Mild Exacerbations and Eosinophilic Inflammation in Patients With Stable, Well-Controlled Asthma After 1 Year of Follow-up*

Transcription

1 Mild Exacerbations and Eosinophilic Inflammation in Patients With Stable, Well-Controlled Asthma After 1 Year of Follow-up* Jose Belda, MD, PhD; Jordi Giner, RN; Pere Casan, MD, PhD; and Joaquin Sanchis, MD, PhD Objectives: To determine the time to exacerbation and probability of a mild exacerbation of asthma, and the impact of eosinophilic inflammation on these parameters in patients with stable, well-controlled asthma. Patients and methods: A cohort of 31 patients with stable, well-controlled asthma receiving inhaled steroid treatment regularly were followed up for 1 year or until a mild exacerbation occurred. Mild exacerbation was defined as symptoms of asthma lasting > 48 h with a fall in peak expiratory flow > 20%. FEV 1, provocative concentration of methacholine causing a 20% fall in FEV 1, eosinophil count, and eosinophilic cationic protein (ECP) levels in blood and in sputum were measured at the first visit and every 2 months. Results: At baseline, the mean (SD) eosinophil count was /L ( /L) in blood and 13% (14%) in sputum; ECP was 30 g/l (28 g/l) in blood and 75 g/l (85 g/l) in sputum. Thirteen subjects experienced a mild exacerbation during the 1-year follow-up period. The mean time to mild exacerbation was 293 days (95% confidence interval [CI], 248 to 337 days), and the cumulative probability of not experiencing a mild exacerbation in 1 year was 49% (95% CI, 39 to 59%). An increased risk of mild exacerbation was associated with blood eosinophil count > /L (relative risk 4.5; 95% CI of relative risk, 1.8 to 38.0), blood ECP > 20 g/l (relative risk, 2.1; 95% CI of relative risk, 1.0 to 9.2), and sputum ECP > 40 g/l (relative risk, 2.5; 95% CI of relative risk, 1.2 to 11.2), but was unassociated with other variables. Conclusions: Patient with stable, well-controlled asthma are at risk of mild exacerbation during 1 year of follow-up despite regular inhaled steroid treatment. Eosinophilic inflammation expressed as eosinophil count and ECP is associated with higher risk of mild exacerbation. (CHEST 2001; 119: ) Key words: asthma; eosinophilic cationic protein; eosinophils; exacerbation; risk factors Abbreviations: CI confidence interval; ECP eosinophilic cationic protein; PC 20 provocative concentration of methacholine causing a 20% fall in FEV 1 ; PEF peak expiratory flow *From the Hospital de la Santa Creu i Sant Pau, Barcelona, Catalunya, Spain. Supported in part by a grant from the Sociedad Española de Neumología. Manuscript received July 15, 1999; revision accepted October 27, Correspondence to: Joaquin Sanchis, MD, PhD, Departament de Pneumologia, Hospital de la Santa Creu i de Sant Pau, C/Sant Antoni Maria Claret, Barcelona, Spain; jsanchis@hsp.santpau.es Eosinophils and their products are known to play an important role in asthma, 1 and the measurement of blood and sputum levels of such markers of inflammation may provide information reflecting the evolution and control of disease. 2 However, the clinical usefulness of these measurements is still uncertain. Asthma is characterized by an extremely variable course with unpredictable exacerbations. In a substantial number of studies of exacerbation, the emphasis has mostly been on life-threatening episodes occurring in patients with symptomatic asthma treated in the emergency department. Thus, many risk factors for severe exacerbations have been described, 3,4 such as age, smoking history, poor disease control, previous mechanical ventilation, admission to the ICU, history of worsening asthma, use of air conditioning, deterioration in FEV 1, labile asthma, and others. In comparison, few studies have been published on mild exacerbations in patients with long-term, stable asthma 5 7 ; therefore, there is little information about the frequency and risk factors for mild exacerbation in subjects with well-controlled asthma whose treatment has been clinically optimized. Information on this aspect is important for determining appropriate future management and may have influence on the lives of the patients. CHEST / 119 / 4/ APRIL,

2 We aimed to determine the time to exacerbation, the cumulative probability of mild exacerbation in a cohort of subjects with well-controlled and moderate-treated asthma, and the impact on these two variables of eosinophilia and eosinophilic cationic protein (ECP) levels in induced sputum and peripheral blood. Subjects Materials and Methods We recruited subjects aged 18 to 65 years with at least a 2-year history of asthma but who reported being asymptomatic with regular inhaled steroid treatment (defined as well-controlled asthma, as previously described 8 ). The diagnosis of asthma and assessment of clinical and functional severity were based on the international consensus of All the subjects were classified as having moderate asthma (asthmatics with normal spirometry findings or mild airflow limitation who need regular steroid treatment to remain symptom free) and were allergic to house dust mites, as judged by symptoms and a positive skin prick test result (wheal 3 mm). Subjects were lifelong nonsmokers or ex-smokers for at least 5 years. They had been receiving treatment regularly with the same dose of inhaled steroid during the previous year and were free of exacerbations and chest infections in the 3 months preceding enrollment. Use of an inhaled 2 -agonist such as salbutamol or salmeterol was started before the study, but no other antiasthmatic medications were allowed. Study Design This was a longitudinal observational study with blind assessment of outcome measures. Subjects were recruited from the outpatient clinic of a referral hospital. Approval from the hospital ethics committee was obtained, and all subjects gave written informed consent. The inclusion period was the fall season (September to November) when most subjects were not highly exposed to house dust mites. The end point was defined by the occurrence of mild exacerbation, defined as symptoms of asthma lasting 48 h and preventing subjects from performing their usual activities, in association with a fall in peak expiratory flow (PEF) 20% of the mean morning value established during the run-in period. Eligible subjects were monitored over a 2-month run-in period. Symptoms and lung function stability were rigorously checked. Subjects were then eligible for the study if they complied with the recording of a home diary and if their asthma was under control. Subjects were then interviewed on 2 consecutive days every 2 months for 1 year. On the first day, a clinical history was obtained, and a physical examination and methacholine challenge test were completed; on the second day, venous blood and induced sputum samples were obtained and examined. Laboratory technicians were blinded to the clinical characteristics until the end of the study. In addition, subjects recorded the following data twice daily in a home diary: PEFs, symptom scores, and the number of puffs of rescue medication needed. No changes were made in treatment or follow-up procedures during the study, and the prescribed dose of inhaled steroid remained the same until mild exacerbation occurred. In the event of exacerbation, subjects were instructed to immediately contact the investigators and come to the hospital within 48 h. The investigators assessed each such event according to preestablished exacerbation criteria, and, in case of mild exacerbation, they increased the dose of inhaled steroid threefold (maximum, 3,600 g/d) for a minimum of 4 days or until PEF returned to normal values. Methods Time to Exacerbation: The primary outcome studied was the time to mild exacerbation, which was defined as the number of days elapsed from the beginning of the study to the onset of the first exacerbation or to the end of the study. Spirometry and Methacholine Challenge Test: Standard spirometry and methacholine challenge tests were performed, the latter using a standardized continuous breathing method. 10 Blood Measurements: Peripheral blood was collected into ethylenediaminetetra-acetic acid specimen (Vacutainer) tubes and serum separating tubes (serum separator gel) before sputum induction. Total and differential WBC counts were obtained using a differential cell counter on ethylenediaminetetra-acetic acid tubes. Serum separating tubes were centrifuged at room temperature (3,000 revolutions/min for 20 min) within 2 h, and serum was stored at 70 C. Sputum Samples: Sputum was induced by inhalation of hypertonic saline solution at increasing concentrations (3%, 4%, and 5%) for 10 min. Selected portions from the whole expectorate were processed as previously described 11 and validated in our laboratory. 12 Cytologic Study: The total cell count was obtained using a Nebauer hemocytometer. The relative and absolute number of eosinophils per milligram of processed sputum was calculated. Four hundred inflammatory cells were counted on two slides stained with Giemsa and eosin. ECP Concentration: The concentration of ECP in the thawed supernatant was determined using an ECP FEIA commercial kit on a Pharmacia UNICAP System (Pharmacia Diagnostics; Uppsala, Sweden) by fluoroenzyme immunoassay. Results were expressed in micrograms per liter. Statistical Analysis Sample size was calculated to estimate the probability of mild exacerbation with 10% accuracy assuming an unknown rate of 50%. Cut-off points for inflammatory markers were established a priori based on the available normal values in our laboratory, except for the provocative concentration of methacholine causing a 20% fall in FEV 1 (PC 20 ), for which an arbitrary cut-off of 4 mg/ml was selected to differentiate borderline hyperresponsiveness. However, no sample-size assumptions were made about differences between subgroups (exploratory approach). Differences between unavailable subjects and subjects who were followed up were checked using a Mann-Whitney U test. Statistical analysis of the exacerbations was performed by the Kaplan- Meier survival analysis method, in which baseline data are used to explore relationships and predictions, with survival standing for time to exacerbation. The effect of each variable tested on the risk of exacerbation was determined by sorting the subjects into two groups depending on high or low baseline values; this was performed for every variable, and the time to exacerbation between groups was compared by the Mantel-Haenszel method (log-rank). A p value 0.05 was accepted as significant. Ninetyfive percent confidence intervals (95% CIs) were calculated to determine the effects of sampling variation on the precision of estimated statistics. A double check with multivariate Cox regression analysis was performed to adjust for steroid dose, long-acting -agonist dose, and to rank independent variables Clinical Investigations

3 Results Subject Demographics and Enrollment We identified 31 consecutive subjects (19 female and 12 male) with stable, well-controlled asthma. Six subjects were ex-smokers, and the remaining 25 subjects were lifelong nonsmokers. In the run-in period, 13 subjects had no symptoms and the remaining 18 subjects had very mild symptoms (mainly wheezing or coughing) and needed a mean of 0.75 puffs of bronchodilator per day. Baseline cough, wheeze, and bronchodilator use were occasional and mainly related to exercise. The median dose of inhaled steroid (budesonide) was 800 g/d (5 subjects received 400 g/d, 2 subjects received 600 g/d, 19 subjects received 800 g/d, 1 subject received 1,200 g/d, and 4 subjects received 1,600 g/d). The mean dose of inhaled long-acting 2 - agonist (salmeterol) was 50 g/d (3 subjects received 50 g/d, 7 subjects received 100 g/d, 7 subjects received 200 g/d, and 14 subjects did not receive it). The characteristics of the subjects at the starting point of the study, after the run-in period, are described in Table 1. Survival Analysis Thirteen subjects had at least one exacerbation during the year of follow-up. Of the remaining 18 subjects, 14 subjects finished the year of follow-up with no exacerbation and 4 subjects were unavailable for follow-up. Two subjects left the study because of changes in work status that prevented them from complying with the scheduled visits, 1 subject withdrew for personal reasons, and 1 subject was excluded because of poor compliance; these 4 subjects were similar to the other 27 subjects at baseline, and they participated in the study for 62 days, 34 days, 140 days, and 62 days, respectively. The mean (95% CI) time to mild exacerbation for the whole group was 293 days (248 to 337 days) (range, 6 to 375 days) and 175 days (105 to 245 days) for those who experienced a mild exacerbation. The cumulative probability of a mild exacerbation was 49% (39 to 59%) independent of treatment. Risk Factors An absolute number of eosinophils in blood /L was associated with a shorter time until an exacerbation occurred and a cumulative probability of a mild exacerbation (SD) of 83% (11%) vs 23% (11%) of the opposite group (relative risk, 4.5; 95% CI of relative risk, 1.8 to 38; Fig 1). High blood ECP levels reduced the time to mild exacerbation, and increased the probability of a mild exacerbation (Fig 1 and Table 2). Conversely, sputum eosinophilia did not have a significant impact on time to mild exacerbation, although the time to exacerbation tended to be shorter in subjects with a sputum eosinophil count 5% (Fig 2). Subjects with a sputum ECP level 2 SDs higher than that reported for healthy subjects 12 were more likely to experience mild exacerbations (68% [14%] vs 33% [14%]) for subjects with levels within 2 SDs of healthy subject levels (Fig 2). FEV 1 and PC 20 did not influence mild exacerbation rates significantly (Fig 3 and Table 2). Multivariate Cox regression analysis, Table 1 Lung Function Sputum, and Blood Characteristics of All Subjects, Subjects With Exacerbation, and Subjects Without Exacerbation at the First Visit Variables All Subjects (n 31) Mean (SD) Range Subjects With Exacerbation* (n 13) Subjects Without Exacerbation* (n 18) p Value Age, yr 31 (14) (18) 30 (12) 0.42 History of asthma, yr 14 (8) (8) 15 (8) 0.71 Budesonide dose, g/d 841 (356) (366) 835 (348) ,600 Salmeterol dose, g/d 73 (83) (86) 59 (78) 0.32 FEV 1, L 2.66 (0.95) (0.84) 2.78 (0.97) 0.54 FEV 1, % of predicted value 79.9 (16) (12) 82.8 (19) 0.40 FEV 1 postbronchodilator, % 7.3 (10) (7.9) 8.3 (11) 0.57 PEF in the morning, L/min 385 (89) (87) 386 (93) 0.95 PC 20, mg/ml 2.6 (4.9) (2.58) 2.89 (6) 0.41 Variability of PEF, % 4.4 (3) (3.1) 4.3 (2.3) 0.95 Sputum eosinophils, 10 6 /mg 2.6 (7.2) (10.3) 1.62 (3.8) 0.59 Sputum eosinophils, % 13.2 (14) (16.6) 9.5 (12.3) 0.24 Sputum ECP, g/l 75 (85) (91.0) 47.7 (73) 0.10 Blood eosinophils, 10 9 /L 0.39 (0.21) (0.15) 0.34 (0.24) 0.04 Blood eosinophils, % 3.1 (3.9) (4.4) 2.2 (1.8) 0.05 Blood ECP, g/l 30 (28) (27.8) 25.7 (28.0) 0.08 *Data are presented as mean (SD). CHEST / 119 / 4/ APRIL,

4 Figure 1. Survival functions for absolute number of eosinophils (upper chart) and ECP (lower chart) in blood. Dotted lines represent absolute number of eosinophils /ml or 20 g/l for ECP level. Continuous lines represent /ml or 20 g/l, respectively. Circles represent censored subjects. controlling for the effect of treatment (steroid and long-acting -agonists), confirmed these results and selected blood eosinophils at baseline as the factor with the greatest impact. Discussion Nearly half of the subjects with well-controlled asthma in our study (according to well-established criteria) had at least one mild exacerbation within 12 months of enrollment, with a cumulative estimated exacerbation rate of 49% and mean time to exacerbation of 175 days for those experiencing a mild exacerbation. In addition, those who showed signs of active eosinophilic inflammation (eosinophils or ECP) at the beginning also proved to have higher risk of exacerbation. The aim of current guidelines for asthma treatment is to reach the control of symptoms, airflow limitation, and inflammation, although the last factor is not regularly assessed. 9 Patients with stable asthma are not expected to experience exacerbations when their conditions are well controlled and treatment has been optimized. However, our finding of a 49% cumulative probability of a mild exacerbation means that approximately one half of the subjects followed up for 1 year are likely to suffer at least one exacerbation. At the initial visit, the asthma of our subjects was apparently under control according to guidelines, 8,9 in spite of underlying eosinophilic inflammation that was ultimately associated with a risk of exacerbation. The eosinophilic inflammation reflected poorly controlled disease, and we may speculate that uncontrolled inflammation predisposed subjects to experience mild exacerbations. If we interpret the presence of inflammation to reflect undertreatment, our findings have clinical relevance, supporting the view that asthma treatment should be guided not only by symptoms and airflow limitation but also by eosinophilic inflammatory markers. 13 We found evidence of higher risk of mild exacerbation in subjects with signs of eosinophilic inflammation over the course of a year, but this effect was independent of the dose of budesonide as well as of the addition of salmeterol. We and others 7 assume that the effect of these medications on exacerbation is mediated by their capacity to modify inflammation. Symptoms, need of treatment, and airflow limitation are well-known markers of asthma severity. 6,14 16 We did not find a relevant role of these factors in our well-controlled subjects, because in these subjects they may lose their predictive value, and improved monitoring of inflammatory markers is likely to play a more important role in controlling disease. Similarly, airway responsiveness was not a significant risk factor for mild exacerbation in our study. However, Figure 2 shows different time-to-exacerbation curves for those subjects who had moderateto-severe hyperresponsiveness ( 0.4 mg/ml) compared with the others. Recent data 17 suggest a relationship between exacerbation and the degree of airway hyperresponsiveness because treating asthma with a sufficient dose of inhaled steroid reduces hyperresponsiveness and prevents exacerbations. Although it is possible that our study was underpowered to detect this effect, multivariate Cox regression analysis excluded airway hyperresponsiveness, suggesting its secondary role. Based on our current knowledge of asthma pathophysiology, it seems logical that an indirect marker of inflammation like bronchial hyperresponsiveness had less relevance than a direct one, such as the number of eosinophils Clinical Investigations

5 Table 2 Evaluation of Risk Factors* Risk Factors Subjects (Events), No. Rates (SE) Days Free of Mild Exacerbation (SE) p Value Log- Rank Test Relative Risk 95% CI of Relative Risk %FEV 1 80% 16 (8) 0.40 (0.14) 267 (34) vs vs vs % 15 (5) 0.62 (0.14) 316 (27) PC 20 4 mg/ml 14 (4) 0.62 (0.15) 292 (34) vs vs vs mg/ml 17 (9) 0.42 (0.13) 288 (30) Eosinophils (blood) /L 17 (11) 0.23 (0.11) 224 (34) vs vs vs /L 14 (2) 0.83 (0.11) 368 (5) Eosinophils (sputum) 5% 15 (8) 0.43 (0.14) 268 (33) vs vs vs % 16 (5) 0.60 (0.14) 316 (29) ECP (blood) 20 g/l 16 (9) 0.32 (0.14) 256 (34) vs vs vs g/l 15 (4) 0.69 (0.13) 324 (26) ECP (sputum) 40 g/l 15 (8) 0.33 (0.14) 258 (33) vs vs vs g/l 16 (5) 0.68 (0.14) 316 (29) *Significance level of p Accordingly, Sont et al 17 found that the improvement in airway hyperresponsiveness was not clearly paralleled by the improvement in bronchial biopsy findings. Moreover, Crimi et al 18 also found that nonspecific airway responsiveness was not a predictor of severity of exacerbation. Therefore, it remains unknown whether or not the reduction in exacerbation rates is because of a reduction in PC 20,orin airway inflammation, which indirectly reduces hyperresponsiveness. Subjects with blood eosinophil counts /L had a 5.4-times higher risk of exacerbation than subjects with /L. The relationship between peripheral blood eosinophils and asthma has been known for a long time, 19 although insufficient attention was paid to this observation until interest was revived by Baigelman et al 20 and Virchow et al. 21 They showed a good correlation between blood eosinophilia and FEV 1 after an exacerbation. In addition, Janson and Herala 22 suggested that eosinophilia in blood could be a prognostic factor for exacerbations in asthma. We think that eosinophilia in blood may reflect activation of bone marrow production, as can be inferred from the observations of other authors 23,24 expressing a particular predisposition of a subject to suffer exacerbation. Eosinophilia in induced sputum has been related to exacerbations 25 and is useful in guiding treatment of an exacerbation in severe asthma. 26 Sont et al, 27 following up asthmatic subjects for 2 years, suggested that the degree of airway inflammation can be Figure 2. Survival functions for percentage of eosinophils (upper chart) and ECP (lower chart) in sputum. Dotted lines represent percentage of eosinophils 5% or 40 g/l for ECP level. Continuous lines represent 5% or 40 g/l, respectively. Circles represent censored subjects. CHEST / 119 / 4/ APRIL,

6 trolled by higher inhaled corticosteroid doses. It is also possible that the etiology of an exacerbation plays a determinant role in the importance of sputum eosinophilia. We, like others, 28,29 found most exacerbations to be related to infection, which has been described as noneosinophilic in sputum. 26,28 Finally, our data reinforce that high ECP levels seem to be related to risk of exacerbations, 30 as several authors have previously described. We conclude that subjects with well-controlled asthma have a high risk of mild exacerbation, which is related to underlying eosinophilic inflammation present at the first visit (baseline). Our data indicate that subjects with active eosinophilic inflammation are more likely to experience a mild exacerbation and that, given the great variability in the natural course of the disease, close monitoring of eosinophilic inflammation may be warranted to evaluate the risk of an individual patient. Figure 3. Survival functions for percentage of FEV 1 of predictive value (upper chart) and PC 20 (lower chart). Dotted lines represent percentage of FEV 1 80% or 4 mg/ml for PC 20. Continuous lines represent 80% or 4 mg/ml, respectively. Circles represent censored subjects. a risk factor for future exacerbations. In our study, sputum eosinophil count was not significantly related to spontaneous exacerbation in the following year. This lack of association was unexpected. Two reasons might explain this finding: treatment with inhaled steroids, and the systemic effect of eosinophilic airway inflammation. The former relates to the fact that our patients were in stable condition and receiving an inhaled steroid treatment regularly. This may have partially controlled local eosinophilic inflammation, therefore hiding the association between eosinophil counts and exacerbation. The latter relates to the fact that sputum eosinophilia reflects local events, whereas other markers of eosinophilic inflammation (mainly in blood) might reflect a more systemic state of eosinophilic activation, resulting in long-term risk of exacerbation that could be con- References 1 Koller DY, Herouy Y, Gotz M, et al. Clinical value of monitoring eosinophil activity in asthma. Arch Dis Child 1995; 73: Pizzichini E, Pizzichini MM, Efthimiadis A, et al. Measuring airway inflammation in asthma: eosinophils and eosinophilic cationic protein in induced sputum compared with peripheral blood. J Allergy Clin Immunol 1997; 99: Turner MO, Noertjojo K, Vedal S, et al. Risk factors for near fatal asthma: a case-control study in hospitalized patients with asthma. Am J Respir Crit Care Med 1998; 157: Westerman DE, Benatar SR, Potgieter PD, et al. Identification of the high-risk asthmatic patient. Am J Med 1979; 66: Bellia V, Cibella F, Coppola P, et al. Variability of peak flow rate as a prognostic index in asymptomatic asthma. Respiration 1984; 46: Ellman MS, Viscoli CM, Sears MR, et al. A new index of prognostic severity for chronic asthma. Chest 1997; 112: Pauwels RA, Löfdahl CG, Postma DS, et al. Effect of inhaled formoterol and budesonide on exacerbation of asthma. N Engl J Med 1997; 337: Hargreave FE, Dolovich J, Newhouse MT. The assessment and treatment of asthma: a conference report. J Allergy Clin Immunol 1990; 85: International consensus report on diagnosis and treatment of asthma. Bethesda, MD: National Heart, Lung, and Blood Institute, National Institutes of Health. Publication No: , March Eur Respir J 1992; 5: Sterk PJ, Cockcroft DW, O Byrne PM, et al. Airway responsiveness: standardized challenge testing with pharmacologic, physiological and sensitizing stimuli in man. Eur Respir J 1993; 6: Pin I, Gibson PG, Kolendowicz R, et al. Use of induced sputum cell counts to investigate airway inflammation in asthma. Thorax 1992; 47: Belda J, Giner J, Casan P, et al. Induced sputum in asthma: study of validity and repeatability. Arch Bronconeumol 1997; 33: Hargreave FE. Induced sputum for the investigation of 1016 Clinical Investigations

7 airway inflammation: evidence for its clinical application. Can Respir J 1999; 6: Ogilvie AG. Asthma: a study in prognosis of 1000 patients. Thorax 1962; 17: McFadden ERJ, Warren EL. Observations on asthma mortality. Ann Intern Med 1997; 127: Juniper EF, Kline PA, Vanzieleghem MA, et al. Long-term effects of budesonide on airway responsiveness and clinical asthma severity in inhaled steroid-dependent asthmatics. Eur Respir J 1990; 3: Sont JK, Willems LN, Bel EH, et al. Clinical control and histopathologic outcome of asthma when using airway hyperresponsiveness as an additional guide to long-term treatment: the AMPUL Study Group. Am J Respir Crit Care Med 1999; 159: Crimi E, Balbo A, Voltolini S, et al. Is the allergen-inhalation challenge predictive of the severity of seasonal asthmatic exacerbations? Allergy 1993; 48: Horn BR, Robin EDTJ, Van Kessel A. Total eosinophil counts in the management of bronchial asthma. N Engl J Med 1975; 292: Baigelman W, Chodosh S, Pizzuto D, et al. Sputum and blood eosinophils during corticosteroid treatment of acute exacerbations of asthma. Am J Med 1983; 75: Virchow JCJ, Oehling A, Boer L, et al. Pulmonary function, activated T cells, peripheral blood eosinophilia, and serum activity for eosinophil survival in vitro: a longitudinal study in bronchial asthma. J Allergy Clin Immunol 1994; 94: Janson C, Herala M. Blood eosinophil count as risk factor for relapse in acute asthma. Respir Med 1992; 86: Wood LJ, Inman MD, Watson RM, et al. Changes in bone marrow inflammatory cell progenitors after inhaled allergen in asthmatic subjects. Am J Respir Crit Care Med 1998; 157: Gibson PG, Dolovich J, Girgis-Gabardo A, et al. The inflammatory response in asthma exacerbation: changes in circulating eosinophils, basophils and their progenitors. Clin Exp Allergy 1990; 20: Fabbri LM, Saetta M, Picotti G, et al. Late asthmatic reactions, airway inflammation and chronic asthma in toluene-diisocyanate-sensitized subjects. Respiration 1991; 58: Pizzichini MM, Pizzichini E, Clelland L, et al. Sputum in severe exacerbations of asthma: kinetics of inflammatory indices after prednisone treatment. Am J Respir Crit Care Med 1997; 155: Sont JK, Kluin-Nelemans JC, Willems LN, et al. Exacerbation in patients with atopic asthma: is the severity of airways inflammation a risk factor [abstract]? Am J Respir Crit Care Med 1998; 157:A Fahy JV, Kim KW, Liu J, et al. Prominent neutrophilic inflammation in sputum from subjects with asthma exacerbation. J Allergy Clin Immunol 1995; 95: Chodosh S. Acute bacterial exacerbations in bronchitis and asthma. Am J Med 1987; 82: Ahlstedt S. Clinical application of eosinophilic cationic protein in asthma. Allergy Proc 1995; 16: Wever AM, Wever-Hess J, Hensgens HE, et al. Serum eosinophilic cationic protein (ECP) in chronic asthma: relationship to spirometry, flow-volume curves, PC 20, and exacerbations. Respir Med 1994; 88: Wever AM, Wever-Hess J, Hermans J. The use of serum eosinophil cationic protein (ECP) in the management of steroid therapy in chronic asthma. Clin Exp Allergy 1997; 27: Pedersen B, Dahl R, Karlstrom R, et al. Eosinophil and neutrophil activity in asthma in a one-year trial with inhaled budesonide: the impact of smoking. Am J Respir Crit Care Med 1996; 153: CHEST / 119 / 4/ APRIL,