Outcome and predictors for successful resuscitation in the emergency room of adult patients in traumatic cardiorespiratory arrest

The TraumaRegister DGU® (TR-DGU) of the German Trauma Society (Deutsche Gesellschaft für Unfallchirurgie, DGU) was founded in 1993 [23]. The aim of this multi-centre database is an anonymous and standardised documentation of severely injured patients for the purpose of quality control. Data are collected prospectively at four consecutive time periods from the site of the accident until discharge from hospital: (A) prehospital phase, (B) emergency room and initial surgery, (C) intensive care unit and (D) at discharge. This documentation records detailed information on demographics, injury pattern, comorbidities, pre- and in-hospital management, course on the intensive care unit, relevant laboratory findings including data on transfusions, and each individual's outcome. The inclusion criterion is admission to hospital via the emergency room with subsequent intensive care unit (ICU) care. Patients who reach the hospital with vital signs and die before admission to the ICU are included as well.

The infrastructure for documentation, data management, and data analysis is provided by the Academy for Trauma Surgery (AUC - Akademie der Unfallchirurgie GmbH), a company affiliated with the German Trauma Society. The scientific leadership is provided by the Committee on Emergency Medicine, Intensive Care and Trauma Management (Sektion NIS) of the German Trauma Society. The participating hospitals submit their data to a central database via a web-based application. Scientific data analysis is approved according to a peer-review procedure established by the Sektion NIS.

Most of the participating hospitals are located in Germany (90 %), but a rising number of hospitals in other countries have been contributing data as well (these include currently Austria, Belgium, China, Finland, Luxemburg, Slovenia, Switzerland, the Netherlands, and United Arab Emirates). About 35,000 cases from over 600 hospitals are currently being entered into the database per year. Participation in the TR-DGU is voluntary. For certified hospitals associated with the TraumaNetzwerk DGU®, however, participation is obligatory for reasons of quality assurance.

Data anonymity for scientific analyses is guaranteed for both the individual patient and participating hospital. [23‐25].

The present study is in line with the publication guidelines of the TraumaRegister DGU® and registered as a TR-DGU project ID 2014-023.

We analyzed the 2013 database from 2002 to 2013. Primary admitted adult patients (age ≥16) from Europe with Injury Severity Score (ISS) ≥16 points and available information on CPR (done, or not) both in the prehospital phase and during emergency room (ER) treatment qualified for analysis. Patients transferred in from other hospitals were excluded since prehospital information was missing. Patients declared dead at the accident scene and not transported to a hospital were not recorded in the TR-DGU. The main focus of this survey was the group of patients in cardiac arrest resuscitated in the ER and in the prehospital phase.

A patient’s death can be pronounced outside a hospital and/or clinic setting in Germany, thus the patient is not transported to a medical setting to be declared dead, as happens in some other countries. This requires the diagnosis of at least one definitive sign of death. Patients without return of spontaneous circulation (ROSC) after cardiac arrest were not included in the TR-DGU, according to our inclusion criteria for the registry. This holds true also for patients admitted to the hospital with ongoing CPR but without subsequent ROSC. The registry was established for quality assessment in the acute care hospital and thus did and does not document prehospital deaths.

We performed a descriptive analysis of patients with and without resuscitation in the ER. Then a multivariate logistic regression analysis with resuscitation in the ER as a dependent variable was done to identify independent predictors for survival by calculating odds ratios for each factor. Odds ratios are presented with 95 % confidence intervals (95 % CI). Nagelkerke’s R² was used as an overall goodness-of-fit measure for the model.

The Glasgow Outcome Scale (GOS) was used to classify patients with good or moderate (GOS 4–5) and dead/bad outcome (GOS 1–3) [26]. Furthermore, we carried out a multivariate logistic regression analysis with bad outcome as a dependent variable. Moreover all parameters for the Trauma-Associated Severe Haemorrhage (TASH) score (a score that reliably predicts the probability for mass transfusion after multiple trauma) [27] and coagulopathy (presence of abnormal coagulation parameters upon the patient’s arrival); i.e. prothrombin time test, Quick’s value <70 % and/or platelets <100,000/ml) [28] and acidosis (base excess ≤ -6) were analysed.

A total of 1755 patients died after CPR in the ER; there was a mean time to death of 2.3 days (SD 5.8; median 1). Of these 1558 (73.9 % of all, 88.6 % of those who died) died within the first 24 hours after admission. The 350 surviving patients stayed on the ICU for 20.8 days (SD 10.0, median 16.5); their hospital stay was 42.0 days (SD 34.2; median 33.5).

Results of multivariate logistic regression analysis with resuscitation in the ER as a dependent variable (Nagelkerke’s R² = 0.30) are shown in Table 2.

The highest risk for mortality in this model was associated with preclinical resuscitation (OR 5.74), followed by shock on admission (OR 3.49), coagulopathy (OR 2.18), preclinical thorax drainage (OR 1.54), preclinical catecholamines (OR 1.52), mild hypotension with preclinical blood pressure ≤90mmHG (OR 1.35), preclinical GCS ≤8 (OR 1.28), relevant injuries of the extremities (OR 1.27), the thorax (OR 1.26), and the abdomen (OR 1.23). The overall injury severity is also associated with in-hospital CPR.

Results of the Glasgow Outcome Scale (GOS) [26] after resuscitation in the preclinical phase and/or in the ER are based on data from 3052 patients. Table 3 displays the Glasgow Outcome Scale (GOS) groups of resuscitated patients.

We defined a good or moderate outcome as GOS 5 and 4 and this was analysed in patients without any resuscitation in 68.5. After only preclinical resuscitation the survival rate was with 31.7 % the highest and 14.7 % had a good or moderate outcome (46 % among the survivors). After resuscitation only in the ER, 25.6 % survived in total and 19.2 % had a GOS of 4 or 5 (75.0 % among the survivors). In total 4.8 % in the group with preclinical and ER resuscitation survived and only 2.7 % revealed a good or moderate outcome (56 % among the survivors).

Furthermore, we cross-tabulated the prognostic factors for patients with good and moderate outcome (GOS 4 + 5) after resuscitation in the ER, as illustrated in Table 4:

Results of multivariate logistic regression analysis with bad outcome (GOS 1–3) as a dependent variable in patients with preclinical resuscitation (Nagelkerke’s R² = 0.28) are shown with n = 841 in Table 5:

The 25.6 % survivor rate after resuscitation in the ER we calculated is much higher than the 3.3 % reported in a systematic review containing a meta-analysis reporting a total of 238 survivors out of 5391 patients [13]; our study’s survivor rate is also much higher than that in a retrospective analysis reporting 5.4 % of 1114 patients surviving to hospital discharge after out-of-hospital traumatic cardiopulmonary arrest [33]. The results are similar to the 29 % survivor rate in a prospective observation study from Berlin/Germany [21].

In another study, 5.0 % of the patients underwent resuscitation during trauma room (TR) treatment [14]. In that study, the neurological outcome after only resuscitation in the ER was better than after preclinical resuscitation and after preclinical and ER resuscitation. One reason for this finding could be an optimal setting with an interdisciplinary team of specialists vs. a smaller preclinical team with normally only one physician. Another reason might be that cardiac arrest was witnessed in the ER setting in conjunction with resuscitation being initiated immediately, thus keeping the “downtime” minimal.

On-scene triage is done by physicians and is based on their clinical assessment of the patient. These physicians also initiate the early treatment of haemorrhagic shock, provide respiratory assistance, manage pain and sedation, and insert chest tubes if necessary and can also use catecholamines (for sufficient cerebral perfusion, neurogenic shock, resuscitation, etc.). The results from the initial assessment of the patient's condition and those together with the patient’s response to treatment are used by the dispatching physician to find the closest and best-suited facility for the patient. This organisation ensures the availability of resources at the receiving centre and avoids unnecessary secondary transfers from an overcrowded or poorly suited institution (accessed initially because of its proximity to the trauma scene). The cohort demographics including injury patterns are typical for western countries.

Our data fail to support data from earlier investigations by Fulton et al., who emphasised that the location of arrest (on scene, transport or emergency room) and age did not correlate with survival [34]. Our cohort’s survival rates were different after preclinical resuscitation (31.7 %), resuscitation only in the ER (25.6 %), and after preclinical and ER resuscitation (4.8 %). A limitation of the study is that the patients who were dead at the scene were not included in the investigation, because these data are not available.

Our logistic regression model revealed that preclinical resuscitation followed by ISS, preclinical catecholamines, preclinical thorax drainage, coagulopathy, preclinical GCS ≤8 points, AIS thorax, abdomen and extremities ≥3 points, mild hypotension with ≤90 mmHg during the preclinical or ER treatment are strongly predictive for mortality after resuscitation in the ER, in line with findings by Pickens et al., who reported that respiratory efforts and a GCS ≥4 points at the scene have a positive influence on survival, and Durham et al., who reported the same regarding prehospital intubation [20, 35]. Fulton et al. also noted that a loss of neurological function is associated with mortality [34].

The percentage of thoracic trauma in polytrauma patients is generally quite high, ranging from 34 to 70 % [29, 36‐39]. Huber-Wagner et al. observed that prehospital chest tube insertion was a strong predictor for survival. Also tension pneumothorax is known as the most common reason for definitive preventable traumatic death [17].

They therefore advised on-scene chest decompression of TCRA patients in case of the decision to initiate emergency cardiac care (ECC) [40].

In the European Resuscitation Council Guidelines for Resuscitation 2015, bilateral chest decompression is advised in the algorithm to address one of the reversible causes for resuscitation [41].

A study by Pickens et al. reported that several survivors met criteria for non-treatment according to proposed clinical guidelines, and that insufficient prehospital treatment was the result [20]. Katz et al., for example, analysed a 25 % rate of misplaced tracheal tubes by paramedics (27/108) [42]. Furthermore, Cera et al. found that the time of intubation (on-scene versus in-hospital) did not influence survival [1].

Normally, patients in a clinical setting should never be deprived of sufficient oxygen for a relevant period of time, unlike what can happen in a preclinical setting, where patients probably suffer several minutes without oxygenation before the medical staff reaches the scene. The fact that a patient requires resuscitation in a preclinical and ER setting seems to reflect the trauma’s severity and according to the literature often appears to be associated with a poor outcome.

In our investigation among 31.7 % survivors in the group of preclinical resuscitation 46 % had a good or moderate neurological outcome; after resuscitation in the ER 25.6 % survived, with 19.2 % having a GOS 4 or 5.

GOS has been documented in the TR-DGU since 2002, thus there are no GOS records for patients from 1993 to 2001. Our GOS calculation is therefore based on 38499 patients.

Reliable data describing outcomes in detail and including neurological function are rare in the literature. A recently published literature review covering 1980 to 2011 analysed the outcomes of children treated for out-of-hospital traumatic cardiopulmonary arrest who reached the clinic alive: 29 patients (57 %) were severely disabled; 3 children (6 %) moderately disabled, and 19 patients (37 %) had a normal outcome [43]. Margolin et al provide information on the GOS in six of 13 survivors (46.2 %) out of 67 patients (79.1 % blunt trauma) who had at least a satisfactory outcome as expressed by a GOS of ≥4 [32]. Other studies describe patients’ outcome data in terms of their being “neurologically intact” after TCRA: Fialka et al. observed 100 % of patients surviving with no neurological deficit (4/4, n = 38) [44], Fulton et al. 83.3 % of survivors being neurologically intact (5/6, n = 245) [34], Powell et al. 81.0 % of survivors (21/26, n = 959) [11], Stratton et al. 66.6 % (3/9, n = 79) [8], Battistella et al. 56.3 % (9/16, n = 604) [6], Branney et al. 3.9 % (34/42, n = 950) [45], and Martin et al. no neurologically intact patients (0/1, n = 110) [12].

Yanagawa et al. detected no significant differences between out-of-hospital TCRA patients with or without head injury except for the fact that the blunt traumatic patients with a head injury more often displayed a return to spontaneous circulation (ROSC). However, they observed no head-injury effect on the outcome [3]. In our investigation, a preclinical GCS ≤8 had a significant odds ratio of 2.18 in the multivariate logistic regression analysis with resuscitation in the ER as a dependent variable. Moreover, an AIS head ≥3 was a significant prognostic factor in our cross-tabulation, with a GOS 4 and 5 and resuscitation in the ER and preclinical GCS ≤8 showing a significant odds ratio of 3.886 in multivariate logistic regression analysis with GOS 1–3 as a dependent variable.

Our multivariate logistic regression revealed that injury severity, the preclinical use of catecholamines and preclinical thorax drainage, coagulopathy, a preclinical GCS ≤8, severe injuries to the thorax, abdomen and extremities, and preclinical shock in the ER are strongly predictive of resuscitation in the ER.

Sorted according to a bad prognosis with GOS 1–3, these factors were preclinical resuscitation, ISS, preclinical GCS ≤8, mild hypotension in the ER with Riva Rocchi (RR) ≤90mmHG, blood transfusion and catecholamines, and age ≥60 years.

Base deficit is a rapidly and widely available serum laboratory marker of systemic acidosis that increases with hypoxemia and/or shock. In the trauma setting, the base deficit correlates with the blood-transfusion requirement, risks of multiorgan failure and mortality in adult patients, including those with traumatic brain injury [31‐33].

These findings are in line with the sparse data in the literature reporting that blood pressure = 0 [34] or low blood pressure in the field were prognostic factors [35].

Our study has several limitations. Patients declared dead on-scene without transportation to a hospital were not recorded in the database. Moreover, only patients surviving CPR and transportation to the hospital could be included. That means that our database does not include those patients who required prehospital CPR but did not achieve ROSC. This is because the TraumaRegister DGU® was established as a tool for quality assessment in the acute care hospital. Patients who died before admission were thus not relevant to such an assessment. Furthermore, Germany maintains no documentation system that covers all trauma cases, including all prehospital deaths. Our survival rates thus do not refer to all cases with cardiac arrest after trauma but only to those cases who were admitted to the acute care hospital alive, i.e., with ROSC. This limitation and bias must be kept in mind when interpreting the surprisingly good survival rates.

This and other exclusion criteria (ISS <16 and secondarily transferred patients) were excluded. This could have biased the results.

Due to the fact that ECG data are not documented in the TR-DGU®, no information on cardiac rhythm was available. Furthermore, we had no specific information on the exact duration of chest compression in the pre- and/or in-hospital phase. According to the literature, resuscitation lasting more than 20 minutes seems to be associated with poor neurological outcome [33]. We could only analyse patients who underwent closed-chest compression on-scene or during transport and/or during treatment in the ED. The main limitation is our patient cohort’s preselection. The TR-DGU® only includes patients who have been transported to a hospital after chest compression after TCRA. This might be one reason for the relatively high survival rates compared to other studies. The group investigated in our study represents patients who had been resuscitated because of a subjective assessment by the on-scene emergency physician who perceived a relatively certain chance of survival for those patients.