| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Division of Pediatric Pulmonology, Critical Care and Sleep Medicine (Drs. Sivan, Ben-Ari, and Soferman) and the Pediatric Otolaryngology Unit (Dr. DeRowe), Dana Children’s Hospital, Tel Aviv Medical Center, Tel Aviv, Israel.
Correspondence to: Yakov Sivan, MD, Institute of Pediatric Pulmonology, Critical Care and Sleep Disorders, Dana Children’s Hospital, Tel Aviv Medical Center, 6 Weizman St, Tel Aviv 64239, Israel; e-mail: sivan{at}post.tau.ac.il
Abstract
Rationale: Fiberoptic flexible laryngoscopy (FFL) is the diagnostic procedure of choice in patients with laryngomalacia. Two techniques can be applied, either when the infant is awake or using anesthesia/sedation. The choice of technique may effect the diagnosis.
Study objectives: To compare the two techniques for diagnosing laryngomalacia.
Patients and interventions: A total of 42 infants who underwent awake fiberoptic laryngoscopy for congenital stridor, in whom either laryngomalacia was diagnosed or no cause for stridor was found, underwent a repeat laryngoscopy using anesthesia/sedation. The 84 video recordings of the supraglottic portions were copied onto a videotape along with 25 recordings of normal upper airways without stridor and 31 duplicate cases with stridor. A total of 140 recordings was mixed at random on a videotape. Sound was not included.
Measurements: Three investigators (Y.S., J.B.A., and A.D.) independently scored each recording using a laryngomalacia scoring system (scoring range, 0 to 8).
Results: A threshold score of 2 was the optimal cutoff point for discriminating laryngomalacia from normal condition. The awake technique (WT) missed three cases of laryngomalacia and overdiagnosed one healthy control subject. The anesthesia technique was superior with a sensitivity of 100%, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 100% compared with 93%, 92%, 97%, and 79%, respectively, for the WT.
Conclusions: The diagnosis of laryngomalacia with FFL is more accurate using anesthesia/sedation. The WT may be appropriate for screening or for patients with mild cases having a characteristic presentation.
Key Words: anesthesia • congenital stridor • infant • laryngoscopy • sedation
Laryngomalacia is the most common cause of congenital stridor.1234 It is usually a benign disorder resolving spontaneously by 12 to 18 months of age. The diagnosis of laryngomalacia requires dynamic visualization of the glottic and supraglottic area.
Since the introduction of video-recorded fiberoptic flexible laryngoscopy (FFL), the following two approaches have emerged for the endoscopic evaluation of congenital stridor, especially for laryngomalacia: a clinic-based awake laryngoscopy; and a drug-assisted technique (DAT) using sedation and anesthesia.456789 The awake technique (WT) is performed while the infant is held in a sitting position; topical anesthesia is applied to the nose, and endoscopy is performed to the level of the vocal cords. In the DAT, the patient is supine, and inspections of the subglottic area and the tracheobronchial tree are possible.
The differences between these methods may affect the diagnostic reliability and accuracy of largyngoscopy. Lesions missed during performance of the WT58 have been observed on a following DAT. Body position, level of agitation, crying, and the degree of consciousness may all modify the dynamic motion of the glottic and supraglottic structures, and may shorten the time for careful inspection over several breaths. This influences the physician’s interpretation and ability to make the correct diagnosis. Sedation interferes with the diagnosis by its effects on muscle tone, structural collapsibility, agitation, and increased secretions. The effect of these factors and of the specific technique on the reliability of diagnosing laryngomalacia is unknown. No study has yet compared the two techniques regarding their accuracy in diagnosing laryngomalacia. The purpose of this study was to compare the two techniques for diagnosing laryngomalacia by investigating their sensitivity and specificity, and the rates of false-positive results (ie, overdiagnosis) and false-negative results (ie, missed diagnosis or underscoring) for each technique.
Materials and Methods
Study Protocol
During the study period, all infants with congenital stridor who were referred for evaluation were seen by the pediatric otolaryngologist (A.D.) who performed an awake FFL. When no cause for the stridor was found or when laryngomalacia was diagnosed, the patient underwent a second endoscopy using the DAT. Forty-two consecutive infants met these criteria. Each infant underwent the two procedures < 10 days apart.
The laryngoscopy performed using the WT was performed with the infant held awake in a sitting position by a parent. Topical anesthesia (lidocaine 1% and phenylephrine 0.25%) was applied to the nose, and FFL was performed to the level of the vocal cords. In the DAT, each infant received anesthesia with IV propofol. A dose of 2.5 mg/kg was injected slowly over 15 min as small boluses of 0.5 to 1.0 mL/kg. Additional boluses were added as needed. To prevent pain on injection, 0.3 mL of a 1% lidocaine solution was slowly injected prior to the injection of diluted propofol. All infants received 100% oxygen by a nonrebreathing mask during propofol loading and then a continuous flow of 2 to 3 L/min of 100% oxygen administered directly to the hypopharynx by an 8F feeding tube. The procedures were performed by a pediatric pulmonologist (Y.S.) in the pediatric ICU with the assistance of the pediatric ICU resident and a registered nurse. Oxygen saturation, respirations, BP (automatically every 2 min), and ECG were continuously monitored. The endoscopies were performed using a bronchoscope (model FB-10V or FB-8V; Pentax; Tokyo, Japan) with a distal outside diameter of 3.5 or 2.8 mm, respectively. Initially, only the upper airways were investigated looking for the cause of the stridor. Only when this stage was complete, was a 1% solution of lidocaine applied to the glottis and the bronchoscope advanced to visualize the trachea. All endoscopies were recorded on videotape. The study was approved by the institutional review board of our hospital.
Clinical Scoring
The degree of upper airway obstruction and stridor in each infant was assessed separately by the two investigators who performed the endoscopies using a clinical scoring system based on (1) history and (2) physical examination (Table 1
). This scoring system is a modification of the scoring system for patients with croup10 as applied to laryngomalacia and its chronic nature.
|
|
The accuracy of the DAT and WT in diagnosing laryngomalacia was assessed by comparing the area under the curve (AUC) of the receiver operating characteristics of each test using the methodology described by Hanley and McNeil.12 Models with AUC values between 0.70 and 0.79 are considered as having moderate discriminative properties, and those with AUC values of 
0.80 as having good discriminative properties. Statistical analysis was performed using a statistical software package (SPSS for Windows, version 13.0; SPSS; Chicago, IL).
Results
The mean age for patients in the study group was 4.9 months (55% male patients), and for the control subjects it was 4.4 months (52% male subjects). No procedure-related complications occurred. All patients were discharged from the hospital promptly after undergoing the WT, and 1.5 to 4.5 h after undergoing the DAT (median time, 2.1 h).
There was a high clinical agreement between investigators 1 (Y.S.) and 4 (A.D.) for the laryngomalacia clinical scores with median scores of 8.0 for both investigators (range, 3 to 12 and 4 to 11, respectively; p = 0.36 [Wilcoxon test]). The correlations were high (Kendall coefficient of concordance, 0.932; and Spearman correlation coefficient [r value], 0.863; p < 0.0001). Similar agreements were obtained for the two components of the clinical score (ie, the history and the physical examination scores).
The video-scoring results are presented in Table 3 . No significant differences were found between the original and the repeated video scores (ie, intrainvestigator variability), showing high accuracy and repeatability (Table 4 ). There were no significant differences among the three investigators (ie, interinvestigator differences) regarding all of their total and subcomponent video scores (Table 5 ). There was a high level of agreement among all three investigators for the video scores, including the following: (1) agreement between each two investigators; (2) agreement for arytenoid scores, epiglottic scores, and total scores; (3) agreement for both the study and control groups separately; and (4) agreement for each of the two laryngoscopic techniques separately (ie, Kendall and Spearman tests) [Table 6 ]. These results confirm both a high level of agreement and no significant differences among the investigators for both clinical and video assessments.
|
|
|
|
2 signified laryngomalacia). Further analysis using a cutoff video score of 2 showed that while the DAT scored all laryngomalacia cases as laryngomalacia and all control subjects as not having laryngomalacia, the WT missed three patients with moderate-to-severe cases of laryngomalacia. Their corresponding DAT values were 6.0, 5.3, and 6.3, respectively, and their clinical scores were 8.5, 10, and 11 respectively. The sensitivity, specificity, and positive and negative predictive values for each technique are presented in Table 7
, and show that the DAT was more accurate in comparison with the WT.
|
This study shows that the DAT is more accurate than the WT in the diagnosis of laryngomalacia in infants. Using the DAT, no cases of laryngomalacia were missed and no healthy control subjects were considered to have laryngomalacia. In comparison, the WT missed 3 of 42 cases of laryngomalacia and made an incorrect diagnosis of laryngomalacia in one healthy control subject. The WT is much less time consuming, requires fewer personnel, is much cheaper, and does not require an ICU or operating room facilities.
However, since the DAT was found to be superior as a diagnostic tool, the question that should be addressed is whether these findings indicate that infants with congenital stridor who are evaluated by FFL should undergo the procedure under sedation/anesthesia. A recent study7 presented a series of 66 FFLs performed using the WT. Laryngomalacia was the most common diagnosis (52%). The relatively lower than expected prevalence was probably due to the inclusion of older children. Adequate visualization of the glottis for reaching a diagnosis was achieved in 89% of patients. Eleven percent of the patients did not tolerate the procedure and were, therefore, referred to undergo FFL using the DAT. A further 24% of patients (14 of 59 patients) were referred to undergo FFL using the DAT because of inconclusive findings. Overall, 21 of their 66 patients (32%) were required to undergo FFL using the DAT following the WT. Although this study shows the applicability of the WT, it also highlights the significant failure rate in reaching a diagnosis and the need for a second FFL using anesthesia/sedation. Two publications813 have reported small series of patients in which the results of FFLs performed using the WT were normal and only repeat procedures using the DAT revealed laryngomalacia. The results of these three, albeit uncontrolled, studies7813 strengthen our finding that the most accurate and reliable technique is the DAT. A previous study5 using awake examination also emphasized that the accurate diagnosis of glottic and subglottic abnormalities is hampered by rapid laryngeal motion and lack of patient cooperation, and requires a frame-by-frame reanalysis to improve the diagnostic accuracy of the WT.
Several factors in both the WT and the DAT may affect the accuracy and the ability to reach the correct diagnosis. The infant’s position and the level of consciousness differ. The degree of alertness affects stridor intensity and the dynamic changes in the glottic and supraglottic regions.7813 Mild and moderate stridor are intensified by agitation and crying, and may diminish or even temporarily disappear during sleep. This has been termed state-dependent laryngomalacia.813 Interestingly, this term has been applied to situations in which a conclusive diagnosis was achieved by the DAT after the WT had failed to do so.814
In this study, we did not have a third "gold standard" to which we could compare the two techniques. Therefore, any result that shows a discrepancy between the two methods may be interpreted as overdiagnosis by one technique or as underdiagnosis by the other. We decided to overcome this methodological problem by refining the study group, adding control subjects without laryngomalacia and taking meticulous care to evaluate interinvestigator and intrainvestigator agreements by their correlations, differences, and variabilities. When an infant presents with congenital stridor, a cause must exist. It is unlikely that other causes of congenital stridor were inadvertently included in the study group because all patients with stridor indeed received diagnoses of laryngomalacia, no other causes for congenital stridor were found, all cases followed the natural history of laryngomalacia, and the stridor resolved by 18 months of age. The control subjects were not necessarily healthy infants, but in all control subjects the cause for performing the FFL was not associated with stridor and did not involve the upper airways, which were normal in structure and function.
The weak correlation between the clinical severity and the video score is not surprising. Each part of the supraglottis, the arytenoids or the epiglottis, may be significantly involved, causing moderate-to-severe stridor independently, while the other part is not involved, thus resulting in a low video score with a high clinical score.
One limitation of this study is that we used only one type of drug for the DAT, propofol, which is widely used today for endoscopies in infants and children. Hence, the results may not be generalized to any other drug or combination of drugs. The effect of different drugs and combinations with a variety of dosage protocols should be investigated in a different comparative study. Another possible limitation of the study is that, according to the history score, the patient with intermittent noise whose condition gets worse with agitation would obtain the same score of 1 as the patient with persistent noise whose condition is not getting worse with agitation; nevertheless, the second patient may have a more severe case of laryngomalacia. In the present study, the conditions of all infants with persistent stridor also were worse during agitation, resulting in a higher score. Even though this limitation may not apply to the present study, we suggest that future studies should consider assigning 2 points for "persistent noise."
Based on the results, we conclude that the DAT is superior to the WT because it is more accurate for diagnosing laryngomalacia, at least when propofol is used in the DAT. Having a 100% sensitivity and specificity with 100% positive and negative predictive values, we suggest that the DAT serve as the "gold standard" for the diagnosis of laryngomalacia. Although it is not as accurate, the WT has relatively high sensitivity, specificity, and predictive values, and is useful for screening. When the WT is used for patients with a mild case of laryngomalacia with characteristic presentation, it should be kept in mind and brought up to the parents that a small but not negligible chance of missing laryngomalacia does exist and that the best way to make the diagnosis is by using the DAT for FFL. However, the increased cost and need for sedation can be avoided. In all cases, when a definite diagnosis of the upper airway lesion causing stridor cannot be missed such as in all patients with moderate-to-severe cases, in cases in which the presentation is not characteristic of laryngomalacia, and whenever there is a considerable likelihood that infraglottic visualization should be conducted, we suggest that the patient undergo FFL using the DAT.
Footnotes
Abbreviations: AUC = area under the curve; DAT = drug-assisted technique; FFL = fiberoptic flexible laryngoscopy; WT = awake technique
The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Received for publication October 3, 2005. Accepted for publication April 21, 2006.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
