Epidemiology of prehospital traumatic cardiac arrest in Geneva: a retrospective cohort study

Abstract

Background

Traumatic cardiac arrest (TCA) has traditionally been seen as a non-survivable event. Advances in prehospital resuscitation have improved outcomes, and current guidelines emphasize addressing reversible causes of TCA, namely hypovolemia, hypoxemia, tension pneumothorax and tamponade. Understanding data on the causes and outcomes of TCA in a dense urban area may help improve treatment and prevention.

Methods

We conducted a retrospective cohort study of all patients who experienced a prehospital TCA in Geneva from January 1st, 2008, to December 31st, 2022. Data was extracted from the databases of the prehospital physician-led emergency service (SMUR) and the forensic medicine service.

Results

During the 15-year period, a total of 673 TCA cases were identified, of which 464 (68.9%) patients were managed by the SMUR. The average age was 52 ± 22 years, with 72.1% of males. In 90.5% of cases, the TCA occurred before the arrival of the SMUR, although emergency response times were short (8.7 min ± 4.4). 216 (46.5%) patients underwent resuscitation. Of these patients, return of spontaneous circulation (ROSC) was achieved in 71 (32.9%) and 62 patients (28.7%) survived to hospital admission. 8 patients (6 blunt, 2 penetrating) were alive at 30 days (3.7%). Pediatric TCA (< 18 years old) was rare, representing only 21 (3.1%) cases.

Conclusion

Between 2008 and 2022, over 44 TCA cases per year were recorded in Geneva. Despite rapid emergency response times and ROSC achieved in more than one-third of patients who underwent resuscitation efforts, survival at 30 days was low. Factors associated with ROSC were crystalloid administration, timing of TCA, mechanism of injury, initial rhythm and the presence of an upper airway obstruction. This data underscores the need for the development and implementation of targeted interventions to enhance clinical outcomes and reduce preventable deaths within this specific patient population.

Trial registration

This trial was approved by the regional research ethics committee, the Commission Cantonale d’Ethique de la Recherche sur l’être humain (CCER). (Project-ID: 2023 − 00767).

Introduction

Accidents and suicides are the third most common cause of death in Switzerland [1]. They are the leading cause of death among men aged 15 to 44 and women aged 17 to 28. Globally, trauma accounts for 9% of deaths, surpassing the combined fatalities from HIV, tuberculosis, and malaria [2]. In severe cases, trauma patients experience cardiac arrest before reaching the hospital. The prognosis for prehospital traumatic cardiac arrest (TCA) is grim, with survival rates ranging from 0.0 to 3.8% [3, 4]. In contrast, the survival rates for prehospital medical cardiac arrest (e.g. following a myocardial infarction) are 17.0% in Geneva and 11.0% in Europe [5, 6]. The epidemiology of TCA varies significantly across different regions of the world and findings from some studies may not apply to urbanized areas with a high population density. Compact and densely populated urban areas benefit from short emergency response and transport times and outcomes may be different from more rural settings.

The aim of this trial is to describe the epidemiology and prehospital management of TCA in a dense European urban setting over 15 years using both the physician-led prehospital service and forensic medicine service registries. A secondary aim is to identify factors associated with ROSC. Given this condition’s historically high mortality rate, which primarily affects the young, a comprehensive overview within this setting regarding TCA could identify potential opportunities for improvement in prevention and management with relevance to comparable environments worldwide.

Methods

Setting

Geneva is Switzerland’s second-largest city after Zürich, with the canton (or state) of Geneva having a population of just over half a million inhabitants [7]. Prehospital care in Geneva is provided by ambulance services staffed with paramedics. In the event of a medical emergency, the Emergency Medical Call Center (EMCC) dispatches appropriate resources to the scene of the incident. In life-threatening situations, such as major trauma or cardiac arrest, physicians from the prehospital physician-led emergency service are also dispatched to assist the ambulance crew. This service is known as the Service Mobile d’Urgence et de Réanimation (SMUR) - which translates as mobile emergency and resuscitation service - and is staffed by a prehospital emergency medicine physician and a paramedic. The level of training of SMUR physicians is variable, with most being senior trainees in emergency medicine or anaesthesia. In TCA cases, a supervisor (senior emergency medicine or anaesthesia consultant) on call is usually dispatched to assist with the case. The SMUR is based at the only level-1 trauma center in the canton, the Geneva University Hospital (HUG). This service attends about 5000 calls yearly, and trauma accounts for just over 10% of its workload. A helicopter emergency medical service (HEMS) is also available, but only attends around 300 to 400 calls yearly. In certain instances, when a trauma-related death is obvious, a physician from the forensic medicine service may be sent to the scene to assess the patient and complete a death certification. Consequently, the SMUR was not involved in all prehospital TCA cases.

Study design

A retrospective cohort study was conducted according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement guidelines [8]. Data was collected from two registries, the first being the SMUR’s electronic patient record and the second from the forensic medicine service database.

Patient selection

All cases of prehospital TCA from January 1st, 2008, to December 31st, 2022, were included in the study, thus covering 15 years. TCA was defined as cardiac arrest secondary to blunt, penetrating, or burn injury, according to the Utstein criteria [9]. Two authors (TV and RM) reviewed all prehospital cases to determine inclusion and exclusion criteria. If there was a disagreement, a third author (LS) was consulted.

The study excluded patients who experienced TCA during transfer from another hospital, those with a suspected medical cause of cardiac arrest, cases occurring outside the canton of Geneva, and individuals who had documented their refusal to participate in medical research. Since the forensic medicine service frequently investigates TCA cases, some patients appeared in both registries. In such instances, duplicates were identified by the authors and merged into single cases while integrating data from both registries into the database. Furthermore, patients who appeared in both registries were classified as SMUR patients as their prehospital data was accessible.

Variables

Data regarding patient demographics, trauma mechanism, dispatch times, vital signs, performed interventions, were collected and managed using REDCap electronic data capture tools hosted at HUG (Supplementary Table 1) [10, 11].

Data sources/measurement

When the SMUR attends a call, an electronic patient report is completed and includes patient demographics, clinical data, interventions, and prehospital outcomes. The forensic medicine service maintains a database of all patients managed by their service and details patient demographics, date of death, cause of death, and injuries sustained. The types of injuries were determined based on those observed or suspected by the prehospital physician. When patients appeared in both the forensic medicine service’s database and the SMUR’s database, this information was used to supplement any missed or incomplete injury patterns. Timings were also recorded. SMUR arrival delay was defined as the interval between the time the SMUR received the alarm notification and the time of its arrival at the scene, in minutes, and was considered, in the analysis, as a continuous variable. Initial vital signs were the first vital signs recorded upon the SMUR’s arrival. To assess the severity of patients, prehospital emergency services in Switzerland use the National Advisory Committee for Aeronautics (NACA) severity score [12]. This score ranges from 0 (no injury or disease) to 7 (lethal injuries or disease with or without resuscitation attempts). By definition, all the patients who were recruited for this study had a NACA score of 6 (injuries/diseases transported after successful resuscitation) or 7. We also documented major system changes during the study period.

Outcomes

The primary outcome measure was return of spontaneous circulation (ROSC). In addition, factors associated with ROSC and survival previously identified in prehospital TCA trials were investigated in a sub-analysis [13, 14].

Statistical methods

The statistical analysis was conducted using STATA (version 17.0) for descriptive statistics and logistic regression, while the R programming language (version 4.3.1) was employed to implement the Random Forest (RF) models.

Descriptive data was presented as means and standard deviation (SD) for continuous variables and as numbers with percentages for categorical variables. The normality of continuous variables was assessed using the Shapiro-Wilk test. The independent sample t-test was used to analyze parametric data, a Wilcoxon rank sum test was used to analyze nonparametric continuous data, and a Pearson chi-square or Fisher exact test was used to analyze categorical variables. The significance level was set at 0.05 throughout the paper.

A multivariable logistic regression model was selected as the primary analytical method due to its interpretability and frequent use in clinical research to identify factors associated with the return of spontaneous circulation (ROSC) in patients who received either bystander or professional resuscitation. The outcome variable was ROSC, with predictors encompassing patient demographics, trauma and injury characteristics, physiological indicators, and resuscitation measures. Interaction terms were incorporated to examine the potential combined effects of specific predictors on ROSC. Results were presented as odds ratios (OR) with 95% confidence intervals (CI).

To supplement the logistic regression, a random forest model was applied to explore potential nonlinear relationships and complex interactions among predictors, without the constraints imposed by parametric modelling. This machine learning approach was used to assess variable importance and identify potential predictors that might not have been detected through logistic regression.

The RF models were built using the ‘randomForest’ package (version 4.7–1.2). The algorithm constructs an ensemble of decision trees, each trained on a bootstrapped subset of the data, and the final prediction is determined by majority vote. At each node, a subset of predictor variables (m) is randomly selected to determine the optimal split based on the Gini impurity index. To assess the relative importance of each predictor, we applied the Boruta algorithm (‘Boruta’ package, version 8.0.0). This wrapper method iteratively compares the importance of real variables with shadow variables generated through random permutation, identifying predictors that contribute significantly to the model.

The RF model was implemented with 500 trees, an optimal m = 3 value identified through cross-validation using the ‘caret’ package (version 6.0–94), and default settings for other hyperparameters.

Ethics

This trial was approved by the regional research ethics committee, the Commission Cantonale d’Ethique de la Recherche sur l’être humain (CCER). (Project-ID: 2023 − 00767).

Results

Epidemiology

Of the 673 cases, 464 (68.9%) were managed by the SMUR (Table 1). TCAs represented 0.5 to 0.8% of all SMUR emergency calls, with 30.9 (± 4.2) cases yearly. Blunt trauma was the most common mode of injury (73.5%). The most common mechanism of injury was a fall from a height > 2 m (39.2%), followed by road traffic accidents (24.1%). Head injuries were found in 62.1% of patients, 41.2% experienced haemorrhage, airway obstruction was identified in 11.0%, pneumothorax in 9.1%. Non shockable rhythms were identified in the majority of cases (67.4%) and shockable rhythms were rare (3.7%). Most patients (90.5%) went into cardiac arrest before the arrival of emergency services.

Table 1 Baseline characteristics of patients studied, features and interventions during traumatic cardiac arrest events. (N(%))

SMUR data

Primary outcome

Resuscitation (bystander and/or professional) was attempted in 216 patients (46.5%) and ROSC was obtained in 71 (32.9%) of whom 4 patients had obtained ROSC before the arrival of the SMUR. Of those who obtained a ROSC, 62 (87.3%) survived to hospital admission. (Fig. 1). During the study period, survival to hospital admission ranged from 0 to 32.3% annually. Of the 62 patients who survived to hospital admission, 8 (13%) were alive at 30 days (Fig. 2). Of these patients, 7 (87.5%) had a favorable neurological outcome (CPC score 1 or 2) and one patient (12.5%) had a CPC score of 3. No survivors had a CPC score of 4 or 5.

Fig. 1

TCAs managed by the SMUR and survival to hospital admission

Fig. 2

Kaplan-Meier 30-day survival analysis of patients who survived to hospital admission

Timings

The time from the first call to the EMCC to the SMUR’s arrival was 8 min 40 sec (8.7 min ± 4.4). On-scene time was 34min44 sec (34.7 min ± 18.9), and transport time from the scene to hospital was 6 min 22 sec (6.4 min ± 4.3).

Interventions

A total of 210 (45.3%) patients underwent resuscitation, defined as CPR delivered by medical professionals. Based on SMUR data, resuscitation was more frequently attempted in blunt trauma compared to penetrating trauma (53.7% vs. 25%, p < 0.001). Resuscitation attempts were also more likely among patients aged 65 or less (50.5% vs. 33.6%, p = 0.001). There was no statistically significant difference in resuscitation rate between women and men (42.3% vs. 46.6%, p = 0.426). Of the patients who underwent resuscitation, 168 (80%) had an advanced airway inserted, 77 (36.7%) had some form of chest decompression (Table 2). Concerning pharmacological treatments, 147 (70%) received adrenaline, 169 (80.5%) received crystalloids. In patients with blunt trauma and attempted resuscitation, 76 (41.5%) had a pelvic binder placed.

Table 2 Interventions in patients who underwent resuscitation (n = 210) *among patients with blunt trauma (N = 183)

Vital signs

Most patients (n = 420; 90.5%) went into cardiac arrest before the SMUR’s arrival. Of those who were not in cardiac arrest and had vital signs at the SMUR’s arrival (n = 37), heart rate (n = 32) was 102/min (± 30), systolic blood pressure (n = 22) was 134 mmHg (± 34), respiratory rate (n = 25) was 20/min (± 10), oxygen saturation (n = 17) was 80% (± 20), and the Glasgow Coma Scale (GCS) (n = 34) was 5/15 (± 4).

Asystole

198 patients were in asystole, and resuscitation was attempted in 95 (48%) of these patients. Of these 95 patients in asystole with attempted resuscitation, 16 (16.8%) had ROSC and 14 (14.7%) survived to ED admission. Asystole was associated with decreased odds of ROSC compared to other rhythms (OR: 0.25, 95% CI: 0.12–0.51, p < 0.001). Only 2 patients survived to hospital discharge.

Factors influencing the initiation of resuscitation

Patients who underwent resuscitative efforts were significantly younger than those who did not (p < 0.001). The median age in the resuscitation group was 45 years (IQR: 27–61), compared to 56 years (IQR: 39–71) in the non-resuscitated group. There was no significant association between gender and resuscitation status (χ² = 0.63, p = 0.43), with similar gender distributions observed across both groups. Resuscitation efforts were initiated in 48% of patients presenting with asystole, compared to 43.7% of those with other rhythms. This difference was not statistically significant (χ² = 0.92, p = 0.34). Although the resuscitation attempt rate was higher in the < 18 group (64.3% vs. 45%), this was not statistically significant (χ² = 1.95, p = 0.16).

Factors associated with ROSC

Logistic regression

In patients who underwent resuscitation following a TCA, those who sustained a fall from a height of less than 2 m demonstrated a 13 times higher probability of ROSC, with an odds ratio (OR = 13.2, 95% confidence interval [CI]: 1.7–100.7, p = 0.013). Conversely, patients presenting with hemorrhage exhibited a markedly reduced likelihood of achieving ROSC (OR 0.2, 95% CI: 0.07–0.6, p = 0.005).

Patients with upper airway obstruction demonstrated a more than seven-fold increase in the probability of ROSC (OR 7.1, 95% CI: 2.5–20.4, p < 0.001). SMUR arrival delay was not associated with a probability of ROSC (OR 1.1, 95% CI: 0.95–1.22, p = 0.237). When TCA occurred after EMS arrival, chances of ROSC were increased (OR 6.8, 95% CI: 2.2–21.2, p = 0.001).

When bystander CPR was initiated in patients with a rhythm other than asystole, chances of ROSC were increased eightfold (OR 8.0, 95% CI: 1.2–53.5, p = 0.03).

The total dose of adrenaline administered (mg) was not associated with the likelihood of ROSC (OR 0.85, 95% CI: 0.73–1.002, p = 0.054). In contrast, the administration of crystalloids.

was positively associated with ROSC with, nonetheless, a small OR of 1.002 (95% CI: 1.001–1.003, p < 0.001).

Surprisingly, variables such as age, mode of injury, SMUR arrival delay and the presence of a pneumothorax were not significantly associated with the likelihood of ROSC in the logistic regression model. With regards to age, no single age cutoff in the dataset meaningfully separated patients who achieved ROSC from those who didn’t (Table 3).

Table 3 Elements associated with an increased odds of ROSC in the random forest and logistic regression models

Random forest analysis

To supplement the logistic regression approach and explore potential nonlinear relationships, a random forest (RF) model was applied. This analysis highlighted that the administration of crystalloids, timing of cardiac arrest (prior vs. following arrival of emergency services), type of initial rhythm, mechanism of trauma, upper airway obstruction, mode of trauma and total adrenaline dose as the most influential predictors of ROSC. There were no tentative predictors of ROSC.

A key finding was the divergence in the role of hemorrhage across the two models. While the logistic regression established a significant association, the RF model rejected its predictive importance. The same divergence was found for bystander CPR, which was associated with the likelihood of ROSC in the logistic regression model although this was rejected in the RF model. Other divergent results were found for the mode of injury and adrenaline as they were not significantly associated with the likelihood of ROSC in the linear regression model, although the RF model found it to be an influential predictor. These discrepancies underscore the capacity of RF to capture complex interactions and nonlinear relationships that may not be apparent in parametric models.

The use of crystalloids emerged as a consistent determinant in both models, reinforcing their clinical relevance, as did the timing of cardiac arrest occurrence, mechanism of injury, and upper airway obstruction. Likewise, the identification of the initial rhythm type as a major predictor aligns with the logistic regression results, further supporting its impact on ROSC.

The consistency of certain predictors across both methods strengthens their robustness, while divergences highlight the complementary nature of these analytical approaches in identifying factors influencing ROSC (Table 3).

SMUR and forensic medicine data

Common data

During the 15-year period, a total of 673 TCA cases were identified and included in the study. The average patient age was 52 ± 22 years, with 72.1% of males (Table 1). Relative to population age groups, patients over 65 years old were over-represented and patients under 15 were under-represented (Fig. 3). Blunt trauma was the most common mode of injury (60.6%), followed by penetrating trauma (28.1%) and finally burns (3.1%). In 55 (8.2%) cases, the mode of injury was unknown or undocumented. A fall from > 2 m was the most common mechanism of injury (32.8%, followed by gunshot wounds (22.4%) and finally road traffic accidents in 20.5% of cases. There were no cases of TCAs following falls from standing in patients under 15, but this was the third most common cause of TCA in patients over 65 (Fig. 4.). Regarding specific injuries, the majority of patients (56.8%) sustained head injuries, and 38.6% experienced haemorrhage. Airway obstruction was found in 7.6% of patients. Additionally, 12.4% had either tamponade, pneumothorax, or burns. Finally, nearly one-third of patients (30.9%) presented with other documented injuries.

Fig. 3

Population age groups (% of total population) and traumatic cardiac arrests (% of total cases)

Fig. 4

Mechanism of injury according to age groups

Forensic medicine data

Of the 673 cases, 209 (31.1%) were managed by the forensic medicine service only. Penetrating injuries were the most common mode of injury (46.4%). 94 (45%) patients suffered a head injury and 69 (33%) a hemorrhage, while 3 (1.4%) patients had a tamponade. The proportion of cases seen solely by the forensic medicine service dropped significantly over time, from over 51.1% in 2008 to less than 14.7% in 2022 (p < 0.001) (Fig. 5).

Fig. 5

Case distribution over time

SMUR versus forensic medicine data

The proportion of males was lower in the SMUR group compared to the forensic medicine service group (69.4% vs. 78%, p = 0.016). In the SMUR group, blunt trauma was the most common mode of injury (73.5% vs. 32.1% in the forensic medicine service group, p < 0.001) followed by penetrating trauma (19.8% vs. 46.4%, in the forensic medicine service group p < 0.001).

Pediatric TCA

Of the 673 cases, only 21 (3.1%) were under the age of 18. The mean age was 11 ± 6 years old and the proportion of males was 52.4%. The most common mode of injury was blunt in 13 (61.9%) cases, followed by penetrating in 4 (19%) cases. In 3 cases, the mode was unknown and there was a single case of burns. Mechanism of injury was listed as ‘other’ in 7 (33.3%) cases, followed by fall from height (n = 4;19%), gunshot wound (n = 4; 19%) and RTC (n = 5;23.8%). There was a single case of fall from standing in a 17-year-old. 14 patients were managed by the SMUR while 7 were seen only by the forensic medicine service. Resuscitation was attempted in 9 (64.3%) of these 14 patients, 4 (44.4%) had ROSC and 3 (33.3%) survived to hospital admission. There were no survivors at 30 days.

Discussion

Interpretation

Geneva experienced approximately 45 TCA cases annually for the past 15 years. Over time, the proportion of cases attended by the prehospital medical team has grown, with cases handled solely by the forensic medicine service now becoming the exception rather than the norm (Fig. 5). One possible reason behind this shift may have been due to changes in prehospital care in Geneva during this period. Indeed, in 2014, a medical house call service, SOS Médecins, lost its license to use emergency blue lights on its vehicles [15]. This service was often employed to attend unsalvageable TCA cases to issue death certificates in a timely manner, facilitated by their use of emergency blue lights. However, as this service lost its ability to respond rapidly, the prehospital medical teams were increasingly called upon to assess and manage these cases. This shift may also help explain the observed decrease in survival to hospital admission rates over time. Although the proportion of cases attended by the SMUR increased (despite the overall number of cases remaining stable), the proportion of cases where resuscitation was attempted did not decrease; instead, it showed a slight increase. This may seem counterintuitive, as one might assume that the additional cases handled by SMUR, likely including many unsalvageable ones, would reduce the resuscitation rate. However, the prehospital teams approached these additional cases with the same resuscitation efforts as their typical cases, leading to a stable or even slightly increased resuscitation rate (Fig. 6).

Fig. 6

Proportion of TCA patients with attempted resuscitation over time

With patients in traumatic cardiac arrest, it is essential to address reversible causes simultaneously [16, 17]. The HOTT acronym, introduced in 2013 and representing hypovolemia, oxygenation, tension pneumothorax, and tamponade, highlights reversible causes of TCA that should be promptly and concurrently managed [18, 19]. At the time of the study, prehospital blood products, ultrasound, thoracotomy, or resuscitative endovascular occlusion of the aorta (REBOA) were not available or performed by the SMUR or HEMS and the service did not have a standard operating procedure (SOP) in regards to TCA. In our cohort, of patients who underwent resuscitation by the SMUR, 58 (27.6%) had at least 3 elements from the HOTT acronym addressed. The low compliance with the HOTT concept may be due to multiple aspects, the first being the absence of SOP during the study period. In addition TCA was due to penetrating trauma, many of which were gunshot wound to the head. In these cases, chest decompression is irrelevant. Lastly, in some instances, there might have been a perceived sense of futility that hindered the teams from carrying out invasive procedures; however, this remains hypothetical and is not supported by our data. Concerning drugs, the role of adrenaline in TCA is unclear, however, 70% of patients who underwent resuscitation in our cohort received it. In a recent Japanese retrospective TCA trial, a majority of patients (79.8%) also received adrenaline, and this was associated with decreased 7-day survival (odds ratio = 0.20; 95% CI = 0.08–0.48; P < 0.01) [20]. The negative association between total adrenaline dose and ROSC in our logistic regression model, and its positive association in our RF model does not support or discourage its use in TCA. Placing less emphasis on adrenaline administration and focusing on addressing reversible causes should be the priority in these cases. Almost 80% of patients who underwent resuscitation had an advanced airway inserted. Just over a third (36.7%) of patients who underwent resuscitation had some form of chest decompression, and of those who did, only 6 (11%) had a finger thoracostomy. Needle chest decompression is no longer the recommended approach in TCA, as this technique is associated with a high failure and complication rate [21,22,23]. For the management of cardiac tamponade, resuscitative thoracotomy was not performed by the teams during the study period, and needle pericardiocentesis was very rarely performed (< 5 cases over the 15-year observation period).

Despite a rapid response time averaging 8 min 40 s from the initial call to the SMUR’s arrival, over 90% of patients had already suffered traumatic cardiac arrest by the time the team reached them. The rapid onset of TCA highlights the severity of these cases. In other cohorts, the time between the incident and TCA was slightly longer, such as the 12-minute interval reported in a study by the London Air Ambulance [24]. This severely limits the potential for a favourable outcome improvement, even with enhanced adherence to guidelines, more aggressive management of reversible causes, and the introduction of previously unavailable techniques such as prehospital thoracotomy or prehospital transfusion. Data from the London Air Ambulance demonstrated dismal outcomes for patients who went into cardiac arrest due to exsanguination, a common cause of TCA in our predominantly blunt-trauma cohort, showing no recorded survivors when resuscitative thoracotomy was performed more than 5 min after TCA in these patients [25].

The rate of cardiac tamponade in the SMUR cohort was 0.6%, though this may be underestimated due to the lack of ultrasound availability at the time and the fact that the teams did not perform resuscitative thoracotomies. The rate of tamponades was higher in the forensic medicine database (1.4%), further supporting the possibility of prehospital underdiagnosis. Even though tamponades in blunt trauma (which represents 73.5% of patients in the SMUR cohort) are rare, they may still occur, sometimes even in isolation in patients with chest trauma [26].

Penetrating trauma

Among the 464 SMUR cases, 92 patients (19.8%) had penetrating trauma, which is higher than similar European studies [27]. Penetrating trauma was split between gunshot wounds (75%) and stabbings (25%). Resuscitation was attempted in only 23 (25%) of cases; ROSC and subsequent survival to hospital admission was achieved in 7 of these cases (30.4%). The resuscitation attempt rate of penetrating trauma was almost half the resuscitation rate we observed in the SMUR cohort. This data was explored and it was found that of all the patients with firearm injuries, 124/151 (82.1%) were gunshot wounds to the head, most being presumably self-inflicted. Switzerland has a suicide rate that is slightly below the European average, however it stands out in terms of suicide methods: firearms are used in 33.6% of male suicides, compared to just 9.7% across the rest of Europe. In fact, Switzerland is the only country in Europe where firearm injuries are the most common method of suicide in males [28]. This highlights that a term such as “penetrating trauma” can represent vastly different realities across various systems. While penetrating trauma may predict better outcomes in one context, it can be fatal in another. In our system, penetrating trauma was associated with lower resuscitation rates and poor outcomes, largely due to the overwhelming number of self-inflicted gunshot wounds to the head. In this context, there is only so much prehospital teams can do, underscoring the importance of prevention efforts and stronger gun control policies [29].

Asystole

Only 2 patients with asystole survived to hospital discharge. In both cases, the asystolic cardiac arrest happened after the SMUR’s arrival, during intubation due to hypoxia, in the first case, and a combination of hypoxia, hypovolemia and tension pneumothorax in the second case. There were no survivors when asystole was the first presenting rhythm. This further reinforces the well-established association between asystole, particularly when it is the initial presenting rhythm in traumatic cardiac arrest, and extremely high mortality rates. This consideration should play a critical role in guiding decisions regarding the continuation of resuscitative efforts.

Strengths

A key strength of this study is that, unlike many prehospital trials on traumatic cardiac arrest, we also included cases seen exclusively by the forensic medicine service. Trauma should be regarded as a disease, and for a comprehensive understanding of TCA epidemiology, it is essential to consider all cases, not only those attended by emergency services.

Limitations

This study has several limitations. The first is the retrospective nature of the study which makes the interpretation of any associations difficult, as they may be subjected to bias, including survival bias [30]. The other limitation is that the database does not include data from HEMS. This limitation is mitigated by the fact that these missed cases may have been included in the forensic medicine service database. In addition, the HEMS TCA cases for 2022 were reviewed and only a single case was found.

Another limitation is missing data from the SMUR reports. The unknown or undocumented data included the mode of injury (unknown in 2.6% of cases), the mechanism of injury (undocumented in 0.9% of cases), types of injuries (undocumented in 6.5% of cases), and the initial rhythm (undocumented in 29.1% of cases). Rhythm checks are not required for all TCA patients if other clear signs of death are present, and even experienced clinicians often struggle to accurately identify injuries during prehospital patient assessments [31]. It is unclear whether the missing data reflects cases with high complexity and intense workload or, alternatively, cases where the attending teams considered further investigation futile due to the perceived hopelessness of the situation.

The epidemiology we present provides a snapshot of incidents from 2008 to 2022. Applying this data to address current or future TCA cases may lose relevance if the epidemiology changes, as has occurred in other countries following legislative shifts, such as revisions to gun laws [32].

Conclusion

Between 2008 and 2022 in Geneva, traumatic cardiac arrests (TCAs) occurred at a rate of nearly 45 cases per year. In recent years, the prehospital medical team (SMUR) became increasingly involved in these incidents, while the forensic medicine service managed fewer than 15% of cases independently toward the end of the study period. The most common cause of trauma was blunt-force injury. Although penetrating trauma made up 28.1% of cases, these were predominantly self-inflicted gunshot wounds to the head and were associated with poor prognoses. In over 90% of cases, TCA occurred before the arrival of the prehospital team. Resuscitation efforts were initiated in just under half of these patients; among them, 28.7% survived to hospital admission, and 3.7% remained alive at 30 days. Factors independently associated with ROSC included crystalloid administration, timing of TCA, mechanism of injury, initial cardiac rhythm, and the presence of upper airway obstruction. Enhancing adherence to TCA management guidelines, particularly those emphasizing early treatment of reversible causes may lead to better patient outcomes.