INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Noisy breathing in infants and young children is a common clinical sign (1, 2), invariably indicating some degree of airway obstruction (3). The most common findings are: (1) stridor, a harsh, crowing noise usually resulting from turbulent airflow in a major airway; and (2) wheezing, a high-pitched, whistling sound often interpreted as indicative of peripheral airway obstruction. However, a clinical distinction between upper airway obstruction and diffuse bronchoconstriction is not always obvious (4). Often, the signs of laryngotracheal obstruction may mimic diffuse peripheral airway obstruction, leading to inappropriate diagnostic and therapeutic management. Moreover, multiple airway obstructive conditions may coexist (5, 6), especially in patients who have experienced prolonged endotracheal intubation (4) and those with laryngomalacia (3).

For these reasons, it is widely recommended that all cases of persistent noisy breathing be thoroughly investigated endoscopically (7), at least in the absence of an obvious or very likely diagnosis, since such breathing may indicate a variety of conditions ranging from benign, self-limited diseases to rapidly progressive airway obstruction (10). Indeed, proper management is possible only after a precise diagnosis has been established, and not on the basis of a presumptive or inferential diagnosis (7). On the other hand, an undifferentiated diagnostic approach may lead to invasive procedures in a large group of infants with only mild and self-limiting disease, with possibly increased morbidity and financial expenditure. In this regard, isolated laryngomalacia is by far the most common cause of stridor in infancy (5, 11, 12).

In adulthood, a valuable tool in diagnosing and quantifying upper airway obstruction is analysis of the forced (13) and tidal (17) flow-volume (F-V) relationships. Some reports suggest a similarly useful role of F-V loops at tidal breathing in infants and young children: however, little experience has been reported with these noncooperative patients (18, 19). Therefore, in the early pediatric age group, persistent noisy breathing is still considered an indication for direct visual examination of the airway (7, 8).

In the 6-yr prospective study reported here, we used analysis of the F-V loop at tidal breathing (TB-FV) as a first diagnostic approach to infants and young children with persistent or recurrent noisy breathing (stridor and/or wheezing), evaluating the potential role of this method in addressing these patients' subsequent diagnostic management.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Subjects participating in the study were recruited from among infants and young children aged less than 48 mo, who were admitted to the Department of Pediatrics of the University of Padova between January 1991 and December 1996 with persistent noisy breathing.

Selection Criteria

We enrolled in the study all consecutive patients with stridor or other noisy breathing with a duration of at least 15 d, as well as patients with protracted or frequently relapsing wheezing unresponsive to therapy with ₂-agonists and steroids. Patients with previous endoscopic evaluations were excluded from the study. Nasal and pharyngeal obstruction were ruled out by a preliminary otorhinolaryngologic evaluation. Infants with nasal stenosis or adenoid hyperplasia were excluded from the study. After a general pediatric assessment, the subjects underwent a TB-FV loop study as a first diagnostic step. From January 1991 to December 1993, flexible fiberoptic laryngotracheobronchoscopy (FFB) was systematically performed to establish the definitive diagnosis and was also used to define the accuracy of the functional findings for each patient. The physicians who conducted the TB-FV studies were blinded to the bronchoscopic findings, and those conducting bronchoscopy were blinded to the TB-FV findings. After January 1994, the endoscopic evaluation was done only if a recommended indication was established on the basis of the F-V study (see the subsequent discussion). A prolonged (at least 1 yr) close clinical follow-up was provided for all infants without an endoscopically confirmed diagnosis. When requested, successive airway function examinations were performed to follow the clinical course of a patient's disease. Moreover, TB-FV loop analysis was used whenever possible to evaluate the effectiveness of pharmacologic and/or surgical treatment.

Fifteen infants and young children (age range: 1 to 24 mo) without a past history or present evidence of respiratory disease were studied with TB-FV loop analysis during spontaneous sleep and were used as controls.

The study was approved by our institutional review board. Informed consent to the study was given by the parents of each subject.

TB-FV Loop Analysis

The airway functional study done on the subjects was performed during quiet sleep, as evaluated through classic behavioral criteria (20). Chloral hydrate (50 to 70 mg/kg, orally) was initially used in all subjects to induce sleep. In the last 3 yr of the study, however, in accordance with the wide experience reported in the literature (21), we attempted to study all subjects during spontaneous sleep and used sedation only when measurements during natural sleep failed. Functional testing was not done if there was evidence of acute upper or lower airway infection. All tests were done with subjects in the supine position and with their heads supported in the midline. Special care was used to avoid neck flexion or hyperextension. Whenever possible, TB-FV loops were recorded during noise production. Sa_O2 and heart rates (HR) were monitored continuously during the study with a pulse oximeter (N-180; Nellcor, Inc., Hayward, CA). One of the parents of each subject was present during the study.

TB-FV loops were recorded with a computerized infant pulmonary function device (Model 2600; SensorMedics, Yorba Linda, CA), allowing immediate graphic visualization of F-V loops. An appropriately sized transparent face mask that could be closely fitted with an air-inflated cuff was connected to a pneumotachograph (Lilly type; linear to 10 or 30 L/min, according to infant size) (Hans Rudolph, Kansas City, MO). Airflow was measured directly at the airway opening, and the signal was integrated to yield volume. Before each measurement, the pneumotachograph was calibrated with a precision air syringe, using digital integration techniques (256 samples/s) that enabled both flow and volume to be calibrated by passing known volumes through the pneumotachograph. Linearity was also checked (128-point linearization table).

A morphologic evaluation of the shape of the TB-FV loop was done with the purpose of identifying specific patterns of airway obstruction, and to establish the acceptability of the loop. VT and respiratory rate (RR) were measured and were expressed as means for the loops that were considered. The peak tidal expiratory flow-to-peak tidal inspiratory flow ratio (PEFT/PIFT), midtidal expiratory flow-to-midtidal inspiratory flow ratio (MEFT/MIFT), ratio of the time to peak tidal expiratory flow to total expiratory time (PEFT_t/TE), and expiratory time (TE) and inspiratory time (TI) and the ratio of these latter two measures (TE/TI) were also recorded. Recording periods of at least 30 s, as suggested in the literature (22), were considered adequate to determine the pattern of an average tidal breath. The stability of the morphologic pattern of the TB-FV loop was assessed during noise production by the patients. At least 12 TB-FV loops were analyzed at each testing session, with their morphologic evaluation done independently by two of us (M.F. and E.B.). Acceptability criteria also included an inspiratory-versus-expiratory volume difference of less than 10%.

Bronchoscopies were performed in the procedure suite of our Institute. Subjects were sedated with midazolam followed by meperidine. Since 1995, propofol has been utilized as a single drug for bronchoscopic sedation. Topical anesthesia was obtained with lidocaine solution instilled into the nose and through the bronchoscope. Continuous HR, respiratory rate, and pulse oximetric monitoring were used, and blood pressure was measured every 5 min. Bronchoscopy was performed via the nasal route, using an Olympus BF3C20 pediatric bronchoscope with a 3.5 mm O.D. and a built-in 1.2 mm working channel (Olympus Optical Co., Hamburg, Germany). Airway findings were videotaped and structural airway abnormalities were noted.

Other diagnostic procedures (e.g., X-ray films of the chest or upper airways, esophagography, two-dimensional echocardiography or cardiac catheterization, computed tomographic scanning, magnetic resonance imaging) were performed as suggested by the clinical findings, to confirm functional diagnoses or define the nature of anatomic lesions.

Statistical analysis was done through Kruskal-Wallis one-way analysis of variance on ranks, and with a multiple comparison procedure (Dunn's method) for the evaluation of between-group differences in respiratory parameters. A value of p < 0.05 was considered statistically significant. Data are expressed as mean ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During the 6-yr period of the study, 113 consecutive infants and young children (71 males and 42 females) presented to our institution with noisy breathing and clinical features matching our selection criteria. All of them were enrolled in the study and underwent TB-FV loop analysis as the first step of the diagnostic protocol used to identify their illnesses.

In the first 3 yr of the study, sleep was induced by giving chloral hydrate to all patients before pulmonary function testing was done. Thereafter, TB-FV loops were recorded during spontaneous sleep whenever possible (43 cases), and sedation was used only in subjects in whom the functional study failed during natural sleep. A total of 70 of 113 patients were studied during chloral hydrate-induced sleep. All infants in the control group were studied during spontaneous sleep. Technically acceptable TB-FV loops were obtained from all of the participants in the study. Measurements were easily made on all subjects; no complications were observed during functional testing. FFB was performed without major complications in 95 patients. Minor complications included laryngo- or bronchospasm (three cases) after the procedure, and transient fever (two cases).

The anthropometric characteristics of the patient and control groups are reported in Table 1.

Inspiratory and/or expiratory stridor was the main clinical feature in 72 infants. The remaining 41 subjects presented with long-term or relapsing wheezing. Some subjects, however, presented with unusual clinical noises or associated stridor and wheezing.

A normally shaped TB-FV loop, with both the inspiratory and expiratory limbs showing a roughly round conformation (Figure 1A), was found in all 15 subjects of the control group and in three wheezing infants. Three abnormal morphologic patterns were identified in TB-FV loops, as follows: Pattern 1 (Figure 1B), characterized by irregular, deep fluctuations of the inspiratory flow rate (inspiratory fluttering) and a normal shape of the expiratory limb. Pattern 2 (Figure 1C), consisting of a normal or variably flattened shape of the inspiratory limb and flattening of the expiratory limb of the loop because of a uniform flow rate limitation. Pattern 3 (Figure 1E), characterized by a normal inspiratory limb and an upper concave-shaped expiratory limb of the loop caused by an early expiratory peak flow followed by reduced flow rates at lower volumes. In some cases, however, the expiratory changes seen in Pattern 2 and Pattern 3 were associated with inspiratory fluttering, as shown in Figures 1D and 1F.

View larger version (14K): [in this window] [in a new window]   Figure 1.   TB-FV loop patterns. (A) Normal pattern: round-shaped inspiratory and expiratory limbs. (B) Pattern 1: irregular fluctuations of the inspiratory flow rate (inspiratory fluttering), with normal expiratory shape (associated with laryngomalacia). (C ) Pattern 2: flattening of the expiratory limb of the loop with a normal (or variably flattened) inspiratory shape (associated with an airway obstruction between the gottis and mainstem bronchi). (D) Association between expiratory flattening (Pattern 2) and inspiratory fluttering (Pattern 1) in a child with laryngomalacia and primary tracheomalacia. (E ) Pattern 3: early expiratory peak flow followed by reduced flow rates at lower volumes, with normal inspiration (associated with reactive airway disease). (F ) Association between peripheral airflow limitation (Pattern 3) and inspiratory fluttering (Pattern 1) in a child with laryngomalacia and asthma.

Pattern 1 was found in 44 subjects. In the first 3 yr of the study, this pattern was always associated with endoscopic evidence of isolated laryngomalacia (26 subjects). Therefore, in the remaining 3 yr of the study, endoscopy was withheld from patients with TB-FV Pattern 1. Eighteen further patients with congenital or long-term stridor had a diagnosis of laryngomalacia without endoscopic confirmation (Table 2). All of these subjects had a course consistent with the diagnosis of isolated laryngomalacia in a prolonged (at least 1 yr) clinical follow-up.

Initially, a confirmatory TB-FV loop study was included in the follow-up program of infants with Pattern 1 without endoscopy, and was done on eight of them, always confirming the inspiratory change; thereafter, repeating the functional study was considered unnecessary, and it was not repeated in the remaining patients.

It should also be noted that as mentioned earlier, inspiratory fluttering was found in 10 further subjects, in whom this morphologic pattern coexisted with a change in the expiratory limb of the TB-FV loop (Figures 1D and 1F); this pattern was classified as Pattern 2 (seven cases) or Pattern 3 (three cases). All of these infants underwent FFB, which confirmed an association between laryngomalacia and some secondary airway obstruction (Table 2); as a result, the overall correspondence between inspiratory fluttering in the TB-FV loop and endoscopic evidence of laryngomalacia was verified in 36 infants.

All 46 subjects with Pattern 2 (Figure 1C) were diagnosed with FFB as having an airway obstruction between the glottis and the mainstem bronchi (Table 2). The obstruction was localized to the subglottic region in 21 patients and to the trachea and mainstem bronchi in the remaining 25 patients (Table 2). Thus, Pattern 2 invariably indicated an airway obstruction distal to the glottis. However, an endoscopic diagnosis of primary tracheomalacia was also made on three wheezing infants whose TB-FV loops had a normal shape. In these false-negative tests, we were unable to record TB-FV loops during noise production. The overall sensitivity of TB-FV loop analysis for airway obstruction localized between the glottis and mainstem bronchi was therefore 94%. As discussed earlier, seven of the 46 infants with expiratory flattening in their TB-FV loops showed coexistent inspiratory fluttering (Figure 1D); all were diagnosed as having associated laryngomalacia and subglottic or tracheobronchial obstruction, as confirmed endoscopically.

Twenty patients had TB-FV loops with the changes characteristic of Pattern 3 (Figure 1E); this pattern was associated with the clinical diagnosis of wheezing illness of early childhood (Table 2). None of the infants in this group showed evidence of anatomic airway obstruction. In the three infants with associated inspiratory fluttering (Figure 1F), FFB confirmed the coexistence of laryngomalacia.

Numerical data obtained by recording TB-FV and flow-versus-time parameters are reported in Table 3. Values for Pattern 1 did not show differences from those of the normal pattern. A significant reduction of PEFT/PIFT in infants with Pattern 2 (p < 0.01) and of PEFT_t/TE in infants with Pattern 3 (p < 0.01) were the more specific differences between groups, whereas MEFT/MIFT and TI/TE were significantly reduced both in infants with Pattern 2 and those with Pattern 3 (p < 0.01). Nonetheless, none of the indexes considered showed clinical value or diagnostic utility because of the large intrasubject and intersubject variability.

Serial TB-FV Loop Evaluations

TB-FV loop analysis was done longitudinally in 12 patients with Pattern 2 (and an endoscopically established diagnosis) to evaluate the effects of pharmacologic or surgical interventions. Five cases of subglottic hemangioma were studied before and after systemic or local (direct infiltration) corticosteroid treatment, and one case of postintubation tracheal stenosis was evaluated longitudinally after balloon dilatation. Six cases of secondary tracheobronchomalacia were studied, again after surgical relief of the extrinsic compression (vascular malformation, mediastinal lymphoma, esophageal cyst). In all of these cases, treatment was followed by a marked morphologic improvement in the expiratory limb of the TB-FV loop owing to the posttreatment increase in expiratory flow rates (Figure 2). A similar finding was also made in two cases (tracheomalacia secondary to mediastinal lymphoma and subglottic hemangioma) in which TB-FV loops were recorded just before and shortly after the placement of an endotracheal tube.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Serial analysis of the TB-FV loop in a 3-mo-old infant with a double aortic arch. This infant presented with recurrent wheezing unresponsive to bronchodilator therapy. Left: TB-FV loop appearance at admission. Note the uniform, marked expiratory flattening associated with slight inspiratory flow limitation. Right: TB-FV loop 1 mo after surgical relief of tracheal compression. Note the large increase in the expiratory flow-rate, with morphologic normalization of the loop.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study suggest useful clinical applicability of TB-FV loop analysis in the diagnostic management of infants and young children with clinical evidence of airway obstruction.

Through a period of 6 yr, we studied 113 infants and young children with signs and symptoms of airway obstruction (congenital or long-term stridor and/or persistent-relapsing wheezing) through analysis of the tidal F-V relationship as an initial screening method, in order to provide a functional localization of these infants' airway narrowing. A complete endoscopic evaluation was performed in 95 patients to obtain a conclusive diagnosis and was also used to evaluate the appropriateness of the functional study (pneumotachography).

The graphic representation of the TB-FV relationship provided three easily recognizable patterns, suggesting as many sites of airway obstruction. On the basis of these simple morphologic patterns, infants with noisy breathing were distributed into three groups, showing Pattern 1 (Figure 1B, inspiratory fluttering), consistent with the diagnosis of laryngomalacia; Pattern 2 (Figure 1C, expiratory flattening), which identified an airway obstruction localized from the subglottic region to the mainstem bronchi; and Pattern 3 (Figure 1E, upper concave-shaped expiratory limb), indicating diffuse intrathoracic airways narrowing such as that seen in children with asthma. The described changes appear especially characteristic when compared with the normal TB-FV loop pattern (Figure 1A), obtained in a group of 15 infants without respiratory disease. In this case, an overall round shape indicated the sinusoidal respiratory drive model observed during tidal breathing in the absence of airflow limitation, as described by other authors (18, 21).

TB-FV loop analysis could be used in young children with noisy breathing to assess the localization of the causative airway obstruction, adding useful information to the initial physical examination. Currently, an endoscopic approach is recommended in all such children (3, 7, 8, 9) because of the difficulties in establishing a specific diagnosis. For this reason, bronchoscopy remains by far the most important diagnostic tool in the cases of suspected anatomic airway obstruction. However, TB-FV loop analysis could help reduce the use of endoscopy in some infants with isolated laryngomalacia.

In our study, laryngomalacia was invariably associated with TB-FV loop Pattern 1 in 36 subjects (26 with isolated laryngomalacia, 10 with multiple airway obstruction), as confirmed by endoscopic inspection. The accuracy of the method allowed us to avoid further diagnostic procedures in another 18 infants with Pattern 1, whose clinical course over at least 1 yr was consistent with the functional diagnosis of isolated laryngomalacia. Moreover, it should be emphasized that TB-FV loop analysis was also accurate in addressing the clinical management of infants in whom laryngomalacia was associated with more distal airway obstruction. This situation may occur frequently according to several previous reports (3). In our experience, specific changes in the expiratory limb of the TB-FV loop indicated the presence of a more peripheral airflow limitation in 10 infants with functional evidence of laryngomalacia. In all of these cases, an indication was established for endoscopy, which revealed a coexistent obstructive lesion in the airway distal to the glottis. Figures 1D and 1F show the typical pattern of laryngomalacia in two infants in whom the associated abnormal expiratory limb warranted further assessment. The conclusive diagnoses were primary tracheomalacia (Figure 1D) and early childhood asthma (Figure 1F), respectively, concurrent with laryngomalacia. In consideration of this issue, the diagnosis of isolated laryngomalacia established by direct or indirect laryngoscopy could be considered unsatisfactory, since this procedure does not yield information about the condition of the distal airway.

Although TB-FV loop Pattern 1 was always associated with isolated laryngomalacia in our study, it should be considered that other, less common conditions (e.g., vocal cord papillomatosis) could produce the same pattern of inspiratory fluttering. For this reason, it is important to remember that the best interpretation of TB-FV loop findings (as for other respiratory function tests) always requires careful analysis of the patient's clinical history and a physical examination. However, considering that many of these more rare conditions are very uncommon, it seems reasonable to think that TB-FV loop analysis remains a very specific method for identifying isolated laryngomalacia.

The expiratory flattening of Pattern 2 (Figure 1C) was detected in 46 subjects in our study and was invariably associated with a fixed or dynamic obstruction localized between the glottis and the mainstem bronchi. For this reason, patients with TB-FV loops showing Pattern 2 always need endoscopy, since TB-FV loop analysis has no role in establishing a definite anatomic diagnosis. Twenty-one infants presented with inspiratory and/or expiratory stridor and had confirmed evidence of subglottic obstruction (mainly subglottic hemangioma and postintubation stenosis; Table 2). Most patients with TB-FV loop Pattern 2 (25 cases) presented with chronic or frequently relapsing wheezing, or with coexistent stridor and wheezing. The clinical approach to this kind of patient presents particular difficulties, and it has been reported that these infants often receive a presumptive clinical diagnosis of early childhood asthma, leading to prolonged, inappropriate treatment with steroids and ₂-agonists (23, 24). In a recent study intended to define the natural history of primary tracheobronchomalacia, Finder (23) reported that all of his patients were thought by their primary care physicians to have reactive airway disease, and in no patient was the diagnosis of bronchomalacia considered before referral to his center. Finder emphasizes the importance of including primary tracheobronchomalacia in the differential diagnosis of all persistently wheezing infants, because of its high incidence (23). Similar conclusions were reached by Schellhase and colleagues (24), who found structural airway abnormalities in 57% of a group of young children with recurrent wheezing. We agree with these investigators that this condition should no longer be considered an uncommon finding in these patients. Moreover, the fact that tracheobronchomalacia can also be due to extrinsic compression of the airway should raise consideration of the possible presence of underlying severe diseases or malformations, as we observed in six of 21 cases of tracheobronchomalacia (Table 2).

In our experience, TB-FV loop analysis represents a rapid, noninvasive, and accurate screening method for the diagnostic management of wheezing infants, since peripheral bronchoconstriction and central airway obstruction yield quite different morphologic changes in the expiratory limb of the TB-FV loop (Pattern 3 in Figure 1E, and Pattern 2 in Figure 1C, respectively). The upper concave change in the expiratory limb of the TB-FV loop (Pattern 3) was found in 20 of our 45 wheezing infants, none of whom showed anatomic airway obstruction with FFB.

It should be noted that we failed to recognize well-defined alterations in the morphology of the TB-FV loop in three subjects (2.6% of our patients), in whom successive ascertainments revealed primary tracheomalacia. This is in agreement with the study by Finder (23), in which some patients with primary tracheomalacia did not have an abnormal F-V loop pattern. However, we suggest that this may be due to inadequate study conditions. It is particularly important to protract the duration of F-V loop recording in subjects with intermittent noisy breathing. This is especially important in laryngomalacia and primary tracheomalacia, since infants with these conditions may have prolonged periods of silent breathing (and a normally shaped F-V loop) while asleep, because of the lower driving pressure during sleep. Moreover, although changes in the F-V loop may also be detected during silent breathing in some patients with airway obstruction, this is not always the case. Airway function studies of such patients should be considered unsatisfactory unless measurements are made during noise production. However, a complete endoscopic evaluation is warranted in all infants with an inconclusive functional study.

Useful information may also be derived by comparing serial evaluations done on the same patient. Figure 2 shows the different morphologic patterns of the F-V loop before and after surgical treatment in a case of tracheal compression by a double aortic arch. The morphologic improvement is clearly due to the higher expiratory flow rate following relief of the tracheal compression. Nonetheless, it should be noted that no definite correlation may be established between airflow improvement and airway diameter as evaluated endoscopically.

Additionally, although the F-V and flow-time relationships may yield a large number of parameters, no evidence of their practical utility has been found. Clearly, in terms of normal control values, some parameters show changes consistent with morphology of the TB-FV loop (Table 3), as is the case, for example, with the significant reduction in MEFT/MIFT in Patterns 2 and 3. Moreover, when considered as a group, infants with Pattern 2 showed a significant reduction in PEFT/ PIFT as compared with all other groups in our study, and a similar finding was made for PEFT_t/TE in subjects with Pattern 3. However, because of wide intersubject and intrasubject variability, substantial overlapping of these values occurred among patients with different conditions. For this reason, TB- FV analysis based only on the interpretation of numerical parameters may be misleading. Unfortunately, the lack of objective parameters for describing the shape of the TB-FV loop makes it difficult to assess the reproducibility of morphologic changes, which represents a limitation to TB-FV loop analysis.

In conclusion, our study suggests a useful role for TB-FV loop analysis as a first diagnostic approach to infants and young children with noisy breathing. Study of the F-V relationship during tidal breathing can be easily done as an extension of the physical examination, without any substantial disturbance of these young patients. It may provide interesting pathophysiologic information, and may be useful as a screening method in addressing their diagnostic management.