Cardiac rhythm changes during transfer from the emergency medical service to the emergency department: a retrospective tertiary single-center analysis on prevalence and outcomes

Background Out-of-hospital cardiac arrest (OHCA) is a leading cause of death worldwide. Cardiac rhythms of OHCA patients can change during transportation and transfer from emergency medical services (EMS) to emergency department (ED). Objective Our objective was to study the prevalence of cardiac rhythm changes during transfer from the EMS to the ED in OHCA patients and the possible association with clinical outcomes. Methods We retrospectively studied adult OHCA patients admitted to the ED between January 2017 and December 2019. The primary outcome was the incidence of cardiac rhythm changes during transfer from EMS to the ED. Secondary outcomes were: ED survival, intensive care unit (ICU) survival, hospital survival, and maximum Glasgow Coma Scale (GCS) score during admission. Results We included 625 patients of whom 49 patients (7.8%) in the rhythm change group and 576 patients in the no rhythm change group. ED survival was significantly lower in the rhythm change group (26.5%) versus the no rhythm change group (78.5%, p<0.01). Conclusion Cardiac rhythm changes occur not uncommonly in OHCA patients during transfer from the Emergency Medical Services (EMS) to the Emergency department (ED). Our results showed some evidence that these changes are associated with a lower ED survival.


Introduction
Worldwide, out-of-hospital cardiac arrest (OHCA) is a leading cause of death. (1,2) In Europe, there are approximately 375,000 OHCA cases yearly. Hospital survival rates of OHCA patients are poor, ranging from 3.0% to 9.7%. (3,4) Some important factors contribute to these poor survival rates, such as primary cardiac rhythm, response time of bystander cardiopulmonary resuscitation (CPR), response time of emergency medical services (EMS), duration of CPR, and quality of CPR. (3,(5)(6)(7) In order to improve the outcomes of OHCA patients, many studies have focused on the prehospital factors. In terms of cardiac rhythm, patients with shockable primary cardiac rhythms (ventricular fibrillation; VF or ventricular tachycardia; VT) have higher survival and neurologically favourable survival than patients with non-shockable primary cardiac rhythms (asystole or pulseless electrical activity; PEA). (5,8,9) This cardiac rhythm can change during CPR. (10) In OHCA patients, this change is greatly influenced by the delay in initiation of CPR and its quality.
As a change from a shockable to a non-shockable rhythm increases the risk of unfavourable outcomes, a change from a non-shockable to a shockable rhythm decreases the risk of unfavourable outcomes. (10)(11)(12) It is known that CPR in the pre-hospital setting is often compromised by low performance scores as a result of loading and transportation of patients, leading to lower survival. (13)(14)(15)(16) The decrease in CPR quality during transfer might lead to cardiac rhythm changes and even loss of cardiac output. (15)(16)(17)(18)(19) A recent study demonstrated that during EMS transportation a nonshockable primary rhythm changed to a shockable rhythm during CPR in 5-15%. (11) However, the incidence of rhythm change in OHCA patients, during transfer from EMS to the ED is unclear.
It is expected that the difficulty and inadequacy of CPR during transfer from EMS to the ED could lead to a cardiac rhythm change, especially during time of unloading the patient from the ambulance and transport to the ED. Therefore, the aim of this study is to determine the incidence of cardiac rhythm changes during transfer from the EMS (i.e. last known cardiac rhythm of the EMS) to the ED in OHCA patients, and the possible association with survival.

Study design and population
This is a single-center retrospective cohort study at the institution of the corresponding author.
The Medical Ethics Committee reviewed and approved this study (number MEC-2019-0741). The need for informed consent was waived. We included all adult OHCA patients (≥ 18 years), primarily admitted to the ED between January 1 st 2017 and December 31 st 2019. We retrieved data from patients in this period before the SARS-CoV-2 (COVID-19) pandemic, as we thought that this period prior to the pandemic was more representative for the care under normal circumstances. These patients were identified by retrospectively screening all ED visits. Four researchers extracted the data: one researcher extracted all data, two researchers checked all data, and one researcher took several samples to screen the data on accuracy. All researchers who extracted the data were not blinded for the study hypothesis. Patients with a cardiac arrest after admission to the ED were classified as 'in-hospital cardiac arrest' (IHCA) patients and were therefore excluded. We also excluded patients with a known 'do not resuscitate' declaration, patients that were transferred from surrounding hospitals, and patients with a cardiac arrest due to trauma.

Emergency medical service and cardiopulmonary resuscitation
In the Netherlands a national emergency number will be called in case of a cardiac arrest. They will instruct the bystanders in performing basic life support (BLS) and will text nearby bystanders with CPR education to perform the CPR and get the automatic external defibrillator (AED). In the meanwhile, the closest ambulance will be dispatched. In case this ambulance is not close by, the police or fire department will be alarmed to continue BLS until the ambulance arrives. In case of a cardiac arrest, two ambulances will be send to the patient. Every ambulance has a specialised nurse to start advanced life support (ALS). In case a medical doctor is needed, a helicopter will be send. The focus of the EMS system in the Netherlands is to have an ambulance at scene within 20 minutes. After initiation of ALS, the EMS personal work with protocols, when to transport patients to which hospitals in the neighbourhood.

Study variables
We extracted the following patient characteristics from the electronic patient files: sex, age, medical history, cardiac risk factors (i.e., diabetes mellitus, hypercholesterolemia, hypertension, smoking, family history of cardiovascular disease, and obesity), and alcohol and drug abuse.
Additionally, we extracted data on the following clinical characteristics: cardiac arrest characteristics, causes of the arrest, types of medications used, laboratory values, treatment, and complications. As for cardiac arrest characteristics we collected the following variables: connected, shocks administered, duration of arrest until return of spontaneous circulation (ROSC), pre-hospital time (measured from the start of the arrest until ED arrival). As for cause of the arrest we collected the following variables: coronary artery disease, pulmonary embolism, intoxication, tamponade, tension pneumothorax, hypoxemia, hypovolemia, hypothermia, potassium disorder, neurological events (e.g. intracranial bleeding or ischemic events), and primary rhythm disorder. As for type of medications used we collected the following variables: adrenaline, amiodarone, and magnesium. Type of breathing assistance at ED arrival (e.g., spontaneous breathing, and bag valved mask). As for laboratory values we collected the following variables: initial arterial/venous blood gas values (pH, lactate), admission sodium (Na), potassium (K), hemoglobin (Hb), creatinine kinase myocardial band (CKMB), and high sensitive troponin T (hsTnT). As for the treatment of these patients, we collected the following variables: internal cardiac defibrillator (ICD), percutaneous coronary intervention (PCI), intra-aortic balloon pump (IABP), coronary artery bypass graft (CABG), impella, veno-arterial extracorporeal membrane oxygenation (VA-ECMO), extracorporeal cardiopulmonary resuscitation (ECPR), left-ventricular assist device (LVAD), and inotropic/vasopressor use). As for complications of these patients we collected the following variables: acute kidney injury (AKI), cerebrovascular accident (CVA), rearrest, infectious complications, liver failure, delirium, post-anoxic brain injury, pulmonary embolism, and bleeding.

Primary and secondary outcomes
Our primary outcome was incidence of cardiac rhythm changes between the last known prehospital cardiac rhythm and the first cardiac rhythm at the ED. In order to determine the change in cardiac rhythm from the EMS to the ED, we registered the last know cardiac rhythm at the EMS and the first known cardiac rhythm at the ED. Patients with a change in cardiac rhythm were included in the 'rhythm change group', and patients without a change in cardiac rhythm were included in the 'no rhythm change group'. Our secondary outcomes were: maximum Glasgow Coma Scale (GCS) score during admission, ED survival, ICU survival, and hospital survival.

Statistical analysis
The data was analysed in a pair-wise fashion. Continuous variables were either reported as means and standard deviation (SD) or medians and interquartile ranges (IQR) when normal or skewed distributed variables, respectively. Categorical variables were reported as numbers and percentages. For differences between continuous variables, independent samples t-test were used for normally distributed variables or else a Kruskal Wallis test. Assessing the normal distribution of a variable was done using a Shapiro-Wilk test. For categorical data, we compared the distributions using a Fisher's exact test. Further, we provide the descriptive statistics for each type of rhythm change (e.g., shockable to non-shockable). We also conducted another additional analysis in a similar fashion as our primary and secondary analysis, where we excluded patients that had ROSC at ED arrival. In this analysis, we attempt to compare the most vulnerable patients (i.e., patients in need of CPR at ED arrival) with and without cardiac rhythm change. A p-value of <0.05 was defined as statistically significant.

Results
Out of 765 screened OHCA patients, 625 were included in the analyses. Of the 123 excluded patients: 61 were trauma patients, 20 did not have a cardiac arrest, 11 were not admitted to our ED, 11 were IHCA patients, 7 had a double registration, 7 did not receive CPR, 3 were false registrations, and 3 had a 'do not resuscitate' declaration.

Patient characteristics
Patient characteristics are shown in Table 1

Clinical characteristics
Clinical characteristics of both groups are presented in Table 2 and Supplementary Material Table   A. Low-flow time and the time from arrest to sustained ROSC (≥ 20 min) were significantly higher in the rhythm change group (p<0.01). The most frequent causes of arrest in the total group were: coronary artery disease (25.9%), primary rhythm disorders (13.0%), hypoxemia (10.4%), pulmonary embolism (4.2%), and hypovolemia (2.9%).
The majority of the patient's last known EMS rhythm was ROSC in 59.2% of the rhythm change group versus 73.0% in the no rhythm change group (p=0.046). The majority of the patient's first ED rhythm was ROSC in 12.2% of the patients in the rhythm change group versus in 73.0% in the no rhythm change group (p<0.01).

Primary outcome
We found cardiac rhythm changes in 49 patients (7.8%) in the transfer from EMS to ED, while 576 (92.2%) patients had no rhythm change. Three different categories of rhythm changes occurred most within the rhythm change group, namely: ROSC to non-shockable (57.1%), shockable to non-shockable (26.5%), and non-shockable to ROSC (8.2%). The clinical characteristics of these each subgroup are listed in Supplementary Material Table B.

Secondary outcomes
The secondary outcomes are also listed in Table 2

Additional analysis
We have performed two additional analysis. First, we provide the descriptive statistics for each rhythm change (e.g., shockable to non-shockable) in Supplementary Material Table B, these are provided for reference only. Second, we repeated our initial analyses only including patients who had no ROSC at ED arrival. The results show that the survival differences (ED, ICU and hospital) between rhythm change and no rhythm change group disappeared (p>0.05). The results of this sensitivity analyses are presented in Supplementary Material Table C and D.

Discussion
The aim of the present study was to examine the consequences of cardiac rhythm changes between the last known rhythm at the emergency medical service (EMS) and the first observed rhythm at the emergency department (ED). In addition, we examined patient and clinical characteristics as well as outcomes of these two groups.
We found that in 7.8% of the OHCA patients cardiac rhythm changes between the last known rhythm at the EMS and the first observed rhythm at the ED. The observed changes in cardiac rhythm and output present in the majority of patients in the rhythm change group may partially explain the lower survival rate of these patients. Emergency department survival as well as hospital survival of patients in the rhythm change group is significantly lower, than of patients in the no rhythm change group. In our additional analyses, only including patients without ROSC at ED arrival, we found that the survival differences between the two groups disappeared.
To the best of our knowledge, this is the first study to focus on the occurrence of cardiac rhythm changes in OHCA patients during transfer from EMS to the ED. There are several potential factors that influence these cardiac rhythm changes during the transfer. For EMS personnel resuscitation is an intensive treatment that when it takes long it induces fatigue that may lead to lower quality of the resuscitation and/or attention for the patient. [15][16][17][18][19][20][21] During a transfer from the EMS to ED, there are many processes that come together. The patient and all connected devices have to be carefully unloaded from the ambulance, all intravenous lines, the orotracheal tube and patient' limbs have to be protected, and a proper transfer of medical information to the hospital staff must take place. All these processes may influence the decreased attention to the patient and in turn the effectiveness of the chest compressions.
The unfavourable outcomes in the no rhythm change group can be explained by the degradation of the cardiac rhythm and output. Unfortunately no previous studies reported the influence of change of cardiac rhythms from shockable to non-shockable on survival outcomes.
Two studies did report the influence of a rhythm change from non-shockable to shockable. Luo et al [12] found that change from a non-shockable rhythm to a shockable rhythm increases survival and favourable neurological outcome one month after the OHCA. Further, Cournoyer et al [11] found that change from a non-shockable rhythm to a shockable rhythm increases the chances of sustained prehospital ROSC. While the majority of patients in our study had rhythm changes from a shockable rhythm to a non-shockable rhythm or from ROSC to loss of cardiac output, it is expected that the survival is decreased. However, this lower survival signifies the importance of cardiac rhythm and output change awareness, specifically during transfer from the EMS to the ED.
We found several other significant differences between the patients in the rhythm change group and the patients in the no rhythm change group. First, low-flow time and the total arrest duration are both significantly longer in the rhythm change group than in the no rhythm change group. This difference can possibly be explained because of the higher number of persisting ROSC patients in the no rhythm change group, as well as a deterioration of rhythm could cause a longer need of CPR until ROSC or discontinuation of treatment. Second, the patients in the rhythm change group were significantly older at time of arrest than patients in the no rhythm change group. Third, patients in the rhythm change group had a lower pH and higher lactate on admission than patients in the no rhythm change group, which reflects the worse condition of the patients with rhythm change. Last, not only ED survival was lower in the rhythm change group, but hospital survival was also lower. These factors could also contribute to the lower survival outcome in the rhythm change group.

Limitations
Our study has several limitations. First, our retrospective study presents as much clinically relevant characteristics as possible of the cardiac rhythm change patients admitted to the ED. As this is one of the first studies focusing on this patient group. Even though we present p-values for statistical significance between rhythm change and no rhythm change groups, these should be interpreted with care due to the multiple comparisons issue. Our analyses is merely a starting point to identify clinically relevant characteristics for future research. Second, the exact timing of the last known EMS rhythm was not always clear. For some patients the last known EMS rhythm could possibly be the rhythm during transport instead of the cardiac rhythm just before transfer.
This discrepancy between the timing of the cardiac rhythm measures may have affected the incidence of rhythm changes and the survival. Third, we have missing values mostly with respect to the exact no-flow, low-flow, and pre-hospital times. In general these parameters suffer from recall bias and are less thoroughly stored in a patients record. As the cardiac care before admission in the hospital is about the survival of the patient and transportation to the hospital, these data are not always known or reported. Fourth, the amount of patients receiving interventions is much larger in the no-rhythm change group the rhythm change group. This might be due to the larger amount of ROSC as cardiac rhythm reported at the ED. However, this could have partially affected the outcomes. Last, as in many retrospective studies, the data we used was extracted from electronic patient files. Some of the included factors could be influenced by interpretation of the attending physicians (such as cause of arrest, could be based on history and diagnostic exams) and it is difficult to ascertain with certainty.

Future research
For future research it is important to determine more closely the rhythm changes that occurred.
Due to the retrospective nature of our study this was the best available proxy. In future studies, more closely monitoring of these changes, we are able to find clearer associations as compared to the present study. In our present study, other factors next to rhythm changes may have influenced the survival rate. Further, longer-term survival and quality of life should also be considered for future prospective studies. 14 Cardiac rhythm changes occur not uncommonly in OHCA patients during transfer from the Emergency Medical Services (EMS) to the Emergency department (ED). Our results showed some evidence that these changes are associated with a lower ED survival. ARTICLE SUMMARY:

Why is this topic important?
On an anecdotal level, we noticed that not uncommonly there is a difference between the cardiac rhythm reported by the EMS and what was first observed at the ED. For this reason, we examined the prevalence and clinical outcomes of cardiac rhythm changes in cardiac arrest patients who were transferred from the Emergency Medical Services (EMS) to the Emergency Department (ED).

What does this study attempt to show?
We investigate the prevalence and clinical outcomes of rhythm changes when cardiac arrest patients are transferred from the EMS to ED.

What are the key findings?
We found that rhythm changes occurs in around 8% of the patients when transferred from the EMS to ED. Further, we found that the rhythm changes may result in a lower ED survival of these patients.

How is patient care impacted?
This study may signify the need to keep monitoring the cardiac rhythm of cardiac arrest patients when transferred from the EMS to ED.