INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Laryngomalacia usually is a benign, self-limited condition that has a typical history and physical examination. It is relatively common, accounting for as much as 60 to 75% of laryngeal problems in newborns and infants (1). Although most children presenting with typical findings and history of laryngomalacia require no additional evaluation or treatment, 7 to 20% of patients with laryngomalacia present with complications that might require intervention such as hypoxemia, feeding difficulty, a second airway lesion, failure to thrive, apnea, and even cor pulmonale (4, 5). Patients rarely die of complications of laryngomalacia.

As many as 10 to 15% of patients who present with laryngomalacia eventually will require an intervention (1, 6) such as endoscopic laser aryepiglottoplasty (removal of redundant tissue over the arytenoids and aryepiglottic folds) (6). Although this is a procedure with low morbidity and mortality, it has some complications, including scarring, laryngeal stenosis, dysphagia, aspiration, and recurrent stridor or dyspnea.

When the clinically diagnosed case of laryngomalacia does not follow its expected course of gradual improvement or when it is severe, the affected children often undergo flexible fiberoptic bronchoscopy (FFB) to rule out other causes of stridor, to assess severity, and to examine the lower airways for additional abnormalities. The bronchoscopist commonly observes any or all of the following: a prolapsing epiglottis, large floppy arytenoids prolapsing into the glottis during inspiration, and short aryepiglottic folds (1).

Published descriptions of FFB techniques in children are similar (4, 11). Care during FFB includes continuous monitoring of oxygen saturation, respiratory rate, and heart rate, along with frequent measurement of blood pressure. Typically, the patients receive one or more drugs for conscious sedation. Topical anesthesia is applied to the nasal mucosa and to the upper and lower airway mucosa to reduce pain, prevent laryngospasm, and reduce cough. As a first step in achieving topical anesthesia, several milliliters of 2% preservative-free lidocaine solution are instilled through the nose prior to insertion of the bronchoscope.

However, we have observed in some children that applying lidocaine solution to the upper airway structures for topical anesthesia during FFB results in worsening stridor and clinical signs of upper airway obstruction. If topical lidocaine alters function of the larynx, such an effect would change the assessment of laryngomalacia, presumably resulting in overestimates of its severity. In turn, this increase in laryngeal collapse increases the difficulty of the lower airway examination and consequently might result in missed diagnoses or incorrect assessment of lower airway problems. In addition, more frequent surgical intervention than necessary might result. It will cause discrepancies between assessments of upper airway function of those who examine the upper airway with topical anesthesia and of those who use no topical anesthesia.

In response to our observation of the apparent effect of topical lidocaine on laryngeal structures, we altered our bronchoscopy protocol to confine initial topical anesthesia to the nasal mucosa. With initial anesthesia confined to the nasal mucosa, we documented on videotape the effects of directly applying lidocaine solution to the larynx. We found that in many, especially younger patients, application of lidocaine resulted in an increase in laryngeal collapse and stridor, distorting and exaggerating the severity of their laryngeal dysfunction. This report summarizes a review of our experience with this altered protocol and a blinded comparison of laryngeal function before and after the application topical lidocaine.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Over a 5-yr period, one of the authors (D.N.) performed FFB in 280 children at Primary Children's Medical Center in Salt Lake City, Utah. These children presented with the initial clinical findings listed in Table 1. These diagnoses were taken directly from the operative report or, if the operative report was unavailable, from the history recorded in the patient record. Patients who presented with the finding of "noisy breathing" had airway sounds that did not fit neatly into the categories of stridor (loud, inspiratory crowing) or wheezing (high-pitched, expiratory). Three had more than one airway examination.

Bronchoscopy Protocol

FFB followed a consistent protocol, beginning with informed consent from a parent or guardian before each procedure. After explaining the reasons for the procedure and describing it, we listed its benefits and risks such as fever, infection, bleeding, pneumothorax, hypoxemia and respiratory arrest. We described the precautions taken to prevent complications and to respond to any difficulties.

All airway examinations were performed in the Pediatric Endoscopy Laboratory at Primary Children's Medical Center. Two well-trained nurses assisted the endoscopist with each procedure, continuously evaluating each patient's respiratory and cardiovascular status. Patients were given nothing by mouth for at least 4 h prior to bronchoscopy. After placing a small intravenous catheter in a peripheral vein, we sedated each patient with midazolam and nalbuphine, generally titrating the dose to 0.1 to 0.2 mg/kg, to achieve satisfactory conscious sedation. We continuously monitored patients with a single channel electrocardiograph, an impedance respirograph, and a pulse oximeter. An automated blood pressure monitor provided data every 5 min. Patients received supplemental oxygen as needed to maintain oxygen saturation above 90%.

Initial topical anesthesia was provided by 1 to 2 ml of 2% preservative-free topical lidocaine hydrochloride gel applied to a nasal passage, usually on the right, with a small-diameter swab (Calgiswab; Spectrum Laboratories Inc., Houston, TX). We chose the bronchoscope (BF type P20D, 3C10, 3C20, 27M, or N20; Olympus Corporation, Tokyo, Japan), with a diameter of 2.6, 3.6 or 5.0 millimeters, depending on the size of the patient's airway. A compact camera (currently model EVIS OVC-100 Olympus Optical Co. Ltd., Tokyo, Japan) attached to the bronchoscope and connected to a video recorder (currently model SVO 9600; Sony, Tokyo, Japan) recorded each procedure at standard-play speed on a commercial VHS tape. After advancing the bronchoscope through the anesthetized nasal passage, we examined the larynx. After this initial examination, we positioned the bronchoscope just above the vocal cords and infused 1 ml of 2% preservative-free lidocaine solution through the bronchoscope channel, spraying the lidocaine onto the larynx and vocal cords. This infusion was repeated as needed until we achieved adequate topical anesthesia, usually applying one or two doses in the patients younger than 1 yr of age. The children older than 1 yr of age occasionally required as many as four doses. A diminished cough reflex in response to touching the larynx indicated adequate topical anesthesia. We examined the larynx after each lidocaine infusion. Examination of the lower airways followed. During examination of the lower airway, 1% preservative-free lidocaine solution was infused as needed for cough suppression, observing the total maximum mg/kg dose allowed (5 mg/kg). After examination of the lower airway, the inpatients recovered in their hospital room. Outpatients recovered in the Outpatient Recovery Room adjacent to the Endoscopy Laboratory.

In order to rule out effects not specific to lidocaine, in 10 additional patients we added a saline control to the procedure. Eight of these patients were 5 mo of age or younger, one was 12 mo of age, and the last was 66 mo of age. They were chosen because of their history of stridor and physical findings consistent with laryngomalacia. In these infants we scored laryngeal function before infusion of any solution through the bronchoscope, after a topical dose of 1 ml of preservative-free normal saline, and after a topical dose of 1 ml of preservative-free 2% lidocaine. We observed and video-recorded the larynx for 2 min at each of these points.

Data Scoring and Analysis

In order to evaluate the effects of topical lidocaine on the function of the larynx during bronchoscopy, we re-recorded, in random order, key segments of the videotaped procedures, masking the patient's name and date of the procedure on each new recording. For each patient examined, the first key segment, referred to hereafter as S1, consisted of the video-recording of a period before direct application of lidocaine to the larynx, with the bronchoscope in position above the larynx and vocal cords. The second key segment, referred to hereafter as S2, consisted of a video-recording of a period after the last application of lidocaine to the larynx and before passage of the bronchoscope between the vocal cords. Re-recorded segments included only times during which the patient was calm. The author himself (P.K.) created and edited the re-recording and avoided features that subsequently would allow anyone to distinguish between S1 and S2. The masked re-recordings of these key segments varied in duration from 30 s to 2 min. We made similar masked and randomly ordered recordings from the videotapes of those patients to whose larynx we applied both saline and lidocaine.

In order to quantify the degree of malacia seen in each video segment, we empirically scored laryngeal function, with separate scores for function of the arytenoids and the epiglottis (Table 2). One author (D.N.), who had no knowledge of the order of the re-recorded and masked key segments, scored each segment for function of the arytenoids (AS) and epiglottis (ES). The laryngomalacia score (LS) consisted of the sum of AS and ES for each recording. To quantify possible changes, we subtracted the scores for the recording of the S1 segments from the paired recording of the S2 segments, referring to each of these as AS, ES, and LS. A positive score indicates increased floppiness after topical lidocaine, and a negative score indicates decreased floppiness after topical lidocaine. On the basis of published descriptions and our own experience, we chose a LS equal to or greater than 4, as consistent with laryngomalacia.

We evaluated the effect of age on baseline scores, and compared the frequency with which each score increased and decreased. We examined the relation of the presenting complaints (Table 1) with laryngomalacia. When appropriate, we applied McNemar's chi-square analysis or logistic regression. In the saline control group, we compared the effects of saline and lidocaine on laryngomalacia scores with a paired t test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Of the 280 bronchoscopic procedures recorded, we excluded 98 from this study because of one or more of the following: (1) the key segment recording was not long enough to score, (2) the quality of the recording was too poor to allow scoring, or (3) the original recording could not be found. If one key segment, either S1 or S2, could not be scored, the paired segments were eliminated from consideration. We excluded eight procedures from analysis because the patient's records were unavailable.

Of the 156 cases (312 key segments) successfully scored, 97 were boys and 59 were girls of similar age. Of the 156 patients, 104 (66.7%) were 100 wk of age or younger. Of the 131 patients with a prelidocaine score < 4, 62.7% (52 of 83) were less than 100 wk of age, but of those patients with a prelidocaine score  4, 76% (19 of 25) were < 100 wk of age (odds ratio, 1.83 by logistic regression). Thus, younger patients (< 100 wk) were more likely to have clear signs of laryngomalacia (LS  4) before topical lidocaine, a result that is reflected clearly in the plot of baseline LS versus age (Figure 1).

View larger version (11K): [in this window] [in a new window]   Figure 1.   Initial laryngomalacia score (LS) as a function of age. LS is equal to the sum of AS and ES (see Table 2) (n = 156 patients).

The change in laryngomalacia score after topical lidocaine, LS, ranged from 3 to +7 (Figure 2), but by McNemar's chi-square test, application of topical lidocaine to the larynx was much more likely to result in an increase, as opposed to a decrease, in ES, AS, and LS (p < 0.0001). In 25 cases, LS scores were higher during S1 than during S2 (19 of these by 1, 6 by 2 or 3). In 76 cases, scores were lower during S1 than during S2 (31 of these by 1, 17 by 2, 9 by 3, 19 by 4 to 7). There were 26 patients who had both a prelidocaine LS < 4 and a postlidocaine LS  4. The epiglottis score decreased on 21 occasions (5 by more than 1) and increased 61 times (20 by more than 1). The arytenoid score decreased 14 times (none by more than 1) and increased 58 times (27 by more than 1). There was a tendency for the AS to increase more than the ES, with mean increases of 0.6 and 0.4, respectively. Sixteen patients had an increase in LS of 4 or more points. In eight of these patients ES increased less than AS. In two patients ES increased more than AS. In six, ES and AS changed equally.

View larger version (12K): [in this window] [in a new window]   Figure 2.   The change in laryngomalacia score (LS) as a function of age. LS equals the initial LS minus the LS after application of 2% lidocaine solution to the larynx.

Not only were younger children more likely to have laryngomalacia at baseline (LS  4), but they were also more likely to experience an increase in LS after topical lidocaine (Figures 1 and 2). In the 104 infants younger than 100 wk of age, LS increased by 3 or more in 20.2% and by 4 or more in 13.7%. The LS increased more often (58 of 104) than it decreased (16 of 104) after topical anesthesia. In the 52 children older than 100 wk of age, LS increased by 3 or more in 9.6% and by 4 or more in only 5.6%. The LS increased more often (17 of 52) than it decreased (8 of 52). This difference did not quite achieve statistical significance.

It would have been helpful to compare results of airway examination on two different occasions in the same children, but there were not enough repeat procedures to examine intrasubject variation. Of the three patients who underwent bronchoscopy twice, one had no sign of laryngomalacia. The other two children had their second exams 1 and 2 yr after the first. In these two children the LS decreased from 1 and 3 to zero between the first and second exams, as would be expected with maturation.

The most common indications for FFB were noisy breathing, wheezing, cough, and stridor (Table 1). Croup, stridor, and wheezing were found to be statistically significant positive predictors for laryngomalacia when laryngomalacia was defined as a baseline LS  4 (odds ratio, 5.3 and 4.1, respectively, by logistic regression). Noisy breathing was also a positive predictor but not statistically significant.

To compare the response of the patients to saline and to lidocaine, we selected a group of 10 patients who had a history and presenting signs typical of laryngomalacia (Table 3). Eight were 5 mo of age or younger, which made a change in LS with lidocaine more likely (Figure 2). One was 66 mo and one was 12 mo of age at the time of the exam. The 66-mo-old child was severely neurologically impaired because of Shaken Baby Syndrome. The 12 mo old presented with recurrent stridor and, on baseline exam, had mild laryngomalacia. There was no difference between scores before and after normal saline was sprayed on the vocal cords in these 10 patients. All 10 patients had a positive LS after lidocaine, although one score changed only by one. The increases in AS, ES, and LS after lidocaine all changed significantly as compared with the baseline scores (Table 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our study has clearly demonstrated that topical lidocaine exaggerates the findings associated with laryngomalacia in some patients (Figure 2 and Table 3). After applying lidocaine, the laryngomalacia score and its components changed in a positive direction (increasing severity) much more often than they changed in a negative direction. There was no similar impact of topical saline on the LS, and therefore, it is very likely that the lidocaine effect resulted from its anesthetic properties and not from some nonspecific effect. In addition, the positive changes were greater in magnitude than the negative changes, even though the definition of the LS and its components limited the magnitude of the changes possible. Clearly, there is no reason to believe the LS had a linear correlation with intrathoracic pressure changes. In fact, the score could not change significantly in patients with a LS of 7 or 8 at baseline, but we had the clinical impression that retractions, stridor, and associated airway obstruction increased after lidocaine in these patients. Thus, it is our impression that the LS underestimated changes in airway obstruction near its upper limit.

The lidocaine effect was not an artifact related to time after intravenous sedation. The patients all had adequate time after sedation to reach a steady state before lidocaine was applied. Further, the effect rapidly followed the application of topical lidocaine, and no such effect was seen after saline. Finally, the predictability of the lidocaine effect speaks for its authenticity. It occurs in young infants, the neurologically impaired child, and children with a history of stridor and a physical exam consistent with laryngomalacia. Using these criteria to select subjects for the saline control portion of the study, only one of the ten selected had a small LS of 1. Seven had a change of 3 to 6. None had a negative change.

The two likely causes of laryngomalacia are abnormally compliant cartilage in the upper airway and weakness of muscles supporting the larynx (1, 23). There is an association of laryngomalacia with mental retardation and congenital anomalies (24). Central nervous system insults from drugs or alcohol, seizures, strokes, or hypoxic brain injury can result in acquired laryngomalacia (12, 23). From these reports, the role of abnormal cartilage in laryngomalacia is difficult to prove, but a cause-effect relation between sensorimotor dysfunction and acquired laryngomalacia seems very likely. Topical lidocaine certainly affects sensory nerve function in the upper airway. We do not know from this study whether it has any direct effect on muscle tone. Either sensory or motor effects might have produced the effects we observed. It is not likely that lidocaine had any effect on the integrity of cartilage in the upper airway.

Lidocaine appears to affect the function of the arytenoids more than that of the epiglottis. As noted in the group of 10 (Table 3) and in the larger group, there was a tendency for the AS to increase more than the ES after lidocaine. The arytenoids might depend more on neuromuscular tone for normal function than does the epiglottis. However, the strength of this finding is limited by the arbitrary nature of the definitions of each component of the LS.

Using lidocaine gel to anesthetize the nasal passages seemed to confine its effects and avoid contamination of the larynx by lidocaine in the early phase of bronchoscopy. Although we do not have direct measures of lidocaine concentration in upper airway fluids, we did not observe any increase in stridor or signs of respiratory distress after application of the lidocaine gel. This contrasted with the increase in stridor and signs of increased airway obstruction that rapidly followed application of lidocaine solution to the upper airway through the bronchoscope.

In order to determine positive predictors of laryngomalacia, we associated a LS  4 with laryngomalacia (see figures for scoring system). Although this assignment is arbitrary, we believe it conforms to previously published descriptions (1). We did attempt a more mathematically based definition, but that attempt depended a great deal on obtaining precisely equivalent views on each recording. This rigorous requirement would have eliminated most of our data and probably would have added little to our understanding. The approach we chose had the advantages of simplicity and allowed a reasonable analysis of the question at hand. The blinding procedure we used probably increased scoring errors because the prelidocaine and postlidocaine segments were taken out of context. The changes in airway function we observed were much easier to determine by watching the original recordings in their entirety, but we scored the masked, random segments to avoid bias.

In other reports, stridor is associated with laryngomalacia. In our study a history of recurrent croup, stridor and wheezing on examination were all positive predictors of laryngomalacia. Noisy breathing was also found to be a positive predictor (odds ratio, 1.7) but, because of the sample size, not statistically significant. From these results, it appears that the breath sounds associated with laryngomalacia are not always typical, and they are not always easily distinguished from abnormal breath sounds generally associated with small and large airways disease and obstruction. Thus, this study adds to the growing evidence of the diagnostic value of flexible bronchoscopy in children.

Topical anesthesia is not the only factor that affects observations of the upper airway during bronchoscopy. We have examined the upper airways of several infants with a history of stridor who had obvious signs of laryngomalacia when examined during conscious sedation but whose airway appeared normal during an unsedated exam. A similar observation has been reported previously (27). Certainly the level of agitation during examination of the upper airway affects endoscopic findings.

Consequently, evaluation of the upper airway, especially in infants, requires great care and attention to detail. Any factor that exaggerates the signs of laryngomalacia will lead to a false impression and, perhaps, to overly aggressive treatment. Evaluation of any infant's airways should take into account all of the factors that might affect function during endoscopy, including the effects of topical anesthesia. Ignoring for the moment factors other than topical anesthesia, the child who shows no signs of malacia before and after topical anesthesia clearly does not have laryngomalacia. The child whose signs of malacia appear only after topical anesthesia probably has a mild form of laryngomalacia. In any case, the findings during flexible fiberoptic bronchoscopy must be considered within the clinical context. A child who exhibits some signs of malacia after topical anesthesia, but who has no history of stridor or other abnormal breath sounds, probably does not have significant laryngomalacia.

Applying topical anesthesia to the nasal airway during flexible fiberoptic bronchoscopy is an essentially universal practice. Because topical lidocaine alters laryngeal function in a significant portion of infants and children, we recommend the use of lidocaine gel in the nasal passages for patient comfort and direct observation of the larynx and contiguous structures before applying additional topical anesthetic.