Antibiotics in Addition to Systemic Corticosteroids for
Acute Exacerbations of Chronic Obstructive
Johannes M. A. Daniels1, Dominic Snijders1, Casper S. de Graaff1, Fer Vlaspolder2, Henk M. Jansen3,
and Wim G. Boersma1
1Department of Pulmonary Diseases and2Department of Microbiology, Medical Center Alkmaar, Alkmaar; and3Department of Pulmonary
Diseases, Amsterdam Medical Center, University of Amsterdam, Amsterdam, The Netherlands
sial and their efficacy when added to systemic corticosteroids is
Objectives: We conducted a randomized, placebo-controlled trial to
clinical outcome, microbiological outcome, lung function, and
systemic inflammation in patients hospitalized with an acute exac-
erbation of chronic obstructive pulmonary disease.
Methods: Of223 patients,we enrolled 265 exacerbations definedon
without increased sputum purulence. Patients received 200 mg of
oral doxycycline or matching placebo for 7 days in addition to
to treatment failure, lung function, symptom scores, and serum
C-reactive protein were assessed.
in intention-to-treat patients (odds ratio, 1.3; 95% confidence in-
terval, 0.8 to 2.0) and per-protocol patients. Doxycycline showed
superiority over placebo in terms of clinical success on Day 10 in
1.1 to 3.2), but not in per-protocol patients. Doxycycline was also
use of open label antibiotics, and symptoms. There was no in-
teraction between the treatment effect and any of the subgroup
variables (lung function, type of exacerbation, serum C-reactive
protein, and bacterial presence).
Conclusions: Although equivalent to placebo in terms of clinical
success and clinical cure on Day 10, microbiological success, the use
of open label antibiotics, and symptoms.
Clinical trial registered with www.clinicaltrials.gov (NCT00170222).
Keywords: pulmonary disease; chronic obstructive pulmonary disease;
antibacterial agents; infection
Chronic obstructive pulmonary disease (COPD) constitutes a ma-
jor health problem (1). Acute exacerbations of COPD (AECOPD)
have considerable impact on morbidity, mortality, and quality of
life (2, 3). Common triggers for AECOPD include air pollution
and viral and/or bacterial infection of the airways, but the cause
of approximately one third of severe exacerbations cannot be
identified (4). The role of bacteria in AECOPD is controversial.
In about 50% of exacerbations significant amounts of potential
bacterial pathogens can be isolated from protected brush
specimens obtained by bronchoscopy (5–7). However, the same
pathogens are found in the airways of patients in a stable phase
of the disease (5, 8–11). It is impossible for clinicians to
distinguish infection from colonization. Nonetheless, antibiotics
are widely used to treat patients with AECOPD. Several meta-
analyses have confirmed the value of antimicrobial therapy (12–
14). Antibiotics seem to be most effective in patients with
increased dyspnea, increased sputum volume, and increased
sputum purulence (15). Unfortunately, the placebo-controlled
trials that investigated the efficacy of antibiotics have important
limitations (16–18). Furthermore, these trials were conducted
several decades ago, before systemic steroids were widely
introduced for the treatment of AECOPD (19–22).
It is unclear whether antibiotics have additional benefits
when applied in patients with more severe exacerbations that
are already treated with systemic corticosteroids. Sachs and
colleagues (23) suggested that antibiotics were redundant when
corticosteroids were given, irrespective of sputum color or
bacterial involvement. However, their sample size was small
(n 5 71) and it consisted of both patients with COPD and
patients with asthma. In our opinion, therefore, new placebo-
controlled trials are justified.
We designed a randomized, double-blind, placebo-controlled
trial of doxycycline in addition to systemic corticosteroids for
patients hospitalized with an acute exacerbation of COPD. Our
goal was to assess the effects of doxycycline on clinical and
microbiological response, symptoms, lung function, and sys-
temic inflammation (C-reactive protein). The design of this trial
is unique in that it has incorporated several features that were
missing in other placebo-controlled trials: first, concomitant treat-
ment, including systemic corticosteroids, was fully standardized.
Second, radiographic signs of pneumonia and fever were exclusion
AT A GLANCE COMMENTARY
Scientific Knowledge on the Subject
Although evidence suggests that antibiotics are effective in
acute exacerbations of chronic obstructive pulmonary dis-
ease, most trials were flawed and performed before systemic
corticosteroids were recognized as a beneficial treatment.
What This Study Adds to the Field
This study provides evidence that antibiotics in addition to
systemic corticosteroids have a limited and short-lived effect
on clinical outcome and symptoms and no effect on lung
function and systemic inflammation.
(Received in original form June 4, 2009; accepted in final form October 29, 2009)
Supported by an unrestricted grant from GlaxoSmithKline (The Netherlands).
The funding source had no part in the study design, conduct, or reporting of the
Correspondence and requests for reprints should be addressed to Johannes M. A.
Daniels, M.D., Department of Pulmonary Diseases, VU University Medical Center,
De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands. E-mail: j.daniels@
This article has an online supplement, which is accessible from this issue’s table of
contents at www.atsjournals.org
Am J Respir Crit Care Med
Originally Published in Press as DOI: 10.1164/rccm.200906-0837OC on October 29, 2009
Internet address: www.atsjournals.org
Vol 181. pp 150–157, 2010
criteria to prevent enrollment of patients with pneumonia.
Finally, sputum samples were collected before and after the
intervention to allow a thorough microbiological workup. Some
of the results of these studies have been previously reported in
the form of an abstract (24).
Setting and Participants
Participants were enrolled at the Medical Centre Alkmaar in Alkmaar,
the Netherlands and the Waterland Hospital in Purmerend, the Nether-
lands. The study population consisted of patients 45 years of age or older,
diagnosed with COPD stages I–IV as defined by the Global Initiative
for Chronic Obstructive Lung Disease (GOLD) (25), with an acute
(onset < 14 d) exacerbation (as defined by Anthonisen and colleagues
: type 1 [increased dyspnea, sputum volume, and sputum purulence]
or type 2 [two of three symptoms]) that required hospitalization. Criteria
for hospital admission are described in the online supplement. The most
important exclusion criteria included fever (>38.58C) to prevent enroll-
ment of patients with pneumonia, antibiotic treatment for at least
24 hours, and radiographic signs of pneumonia (not specified). All
exclusion criteria are listed in the online supplement.
Randomization and Intervention
Within 24 hours of admission, patients were randomly assigned to
receive a 7-day course of doxycycline or placebo. Details about the
randomization process and allocation concealment are presented in the
online supplement. Concomitant treatment consisted of intravenously
administered prednisolone (starting with 60 mg/d, tapering by 10 mg
per 2 d to 40 mg/d followed by 30 mg of oral prednisolone on Day 7,
tapering by 5 mg per 2 d to 0 mg or the maintenance dose before
admission), nebulized bronchodilator therapy four to six times daily,
and physiotherapy. Other COPD medication was continued, except for
short-acting bronchodilators. Because all patients received 6 days of
intravenous steroids the minimum length of stay was 7 days.
Throughout the trial, attending physicians were unable to access the
results of sputum cultures. However, when a pathogen was isolated that
is usually resistant to doxycycline, the attending physician was in-
formed. In case of clinical treatment failure the attending physician was
allowed to access the culture results and to replace the study drug by an
open label antibiotic. Rerandomization was allowed if a new exacer-
bation occurred at least 3 months after the first enrollment. Safety was
monitored daily with the help of adverse event reports.
Outcomes and Follow-Up
On Days 1, 10, and 30, patients were assessed clinically: blood samples
were drawn for measurement of C-reactive protein (CRP, Beckman
Coulter Inc., Fullerton, CA) and serologic testing, spirometry was
performed and expectorated sputum samples were collected. Symptom
scores consisted of visual analog scales (VAS) for dyspnea, cough,
fatigue, and sputum purulence. For each symptom the minimal score
was 1 and the maximal score was 10. Separate and total scores were
calculated. Microbiological procedures are described in detail in the
The primary end point was clinical response on Day 30 as defined by
Chow and colleagues (26) Treatment success was defined as cure (a
complete resolution of signs and symptoms associated with the exacer-
bation) or improvement (a resolution or reduction of the symptoms and
signs without new symptoms or signs associated with the infection).
Treatment failure was defined as absence of resolution of symptoms and
signs, worsening of symptoms and signs, occurrence of new symptoms
and signs associated with the primary or a new infection or death.
Secondary end points included clinical success on Day 10, clinical cure
on Days 10 and 30, antibiotic treatment for lack of efficacy, lung function
(DFEV1), time to treatment failure (defined previously), serum C-reactive
protein (CRP), symptoms, and microbiological response. Criteria for
microbiological response are described in the online supplement.
Our sample size calculation was based on the results of Anthonisen and
colleagues (15), who found a clinical success rate of 52% for the
placebo group and 67% for the antibiotic group (type 1 and type 2
AECOPD combined). We calculated that 167 exacerbations were
needed in both arms to detect the previously mentioned difference
between antibiotic and placebo treatment on Day 30 with a power of
80% and a two-sided a level of 0.05. During the trial we discovered that
the percentage of type 1 exacerbations in our trial was higher than
expected. Therefore we expected a higher treatment effect (16.7%).
Recalculation of the sample size showed that 132 exacerbations were
needed in each arm.
SPSS 16.0 for Windows (SPSS Inc, Chicago, IL) and Stata version
11.0 (StataCorp, College Station, TX) were used for data management
and statistical analysis. Differences between the treatment groups were
analyzed by logistic regression analysis, correcting for within-patient
clustering with generalized estimating equations. Differences in time to
treatment failure were compared by Cox proportional hazards re-
gression, adjusting for within-patient clustering by robust standard
error estimation. Subgroups were specified according to type of exac-
erbation, bacterial presence, serum CRP, and lung function. Heteroge-
neity of treatment effect between subgroups was examined by logistic
regression analysis. A significance level of 0.05 was specified for all
comparisons. We did not correct for multiple comparisons because
several sets of end points are closely related and do not represent
One planned interim analysis, regarding the primary end point and
mortality, that was performed by an independent statistician after
enrollment of 140 exacerbations showed no significant differences and
the study was therefore continued as planned.
Of the 367 exacerbations that were screened, 265 exacerbations
of 223 patients were enrolled and randomly assigned to placebo
(137 exacerbations) or doxycycline (128 exacerbations). After two
enrollments, inclusion at the Waterland Ziekenhuis in Purmerend
was terminated for logistical reasons. The majority of the patients
in the placebo group (110 of 137 [80%]) and the doxycycline
group (111 of 128 [87%]) completed the trial (Figure 1). The most
common reason for withdrawal was lack of efficacy, which was
more common in the placebo group (23 of 137 [17%]) than in the
doxycycline group (8 of 128 [6%]). The baseline characteristics
are shown in Table 1. Of the 265 enrolled exacerbations, 178
(67%) were type 1 and 87 (33%) were type 2.
On Day 30, clinical success was observed in 78 patients (61%)
from the doxycycline group and 72 patients (53%) from the
placebo group (odds ratio [OR], 1.3; 95% confidence interval
[95% CI], 0.8 to 2.0; P 5 0.32) (Table 2). In the per-protocol
population we found similar results (OR, 1.2; 95% CI, 0.7 to 1.9;
P 5 0.47).
Clinical outcome on Day 10. On Day 10, clinical success was
observed in 103 patients (80%) from the doxycycline group and
94 patients (69%) from the placebo group (OR, 1.9; 95% CI, 1.1
to 3.2; P 5 0.03). This significant difference was lost in the per-
protocol population (OR, 1.8; 95% CI, 1.0 to 3.1; P 5 0.05).
Clinical cure. Clinical cure was observed in 86 patients
(67%) from the doxycycline group and 69 patients (51%) from
the placebo group on Day 10 (OR, 1.9; 95% CI, 1.2 to 3.2; P 5
0.01). On Day 30 clinical cure was observed in 65 patients
(51%) from the doxycycline group and 56 patients (41%) from
the placebo group (OR, 1.4; 95% CI, 0.9 to 2.3; P 5 0.15).
Time to treatment failure. Time to treatment failure (defined
as absence of resolution of symptoms and signs, worsening of
symptoms and signs, occurrence of new symptoms and signs
associated with the primary or a new infection or death) was not
Daniels, Snijders, de Graaff, et al.: Antibiotic Therapy for Acute Exacerbations of COPD151
significantly longer in the doxycycline group than in the placebo
group (P 5 0.19 by the log-rank test) (Figure 2). Forty-six
patients (37%) in the doxycycline group and 62 patients (46%)
in the placebo group had treatment failure.
Antibiotic treatment for lack of efficacy. Open label antibiotic
treatment for lack of efficacy was applied in 19 patients (15%)
of the doxycycline group and 38 patients (28%) in the placebo
group by Day 10 (OR, 0.5; 95% CI, 0.3 to 0.9; P 5 0.01) and 42
patients (33%) in the doxycycline group and 61 patients (45%)
of the placebo group by Day 30 (OR, 0.7; 95% CI, 0.4 to 1.1;
P 5 0.13) (Table 2).
Lung function. Paired lung function data were available for
224 patients (85%) on Days 1 and 10 and in 189 patients (71%)
on Days 1 and 30. The mean increase in FEV1on Day 10 was
0.16 6 0.26 L in the doxycycline group and 0.11 6 0.26 L in the
placebo group (mean difference, 0.05 L; 95% CI, –0.02 to 0.12;
P 5 0.16) (Table 2). On Day 30, the mean increase was 0.15 6
0.33 L in the doxycycline group and 0.08 6 0.25 L in the placebo
group (mean difference, 0.07 L; 95% CI, 20.03 to 0.13; P 5
0.22) (Table 2).
Serum C-reactive protein. The mean change in serum CRP
on Day 10 was 256.4 6 65.5 mg/L in the doxycycline group and
238.9 6 72.7 mg/L in the placebo group (P 5 0.07) (Table 2).
This trend was no longer present on Day 30.
Symptom scores. The mean change in total symptoms score
on Day 10 was 210.1 6 9.0 in the doxycycline group and 26.2 6
8.6 in the placebo group (mean difference, 22.3; 95% CI, 23.9
to 20.8; P 5 0.003) (Table 2). On Day 30, the mean change was
29.4 6 9.7 in the doxycycline group and 28.3 6 8.6 in the
placebo group (mean difference, 21.0; 95% CI, 23.7 to 1.8; P 5
0.50). Separate mean symptom scores of cough and sputum
purulence were significantly more reduced in those treated with
doxycycline on Day 10, but not on Day 30 (Table 2).
Microbiological outcome. Two hundred and fourteen poten-
tial bacterial pathogens were isolated in 158 exacerbations. The
most predominant bacteria were Haemophilus influenzae
(41%), Streptococcus pneumoniae (24%) and Moraxella catar-
rhalis (22%). A viral infection was serologically diagnosed in
20 patients (influenza A virus, n 5 6; parainfluenza virus, n 5 5;
respiratory syncytial virus, n 5 5; adenovirus, n 5 3; influenza B
virus, n 5 1). Resistance to tetracycline was observed in 1% of
H. influenzae isolates, 7% of S. pneumoniae isolates, 7% of
M. catarrhalis isolates, 0% of Staphylococcus aureus isolates,
and 48% of Pseudomonas spp. isolates. We were able to evaluate
low-up of patients. Shows the
dom assignment, and follow-
up of patients.
Enrollment and fol-
152AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 1812010
the bacteriological response in 151 patients (Table 3). In the
doxycycline group bacteriological success was accomplished in
52 of 78 patients (67%) and in the placebo group in 25 of 73
patients (34%) (OR, 3.8; 95% CI, 1.9 to 7.5; P , 0.001) (Table
3). For the three most predominant pathogens (H. influenzae,
S. pneumoniae, and M. catarrhalis) the success rates were
significantly better in the doxycycline group (Table 3). In the
doxycycline group bacterial persistence rates were 31% for H.
influenzae, 17% for S. pneumoniae, and 9% for M. catarrhalis.
On Day 10 doxycycline showed superiority in patients with type
1 AECOPD and patients with a CRP value of 50 mg/L or more
(Figure 3a). However, the treatment effect did not differ among
these subgroups. On Day 30 doxycycline was still superior to
placebo in patients with a CRP of at least 50 mg/L, but not in
patients with type 1 AECOPD (Figure 3B). Again, no in-
teraction between the treatment effect and any of the subgroup
variables was found.
Adverse reactions that were considered to be related to the
study medication occurred in four patients (3%) in the doxy-
cycline group and five patients (4%) in the placebo group.
Adverse reactions in both groups were heartburn, diarrhea, and
nausea. All reactions were mild and self-limiting. In only one
patient was the study medication discontinued, because of an
adverse reaction (placebo group, complaints of heartburn).
Serious adverse events occurred in 11 patients (9%) from the
doxycycline group (11 events, including 7 deaths) and 7 patients
(5%) from the placebo group (7 events, including 3 deaths).
Serious adverse events in the doxycycline group included
pneumonia, urinary tract infection, myocardial infarction, and
hypoglycemia. One patient died of gram-negative sepsis, four
patients died of respiratory failure, and two patients died of
acute heart failure. Serious adverse events in the placebo group
included pneumonia and stomach perforation. One patient died
of pneumonia and two died of respiratory failure.
We found no significant difference in clinical outcome on Day 30
among patients with AECOPD who were randomly assigned to
doxycycline as compared with those who were assigned to
placebo. Doxycycline was superior to placebo in terms of clinical
success and clinical cure on Day 10 as well as microbiological
success. In addition, open label antibiotic therapy for lack of
efficacy occurred significantly less often in patients assigned to
doxycycline. Finally, patients taking doxycycline had a greater
reduction in symptoms on Day 10. We did not see a difference in
recovery of lung function, resolution of systemic inflammation,
and time to treatment failure between the treatment groups.
This is the first placebo-controlled trial of antibiotics in
addition to systemic corticosteroids for acute exacerbations of
COPD. We did not find a significant treatment effect at the
primary end point (clinical success on Day 30). This observation
could be explained by several factors: first, systemic steroids
have proven to be highly beneficial in hospitalized patients (19–
22). The benefit of antibiotics on top of systemic steroids might
be smaller than that observed in other studies where systemic
steroids were often withheld or applied only in a minority of
patients. Second, we excluded patients with fever and patients
with chest radiographs suggestive of pneumonia. This was not
done in other placebo-controlled trials. Consequential enroll-
ment of patients with pneumonia might have inflated the
observed treatment effects. Third, we studied patients with
moderate to severe COPD with exacerbations that required
hospitalization. The severe nature of these exacerbations might
have caused a large proportion to relapse early, thereby
attenuating the treatment effect. A fourth explanation could
be insufficient antibacterial activity of doxycycline. Tetracy-
clines were used in the majority of placebo-controlled trials so
far and the resistance rates of the commonly isolated bacterial
pathogens in our region are low. Nonetheless, although persis-
tence rates for S. pneumoniae and M. catarrhalis were low, we
did prove persistence in 31% of patients with H. influenzae who
received doxycycline whereas the in vitro resistance rate of
H. influenzae was only 1%.
TABLE 1. BASELINE PATIENT CHARACTERISTICS
(n 5 107)
(n 5 116)
Male sex, no. (%)
Smokers, no. (%)
Current smokers, no. (%)
Pack-years, median (IQR)
Body mass index*
Percent predicted FEV1, %†
Percent predicted FVC, %†
Comorbidities, no. (%)
Ischemic heart disease
C-reactive protein (mg/L), median (IQR)
PaO2(mm Hg), median (IQR)
PaCO2(mm Hg), median (IQR)
Breathing frequency, median (IQR)
ICS, no. (%)
SCS, for maintenance use, no. (%)
SCS, course for current AECOPD, no. (%)
71.0 6 10.2
25.0 6 5.7
43.9 6 17.2
71.1 6 17.7
72.8 6 9.2
24.7 6 6.8
46.9 6 18.5
72.7 6 18.6
Definition of abbreviations: AECOPD 5 acute exacerbation of COPD; FEV15 postbronchodilator forced expiratory volume in 1
second; ICS 5 inhalation corticosteroids; IQR 5 interquartile range; SCS 5 systemic corticosteroids.
Plus–minus values represent means 6 SD.
* The body mass index is the weight in kilograms divided by the square of the height in meters.
†Last recorded postbronchodilator value in a stable state before admission.
Daniels, Snijders, de Graaff, et al.: Antibiotic Therapy for Acute Exacerbations of COPD153
Persistence of H. influenzae strains after antibiotic therapy,
even if the strain is susceptible to the prescribed antibiotic, is
a well-known phenomenon (27, 28). An in vitro study showed
that penetration of H. influenzae between epithelial cells pro-
tects the bacteria from antibody-mediated defense mechanisms
and antibiotics (29). There is sufficient evidence to suggest that
fluoroquinolones such as moxifloxacin outperform conventional
antibiotics in terms of bacterial eradication, especially of H.
influenzae. In spite of this there is no evidence of clinical supe-
riority over conventional antibiotics and only limited evidence
of superiority in long-term outcomes such as time to the next
exacerbation (30). In light of this evidence, it seems unjustified to
advocate the use of quinolones for AECOPD in regions with
acceptable resistance rates to conventional antibiotics.
As doxycycline was not superior to placebo in the overall
analysis, it is important to assess whether certain subgroups
benefit from antibiotics. An important finding of Anthonisen
and colleagues (15) was that antibiotics are most effective in
patients with increased sputum purulence (type 1 AECOPD).
In the current study, we found that doxycycline was superior on
Day 10 in patients with a type 1 exacerbation and equivalent in
patients with a type 2 exacerbation. The treatment effect, how-
ever, did not differ significantly between these groups (P 5
0.14). We therefore cannot claim interaction between the
treatment effect of doxycycline and the type of exacerbation.
Because sputum purulence is a marker for bacterial infection,
TABLE 2. EFFECTS OF INTERVENTION ON PRIMARY AND SECONDARY END POINTS IN THE INTENTION-TO-TREAT POPULATION
(n 5 128)
(n 5 137)
Odds Ratio or Mean Difference
(95% CI)* End PointP Value*
Clinical success on Day 30, no. (%)
Clinical success on Day 10, no. (%)
Clinical cure on Day 10, no. (%)
Clinical cure on Day 30, no. (%)
Open-label antibiotic treatment for lack of efficacy, no. (%)
On Day 10
On Day 30
Change on Day 10
Change on Day 30
Serum CRP, mg/L
Change on Day 10
Change on Day 30
Change on Day 10
Change on Day 30
Change on Day 10
Change on Day 30
Change on Day 10
Change on Day 30
Change on Day 10
Change on Day 30
Change on Day 10
Change on Day 30
78 (61)72 (53)1.3 (0.8 to 2.0) 0.32
1.9 (1.1 to 3.2)
1.9 (1.2 to 3.2)
1.4 (0.9 to 2.3)
0.5 (0.3 to 0.9)
0.7 (0.4 to 1.1)
0.05 (20.02 to 0.12)
0.07 (20.03 to 0.13)
215.9 (232.9 to 1.1)
10.3 (212.8 to 33.4)
22.3 (23.9 to 20.8)
21.0 (23.7 to 1.8)
20.3 (20.9 to 0.3)
20.5 (21.3 to 0.3)
20.7 (21.6 to 0.1)
20.3 (21.3 to 0.7)
21.1 (21.8 to 20.3)
20.4 (21.3 to 0.6)
21.3 (22.1 to 20.5)
20.1 (21.2 to 0.9)
Definition of abbreviations: 95% CI 5 95% confidence interval; CRP 5 C-reactive protein.
Values are listed as mean 6 SD unless stated otherwise.
* Corrected for within-patient clustering.
†Symptoms scores were assessed with a visual analog scale (scale 1–10).
Figure 2. Kaplan-Meier curves showing the effect of the intervention
on time to treatment failure in the intention-to-treat population.
*Corrected for within-patient clustering.
154 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINEVOL 1812010
we also established subgroups according to the presence of
a bacterial pathogen in expectorated sputum. Again, we found
no interaction with the treatment effect. This finding suggests
that the presence of bacteria in the sputum of patients with
AECOPD does not necessarily represent a new infection that
warrants treatment with antibiotics. It is evident that the
airways of patients with COPD are often colonized with bac-
teria and that increase in bacterial load or even acquisition of a
TABLE 3. BACTERIOLOGICAL RESPONSE ON DAY 10 IN SUBJECTS FROM THE INTENTION-TO-TREAT
POPULATION WITH BACTERIAL INFECTION
Doxycycline (n 5 78)
Placebo (n 5 73)
(95% CI)*End PointnP Value*
Overall success, no. (%)
Success per pathogen
15152/78 (67)25/73 (34) 3.8 (1.9 to 7.5)
3.5 (1.4 to 8.5)
6.9 (1.9 to 24.8)
5.1 (1.4 to 18.9)
0/1 (0)1/2 (50)
A potential pathogen was identified in 158 exacerbations. The bacteriological response could be evaluated in 151
* Corrected for within-patient clustering.
†Serologic tests were used for the diagnosis of M. pneumoniae and C. pneumoniae infection.
on (A) Day 10 and (B) Day
30. Shown are treatment ef-
fect (solid squares), 95% confi-
lines), P values, and P values
for interaction between the
treatment effect and the sub-
group variable. *Type of exac-
erbation was defined according
to Anthonisen and colleagues
prespecified at 10 mg/L. In-
clearly altered the results, with
the best results at 50 mg/L;
‡predicted FEV1% 5 postbron-
chodilator forced expiratory
volume in 1 second, % of
predicted (last recorded value
in a stable state before admis-
sion). The cutoff value of pre-
dicted FEV1% was prespecified
at 50%. Lowering the cutoff
value did not alter the results.
AECOPD 5 acute exacerba-
tion of COPD;xcorrected for
†the cutoff value was
Daniels, Snijders, de Graaff, et al.: Antibiotic Therapy for Acute Exacerbations of COPD155
new strain does not necessarily lead to an exacerbation (5, 8–11,
31–33). We also constructed subgroups according to serum CRP
levels. We hypothesized that infection with a new strain, as op-
posed to colonization, results in more pronounced systemic
inflammation, which was confirmed by others (34). Whereas we
observed that doxycycline was superior on Day 10 in patients
with a CRP value equal to or exceeding 50 mg/L, it was equiv-
alent in patients with a CRP value less than 50 mg/L. Although
we were unable to prove interaction between serum CRP levels
and treatment effect, this might suggest that only patients with
marked systemic inflammation harbor a bacterial infection that
requires antimicrobial therapy. These findings bear resemblance
with evidence that antibiotic treatment guided by levels of
procalcitonin, a systemic marker of bacterial infection, safely
reduces antibiotic use (35, 36). Although CRP has been pro-
posed as a marker of infection in AECOPD (37, 38), this is the
first randomized controlled trial that indicates its possible value
in selecting patients for antibiotic therapy. CRP-guided therapy
would reduce the use of antibiotics because in our population
CRP was at least 50 mg/L in only 41% of exacerbations, whereas
67% of exacerbations were type 1. Because we observed only
a trend, possibly because the current trial was not powered to
investigate a difference of treatment effect between subgroups
according to CRP value, additional studies are needed to further
investigate the role of CRP in the management of AECOPD.
The finding of the present study must be interpreted in the
context of several potential limitations. The two-center design
and the absence of advanced antimicrobial resistance in our
region might affect the generalizability to other populations.
Advantages of the two-center design include efficiency and
influence outcome such as disease severity. In the subgroup
analysis, however, we found that the treatment effect was not
affected by disease severity. Finally, the symptom scores assessed
there is evidence that the VAS can reproducibly measure symp-
In conclusion, doxycycline is equivalent to placebo in terms
of clinical response on Day 30, but superior in terms of clinical
success and clinical cure on Day 10 as well as microbiological
success on Day 10. Additional effects of doxycycline include
reduction of open label antibiotic therapy for lack of efficacy
and a greater reduction in symptoms. Subgroup analysis revealed
no interaction between the subgroup variables and the treatment
effect, although there was a trend for patients with a high CRP
value to benefit from antibiotic therapy.
Conflict of Interest Statement: None of the authors has a financial relationship
with a commercial entity that has an interest in the subject of this manuscript.
Acknowledgment: The authors thank the patients who enrolled in this study, and
the physicians, nurses, and secretaries of the Department of Pulmonary Diseases
for important contributions. The authors thank T. van der Ploeg and D. L. Knol
for work on the statistical analysis and T. Tossijn-Groot and P. Singer of the
Department of Microbiology for their hard work.
1. Murray CJ, Lopez AD. Evidence-based health policy: lessons from the
Global Burden of Disease Study. Science 1996;274:740–743.
2. Seemungal TA, Donaldson GC, Bhowmik A, Jeffries DJ, Wedzicha JA.
Time course and recovery of exacerbations in patients with chronic
obstructive pulmonary disease. Am J Respir Crit Care Med 2000;161:
3. Fletcher CM, Peto R, Tinker CM, Speizer FE. Natural history of chronic
bronchitis and emphysema. Oxford: Oxford University Press; 1976.
4. White AJ, Gompertz S, Stockley RA. Chronic obstructive pulmonary
disease. 6. The aetiology of exacerbations of chronic obstructive pul-
monary disease. Thorax 2003;58:73–80.
5. Monso E, Ruiz J, Rosell A, Manterola J, Fiz J, Morera J, Ausina V.
Bacterial infection in chronic obstructive pulmonary disease: a study
of stable and exacerbated outpatients using the protected specimen
brush. Am J Respir Crit Care Med 1995;152:1316–1320.
6. Pela R, Marchesani F, Agostinelli C, Staccioli D, Cecarini L, Bassotti C,
Sanguinetti CM. Airways microbial flora in COPD patients in stable
clinical conditions and during exacerbations: a bronchoscopic investiga-
tion. Monaldi Arch Chest Dis 1998;53:262–267.
7. Soler N, Torres A, Ewig S, Gonzalez J, Celis R, El-Ebiary M, Hernandez
C, Rodriguez-Roisin R. Bronchial microbial patterns in severe
exacerbations of chronic obstructive pulmonary disease (COPD)
requiring mechanical ventilation. Am J Respir Crit Care Med 1998;
8. Riise GC, Larsson S, Larsson P, Jeansson S, Andersson BA. The
intrabronchial microbial flora in chronic bronchitis patients: a target
for N-acetylcysteine therapy? Eur Respir J 1994;7:94–101.
9. Cabello H, Torres A, Celis R, El-Ebiary M, Puig de la Bellacasa J,
Xaubet A, Gonza ´lez J, Agustı ´ C, Soler N. Bacterial colonization of
distal airways in healthy subjects and chronic lung disease: a broncho-
scopic study. Eur Respir J 1997;10:1137–1144.
10. McHardy VU, Inglis JM, Calder MA, Crofton JW, Gregg I, Ryland DA,
Taylor P, Chadwick M, Coombs D, Riddell RW. A study of infective
and other factors in exacerbations of chronic bronchitis. Br J Dis
11. Gump DW, Phillips CA, Forsyth BR, McIntosh K, Lamborn KR, Stouch
WH. Role of infection in chronic bronchitis. Am Rev Respir Dis 1976;
12. Ram FS, Rodriguez-Roisin R, Granados-Navarrete A, Garcia-Aymerich
J, Barnes NC. Antibiotics for exacerbations of chronic obstructive
pulmonary disease. Cochrane Database Syst Rev 2006;19:CD004403.
13. McCrory DC, Brown C, Gelfand SE, Bach PB. Management of acute
exacerbations of COPD: a summary and appraisal of published
evidence. Chest 2001;119:1190–1209.
14. Saint S, Bent S, Vittinghoff E, Grady D. Antibiotics in chronic obstruc-
tive pulmonary disease exacerbations: a meta-analysis. JAMA 1995;
15. Anthonisen NR, Manfreda J, Warren CP, Hershfield ES, Harding GK,
Nelson NA. Antibiotic therapy in exacerbations of chronic obstruc-
tive pulmonary disease. Ann Intern Med 1987;106:196–204.
16. Hirschmann JV. Do bacteria cause exacerbations of COPD? Chest 2000;
17. Wilson R. Treatment of COPD exacerbations: antibiotics. Eur Respir
18. Hirschmann JV. Bacteria and COPD exacerbations redux. Chest 2001;
19. Quon BS, Gan WQ, Sin DD. Contemporary management of acute
exacerbations of COPD: a systematic review and metaanalysis. Chest
20. Niewoehner DE, Erbland ML, Deupree RH, Collins D, Gross NJ, Light
RW, Anderson P, Morgan NA. Department of Veterans Affairs Co-
operative Study Group. Effect of systemic glucocorticoids on exac-
erbations of chronic obstructive pulmonary disease. N Engl J Med
21. Davies L, Angus RM, Calverley PM. Oral corticosteroids in patients
admitted to hospital with exacerbations of chronic obstructive pul-
monary disease: a prospective randomised controlled trial. Lancet
22. Maltais F, Ostinelli J, Bourbeau J, Tonnel AB, Jacquemet N, Haddon J,
Rouleau M, Boukhana M, Martinot JB, Duroux P. Comparison of
nebulized budesonide and oral prednisolone with placebo in the
treatment of acute exacerbations of chronic obstructive pulmonary
disease: a randomized controlled trial. Am J Respir Crit Care Med
23. Sachs AF, Koe ¨ter GH, Groenier KH, van der Waaij D, Schiphuis J,
Meyboom-de Jong B. Changes in symptoms, peak expiratory flow,
and sputum flora during treatment with antibiotics of exacerbations in
patients with chronic obstructive pulmonary disease in general
practice. Thorax 1995;50:758–763.
24. Daniels JM, Snijders D, Vlaspolder F, de Graaff CS, Jansen HM,
Boersma WG. Antibiotics versus placebo for the treatment of severe
acute exacerbations of COPD [abstract]. Am J Respir Crit Care Med
156 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINEVOL 1812010
25. Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Global Download full-text
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.
26. Chow AW, Hall CB, Klein JO, Kammer RB, Meyer RD, Remington JS.;
budesonide Infectious Diseases Society of America and the Food
and Drug Administration. Evaluation of new anti-infective drugs for
the treatment of respiratory tract infections. Clin Infect Dis 1992;15:
27. Groeneveld K, van Alphen L, Eijk PP, Visschers G, Jansen HM, Zanen
HC. Endogenous and exogenous reinfections by Haemophilus influ-
enzae in patients with chronic obstructive pulmonary disease: the
effect of antibiotic treatment on persistence. J Infect Dis 1990;161:
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. van Schilfgaarde M, Eijk P, Regelink A, van Ulsen P, Everts V,
Dankert J, van Alphen L. Haemophilus influenzae localized in epi-
thelial cell layers is shielded from antibiotics and antibody-mediated
bactericidal activity. Microb Pathog 1999;26:249–262.
30. Siempos II, Dimopoulos G, Korbila IP, Manta K, Falagas ME. Macro-
lides, quinolones and amoxicillin/clavulanate for chronic bronchitis:
a meta-analysis. Eur Respir J 2007;29:1127–1137.
31. Sethi S, Evans N, Grant BJB, Murphy TF. New strains of bacteria and
exacerbations of chronic obstructive pulmonary disease. N Engl J
32. Bresser P, van Alphen L, Lutter R. New strains of bacteria and
exacerbations of COPD. N Engl J Med 2002;347:2077–2079.
33. Sethi S, Sethi R, Eschberger K, Lobbins P, Cai X, Grant BJ, Murphy TF.
Airway bacterial concentrations and exacerbations of chronic ob-
structive pulmonary disease. Am J Respir Crit Care Med 2007;176:
34. Sethi S, Wrona C, Eschberger K, Lobbins P, Cai X, Murphy T.
Inflammatory profile of new bacterial strain exacerbations of chronic
obstructive pulmonary disease. Am J Respir Crit Care Med 2008;177:
35. Stolz D, Christ-Crain M, Bingisser R, Leuppi J, Miedinger D, Mu ¨ller C,
Huber P, Mu ¨ller B, Tamm M. Antibiotic treatment of exacerbations
of COPD: a randomized, controlled trial comparing procalcitonin-
guidance with standard therapy. Chest 2007;131:9–19.
36. Christ-Crain M, Jaccard-Stolz D, Bingisser R, Gencay MM, Huber PR,
Tamm M, Mu ¨ller B. Effect of procalcitonin-guided treatment on
antibiotic use and outcome in lower respiratory tract infections:
cluster-randomised, single-blinded intervention trial. Lancet 2004;
37. Weis N, Almdal T. C-reactive protein—can it be used as a marker of
infection in patients with exacerbation of chronic obstructive pulmo-
nary disease? Eur J Intern Med 2006;17:88–91.
38. Dev D, Wallace E, Sankaran R, Cunniffe J, Govan JR, Wathen CG,
Emmanuel FX. Value of C-reactive protein measurements in exac-
erbations of chronic obstructive pulmonary disease. Respir Med 1998;
39. Grant S, Aitchison T, Henderson E, Christie J, Zare S, McMurray J,
Dargie H. A comparison of the reproducibility and the sensitivity to
change of visual analogue scales, Borg scales, and Likert scales in
normal subjects during submaximal exercise. Chest 1999;116:1208–1217.
Daniels, Snijders, de Graaff, et al.: Antibiotic Therapy for Acute Exacerbations of COPD 157