| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Eur J Cardiothorac Surg 2004;26:377-385
© 2004 Elsevier Science NL
Department of Cardiothoracic Surgery, St Mary's Hospital, 70 St Olafs Road, Fulham, London SW6 7DN, UK
Received 11 October 2003; received in revised form 17 January 2004; accepted 4 March 2004.
* Corresponding author
e-mail: tathan5253{at}aol.com
 |
Abstract |
|---|
 Top Abstract 1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
Key Words: Thoracic trauma • Chest injury • Emergency thoracotomy
|
1. Background |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
ET is a procedure often carried out in patients presenting in extremis following severe thoracic trauma. Because of the severe nature of thoracic trauma, some of the patients that undergo ET have already sustained severe pathophysiological derangement, making them unsalvageable. In this group of patients ET is of limited value as it unnecessarily compromises personal dignity. There are also the issues of wasting valuable health care resources and the significant risk of injury to both hospital staff (needle stick injuries) and paramedic staff (accidents in accessing and transferring patients) [2–4]. In order to help physicians decide when to withhold or terminate further resuscitation attempts in a pre-hospital traumatic cardiopulmonary arrest, the National Association of Emergency Medical Service physicians and the American College of Surgeons Committee on Trauma have recently issued a set of guidelines [5].
Few studies have been performed in an aero-medical transport population [6,7]. The indications for performing a thoracotomy for severe thoracic injury especially in the pre-hospital setting have not clearly defined. There are limited reports on the role of potential independent pre-hospital predictors on mortality [8,9].
The aim of this study is to investigate the influence of ET on mortality in a group of patients suffering from severe thoracic trauma requiring Helicopter Emergency Medical Service (HEMS) transfer to hospital. This is not clearly defined in the literature especially when thoracotomy takes place in the pre-hospital setting. Several factors (time related variables, physiological factors, on-scene interventions, patient characteristics and procedure related factors) can influence efficacy of the emergency thoracotomy procedure.
|
2. Material and methods |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
Emergency thoracotomy in our study was defined as the operative procedure performed as an integral part of pre-hospital cardiopulmonary arrest management. This is often done simultaneously with other manoeuvres such as airway control and volume restoration. In the trauma system of the London HEMS there is a policy to perform an on-scene pre-hospital thoracotomy to a pulseless patient after penetrating thoracic injury especially if the nearest trauma center is more than 10 min away (measured from loss of pulse to surgical intervention). There are no specific guidelines for the pre-hospital physician when the patient is pulseless for a period more than 10 min. In our study pre-hospital thoracotomy was performed according to the London HEMS algorithm, by highly trained medical personnel according to ATLS guidelines (Fig. 1) .
|
In this group of 670 patients the effect of ET on mortality was evaluated using both uni- and multivariate analysis. Finally subgroup analysis was performed on the ET group of 53 patients, to evaluate the effect of several factors on mortality (for example, the absence of signs of life on arrival, the setting in which the thoracotomy was performed, and the type of injury).
2.1. Characteristics of the London Helicopter Emergency Medical Service
Listed below, are the important features of the HEMS thoracic trauma system:
2.2. Data collection
The anatomical injuries are measured according to the Abbreviated Injury Scale (AIS 6–1, highest fatal, lowest minor) [10,11]. The six scales are calculated in each of nine anatomical services: (1) head; (2) face; (3) neck; (4) thorax; (5) abdomen and pelvic contents; (6) spine and vertebrae; (7) upper extremity; (8) bony pelvis and lower extremity; (9) external including burns. Such injuries can also be coded according to the International Classification of Diseases (ICD-9). The original nine AIS subgroups were then placed into six body regions and the highest score was taken for each of these anatomical groups. The sum of the squares of the three highest scores were added together to calculate the Injury Severity Score (ISS) and put in order from 75 to 0 (highest usually fatal, lowest less so). Finally, the Trauma Injury Severity Score was used in order to calculate the probability of survival according to the pathophysiological data from the patient.
This data included the Cardiovascular-Respiratory status (CVRS) and the Glasgow Coma Scale. The CVRS reflects individual elements of blood pressure, respiratory rate, heart rate and saturation on scene and on A&E department. In order to monitor these parameters, HEMS uses the Propac 106EL device (Protocol Systems, Inc., Beaverton, OR). It has a minimum battery life of 4 h and has electrocardiogram, non-invasive blood pressure, oxygen saturation , end-tidal CO2 and temperature facilities.
2.3. Decision algorithm used by HEMS for pre-hospital resuscitative thoracotomy for cardiac arrest following penetrating trauma
During the period 1991–1992 from a total of 1084 HEMS patients, 16 underwent ET (nine pre-hospital, seven A&E). The mean ISS was 30 and the mean time spent on the scene of injury attending patients who underwent ET (35 min) was significantly longer in comparison to those in whom the ET was performed in A/E (16 min). None of the patients survived with either ‘scoop and run’ or ‘stay and play’ strategies demonstrating the need for a more selective ET decision algorithm [12]. Since 1993 HEMS introduced a decision algorithm (Fig. 1) for management of penetrating thoracic trauma [13].
2.4. Indications and surgical technique of emergency thoracotomy
The objectives of ET are to release pericardial tamponade, to prevent air embolism, to control massive intrathoracic or intraabdominal haemorrhage, and to provide access for open cardiac massage and descending thoracic aortic cross-clamping.
The airway was established by either endotracheal intubation or cricothyrotomy and vascular access gained, always prior to ET. The initial incision for an emergency thoracotomy was a left antero-lateral approach just below the nipples in males or just beneath the inframammary fold in females (usually in the fifth intercostal space). The muscle and soft tissue was divided with heavy scissors along the superior rib margin taking care upon entering the thorax not to lacerate the lung. The pericardium was opened taking care not to injure the phrenic nerve or the coronary arteries. If pericardial tamponade was noted a cardiac wound was sought and bleeding controlled by digital pressure and/or insertion of a Foley catheter, staples or sutures.
The equipment pack used for ET has been designed to be used by non-cardiothoracic surgeons and consists of a Finochietto rib spreader, a large scalpel, ‘Tuffcut’ scissors, Spencer–Wells forceps and a silk suture on a round-bodied needle. A back-up pack of cardiothoracic instruments is also available in case of need.
|
3. Statistical analysis |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
We performed univariate analysis in order to identify potential predictors of mortality (categorical variables that differed significantly among survivors and non-survivors) and we calculated the unadjusted odds ratio, and a 95% confidence interval for the odds ratio. If the value 1 was not in the range of the confidence interval, it was concluded that there was an increased relative risk in one group compared to the other. Those variables proving to be significant were included in the logistic regression multivariate analysis. All statistical analysis was performed using the SPSS 11.0 for Windows software package (SPSS, Inc., Chicago, IL, USA).
|
4. Results |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
4.1. Uni- and multivariate analysis
Twenty-nine variables including demographics, time related variables, type of injury, injury related variables, associated injuries and pathophysiological data were included in a univariate analysis. Sixteen variables were identified to be potential predictors of survival. Results of this analysis are presented in Table 1. All these variables were included in the logistic regression model and the independent predictors of mortality are presented in Table 2.
|
|
The following variables were identified as independent predictors of mortality: Age>60 years (OR 5.57, 95% CI 2.19–14.16), GCS<8 at the scene (OR=7.4, 95% CI=3.15–17.46), ISS>25(OR 5.3, 95% CI=1.64–17.11), need for intubation at the scene (OR=2.80, 95% CI=1.022–7.69), saturation on A/E (<89%) (OR=2.39, 95% CI=1.13–5.05) and haemothorax (OR=3.30, 95% CI=1.53–7.13) or bilateral injury (OR=3.1, 95% CI=1.51–6.61).
|
5. Emergency thoracotomy—subgroup analysis |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
The incidence of blunt injury in the ET group was almost similar (27–51%) in comparison to the incidence of penetrating injury (26–49%). Also, 25 patients (47.1%) had no signs of life on arrival on the scene and 28 patients (52.8%) had no signs of life on arrival in the A&E.
The most common abdominal injury involved the liver or the spleen, seen in seven patients (13.2%). Bilateral chest thoracic injury was observed in seven patients (13.2%), fracture of the ribs in 23 (43%), lung injury in 19 (36%), and heart injury in 16 (30%). Flail chest seen in 4 (7.5%), sternal fracture in 2 (3.7%) and injury of the diaphragm in 6 (1.3%). Vascular injury involving the inferior vena cava was found in patients (11.3%), the descending aorta in five patients (9.4%) and the superior vena cava in 2 (3.7%). In the ET group there were commonly other orthopaedic injuries found, mainly affecting the limbs, scapula, and clavicle (23 patients, 43.4%). Neurological injuries including head injuries occurred in nine patients (16.9%) and spinal injuries in 7 (13.2%).
The pathophysiological status and the description of the associated injuries in patients undergoing ET are presented in Table 3. Ten patients survived in the ET group (18.8%). Graphical presentation of the mortality following ET between 1994 and 1995 is presented in Fig. 2 . The distribution of factors that can affect outcome following ET is shown in Table 4.
|
|
|
The indications for emergency thoracotomy were the following: suspicion of obstructive cardiac tamponade in patients 36 patients (68%), control of massive bleeding 14 (26%) and significant air leak in three patients (6%). In all the pre-hospital thoracotomy cases the indications recorded were suspicion of obstructive cardiac tamponade and need for more effective internal cardiac massage which was performed in 20 cases as mentioned earlier.
Most of the ET (37–70%) was performed by non-surgeons (Anaesthetists and A&E doctors). Five of them survived (13.5%), which was not significantly different to the survival rate when the thoracotomy performed by surgeons 5/16 (31.2%) (P=0.14).
The majority of ET procedures were performed by specialist registrars (79%) in the later years of their training. When the procedure was performed by specialist registrar the survival was 6/42 (14%) was not significantly different (P=0.18) compared to those performed by consultants 4/11 (36%) (Table 4).
Univariate logistic regression was performed between ET mortality and the ISS, Arrival on scene time (AOS), Stay on scene time (SOS). There was a significant association between mortality and ISS (OR=1.07,95% CI=1.01–1.14, P=0.01), but not with the time-related variables AOS (OR=1.08,95% CI=0.88–1.32, P=0.43), and SOS (OR=0.96,95% CI=0.90–1.03).
When we compared the physiological variables on scene to those in A&E in the ET group with the Wilcoxon Signed Ranks test, it was noted that there was a significant reduction of the heart rate (z=–3.7, P<0.005), respiratory rate (z=–2.3, P=0.01) and GCS (z=–4.17, P<0.005) but not in the systolic blood pressure (z=–0.7, P=0.4) and oxygen saturation (z=–1.6, P=0.09).
Linear regression analysis did not show any significant association between age and ISS in the ET group (beta coefficient=–0.21, standard error=0.1, P=0.14).
|
6. Discussion |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
This study has demonstrated that using multivariate analysis of our data, ET is no longer an independent predictor of mortality in patients with major chest trauma following adjustment for confounding pre-hospital factors, and should therefore be considered in their management algorithm. We have identified that the combination of neurological deterioration (GCS<8 on the scene), increased age (>60), and extensive injury with significant blood loss (ISS>25, bilateral injury, haemothorax) correlates with very poor survival.
There have been dynamic changes in trauma characteristics and management over the last four decades, with an increase in incidence and severity of chest trauma leading to the development of trauma delivery care in the pre- and hospital level. Despite this, the role of ET as an emergency life-saving procedure remains unclear. Prior to the late 1960s, attitudes regarding thoracotomy were conservative, and as a result this was only carried out in the operating theatre. Since then, the concept of ET has been advocated for certain patients with penetrating thoracic trauma [16], and extended to patients with either penetrating trauma outside the thorax, or blunt thoracic injury. The latest part of this progression is the ability to perform ET in the pre-hospital phase advocated by the HEMS system at the Royal London Hospital. The clear aim of the Royal London Algorithm (Fig. 1) is to target patients with obstructive cardiac tamponade following penetrating injury. This specific pathology can be treated on the scene. Other more complex injuries associated with significant blood loss can be treated more effectively in a hospital setting.
It is noteworthy that at present there is no research available on survival after pre-hospital ET in a trauma system where the surgeon has a specialist cardiothoracic training.
Our study reveals a trend of improved survival in patients undergoing ET in hospital as compared to that performed on the scene (P=0.07), a finding that has been previously reported [8]. It is also important to note that we, as in a previous prospective study by Asensio et al. [8], did not identify the time-related variables as significant predictors of survival. It is encouraging that the survival rates reported in this study (16%) were very similar to our results (18.8%), and whereas our patient group included both blunt and penetrating trauma, Asensio et al. only looked at patients with penetrating cardiac injuries.
In our study, the median time taken to arrive on the scene was 8 min and the median time spent on-scene was 17 min. This is compatible with a recent report emphasizing the importance of performing ET within the first 30 min of thoracic injury [17]. The combination of quick delivery of on-the-scene care, and implementation of the HEMS decision algorithm (derived in 1993), has improved survival from 0% (1991–1992) to 18.8% (1993–2002) [18]. We feel that time spent on-the-scene should not be considered as being ‘wasted’, but instead should be ‘maximized’ to provide the right care at the right time. This is particularly relevant to the ET procedure.
Several studies have emphasized the ‘dismal’ effect of patients with ‘absence of signs of life’ on survival. This can be a confounding factor on the survival rates, and result in heterogeneity between studies [19]. Our study demonstrated a statistically significantly improved survival when out-of-hospital ET was performed in patients with penetrating rather than blunt trauma, and in patients without absence of signs of life on arrival. It is worth mentioning that half had signs of life on arrival of HEMS on the scene. The 1:1 ratio of blunt versus penetrating injury and 1:1 ratio of patients with and without signs of life on arrival of the HEMS shows that ET was over performed in this group of patients and better selection of candidates for this procedure could have improved its efficacy. Also the fact that the majority of them performed on the scene shows that there is a significant impact on mortality when the doctor on the scene decides to perform the procedure based on subjective rather than objective criteria. The data shown in Table 4 suggests that the group of patients with penetrating trauma involving only the thorax (P=0.03) and with signs of life on arrival (P=0.01) represents the target group that ET is most likely to benefit.
Pathophysiological variables often change significantly during transport. We found that a decrease in the heart rate, respiratory rate, and GCS occurs more often in patients who have required ET. Our findings are in accordance with a previous report which has shown that initial abnormal electrocardiographic rhythm (bradycardia and asystole) [5], is associated with a worse survival outcome. With regards to respiratory support, on the scene intubation is an effective way of maintaining respiratory function and thus increasing cerebral oxygen delivery, translating to a longer tolerance of ischaemia during the pre-hospital arrest period. Our study did not show a significant effect on survival by intubating on-the-scene (P=0.49). Finally aggressive volume restoration, use of inotropic drugs, intubation and positive ventilation, and hypothermia can all bias blood pressure measurements, making them unreliable predictors of outcome. Although invasive and more accurate blood pressure monitoring would be ideal, this is practically very difficult to achieve in the pre-hospital setting or during transfer.
A significant predictor of adverse early and late outcome in major trauma is the delay in diagnosis and intervention to manage potentially lethal but reversible conditions [20]. These delays and errors are initiated at the scene of the incident and can be continued in the operating theatre. A rational decision may be the immediate delivery of ET in selected patients (‘stay and play’) with an immediate transfer (‘scoop and swoop’) strategy in the rest. In order to make this decision, well-defined guidelines for ET in the management of trauma are needed. The two management strategies can be used in a complementary fashion.
In a review study of 4620 cases from institutions reporting ET for both blunt and penetrating trauma, it was demonstrated that the overall survival rates ranged from 1.8 to 27.5% [19]. This range in survival rates occurs for several reasons:
Although it is obvious that further research into this field is required, major thoracic trauma (ISS>15) occurs infrequently, and therefore there are very few prospective studies showing the outcomes of these patients. Only a prospective randomized controlled trial would be able to prove the efficacy of the ET. As previously mentioned, a review of data from 24 studies including 4620 ET-patients cases, demonstrated an overall survival of 7.4% [19]. In patients who did not undergo ET (control), the probability of survival was extremely low (approximately 1:1000). A traditional randomized controlled trial with 1 to 1 ratio (ET: Control) and a 5% significance level with 80% power would require 132 patients in each arm. A trial like this is not only impractical, involving significant ethical implications, but would also require decades to be completed.
Finally, a limitation of our study is that the sample for the ET-subgroup analysis is relatively small, meaning that every conclusion should be treated with caution. There is also the issue of incomplete documentation of pre-hospital interventions and therefore a potential underestimation of their impact on patient mortality. Further research is required to clarify the role of ET on outcome especially in the pre-hospital setting.
|
7. Conclusion |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
|
Footnotes |
|---|
|
Appendix A. Conference discussion |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
Dr Athanasiou: In this group of patients, median sternotomy was not performed in any of these patients. Of course your question is very interesting because, as we know, in bilateral trauma the use of a median sternotomy is a valid option. Especially when you have severe cardiac trauma, in order to control the bleeding, the first thing to do is to have access through a sternotomy because you can use bypass to fix the injury.
Dr O. Kshivets (Siauliai, Lithuania): As I understood you, you didn't find any predictors of mortality. The first question is, did you use only multivariate analysis? Maybe it would be more useful to use neural networks computing, genetic algorithms selection and complex system analysis. The second question is, did you try to create a decision-making system or better—expert system?
Dr Athanasiou: I didn't understand your first question. I only understand the question regarding the neural networks.
Dr Kshivets: Sometimes the traditional method, biometrics matches the system, although you didn't find any predictors, but using neural networks computing, and at last simply knowledge engineering and expert system technology you may find these predictors.
Dr Athanasiou: Yes.
Dr Kshivets: And these predictors you can verify on learning samples.
Dr Athanasiou: I have an answer. You asked me if we can use neural networks in our series. It has been published already by Tim Coats, who is Professor of Emergency Medicine in the Royal London and it was found that didn't have any better ROC curve performance in comparison to multivariate models. The multivariate models have a big disadvantage in that they can identify only patients at very high risk, with high incidence. So basically there is not an efficient way to identify the patients who will survive after emergency thoracotomy. We are concentrating on other types of models to analyze these data.
Dr F. Vyhnanek (Prague, Czech Republic): We know it is a problem, emergency thoracotomy in blunt injury. Do you have some experience with emergency thoracotomy in serious blunt injury of the abdomen?
Dr Athanasiou: We know that blunt trauma has a very bad outcome when we perform emergency thoracotomy, and what we have found as well is that between the years 1991 to 1993 we performed many thoracotomies and the majority of them were from blunt injury, and the results showed that we didn't have any survivors. So we introduced a new algorithm for penetrating injury which has been published in the Journal of Trauma.
|
References |
|---|
|
TopAbstract1. Background2. Material and methods3. Statistical analysis4. Results5. Emergency...6. Discussion7. ConclusionAppendix A. Conference...References |
|---|
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
