INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway inflammation is now established as a fundamental feature of asthma (1). Typically there is a mucosal infiltrate consisting of eosinophils (4), mast cells (3), and activated lymphocytes (1) that contribute to structural changes of epithelial damage, subepithelial fibrosis, and smooth muscle hypertrophy (1). Airway inflammation occurs in both allergic and nonallergic forms of asthma (3) and is a feature of all grades of asthma severity (4). These observations mainly derive from studies of adults with asthma and form the basis of the modern therapeutic approach to asthma as an inflammatory disease (2). There are, however, several important questions that remain to be answered regarding airway inflammation in asthma. First, is airway inflammation as important in childhood asthma as it is in adult asthma, and second, can the role of airway inflammation that has been established in clinical studies be confirmed by epidemiologic studies? The answers to these questions have important implications for both the treatment of asthma and for understanding its mechanisms. The prevalence of childhood asthma is increasing (5). Although children may have the most to gain from early anti-inflammatory therapy, they are also at greatest risk from its adverse effects on growth and bone demineralization.

In order to address these questions, it is necessary to use a noninvasive marker of airway inflammation that is suitable for use in children and in epidemiologic studies. Recent improvements in methods for collecting and analyzing sputum have made induced sputum cell counts feasible for studies of airway inflammation in children (6). Sputum can be induced by the inhalation of hypertonic saline and processed to yield reproducible results that reflect disease activity and correlate with samples obtained at bronchoscopy (7). In this study, we use induced sputum to examine airway inflammation in children and to identify whether airway inflammation is an important determinant of asthma symptoms and airway hyperresponsiveness (AHR). The research question was: Are there associations among airway inflammation, AHR, and asthma symptoms in children studied in an epidemiologic setting?

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject

A cohort of 263 healthy, full-term infants was recruited between 1979 and 1984 for a longitudinal study of the development of mucosal immunity and the occurrence of allergy and respiratory disease, especially asthma (8, 9). In this study, 245 of these children, then aged between 8 and 14 yr, were identified for ongoing study and invited to participate in a follow-up survey. A postal questionnaire containing validated items relating to respiratory symptoms, asthma diagnosed by a doctor, asthma therapy, and family background of respiratory and allergic diseases (9) was completed, and 170 children (69%) attended for testing (Table 1). The reasons for nonattendance included these: they declined participation (44 children); family moved out of area (15 children); or they were unable to be contacted (16 children). Written informed consent was obtained from parents and children, and the study received approval from the Hunter Area Health Service Research Ethics Committee.

Methods

Demographic details, asthma symptoms, and therapy were recorded at the testing visit. The following sequence of tests was then performed: allergy skin-prick tests, spirometry, a combined hypertonic saline inhalation challenge to assess airway responsiveness and to obtain induced sputum for cytological analysis, and a nasal smear.

Asthma symptom questionnaire. Symptom severity in the past 2 wk was rated using a 7-point Likert scale where 1 = absent, no discomfort at all; 4 = mild discomfort; and 7 = most severe discomfort ever (10). A value greater than 1 was considered to be positive for that symptom. The symptom items (11) included itchy or red eyes, a stuffy or runny nose, cough, wheeze, shortness of breath, tightness in the chest, wheeze or chest tightness on walking, chest symptoms during sleep, and asthma symptoms overall.

Allergy tests. Allergy skin test were performed using the modified prick technique with allergen extracts (Dome/Hollister-Steir; Bayer Pharmaceuticals, Sydney, Australia) for house dust mites (Dermatophagoides pteronyssinus and Dermatophagoides farinae), mold mix (Alternaria, Aspergillus mix, Hormodendrum, Penicillium mix), mixed grasses, cat fur (cat hair and epithelium), and cockroach, together with positive (histamine) and negative (glycerine) controls. A skin-prick test was defined as positive if the wheal diameter was 3 mm or greater at 15 min. A child was considered atopic if a positive skin-prick test was recorded for any allergen.

Hypertonic saline challenge. Children withheld bronchodilators for their duration of action, antihistamines for 48 h, and inhaled corticosteroid for 72 h before testing. After spirometry, children with FEV₁ values more than 80% of the predicted value proceeded to hypertonic saline challenge and sputum induction. Saline (4.5%) was inhaled for doubling time periods (30 s, 1 min, 2 min, 4 min) from a Timeter MP500 ultrasonic nebulizer (Oregon Scientific, Pike, PA) with 23-cm corrugated tubing and a Hans Rudolph 2700 two-way nonrebreathing valve box (Hans Rudolph, Inc., Kansas City, MO) with rubber mouthpiece and nose clips (12, 13). The FEV₁ was measured, in duplicate, 60 s after each saline dose, after which children were asked to expectorate into a sterile container. The test was stopped when either the FEV₁ had fallen by more than 20% or 15.5 cumulative minutes nebulization time had elapsed. If the FEV₁ fell by more than 20% during the challenge, then 200 µg salbutamol was administered using a pressurized metered-dose inhaler and valved holding chamber (Breath-A-Tech; Scott-Dibben, Newcastle, Australia). If a satisfactory sputum sample was not obtained by the time the FEV₁ had fallen by more than 20%, nebulization with 4.5% saline was continued for 4-min periods after the FEV₁ had returned to within 10% of baseline. The dose of 4.5% saline delivered to the mouth was assessed by weighing the nebulizer cup, tubing, and valve-box before and after each challenge.

Sputum analysis. Sputum was processed as described by Pin (6) and Popov (14) and their colleagues. Briefly, the sputum volume and macroscopic characteristics were recorded. Opaque, mucocellular clumps were identified by examination against a black background. At inverted microscopy, these portions were rich in lower respiratory cells, and had minimal squamous epithelial cell content. These portions were selected using forceps, and a 300 µl aliquot was aspirated from the Petri dish using a positive displacement pipette. The aliquot was added to 2,700 µl of dithiothreitol (Sputolysin, Calbiochem, La Jolla, CA), mixed by rotating for 30 min at room temperature, and filtered through 50 µm nylon gauze. A total cell count was performed and cytocentrifuge slides were prepared (Shandon Cytospin II, Sewickey, PA).

The quality of induced sputum samples was assessed based upon a slide quality assessment procedure that evaluated the presence of three parameters: (1) adequate number of cells for enumeration, (2) the presence of pulmonary macrophages on the slide, and (3) the proportion of squamous epithelial cells. Cell number was scored as 0 if there were fewer than 200 cells, 1 if there were 200-399 cells, and 2 if 400 or more cells were present. Pulmonary macrophages were scored as present (2) or absent (1). The proportions of squamous epithelial cells was scored as 2 if less than 20%, and 1 if 20% or greater. This gave a quality score ranging from 0 (poor quality) to 6 (good quality). The slide-based quality assessment technique was compared with previously published methods (6, 15), based upon macroscopic size of sputum mucocellular clumps in 78 consecutive children.

Eosinophil counts were expressed as the percentage of 400 nucleated cells on each of two slides fixed with methanol and stained with Chromotrope 2R. Mast cells were identified in two further slides that were fixed with Carnoy's solution and stained with 0.5% toluidine blue in 0.7 N hydrochloric acid at pH 0.1; 1,500 nucleated cells were counted on each slide. Cells containing the cytokine granulocyte-macrophage colony-stimulating factor (GM-CSF) were identified using a monoclonal antibody to GM-CSF (IgG₁, 10 µg/ml; Genzyme, Cambridge, MA), which was detected using the alkaline phosphatase-antialkaline phosphatase technique (APAAP) in samples from 91 subjects, comprising all the symptomatic children and a random sample of asymptomatic children. Positive control slides (RC2A cells), an isotype control, and substrate controls were included with each staining run.

Nasal smears. Nasal smears (16) were obtained using saline-moistened cotton-tipped swabs placed in the nose, then immediately smeared onto glass slides for staining as described above.

Statistical Analysis

Eosinophil and mast cell counts were used as the outcome variables for the analysis, and their association with the explanatory variables of age, sex, presence of chest symptoms, presence of nasal symptoms, and AHR was described by the odds ratio (OR) with its 95% confidence intervals. Neither eosinophil nor mast cell counts were normally distributed. There was a large proportion of zero counts and positive counts tended to be skewed. Examination of residuals from models based on the normal and Poisson distributions indicated a poor fit; hence, the analysis was conducted using an ordinal regression model with cell counts divided into three categories. The first category contained all cell counts of zero, with the cut-point for the two remaining categories being the median of the nonzero counts. The degree of eosinophilia was categorized as 0%, 0-2.5%, and more than 2.5%. Sputum mast cell counts were categorized as 0%, 0-0.27%, and more than 0.27%. Asthma symptoms were categorized as present or absent. The symptom scores for wheeze, shortness of breath, and chest tightness were summed. A score of three indicated no symptoms, with values greater than 3 indicating that the child had current asthma symptoms. Analysis using additional symptom categories (3, 4-15, > 15) gave similar results. All explanatory variables were included in the model initially, and a backward elimination procedure was used to remove terms that did not contribute significantly to the fit of the model. A chi-square test of the proportional odds (parallel lines) assumption indicated that the ordinal model was appropriate for the data. The SAS statistical software program (LOGISTIC procedure) were used for data analysis. Significance was accepted at the p < 0.05 level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

A prior diagnosis of asthma was reported by 61 (36%) children with a mean (range) age of onset of 3.4 (1) yr. A paternal history of asthma was reported for 21 (12%) children, and a maternal history of asthma for 30 (18%). There was a high prevalence of atopy (52%) (Table 1). Twenty-five (15%) of the children were currently using ₂ agonists for asthma symptoms, and 7% were using inhaled anti-inflammatory therapy (cromoglycate or inhaled corticosteroid). Current chest symptoms (cough, wheeze, breathlessness, or chest tightness) were reported by 56 (34%) of the children in the 2 wk before the study. Airway hyperresponsiveness to 4.5% saline was present in 23 (14%) of 165 children, of whom 18 were atopic. The mean (SD) volume of hypertonic saline delivered to the mouth was 14.3 (6.1) ml.

Sputum Quality

Sputum samples of adequate quality were obtained from 157 (92%) children. The median (interquartile range) sputum quality score was 6 (4) and the median volume of the expectorated sample was 2.0 ml, ranging from 0.5 to 10 ml.

The slide quality assessment technique proved superior to the quality assessment based upon macroscopic features of the sputum mucocellular clumps. The two methods were compared in a subset of 78 consecutively tested children (Table 2). Sixty-five (83%) of the samples were judged satisfactory (score = 4 or greater) using the slide quality assessment technique. Forty-five (78%) of the samples had macroscopically visible sputum mucocellular clumps with dimensions of 4.5-9 mm (6). Adequate slide quality assessment was an important determinant of sputum eosinophil counts. Samples with satisfactory slide assessments had significantly higher eosinophil counts (Table 2, p = 0.001) and were more likely to have eosinophils present on the slide (OR 25, p = 0.002). Consequently, samples with inadequate slide quality assessments were excluded. In contrast, the presence (or absence) of macroscopically adequate sputum mucocellular clumps did not have an important influence on sputum eosinophil counts. Samples without visible mucocellular clumps were just as likely to contain eosinophils as those with them (45% versus 65%, p = 0.14). The presence and average count of mast cells was also similar in samples with and without adequate plugs (p > 0.05).

Sputum Cytology

Sputum samples had a median total cell count of 1.55 × 10⁶/ml, with a range of cell counts from 0.2 to 88.7 × 10⁶ cells/ml. The results of sputum and nasal cytology are shown in Table 3. Eosinophils were present in the sputum of 70% of the children with current chest symptoms and 52% of children without current symptoms. Both current asthma symptoms (OR 2.25; 95% CI, 1.19-4.16) and AHR (OR 4.36; 95% CI, 1.7-11.2) were significantly and independently related to sputum eosinophils (Table 4, Figure 1). There was no significant confounding by current symptoms, AHR, age, and sex. The highest eosinophil counts were in children with both symptoms and AHR (Figure 2). There was no significant association between nasal symptoms and sputum eosinophils.

View larger version (7K): [in this window] [in a new window]   View larger version (7K): [in this window] [in a new window]   Figure 1.   Odds ratios (with 95% confidence intervals) for the association among sputum eosinophils (top panel ), sputum mast cells (bottom panel ), and asthma symptoms (p < 0.05), AHR (p < 0.05), and sex (p < 0.05 for mast cells and male sex).

View larger version (8K): [in this window] [in a new window]   View larger version (7K): [in this window] [in a new window]   Figure 2.   Sputum eosinophils (top panel ) and mast cells (bottom panel ) displayed as median with interquartile range in children grouped according to the presence or absence of asthma symptoms and AHR. Eosinophils were highest in children with both symptoms and AHR, whereas mast cells were highest in children with AHR.

Mast cells were present in the sputum of 16% of children with symptoms and 10% of children with no symptoms. A higher proportion of children with AHR had mast cells in their sputum (38%) compared to children without AHR (9%). Regression analysis identified that male sex (OR 3.23, 1.11- 5.88) and AHR (OR 10.00, 3.33-33.33) were significantly related to the presence of sputum mast cells (Figure 1).

GM-CSF-positive cells were increased in children with AHR (mean 31.3%, SE 7.7) compared to children without AHR (17.9%, 2.6, p < 0.05). There was no relationship between GM-CSF-positive cells and diagnosed asthma (22.5% versus 18.5%, p > 0.05) or asthma symptoms (p > 0.05).

Nasal Cytology

Nasal smears were obtained from 166 children. Nasal eosinophils ranged from 0% to 44.3%, with a mean eosinophil count of 3.7% (median, 0.3%). Eosinophils were found in 63% of samples from children with current chest symptoms and from 52% of samples from children without symptoms. Males (OR 3.23, 1.72-5.88) and children with AHR (OR 5.55, 2.17-14.29) had significantly higher degrees of nasal eosinophilia (Table 4). Nasal mast cells comprised an average of 0.07% of cells (median 0%), and ranged from 0 to 3.7% of total cells. Mast cells were present in the nasal smears of 28% of children and were not significantly related to any of the explanatory variables in the model.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has established that in children, airway inflammation is associated with both symptoms of asthma and airway hyperresponsiveness. Distinct differences were noted between the strength of these associations and specific inflammatory cell types. Airway eosinophils were linked to both current asthma symptoms and AHR, whereas mast cells were linked strongly to AHR but showed no association with current asthma symptoms. There was also a positive association between male sex and sputum mast cells. This interesting observation suggests that there may be an inflammatory basis for the increased prevalence of asthma in males in this age group.

In order to assess any response bias in this cross-sectional survey, the characteristics of the children who were tested were compared to a random sample drawn from the same geographic area (5). The study population had a similar prevalence of diagnosed asthma to the random population sample (36% versus 38%, p > 0.05). The prevalence of histamine AHR was 20% when this cohort was previously tested (9), which is comparable to the prevalence of 19.7% reported by Peat and colleagues (5) (p > 0.05). These comparisons argue against any response bias confounding the results.

The measurements of asthma symptoms, airway responsiveness, and airway inflammation that were used in this study were chosen to overcome difficulties noted in previous epidemiologic surveys of childhood asthma (17). In particular, symptoms were assessed for the 2 wk before the study visit in order to obtain a short-period prevalence. By contrast, previous studies have assessed symptoms in the 12 mo before measurement of airway responsiveness, which may obscure a relationship between symptoms and AHR because of the variable nature of asthma symptoms and AHR (17, 18). Airway responsiveness was assessed using hypertonic saline, since this test has a close relationship to clinical asthma (12, 13). Hypertonic saline also had the important advantage that it could be used to induce sputum (6, 7).

Airway inflammation is now considered integral to the pathogenesis of asthma and has been added to definitions of asthma in adults. In children, however, asthma continues to be defined in term of symptoms (19). Studies of airway inflammation in children are limited, and there is a need to evaluate the relationship between airway inflammatory cells and asthma symptoms in children. Bronchoscopy with bronchoalveolar lavage has demonstrated an eosinophil infiltrate in children with asthma, with the degree of AHR related to the eosinophil infiltrate (20). The use of hypertonic saline to induce sputum is an important advance to the study of airway inflammation in children (6). Clinical studies have shown that children with asthma demonstrate airway inflammation with eosinophils and metachromatic cells to a similar degree as adults (6). This study serves to validate this concept by providing epidemiologic evidence for the role of airway inflammation in childhood asthma.

The study has confirmed the association between current asthma symptoms and airway eosinophilia and demonstrated that AHR is linked to both airway mast cells and eosinophils. The presence of asthma symptoms was strongly linked with higher degrees of sputum eosinophilia. The practical implications of this observation are that a child who has chest symptoms suggestive of asthma has 2.25-fold increased odds of having eosinophilic inflammation of the lower airways. The likelihood of airway inflammation is further increased (five-fold) by finding increased airway responsiveness. These findings emphasize the role for anti-inflammatory therapy in symptomatic childhood asthma. Induced sputum cell counts may also have a role in monitoring therapy.

We could not demonstrate a relationship between the intensity of asthma symptoms (severe, less severe, nil) and airway inflammation. This is likely to be due to insufficient power of the study to test this relationship and the narrow spectrum of symptom intensity in this community-based survey. The relationship between symptom severity and degrees of airway inflammation warrants further study.

The results indicate a different role for the mast cell in childhood asthma than that of the eosinophil. There was no clear association between mast cells and recent asthma symptoms; however, mast cells were strongly linked to AHR to hypertonic saline. This demarcation between inflammatory cell types and the clinical manifestations of asthma is a novel finding, and it has not been seen in adult asthma (1, 3, 4, 6). Hypertonic saline is believed to indirectly cause bronchoconstriction by causing histamine release from airway mast cells. This close association between the stimulus and its mechanism of action may explain the strong relationship between mast cells and AHR in this study. If this is the case, then such a relationship may not exist for other bronchoconstrictor stimuli such as methacholine. Future studies are needed to define whether the association of saline AHR with airway mast cells is a feature peculiar to the bronchoconstrictor stimulus, peculiar to asthma in childhood, or typical of epidemiologic studies of asthma in general. It will be important to study larger numbers of children to exclude the possibility of a type II error.

The eosinophilic airway infiltrate in asthma is believed to be mediated by cytokines including GM-CSF and interleukin-5 that prime eosinophils for activation and prolong their survival at sites of inflammation. This study extends prior work by finding increased expression of GM-CSF immunoreactivity in airway cells in children with AHR to hypertonic saline. The dominant cell expressing GM-CSF had the morphological features of a macrophage. This finding supports a relationship between GM-CSF expression, airway inflammation, and AHR.

Gender differences in asthma prevalence vary by age and raise important issues concerning causes of asthma. In childhood, asthma and wheeze are more common in males, whereas this trend is reversed after puberty when asthma prevalence and morbidity are greater in females (21). Studies examining wheeze alone suggest that diagnostic bias is not responsible for the sex difference (22). Some but not all studies (23) report increased atopy in males. Male children do have increased bronchial lability (21) and a greater decrement in lung function associated with current wheeze (24). The current study indicates that male children are also more likely to have airway inflammation with mast-cell infiltration. A gender difference in the inflammatory reaction in children has also been reported with involuntary tobacco smoke exposure (25), where exposed males had greater peripheral blood eosinophil counts. These results indicate that in addition to playing a role in current asthma symptoms, airway inflammation may be implicated in the prevalence of AHR, which is a key feature of asthma.

Atopy was common in this population, and in previous studies it has been closely associated with asthma and AHR in childhood (9). This study did not have sufficient power to identify separate effects of atopy and AHR on airway inflammation, as virtually all of the children with AHR were also atopic. For this reason, atopy was not included as a variable in the regression analysis. A larger study would be required to examine this issue in more detail.

Nasal inflammation with eosinophils was also found to be related to symptomatic AHR (asthma), similar to the results for sputum eosinophils. Unlike sputum mast cells, nasal mast cells were not significantly higher in the groups with AHR. The strong relationship between nasal eosinophilia and symptoms confirms the association described by Williams and McNicol (26), and supports the high concordance between nasal disease and asthma.

In conclusion, this study describes the relationship between recent asthma symptoms, airway inflammation, and AHR in a epidemiologic study of children. There were important associations among airway eosinophils, asthma symptoms, and AHR, and a strong association between airway mast cells and AHR. The results demonstrate that it is possible to assess airway inflammation in children in an epidemiologic setting and that the use of induced sputum may be an important adjunct to the study of airway disease in children. The findings support the concept that airway inflammation is an important determinant of asthma symptoms and AHR in children.