Diagnosis and management of inhalation injury: an updated review

Inhalation injury complicates burns in approximately 10 to 20 % of patients and significantly increases morbidity and mortality [2‐5]. Other factors associated with a significant effect on mortality include burn size and age [6, 7] and the incidence of inhalation injury is correlated with an increase in both these factors [6, 7]. Inhalation injury has also been found to be an independent predictor of mortality in burn patients [8] and worsens survival even among patients with similar age and burn size [8]. Thermal airway injury is generally limited to supraglottic structures, whereas injury to the lower airway is chemical in nature. In the setting of steam, however, the injury is pervasive, causing damage to both upper airways and direct thermal injury to the lungs [9]. The degree of inhalation injury is variable and is dependent on several factors: the gas components inhaled, the presence of particulate matter (soot), the magnitude of the exposure, and individual host factors such as underlying lung disease and inability to flee the incident.

Historically, it was speculated that the combustion of certain materials, such as noncommercial polyurethane foam, resulted in the formation of a neurotoxin [10]; however, newer material testing methods revealed most smoke toxicity can be explained by a small number of toxic gases exerting their effects through asphyxiation, systemic toxicity, or direct effects on respiratory tissue [11]. Many products of combustion, such as carbon dioxide, function as simple asphyxiants by displacing oxygen at the alveolar level. This is further exacerbated by the hypoxic fire environment. CO functions systemically as an asphyxiant by (1) competitively displacing oxygen from hemoglobin and (2) binding to cytochrome oxidase at the mitochondrial level. By contrast, hydrogen cyanide binds only to cytochrome oxidase [1].

Other combustion products, such as halogen acids, formaldehyde, and unsaturated aldehydes (for example, acrolein), function as respiratory irritants. The chemical irritation causes denuding of the respiratory mucosa, leading to sloughing within the airways, and induces the host inflammatory response. Furthermore, the chemical injury stimulates vasomotor and sensory nerve endings to produce neuropeptides. Preclinical studies have shown these neuropeptides can induce an inflammatory response [12]. Substance P and calcitonin gene-related peptide are two suggested neuropeptides inducing tissue injury after inhalation injury [12, 13]. Lange et al. [13] found antagonists to calcitonin gene-related peptide and substance P attenuated the fluid shifts/inflammation in an ovine model subjected to smoke and inhalation injury. These neuropeptides then induce bronchoconstriction and nitric oxide synthase (NOS) to generate reactive oxygen species (ROS) [14]. As described by Kraneveld and Nijkamp [14], these neuropeptides can function as tachykinins, inducing a robust inflammatory response with the downstream effects of bronchoconstriction, increased vascular permeability, and vasodilation. Furthermore, tachykinins like substance P and neurokinin A can modulate immune cells and stimulate neutrophil and eosinophil chemotaxis [14].

Overall, these factors potentiate local cellular damage and the loss of hypoxic pulmonary vasoconstriction. The loss of hypoxic pulmonary vasoconstriction causes bronchial blood flow to increase by a factor of 10 within 20 min of inhalation injury. ROS may also induce mitochondrial dysfunction and cellular apoptosis [15]. Tissue factor expressed by damaged respiratory epithelial cells and alveolar macrophages initiates the extrinsic coagulation cascade, disrupting pro- and anti-coagulant alveolar homeostasis. We found, for example, that smoke inhalation injury contributes to a hypercoagulable state in the lung by inducing plasminogen activator inhibitor 1 and stabilizing its mRNA [16].

In addition, the increased bronchial blood flow delivers activated polymorphonuclear leukocytes and cytokines to the lung, potentiating the host inflammatory response. The loss of an intact bronchial epithelium and the effects of ROS result in a loss of plasma proteins and fluid from the intravascular space into the alveoli and bronchioles [17]. The transvascular shift of protein causes exudate and cast formation within the airways, leading to alveolar collapse or complete occlusion of the airways [17]. Experimental measures to decrease bronchial blood flow show attenuation of airway obstruction, pulmonary edema, and improved oxygenation [18]. These processes - loss of hypoxic vasoconstriction, increased blood flow to injured lung segments, decreased ventilation of collapsed segments - contribute to ventilation-perfusion mismatch as a primary mechanism of hypoxemia following the smoke inhalation injury [19]. Atelectasis, dysfunction of the immune system, and mechanical ventilation, in turn, predispose to pneumonia as a common complication of inhalation injury.

Many studies are now being completed to better understand the role pro- and anti-inflammatory mediators, or other immune modulators, play in patient outcomes. Albright et al. [1] demonstrated a graded increase in inflammatory cytokines from bronchoalveolar lavage fluid (IL-4, IL-6, IL-9, IL-15, interferon-gamma, granulocyte macrophage colony-stimulating factor, monocyte chemoattractant protein-1) that correlates with the severity of inhalation injury noted on bronchoscopic evaluation (grades 3 or 4 versus 1 or 2). They also found a significant shift from a macrophage-predominant population of cells in lavage fluid to one dominated by neutrophils [1]. This is thought to contribute to the later immune dysfunction, bacterial overgrowth, and pneumonia [1]. The source of the cytokines identified in inhalation is thought to be secondary to complement activation by heat denatured proteins [20]. The stimulation of the complement cascade releases histamine, resulting in xanthine oxidase upregulation and ROS formation [20, 21].

Davis et al. [22] showed several plasma immune mediators were associated with increased inhalation injury severity, even after adjusting for age and percentage of total body surface area burned. IL-1 receptor antagonist (IL-1RA), an anti-inflammatory immune mediator, had the strongest correlation with injury severity and outcome measures, including mortality [22]. The authors also found a much lower IL-1β to IL-1RA ratio in patients with inhalation injury who died. Given that IL-1β is an essential component of the host defense, they hypothesized that insufficient IL-1β or excessive IL-1RA results in systemic immune dysfunction.

The formation of ROS, such as superoxide anions (O₂ ⁻), hydrogen peroxide (H₂O₂), and hydroxyl radicals (OH⁻; the most unstable and reactive), appears to play a major role in numerous injury models [23]. Under normal circumstances the body has compensatory antioxidant mechanisms to mitigate the effects of ROS. Following reperfusion or injury, however, there is a large burst of ROS, which overwhelm the body’s protective measures. As a result, the ROS can lead to cell injury through neutrophil attraction and cytokine production. Indeed, evidence of oxidative stress is found in plasma and lung tissue following smoke inhalation injury [24]. IL-8, a potent chemokine, has been suggested to play a vital role in the initiation and progression of lung inflammation after smoke inhalation [25]. NOS-dependent formation of ROS has been studied in inhalation injury. Activated neutrophils produce large quantities of superoxide that combine with nitric oxide to produce peroxynitrite, which can damage DNA [20]. Peroxynitrite and the resultant DNA damage stimulates poly-(ADP ribose) polymerase-1 (PARP-1), a nuclear repair enzyme and coactivator of NF-ĸB, which can deplete ATP and produce cell damage [26, 27]. Combined, these factors contribute to an increase in IL-8 [27], induce NOS, neutrophil chemotaxis, and increase ROS production [20]. For this reason, heparin/acetylcysteine combinations are being utilized as ROS scavengers in inhalation injury [20].

Classically, the diagnosis of inhalation injury was subjective and made on the basis of clinical findings. When evaluating a patient with suspected inhalation injury, a clinician first reviews the history and reported mechanism to determine the likelihood of an inhalation injury. Pertinent information includes exposure to flame, smoke, or chemicals (industrial and household), duration of exposure, exposure in an enclosed space, and loss of consciousness or disability. Pertinent physical exam findings include facial burns, singed facial or nasal hair, soot or carbonaceous material on the face or in the sputum, and signs of airway obstruction including stridor, edema, or mucosal damage [3]. Older patients, and those with more extensive burns, are at increased risk of inhalation injury because of prolonged exposure to the fire environment [8].

There are several modalities for confirming inhalation injury to include fiberoptic bronchoscopy (FOB), chest computed tomography (CT), carboxyhemoglobin measurement, radionuclide imaging with ¹³³Xenon, and pulmonary function testing. Many of these modalities lack sensitivity, are invasive, or are subject to significant variability between institutions. In studies by Shirani et al. [8], the following gradation of morbidity and mortality risk was seen in order of increasing risk: (1) patients without inhalation injury; (2) patients with inhalation injury by ¹³³Xenon scan only, but not by FOB; and (3) patients with inhalation injury by FOB. Also, presence of inhalation injury on FOB predicted risk of acute lung injury and the need for increased fluid resuscitation volumes. More recent studies have found a significant correlation between the severity of inhalation injury on FOB and mortality [28].

There are several difficulties in diagnosing the presence and severity of inhalation injuries. Although several laboratories have developed dose–response models of inhalation injury in large animals [29], the characteristics of the material inhaled are important in determining the degree of respiratory failure. In addition, differences in the individual host inflammatory response may lead to a heterogeneous clinical presentation [30]. FOB is unable to assess distal airways and respiratory bronchioles; therefore, damage to this portion of the lung has been proposed as an explanation for the discordance between bronchoscopic severity of injury and mortality. Despite these limitations, FOB continues to be the standard technique used to assess the presence and severity of inhalation injury. Its relative ease and availability allows the initial diagnosis to be made (Fig. 1), and allows the inhalation injury to be followed serially (Figs. 2 and 3).

Given the lack of a widely standardized and validated method for scoring inhalation injury severity, Woodson [30] has proposed a large multicenter study to create such a scoring system to allow for more reliable prognostic estimations. To date, no such study has been done, though one is currently underway based on clinical, radiographic, bronchoscopic, and biochemical parameters (ClinicalTrials.gov identifier NCT01194024).

Multiple studies have demonstrated that inhalation injury is a graded phenomenon with severity correlating with outcome. The Abbreviated Injury Score grading scale for inhalation injury on bronchoscopy has been shown to correlate with an increase in mortality as well impaired gas exchange [1, 28, 31]. This scale is shown in Table 1. Endorf and Gamelli [3] found that patients with more severe inhalation injury on initial bronchoscopy (grades 2, 3, 4) had worse survival rates than patients with lower scores (grades 0 or 1) (P = 0.03). They also noted the highest-grade inhalation injuries were not necessarily associated with an increased fluid requirement, contrary to prior data. Lastly, they found patients with an arterial partial pressure of oxygen (P_aO₂)/fraction of inspired oxygen (F_iO₂) ratio <350 upon presentation had a statistically significant increase in fluid resuscitation needs compared with patients with a ratio >350 (P = 0.03) [3]. Ryan et al. [32] have stated that, at this time, the most reliable indicator of the impact of inhalation injury is the P_aO₂/F_iO₂ ratio after the resuscitation has started. This is based on a retrospective review by Hassan et al. [28] of 105 patients admitted with inhalation injury. They assessed respiratory function by using the P_aO₂/F_iO₂ ratio from 0 to 192 h after injury. Their study showed a significant difference (P < 0.01) in P_aO₂/F_iO₂ ratios between patients who died (mean P_aO₂/F_iO₂ ratio 20.17) and those who survived (mean P_aO₂/F_iO₂ ratio 32.24). Ultimately, they propose to use P_aO₂/F_iO₂ ratio as a predictor of survival once the initial burn resuscitation has been completed and a full response to injury is able to be mounted [28]. Similarly, Cancio et al. [33] found that the mean alveolar-arterial oxygen gradient during the first 2 days was an independent predictor of mortality in mechanically ventilated burn patients. It is important to note that P_aO₂/F_iO₂ can be arbitrarily high or low depending on the choice of ventilator mode. In addition, P_aO₂/F_iO₂ may be affected by the volume of resuscitation. Therefore, we do not use P_aO₂/F_iO₂ as a basis for diagnosis of inhalation injury, but use it to trend the patient’s oxygenation and potential need for nonconventional ventilation.

Other means of evaluating the severity of inhalation injury include chest CT. First, a scoring system for severity of CT scan findings has been developed [29]. Our group studied 25 patients with inhalation injury and 19 patients without inhalation injury who received a chest CT within 24 h of admission [34]. The severity of radiographic findings was calculated by looking at 1-cm axial slices from the chest CT and these were scored by adding the highest radiologist’s score (RADS) for each quadrant. The RADS scoring system is shown in Table 2, and the various RADS findings are shown in Fig. 4. Our group assessed a composite endpoint of pneumonia, acute lung injury/ARDS, and death. We found that the detection of inhalation injury on bronchoscopy was associated with an 8.3-fold increase in the composite endpoint. A high RADS score (>8 per slice) in addition to a positive bronchoscopy was associated with a 12.7-fold increase, thus showing the potential for chest CT to complement bronchoscopy in detecting clinically significant inhalation injury [34].

Second, Yamamura et al. [35] used CT imaging to measure bronchial wall thickness 2 cm distal to the tracheal bifurcation in patients who had sustained an inhalation injury. The authors noted a statistically significant correlation between bronchial wall thickness and the development of pneumonia, total number of ventilator days, and ICU length of stay. They also found that a bronchial wall thickness value of >3.0 mm predicted the development of pneumonia with a sensitivity of 79 % and specificity of 96 %. Interestingly, this study was not able to replicate the association between Abbreviated Injury Score bronchoscopic scoring and clinical outcomes as described above [35].

A third approach to using the CT scan is virtual bronchoscopy. A three-dimensional reconstructed image is presented in such a way that the viewer navigates through the lung as if using a bronchoscope. We found that virtual bronchoscopy agrees best with FOB in the detection of airway narrowing, and less so in the detection of blistering or necrosis [36].

Problems with using chest CT as part of a diagnostic algorithm for inhalation injury include determining the optimal timing of the test and how to interpret abnormal radiographic findings in the setting of a negative bronchoscopy. Putman et al. [37] found that chest radiography on admission was rarely helpful in determining the presence or severity of inhalation injury, but its use is helpful as a baseline for determining future changes.

Given the limited availability of targeted therapies for inhalation injury, one of the fundamental tenets is supportive respiratory care. This includes aggressive pulmonary toilet and mechanical ventilation when indicated. It should be noted that approximately 20 to 33 % of patients hospitalized with inhalation injury experience some degree of upper airway obstruction due to pharyngeal edema that can progress rapidly [38]. As thermal injury increases airway edema and can lead to airway obstruction, early intubation is favored [4]. This is of particular concern in patients who receive large amounts of intravenous fluids during resuscitation. Generally speaking, the most experienced clinician in airway management should perform endotracheal intubation with the largest available, age-appropriate endotracheal tube for patients with suspected or impending upper airway obstruction in the setting of inhalation injury. One study suggests prophylactic intubation can decrease mortality related to pulmonary-related death in patients with inhalation injury [39].

Maintaining bronchial hygiene is paramount in patients who have suffered inhalation injury. Early ambulation, chest physiotherapy, airway suctioning, and therapeutic bronchoscopy are adjunctive tools [38]. Reper et al. [40] demonstrated that intrapulmonary percussive ventilation administered through a face mask to spontaneously breathing patients with smoke inhalation injury, hypoxia, and persistent atelectasis can result in a significant improvement in P_aO₂/F_iO₂ ratio.

A low threshold should be maintained for intubation and mechanical ventilation in inhalation injury due to the progressive nature of the airway edema. Interestingly, a study by Mackie et al. [41] showed an increased use of mechanical ventilation in patients at a Dutch burn center from 1997 to 2006 (76 %) compared with 1987 to 1996 (38 %) despite a decrease in the incidence of inhalation injury (34 % versus 27 %). The authors hypothesized that this was related to the institution of Advanced Trauma Life Support principles in the mid-1990s in the Netherlands. Mackie [42] also suggested mechanical ventilation may be a significant contributor to mortality in burn patients independent of inhalation injury. He proposed increased intrathoracic pressure from positive pressure ventilation led to decreased venous return, followed by decreased cardiac and urine output. The typical clinical reaction is to increase intravenous fluid administration, resulting in higher volumes of infused fluid, a known risk factor for adverse outcomes in burn patients [42].

Multiple challenges, to include concern for ventilator-induced lung injury in patients with inhalation injury, have led to the use of unconventional ventilator modes [43].

Conventional mechanical ventilation is limited in the patient with inhalation injury. In a patient with fibrin casts, extensive chest wall thermal injuries, or high volumes of resuscitative fluid maintaining the recommended tidal volumes of less than 7 ml/kg body weight and plateau pressures of less than 30 cm water [44], can prove difficult with conventional techniques. Therefore, in order to apply lung-protective ventilation in patients with inhalation injury, nonconventional ventilator modes are employed.

High-frequency percussive ventilation (HFPV) was first described in patients with inhalation injury as a means of assisting with clearance of sloughed respiratory mucosa and plugs, as well as decreasing iatrogenic barotrauma and the incidence of pulmonary infection [45]. Further studies have demonstrated benefits from using HFPV prophylactically (that is, not as a salvage mode) in both adult [46, 47] and pediatric populations [48]. Our group performed a randomized controlled trial to compare HFPV versus conventional low tidal volume (LTV) ventilation [49]. While we detected no difference in ventilator-free days between patients randomized to HFPV compared with LTV ventilation, there was a statistically significant increase in the P_aO₂/F_iO₂ ratio for the HFPV cohort on days 0 to 3 (Fig. 5). We also found that less patients in the HFPV cohort required conversion to a rescue mode of ventilation compared with the LTV ventilation cohort [49]. Also, HFPV was associated with a decrease in the incidence of pneumonia from 45 to 26 % (P < 0.005) and resulted in an improvement in survival [46]. Although HFPV cannot reverse the effects of inhalation injury, it can improve the clearance of secretions, provide positive pressure throughout the ventilator cycle, allow for lower airway pressures, and increase functional reserve capacity [46].

Interestingly, a recent retrospective study by Sousse et al. [50] compared high tidal volume (HTV) ventilation (15 ± 3 ml/kg, n = 190) with LTV ventilation (9 ± 3 ml/kg, n = 501) in pediatric patients suffering from inhalation injury. Patients on HTV had fewer days on the ventilator (P < 0.005), increased maximum peak inspiratory pressure (P < 0.02), and plateau pressures (P < 0.02) compared with those on LTV ventilation. Furthermore, the incidence of atelectasis and ARDs was significantly lower in the cohort receiving HTV ventilation (P < 0.0001 and P < 0.02, respectively). However, the HTV ventilation group were not without complications and had a significantly higher rate of pneumothorax compared with the LTV ventilation group (P < 0.03). For pediatric patients suffering from inhalation injury, HTV ventilation may be better than traditional LTV ventilation [50]. The mechanisms for this observation are unclear, but this study, much like the previous study, suggests that we must be cautious when extrapolating LTV ventilation to all patient populations, especially in those with different pathophysiologies. A randomized controlled trial may be necessary to tease out the true impact of these divergent strategies. Our group also looked at airway pressure release ventilation (APRV) in a prospective animal model study. We found that P_aO₂/F_iO₂ ratios were initially lower in pigs with inhalation injury on APRV compared with conventional mechanical ventilation, although this equilibrated at 48 h. Higher mean airway pressures were necessary to maintain oxygenation in APRV, and, in the end, no survival difference was seen between APRV and conventional mechanical ventilation [51].

Other nonventilator adjuncts to consider for inhalation injury include prone positioning and extracorporeal membrane oxygenation (ECMO). Our group showed that prone positioning led to a statistically significant increase in P_aO₂/F_iO₂ ratio in patients with inhalation injury and refractory ARDS whose initial ratio was an average of 87 ± 38 [52].

A systematic review and meta-analysis on the use of ECMO in inhalation injury was limited by number of studies and total patients available. There was a tendency towards increased survival in burn patients with acute hypoxemic respiratory failure treated with ECMO. ECMO use of <200 h was correlated with higher survival compared with time >200 h. There was no improvement in survival if ECMO was initiated once the P_aO₂/F_iO₂ ratio was <60 [53].

CO has an affinity for hemoglobin 200 to 250 times greater than oxygen and exposure results in hypoxia and ischemia. Unlike inhalation injury, CO has deleterious effects at the level of hemoglobin and more specifically the ability for oxygen delivery. CO acts to displace oxygen from hemoglobin (forming carboxyhemoglobin (COHb)) and binds to cytochrome c oxidase. COHb shifts the oxygen dissociation curve to the left, ultimately leading to decreased oxygen delivery at the tissue level and interfering with cellular respiration at the mitochondrial level [73]. Symptoms of CO toxicity include confusion, stupor, coma, seizures, and myocardial infarction [74]. CO diagnosis requires the use of a Co-oximeter (not available in every blood gas lab), since elevated COHb levels may be present despite normal P_aO₂ and oxygen saturation readings. Available since 2005, newer, non-invasive CO-oximetry monitors permit more rapid diagnosis [75]. CO poisoning is associated with an increased risk of mortality even at long-term follow-up (median of 7.6 years) [74]. Complications of CO poisoning include persistent and delayed neurologic sequelae and myocardial injury, as well as functional effects on leukocytes, platelets, and vascular endothelium [73]. Treatment of CO poisoning involves providing 100 % oxygen, which shortens the half-life of COHb to about 45 min.

Hyperbaric oxygen therapy (HBO) has been used to treat CO poisoning and can further reduce the COHb half-life to about 20 min. The benefits of HBO were evident in a study of 75 patients with acute CO poisoning. Three treatment sessions were administered within 24 h and neuropsychological tests were administered at various points throughout the study. They found the rates of cognitive sequelae were reduced at 6 weeks and 12 months after CO poisoning in patients with three HBO sessions [76]. It should be pointed out that the theoretical basis for HBO up to 24 h after exposure is to facilitate the clearance of CO from cytochrome c oxidase in the brain, rather than to increase its clearance from hemoglobin in the blood. Logistical factors have limited the utilization of HBO. A systematic review found that not enough evidence exists at this point to determine definitively whether HBO reduces adverse neurologic outcomes after CO poisoning [77]. Continued advancements in HBO technology combined with increased ICU accessibility will likely result in the generation of more clinical studies in this area.

The gaseous form of cyanide, hydrogen cyanide (HCN), is formed in fire atmospheres from the thermal decomposition of nitrogen-containing polymers, both natural (wool, silk, and paper), and synthetic (nylon and polyvinyl chloride). The significance of HCN in fire environments is unclear [11]. The lack of a rapid and reliable test to detect cyanide poisoning limits our understanding of the role of HCN in inhalation injury [78]. Additionally, symptoms can mimic CO poisoning. Dumestre et al. [79] found that most burn centers do not test for cyanide poisoning on admission and do not administer an antidote on the basis of clinical suspicion alone. Lactate has been suggested as a marker for severity of cyanide poisoning without other comorbidities [80], but its role in inhalation injury is less clear in a population at risk for CO poisoning and with coexisting hypovolemic shock. Hydroxocobalamin, the most commonly available antidote (sold as Cyanokit®), binds to HCN to form cyanocobalamin, which is nontoxic and excreted in the urine. The standard dose of 5 g is infused intravenously over 15 min. A second dose of 5 g can be administered in patients with severe toxicity or poor clinical response. It is generally regarded as safe. Red discoloration of the skin and urine is common, which may interfere with colorimetric assays.

Sodium nitrite (300 mg) and sodium thiosulfate (12.5 g) are also commercially available (sold as Nithiodote™). Prior to 2007, when hydroxocobalamin became available, sodium nitrite and sodium thiosulfate were used primarily for cyanide poisoning despite limited data as to their efficacy. In 2012, Bebarta et al. [81] evaluated sodium thiosulfate versus hydroxocobalamin in a swine model of severe cyanide poisoning. They found that sodium thiosulfate failed to reverse cyanide-induced cardiovascular collapse. Further, sodium thiosulfate was not found to be effective when added to hydroxocobalamin. Hydroxocobalamin alone was found to be effective for severe cyanide toxicity. Treatment with nitrites carries significant risk of hypotension and methemoglobinemia, which can further jeopardize tissue oxygen delivery.

Complications from inhalation injuryInhalation injury can be divided into anatomic levels and the mechanism of injury - direct thermal injury to the upper airways or chemical injury to the subglottic region and tracheobronchial tree [82‐84]. The associated complications vary with the level of injury and are also effected by intubation, infection, and chronic inflammation [82]. In addition, the complications from injury may be acute or delayed. Pneumonia and airway obstruction are early complications of inhalation injury and have been well described in the literature [4, 8]. However, there is a paucity of data on the long-term or delayed complications from inhalation injury [84].