Airway Management for Penetrating Head and Neck Trauma, Experience of a Trauma Center from 2012 to 2020

Background From a series of penetrating head and neck trauma managed in a level-1 Trauma Center, the main aim of this study was to determine predictive factors for early denitive airway management, during pre-hospital time or in the emergency room. The secondary objective was to perform a descriptive epidemiological analysis of the series. Methods A single-center retrospective study was conducted between January 1 2012 and June 30 2020. in a French Level 1 Trauma Center. The patients included were adults treated for penetrating head and neck trauma, regardless of the mechanism and the causal agent. Results 56 patients were included. Ballistic origin, Shock Index >0.9 and active bleeding in the emergency room were predictive criteria for denitive airway management during pre-hospital time or in the emergency room. 78.6% of patients were male. Median age was 54 years. The trauma followed a suicide attempt in 50% of cases, an accident in 26.7% and an assault in 23.2%, with use of a knife in 42.9% and rearm in 26.8%. Mortality was 10.7%. 16.1% of patients had undergone pre-hospital intubation and 19.1% intubation in the emergency room. CT scan was performed in 87.5% of cases, surgery in 96.4% and tracheotomy in 37.5%. A laryngotracheal lesion was seen in 14.2%. In 50% of patients, primary admission was to intensive care.


Abstract
Background From a series of penetrating head and neck trauma managed in a level-1 Trauma Center, the main aim of this study was to determine predictive factors for early de nitive airway management, during pre-hospital time or in the emergency room. The secondary objective was to perform a descriptive epidemiological analysis of the series.

Methods
A single-center retrospective study was conducted between January 1 2012 and June 30 2020. in a French Level 1 Trauma Center. The patients included were adults treated for penetrating head and neck trauma, regardless of the mechanism and the causal agent.
Results 56 patients were included. Ballistic origin, Shock Index >0.9 and active bleeding in the emergency room were predictive criteria for de nitive airway management during pre-hospital time or in the emergency room. 78.6% of patients were male. Median age was 54 years. The trauma followed a suicide attempt in 50% of cases, an accident in 26.7% and an assault in 23.2%, with use of a knife in 42.9% and rearm in 26.8%. Mortality was 10.7%. 16.1% of patients had undergone pre-hospital intubation and 19.1% intubation in the emergency room. CT scan was performed in 87.5% of cases, surgery in 96.4% and tracheotomy in 37.5%. A laryngotracheal lesion was seen in 14.2%. In 50% of patients, primary admission was to intensive care.

Conclusions
Ballistic origin, Shock Index >0.9 and active bleeding in the emergency room were predictive criteria for early de nitive airway management. This study established the pro le of patients suffering from penetrating head and neck trauma managed in a Trauma Center over a period of 9 years. Background Penetrating head and neck trauma (PHNT) is de ned as a wound crossing the cervical platysma muscle or facial super cial muscular-aponeurotic system (SMAS), whatever the mechanism or instrument. PHNT is rare in civilian contexts, but life-threatening due to the anatomic complexity and high likelihood of vascular and airway lesions [1]. The di culty of controlling and securing the airway is a major challenge [2]. It is the second most frequent cause of avoidable death after hemorrhagic shock [3,4].
From a series of PHNT managed in a level-1 Trauma Center over a 9-year period, the main aim of the present study was to determine predictive factors for early de nitive airway management, in the pre-hospital phase or in the emergency room. The secondary objective was to perform a descriptive epidemiological analysis of the series.

Methods
A single-center retrospective study was conducted using the prospective database of all cases of severe trauma managed in a level-1 Trauma Center, for the period January 1st, 2012 to June 30th, 2020.
According to French recommendations, all patients meeting Vittel criteria are considered as severe trauma [5]. Inclusion criteria comprised adult patient, with PHNT; exclusion criteria comprised craniocerebral wounds by temporal ballistic impact, and death during the pre-hospital phase. Analysis relied on the usual hospital database, with authorization from the CNIL data protection commission and institutional review board approval (CNIL ref.: 911461V2; IRB n°: 0011873-2020-14).
Two groups were formed: PHNT requiring de nitive airway management in the pre-hospital phase or emergency room (group 1), and patients with no (surveillance or surgery under local anesthesia) or elective de nitive airway management (group 2). De nitive airway management is de ned as the presence of a tracheal intubation with a subglottic in ated balloon connected up to the ventilation device. Airway management data comprised method (orotracheal intubation or alternatives), ease of implementation (subjective operator appreciation), e ciency and indication among uncontrolled bleeding, consciousness disorder or agitation and dyspnea. In case of orotracheal intubation failure, the alternative techniques (cricothyroidotomy, intubation via the cervical wound) were recorded.
Epidemiological data comprised age, gender, trauma mechanism, place of trauma, causal agent, mortality. Injury Severity Score (ISS) was calculated [6]. Pre-hospital transport data comprised type of transfer, type and duration of transport.
In the emergency room, the previous data were reassessed. Any changes in hemostasis were recorded.
Functional and clinical signs were assessed. PHNT was standardly described as: cervical zone I (clavicle to cricoid cartilage), II (cricoid cartilage to mandibular angle) or III (mandibular angle to skull base) according to the classical 1969 Monson classi cation [8]; medial or lateral with respect to the anterior edge of the sternocleidomastoid muscle; superior, middle or inferior facial zone; and right or left side. Biological data comprised: HemoCue, hemoglobinemia, blood alcohol, and arterial blood gas. Posttraumatic coagulation disorder was de ned by at least one of the following: prothrombin time < 70%, platelets < 150,000/mm 3 , or brinogen < 1.5 g/L. Other data comprised: imaging, endoscopy, surgery, hospital stay and department. Surgical data comprised: observed lesions and secondary tracheotomy with time to decannulation. Data were reported as median, range and interquartile range for quantitative data, and as percentage for qualitative data. Mann-Whitney U test for independent samples was used, with the signi cance threshold set at p < 0.05. Quantitative data were compared between groups. Qualitative data were compared between groups on chi² test. Data identi ed on univariate analysis or otherwise considered clinically relevant were selected for multivariate analysis and stepwise descending logistic regression. Analyses used IBM® SPSS® Statistics 25.0 software.

Results
Fifty-six of the 2,760 severe trauma patients admitted to the emergency department were included ( owchart: Fig. 1). PHNT accounted for an annual median 2% (range, 1.27-3.33%) of this population. Table 1 shows airway management data.  cervical zone III and lower face involvement (p = 0.033), active bleeding (p = 0.004), Shock Index > 0.9 (p = 0.033), Glasgow score < 7 (p < 0.001), and hemoglobinemia < 10 g/dL (p < 0.001). Multivariate analysis identi ed a signi cant association between ballistic trauma, Shock Index > 0.9 and active bleeding in emergency room and the need for early de nitive airway management.

Discussion
The present study has the particular interest of being the rst descriptive report of a series of PHNT in civilian practice in a French Trauma Center, over a long 9-year period at the level of a whole administrative area (Département). The single-center design ensured homogeneity of practices and thus of data.
Although conducted retrospectively, the study was based on prospectively collected data from a database allowing epidemiological comparison between PHNT and the severe trauma population as a whole. It also provided precise analysis of the issue of airway management in PHNT, which has been rarely addressed speci cally in the literature. Predictive factors for early de nitive airway management were identi ed.

Airway Management
Acute respiratory failure is seen in 10-50% of cases, by direct obstruction, extrinsic compression, partial or total airway sectioning, or of neurologic origin. It is consensual that airway control, despite the di culty of the approach, should be anticipated due to the risk of rapid respiratory distress [9]. In the present series, orotracheal intubation was di cult in 12.5% of cases, and in 25% when performed in emergency in group 1, versus 5.6% in group 2. In case of dyspnea in the emergency room (n = 6), intubation failed in 33.4% of cases, and in 100% in case of associated cervical emphysema or sucking wound. In combat, upper airway obstruction is the second cause of preventable death, after hemorrhage and ahead of compressive pneumothorax [2][3][4]. According to Demetriades [10] about 16% of rearm injuries and 14% of knife wounds are associated with pneumothorax; the rate was 8.9% in the present series. Cases where both chest X-ray and FAST proved negative involved anteroinferior pneumothorax following sub-xyphoid knife wound. This association is more frequent in zone I trauma (80% in the present series).
When release maneuvers and mask oxygen therapy fail, de nitive airway treatment is required, with tracheal intubation with a subglottic in ated balloon connected up to the ventilation device. Even with a cricothyroidotomy kit, a balloon should be used. The present series con rmed frequent radiological signs of pulmonary inhalation. Once the airway is secured, oronasal packing is possible and effective. Prehospital and in-hospital algorithms determine the roles of 3 types of de nitive airway: orotracheal, nasotracheal, and cervical intubation (cricothyroidotomy, tracheotomy, intubation via sucking cervical wound). Urgency and circumstances determine the method to be used. Laryngoscopic orotracheal intubation is the method of choice, with gum elastic bougie in case of di culty. In case of failure, supraglottic airway and videolaryngoscopy can be attempted by experienced operator [11], but are likely to fail due to anatomical distorsion and/or soiling of the airway by blood. Prompt escalation to a surgical airway must be quickly considered. Cricothyroidotomy is the technique of choice. Tracheotomy is an option, but is a surgical procedure, di cult to implement in emergency [12], with dissection liable to incur blood loss for several minutes, thus not an adapted salvage procedure. It is, however, indicated secondarily to orotracheal intubation to free the operative eld, improving patient comfort with earlier awakening, despite the post-traumatic pharyngolaryngeal edema. Tracheotomy is performed within hours of cricothyroidotomy, to avoid secondary laryngeal stenosis. Thus, in the present series, secondary tracheotomy was performed in 37.5% of cases. All were subsequently decannulated. In rare cases of laryngeal or tracheal sectioning with complete airway obstruction, direct insertion of the intubation probe or tracheotomy cannula through the lesion is the optimal means of airway control. We recommend systematically having Laborde forceps in the pre-hospital kits and emergency trolleys.
Optimal airway control makes orotracheal intubation di cult or even risky when performed at the accident site: risk of inhalation on a full stomach, risk of spinal instability, presence of a cervical collar, altered anatomic relations, mechanical limitation of oral opening, and blood loss into the oropharynx [13].
Nevertheless, orotracheal intubation remains mandatory in case of respiratory distress or severe cyanosis. In contrast, indications are trickier in apparently stable patients. But the literature reports no signi cant predictive criteria for early de nitive airway management. Despite its limitations, the present study identi ed three: ballistic cause of trauma, hemodynamic impact with Shock Index > 0.9, and active bleeding in the emergency room. It also identi ed trends for certain criteria: submental impact with involvement of cervical zone III and the lower face, neurologic impact with Glasgow score < 7, and hemoglobinemia < 10 g/dL. Irretrievable loss of substance leads to di culties hemostasis with blood loss and di culties temporary airway control maneuvers. Prolonged transfer to hospital is a classical factor in favor of early preventive airway salvage, perhaps in conditions such as helicopter transport in which intubation is di cult. In the present series, only one cricothyroidotomy was performed in the emergency room, in a patient with dyspnea and progressive cervical hematoma following a submental ballistic suicide attempt, who had not been intubated ahead of helicopter transport that took more than an hour (Fig. 4).
Even so, the decision to intubate is always very di cult; if considered, it should be performed as quickly as possible while conditions are still favorable. A further multicenter study based on a national registry could increase statistical power and help determine optimal de nitive airway management timing, notably according to transport time.
Descriptive epidemiological analysis of the series PHNT is rare, at 5-10% of traumas treated in emergency [14], and 2% in the present series. Incidence varies geographically, and is higher in American and South African series [10,13,15]. In Europe, incidence is lower: 4.3/100,000 per year in London, UK [16]], and 1.3/100,000 per year in Finland [17]. Over the present study period, the median annual emergency department turnover was 34,600 patients, with 310 cases of severe trauma and 1 of PHNT every 2 months. Management is familiar in military contexts: warscene incidence is 5-10% [18]; poor head and neck ballistic protection and asymmetric con icts featuring improvised explosive devices (IEDs) led to increased incidence, reaching 36-55%, in French troops in Afghanistan [19,20]. In line with the literature [16], we found male predominance and a median age around 50 years.
Frequency also varies according to circumstances and instrument. In Los Angeles [10] PHNT is caused by aggression or attempted suicide by rearms in 48% of cases, by knife wounds in 40% and blunt objects (road accidents or falls) in 10%. In the present series, 50% of cases also concerned attempted suicide, by violent means inasmuch as 86.7% of ballistic PHNTs were suicide attempts. PHNT accounts for 1.6-3% of suicide attempts as a whole [21], but is one of the most lethal means [22]. The instrument was more often a knife (42.9%) than a rearm (26.8%), probably because the latter are less readily available in France than in the USA. In wartime, the distribution is different, with 62-98% of cases involving highvelocity weapons [19,23].
Mortality is between 3% and 6%, mainly due to massive blood loss from large vessels [1]; it is generally underestimated, as most studies exclude pre-hospital mortality, for lack of data; if this is included, mortality can be as high as 11% [19]. Nevertheless, in our series, no deaths by exsanguination were found. These patients at high risk of rapid hemodynamic and respiratory decompensation need urgent medical transport to the nearest Trauma Center [24]. In the present series, despite a mean transport time of 46 minutes (72.4% with respect to the "Golden Hour" [25]) and pre-hospital medical care in 82.1% of cases, mortality was 10.7% (not including pre-hospital deaths). In severe trauma as a whole, mortality is higher than speci cally in PHNT. The present high mortality may be due to the small sample and to a recruitment bias with two treatment interruptions prescribed due to the severity of neurologic lesions in a context of advanced cancer. In 83.3% of cases, death was due to ballistic trauma in attempted suicide, impacting the submental central compartment in zone III. The only patient who died after isolated cervical trauma (zone II) showed total sectioning of the internal carotid artery and pharynx, after falling onto a metal picket (Fig. 3).
As in the literature [10], zone II predominated in purely cervical trauma, at 56.2%. The "central/lateral" classi cation separates the vascular axis and aerodigestive tract [19]. Treatment in PHNT used to involve systematic surgical exploration of any wound crossing the platysma, but surgery has now become more selective [27]. Several prospective series supported selective surgery, and described algorithms [1,15,28]. In the present series, 87.5% of patients underwent contrast-enhanced CT, and 96.4% underwent surgery. Only 1 patient was treated non-operatively, by in-hospital surveillance.
In some patients, surgical exploration con rmed the absence of life-threatening lesions, and CT ndings were never falsi ed. Even so, simple surgical damage control with exploration to check hemostasis (especially in dry venous wounds following knife attacks), prevention of superinfection by lavage, wound care and debridement, extraction of any foreign bodies, possibly with drainage, and suturing in 2 planes was always essential. The sole indication for interventional neuroradiology was for a lateral zone III knife wound with internal carotid thrombosis. According to some authors [29][30][31], algorithms based on entry point are too rigid, leading to unproductive surgery, overlooked lesions with poor correlation between the location of the external wound and the internal lesions, increased hospital stay, and higher rates of complications. In 2018, Nowicki [24] described a selective attitude, independent of cervical zone, taking the whole neck as a single entity. Even so, we consider these classi cations useful, especially when there is an in ux of injured patients and imaging is not available, as often happens in military surgery overseas, where the surgeon is seldom specialized in head and neck. The classi cations help mentalize the lesion trajectory, disclosing vascular involvement in lateral lesions or respiratory involvement in central lesions, and determining optimal strategy in borderline zones I and III.

Conclusion
The present study, conducted at the level of a French regional Trauma Center, con rmed that penetrating head and neck trauma is rare, but with a high morbi-mortality, especially in case of rearm trauma.
Securing the airway is a major challenge, especially before a prolonged transport. Ballistic origin, Shock Index > 0.9 and active bleeding in the emergency room were predictive criteria for early de nitive airway management. Prompt intubation should be considered.

Declarations
Ethics approval and consent to participate Analysis relied on the usual hospital database, with authorization from the CNIL data protection commission and institutional review board approval (CNIL ref.: 911461V2; IRB n°: 0011873-2020-14).

Consent for publication
Consent for publication was obtained for individual person's data in gures.
Availability of data and materials Population ow chart (ND: no data).

Figure 3
Cervical angio-CT-scan in axial section and intraoperative view in suspension laryngoscopy. Internal carotid artery rupture and massive hemorrhagic inhalation through a hypopharyngeal wound after fall on a metal stake with lateral zone 2 PHNT.

Figure 4
Rescue cricothyroidotomy in the emergency room, after failure of two orotracheal intubation attempts in a patient in hypoxic cardiorespiratory arrest, with ballistic PHNT with expansive cervical hematoma in central zone III.