The Role of Airway Fluoroscopy in the Evaluation of Stridor in Children

Objective  To determine the role of airway fluoroscopy in comparison with other diagnostic modalities in diagnosing the site of partial airway obstruction in children with stridor.

Design  Prospective study comparing direct laryngoscopy and bronchoscopy with nasopharyngoscopy, airway fluoroscopy, and plain films. Children with stridor or partial airway obstruction were evaluated by the Department of Otolaryngology at Columbus Children's Hospital, Columbus, Ohio. A history review and physical examination, including flexible fiberoptic laryngoscopy, plain films, airway fluoroscopy, and direct laryngoscopy and bronchoscopy, were performed for all children.

Setting  Tertiary care children's hospital.

Patients  From November 1996 to September 1999, 64 children aged 1 week to 12 years, with a mean age of 1.8 years and male-female ratio of 3:2, were evaluated for stridor.

Main Outcome Measures  The sensitivity and specificity of airway fluoroscopy in diagnosing the site of partial airway obstruction in comparison with nasopharyngoscopy and plain films.

Results  Airway fluoroscopy had a sensitivity of 80% for subglottic, 73% for tracheal, and 80% for bronchial sites of obstruction. It was less sensitive for supraglottic and glottic sites—33% and 14%, respectively. Nasopharyngoscopy was more sensitive for supraglottic and glottic sites of obstruction. Overall, airway fluoroscopy was far more sensitive than plain films for diagnosing site of obstruction.

Conclusions  Airway fluoroscopy is a quick, noninvasive, and dynamic study of the entire airway that provides important additional information to the history review and physical examination and is a valuable adjunct to flexible fiberoptic laryngoscopy. It was far superior to plain films and may serve as a cost-effective screening tool in the evaluation of stridor in children, especially for lesions of the lower airway.

STRIDOR IS NOT a diagnosis but a symptom produced by turbulent airflow related to partial airway obstruction.^1,2 The obstruction may be fixed or dynamic and there are many causes: congenital, traumatic, iatrogenic, inflammatory, and neoplastic. History review and physical examination aid in the diagnosis. Radiologic modalities can provide additional information that enhances the history review and physical examination in certain instances, but their routine use is controversial. Plain films are obtained in many instances to add to the diagnosis; however, retrospective studies have shown them to be of little use, especially for functional causes of stridor.³ Computed tomography (CT) and magnetic resonance imaging (MRI) are useful in select cases, but their expense does not make them cost-effective as a screening tool. Flexible fiberoptic laryngoscopy (FFL) provides visualization from the nasopharynx down to the vocal folds but cannot accurately evaluate the subglottis or trachea, which may represent the primary site of obstruction or a synchronous airway lesion.^4,5 Direct laryngoscopy and bronchoscopy (DLB) is the gold standard of diagnosis but entail the risk of anesthesia and greater costs.⁶ While airway fluoroscopy is a dynamic study that can evaluate multiple sites of obstruction simultaneously, it is not routinely used in the initial evaluation of stridor. There is little information in the otolaryngology literature about the role of airway fluoroscopy in the evaluation of stridor in children

The study population comprised 95 children who presented with stridor or partial upper airway obstruction to the Department of Otolaryngology at Columbus Children's Hospital, Ohio State University, Columbus, from November 1996 to September 1999. Sixty-four patients met the inclusion criteria. The patients were aged 1 week to 12 years, with a mean age of 1.8 years and a male-female ratio of 3:2. A thorough history review and physical examination, which included FFL with either a neonatal (1.8-mm) or pediatric (3.2-mm) endoscope, were performed on all patients by a staff physician in the Department of Otolaryngology.

Patients were given general consent for radiographic studies, and informed consent was obtained prior to any surgical procedures. The institutional review board of Columbus Children's Hospital authorized the waiver of consent for enrollment into this study.

Anteroposterior and lateral radiographs of the soft tissues of the neck and chest were performed using standard protocols created by the Department of Radiology at Columbus Children's Hospital. This included low-kilovolt technique and imaging during inspiration and expiration. Airway fluoroscopy was performed by an attending pediatric radiologist versed in imaging of the airway. Careful attention to limited dosimetry was undertaken. Fluoroscopic times ranged from 1 to 2 minutes at 4 to 7 mR/min. After evaluating movements of the hemidiaphragms and observing for focal air trapping, imaging of the airway from nasopharynx to main bronchi was performed in anteroposterior, oblique, and lateral projections. Spot images and videofluoroscopy were obtained.

Direct laryngoscopy and bronchoscopy was performed on this group of patients because of severity, progression of symptoms, clinical features, or certain radiographic abnormalities. Endoscopy was performed with a Hopkins system of rigid telescopes and a camera attachment for video recording. General anesthesia was induced via inhalational technique while maintaining spontaneous ventilation to allow the assessment of dynamic airway characteristics such as respiratory phase variability and vocal cord mobility. This technique was augmented by the use of topical anesthesia using 4% lidocaine hydrochloride atomization of the airway.

Data sheets were used to record history, associated symptoms, physical examination findings (including those from FFL), radiography results, DLB results, diagnosis, and therapeutic interventions. All information was recorded onto data sheets with 120 data fields and entered into a Microsoft Excel database (Microsoft Corp, Redmond, Wash). Medical record review was used to supplement and verify information in some instances.

Comparisons between FFL, plain films, airway fluoroscopy, and DLB were undertaken to determine the accuracy of diagnosis by site of obstruction. Statistics were performed using Statistical Package for the Social Science (SPSS 9.0) software (SPSS Inc, Chicago, Ill). Pearson χ² analysis was used to measure the association between DLB (the gold standard) and other diagnostic tests. Sensitivity and specificity were determined by site of obstruction. Descriptive statistics were also used to characterize the patient population.

Sixty-four children with stridor or partial airway obstruction were evaluated with a thorough history review and physical examination, including FFL, plain films, airway fluoroscopy, and DLB. From this patient population, 52 (81%) had a history of intubation, 53 (83%) had a history of gastroesophageal reflux disease, and 95% had progression of symptoms. Significant associated symptoms included 31 patients (48%) with retractions, 44 (69%) with shortness of breath, and 63 (98%) with active stridor. The primary diagnoses are listed below.

Of the patients, 17% had significant secondary diagnosis as given below.

Treatment of stridor was medical in 27 patients (42%) and surgical in 37 (59%).

Flexible fiberoptic laryngoscopy, airway fluoroscopy, and plain films were compared with DLB for site of obstruction at the supraglottis, glottis, subglottis, trachea, and bronchi. Nasopharyngoscopy was more sensitive for supraglottic and glottic lesions, detecting 89% (8 of 9) and 71% (5 of 7), respectively, but was less sensitive in detecting subglottic lesions, detecting only 33% (10 of 30). Airway fluoroscopy was less sensitive to supraglottic and glottic lesions, detecting only 33% (3 of 9) and 14% (1 of 7), respectively, but was much more sensitive in detecting subglottic, tracheal, and bronchial lesions, detecting 80% (24 of 30), 73% (8 of 11), and 80% (8 of 10), respectively. Plain films were the least sensitive of any diagnostic modality, with a sensitivity of 0% (0 of 9) for supraglottic, 0% (0 of 7) for glottic, 33% (10 of 30) for subglottic, 0% (0 of 11) for tracheal, and 40% (4 of 10) for bronchial sites of obstruction. Oropharyngeal collapse was present in 2 patients, and both were detected by airway fluoroscopy, whereas 1 was missed by FFL and plain films failed to detect any abnormalities. Because of the small sample of oropharyngeal collapse as a cause of stridor in this group of patients, sensitivity and specificity could not be determined. However, it would appear that fluoroscopy is much more sensitive to oropharyngeal collapse than plain films or FFL.

In 1978, Smith et al⁷ were the first to report on the value of fluoroscopy in diagnosing upper airway obstruction. Obstructive sleep apnea was demonstrated in a 12½-year-old patient with achondroplastic dwarfism on fluoroscopic examination. Since then, a few studies have used fluoroscopy to diagnose obstructive sleep apnea, but little has been reported on its usefulness in stridor.^8,9 Airway fluoroscopy is used as an adjunct to standard imaging if further dynamic information is needed and if findings of plain films are within normal limits.¹⁰

While plain radiographs are usually the first and only radiographic studies ordered for the evaluation of children with stridor, their usefulness is controversial. In a retrospective analysis of 100 infants, Tostevin et al³ found that plain radiographs and barium swallow were of little value in the preendoscopic assessment of children with stridor. When comparing DLB with preendoscopic plain films, their data showed that plain films failed to make the diagnosis in most cases of stridor. The correlation of plain films with the gold standard was extremely poor in instances in which stridor was functional, but improved when stridor was due to fixed lesions. Laryngomalacia is the most common cause of stridor in infants—up to 60% in most series.¹ It is not surprising that plain films would fail to detect dynamic abnormalities. It should be noted that fluoroscopy was not one of the modalities used in their series. Tostevin et al³ recommended chest radiographs before general anesthesia to identify any major vascular or pulmonary malformations but did not assess their value in the diagnosis of stridor. They conclude that radiology had a limited screening role in the evaluation of stridor and further imaging and a barium swallow should be performed only if an abnormality is found at DLB.³ However, these examinations will have a low diagnostic yield after DLB has already been performed.

Friedman et al,⁴ in a retrospective analysis, found that only 18.5% of plain films were positive in endoscopically proven lesions. Our prospective analysis concurs with the results of Friedman et al, since plain films were only able to detect 20% of endoscopically proven lesions in our series. Airway fluoroscopy was far more sensitive than plain films for all sites of airway lesions and correlated with 67% of endoscopically proven lesions.

Fluoroscopy has had significant role in evaluating foreign bodies. Blazer et al,¹¹ in a review of 200 cases, found that fluoroscopy contributed to the diagnosis in 90% of bronchial foreign bodies but in only 32% of those cases above the carina. In our series, airway fluoroscopy detected 80%, whereas plain films only detected 40% of bronchial lesions. Whereas plain films can easily reveal radiopaque foreign bodies, radiolucent foreign bodies pose a diagnostic dilemma and account for most bronchial foreign bodies. Airway fluoroscopy can evaluate movements of the hemidiaphragms and check for focal air trapping. Imaging of the entire airway from nasopharynx to main bronchi can be performed in anteroposterior, oblique, and lateral projections. Spot images and videofluoroscopy can be obtained to help discern fixed from functional lesions. Although airway fluoroscopy provides valuable information in the evaluation of airway foreign bodies, a history suggestive of airway foreign body precludes the use of fluoroscopy as a diagnostic modality because the patient will require endoscopy regardless of fluoroscopy results.

The distinct advantage of evaluating the entire airway in a dynamic, not static, manner allows physicians to see in real time what is causing partial airway obstruction. Fluoroscopy is especially helpful in evaluating dynamic lesions and better at correlating their severity, extent, and location than other radiographic modalities. Fluoroscopy also has an advantage of not requiring the same degree of cooperation as plain films. The radiologist can position the patient and the fluoroscopy machine so that technical problems, which occur with plain films, become virtually nonexistent. Images can be captured even if a position is only momentary. A fixed lesion, such as a subglottic web, can be seen throughout the respiratory cycle, whereas a functional lesion, such as laryngomalacia, changes. Airway fluoroscopy provides a safe, quick, and noninvasive way of evaluating the entire airway, and information about multiple sites of obstruction can be obtained.

Synchronous or multiple airway lesions present a diagnostic dilemma for the otolaryngologist. Holinger¹ reported in 1980 on a series of 219 children presenting with stridor in which 45% of patients had more than 1 anomaly. He recommended early endoscopy to accurately diagnose the cause of the stridor.¹ In 1987, Gonzalez et al¹² reported on a series of 103 patients with 17.5% synchronous airway lesions and recommended that workup of the entire respiratory tract (ie, bronchoscopy and radiographic studies) was essential because laryngoscopy alone may fail to detect concurrent disorders in infants with airway obstruction. Friedman et al,⁴ in a retrospective series of 60 patients, only found 12% of synchronous airway lesions, but 65% of them were below the vocal cords. Flexible fiberoptic laryngoscopy is less accurate below the glottis and would not be able to adequately evaluate children for synchronous airway lesions.

Most recently in 1996, the recommendations of Mancuso et al¹³ differed from those of previous authors. A retrospective series of 233 patients diagnosed with laryngomalacia by FFL (90 patients underwent DLB) discovered synchronous airway lesions in 18.9% of the patients.¹³ They believed that rigid endoscopy was rarely necessary in evaluating laryngomalacia in infants owing to the relatively benign clinical nature of synchronous airway lesions and because FFL and plain films usually sufficed in diagnosing laryngomalacia and synchronous airway lesions.¹³ However, of the 44 patients in whom synchronous airway lesions were found, only 9 (22.5%) of the 40 plain films were abnormal.¹³ Mancuso et al stated that plain films were useful, but their data did not support this claim. It is interesting that airway fluoroscopy was obtained in 16 patients. These tests were not ordered by otolaryngologists, and the authors could not glean from the medical record the specific indication for ordering airway fluoroscopy.¹³ Fluoroscopy identified tracheal lesions in 12 patients, whereas DLB identified tracheomalacia in only 9 of these patients. It appears that fluoroscopy was more sensitive in diagnosing dynamic tracheal lesions than DLB. Mancuso et al noted that the utility of radiologic studies in identifying synchronous airway lesions has not been well addressed in the literature and that they could draw no specific conclusions about the utility of airway fluoroscopy from this small sampling.¹³

Gibson et al¹⁴ found sleep fluoroscopy to be a valuable adjunct to endoscopy in the identification and management of pediatric upper airway obstruction when hypopharyngeal collapse or multiple levels of obstruction are suspected. Their study was done prospectively with a population of 50 children who had complex airways and possibly multiple sites of upper airway obstruction.¹⁴ The technique of sleep fluoroscopy was used after all patients had already underwent DLB. They found the sensitivity for sleep fluoroscopy to be 100% (9 of 9) for supraglottic lesions, 92% (12 of 13) for tracheal lesions, 70% (7 of 10) for subglottic lesions, and 67% (2 of 3) for vocal cord lesions.¹⁴ The correlation of fluoroscopy with endoscopy was exactly the same in 44% of the patients.¹⁴ Fluoroscopy showed multiple of sites of obstruction in 16 patients (whereas DLB found only single sites) and found abnormalities in 11 patients (whereas the findings from DLB were normal).¹⁴ Furthermore, fluoroscopy altered the course of treatment in 52% of the study group (influencing a procedure not anticipated in 17 patients and obviating the need for surgery in 3 patients) and confirmed the treatment plan in the remaining 48%.¹⁴ One of the main disadvantages of sleep fluoroscopy was the need to sedate children with potentially tenuous airways, but overall it possessed significant advantages in diagnosing upper airway obstruction and had a significant impact on patient treatment.

Awake airway fluoroscopy has advantages over sleep fluoroscopy. Sleep fluoroscopy requires sedation and staffing by an otolaryngologist and takes about an hour (30 minutes until the patient is adequately sedated, 1 minute of fluoroscopy time, and then 30 minutes of recovery time). Awake airway fluoroscopy requires no sedation, is performed by staff radiologists (usually in about 1-2 minutes), and does not require the presence of an otolaryngologist to manage the airway. Furthermore, awake fluoroscopy can better evaluate vocal cord mobility, even in infants, since crying serves as adequate phonation.

The main disadvantage of fluoroscopy is increased radiation exposure. While fluoroscopy creates additional radiation exposure compared with plain films only, it poses no immediate risk to patients and is currently used at Columbus Children's Hospital for diagnostic purposes. To our knowledge, there is no literature to support an increased risk of malignancy after airway fluoroscopy, which is usually less than 1 minute and equivalent to 10 rads (0.1 Gy) of radiation. However, all efforts should be made to limit doses to what is diagnostically necessary.

Costs of different diagnostic modalities for stridor are given below.

These represent costs of radiographs, including professional fees. The cost of DLB includes anesthesia fees, otolaryngology fees, and facility fees, including operating room time, supplies, and personnel. Airway fluoroscopy is less expensive and far more sensitive than a battery of plain films. If radiographic information is needed to help in the evaluation of stridor in children, airway fluoroscopy is the best test to order and would be more cost-effective than plain films.

The ideal study would prospectively perform DLB on all patients undergoing airway fluoroscopy and be conducted so that the otolaryngologist and radiologist would be blinded to their results. While the symptom of stridor can represent serious airway problems, most cases of stridor are benign and self-limiting, and DLB is not indicated in most circumstances. Performing DLB for the workup of every patient with stridor would not be cost-effective and would place children at risk from general anesthesia. While performing a DLB on all patients with stridor would be good science, it would not be good medical practice; thus, we believe it was not justified for the purposes of our study. Although this places some limitations on our study, important information was gained on endoscopic and radiographic correlation. This patient population was biased toward more serious causes of stridor, with 46% having subglottic pathological features. Most of our patients had progressive symptoms with active stridor for which DLB was indicated. This could make the sensitivity higher for airway fluoroscopy, since lesions that are more clinically significant may be more readily apparent on radiographic studies; however, plain films lacked sensitivity in spite of the same clinical circumstances.

Recently, there have been pilot studies of newer technologies, such as ultrafast CT and MRI, to evaluate obstructive sleep apnea and upper airway anatomy.^15-18 Although these techniques provide dynamic information similar to fluoroscopy, they are much more expensive and much less available than fluoroscopy, which severely limits their usefulness. Fluoroscopy is widely available and relatively inexpensive compared with these newer technologies. While a few reports have alluded to the possible benefits and advantages of using fluoroscopy in the diagnosis of stridor, literature that evaluates the utility of this imaging modality, especially in a prospective manner, is lacking.

Airway fluoroscopy appears to aid in the diagnosis of oropharyngeal collapse and has a sensitivity of 80% for subglottic lesions, 73% for tracheal lesions, and 80% for bronchial causes of upper airway obstruction. Although it is less sensitive for supraglottic and glottic lesions, FFL is the most sensitive for these sites. A thorough history review and physical examination, with FFL to evaluate the upper airway and airway fluoroscopy to provide information about the lower airway, appears to be the most cost-effective evaluation of stridor in children. Airway fluoroscopy provides important additional information to the history review and physical examination and is a useful adjunct to endoscopy.

Corresponding author and reprints: Gregory J. Wiet, MD, Children's Hospital Outpatient Care Center, 555 S 18th St, Suite 6B, Columbus, OH 43205.

Accepted for publication August 23, 2002.

This study was presented at the American Society of Pediatric Otolaryngology at the Combined Otolaryngological Spring Meeting, Orlando, Fla, May 18, 2000, and was first place recipient of the Charles F. Ferguson Clinical Research Award, May 18, 2000.

We would like to acknowledge the support of Pamela Walters for her assistance in the preparation of the manuscript and John Hayes for his help in the statistical analysis.