Background
Out-of-hospital cardiac arrest (OHCA) is often fatal and affects some 300,000 people in Europe each year [
1]. The majority of all reported cases are of a presumed cardiac aetiology [
2‐
4], however non-cardiac OHCAs include cardiac arrest in a trauma patient [
5,
6]. Trauma, resulting in cardiac arrest, is the main cause of death among young adults [
7,
8]. There is no international consensus for a definition of traumatic cardiac arrest (TCA) but incidents traumatic in origin can clinically be diagnosed as cardiac arrests in the case of agonal or absent spontaneous respiration and absence of a central pulse [
8,
9]. Resuscitation attempts including cardiopulmonary resuscitation (CPR) in TCAs have previously been considered futile and an inappropriate use of resources due to survival ratios as low as 0% and poor neurological outcome in survivors [
7,
10,
11].
However, recent studies have presented survival ratios similar to OHCA linked to other aetiologies or even better, i.e. up to 17% [
7‐
10,
12‐
17]. In addition, specific algorithms including one released by the European Resuscitation Council (ERC) has stressed the reversibility of aetiology behind the trauma [
8,
9]. So far, large population-based studies on prevalence, prognostic factors of certain importance in cardiac arrests and outcomes are sparse for this relatively young group of patients.
Therefore, we conducted a national population-based cohort study with the primary aim of describing characteristics and 30-day survival following a TCA as compared with a medical out-of-hospital cardiac arrest (medical CA). The secondary aim was to describe factors associated with 30-day survival among TCA patients.
Discussion
This national population-based cohort study identifies poor overall 30-day survival of 3.7% for TCAs compared to 8.1% for medical CAs. However, survival for TCA gradually increased up to 8% in 2015 and shockable rhythm was an important prognostic factor. However, even though survival seems low, still 65 young individuals were successfully resuscitated with mostly good neurological outcome.
This study is an observational registry-based cohort study and one needs to keep in mind that association does not imply causality. However, the reason for the finding that survivors had good neurological outcome might be that many TCAs were witnessed, there were relatively short time intervals within the chain of survival and there were equally high rates of bystander-CPR as for medical CAs.
The prevalence of TCAs in our study indicates that TCA is the most common non-medical etiology [
20] among all OHCAs although our figures were only about half that of recently reported data from Australia [
15]. When assessing prevalence in our study, we must stress that SRCR is primary a cardiac arrest registry and it is possible that only a fraction of all TCA in Sweden is included. However Beck et al. included only OHCAs from metropolitan Perth, and traumas fulfilling the inclusion criteria might be more common in metropolitan areas than in areas on a national level which include rural areas as well [
17]. Still both the Australian and Asian studies as well as ours indicate how rare TCAs are, which emphasises the need for simple and clear guidelines. Potentially survivable trauma deaths are most often due to extensive hemorrhage [
21] demanding different treatments than standard ALS. In addition, an algorithm for TCAs was introduced by the ERC in 2015 [
6]. The algorithm might constitute a useful guideline for the EMS to follow when confronted with a TCA.
Another rough indication of the prevalence of TCAs receiving CPR could be estimated based on national data for deaths and OHCAs in Sweden. Namely, during 1992-2014 a total of 62,697 individuals aged 35-79 years were reported dead due to either suicide, intoxications or trauma in Sweden [
22]. During the same period of time 4664 cases of OHCA due to TCA, intoxication, drowning or suicide [
20] receiving CPR were reported to the SRCR. So even if it is not possible to distinguish who suffered a trauma or not it indicates that CPR was initiated in approximately 7% of similar cases. Reasons for the low number of initiated CPR incidents in similar cases might be related to injuries not compatible with life or a late identification of the victim. Reasons may also be related to the concept that resuscitation attempts have been seen as futile, especially in blunt traumas [
7,
10,
11] These concepts and figures might still challenge the implementation of algorithms for TCAs in clinical practice.
The overall survival of 3.7% found within this study is fairly poor and a comparison to previous studies is difficult to make since the range varies between 0 and 17% [
14,
15] mainly due to differences in inclusion criteria and the lack of a national and population-based approach.
In Sweden, the general survival ratio for OHCAs has more than doubled during the study period [
20]. This is probably due to a greater awareness and knowledge in society as well as the optimisation of each link in the chain of survival, e.g. dispatch with earlier recognition of OHCA, a greater focus on recognising agonal breathing alongside dispatcher-assisted CPR as well as dual dispatch using the EMS and first responders such as the fire department or police [
23,
24].
The population suffering a TCA in our study is comparable to those described in previous literature, .i.e. consisting mainly of young males [
14,
15,
17]. However, characteristics related to cardiac arrest differ in several aspects which might explain our higher survival ratio than both the Qatari and the Australian ratio. Both the Qatari and the Australian study found a much lower proportion with a shockable rhythm both for TCAs and medical OHCAs, and TCAs were less likely to receive bystander-CPR than medical OHCAs [
15].
Interestingly, in our study, patients with TCAs received CPR before the arrival of the EMS slightly more often than medical CAs. In general, ratios for CPR before the arrival of the EMS in Sweden are high [
2], perhaps due to general awareness, a simplified CPR-training with shorter courses and one manikin per participant, thereby making these basic lifesaving skills more widely available in the community. Further, the introduction of Automated External Defibrillators (AEDs) and Public Access Defibrillation (PAD) programmes alongside general awareness in society have certainly played a role for CPR before the arrival of the EMS [
25]. The specific reasons for our ratio of bystander-CPR in TCAs are unknown but it is possible that relatives and others involved in the same trauma act as resuscitators.
The only significant factors associated with survival were whether the case was bystander witnessed, whether the first recorded rhythm was shockable and whether the patient was treated with adrenaline. However, it is possible that in cases with shockable rhythm the cardiac arrest came first and the trauma was a result of it, for example a patient having chest pain or arrhythmia might have braked in the car and thereby reduced the trauma. Our material includes patients with a cardiac arrest and trauma no matter if the trauma caused the cardiac arrest or if the cardiac arrest caused the trauma. However, EMS crews synthesise the whole picture at the scene including known preceding signs and the type of accident when categorising the cardiac arrest’s most likely etiology, i.e. cardiac or traumatic.
Regarding the poorer outcome among those given adrenaline the same finding has previously been shown in a registry-based study [
26] and needs to be scrutinised in the light of confounding. Likely, adrenaline can be seen as a proxy for a longer time to ROSC and/or arrival to hospital. However, such information is missing in the current registry but the findings demand future studies. Also, this study was observational and we need to bear in mind that associations do not imply causality.
The good neurological outcome among survivors of TCA in this study, even if survivors are few in number, is in line with previous studies [
14,
15] and is worth highlighting while implementing new algorithms for TCAs in clinical practice. Except for the constant improvement in the chain of survival, it is important to bear in mind that during the study period, general care of trauma has improved both before and after arrival in hospital. Many bigger cities in Sweden have trauma centers with high competence around the clock [
21] and a new algorithm for management of TCAs was released in 2015 by European Resuscitation Council [
9] stressing that TCAs are a state with a very low output, requiring immediate advanced directed interventions such as thoracotomy and external compression to stop bleeding. However, during the studied period there was no specific algorithm for prehospital management of TCAs in Sweden and thoracotomy in the field is still rare. However general trauma care has improved significantly between 1990 and 2016. The strengths of this study include the national and population-based design as well as the almost complete inclusion and constant validation of variables in the registry.
Limitations of this study include the design of the registry as a cardiac arrest registry rather than TCA registry and thereby the lack of information about the type of trauma and the severity index as well as information on the specific interventions undertaken such as thoracotomy [
13]. Further, cases of TCA who have not received resuscitation in the form of chest compressions and/or defibrillation have not been included in the registry, due to the registry’s specific inclusion criteria. In Sweden, only certain units staffed with doctors or ambulance helicopters perform thoracotomy, so likely it is an uncommon procedure. Further, the registry we used has a main focus on cardiac arrests, not trauma, therefore inclusion might be affected by selection bias, i.e. patients initially dispatched as traumatic in origin might have been missed. However, efforts are made on a routine basis to validate and to find missing cases in EMS case records and hospital journals. Finally, in some of the variables the proportion of patients with missing information was not minor, only 22 out of the 41 TCA-survivors since the start of CPC recording had a CPC score recorded. Therefore the results regarding neurological outcome should be interpreted with caution.