INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: asthma; lung function; methacholine; airway remodeling; bronchiolitis; infantile asthma; bronchial hyperreactivity

Infants often display wheezing with acute lower respiratory tract illness. The relationship between symptoms of wheezing in infancy and asthma remains unclear, and most infants who wheeze stop wheezing after 3 yr of age (1). Many authors have tried to relate illnesses involving wheezing in infants to bronchial hyperresponsiveness, by analogy with the strong association between nonspecific bronchial hyperresponsiveness and asthma in older children. However, the interpretation of these studies is controversial. Some studies have reported that bronchial hyperresponsiveness is present in all young infants, especially in those younger than 1 yr, independently of the presence of illnesses involving wheezing (2). If this is true, then the detection of bronchial hyperresponsiveness is less useful as a marker of infantile asthma. However, if bronchial hyperresponsiveness is inversely proportional to age during childhood (5), it may be possible to identify infants with asthma by the persistence of increased bronchial hyperresponsiveness with age. We therefore carried out a prospective study on the natural history of bronchial hyperresponsiveness in infants who wheeze. We investigated whether infants who wheeze and subsequently develop persistent asthma differed from infants who wheeze and who became asymptomatic either in the initial degree of bronchial responsiveness or in the persistence of bronchial hyperresponsiveness with age. Two reliable indices of airway response are used in infants: maxFRC and transcutaneous oxygen tension (PtcO₂) (6). We previously showed that both indices reproducibly detect bronchial hyperresponsiveness in infants who wheeze (9). PtcO₂ has the advantage of being easy to measure, even for infants who are awake, and therefore can be used for repeated testing without repeated sedation. We therefore used this variable to evaluate airway reactivity in infants who wheeze repeatedly over a 4-yr period.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and Study Design

One hundred and twenty-nine infants were recruited from the patients attending the Paediatric Pneumology Unit at Necker-Enfants Malades Hospital in Paris. Eighty-nine were boys and 40 were girls. All infants had suffered at least three episodes of wheezing. Their mean age was 16 ± 7 mo (range 11 to 24 mo). Forty-six percent of children had their mother and/or mother with asthma, allergic rhinitis, or atopic eczema. At the initial visit (V1), complete pulmonary function tests and methacholine challenge were performed under sedation. During the 4-yr follow-up study, clinical evaluation and methacholine challenge were performed every 6 mo until the child was 4 yr old and once per year thereafter. The response to methacholine inhalation was assessed by determining transcutaneous oxygen tension (PtcO₂), thereby avoiding the necessity of repeated sedation. All infants were asymptomatic for at least 1 wk before the day of the test. If any respiratory symptoms were observed during the week before the planned test, the test was delayed. At each visit, clinical progression was evaluated using a standardized score taking into account any coughing, wheezing, and acute exacerbation during the past 6 mo (Table 1). This study was approved by the local ethical committee and informed consent was obtained from the parents of all children.

Lung Function Tests

For their first test, infants were sedated with chloral hydrate (75 mg kg¹). Maximal partial expiratory flow volume (PEFV) was determined using the squeeze technique by rapidly inflating a thoracoabdominal jacket at the beginning of expiration (Medical Engineering Department, Royal Postgraduate Medical School, Hammersmith Hospital, London), according to a previously described technique (8, 9). The jacket was wrapped around the infant's chest and abdomen with the arms extended outside the jacket. The neck was extended to minimize airway and glottic obstruction. All measurements and calculations were performed using a pediatric mobile measurement module (SensorMedics Corporation 2600, Yorba Linda, CA), containing pressure transducers, electronics modules, and a 14-bit analogue-to-digital signal converter. Flow was measured at the infant's mouth using a face mask attached to a 0-30 LPM triple screen pneumotachograph with a flow resolution of 0.06 ml/s, a volume resolution of 0.12 ml, and a volume range of 0-255 ml. Silicone putty was applied around the mouth and nose to provide an airtight seal with the face mask. Forced expiration was measured as the maximum expiratory flow at functional residual capacity (maxFRC). Partial forced expiratory tests were performed after FRC was found to be stable for at least 10 consecutive regular tidal breaths. The FRC level was defined as the end-expiratory level obtained from the respiratory cycle preceding the forced expiratory test. As we had found with our equipment that the jacket pressure required to obtain maximal flow was usually between 60 and 80 cm H₂O, we used these two pressures in succession. Maximal flow was obtained at 60 cm H₂O for 33% of infants and at 80 cm H₂O for 67% of children. The pressure transmitted from the jacket to the infant was assessed for each infant by an occlusion test and found to be around 50% of the jacket pressure (range: 44-58%). Three PEFV curves were produced for each pressure and a mean baseline value was determined from the three highest of six technically acceptable values obtained. Criteria for an acceptable PEFV curve included a rapid rise in forced expiratory flow so that peak flow occurred before 50% of the tidal volume was expired; a smooth curve with no transients in the region of FRC; and forced expiration post FRC. At V6 and V7, the specific resistance of airways (SRaw) was measured by plethysmography (Sensor Medics, Yorba Linda, CA). Plethysmographic airway resistance was measured during relative panting to a frequency of 40 to 50/min, with a tidal volume of 120 to 150 ml. Airway resistance was measured during the first half of both the inspiratory and the expiratory time. The final value was the mean of at least five technically satisfactory measurements.

Transcutaneous Oxygen Tension

PtcO₂ was measured with a Roche electrode calibrated at room temperature, heated to 45° C and placed on the volar side of the forearm. Transcutaneous oxygen tension was continuously recorded with a chart plotter. Baseline value was determined after PtcO₂ had reached a stable maximum. A 15- to 20-min period was required to obtain a stable value in sleeping infants. A baseline value of less than 75 mm Hg was considered as being likely due to technical factors and resulted in a new preparation of the electrode. The lowest PtcO₂ value after methacholine inhalation was recorded by direct lecture on the monitor.

Methacholine Challenge

The aerosol was administered with a dosimeter (MFDC 88, Mediprom, Paris, France) attached to a nebulizer (De Vilbis 5610 D). The size of the particles generated by the nebulizer was 1.9 µm MMAD. The apparatus was programmed to deliver a dose of 50 µg of methacholine in a volume of 40 ml of air in 0.5 s. Under these conditions, the duration and volume of each aerosol dose did not exceed the inspiratory time and volume of the infants, so each delivered dose was completely inhaled. The dosimeter was triggered by the inspiratory negative buccal pressure produced in the face mask applied to the infant's face. The infants initially inhaled normal saline. Two minutes later, PtcO₂ and lung function were measured. This sequence was repeated after data collection. The initial methacholine dose was 50 µg and the dose was doubled for each sequence until PtcO₂ decreased by at least 15% from baseline (10) or a maximal methacholine dose of 1600 µg was inhaled. The change in PtcO₂ was taken as the maximum deviation from the baseline 2 min after each dose, before any lung function tests were performed.

The provocative concentration for a 15% fall in PtcO₂ (PD₁₅ PtcO₂) was derived from the plot of the log dose of methacholine against PtcO₂, by linear interpolation between the last two points on the semilogarithmic dose-response graph. Infants who had not responded by the maximal dose of methacholine were assigned the value PD₁₅ = 3200 µg (i.e., twice the maximal dose).

Data Analysis

At the end of the 4-yr follow-up period, children were classified into four groups:

• persistent wheezers (n = 22): children with wheezing between V6 and V7 and no remission period (12 mo or more) between V1 and V6;
• intermittent wheezers (n = 20): children with wheezing between V6 and V7 with remission periods of 12 consecutive mo or more between V1 and V6;
• coughers (n = 18): children with coughing but not wheezing between V6 and V7; and
• asymptomatics (n = 52): children without respiratory symptoms between V6 and V7.

All data are expressed as means ± standard error of the mean (SEM). Differences between visits and in changes over time for these four groups were compared by analysis of variance (ANOVA) followed by Fisher's PLSD test. Logistic regression analysis was used to evaluate the influence of family history of atopy on clinical outcome.

The sensitivity and specificity of possible cut-off points for PD₁₅ PtcO₂ measured at V1 or V3 for discriminating between children with persistent wheezing and those who became asymptomatic were determined using receiver operator characteristic (ROC) curves (11). ROC curves plot all possible combinations between the true-positive ratio (sensitivity; y axis) and the false-positive ratio (1  specificity; x axis) as one varies the definition of positivity. Different points were therefore obtained by varying the PD₁₅ PtcO₂ values used as the cut-off for positivity. The cut-off value corresponding to the best compromise between sensitivity and specificity is defined as the point of the curve closest to the upper left-hand corner.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One hundred and twenty-nine infants were included in this prospective study. One hundred and twelve were followed until the final visit (V7) 4 yr later. Mean anthropometric data at each visit are summarized in Table 2.

Clinical Progression

The clinical score significantly improved as early as the second visit (V2; Table 2) and most children had remission periods of 12 mo or more without wheezing (Figure 1). After 2 yr of follow-up (V5), only 31% of children had persistent wheezing illness without remission periods and after 4 yr (V7) only 20% had persistent wheezing without remission. However, only 63% of children who did not experience wheezing between V1 and V3 remained asymptomatic throughout the study.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Clinical progression in 112 infants followed from V1 to V7. Black parts of the bars represent children with wheezing episodes during the past year. White parts of the bars represent children without wheezing episodes during the past year.

Anthropometric data at V1 did not differ significantly among the four groups defined according to clinical status at the end of the 4-yr follow-up period (Table 3). The clinical score decreased over time in all groups (Figure 2), so that even persistent wheezers had lower clinical scores at the end of follow-up than at the start: 5.7 ± 0.7 versus 21.5 ± 1.9 at V1. However, the clinical scores of persistent wheezers were higher than those of asymptomatic children from V1 to V7, but the significance was reached only from V2 to V7. Underpowered analysis at V1 (50%) may contribute to the absence of significance.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Mean clinical score (± SEM) measured at each visit for the four subgroups of children: persistent wheezers (solid line and squares), intermittent wheezers (dashed line and diamonds), coughers (dashed line and triangles), and asymptomatics (solid line and circles). *Significant difference between persistent wheezers and asymptomatic children (p < 0.05).

Familial atopy was associated with a significantly higher risk of being a persistent wheezer (p < 0.02). Between 3 and 4 yr of age, 39 children had skin tests with inhaled allergens. Among these 39 children, 50% of persistent wheezers had at least one positive skin test, defined by a wheal at least 2 mm larger than the negative control, compared with 24% of all other children, and 13% of asymptomatic children. Because of the small number of children tested, the difference was not significant.

During infancy, only two children were treated with inhaled corticosteroids for more than 3 consecutive mo. Thus, treatment was not considered as a potential bias in our study.

Lung Function Measures

Lung function was evaluated under sedation at V1. Persistent wheezers had significantly lower maxFRC values at V1 than infants who became asymptomatic (p < 0.01) (Table 3). They also had higher SRaw values at the end of follow-up than did asymptomatic children (Figure 3).

View larger version (11K): [in this window] [in a new window]   Figure 3.   Mean SRaw values (± SEM) in persistent wheezers (solid bars) and asymptomatic children (open bars) at visits V6 and V7.

Methacholine Challenge

Methacholine challenge was performed at each visit. Eight infants showed no significant fall in PtcO₂ after 1600 µg of methacholine at one test. None of these infants had repeated negative tests.

In this population, mean log PD₁₅ PtcO₂ was 2.41 ± 0.05 at V1 and did not vary significantly from V1 to V7 (Table 2). PD₁₅ PtcO₂ at V1 was not correlated with clinical score. No significant difference among groups with subsequently different clinical progression was observed at V1 (Table 3). However, the changes in PD₁₅ PtcO₂ over time differed significantly among groups (Figure 4). Persistent wheezers had significantly lower PD₁₅ PtcO₂ than asymptomatic children from V3 to V7. Despite these differences, ROC analysis did not identify an acceptable early PD₁₅ PtcO₂ cut-off point predictive for subsequent clinical progression. The cut-off value, corresponding to the best compromise between sensitivity and specificity for discriminating between infants with persistent wheezing and those who became asymptomatic, was defined as the point of the curve closest to the upper left-hand corner. At V3, it corresponded to a PD₁₅ PtcO₂ value  2.3, which identified persistent wheezers with a sensitivity of 69% and a specificity of 59%. Multiple regression analysis performed with familial history of atopy and airway responsiveness level at V3 as independent variables showed that only a familial history of atopy was a significant risk factor for persistent wheezing.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Mean PD15 PtcO2 log value (± SEM) measured at each visit for the four subgroups of children: persistent wheezers (solid line and squares), intermittent wheezers (dashed line and diamonds), coughers (dashed line and triangles), and asymptomatics (solid line and circles). *A significant difference between persistent wheezers and asymptomatic children (p < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recurrent illness involving wheezing is one of the most frequent causes of medical consultation during the first years of life. Only a minority of these infants who wheeze still have asthma at school age (1). Identifying factors predicting clinical progression in infants who wheeze would be of value as it would make it possible to give early appropriate treatment. Because increased airway responsiveness is a common manifestation of asthma in school-age children, we investigated whether infants who wheezed and subsequently developed persistent asthma differed from infants who wheezed and who later became asymptomatic either in the initial degree of bronchial hyperresponsiveness or in the persistence of bronchial hyperresponsiveness with age. The results of this prospective study show that infants who wheeze and subsequently develop persistent asthma had lower PD₁₅ PtcO₂ values but that no early predictive cut-off value could be identified.

All the infants included in our study were recurrent wheezers at their initial evaluation. Four years later, only 38% still were wheezing, with or without periods of remission. This is consistent with previous epidemiological data. Martinez and coworkers reported that 41% of infants who wheeze still were wheezing at 6 yr of age (1). Similarly, Brooke and coworkers found that 53% of infants who wheeze still had persistent wheezing 3 yr later (12). Our study adds that half the infants who wheezed who still wheezed at 5 yr of age were in fact intermittent wheezers, with clinical remission of their wheezing for periods of 12 mo or more. Initial clinical severity (high clinical score) was not predictive of subsequent progression. These results conflict with those of Park and coworkers who found that a high frequency of symptoms during infancy was a significant risk factor for the persistence of asthma at school age (13). Differences between the two studies in the grading symptom severity probably account for most of this difference. We found that lower baseline lung function at initial evaluation, unlike clinical score, was associated with subsequent persistent wheezing. However, we were unable to determine a maxFRC cut-off value that discriminated between children with persistent wheezing and those who became asymptomatic.

It has been demonstrated that lower levels of lung function precede and predict the development of respiratory illnesses involving wheezing during infancy (14, 15). Once recurrent wheezing is established, our study shows that a low maxFRC value reflects a higher risk of persistent wheezing 4 yr later, but the overlap of maxFRC values between subgroups of infants with different clinical progressions renders this variable useless in clinical practice as a predictor of persistent wheezing for a given infant. We found that reduced airway caliber, reflected by high SRaw values, is also associated at school age with persistent wheezing. This finding may be explained by either constitutional or acquired narrow airways.

Martinez and coworkers showed that low maxFRC in healthy newborns is predictive of illness involving wheezing only during infancy and not at school age (1). Once infants become wheezy, our data suggest that the presence of small airways constitutes a higher risk for persistence of their symptoms at school age. However, constitutional narrow airways may imply that reduced caliber airways is constantly found in these infants who wheeze, whereas we previously demonstrated a high individual variability in max values (9). Thus, acquired narrow airways appears as a more likely hypothesis. However, our data do not permit us to argue for a fixed or a reversible acquired reduced airway caliber. Changes in baseline parameters induced by a bronchodilator could not be adequately tested, because of methacholine challenge tests. However, persistent wheezing is likely to interfere with airway growth. Indeed, it has been shown that persistent asthma throughout childhood leads to low FEV₁ values in adulthood (16). Similarly, Palmer and coworkers showed that early airway hyperresponsiveness was associated with wheezing during the first 6 yr of life and with decreased baseline FEV₁ by school age (17). Our data may therefore suggest an underestimated and early airway remodeling in these infants with persistent wheeze. Such a progression would be independent of the initial severity of symptoms because the decrease in clinical score at each visit was similar to that observed in the subgroup of asymptomatic children.

All children underwent repeated methacholine inhalation challenge. We used PtcO₂ to measure the response to methacholine. PD₁₅ PtcO₂ is now widely used as a sensitive index for detecting bronchial responsiveness during methacholine challenge (18), even in infants (6). We have shown that it is as sensitive as PD₃₀ maxFRC in infants who wheeze (9). Furthermore, PtcO₂ is easy to measure, even in infants who are awake, thus facilitating repeated testing without repeated sedation.

At the initial evaluation of our population, illness involving wheezing was associated with a significant response to methacholine inhalation challenge. The degree of bronchial responsiveness was however not correlated with the intensity of symptoms. Above all, the initial level of bronchial responsiveness was not predictive of subsequent persistence of asthma. The relationship between bronchial responsiveness and wheezing in infancy remains unclear. Previous studies have concluded that established recurrent wheezing in infants is independent of bronchial reactivity (3, 21). Furthermore, the airways of normal healthy infants are also able to constrict in response to inhaled methacholine (2), histamine (4), or cold air (22). However, this absence of difference in nonspecific bronchial challenge response between children who wheeze and healthy children may rapidly disappear with age because airway reactivity in healthy infants and young children has been shown to decrease with increasing age (5). Indeed, we found that children with persistent wheezing had a significantly lower mean PD₁₅ PtcO₂ value than asymptomatic children as early as V3, that is, 30 mo of age. This finding is consistent with Palmer and coworkers who found that early airway responsiveness level, at 1 mo of age, was predictive of persistent wheezing by school age (17). However, the ROC curve did not identify a PD₁₅ PtcO₂ cut-off value at V3 that discriminated between children with persistent wheezing and those who became asymptomatic.

The persistence of bronchial hyperresponsiveness with higher PD₁₅ PtcO₂ values in asymptomatic children may indicate a slow progressive return to the values observed in healthy children. The time for which bronchial hyperresponsiveness persists in this asymptomatic group of children is unknown. Kolnaar and coworkers showed that asymptomatic bronchial hyperresponsiveness in adolescents and young adults is not related to the incidence of lower respiratory tract infections in early childhood (23). Together with our data, this suggests that acquired bronchial hyperresponsiveness during early life may initially persist without symptoms before disappearing and is not linked to bronchial hyperresponsiveness measured later in childhood. Similarly, Palmer and coworkers found no association between airway responsiveness at 1 mo and at 6 yr and suggested that factors contributing to airway responsiveness in early life may be different from those contributing to airway responsiveness in later childhood (17). Longer prospective evaluation of our group of children is needed to confirm this hypothesis.