Azithromycin for acute bronchitis: a randomised, double-blind, controlled
trial
Lancet 2002; 359: 1648-54
Arthur T Evans, Shahid Husain, Lakshmi Durairaj, Laura S Sadowski, Marjorie
Charles-Damte, Yue Wang
  _____

Collaborative Research Unit, Department of Medicine, Cook County Hospital
and Rush Medical College, Chicago, IL (A T Evans MD, L S Sadowski MD, M
Charles-Damte RN, Y Wang PhD); Division of Infectious Diseases, University
of Pittsburgh Medical Center, Pittsburgh, PA (S Husain MD); and Division of
Pulmonary, Critical Care and Occupational Medicine, University of Iowa
College of Medicine, Iowa City, IA, USA (L Durairaj MD)
  _____

Correspondence to: Dr Arthur T Evans, Collaborative Research Unit, Cook
County Hospital, Room 1600, Administration Building, 1900 W Polk St,
Chicago, IL 60612, USA (e-mail: [log in to unmask] <mailto:[log in to unmask]> )
Summary <http://www.thelancet.com/search/#summary>
Introduction <http://www.thelancet.com/search/#introduction>
Methods <http://www.thelancet.com/search/#methods>
Results <http://www.thelancet.com/search/#results>
Discussion <http://www.thelancet.com/search/#discussion>
References <http://www.thelancet.com/search/#references>
Summary
Background The value of azithromycin for treatment of acute bronchitis is
unknown, even though this drug is commonly prescribed. We have investigated
this question in a randomised, double-blind, controlled trial.
Methods Adults diagnosed with acute bronchitis, without evidence of
underlying lung disease, were randomly assigned azithromycin (n=112) or
vitamin C (n=108) for 5 days (total dose for each 1·5 g). All individuals
were also given liquid dextromethorphan and albuterol inhaler with a spacer.
The primary outcome was improvement in health-related quality of life at 7
days; an important difference was defined as 0·5 or greater. Analysis was by
intention to treat.
Findings The study was stopped by the data-monitoring and safety committee
when 220 patients had been recruited. On day 7, the adjusted difference in
health-related quality of life was small and not significant (difference
0·03 [95% CI -0·20 to 0·26], p=0·8). 86 (89%) of 97 patients in the
azithromycin group and 82 (89%) of 92 in the vitamin C group had returned to
their usual activities by day 7 (difference 0·5% [-10% to 9%], p>0·9). There
were no differences in the frequency of adverse effects; three patients in
the vitamin C group discontinued the study medicine because of perceived
adverse effects, compared with none in the azithromycin group. Most patients
(81%) reported benefit from the albuterol inhaler.
Interpretation Azithromycin is no better than low-dose vitamin C for acute
bronchitis. Further studies are needed to identify the best treatment for
this disorder.
Lancet 2002; 359: 1648-54
Top <http://www.thelancet.com/search/#top>

Introduction
Every year, more than ten million US adults visit physicians for acute
bronchitis, and most of them receive antibiotics.1,2 Many experts condemn
such treatment, citing three main reasons: weak or conflicting experimental
evidence of clinical benefit, lack of a strong biological rationale (the
causative pathogens are viruses in most cases), and increasing societal
concern about widespread antibiotic resistance.2-5
Nine randomised controlled trials of three different antibiotics for acute
bronchitis have been published;6-14 four demonstrated clinical benefit.9-12
Meta-analyses of these studies15-18 have been beset by fundamental
problems:2,15-18 different studies measured different outcomes; the
reliability and validity of outcome measurement was uncertain; and some
outcomes had questionable clinical significance (for example, duration of
purulent sputum). Equally important, no published study has measured the
effect of antibiotic therapy on quality of life, although four studies8-11
reported on limitations in work or other activities.18 Finally, no study has
assessed any of the newer antibiotics widely promoted and prescribed today,
including quinolones and the newer macrolides.
Azithromycin is a macrolide antibiotic commonly prescribed for acute
bronchitis. It has a broad spectrum of activity and infrequent adverse
effects, and it is easy to take (once daily for 5 days). However, it is also
expensive, and the only published evidence of its efficacy in acute
bronchitis comes from several small equivalence studies19-22 and one large,
open, uncontrolled, drug-company-sponsored case series.23,24
In a randomised, double-blind, controlled clinical trial, we tested whether
patients prescribed azithromycin for acute bronchitis returned to work (or
other usual activities) sooner and whether they experienced greater
improvements in health-related quality of life than patients prescribed
low-dose vitamin C.
Top <http://www.thelancet.com/search/#top>

Methods

Participants
After approval of the study protocol by the institutional review board, we
recruited adult patients without chronic lung disease who presented with
cough of 2-14 days duration (with or without sputum production) and were
diagnosed with acute bronchitis by attending physicians in the adult
ambulatory screening clinic of Cook County Hospital, Chicago, Illinois, USA,
during weekdays from December, 1999, until March, 2000. Patients were
enrolled after giving written informed consent. Reasons for exclusion were:
pregnancy; other infectious diseases necessitating antimicrobial therapy;
chronic lung disease (including asthma) or current treatment with
bronchodilators or glucocorticoids; angiotensin-converting-enzyme inhibitor
started within the previous 4 weeks; antibiotic treatment within the
previous 2 weeks; need for hospital admission; allergies to macrolides,
bronchodilators, or vitamin C; duration of cough less than 2 days or longer
than 2 weeks; or any clinical characteristic suggesting pneumonia, including
oral temperature above 38·9°C (102°F), respiratory rate greater than 25 per
min, or infiltrates or other abnormalities on chest radiograph.
Immediately before the study, physicians who would be seeing patients with
cough in the ambulatory screening clinic during the study period took a 2-h
training course that covered the assessment of patients presenting with
acute cough, the diagnosis of acute bronchitis, the evidence supporting
treatment options for cough25 and acute bronchitis, and the rationale for
the planned study. The physicians were reminded that the diagnosis of acute
bronchitis is largely one of exclusion, and therefore other diagnoses should
be carefully considered--such as pneumonia, viral upper-respiratory-tract
illness, influenza, gastro-oesophageal reflux, sinusitis, postnasal drip,
heart failure, and asthma--before a final diagnosis of acute bronchitis is
made. All laboratory and radiographic studies were left to the discretion of
the treating physicians, but if a chest radiograph was done before
enrolment, any pulmonary abnormality was grounds for exclusion.
Design and procedures
Patients were randomly assigned one of two study drugs, azithromycin (total
dose 1·5 g) or vitamin C (total dose 1·5 g). All study drugs were prepared
at one site by an independent pharmaceutical consultant (MediDerm, Chicago,
IL, USA) in opaque sealed capsules that were identical in appearance, taste,
and smell. Six capsules were packaged in bottles that were identical except
for labels that differed only by a unique identification number. Each
azithromycin capsule contained the equivalent of a 250 mg caplet. Each
vitamin C capsule contained 250 mg vitamin C and enough dextrose to fill the
capsule to a volume equivalent to the azithromycin capsule.
A research associate, not a member of the research team, prepared the
randomisation code using a computer-generated (Arcus QuickStat Biomedical,
Cambridge, UK) random allocation schedule in three unequal blocks. The
external pharmaceutical company then sealed the study drugs in sequentially
numbered identical containers according to the allocation sequence. All
members of the research team--investigators, project coordinator,
physicians, patient educators, research assistants, data collectors,
data-entry staff, and the research pharmacist who dispensed all study drugs
in sequence--were unaware of the allocation schedule. Masking was maintained
until data analysis except on two occasions, when physicians specifically
requested knowledge of the study drug because of patients' lack of
improvement.
Before leaving the clinic, each patient was directly observed taking two
study drug capsules. Participants were then instructed to take one study
drug capsule in the morning on an empty stomach 1 h before eating breakfast,
for the next 4 days (ie, a total of six capsules).
We chose vitamin C instead of a traditional placebo compound, such as
dextrose, because the results of several focus groups we undertook a year
earlier with members of the target population showed that they would
probably refuse to take part in any randomised trial of acute bronchitis if
the placebo group received a "sugar pill". However, they suggested that a
randomised trial of antibiotic and a multivitamin or some other vitamin pill
would be acceptable. We followed this recommendation and used vitamin C as
the comparison drug, since it can be administered in a capsule without any
identifying taste or smell and because there is no evidence of efficacy in
acute bronchitis or any other respiratory illness at the doses given.26
Patients in both treatment groups all received aggressive standard
symptomatic therapy of proven benefit. All participants were given 240 mL (8
ounces) of dextromethorphan syrup and were instructed to take 10 mL by mouth
every 6 h as necessary for cough during the day and 15 mL at bedtime.25 All
participants also received one albuterol inhaler with a spacer and were
instructed to inhale two puffs with the spacer every 6 h as necessary for
cough.27,28 A research nurse or trained research assistant gave instructions
to all participants on taking all medications and provided supervised
practice in use of the inhaler and spacer. Patients were instructed not to
take any other medications for their illness during the study period.
The primary endpoint was health-related quality of life on day 7 of
follow-up. Secondary endpoints were return to usual daily activities at
follow-up, domain scores that comprise health-related quality of life, and
adverse effects.
On enrolment, trained research assistants interviewed study participants to
establish baseline measurements of health-related quality of life and to
describe the activity limitations attributed to the symptoms of acute
bronchitis. In addition, the examining physicians completed a standard
data-collection instrument for recording clinical data from their history,
physical examination, clinical assessment, and any tests done.
On day 3 (about 48 h after enrolment), a research assistant telephoned all
participants for a follow-up interview to inquire whether the patient had
returned to his or her usual activities at work, home, or school and to
reassess the patient's health-related quality of life. In addition, patients
were interviewed about use of all medications and any possible adverse
effects. On day 7, the interview was repeated, with additional questions
included at the end to assess the adequacy of masking and patients'
subjective impressions about the effectiveness of all medications. If a
patient was not available for interview on day 7 at the scheduled time, the
interviewers attempted to contact him or her up to six times daily for 3
consecutive days before declaring the patient lost to follow-up.
The acute bronchitis health-related quality-of-life interview was adapted
from similar instruments developed at McMaster University to measure changes
in health-related quality of life for patients treated for asthma,29
rhinitis,30 chronic lung disease,31 or congestive heart failure.32 A score
is obtained by taking the mean of 22 equally weighted items representing
four domains: effects on daily activities (three activities specified by the
individual and three general activities--sleep, recreational activities, and
regular activities at home and at work); effects of coughing, including
chest pain and dyspnoea (eight items); general symptoms (four items); and
emotional functioning (four items; panel). For each item, patients are asked
to indicate on a 7-point scale how troubled they have been during the
previous few days as a result of their bronchitis symptoms, from not
troubled at all (0) to extremely troubled (6).
We assessed the adequacy of masking by asking all participants at the day 7
interview to guess whether they received azithromycin or vitamin C.
Statistical analysis
We used multivariate ANCOVA with repeated measures to assess the effect of
treatment group on health-related quality-of-life scores during the
follow-up period, while controlling for any differences at baseline. In
addition, we used the simpler ANCOVA to assess differences between groups on
day 3 and day 7, separately.
A difference between groups of 0·5 points on the quality-of-life scale was
selected as the smallest important difference based on published research of
similar health-related quality-of-life scales.33-36 However, the previous
research that identified 0·5 as a reasonable "smallest clinically important
difference" involved chronic illnesses--allergic rhinitis, asthma,
congestive heart failure, chronic obstructive lung disease--in which a small
difference over a long period might be judged important. For self-limited
illnesses of shorter duration, such as acute bronchitis, the smallest
important difference might therefore be greater than 0·5. A difference of
0·5 points represents a change, for example, from 2 (somewhat troubled) to 1
(hardly troubled at all) for half of the symptoms or activities addressed in
the questionnaire (on the assumption the remaining items are unchanged).
In analysing secondary endpoints, we compared the proportion of participants
who had returned to their usual daily activities on day 3 and day 7 in the
azithromycin and vitamin C groups. We used *2 or Fisher's exact tests and
constructed near-exact 95% CI for the difference in proportions by the
method of Miettinen.37
We had originally planned for a sample size of 400 patients so that we would
have sufficient power to test treatment effects within several clinical
subgroups that previous research had suggested as potentially important
(based on age, duration of cough, magnitude of fever, and severity of
symptoms). However, an interim analysis was done at the end of the fourth
month to help decide whether to continue recruiting. The data-monitoring and
safety committee, independent of the investigators, was presented with
interim analyses addressing primary and secondary endpoints, as well as
adverse effects, after the first 220 patients had been enrolled. Unaware of
treatment-group identity, the committee concluded that the study should stop
enrolment because the outcomes were equivalent and there was sufficient
precision to be confident that the likelihood of detecting a clinically
meaningful difference with a larger sample was so small that continued
enrolment of patients would be inappropriate.
The sample size of 220 provided power of 95-99% to detect a difference
between the groups of 0·5 points in health-related quality of life, on the
assumption of a population SD of 0·8-1·0 points and alpha=0·05. This sample
size also provided 80% power to detect a difference of 10% in the proportion
who had returned to their usual activities by day 7.
Statistical analyses were done with SPSS (version 10), Stata (version 6),
and Arcus Quickstat Biomedical (version 1). We updated a previously
published meta-analysis18 using the DerSimonian-Laird random-effects model
for risk differences (Arcus Quickstat Biomedical). This meta-analysis model
was used instead of other methods that pool odds ratios18 because the
results have a more transparent clinical interpretation and because the
study by Williamson and colleagues11 can be included in the analysis,
despite an odds ratio that is undefined.
Top <http://www.thelancet.com/search/#top>

Results
Of the 340 patients who met the inclusion criteria, 91 were ineligible
because of one or more exclusion criteria and 29 declined to participate; we
enrolled the remaining 220 patients (figure 1). Table 1 shows the baseline
demographic and clinical characteristics of study participants. 93 (96%)
patients in the azithromycin group and 87 (95%) in the vitamin C group
reported taking at least five of the six study drug capsules (p=0·7).
<javascript:%7B%7D>  01art/7489(1) <javascript:%7B%7D>   <javascript:%7B%7D>
Figure 1: Trial profile
  _____

Characteristic
Azithromycin
Vitamin C

(n=112)
(n=108)
Demography


Mean (range) age, years
47 (20-88)
44 (18-80)
Men
55 (49%)
45 (42%)
Women
57 (51%)
63 (58%)
Race
African American/Black
71 (63%)
76 (70%)
Hispanic/Latino
16 (14%)
14 (13%)
Asian
14 (13%)
9 (8%)
Other
11 (10%)
9 (8%)
Clinical
Median (10th, 90th percentiles)
5 (3, 10)
4 (3, 10)
duration of cough, days
Median (10th, 90th percentiles)
37·0 (36·3, 37·7)
37·0 (36·3, 37·8)
temperature, ºC
Wheezing
15 (13%)
11 (10%)
Fine crackles
4 (4%)
1 (1%)
Coarse crackles
1 (1%)
1 (1%)
Purulent nasal discharge
8 (7%)
8 (7%)
Exudates of tonsils or pharynx
0
0
Smoking status


Ever
66 (59%)
59 (55%)
Current
43 (38%)
37 (34%)
History


Cocaine or heroin use
18 (16%)
13 (12%)
Childhood asthma
3 (3%)
5 (5%)
ACE inhibitor* use for ge30 days
5 (5%)
3 (3%)
Diagnoses at study entry†


Upper respiratory tract infection
20 (18%)
12 (11%)
Rhinitis
2 (2%)
10 (9%)
Diabetes mellitus
5 (5%)
3 (3%)
Influenza
1 (1%)
3 (3%)
Congestive heart failure
1 (1%)
1 (1%)
Gastro-oesophageal reflux
1 (1%)
0
Sarcoidosis‡
1 (1%)
0
Data are number of participants unless otherwise stated. *Use of an
angiotensin-converting-enzyme inhibitor for at least 30 days. Patients
recently started on these drugs (within 30 days) were excluded before
randomisation. †All enrolled patients were diagnosed by an attending
physician as having acute bronchitis, although some had other concurrent
diagnoses. ‡Sarcoidosis, along with other chronic lung diseases, was a
reason for exclusion, but one patient with sarcoidosis was nevertheless
enrolled and randomised and therefore included in data analysis.
Table 1: Demographic and clinical characteristics of the 220 participants at
baseline
  _____

On day 7, the proportion of participants who guessed they were taking
azithromycin was similar in the azithromycin and vitamin C groups (34 of 97
[35%] vs 36 of 92 [39%], p=0·6). We unmasked treatment assignment before the
end of follow-up for two patients (both in the vitamin C group) at the
request of their treating physicians. In one of these patients, pneumonia
was subsequently diagnosed because infiltrates were seen on chest radiograph
and successfully treated on an outpatient basis with a quinolone antibiotic.
The other patient improved without any additional treatment. The
azithromycin and vitamin C groups did not differ significantly in their
crude or adjusted mean health-related quality-of-life scores on day 3 or day
7 (p>0·2; multivariate ANCOVA with two repeated measures; table 2 and figure
2). The rate of improvement was the same for both groups (p=0·3). There was
no interaction between baseline health-related quality-of-life score and
treatment effect (p>0·3).
  _____


Azithromycin (n=97)
Vitamin C (n=92)
Difference vitamin C-azithromycin (95% CI)*
p†
Day 1 baseline QOL (summary score)‡
2·9 (1·2)
2·6 (1·2)
-0·3 (-0·6 to 0·03)
0·07
Day 3 return to usual activities
64/96 (67%)
58/92 (63%)
4% (-10% to 17%)
0·6
Day 3 QOL (summary score)‡
1·4 (1·1)
1·7 (1·1)
0·3 (-0·03 to 0·55)§
0·08
Day 3 QOL domain
Activity limitations
1·7 (1·5)
2·1 (1·5)
0·4 (0 to 0·8)
0·05
Cough-related problems
1·4 (1·1)
1·6 (1·1)
0·2 (-0·1 to 0·5)
0·16
General symptoms
1·4 (1·3)
1·5 (1·4)
0·1 (-0·2 to 0·5)
0·4
Emotional functioning
1·1 (1·2)
1·3 (1·3)
0·2 (-0·1 to 0·6)
0·15
Day 7 return to usual activities†
86/97 (89%)
82/92 (89% )
0·5% (-9% to 10%)
>0·9
Day 7 QOL summary score‡
0·9 (0·76)
0·9 (0·8)
0·03 (-0·20 to 0·26)
0·8
Day 7 QOL domain
Activity limitations
0·9 (0·9)
1·1 (1·1)
0·2 (-0·1 to 0·5)
0·2
Cough-related problems
0·9 (0·8)
0·9 (0·9)
0·01 (-0·2 to 0·3)
0·9
General symptoms
0·8 (0·9)
0·8 (0·9)
-0·06 (-0·3 to 0·2)
0·6
Emotional functioning
0·7 (1·0)
0·8 (1·1)
0·06 (-0·2 to 0·3)
0·7
*Differences in health-related quality of life on day 3 and day 7 are
adjusted for baseline scores on day 1 (ANCOVA). Differences for specific
domains of quality of life are adjusted for the domain-specific scores on
day 1. A positive value for the difference means that there was a greater
improvement in the azithromycin group. A difference of 0·5 points is taken
to be the smallest important difference. †Fisher's exact test was used to
compare the proportions in each group who had returned to usual activities
on days 3 and 7. All other p values are based on ANOVA (day 1) or ANCOVA
(days 3 and 7). ‡Mean (SD). §A difference of 0·3 points on day 3 represents
a change from 2=somewhat troubled to 1=hardly troubled at all for one in
three of the symptoms or activities addressed in the questionnaire (if the
scores for the other items remain unchanged).
Table 2: Baseline and follow-up results for the primary and secondary
outcomes: mean health-related quality-of-life (QOL) scores and the
proportions who had returned to usual activities on day 3 and day 7
  _____

<javascript:%7B%7D>  01art/7489(2?) <javascript:%7B%7D>
<javascript:%7B%7D>
Figure 2: Boxplots of the baseline and follow-up results for the primary
outcome: health-related quality-of-life scores
Boxes extend from the 25th to the 75th percentile, the black horizontal
line=median score. Whiskers extend to the most extreme value or to 1·5 times
IQR, whichever is closer. Small open circles=outliers.
One patient in the vitamin C group was diagnosed with pneumonia during
follow-up, compared with none in the azithromycin group. No patient required
hospital admission during the study period.
The measurement of health-related quality of life was internally consistent,
since Cronbach's alphawas greater than 0·92 for each of the 3 days on which
it was measured (baseline, day 3, and day 7). The measure was also
responsive to change; the average change in score over 7 days was roughly 2
points, equivalent to more than 1·5 SDs of the baseline score.
The effects of treatment within specific domains of health-related quality
of life approached clinical and statistical significance on day 3 but not on
day 7. On day 3, the azithromycin group was less troubled doing daily
activities and less troubled with cough and cough-related problems than the
vitamin C group (table 2), but the between-group differences were less than
0·5, the smallest clinically important difference. By day 7, however,
differences between groups for all domains of health-related quality of life
were smaller and all CI excluded any clinically important difference. For
both the activity limitation domain and the cough domain, there was a
greater treatment effect among the most troubled patients, although the
statistical test for interaction was not significant (p>0·2).
The azithromycin and vitamin C groups did not differ significantly in the
proportion who had returned to their usual activities at work, home, or
school by day 3 or by day 7 (table 2, figure 3). By day 7, five patients in
the azithromycin group and four in the vitamin C group had made a second
visit to a physician because they were not feeling better.
<javascript:%7B%7D>  01art/7489(2) <javascript:%7B%7D>   <javascript:%7B%7D>
Figure 3: Cumulative proportion of patients who had returned to their usual
daily activities
These results are a summary of four sources of data: day 1 face-to-face
interview (baseline), day 3 and day 7 telephone interviews ("Have you
returned to your usual activities at work, home, or school?" If yes: "When
was that?"), and a patient's symptom diary (returned by mail at the end of
the follow-up period).
Among the participants contacted on day 3, 18 of 96 (19%) in the
azithromycin group reported adverse effects from study medications, compared
with 19 of 92 (21%) in the vitamin C group. The most common complaints were
diarrhoea (nine vs seven) and nausea (seven vs three). Four participants had
stopped taking the albuterol inhaler by the time of the interview on day 3
because of perceived adverse effects, and an additional 14 had stopped
because they felt sufficiently better.
On day 7, perceived adverse effects were reported by 24 of 97 (25%) in the
azithromycin group and 19 of 92 (21%) in the vitamin C group (p>0·2). The
most commonly reported adverse effects on day 7 were diarrhoea (11 vs six)
and nausea (six vs four). Only three participants reported that they stopped
the study drug because of perceived adverse effects; all three were in the
vitamin C group.
The albuterol inhaler was considered effective by 81%, ineffective by 10%,
and of uncertain value by 9%. There were no differences between study
groups. By day 7, 84 (44%) of the 189 patients reported still using the
inhaler; 11 (6%) had stopped using it because of adverse effects, three (2%)
because of perceived ineffectiveness, 64 (34%) because they were
sufficiently better, and 27 (14%) for other reasons.
Although we did an intention-to-treat analysis of all available
participants, 15 patients in the azithromycin group and 16 in the vitamin C
group were not available at follow-up. We therefore undertook further
analyses to assess the sensitivity of our results to various assumptions
about the missing data. The 31 patients lost to follow-up were more likely
than those interviewed at day 7 to have a baseline respiratory rate of 20
per min or greater (odds ratio 2·9), a baseline self-report of cocaine or
heroin use (2·7), and a lower (less troubled) baseline quality-of-life score
(0·7; each p<0·05 in a multivariate model). There were no differences by
study group. We repeated the analysis of our primary endpoint with all 220
study participants with three different methods to impute missing values:
the last value carried forward method; a best-case scenario for
azithromycin; and a worst-case scenario for azithromycin. The point
estimates for the main outcome were 0·1, 0·4, and -0·2 for the three
methods. Since the three point estimates were within 0·5 points, the
smallest clinically important difference, the study's conclusions were not
considered sensitive to the method for dealing with loss to follow-up.
Top <http://www.thelancet.com/search/#top>

Discussion
These results show that azithromycin is no more effective than low-dose
vitamin C for treatment of acute bronchitis. Given the lack of evidence that
low-dose vitamin C is beneficial, we conclude that azithromycin is
ineffective and should not be prescribed for patients with acute bronchitis.
Our study improves on the previous evidence in several ways. It measured the
clinical outcomes most relevant to patients: health-related quality of life
and return to usual daily activities. Our measure of disease-specific
health-related quality of life was adapted from validated instruments used
for other respiratory illnesses29-31,34 and was assessed reliably with
a-priori definitions of clinical significance. Our eligibility criteria kept
to a minimum diagnosis misclassification bias;5 only one of 108 patients in
the control group required antibiotics for pneumonia (apparently
misdiagnosed as acute bronchitis on study entry). Patients in both groups
received intensive symptomatic treatment, including inhaled
bronchodilators.5,28 Finally, we studied a widely prescribed, newer agent,
the infrequent side-effects of which mean not only that it allows a "best
case" for antibiotic therapy to be investigated but also that it protects
concealment of study drug allocation.5 Indeed, our exit interviews confirmed
that patients did not know whether or not they had received the antibiotic.
These features of our study strengthen its internal and external validity.
The rate of improvement was more rapid among our patients than in other
studies,38-40 perhaps because the range of illness was milder (shorter mean
duration of cough at presentation and fewer patients with abnormal chest
signs) or because all patients received aggressive symptomatic therapy with
inhaled bronchodilators.5,28,
The most comprehensive meta-analysis of antibiotics for acute bronchitis18
concluded that antibiotics decrease cough but at the expense of drug
side-effects. Although no previous trial has tested azithromycin or measured
health-related quality of life, four other studies measured the related
endpoint of return to usual activities.18 Our results for this outcome can
be evaluated in the context of the other published evidence by updating the
meta-analysis (figure 4). The pooled difference of 2% favours antibiotics,
but the aggregate data are compatible with a range of values that include
both benefit (up to 6%) and harm (up to 3%) from antibiotics.
<javascript:%7B%7D>  01art/7489(3) <javascript:%7B%7D>   <javascript:%7B%7D>
Figure 4: Meta-analysis of randomised trials that assess the difference in
the proportion of patients who have returned to their usual daily activities
in antibiotic and control groups
For each study, the difference in the proportion of patients who have
returned to their usual activities at follow-up (control group minus the
antibiotic group) is represented by a thin vertical line, which transects a
horizontal open rectangle depicting the 95% CI. Results less than zero
favour the antibiotic group. The solid diamonds are proportional to the
studies' sample sizes. The pooled results and 95% CI are described by the
final (open) diamond. The DerSimonian-Laird (DL) random-effects model of
risk differences was used to calculate the summary point estimate for the
difference in proportions between treatment.
Our findings do not prove that azithromycin provides absolutely no benefit
to patients with acute bronchitis. The 95% CI for differences between
azithromycin and vitamin C on study day 3 included 0·5 for the summary
quality-of-life score and for each of its component domains. Although none
of these effects persisted on day 7, our results do not exclude the
possibility that azithromycin may provide transient benefit of little
clinical significance to patients with acute bronchitis. In addition, our
sample size did not permit precise subgroup comparisons to analyse whether
azithromycin is effective for selected subgroups (for example, elderly
patients with severe symptoms). We discontinued enrolment when our
independent data-monitoring and safety committee, unaware of study group
identity, concluded that the likelihood of showing a clinically meaningful
treatment effect with a larger sample was too small to warrant continuation
of the study. Notwithstanding these reservations, our findings shift the
burden of proof to proponents of antibiotic therapy for patients with acute
bronchitis.
Our study is limited in two important ways. First, although our follow-up
rate was similar in both study groups, 17% (31 of 220) of enrolled patients
did not complete the full evaluation of all study outcomes. None of these
patients was admitted to our hospital, none was seen in the emergency
department or in one of the clinics, and none died within the month after
the study period, but we cannot be certain that none had an adverse event
and either did not seek medical attention or sought attention elsewhere.
Second, our study shows that azithromycin is ineffective, but it does not
identify the best treatment for acute bronchitis. Further studies are needed
to identify more effective management strategies,41,42 because many patients
with acute bronchitis require their physicians to "do something".2,3,43
Those very reasonable demands should be met with rigorous clinical trials,
not defensive use of ineffective antibiotics.
Contributors
Arthur Evans, Shahid Husain, and Lakshmi Durairaj designed the study,
coordinated data collection, interpreted results, and wrote the report.
Laura Sadowski helped design the study, develop the data-collection
instruments, interpret results, and prepare the report. Arthur Evans and Yue
Wang analysed the data. Marjorie Charles-Damte supervised recruitment and
assessment of patients and helped design the study, develop the
data-collection instruments, interpret results, and prepare the report.
Conflict of interest statement
None declared.
Acknowledgments
We thank Brendan Reilly and Robert Weinstein for their help in initiating
the study, monitoring its quality and safety, interpreting results, and
writing the report; and the staff and patients of the Ambulatory Screening
Clinic, Cook County Hospital. The Department of Medicine of Cook County
Hospital funded the project from internal funds.
Top <http://www.thelancet.com/search/#top>

References
1 Schappert SM. Ambulatory care visits to physician offices, hospital
outpatient departments, and emergency departments: United States, 1997.
Vital and Health Statistics. Series 13, No 143. Hyattsville, MD: US
Department of Health and Human Services, Centers for Disease Control and
Prevention, National Center for Health Statistics; 1999. DHHS publication no
2000-1714.
2 Gonzales R, Bartlett JG, Besser RE, et al. Principles of appropriate
antibiotic use for treatment of uncomplicated acute bronchitis: background.
Ann Intern Med 2001; 134: 521-29. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Ann+Intern+Me
d&volume=134&year=2001&page=521&display=abstract> ]
3 Gonzales R, Steiner JF, Sande MA. Antibiotic prescribing for adults with
colds, upper respiratory tract infections, and bronchitis by ambulatory care
physicians.  JAMA 1997; 278: 901-04. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=JAMA&volume=2
78&year=1997&page=901&display=abstract> ]
4 Huovinen P, Cars O. Control of antimicrobial resistance: time for action.
BMJ 1998; 317: 613-14. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=BMJ&volume=31
7&year=1998&page=613&display=abstract> ]
5 Gonzales R, Sande M. Uncomplicated acute bronchitis.  Ann Intern Med 2001;
133: 981-91. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Ann+Intern+Me
d&volume=133&year=2001&page=981&display=abstract> ]
6 Stott NC, West RR. Randomized controlled trial of antibiotics in patients
with cough and purulent sputum.  BMJ 1976; 2: 556-59. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=BMJ&volume=2&
year=1976&page=556&display=abstract> ]
7 Brickfield FX, Carter WH, Johnson RE. Erythromycin in the treatment of
acute bronchitis in a community practice.  J Fam Pract  1986; 23: 119-22.
[ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=23&year=1986&page=119&display=abstract> ]
8 Franks P, Gleiner JA. The treatment of acute bronchitis with trimethoprim
and sulfamethoxazole.  J Fam Pract  1984; 19: 185-90. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=19&year=1984&page=185&display=abstract> ]
9 Dunlay J, Reinhardt R, Roi LD. A placebo-controlled, double-blind trial of
erythromycin in adults with acute bronchitis.  J Fam Pract  1987; 25:
137-41. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=25&year=1987&page=137&display=abstract> ]
10 Verheij T, Hermans J, Mulder J. Effects of doxycycline in patients with
acute cough and purulent sputum: a double-blind placebo-controlled trial.
Br J Gen Pract  1994; 44: 400-04. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Br+J+Gen+Prac
t+&volume=44&year=1994&page=400&display=abstract> ]
11 Williamson HA. A randomized, controlled trial of doxycycline in the
treatment of acute bronchitis.  J Fam Pract  1984; 19: 481-86. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=19&year=1984&page=481&display=abstract> ]
12 King DE, Williams CW, Bishop L, Schecter A. Effectiveness of erythromycin
in the treatment of acute bronchitis.  J Fam Pract  1996; 42: 601-05. [
PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=42&year=1996&page=601&display=abstract> ]
13 Scherl ER, Riegler SL, Cooper JK. Doxycycline in acute bronchitis: a
randomized double-blind trial.  J Ky Med Assoc  1987; 85: 539-41. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Ky+Med+Asso
c+&volume=85&year=1987&page=539&display=abstract> ]
14 Hueston WJ. A comparison of albuterol and erythromycin for the treatment
of acute bronchitis.  J Fam Pract  1991; 33: 476-80. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=33&year=1991&page=476&display=abstract> ]
15 Fahey T, Stocks N, Thomas T. Quantitative systematic review of randomized
controlled trials comparing antibiotic with placebo for acute cough in
adults.  BMJ 1998; 316: 906-10. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=BMJ&volume=31
6&year=1998&page=906&display=abstract> ]
16 Smucny JJ, Becker LA, Glazier RH, McIsaac. Are antibiotics effective
treatment for acute bronchitis? A meta-analysis.  J Fam Pract  1998; 47:
453-60. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=47&year=1998&page=453&display=abstract> ]
17 Bent S, Saint S, Vittinghoff E, Grady D. Antibiotics in acute bronchitis:
a meta-analysis.  Am J Med 1999; 107: 62-67. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Am+J+Med&volu
me=107&year=1999&page=62&display=abstract> ]
18 Smucny J, Fahey T, Becker L, Glazier R, McIsaac W. Antibiotics for acute
bronchitis (Cochrane Review). In: The Cochrane Library, Issue 4, 2000.
Oxford: Update Software.
19 Soepandi P, Mangunnegoro H, Yunus F, Gunawan J. The pattern of
micro-organisms and the efficacy of new macrolides in acute lower
respiratory tract infections.  Respirology 1998; 3: 113-17. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Respirology&v
olume=3&year=1998&page=113&display=abstract> ]
20 Hoepelman IM, Mollers MJ, van Schie MH, et al. A short (3-day) course of
azithromycin tablets versus a 10-day course of amoxycillin-clavulanic acid
(co-amoxiclav) in the treatment of adults with lower respiratory tract
infections and effects on long-term outcome.  Int J Antimicrob Agents 1997;
9: 141-46. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Int+J+Antimic
rob+Agents&volume=9&year=1997&page=141&display=abstract> ]
21 Biebuyck XA. Comparison of azithromycin and co-amoxiclav in the treatment
of acute tracheobronchitis and acute infectious exacerbations of chronic
bronchitis in adults.  J Int Med Res  1996; 24: 407-18. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Int+Med+Res
+&volume=24&year=1996&page=407&display=abstract> ]
22 Sternon J, Leclerq P, Knepper C, Blot K. Azithromycin compared with
clarithromycin in the treatment of adult patients with acute purulent
tracheobronchitis: a cost of illness study.  J Int Med Res  1995; 23:
413-22. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Int+Med+Res
+&volume=23&year=1995&page=413&display=abstract> ]
23 Salit IE, Mederski B, Morisset R, Trottier S. Azithromycin for the
treatment of acute LRTIs: a multicenter, open-label study.  Infect Med
1998; 15: 773-77. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Infect+Med+&v
olume=15&year=1998&page=773&display=abstract> ]
24 http://www.medscape.com/viewarticle/417417
<http://www.medscape.com/viewarticle/417417> (accessed Feb 20, 2002).
25 Irwin RS, Boulet L-P, Cloutier MM, et al. Managing cough as a defense
mechanism and as a symptom: a consensus panel report of the American College
of Chest Physicians.  Chest 1998; 114: 133S-90S. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Chest&volume=
114&year=1998&page=133S&display=abstract> ]
26 Douglas RM, Chalker EB, Treacy B. Vitamin C for preventing and treating
the common cold (Cochrane Review). In: The Cochrane Library, Issue 4, 2000.
Oxford: Update Software.
27 Melbye H, Aasebo U, Straume B. Symptomatic effect of inhaled fenoterol in
acute bronchitis: a placebo-controlled double-blind study.  Fam Pract 1991;
8: 216-22. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Fam+Pract&vol
ume=8&year=1991&page=216&display=abstract> ]
28 Hueston WJ. Albuterol delivered by metered-dose inhaler to treat acute
bronchitis.  J Fam Pract  1994; 39: 437-40. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Fam+Pract+&
volume=39&year=1994&page=437&display=abstract> ]
29 Juniper EF, Guyatt GH, Ferrie PJ, Griffith LF. Measuring quality of life
in asthma.  Am Rev Respir Dis 1993; 147: 832-38. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Am+Rev+Respir
+Dis&volume=147&year=1993&page=832&display=abstract> ]
30 Juniper EF, Guyatt GH. Development and testing of a new measure of health
status for clinical trials in rhinoconjunctivitis.  Clin Exp Allergy 1991;
21: 77-83. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Clin+Exp+Alle
rgy&volume=21&year=1991&page=77&display=abstract> ]
31 Guyatt GH, Berman LB, Townsend M, Pugsley SO, Chambers LW. A measure of
quality of life for clinical trials in chronic lung disease.  Thorax 1987;
42: 773-78. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Thorax&volume
=42&year=1987&page=773&display=abstract> ]
32 Guyatt GH, Nogradi S, Halcrow S, et al. Development and testing of a new
measure of health status for clinical trials in heart failure.  J Gen Intern
Med  1989; 4: 101-07. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Gen+Intern+
Med+&volume=4&year=1989&page=101&display=abstract> ]
33 Juniper EF, Guyatt GH, Willan A, Griffith LE. Determining a minimal
important change in disease-specific quality of life questionnaire.  J Clin
Epidemiol  1994; 47: 81-87. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Clin+Epidem
iol+&volume=47&year=1994&page=81&display=abstract> ]
34 Juniper EF, Guyatt GH, Griffith LE, Perrie PJ. Interpretation of
rhinoconjunctivitis quality of life questionnaire data.  J Allergy Clin
Immunol  1996; 98: 843-45. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Allergy+Cli
n+Immunol+&volume=98&year=1996&page=843&display=abstract> ]
35 Guyatt GH, Juniper EF, Walter SD, Griffith LE, Goldstein RS. Interpreting
treatment effects in randomised trials.  BMJ 1998; 316: 690-93. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=BMJ&volume=31
6&year=1998&page=690&display=abstract> ]
36 Redelmeier DA, Goldstein RS, Guyatt GH. Assessing the minimal important
difference in symptoms: a comparison of two techniques.  J Clin Epidemiol
1996; 49: 1215-19. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=J+Clin+Epidem
iol+&volume=49&year=1996&page=1215&display=abstract> ]
37 Miettinen OS, Nurminen M. Comparative analysis of two rates.  Stat Med
1985; 4: 213-26. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Stat+Med+&vol
ume=4&year=1985&page=213&display=abstract> ]
38 Verheij T, Hermans J, Kaptein A, Mulder J. Acute bronchitis: course of
symptoms and restrictions in patients' daily activities.  Scand J Primary
Health Care 1995; 13: 8-12. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Scand+J+Prima
ry+Health+Care&volume=13&year=1995&page=8&display=abstract> ]
39 Holmes W, Macfarlane JT, Macfarlane R, Hubbard R. Symptoms, signs and
prescribing for acute lower respiratory tract illness.  Br J Gen Pract
2001; 51: 177-81. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Br+J+Gen+Prac
t+&volume=51&year=2001&page=177&display=abstract> ]
40 Macfarlane J, Holmes W, Gard P, et al. Prospective study of the
incidence, aetiology and outcome of adult lower respiratory tract illness in
the community.  Thorax 2001; 56: 109-14. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Thorax&volume
=56&year=2001&page=109&display=abstract> ]
41 Little P, Williamson I, Warner G, et al. Open randomised trial of
prescribing strategies in managing sore throat.  BMJ 1997; 314: 722-27. [
PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=BMJ&volume=31
4&year=1997&page=722&display=abstract> ]
42 Macfarlane JT, Holmes WF, Macfarlane RM. Reducing reconsultations for
acute lower respiratory tract illness with an information leaflet: a
randomised controlled study of 1006 patients in primary care.  Br J Gen
Pract  1997; 47: 719-22. [ PubMed
<http://www.ncbi.nlm.nih.gov/htbin-post/PubMed/wgetcit?journal=Br+J+Gen+Prac
t+&volume=47&year=1997&page=719&display=abstract> ]
43 Gonzales R, Sande M. What will it take to stop physicians from
prescribing antibiotics in acute bronchitis? Lancet 1995;

Edward E. Rylander, M.D.
Diplomat American Board of Family Practice.
Diplomat American Board of Palliative Medicine.