution and favorable neurological outcomes by adjusting for within-hospital clustering and sufficient covariates including post-hospital treatments.
2
Methods
2.1
Study design and setting
We conducted a multicenter, prospective observational study using the OHCA registry of Survey of Survivors after Cardiac Arrest in the Kanto Area in 2017 (SOS-KANTO, 2017). Our research theme was predetermined before the data collection of SOS-KANTO, 2017; therefore, our study was a prospective observational study. Specifically, this was a registry of OHCA patients who were transported to one of the 47 hospitals in the Kanto area of Japan between September 2019 and March 2021. This registry collected prehospital and post-hospital information. Prehospital information was collected by emergency medical service (EMS) personnel and post-hospital information was collected by medical professionals, including physicians at each institution. The information was registered in a web-based system by physicians supervising study implementation at each institution. The method of data collection was visible, and the outcome assessors were not blinded.
Each participating institution obtained approval for the collection of SOS-KANTO 2017 information from their ethics committees. Approval for the current study (secondary analysis) was obtained from the ethics committee of the Jichi Medical University Saitama Medical Center (approval number: S19–012). Although informed consent was waived by the ethics review committee of each participating institution, as no interventions deviated from general CPR practices, we provided an opt-out procedure on the website of the Department of Emergency Medicine of the Jichi Medical University Saitama Medical Center.
This study was conducted based on the principles of the Declaration of Helsinki and its later amendments and was reported in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement guidelines [ ] (Supplementary Table 1).
2.2
Participants
This study included all registry patients who met the following criteria: 1) OHCA and 2) resuscitation performed by EMS personnel. Patients who met any of the following criteria were excluded: 1) traumatic cause of OHCA; 2) age < 18 years; and 3) primary or secondary outcome data was missing. Enrolled patients were followed up for 90 days after cardiac arrest.
2.3
Data collection
The following data were collected: age, sex, activities of daily living (good, moderate impairment, severe impairment), location of cardiac arrest (home, public space, others), transportation by vehicular or air ambulance with a physician present, witness status (EMS personnel, others, none), bystander CPR, the initial monitored cardiac rhythm (ventricular fibrillation [VF]/pulseless ventricular tachycardia [VT], pulseless electrical activity [PEA]/asystole, others), prehospital advanced airway management, prehospital administration of adrenaline, administration of antiarrhythmic drug (amiodarone, lidocaine, nifekalant, magnesium), the presence of shock delivery, the cause of cardiac arrest (cardiogenic, non-cardiogenic), the time from the emergency call to EMS contact, the time from the emergency call to hospital arrival, the duration of CPR, the cardiac rhythm at hospital arrival (VF/pulseless VT, PEA/asystole, others), Glasgow Coma Scale (GCS) score at hospital arrival, lactic acid at hospital arrival, treatment after hospital arrival (V-A ECMO, CAG, IABP, TTM), presence or absence of return of spontaneous circulation (ROSC; non-sustained, sustained), death declaration at emergency department, 30-day survival, and cerebral performance category (CPC) [ ]. When considering OHCA cases per institution, no limitations were set for the presence or absence of ROSC or hospitalization. Based on a previous study, we classified the status of activities of daily living into good, moderate impairment, and severe impairment; location of cardiac arrest into home, public space, and others; and cardiac rhythm at hospital arrival into VF/pulseless VT, PEA/asystole, and others [ , ].
2.4
Outcome measures
The primary outcome evaluated was 30-day favorable neurological outcome, which was defined as a CPC score of 1 or 2 [ ]. The secondary outcome was 30-day survival.
2.5
Statistical analyses
Continuous variables are described using mean and standard deviation (SD); categorical variables are described using absolute count and percentage (%). Analysis of variance and the chi-squared test were used for univariate analyses.
We calculated the number of eligible OHCA cases in each institution. Similar to previous studies, the cutoff values for each group were determined to trisect the number of patients as equally as possible and we classified the highest, middle, and lowest one-third of institutions as high-volume, middle-volume, and low-volume institutions, respectively [ , ]. Regarding the duration of CPR, values greater than twice the third quartile range were considered outliers and were treated as missing values.
Multiple imputation by generating 50 datasets using the chained equations method was performed to complement missing data for the generalized estimating equations (GEE). After multiple imputations, we performed GEE to adjust for multiple covariates as well as within-hospital and within-patient clustering. We set a hierarchical structure at the level of each of the 47 institutions and at the patient level. In addition, we considered the correlation matrix as “exchangeable.” The response variable was 30-day favorable neurological outcomes or 30-day survival, and the following explanatory variables were identified based on previous studies: volume of each institution (low-volume institution, middle-volume institution, high-volume institution), age, sex, activities of daily living (good, moderate impairment, severe impairment), location of cardiac arrest (home, public space, others), transportation by vehicular or air ambulance with a physician present, witness status (EMS personnel, others, none), bystander CPR, initial monitored cardiac rhythm (VF/pulseless VT, PEA/asystole, others), presence of shock delivery, prehospital advanced airway management, prehospital administration of adrenaline, administration of antiarrhythmic drugs, cause of cardiac arrest (cardiogenic, non-cardiogenic), time from the emergency call to EMS contact, duration of CPR, ECMO, CAG, IABP, and TTM [ , , , ]. Effect estimates were described using odds ratios (OR) and 95% confidence intervals [CI].
As a sensitivity analysis, we performed a logistic regression analysis after multiple imputations to consider the case where within-hospital clustering was not taken into account. We performed logistic regression analysis using the same response and explanatory variables used for GEE.
Statistical analyses were performed using SPSS software (version 24.0; IBM Corp, Armonk, NY). A two-sided p -value of <0.05 was considered statistically significant.
3
Results
3.1
Patient enrollment
A total of 9909 patients were registered; the records of 7857 were analyzed ( Fig. 1 ). Patient inclusion and exclusion criteria are presented in Fig. 1 .
3.2
Distribution and grouping of institutional volume
In the study population, 2679, 2657, and 2521 patients were transported to low-volume institutions (27 institutions), middle-volume institutions (10 institutions), and high-volume institutions (5 institutions), respectively, ( Fig. 2 ) during the 19-month study period. Five of the participating hospitals had no eligible cases. The median numbers of eligible patients per institution in 19 months of study periods were 82 (range, 1–207), 252 (range, 210–353), and 463 (range, 390–701) in the low-, middle-, and high-volume institutions, respectively ( Table 1 ).
| Variables | Overall n = 7857 | Low-volume institutions ( n = 2679 in 27 institutions) | Middle-volume institutions ( n = 2657 in 10 institutions) | High-volume institutions ( n = 2521 in 5 institutions) | p value |
|---|---|---|---|---|---|
| OHCA cases per institution | 187.1 ± 154.3 | 99.2 ± 69.0 | 265.7 ± 49.5 | 504.2 ± 119.5 | <0.01 |
| Age, years | 74.0 ± 14.6 | 74.6 ± 14.6 | 74.9 ± 14.2 | 72.6 ± 14.9 | <0.01 |
| Male, n (%) | 4875 (62.1) | 1663 (62.1) | 1641 (61.8) | 1571 (62.3) | 0.92 |
| Activities of daily living, n (%) Good Moderate impairment Severe impairment | 3033 (38.6) 2394 (30.5) 362 (4.6) | 1133 (42.3) 787 (29.4) 148 (5.5) | 921 (34.7) 794 (29.9) 134 (5.0) | 979 (38.8) 813 (32.2) 80 (3.2) | <0.01 0.80 <0.01 |
| Location of cardiac arrest, n (%) Home Public space Others | 5296 (67.4) 1394 (17.7) 810 (10.3) | 1810 (67.6) 457 (17.1) 248 (9.3) | 1757 (66.1) 484 (18.2) 269 (10.1) | 1729 (68.6) 453 (18.0) 293 (11.6) | 0.38 0.54 0.15 |
| Transportation by vehicular or air ambulance with a physician, n (%) | 501 (6.4) | 207 (7.7) | 90 (3.4) | 204 (8.1) | <0.01 |
| Witness status, n (%) EMS personnel Others None | 460 (5.9) 3360 (42.8) 4037 (51.4) | 116 (4.3) 1116 (41.7) 1447 (54.0) | 183 (6.9) 1136 (42.8) 1338 (50.4) | 161 (6.4) 1108 (44.0) 1252 (49.7) | <0.01 0.83 <0.01 |
| Bystander CPR, n (%) | 3300 (42.0) | 1155 (45.6) | 1249 (51.7) | 896 (36.3) | <0.01 |
| Initial cardiac rhythm monitored, n (%) VF/pulseless VT PEA/asystole Others | 674 (8.6) 6166 (78.5) 477 (6.1) | 194 (7.2) 2076 (77.5) 141 (5.3) | 217 (8.2) 2073 (78.0) 132 (5.0) | 263 (10.4) 2017 (80.0) 204 (8.1) | <0.01 0.58 <0.01 |
| Shock deliverry, n (%) | 1451 (18.5) | 475 (19.1) | 459 (18.5) | 517 (20.7) | 0.12 |
| Prehospital advanced airway management, n (%) | 4718 (60.1) | 1366 (93.6) | 1601 (93.6) | 1751 (90.2) | <0.01 |
| Prehospital administration of adrenaline, n (%) | 2365 (30.1) | 781 (30.9) | 817 (33.2) | 767 (30.8) | 0.13 |
| Administration of antiarrhythmic drugs, n (%) | 503 (6.4) | 162 (6) | 163 (6.1) | 178 (7.1) | 0.27 |
| Cardiac OHCA, n (%) | 5831 (74.2) | 1847 (68.9) | 1879 (70.7) | 2105 (83.5) | <0.01 |
| Time from call to EMS contact, mins | 8.3 ± 3.8 | 8.7 ± 3.9 | 8.1 ± 4.2 | 8.2 ± 3.4 | <0.01 |
| Time from call to hospital arrival, mins | 36.7 ± 14.1 | 36.3 ± 16.0 | 35.1 ± 15.4 | 38.7 ± 9.7 | <0.01 |
| Duration of CPR, mins | 45.3 ± 19.7 | 46.2 ± 20.4 | 43.9 ± 20.2 | 45.8 ± 18.3 | <0.01 |
| Cardiac rhythm at hospital arrival, n (%) VF/pulseless VT PEA/asystole Others | 347 (4.4) 6843 (87.1) 667 (8.5) | 103 (3.8) 2368 (88.4) 208 (77.6) | 122 (4.6) 2326 (87.5) 209 (78.7) | 122 (4.8) 2149 (85.2) 250 (9.9) | 0.35 <0.01 0.08 |
| Glasgow coma scale score at hospital arrival k | 3.2 ± 1.2 | 3.2 ± 1.2 | 3.1 ± 1.1 | 3.2 ± 1.3 | 0.73 |
| Lactic acid at hospital arrival, mmol/l | 14 (10.5–17.7) | 14.3 (10.7–18.3) | 14.3 (10.6–17.6) | 13.5 (10.2–17) | <0.01 |
| CAG, n (%) | 492 (6.3) | 144 (5.4) | 142 (5.3) | 206 (8.2) | <0.01 |
| ECMO, n (%) | 310 (4.0) | 86 (3.2) | 82 (3.1) | 142 (5.6) | <0.01 |
| IABP, n (%) | 196 (2.5) | 64 (2.4) | 38 (1.4) | 94 (3.7) | <0.01 |
| TTM, n (%) | 434 (5.5) | 112 (4.2) | 111 (4.3) | 211 (8.4) | <0.01 |
| ROSC, n (%) Non-sustained Sustained | 934 (11.9) 1377 (17.5) | 314 (11.7) 419 (15.6) | 386 (14.5) 407 (15.3) | 234 (9.3) 551 (21.9) | <0.01 <0.01 |
| Death at ED, n (%) | 6431 (81.9) | 2240 (83.6) | 2226 (83.8) | 1965 (77.9) | <0.01 |
Related posts:
CT-based diagnosis in patients presenting with throat pain: A single institutional review
Droperidol undermining gastroparesis symptoms (DRUGS) in the emergency department
The long and the short of pediatric emergency department antibiotic prescribing: A retrospective observational study
Efficacy and safety of corticosteroid therapy in patients with cardiac arrest: A meta-analysis of randomized controlled trials
Further considerations regarding acute limb ischemia
Improving CT scanner efficiency for trauma team activations in the emergency department
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
