INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Airway inflammation is important in asthma because it is considered to be the primary cause of symptoms, physiologic abnormalities, and exacerbations. Different causes of inflammation lead to distinct types of inflammatory events, which in turn result in different clinical and physiologic effects. Upper respiratory viral infections are the major cause of common colds and have been identified, by epidemiologic cohort studies (1, 2), to be commonly associated with asthma exacerbations. It has been proposed that acute respiratory viral infections cause asthma exacerbations by damaging the airway epithelium and changing the airway cytokine and cellular profiles (3). Only two reports (4, 5) have documented the effects of common colds on the lower airways of asthmatics. Neither study, however, produced an asthma exacerbation.

Experimental colds have been preferred to study the airway inflammatory response to acute viral infections because the time of infection, the amount of the inoculum, the virus serotype, and, potentially, the pathogenicity level can be controlled. However, the study of natural colds may provide insights into the effects of a variety of viral infections. The use of induced sputum examination provides a noninvasive method to follow the inflammatory events.

In this study we have used induced sputum to investigate indices of airway inflammation in asthmatic and healthy subjects with a natural cold at the time of the cold and during convalescence. Our main objectives were: (1) to evaluate the feasibility of using this model to study the inflammatory response to acute respiratory virus infections, and (2) to determine the pattern of inflammatory cells and fluid-phase markers in the sputum.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Eight asthmatics and nine healthy adults with, at most, 4 d of symptoms of a cold were simultaneously recruited from the Firestone Regional Chest and Allergy Clinic, other physicians, and advertisement between December 1995 and April 1996 (Table 1). The type and severity of cold symptoms (nasal congestion, discharge or obstruction, sore throat, cough, fever, and headache) was similar in asthmatic and healthy subjects. Subjects with asthma had a history of asthma symptoms and a previous physician diagnosis. The asthma was objectively confirmed (within the past year) in five subjects who had an FEV₁/VC of  70% by methacholine airway hyperresponsiveness (PC₂₀ < 4 mg/ml) and in the remainder who had an FEV₁/VC of < 70%, by an improvement in FEV₁ from predicted of  15% or  0.2 L after salbutamol (200 µg). The asthma was exacerbated in six asthmatics subjects as defined by: (1) a history of recent increase in symptoms (dyspnea, n = 6; wheeze, n = 5; and chest tightness, n = 4), (2) nocturnal symptoms disturbing sleep (n = 5), and/or (3) an increased need for an inhaled short-acting ₂-agonist (n = 6), and (4) an FEV₁ after bronchodilator of < 80% of predicted or previous best (n = 6). All were receiving treatment with inhaled ₂-agonist when needed, and six were receiving inhaled steroid (dose  1,600 µg/d budesonide or equivalent). None had a prior history of colds or respiratory infection or asthma exacerbation for at least 4 wk. Healthy subjects had no history of asthma or other respiratory diseases or symptoms than those caused by the cold, an FEV₁ > 80% predicted, an FEV₁/VC > 80%, and normal methacholine airway responsiveness (PC₂₀ > 16 mg/ml). The study was approved by the hospital research committee, and all subjects gave written informed consent.

Study Design

This was a descriptive cohort study. The subject selection was based only on the severity of the cold score; the subjects selected had a cold symptoms score of severity  7. They were seen at the time of the cold (Day 4) and during recovery (Day 21). On Day 4 subject characteristics were documented, a questionnaire of cold symptoms was applied, reversibility of spirometry and/or methacholine inhalation test was performed, and induced sputum was collected for inflammatory indices. A nasal pharyngeal swab was obtained for viral identification on the day of presentation (n = 4) or on Day 4. The subjects were instructed to maintain their current medications or to comply with any changes made by their physicians. At Day 21 the questionnaire of cold symptoms was applied, spirometry was performed, and sputum was obtained for inflammatory indices. In addition, peripheral blood was collected on both days for viral serology. The main outcomes were sputum total (TCC) and differential cell counts, fluid phase IL-2, IL-5, IL-8, interferon gamma (IFN-), fibrinogen, albumin, and ECP. Secondary outcomes were cold symptoms score and FEV₁. Sputum and serum albumin for calculation of sputum/serum were measured only on Day 4. To avoid bias, measurements in sputum were performed blind to the clinical characteristics and vice versa.

Clinical Methods

Subject characteristics were documented by a structured questionnaire. The presence of a cold was established by using a standard cold questionnaire (6). The severity of the cold was based on at least two or more symptoms of nasal discharge, sneezing, nasal congestion, sore throat, cough, headache, malaise, chills, and/or fever. The severity of each symptom was assessed by a severity scale: 0 = absent, 1 = mild, 2 = moderately severe, and 3 = severe (7). At least one of the symptoms had to be classified as 2 on the severity scale. Spirometry, methacholine inhalation tests, and allergy skin prick tests were performed by standard methods (8). Sputum was induced by the inhalation of an aerosol of hypertonic saline as previously described (12); in asthmatics with more severe airflow limitation a modified procedure was used (13).

Virus Detection Methods

The nasal pharyngeal swabs were stored at 70° C until screened by PCR for the following virus types: rhinoviruses, coronaviruses 229E and OC43, influenza virus types A and B, parainfluenza virus 1-3, adenoviruses, and RSV, and for atypical bacteria (Chlamydia and Mycoplasma pneumoniae). The PCRs used were reverse transcriptase PCRs to detect either genomic or mRNA specific for the virus or atypical bacteria. The assays included single round or nested amplification methods ± internal probe hybridization as described (14). Nasal pharyngeal swabs were also screened by standard virologic techniques of immunofluorescent staining. Paired serology for influenza virus types A and B, parainfluenza virus 1-3, adenoviruses, and RSV were performed.

Sputum and Blood Examination

Sputum selected from the expectorate of sputum plus saliva was processed as previously described (15). The success of the selection was examined by the proportion of squamous epithelial cells of  5%. The concentration of IL-8, IL-2, IL-5, and IFN- in the thawed supernatant were determined by quantitative "sandwich" enzyme immunoassay (Quantikine; R&D Systems, Inc., Minneapolis, MN). ECP (µg/L) was measured by a sensitive radioimmunoassay (RIA; Kabi Pharmacia Diagnostics AB, Uppsala, Sweden) and fibrinogen by a "sandwich" ELISA assay using a rabbit antihuman fibrinogen antibody (A080; Dako Ltd, High Wycome, UK). The limit of detection for the fluid-phase IL-8, IL-2, IL-5, IL-8, IFN-, ECP, and fibrinogen were 25.6, 38.4, 7.8, and 15.6 pg/mL and 2.0, and 0.79 µg/L, respectively. Venous blood was collected into a 5.0 ml EDTA (K₃ Vacutainer; BD, Rutherford, NJ). Sputum and serum albumin were determined by standard procedures.

Data Analysis

Results are reported as median and interquartile range (IQR) unless otherwise specified. All statistical tests were two-sided, and significance was accepted at the level of 95%. Dependent variables with a non-normal distribution (PC₂₀ methacholine, total cell counts, the percentage and the absolute numbers of eosinophils and lymphocytes, the absolute numbers of neutrophils, ECP, IL-8, and fibrinogen) were log transformed before analysis. Within-group differences on clinical and inflammatory outcomes on Days 4 and 21 were examined by paired t tests. Unpaired t tests were used for between-group comparisons; p values were adjusted for multiple comparisons. The correlations between variables were examined by Spearman's rank correlation coefficient (rs). To reduce the possibility of correlations occurring by chance, only those correlations with rs values above 0.50 that were significant at the 0.01 level were considered relevant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Six of the eight asthmatic and six of the nine healthy subjects had a confirmed viral infection (Table 1). The viruses differed in the two groups, being only influenza in the asthmatics and mainly other viruses in the healthy subjects. Two of the healthy subjects had dual infections (rhinovirus plus coronavirus and rhinovirus plus influenza A). We examined clinical parameters (cold score symptoms and spirometry) and inflammatory indices in sputum (sputum appearance, total and differential cell count, and fluid-phase markers) on Days 4 and 21 in the 12 subjects with confirmed viral infection (Table 2) and in the five subjects without proven viral infection (Table 3).

Clinical and Sputum Indices in Subjects with Confirmed Viral Infection

Clinically, five of the six asthmatics had an exacerbation with a mean (SD) postbronchodilator FEV₁ of 1.4 (0.6) L and of 52 (20)% predicted, and four of them had their baseline treatment changed after measurements on Day 4 (one increased the inhaled steroid dose, one increased the inhaled steroid dose plus added an antibiotic for 7 d, and two added prednisone 30 mg/d for 7 d). The severity of the cold score symptoms on Day 4 was similar in asthmatic and healthy subjects. On Day 21 the cold symptoms had improved significantly in both groups. The FEV₁ improved in asthmatics but remained normal and unchanged in healthy subjects.

The appearance of the sputum on both days was mucoid and the cell viability was good. On Day 4 the sputum inflammatory response was greater in asthmatic than in healthy subjects, but it was similar in characteristics (Table 2 and Figures 1 and 2). Only asthmatics showed a marked increase in total cell count. Both groups had a modest neutrophilia and increases in fluid phase levels of IL-8, ECP, fibrinogen, and albumin. The fibrinogen levels were particularly high in the asthmatics. The proportion and the absolute number of eosinophils were low (Table 2 and Figure 1). IL-2 levels were within the normal range, and IL-5 and IFN- were under the limit of detection of the assays. Sputum IL-8 on Day 4 was strongly correlated with neutrophils (rs = 0.8, p < 0.001) (Figure 3). This correlation was also significant when each group was analyzed separately (rs = 0.8, p < 0.01 and rs = 0.9, p = 0.005 in asthmatics and healthy subjects, respectively). On Day 21 there was improvement in neutrophilia (p = 0.001) and in fibrinogen (p = 0.01) and ECP levels (p = 0.07) in asthmatic subjects. These parameters also improved in healthy subjects, but the differences did not reach significance. Most measurements were still abnormal in both groups.

View larger version (23K): [in this window] [in a new window]   Figure 1.   Individual values of sputum eosinophils and neutrophils (percentage and absolute numbers) from asthmatic and healthy subjects with proven viral infection on Day 4 and on Day 21. Closed circles represent asthmatic subjects, lighter circles represent healthy subjects, and the open star represents the healthy subject with dual viral infection (Influenza A + rhinovirus); *p = 0.05, **p  0.01. Horizontal bars are median values. Dashed line represents the median value for healthy subjects without a cold. At the time of the cold, both groups have a modest sputum neutrophilia, which improved at convalescence.

View larger version (14K): [in this window] [in a new window]   Figure 2.   Individual values of fluid-phase fibrinogen and ECP from asthmatic and healthy subjects with proven viral infection. Symbols and lines as in Figure 1. The sputum on Day 4 showed very high levels of fibrinogen, which improved significantly at convalescence.

View larger version (16K): [in this window] [in a new window]   Figure 3.   Correlations between sputum IL-8 and neutrophils, fibrinogen, eosinophils, and ECP in subjects with proven viral infection. Symbols as in Figure 1. rs = Spearman's rank correlation coefficient. Solid line is the calculated regression line for linear regression. Strong and highly significant correlations were observed between IL-8 and neutrophils, fibrinogen, and ECP. The correlation between IL-8 and eosinophils was weaker.

In view of the unusually high fibrinogen levels on Day 4, we also examined the relationship between sputum albumin and fibrinogen. In the asthmatic group sputum albumin was increased by 2.6-fold in contrast with an increase in fibrinogen levels by 625-fold. In healthy subjects sputum albumin levels were within the normal range compared with the 16.4-fold increase in the fibrinogen levels.

Clinical and Sputum Indices in Subjects without Confirmed Viral Infection

The clinical and inflammatory indices were within the range of those seen in subjects with confirmed viral infection (Table 3). One of the two asthmatic subjects had an exacerbation but did not receive extra treatment. The other was treated after Day 4 (with an antibiotic).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has demonstrated the feasibility of using induced sputum to investigate the effects of natural colds and influenza on the airways of the lungs, as identified by its ability to distinguish acute and convalescent airway inflammatory changes in asthmatic and in healthy subjects. It also indicates that this model is suited to examine asthmatics with or without an asthma exacerbation caused by the cold, although it is limited by the heterogeneity of the recovered viruses and by the difficulty to prove a viral infection in all subjects who met the criteria for a naturally acquired cold. All viruses involved were associated with lower airway inflammation in both asthmatic and healthy subjects and the inflammatory response was predominantly neutrophilic. The inflammatory changes were most marked in asthmatics when trivial sputum eosinophilia was present and when only influenza viruses were identified as opposed to other viruses in the healthy subjects. Although the effects of the inflammation caused an exacerbation with mild to moderate airflow limitation in six of the eight asthmatic subjects, it caused only cold symptoms in healthy subjects without any change in FEV₁. The results of this study are relevant because they illustrate that the examination of sputum in naturally acquired colds can be used to generate hypotheses of virus-induced exacerbations of asthma and, perhaps, to the most appropriate treatment of these.

This feasibility study has a number of strengths and weaknesses. The strengths include the subject selection, which was based only on the severity of the cold score, the confirmation of virus infection in 76.5% of subjects, a similar time of examination between subjects (4 d), and the use of a reliable, valid (15), and responsive (13) method for the examination of inflammatory indices in induced sputum. The weaknesses are mostly a result of the use of naturally acquired colds. There were a number of different types of virus involved and different types of virus between healthy and asthmatic subjects. The reason for the latter is probably chance. The results, therefore, need to be interpreted cautiously. It is possible that the more severe involvement in asthmatics was due to the infection with influenza, which only occurred in one of the healthy subjects. In this person, the changes in sputum indices tended to be quantitatively higher than in most of the other healthy subjects with an increase in albumin of 3.6-fold and in fibrinogen of 36-fold.

Control nasal pharyngeal swabs were not a part of this feasibility study. Although the mere presence of virus genome by PCR does not indicate a casual link with the cold, the same can be said for most virus detection methods. Given that many of our primer sequences are complementary to viral mRNA sequences and that the bioactivity of most mRNA is short-lived, the presence of viral mRNA detected by PCR implies active viral transcription at the time of sampling. In addition, the results are unlikely to be false positive for several reasons, including the adoption of sensitive and specific PCR assays with high positive and negative predictive values, the use of standard protocols to avoid carryover contamination, and the always negative blank noninfected controls (14). Finally, all samples were tested in duplicate and the results were consistent.

This is the first study to investigate the effect of natural respiratory virus infections on lower airway inflammatory indices of asthmatics. Previously, only experimental colds with rhinovirus have been investigated, and these have not been associated with asthma exacerbations (4, 5). Fraenkel and colleagues (4) found that at the time (Day 4) of an experimental rhinovirus cold in six atopic asthmatics and in 11 healthy subjects, there was an infiltration of bronchial mucosa with lymphocytes and activated eosinophils; after recovery (6 to 10 wk after the cold) the airway eosinophilia persisted only in the asthmatics. Recently, Grünberg and colleagues (5), examining induced sputum of 19 atopic asthmatics with an experimental rhinovirus cold, reported no changes in the proportion of different inflammatory cells, whereas there was an increase in the ECP, IL-6, and IL-8 levels. Our results differ from these reports by showing that the cellular airway inflammatory response in asthmatics and in healthy subjects was neutrophilic, as demonstrated by a modest increase in the proportion and in the absolute number of neutrophils. These differences may be due to different viruses and/or to differences in the severity of the inflammatory response.

The effects of rhinovirus (16, 17) and coronavirus (18) colds on the cellular inflammatory response have been studied in the nose and, like our findings in the airways of the lungs, neutrophilic inflammation was also observed (16, 17). Another study (19), examining changes caused by naturally acquired colds in bronchial mucosa from atopic and nonatopic healthy subjects showed that in comparison with baseline, only those subjects with a virologic proved cold (40%) had a relative rise in bronchial mucosal neutrophils. In the present study, both subjects with proven and nonproven viral infection had similar sputum findings. The lack of differences in the inflammatory indices between the two groups may be due to the small number of subjects in whom virus infection was not proven.

Interleukin-8 is a chemoattractant for neutrophils produced by alveolar macrophages, lymphocytes, epithelial cells, and neutrophils (20). Evidence for the participation of IL-8 in virus infections has been obtained from in vitro studies using both epithelial cell lines and bronchial epithelial cells infected with influenza A (21), RSV (22), or rhinovirus (23). However, it is unlikely that these cell lines are an exclusive source of IL-8 in acute respiratory viral infections. It could be speculated that the initial production of IL-8 during an acute respiratory virus infection is from epithelial cells and that once the recruitment of neutrophils has begun, in response to the epithelial IL-8 production, a great deal of IL-8 production can be provided by neutrophils. In fact, Grünberg and colleagues (5), using intracellular staining to investigate the origin of IL-8 in the sputum of 19 asthmatics with experimental rhinovirus colds, demonstrated that most of the cells positively stained were neutrophils. Furthermore, IL-8 has been shown to be associated with increased numbers of neutrophils and/or with levels of myeloperoxidase in nasal aspirates from asthmatic children with a naturally acquired cold (17) and in nasal lavage from adults with asthma and a rhinovirus experimental cold (24). Our results support and extend these previous findings by showing that the elevated levels of IL-8 in induced sputum of the subjects with a cold were highly correlated with neutrophilia.

The sputum neutrophilia could also be related to a Th1 response, but we could not detect IFN- or increased levels of IL-2. IFN- might not be detected because it peaks at Day 2 (25). An alternative explanation for the lack of detection of IFN- is methodological; it has not been investigated in sputum in other studies. IL-2 (26, 27), but not IL-4 (27), production has been identified in in vitro studies using antigen-stimulated blastogenesis during experimental rhinovirus infection (26) or after stimulation of T cells with purified human rhinovirus (27); this is consistent with a predominant Th1 response during virus infections.

The elevated levels of ECP in both asthmatic and healthy subjects might be indicative of an increased Th2 response. However, the proportion of eosinophils was not increased, IL-5 could not be detected, and ECP does not seem to be specific for eosinophilic activation since it is also found in neutrophilic inflammation (28). This inability to demonstrate IL-5 could be a result of the dithiothreitol treatment of the sputum. However, levels have been measured using the same methods in uncontrolled and exacerbated asthma (which was present in six of the eight asthmatics in this study) (13). Therefore, there is no good evidence in this study that the virus infection turned on the Th2 pathway either. It is possible, however, that the baseline Th2 response influenced and magnified the effects of the Th1 response or a nonspecific T-cell recruitment (31).

Sputum fibrinogen was increased in both healthy and asthmatic subjects, but this increase was extraordinary in the asthmatic subjects. In both healthy and asthmatic subjects the increase of fibrinogen was quite different to the increases in albumin, suggesting that local production of fibrinogen rather than just plasma exudation was occurring. Marked increases in fibrinogen have also been observed in experimental coronavirus infection in the nose of healthy subjects with rhinitis (32).

Sputum examination was also performed in convalescence. In the healthy subjects there was a decrease in the total number of neutrophils and in the levels of ECP and fibrinogen, although the latter was still raised above the normal range. In asthmatic subjects there was also a significant improvement, but this may have been influenced by the extra treatment that was given to six subjects. The course of inflammatory and clinical effects can be studied in the future.

We conclude that this study supports the feasibility of using induced sputum examination to study the inflammatory effects of natural virus infections. The results illustrate the importance of doing this. They suggest that exacerbations of asthma are associated with a neutrophilic inflammatory process, which may be magnified in asthmatic subjects because of an associated eosinophilic process. This may have implications for treatment since accumulating evidence suggests that glucocorticoid treatment improves eosinophilic inflammation but not noneosinophilic inflammation (33). The lack of effect on neutrophilic inflammation is supported by the results of Doull and coworkers (34) who observed that regular inhaled steroid treatment "offers no clinically significant benefit in school-aged children with wheezing episodes associated with viral infection." Further studies are required to validate these interpretations.