INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: asthma; eosinophil; induced sputum wheeze; cough; children

Although asthma is typically characterized by episodic wheezing and variable airway obstruction, variant presentations can occur where persistent cough or recurrent "chest colds" are the dominant symptoms and wheeze is minimal or absent. Some guidelines recommend the diagnosis of asthma when cough is an isolated symptom (1), but others advise caution in attributing a diagnosis of asthma to patients with recurrent cough (2). Eosinophilic inflammation of the airways is characteristic of typical asthma with wheezing in both adults and children (3) and is the focus of treatment with inhaled corticosteroids (5, 6). The role of airway inflammation in the variant asthma syndromes of chronic cough and recurrent chest colds is not well defined. There is evidence of eosinophilic bronchitis in adults with chronic cough (7, 8), but the subjects in these studies were selected from a tertiary referral setting, and chosen because of corticosteroid responsiveness. Further, other studies of cough in adults do not identify eosinophilic inflammation, but rather report a neutrophil response in chronic cough (9). The pattern of airway inflammation in children with persistent cough and recurrent chest colds is not well described, but is of relevance to both the mechanisms and treatment of these clinical syndromes.

The aim of this study was to assess whether children with persistent cough and recurrent chest colds had evidence of airway inflammation as reflected by increased eosinophils and eosinophil cationic protein (ECP) using induced sputum analysis. To avoid the selection bias that has occurred in other studies, we identified subjects from a community survey and tested the hypothesis that children with persistent cough and chest colds have increased sputum eosinophils, similar to those with current wheeze.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

The parents of 569 children in Years 3, 4, and 5 at five metropolitan primary schools were invited to complete a questionnaire about the child's and family's respiratory health (10). A total of 390 questionnaires were returned (68%). The selection criteria for the case groups and the control group were as follows:

1. Wheezechildren reporting two or more attacks of wheezing in the last 12 mo, irrespective of other symptoms.

2. Cough alonechildren reporting, in the last 12 mo, a dry cough at night (without a cold or chest infection) that lasted for more than 2 wk and no wheeze or recurrent chest colds.

3. Chest coldchildren reporting two or more chest colds ("colds that go to the chest") in the last 12 mo, but with no associated wheeze.

4. Controlchildren ineligible to be included in the three case groups.

All of the cough (n = 21) and wheeze (n = 46) groups were invited to participate, while the chest cold (n = 32) and control (n = 40) groups were randomly sampled at a ratio of 1 in 2. Eighty-five children agreed to take part, 83 attended for testing, and 80 completed sputum induction and hypertonic saline challenge (Table 1). Peak expiratory flow (PEF) symptom diaries were available for 66 of the 76 children who agreed to complete diaries. Nine children were excluded from the diary analysis because they completed less than 70% of the 56 d of the diary and one was unable to be contacted for follow-up.

Design and Measurements

Children attended a single visit for testing, and then completed a 2-mo period of home monitoring of symptoms and PEF (10). After informed consent was obtained, the following tests were performed: allergy skin prick tests (3), spirometry, hypertonic saline inhalation challenge to assess airway responsiveness (3), and sputum induction (3) for cytological analysis (11, 12). Sputum portions were selected, smears were made (11), and a further 100 µL of sputum was mixed with dithiothreitol in lysis buffer for assessment of cell-associated ECP (12).

Statistical Analysis

Individual plots of PEF and symptoms were examined. The first week of each diary was excluded from the analysis in order to remove any learning effect associated with completion of the diary. The frequency of symptomatic days was calculated as the number of days in which the morning symptoms score was > 0, expressed as a percentage of the total number of days completed in the diary. Cough and wheeze were converted to dichotomous variables based on their median frequencies. For morning wheeze, the cut-point was 0% of days, and for morning cough, it was  29% of days.

PEF and baseline FEV₁ were expressed as a percentage of predicted (13). PEF variability was expressed as the standard deviation as a percentage of mean PEF for each day (14). Eosinophilic bronchitis was defined as a sputum eosinophil percentage > 2.5% (15). ANOVA with Tukey post hoc testing was used to compare normally distributed continuous variables and significance was accepted at the p < 0.05 level.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

We studied 83 children aged between 8 and 11 yr, of whom 55% were male. The children agreeing to participate were compared with those who were eligible but did not participate. The study children were representative of their respective symptom groups within the larger population in terms of age, sex, and other characteristics. Asthma diagnosis tended to be more common and rhinitis less common in children studied from the chest cold group (p > 0.05). A family history of asthma was reported more often in the children studied in the wheeze group. In the wheeze and control groups, those not included had more environmental tobacco smoke (ETS) exposure (p = 0.04). The prevalence of atopy ranged between 35% and 66% in the four symptom groups. The cough and wheeze groups had a significantly higher prevalence of house dust mite sensitization (Table 1, p < 0.05).

Sputum Analysis

Adequate sputum samples were obtained from 67 children (81%, Table 2). The quality of the sputum samples was good, with a mean (SD) quality score of 3.8 (0.6). Sputum eosinophils were significantly higher in children with wheeze, when compared with the other three groups (Table 2, Figure 1, p = 0.03). The three symptom groups had similar cell differentials for other cell types. No significant difference was seen in sputum ECP levels, although median values differed especially between wheeze and control groups. Sputum cell counts and ECP were comparable between children with airway hyperresponsiveness (AHR) and those not reactive to hypertonic saline (Table 3). There was no difference in inflammatory markers between these two groups. Sputum eosinophils and ECP were not significantly correlated with peak flow variability or dose-response slope.

View larger version (8K): [in this window] [in a new window]   Figure 1.   Sputum eosinophils (%) in children with wheeze (n = 20), persistent cough (n = 10), recurrent chest colds (n = 13), and asymptomatic control subjects (control, n = 21). Eosinophils were elevated in children with wheeze (p = 0.03).

The prevalence of eosinophilic bronchitis (sputum eosinophils > 2.5%) was 45% in the wheeze group, which was significantly higher than that of the control group (9.35%, p = 0.04, Figure 1). In the wheeze group, all of the children with AHR also had eosinophilic bronchitis (p = 0.008). Eosinophilic bronchitis was present in two children from the cough group (20%) and two from the chest cold group (15%, p > 0.05 versus control). In these groups, eosinophilic bronchitis was not associated with AHR (p > 0.05). The two control children with eosinophilic bronchitis were both atopic and reported frequent nasal symptoms in their symptom diaries on over 80% of the days. Neither of these children had AHR, although one child had a peak flow variability of 20%. Similarly, the children with chest colds and eosinophilic bronchitis recorded frequent nasal symptoms (72% and 37% of days, respectively), as did one of the children with cough and eosinophilic bronchitis who recorded nasal symptoms on 80% of days. Each of the children with eosinophilic bronchitis in the cough and chest cold groups also reported exposure to ETS.

Symptoms

Prospective diary recording of symptoms was used to validate the questionnaire classification. Wheeze was a good discriminator between the groups, as children in the wheeze group recorded wheeze in their diary on an average (SEM) of 30 (6.2)% of days, whereas wheeze was recorded for 8 (5.5)% of days in the cough group, 1.04 (0.4)% of days in the chest cold group, and 1 (2.2)% of days in the control group. Cough was recorded on 52 (8.9)% of days in the cough group and 45 (7.8)% of days in the wheeze group, which was significantly greater than cough frequency in the control group (23 [5.4]%) and chest cold (28 [6.1]%) groups (p = 0.01). Nasal symptoms were recorded in the diaries on 47 (8.5)% of days in the wheeze group and 49 (9.7)% of days in the cough group, compared with 26 (5.5)% of days in the control group and 29 (5.3)% of days in the chest cold group (p = 0.04).

The frequency of morning cough was negatively correlated with morning peak flow (r = 0.31, p = 0.006) and evening peak flow (r = 0.33, p = 0.003), indicating that peak flow values tended to be lower in children who reported more frequent cough in their diary. There was also a correlation between the frequency of cough and wheeze (r = 0.48, p = 0.001). Peak flow variability was correlated with wheeze (r = 0.69, p < 0.05) and with the percentage of days with nasal symptoms (r = 0.44, p < 0.05). Children with wheeze had significantly lower PEF compared with the cough, chest cold, and control groups (p = 0.0001, Table 4). Airway responsiveness was similar between the groups, although there was a trend for more children with wheeze to have increased airway responsiveness.

Children with wheeze reported greater use of asthma medications (short-acting ₂-agonists, inhaled corticosteroids) in the 12 mo prior to testing (Table 1, p < 0.001). Short-acting ₂- agonists were used by 22 (88%) of the children in the wheeze group. Antiinflammatory asthma medications had been used by two children in the cough group and two in the chest cold group. During the diary period, these children in the cough and chest cold groups were not using inhaled corticosteroid or cromoglycate.

Environmental Smoke Exposure

Exposure to ETS was recorded by 30 (35%) children, and was more frequent in children with persistent cough (67%) and chest colds (44%) than those with wheeze (24%) or control subjects (27%, p = 0.04, Figure 2). Children exposed to ETS reported a greater frequency of cough and nasal symptoms, but not wheeze (Table 5). They also had lower lung function (p = 0.03) and tended to have higher sputum ECP and eosinophil counts.

View larger version (11K): [in this window] [in a new window]   Figure 2.   Exposure to environmental tobacco smoke (ETS) at home was more prevalent in the children with cough and chest colds than control subjects (*p = 0.04).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There has been considerable debate as to whether variant asthma can present as cough or recurrent chest colds. This study represents the first examination using induced sputum to examine the characteristics of airway inflammation in a community sample of children with these syndromes. We found that eosinophilic airway inflammation was associated with the wheezing phenotype where it was present in almost half of the children with wheeze and in all of the children with wheeze and AHR. Persistent cough or recurrent chest colds were not strongly associated with sputum eosinophilia but were associated with exposure to ETS.

To minimize selection bias, we conducted a cross-sectional community survey and selected children to form the different symptoms and control groups. The response rates to this process (68% and 61%) were not ideal, but satisfactory. It is difficult to estimate the direction of bias associated with the response rate; however, the children who were studied were representative of the population base. The use of symptom questionnaires in the cross-sectional survey may result in a potential problem with recall bias, as some studies identify a poor correlation between cough reported by parents and by children (16). To evaluate the importance of this, we asked the children to prospectively record their symptoms in a diary. This established that wheeze reported in the cross-sectional questionnaire was a good discriminator between the groups, and that most children allocated to this group were currently symptomatic during the testing period.

Airway responsiveness was assessed using hypertonic saline. This indirect stimulus is correlated with airway inflammation (3), and although it is less sensitive than histamine, it is highly specific and performs well in community surveys of asthma (3, 17). The other advantage of hypertonic saline is that it can be used to induce sputum for the assessment of airway inflammation. Induced sputum has emerged as a reproducible and acceptable assessment tool for the investigation of inflammation in airway diseases. Sputum cell counts analyzed from smears (11) are reproducible, although time consuming to evaluate. We also included measurement of ECP in lysed sputum cells as a biochemical marker of the number of eosinophils present (12). Although sputum ECP was higher in the wheeze group, this was not significant. This may be due to the reduced specificity of this assay when compared with sputum eosinophils (18). ECP is present not only in eosinophils, but also in saliva (19), and can be detected in neutrophils, possibly due to uptake by these cells (20). These other sources of ECP reduce its specificity for eosinophilic bronchitis.

Sputum eosinophilia occurred in almost half of the children with a history of wheezing and all of the children with wheeze and AHS. This observation concurs with prior studies of airway inflammation in childhood asthma (3, 15) and adults (5, 11, 12, 18, 19) that demonstrate increased sputum eosinophils in asthma. Eosinophilic bronchitis was less common in children with chronic cough and recurrent chest colds. We found that about 20% of children in these groups had eosinophilic bronchitis, which is similar to the prevalence of eosinophilic bronchitis in persistent cough reported by others (21, 22). Although eosinophilic bronchitis was less common in children with cough than those with wheeze, it is still an important characteristic to identify. When eosinophilic bronchitis is present, it is associated with a favorable response to corticosteroid in chronic cough (8, 21), asthma (5), and chronic obstructive airway disease (5, 23). Furthermore, there is little response to corticosteroid when eosinophilic bronchitis is absent (5, 21, 23). Together, these data indicate that labeling all children with persistent cough or recurrent chest colds as variant asthma is not appropriate. There is a subset of patients with cough who have a pattern of airway inflammation similar to asthma, which responds to corticosteroid therapy, but which is not associated with increased variability of airway obstruction. In the absence of variable airflow obstruction, the label of asthma is not appropriate, and they are probably best referred to as cough with eosinophilic bronchitis, or corticosteroid-responsive cough.

Cough is a common symptom in the community. Nighttime cough was reported by up to 28% of children in Australia in the International Study of Asthma and Allergies in Childhood (ISAAC) asthma prevalence survey (24). Children with cough can have frequent episodes (25) leading to loss of time from school (26). The majority of these children have clinical features different from those with asthma (27). They tend to have more features (risk factors) in common with children without asthma (28), and on prospective follow-up seldom progress to develop wheeze and asthma (29). Persistence of cough in longitudinal epidemiological surveys is not associated with AHR or a family history of asthma, but is associated with hay fever and environmental tobacco smoke exposure (27). Our data are consistent with these observations, as those with persistent cough did not have the same pattern of airway inflammation as those children with wheeze, and when eosinophilic inflammation occurred in the children with cough, it tended to be in association with nasal symptoms.

The promotion of persistent cough as a variant of asthma has led to many children with persistent cough being treated with antiinflammatory medications (30, 31). Although this may be appropriate for the small number of children in a referral setting who may have asthma and/or airway inflammation with eosinophils as the cause of their cough, this is uncommon in the community and in primary care. In primary care, isolated cough is not a predictor for asthma (32) and the general application of asthma therapy is seldom helpful (33). The results of the present study suggest that only a small proportion of children with the variant asthma syndromes of cough and recurrent colds have eosinophilic bronchitis, and that inhaled corticosteroids are probably not indicated as initial therapy for the majority of children with chronic cough or recurrent chest colds. The optimal approach to these patients is not defined, but could include objective assessment of airway function and airway inflammation to determine whether asthma or eosinophilic bronchitis is present, investigation for other associations of cough such as rhinitis and gastroesophageal reflux, or empiric therapy for these disorders. Systematic investigations are required to define these issues.

Other causes of chronic cough in these children could include exposure to ETS (27, 34) or air pollution (10). We have previously defined an association between passive smoke exposure and cough in Hunter school children (10), and also between exposure to particulates (PM10) and respiratory symptoms. This study confirms these results, and demonstrates that cough and nasal symptoms were more frequent in children exposed to ETS. There was a tendency to greater eosinophil degranulation in children with ETS exposure, reflected by higher sputum ECP levels in these children. Other cell counts were not different. When children with cough or chest colds did have increased sputum eosinophils, it tended to be associated with allergic rhinitis. The sputum eosinophilia in these subjects could represent expectoration of nasal secretions or eosinophilic lower airway inflammation occurring as part of the rhinitis syndrome (35). Because the sputum collection method avoids collection of nasal discharge, we consider the sputum eosinophilia in these children to be a feature of lower airway inflammation, as previously reported in allergic rhinitis (35). Nasal steroids would be appropriate for those with allergic rhinitis, although it is unclear whether this would improve the lower respiratory symptoms.

In conclusion, this community-based survey of children with chronic respiratory symptoms has shown that wheeze is a good discriminator for the presence of eosinophilic bronchitis, and that the presence of persistent cough and recurrent chest colds without wheeze should not be considered a variant of asthma. However, eosinophilic bronchitis did occur in a significant minority of these children. These results have therapeutic implications that deserve further research.