American Journal of Respiratory and Critical Care Medicine

Rationale: Severe exacerbations of chronic obstructive pulmonary disease (COPD) are major causes of health care costs mostly related to hospitalization. The role of infections in COPD exacerbations is controversial.

Objectives: We investigated whether COPD exacerbations requiring hospitalization are associated with viral and/or bacterial infection and evaluated relationships among infection, exacerbation severity, assessed by reduction of FEV1, and specific patterns of airway inflammation.

Methods: We examined 64 patients with COPD when hospitalized for exacerbations, and when in stable convalescence. We measured lung function, blood gases, and exhaled nitric oxide, and examined sputum for inflammation and for viral and bacterial infection.

Results: Exacerbations were associated with impaired lung function (p < 0.01) and increased sputum neutrophilia (p < 0.001). Viral and/or bacterial infection was detected in 78% of exacerbations: viruses in 48.4% (6.2% when stable, p < 0.001) and bacteria in 54.7% (37.5% when stable, p = 0.08). Patients with infectious exacerbations (29.7% bacterial, 23.4% viral, 25% viral/bacterial coinfection) had longer hospitalizations (p < 0.02) and greater impairment of several measures of lung function (all p < 0.05) than those with noninfectious exacerbations. Patients with exacerbations with coinfection had more marked lung function impairment (p < 0.02) and longer hospitalizations (p = 0.001). Sputum neutrophils were increased in all exacerbations (p < 0.001) and were related to their severity (p < 0.001), independently of the association with viral or bacterial infections; sputum eosinophils were increased during (p < 0.001) virus-associated exacerbations.

Conclusions: Respiratory infections are associated with the majority of COPD exacerbations and their severity, especially those with viral/bacterial coinfection. Airway neutrophilia is related to exacerbation severity regardless of viral and/or bacterial infections. Eosinophilia is a good predictor of viral exacerbations.

Chronic obstructive pulmonary disease (COPD) is the fourth leading cause of mortality worldwide (1). The estimated annual costs of COPD are $24 billion, and 70% are related to exacerbations requiring hospitalization (2). In addition to their enormous acute morbidity, mortality, and cost, exacerbations are also associated with major reductions in long-term quality of life and lung function (3, 4). Despite the clinical and economic importance of severe COPD exacerbations, their etiology and mechanisms are poorly understood. Independently, both bacterial (5, 6) and viral infections (7, 8) have been detected at increased frequencies during exacerbations, and acquisition of new bacterial strains is associated with increased risk of exacerbation (9). However, the importance of infection overall, and the relative importance of viral versus bacterial infections, to the etiology of COPD exacerbations is unknown, as no study has comprehensively investigated both bacteria and viruses during the same exacerbations.

Antibiotic therapy is already widely, but not always appropriately, used and antiviral therapy for influenza viruses is available, whereas antirhinoviral agents are in clinical development. Determining the etiology of exacerbations will inform appropriate antibiotic and antiviral therapy.

COPD exacerbations are associated with increased airway inflammation; however, relationships between airway inflammation, etiology, and exacerbation severity have not been established. There are also conflicting data on the nature of the inflammation. Some studies report increased airway eosinophilia (1013), whereas others document increased markers of neutrophilic inflammation (1416) and increased numbers of neutrophils in bronchial biopsies (17). No study has so far investigated whether the type of infection (virus or/and bacteria) influences the inflammatory profile. If relationships with exacerbation severity and etiology were established, such information would be relevant to guide therapeutic intervention in COPD exacerbations.

Given the clinical and economic burden of severe COPD exacerbations, this study examined only patients with severe exacerbations admitted to the Department of Pulmonary Medicine of our hospital. We have therefore performed a prospective controlled study investigating viral and bacterial detection during severe COPD exacerbations requiring hospitalization and in stable convalescence and have examined relationships among etiology, airway inflammation, and exacerbation severity.

Study Design

Patients were recruited from a cohort of patients with COPD at the University of Ferrara, Italy. COPD was defined according to guidelines (Global Initiative for Chronic Obstructive Lung Disease, GOLD) (1), with a post-bronchodilator FEV1/FVC ratio of less than 70%.

From September 2000 to August 2002, patients were instructed to contact the university's research center on the appearance of signs/symptoms of acute exacerbations. Exacerbation was defined as increased dyspnea, cough, or sputum expectoration (quality or quantity) that led the subject to seek medical attention (1). A clinician saw patients within 24 h to confirm the diagnosis (1) via medical history and physical examination, and to perform blood gas analysis and administer oxygen as required. After initial treatment with inhaled bronchodilators, when clinical condition permitted, pulmonary function was assessed and nitric oxide (NO) measurements and peripheral blood and sputum samples were obtained. Only patients requiring hospitalization (1) and who, at the time of recruitment and all biological sampling, had not received any antibiotic or systemic glucocorticoid therapy were enrolled.

Criteria for a patient's discharge from hospital were adopted according to international guidelines (1) and included clinical stability and normal blood gases for 24 h. Patients were seen in convalescence, when clinically stable, 8 to 10 wk later, and all clinical assessments and sampling were repeated.

All patients remained on their regular treatment for the duration of the study. After recruitment and sampling, exacerbations were treated with standardized therapy (systemic glucocorticoids and antibiotics) and additional bronchodilators (1). Treating clinicians were blinded to results of investigations.

In all patients, chest radiographs were performed during exacerbations to exclude concomitant pneumonia. Also, all patients at some stage of the study underwent high-resolution computed tomography (HRCT) scan. Patients with evidence of bronchiectasis based on standard HRCT scan criteria (18, 19) were excluded. (Additional details on methods are provided in the online supplement.)

Pulmonary Function and Blood Analyses

Pulmonary function tests were performed with previously described methods (20, 21), and blood taken for differential white cell count by standard Coulter counting.

Sputum Collection and Analysis

Sputum was induced and examined as previously described (20). Total neutrophil elastase (NE) and eosinophil cationic protein (ECP) in sputum supernatants were measured by commercial ELISA and fluoroenzyme immunoassay UniCAP ECP, respectively (see the online supplement for details).

Quantitative Bacteriology

Sputum samples were processed using sputolysin; serial dilutions were made and cultured on appropriate media for identification by standard methods (22). Colony-forming units (cfu)/ml were calculated.

Respiratory Virus and Atypical Bacteria Detection

A panel of reverse transcriptase–polymerase chain reactions was used to screen sputum samples for the commonest respiratory viruses and for Chlamydia pneumoniae and Mycoplasma pneumoniae (2325). (See the online supplement for details.)

Statistical Analysis

Normally distributed data are reported as means ± SE and skewed by medians. All exacerbation/convalescent comparisons were performed by parametric (paired Student's t) or nonparametric (Wilcoxon signed rank) tests. Comparisons of continuous variables among subgroups were made using analysis of variance (ANOVA) and, if significant, Student's t tests or Kruskal-Wallis and Mann-Whitney U tests, where appropriate. Bonferroni's correction was applied for multiple comparisons. Categoric binary variables were compared using χ2 tests. Correlation coefficients were calculated using Spearman's rank method. All tests were two-sided; type I error was 0.05.

Receiver operating characteristic (ROC) curve analysis was performed for sputum eosinophil counts at exacerbation and change from convalescence to define the predictive value of eosinophil numbers for exacerbations associated with viral infection. The area under the ROC curves was determined, and a value above 0.80 was considered good discrimination (20, 26).

Seventy-five patients fulfilled inclusion criteria; 69 agreed to participate and sputum was obtained from 65. One patient was rehospitalized for COPD exacerbation during the study and withdrew. Sixty-four patients completed both exacerbation and stable convalescence visits.

Patient characteristics are summarized in Table 1. The cause of hospitalization was acute respiratory failure (PaO2 < 60 mm Hg) in 58 of 64 patients; in six markedly intensified symptoms, no patients required admission to intensive care, had chronic respiratory failure, or were on long-term oxygen therapy when stable. All were receiving regular inhaled long-acting β2-agonists, 97% of patients were receiving regular inhaled glucocorticoids (mean beclomethasone equivalent dosage, 980 ± 70 μg/d) and 10 patients (15.6%) were receiving oral theophyllines.


Patient Characteristics (n = 64)

Age, yr70.6 ± 2.5
Male/female, n56/8
Smoking history, n61 ex smokers 3 current smokers
 Pack/yr48.3 ± 5.7
Chronic bronchitis/no chronic bronchitis, n43/21
Pulmonary Function
p Value
 FEV1, L0.96 ± 0.051.18 ± 0.07< 0.001
 FEV1, % pred39.4 ± 2.249.5 ± 2.3< 0.01
 FEV1/FVC, %41.7 ± 1.3248.6 ± 1.6< 0.01
 RV, % pred157.1 ± 6.3131 ± 4.6< 0.05
 TLC, % pred111.9 ± 8.3108.2 ± 5.1NS
 Kco, % pred45.23 ± 2.658.7 ± 2.9< 0.001
 NO, ppb15.18 ± 1.8510.32 ± 1.6< 0.05
 Blood gases
  PaO2′, mm Hg54.7 ± 1.569.3 ± 1.4< 0.001
  PaCO2′, mm Hg43.58 ± 1.1240.93 ± 0.970.07
  pH7.388 ± 0.0077.403 ± 0.005< 0.001
Sputum Cell TypeCell Number (× 106/g)
 Neutrophils26.7 (7.7–32.9)9.5 (3.8–18.9)< 0.001
 Macrophages3.6 (0.7–4.2)2.4 (0.5–4.1)NS
 Eosinophils1.65 (0–3.0)1.01 (0–2.1)NS
0.11 (0–0.16)
0.04 (0–0.06)

Definition of abbreviations: NO = exhaled nitric oxide; NS = not significant; RV = residual volume; TLC = total lung capacity. Values are mean ± SE.

Symptoms characterizing the exacerbations were as follows: increased breathlessness in 61 patients (95.3%), increased cough and sputum production in 48 (75%), sputum color change in 33 (51.5%), and fever in 23 patients (35.9%).

In the convalescence visit, patients returned to preexacerbation baseline lung function values (obtained within 3 mo [mean, 51.2 ± 3.2 d] before severe COPD exacerbation; Table E1 of the online supplement; changes from baseline ⩽ 3%, p > 0.3 for all parameters) and to baseline requirements for as-needed medication (1.5 ± 0.9 puffs/d at baseline vs. 1.3 ± 1.1 puffs/d at convalescence, p = 0.5). This was true also within the different subgroups analyzed below.

Changes in Lung Function and Sputum and Blood Inflammatory Cells at Exacerbation

Severe COPD exacerbations were associated with significantly lower FEV1, % predicted FEV1, FEV1/FVC, Kco, PaO2 and arterial pH, and a significantly higher residual volume compared with stable convalescence (Table 1). Paired exhaled NO measurements were obtained in 39 patients; levels were also significantly higher during exacerbations (Table 1).

Subjects with and without chronic bronchitis underwent the same protocol of sputum induction at exacerbation and convalescence. Sputum cell counts were performed in samples from all 64 patients both at exacerbation and in stable convalescence

Sputum neutrophils were markedly increased during exacerbations (p < 0.001; Table 1). A significant relationship between neutrophil infiltration and exacerbation severity was demonstrated by significant correlations between severity of reductions in lung function (% predicted FEV1) at exacerbation and both increase in sputum neutrophils (Figure 1; r = 0.325, p < 0.01) and absolute numbers of sputum neutrophils at exacerbations (r = 0.445, p < 0.001). In addition, patients with more severe hypoxia (reductions in PaO2 of ⩾ 10 mm Hg, n = 39) had greater airway neutrophilia (19.8 [5.2–33.6] cells × 106/g) versus those with reductions in PaO2 of less than 10 mm Hg (n = 25, 7.5 cells × 106/g [1–17.4], p < 0.05). At convalescence, a relationship was found between neutrophil numbers and the absolute FEV1 (r = 0.40, p = 0.001).

Peripheral blood neutrophils were also significantly higher during exacerbations (9.9 ± 0.5 vs. 4.9 ± 0.3 × 103/dl, p < 0.001). A significant relationship between peripheral blood neutrophil responses and exacerbation severity (reduction in % predicted FEV1) was also demonstrated (r = 0.5518, p < 0.001; Figure 2).

Sputum NE was increased at exacerbation (3.59 [1.12–5.54] μg/ml supernatant) as compared with stable convalescence (0.64 [0.29–1.1] μ/ml supernatant, p < 0.001). Sputum NE at exacerbation correlated with sputum neutrophil counts at exacerbation (r = 0.62, p < 0.001) and at convalescence with counts in stable convalescence (r = 0.56, p < 0.001).

Detection of Viruses and Bacteria in Exacerbation and Stable Convalescence Sputum Samples

Bacterial and viral analyses were performed in samples from all 64 patients both at exacerbation and in stable convalescence.

Of the 64 samples analyzed, respiratory viruses were detected in 31 (48.4%) sputum samples during exacerbations: 17 rhinoviruses, seven influenza viruses, four respiratory syncytial viruses (RSV), two parainfluenza viruses, two coronaviruses, three human metapneumoviruses (HMPV; three HMPV + rhinovirus, one RSV + rhinovirus); and respiratory viruses were detected in four (6.25%) sputum samples in stable convalescence: two rhinoviruses and two RSV (p < 0.001 vs. Exacerbations).

Of the 64 samples analyzed, positive bacterial cultures were obtained from 35 (54.7%) sputum samples during exacerbations: nine Haemophilus influenzae, eight Streptococcus pneumoniae, seven Moraxella catarrhalis; four Staphylococcus aureus; four Pseudomonas aeruginosa, three Enterobacter spp.; and 24 (37.5%) positive bacterial cultures were obtained in convalescence: six H. influenzae, five S. pneumoniae, four M. catarrhalis; four S. aureus; three Enterobacter spp., two P. aeruginosa (p = 0.08 vs. exacerbations). The bacterial load in positive samples was 106 cfu/ml or greater. Samples yielding a bacterial growth of 107 cfu/ml or more were as follows: 27 at exacerbation (77.1% of the positive samples) and 12 in stable conditions (50% of the positive samples; p < 0.01). The frequency of positive bacterial growth was not different in the subgroup of subjects with chronic bronchitis as compared with those without chronic bronchitis both at exacerbation and in stable convalescence comparisons. However, in the subset of patient with chronic bronchitis, but not in those without, the frequency of positive samples at exacerbation was significantly higher than in stable conditions (p < 0.05). (For detailed description of these data, see the online supplement.)

Purulent sputum at exacerbation was more frequent in infective exacerbations (30/50) as compared with noninfective exacerbations (3/14; p < 0.05). Interestingly, the purulence of sputum was not different in viral versus bacterial infections. (For detailed description of these data, see the online supplement.)

Increased Severity of Exacerbations Associated with Infection

Having detected infection in the majority of exacerbations, we next determined whether infection-associated exacerbations were more severe than those in which no pathogen was detected. Exacerbations associated with identified pathogens (bacteria and/or viruses, n = 50; 78%) when compared with noninfective exacerbations (n = 14; 22%) had longer hospitalizations (11.6 ± 0.6 vs. 8.8 ± 0.8 d, p < 0.02), greater decreases in FEV1 (0.25 ± 0.03 vs. 0.12 ± 0.04 L, p < 0.05), FEV1% predicted (11.44 ± 1.7 vs. 5.3 ± 2.1%, p < 0.05), FEV1/FVC% (7.32 ± 0.7 vs. 5.47 ± 0.9%, p < 0.05), and diffusion capacity (Kco% predicted: 14.5 ± 1.9 vs. 7.6 ± 3.2%, p < 0.01). There was also a trend for a greater decrease in PaO2 (15.3 ± 1.35 vs. 11.92 ± 1.86 mm Hg, p = 0.078).

Exacerbations associated with infection also had greater blood neutrophil responses compared with noninfective exacerbations (increase of 5.3 ± 0.4 vs. 4.0 ± 0.3 × 103/dl, p = 0.02).

Relationship among Viral and/or Bacterial Etiology, Airway Inflammation, and Lung Function Impairment

To investigate relative differences among viral and/or bacterial etiology, lung function impairment, and specific patterns of airway inflammation, patients were divided into four subgroups on the basis of the virologic and bacteriologic detection at exacerbation: virus infection alone (V; n = 15; 23.4%), viral and bacterial coinfection (VB; n = 16; 25%), bacterial infection alone (B; n = 19; 29.7%), and no infectious agent detected (N; n = 14; 21.8%).

Lung function impairment and length of hospitalization.

Kco was significantly decreased at exacerbation in all subgroups (p < 0.01); values were significantly lower in the VB (40.1 ± 1.4% predicted) compared with the N subgroup (48.6 ± 1.6% predicted, p = 0.02). Similarly, FEV1/FVC% was significantly lower at exacerbation in all subgroups (p < 0.01), with a decrease that was significantly greater in the VB (8.0 ± 1.1% predicted) compared with the N subgroup (5.5 ± 0.9% predicted, p < 0.02,). Although the length of hospitalization was not different among the groups (V: 11.3 ± 0.9 d; VB: 11.18 ± 0.95 d; B: 11.15 ± 0.85 d; N: 8.8 ± 0.8 d; p = 0.2 by ANOVA), numbers of patients with hospitalizations of 10 d or more were significantly greater in the VB (13/16, 81%) compared with the N subgroup (4/14, 29%; p = 0.001).

There were no significant differences between subgroups in any other clinical parameter at exacerbation (data not shown).

Airway inflammation.

Sputum neutrophils were significantly increased at exacerbation in all subgroups (p < 0.01), but there were no significant differences between subgroups (Figure 3). In addition, changes in sputum neutrophil counts from exacerbation to stable convalescence were not different among subgroups (14.6 ± 5.8 106/g, 17.1 ± 5.4 106/g, 29.6 ± 13.5 106/g, 24.9 ± 12.3 106/g for groups V, VB, B, and N, respectively; p ⩾ 0.36 for all comparisons). Similarly, the increased sputum NE found at exacerbation occurred in all subgroups (sputum NE at exacerbation vs. stable convalescence, p < 0.01 in all subgroups except N where p < 0.05; Table 2). In sputum, the ratio between NE (expressed as μg/ml supernatant) and neutrophil cell counts (expressed as 106/g) at exacerbation in the N group (80 [44.2–141]) was significantly lower as compared with the infective group (156 [84–295.3], p = 0.05).



Group V

Group VB

Group B

Group N

p Value
NE, μg/ml
 Exacerbation,4.1* (2.7–5.88)3.4* (1.7–5.1)3.3* (1.2–8.1)2.4 (1.1–5.6)NS
 Convalescence0.78 (0.5–0.86)0.62 (0.35–1.1)0.61 (0.2–1.12)0.41 (0.23–1.36)NS
ECP, μg/ml
 Exacerbation3.79* (0.86–5.32)4.08* (0.5–5.56)0.92 (0.06–1.78)1.24 (0.2–2.48)p < 0.05, V vs. N p < 0.01, VB vs. B
0.70 (0.02–0.96)
0.74 (0.02–0.92)
0.78 (0.03–0.94)
0.88 (0.1–1.08)

Definition of abbreviations: B = bacteria at exacerbation; ECP = eosinophil cationic protein; N = no pathogen at exacerbation; NE = neutrophil elastase; NS = not significant; V = viruses at exacerbation; VB = viruses + bacteria at exacerbation.

*p < 0.05 versus convalescence.

p < 0.01 versus convalescence.

At variance with neutrophils, sputum eosinophils were significantly elevated at exacerbation only in the subgroups with viral infections (3.50 [1.47–6.00] and 3.65 [1.77–6.85] during exacerbations vs. 0.96 [0–1.80] and 0.75 [0–1.35] × 106/g in stable convalescence in V and VB, respectively; p < 0.001; Figure 4). Eosinophils at exacerbation were significantly greater in the V and VB groups compared with both the B group (1.00 [0.20–1.65] × 106/g; p < 0.001 vs. V and VB) and N group (1.30 [0–1.95] × 106/g, p < 0.01 vs. V and p < 0.001 vs. VB; Figure 4). Significant correlation was found between the increase in exhaled NO and sputum eosinophils during exacerbations in the V subgroup (r = 0.67, p = 0.038).

By considering the acute episodes where respiratory viruses were found (i.e., V and VB subgroups), we found that sputum eosinophils at exacerbation are good predictors of virus-associated exacerbations with or without a concomitant bacterial coinfection. By ROC curve analysis, the area under the ROC curve was 0.83 for sputum eosinophils at exacerbation and 0.91 for the difference between sputum eosinophils at exacerbation and at stable convalescence. For sputum eosinophil numbers, values of 1.68 × 106/g or greater (best cut-off point) predicted a virus-associated COPD exacerbation (sensitivity of 0.82 and a specificity of 0.77) with or without a concomitant bacterial coinfection. Similarly, for sputum eosinophil increase from exacerbation to stable convalescence, values of 0.3 × 106/g or more predicted a virus-associated COPD exacerbation (sensitivity of 0.87 and a specificity of 0.81) with or without a concomitant bacterial coinfection.

Overall, sputum ECP was increased at exacerbation (2.4 [0.78–3.79] μg/ml) as compared with stable convalescence (0.60 [0.04–1.08] μg/ml, p < 0.001). Sputum ECP positively correlated with sputum eosinophil number both at exacerbation (r = 0.68, p < 0.001) and in stable convalescence (r = 0.61, p < 0.001). At exacerbation, sputum ECP was significantly higher as compared with stable convalescence in the V and VB groups (p < 0.01; Table 2). Comparisons between subgroups at exacerbations showed significant differences between groups V and N (p < 0.05) and VB and B (p < 0.01; Table 2). Sputum ECP levels in the supernatant of samples obtained from exacerbations associated with viral infection, either with or without a bacterial coinfection (3.88 [0.62–5.48] μg/ml), were significantly higher as compared with virus-free exacerbations (1.12 [0.08–2.02] μg/ml, p < 0.01).

In this study, we observed that 78% of patients with severe COPD exacerbations admitted to the hospital are associated with respiratory virus and/or bacterial infection. The presence of infection with viruses (48.4%) and/or bacteria (54.7%) at exacerbation was associated with both clinical and lung function indicators of increased exacerbation severity. Importantly, the presence of coinfection with both viruses and bacteria was found in 25% of exacerbations and these patients had more severe functional impairment and longer hospitalization. These data clearly indicate that the great majority of severe COPD exacerbations requiring hospitalization are infectious in etiology. The study also shows sputum and blood neutrophilia during exacerbations, which is significantly related to exacerbation severity independently of the presence of detectable virus and/or bacteria in sputum. Finally, our study also showed that sputum eosinophilia is associated with viral infection and with viral/bacteria coinfection at exacerbations, and that sputum eosinophilia can represent a good predictor of viral exacerbations.

This is the first study in which both viral and bacterial infection and inflammation were simultaneously assessed at hospital admission and convalescence in sputum obtained from patients with severe exacerbations of COPD.

The design of the study was important in enabling the clinical characterization of the exacerbations to be extremely detailed, as were assessments when stable. Stability of the disease when patients were evaluated at convalescence is indicated by the fact that patients regained preexacerbation baseline conditions both clinically and in terms of lung function values (Table E1).

The exacerbations were associated with significant impairment of most lung function parameters assessed (Table 1), suggesting their primary respiratory nature.

To improve the management and the outcomes of these severe exacerbations, information is required on the etiology and pathogenesis. In agreement with previous studies that have investigated only respiratory viruses (27), viruses were detected in around half of exacerbations, and were detected significantly more often at exacerbation than when stable (p < 0.001), and rhinoviruses (54.8% of virus-positive cases) and influenza viruses were the most frequently identified viruses (27). Consistent with studies that have only performed bacteriologic evaluations, bacterial infections were also found in around half of exacerbations (5). However, bacterial infection in our setting only tended to be significantly more frequent in exacerbation than when stable (p = 0.08) because, as previously reported, significant bacterial infection was also detected in almost 40% of patients when stable (27). These data are consistent with bacterial infection playing a significant role in the pathogenesis of stable COPD as well as in acute exacerbations; however, stable COPD was not the focus of this study—the role of bacteria in stable COPD is discussed elsewhere (27). To underline complexity of the issue of bacterial infection versus colonization in COPD and of the relationship with airway inflammation, we found, in accordance with previous studies limited to chronic bronchitis (14, 28), that samples yielding a bacterial growth of 107 cfu/ml or more were significantly more frequent at exacerbation as compared with stable convalescence. We confirmed this finding in our population with chronic bronchitis. Furthermore, when we looked at exacerbation associated with bacterial infection (with and without a viral coinfection), we found that higher bacterial loads were more frequently associated with increased sputum neutrophils (see the online supplement for details), supporting the concept that when bacteria are present at exacerbation, airway inflammation is higher in those samples where bacterial load is higher. A subanalysis of the presence of sputum purulence was performed in the groups subdivided according to pathogens identified at exacerbation. Although the frequency of purulent sputum was higher in infective exacerbations, the purulence of the sputum was unable to identify the subgroup of samples with positive bacterial growth. This is partially in contrast with published studies in chronic bronchitis (14, 29) where viral analysis was not performed, and therefore the subgroup with double viral and bacterial coinfection was only counted among the bacterially positive samples.

In previous studies, bacterial strain typing has been shown to be important in discriminating between infective and colonizing strains (9, 30, 31). However, the setting of our study was not adequate for such evaluation as no samples were obtained before the start of the exacerbation.

In addition to finding a very high frequency of infection in severe COPD exacerbations, our observation that exacerbations associated with infection are more severe in terms of all of length of hospitalization, decrease in FEV1, FEV1% predicted, FEV1/FVC%, and diffusion capacity, with a trend for a greater decrease in PaO2, is of major clinical importance. Importantly, we observed that coinfection with both viruses and bacteria occurs in 25% of exacerbations. Exacerbations with coinfection were associated with greater impairment in terms of lung function (FEV1/FVC% and Kco) and longer length of hospitalization. These data indicate that infection is the major driver of severity of COPD exacerbation and that therapeutic strategies need to be focused on treating infection and its consequences (1).

These observations also highlight the need to develop therapeutic and/or preventive strategies to combat virus infections in COPD exacerbations. Antiviral therapy may have a role in at least half and probably two-thirds of exacerbations (as 65% of patients report common cold symptoms in the week preceding exacerbation [23]), and a combined antibacterial and antiviral strategy in at least one-quarter of exacerbations (i.e., those with viral and bacterial coinfections). Use of antiinfluenza agents in COPD exacerbations accompanied by appropriate symptoms during a known influenza epidemic appears appropriate (32), although further controlled clinical trails are needed to confirm therapeutic benefit. Antiviral agents for rhinoviruses and RSV are in development—this development needs accelerating. Relationships between viral and bacterial infection, especially when combined together, and whether viral infections can lead to bacteriologic exacerbation (of previously colonizing bacteria) are still open questions that deserve a properly designed longitudinal study.

Antiinflammatory therapy also may have a therapeutic role in COPD exacerbations (15). However, the nature of the inflammatory response in COPD exacerbations is highly controversial (1017). The variability of inflammatory responses described in studies of COPD exacerbations so far could theoretically be explained by differing etiologies. Most of the previous studies, particularly those seeking correlations between airway inflammation and bacterial infections, have investigated indirect markers of inflammatory cell recruitment and activation within the airways. Because the main aim of this study was clinical, we examined sputum cell count as a marker of airway inflammation, although this method is not yet a routine laboratory test (33).

This enabled us to investigate inflammatory changes in four subgroups of exacerbations: viral infection alone, bacterial infection alone, viral and bacterial coinfection, and no pathogen detected. We demonstrate that sputum neutrophilia is present during exacerbations in all subgroups independent of etiology. Similar results were obtained from the analysis of sputum NE, a soluble mediator released from neutrophils, which was increased at exacerbation as already documented (13). Previous studies have found increased markers of neutrophilic inflammation in sputum supernatant in exacerbation associated with bacterial infection; however, viral and viral/bacterial coinfections were not included in those comparisons (16, 29, 34). There are some important differences in the patient population in this study as compared with earlier studies. For example, earlier studies involved outpatients, whereas in this study, hospitalized patients were considered (14, 16, 28, 29). Also, in contrast with the present study, previous studies were restricted to patients with chronic bronchitis (mainly COPD), with the exclusion of those patients with COPD not belonging to the chronic bronchitis phenotype (14, 16, 28, 29). In contrast with previous studies that have measured free, active NE in the sputum supernatant of patients with COPD (16, 28, 29), we measured total human NE, since our aim was to evaluate neutrophilic release of NE and not NE activity, which is influenced by the presence of NE inhibitors in sputum (35, 36). Interestingly, although no difference in sputum NE was found between subgroups, the ratio between NE and neutrophil cell counts at exacerbation was significantly lower in the group where no pathogen was found at exacerbation as compared with the infective group, indicating less release of NE from neutrophils, and possibly less activation of neutrophils, when a pathogen is not present. Notably, bacterial lipopolysaccharide is a well-known stimulus for NE release from neutrophils (37).

Previous studies have found correlations between indirect soluble markers of sputum neutrophilic inflammation and clinical severity of exacerbations (16). Our results substantiate and extend those results in showing for the first time a significant correlation with sputum and circulating neutrophil cell counts (Figures 1 and 2). These data are important as they may indicate a cause–effect relationship between neutrophilic inflammation and exacerbation severity.

Interestingly, the frequency of bacterial infections during COPD exacerbations contrasts with the marginal effect of antibiotic therapy documented in the past (38) and the significant effect of antiinflammatory steroid treatment (1), suggesting that suppression of inflammation might be more important than removal of bacteria for the improvement of symptoms and functional abnormalities associated with COPD exacerbations. However, because the data showing a marginal beneficial effect of antibiotics on exacerbation clinical outcomes may have been influenced by the fact that poor antibiotics were available at the time when those studies were performed, a reevaluation of this hot topic with currently available antibiotic treatment deserves study.

This study also provides an explanation for the controversy regarding lower airway eosinophilia during COPD exacerbations (1017). We observed increased sputum eosinophils only in those exacerbations associated with viral infections (Figure 4). The sputum eosinophilia observed was similar to that reported in previous studies showing airway eosinophilia during exacerbations (11). Consistent with our results, experimental rhinovirus infection (the most frequently identified virus type) induces lower airway eosinophilia (39, 40) and rhinovirus infection of respiratory epithelial cells induces several proinflammatory mediators promoting eosinophil recruitment (41).

The results of our study show that sputum eosinophils at exacerbations are a good predictor of viral exacerbation, either in the presence or in the absence of a bacterial coinfection. This finding suggests the possible use of noninvasive measurements (e.g., sputum eosinophils) in clinical practice to provide clinicians relevant etiologic information. Further studies are required to confirm and extend this pivotal observation. In agreement with previous studies, we found that sputum ECP levels were higher at exacerbation (13). We could document that this increase is specifically correlated with virus-associated exacerbations, either in the presence or in the absence of a bacterial coinfection.

One possible bias in the interpretation of the results of our study is the fact that stable data were acquired in stable convalescence (i.e., after pharmacologic treatment). However, although we cannot completely exclude prolonged effects for some of the treatment, this possibility does not seem to interfere with the interpretation of results in our setting because clinical/lung function evaluations and sampling at exacerbation were performed before any treatment was started and because exacerbation treatment was standardized. Furthermore, lung function, blood gases, and use of rescue medication at convalescent assessment were not different from preexacerbation values (Table E1), recorded before exacerbation from all subjects because they were regularly followed up by our center every 3 mo, suggesting complete recovery to preexacerbation conditions. Lung function parameters for which a significant difference was observed from convalescence to exacerbation also showed a significant difference in the comparisons between preexacerbation baseline values and exacerbation values (Table E2). No difference was found in lung function changes from baseline to exacerbation when compared with those observed between exacerbation and convalescence (p > 0.57 for all comparisons).

No difference between subgroups was observed at preexacerbation baseline in lung function and rescue medication use (data not shown). It was not possible to collect microbiological data and sputum cell counts at baseline, before the study entry, from the entire cohort of patients with COPD referred to our center.

We conclude that viral and bacterial infections are associated with the vast majority of severe COPD exacerbations requiring hospitalization, and that presence of infection is related to exacerbation severity. Antiviral and antibacterial therapy may be beneficial in treating COPD exacerbations, and in around one-fourth of exacerbations, combined therapy may be indicated (i.e., those with bacterial and viral coinfection). Sputum neutrophils are increased at exacerbation, independent of etiology, and are related to exacerbation severity; development of therapies targeting neutrophilic inflammation therefore may have therapeutic benefit. Finally, sputum eosinophilia at exacerbation could be used as predictor of viral etiology.

The authors thank Elisa Veratelli for scientific secretarial assistance.

1. Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease. NHLBI/WHO Global Initiative for Chronic Obstructive Lung Disease (GOLD) workshop summary. Am J Respir Crit Care Med 2001;163:1256–1276.
2. Sullivan SD, Ramsey SD, Lee TA. The economic burden of COPD. Chest 2000;117:5S–9S.
3. Seemungal TA, Donaldson GC, Paul EA, Bestall JC, Jeffries DJ, Wedzicha JA. Effect of exacerbation on quality of life in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1998;157:1418–1422.
4. Donaldson GC, Seemungal TA, Bhowmik A, Wedzicha JA. Relationship between exacerbation frequency and lung function decline in chronic obstructive pulmonary disease. Thorax 2002;57:847–852.
5. Sethi S, Murphy TF. Bacterial infection in chronic obstructive pulmonary disease in 2000: a state-of-the-art review. Clin Microbiol Rev 2001;14:336–363.
6. White AJ, Gompertz S, Stockley RA. Chronic obstructive pulmonary disease. 6: the aetiology of exacerbations of chronic obstructive pulmonary disease. Thorax 2003;58:73–80.
7. Rohde G, Wiethege A, Borg I, Kauth M, Bauer TT, Gillissen A, Bufe A, Schultze-Werninghaus G. Respiratory viruses in exacerbations of chronic obstructive pulmonary disease requiring hospitalisation: a case-control study. Thorax 2003;58:37–42.
8. Seemungal TA, Wedzicha JA. Viral infections in obstructive airway diseases. Curr Opin Pulm Med 2003;9:111–116.
9. Sethi S, Evans N, Grant BJ, Murphy TF. New strains of bacteria and exacerbations of chronic obstructive pulmonary disease. N Engl J Med 2002;347:465–471.
10. Saetta M, Di Stefano A, Maestrelli P, Turato G, Ruggieri MP, Roggeri A, Calcagni P, Mapp CE, Ciaccia A, Fabbri LM. Airway eosinophilia in chronic bronchitis during exacerbations. Am J Respir Crit Care Med 1994;150:1646–1652.
11. Maestrelli P, Saetta M, Di Stefano A, Calcagni PG, Turato G, Ruggieri MP, Roggeri A, Mapp CE, Fabbri LM. Comparison of leukocyte counts in sputum, bronchial biopsies, and bronchoalveolar lavage. Am J Respir Crit Care Med 1995;152:1926–1931.
12. Zhu J, Qiu YS, Majumdar S, Gamble E, Matin D, Turato G, Fabbri LM, Barnes N, Saetta M, Jeffery PK. Exacerbations of bronchitis: bronchial eosinophilia and gene expression for interleukin-4, interleukin-5, and eosinophil chemoattractants. Am J Respir Crit Care Med 2001;164:109–116.
13. Fujimoto K, Yasuo M, Urushibata K, Hanaoka M, Koizumi T, Kubo K. Airway inflammation during stable and acutely exacerbated chronic obstructive pulmonary disease. Eur Respir J 2005;25:640–646.
14. Stockley RA, O'Brien C, Pye A, Hill SL. Relationship of sputum color to nature and outpatient management of acute exacerbations of COPD. Chest 2000;117:1638–1645.
15. Aaron SD, Angel JB, Lunau M, Wright K, Fex C, Le Saux N, Dales RE. Granulocyte inflammatory markers and airway infection during acute exacerbation of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163:349–355.
16. Sethi S, Muscarella K, Evans N, Klingman KL, Grant BJ, Murphy TF. Airway inflammation and etiology of acute exacerbations of chronic bronchitis. Chest 2000;118:1557–1565.
17. Qiu Y, Zhu J, Bandi V, Atmar RL, Hattotuwa K, Guntupalli KK, Jeffery PK. Biopsy neutrophilia, neutrophil chemokine and receptor gene expression in severe exacerbations of chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2003;168:968–975.
18. Barker AF. Bronchiectasis. N Engl J Med 2002;346:1383–1393.
19. Hansell DM. Bronchiectasis. Radiol Clin North Am 1998;36:107–128.
20. Fabbri LM, Romagnoli M, Corbetta L, Casoni G, Busljetic K, Turato G, Ligabue G, Ciaccia A, Saetta M, Papi A. Differences in airway inflammation in patients with fixed airflow obstruction due to asthma or chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2003;167:418–424.
21. Quanjer PH, Tammeling GJ, Cotes JE, Pedersen OF, Peslin R, Yernault JC. Lung volumes and forced ventilatory flows. Report Working Party Standardization of Lung Function Tests, European Community for Steel and Coal. Official statement of the European Respiratory Society. Eur Respir J Suppl 1993;16:5–40.
22. Barrow GI, Feltham RK. Cowan and Steel's manual for identification of medical bacteria. Cambridge, UK: Cambridge University Press; 1993.
23. Seemungal T, Harper-Owen R, Bhowmik A, Moric I, Sanderson G, Message S, Maccallum P, Meade TW, Jeffries DJ, Johnston SL, et al. Respiratory viruses, symptoms, and inflammatory markers in acute exacerbations and stable chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;164:1618–1623.
24. Papadopoulos NG, Hunter J, Sanderson G, Meyer J, Johnston SL. Rhinovirus identification by BglI digestion of picornavirus RT-PCR amplicons. J Virol Methods 1999;80:179–185.
25. Maertzdorf J, Wang CK, Brown JB, Quinto JD, Chu M, de Graaf M, van den Hoogen BG, Spaete R, Osterhaus AD, Fouchier RA. Real-time reverse transcriptase PCR assay for detection of human metapneumoviruses from all known genetic lineages. J Clin Microbiol 2004;42:981–986.
26. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29–36.
27. Wedzicha JA. Exacerbations: etiology and pathophysiologic mechanisms. Chest 2002;121:136S–141S.
28. White AJ, Gompertz S, Bayley DL, Hill SL, O'Brien C, Unsal I, Stockley RA. Resolution of bronchial inflammation is related to bacterial eradication following treatment of exacerbations of chronic bronchitis. Thorax 2003;58:680–685.
29. Gompertz S, O'Brien C, Bayley DL, Hill SL, Stockley RA. Changes in bronchial inflammation during acute exacerbations of chronic bronchitis. Eur Respir J 2001;17:1112–1119.
30. Bandi V, Apicella MA, Mason E, Murphy TF, Siddiqi A, Atmar RL, Greenberg SB. Nontypeable Haemophilus influenzae in the lower respiratory tract of patients with chronic bronchitis. Am J Respir Crit Care Med 2001;164:2114–2119.
31. Murphy TF, Brauer AL, Aebi C, Sethi S. Antigenic specificity of the mucosal antibody response to Moraxella catarrhalis in chronic obstructive pulmonary disease. Infect Immun 2005;73:8161–8166.
32. Murphy KR, Eivindson A, Pauksens K, Stein WJ, Tellier G, Watts R. Efficacy and safety of inhaled Zanamivir for the treatment of influenza in patients with asthma or chronic obstructive pulmonary disease: a double-blind, randomised, placebo-controlled, multicentre study. Clin Drug Invest 2000;20:333–349.
33. Pavord ID, Sterk PJ, Hargreave FE, Kips JC, Inman MD, Louis R, Pizzichini MM, Bel EH, Pin I, Grootendorst DC, et al. Clinical applications of assessment of airway inflammation using induced sputum. Eur Respir J Suppl 2002;37:40s–43s.
34. Crooks SW, Bayley DL, Hill SL, Stockley RA. Bronchial inflammation in acute bacterial exacerbations of chronic bronchitis: the role of leukotriene B4. Eur Respir J 2000;15:274–280.
35. Travis J. Structure, function, and control of neutrophil proteinases. Am J Med 1988;84:37–42.
36. Sallenave JM. The role of secretory leukocyte proteinase inhibitor and elafin (elastase-specific inhibitor/skin-derived antileukoprotease) as alarm antiproteinases in inflammatory lung disease. Respir Res 2000;1:87–92.
37. Fittschen C, Sandhaus RA, Worthen GS, Henson PM. Bacterial lipopolysaccharide enhances chemoattractant-induced elastase secretion by human neutrophils. J Leukoc Biol 1988;43:547–556.
38. Saint S, Bent S, Vittinghoff E, Grady D. Antibiotics in chronic obstructive pulmonary disease exacerbations: a meta-analysis. JAMA 1995;273:957–960.
39. Fraenkel DJ, Bardin PG, Sanderson G, Lampe F, Johnston SL, Holgate ST. Lower airways inflammation during rhinovirus colds in normal and in asthmatic subjects. Am J Respir Crit Care Med 1995;151:879–886.
40. Trigg CJ, Nicholson KG, Wang JH, Ireland DC, Jordan S, Duddle JM, Hamilton S, Davies RJ. Bronchial inflammation and the common cold: a comparison of atopic and non-atopic individuals. Clin Exp Allergy 1996;26:665–676.
41. Papi A, Message SD, Papadopoulos NG, Casolari P, Ciaccia A, Johnston SL. Respiratory viruses and asthma. In: Chung KF, Fabbri LM, editors. European Respiratory monograph: asthma. Sheffield, UK: European Respiratory Society; 2003. pp. 223–238.
Correspondence and requests for reprints should be addressed to Leonardo M. Fabbri, M.D., Department of Respiratory Diseases, University of Modena and Reggio Emilia, Via del Pozzo 71, I-41100 Modena, Italy. E-mail:


No related items
American Journal of Respiratory and Critical Care Medicine

Click to see any corrections or updates and to confirm this is the authentic version of record