What are the predictors of scapula fractures in high-impact blunt trauma patients and why do we miss them in the emergency department?

Transcription

1 Eur J Trauma Emerg Surg (2012) 38: DOI /s ORIGINAL ARTICLE What are the predictors of scapula fractures in high-impact blunt trauma patients and why do we miss them in the emergency department? M. Uzkeser M. Emet M. Kılıç M. Işık Received: 17 February 2011 / Accepted: 16 July 2011 / Published online: 9 August 2011 Ó Springer-Verlag 2011 Abstract Aim To find out the predictors of scapula fractures in high-impact blunt trauma patients and the predictors of missing them in the emergency department (ED). Methods This is a retrospective descriptive review of patients treated at a single hospital over a 3.5-year period. The study included 1,039 high-impact blunt trauma patients of any age. The patients were divided into two groups: patients with scapula fractures (the patient group) and others (the control group). The patient group was reassessed into two subgroups: patients with and without missed scapula fractures. Logistic regression analysis was used to assess the predictors of a scapular fracture in the ED. Results The incidence of scapula fractures was 4%. The Injury Severity Scores (ISS) of the scapula fracture group were significantly higher than those of the control group (P \ ). Patients with scapula fractures were significantly older (P \ 0.001). Patients with scapular body fractures stayed for significantly longer in the intensive M. Uzkeser (&) M. Emet Department of Emergency Medicine, Ataturk University Medical Faculty, Erzurum, Turkey mustafauzkeser@gmail.com M. Kılıç Department of Orthopedics, Oltu State Hospital, Erzurum, Turkey M. Işık Department of Family Medicine, Ataturk University Medical Faculty, Erzurum, Turkey care unit (ICU) (P = 0.039) and their mean Abbreviated Injury Scale (AIS) face score was higher (P = 0.04) compared to scapular fractures without body parts. Scapula fractures were missed in 35.7% of patients. All bilateral scapula fractures were missed. Computerized tomography showing shoulder region or scapula was obtained in 26.7% of missed injuries. Conclusion After matching for ISS, the mortality of patients with scapula fracture was significantly higher than the control group. We observed that fractures in the glenoid and scapular neck occurred higher than in the body region. We found that the seniority of the ED doctor, consultation ratio for orthopedics in ED, and the existence of brain contusion were important parameters for missing scapula fractures in ED. Keywords Scapula fractures Emergency medicine Missed injuries Introduction The rare occurrence of scapula fractures makes them understudied and missed injuries. High-energy trauma is the leading cause of scapular fractures, ranging between 85 and 93% [1, 2]. Besides, they are usually associated with additional lesions, varying from 78 to 88% [1 6]. For that reason, these fractures are even suggested as sentinel injuries which alert emergency physicians to the presence of concomitant injury in other body regions [7]. In recent years, multicentric extended studies have been carried out with a focus on scapula fractures. Giving important information, they also provoked additional questions: after adjusting for the Injury Severity Score (ISS), do morbidity and mortality change between the patients with and without

2 158 M. Uzkeser et al. scapula fracture? Does age play a role in the incidence of scapula fracture? What are the predictors of missing scapula fractures in emergency department (ED)? We aimed to ascertain the predictors of scapula fractures in high-impact blunt trauma patients and the predictors of missing them in the ED. Patients and methods Inclusion and exclusion criteria The study is a retrospective evaluation of a prospectively collected data set. The study includes 1,039 high-impact blunt trauma patients (HIBTP) of any age admitted to the ED of Ataturk University Aziziye Hospital in Erzurum providence, Turkey, between July 2006 and January We recorded patients demographics, organ and musculoskeletal injuries, consultations in the ED, Glasgow Coma Scale (GCS), Abbreviated Injury Scale (AIS) of six body regions, ISS, and length of stay (LOS) in the ED in a chart developed for study purposes. Low-impact injuries such as ground-level falls, blow injuries (assaults), isolated eye traumas, occupational accidents causing only extremity injury, and penetrating injuries like stabbing or firearm injuries were excluded. Burns, inhalation injuries, electrical and lightning injuries, and patients transferred to other hospitals within 24 h of admission were also excluded. Patient groups After the patient s discharge, the patient s data were reexamined from the review of hospital records so as to define operations, length of hospital and intensive care unit (ICU) stay, mortality, and whether there was any missed scapula fracture. First, HIBTP were divided into two groups: patients with scapula fractures (the patient group) and others (the control group). Secondly, patients with scapula fractures were re-assessed into two subgroups: patients with and without missed scapula fractures. Definitions The Seniority of the ED doctor was divided up into five senior house officer levels (postgraduate years) of training emergency doctors. Fractures detected after onward transfer from the ED to the inpatient bed unit and were not documented in emergency records, admission notes, or consultation notes in the ED were called missed fractures. High-impact trauma patients included fall from height, motor vehicle accidents (MVA), and others (dent to remain under heavy objects, animal-related injuries). Scapula fractures were classified into three groups following Zdravkovic and Damholt, who divided the scapula anatomically into: Type I, scapula body; Type II, apophyseal fractures, including the acromion and coracoids; and Type III, superolateral angle fractures, including the glenoid and scapular neck [8]. Outcome measures We compared groups with and without scapular fractures and groups with or without missed fractures with respect to outcomes such as AIS of six anatomic areas, ISS, LOS in the ED, in hospital, and in the ICU, trauma consultations, GCS, mortality, operation, frequency of skeletal injuries (skull, maxillofacial, rib, pelvis, hand/foot, forearm, forefoot, femur, humerus, clavicle, sternum, cervical, thoracal, and lumbar spine), frequency of soft tissue injuries (brain contusion, intracranial hemorrhage, thoracic, pulmonary, cardiac, retroperitoneal hemorrhage, kidney, diaphragm, spleen, bowel, liver, gastropancreatic, vascular, and nerve), and the total number of injuries. Statistics As the ISS score is a confounding factor for our outcome measures, adjustment was made by weighting the data for ISS using SPSS. Since our study design does not permit for matching, we have chosen adjustment in order to eliminate ISS as a confounding factor. To compare two independent groups, the Mann Whitney U-test or Student s t-test was used. v 2 or Fisher s exact test was used to compare categorical data. Logistic regression analysis was conducted in order to assess the predictors of a scapular fracture in HIBTP. The data were analyzed with the SPSS 11.0 software package (SPSS Inc., Chicago, IL). A P-value \ 0.05 was considered to be statistically significant. Results Of the study population, 63.8% (n = 663) were due to MVA, 25.1% (n = 261) were due to falls from height, and 11.1% (n = 115) were due to other high-impact blunt trauma. The incidence of scapula fractures among HIBTP was 4% (n = 42). Scapula fractures occurred in 5% (n = 33) of MVA, 2.3% (n = 6) of falls from height, and 2.6% (n = 3) of other high-impact blunt trauma (v 2 ; P = 0.105). Fractured scapula was on the left in 50% (n = 21) and on the right in 42.9% (n = 18). Bilateral scapular fracture was observed in 7.1% (n = 3) of scapular fractures (two were due to MVA, one was due to fall from height). In only five of the patients (11.9%) was scapular fracture the only injury.

3 What are the predictors of scapula fractures in high impact? 159 Characteristics of patients with scapula fractures There was no difference between the two groups with respect to gender (33.3% vs. 25.8% females with and without scapula fractures, respectively; v 2 ; P = 0.275). No age difference was observed between genders in patients with scapula fractures (44.1 ± 21.9 for males vs ± 18.9 for females). The ISS scores of the scapula fracture group were significantly higher than those of the control group (19.0 ± 10.2 vs ± 8.7; t-test; P \ ). Patients with scapula fractures were significantly older (37.6 ± 20.7 years vs ± 20.7 years, t-test; P \ 0.001) and they were consulted more by orthopedics and thorax surgery by emergency doctors compared to patients without fracture (81% vs. 52.1% for orthopedics, v 2 ; P \ ; 57.1% vs. 29.9% for chest surgery, v 2 ; P \ ). Regarding accompanying injuries in the two groups, the significance of accompanying injuries and ISS adjusted ratios of accompanying injuries are analyzed in Table 1, the outcome measures of the two groups are summarized in Table 2, and the predictors of scapula fracture in HIBTP are shown in Table 3. Characteristics of scapula fractures The fractures occurred in the body (45.2%, n = 19), in the glenoid and scapular neck (52.4%, n = 22), and in the acromion and coracoid (9.5%, n = 4). Three patients (7.1%) had fractures in two different anatomical sites. No difference was observed between patients with and without glenoid fractures with respect to the outcome measures. Patients with scapular body fractures stayed for significantly longer in the ICU (6.5 ± 19.5 days vs ± 0.65 days; Mann Whitney U; P = 0.039) and their mean AIS face was higher (0.42 ± 0.77 vs ± 0.21; t-test; P = 0.04) compared to scapular fractures without body parts. Characteristics of patients with missed scapula fractures Scapula fractures were missed in 35.7% (n = 15) of patients. All bilateral scapula fractures were missed. Computerized tomography showing the shoulder region or scapula was obtained in 26.7% of missed injuries. Forward stepwise logistic regression was applied using missed scapula fractures as the dependent variable (Table 3). Values entered to the regression were seniority of the ED doctor, ISS, consultation to orthopedics (categorical), brain contusion (categorical), and lumbar fracture and/or dislocation (categorical). The sensitivity and specificity of the model to predict missing scapula fractures were 89 and 73%, respectively. Discussion Previous reports mentioned that high-energy trauma was the leading cause of scapular fractures, but none of them examined these fractures by means of trauma mechanism. Superior methodological features of our study were the presence of a control group, a more detailed evaluation of associated injuries, and the ISS of the patient and control groups were not excessively high or low. Baldwin et al. [7] did not mention about the mechanism of injuries and the ISS of the control group was too low. Stephens et al. [9] declared relatively lower ISS scores for both cases and controls, while Veysi et al. [10] included groups with relatively higher ISS scores. The incidence of scapula fractures range between 0.47 and 6.7% [10 13]. Contrary to our study, all previous studies were conducted among adults in the literature. They found that these fractures were more common in young and middle-aged men [14 16]. We included all ages of patients and 14.3% of our cohort were B18 years. Our study is the first one which finds a difference between patients with and without scapula fracture with respect to age. Baldwin et al. [7] found that rib fractures, clavicle fractures, spinal fractures, pneumothorax, and lung injury were significantly more common in scapular fracture patients compared to controls. The odds of rib fracture, clavicle fracture, pelvic fracture, spinal fracture, sternal fracture, pneumothorax, lung injury, upper extremity injury, and humerus fracture were significantly higher in the scapula fracture group than in the controls after accounting for the ISS [7]. They concluded that a physician should suspect concomitant thoracic and upper extremity injuries in a patient with a scapula fracture [7]. Veysi et al. [10] found a significantly higher incidence of rib and clavicle fractures, but there were no statistical differences in the incidence of humerus, femur, and pelvis fractures. They concluded that the presence of a scapula fracture indicates trauma to the central and upper parts of the body, sparing the lower regions [10]. Weening et al. [11] reported that significantly increased incidence of skeletal injuries when compared with controls were flail chest, clavicle fractures, rib fractures, spine fractures, and tibia and fibula fractures. Injuries to the chest, either a pneumothorax or a pulmonary contusion, or injuries to the spleen and liver were also significantly more common in patients with scapular fractures when compared with those without them. They reported a higher incidence of spinal column fractures (44.7%) and hepatic (13%) and splenic injuries (15%) than previous studies [11]. In a recent study by Coimbra et al. [13], motor vehicle occupants with scapula fractures had three times the odds of having thoracic injury and twice the odds for spinal injuries compared to occupants with other shoulder injuries. When we compared patients

4 160 M. Uzkeser et al. Table 1 Accompanying injuries in the two groups Diagnosis ISS not adjusted P-value (v 2 ) ISS adjusted P-value (v 2 ) Scapula frx (?) Scapula frx (-) Scapula frx (?) Scapula frx (-) % (95% CI) [n = 42] % (95% CI) [n = 997] % (95% CI) [n = 42] % (95% CI) [n = 997] Skull frx 19 ( ) [8] 22 ( ) [219] ns 26.2 ( ) [8] 28.3 ( ) [219] ns Maxillofacial frx 14.3 ( ) [6] 10.6 ( ) [106] ns 21 ( ) [6] 14.7 ( ) [106] \ Intracranial hemorrhage 11.9 ( ) [5] 15 ( ) [150] ns 16.3 ( ) [5] 23.4 ( ) [150] \ Brain contusion 14.3 ( ) [6] 22.9 ( ) [228] ns 24.5 ( ) [6] 32.9 ( ) [228] \ Clavicle frx and/or disl 26.2 ( ) [11] 6.2 ( ) [62] \ ( ) [11] 8.5 (8 8.9) [62] \ Rib frx 50 ( ) [21] 20.8 ( ) [207] \ ( ) [21] 26.8 ( ) [207] \ Hemo- and/or pneumothorax 35.7 ( ) [15] 15 ( ) [150] \ ( ) [15] 22.3 ( ) [150] \ Pulmonary contusion 47.6 ( ) [20] 13.6 ( ) [136] \ ( ) [20] 21.6 ( ) [136] \ Cardiac contusion 2.4 (-2.2 to 7) [1] 2 ( ) [20] ns 1.8 ( ) [1] 3 ( ) [20] ns Splenic injury 7.1 (-0.7 to 14.9) [3] 5.2 ( ) [52] ns 13.5 ( ) [3] 8.2 ( ) [52] \ Liver injury 11.9 ( ) [5] 6.9 ( ) [69] ns 16.7 ( ) [5] 11.1 ( ) [69] \ Kidney injury 2.4 (-2.2 to 7) [1] 3.1 (2 4.2) [31] ns 1.5 ( ) [1] 4 ( ) [31] \ Retroperitoneal hematoma 4.8 (-1.7 to 11.2) [2] 4.6 ( ) [46] ns 5.9 ( ) [2] 7 ( ) [46] ns Pelvic frx and/or disl 23.8 ( ) [10] 15.4 ( ) [154] ns 29 ( ) [10] 17.2 ( ) [154] \ Tibia and/or fibula frx and/or disl 11.9 ( ) [5] 12.1 ( ) [121] ns 10 ( ) [5] 11.3 ( ) [121] ns Femur frx and/or disl 7.1 (-0.7 to 14.9) [3] 11.5 ( ) [115] ns 10 ( ) [3] 13.1 ( ) [115] Humerus frx and/or disl 16.7 ( ) [7] 7.8 ( ) [78] ns 19.2 ( ) [7] 6.5 ( ) [78] \ Radius and/or ulna frx and/or disl 4.8 (-1.7 to 11.2) [2] 7.7 ( ) [77] ns 6.3 (4.6 8) [2] 8.1 ( ) [77] ns Cervical vertebra frx and/or disl 7.1 (-0.7 to 14.9) [3] 3.5 ( ) [35] ns 9.4 ( ) [3] 5.1 ( ) [35] \ Thoracal vertebra frx and/or disl 14.3 ( ) [6] 5.4 (4 6.8) [54] ( ) [6] 7.1 ( ) [54] \ Lumbar vertebra frx and/or disl 7.1 (-0.7 to 14.9) [3] 7.5 ( ) [75] ns 9.6 ( ) [3] 8.4 ( ) [75] ns Vascular injury 2.4 (-2.2 to 7) [1] 2.5 ( ) [25] ns 4.5 (3 5.9) [1] 2.9 ( ) [25] Nerve injury 4.8 (-1.7 to 11.2) [2] 5.4 (4 6.8) [54] ns 8.8 ( ) [2] 7.1 ( ) [54] ns frx fracture, disl dislocation, ns not significant

5 What are the predictors of scapula fractures in high impact? 161 Table 2 Outcome measures of the groups with and without scapula fractures Variable ISS not adjusted ISS adjusted Scapula frx (?) Scapula frx (-) P-value Scapula frx (?) Scapula frx (-) P-value LOS in ED (min) 215 ± ± 88 \ c 211 ± ± 89 \ a Hospitalization day 12.6 ± ± 11.3 ns 15.5 ± ± 13.9 \ a ICU admission rate (%) (n) 19 (8) 10.4 (104) ns 30 (8) 17.9 (104) \ b LOS in ICU (day) 3.2 ± ± 6.2 ns 5.5 ± ± 8.6 \ c Total number of injuries 4.5 ± ± 1.7 \ a 5.4 ± ± 2 \ a AIS (head and neck) 1.4 ± ± 1.7 ns 1.7 ± ± 1.9 ns AIS (face) 0.2 ± ± 0.6 ns 0.3 ± ± 0.7 ns AIS (chest) 2.1 ± ± 1.4 \ a 2.6 ± ± 1.6 \ a AIS (abdomen) 0.7 ± ± 1.3 ns 1.0 ± ± 1.4 ns AIS (extremity) 2.1 ± ± 1.3 \ c 2.2 ± ± 1.4 \ c AIS (skin) 1.0 ± ± 0.4 ns 1.0 ± ± 0.3 ns ISS 19.0 ± ± 8.7 \ a GCS 13.3 ± ± 3.1 ns 12.4 ± ± 3.8 ns Operation (%) (n) 41.4 (11) 26.2 (412) 0.05 b 33.1 (11) 43.9 (412) \ b Died (%) (n) 8.7 (6) 14.3 (87) ns 21.5 (6) 16.9 (87) b ns not significant a Independent samples t-test b v 2 c Mann Whitney U-test Table 3 Predictors of scapula fracture in high-impact blunt trauma patients (HIBTP) using logistic regression analysis Characteristics P-value Odds ratio 95% CI Age Rib fracture Hemo and/or pneumothorax Pulmonary contusion Clavicle fracture and/or dislocation Humerus fracture and/or dislocation Variables entered in step 1: age, rib fracture, hemo and/or pneumothorax, pulmonary contusion, humerus fracture and/or dislocation, clavicle fracture and/or dislocation with scapula fracture to those without fracture, significant accompanying injuries were rib fractures, hemo and/or pneumothorax, pulmonary contusion, clavicle fracture and/ or dislocation, humerus fracture and/or dislocation, and thoracal fracture and/or dislocation (Table 1). After adjusting for the ISS, additional to these injuries, patients with scapula fracture have significantly more maxillofacial fractures, splenic injury, liver injury, pelvic fracture, cervical fracture and/or dislocation, and vascular injury. On the other hand, intracranial hemorrhage, brain contusion, kidney injury, and femur fracture and/or dislocation were lower in patients with scapula fracture compared to those without fracture (Table 1). We also found that predictors of scapula fracture were age, pulmonary contusion, clavicle fracture, and humerus fracture (Table 3). Veysi et al. [10] demonstrated that patients with scapula fractures have more severe underlying chest injuries and overall ISS, but scapula fractures were not related to outcomes such as mortality, hospital stay, ICU admission, or length of ICU stay. Stephens et al. [9] and Weening et al. [11] found 77 and 44% reductions in the risk of mortality with the presence of a scapular fracture, respectively. Weeinig et al. [11] found this result surprising, given the increased risk of chest, spleen, and liver injuries in patients with concomitant scapula fractures. They contributed lower mortality to the lower incidence of head injuries in the scapular fracture group [11]. Baldwin et al. [7] later expressed that, when the ISS was taken into account, there was less mortality in the scapular fractures compared to the controls. Recently in a research among patients injured during motor vehicle crashes, it was suggested that occupants with a shoulder injury other than the scapula were significantly more likely to have a fatal outcome, although there was no difference in the ISS [13]. We demonstrated 39.2% reduction in the risk of mortality in patients with scapular fracture compared to controls (Table 2). However, after the ISS was factored in, the mortality was significantly raised in the group with scapula fracture compared to the group without fracture, and this caused a 27.2% increase in the risk of mortality in patients with scapular fracture compared to controls (Table 2).

6 162 M. Uzkeser et al. The most common area of the scapula to be fractured is the body, ranging from 56 to 80%; % sustained fractures of the glenoid and neck of the scapula and 2 9.8% had fractures of the coracoids process and acromion [2, 10, 12]. Baldwin et al. [7] explained that the acromion and coracoid process fractures were associated with greater rates of upper extremity injury. However, these differences were not statistically significant after correction [7]. We observed that fractures in the glenoid and scapular neck occurred higher than in the body region. We also found that patients with scapular body fractures stayed for significantly longer in the ICU and that their rate of maxillofacial injury was also higher. There are only a few studies investigating missed scapula fractures. The incidence of missed scapula fractures ranges between 12.5 and 43% [3, 6, 17]. In a previous study in which only chest X-ray was available for diagnosing, it was found that the scapular fracture was not included on the initial chest radiograph in 19%, the scapular fracture was present on the initial supine chest radiograph but simply overlooked in 72%, and it was obscured by superimposed structures or artifacts in 9% [3]. Tadros et al. [6] found that the AIS for the chest scored significantly higher among the missed scapula fractures group than among the control group, but there was no statistically significant difference in the ISS between the two groups. The missed scapula fractures group had significantly higher numbers of associated serious chest injuries, pulmonary complications, and days spent in the ICU. They attributed the delayed diagnosis of scapular fractures to the extensive chest injuries overshadowing the scapula on the chest trauma radiographs, inappropriately performed computer tomography, or an unusual mechanism of injury [6]. We found that patients with and without missed fractures differed significantly in these parameters: the seniority of the ED doctor and the consultation ratio for orthopedics were lower, AIS extremity, ISS, and brain contusion were higher in patients with missed scapula fractures. Our main limitation is that our study may not have been sufficiently large to make a conclusion about associations between scapular fractures and other injuries. The patient population was limited to the catchment area of one hospital. So, the differences may not be generalized to the overall trauma population. Although we sought for missed scapula fractures, there is a possibility that some missed fractures actually remained missed. In conclusion, our findings that arouse intention as follows. (1) Patients with scapula fractures were significantly older than those without fractures. (2) Without adjusting for the ISS, thoracal region injuries were significantly higher in patients compared to the control group. After the ISS was factored in, significant differences between injuries of these two groups increased, especially for injuries of body parts other than the thorax. (3) Previous studies reported that the presence of a scapular fracture was associated with a lower mortality [11, 12]. Our evaluations showed that, after matching for the ISS, the mortality of patients with scapula fracture was significantly higher than the control group. (4) We observed that fractures in the glenoid and scapular neck occurred higher than in the body region. (5) We found that the seniority of the ED doctor, consultation ratio for orthopedics in the ED, and the existence of brain contusion were important parameters for missing scapula fractures in the ED. Conflict of interest References None. 1. Armstrong CP, Van der Spuy J. The fractured scapula: importance and management based on a series of 62 patients. Injury. 1984;15: Scavenius M, Sloth C. Fractures of the scapula. Acta Orthop Belg. 1996;62(3): Harris RD, Harris JH Jr. The prevalence and significance of missed scapular fractures in blunt chest trauma. AJR Am J Roentgenol. 1988;151(4): Wilber MC, Evans EB. Fractures of the scapula. An analysis of forty cases and a review of the literature. J Bone Joint Surg Am. 1977;59(3): Imatani RJ. Fractures of the scapula: a review of 53 fractures. J Trauma. 1975;15(6): Tadros AM, Lunsjo K, Czechowski J, Abu-Zidan FM. Causes of delayed diagnosis of scapular fractures. Injury. 2008;39(3): Baldwin KD, Ohman-Strickland P, Mehta S, Hume E. Scapula fractures: a marker for concomitant injury? A retrospective review of data in the National Trauma Database. J Trauma. 2008;65(2): Zdravkovic D, Damholt VV. Comminuted and severely displaced fractures of the scapula. Acta Orthop Scand. 1974;45(1): Stephens NG, Morgan AS, Corvo P, Bernstein BA. Significance of scapular fracture in the blunt-trauma patient. Ann Emerg Med. 1995;25: Veysi VT, Mittal R, Agarwal S, Dosani A, Giannoudis PV. Multiple trauma and scapula fractures: so what? J Trauma. 2003;55: Weening B, Walton C, Cole PA, Alanezi K, Hanson BP, Bhandari M. Lower mortality in patients with scapular fractures. J Trauma. 2005;59(6): Salimi J, Khaji A, Karbakhsh M, Saadat S, Eftekhar B. Scapular fracture: lower severity and mortality. Sao Paulo Med J. 2008;126(3): Coimbra R, Conroy C, Tominaga GT, Bansal V, Schwartz A. Causes of scapula fractures differ from other shoulder injuries in occupants seriously injured during motor vehicle crashes. Injury. 2010;41(2): Cole PA. Scapula fractures. Orthop Clin North Am. 2002;33: McGahan JP, Rab GT, Dublin A. Fractures of the scapula. J Trauma. 1980;20: Ideberg R, Grevsten S, Larsson S. Epidemiology of scapular fractures. Incidence and classification of 338 fractures. Acta Orthop Scand. 1995;66: Emet M, Saritas A, Acemoglu H, Aslan S, Cakir Z. Predictors of missed injuries in hospitalized trauma patients in the emergency department. Eur J Trauma Emerg Surg. 2010;36(6):