Traumatic cardiac arrest in Sweden 1990-2016 - a population-based national cohort study

This is a cohort study based on data from the Swedish Registry for Cardiopulmonary Resuscitation (SRCR) between 1990 and 2016. The SRCR is a nationwide population-based registry including OHCAs since 1990. From the start of the registry, the data for all Emergency Medical Service (EMS) managed OHCAs have been reported simultaneously, and according to Utstein methodology [18]. Since 2007, the SRCR has been web-based, today incorporating all EMS organisations in Sweden. The estimated coverage is close to 100% of all OHCAs since 2011(15). Further details have been described elsewhere [18].

Sweden has a population of approximately 9,700,000 people within an area of 447 435km². The EMS consisted in 2011 of 849 ambulances in 21 counties dispatched by 15 national dispatch centres reached by people in Sweden by one telephone number, 112. The call to the dispatch centre is free of charge, and both the use of EMS and acute care comes with a maximum fee of approximately 200 Euro per year for all inhabitants in Sweden. The ambulances are manned with two personnel each; in the majority of cases two registered nurses with additional paramedic training, all educated in advanced life support. Further, since 2005 according to a new law, all Swedish ambulances are staffed with at least one RN with special education in pre-hospital care. When responding to an OHCA, a dual-dispatch of two EMS units is used for quantitative reasons, providing advanced life support.

This was a study based on register data. Survivors are informed about their participation and can afterwards withdraw their inclusion in the registry. Since the start of the registry in 1990 only a handful of patients have withdrawn their participation. Deceased patients are included without consent. The Regional Ethical Review Board in Stockholm, Sweden approved the study, Dnr 2013/1959-31/4.

The manuscript does not contain individual person’s data, therefore consent for publications is not applicable.

The definition of OHCA in SRCR is, a patient who is unresponsive where cardiopulmonary resuscitation including external chest compressions with or without ventilation (CPR) and/or defibrillation have been either initiated by a first responder and continued by the EMS or initiated by the EMS [18]. In close connection to the resuscitation attempt, the EMS report presumed aetiology of the cardiac arrest. The aetiology is based on the EMS crew’s best assessment made by synthesising the chain of events in the following categories: heart disease, pulmonary disease, trauma, choking, drowning, intoxication by prescribed drugs, suicide, sudden infant death syndrome or other. Therefore, for the purposes of this study, TCA was defined as an OHCA [18] with traumatic aetiology which, according to Utstein, is defined as “cardiac arrest directly caused by blunt, penetrating or burn injury”. Therefore, only OHCAs with the EMS crew’s best assessment as “trauma” were included, i.e. hypoxic cases such as hanging and drowning were excluded. Medical CA was used as a comparator group and included all the remaining cases with the exception of choking, drowning, suicide and drug overdose. All cases of TCA occurring between January 1st 1990 and December 31st 2016 were eligible for inclusion in the study and identified through the SRCR [18]. No further exclusion criteria were applied. In the case of multiple CAs within the same pre-hospital episode, only the first event was included.

Patients were identified through the SRCR report sheet [19], where data on the following variables were collected: sex, age (in years), witnessed (yes by crew, yes by bystander or no), CPR before arrival of the EMS (yes or no), first documented heart rhythm (shockable (–Ventricle takycardia/Ventricle fibrillation) or non-shockable (Pulseless electric activity /asystole), place of OHCA (home, public or other), time between collapse and call (in minutes), time between collapse and start of CPR (in minutes), time between collapse and defibrillation (in minutes), time between collapse and arrival of the EMS (in minutes) as well as information on treatment provided during CPR (yes or no for the following: mechanical chest compressions, intubation, adrenaline, amiodarone; information on mechanical chest compression and amiodarone was available only in the web-based version). Information on the outcome, i.e. 30-day survival (yes or no) as well as information on Cerebral Performance Category (CPC) score at discharge from hospital among 30-day survivors was gathered from the SRCR and categorised as good (1-2) or poor (3-5) [19]. Data on CPC score was available only in the web-based version.

Descriptive statistics such as median with interquartile range (IQR) and mean with standard deviation (mean+/− SD) were used for patient characteristics. Characteristics of medical CA patients and TCA patients, as well as between survivors and non-survivors among TCAs were compared using Chi2 for dichotomous variables, and Student’s t-test for continuous variables. Two-sided tests were applied and a p-value ≤0.05 was interpreted as statistically significant. Missing data were kept missing, i.e. not imputed or estimated, and are presented in Table 1. Percentages in Table 2, Figs. 1 and 2 were based on numbers of patients with available information. A multivariable regression model was built based on significant variables, i.e. including witness status, shockable rhythm, given adrenaline, CPR before arrival of EMS, and time from call for help to arrival of EMS were used to estimate the associations between known prognostic factors and 30-day survival, expressed as odds ratios (OR) with 95% confidence intervals (95% CI). All statistical analyses were performed in SAS (version 9), SAS Inc., Cary, NC, USA.

In all, between 1990 and 2016 there were 1774 (2.4%) defined TCA cases within this study and 72,547 medical CA cases (Table 1). There were more men among TCA patients than among medical CA patients (78% vs 69%, respectively p-value < 0.01) and TCA patients were approximately 20 years younger than medical CA patients. TCAs were more likely not to have had anyone witness their arrest compared to medical CAs (36 and 29%, p-value < 0.01) but TCA patients received CPR before arrival of the EMS in 56% of cases versus 53% for medical CA, p-value 0.01). The heart rhythm was shockable in 10% of TCA cases compared to 30% of medical CA cases (p-value < 0.01). Delay from collapse to call, and from collapse to start of CPR were shorter among patients with witnessed TCA, whereas the opposite was found with regard to delay from call until arrival of the EMS, i.e. prolonged among patients with TCA.

Overall, 30-day survival was 3.7% (65 patients) in TCA cases versus 8.1% (5941 patients) in medical CA cases (p-value < 0.01). The survival ratio gradually increased for both TCAs and medical CAs and more than doubled from 1990 to 2015 (Fig. 1). No significant difference was found between medical CA cases and TCAs found in a shockable rhythm (p-value 0.23, data not shown).

A good neurological outcome, i.e. CPC 1-2, was found in 22 out of 28 (79%) patients with available information regarding survival after TCA as compared to 2380 out of 2644 (90%) with available information regarding survival after medical CA. However, the best CPC, i.e.1, was more common among medical CAs than TCAs (54% and 80% respectively, p-value < 0.01, Fig. 2).

Patients surviving a TCA were more often witnessed by bystanders than non-survivors (p-value < 0.01, Table 2). Two main differences were found between survivors and non-survivors; survivors more often received CPR before EMS arrival (57% compared to 41%, p-value < 0.01, Table 2) and they also had a shockable initial rhythm more frequently (44% compared to 8%, p-value< 0.01, Table 2). Further, survivors had a significantly shorter delay from collapse to start of CPR, to defibrillation and from call to arrival of the EMS (Table 2). Regarding treatment, survivors significantly less often received adrenaline than non-survivors (43% and 61% respectively, p-value < 0.01).

An independent association with increased survival to 30 days was found for bystander witnessed TCAs and a shockable rhythm, (adjusted OR 2.67, (95% C.I. 1.15 – 6.22) and OR 8.94 (4.27-19.67), respectively) while those given adrenaline had a poorer survival (adjusted OR 0.35, 95% C.I. 0.17 - 0.73, Table 3).