Discussion
The analysis of physical findings across the included studies demonstrates considerable variability in their diagnostic performance for inhalation injury in burn patients. The 10 evaluated signs—facial and neck burns, singed nasal hair, carbonaceous sputum, cough, sore throat, hoarseness, dyspnea, stridor, and flame exposure—showed high specificity but generally low sensitivity. This pattern indicates that while the presence of certain findings, particularly those with high DOR, strongly suggests inhalation injury, their absence does not reliably exclude the condition. Relying solely on physical examination, therefore, may result in missed diagnoses, especially in the absence of bronchoscopy. Although bronchoscopy remains the most widely accepted reference standard for confirming inhalation injury, other diagnostic approaches have also been explored. Radiological findings, including bronchial wall thickening on cross-sectional imaging, have been proposed as potential indicators of inhalation injury severity.2 While these modalities may provide additional diagnostic information, their availability and practicality may be limited in certain emergency or resource-constrained environments.
High specificity, low sensitivity means that a diagnostic test is very good at correctly identifying people who do not have a disease (few false positives), but may miss a significant number of people who actually do have the disease (many false negatives). Essentially, it is very accurate at ruling out a disease but not very good at ruling it in.
Dyspnea was defined as subjective or clinically observed respiratory distress, including labored breathing, tachypnea, or patient-reported shortness of breath.12 Dyspnea exhibited the highest diagnostic accuracy, with a DOR of 7.50 (95% CI 0.48 to 119.17) and a specificity of 0.97, making it a valuable predictor of airway compromise when present. Stridor (DOR 6.22) and hoarseness (DOR 4.14) followed closely, both associated with high specificity but low sensitivity. These signs are clinically significant as they often reflect upper airway edema and thermal injury, conditions that typically necessitate early intubation to prevent airway obstruction. Conversely, superficial indicators such as singed nasal hair (DOR 0.93) and facial burns (DOR 1.71) ranked lowest in diagnostic accuracy, reflecting their limited reliability as standalone markers. Flame exposure and carbonaceous sputum showed intermediate diagnostic performance, but the wide CIs of some findings highlight uncertainty due to limited data.
Overall, these results underscore the importance of using a combination of physical findings rather than relying on any single sign. Prioritizing high-specificity indicators such as dyspnea, stridor, and hoarseness may enhance early diagnosis and guide timely intubation decisions, particularly in settings where bronchoscopy is not readily available. Across the included studies, the number of patients diagnosed with inhalation injury varied widely, ranging from 16 to 127, while the number of intubated patients ranged from 34 to 274 per study, suggesting that a considerable proportion underwent intubation without confirmed inhalation injury. Such unnecessary intubations carry significant clinical risks, including ventilator-associated respiratory complications, prolonged hospital stays, and increased healthcare costs, highlighting the need for more accurate diagnostic tools and criteria to guide intubation decisions.
The variability in the diagnostic performance of physical findings observed in this review aligns with previous research demonstrating the limited accuracy of commonly cited clinical signs. Several studies have reported that findings such as singed nasal hair, carbonaceous sputum, and facial burns correlate poorly with bronchoscopy-confirmed inhalation injury, reflecting low sensitivity and limited reliability as standalone indicators.12 13 Conversely, dyspnea, stridor, and hoarseness showed higher DORs and greater specificity, consistent with evidence identifying respiratory distress and upper-airway symptoms as more reliable markers of clinically significant airway injury.14
Despite their high specificity, the generally low sensitivity of physical findings underscores the risk of missed diagnoses when bronchoscopy is unavailable. This limitation is clinically important, as inhalation injury is associated with increased mechanical ventilation requirements, higher pneumonia rates, prolonged intensive care unit (ICU) stay, and greater mortality.15 Furthermore, the discrepancy between the number of patients intubated and those with confirmed inhalation injury across studies indicates a tendency toward overintubation, exposing patients to additional complications and increased healthcare burden.16 Taken together, these findings emphasize the need for improved diagnostic precision and a cautious, evidence-based approach to airway management in resource-limited settings.
Clinical decision aids
Clinical decision aids have emerged as valuable tools to support clinicians in the early recognition of inhalation injuries and the decision to initiate intubation, particularly in settings where bronchoscopy is not accessible. Recent efforts have focused on combining multiple physical findings and patient history factors, for example, variables such as shortness of breath, hoarseness, facial burns, enclosed-space flame exposure, soot in the mouth, and dysphonia, to improve predictive accuracy for airway compromise.17 18 One such prognostic model developed from prehospital and burn-center data showed strong discrimination for bronchoscopy-confirmed inhalation injury, incorporating factors like face burns, soot, hoarse voice, enclosed-space burns, and dyspnea.16 Another more recent score, the Predicting Delayed Intubation scale, identified predictors such as respiratory distress, facial burns, and pharyngolaryngeal swelling, showing excellent discrimination for later airway obstruction.18
These multiparameter tools attempt to mitigate the limitations of single physical signs, which individually suffer from low sensitivity and risk both overintubation and underintubation. By integrating several clinical variables, decision aids provide a more comprehensive risk assessment.19 Patients showing a combination of high-specificity signs, such as stridor or dyspnea along with moderately sensitive ones (facial burns, soot exposure) may be prioritized for airway protection, while those lacking multiple risk factors may be safely observed.19 Use of such structured decision aids can reduce unnecessary intubations, thereby minimizing ventilator-associated complications, reducing ICU stay, and lowering healthcare costs.19
However, the current evidence base for these aids remains limited. Most published models are retrospective, single-center, or derived from specific populations, and very few have been prospectively validated across different burn centers or in low-resource settings. As a result, their generalizability and reliability remain uncertain. Moreover, studies consistently emphasize the need for serial assessment and close airway monitoring, especially when initial signs are absent but risk remains, rather than relying on a one-time evaluation.20
Indication for early intubation
Clinical decisions regarding early intubation in suspected inhalation injury must balance the need for timely airway protection against the risks of unnecessary invasive intervention. Current guidelines provide varying but complementary criteria. The Emergency Management of Severe Burns 2024 guidelines recommend intubation in cases of impending airway obstruction, reduced consciousness, hemodynamic instability, significant facial burns or trauma, need for secretion clearance, and for safe transport or investigation.21 Similarly, the ISBI (International Society of Burn Injuries) states that early intubation is warranted in patients with symptomatic inhalation injury or any thermal injury to the face, mouth, or oropharynx that threatens airway patency.22
Despite these criteria, determining which patients truly require early intubation remains challenging. Some physical signs, such as singed nasal hair, hoarseness, or carbonaceous sputum, are commonly present but lack specificity, meaning many patients displaying these signs do not ultimately have significant inhalation injury.15 Overintubation exposes patients to risks such as ventilator-associated pneumonia, increased resource use, and complications during extubation, emphasizing the need for careful diagnostic judgment.23 Conversely, delayed intervention may lead to airway obstruction as edema progresses, making subsequent intubation more hazardous.23
Strong indicators such as intraoral burns and stridor remain reliable triggers for immediate intubation due to their close association with impending airway compromise.14 Ultimately, the decision hinges on clinical expertise, available resources, and ongoing assessment of airway stability. In resource-limited settings, where bronchoscopy is inaccessible, prioritizing physical findings with the highest diagnostic accuracy, such as dyspnea, stridor, and hoarseness, may help guide safer and more judicious intubation practices.
Strengths and limitations
This study provides a systematic comparison of commonly reported physical examination findings used in the assessment of inhalation injury and identifies which clinical signs demonstrate the highest diagnostic value. By synthesizing available evidence and ranking findings according to DOR, this review offers practical insight that may assist clinicians in making early airway management decisions, particularly in environments where advanced diagnostic tools such as bronchoscopy may not be readily available.
However, several limitations should be considered. First, the overall number of included studies was relatively small, which may limit the robustness of pooled estimates. Second, heterogeneity among studies in terms of patient populations, diagnostic protocols, and clinical settings may influence the observed diagnostic performance of individual physical findings. In addition, most of the available evidence originates from tertiary care centers in high-resource settings, which may limit generalizability to rural or resource-limited environments.
