new england

The
journal of medicine
established in 1812 august 25, 2011 vol. 365  no. 8

Azithromycin for Prevention of Exacerbations of COPD

Richard K. Albert, M.D., John Connett, Ph.D., William C. Bailey, M.D., Richard Casaburi, M.D., Ph.D.,
J. Allen D. Cooper, Jr., M.D., Gerard J. Criner, M.D., Jeffrey L. Curtis, M.D., Mark T. Dransfield, M.D.,
MeiLan K. Han, M.D., Stephen C. Lazarus, M.D., Barry Make, M.D., Nathaniel Marchetti, M.D.,
Fernando J. Martinez, M.D., Nancy E. Madinger, M.D., Charlene McEvoy, M.D., M.P.H.,
Dennis E. Niewoehner, M.D., Janos Porsasz, M.D., Ph.D., Connie S. Price, M.D., John Reilly, M.D.,
Paul D. Scanlon, M.D., Frank C. Sciurba, M.D., Steven M. Scharf, M.D., Ph.D., George R. Washko, M.D.,
Prescott G. Woodruff, M.D., M.P.H., and Nicholas R. Anthonisen, M.D., for the COPD Clinical Research Network

A BS T R AC T

Background
Acute exacerbations adversely affect patients with chronic obstructive pulmonary The affiliations of the authors are listed in
disease (COPD). Macrolide antibiotics benefit patients with a variety of inflamma- the Appendix. Address reprint requests to
Dr. Albert at Denver Health, 777 Bannock
tory airway diseases. St., MC 4000, Denver, CO 80204-4507, or
Methods at [email protected].
We performed a randomized trial to determine whether azithromycin decreased the This article (10.1056/NEJMoa1104623) was
frequency of exacerbations in participants with COPD who had an increased risk of updated on April 5, 2012, at NEJM.org.
exacerbations but no hearing impairment, resting tachycardia, or apparent risk of pro-
N Engl J Med 2011;365:689-98.
longation of the corrected QT interval. Copyright © 2011 Massachusetts Medical Society.
Results
A total of 1577 subjects were screened; 1142 (72%) were randomly assigned to receive
azithromycin, at a dose of 250 mg daily (570 participants), or placebo (572 participants)
for 1 year in addition to their usual care. The rate of 1-year follow-up was 89% in the
azithromycin group and 90% in the placebo group. The median time to the first exac-
erbation was 266 days (95% confidence interval [CI], 227 to 313) among participants
receiving azithromycin, as compared with 174 days (95% CI, 143 to 215) among par-
ticipants receiving placebo (P<0.001). The frequency of exacerbations was 1.48 exacerba-
tions per patient-year in the azithromycin group, as compared with 1.83 per patient-year
in the placebo group (P = 0.01), and the hazard ratio for having an acute exacerbation
of COPD per patient-year in the azithromycin group was 0.73 (95% CI, 0.63 to 0.84;
P<0.001). The scores on the St. George’s Respiratory Questionnaire (on a scale of 0 to
100, with lower scores indicating better functioning) improved more in the azithro-
mycin group than in the placebo group (a mean [±SD] decrease of 2.8±12.1 vs.
0.6±11.4, P = 0.006); the percentage of participants with more than the minimal clini-
cally important difference of −4 units was 43% in the azithromycin group, as compared
with 36% in the placebo group (P = 0.03). Hearing decrements were more common in
the azithromycin group than in the placebo group (25% vs. 20%, P = 0.04).
Conclusions
Among selected subjects with COPD, azithromycin taken daily for 1 year, when added
to usual treatment, decreased the frequency of exacerbations and improved quality of
life but caused hearing decrements in a small percentage of subjects. Although this
intervention could change microbial resistance patterns, the effect of this change is not
known. (Funded by the National Institutes of Health; ClinicalTrials.gov number,
NCT00325897.)

n engl j med 365;8  nejm.org  august 25, 2011 689

The New England Journal of Medicine
Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

A 
cute exacerbations of chronic ob- of the data and safety monitoring committee.
structive pulmonary disease (COPD) result All the authors made the decision to submit the
in frequent visits to physicians’ offices and manuscript for publication. There were no confi-
emergency rooms and numerous hospitalizations dentiality agreements with the sponsor. The study
and days lost from work; they also account for a drugs (both azithromycin and placebo) were pur-
substantial percentage of the cost of treating chased by the investigators. All the authors assume
COPD.1-5 Patients who have acute exacerbations of responsibility for the data and analyses and vouch
COPD, as compared with patients with COPD who for the fidelity of the study to the protocol.
do not have acute exacerbations, have an increased
risk of death, a more rapid decline in lung func- Study Participants
tion, and reduced quality of life.6-11 Although in- Eligible participants were at least 40 years of age,
haled glucocorticoids, long-acting beta2-agonists, had a clinical diagnosis of COPD (defined as having
and long-acting muscarinic antagonists reduce the a smoking history of at least 10 pack-years, a ratio
frequency of acute exacerbations of COPD,12-24 pa- of postbronchodilator forced expiratory volume in
tients receiving all three of these medications may 1 second [FEV1] to forced vital capacity of <70%,
still have as many as 1.4 acute exacerbations, on and a postbronchodilator FEV1 of <80% of the pre-
average, each year.23 dicted value), were either using continuous supple-
Macrolide antibiotics have immunomodulatory, mental oxygen or had received systemic glucocor-
antiinflammatory, and antibacterial effects.25 Sev- ticoids within the previous year, had gone to an
en small studies that tested whether macrolides emergency room or had been hospitalized for an
decrease the frequency of acute exacerbations of acute exacerbation of COPD,19 and had not had an
COPD reported conflicting results.26-32 According- acute exacerbation of COPD for at least 4 weeks
ly, we conducted a large, randomized trial to test before enrollment.
the hypothesis that azithromycin decreases the Exclusion criteria were asthma, a resting heart
frequency of acute exacerbations of COPD when rate greater than 100 beats per minute, a pro-
added to the usual care of these patients. longed corrected QT (QTc) interval (>450 msec),
the use of medications that prolong the QTc inter-
Me thods val or are associated with torsades de pointes (with
the exception of amiodarone),33 and hearing im-
Study Design and Oversight pairment documented by audiometric testing.
We used a prospective, parallel-group, placebo-
controlled design. Participants were randomly as- Outcomes
signed, in a 1:1 ratio, to receive azithromycin, at a The primary outcome was the time to the first acute
dose of 250 mg orally, or an identical-appearing exacerbation of COPD, with acute exacerbation of
placebo, once daily. Participants were recruited COPD defined as “a complex of respiratory symp-
from 17 sites associated with 12 academic health toms (increased or new onset) of more than one of
centers in the United States. Written informed con- the following: cough, sputum, wheezing, dyspnea,
sent was obtained from all participants. or chest tightness with a duration of at least 3 days
The study was approved by the institutional re- requiring treatment with antibiotics or systemic
view board at each participating institution. The steroids.”19 At each clinic visit and telephone con-
protocol was designed by the first author and was tact, study personnel determined whether an acute
modified on the basis of input from the remaining exacerbation of COPD had occurred in the previous
authors. The complete protocol, including the sta- month. The date of each acute exacerbation was
tistical analysis plan, is available with the full text taken as the date treatment was prescribed.
of this article at NEJM.org. The data were gathered Secondary outcomes included quality of life,
by study personnel at each participating site, over- nasopharyngeal colonization with selected respi-
seen by one of the authors at each site. The data ratory pathogens (i.e., Staphylococcus aureus, Strepto-
were analyzed by the second author, who is a stat- coccus pneumoniae, haemophilus species, and morax-
istician, together with the first author. All the ella species), and adherence to taking the study
authors participated in interpreting the data. The drug as prescribed. The St. George’s Respiratory
first and final drafts of the manuscript were writ- Questionnaire (SGRQ, in which scores range from
ten by the first author and revised on the basis of 0 to 100, with lower scores indicating better func-
input from the other authors and from members tioning) and the Medical Outcomes Study 36-Item

690 n engl j med 365;8  nejm.org  august 25, 2011

The New England Journal of Medicine

Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 Azithromycin for Prevention of COPD Exacerbations

Short-Form Health Survey (SF-36) were adminis- the data were analyzed with the use of group se-
tered at the time of enrollment and at study quential testing36 that allowed “spending” a little
months 6 and 12. The minimal clinically impor- of the alpha at each interim analysis such that, at
tant difference in the SGRQ score was considered the end of the study, the total type I error did not
to be −4 units, and the percentage of patients with exceed 0.05.
a change of at least 4 units from the time of enroll-
ment was a prespecified end point.34 Deep naso- R e sult s
pharyngeal swabs were obtained at the time of
enrollment and every 3 months thereafter, and Study Participants
selected respiratory pathogens were assessed for The screening, randomization, and follow-up of pa-
resistance to macrolides. Adherence to the study tients are shown in Figure 1. The first site started
drug was assessed at each clinic visit by means of enrolling participants in March 2006, and the last
pill counts performed by the staff. Hearing was patient finished the 1-year follow-up assessment
assessed by means of audiometry at the time of on June 30, 2010. The characteristics of the partici-
enrollment and at 3 and 12 months, or whenever a pants at the time of enrollment are summarized in
patient reported worsening hearing or tinnitus. Table 1. All reported results were prespecified.

Statistical Analysis Primary Outcome

We estimated that with enrollment of 1130 sub- A life-table analysis showed that the risk of acute
jects, the study would have 90% power to show a exacerbations of COPD was reduced among par-
significant difference between the two groups in ticipants receiving azithromycin (P<0.001) (Fig. 2).
the time to the first acute exacerbation of COPD, The median time to the first acute exacerbation of
assuming that 50% of the participants in the con- COPD was 266 days (95% confidence interval [CI],
trol group19 and 40% in the azithromycin group25 227 to 313) in participants receiving azithromycin,
would have an acute exacerbation, that the rate of as compared with 174 days (95% CI, 143 to 215) in
nonadherence would be 20%, and that 6% of par- participants receiving placebo (P<0.001). The haz-
ticipants would die or be lost to follow-up during ard ratio of having an acute exacerbation of COPD
the study (extrapolated from Niewoehner et al.19), per patient-year in the azithromycin group as com-
with a two-sided type I error of 0.05. pared with the placebo group was 0.73 (95% CI,
The groups were compared with the use of an 0.63 to 0.84; P<0.001). These differences remained
intention-to-treat survival analysis. The primary significant after adjustment with the use of Cox
analysis was based on a log-rank test of the differ- regression for differences in sex, FEV1, age, smok-
ence between the two treatment groups in the time ing status, and study center. The rates of acute ex-
to the first exacerbation, with no adjustments for acerbations of COPD differed according to center,
baseline covariates. A Cox proportional-hazards but the hazard ratio for the time to the first acute
model was used to adjust for differences in pre- exacerbation of COPD, stratified according center,
specified, prerandomization factors that might was 0.71 (95% CI, 0.61 to 0.83; P<0.001).
predict the risk of acute exacerbations of COPD. A total of 1641 acute exacerbations of COPD
Bootstrap methods were used to compute confi- occurred during the study — 741 among the 558
dence intervals for median times to the first ex- participants who received at least one dose of
acerbation and for the difference in median times azithromycin and 900 among the 559 who received
between the two groups.35 The rates of acute ex- at least one dose of placebo, and the rates of acute
acerbations of COPD were determined by dividing exacerbations of COPD per patient-year were 1.48
the number of acute exacerbations by the person- and 1.83, respectively (P = 0.01; rate ratio from
years of follow-up and were compared with the use negative binomial analysis, 0.83; 95% CI, 0.72 to
of both Poisson and negative binomial analyses. 0.95). The frequency of acute exacerbations was
The data and safety monitoring board met ap- lower among participants receiving azithromycin
proximately every 6 months and had the authority than among those receiving placebo regardless of
to stop the study prematurely on the basis of any the rate of acute exacerbations per patient-year
of the interim analyses and on the basis of calcu- (P = 0.008 by both Poisson and negative binomial
lations of conditional power derived from a futil- models) (Fig. 3). The number needed to treat to
ity analysis supplied at each analysis. Accordingly, prevent one acute exacerbation of COPD was 2.86.

n engl j med 365;8  nejm.org  august 25, 2011 691

The New England Journal of Medicine
Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

1577 Patients were screened

435 Were excluded

102 Had cardiac issues
59 Had FEV1:FVC >70%
47 Had FEV1 >80%
33 Were not willing to return
27 Had hearing deficit
27 Declined to participate
140 Had other reason

1142 Underwent randomization

570 Were assigned to receive 572 Were assigned to receive

azithromycin placebo

12 Had no follow-up visit 13 Had no follow-up visit

558 Were included in primary analysis 559 Were included in primary analysis

32 Withdrew 28 Withdrew
6 (1%) Lost interest 1 (0.2%) Lost interest
2 (0.3%) Were unwilling to follow 3 (0.5%) Were unwilling to follow
protocol protocol
3 (0.5%) Had clinic-access issues 1 (0.2%) Had clinic-access issues
6 (1%) Had medical conditions 3 (0.5) Moved out of area
2 (0.4%) Withdrew consent 8 (0.2%) Had medical conditions
9 (1%) Had adverse event 1 (0.2%) Withdrew consent
4 (0.7%) Had other reason 4 (0.7%) Had adverse event
13 (2.3%) Were lost to follow-up 7 (1.3%) Had other reason
18 (3%) Died 9 (1.6%) Were lost to follow-up
20 (3.5%) Died

495 (89%) Completed 12-month 502 (90%) Completed 12-month

or washout visit or washout visit
19 (3%) Attended washout visit 16 (3%) Attended washout visit
but not 12-month visit but not 12-month visit

Figure 1. Screening, Randomization, and Follow-up.

Patients who completed the 12-month course of the study drug were asked to return 1 month later for a washout visit. FEV1 denotes
forced expiratory volume in 1 second, and FVC forced vital capacity.

Secondary Outcomes mentary Appendix, available at NEJM.org. The to-

The effect of azithromycin on the secondary out- tal SGRQ scores recorded at 1 year decreased a
comes is summarized in Table 2. The effect of the mean (±SD) of 2.8±12.1 units in the azithromycin
Global Initiative for Chronic Obstructive Lung Dis- group, as compared with a mean of 0.6±11.4 units
ease (GOLD) stage (which ranges from stage I in the placebo group (P = 0.006) (see Section C in
COPD, indicating mild disease, to stage IV COPD, the Supplementary Appendix for individual scale
indicating very severe disease) on the rate of acute scores). Although this mean change did not exceed
exacerbations of COPD and on frequency of hospi- the minimal clinically important difference of at
talizations is presented in Section B in the Supple- least 4 units, more participants in the azithromycin

692 n engl j med 365;8  nejm.org  august 25, 2011

The New England Journal of Medicine

Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 Azithromycin for Prevention of COPD Exacerbations

group than in the placebo group had a decrease of

Table 1. Baseline Clinical and Demographic Characteristics of the Patients.*
at least 4 units in their SGRQ score (43% vs. 36%,
P = 0.03). No consistent changes were seen in the Azithromycin Placebo
scores on the SF-36 (see Section D in the Supple- Characteristic (N = 558) (N = 559)
mentary Appendix). The mean rate of adherence to Age — yr 65±9 66±8
the study medication was 67.3% in the azithromy- Female sex — no. (%) 229 (41) 227 (41)
cin group and 66.9% in the placebo group (P = 0.84). Race or ethnic group — no. (%)†
White 456 (82) 449 (80)
Subgroup Analyses
Black 75 (13) 86 (15)
Analyses were performed according to 22 sub-
Other 19 (3) 22 (4)
groups; the results are provided in Section I in the
Supplementary Appendix. The subgroup analyses Multiethnic 19 (3) 16 (3)
showed that the response to azithromycin seemed Postbronchodilator FEV1
to vary according to age (≤65 vs. >65 years), smok- Liters 1.10±0.50 1.12±0.52
ing status (former smoker vs. current smoker), use % of predicted value 39±16 40±16
or nonuse of oxygen, GOLD stage, and use or non- Ratio of FEV1 to FVC — % 42±13 43±13
use of inhalers. GOLD stage — no. (%)‡
II 144 (26) 148 (26)
Adverse Events
III 225 (40) 226 (40)
The rate of death from any cause was 3% in the
azithromycin group and 4% in the placebo group IV 188 (34) 182 (33)
(P = 0.87). The rate of death from respiratory causes Smoking history — pack-yr 58±32 59±32
was 2% and 1% in the two groups, respectively Current smoker — no. (%) 119 (21) 127 (23)
(P = 0.48), and the rate of death from cardiovas- Medications for COPD — no. (%)
cular causes was 0.2% in both groups (P = 1.00) Inhaled glucocorticoids only 21 (4) 36 (6)
(see Section H in the Supplementary Appendix). LAMAs only 34 (6) 43 (8)
No significant differences were observed in the LABAs only 15 (3) 6 (1)
frequency of serious adverse events or of adverse
Inhaled glucocorticoids and LABAs 104 (19) 125 (22)
events leading to discontinuation of the study
Inhaled glucocorticoids and LAMAs 23 (4) 28 (5)
drug, but an audiogram-confirmed hearing decre­
ment occurred in 142 of the participants receiv- LABAs and LAMAs 30 (5) 23 (4)
ing azithromycin (25%), as compared with 110 of Inhaled glucocorticoids, LABAs, and LAMAs 273 (49) 255 (46)
those receiving placebo (20%) (P = 0.04). A small None 58 (10) 43 (8)
but significant between-group difference was ob- Entry criteria
served in the mean age-adjusted hearing thresholds Acute exacerbation of COPD in previous 278 (50) 283 (51)
for the four sound frequencies from enrollment to 12 mo requiring hospitalization or
month 3, with patients in the azithromycin group ­emergency room visit — no. (%)
having more pronounced hearing decrements (see Systemic glucocorticoids for acute exacerbation 467 (84) 477 (85)
of COPD in previous 12 mo — no. (%)
Section F in the Supplementary Appendix). In
80 participants receiving azithromycin and in 45 Long-term oxygen — no. (%) 334 (60) 328 (59)
receiving placebo, the hearing decrement occurred Long-term oxygen as the only criterion — no. (%) 72 (13) 65 (12)
before the 12-month visit, providing the opportu-
* Plus–minus values are means ±SD. The total numbers in the two groups are
nity to determine whether hearing returned within the number of participants who were randomly assigned to the group (570 to
a minimum of 1 month after discontinuation of the azithromycin group and 572 to the placebo group) minus those who did
the study drug. Although all of these participants not return for any follow-up assessment (12 in the azithromycin group and the
13 in the placebo group). There were no significant between-group differences
should have had their study drug discontinued, (P>0.05 for all comparisons). COPD denotes chronic obstructive pulmonary
the drug was discontinued in only 61 participants disease, FEV1 forced expiratory volume in 1 second, FVC forced vital capacity,
in the azithromycin group (76%) and 37 in the pla- LABA long-acting beta2 agonist, and LAMA long-acting muscarinic antagonist.
† Race or ethnic group was self-reported. Other groups included Native Americans,
cebo group (82%), owing to protocol errors. Subse- Asians, Hispanics, and Pacific Islanders. A breakdown according to these groups
quent audiograms showed that hearing improved is provided in Section A in the Supplementary Appendix, available at NEJM.org.
to the baseline level in 21 of the 61 participants ‡ The Global Initiative for Chronic Obstructive Lung Disease (GOLD) stage ranges
from stage I COPD, indicating mild disease, to stage IV COPD, indicating very
(34%) who discontinued azithromycin and in 6 of severe disease.
the 19 (32%) who did not, as well as in 14 of the

n engl j med 365;8  nejm.org  august 25, 2011 693

The New England Journal of Medicine
Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

1.0

0.9

0.8
Proportion Free of COPD Exacerbations

0.7

0.6

0.5
Azithromycin

0.4

0.3 Placebo

0.2

0.1 P<0.001 by log-rank test and Wilcoxon signed-rank test

0.0
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360

Follow-up (days)

Figure 2. Proportion of Participants Free from Acute Exacerbations of Chronic Obstructive Pulmonary Disease
(COPD) for 1 Year, According to Study Group.
The analyses were based on the participants who were randomly assigned to the group minus those who did not
return for any follow-up assessment — 558 participants in the azithromycin group, of whom 317 (57%) had an acute
exacerbation, and 559 in the placebo group, of whom 380 (68%) had an acute exacerbation.

37 participants (38%) who discontinued placebo colonization either at the time of enrollment or
and in 2 of the 8 (25%) who did not. any time thereafter and the occurrence of acute
exacerbations of COPD (P = 0.31).
Nasopharyngeal Colonization Cultures from 56% of the participants in the
and Resistance to Macrolides azithromycin group and 59% in the placebo group
Nasopharyngeal swabs were obtained at 85% and who had selected respiratory pathogens cultured
84% of the clinic visits attended by participants in from their nasopharyngeal swabs at the time of
the azithromycin group and placebo group, re- enrollment were available for susceptibility testing
spectively. A total of 79 of the 558 participants (P = 0.68); the remaining cultures were not tested
who were randomly assigned to receive azithro- because of protocol errors. The prevalence of re-
mycin and had at least one follow-up visit (14%) sistance to macrolides was 52% and 57% in the
and 83 of the 559 participants who were random- two groups, respectively (P = 0.64). Cultures from
ly assigned to receive placebo and had at least one 68% of the participants in the azithromycin group
follow-up visit (15%) were colonized with selected and 70% in the placebo group who were not colo-
respiratory pathogens at the time of enrollment nized with selected respiratory pathogens at the
(P = 0.81). A total of 66 of the participants in the time of enrollment but who became colonized
azithromycin group (12%) and 172 in the placebo during the course of the study were available for
group (31%) who had not had nasopharyngeal susceptibility testing (P = 0.76), and the incidence of
colonization at the time of enrollment became resistance to macrolides was 81% and 41% in the
colonized during the course of the study (P<0.001). two groups, respectively (P<0.001) (Section G in
No association was seen between nasopharyngeal the Supplementary Appendix).

694 n engl j med 365;8  nejm.org  august 25, 2011

The New England Journal of Medicine

Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 Azithromycin for Prevention of COPD Exacerbations

Discussion
250
Azithromycin Placebo
Among subjects at increased risk for acute exacer-
bations of COPD who received azithromycin, at a
200 P<0.001 by Poisson analysis
dose of 250 mg once daily, for 1 year in addition to P=0.004 by negative binomial analysis
their usual care, the frequency of acute exacerba-

No. of Participants
tions was decreased. This decrease was accompa- 150

nied by a decrease in the incidence of colonization

with selected respiratory pathogens and improved 100
quality of life, but also an increase in the incidence
of colonization with macrolide-resistant organisms
50
and an excess rate of hearing decrements of ap-
proximately 5%.
Seven previous studies have evaluated whether 0
0 0.01–1.50 1.51–3.00 3.01–4.50 4.51–6.00 6.01–7.50 ≤7.51
macrolide antibiotics decrease the risk of acute
Rates of Exacerbations per Person-Yr
exacerbations of COPD. Two of the studies showed
no effect, but one of these used a retrospective
Figure 3. Rates of Acute Exacerbations of Chronic Obstructive Pulmonary
design29 and the other was conducted for only Disease per Person-Year, According to Study Group.
3 months, and very few acute exacerbations of
COPD occurred in either group.26 Five studies
have reported that macrolides decrease acute ex- plication.26-32 We found that more participants
acerbations of COPD, but one of these was not a receiving azithromycin met the criteria for devel-
blinded study,27 two did not include concurrent opment of a hearing decrement than did those
controls,31,32 and one involved only 35 patients.30 receiving placebo, but the improvements in hear-
Seemungal and colleagues performed a well- ing that occurred on repeat testing, regardless of
designed, randomized, 1-year trial of erythromycin, whether the study drug was discontinued, suggest
at a dose of 250 mg twice daily, in 109 patients (a that our criteria were too stringent and that the
study that was stopped before the target enroll- incidence of hearing decrements was overestimat-
ment was met).28 The relative rate of acute exac- ed in both groups.
erbations of COPD among the treated participants Participants receiving azithromycin were less
was 0.65. The median time to the first acute ex- likely to become colonized with respiratory patho-
acerbation among the participants who received gens but were more likely to become colonized
erythromycin was 271 days, similar to the 266 with macrolide-resistant organisms (contrary to
days we found among the participants in our study the findings of Seemungal and colleagues28). De-
who received azithromycin. However, Seemungal spite this, we found no evidence suggesting that
and colleagues found that in the control group the colonization increased the incidence of acute exac-
median time to the first acute exacerbation was erbations of COPD or pneumonia, consistent with
89 days, whereas in our study it was 174 days, prior observations in patients with cystic fibro-
perhaps because nearly 40% of the participants in sis.37,38 Although we saw no adverse cardiac effects
their study had had three or more acute exacerba- of azithromycin, risk factors for the prolongation
tions in the year before enrollment, more were of the QTc interval were assessed before enroll-
current smokers, and fewer were receiving long- ment and participants who were thought to be at
acting muscarinic antagonists. risk were excluded.
Since approximately 80% of our participants We chose a 250-mg dose of azithromycin be-
were taking inhaled glucocorticoids with or with- cause we thought that it was high enough to limit
out long-acting beta2-agonists or long-acting mus- the possibility that a negative result might occur
carinic antagonists throughout the study, the abil- because of insufficient dosing. We administered
ity of azithromycin to decrease the frequency of the dose daily, rather than less frequently, to fa-
acute exacerbations would seem to be additive to cilitate adherence. It is possible that lower doses
these other therapies. or less frequent administration could have pro-
None of the previous studies of the effect of duced similar results.
macrolides on acute exacerbations of COPD either Sputum samples are preferred for the assess-
assessed or reported hearing problems as a com- ment of bacterial colonization. When we began the

n engl j med 365;8  nejm.org  august 25, 2011 695

The New England Journal of Medicine
Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 2. Effect of Treatment for Chronic Obstructive Pulmonary Disease (COPD) on Hospitalization Rates, Emergency Department
or Urgent Care Visits, and Unscheduled Office Visits.

Hazard Ratio
Event Azithromycin Placebo P Value* (95% CI)† P Value†
mean events/ mean events/
patient-yr patient-yr
no. of events (95% CI) no. of events (95% CI)
Hospitalization for any 323 0.74 (0.60–0.89) 329 0.95 (0.76–1.18) 0.13 0.94 (0.76–1.15) 0.52
cause
Hospitalization related to 156 0.34 (0.26–0.43) 200 0.49 (0.31–0.67) 0.14 0.82 (0.64–1.07) 0.15
COPD
Emergency department or 199 0.43 (0.34–0.53) 257 0.48 (0.39–0.57) 0.47 0.81 (0.63–1.04) 0.09
urgent care visit
Unscheduled office visit 1202 2.46 (2.08–2.48) 1345 2.57 (2.21–2.60) 0.048 0.85 (0.74–0.98) 0.02
Intubations 11 0.02 (0.01–0.04) 16 0.04 (0.01–0.06) 0.23 0.79 (0.04–1.75) 0.56

* The P value is for the rate of events per patient-year.

† The hazard ratio and P value are for the time to the first event in the azithromycin group as compared with the placebo group.

study, we obtained both expectorated sputum hearing abnormalities that place them below the
samples and nasopharyngeal swabs because we 95th percentile of patients of similar age.
knew that some patients would not be able to pro- In summary, we found that adding azithromy-
duce sputum. By the third month, however, less cin, at a dose of 250 mg daily, for 1 year to the
than 15% of participants had been able to produce usual treatment of patients who have an increased
sputum, causing us to restrict our subsequent as- risk of acute exacerbations of COPD but no hear-
sessments of colonization to an assessment of ing impairment, resting tachycardia, or apparent
deep nasopharyngeal swabs. Patel and colleagues39 risk of QTc prolongation decreased the frequency
found that 52% of patients with frequent acute of acute exacerbations of COPD and the incidence
exacerbations of COPD had induced sputum that of colonization with selected respiratory pathogens
was most commonly colonized with Haemophilus and improved quality of life but increased the inci-
influenzae, S. pneumoniae, H. parainfluenzae, and Mo- dence of colonization with macrolide-resistant or-
raxella catarrhalis. Although we found a much lower ganisms and decreased hearing in a small percent-
rate of nasopharyngeal colonization than did Patel age of participants. Given the deleterious effects of
et al. (approximately 15% of patients at the time acute exacerbations of COPD with respect to the
of enrollment), the most common pathogens were risk of death, quality of life, loss of lung function,
similar (with the exception that S. aureus was cul- and cost of care, adding azithromycin to the treat-
tured more frequently in our study, as would be ment regimen of patients who have had an acute
expected from nasopharyngeal sampling), and the exacerbation of COPD within the previous year or
effect of taking macrolides on colonization with who require supplemental oxygen is a valuable op-
macrolide-resistant respiratory pathogens was still tion; however, the patients should be screened for
clearly apparent. the presence of QTc prolongation and the risk of
We cannot comment on the safety profile of QTc prolongation and their hearing should be
azithromycin when it is taken for longer than monitored. In addition, it should be recognized
1 year, and we have no information pertaining to that the long-term effects of this treatment on
potential effects of long-term macrolide admin- microbial resistance in the community are not
istration on bacterial resistance patterns in the known.
community. Our results should be applied only to Supported by grants (U10 HL074407, U10 HL074408, U10
HL074409, U10 HL074416, U10 HL074418, U10 HL074422, U10
patients with COPD who either require supplemen- HL074424, U10 HL074428, U10 HL074431, U10 HL074439, and
tal oxygen or have had acute exacerbations and U10 HL074441) from the National Heart, Lung, and Blood Insti-
who do not have resting tachycardia or prolonga- tute (NHLBI) of the National Institutes of Health. The COPD
Clinical Research Network is supported by a Cooperative Agree-
tion of the QTc interval, are not taking medications ment from the Division of Lung Diseases of the NHLBI. At some
associated with QTc prolongation, and do not have sites, General Clinical Research Center facilities were used; their

696 n engl j med 365;8  nejm.org  august 25, 2011

The New England Journal of Medicine

Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 Azithromycin for Prevention of COPD Exacerbations

M01 grants from the National Center for Research Resources are boards of GlaxoSmithKline, MedImmune–Astra Zeneca, Merck,
listed in the Supplementary Appendix. Pearl Therapeutics, Novartis, UBC, Mpex Pharmaceuticals and
Dr. Albert reports receiving consulting fees from Gilead Sci- Ikaria, consulting fees from Forest–Almirall, Boehringer Ingel-
ences, fees for expert testimony from the Bruce Fagel Law Firm, heim, Nycomed–Forest, Roche, Bayer, Schering-Plough (Merck),
and royalties from Elsevier, and being named on a patent pending HLS, Talecris Biotherapeutics, Comgeniz, fb Communications,
for a device that provides continuous monitoring of the elevation BoomComm, Actelion Pharmaceuticals, Elan, Genzyme, Quark
of the head of the bed (Denver Health and the University of Colo- Pharmaceuticals, Merck, Pfizer, and Sanofi-Aventis, royalties
rado); Dr. Casaburi, receiving payment for service on the advisory from Associates in Medical Marketing and Castle Connolly, lec-
boards of Novartis Pharmaceuticals and Forest Pharmaceuticals, ture fees from GlaxoSmithKline, National Association for Con-
consulting fees from Theratechnologies, Breathe Technologies, tinuing Education, Med-Ed, Potomac Pharma, Pfizer, Boehringer
Medtronic Spinal and Biologics, Boehringer Ingelheim, Philips Ingelheim, Schering-Plough (Merck), Vox Medica, WebMD, Ep-
Respironics, Novartis, and Actelion Pharmaceuticals, lecture fees ocrates, AstraZeneca, and Altana–Nycomed, payment for develop-
from Boehringer Ingelheim, AstraZeneca, and Pfizer, holding ment of educational presentations for HIT Global and UpToDate,
stock in Inogen, and receiving grant support from Novartis, and grant support from Boehringer Ingelheim, Gilead, Johnson &
Roche, Boehringer Ingelheim, Osiris Therapeutics, Forest Phar- Johnson–Centocor Ortho Biotech, and Actelion Pharmaceuticals;
maceuticals, and GlaxoSmithKline, Breathe Technologies (pend- Dr. McEvoy, receiving grant support from Boston Scientific and
ing), and Theratechnologies (pending); Dr. Cooper, receiving fees GlaxoSmithKline and lecture fees from Boehringer Ingelheim
for expert testimony from Watkins, Louri, Roll and Change; Law- and GlaxoSmithKline; Dr. Niewoehner, receiving consulting fees
rence R. Dry & Associates; Starnes Davis Florie LLP; Walker, Tipps from Boehringer Ingelheim, AstraZeneca, GlaxoSmithKline, For-
& Malone PLC; Moore, Ingram Johnson & Steele; and Farris, Riley est Research, Merck, Sanofi-Aventis, Bayer Schering Pharma, Ny-
& Pitt, LLP; and grant support from GlaxoSmithKline and Novar- comed, ProtAffin, and Pfizer; Dr. Porszasz, receiving consulting
tis; Dr. Criner, receiving consulting fees from Phillips Pharmaceu- fees from Breathe Technologies, Boehringer Ingelheim, Novartis,
ticals, GlaxoSmithKline, Uptake Medical, and Dey, grant support and Forest Pharmaceuticals, and grant support from Boehringer
from Philips Respironics, GlaxoSmithKline, Boehringer Ingel- Ingelheim, Forest Pharmaceuticals, Breathe Technologies (pend-
heim, Novartis, and AstraZeneca, and royalties from Springer; Dr. ing), and Novartis; Dr. Price, receiving consulting fees from Astel-
Curtis, receiving grant support from Boehringer Ingelheim; Dr. las Pharma, Cubist Pharmaceuticals, and Merck, providing expert
Dransfield, receiving consulting fees from GlaxoSmithKline, testimony on transmission of hospital-associated infections,
Boehringer Ingelheim, and Forest Pharmaceuticals, grant support management of musculoskeletal infections, aminoglycoside tox-
from GlaxoSmithKline, Boehringer Ingelheim, and Boston Scien- icity, and diagnosis of infection, receiving grant support from
tific, and lecture fees from GlaxoSmithKline and Boehringer In- MicroPhage, Cubist Pharmaceuticals, Quintiles, Sanofi Pasteur,
gelheim; Dr. Han, receiving payment for service on the advisory BioCryst Pharmaceuticals, and Accelr8 Technology, and lecture
boards of CSL Behring, GlaxoSmithKline, and Boehringer Ingel- fees from Robert Michael (a CME vendor), Cubist Pharmaceuti-
heim, consulting fees from Genentech and Novartis, lecture fees cals, and Baxter Healthcare; Dr. Scanlon, receiving grant support
from GlaxoSmith­Kline, CSL Behring, Boehringer Ingelheim, and from Altana, Boehringer Ingelheim, Dey, Forest Pharmaceuticals,
Pfizer, royalties from UpToDate, and meeting expenses from As- GlaxoSmithKline, Novartis, and Pfizer and royalties from Lip-
traZeneca; Dr. Madinger, receiving honoraria in conjunction with pincott Williams & Wilkins and Wolters Kluwer; Dr. Sciurba, re-
the Merck Study Monitoring Antimicrobial Resistance Trends ceiving consulting fees from Boehringer Ingelheim, AstraZeneca,
(SMART), the JMI Laboratories Sentry study, and the Eurofins and GlaxoSmithKline and grant support from Pfizer and GlaxoS-
Medinet Trust study; Dr. Make, receiving payment for service on mithKline; Dr. Washko, receiving consulting fees from MedIm-
the advisory boards of Forest Pharmaceuticals, AstraZeneca, No- mune and Spiration and being married to an employee of Merck
vartis, Dey, Nycomed, Philips Respironics, Schering-Plough (now Research Laboratories Division of Clinical Pharmacology; and Dr.
Merck), SeQual, Embryon, Boehringer Ingelheim, Pfizer, and Woodruff, receiving consulting fees from MedImmune, grant
GlaxoSmithKline, consulting fees from Astellas Pharma, Talecris support from Genentech, and being a coinventor on a patent held
Biotherapeutics, and Chiesi Pharmaceuticals, lecture fees from jointly by his institution and Genentech. No other potential con-
GlaxoSmithKline, Boehringer Ingelheim, Pfizer, and Forest Phar- flict of interest relevant to this article was reported.
maceuticals, payment for manuscript preparation from AstraZen- Disclosure forms provided by the authors are available with the
eca, payment for video presentation preparation from Boehringer full text of this article at NEJM.org.
Ingelheim and Pfizer, payment for document reviews from Spira- We thank Dr. Peter Henson for first suggesting the hypothe-
tion, and grant support from AstraZeneca, GlaxoSmithKline, sis tested in this study; Angela Keniston, M.S.P.H., for assistance
Pfizer, Nabi Biopharmaceuticals, Boehringer Ingelheim, and Sun- with the statistical analysis; and Marianne Dieterich and Kris
ovion; Dr. Martinez, receiving payment for service on the advisory Richardson for assistance with susceptibility testing.

Appendix
The authors’ affiliations are as follows: the Medicine Service, Denver Health and Department of Medicine (R.K.A., C.S.P.), the University of
Colorado Denver Health Sciences Center (R.K.A., B.M., C.S.P.), the Division of Pulmonary Medicine, National Jewish Health and Department
of Medicine (B.M.), and the Division of Infectious Diseases, Department of Medicine, University of Colorado Denver (N.E.M., C.S.P.) — all
in Denver; the Division of Biostatistics, School of Public Health (J.C.), and the Department of Medicine (D.E.N., J.R.), University of Minne-
sota, and the Division of Pulmonary Medicine, Medicine Service, Minneapolis Veterans Affairs (VA) Medical Center (D.E.N., J.R.) — both in
Minneapolis; the Division of Pulmonary, Allergy, and Critical Care Medicine, Department of Medicine, University of Alabama at Birmingham
(W.C.B., J.A.D.C., M.T.D.), and the Pulmonary Section, Birmingham VA Medical Center (J.A.D.C., M.T.D.) — both in Birmingham; the
Division of Respiratory and Critical Care Physiology and Medicine, Department of Medicine, Harbor–UCLA Medical Center, Torrance, CA
(R.C., J.P.); the Division of Pulmonary and Critical Care Medicine, Department of Medicine, Temple University, Philadelphia (G.J.C., N.M.);
the Division of Pulmonary and Critical Care Medicine, Department of Medicine, University of Michigan (J.L.C., M.K.H., F.J.M.), and the
Division of Pulmonary and Critical Care Medicine, Ann Arbor VA Medical Center (J.L.C.) — both in Ann Arbor; the Division of Pulmonary
and Critical Care Medicine, Department of Medicine, and the Cardiovascular Research Institute, University of California, San Francisco, San
Francisco (S.C.L., P.G.W.); the Pulmonary, Critical Care, and Sleep Department, HealthPartners Research Foundation, St. Paul (C.M.), and
the Division of Pulmonary and Critical Care Medicine, Mayo Clinic, Rochester (P.D.S.) — both in Minnesota; the Division of Pulmonary and
Critical Care Medicine, Department of Medicine, University of Pittsburgh, Pittsburgh (F.C.S.); the Division of Pulmonary and Critical Care,
Department of Medicine, University of Maryland, Baltimore (S.M.S.); the Division of Pulmonary and Critical Care Medicine, Department of
Medicine, Brigham and Women’s Hospital, Boston (G.R.W.); and Respiratory Hospital, Winnipeg, MB, Canada (N.R.A.).

n engl j med 365;8  nejm.org  august 25, 2011 697

The New England Journal of Medicine
Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.
 Azithromycin for Prevention of COPD Exacerbations

References
1. Strassels SA, Smith DH, Sullivan SD, 14. Calverley PMA, Anderson JA, Celli B, Erythromycin and common cold in COPD.
Mahajan PS. The costs of treating COPD in et al. Salmeterol and fluticasone propio- Chest 2001;120:730-3.
the United States. Chest 2001;119:344-52. nate and survival in chronic obstructive 28. Seemungal TAR, Wilkinson TMA,
2. Andersson F, Borg S, Jansson SA, et al. pulmonary disease. N Engl J Med 2007; Hurst JR, Perera WR, Sapsford RJ, Wedzi-
The costs of exacerbations in chronic ob- 356:775-89. cha JA. Long-term erythromycin therapy
structive pulmonary disease (COPD). Respir 15. Mahler DA, Donohue JF, Barbee RA, is associated with decreased chronic ob-
Med 2002;96:700-8. et al. Efficacy of salmeterol xinafoate in structive pulmonary disease exacerbations.
3. Druss BG, Marcus SC, Olfson M, Pin- the treatment of COPD. Chest 1999;115: Am J Respir Crit Care Med 2008;178:1139-
cus HA. The most expensive medical con- 957-65. 47.
ditions in America. Health Aff (Millwood) 16. Vincken W, van Noord JA, Greefhorst 29. Yamaya M, Azuma A, Tanaka H, et al.
2002;21:105-11. APM, et al. Improved health outcomes in Inhibitory effects of macrolide antibiotics
4. Miller JD, Foster T, Boulanger L, et al. patients with COPD during 1 yr’s treat- on exacerbations and hospitalization in
Direct costs of COPD in the U.S.: an anal- ment with tiotropium. Eur Respir J 2002; chronic obstructive pulmonary disease in
ysis of Medical Expenditure Panel Survey 19:209-16. Japan: a retrospective multicenter analy-
(MEPS) data. COPD 2005;2:311-8. 17. Casaburi R, Mahler DA, Jones PW, et sis. J Am Geriatr Soc 2008;56:1358-60.
5. Lindenauer PK, Pekow P, Gao S, Craw- al. A long-term evaluation of once-daily 30. He ZY, Ou LM, Zhang JQ, et al. Effect
ford AS, Gutierrez B, Benjamin EM. Qual- inhaled tiotropium in chronic obstructive of 6 months of erythromycin treatment
ity of care for patients hospitalized for pulmonary disease. Eur Respir J 2002; on inflammatory cells in induced sputum
acute exacerbations of chronic obstructive; 19:217-24. and exacerbations in chronic obstructive
144:894-903. 18. Brusasco V, Hodder R, Miravitlles M, pulmonary disease. Respiration 2010 Sep-
6. Seemungal TA, Donaldson GC, Paul Korducki L, Towse L, Kesten S. Health tember 28 (Epub ahead of print).
EA, Bestall JC, Jeffries DJ, Wedzicha JA. outcomes following treatment for six 31. Blasi F, Bonardi D, Aliberti S, et al.
Effect of exacerbation on quality of life in months with once daily tiotropium com- Long-term azithromycin use in patients
patients with chronic obstructive pulmo- pared with twice daily salmeterol in pa- with chronic obstructive pulmonary dis-
nary disease. Am J Respir Crit Care Med tients with COPD. Thorax 2003;58:399- ease and tracheostomy. Pulm Pharmacol
1998;157:1418-22. 404. [Erratum, Thorax 2005;60:105.] Ther 2010;3:200-7.
7. Kanner RE, Anthonisen NR, Connet 19. Niewoehner DE, Rice K, Cote C, et al. 32. Gómez J, Baños V, Simarro E, et al.
JE. Lower respiratory illnesses promote Prevention of exacerbations of chronic Prospective, comparative study (1994-1998)
FEV1 decline in current smokers but not obstructive pulmonary disease with tiotro- of the influence of short-term prophylac-
ex-smokers with mild chronic obstructive pium, a once-daily inhaled anticholiner- tic treatment with azithromycin on pa-
pulmonary disease: results from the lung gic bronchodilator: a randomized trial. Ann tients with advanced COPD. Rev Esp Qui-
health study. Am J Respir Crit Care Med Intern Med 2005;143:317-26. mioter 2000;13:379-83. (In Spanish.)
2001;164:358-64. 20. Tashkin DP, Celli B, Senn S, et al. 33. Arizona Center for Education and Re-
8. Mannino DM, Homa DM, Akinbami A 4-year trial of tiotropium in chronic ob- search on Therapeutics. QT drug lists
LJ, Ford ES, Redd SC. Chronic obstructive structive pulmonary disease. N Engl J Med by risk groups. (http://www.azcert.org/
pulmonary disease surveillance — United 2008;359:1543-54. medical-pros/drug-lists/drug-lists.cfm.)
States, 1971–2000. MMWR Surveill Summ 21. Szafranski W, Cukier A, Ramirez A, et 34. Jones PW, Quirk FH, Baveystock CM.
2002;51:1-16. al. Efficacy and safety of budesonide/ The St. George’s Respiratory Question-
9. Tsai CL, Sobrino JA, Carmago CA Jr. formoterol in the management of chronic naire. Respir Med 1991;85:Suppl B:25-31.
National study of emergency department obstructive pulmonary disease. Eur Respir 35. Efron B, Tibshirani RJ. An introduc-
visits for acute exacerbation of chronic J 2003;21:74-81. [Erratum, Eur Respir J tion to the bootstrap. New York: Chapman
obstructive pulmonary disease, 1993-2005. 2004;21:912.] & Hall, 1993.
Acad Emerg Med 2008;15:1275-83. 22. Welte T, Miravitlles M, Hernandez P, 36. Kim K, DeMets DL. Design and analy-
10. Donaldson GC, Seemungal TA, et al. Efficacy and tolerability of budeso­ sis of group sequential tests based on the
Bhowmik A, Wedzicha JA. Relationship nide/formoterol added to tiotropium in type I error spending rate function.
between exacerbation frequency and lung patients with chronic obstructive pulmo- Biometrika 1987;74:149-54.
function decline in chronic obstructive nary disease. Am J Respir Crit Care Med 37. Phaff SJ, Tiddens HA, Verbrugh HA,
pulmonary disease. Thorax 2002;57:847- 2009;180:741-50. Ott A. Macrolide resistance of Staphylo-
52. [Erratum, Thorax 2008;63:753.] 23. Aaron SD, Vandemheen KL, Fergus- coccus aureus and Haemophilus species as-
11. Soler-Cataluña JJ, Martínez-García son D, et al. Tiotropium in combination sociated with long-term azithromycin use
MA, Román Sánchez P, Salcedo E, Navarro with placebo, salmeterol, or fluticasone- in cystic fibrosis. J Antimicrob Chemother
M, Ochando R. Severe acute exacerbations salmeterol for treatment of chronic ob- 2006;57:741-6.
and mortality in patients with chronic ob- structive pulmonary disease: a randomized 38. Saiman L, Anstead M, Mayer-Hamblett
structive pulmonary disease. Thorax 2005; trial. Ann Intern Med 2007;146:545-55. N, et al. Effect of azithromycin on pulmo-
60:925-31. 24. Sin DD, McAlister FA, Man SFP, An- nary function in patients with cystic fibro-
12. Burge PS, Calverley PMA, Jones PW, thonisen NR. Contemporary manage- sis uninfected with Pseudomonas aerugi-
Spencer S, Anderson JA, Maslen TK. Ran- ment of chronic obstructive pulmonary nosa: a randomized controlled trial. JAMA
domised, double blind, placebo controlled disease: scientific review. JAMA 2003;290: 2010;303:1707-15.
study of fluticasone propionate in patients 2301-12. 39. Patel IS, Seemungal TAR, Wilks M,
with moderate to severe chronic obstruc- 25. Martinez FJ, Curtis JL, Albert R. Role Lloyd-Owen SJ, Donaldson GC, Wedzicha
tive pulmonary disease: the ISOLDE trial. of macrolide therapy in chronic obstruc- JA. Relationship between bacterial coloni-
BMJ 2000;320:1297-303. tive pulmonary disease. Int J Chron Ob- sation and the frequency, character, and
13. Calverley P, Pauwels R, Vestbo J, et al. struct Pulmon Dis 2008;3:331-50. severity of COPD exacerbations. Thorax
Combined salmeterol and fluticasone in 26. Banerjee D, Khair OA, Honeybourne 2002;57:759-64.
the treatment of chronic obstructive pul- D. The effect of oral clarithromycin on Copyright © 2011 Massachusetts Medical Society.
monary disease: a randomised controlled health status and sputum bacteriology in
trial. Lancet 2003;361:449-56. [Erratum, stable COPD. Respir Med 2005;99:208-15.
Lancet 2003;361:1660.] 27. Suzuki T, Yanai M, Yamaya M, et al.

698 n engl j med 365;8  nejm.org  august 25, 2011

The New England Journal of Medicine

Downloaded from nejm.org at UFC on June 14, 2016. For personal use only. No other uses without permission.
Copyright © 2011 Massachusetts Medical Society. All rights reserved.