(Chest. 2001;119:1190-1209.)
© 2001
American College of Chest Physicians
Management of Acute Exacerbations of COPD*
A Summary and Appraisal of Published Evidence
Douglas C. McCrory, MD, MHSc;
Cynthia Brown, MD;
Sarah E. Gelfand, BA and
Peter B. Bach, MD
*
From the Center for Clinical Health Policy Research (Drs. McCrory and Brown), Duke Evidence-Based Practice Center and Duke University Medical Center, Durham, NC; and the Department of Epidemiology and Biostatistics (Ms. Gelfand and Dr. Bach), Health Outcomes Research Group, Memorial Sloan-Kettering Cancer Center, New York, NY. This paper also appeared in Annals of Internal Medicine 2001; 134:600–620.
Correspondence to: Peter B. Bach, MD, MAPP, Health Outcomes Research Group, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, Box 221, New York, NY 10021.
 |
Abstract
|
|---|
Study objectives: To critically review the
available data on the diagnostic evaluation, risk stratification, and
therapeutic management of patients with acute exacerbations of
COPD.
Design, setting, and participants:
English-language articles were identified from the following databases:
MEDLINE (from 1966 to week 5, 2000), EMBASE (from 1974 to week 18,
2000), HealthStar (from 1975 to June 2000), and the Cochrane Controlled
Trials Register (2000, issue 1). The best available evidence on each
subtopic then was selected for analysis. Randomized trials, sometimes
buttressed by cohort studies, were used to evaluate therapeutic
interventions. Cohort studies were used to evaluate diagnostic tests
and risk stratification. Study design and results were summarized in
evidence tables. Individual studies were rated as to their internal
validity, external validity, and quality of study design. Statistical
analyses of combined data were not performed.
Measurement
and results: Limited data exist regarding the utility of most
diagnostic tests. However, chest radiography and arterial blood gas
sampling appear to be useful, while short-term spirometry measurements
do not. In terms of the risk of relapse and the risk of death after
hospitalization for an acute exacerbation, there are identifiable
clinical variables that are associated with these outcomes. Therapies
for which there is evidence of efficacy include bronchodilators,
corticosteroids, and noninvasive positive-pressure ventilation. There
is also support for the use of antibiotics in patients with more severe
exacerbations. Based on limited data, mucolytics and chest
physiotherapy do not appear to be of benefit, and oxygen
supplementation appears to increase the risk of respiratory failure in
an identifiable subgroup of patients.
Conclusions:
Although suggestions for appropriate management can be made based on
available evidence, the supporting literature is spotty. Further
high-quality research is needed and will require an improved, generally
acceptable, and transportable definition of the syndrome "acute
exacerbation of COPD" and improved methods for observing and
measuring outcomes.
|
Introduction
|
|---|
This
article describes the background evidence for the clinical practice
guidelines entitled "The Evidence Base for Management of Acute
Exacerbations of COPD." A joint panel from the American College of
Physicians (ACP)-American Society for
Internal Medicine (ASIM) and the American College of Chest
Physicians (ACCP) assisted in the design, conduct, and development of
this summary, which is based in large part on the evidence report
produced by the Evidence-Based Practice center at Duke University,
Durham, NC.1
The primary aims of this article are to summarize and evaluate the
published data addressing the care of patients with acute exacerbations
of COPD and to improve the care that these patients receive by
identifying efficacious and inefficacious treatment strategies. We
first review the health impact of COPD. We then define the entity
acute exacerbation and describe the methods that we used to
identify and grade the available data on the care of patients with this
condition. In the "Results" section, we assess studies that
evaluate diagnostic techniques, prognostic and risk stratification
models, and an array of therapies and interventions. In the concluding
sections, we review important elements of postexacerbation management,
with special attention to follow-up care, and gradual titration of
therapeutic agents such as oxygen and corticosteroids. Last, we comment
on domains of management for patients with acute exacerbations that
would most benefit from further research.
|
COPD
|
|---|
In the United States at present, > 16 million adults are
afflicted with COPD, a slowly progressive condition that typically
becomes symptomatic in the fifth and sixth decade of life. As the US
population ages, the prevalence of this disease is expected to
climb.2
COPD currently accounts for approximately 110,000
deaths per year, making it, after heart disease, cancer, and stroke,
the fourth leading cause of death. Nonasthma COPD in the United States
annually accounts for 16,367,000 office visits, 500,000
hospitalizations, and direct health-care costs of $18
billion.3
4
The term COPD is used to describe a range of pathophysiologic entities
that are characterized by airflow obstruction, including chronic
bronchitis, emphysema, asthma, and bronchiectasis. In this article, and
in our guidelines, we focus our attention on the care of patients with
the chronic bronchitis and emphysema, an approach consistent with the
National Heart, Lung, and Blood Institute definition of COPD as an
"umbrella term used to encompass several more specific respiratory
conditions" including chronic (obstructive) bronchitis and
emphysema.5
In fact, separating these entities is
difficult both when evaluating clinical studies and when practicing
clinical medicine.
Causes of COPD include smoking (85 to 90% of all cases), genetic
factors (including 
1-antitrypsin deficiency),
passive smoking, occupational exposures, air pollution, and possibly
hyperresponsive airways. Although the precise distinctions between
chronic bronchitis and emphysema are a subject of debate, tradition
holds that chronic bronchitis is responsible for 85% of COPD. Patients
with chronic bronchitis experience intermittent airway inflammation
that leads to frequent, prolonged episodes of productive cough. In
contrast, 15% of patients with COPD suffer primarily from emphysema, a
disease in which destruction of the infrastructure of alveoli and
distal airspaces, and thus the portion of the lung that provides
elastic recoil, occurs. Both conditions predispose patients to a common
constellation of symptoms and signs, and to a collection of
derangements in respiratory function.
Spirometric testing is used to confirm the diagnosis of COPD.
Typical abnormalities include a decrease in FEV1
and a decrease in the ratio of FEV1 to FVC. Other
abnormalities include an increased residual volume and total lung
capacity, and a limited and incomplete response in
FEV1 to bronchodilators (incomplete
reversibility). A diminished diffusing capacity of the lung for carbon
monoxide is often seen in patients with emphysema, and a response to
bronchodilators can be seen in patients with concomitant asthma.
Several staging systems are available for patients with stable COPD.
Both the European Respiratory Society and the American Thoracic Society
systems use FEV1, which correlates most closely
with mortality and frequency of acute exacerbation, as the sole staging
characteristic. The British Thoracic Society staging definition also
includes clinical features of a patient’s cough, sputum, dyspnea, and
lung sounds (Table 1
).
|
What Is an Acute Exacerbation of COPD?
|
|---|
In evaluating the published literature, and in developing practice
guidelines, we have attempted to adhere to a generally accepted and
useful concept of an acute exacerbation or flare
of COPD. Unfortunately, many definitions exist, many authors employ
substantively different criteria, and many studies poorly describe
their inclusion criteria. As a generalization, however, most published
definitions embrace some combination of the following three clinical
findings: worsening of dyspnea; increase in sputum purulence; and
increase in sputum volume. Unlike the staging systems for stable COPD,
there are no standardized, validated grading systems for the severity
of an acute exacerbation. Probably the most commonly used system was
developed by Anthonisen and colleagues6
and is based on
these and other symptoms. Patients with type 1 (severe) exacerbations
have all three of the above symptoms, and those with type 2 (moderate)
exacerbations have two of three of the symptoms. Patients with type 3
(mild) exacerbations have at least one of these symptoms, as well as
one of the following clinical criteria: an upper respiratory tract
infection in the past 5 days; fever without another apparent cause;
increased wheezing; increased cough; or increase in respiratory rate or
heart rate by 20% above baseline (Table 1)
.6
Clinicians
should be aware that other conditions such as heart failure and
pulmonary embolism can mimic an acute exacerbation.
Tracheobronchial infections are believed to be a common inciting cause
of acute exacerbations of COPD; however, controversy exists regarding
the nature of the infectious agent, as well as its exact role. Sputum
obtained from patients with mild to moderately severe chronic
bronchitis routinely grow a variety of bacteria in cultures, including
Haemophilus influenzae (22%), Pseudomonas
aeruginosa (15%), Streptococcus pneumoniae (10%), and
Moraxella catarrhalis (9%).7
Nonpathogenic
bacteria, such as Haemophilus parainfluenzae, account for up
to one third of all isolates. Also, the following certain groups of
patients are more likely to be colonized with resistant organisms such
as Pseudomonas: patients from nursing homes; patients recently treated
with antibiotics; and patients admitted to ICUs. The role of these
colonizers in the pathogenesis of acute exacerbation remains unclear,
and their presence makes the interpretation of any sputum culture
difficult. Some investigators8
9
10
also have proposed that
Mycoplasma pneumoniae or Chlamydia pneumoniae may
precipitate between 1% and 10% of exacerbations, and
others11
12
have pointed out that the presence of
eosinophilic inflammation in bronchial biopsy specimens of patients
with exacerbations is consistent with viruses (notably rhinovirus)
playing an important role.
Acute exacerbations are clearly associated with environmental exposures
as well, as significant correlations between levels of respirable
particles (diameter, < 10 µm) and ozone have been linked to
hospital admission rates.13
Finally, severe exacerbations
may be precipitated by other serious clinical conditions, such as heart
failure, nonpulmonary infections, pulmonary embolism, and
pneumothorax.14
The outcomes of COPD exacerbations are similarly heterogeneous. While
nearly 50% of exacerbations are not reported to
physicians,15
16
exacerbations requiring hospitalization
are associated with an inpatient mortality of 3 to 4%.17
For those patients requiring treatment in an ICU for an acute
exacerbation, mortality rates are substantially higher (in hospital, 11
to 24%; by 1 year, 43 to 46%). 14
18
19
20
21
After an acute
exacerbation, most patients are expected to experience at least a
temporary decrement in functional status and quality of
life,16
22
23
and half of those patients who are
hospitalized are expected to be readmitted at least once in the ensuing
6 months.14
24
|
Materials and Methods
|
|---|
Identification of Topics for Literature Search
Topics to be covered in this article and in the practice
guideline were determined through a consensus process that involved
both the ACP-ASIM/ACCP expert panel and the technical advisory panel of
the Evidence-Based Practice Center at Duke University (Durham, NC). The
topic list was generated to address the following three questions: (1)
what information is available to aid clinicians in predicting the
clinical course of a patient with an acute exacerbation?; (2) what
information is available about the utility of diagnostic tests used to
evaluate patients with symptoms of acute exacerbation?; and (3) what
information is available to help guide clinicians in using available
therapies and interventions? In this article, we do not consider the
care of patients in stable condition with chronic COPD, experimental
(Exp) therapies that are not widely available, or the provision of
invasive mechanical ventilation.
Search Strategy
The information presented in this report was gathered through
systematic searches and ongoing surveillance of the MEDLINE (from 1966
to week 5, 2000), EMBASE (from 1974 to week 18, 2000), and HealthStar
(from 1975 to June 2000) databases and of the Cochrane Controlled
Trials Register (2000, issue 1). Search strategies included index terms
and text words for "COPD" and "acute exacerbation" and specific
terms relating to the interventions and outcomes discussed in ensuing
sections. Variations on several search strategies were tested in order
to locate the greatest number of relevant articles. The abstracts of
relevant articles were reviewed against predetermined criteria,
appropriate articles were retrieved, and the reference lists of those
were examined. Seven hundred seven full-text articles were obtained
through this process, and those that were eligible for analysis were
summarized in evidence tables. The data, study methods, and evidence
available in each article then were evaluated in the manner described
below.
Assessment of the Quality of Available Evidence
Each retrieved study was evaluated along the following two
dimensions: to what extent did the study enroll the patients in whom we
were interested (external validity [EV])?; and to what
extent did the study follow the optimal design (internal
validity)? Our criteria for EV hinged on the following two
questions: did the study enroll patients who had COPD by a conventional
definition (Table 1)
?; and did the study enroll patients with acute
exacerbations of COPD, as documented both by a description of the
cohort symptomatology and by a description of the diagnostic testing
that was used to exclude other etiologies? We generated a scoring
system for EV (Table 2
) that ranged from 0 (poorest quality) to 5 (highest quality) based on
the adequacy of the documentation of each study for these two concerns.
Sample size was not taken into consideration, and all comments about
the "significance" of results reflect that the authors reported
statistical significance at the p < 0.05 level.
Our criteria for internal validity differed when we evaluated Exp vs
observational (Obs) studies. To evaluate Exp studies, we employed the
scoring system described by Jadad and colleagues25
that
assigns scores based on the quality of design of randomized controlled
trials (RCTs) (Table 3
). Specifically, scores range from 0 to 5, and points are earned for
adequate randomization, blinding, and assessment of withdrawals and
dropouts. To evaluate Obs studies, we used the hierarchy of evidence
proposed by Ball et al26
(Table 4
). Unlike the Jadad scale for Exp designs, lower scores for the internal
validity of Obs studies denote a higher level of evidence. For studies
that presented prognostic models, clinical prediction rules, or
severity-of-illness algorithms, we assessed the extent of model
validation reported using the system proposed by Justice and
colleagues27
(Table 5
). This scoring system ranges from 0 to 5; higher scores reflect that
the prediction model presented in the article has been more extensively
evaluated on independent populations of patients.
For studies that appear in the tables, these scores are recorded in
those tables. For studies that are referenced only in the text, these
assessments are recorded in parentheses the first time the study is
mentioned in the following manner. EV is documented as a ratio of the
total number of points earned to the number of points possible
(eg, 3:5 [Table 2
]). For internal validity, the type of
study (Exp or Obs) is documented, followed by the score on the relevant
scale (see Tables 3
and 4
). The degree of validation of prognostic
models is relevant only to the studies presented in Tables 7
and 8
, and
are reported there.
Choice of Inclusion of Studies for Reporting and Analysis
The minimum threshold for inclusion of studies of different
design types was driven by the relative availability of studies in each
category. Randomized, placebo-controlled studies are considered to
produce the highest level of evidence, but for some treatment and
diagnostic modalities, information from these types of studies was
either scanty or lacking. Ultimately, we chose a different threshold of
inclusion for each topic based on the availability of relevant data
(Table 6
). The varying quality of the assessed studies is taken into account in
the evaluation.
|
Approach to the Patient With an Acute Exacerbation of COPD
|
|---|
In the following section, we discuss our recommendations and
findings for the following three domains of care for patients with
acute exacerbations of COPD: risk stratification of patients
(specifically, data on predictors of outpatient relapse) and predictors
of inpatient mortality; choice of diagnostic tests; and benefits and
risks of therapeutic interventions, including mucus clearance
strategies, bronchodilating agents, corticosteroids, antibiotics,
oxygen, and noninvasive mechanical ventilation. Three methodological
problems hindered our analysis. First, the care of patients with acute
exacerbations of COPD is sometimes characterized as "shotgun
therapy"; that is, most patients receive most available therapies. As
such, many studies designed to evaluate one intervention include
patients receiving other interventions, and these cointerventions make
analysis of the effects of single therapies more difficult, especially
when cointerventions are not standardized. Second, many studies
evaluate changes in FEV1 as the primary outcome
of interest because it can be safely and easily measured. This measure
of respiratory function, although a reliable predictor of other
clinical outcomes, is relatively insensitive to changes in clinical
condition when compared both to other quantitative measures (such as
arterial blood gas values) and to qualitative evaluations of
symptoms.15
28
Last, the majority of studies that we found
address the care of patients in emergency departments or inpatient
settings, while many patients with milder acute exacerbations do not
receive care in these settings. As such, our conclusions are more
focused on the care of patients with more severe exacerbations.
|
Risk Stratification
|
|---|
Prediction of Outpatient Relapse
Based on 10 studies that evaluated patients with acute
exacerbations of COPD in emergency departments (7 studies) and in the
outpatient setting (3 studies), we concluded that certain
characteristics are associated with patients returning for more
treatment rather than with those experiencing gradual improvement
(Table 7
). The ability to identify patients at high risk for relapse
should improve decisions about hospital admissions and follow-up
appointments. Several investigators have confirmed that patients who
have lower pretreatment or posttreatment FEV1
levels, who receive more bronchodilator treatments or corticosteroids
during their visit or have higher rates of prior relapse, are more
likely to return (ie, relapse) than are patients with more
favorable values of these characteristics. At present, the available
prediction models can provide clinical guidance based on these
identified predictors, and those patients with these characteristics
are at higher risk of relapse. It should be noted that these models,
however, show only moderately good discriminatory power. For example,
the best model derived to predict relapse (defined as a return to the
emergency department within 14 days of initial presentation) had a
sensitivity of 0.57 and a specificity of 0.72.29
Prediction of Inpatient Mortality
Based on 11 studies, we concluded that certain physiologic
characteristics are associated with a higher likelihood of inpatient
mortality. Prediction models containing these characteristics are
potentially useful for risk stratification in the context of
population-based and randomized studies. To the extent that these
characteristics are used to influence decisions about instituting,
continuing, or withdrawing life-sustaining therapies, caution should be
exercised. We identified no prediction models that were able to
identify patients who were virtually certain to die (for example, those
with a likelihood of death of 
90%) during their inpatient stay. It
should be noted also that in these studies, there is substantial
variability in the inclusion criteria, raising concerns about the EV of
some of these results. Of the 11 studies, 8 (Table 8
) documented an association between specific clinical predictors and
mortality, while the other 3 studies did not report significant
predictors.17
30
31
The two largest studies examining this
outcome are summarized below.
The Study to Understand Prognoses and Preferences for Outcomes and
Risks of Treatments enrolled 1,016 patients with acute exacerbations of
COPD on hospital admission.14
The patients
hada variety of etiologies for exacerbation,
includingrespiratory tract infection (including pneumonia) (48%),
congestive heart failure (26%), lung cancer (3.3%), pulmonary embolus
(1.4%), and pneumothorax (1%). The outcome of interest was mortality
by 180 days, which was 33% (2-year mortality, 49%). Significant
predictors were worse acute physiology score from the acute physiology
and chronic health evaluation (APACHE) III algorithm,32
lower body mass index, older age, worse functional status 2 weeks
before hospital admission, lower
PO2/fraction of inspired oxygen
ratio, history of congestive heart failure, lower serum albumin level,
presence of cor pulmonale, lower activities of daily living scores, and
lower Duke activity status index score. Predictions from the model that
included these variables showed good calibration (calibration index,
0.0016) and fair discrimination (area under receiver operating
characteristics curve, 0.731) in a validation cohort.
Another large prospective cohort study enrolled 362 patients who
were admitted to ICUs with respiratory failure because of COPD.
Patients with pneumonia, pulmonary edema, or pulmonary embolus were
excluded. The in-hospital mortality of 23.8% was predicted by the
number of pre-ICU hospital days and the nonrespiratory
component of the APACHE III score. A separate analysis identified the
following three predictors of 180-day mortality: acute physiology
score; old age; and a higher number of pre-ICU hospital days.
Activities of daily living were also a significant predictor on
univariable analysis.20
|
Diagnostic Testing
|
|---|
General Approach
Many assessment techniques frequently are used in evaluating
patients with acute exacerbations of COPD. These include measuring
routine laboratory values, performing a physical examination, obtaining
an ECG, assessing cardiac function, and instituting an empiric trial of
diuretics. We found no published evidence that could help us to
determine the utility of these approaches. For another commonly used
assessment (arterial blood gas sampling), we found indirect evidence in
a number of studies supporting its clinical utility. These studies,
which are covered in detail in other parts of this report, demonstrate
that arterial blood gas analysis is helpful both in terms of gauging
the severity of an exacerbation, and in identifying those patients
currently in need of oxygen therapy and those potentially requiring
mechanical ventilatory support. Two other diagnostic modalities, chest
roentgenography and spirometric testing, have been assessed and are
discussed below.
Chest Roentgenography in Establishing Causes/Coexisting Illnesses
in Acute Exacerbation of COPD
Based on three Obs studies, we concluded that for patients treated
in emergency departments or hospitals, a chest radiograph (CXR) is a
useful diagnostic test. A substantial rate of CXR abnormalities was
documented in the following two retrospective studies: 16% abnormality
rate from a study of 685 episodes occurring in a single urban emergency
department (EV, 0/4; internal validity, Obs 2b)33
; and
16% (7% judged as "clinically significant") occurring in 107
patients admitted to a single hospital (EV, 0/4; internal validity, Obs
2b).34
In a prospective cohort study of 128 hospital
admissions for asthma or COPD, 21% of patients had a change in
management that was prompted by their CXR result (the majority of these
patients had new pulmonary infiltrates or evidence of congestive heart
failure) (EV, 1/4; internal validity, Obs 1b).35
Models presented by these authors for predicting CXR abnormalities were
not sufficiently reliable to be clinically useful.
Spirometric Testing
Although several studies have shown that
FEV1 is loosely correlated with relapse rate,
based on three Obs studies, we concluded that spirometric assessment at
the time of presentation or during the course of treatment is of
limited usefulness in the care of patients with acute exacerbations of
COPD. Changes in clinical status do not correlate well, in general,
with changes in spirometric measures in patients with this disease. A
study performed in one urban emergency department (EV, 3:4; internal
validity, Obs 1b) enrolling 70 patients demonstrated that
FEV1 at the time of presentation was weakly, but
statistically significantly, correlated with both
PCO2 (r = -0.46;
p < 0.001) and pH (r = 0.33; p < 0.01) but was
uncorrelated with arterial
PO2. These results are
different from those seen in studies of patients with asthma presenting
to the emergency department, in which spirometry and arterial blood gas
levels are highly correlated.36
Another study enrolling
199 patients presenting with acute exacerbation of COPD in an urban
emergency department demonstrated that peak expiratory flow rate (PEFR)
and FEV1 are correlated (r = 0.84;
p < 0.001); the clinical implication of this finding, however, is
unclear (EV, 1:4; internal validity, Obs 1b).37
This
latter study also noted that for a minority of patients, the
absolute difference between the percent predicted values based on
FEV1 and those based on PEFR was > 10%.
|
Therapeutic Interventions
|
|---|
Bronchodilating Agents
Based on 14 randomized studies, we concluded the following: that
short-acting ß-agonist-type and anticholinergic-type inhaled
bronchodilators have comparable effects on spirometry and a greater
effect than all parenterally administered bronchodilators
(ie, parenteral methylxanthines and sympathomimetics); that
the toxicity profile of the methylxanthine agents makes them
potentially harmful; and that there may be an additional benefit in
some patients when a second bronchodilating agent is administered once
the maximal dose of the initial inhaled bronchodilator is reached.
These generalizations are limited by the small number of analyzable
trials38
39
that have been published, the substantial
differences in inclusion and exclusion criteria between them, and the
variability in drug dosages that were studied.
Efficacy of Bronchodilators:
There were five RCTs that
compared individual bronchodilating agents. Two RCTs39
40
compared the efficacy of inhaled ipratropium bromide to that of
short-acting ß-agonists (EV, 2:5; internal validity, Exp
3:539
; EV, 3:5; internal validity, Exp
5:540
). The first study enrolled 40 patients and observed
that FEV1 among those receiving ipratropium showed
statistically significant improvement from day 1 to day 7 at 15 and 30
min after administration, while no significant differences were seen at
0, 5, 10, 60, 120, and 240 min after administration. Similarly, the
only significant improvement observed in patients receiving fenoterol
was at 60 min after treatment on day 7 (p < 0.05).39
The second study involved 32 patients in a crossover design comparing
ipratropium and metaproterenol. At 30 min after administration,
patients receiving ipratropium had a significant rise in
PaO2, while those receiving metaproterenol had
a significant fall in PaO2. At 90 min, these
differences had disappeared, and both patient groups showed a
significant improvement in FEV1. However, no additional
improvement was seen after the patients were crossed over to treatment
with the second drug.40
In a study of 90 patients with
asthma and/or COPD during transport to an emergency department,
treatment with nebulized albuterol was compared to treatment with
subcutaneous terbutaline. Patient-perceived improvement, respiratory
rate, and dyspnea rating showed significant improvements only in the
group receiving albuterol (p < 0.05) (EV, 0:5; internal validity,
Exp 5:5).41
In a dosing study42
involving 86
patients, there were no significant differences in FEV1 at
2 h between patients receiving nebulized albuterol, 2.5 mg, given
every 20 min and those receiving nebulized albuterol, 2.5 mg, given
every hour, although there was a suggestion that patients with lower
FEV1 benefited from the former regimen (EV, 1:5; internal
validity, Exp 4:5).
Incremental Benefit of a Second Bronchodilator:
The addition
of a methylxanthine to inhaled bronchodilators was examined in three
randomized studies. One study43
involving 143 patients
with asthma and COPD receiving care in an emergency department reported
a trend toward lower hospitalization rates for patients given
aminophylline in addition to short-acting ß-agonists and
corticosteroids (EV, 1:5; internal validity, Exp 3:5). Two
studies44
45
found no significant differences in measured
changes in FEV1 between patients receiving standard therapy
(including short-acting ß-agonists) and those who also received
aminophylline (EV, 4:5; internal validity, Exp 5:544
; EV,
1:5; internal validity, Exp 4:545
).
The effect of adding a second class of bronchodilator (ie,
anticholinergic or short-acting ß-agonists) to a full-dose regimen of
the other agent has been examined in seven randomized studies. Six of
these studies38
46
47
48
49
50
specifically examined the impact of
an anticholinergic added to a short-acting ß-agonist for treatment of
acute exacerbations of COPD. In a study46
of 57 emergency
department patients, the addition of glycopyrrolate to albuterol
resulted in a proportionally larger increase in
FEV1 than that experienced by patients treated
with albuterol alone. (EV, 2:5; internal validity, Exp 4:5). A
study47
of 68 emergency department patients found that the
addition of ipratropium to isoetharine resulted in significantly lower
lengths of stay but that admission rates to the hospital were similar
(EV, 1:5; internal validity, Exp 5:5). Three other
studies38
48
49
were unable to detect a difference in
spirometry (FEV1 and/or FVC) in patients treated
with short-acting ß-agonists alone when compared to those who also
were given anticholinergic agents (EV, 3:5; internal validity, Exp
4:548
; EV, 1:5; internal validity, Exp 2:538
;
and EV, 1:5; internal validity, Exp 4:549
). A three-armed
study50
examined 52 emergency department patients
receiving a short-acting ß-agonist alone (fenoterol), an
anticholinergic alone (ipratropium), or both agents. At 90 min,
patients in all three groups experienced similar improvements in
FEV1. Patients receiving ipratropium alone had
the lowest rate of reported side effects (EV, 2:5; internal validity,
Exp 5:5).
Adverse Effects:
The adverse effects of bronchodilators are
varied. The side effects of ipratropium bromide are generally fewer and
milder. Three RCTs39
47
49
did not report any adverse
effects with ipratropium bromide. Other effects include increased
incidence of tremors and dry mouth,40
50
and urinary
retention when used in combination with albuterol.48
The
adverse effects of albuterol include tremors, headache, nausea,
vomiting, and palpitations. Adverse cardiovascular effects such as
changes in heart rate, BP, and ECG tracings are also possible but
rare.51
Adverse effects associated with theophylline
include nausea, vomiting, headache, arrhythmias, and
seizures.44
52
The effects are more significant among
those patients with higher levels of theophylline.
Bronchodilating Agent Delivery Devices
Based on eight RCTs53
54
55
56
57
58
59
60
comparing metered-dose
inhalers (MDIs) and nebulizers in patients with acute exacerbations of
COPD, we concluded that there is insufficient evidence to support the
conclusion that one delivery modality is superior to the other. Of the
eight studies, six53
54
55
57
59
60
described using spacer
devices with the MDIs, one56
specifically mentioned using
an MDI without a spacer, and one (an abstract)58
did not
describe whether or not a spacer was used. The percentage improvement
in the FEV1 was significantly larger after
treatment with the nebulizer than with the MDI in two
studies57
58
but was not significantly different in the
other six.53
54
55
56
59
60
A meta-analysis61
of bronchodilator delivery devices in acute airflow obstruction
included these studies of COPD and additional studies of patients with
asthma. The meta-analysis found a negligible effect of nebulizers vs
MDI that is neither clinically nor statistically significant. The doses
of the bronchodilator administered by MDIs in these studies were lower
than those delivered by nebulizer and were lower than those often used
in clinical practice, and, thus, the few positive results may reflect
differences in the dose of the bronchodilator actually received.
Furthermore, the studies were all rather small, resulting in imprecise
estimates of the efficacy of MDI vs nebulizer delivery.
Corticosteroid Drugs
Based on six randomized, placebo-controlled studies, we concluded
that a short course of systemic corticosteroid therapy given to
patients with acute exacerbations of COPD improves spirometry and
decreases the relapse rate (Table 9
). However, the optimal dose and duration of treatment remain uncertain,
and few data exist documenting the efficacy of corticosteroids for
patients cared for in outpatient settings. There was a great deal of
variability in the dosage, length of treatment, administration, and
setting among the studies evaluated.62
63
64
65
66
67
In the largest
study, the Systemic Corticosteroids in COPD Exacerbations (SCCOPE)
trial, 271 patients admitted for acute exacerbations of COPD at one of
25 Veterans Administration hospitals were assigned to receive placebo
or 3 days of IV methylprednisolone followed by a course of oral
prednisone.67
For the combined glucocorticoid group, the
risk of treatment failure was reduced by 10% (33% vs 23%), and
FEV1 showed an improvement averaging
approximately 0.1 L in the first 3 days of treatment. The change in
FEV1 is similar to the magnitude of benefit
reported in smaller trials. The SCCOPE trial demonstrated equivalence
between an 8-week regimen and a 2-week regimen, the latter consisting
of the following: methylprednisolone, 125 mg IV every 6 h (on days
1 to 3); oral prednisone, 60 mg each day (on days 4 to 7); oral
prednisone, 40 mg each day (on days 8 to 11); and oral prednisone, 20
mg each day (on days 12 to 15).
Several trials have examined the time course of improvement in
FEV1 during treatment with systemic
corticosteroids. The difference in FEV1 between
glucocorticoid-treated and placebo-treated patients in the SCCOPE trial
was highest after the first day of treatment, remained statistically
significant after the second and third days, and was no longer
significant at 2 weeks. Of two trials63
64
that considered
short-term outcomes of emergency department treatment,
one64
observed similar improvements in
FEV1 in patients receiving corticosteroids and
placebo, and the other63
demonstrated a significant
improvement in FEV1 over time for patients
receiving corticosteroids but did not compare these patients to those
receiving placebo. Those trials that measured
FEV1 changes over longer periods of time, in
contrast, have shown more conclusive results.
The most common adverse effect associated with systemic corticosteroids
for acute exacerbation of COPD was hyperglycemia.66
67
In
the SCCOPE trial, two thirds of the episodes of hyperglycemia requiring
treatment occurred in patients who were known to have diabetes
mellitus. Nearly all episodes occurred in the first 30 days, and
whether hyperglycemia was more frequent or severe in the 8-week or the
2-week course of therapy was not described.67
Antibiotics
Based on 11 randomized, placebo-controlled studies of antibiotic
treatment, we concluded that antibiotics are beneficial in the
treatment of patients with acute exacerbations of COPD (Table 10
).6
68
69
70
71
72
73
74
75
76
77
Patients with more severe exacerbations are more
likely to experience benefit than those who are less ill. These
conclusions are consistent with those of a recent meta-analysis that
included many of the trials reviewed herein.78
It should
be noted that many of the studies that do show benefit were performed
before the emergence of respiratory pathogens that are resistant to
multiple antibiotics.
In their meta-analysis, Saint and colleagues78
included
nine RCTs of antibiotics. These trials used the following variety of
outcome measures: PEFR; duration of exacerbation;
PaO2; symptom score; and overall
severity score as determined by a physician. Three of nine
studies6
68
69
70
71
72
73
74
75
found a statistically significant benefit
for antibiotics, three found a trend favoring antibiotics, and three
failed to show any difference from placebo. The most consistently
measured end point across studies, improvement in PEFR, was estimated
to improve a mean of 10.75 L/min more in patients treated with
antibiotics than in patients treated with placebo (95% confidence
interval [CI], 4.96 to 16.54).
Three of these studies6
68
75
analyzed the efficacy of
antibiotics within subgroups defined either by evidence of bacterial
infection or by severity of illness. One trial6
found that
an a priori selection of patients with more severe
exacerbations (using the above-mentioned grading system [Table 1
])
identified those more likely to benefit from antibiotic treatment.
Patients with type 1 exacerbations (severe) experienced a benefit
(antibiotic-treated patients, 63%; placebo-treated patients, 43%).
The benefit of antibiotic treatment was less apparent in less severe
exacerbations (type 1 vs type 2 exacerbations, 70% vs 60%; type 1 vs
type 3 exacerbations, 74% vs 70%). Another study68
demonstrated that physician-assigned severity was correlated with a
likelihood of benefit from antibiotics. Among patients with "mild"
attacks, there were no significant differences between patients treated
with antibiotics and those treated with placebo. Among patients with
"moderate" or "severe" attacks, patients treated with
antibiotics had significantly fewer severe symptoms on days 2 and 7. A
third study75
demonstrated a similar relationship between
the severity of the exacerbation and the benefit from antibiotics.
However, this study included patients with clinical evidence of
pneumonia among those with severe exacerbations.
There is little evidence regarding the appropriate duration of
administration of antibiotics. Typical administration periods range
from 3 to 14 days in both placebo-controlled and head-to-head
comparisons of antibiotics for this condition. A single retrospective
study of patients receiving amoxicillin for acute exacerbations of COPD
found a clinically favorable response in 70% of patients who received
between 6 and 10 days of treatment. No follow-up assessment was
performed.79
Oxygen Therapy
Oxygen therapy provides enormous benefits to patients with acute
exacerbations of COPD who are hypoxemic (ie, the
PO2 level in arterial blood
is reduced). Oxygen relieves pulmonary vasoconstriction and right heart
strain, and lessens myocardial ischemia, thereby improving cardiac
output and oxygen delivery to the CNS and to other critical organs.
There is also a substantial amount of evidence supporting the
hypothesis that improved oxygen delivery to the lung enhances pulmonary
defenses and augments mucociliary transport. The major concern for most
clinicians administering oxygen to patients with acute exacerbations of
COPD is the risk that oxygen supplementation will lead to hypercarbia
and subsequent respiratory failure. Various mechanisms have been
advanced to explain this observation, including depression of
respiratory drive, alteration in ventilation/perfusion matching, and
the Haldane effect (ie, oxygenated erythrocytes have lower
capacity for CO2 than deoxygenated erythrocytes).
Based on four Obs studies,80
81
82
83
we concluded that oxygen
administration in patients with acute exacerbations of COPD may result
in hypercarbia but that there are methods for identifying the patients
at highest risk for developing respiratory failure associated with
oxygen administration.
A study80
of 23 patients with respiratory failure
accompanying COPD (EV, 3:4; internal validity, Obs 1b) who were given
28% oxygen demonstrated that arterial
PCO2 increased in 17 patients, with a
mean rise of 4 mm Hg (range, -2 to 11 mm Hg). The authors
stated that in no patient was serious CO2
retention encountered. A study83
of seven patients (EV,
3:4; internal validity, Obs 4) with acute exacerbations given both
24.5% and 28% oxygen demonstrated that
PCO2 increased in six of the seven
patients. A study82
of 53 patients (EV, 1:4; internal
validity, Obs 2b) with acute exacerbations who were given graded oxygen
therapy to raise oxygen saturation had similar findings. All but three
patients had elevations in PCO2, and
the greatest rise was observed in patients who presented with the
lowest PaO2 levels. The largest
study81
(EV, 2:4; internal validity, Obs 1b) to address
this issue enrolled 50 patients with acute exacerbations of COPD and
patients received 24% oxygen, followed by 28% oxygen if hypoxemia
persisted. Thirteen of the patients (26%) developed hypercarbia and
subsequently required mechanical ventilation. These 13 patients did not
differ from the 37 who did not require mechanical ventilation in terms
of age, baseline pulmonary function test results, or initial response
to therapy. Notably, the relationship between arterial pH and
PaO2 at the time of presentation was
predictive of respiratory failure but resting
PCO2was not. The authors
derived a discriminant function (Fig 1
) for predicting respiratory failure (pH =7.66–0.00910
PaO2) that had a sensitivity
of 77%. The authors then validated this predictive function in a
cohort of 76 subsequent patients, 16 of whom (21%) required mechanical
ventilation. Of these 16 patients, 13 had values of pH and
PaO2 that intersected below the
discriminant line (sensitivity, 81%). Although, to our
knowledge, this predictive model does not currently see heavy use, it
does emphasize that patients with simultaneous hypercarbia and
hypoxemia are at the greatest risk of developing respiratory failure.
View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. The discriminant function,
pH = 7.66–0.00919 (PaO2), helps to identify
patients at risk for carbon dioxide retention after the administration
of supplemental oxygen. When the patient’s observed
PaO2 is entered into the equation, the pH that
has been calculated can be compared with the measured pH to distinguish
between high-risk and low-risk patients. If a patient is at high risk,
the value calculated will be greater than that observed in the arterial
blood gas. The symbols represent PaO2 and pH
values on hospital admission of patients who were eventually intubated
(triangles) or not nonintubated (circles) in a study evaluating this
predictive model. The diagonal line reflects values of the discriminant
function and separates high-risk and low-risk patients. Adapted with
permission from Wysocki et al.98
|
|
To our knowledge, there are no available data directly addressing the
titration of oxygen after an acute exacerbation of COPD. Perhaps the
best data can be extrapolated from the Nocturnal Oxygen Therapy Trial,
which found that 20% of the 800 patients studied no longer required
oxygen 3 weeks after hospital discharge after acute exacerbations of
COPD.84
Mucus Clearance Strategies
Expectorants, Mucolytics, and Mucokinetics:
Based on five
RCTs73
85
86
87
88
involving five different drugs, we concluded
that pharmacologic mucus clearance strategies have not been
demonstrated to shorten the course of treatment for patients with acute
exacerbations of COPD, although there is a possibility that these
agents improve symptoms. There were no statistically significant
differences reported in mean FEV1 between treatments in any
study. Comparisons tested included domiodol vs control (EV, 1:5;
internal validity, Exp 1:5),85
bromhexine vs placebo (EV,
2:5; internal validity, Exp 5:5),86
ambroxol vs control
(EV, 2:5; internal validity, Exp 3:5),87
S-carboxymethylcysteine vs bromhexine (EV, 3:5; internal validity, Exp
4:5),88
and potassium iodide vs chloramphenicol,
physiotherapy, and placebo (EV, 2:5; internal validity, Exp
1:5).73
Of the five trials measuring subjective symptom
scores on difficulty with expectoration, only two85
87
reported significant differences (p < 0.01) favoring the mucolytic
drug over the control.
Physical and Respiratory Therapies:
Based on three
RCTs73
89
90
of chest physiotherapy and one Obs
study,91
we conclude that mechanical percussion of
the chest as applied by physical/respiratory therapists is ineffective
and perhaps even detrimental in the treatment of patients with acute
exacerbations of COPD. None of the randomized trials (EV, 3:5, internal
validity, Exp 3:589
; EV, 2:5; internal validity, Exp
1:573
; and EV, 2:5; internal validity, Exp
1:590
) reported any improvement in ventilatory function
(either FEV1 or FVC). One RCT90
described a
significantly lower FEV1 in patients who received chest
percussion therapy compared with control subjects. A similar transient
decrease in FEV1 after chest percussion was previously
described in an uncontrolled study.91
No other adverse
effects were reported.
Noninvasive Positive-Pressure Ventilation
Based on five RCTs92
93
94
95
96
and five Obs
studies,99
100
101
102
103
we concluded that noninvasive
positive-pressure ventilation (NPPV) is a beneficial support strategy
that, in selected hospitalized patients with respiratory failure,
decreases the likelihood of requiring invasive mechanical ventilation
and, possibly, improves survival time (Table 11
). In some of these studies, the exclusion criteria were omitted from
the reports, while in others, exclusion criteria included significant
cardiovascular disease, lack of mental capacity, presence of pneumonia,
and concern about upper airway narrowing or obstruction. As such, the
selection criteria for this therapy remain unclear.
Among the four RCTs92
93
94
95
that compared NPPV to a standard
therapy control, a significant difference in need for intubation was
found in two trials,94
95
with reduced need for
intubation in the NPPV groups (26% vs 74% in a study involving 85
patients94
; 9% vs 67% in a study involving 23
patients95
). A fifth trial, comparing NPPV to a
respiratory stimulant medication (doxapram) demonstrated a mortality
benefit associated with NPPV that was not statistically
significant.96
A meta-analysis97
published in
1996 that included three of the above trials, as well as three
published abstracts97a
97b
97c
and one other published
study,98
concluded that the risk of death was lower in
patients who were randomized to receive NPPV (odds ratio, 0.29; 95%
CI, 0.15 to 0.59), as was the risk of requiring invasive mechanical
ventilation (odds ratio, 0.20; 95% CI, 0.11 to 0.36). The results from
four prospective case series99
100
101
102
were similar to those
from the RCTs when NPPV-treated patients were compared to historical
control subjects. One Obs study103
found no increased
effectiveness of NPPV over more conventional treatment and observed a
large number of adverse effects associated with the use of NPPV.
Additional questions addressed in the literature include comparisons
between NPPV and invasive ventilation, optimal NPPV delivery methods,
and predictors of the successful application of NPPV. Four prospective
controlled studies compared types of NPPV delivery methods (EV, 3:5;
internal validity, Exp 0:5104
; EV, 1:5; internal validity,
Exp 1:5105
; EV, 1:5; internal validity, Exp
1:5106
; and EV, 4:5; internal validity, Exp
2:5107
). Outcomes of interest were the effect on gas
exchange, the need for intubation, mortality, adverse effects/side
effects, and the comfort with which the devices may be used. No
significant differences in these parameters were seen among the various
modes of ventilation. A retrospective study attempting to identify
parameters that could predict a successful outcome with the use of NPPV
looked at anthropometric and demographic characteristics, nutritional
status, spirometry, blood gas levels, and causes of acute exacerbation
of COPD. Factors that predicted success included higher pH, lower
PaCO2, and higher FVC (p < 0.05).
Poor outcomes were associated with a diagnosis of pneumonia, poor
nutritional status, and decreased compliance with the
apparatus.108
124
125
|
Research Priorities
|
|---|
In a disease held responsible for 5% of all deaths in the United
States, enormous disability, and $18 billion dollars in annual
health-care costs, the paucity of primary data on therapeutics
is startling. We found that in more than 40 years of research, fewer
than 1,100 patients had been enrolled in randomized, placebo-controlled
trials of antibiotics, fewer than 650 patients had been enrolled in
studies of corticosteroids vs placebo (before the 1999 SCOPPE trial,
the count was less than 400), and virtually no controlled trials (to
our knowledge) have enrolled patients with milder (outpatient)
exacerbations. Certainly, more in-depth research into therapeutics and
management would greatly benefit patients with this disease.
To be maximally beneficial, however, more groundwork is required. At
present, we lack a reproducible, transportable definition of acute
exacerbation, and we also lack an objective rating system for severity.
Equally important, there is no consensus on the outcomes that should be
measured and reported in clinical studies, although there is an
emerging recognition that nonphysiologic outcomes such as
symptomatology, quality of life, and interval before subsequent relapse
are all important to patients. Given these opportunities, our first
research objectives must include untangling the questions surrounding
the selection of patients for antibiotic and corticosteroid treatment,
identifying optimal dosing and durations for these agents, and
determining to what degree broad-spectrum and narrow-spectrum
antibiotics have similar efficacy.
There are a number of potentially promising new research directions as
well, including the following: (1) the components of mucus formation,
content, release, and transport; (2) strategies for improving muscle
strength and reducing muscle fatigue; (3) therapies aimed at aborting
the exacerbation cycle, including arrest of the inflammatory cascade;
(4) strategies aimed at preventing infectious exacerbations, perhaps
through reducing bacterial adherence or limiting cellular damage in the
presence of microorganisms; and (5) the determination of biological
markers of infection and inflammation (eg, antielastase,
antioxidant, and cytokine release or action) in the blood and/or
sputum.
|
Acknowledgements
|
|---|
We gratefully acknowledge the assistance of the
combined ACP-ASIM and ACCP expert panel, the Evidence-Based Center peer
review and technical advisory panels, and the efforts of Ruth E.
Goslin, MAT, and Rebecca N. Gray, DPhil.
|
Footnotes
|
|---|
Abbreviations:
ACCP = American College of Chest Physicians; ACP = American College
of Physicians; APACHE = acute physiology and chronic health
evaluation; ASIM = American College of Physicians-American Society
for Internal Medicine; CI = confidence interval; CXR = chest
radiograph; EV = external validity; Exp = experimental;
MDI = metered-dose inhaler; NPPV = noninvasive positive-pressure
ventilation; Obs = observational; PEFR = peak expiratory flow rate;
RCT = randomized controlled trial; SCCOPE trial = Systemic
Corticosteroids in COPD Exacerbations trial
This article is based on research conducted by investigators at
Memorial Sloan-Kettering Cancer Center, New York, NY, under contract
with the ACP–ASIM and the ACCP, and by investigators at Duke
University, Durham, NC, under contract with the Agency for Healthcare Research and Quality (contract No. 290–97-0014).
The authors of this article are responsible for its contents, including
any clinical or treatment recommendations. No statement in this article
should be construed as an official position of the Agency for
Healthcare Research and Quality of the US Department of Health and
Human Services.
Received for publication August 1, 2000.
Accepted for publication December 8, 2000.
|
References
|
|---|
-
Agency for Healthcare Research and Quality. Management of Acute Exacerbations of Chronic Obstructive Pulmonary Disease Report No. 19, 2000; Agency for Healthcare Research and Quality publication No. 00-E020. Also available at: http://www.ahrq.gov/clinic/copdsum.htm
-
. The National Lung Health Education Program. (1998) Strategies in preserving lung health and preventing COPD and associated diseases. Chest 113,123S-163S[ISI][Medline]
-
. Department of Commerce (1997) Statistical abstract of the United States 1997; US Department of Commerce, Bureau of the Census Department of Commerce Washington, DC.
-
Agency for Health Care Research and Policy. Healthcare cost and utilization project: nationwide inpatient sample for 1997. Available at: http://www.ahcpr.gov/data/hcup/hcupnet.htm
-
Petty, TL, Weinmann, GG (1997) Building a national strategy for the prevention and management of and research in chronic obstructive pulmonary disease: National Heart, Lung, and Blood Institute Workshop Summary; Bethesda, Maryland, August 29–31, 1995. JAMA 277,246-253[CrossRef][ISI][Medline]
-
Anthonisen, NR, Manfreda, J, Warren, CP, et al (1987) Antibiotic therapy in exacerbations of chronic obstructive pulmonary disease. Ann Intern Med 106,196-204
-
Miravitlles, M, Espinosa, C, Fernandez-Laso, E, et al (1999) Relationship between bacterial flora in sputum and functional impairment in patients with acute exacerbations of COPD. Chest 116,40-46[Abstract/Free Full Text]
-
Gump, DW, Phillips, CA, Forsyth, BR, et al (1976) Role of infection in chronic bronchitis. Am Rev Respir Dis 113,465-474[ISI][Medline]
-
Beaty, CD, Grayston, JT, Wang, SP, et al (1991) Chlamydia pneumoniae, strain TWAR, infection in patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 144,1408-1410[ISI][Medline]
-
Smith, CB, Golden, C, Klauber, MR, et al (1976) Interactions between viruses and bacteria in patients with chronic bronchitis. J Infect Dis 134,552-561[ISI][Medline]
-
Schaberg, T, Gialdroni-Grassi, G, Huchon, G, et al (1996) An analysis of decisions by European general practitioners to admit to hospital patients with lower respiratory tract infections. Thorax 51,1017-1022[Abstract]
-
Ball, P (1995) Epidemiology and treatment of chronic bronchitis and its exacerbations. Chest 108,43S-52S[Medline]
-
Schwartz, J (1994) Air pollution and hospital admissions for the elderly in Detroit, Michigan. Am J Respir Cr
TOP
Abstract
Introduction
