A review of traumatic airway injuries: Potential implications for airway assessment and management☆

Introduction 

The absolute priority in the initial resuscitation of any trauma is to secure a patent airway and provide adequate ventilation.1 Even under the best of conditions, securing a patent airway may be challenging in a trauma patient, but it is particularly difficult in patients with traumatic injuries to the airway,2, 12 in whom complex airway management techniques may be required.11, 13 Traumatic airway injury (TAI) requires high levels of awareness to be diagnosed and advanced airway management skills. Compounding the technical challenges of securing a definitive airway (defined as a tube present in trachea with the cuff inflated)1 in these patients, any airway management intervention may prove useless if the injury is located distal to the inflated cuff. In fact, efforts to intubate the trachea can themselves augment the airway injury and lead to a fatal outcome.2, 10

In spite of the clinical relevance, there is a lack of literature reporting large series of patients with TAI, data on its frequency and the specific airway management interventions used to treat these patients. While some studies report that airway trauma is rare and occurs in less than 1% of all trauma patients,6 many of these patients present with subtle clinical findings that may be easily missed during the initial presentation. This is important because undiagnosed airway injuries are sometimes fatal.7, 13 Due to the recent improvements in pre-hospital care, the initial airway assessment and management is frequently performed by paramedics and ambulance personnel at the scene of the accident, which may play a role in the outcome of these patients.

Many concerns on airway trauma remain under investigated. This study was designed to retrospectively evaluate all traumatic airway injuries seen at a large Regional Trauma Centre and focus in determining how urgent these cases were considered initially, the airway management interventions used during the initial assessment to secure a definitive airway and the potential impact of these interventions in determining patients’ outcome.

Methods 

After approval by the Research and Ethical Boards, patients were identified from the Trauma Database of Sunnybrook Health Sciences Centre, Toronto, Ontario, Canada. Charts were reviewed from all trauma patients presenting with airway injuries from July 1989 to June 2004. Inclusion criteria consisted of the diagnosis of traumatic airway injury (pharynx, larynx, trachea and main bronchi). Patients deemed dead on arrival were excluded. Details regarding the mechanism of injury, clinical presentation, urgency of presentation, requirement and type of definitive airway obtained, injury severity scores, associated injuries, diagnostic tests and initial treatment were collected. Pre-hospital notes, referring hospital records, anaesthesia and ICU charts were revised.

For the purpose of this study, pharynx, larynx and cervical trachea were considered upper airways while thoracic trachea and main stem bronchi were defined as lower airways. We defined death as related with the airway injuries when this was stated in the patients’ Death Certificate or chart, or when upon reviewing all circumstances reported in the patients’ files, the authors of this study (all trauma experts) concluded that the TAI had an important causal role.

Data was presented as mean±standard deviation, or percentage, as indicated. Differences in outcome (survival or complications) associated with mechanism, clinical presentation and airway management were analysed by Chi-square test, or t-test as appropriate, accepting p<0.05 as significant.

Results 

During the study period, 12,187 trauma patients were treated at Sunnybrook Health Sciences Centre, of which 104 were diagnosed as having TAI. Of these patients, 36 suffered blunt trauma, and 68 were victims of a penetrating mechanism (incidence of 0.4% for blunt and 4.5% for penetrating trauma—p<0.001). Specific details on the mechanisms of injuries are shown in Table 1. Patients’ demographics and clinical characteristics are listed in Table 2.

Table 1. Underlying mechanisms resulting in traumatic airway injury

	Type of mechanism	Number
	Penetrating	68 (65%)
	Stab wounds	44
	GSW	15
	Shotgun wounds	4
	Other	5
	Blunt	36 (35%)
	MVC	17
	Pedestrian	6
	Bicycle	3
	Recreational vehicle	3
	Industrial accident	2
	Motorcycle crash	2
	Fall	1
	Other	2

GSW: gunshot wounds; MVC: motor vehicle collision.

Table 2. TAI patients’ demographics and clinical presentation

		Blunt	Penetrating	Overall
	Male	31 (86%)	56 (82%)	87 (82%)
	Age	37±16	36±14	36±16
	ISS	33±16*	24±10	27±13
	Symptoms at presentation
	SC emphysema	19 (53%)*	17 (25%)	36 (38%)
	External bleeding	2 (6%)*	30 (44%)	32 (30%)
	Air escape	4 (11%)*	22 (32%)	26 (24%)
	Cervical hematoma	3 (8%)	8 (12%)	11 (10%)
	Coma/VSA	5 (14%)	6 (9%)	11 (10%)
	Stridor	6 (17%)*	2 (3%)	8 (8%)
	Hoarseness	7 (19%)*	1 (1%)	8 (8%)
	Asymptomatic	0	4 (6%)	4 (4%)
	Other	7 (19%)	6 (9%)	13 (12%)
	Suicide attempts	5 (14%)*	20 (29%)	25 (24%)
	Mortality	13 (36%)*	11 (16%)	24 (23%)

* p<0.05 comparing blunt and penetrating.

The diagnosis of TAI was definitively established by surgery in 78 patients (18 blunt, 59 penetrating), fiberoptic bronchoscopy (FOB) in 11 patients (8 blunt, 3 penetrating), CT scan of the neck in 5 patients (5 blunt, 1 penetrating), FOB and CT scan in 3 (3 blunt), Coroner's autopsy report in 5 patients (1 blunt, 4 penetrating) and by direct laryngoscopy examination in 2 patients (2 blunt).

In 54 patients (52%), the definitive airway (DA) was secured before arrival to Sunnybrook by pre-hospital care personnel either at the scene of the accident or en route in 32 patients (31%), or at a referring hospital in 22 patients (21%). Forty-seven patients (45%) had a DA established at Sunnybrook, 17 patients (16%) in the Trauma Room during initial assessment and resuscitation and 30 (29%) in the Operating Room. Among the latter, DA was established either as part of the treatment of the TAI or because the patient needed surgery for injuries unrelated to the airway injury. Overall, 71 patients (68%) required a DA during initial assessment either at the scene of the accident or en route to the hospital.

The interventions used to establish a definitive airway are shown in Table 3. Difficulties in intubating the patients were charted in only 15 patients. Orotracheal intubation was the most common intervention regardless of the location in which the DA was obtained, before or after arrival to a hospital. Thirty patients (29%) required more complex interventions than simple intubation under laryngoscopic vision to secure DA, including FOB in 11 patients, surgical airway in 10 patients and temporary intubation through the wound in 9 patients. Patients that required DA management in the pre-hospital setting had a significantly higher mortality (19 deaths among 32 patients—p<0.0001) and significantly more severe injuries (ISS of 32 compared to 22—p=0.01) than the patients receiving DA in the operating room.

Table 3. Distribution of the initial airway intervention following location of intervention

	Place DA was established (n)	Oral TT	Nasal TT	FOB	Surgical airway	Wound TT	Mortality
	Field/route (32)	17	9	–	1	5	19 (60%)*
	Referring hospital (22)	17	2	–	3	–	4 (17%)*
	Trauma Room (17)	10	–	6	–	1	1 (6%)
	Operating Room (30)	15	1	5	6a	3	–
	Total (101)	59	12	11	10	9	24 (23%)

DA: definitive airway; TT: tracheal tube; FOB: fiberoptic bronchoscopy; *p<0.05 when compared to Operating Room.

a One of the patients who received initial surgical airway was operated only 48h after admission.

Three patients never required any intervention for treatment of their airway injuries which were managed expectantly. One of them had a blunt trauma of the neck with a mountain bike, and presented hoarseness, blood in oropharynx and pain in swallowing. The CT scan showed only hyoid bone and thyroid cartilage fractures. The other two patients had penetrating trauma: the first had a small gunshot wound just above the sternal notch and underwent fiberoptic bronchoscopy, which showed a small hole in the anterior trachea with exit in the antero-lateral tracheal wall. The second patient was shot in the face and the bullet crossed the posterior pharynx. The pharyngeal wound was diagnosed by CT scan. All these patients were stable, had isolated injuries, were admitted for monitoring, and were discharged home in 72h.

A total of 61 (59%) patients required tracheostomy as part of the TAI management, 18 patients after blunt trauma (50% of all blunt traumas) and 43 after penetrating trauma (63% of all victims of penetrating trauma). Fifteen patients did not undergo any form of surgical repair of their TAI, even when submitted to surgery for indications other than the airway injury. These patients were treated with tracheostomy or endotracheal intubation with the cuff inflated distal to the airway lesion.

Upper airway injuries (UAI) occurred more frequently than lower airway injuries (LAI) in both blunt and penetrating trauma groups. The blunt group had 28 patients with UAI and 8 with LAI while the penetrating group had 62 patients with UAI and 6 with LAI. The anatomical location of the airway injuries as well as overall mortality and mortality related with airway injuries is shown in Table 4. LAI had the highest mortality rate in both groups, blunt and penetrating (p<0.05).

Table 4. Distribution of mortality by mechanism type and segment injured

	Type	Location of TAI (n)	Mortality from TAI	Overall mortality
	Blunt	Upper airway injuries (28)	0	8 (29%)
		Lower airway injuries (8)	5 (63%)	5 (63%)
	Penetrating	Upper airway injuries (62)	3 (5%)	7 (11%)
		Lower airway injuries (6)	2 (33%)	4 (66%)

Mortality TAI: mainly caused by airway injury. For both groups (blunt and penetrating), p<0.05 when comparing upper and lower airways in each category of mortality.

Ten of the 24 deaths (42%) were deemed as related with the TAI. These cases are summarized in Table 5. Seven patients had LAI and all had major problems with regard to adequate ventilation and oxygenation. Of the patients who died following UAI, three were victims of penetrating trauma and arrived at Sunnybrook with absent vital signs, profound hypoxia and shock. All three patients were described as being very difficult to secure patent, definitive airways.

Table 5. Summary of 10 deaths related to airway injuries

	Mechanism of trauma	Airway injury	Main admission findings	Placement of DA	Difficulty in obtaining DA	Difficulty in ventilation	Main associated injuries	ISS	LOS
	MVC	R bronchus	Massive hemothorax, ↑ air leak in CT	Field/route	No	Yes	CHI, flail chest	38	<1d
	Crushed by bulldozer	Tracheo/bronchial	Hypoxia, ↑ air leak in CT	Field/route	Yes—multiple attempts	Yes	Flail chest	26	20d
	MVC	R bronchus	↑ Air leak in CT, desaturation	Referring hospital	No	Yes	CHI, flail chest	51	<1d
	Boat collision	R bronchus	Severe hemoptysis	Field/route	No	Yes	CHI, R massive hemothorax	50	<1d
	Pedestrian struck by train	L bronchus	Hypoxia and shock	SB TR	No	Yes	CHI, flail chest, L hemothorax	50	<1d
	GSW	Thoracic trachea	VSA	Field/route	No	Not recorded	Aortic arch	19	<1d
	Stab wound	Larynx	Hypoxia, shock and coma	Field/route	Yes—tube through the wound	No	L carotid artery and jugular vein	35	<1d
	GSW	Pharynx	VSA	Field/route	Yes—hypoxic	No	Hemo/pneumothorax and intra-abdominal injuries	17	<1d
	Stab wound	Cervical trachea	VSA	Field/route	Yes—hypoxic with intra-oral bleeding	Yes	Hypoxic brain injury and jugular vein	17	3d
	GSW	Thoracic trachea	VSA	Field/route	Not recorded	No	Innominate artery injury	21	<1d

MVC: motor vehicle collision; GSW: gunshot wound; R: right; L: left; ↑: significant; CT: chest tube; VSA: vital signs absence; SB TR: Sunnybrook Trauma Room; CHI: closed head injury.

Discussion 

There are not many studies on traumatic airway injuries. The available studies are not homogenous in their approaches. Therefore, rendering comparisons and recommendations is very challenging. While some trauma-related studies consider the pharynx part of the airways,6, 9 others restrict the airways to the larynx, trachea and bronchi.4, 5 In the present study, the pharynx was included as part of the upper airways. In clinical practice, it may be difficult during the initial assessment of a cervical trauma to accurately diagnose whether the patient had an injury to the larynx, pharynx or cervical trachea. Likewise, it might also very difficult to determine whether the clinical presentation of a TAI is due to an injury to the thoracic trachea or main bronchi.

Furthermore, to facilitate analysis in the present review, we arbitrarily opted to divide the airway injuries into upper and lower airway injuries. This classification allowed a better evaluation of the impact of the anatomical location of the injuries during the two initial steps of the ATLS protocol assessment for trauma patients: (1) secure the airway/neck stabilization and (2) provide satisfactory lung ventilation and oxygenation.

In the present study, TAI was identified in 0.9% of all trauma patients, in 0.4% of the blunt and 4.5% of the penetrating trauma patients. Similarly, the literature reports that TAI is rare and occurs in less than 1% of all trauma patients, being more frequent in penetrating trauma.4, 8 The relatively high incidence of penetrating TAI in our population might be the result of referring bias to our Trauma Centre. Of note, 20 of the penetrating TAI cases were self-inflicted injuries (29% of penetrating TAI), 18 of them being stab wounds.

In this series of patients, the initial clinical presentation of the patients with TAI varied accordingly to the underlying mechanism and location of the airway injuries. Subcutaneous emphysema was the most common presentation related to TAI in blunt trauma, followed by oral bleeding and hoarseness. In penetrating trauma, the three most common presentations were external bleeding, air leak through the wound, and subcutaneous emphysema. Although the retrospective nature of the data collection may fail to capture less prominent signs and symptoms, the different clinical presentations are consistent with the mechanism of injury. If a wound is present, then external bleeding is frequently seen. The wound itself will also allow air leak from the underlying injured structures. In contrast, in blunt traumas, often without external wounds, the air leaking from the underlying traumatised airway will collect in the subcutaneous tissues resulting in subcutaneous emphysema, common among these patients.

In the penetrating group, stab wounds represented the most common cause of TAI, followed by gunshot wounds. We noticed a high number of self-inflicted injuries associated with stab wounds to the neck in our hospital. The most frequent mechanism of blunt trauma was motor vehicle collision, followed by pedestrians.

As expected, patients sustaining blunt TAI had a higher ISS compared to the penetrating group.4 The higher severity and often multiple associated injuries account for the higher mortality in this group.

In 32 patients, a definitive airway was established at the scene of the accident. These patients suffered the most severe injuries, thus had the highest injury severity score (ISS) and predictably the highest mortality rate. These findings are consistent with a previous study by Bhojani et al. demonstrating that the urgent requirement for a patent airway is an independent predictor of mortality.4

Orotracheal intubation was the most common procedure in both the pre- and in-hospital settings. The number of nasotracheal intubations was higher in the pre-hospital setting. Surgical airways and intubation through the wound were done in both, pre- and in-hospital settings as the first approach to secure a patent airway when intubation attempts were unsuccessful.

Three patients did not require any type of airway management for TAI and were discharged from the hospital after a short admission period. Upon admission, these patients were not in respiratory distress and the definitive diagnosis of TAI was done by CT and/or fiberoptic bronchoscopy. Among the 78 patients undergoing any type of surgery, 15 patients (19%) were tracheally intubated (oral or nasatracheal) for the surgery but did not have the airway injury itself surgically repaired. These findings corroborate the principle that in selected cases, non-operative management with oxygen supplementation or a tracheal tube placed below the level of the injury is effective to treat TAI.3 However, most patients (81%) required some form of direct surgical repair of the injured airway.

Tracheostomy was performed as part of the definitive management of the TAI in 61 patients (59% of all patients). It remains the most common form of surgical treatment for patients with TAI in our institution, in isolation or in association with other airway surgical procedures.

In this study, the mortality varied according to the anatomical location of the airway injury, with lower injuries (LAI) having the highest mortality rate. Because the injury may extend beyond the end of the tracheal tube, patients with LAI, particularly when extensive, are often difficult to ventilate/oxygenate, even after tracheal intubation and mechanical ventilation. Other investigators have suggested that extra corporeal membrane oxygenation (ECMO) might be beneficial for the treatment of such complex airway injuries.13

Of the nine patients that died of a LAI, seven were related with the airway injury. In this series, patients with LAI had more severe injuries besides the airway one, the airway injury was considered as being of difficult surgical access and consequently had the highest mortality. Of the seven patients that died, difficulties in ventilation were reported in five, and there was enough evidence to suspect that the other two patients also had a similar problem.

As previously mentioned, of the 24 deaths in this study, 10 patients (42% of deaths) died as consequence of the TAI, 7 with a lower and 3 with an upper airway injury. Most of those deaths occurred early in admission, less than 24h in eight patients, confirming the immediate physiological consequences of the lack of a patent airway and failure of oxygenation and ventilation. The other two deaths occurred following hypoxic brain injury and sequential failures in the suture line of an extensive tracheo-bronchial injury. This last patient was transiently treated with ECMO but died in consequence of complications of this injury.

This study has limitations, many common to all retrospective studies. The data collection depended on the information contained in the chart. Details about the airway management at the scene of the accident, the referring hospital or en route were limited. It was not possible to evaluate the training of all those involved in managing the airway of these patients or even the number of difficult airways. Such information could be important in preparing a trauma team to receive a patient with known TAI. Furthermore, the link between death and the airway injury was determined subjectively when there was no clear statement in the patients’ charts.

Conclusion 

The overall incidence of TAI remains low. Airway injury from blunt trauma is less common, has a different clinical presentation but higher ISS and mortality compared to penetrating TAI. The early assessment of airways is crucial and a definitive airway is required in most patients with TAI. Alternative techniques for obtaining DA are frequently required and include: tracheal intubation through the neck wound, surgical airway and intubation using fiberoptic bronchoscopy. Patients that required a definitive airway at the scene of the accident, en route and/or sustain LAI have a higher mortality. Surgery and tracheostomy are commonly used for the definitive treatment of patients sustaining TAI. Even though most patients died as a result of their other injuries besides the TAI, causative factors in UAI patients included difficulty in establishing a definitive airway and ventilation/oxygenation problems were contributory in LAI patients.