METHODS

Children with classical clinical asthma (recurrent airways obstruction manifested by wheeze and dyspnoea relieved by bronchodilator therapy) aged 5.5–16 years were recruited from the outpatient department. Children with any chronic disease (other than asthma) were excluded. The children were tested when well (no respiratory tract infection or exacerbation of asthma for at least four weeks) during baseline. A clinical history was obtained and a repeatable question of cough dominant asthma was utilised.¹⁴ At baseline, the children underwent: venipuncture for ECP (serum ECP), skin prick test, spirometry, capsaicin CRS test,¹⁶ hypertonic saline (HS) test, induced sputum, and FEV₁ measured twice a day on a pocket digital spirometer (Vitalograph 2110 Electronic Diary, Buckingham, UK) for a week. The children and parents also kept separate daily daytime diaries for cough and asthma symptoms. The validated cough scale utilised a verbal descriptive score of daytime coughs, ranging from 0 (no cough) to 5 (cannot perform most usual daytime activity because of severe coughing).¹⁷ The asthma score was the average of four questions,¹¹ and the children and parents excluded the cough component of their asthma symptoms in the asthma diary score. Children and parents completed the child and carer's QOL questionnaire respectively.^18,19 History of environmental tobacco smoke (ETS) exposure was obtained; urine cotinine levels were measured using gas chromatography with deuterated cotinine as an internal standard.

Subjects were retested on days 1 (D1), 3 (D3), 7 (D7), and 28 (D28) of an asthma exacerbation (increased wheeze, chest tightness, and requirement for additional bronchodilator medications for more than eight hours). The study was carried out between February 1999 and January 2000. During repeat visits, all tests except the skin prick test were repeated. D1 tests were done before any additional steroids were taken. Exacerbation surveillance was done on diary cards and telephone reminders. On a monthly basis, new diary cards were sent, and the preceding month's diary card returned. Written consent was obtained from the parent(s) and the study was approved by the local hospital's Ethics Committee for Human Research.

Tests

Serum ECP was processed by a dedicated service on site using radioimmunoassay (Pharmacia & Upjohn, Uppsala, Sweden). Skin prick test was performed using six allergens, and negative and positive controls.²⁰ Medications were withheld prior to all tests.²¹ Spirometry (Cosmed Kit, Italy) was performed in a sitting erect position with a nose clip. Outcome measures of capsaicin CRS test performed with regulation of the inspiratory flow¹⁶ were C2 and C5 (lowest concentration of capsaicin that stimulated ≥2 coughs and ≥5 coughs respectively). The standardised 4.5% HS challenge was used.²¹ Presence of AHR (AHR positive) was defined as a fall in FEV₁ of ≥15% from baseline. Children were encouraged to cough and expectorate after each HS nebulisation. Sputum collected was processed immediately in the standardised manner.⁴

Statistical analysis

Statistical software package SPSS version 7.5 (Chicago, USA) was utilised. For continuous variables, the Mann–Whitney test was used for unpaired data and the Wilcoxon test for paired data. When repeated measurements were made, the Friedman and Kruskal–Wallis tests were utilised for paired and unpaired data respectively. Spearman correlation (r_s) was used for correlation. A two tailed p value of ≤0.05 was considered significant. Logistic regression was utilised for prediction of the development of an asthma exacerbation within the follow up period.

RESULTS

Twenty one children were recruited (16 boys, five girls; median age 10.8 years, interquartile range (IQR) 3.9). Only 11 of the 21 children had an exacerbation of their asthma during the observation period of 175 person months. The median age of these 11 children (nine boys, two girls) was 10.53 years (IQR 3.9). All but one were atopic, nine were on regular inhaled corticosteroids, three on cromolyns, and four on long acting β₂ agonists. Six children had passive tobacco smoke exposure and five did not. All but one child had cough as a significant symptom with their asthma episodes. None were hospitalised for their asthma exacerbation, four children were treated with a course of oral steroids, three with doubling their inhaled steroid dose, and the remaining four were managed with increased use of salbutamol (3–4 hourly for several days). The median length of time from baseline to D1 was 1.5 months (IQR 1.0). There was no significant difference in the length of follow up between those with (median 21.3 months, IQR 0.9) and without an asthma exacerbation (median 20.9 months, IQR 2.2).

At the onset of the exacerbation (D1), cough and asthma symptoms significantly increased (p = 0.00001). Sputum ECP also significantly increased and sputum eosinophils were increased (>2.5%) in all but one child. Trends were seen in FEV₁ fall and FEV₁ variability but were not significant. Table 1 shows the course of asthma scores, cough scores, QOL, spirometry, and AHR indices at baseline and during the exacerbation. Table 2 shows those of CRS, serum, and sputum indices. The trend of the various markers shows a non-concurrent peak; serum ECP and C2 peaked on D1, but non-significantly when compared to baseline. These peaks were concurrent with diary symptoms asthma and cough scores on D1. Sputum ECP significantly peaked on D3 by an increase of more than 10-fold. IL-8 and MPO did not significantly change during asthma exacerbation. Cough and asthma scores between baseline and D1 for both child and parent scores were significantly different. Comparing D3 and baseline, significant differences were found for both scores in parent recorded diaries and child recorded cough score but not in child recorded asthma score. There was no significant difference between exacerbation and baseline scores in QOL, all spirometry indices, AHR, and CRS.

CRS outcome measures (C2 and C5) did not correlate to any marker of clinical severity (asthma score, cough score, QOL), pulmonary function indices (FEV₁, forced vital capacity (FVC), FEV₁ variability) or inflammatory marker (IL-8, serum ECP, sputum ECP, serum eosinophils, sputum eosinophils, sputum neutrophils, MPO) of asthma during any of the test days (baseline, D1, D3, D7, and D28). At baseline and D1, there was a strong correlation between cough and asthma score, as recorded by the child and parent(s) (tables 3 and 4). On D3 and D7, this relation was lost except for that between child recorded asthma score with parent recorded cough score on D3, and between parent recorded asthma score with child recorded cough score on D7.

On logistic regression, no baseline factor could significantly predict the asthma exacerbation.

The mean urine cotinine in ETS exposed children was 94.3 (SEM 91) ng/ml and that in ETS non-exposed was 1 (SEM 0) ng/ml (mean difference 93, p = 0.037). Smoke exposure did not significantly influence any inflammatory marker on any test day. At baseline but not during exacerbation test days, child recorded asthma and cough scores and parent recorded cough scores were significantly higher in ETS exposed children when compared to ETS non-exposed children (table 5).

DISCUSSION

This study has described the course of various markers of cough and asthma, spirometry indices, and airway inflammation in childhood asthma, from baseline to an acute mild exacerbation and resolution state. The mild exacerbations were characterised by an increase in symptoms and airway inflammation, with non-significant changes in lung function. Only sputum ECP showed a significant increase during the exacerbation, and peaked later than clinical scores. Asthma scores consistently related to cough score only in the very early stages of the exacerbation. We have also shown that neither CRS nor cough scores relate to any of the inflammatory markers measured, and that spirometry, AHR, and FEV₁ variability were insensitive markers of mild asthma exacerbations. Lastly, baseline measurements could not predict the development of an exacerbation within the follow up time of 175 person months.

We have previously shown that in acute severe asthma, airway inflammation is characterised by infiltration of both eosinophils and neutrophils.⁴ In the present study where mild exacerbations were studied, eosinophilic but not neutrophilic inflammation was present (evidenced by D1 sputum eosinophils >2.5%, increase in sputum ECP but not IL-8 or MPO). This raises the question of whether the more common occurrence of mild asthma exacerbations in childhood differs from the less common severe exacerbations. In adults, neutrophilic inflammation occurs in the more severe forms of asthma,^22,23 and typically reflects an infective cause.²⁴ Eosinophilic exacerbations can be triggered by allergen exposure or non-compliance with corticosteroid therapy.²⁵

Poor agreement between clinical judgement (symptoms, spirometry) and airway inflammatory component assessed by sputum cell counts has been recently described in adults with asthma.³ We have also shown that asthma symptoms related poorly to most inflammatory markers in mild exacerbations of childhood asthma. The relation between sputum ECP and asthma was significant but negatively associated; as asthma symptoms were subsiding, sputum ECP was at its peak. In cross sectional studies, sputum ECP and serum ECP related to FEV₁² but in this small prospective study, the relation was not found. In this study of mild asthma exacerbations, changes in AHR, FEV₁, and FEV₁ variability were insensitive measures, when compared to sputum ECP and symptom perception. Although the trend of QOL in both child and parents reflected asthma scores, the changes were insignificant. This would be expected for this acute setting as QOL is insensitive for short term effects. We were unable to define any factors that could predict the development of an asthma exacerbation within the follow up period. As viral infections are the commonest cause of asthma exacerbations in children,²⁶ this finding is not surprising. However, this study is limited by small numbers and a relatively short follow up period. Martinez argues that in an individual patient (in contrast to research), single markers for childhood asthma with its various phenotypes are unlikely to be useful as they are context and time variant dependent.⁷ This study cannot answer the question of the usefulness of the various markers of asthma in an individual patient, but the data presented support the sensitivity of sputum ECP when compared to other indices of asthma severity.

Although cough features prominently in asthma management plans, whether the presence of cough does indeed herald the onset of an asthma exacerbation has never been shown prospectively. This study has shown this is indeed true in a selected group of children, whose parents had previously indicated that cough was a significant symptom with their asthma episodes. We were unable to compare coughers versus non-coughers because of the small sample size. Although we used a validated cough score chart,¹⁷ this finding is also limited by the lack of an objective cough monitor, as symptoms reflect perception, which is known to be dependent on the population studied.²⁷ Nevertheless we found no relation between CRS and asthma inflammatory markers, again raising the question whether cough should be used as a marker of asthma severity in the non-acute asthma. The occurrence of cough without an increase in CRS may be related to an irritant effect or cough receptor activation by eosinophilic activates.²⁸ The reduction in cough symptoms while sputum ECP was still raised (D3) could be a result of temporal adaptation to the cough stimulus.^29,30 In this study CRS measures did not change, unlike our previous study which described increased cough sensitivity during the coughing period in hospitalised children with asthma.¹⁴ We postulate this difference relates to the different asthma exacerbations studied (mild versus severe exacerbation where neutrophilic airway inflammation occurs⁴), and possibly to the small sample size.

Children with ETS exposure were more likely to have higher asthma and cough scores at baseline but not during an exacerbation. We did not find any difference in the inflammatory markers. A three hour ETS exposure did not change cellular and inflammatory markers (ECP, MPO, tryptase, prostanoids, and leukotrienes) in bronchoalveolar lavage fluid and FEV₁, but increased symptoms in adults with mild asthma.³¹ The lack of difference could be related to the small subject numbers or more likely, the appropriate inflammatory marker was not measured. In animal studies, ETS exposure primarily causes neurogenic inflammation,³² which is arguably important in the symptom perception of cough and chest tightness.³³

We conclude that in mild asthma exacerbations in children, eosinophilic inflammation is dominant, and that the commonly used clinical indices of asthma, FEV₁, FEV₁ variability, and AHR are insensitive markers. These markers could not predict the development of an asthma exacerbation. In the subgroup of children with asthma (those with cough dominant asthma), the presence of cough heralds the onset of an exacerbation and increased eosinophilic inflammation, but cough scores and CRS do not reflect eosinophilic inflammation during the course of the exacerbation. Whether cough scores and CRS reflect neutrophilic airway inflammation in severe childhood asthma exacerbations is unknown.