Factors that predict the need for intubation in patients with smoke inhalation injury.

Abstract

Early identification of smoke inhalation patients who will require intubation is crucial. We conducted a retrospective chart review to identify predictors of respiratory distress in patients who present with smoke inhalation injury. Our study involved 41 patients who had been treated in the emergency room at a regional burn center. Eight of these patients required intubation. Intubation was positively correlated with physical examination findings of soot in the oral cavity (p < 0.001), facial burns (p = 0.025), and body burns (p = 0.025). The need for intubation was also predicted by fiberoptic laryngoscopic findings of edema of either the true vocal folds (p < 0.001) or the false vocal folds (p < 0.01). No statistically significant correlation was found between intubation and any of the classic symptoms of smoke inhalation: stridor, hoarseness, drooling, and dysphagia (all p = 1.0). Also, multivariate analysis revealed that facial burns correlated significantly with edema of the true vocal folds (p = 0.01) and body burns correlated significantly with edema of both the true (p = 0.047) and false (p = 0.003) vocal folds. We conclude that patients with soot in the oral cavity, facial burns, and/or body burns should be monitored closely because these findings indicate a higher likelihood of laryngeal edema and the need for intubation.

Introduction

The clinical spectrum of smoke inhalation injuries is varied. Many patients succumb to their injuries prior to hospitalization. Those who do reach the emergency room alive represent an interesting challenge for the entire burn team. The diagnosis of smoke inhalation injury is based on clinical findings such as singed nasal hairs, intraoral soot, and signs of respiratory distress, including stridor, hoarseness, drooling, and dysphagia. The severity of smoke inhalation injuries varies greatly. Some patients are asymptomatic and are discharged home while others progress to respiratory distress and require intubation and intensive care. The challenge for emergency room physicians and consultant otolaryngologists lies in predicting which patients require intubation.

Many different methods of predicting the clinical course of smoke inhalation patients have been proposed. The low sensitivity and specificity of chest x-rays and clinical symptoms render these evaluations unreliable. (1) Other methods range from basic procedures such as measuring arterial blood gas and performing pulse oximetry to more elegant options such as the [sup.133]xenon ventilation scan. Otolaryngologists in the acute care setting are often called on to evaluate a patient's airway via fiberoptic laryngoscopy. However, reports in the literature on the value of this and other fiberoptic endoscopic examinations are contradictory. (1-4) In 1976, Agee et al reported that the diagnostic accuracy of fiberoptic endoscopy could be improved by the addition of a [sup.133]xenon perfusion scan. (2) Moreover, fiberoptic bronchoscopy has been recommended as part of the initial assessment in smoke inhalation cases to determine the extent of airway injury. (1,3)

In addition to diagnosis, fiberoptic endoscopy is used to determine the need for intervention but, again, results are not clear-cut. Clark found that bronchoscopic findings were rarely the deciding factor in determining whether to intubate a patient. (4) Rather, the history and clinical presentation were most useful in guiding diagnosis and treatment.

At our institution, the otolaryngology service is routinely consulted to perform fiberoptic laryngoscopy as part of the initial assessment of patients with possible smoke inhalation injury. Symptoms of stridor, hoarseness, drooling, and dysphagia are classic indicators of airway compromise. The presence or absence of these symptoms, combined with findings on fiberoptic laryngoscopy, is used to determine the status of respiratory compromise. In this article, we describe our attempt to identify signs and symptoms that can indicate an elevated risk of impending respiratory distress and thus the need for intubation.

Patients and methods

We conducted a retrospective chart review of all smoke inhalation cases that included a consultation by a member of the Department of Otolaryngology at the Jacobi Medical Center between Sept. 1, 1998, and March 31, 2003. We initially identified 48 cases. Seven of these cases were excluded from the study for various reasons: 4 patients had been intubated in the field, 2 patients had been transferred to a pediatric burn unit at another institution, and follow-up data on 1 patient were inadequate. Therefore, our study population was made up of 41 patients--28 males and 13 females, aged 3 to 78 years (mean: 36).

In addition to intubation status, four categories of data were recorded: each patient's age, clinical symptoms, findings on physical examination, and findings on fiberoptic laryngoscopy. Clinical symptoms included the aforementioned stridor, hoarseness, drooling, and dysphagia. Physical examination findings included the presence or absence of soot in the nose or oral cavity and the presence or absence of facial and body burns. Laryngoscopic findings included the presence or absence of soot in the vocal folds and the presence or absence of edema of either the true or false vocal folds.

A chi-squared ([chi square]) test was used to ascertain correlations between intubation and the different variables. Statistical analysis was performed with a Web ([chi square]) calculator (available on the Internet at www.georgetown.edu/cball/webtools/web_chi.html). The threshold for statistical significance was p < 0.05. In addition, to determine if findings on physical examination can predict findings on laryngoscopic examination, we also performed a multivariate analysis with a Pearson's correlation model.

Results

Only 8 of the 41 patients (19.5%) required intubation (table 1). Four patients had been intubated by an emergency room physician after developing worsening hoarseness; 2 patients had been intubated after undergoing a change in mental status secondary to hypoxia (oxygen saturation: <80% on a 100% nonrebreather mask); 1 had been intubated after developing stridor; and 1 had been intubated for airway protection after she became increasingly anxious and agitated.

Symptoms. Intubation was required for 1 of 3 patients with stridor, 4 of 16 with hoarseness, 1 of 2 with drooling, and 1 of 2 with dysphagia. No statistically significant correlations were observed.

Physical examination findings. Statistically significant correlations were observed between intubation and the presence of soot in the oral cavity, facial burns, and body burns. Soot in the oral cavity was seen in 11 patients, 6 of whom were intubated (p < 0.001). Likewise, intubation was performed on 7 of 19 patients with facial burns (p = 0.025) and 7 of 19 with body burns (p = 0.025). (With respect to burns, there was no significant correlation when the percentage of the total body surface area that was burned was analyzed). No significant correlation was seen between intubation and the presence of soot in the nose and/or singed nasal hairs, as only 6 of 27 patients required intubation.

Laryngoscopic examination. Significant correlations were seen between intubation and edema of the true vocal folds (6 of 8 patients; p < 0.001) and edema of the false vocal folds (4 of 7 patients; p < 0.01). Intubation was not significantly associated with soot in the vocal folds.

Multivariate analysis. Facial burns correlated significantly with true vocal fold edema (p = 0.01), and body burns correlated significantly with both true vocal fold edema (p = 0.047) and false vocal fold edema (p = 0.003) (table 2). Soot in the oral cavity did not correlate with edema of the true vocal folds or the false vocal folds.

Discussion

Most victims of smoke inhalation do not require intubation, but it is important to quickly identify those who do. Despite the lack of consensus in the literature on the need for fiberoptic endoscopic examination, we perform it routinely during the initial evaluation at our institution's burn center.

Our study did not demonstrate that the classic symptoms of smoke inhalation injury--stridor, hoarseness, drooling, and dysphagia--are associated with the need for intubation. Nor were these symptoms correlated with laryngeal edema. This finding might be explained by the fact that only a small number of patients presented with these symptoms, which limited the statistical power of this analysis. Another factor that limited the power of our study is the inherent subjectivity of the clinical evaluation and the descriptive terms used to document physical findings. Examinations were performed by different otolaryngology residents, and different emergency room physicians were responsible for determining which patients required intubation. Also, a few patients had to be excluded from the study because of inadequate follow-up or incomplete data on chart notes. But even if the charts had included data on all variables, patients were still seen in the emergency room at different times following the onset of injury. Likewise, the length of time between admission and otolaryngologic consultation varied depending on the status of the emergency room, the severity of the case, and the availability of the otolaryngologist. Therefore, the condition of any given patient may have improved or worsened prior to consultation. Finally, the condition of all patients in this study was serious enough to warrant an otolaryngologic consultation; presumably, other patients with less severe injuries were examined by an emergency room physician and deemed safe for discharge without an otolaryngologic evaluation.

Yet despite these limitations, we believe that our findings are significant enough to have an impact on the routine management of smoke inhalation. We conclude that patients who present with soot in the oral cavity, facial burns, or body burns should undergo fiberoptic laryngoscopy by an otolaryngologist to look for laryngeal edema because they are much more likely to require intubation. Patients without these signs rarely require intubation, and they should not routinely undergo fiberoptic laryngoscopy.

References

(1.) Pruitt BA Jr., Cioffi WG. Diagnosis and treatment of smoke inhalation. J Intensive Care Med 1995;10:117-27.

(2.) Agee RN, Long JM III, Hunt JL, et al. Use of 133xenon in early diagnosis of inhalation injury. J Trauma 1976; 16:218-24.

(3.) Weil RB, Capozzi A, Falces E, Ghatan J. Smoke inhalation study. Ann Plast Surg 1980;4:121-7.

(4.) Clark WR Jr. Smoke inhalation: Diagnosis and treatment. World J Surg 1992;16:24-9.

Dilip D. Madnani, MD; Natalie P. Steele, MD; Egbert de Vries, MD

From the Department of Otolaryngology, Jacobi Medical Center, Albert Einstein College of Medicine, New York City.

Reprint requests: Egbert de Vries, MD, Department of Otolaryngology, Jacobi Medical Center, 1400 Pelham Parkway South. Room 5N49, Bronx, NY 10461. Phone: [phone]; fax: [phone]; e-mail: [e-mail]

Originally presented at the Eastern Section Meeting of the Triological Society: Jan. 23-25, 2004: Boston.

Table 1. Analysis of variables associated
with intubation in the 41 patients

                                    No. patients

Variable                Intubated   Not intubated   p Value

Age
  <50 yr                   5            24           1.0
  [greater than or
    equal to] 50 yr        3             9           1.0

Symptom
  Stridor                  1             2           1.0
  Hoarseness               4            12           1.0
  Drooling                 1             1           1.0
  Dysphagia                1             1           1.0

Physical exam
  Soot in nose             6            21           0.1
  Soot in oral cavity      6             5          <0.001
  Facial burns             7            12           0.025
  Body burns               7            12           0.025

Laryngoscopic exam
  Soot in vocal folds      4            10           0.1
  True fold edema          6             2          <0.001
  False fold edema         4             3          <0.01

The threshold for statistical significance is p < 0.05.

Table 2. Multivariate analysis

                     Soot in   True    False   Facial   Body
Variable              oral     fold    fold    burns    burns
                     cavity    edema   edema

Soot in oral cavity
  Pearson              N/A     0.257   0.252   0.087    0.262
  p Value                      0.104   0.117   0.594    0.098
  N                             41      40       40      41

True fold edema
  Pearson             0.257     N/A    0.428   0.401    0.312
  p Value             0.104            0.006    0.01    0.047
  N                    41               40       40      41

False fold edema
  Pearson             0.252    0.428    N/A    0.294    0.452
  p Value             0.117    0.006           0.069    0.003
  N                    40       40               39      40

Facial burns
  Pearson             0.087    0.401   0.294    N/A     0.716
  p Value             0.594    0.01    0.069            0.000
  N                    40       40      39               40

Body burns
  Pearson             0.262    0.312   0.452   0.716     N/A
  p Value             0.098    0.047   0.003   0.000
  N                    41       41      40       40

The threshold for statistical significance is p < 0.05.