- Email: [email protected]
Significance of Scapular Fracture in the Blunt-Trauma Patient?
GENERAL C L I N I C A L I N V E S T I G A T I O N / B R I E F
REPORT
ignificance of Scapular Fracture in the Blunt-Trauma Patient From the University of Connecticut Ir~tegrated Residency Program* and the Departments of Surgery '~and Pediatrics~, University of Connecticut 5chooI of Medicine, and 5t Francis Hospital and Medical Center II, Hartford, Connecticut; and the Stamford Hospital~, Stamford,
Natalie G Stephens, MD*
Study objective: Todetermine the significance of scapular
Anthony S Morgan, MD, FACS*"
fractures in blunt-trauma patients compared with blunt-trauma patients without scapular fractures.
Phil Corvo, MD ~ Bruce A Bernstein, PhD*§
Design: Retrospective chart review of 11,500 blunt-trauma patients with a control group matched for age, sex, and mochanism of injury.
Connecticut.
Receivedfor publication September30, 1993. Revisions received April 29, 1994, and 'April 19, 1995. Accepted for publication May II, 1995. Presented at the Connecticut 5ociety of the American Board of Surgeons
Annual Meeting, Cromwell, Connecticut, December I992. Copyright © by the American College of Emergency Physicians.
Setting: Two Level I trauma centers.
Participants: Ninety-two blunt-trauma patients with scapular fractures and 81 control patients. Results: Mortality, neurovascular injury, and injury severity scores were compared for blunt-trauma patients with scapular fractures with those of the control group. Analysis revealed a 1% incidence of scapular fractures in blunt trauma with no neurovascular injury and no mortality. Scapular fractures were associated with thoracic injury in 49% of the patients, compared with 6% in the control group (difference, 43%; 95% confidence interval, 31.6 to 51.4; P<.001, Fisher's exact test). Conclusion: Scapular fractures are not a significant marker of greater mortality or of neurovascular morbidity in blunt-trauma patients. [Stephens NG, Morgan AS, Corvo P, Bernstein BA: Significance of scapular fracture in the blunt-trauma patient. Ann EmergMed October 1995;26:439-442.]
OCTOBER 1995 26:4 ANNALS OF EMERGENCYMEDICrNE
439
SCAPULAR FRACTURES
5tephens et al
INTRODUCTION
Scapular fractures are uncommon and usually result from major blunt trauma. The scapulae are well protected by muscle groups, so the force required to break them is great. Scapular fracture was shown to be associated with ipsilateral chest-wall injury by Imatani. s Data from studies by McGahan, Rub, and Dublin2; Tomaszek3; and Thompson et al 4 suggest that scapular fractures are associated with neurovascular injury. Armstrong and Van der Spuy 5 reported that scapular fracture in blunt trauma carried a 9.7% mortality rate. We conducted this study to determine the mortality and morbidity associated with scapular fracture in blunt-trauma patients.
Objective comparisons of these groups were based on associated injuries and Injury Severity Scores (ISSs). ISS was determined by a trauma registry nurse during each patient's hospitalization. In the calculation of ISS, lacerations were not included because they were not consistently documented. The difference in mean ISS between the scapular-fracture and control groups was tested with a two-sided Student t test. Differences in mortality and percentage of associated injuries between the two groups were explored with Fisher's exact test. In addition, fractures were categorized in six groups based on injury location (acromion, body, glenoid, neck, spine, coracoid process). The difference in mean ISS according to fracture site was tested with one-way ANOVA. P values of less than .05 were considered statistically significant for all tests.
MATERIALS AND METHODS
A retrospective chart review of 11,500 blunt-trauma cases identified 94 patients with scapular fractures at two Level I trauma centers from January 1, 1985, to October 1, 1991. The charts of 92 of these patients were available for review, along with select radiographs. Special attention was paid to age, sex, mechanism of injury, fracture site, and associated injuries. A control group matched for age, sex, and mechanism of injury was randomly selected from the trauma registry from the same time period. Eighty-one controls were obtained for motor vehicle, motorcycle, and fall injuries for comparison with the scapular-fracture patients. Controls were not available for bicycle, crush, or assault victims. Table.
Associated injuhes, ISSs, and deaths in the study groups. Associated Injuries
No. from ScapularFracture Group (%)*
No. from Control Group (%P
Rib fracture 34 (37) 3 (4) Extremity fracture 19 (21) 28 (35) Hemothorax/ 18 (20) 3 (4) pneumothorax/ pulmonary contusion Head/neck 17 (18) 35 (43) Clavicle 13 (16) 1 (1) Pelvic girdle 4 (5) 8 (10) Intraabdomina] 2 (3) 2 (3) Combined thoracic 45 (49) 5 (6) injury ISS (mean+SD) 5.9+5.7 5.6+6.0 Death 0 3 *Seventyof the 92 scapular-fracturepatients(76%)had associatedinjuries. tSkty-aight of the 81 controlpatientshad associatedinjuries, *Fisher'sexacttest. §Student'sttest.
44 0
P
<001' NS* <.01'
NS* <.01' NS * NS ~ <.001' NS § NS ~
RESULTS
The incidence of scapular fractures in blunt trauma was .8% (94 of 11,500). Of the 92 patients studied, 65 were men. Thirty-eight fractures were on the right and 54 on the left; there were no bilateral fractures. The patients ranged in age from 13 to 92 years; however, most patients were younger than 25 years. Motor vehicle, pedestrian, and motorcycle accidents were the most common causes of injury (40, 9, and 7, respectively). These were followed by falls, bicycle accidents, and assaults (27, 13, and 6, respectively). Seventy of the 92 patients had associated injuries (76%). The most common of these injuries was rib fracture (49%), followed by extremity fracture (27%), hemothorax/pneumothorax/pulmonary contusion (26%), head and neck injury (24%), clavicle fracture (19%), and pelvic girdle (6%) and intraabdominal injuries (3%). The Table shows our comparison of the associated injuries of the scapular-fracture patients and the matched controls. Compared with the scapular-fracture group, the control group had an 84% associated-injury rate with many fewer thoracic injuries (P<.001), although these patients had a higher head- and neck-injury ratio. Peripheral neurotogic injury, great-vessel injury, and peripheral-artery injury were not found in the scapular-fracture group. In this study ISSs ranged from 0 to 32. The mean ISS (±SD) for the scapular-fracture patients was 5.9+5.7, compared with 5.6+6.0 for the control group, a nonsignificant difference. There were three deaths in the control group and none in the scapular-fracture patients. There were no neurovascular injuries among the scapular-fracture patients. Three patients underwent aortography for investigation of great-vessel inju® Each of these
ANNALS OF EMERGENCYMEDICINE 26:4 OCTOBER1995
SCAPULAR FRACTURES
5tephens et al
patients had a widened mediastinum or an apical cap along with the scapular fracture as an indication for aortography. There were no aortic injuries. No aortograms were performed[ in the control group. The distribution of fracture sites was similar to those reported by Armstrong. The fractures occurred in the body (n=53), the glenoid (n=17), the acromion (n=10), the coracoid (n=5), the neck (n=4), and the spine (n=3). ISS scores were not statistically different by fracture location (one-way ANOVA, P>.67); therefore no scapula>fracture location correlated with a greater ISS. DISCUSSION
These results demonstrate that scapular fractures are rare in blunt trauma, with an incidence of .2% to 3%. 1'2'4'5 This study and others have shown a likelihood that 76% to 100% of patients will have associated injury. Imatani 1 and McGahan, Rab, and Dublin 2 found rib and clavicle fractures to be the most common associated injuries in cases of scapular fracture. The data of this study document thoracic injury, rib fracture, clavicle fracture, pulmonary con>usion, and hemopneumothorax to be the most common associated injuries in cases of scapular fracture. With the low morbidity and mortality of scapular fracture, associated injury is what must be carefully defined. This conclusion differs from those presented by Thompson et al4 and Armstrong and Van der Spuy. 5 The prospective study by Armstrong of 62 patients reported a 9 . 7 % mortality rate associated with scapular fractures and one vascular injury. In the retrospective analysis by Thompson et al of 56 patients with scapular fractures resulting from blunt trauma, eight died (14.3% mortality), seven ~ad brachial plexus injury, and six had subclavian, axil!ary or brachial arterial injury. The findings of Thompso n et al are not supported when the data from this study, and those of Imatanil; McGahan, Rab, and Dublin2; and Armstrong and Van der Spuy5 are analyzed. Collectively, these four papers document only two vascular injuries in 328 scapular-fracture patients. Thompson et al also failed to break down the associated injuries of patients with arterial injuries in their search for a concomitant marker of vascular injury. On the basis of these data, it maybe concluded that no association of scapular fracture wit h vascular injury exists. Brachial plexus injuries must be considered in patients with scapular fracture. Imatani 1 reported four brachial plexus injur!es and Thompson et al 4 reported seven, but in neither study was "brachial plexus injury" defined.
OCTOBER 1995 26:4
ANNALS OF EMERGENCY MEGrCINE
McGahan, Rab, and Dublin 2 had the largest series: 121 patients, 5 with permanent peripheral-nerve injury. Four of these patients had acromion fractures. McGahan and colleagues2 concluded that the mechanism of injury required to break the acromion involves depression of the shoulder and contralateral flexion of the neck, creating a strain on the brachial plexus. Their results differ from those of this study. No neurologic injuries were found in connection with the 10 acromion fractures in this study, compared with four neurologic deficits in 11 acromion fractures in the McGahan data. However, this study has the lowest incidence of neurologic injury (0 of 92) when compared with those of Thompson (7 of 56), Imatani (4 of 52) and Armstrong (2 of 62). In two case reports of neurovascular injury with scapular fracture, the arterial and neurologic deficits were apparent on initial examination. 5,6 Loss of pulses and sensorimotor deficit prompted arteriography or surgical exploration, not the radiographic finding of a scapular fracture. The idea that scapular fracture location is associated with neurologic injury previously suggested by McGahan was also not reproduced in this study. Fractures were divided into six locations with distributions similar to those of previous studies, r Thompson et al4 used a different classification of scapular-fracture severity and also found no statistically significant difference among the three classes of fractures described with regard to ISS, neurovascular injury, and mortality. This series does not demonstrate scapular fractures to be as severe as Armstrong and colleagues concluded. Armstrong described patients with first rib fractures and scapular fractures to be a marker for severe injury in a 1985 prospective study. The Armstrong data are from 62 patients with both blunt and penetrating trauma. Five of 62 patients in that study had first-rib and scapular fractures. Among these five patients there were two deaths, one subclavian-artery injury, and two brachial plexus injuries. However, mechanism of injury, blunt or penetrating, is not clear. Neither Imatani nor McGahan et al separate first-rib fractures from all other rib fractures. In our series, four patients sustained first-rib and scapular fracture; three each had an ISS of 9, and the score in the fourth patient was 24. Falls accounted for the three less severe injuries, and multiple rib fractures were their only associated injuries. The most severely injured patient was involved in a motor vehicle accident and sustained a closed-head injury and clavicle, scapular and rib fractures. He underwent aortography for a widened mediastinum; it showed no abnormality. He was discharged home in good condition after hospitafiration.
4 4 1
SCAPULAR FRACTURES
5teFhens et al
The retrospective nature of this study is its basic weakness. We relied solely on chart documentation for accuracy. Subtle neurologic injury that resolved during hospitalization may account for the absence of such injuries in this study. Intimal flap injuries that may have required subsequent intervention are not identified because all patients did not undergo arteriography. New prospective studies with control groups may offer a promising marker for physicians.
Reprint no. 47/1/67354 Address for reprints: AnthonyS Morgan,MD, FACS SaintFrancisHospitaland MedicalCenter 114WoodlandStreet Hartford,Connecticut06105-1299
CONCLUSION
This study demonstrates that scapular fracture is clearly associated with thoracic injury and that the associated thoracic injuries are clinically more significant than the scapular fracture. Scapular fracture alone is not a significant marker for mortality or neurovascular injury. We found that the location of the scapular fracture did not correlate with a greater ISS score. No data, in this study or in most other studies reviewed, supported an association of vascular injury with scapular fracture. Brachial plexus injury is perhaps more common than demonstrated in this study. REFERENCES 1. Imataei RJ: Fracturesof the scapula:A review of 53 fractures. J Trauma1975;15:473-478. 2. McGabanJP, Rab GT, Dublin A: Fracturesof the scapula. J Trauma1980;20:880-883. 3. TomaszakDE: Combinedsubclavian artery and brachial plexus injuries from blunt upperextremity trauma. J Trauma1984;24: 161-t63. 4. ThompsonDA, FlynnTC, Mil[er PW, et al: The significance of scapularfractures. J Trauma 1985;25:974-977. 5. ArmstrongCP,Van der Spuy J: The fractured scapula: [mportanceand managementbasedon a series of 62 patients. Injury1984;15:324-329. 5. HalpernAA, Joseph R, PageJ, et al: Subclavianartery injury and fracture of the scapula. JACEP1979;8:19-20. 7. Harris RD, Harris JH: The prevalenceand significance of missed scapularfractures in blunt chest trauma. AJRAm J Roentgenal1988;151:747-758.
442
ANNALS OF EMERGENCY MEDICINE
28:4
OCTOBER1995
REPORT
ignificance of Scapular Fracture in the Blunt-Trauma Patient From the University of Connecticut Ir~tegrated Residency Program* and the Departments of Surgery '~and Pediatrics~, University of Connecticut 5chooI of Medicine, and 5t Francis Hospital and Medical Center II, Hartford, Connecticut; and the Stamford Hospital~, Stamford,
Natalie G Stephens, MD*
Study objective: Todetermine the significance of scapular
Anthony S Morgan, MD, FACS*"
fractures in blunt-trauma patients compared with blunt-trauma patients without scapular fractures.
Phil Corvo, MD ~ Bruce A Bernstein, PhD*§
Design: Retrospective chart review of 11,500 blunt-trauma patients with a control group matched for age, sex, and mochanism of injury.
Connecticut.
Receivedfor publication September30, 1993. Revisions received April 29, 1994, and 'April 19, 1995. Accepted for publication May II, 1995. Presented at the Connecticut 5ociety of the American Board of Surgeons
Annual Meeting, Cromwell, Connecticut, December I992. Copyright © by the American College of Emergency Physicians.
Setting: Two Level I trauma centers.
Participants: Ninety-two blunt-trauma patients with scapular fractures and 81 control patients. Results: Mortality, neurovascular injury, and injury severity scores were compared for blunt-trauma patients with scapular fractures with those of the control group. Analysis revealed a 1% incidence of scapular fractures in blunt trauma with no neurovascular injury and no mortality. Scapular fractures were associated with thoracic injury in 49% of the patients, compared with 6% in the control group (difference, 43%; 95% confidence interval, 31.6 to 51.4; P<.001, Fisher's exact test). Conclusion: Scapular fractures are not a significant marker of greater mortality or of neurovascular morbidity in blunt-trauma patients. [Stephens NG, Morgan AS, Corvo P, Bernstein BA: Significance of scapular fracture in the blunt-trauma patient. Ann EmergMed October 1995;26:439-442.]
OCTOBER 1995 26:4 ANNALS OF EMERGENCYMEDICrNE
439
SCAPULAR FRACTURES
5tephens et al
INTRODUCTION
Scapular fractures are uncommon and usually result from major blunt trauma. The scapulae are well protected by muscle groups, so the force required to break them is great. Scapular fracture was shown to be associated with ipsilateral chest-wall injury by Imatani. s Data from studies by McGahan, Rub, and Dublin2; Tomaszek3; and Thompson et al 4 suggest that scapular fractures are associated with neurovascular injury. Armstrong and Van der Spuy 5 reported that scapular fracture in blunt trauma carried a 9.7% mortality rate. We conducted this study to determine the mortality and morbidity associated with scapular fracture in blunt-trauma patients.
Objective comparisons of these groups were based on associated injuries and Injury Severity Scores (ISSs). ISS was determined by a trauma registry nurse during each patient's hospitalization. In the calculation of ISS, lacerations were not included because they were not consistently documented. The difference in mean ISS between the scapular-fracture and control groups was tested with a two-sided Student t test. Differences in mortality and percentage of associated injuries between the two groups were explored with Fisher's exact test. In addition, fractures were categorized in six groups based on injury location (acromion, body, glenoid, neck, spine, coracoid process). The difference in mean ISS according to fracture site was tested with one-way ANOVA. P values of less than .05 were considered statistically significant for all tests.
MATERIALS AND METHODS
A retrospective chart review of 11,500 blunt-trauma cases identified 94 patients with scapular fractures at two Level I trauma centers from January 1, 1985, to October 1, 1991. The charts of 92 of these patients were available for review, along with select radiographs. Special attention was paid to age, sex, mechanism of injury, fracture site, and associated injuries. A control group matched for age, sex, and mechanism of injury was randomly selected from the trauma registry from the same time period. Eighty-one controls were obtained for motor vehicle, motorcycle, and fall injuries for comparison with the scapular-fracture patients. Controls were not available for bicycle, crush, or assault victims. Table.
Associated injuhes, ISSs, and deaths in the study groups. Associated Injuries
No. from ScapularFracture Group (%)*
No. from Control Group (%P
Rib fracture 34 (37) 3 (4) Extremity fracture 19 (21) 28 (35) Hemothorax/ 18 (20) 3 (4) pneumothorax/ pulmonary contusion Head/neck 17 (18) 35 (43) Clavicle 13 (16) 1 (1) Pelvic girdle 4 (5) 8 (10) Intraabdomina] 2 (3) 2 (3) Combined thoracic 45 (49) 5 (6) injury ISS (mean+SD) 5.9+5.7 5.6+6.0 Death 0 3 *Seventyof the 92 scapular-fracturepatients(76%)had associatedinjuries. tSkty-aight of the 81 controlpatientshad associatedinjuries, *Fisher'sexacttest. §Student'sttest.
44 0
P
<001' NS* <.01'
NS* <.01' NS * NS ~ <.001' NS § NS ~
RESULTS
The incidence of scapular fractures in blunt trauma was .8% (94 of 11,500). Of the 92 patients studied, 65 were men. Thirty-eight fractures were on the right and 54 on the left; there were no bilateral fractures. The patients ranged in age from 13 to 92 years; however, most patients were younger than 25 years. Motor vehicle, pedestrian, and motorcycle accidents were the most common causes of injury (40, 9, and 7, respectively). These were followed by falls, bicycle accidents, and assaults (27, 13, and 6, respectively). Seventy of the 92 patients had associated injuries (76%). The most common of these injuries was rib fracture (49%), followed by extremity fracture (27%), hemothorax/pneumothorax/pulmonary contusion (26%), head and neck injury (24%), clavicle fracture (19%), and pelvic girdle (6%) and intraabdominal injuries (3%). The Table shows our comparison of the associated injuries of the scapular-fracture patients and the matched controls. Compared with the scapular-fracture group, the control group had an 84% associated-injury rate with many fewer thoracic injuries (P<.001), although these patients had a higher head- and neck-injury ratio. Peripheral neurotogic injury, great-vessel injury, and peripheral-artery injury were not found in the scapular-fracture group. In this study ISSs ranged from 0 to 32. The mean ISS (±SD) for the scapular-fracture patients was 5.9+5.7, compared with 5.6+6.0 for the control group, a nonsignificant difference. There were three deaths in the control group and none in the scapular-fracture patients. There were no neurovascular injuries among the scapular-fracture patients. Three patients underwent aortography for investigation of great-vessel inju® Each of these
ANNALS OF EMERGENCYMEDICINE 26:4 OCTOBER1995
SCAPULAR FRACTURES
5tephens et al
patients had a widened mediastinum or an apical cap along with the scapular fracture as an indication for aortography. There were no aortic injuries. No aortograms were performed[ in the control group. The distribution of fracture sites was similar to those reported by Armstrong. The fractures occurred in the body (n=53), the glenoid (n=17), the acromion (n=10), the coracoid (n=5), the neck (n=4), and the spine (n=3). ISS scores were not statistically different by fracture location (one-way ANOVA, P>.67); therefore no scapula>fracture location correlated with a greater ISS. DISCUSSION
These results demonstrate that scapular fractures are rare in blunt trauma, with an incidence of .2% to 3%. 1'2'4'5 This study and others have shown a likelihood that 76% to 100% of patients will have associated injury. Imatani 1 and McGahan, Rab, and Dublin 2 found rib and clavicle fractures to be the most common associated injuries in cases of scapular fracture. The data of this study document thoracic injury, rib fracture, clavicle fracture, pulmonary con>usion, and hemopneumothorax to be the most common associated injuries in cases of scapular fracture. With the low morbidity and mortality of scapular fracture, associated injury is what must be carefully defined. This conclusion differs from those presented by Thompson et al4 and Armstrong and Van der Spuy. 5 The prospective study by Armstrong of 62 patients reported a 9 . 7 % mortality rate associated with scapular fractures and one vascular injury. In the retrospective analysis by Thompson et al of 56 patients with scapular fractures resulting from blunt trauma, eight died (14.3% mortality), seven ~ad brachial plexus injury, and six had subclavian, axil!ary or brachial arterial injury. The findings of Thompso n et al are not supported when the data from this study, and those of Imatanil; McGahan, Rab, and Dublin2; and Armstrong and Van der Spuy5 are analyzed. Collectively, these four papers document only two vascular injuries in 328 scapular-fracture patients. Thompson et al also failed to break down the associated injuries of patients with arterial injuries in their search for a concomitant marker of vascular injury. On the basis of these data, it maybe concluded that no association of scapular fracture wit h vascular injury exists. Brachial plexus injuries must be considered in patients with scapular fracture. Imatani 1 reported four brachial plexus injur!es and Thompson et al 4 reported seven, but in neither study was "brachial plexus injury" defined.
OCTOBER 1995 26:4
ANNALS OF EMERGENCY MEGrCINE
McGahan, Rab, and Dublin 2 had the largest series: 121 patients, 5 with permanent peripheral-nerve injury. Four of these patients had acromion fractures. McGahan and colleagues2 concluded that the mechanism of injury required to break the acromion involves depression of the shoulder and contralateral flexion of the neck, creating a strain on the brachial plexus. Their results differ from those of this study. No neurologic injuries were found in connection with the 10 acromion fractures in this study, compared with four neurologic deficits in 11 acromion fractures in the McGahan data. However, this study has the lowest incidence of neurologic injury (0 of 92) when compared with those of Thompson (7 of 56), Imatani (4 of 52) and Armstrong (2 of 62). In two case reports of neurovascular injury with scapular fracture, the arterial and neurologic deficits were apparent on initial examination. 5,6 Loss of pulses and sensorimotor deficit prompted arteriography or surgical exploration, not the radiographic finding of a scapular fracture. The idea that scapular fracture location is associated with neurologic injury previously suggested by McGahan was also not reproduced in this study. Fractures were divided into six locations with distributions similar to those of previous studies, r Thompson et al4 used a different classification of scapular-fracture severity and also found no statistically significant difference among the three classes of fractures described with regard to ISS, neurovascular injury, and mortality. This series does not demonstrate scapular fractures to be as severe as Armstrong and colleagues concluded. Armstrong described patients with first rib fractures and scapular fractures to be a marker for severe injury in a 1985 prospective study. The Armstrong data are from 62 patients with both blunt and penetrating trauma. Five of 62 patients in that study had first-rib and scapular fractures. Among these five patients there were two deaths, one subclavian-artery injury, and two brachial plexus injuries. However, mechanism of injury, blunt or penetrating, is not clear. Neither Imatani nor McGahan et al separate first-rib fractures from all other rib fractures. In our series, four patients sustained first-rib and scapular fracture; three each had an ISS of 9, and the score in the fourth patient was 24. Falls accounted for the three less severe injuries, and multiple rib fractures were their only associated injuries. The most severely injured patient was involved in a motor vehicle accident and sustained a closed-head injury and clavicle, scapular and rib fractures. He underwent aortography for a widened mediastinum; it showed no abnormality. He was discharged home in good condition after hospitafiration.
4 4 1
SCAPULAR FRACTURES
5teFhens et al
The retrospective nature of this study is its basic weakness. We relied solely on chart documentation for accuracy. Subtle neurologic injury that resolved during hospitalization may account for the absence of such injuries in this study. Intimal flap injuries that may have required subsequent intervention are not identified because all patients did not undergo arteriography. New prospective studies with control groups may offer a promising marker for physicians.
Reprint no. 47/1/67354 Address for reprints: AnthonyS Morgan,MD, FACS SaintFrancisHospitaland MedicalCenter 114WoodlandStreet Hartford,Connecticut06105-1299
CONCLUSION
This study demonstrates that scapular fracture is clearly associated with thoracic injury and that the associated thoracic injuries are clinically more significant than the scapular fracture. Scapular fracture alone is not a significant marker for mortality or neurovascular injury. We found that the location of the scapular fracture did not correlate with a greater ISS score. No data, in this study or in most other studies reviewed, supported an association of vascular injury with scapular fracture. Brachial plexus injury is perhaps more common than demonstrated in this study. REFERENCES 1. Imataei RJ: Fracturesof the scapula:A review of 53 fractures. J Trauma1975;15:473-478. 2. McGabanJP, Rab GT, Dublin A: Fracturesof the scapula. J Trauma1980;20:880-883. 3. TomaszakDE: Combinedsubclavian artery and brachial plexus injuries from blunt upperextremity trauma. J Trauma1984;24: 161-t63. 4. ThompsonDA, FlynnTC, Mil[er PW, et al: The significance of scapularfractures. J Trauma 1985;25:974-977. 5. ArmstrongCP,Van der Spuy J: The fractured scapula: [mportanceand managementbasedon a series of 62 patients. Injury1984;15:324-329. 5. HalpernAA, Joseph R, PageJ, et al: Subclavianartery injury and fracture of the scapula. JACEP1979;8:19-20. 7. Harris RD, Harris JH: The prevalenceand significance of missed scapularfractures in blunt chest trauma. AJRAm J Roentgenal1988;151:747-758.
442
ANNALS OF EMERGENCY MEDICINE
28:4
OCTOBER1995