TRAUMA/SYSTEMATIC REVIEW–META-ANALYSIS
To Be Blunt: Are We Wasting Our Time? Emergency Department Thoracotomy Following Blunt Trauma: A Systematic Review and Meta-Analysis David Slessor, FCEM*; Simon Hunter, FCEM *Corresponding Author. E-mail: [email protected], Twitter: @davidslessor.
Study objective: The role of emergency department (ED) thoracotomy after blunt trauma is controversial. The objective of this review is to determine whether patients treated with an ED thoracotomy after blunt trauma survive and whether survivors have a good neurologic outcome. Methods: A structured search was performed with MEDLINE, EMBASE, CINAHL, and PubMed. Inclusion criteria were ED thoracotomy or out-of-hospital thoracotomy, cardiac arrest or periarrest, and blunt trauma. Outcomes assessed were mortality and neurologic result. The articles were appraised with the system designed by the Institute of Health Economics of Canada. A fixed-effects model was used to meta-analyze the data. Heterogeneity was assessed with the I2 statistic. Results: Twenty-seven articles were included in the review. All were case series. Of 1,369 patients who underwent an ED thoracotomy, 21 (1.5%) survived with a good neurologic outcome. All 21 patients had vital signs present on scene or in the ED and a maximum duration of cardiopulmonary resuscitation of 11 to 15 minutes. Thirteen studies were included in the meta-analysis. If there were either vital signs or signs of life present in the ED, the probability of a poor outcome was 99.2% (95% confidence interval 96.4% to 99.7%). Conclusion: There may be a role for ED thoracotomy after blunt trauma, but only in a limited group of patients. Good outcomes have been achieved for patients who had vital signs on admission and for patients who received an ED thoracotomy within 15 minutes of cardiac arrest. The proposed guideline should be used to determine which patients should be considered for an ED thoracotomy, according to level 4 evidence. [Ann Emerg Med. 2015;65:297-307.] Please see page 298 for the Editor’s Capsule Summary of this article. A podcast for this article is available at www.annemergmed.com. 0196-0644/$-see front matter Copyright © 2014 by the American College of Emergency Physicians. http://dx.doi.org/10.1016/j.annemergmed.2014.08.020
INTRODUCTION Background Emergency department (ED) thoracotomy is an established treatment for patients who have a cardiac arrest after penetrating trauma.1 The role of ED thoracotomy after blunt trauma is less clear and remains controversial. The primary goals of ED thoracotomy are to2 relieve a cardiac tamponade, gain direct control of intrathoracic hemorrhage, allow internal cardiac massage, and compress the aorta so that blood flow is prioritized to vital organs while limiting blood loss from hemorrhage below the diaphragm. The Western Trauma Association3 recommends ED thoracotomy after blunt trauma if there are any signs of life, including electrical activity, or when there are no signs of life, providing there has been less than 10 minutes of cardiopulmonary resuscitation (CPR), whereas Advanced Trauma Life Support4 states that patients who arrive pulseless after blunt trauma are not candidates for ED thoracotomy and that ED thoracotomy for Volume 65, no. 3 : March 2015
cardiac arrest after blunt trauma is rarely effective. This contrasts with penetrating trauma, in which ATLS states that patients who arrive pulseless but with myocardial electrical activity may be candidates for ED thoracotomy; the Western Trauma Association recommends ED thoracotomy for penetrating torso trauma, providing there has been less than 15 minutes of CPR. For this review, the hypothesis statement was the following: an ED thoracotomy for patients in cardiac arrest or periarrest after blunt trauma is a futile procedure. The study outcomes assessed were mortality and neurologic outcome. The intervention assessed was a thoracotomy performed either in the ED or out of hospital. The study population was patients in cardiac arrest or periarrest after blunt trauma. Objective The primary objective was to answer the 3-part question, Do adult patients in cardiac arrest or periarrest after blunt trauma (part 1) who are treated with an ED thoracotomy (part 2) survive and have a good neurologic outcome (part 3)? Annals of Emergency Medicine 297
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Editor’s Capsule Summary
What is already known on this topic Patients with blunt traumatic arrest have a poor outcome. What question this study addressed Is there a subset of patients with traumatic arrest secondary to blunt trauma who are candidates for emergency department (ED) thoracotomy? What this study adds to our knowledge In a review of a 27-case series and meta-analysis of 13 studies, only 1.5% of patients had good neurologic recovery. Without signs of life or vital signs in the ED and a thoracotomy within 15 minutes of arrest, the likelihood of survival with good neurological outcome is very low. How this is relevant to clinical practice These data confirm the limited benefit of ED blunt trauma thoracotomy.
Secondary objectives were to determine survival rates based on the presence or absence of vital signs or signs of life at scene, or in the ED, to determine the duration of CPR after which good outcomes have been reported and to determine which injuries patients have been reported to have survived.
MATERIALS AND METHODS A comprehensive literature search was completed by a single author (D.S.). A librarian reviewed the search strategies. The following sources were used: databases, using NHS Healthcare Databases interface (MEDLINE, 1950 to April 15, 2014; EMBASE, 1980 to April 15, 2014; CINAHL, 1981 to April 15, 2014; Medical Subject Headings terms and results detailed in Appendix E1, available online at http://www.annemergmed. com), Cochrane Library, clinicaltrials.gov, TRIP Database, Google Scholar, PubMed, MDConsult, Bibliographic search, liaison with experts through e-mail correspondence, and gray literature through Google Internet search. There were no language restrictions. Articles were selected with the following criteria. Inclusion criteria were ED or out-of-hospital thoracotomy, cardiac arrest or periarrest, and blunt trauma. Exclusion criteria were operating room thoracotomy, penetrating trauma, pediatric patients only, data on blunt trauma and penetrating trauma patients not clearly differentiated, data on ED thoracotomy not clearly differentiated from operating room thoracotomy or patients who did not have ED thoracotomy, data included in another study, case series of specific injury (eg, splenic injury), case report or case series with only 1 ED thoracotomy for blunt trauma, timings of loss of signs 298 Annals of Emergency Medicine
of life or vital signs not clearly reported (excluded if the number of patients undergoing ED thoracotomy and the number of survivors were not reported for at least 1 of the following subgroups: no vital signs/signs of life at scene, vital signs/signs of life at scene but not in the ED, and vital signs/signs of life in ED), and meta-analysis or review article with no original work. A single author (D.S.) performed the study selection. A second author (S.H.) performed a literature search to ensure that all potentially eligible studies had been included. A system designed by the Institute of Health Economics of Canada was used to critically appraise the articles (personal communication, C. Moga, Institute of Health Economics, October 2012).5 This consists of a 20-point checklist and is designed specifically for case series. The data were extracted and coded in a standardized manner by a single reviewer (D.S.) in an unblinded manner. A meta-analysis was performed according to the Meta-analysis of Observational Studies in Epidemiology Group guideline.6 The study selection criteria were as specified for the systematic review, with the addition of 1 further exclusion criterion: There was not a predetermined protocol or guideline for the use of ED thoracotomy after blunt trauma.7-20 Without a protocol, the decision to perform an ED thoracotomy would have been made on a case-by-case basis, which would have led to a selection bias because there would have been differences between patients for whom the decision was to perform an ED thoracotomy as opposed to declaring the patient dead or prognosis hopeless. This may have led to overestimation of the benefits of ED thoracotomy. The justification for all of the exclusion criteria is shown in Appendix E2, available online at http://www. annemergmed.com. For the meta-analysis, if the number of patients included in a subgroup was unclear,21,22 then the highest possible number of patients was recorded, which therefore resulted in the lowest possible survival rate. To identify heterogeneity, I2 was calculated and a forest plot was produced, both with the software program MIX (version 2.0, BiostatXL, Mountain View, CA).23 The mortality rate and sample size were converted to Freeman-Tukey double arcsines with standard errors.24 This was then converted to a probability scale. Forest plots were also produced to determine the probability of a poor outcome according to the presence or absence of signs of life or vital signs at various timeframes.
RESULTS Twenty-seven articles were included in the review; all were case series. Thirteen studies were analyzed for the meta-analysis.21,22,25-35 The search outline is documented in Figure 1. The studies identified through the search strategy were appropriate for testing the hypothesis because they all included the target study population. In addition, they all used the intervention of interest and recorded mortality. The research performed about the topic question is limited and is based on case series only. They are classified as low-quality Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.40 EDT, Emergency department thoracotomy; PT, penetrating trauma; BT, blunt trauma.
evidence because of a number of biases, including the lack of randomization, blinding, and a control group.5 Appendix E3, available online at http://www.annemergmed. com, shows a summary of the studies included, their important strengths and weaknesses, and the Institute of Health Economics Appraisal Score. The studies are ordered according to the presence or absence of a protocol for the use of ED thoracotomy and the appraisal score. Appendices E4 and E5, available online at http://www.annemergmed.com, show the template marking sheet for the appraisal score and the completed marking sheets, respectively. Appendix E6, available online at http://www. annemergmed.com, reports the inclusion and exclusion criteria for each study and details of any protocol that each study used for deciding when to perform an ED thoracotomy after blunt trauma. The majority of studies were case series of patients who underwent an ED thoracotomy. Only 13 studies had a protocol for the use of ED thoracotomy after blunt trauma. Of those studies that used a protocol, the majority performed an ED Volume 65, no. 3 : March 2015
thoracotomy if the patient was in cardiac arrest, providing that either signs of life or vital signs had been documented as present either on scene or in the ED. The Table shows the results achieved for the patients with blunt trauma who had an ED thoracotomy according to the presence or absence of signs of life or vital signs at different timeframes.
WHO HAD A GOOD OUTCOME AFTER AN ED THORACOTOMY? In studies that reported neurologic outcomes of any survivors, of 1,369 patients with blunt trauma who underwent an ED thoracotomy, only 21 were recorded as having a good neurologic outcome, ie, a rate of 1.5%. A further 4 survivors of 141 patients had a neurologic outcome that was either unclear or not reported. Two of these survivors had signs of life and vital signs present at the scene but lost them before arrival at the ED (personal communication, M. Kalina, Christiana Care, November 2012).11 The other 2 Annals of Emergency Medicine 299
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma Table. Results for patients with blunt trauma who had an ED thoracotomy. Results*
Author Pahle25; personal communication, B. L. Pedersen, Ulleval University Hospital, Norway, May 2012 and February 2013 Ivatury26
Boyd27
Lustenberger28; personal communication, T. Lustenberger, Goethe University, Germany, January 2013 Durham29
Total Number of BT Patients Who Had an ED Thoracotomy
Patients With No SOL or VS on Scene
Patients With SOL or VS on Scene, but Not in the ED
Patents With SOL or VS in ED
SOL (pupillary response, eye movement, respirations, cardiac activity, movement)
82 10 9
13 0 0
14 2 2
55 8 7
VS (pulse, BP)
29 0 0 17 0 0 39 3 3 69 0 0 176 1 0 385 8 4 28 2 2 165 1 0 7–8 0 0 88 1 0 13 0 0 14 0 0 27 0 0 2 0 0 7 0 0 81 0 0 3 0 0 125 2 NR
15 0 0 2 0 0 3 0 0 0 0 0 0 0 0 269 2 0 0 0 0 97 0 0 0 0 0 64–76 0 0 NR
5 0 0 6 0 0 10 0 0 27 0 0 NR
9 0 0 9 0 0 26 3 3 42 0 0 NR
40 NR 1 NR
86 NR 3 NR
NR
NR
NR
NR
1–13 1 0 NR
11 0 0 5 0 0 NR
Results Based on SOL or VS (Definition Used)
SOL (pupil reaction, respiration, movement, supraventricular electrical activity) SOL (BP, respiratory or motor effort, cardiac electrical activity, pupillary response) VS (not defined; if no VS present, then patient received CPR)
Velmahos30
SOL (pupil reactivity, agonal respirations)
Branney31
VS (pulse/BP)
Fialka32
VS (not defined; if no VS present, then patient received CPR)
Lorenz33
SOL (pupillary, corneal, or gag reflex; pulse; BP)
Brautigan21
SOL (not defined)
Esposito22
VS (not defined; if no VS present, then patient received CPR)
Lewis34
VS (BP >60 mm Hg)
Schwab35
SOL (pulse, BP, respirations, reactive pupils, movement, communication)
Passos19
VS (not defined; if no VS present, then patient received CPR)
Johannesdottir18
SOL (respiratory effort, pupillary response, movement, pulse, BP, cardiac electrical activity) SOL (cardiac electrical activity, respirations, pulse, BP, pupillary response, movement) VS (not defined; if no VS present, then patient received CPR)
Edens17
Matsumoto20
Søreide16
SOL (pupillary response, respiratory effort, cardiac electrical activity)
Shimazu15
VS (BP, pulse, respirations, cardiac electrical activity)
300 Annals of Emergency Medicine
0 0 0 3 0 0 1 0 0 NR
NR
24 0 0 1 0 0 NR
52 0 0 1 0 0 NR
29 0 0 2 0 0 NR
0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Table. Continued. Results*
Author
Total Number of BT Patients Who Had an ED Thoracotomy
Results Based on SOL or VS (Definition Used)
Purkiss14
VS (not defined)
Shorr13
VS (pulse, BP)
Bodai12
VS (BP)
Kalina11; personal communication, M. Kalina, Christiana Care, November 2012 Martin10
SOL (pupillary response, breathing, cardiac electrical activity) VS (pulse or BP)
Flynn9
VS (BP)
Mazzorana7
SOL (pulse, BP, respiration, Glasgow Coma Scale score >3, pupillary response, cardiac electrical activity) VS (pulse rate, BP)
Aihara8
10 0 0 39 0 0 38 0 0 16 2 NR 6 0 0 20 0 0 21 0 0 2 0 0
Patients With No SOL or VS on Scene
Patients With SOL or VS on Scene, but Not in the ED
Patents With SOL or VS in ED
2 0 0 NR
NR
NR
NR
NR
NR
0 0 0 NR
4 2 NR NR
NR
NR
12 0 0 1 0 0
3 0 0 NR
7 0 0 5 0 0 12 0 0 4 0 0 6 0 0 6 0 0 NR
SOL, Signs of life; VS, vital signs; BP, blood pressure; NR, not reported. *Results are presented in columnar form as number of patients, number who survived, and number who survived with good neurologic outcome.
survivors, whose neurologic outcome was unclear, had an ED thoracotomy after out-of-hospital cardiac arrest.15 However, because of their injuries, an ED thoracotomy could not have been used to stop any hemorrhage or repair any injury. The highest survival rates were found in patients who had vital signs or signs of life present in the ED, with a smaller number of patients surviving if vital signs or signs of life were present on scene but not on arrival at the ED. All patients who were reported to survive with a good neurologic outcome had vital signs present in the ED or on scene (personal communication, B. L. Pedersen, Ulleval University Hospital, Norway, February 2013).25,28,31,32
WHO HAD A POOR OUTCOME AFTER AN ED THORACOTOMY? The majority of patients did not survive after an ED thoracotomy. This was still the case in patients who had vital signs or signs of life present in the ED. The lowest rates of survival were in patients with no signs of life or vital signs on scene, with only 2 of more than 500 patients surviving, but both had a poor neurologic outcome.31 Eight survivors were reported to have a poor neurologic outcome.21,25,30,31,33,36 The majority of these patients had severe head injuries. Volume 65, no. 3 : March 2015
WHAT INJURIES DID PATIENTS SURVIVE? Seven articles described the injuries survived after an ED thoracotomy (personal communication, M. Kalina, Christiana Care, January 2013).15,22,28,31-33 Six of these articles reported neurologic outcome,15,22,28,31-33 although none used an objective measure such as the Glasgow Outcome Scale. Ten patients, of whom at least 8 survived neurologically intact, had injuries and hemorrhage that the ED thoracotomy may have been helpful for repairing or stopping.11,28,31,32 Four of these patients had cardiac injuries or pericardial tamponade, and at least 5 of these patients had abdominal or pelvic injuries. Eight patients were reported to survive injuries and hemorrhage that an ED thoracotomy would not have been helpful for repairing or stopping.22,28,31,33 At least 6 of these patients had severe neurologic deficits after head injuries. It comes as no surprise that ED thoracotomy has such poor results after severe head injury. However, when the patient arrives in critical condition after blunt trauma, it is difficult to determine whether it was due to a head injury or another injury.
WHAT OTHER FACTORS MADE A DIFFERENCE? Duration of CPR The duration of CPR before ED thoracotomy is a confounding variable that is likely to have affected the results presented. A number of studies have not provided this information. In those Annals of Emergency Medicine 301
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 2. Forest plot for the probability of overall mortality. For each forest plot the studies are ordered according to the Institute of Health Economics Appraisal score 5, with the study with the highest score at the top. See Appendix E5 at www.annemergmed.com for reports of the appraisal score for each study.
that did, the duration of CPR was variable. The longest duration of CPR recorded before ED thoracotomy was 136 minutes.9 Information on the duration of CPR was reported for only 5 patients who survived neurologically intact25,31,32; of these, the longest time was between 11 and 15 minutes for 2 survivors in the study by Fialka et al.32 For patients who survived with a poor neurologic outcome after CPR, we are clearly told the duration for only 1 patient, who had CPR for 15 minutes.22 One limitation is the difficulty in determining accurate timings for the duration of CPR. When critically ill patients were dealt with, recording of accurate timings may not have been prioritized. No study reported methods to ensure this was done accurately. The difficulties in accurately assessing the presence of a pulse in critically ill patients are also well known.
ED THORACOTOMY PRE– OR POST–CARDIAC ARREST? The majority of studies performed ED thoracotomy after cardiac arrest. Pahle et al25 took a different approach, with the majority of patients having vital signs present when the ED thoracotomy was started. Their indications for ED thoracotomy after blunt trauma were exsanguinated patients without immediate response to fluid resuscitation, obviously large abdominal bleeding, and decreasing blood pressure with no response to fluid resuscitation before laparotomy. The median baseline blood pressure for patients in the study by Pahle et al25 was 40 mm Hg, which suggests that the majority of patients were too unstable to be transferred to the operating room. However, 25% of patients in their study had a median baseline blood pressure of greater than or equal to 85 mm Hg, and therefore in other studies similar patients may have been transferred to the operating room. Pahle et al25 reported the highest survival rates for patients who arrived at the ED with vital signs or signs of life present, 302 Annals of Emergency Medicine
which may be due to a selection bias, with the inclusion of a less unwell population that in other studies would have had an operating room thoracotomy. In contrast, the better results may be due to performing an ED thoracotomy on patients with vital signs still present; they may have been able to prevent exsanguination from occurring. Additionally, valuable time may not have been wasted transferring patients to the operating room.
META-ANALYSIS I2 was calculated with MIX (version 2.0). For the 13 studies included in the meta-analysis, I2¼57.88%, indicating moderate heterogeneity.37 Figure 2 demonstrates the forest plot for the probability of overall mortality, P¼.99 (95% confidence interval [CI] 0.98 to 1.0). The results for the study by Pahle et al25 lie outside of this CI, at P¼.88 (95% CI 0.8 to 0.94). The study by Pahle et al25 is likely to be clinically heterogenic because they performed a number of ED thoracotomies on patients who had vital signs present, which is likely to have a significant influence on the overall results reported. Because this practice was not representative of any of the other studies included, the decision was made to exclude the study from the meta-analysis. Excluding the study, I2¼6.84%, indicating minimal heterogeneity. With the other studies having minimal heterogeneity, the data were analyzed with a fixed-effects model. All further forest plots include data from only the 12 remaining studies. Figure 3 shows the forest plot for the overall probability of mortality after an ED thoracotomy when the study by Pahle et al25 was excluded. Figures 4-7 are forest plots of analyses according to the presence or absence of signs of life or vital signs at various timeframes. A poor outcome is defined as death or poor neurologic outcome. For each forest plot, the vertical axis lists the studies according to their appraisal score, with the study with the highest score at the top. Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 3. Forest plot for the probability of overall mortality (after exclusion of the study by Pahle et al25).
The results demonstrate that the probability of a poor outcome was very high. All patients had a poor outcome if there were no signs of life or vital signs on scene, or if there were signs of life on scene but not in the ED. If there were either vital signs or signs of life present in the ED, the probability of a poor outcome was still 99.2% (95% CI 96.4% to 99.7%). The narrow CIs give us confidence in the result. However, if the true result is at either limit of the CI, then the decision about whether to perform an ED thoracotomy will be greatly influenced.
Limitations in data reported meant that survivors from the studies by Fialka32 and Lustenberger28 could not be included in the vital signs subgroups. Therefore, the benefits reported for ED thoracotomy in these subgroups may have been underestimated.
LIMITATIONS The evidence is based on case series only. Because of the lack of blinding, randomization, and a control group, this is classed as
Figure 4. Probability of a poor outcome if vital signs are present on the scene but not in the ED. Volume 65, no. 3 : March 2015
Annals of Emergency Medicine 303
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 5. Probability of a poor outcome if vital signs are present in ED.
level 4 evidence.38 Without a comparison group, we do not know how many patients would have survived without an ED thoracotomy. Transferring patients to the operating room or treating patients with a pericardiocentesis may have resulted in similar or improved outcomes. One of the major weaknesses is the heterogeneity between studies. There was heterogeneity with regard to patient selection, differing definitions of signs of life or vital signs, and the practice of performing additional procedures such as laparotomies in the ED. Patient selection is of particular relevance in studies that did not have a protocol for the use of ED thoracotomy because it is unclear how the decision was made to perform an ED
thoracotomy, as opposed to declaring that the patient was dead or the prognosis hopeless. In studies that did have a protocol, heterogeneity between patient selection remained, with some studies including patients only in cardiac arrest and others including those who had an ED thoracotomy performed with vital signs present. Further heterogeneity was introduced because studies used different definitions for signs of life or vital signs, which makes comparisons difficult. The studies have also inconsistently recorded information on the presence or absence of signs of life or vital signs on scene and on arrival at the ED. The majority of studies were carried out in trauma centers with substantial experience in performing ED thoracotomies. It
Figure 6. Probability of a poor outcome if signs of life are present in the ED.
304 Annals of Emergency Medicine
Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 7. Probability of a poor outcome if either vital signs or signs of life are present in the ED.
is likely that with experience, better outcomes are achieved. Only 1 study included emergency physicians performing the ED thoracotomy, in which there were no survivors. Both of these factors may limit the external validity of the analysis. Including all of the studies, the number of survivors reported is low but the information they provide is vital. Because of the very nature of the intervention, it is unlikely that a randomized controlled trial will ever be performed on the topic, and therefore our decisions have to be made with weak evidence only. Weaknesses of this review include the fact that only a single reviewer performed study selection, data extraction, and quality appraisal. The review was not registered with the International Prospective Register of Systematic Reviews.
DISCUSSION Despite the overall poor results observed after an ED thoracotomy, there are patients who survived with a good neurologic outcome. A number of these patients had lifesaving procedures made possible as a result of the ED thoracotomy, and therefore its use after blunt trauma may be recommended in specific situations. Better results may be possible if stricter criteria are used for determining who undergoes an ED thoracotomy. If stricter criteria had been applied, an ED thoracotomy may not have been performed on a number of patients with prolonged CPR, including patients in the study by Esposito et al22 who received up to 102 minutes of CPR before ED thoracotomy; head injuries, including 70% and 31% of patients in the studies by Durham et al29 and Branney et al,31 respectively; and patients with no vital signs present on scene.22,26,28,31,33 The use of ultrasonography may help identify which patients are likely to benefit from an ED thoracotomy. Unfortunately, no studies used this approach. In patients who have vital signs present in the ED but are too unstable to take to the operating room, results may be better if Volume 65, no. 3 : March 2015
the approach taken by Pahle et al25 is used, when ED thoracotomies are performed with vital signs still present. However, in these patients the same outcomes, or better, may have been achieved by operating room thoracotomy or Resuscitative Endovascular Balloon Occlusion of the Aorta. With these patients, further research to compare ED thoracotomy with operating room thoracotomy may be possible. Figure 8 shows the guideline that was developed to determine when an ED thoracotomy should be considered. This will limit the use of ED thoracotomy in situations that are clearly inappropriate while allowing it in patients for whom it may have a benefit. Consideration will always need to be made with regard to whether transfer to the operating room may be more appropriate or whether the procedure is futile. The duration of “prolonged CPR” has not been defined because of low numbers of survivors after CPR. All survivors in this analysis had less than or equal to 15 minutes of CPR. Operators will always need to consider the potential hazards of ED thoracotomy on health care provider safety. Operators also need to be aware that the guideline is based on level 4 evidence only. One of the major strengths of the literature is that similar results have been repeated across different health care settings. There have been survivors with a good neurologic outcome, and survivors have had common features across all studies.
CONCLUSIONS There may be a role for ED thoracotomy after blunt trauma in a limited group of patients. Good outcomes have been achieved in patients who had vital signs on admission and in patients who received an ED thoracotomy within 15 minutes of cardiac arrest. An ED thoracotomy should not be performed if there are no vital signs at the scene, there has been prolonged CPR, or there is an obvious head injury that is incompatible with a good outcome. Annals of Emergency Medicine 305
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 8. When to consider an ED thoracotomy after blunt trauma.
RECOMMENDATIONS The guideline developed should currently be used to determine which patients to consider performing an ED thoracotomy on after blunt trauma. Future studies should use a standard definition for signs of life or vital signs. The presence or absence of these should be recorded on scene and on arrival at the ED, as should the duration of cardiac arrest and the cardiac rhythm. Studies should report results according to findings on bedside ultrasonography, which may help determine which patients may benefit from an ED thoracotomy. Information should be given on injuries survived to enable readers to determine which patients benefit from an ED thoracotomy. Future studies should compare ED 306 Annals of Emergency Medicine
thoracotomy with operating room thoracotomy in patients who are periarrest. Ideally, these studies should be prospective and multicenter. They should also help us further define which, if any, patients benefit from ED thoracotomy and for which patients it is a futile procedure. The authors acknowledge Julia Harris, FCEM, FRCS, Emergency Department, Southampton General Hospital, UK, for reviewing the article and providing constructive feedback. Supervising editor: Judd E. Hollander, MD Author affiliations: From North Hampshire Hospital, Hampshire, UK (Slessor); and Queen Alexandra Hospital, Portsmouth, UK (Hunter). Volume 65, no. 3 : March 2015
Slessor & Hunter Author contributions: DS conceived the study, performed the systematic review and the meta-analysis. SH supervised the study, performed a literature review, reviewed the manuscript and provided constructive feedback. DS takes responsibility of the article as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist. Publication dates: Received for publication April 28, 2014. Revision received July 28, 2014. Accepted for publication August 15, 2014. Available online October 23, 2014.
REFERENCES 1. Asensio J, Wall M, Minei J. Working Group; Ad Hoc Subcommittee on Outcomes, American College of Surgeons–Committee on Trauma. Practice management guidelines for emergency department thoracotomy. J Am Coll Surg. 2001;193:303-309. 2. Hunt P, Greaves I, Owens W. Emergency thoracotomy in thoracic trauma? a review. Injury. 2006;37:1-19. 3. Burlew CC, Moore EE, Moore FA, et al. Western Trauma Association critical decisions in trauma: resuscitative thoracotomy. J Trauma Acute Care Surg. 2012;73:1359-1363. 4. American College of Surgeons Committee on Trauma, Advanced Trauma Life Support (ATLS). 9th ed., American College of Surgeons, Chicago, IL: 2012. 5. Moga C, Guo B, Schopflocher D, Harstall C. Development of a Quality Appraisal Tool for Case Series Studies Using a Modified Delphi Technique. Edmonton, Canada: Institute of Health Economics; 2012. 6. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology. JAMA. 2000;283:2008-2012. 7. Mazzorana V, Smith RS, Morabito DJ, et al. Limited utility of emergency department thoracotomy. Am Surg. 1994;60:516-520. 8. Aihara R, Millham FH, Blansfield J, et al. Emergency room thoracotomy for penetrating chest injury: effect of an institutional protocol. J Trauma Acute Care Surg. 2001;50:1027-1030. 9. Flynn TC, Ward RE, Miller PW. Emergency department thoracotomy. Ann Emerg Med. 1982;11:413-416. 10. Martin SK, Shatney CH, Sherck JP, et al. Blunt trauma patients with prehospital pulseless electrical activity (PEA): poor ending assured. J Trauma Acute Care Surg. 2002;53:876-881. 11. Kalina M, Teeple E, Fulda G. Are there still selected applications for resuscitative thoracotomy in the emergency department after blunt trauma? Del Med J. 2009;81:195-198. 12. Bodai BI, Smith JP, Blaisdell FW. The role of emergency thoracotomy in blunt trauma. J Trauma. 1982;22:487-491. 13. Shorr RM, Crittenden M, Indeck M, et al. Blunt thoracic trauma. Analysis of 515 patients. Ann Surg. 1987;206:200-205. 14. Purkiss S, Williams M, Cross F, et al. Efficacy of urgent thoracotomy for trauma in patients attended by a helicopter emergency medical service. J R Coll Surg Edinb. 1994;39:289-291. 15. Shimazu S, Shatney CH. Outcomes of trauma patients with no vital signs on hospital admission. J Trauma. 1983;23:213-216. 16. Søreide K, Søiland H, Lossius H, et al. Resuscitative emergency thoracotomy in a Scandinavian trauma hospital—is it justified? Injury. 2007;38:34-42. 17. Edens JW, Beekley AC, Chung KK, et al. Longterm outcomes after combat casualty emergency department thoracotomy. J Am Coll Surg. 2009;209:188-197.
Volume 65, no. 3 : March 2015
Emergency Department Thoracotomy After Blunt Trauma 18. Johannesdottir BK, Mogensen B, Gudbjartsson T. Emergency thoracotomy as a rescue treatment for trauma patients in Iceland. Injury. 2013;44:1186-1190. 19. Passos EM, Engels PT, Doyle JD, et al. Societal costs of inappropriate emergency department thoracotomy. J Am Coll Surg. 2012;214: 18-25. 20. Matsumoto H, Mashiko K, Hara Y, et al. Role of resuscitative emergency field thoracotomy in the Japanese helicopter emergency medical service system. Resuscitation. 2009;80:1270-1274. 21. Brautigan MW, Tietz G. Emergency thoracotomy in an urban community hospital: initial cardiac rhythm as a new predictor of survival. Am J Emerg Med. 1985;3:311-315. 22. Esposito TJ, Jurkovich GJ, Rice CL, et al. Reappraisal of emergency room thoracotomy in a changing environment. J Trauma Acute Care Surg. 1991;31:881-887. 23. Bax L. MIX 2.0—professional software for meta-analysis in Excel. Version 2.0.1.4. BiostatXL. 2011. Available at: http://www.metaanalysis-made-easy.com. Accessed November 20, 2013. 24. Freeman MF, Tukey JW. Transformations related to the angular and the square root. Ann Math Statist. 1950;21:607-611. 25. Pahle AS, Pedersen BL, Skaga NO, et al. Emergency thoracotomy saves lives in a Scandinavian hospital setting. J Trauma Acute Care Surg. 2010;68:599-603. 26. Ivatury R, Kazigo J, Rohman M, et al. “Directed” emergency room thoracotomy: a prognostic prerequisite for survival. J Trauma. 1991;31:1076-1081. 27. Boyd M, Vanek VW, Bourguet CC. Emergency room resuscitative thoracotomy: when is it indicated? J Trauma. 1992;33:714-721. 28. Lustenberger T, Labler L, Stover J, et al. Resuscitative emergency thoracotomy in a Swiss trauma centre. Br J Surg. 2012;99:548-549. 29. Durham L III, Richardson RJ, Wall MJ Jr, et al. Emergency center thoracotomy: impact of prehospital resuscitation. J Trauma. 1992;32:775-779. 30. Velmahos GC, Degiannis E, Souter I, et al. Outcome of a strict policy on emergency department thoracotomies. Arch Surg. 1995;130:774-777. 31. Branney SW, Moore EE, Feldhaus KM, et al. Critical analysis of two decades of experience with postinjury emergency department thoracotomy in a regional trauma center. J Trauma Acute Care Surg. 1998;45:87-94. 32. Fialka C, Sebök C, Kemetzhofer P, et al. Open-chest cardiopulmonary resuscitation after cardiac arrest in cases of blunt chest or abdominal trauma: a consecutive series of 38 cases. J Trauma Acute Care Surg. 2004;57:809-814. 33. Lorenz HP, Steinmetz B, Lieberman J, et al. Emergency thoracotomy: survival correlates with physiologic status. J Trauma. 1992;32: 780-785. 34. Lewis G, Knottenbelt J. Should emergency room thoracotomy be reserved for cases of cardiac tamponade? Injury. 1991;22:5-6. 35. Schwab CW, Adcock O, Max M. Emergency department thoracotomy (EDT). A 26-month experience using an “agonal” protocol. Am Surg. 1986;52:20-29. 36. Timothy Baxter B, Moore EE, Moore JB, et al. Emergency department thoracotomy following injury: critical determinants for patient salvage. World J Surg. 1988;12:671-674. 37. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557-560. 38. Phillips B, Ball C, Sackett D, et al. Levels of evidence and grades of recommendation. 2001. Available at: http://www.cebm.net. Accessed November 1, 2013. 39. Kissoon N, Dreyer J, Walia M. Pediatric trauma: differences in pathophysiology, injury patterns and treatment compared with adult trauma. CMAJ. 1990;142:27-34. 40. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151:264-269.
Annals of Emergency Medicine 307
Emergency Department Thoracotomy After Blunt Trauma
Slessor & Hunter
APPENDIX Appendix E1. Database searches. Line EMBASE 1 2 3 4 5 6 7 8 9 10 11 MEDLINE 1 2 3 4 5 6 7 8 9 10 11 CINAHL 1 2 3 4 5 6 7 8 9
Search Term
Results
exp THORACOTOMY/ thoracotom*.ti,ab 1 OR 2 exp BLUNT TRAUMA/ blunt.ti,ab 4 OR 5 exp EMERGENCY/ exp RESUSCITATION/ (emerg* OR resus* OR immediate OR “pre-hosp*”).ti,ab 7 OR 8 OR 9 3 AND 6 AND 10
21,512 20,567 29,954 18,816 26,427 32,513 34,530 68,838 746,434
exp THORACOTOMY/ thoracotom*.ti,ab 1 OR 2 exp WOUNDS, NONPENETRATING/ blunt.ti,ab 4 OR 5 exp EMERGENCIES/ exp RESUSCITATION/ (resus* OR emerg* OR immediate OR “pre-hosp*”).ti,ab 7 OR 8 OR 9 3 AND 6 AND 10 exp THORACOTOMY/ thoracotom*.ti,ab 1 OR 2 blunt.ti,ab exp EMERGENCIES/ exp RESUSCITATION/ (resus* OR emerg* OR immediate OR “pre-hosp*”).ti,ab 5 OR 6 OR 7 3 AND 4 AND 8
307.e1 Annals of Emergency Medicine
783,870 476 8,395 17,358 21,012 28,002 23,346 40,506 34,103 72,665 622,202 680,236 405 616 881 1,203 2,943 5,420 20,701 107,934 124,340 30
Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Appendix E2. Justification for exclusion criteria. Exclusion Criteria Operating room thoracotomy
Penetrating trauma
Study on pediatric patients only
Data on BT and penetrating trauma patients not clearly differentiated, or data on ED thoracotomy not clearly differentiated from operating room thoracotomy/ patients who did not have ED thoracotomy Case series of specific injury, eg, splenic injury
Case report or case series with only 1 case of ED thoracotomy for BT Timings of loss of SOL or VS not clearly reported Meta-analyses or review article with no original work Data included in another study Studies that did not have a predetermined protocol or guideline for the use of ED thoracotomy after BT
Volume 65, no. 3 : March 2015
Reason for Exclusion Patients who have an operating room thoracotomy are likely to differ from those who have an ED thoracotomy. One of the main differences is likely to be the hemodynamic stability. Patients who are transferred to the operating room are likely to be more hemodynamically stable than patients who have the thoracotomy performed in the ED. Patients with penetrating injuries were excluded because good outcomes have already been demonstrated after penetrating trauma, in part because of the high rate of pericardial tamponade.1,2 Pediatric trauma patients have different injury patterns in comparison with adult trauma patients.39 Pediatric patients are also known to compensate well and deteriorate late.39 Therefore, the results may be different in the pediatric population compared with the adult population, and a separate meta-analysis is advised. To identify the performance of ED thoracotomy for BT patients, the results for these patients need to be reported independently.
These studies would give results for a specific injury. When patients arrive in the ED, their specific injuries are unknown, and a decision has to be made about whether to perform an ED thoracotomy without knowing this information. If data can be analyzed from only 1 patient, then limited conclusions can be made. This variable is a confounding variable and if not reported may invalidate the results found. These articles would bring no additional information for this analysis. Bibliography searches were completed if the other inclusion and exclusion criteria were met. This ensures that patients were not counted twice. Without a protocol, the decision about whether to perform an ED thoracotomy would be made on a case-by-case basis, and this would introduce significant bias and heterogeneity. There are likely to be differences between patients who had an ED thoracotomy and those who did not.
Annals of Emergency Medicine 307.e2
Author, Year, Country
Design and Setting
Participants
Intervention (Technique of ED Thoracotomy for BT Patients)
Predetermined Outcomes
Strengths
Weaknesses 52 patients had VS at start of ED thoracotomy.2 In other studies, these patients may have had an operating room thoracotomy. Survival reported at only 30 days 1 survivor lost to follow-up ED thoracotomy delayed for up to 20 min in some patients Duration of CPR not reported
Prospective case series Single trauma center
ED thoracotomy (n¼109), which included ED thoracotomy for BT (n¼82)
66 anterolateral, 3 sternotomy, 13 combined
Mortality
Predetermined protocol for use of ED thoracotomy Data collected prospectively
Ivatury,3 1991, USA
Retrospective case series Single trauma center
ED thoracotomy (n¼163), which included ED thoracotomy for BT (n¼29)
Not recorded
Mortality Neurologic outcome
Boyd,4 1992, USA
Retrospective case series Single Level I trauma center
ED thoracotomy for trauma (n¼28), which included ED thoracotomy for BT (n¼17)
Anterolateral
Mortality Neurologic outcome
Predetermined protocol for use of ED thoracotomy Report subgroup results according to presence/ absence of VS on scene and in ED Planned to assess neurologic outcome Predetermined protocol for use of ED thoracotomy Report subgroup results according to presence/ absence of SOL on scene and in ED Reported neurologic outcome
Lustenberger,5 2012, Switzerland
Retrospective case series Single Level I trauma center
ED thoracotomy or operating room thoracotomy for trauma (n¼121), which included ED thoracotomy for BT (n¼39)
Including penetrating trauma (n¼10): 24 anterolateral, 21 sternotomy, 2 clamshell, 2 combined
Mortality
Predetermined protocol for use of ED thoracotomy Results for ED thoracotomy/operating room thoracotomy clearly differentiated Reported injuries that survivors sustained
Duration of CPR not reported. Some patients had prolonged transport times. Unclear whether they also had prolonged CPR. It appears that 2 patients had an ED thoracotomy who according to protocol should not have. Patients had sternotomies and laparotomies in the ED; may not apply to UK setting Included patients who according to protocol should not have had an ED thoracotomy6
13.5
12.5
12
11
Slessor & Hunter
Volume 65, no. 3 : March 2015
Pahle,1 2010, Norway
Appraisal Score
Emergency Department Thoracotomy After Blunt Trauma
307.e3 Annals of Emergency Medicine
Appendix E3. Summary of articles.
ED thoracotomy (n¼389), which included ED thoracotomy for BT (n¼69)
Anterolateral and if required extended across midline
Mortality
Predetermined protocol for use of ED thoracotomy Duration of out-of-hospital CPR reported Prospective log of patients Detailed report of technique for ED thoracotomy
Velmahos,8 1995, South Africa
Retrospective case series Single trauma center
ED thoracotomy (n¼846), which included ED thoracotomy for BT (n¼176)
Anterolateral Occasionally sternotomy or transsternal extension
Mortality Neurologic outcome
Predetermined protocol for use of ED thoracotomy Large number of patients Aimed to assess neurologic outcome
Branney,9 1998, USA
Retrospective case series Single Level I trauma center
ED thoracotomy for trauma (n¼950), which included ED thoracotomy for BT (n¼385)
Not recorded
Mortality
Predetermined protocol for use of ED thoracotomy Large number of patients Report subgroup results according to presence/ absence of VS on scene and in ED
Fialka,10 2004, Italy
Retrospective case series Single Level I trauma center
ED thoracotomy or operating room thoracotomy for chest/ abdominal BT after witnessed cardiac arrest 70 y Cardiac arrest not witnessed by medical staff None stated
Brautigan10
None stated
None stated
Esposito11 Lewis12
ED thoracotomy for trauma ED thoracotomy
None stated Operating room thoracotomy
Schwab13
ED thoracotomy
Passos14
ED thoracotomy
Johannesdottir15
ED thoracotomy for major thoracic trauma
Edens16 Matsumoto17
ED thoracotomy ED thoracotomy and out-of-hospital thoracotomy for BT
Operating room thoracotomy Patients who were stabilized with pericardiocentesis Patients transferred from other hospitals, or who sustained burns ED thoracotomy performed for nontraumatic indications, eg, ruptured aortic aneurysm Operating room thoracotomy Thoracotomy before cardiac arrest Cardiac arrest after arrival in ED
307.e15 Annals of Emergency Medicine
Predetermined Protocol for ED Thoracotomy After BT
Exclusion Criteria of Study
Exsanguinated patients without immediate response to fluid resuscitation Obviously large abdominal bleeding and decreasing BP with no response to fluid resuscitation Contraindicated if without SOL at admission Extremis not responding to intubation/fluids If lose VS and had VS present either in ED or on scene Lost SOL in hospital or immediately before arrival Exsanguinated patients with no immediate response to fluid resuscitation Sudden hemodynamic deterioration in the ED after thoracic/abdominal trauma If became pulseless en route and had organized cardiac rhythm on arrival to ED Loss of VS around time of admission or cardiac arrest in ED Absence of VS during transport but preservation of SOL on admission Without VS posttrauma Or in select patients (n¼9 including penetrating trauma patients) if cannot maintain BP >70 mm Hg despite aggressive volume resuscitation Traumatic cardiac arrest with
To Be Blunt: Are We Wasting Our Time? Emergency Department Thoracotomy Following Blunt Trauma: A Systematic Review and Meta-Analysis David Slessor, FCEM*; Simon Hunter, FCEM *Corresponding Author. E-mail: [email protected], Twitter: @davidslessor.
Study objective: The role of emergency department (ED) thoracotomy after blunt trauma is controversial. The objective of this review is to determine whether patients treated with an ED thoracotomy after blunt trauma survive and whether survivors have a good neurologic outcome. Methods: A structured search was performed with MEDLINE, EMBASE, CINAHL, and PubMed. Inclusion criteria were ED thoracotomy or out-of-hospital thoracotomy, cardiac arrest or periarrest, and blunt trauma. Outcomes assessed were mortality and neurologic result. The articles were appraised with the system designed by the Institute of Health Economics of Canada. A fixed-effects model was used to meta-analyze the data. Heterogeneity was assessed with the I2 statistic. Results: Twenty-seven articles were included in the review. All were case series. Of 1,369 patients who underwent an ED thoracotomy, 21 (1.5%) survived with a good neurologic outcome. All 21 patients had vital signs present on scene or in the ED and a maximum duration of cardiopulmonary resuscitation of 11 to 15 minutes. Thirteen studies were included in the meta-analysis. If there were either vital signs or signs of life present in the ED, the probability of a poor outcome was 99.2% (95% confidence interval 96.4% to 99.7%). Conclusion: There may be a role for ED thoracotomy after blunt trauma, but only in a limited group of patients. Good outcomes have been achieved for patients who had vital signs on admission and for patients who received an ED thoracotomy within 15 minutes of cardiac arrest. The proposed guideline should be used to determine which patients should be considered for an ED thoracotomy, according to level 4 evidence. [Ann Emerg Med. 2015;65:297-307.] Please see page 298 for the Editor’s Capsule Summary of this article. A podcast for this article is available at www.annemergmed.com. 0196-0644/$-see front matter Copyright © 2014 by the American College of Emergency Physicians. http://dx.doi.org/10.1016/j.annemergmed.2014.08.020
INTRODUCTION Background Emergency department (ED) thoracotomy is an established treatment for patients who have a cardiac arrest after penetrating trauma.1 The role of ED thoracotomy after blunt trauma is less clear and remains controversial. The primary goals of ED thoracotomy are to2 relieve a cardiac tamponade, gain direct control of intrathoracic hemorrhage, allow internal cardiac massage, and compress the aorta so that blood flow is prioritized to vital organs while limiting blood loss from hemorrhage below the diaphragm. The Western Trauma Association3 recommends ED thoracotomy after blunt trauma if there are any signs of life, including electrical activity, or when there are no signs of life, providing there has been less than 10 minutes of cardiopulmonary resuscitation (CPR), whereas Advanced Trauma Life Support4 states that patients who arrive pulseless after blunt trauma are not candidates for ED thoracotomy and that ED thoracotomy for Volume 65, no. 3 : March 2015
cardiac arrest after blunt trauma is rarely effective. This contrasts with penetrating trauma, in which ATLS states that patients who arrive pulseless but with myocardial electrical activity may be candidates for ED thoracotomy; the Western Trauma Association recommends ED thoracotomy for penetrating torso trauma, providing there has been less than 15 minutes of CPR. For this review, the hypothesis statement was the following: an ED thoracotomy for patients in cardiac arrest or periarrest after blunt trauma is a futile procedure. The study outcomes assessed were mortality and neurologic outcome. The intervention assessed was a thoracotomy performed either in the ED or out of hospital. The study population was patients in cardiac arrest or periarrest after blunt trauma. Objective The primary objective was to answer the 3-part question, Do adult patients in cardiac arrest or periarrest after blunt trauma (part 1) who are treated with an ED thoracotomy (part 2) survive and have a good neurologic outcome (part 3)? Annals of Emergency Medicine 297
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Editor’s Capsule Summary
What is already known on this topic Patients with blunt traumatic arrest have a poor outcome. What question this study addressed Is there a subset of patients with traumatic arrest secondary to blunt trauma who are candidates for emergency department (ED) thoracotomy? What this study adds to our knowledge In a review of a 27-case series and meta-analysis of 13 studies, only 1.5% of patients had good neurologic recovery. Without signs of life or vital signs in the ED and a thoracotomy within 15 minutes of arrest, the likelihood of survival with good neurological outcome is very low. How this is relevant to clinical practice These data confirm the limited benefit of ED blunt trauma thoracotomy.
Secondary objectives were to determine survival rates based on the presence or absence of vital signs or signs of life at scene, or in the ED, to determine the duration of CPR after which good outcomes have been reported and to determine which injuries patients have been reported to have survived.
MATERIALS AND METHODS A comprehensive literature search was completed by a single author (D.S.). A librarian reviewed the search strategies. The following sources were used: databases, using NHS Healthcare Databases interface (MEDLINE, 1950 to April 15, 2014; EMBASE, 1980 to April 15, 2014; CINAHL, 1981 to April 15, 2014; Medical Subject Headings terms and results detailed in Appendix E1, available online at http://www.annemergmed. com), Cochrane Library, clinicaltrials.gov, TRIP Database, Google Scholar, PubMed, MDConsult, Bibliographic search, liaison with experts through e-mail correspondence, and gray literature through Google Internet search. There were no language restrictions. Articles were selected with the following criteria. Inclusion criteria were ED or out-of-hospital thoracotomy, cardiac arrest or periarrest, and blunt trauma. Exclusion criteria were operating room thoracotomy, penetrating trauma, pediatric patients only, data on blunt trauma and penetrating trauma patients not clearly differentiated, data on ED thoracotomy not clearly differentiated from operating room thoracotomy or patients who did not have ED thoracotomy, data included in another study, case series of specific injury (eg, splenic injury), case report or case series with only 1 ED thoracotomy for blunt trauma, timings of loss of signs 298 Annals of Emergency Medicine
of life or vital signs not clearly reported (excluded if the number of patients undergoing ED thoracotomy and the number of survivors were not reported for at least 1 of the following subgroups: no vital signs/signs of life at scene, vital signs/signs of life at scene but not in the ED, and vital signs/signs of life in ED), and meta-analysis or review article with no original work. A single author (D.S.) performed the study selection. A second author (S.H.) performed a literature search to ensure that all potentially eligible studies had been included. A system designed by the Institute of Health Economics of Canada was used to critically appraise the articles (personal communication, C. Moga, Institute of Health Economics, October 2012).5 This consists of a 20-point checklist and is designed specifically for case series. The data were extracted and coded in a standardized manner by a single reviewer (D.S.) in an unblinded manner. A meta-analysis was performed according to the Meta-analysis of Observational Studies in Epidemiology Group guideline.6 The study selection criteria were as specified for the systematic review, with the addition of 1 further exclusion criterion: There was not a predetermined protocol or guideline for the use of ED thoracotomy after blunt trauma.7-20 Without a protocol, the decision to perform an ED thoracotomy would have been made on a case-by-case basis, which would have led to a selection bias because there would have been differences between patients for whom the decision was to perform an ED thoracotomy as opposed to declaring the patient dead or prognosis hopeless. This may have led to overestimation of the benefits of ED thoracotomy. The justification for all of the exclusion criteria is shown in Appendix E2, available online at http://www. annemergmed.com. For the meta-analysis, if the number of patients included in a subgroup was unclear,21,22 then the highest possible number of patients was recorded, which therefore resulted in the lowest possible survival rate. To identify heterogeneity, I2 was calculated and a forest plot was produced, both with the software program MIX (version 2.0, BiostatXL, Mountain View, CA).23 The mortality rate and sample size were converted to Freeman-Tukey double arcsines with standard errors.24 This was then converted to a probability scale. Forest plots were also produced to determine the probability of a poor outcome according to the presence or absence of signs of life or vital signs at various timeframes.
RESULTS Twenty-seven articles were included in the review; all were case series. Thirteen studies were analyzed for the meta-analysis.21,22,25-35 The search outline is documented in Figure 1. The studies identified through the search strategy were appropriate for testing the hypothesis because they all included the target study population. In addition, they all used the intervention of interest and recorded mortality. The research performed about the topic question is limited and is based on case series only. They are classified as low-quality Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.40 EDT, Emergency department thoracotomy; PT, penetrating trauma; BT, blunt trauma.
evidence because of a number of biases, including the lack of randomization, blinding, and a control group.5 Appendix E3, available online at http://www.annemergmed. com, shows a summary of the studies included, their important strengths and weaknesses, and the Institute of Health Economics Appraisal Score. The studies are ordered according to the presence or absence of a protocol for the use of ED thoracotomy and the appraisal score. Appendices E4 and E5, available online at http://www.annemergmed.com, show the template marking sheet for the appraisal score and the completed marking sheets, respectively. Appendix E6, available online at http://www. annemergmed.com, reports the inclusion and exclusion criteria for each study and details of any protocol that each study used for deciding when to perform an ED thoracotomy after blunt trauma. The majority of studies were case series of patients who underwent an ED thoracotomy. Only 13 studies had a protocol for the use of ED thoracotomy after blunt trauma. Of those studies that used a protocol, the majority performed an ED Volume 65, no. 3 : March 2015
thoracotomy if the patient was in cardiac arrest, providing that either signs of life or vital signs had been documented as present either on scene or in the ED. The Table shows the results achieved for the patients with blunt trauma who had an ED thoracotomy according to the presence or absence of signs of life or vital signs at different timeframes.
WHO HAD A GOOD OUTCOME AFTER AN ED THORACOTOMY? In studies that reported neurologic outcomes of any survivors, of 1,369 patients with blunt trauma who underwent an ED thoracotomy, only 21 were recorded as having a good neurologic outcome, ie, a rate of 1.5%. A further 4 survivors of 141 patients had a neurologic outcome that was either unclear or not reported. Two of these survivors had signs of life and vital signs present at the scene but lost them before arrival at the ED (personal communication, M. Kalina, Christiana Care, November 2012).11 The other 2 Annals of Emergency Medicine 299
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma Table. Results for patients with blunt trauma who had an ED thoracotomy. Results*
Author Pahle25; personal communication, B. L. Pedersen, Ulleval University Hospital, Norway, May 2012 and February 2013 Ivatury26
Boyd27
Lustenberger28; personal communication, T. Lustenberger, Goethe University, Germany, January 2013 Durham29
Total Number of BT Patients Who Had an ED Thoracotomy
Patients With No SOL or VS on Scene
Patients With SOL or VS on Scene, but Not in the ED
Patents With SOL or VS in ED
SOL (pupillary response, eye movement, respirations, cardiac activity, movement)
82 10 9
13 0 0
14 2 2
55 8 7
VS (pulse, BP)
29 0 0 17 0 0 39 3 3 69 0 0 176 1 0 385 8 4 28 2 2 165 1 0 7–8 0 0 88 1 0 13 0 0 14 0 0 27 0 0 2 0 0 7 0 0 81 0 0 3 0 0 125 2 NR
15 0 0 2 0 0 3 0 0 0 0 0 0 0 0 269 2 0 0 0 0 97 0 0 0 0 0 64–76 0 0 NR
5 0 0 6 0 0 10 0 0 27 0 0 NR
9 0 0 9 0 0 26 3 3 42 0 0 NR
40 NR 1 NR
86 NR 3 NR
NR
NR
NR
NR
1–13 1 0 NR
11 0 0 5 0 0 NR
Results Based on SOL or VS (Definition Used)
SOL (pupil reaction, respiration, movement, supraventricular electrical activity) SOL (BP, respiratory or motor effort, cardiac electrical activity, pupillary response) VS (not defined; if no VS present, then patient received CPR)
Velmahos30
SOL (pupil reactivity, agonal respirations)
Branney31
VS (pulse/BP)
Fialka32
VS (not defined; if no VS present, then patient received CPR)
Lorenz33
SOL (pupillary, corneal, or gag reflex; pulse; BP)
Brautigan21
SOL (not defined)
Esposito22
VS (not defined; if no VS present, then patient received CPR)
Lewis34
VS (BP >60 mm Hg)
Schwab35
SOL (pulse, BP, respirations, reactive pupils, movement, communication)
Passos19
VS (not defined; if no VS present, then patient received CPR)
Johannesdottir18
SOL (respiratory effort, pupillary response, movement, pulse, BP, cardiac electrical activity) SOL (cardiac electrical activity, respirations, pulse, BP, pupillary response, movement) VS (not defined; if no VS present, then patient received CPR)
Edens17
Matsumoto20
Søreide16
SOL (pupillary response, respiratory effort, cardiac electrical activity)
Shimazu15
VS (BP, pulse, respirations, cardiac electrical activity)
300 Annals of Emergency Medicine
0 0 0 3 0 0 1 0 0 NR
NR
24 0 0 1 0 0 NR
52 0 0 1 0 0 NR
29 0 0 2 0 0 NR
0 0 0 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0
Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Table. Continued. Results*
Author
Total Number of BT Patients Who Had an ED Thoracotomy
Results Based on SOL or VS (Definition Used)
Purkiss14
VS (not defined)
Shorr13
VS (pulse, BP)
Bodai12
VS (BP)
Kalina11; personal communication, M. Kalina, Christiana Care, November 2012 Martin10
SOL (pupillary response, breathing, cardiac electrical activity) VS (pulse or BP)
Flynn9
VS (BP)
Mazzorana7
SOL (pulse, BP, respiration, Glasgow Coma Scale score >3, pupillary response, cardiac electrical activity) VS (pulse rate, BP)
Aihara8
10 0 0 39 0 0 38 0 0 16 2 NR 6 0 0 20 0 0 21 0 0 2 0 0
Patients With No SOL or VS on Scene
Patients With SOL or VS on Scene, but Not in the ED
Patents With SOL or VS in ED
2 0 0 NR
NR
NR
NR
NR
NR
0 0 0 NR
4 2 NR NR
NR
NR
12 0 0 1 0 0
3 0 0 NR
7 0 0 5 0 0 12 0 0 4 0 0 6 0 0 6 0 0 NR
SOL, Signs of life; VS, vital signs; BP, blood pressure; NR, not reported. *Results are presented in columnar form as number of patients, number who survived, and number who survived with good neurologic outcome.
survivors, whose neurologic outcome was unclear, had an ED thoracotomy after out-of-hospital cardiac arrest.15 However, because of their injuries, an ED thoracotomy could not have been used to stop any hemorrhage or repair any injury. The highest survival rates were found in patients who had vital signs or signs of life present in the ED, with a smaller number of patients surviving if vital signs or signs of life were present on scene but not on arrival at the ED. All patients who were reported to survive with a good neurologic outcome had vital signs present in the ED or on scene (personal communication, B. L. Pedersen, Ulleval University Hospital, Norway, February 2013).25,28,31,32
WHO HAD A POOR OUTCOME AFTER AN ED THORACOTOMY? The majority of patients did not survive after an ED thoracotomy. This was still the case in patients who had vital signs or signs of life present in the ED. The lowest rates of survival were in patients with no signs of life or vital signs on scene, with only 2 of more than 500 patients surviving, but both had a poor neurologic outcome.31 Eight survivors were reported to have a poor neurologic outcome.21,25,30,31,33,36 The majority of these patients had severe head injuries. Volume 65, no. 3 : March 2015
WHAT INJURIES DID PATIENTS SURVIVE? Seven articles described the injuries survived after an ED thoracotomy (personal communication, M. Kalina, Christiana Care, January 2013).15,22,28,31-33 Six of these articles reported neurologic outcome,15,22,28,31-33 although none used an objective measure such as the Glasgow Outcome Scale. Ten patients, of whom at least 8 survived neurologically intact, had injuries and hemorrhage that the ED thoracotomy may have been helpful for repairing or stopping.11,28,31,32 Four of these patients had cardiac injuries or pericardial tamponade, and at least 5 of these patients had abdominal or pelvic injuries. Eight patients were reported to survive injuries and hemorrhage that an ED thoracotomy would not have been helpful for repairing or stopping.22,28,31,33 At least 6 of these patients had severe neurologic deficits after head injuries. It comes as no surprise that ED thoracotomy has such poor results after severe head injury. However, when the patient arrives in critical condition after blunt trauma, it is difficult to determine whether it was due to a head injury or another injury.
WHAT OTHER FACTORS MADE A DIFFERENCE? Duration of CPR The duration of CPR before ED thoracotomy is a confounding variable that is likely to have affected the results presented. A number of studies have not provided this information. In those Annals of Emergency Medicine 301
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 2. Forest plot for the probability of overall mortality. For each forest plot the studies are ordered according to the Institute of Health Economics Appraisal score 5, with the study with the highest score at the top. See Appendix E5 at www.annemergmed.com for reports of the appraisal score for each study.
that did, the duration of CPR was variable. The longest duration of CPR recorded before ED thoracotomy was 136 minutes.9 Information on the duration of CPR was reported for only 5 patients who survived neurologically intact25,31,32; of these, the longest time was between 11 and 15 minutes for 2 survivors in the study by Fialka et al.32 For patients who survived with a poor neurologic outcome after CPR, we are clearly told the duration for only 1 patient, who had CPR for 15 minutes.22 One limitation is the difficulty in determining accurate timings for the duration of CPR. When critically ill patients were dealt with, recording of accurate timings may not have been prioritized. No study reported methods to ensure this was done accurately. The difficulties in accurately assessing the presence of a pulse in critically ill patients are also well known.
ED THORACOTOMY PRE– OR POST–CARDIAC ARREST? The majority of studies performed ED thoracotomy after cardiac arrest. Pahle et al25 took a different approach, with the majority of patients having vital signs present when the ED thoracotomy was started. Their indications for ED thoracotomy after blunt trauma were exsanguinated patients without immediate response to fluid resuscitation, obviously large abdominal bleeding, and decreasing blood pressure with no response to fluid resuscitation before laparotomy. The median baseline blood pressure for patients in the study by Pahle et al25 was 40 mm Hg, which suggests that the majority of patients were too unstable to be transferred to the operating room. However, 25% of patients in their study had a median baseline blood pressure of greater than or equal to 85 mm Hg, and therefore in other studies similar patients may have been transferred to the operating room. Pahle et al25 reported the highest survival rates for patients who arrived at the ED with vital signs or signs of life present, 302 Annals of Emergency Medicine
which may be due to a selection bias, with the inclusion of a less unwell population that in other studies would have had an operating room thoracotomy. In contrast, the better results may be due to performing an ED thoracotomy on patients with vital signs still present; they may have been able to prevent exsanguination from occurring. Additionally, valuable time may not have been wasted transferring patients to the operating room.
META-ANALYSIS I2 was calculated with MIX (version 2.0). For the 13 studies included in the meta-analysis, I2¼57.88%, indicating moderate heterogeneity.37 Figure 2 demonstrates the forest plot for the probability of overall mortality, P¼.99 (95% confidence interval [CI] 0.98 to 1.0). The results for the study by Pahle et al25 lie outside of this CI, at P¼.88 (95% CI 0.8 to 0.94). The study by Pahle et al25 is likely to be clinically heterogenic because they performed a number of ED thoracotomies on patients who had vital signs present, which is likely to have a significant influence on the overall results reported. Because this practice was not representative of any of the other studies included, the decision was made to exclude the study from the meta-analysis. Excluding the study, I2¼6.84%, indicating minimal heterogeneity. With the other studies having minimal heterogeneity, the data were analyzed with a fixed-effects model. All further forest plots include data from only the 12 remaining studies. Figure 3 shows the forest plot for the overall probability of mortality after an ED thoracotomy when the study by Pahle et al25 was excluded. Figures 4-7 are forest plots of analyses according to the presence or absence of signs of life or vital signs at various timeframes. A poor outcome is defined as death or poor neurologic outcome. For each forest plot, the vertical axis lists the studies according to their appraisal score, with the study with the highest score at the top. Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 3. Forest plot for the probability of overall mortality (after exclusion of the study by Pahle et al25).
The results demonstrate that the probability of a poor outcome was very high. All patients had a poor outcome if there were no signs of life or vital signs on scene, or if there were signs of life on scene but not in the ED. If there were either vital signs or signs of life present in the ED, the probability of a poor outcome was still 99.2% (95% CI 96.4% to 99.7%). The narrow CIs give us confidence in the result. However, if the true result is at either limit of the CI, then the decision about whether to perform an ED thoracotomy will be greatly influenced.
Limitations in data reported meant that survivors from the studies by Fialka32 and Lustenberger28 could not be included in the vital signs subgroups. Therefore, the benefits reported for ED thoracotomy in these subgroups may have been underestimated.
LIMITATIONS The evidence is based on case series only. Because of the lack of blinding, randomization, and a control group, this is classed as
Figure 4. Probability of a poor outcome if vital signs are present on the scene but not in the ED. Volume 65, no. 3 : March 2015
Annals of Emergency Medicine 303
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 5. Probability of a poor outcome if vital signs are present in ED.
level 4 evidence.38 Without a comparison group, we do not know how many patients would have survived without an ED thoracotomy. Transferring patients to the operating room or treating patients with a pericardiocentesis may have resulted in similar or improved outcomes. One of the major weaknesses is the heterogeneity between studies. There was heterogeneity with regard to patient selection, differing definitions of signs of life or vital signs, and the practice of performing additional procedures such as laparotomies in the ED. Patient selection is of particular relevance in studies that did not have a protocol for the use of ED thoracotomy because it is unclear how the decision was made to perform an ED
thoracotomy, as opposed to declaring that the patient was dead or the prognosis hopeless. In studies that did have a protocol, heterogeneity between patient selection remained, with some studies including patients only in cardiac arrest and others including those who had an ED thoracotomy performed with vital signs present. Further heterogeneity was introduced because studies used different definitions for signs of life or vital signs, which makes comparisons difficult. The studies have also inconsistently recorded information on the presence or absence of signs of life or vital signs on scene and on arrival at the ED. The majority of studies were carried out in trauma centers with substantial experience in performing ED thoracotomies. It
Figure 6. Probability of a poor outcome if signs of life are present in the ED.
304 Annals of Emergency Medicine
Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 7. Probability of a poor outcome if either vital signs or signs of life are present in the ED.
is likely that with experience, better outcomes are achieved. Only 1 study included emergency physicians performing the ED thoracotomy, in which there were no survivors. Both of these factors may limit the external validity of the analysis. Including all of the studies, the number of survivors reported is low but the information they provide is vital. Because of the very nature of the intervention, it is unlikely that a randomized controlled trial will ever be performed on the topic, and therefore our decisions have to be made with weak evidence only. Weaknesses of this review include the fact that only a single reviewer performed study selection, data extraction, and quality appraisal. The review was not registered with the International Prospective Register of Systematic Reviews.
DISCUSSION Despite the overall poor results observed after an ED thoracotomy, there are patients who survived with a good neurologic outcome. A number of these patients had lifesaving procedures made possible as a result of the ED thoracotomy, and therefore its use after blunt trauma may be recommended in specific situations. Better results may be possible if stricter criteria are used for determining who undergoes an ED thoracotomy. If stricter criteria had been applied, an ED thoracotomy may not have been performed on a number of patients with prolonged CPR, including patients in the study by Esposito et al22 who received up to 102 minutes of CPR before ED thoracotomy; head injuries, including 70% and 31% of patients in the studies by Durham et al29 and Branney et al,31 respectively; and patients with no vital signs present on scene.22,26,28,31,33 The use of ultrasonography may help identify which patients are likely to benefit from an ED thoracotomy. Unfortunately, no studies used this approach. In patients who have vital signs present in the ED but are too unstable to take to the operating room, results may be better if Volume 65, no. 3 : March 2015
the approach taken by Pahle et al25 is used, when ED thoracotomies are performed with vital signs still present. However, in these patients the same outcomes, or better, may have been achieved by operating room thoracotomy or Resuscitative Endovascular Balloon Occlusion of the Aorta. With these patients, further research to compare ED thoracotomy with operating room thoracotomy may be possible. Figure 8 shows the guideline that was developed to determine when an ED thoracotomy should be considered. This will limit the use of ED thoracotomy in situations that are clearly inappropriate while allowing it in patients for whom it may have a benefit. Consideration will always need to be made with regard to whether transfer to the operating room may be more appropriate or whether the procedure is futile. The duration of “prolonged CPR” has not been defined because of low numbers of survivors after CPR. All survivors in this analysis had less than or equal to 15 minutes of CPR. Operators will always need to consider the potential hazards of ED thoracotomy on health care provider safety. Operators also need to be aware that the guideline is based on level 4 evidence only. One of the major strengths of the literature is that similar results have been repeated across different health care settings. There have been survivors with a good neurologic outcome, and survivors have had common features across all studies.
CONCLUSIONS There may be a role for ED thoracotomy after blunt trauma in a limited group of patients. Good outcomes have been achieved in patients who had vital signs on admission and in patients who received an ED thoracotomy within 15 minutes of cardiac arrest. An ED thoracotomy should not be performed if there are no vital signs at the scene, there has been prolonged CPR, or there is an obvious head injury that is incompatible with a good outcome. Annals of Emergency Medicine 305
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Figure 8. When to consider an ED thoracotomy after blunt trauma.
RECOMMENDATIONS The guideline developed should currently be used to determine which patients to consider performing an ED thoracotomy on after blunt trauma. Future studies should use a standard definition for signs of life or vital signs. The presence or absence of these should be recorded on scene and on arrival at the ED, as should the duration of cardiac arrest and the cardiac rhythm. Studies should report results according to findings on bedside ultrasonography, which may help determine which patients may benefit from an ED thoracotomy. Information should be given on injuries survived to enable readers to determine which patients benefit from an ED thoracotomy. Future studies should compare ED 306 Annals of Emergency Medicine
thoracotomy with operating room thoracotomy in patients who are periarrest. Ideally, these studies should be prospective and multicenter. They should also help us further define which, if any, patients benefit from ED thoracotomy and for which patients it is a futile procedure. The authors acknowledge Julia Harris, FCEM, FRCS, Emergency Department, Southampton General Hospital, UK, for reviewing the article and providing constructive feedback. Supervising editor: Judd E. Hollander, MD Author affiliations: From North Hampshire Hospital, Hampshire, UK (Slessor); and Queen Alexandra Hospital, Portsmouth, UK (Hunter). Volume 65, no. 3 : March 2015
Slessor & Hunter Author contributions: DS conceived the study, performed the systematic review and the meta-analysis. SH supervised the study, performed a literature review, reviewed the manuscript and provided constructive feedback. DS takes responsibility of the article as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist. Publication dates: Received for publication April 28, 2014. Revision received July 28, 2014. Accepted for publication August 15, 2014. Available online October 23, 2014.
REFERENCES 1. Asensio J, Wall M, Minei J. Working Group; Ad Hoc Subcommittee on Outcomes, American College of Surgeons–Committee on Trauma. Practice management guidelines for emergency department thoracotomy. J Am Coll Surg. 2001;193:303-309. 2. Hunt P, Greaves I, Owens W. Emergency thoracotomy in thoracic trauma? a review. Injury. 2006;37:1-19. 3. Burlew CC, Moore EE, Moore FA, et al. Western Trauma Association critical decisions in trauma: resuscitative thoracotomy. J Trauma Acute Care Surg. 2012;73:1359-1363. 4. American College of Surgeons Committee on Trauma, Advanced Trauma Life Support (ATLS). 9th ed., American College of Surgeons, Chicago, IL: 2012. 5. Moga C, Guo B, Schopflocher D, Harstall C. Development of a Quality Appraisal Tool for Case Series Studies Using a Modified Delphi Technique. Edmonton, Canada: Institute of Health Economics; 2012. 6. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology. JAMA. 2000;283:2008-2012. 7. Mazzorana V, Smith RS, Morabito DJ, et al. Limited utility of emergency department thoracotomy. Am Surg. 1994;60:516-520. 8. Aihara R, Millham FH, Blansfield J, et al. Emergency room thoracotomy for penetrating chest injury: effect of an institutional protocol. J Trauma Acute Care Surg. 2001;50:1027-1030. 9. Flynn TC, Ward RE, Miller PW. Emergency department thoracotomy. Ann Emerg Med. 1982;11:413-416. 10. Martin SK, Shatney CH, Sherck JP, et al. Blunt trauma patients with prehospital pulseless electrical activity (PEA): poor ending assured. J Trauma Acute Care Surg. 2002;53:876-881. 11. Kalina M, Teeple E, Fulda G. Are there still selected applications for resuscitative thoracotomy in the emergency department after blunt trauma? Del Med J. 2009;81:195-198. 12. Bodai BI, Smith JP, Blaisdell FW. The role of emergency thoracotomy in blunt trauma. J Trauma. 1982;22:487-491. 13. Shorr RM, Crittenden M, Indeck M, et al. Blunt thoracic trauma. Analysis of 515 patients. Ann Surg. 1987;206:200-205. 14. Purkiss S, Williams M, Cross F, et al. Efficacy of urgent thoracotomy for trauma in patients attended by a helicopter emergency medical service. J R Coll Surg Edinb. 1994;39:289-291. 15. Shimazu S, Shatney CH. Outcomes of trauma patients with no vital signs on hospital admission. J Trauma. 1983;23:213-216. 16. Søreide K, Søiland H, Lossius H, et al. Resuscitative emergency thoracotomy in a Scandinavian trauma hospital—is it justified? Injury. 2007;38:34-42. 17. Edens JW, Beekley AC, Chung KK, et al. Longterm outcomes after combat casualty emergency department thoracotomy. J Am Coll Surg. 2009;209:188-197.
Volume 65, no. 3 : March 2015
Emergency Department Thoracotomy After Blunt Trauma 18. Johannesdottir BK, Mogensen B, Gudbjartsson T. Emergency thoracotomy as a rescue treatment for trauma patients in Iceland. Injury. 2013;44:1186-1190. 19. Passos EM, Engels PT, Doyle JD, et al. Societal costs of inappropriate emergency department thoracotomy. J Am Coll Surg. 2012;214: 18-25. 20. Matsumoto H, Mashiko K, Hara Y, et al. Role of resuscitative emergency field thoracotomy in the Japanese helicopter emergency medical service system. Resuscitation. 2009;80:1270-1274. 21. Brautigan MW, Tietz G. Emergency thoracotomy in an urban community hospital: initial cardiac rhythm as a new predictor of survival. Am J Emerg Med. 1985;3:311-315. 22. Esposito TJ, Jurkovich GJ, Rice CL, et al. Reappraisal of emergency room thoracotomy in a changing environment. J Trauma Acute Care Surg. 1991;31:881-887. 23. Bax L. MIX 2.0—professional software for meta-analysis in Excel. Version 2.0.1.4. BiostatXL. 2011. Available at: http://www.metaanalysis-made-easy.com. Accessed November 20, 2013. 24. Freeman MF, Tukey JW. Transformations related to the angular and the square root. Ann Math Statist. 1950;21:607-611. 25. Pahle AS, Pedersen BL, Skaga NO, et al. Emergency thoracotomy saves lives in a Scandinavian hospital setting. J Trauma Acute Care Surg. 2010;68:599-603. 26. Ivatury R, Kazigo J, Rohman M, et al. “Directed” emergency room thoracotomy: a prognostic prerequisite for survival. J Trauma. 1991;31:1076-1081. 27. Boyd M, Vanek VW, Bourguet CC. Emergency room resuscitative thoracotomy: when is it indicated? J Trauma. 1992;33:714-721. 28. Lustenberger T, Labler L, Stover J, et al. Resuscitative emergency thoracotomy in a Swiss trauma centre. Br J Surg. 2012;99:548-549. 29. Durham L III, Richardson RJ, Wall MJ Jr, et al. Emergency center thoracotomy: impact of prehospital resuscitation. J Trauma. 1992;32:775-779. 30. Velmahos GC, Degiannis E, Souter I, et al. Outcome of a strict policy on emergency department thoracotomies. Arch Surg. 1995;130:774-777. 31. Branney SW, Moore EE, Feldhaus KM, et al. Critical analysis of two decades of experience with postinjury emergency department thoracotomy in a regional trauma center. J Trauma Acute Care Surg. 1998;45:87-94. 32. Fialka C, Sebök C, Kemetzhofer P, et al. Open-chest cardiopulmonary resuscitation after cardiac arrest in cases of blunt chest or abdominal trauma: a consecutive series of 38 cases. J Trauma Acute Care Surg. 2004;57:809-814. 33. Lorenz HP, Steinmetz B, Lieberman J, et al. Emergency thoracotomy: survival correlates with physiologic status. J Trauma. 1992;32: 780-785. 34. Lewis G, Knottenbelt J. Should emergency room thoracotomy be reserved for cases of cardiac tamponade? Injury. 1991;22:5-6. 35. Schwab CW, Adcock O, Max M. Emergency department thoracotomy (EDT). A 26-month experience using an “agonal” protocol. Am Surg. 1986;52:20-29. 36. Timothy Baxter B, Moore EE, Moore JB, et al. Emergency department thoracotomy following injury: critical determinants for patient salvage. World J Surg. 1988;12:671-674. 37. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557-560. 38. Phillips B, Ball C, Sackett D, et al. Levels of evidence and grades of recommendation. 2001. Available at: http://www.cebm.net. Accessed November 1, 2013. 39. Kissoon N, Dreyer J, Walia M. Pediatric trauma: differences in pathophysiology, injury patterns and treatment compared with adult trauma. CMAJ. 1990;142:27-34. 40. Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151:264-269.
Annals of Emergency Medicine 307
Emergency Department Thoracotomy After Blunt Trauma
Slessor & Hunter
APPENDIX Appendix E1. Database searches. Line EMBASE 1 2 3 4 5 6 7 8 9 10 11 MEDLINE 1 2 3 4 5 6 7 8 9 10 11 CINAHL 1 2 3 4 5 6 7 8 9
Search Term
Results
exp THORACOTOMY/ thoracotom*.ti,ab 1 OR 2 exp BLUNT TRAUMA/ blunt.ti,ab 4 OR 5 exp EMERGENCY/ exp RESUSCITATION/ (emerg* OR resus* OR immediate OR “pre-hosp*”).ti,ab 7 OR 8 OR 9 3 AND 6 AND 10
21,512 20,567 29,954 18,816 26,427 32,513 34,530 68,838 746,434
exp THORACOTOMY/ thoracotom*.ti,ab 1 OR 2 exp WOUNDS, NONPENETRATING/ blunt.ti,ab 4 OR 5 exp EMERGENCIES/ exp RESUSCITATION/ (resus* OR emerg* OR immediate OR “pre-hosp*”).ti,ab 7 OR 8 OR 9 3 AND 6 AND 10 exp THORACOTOMY/ thoracotom*.ti,ab 1 OR 2 blunt.ti,ab exp EMERGENCIES/ exp RESUSCITATION/ (resus* OR emerg* OR immediate OR “pre-hosp*”).ti,ab 5 OR 6 OR 7 3 AND 4 AND 8
307.e1 Annals of Emergency Medicine
783,870 476 8,395 17,358 21,012 28,002 23,346 40,506 34,103 72,665 622,202 680,236 405 616 881 1,203 2,943 5,420 20,701 107,934 124,340 30
Volume 65, no. 3 : March 2015
Slessor & Hunter
Emergency Department Thoracotomy After Blunt Trauma
Appendix E2. Justification for exclusion criteria. Exclusion Criteria Operating room thoracotomy
Penetrating trauma
Study on pediatric patients only
Data on BT and penetrating trauma patients not clearly differentiated, or data on ED thoracotomy not clearly differentiated from operating room thoracotomy/ patients who did not have ED thoracotomy Case series of specific injury, eg, splenic injury
Case report or case series with only 1 case of ED thoracotomy for BT Timings of loss of SOL or VS not clearly reported Meta-analyses or review article with no original work Data included in another study Studies that did not have a predetermined protocol or guideline for the use of ED thoracotomy after BT
Volume 65, no. 3 : March 2015
Reason for Exclusion Patients who have an operating room thoracotomy are likely to differ from those who have an ED thoracotomy. One of the main differences is likely to be the hemodynamic stability. Patients who are transferred to the operating room are likely to be more hemodynamically stable than patients who have the thoracotomy performed in the ED. Patients with penetrating injuries were excluded because good outcomes have already been demonstrated after penetrating trauma, in part because of the high rate of pericardial tamponade.1,2 Pediatric trauma patients have different injury patterns in comparison with adult trauma patients.39 Pediatric patients are also known to compensate well and deteriorate late.39 Therefore, the results may be different in the pediatric population compared with the adult population, and a separate meta-analysis is advised. To identify the performance of ED thoracotomy for BT patients, the results for these patients need to be reported independently.
These studies would give results for a specific injury. When patients arrive in the ED, their specific injuries are unknown, and a decision has to be made about whether to perform an ED thoracotomy without knowing this information. If data can be analyzed from only 1 patient, then limited conclusions can be made. This variable is a confounding variable and if not reported may invalidate the results found. These articles would bring no additional information for this analysis. Bibliography searches were completed if the other inclusion and exclusion criteria were met. This ensures that patients were not counted twice. Without a protocol, the decision about whether to perform an ED thoracotomy would be made on a case-by-case basis, and this would introduce significant bias and heterogeneity. There are likely to be differences between patients who had an ED thoracotomy and those who did not.
Annals of Emergency Medicine 307.e2
Author, Year, Country
Design and Setting
Participants
Intervention (Technique of ED Thoracotomy for BT Patients)
Predetermined Outcomes
Strengths
Weaknesses 52 patients had VS at start of ED thoracotomy.2 In other studies, these patients may have had an operating room thoracotomy. Survival reported at only 30 days 1 survivor lost to follow-up ED thoracotomy delayed for up to 20 min in some patients Duration of CPR not reported
Prospective case series Single trauma center
ED thoracotomy (n¼109), which included ED thoracotomy for BT (n¼82)
66 anterolateral, 3 sternotomy, 13 combined
Mortality
Predetermined protocol for use of ED thoracotomy Data collected prospectively
Ivatury,3 1991, USA
Retrospective case series Single trauma center
ED thoracotomy (n¼163), which included ED thoracotomy for BT (n¼29)
Not recorded
Mortality Neurologic outcome
Boyd,4 1992, USA
Retrospective case series Single Level I trauma center
ED thoracotomy for trauma (n¼28), which included ED thoracotomy for BT (n¼17)
Anterolateral
Mortality Neurologic outcome
Predetermined protocol for use of ED thoracotomy Report subgroup results according to presence/ absence of VS on scene and in ED Planned to assess neurologic outcome Predetermined protocol for use of ED thoracotomy Report subgroup results according to presence/ absence of SOL on scene and in ED Reported neurologic outcome
Lustenberger,5 2012, Switzerland
Retrospective case series Single Level I trauma center
ED thoracotomy or operating room thoracotomy for trauma (n¼121), which included ED thoracotomy for BT (n¼39)
Including penetrating trauma (n¼10): 24 anterolateral, 21 sternotomy, 2 clamshell, 2 combined
Mortality
Predetermined protocol for use of ED thoracotomy Results for ED thoracotomy/operating room thoracotomy clearly differentiated Reported injuries that survivors sustained
Duration of CPR not reported. Some patients had prolonged transport times. Unclear whether they also had prolonged CPR. It appears that 2 patients had an ED thoracotomy who according to protocol should not have. Patients had sternotomies and laparotomies in the ED; may not apply to UK setting Included patients who according to protocol should not have had an ED thoracotomy6
13.5
12.5
12
11
Slessor & Hunter
Volume 65, no. 3 : March 2015
Pahle,1 2010, Norway
Appraisal Score
Emergency Department Thoracotomy After Blunt Trauma
307.e3 Annals of Emergency Medicine
Appendix E3. Summary of articles.
ED thoracotomy (n¼389), which included ED thoracotomy for BT (n¼69)
Anterolateral and if required extended across midline
Mortality
Predetermined protocol for use of ED thoracotomy Duration of out-of-hospital CPR reported Prospective log of patients Detailed report of technique for ED thoracotomy
Velmahos,8 1995, South Africa
Retrospective case series Single trauma center
ED thoracotomy (n¼846), which included ED thoracotomy for BT (n¼176)
Anterolateral Occasionally sternotomy or transsternal extension
Mortality Neurologic outcome
Predetermined protocol for use of ED thoracotomy Large number of patients Aimed to assess neurologic outcome
Branney,9 1998, USA
Retrospective case series Single Level I trauma center
ED thoracotomy for trauma (n¼950), which included ED thoracotomy for BT (n¼385)
Not recorded
Mortality
Predetermined protocol for use of ED thoracotomy Large number of patients Report subgroup results according to presence/ absence of VS on scene and in ED
Fialka,10 2004, Italy
Retrospective case series Single Level I trauma center
ED thoracotomy or operating room thoracotomy for chest/ abdominal BT after witnessed cardiac arrest 70 y Cardiac arrest not witnessed by medical staff None stated
Brautigan10
None stated
None stated
Esposito11 Lewis12
ED thoracotomy for trauma ED thoracotomy
None stated Operating room thoracotomy
Schwab13
ED thoracotomy
Passos14
ED thoracotomy
Johannesdottir15
ED thoracotomy for major thoracic trauma
Edens16 Matsumoto17
ED thoracotomy ED thoracotomy and out-of-hospital thoracotomy for BT
Operating room thoracotomy Patients who were stabilized with pericardiocentesis Patients transferred from other hospitals, or who sustained burns ED thoracotomy performed for nontraumatic indications, eg, ruptured aortic aneurysm Operating room thoracotomy Thoracotomy before cardiac arrest Cardiac arrest after arrival in ED
307.e15 Annals of Emergency Medicine
Predetermined Protocol for ED Thoracotomy After BT
Exclusion Criteria of Study
Exsanguinated patients without immediate response to fluid resuscitation Obviously large abdominal bleeding and decreasing BP with no response to fluid resuscitation Contraindicated if without SOL at admission Extremis not responding to intubation/fluids If lose VS and had VS present either in ED or on scene Lost SOL in hospital or immediately before arrival Exsanguinated patients with no immediate response to fluid resuscitation Sudden hemodynamic deterioration in the ED after thoracic/abdominal trauma If became pulseless en route and had organized cardiac rhythm on arrival to ED Loss of VS around time of admission or cardiac arrest in ED Absence of VS during transport but preservation of SOL on admission Without VS posttrauma Or in select patients (n¼9 including penetrating trauma patients) if cannot maintain BP >70 mm Hg despite aggressive volume resuscitation Traumatic cardiac arrest with
