Aus der Abteilung Allgemeinmedizin

(Prof. Dr. med. M. M. Kochen, MPH, FRCGP)

im Zentrum Innere Medizin

der Medizinischen Fakultät der Universität Göttingen

The treatment of community-acquired pneumonia

in ambulatory patients

- A systematic review and meta-analysis -

(Behandlung der ambulant erworbenen Pneumonie bei ambulanten Patienten

– eine systematische Übersicht und eine Meta-Analyse)

INAUGURAL–DISSERTATION

zur Erlangung des Doktorgrades

der Medizinischen Fakultät

der Georg-August-Universität zu Göttingen

vorgelegt von

Lise M. Bjerre, MSc, MD

aus

Montréal, Canada

Göttingen 2002
 2

Dekan: Prof. Dr. med. M. Droese

I. Berichterstatter: Prof. Dr. med. M.M. Kochen, MPH, FRCGP

II. Berichterstatter/in: Prof. Dr. med. H. Hartmann

III. Berichterstatter/in:

Tag der mündlichen Prüfung: 24. Juni 2003

 3

Table of contents

1. INTRODUCTION ................................................................................................. 5
1.1 FRAMEWORK OF THE REVIEW: THE COCHRANE COLLABORATION .............................. 6
1.2 STATEMENT OF AUTHORSHIP ....................................................................................... 7
2. BACKGROUND AND LITERATURE REVIEW................................................... 9
2.1 COMMUNITY-ACQUIRED PNEUMONIA .......................................................................... 9
2.1.1 Incidence ............................................................................................................ 9
2.1.2 Etiology ............................................................................................................ 10
2.1.3 Risk factors ....................................................................................................... 11
2.1.4 Diagnosis.......................................................................................................... 12
2.1.5 Treatment ......................................................................................................... 13
2.1.5.1 Antibiotic resistance ................................................................................................ 14
2.1.6 Prognosis.......................................................................................................... 15
2.2 PRACTICE GUIDELINES FOR THE TREATMENT OF COMMUNITY-ACQUIRED PNEUMONIA ..
.................................................................................................................................. 15
2.2.1 American Thoracic Society (2001)................................................................... 16
2.2.2 British Thoracic Society (2001) ....................................................................... 19
2.2.3 Canadian Infectious Diseases Society / Canadian Thoracic Society (2000) ... 21
2.2.4 Infectious Diseases Society of America (2000) ................................................ 22
2.3 CURRENT BEST EVIDENCE IN THE TREATMENT OF COMMUNITY-ACQUIRED PNEUMONIA
.................................................................................................................................. 24
3. OBJECTIVES.................................................................................................... 27
4. METHODS ........................................................................................................ 28
4.1 STUDY SELECTION CRITERIA ..................................................................................... 28
4.1.1 Types of studies ................................................................................................ 28
4.1.2 Types of participants ........................................................................................ 28
4.1.3 Clinical Signs and Symptoms ........................................................................... 29
4.1.4 Types of interventions....................................................................................... 29
4.1.5 Types of outcome measures.............................................................................. 30
4.2 SEARCH STRATEGY FOR IDENTIFICATION OF STUDIES ................................................ 31
4.3 SELECTION PROCESS .................................................................................................. 31
4.4 DATA EXTRACTION ................................................................................................... 35
4.5 ANALYSES ................................................................................................................. 35
5. RESULTS.......................................................................................................... 36
5.1 DESCRIPTION OF STUDIES .......................................................................................... 36
5.1.1 Number of trials and trial size.......................................................................... 38
5.1.2 Diagnoses ......................................................................................................... 38
5.1.3 Diagnostic criteria ........................................................................................... 39
5.1.4 Out- vs Inpatients ............................................................................................. 39
5.1.5 Patient inclusion and exclusion criteria........................................................... 39
5.1.6 Antibiotics......................................................................................................... 40
5.1.7 Methodological quality of included studies ..................................................... 40
 4

5.2 STUDY RESULTS ........................................................................................................ 41

5.2.1 Efficacy analysis (“primary analysis”)............................................................ 41
5.2.1.1 Comparisons across antibiotic groups .................................................................... 46
5.2.1.2 Side-effects ............................................................................................................. 46
5.2.1.3 Bacterial pathogens ................................................................................................ 49
5.2.1.4 Serologically identified pathogens .......................................................................... 51
5.2.2 Effectiveness analysis (open-label studies) ...................................................... 52
5.2.2.1 Effect of administration schedule on adherence to treatment................................. 52
6. DISCUSSION .................................................................................................... 54
6.1 EVIDENCE.................................................................................................................. 54
6.2 METHODS .................................................................................................................. 54
6.3 IMPLICATIONS FOR PRACTICE .................................................................................... 55
6.4 IMPLICATIONS FOR RESEARCH ................................................................................... 56
7. SUMMARY........................................................................................................ 57
8. ZUSAMMENFASSUNG (SUMMARY IN GERMAN) ......................................... 60
9. APPENDICES ................................................................................................... 63
9.1 LIST OF ABBREVIATIONS ............................................................................................ 63
9.2 LIST OF FIGURES ........................................................................................................ 64
9.3 LIST OF TABLES ......................................................................................................... 65
10. REFERENCES .............................................................................................. 66
 5

1. Introduction

Community-acquired pneumonia (CAP) is defined as an acute infection of the

pulmonary parenchyma that is associated with at least some symptoms of acute in-

fection, a new infiltrate on chest x-ray or auscultatory findings such as altered breath

sounds and/or localized rales in community-dwelling patients (Infectious Diseases

Society of America 2000). It is a common condition that carries a high burden of

mortality and morbidity, particularly in elderly populations. Although most patients

recover without sequellae, CAP can take a very severe course, requiring admission

to an intensive care unit (ICU) and even leading to death. According to US data, it is

the most important cause of death from infectious causes and the sixth most impor-

tant cause of death overall (Adams et al. 1996). Even though the mortality from

pneumonia decreased rapidly in the 1940s after the introduction of antibiotic therapy,

it has remained essentially unchanged since then or has even increased slightly

(MMWR 1995).

Furthermore, significant costs are associated with the diagnosis and manage-

ment of CAP. Between 22% and 42% of adults with CAP are admitted to hospital,

and of those, 5% to 10% need to be admitted to an ICU (British Thoracic Society

2001). In the US, it is estimated that the total cost of treating an episode of CAP in

hospital is about USD $ 7500, which is approximately 20 times more than the cost of

treating a patient on an outpatient basis (Lave et al. 1999). CAP also contributes

significantly to antibiotic use, which is associated with well-known problems of resis-

tance.

In treating patients with CAP, the choice of antibiotic is a difficult one. Factors

to be considered are the possible etiologic pathogen, the efficacy of the substance,

potential side-effects, the treatment schedule and its effect on adherence to treat-
 6

ment as well as the particular regional resistance profile of the causative organism

and the co-morbidities that might influence the range of potential pathogens (such as

in cystic fibrosis) or the dosage (as in the case of renal insufficiency). Although many

studies have been published concerning CAP and its treatment, there is no concise

summary of the available evidence and only few guidelines (British Thoracic Society

2001, American Thoracic Society 2001, Canadian Community-Acquired Pneumonia

Working Group 2000, Infectious Diseases Society of America 2000) that can help

clinicians in choosing the most appropriate antibiotic. The applicability of such guide-

lines is furthermore limited by wide variations in regional resistance profiles and by

their focus on hospitalized patients.

1.1 Framework of the review: the Cochrane Collaboration

This review addresses the treatment of community-acquired pneumonia (CAP)

in adolescent and adult outpatients. It was conducted within the framework of the

Cochrane Collaboration, a worldwide network of researchers whose aim is to “pre-

pare, maintain and promote the accessibility of systematic reviews of the effects of

health care interventions” (see www.cochrane.org, www.cochrane.de) using the evi-

dence available from randomized controlled trials (RCT). Reviews are initiated by

researchers with an interest in a particular clinical question. A protocol in which the

goals and methods of the review are described is then written and published in elec-

tronic form in the Cochrane Library (accessible for a fee at www.cochranelibrary.net).

The review is then carried out independently by at least two reviewers who may be

assisted by others, particularly when initially screening study reports for inclusion into

the study. Having reached their own individual conclusions about which studies to
 7

include in the review, the two reviewers compare their results and resolve any differ-

ences by discussion and consensus. One or both of the reviewers then proceeds to

analysing the data of the selected studies and writing the final review. The review

then undergoes peer review within the framework of the Cochrane Collaboration and

is published as a citable peer-reviewed publication in the Cochrane Library of Sys-

tematic Reviews (accessible for a fee at www.cochranelibrary.net).

In order to disseminate the results of reviews more broadly and to circumvent

the problems arising from the fact that the Cochrane Library is only accessible to pay-

ing subscribers, the Lancet has recently made a commitment to publishing Cochrane

Reviews and has encouraged review authors to submit their reviews for publication in

the Lancet (Clark and Horton 2001).

1.2 Statement of authorship

This review was initiated and published as a Cochrane protocol by Prof. Mi-

chael Kochen (MMK) in collaboration with other colleagues from the UK and the

Netherlands, in particular Prof. Theo JM Verheij (TJMV) of Utrecht University (Verheij

2001). MMK co-wrote the protocol and screened abstracts and full articles for inclu-

sion into the review. TJMV co-wrote the protocol, screened abstracts and full articles

for inclusion into the study and decided, in agreement with me, which articles to in-

clude into the review.

As the main reviewer, I carried out a review of the background literature,

screened abstracts and full articles for inclusion into the study, decided, in conjunc-

tion with TJMV, which articles to include, extracted the data from these articles, per-

formed the quantitative analyses and wrote the text, tables and figures of the review.
 8

The present dissertation is a report of the work that I carried out myself and is

not the final text of the Cochrane review, which is less extensive in both scope and

length. In particular, the literature review included in the present dissertation is much

more extensive. Furthermore, the efficacy analyses as well as all the data pertaining

to the open-label studies are unique to this dissertation. Except for discussions with

TJMV concerning the choice of studies to be included into the review, I have carried

out this study independently. As such, I take full responsibility for the contents of this

study as well as for any shortcomings or errors that may remain.

 9

2. Background and literature review

2.1 Community-acquired pneumonia

2.1.1 Incidence

In the industrialized world, the annual incidence of CAP in community-

dwelling adults is estimated at 5 to 11 cases per 1000 adult population (British Tho-

racic Society 2001). The incidence is known to vary markedly with age, being higher

in the very young and the elderly. In one Finnish study, the annual incidence for

people aged 16-59 years was 6 cases per 1000 population, for those 60 years and

older it was 20 per 1000, and for people aged 75 and over, 34 per 1000 (Jokinen et

al. 1993). Annual incidences of 30-50 per 1000 population have been reported for

infants below 1 year of age (Marrie 2001). Seasonal variations in incidence are also

significant, with a peak in the winter months (Marrie 2001).

The annual incidence of CAP requiring hospitalisation has been estimated at

1 to 4 patients per 1000 population (Marrie 1990, Fine et al. 1996). The proportion of

patients requiring hospitalisation varies from country to country and across studies

and has been estimated as ranging anywhere between 15% and 56% (Foy et al.

1973, Minogue et al. 1998). Of those, 5% to 10% required admission to an intensive

care unit (ICU) (British Thoracic Society Research Committee and Public Health

Laboratory Service 1992, Torres et al. 1991). Conversely, about 8% to 10% of ad-

missions to a medical ICU are due to severe CAP (Woodhead et al. 1985).
 10

2.1.2 Etiology

More than 100 microorganisms have been identified so far as potential causa-

tive agents of CAP (Marrie 2001). They can be classified according to their biological

characteristics as either bacteria, mycoplasma and other intracellular organisms, vi-

ruses, fungi and parasites. The most common causative agent of CAP is the bacte-

rium Streptococcus pneumoniae, which is implicated in 20% to 75% of cases of CAP

(Marrie 2001) and about 66% of bacteremic pneumonia (Infectious Diseases Society

of America 2000). Another causative bacterium is Haemophilus influenzae. So-

called “atypical” organisms have also been implicated as causal agents. These in-

clude Chlamydia pneumoniae, Mycoplasma pneumoniae and Legionella pneumo-

phila (Marrie 2001).

Influenza is the most common serious viral pathogen causing airway infections

in adults (Infectious Diseases Society of America 2000). Although it does not itself

cause pneumonia, its most common serious complication is bacterial superinfection,

usually due to Streptococcus pneumoniae. Affected patients are primarily older than

65 years and/or residents of chronic care facilities. Effective prevention is possible

through yearly vaccination of populations at risk and should include vaccination of

those who care for such patients (Potter et al. 1997).

The identification of the causal organism is a challenging task: since lung tis-

sue cannot be routinely obtained, clinicians must rely on sputum samples – which

can only be obtained successfully in about 33% of patients – or on blood cultures,

that are positive in only 6% to 10% of patients with CAP (Canadian Community-

Acquired Pneumonia Working Group 2000). Furthermore, it takes a minimum of 2 to

3 working days to obtain culture results, be it from sputum samples or blood. Conse-

quently, it is often necessary to initiate therapy on empiric grounds alone. Further-

 11

more, routine surveillance of samples sent to microbiology labs by primary care phy-

sicians does not provide an accurate picture of the actual situation in the community,

as samples are often sent only when a first, empirical therapeutic attempt has failed.

2.1.3 Risk factors

A variety of risk factors predisposing a patient to CAP have been identified.

These include host factors, such as chronic obstructive pulmonary disease (COPD),

alcoholism or immune suppression, environmental factors, such as exposure to cer-

tain animals, for example parrots (Chlamydia psittaci), parturient cats and sheep

(Coxiella burnetii), and rabbits (Francisella tularensis), recent hotel stay (Legionella

pneumophila), travel abroad or in endemic regions (Coccidioides immitis, Histo-

plasma capsulatum), and occupational factors, such as contact with body fluids con-

taining infective agents (Mycoplasma tuberculosis) (Canadian Community-Acquired

Pneumonia Working Group 2000). Smoking is also thought to be an important risk

factor for acquiring CAP (Marrie 2001). As outlined above, a number of risk factors

are related to particular causative organisms and enquiring about their presence may

improve diagnostic accuracy with respect to the etiologic agent, however the British

Thoracic Society (2001) cautions that due to the low frequency of some of these or-

ganisms in patients with CAP and the high frequency of the risk factors for exposure

to these organisms in the population, routine questioning about such risk factors may

be misleading.
 12

2.1.4 Diagnosis

The diagnosis of CAP remains a challenge for clinicians. There is no single

finding that is pathognomonic of CAP, and even the gold-standard chest x-ray may

fail to provide the necessary information to make the correct diagnosis. However,

there is good evidence supporting the view that the diagnosis of CAP is inaccurate

without a chest x-ray (British Thoracic Society 2001).

It is important to differentiate between CAP and other lower respiratory tract in-

fections, such as acute bronchitis, and to differentiate between these entities and

other potential causes of similar symptom complexes, such as pulmonary neoplasia,

congestive heart failure or pulmonary embolism, as the subsequent management of

such patients differs greatly. Most cases of upper respiratory tract infections and

acute bronchitis are caused by viruses, and therefore do not require antibiotic treat-

ment (Infectious Diseases Society of America 2000). The use of antibiotics in such

cases is inappropriate and should be avoided.

Despite the importance accorded to history, physical examination, chest x-rays

and some laboratory investigations in the assessment of patients suspected of hav-

ing CAP, only very few studies have attempted to assess the validity of such ap-

proaches (Canadian Community-Acquired Pneumonia Working Group 2000). Fur-

thermore, none of these studies relied on autopsies as a diagnostic gold standard.

Instead, they used chest x-rays or even clinical suspicion to decide whether pneumo-

nia was present, thus making the validity of their conclusions highly questionable.

Another factor that further complicates the diagnosis of CAP are inter-observer

variations in the identification of symptoms and signs in patients suspected of having

CAP. The reliability of physical signs has been studied and found to be highly vari-

able (Spiteri et al. 1988, Schilling et al. 1955). As for symptoms, inter-observer reli-
 13

ability has not been studied, but it is known from studies of other respiratory condi-

tions that there is significant variation between observers (Cochrane et al. 1951,

Fletcher 1964).

Recently, a urine test for rapid detection of Streptococcus pneumoniae has

been approved by the American Food and Drug Administration (FDA) (Henney

1999). The test can be carried out in the physician’s office or in the emergency room,

requires only 5 ml of urine and results are available within 15 minutes. The test is

reported to have a sensitivity of 86% to 90% and a specificity of 71% to 94%. It is

intended as an adjunct to the usual clinical, laboratory and radiological investigations

for suspected CAP. Whether it will become part of the diagnostic armamentarium in

actual practice remains to be seen.

Consequently, the diagnosis of CAP should be made based on a combination

of physical, laboratory, microbiologic and radiographic findings, keeping in mind that

none of them is perfectly reliable for diagnosis (Canadian Community-Acquired

Pneumonia Working Group 2000).

2.1.5 Treatment

Since the majority of cases CAP are caused by organisms amenable to treat-

ment with an antibiotic drug, rapid initiation of antibiotic treatment is indicated in the

vast majority of cases. Difficulties arise when a clinician is confronted with the need

to choose an antibiotic drug for a particular patient. The appropriate choice of antibi-

otic for the ambulatory treatment of CAP in adults and adolescents is the focus of the

present study and I will attempt to provide an answer to this question on the basis of

the currently available evidence.

 14

2.1.5.1 Antibiotic resistance

The problem of antibiotic resistance has received increasing attention in recent

years. The problem is not confined to community-acquired pneumonia, however,

since CAP can take a very severe course even leading to death, it is a condition for

which the issue of antibiotic resistance takes on even greater importance.

Traditionally, the preferred antimicrobial agent for the treatment of Streptococ-

cus pneumoniae was penicillin G (Infectious Diseases Society of America 2000).

However, widespread penicillin use for a variety of infectious conditions has lead to

the emergence and rise of penicillin resistance. Recent studies estimate the propor-

tion of penicillin resistant Streptococcus pneumoniae at around 25% (Marrie 2001).

Similarly, the use of other antibiotics has led to the emergence of resistance against

these agents in a variety of microorganisms. For example, it is estimated that ap-

proximately 30% of Haemophilus influenzae isolates are resistant to amoxicillin (Mar-

rie 2001).

Furthermore, the phenomenon of antibiotic resistance is subject to wide re-

gional and international variations, as access to and patterns of use of antibiotics

vary widely.

As is the case with the identification of pathogens, it is also difficult to estimate

the exact prevalence of antibiotic resistance in a particular area by simple surveil-

lance of specimen sent to microbiology laboratories because these come from pre-

selected patients, some them having already failed a first empirical treatment and

therefore being more likely to carry a resistant pathogen.

For this reason, some practice guidelines emphasize the importance of obtain-

ing baseline microbiologic specimen – the minimum being a Gram stain, with or with-
 15

out culture – before initiation of empiric therapy (Infectious Diseases Society of Amer-

ica 2000).

2.1.6 Prognosis

The prognosis of CAP ranges from full recovery without sequellae to death on

an intensive care unit within a few days of disease onset. Because of this broad and

dramatic spectrum, prognostic factors for the identification of high-risk patients have

been the subject of much research.

The Pneumonia Patient Outcome Research Team (Pneumonia PORT) has

developed a clinical prediction rule to identify patients at risk of short-term mortality

from CAP that is intended as a tool to assist clinicians in making decisions about the

initial location and intensity of treatment (Fine et al. 1997). This prediction rule has

gained wide acceptance and has been included into recent clinical practice guide-

lines (Infectious Diseases Society of America 2000, Canadian Community-Acquired

Pneumonia Working Group 2000).

In ambulatory patients, the mortality rate from pneumonia is low, probably be-

low 1% (British Thoracic Society 2001), but some estimates go as high as 5%

(American Thoracic Society 2001). In hospitalized patients, it hovers around 12%,

increasing to close to 40% (American Thoracic Society 2001) or even 50% (British

Thoracic Society 2001) in patients requiring admission to an ICU.

2.2 Practice guidelines for the treatment of community-acquired

pneumonia

In recent years, there has been an explosion in the number of clinical practice

guidelines being produced and published. The field of infectious diseases is no ex-
 16

ception, and a few guidelines for the diagnosis and treatment of community-acquired

pneumonia have been published over the past decade. Most recently, four major

professional societies have updated the guidelines they had published in the early

1990s. These guidelines are based on a combination of literature review and expert

opinions. As such, they represent an attempt at synthesizing the available evidence

and aim at providing clinicians with diagnosis and treatment strategies that are based

as much as possible on the current state of knowledge. In the following sections,

these guidelines will be examined in further detail in an attempt to get an overall view

of the current state of knowledge with respect to CAP.

These guidelines were identified in the course of searching the literature for

studies and reviews concerned with community-acquired pneumonia. This was done

using the search strategy reported in section 4.2. No language restrictions were ap-

plied. In an effort to broaden the search, the Internet was searched for websites list-

ing guidelines about the treatment of CAP. This search was conducted both in Eng-

lish and in German. No German-language guidelines could be identified using this

search strategy. A few German language review articles dealing with CAP were

identified (Dusch and Täuber 2001, Gillissen and Ewig 2000 a, Gillissen and Ewig

2000 b, Rosseau and Suttorp 2000, Ruef 2001), however these were all “secondary

literature” article summarizing evidence from other studies and guidelines in a non-

systematic way and intended for a general medical readership.

2.2.1 American Thoracic Society (2001)

In 2001, the American Thoracic Society published an update of its original

1993 statement on community-acquired pneumonia (American Thoracic Society

2001). These guidelines were developed by a committee composed of pulmonary,

 17

critical care, infectious disease and general internal medicine specialists. Ambulatory

care physicians, general practitioners in particular, appear to have been left out. This

raises concerns that the ambulatory care perspective may have been neglected.

The guidelines development strategy is described in detail, however there is

no detailed account of how the literature was searched.

The American Thoracic Society claims that its guidelines are evidence-based

and reports using a classification system based on the system used by the Canadian

Infectious Diseases Society and Canadian Thoracic Society in their CAP guidelines

update (Canadian Community-Acquired Pneumonia Working Group 2000), however,

they do not state the level of evidence for each of their therapeutic recommendations,

nor do they give any specific references supporting those recommendations. Finally,

the committee reports that they focused on “studies that included an extensive diag-

nostic approach to define the etiologic pathogen” and that “most [studies] involved

hospitalized patients” (American Thoracic Society 2001, p. 1733). This raises con-

cerns that the evidence-base on which the recommendations for outpatients were

made may have been insufficient.

The new statement includes a summary of the available literature as well as

“evidence-based recommendations for patient management” (American Thoracic So-

ciety 2001, p. 1730). The guidelines recommend that all patients with suspected

CAP should have a chest radiograph to confirm the diagnosis, yet they recognize that

this may not be feasible in some ambulatory settings. Sputum Gram stain and cul-

ture are recommended only if drug-resistant bacteria or an organism not covered by

the usual empiric therapy are suspected.

As for therapy and management, the Society advocates an empiric approach

based on likely pathogens. Patients are to be classified into one of four groups de-
 18

pending on factors thought to influence the spectrum of potential pathogens, namely:

1) the place of therapy (outpatient, inpatient regular ward, inpatient ICU), 2) the pres-

ence of cardiopulmonary disease (COPD, heart failure), and 3) the presence of modi-

fying factors, which include risk factors for drug-resistant Streptococcus pneumoniae

(DRSP), enteric gram-negatives (nursing home residence) and Pseudomonas

aeruginosa. Using these factors, the guidelines define four patient groups: 1) outpa-

tients with no history of cardiopulmonary disease and no modifying factors, 2) outpa-

tients with cardiopulmonary disease and/or other modifying factor, 3) inpatients not

admitted to the ICU; this group is further subdivided into those with and without car-

diopulmonary disease and/or other modifying factors, and 4) ICU-admitted patients,

who are further subdivided into those with or without risk factors for Pseudomonas

aeruginosa.

For each group, the available evidence was reportedly combined to identify

the most likely pathogens, and recommendations for empiric therapy were made on

this basis. For group 1 (outpatients without additional risk factors), the recommended

therapy is an advanced generation macrolide, such as azithromycin or clarithromycin,

or doxycycline. The advanced generation macrolides were recommended on the

grounds that erythromycin does not cover Haemophilus influenzae and is not toler-

ated as well. In group two (outpatients with cardiopulmonary disease and/or other

modifying factors), a combination of a beta-lactam with either one of the above-

mentioned macrolides or doxycycline is recommended. The beta-lactams mentioned

include cefpodoxine, high-dose amoxicillin and amoxi/clavulanate. The combination

treatment is advocated because amoxicillin does not offer adequate coverage for H.

influenzae. Furthermore, it is recommended that all patients, regardless of what

group they belong to, should be treated for “atypical” organisms (Chlamydia pneu-
 19

moniae, Mycoplasma pneumoniae, Legionella species). This is usually done by in-

cluding a macrolide antibiotic in the recommended treatment plan.

2.2.2 British Thoracic Society (2001)

The British Thoracic Society also recently updated its 1993 guidelines for the

treatment of CAP in adults admitted to hospital to include patients treated in an am-

bulatory setting (British Thoracic Society 2001). The British Thoracic Society guide-

lines committee was composed of 12 members, of which 6 were general practitio-

ners, four of them with a special interest in respiratory medicine and an “active re-

search interest” in respiratory infectious diseases. The other members of the

committee were a clinical microbiologist, two infectious disease specialists, a

registrar in respiratory medicine, a clinical epidemiologist and a medical librarian.

The search and study selection strategy employed is described in details, and a level

of evidence is explicitly given for every recommendation made by the committee.

The guidelines do not advocate the routine use of chest radiographs or spu-

tum culture for the majority of patients with CAP who are managed on an outpatient

basis. The diagnosis of CAP is to be made on clinical grounds, and severity as-

sessment is emphasized as the key to appropriate management, whether the pa-

tients are to be treated in the community or in hospital. The choice of antibiotic

treatment for outpatients is empiric and the main target organism remains S. pneu-

moniae.

The authors emphasize the fact that their literature search for the period 1981-

99 yielded only 16 articles judged relevant to the antibiotic treatment of CAP and that

few of these studies were conducted within a setting comparable to those of UK prac-

tices. Nonetheless, and despite explicitly acknowledging that the currently available
 20

evidence forms an “unsatisfactory basis” for making solid evidence-based recom-

mendations, the British Thoracic Society continues to recommend amoxicillin as the

preferred agent on the grounds of cost, current practice, “wide experience”, safety

and drug tolerance, but recommends a higher dose (500 mg to 1000 mg po tid) than

used commonly in practice. The fact that clinical treatment failures have rarely been

documented when penicillin-resistant strains are treated with higher doses of amox-

icillin and that penicillin resistant pneumococci are still relatively rare in the UK is

given as the rationale for recommending higher doses of amoxicillin. Erythromycin

(500 mg po qid) is recommended as the alternative treatment for patients who do not

tolerate amoxicillin. Clarithromycin (500 mg po bid) is suggested as the alternative

agent for the sub-group of these patients who do not tolerate erythromycin, usually

due to gastrointestinal side-effects.

Interestingly, the guidelines committee considered tetracyclines (doxycycline)

as an agent of first choice because resistance rates for pneumococci are lower that

for penicillins or erythromycin and it is also active against “atypical” agents, however

they refrained from making it a first choice recommendation in their guidelines due to

a presumed reluctance of physicians to change their current practice that would “limit

compliance with recommendations”. This is an interesting example of how perceived

inertia on the part of practitioners (whether real or only imagined by the guidelines

committee) can significantly influence the content of practice recommendations

(Keeley 2002).
 21

2.2.3 Canadian Infectious Diseases Society / Canadian Thoracic Society

(2000)

In 2000, the Canadian Infectious Diseases Society and the Canadian Thoracic

Society updated their 1993 guidelines for the treatment of CAP (Canadian Commu-

nity-Acquired Pneumonia Working Group 2000). Members of the guidelines commit-

tee are listed at the end of the report, however there is no mention of the members’

area of specialty, so it is unclear whether physicians primarily involved in the care of

ambulatory patients were involved in the guidelines formulation process.

The literature search strategy is described in reasonable detail and a hierar-

chical evaluation of the strength of evidence was carried out. Accordingly, a level of

evidence is explicitly given for each recommendation made by the committee, unfor-

tunately these are included only in the text of the guidelines and not in the tables

where the recommendations are also summarized.

The committee bases its recommendations on a classification of patients ac-

cording to the place of treatment (outpatient, inpatient, nursing home). The guide-

lines also provide a scoring system that uses objective criteria to assist physicians in

deciding whether a patient should be hospitalized or not.

With respect to chest radiography, the committee points out that a number of

infectious and non-infectious conditions may present a radiographic picture that is

indistinguishable from that of pneumonia and that only one small study has assessed

the ability of chest radiography to detect pulmonary infiltrates in patients suspected of

having CAP (the gold standard used was high resolution CT scanning). They also

point out that expert opinions are divided concerning the necessity of performing rou-

tine chest x-rays in patients suspect of having CAP. Nonetheless, the committee

recommends that chest x-rays be performed routinely “under most circumstances” in

 22

such patients because the diagnosis of pneumonia is strengthened (although not

confirmed) by the presence of an infiltrate.

As for microbiological studies, no specific investigations are recommended for

the majority of patients treated on an outpatient basis.

For outpatients without modifying risk factors, the treatment of choice is a

macrolide (erythromycin, azithromycin or clarithromycin), the second choice treat-

ment being doxycycline. Outpatients with modifying factors are further subdivided

into three groups: those with chronic obstructive pulmonary disease (COPD) who did

not receive antibiotics or steroids within the past 3 months, COPD patients who did

get antibiotics or steroids within the past three months, and patients in whom

macroaspiration is suspected (alcoholics, patients with impaired consciousness, im-

paired gag reflex or other deglutitional dysfunction).

In the first group (COPD, no antibiotics or steroids in past 3 months), the first

choice is a so-called “newer” macrolide, namely azithromycin or clarithromycin, the

second choice being doxycycline. In patients with COPD who received an antibiotic

or steroids in the past three months, a “respiratory” quinolone (levofloxacin, gatiflox-

acin or moxifloxacin) is recommended, the second line choice being amoxicillin-

clavulanate plus a macrolide, or alternatively a second-generation cephalosporin plus

a macrolide. In cases of suspected macroaspiration, the first choice recommendation

is amoxicillin-clavulanate plus a macrolide, the second choice being levofloxacin plus

either clindamycin or metronidazole.

2.2.4 Infectious Diseases Society of America (2000)

In 2000, the Infectious Diseases Society of America (IDSA) updated their 1998

guidelines for the treatment of CAP in adults (Infectious Diseases Society of America
 23

2000). Members of the guidelines committee are listed as co-authors of the report

together with their affiliated institution, however there is no mention of the members’

area of specialty, so it is unclear whether primary care physicians were involved in

the guidelines formulation process.

The literature search strategy is not described, however, the committee used a

grading system to assess the quality of the evidence provided by the research stud-

ies that they reviewed, as well as another grading system to classify the strength of

the recommendations they made. The grades for quality of evidence and strength of

recommendation are explicitly stated with each recommendation presented in the

guidelines.

The IDSA guidelines emphasize the clinical importance of the decision to hos-

pitalize a patient or to treat on an outpatient basis. They recommend the use of the

clinical prediction rule for short-term mortality developed and validated by the Pneu-

monia Patient Outcome Research Team (Pneumonia PORT) (Fine et al. 1997) as a

basis for deciding whether or not to hospitalize a patient.

The IDSA guidelines state that the diagnosis of CAP is based on a combina-

tion of clinical and laboratory data, adding that a chest x-ray is usually necessary to

establish the diagnosis. The guidelines recommend that posteroanterior and lateral

chest radiography be part of the routine workup of patient in whom CAP is consid-

ered a likely diagnosis and they discourage the initiation of empiric therapy without

radiographic confirmation, although they acknowledge that obtaining chest x-rays

“may not always be practical” (Infectious Diseases Society of America 2000, p. 370).

For outpatients, sputum collection for Gram stain and culture are deemed op-

tional, however the IDSA panel makes a strong case in favour of establishing an etio-
 24

logic diagnosis for all patients. For outpatients, the guidelines state that it is desirable

to perform at least a Gram stain, with or without culture.

Treatment recommendations emphasize a pathogen-directed antimicrobial

therapy and prompt antimicrobial treatment. Treatment recommendations are made

based on suspected pathogens. Recommendations for empiric antibiotic selection in

the absence of an etiologic diagnosis, i.e. when Gram stain and culture are not diag-

nostic, are also made. Drugs of first choice are recommended in “no particular order”

and include doxycycline, a macrolide (erythromycin, clarithromycin or azithromycin)

or a fluoroquinolone (levofloxacin, moxifloxacin or gatifloxacin). For older patients or

patients with co-morbidities, a fluoroquinolone is to be preferred. When S. pneumo-

niae or H. influenzae are the suspected etiologic agents, amoxicillin-clavulanate or

some second-generation cephalosporins (cefuroxime, cefpodoxime and cefprozil) are

considered appropriate alternatives.

2.3 Current best evidence in the treatment of community-acquired

pneumonia

The guidelines reviewed above differ in many respects: in the composition of

the guidelines committees, in the extent of reporting about the literature review and

guidelines formulation process, in the use of classification systems to assess evi-

dence and rank recommendations, and in the content of the recommendations, be it

in terms of diagnosis or treatment. Table 1 summarizes the main features of the

guidelines.

A few common points also emerge from the above guideline review process:

firstly, the importance of assessing the severity of disease and of the resulting deci-

sion to hospitalize was emphasized in all guidelines. Furthermore, all guidelines ac-
 25

knowledge – with varying degrees of openness - that there is little evidence on which

to base treatment recommendations, particularly concerning ambulatory patients.

However, in the end, in most cases it remains unclear on what basis the specific

treatment recommendations were made and exactly what evidence was used to sub-

stantiate these recommendations.

Consequently, the current project aims at identifying, evaluating and summa-

rizing the evidence available from RCT with respect to the treatment of CAP in adult

and adolescent outpatients.

Table 1: Characteristics and recommendations of clinical practice guidelines for the treatment of CAP
Guideline Guideline formulation process Routine diagnostic Treatment of outpatients
studies recommended
Primary care Search strat- Level of evidence Chest Gram Sputum Without modifying With modifying
physician on egy reported stated for each x-ray stain culture risk factors risk factors
guidelines in detail recommendation
committee
American Tho- no no no yes no no macrolide (azi, clari) beta-lactam plus
racic Society or doxycycline macrolide (azi, cla-
(2001) ri) or doxycycline
British Tho- yes yes yes no no no high-dose amoxicillin no separate rec-
racic Society (500-1000mg po tid) ommendation
(2001) 2nd choice: ery or
clari
Canadian ? yes yes yes no no macrolide (ery, azi, newer macrolide
Community- clari) 2nd choice: (azi, clari) or "respi-
Acquired doxycycline ratory" quinolone
Pneumonia (see section 2.2.3
Working for details) 2nd
Group (2000) choice: doxycycline

Infectious Dis- ? no yes yes yes optional "in no particular or- fluoroquinolones
eases Society der": macrolide (ery, (levo, moxi, gati)
of America azi, clari), doxycy-
(2000) cline, fluoroquinolone
(levo, moxi, gati)

Abbreviations: azi=azithromycin, clari=clarithromycin, ery=erythromycin, gati=gatifloxacin, levo=levofloxacin, moxi=moxifloxacin

 27

3. Objectives

The objectives of this review are:

1. To assess and compare the efficacy of individual antimicrobial therapies with re-

spect to clinical, radiological and bacteriological outcomes in adult outpatients with

CAP;

2. To assess and compare the efficacy of drugs across drug groups;

3. To make evidence-based practice recommendations if possible.

4. To assess the effect of drug schedule on adherence to treatment by carrying out

an effectiveness analysis with open-label studies using different drug administration

regimen.
 28

4. Methods
Randomized controlled trials (RCT) are considered the gold standard in clini-

cal research when it comes to establishing the efficacy of a treatment. This favoured

status is attributable to the fact that RCT most closely mirror a scientific experiment in

the classical sense and are least susceptible to bias. When properly conducted, they

can lead to clear-cut conclusions about the relative efficacy of two treatments in a

defined group of patients. Consequently, we chose to focus our review exclusively

on RCT.

The following criteria were applied in selecting studies for inclusion into this

review. The rationale underlying the decision to use each criterion is detailed in the

following sections.

4.1 Study selection criteria

4.1.1 Types of studies

All randomized trials of antibiotics in adolescent and adult outpatients with

CAP reporting on clinical parameters, cure rates or mortality were considered for in-

clusion.

4.1.2 Types of participants

Trials that included outpatients of either gender over 12 years of age in which

pre-defined criteria for CAP were met as defined by the British Thoracic Society

(2001) were included in this review.

 29

4.1.3 Clinical Signs and Symptoms

CAP in outpatients is defined as follows by the British Thoracic Society (2001):

· Clinical definition, in the absence of a chest X-ray:

- Symptoms of an acute lower respiratory tract (LRT) infection (cough and at

least one other LRT symptom);

- New focal chest signs on physical examination;

- at least one systemic feature (either a symptom complex of sweating, fevers,

shivers, aches and pains and/or temperature >38.0°C).

- No other explanation for the illness, which is treated as pneumonia with an-

tibiotics.

· Definition when a chest X-ray is available:

- Symptoms and signs consistent with an acute lower respiratory tract infec-

tion associated with new radiographic shadowing for which there is no other

explanation (e.g. not pulmonary oedema or infarction)

- The illness is the primary clinical problem and is managed as pneumonia

4.1.4 Types of interventions

All double-blind randomized controlled comparisons of at least two antibiotics

(or one antibiotic and a placebo) used to treat community-acquired pneumonia were

included. Trials comparing two doses of the same drug were not included.
 30

The perceived efficacy of antibiotics in pneumonia means that most of the

available research deals with comparisons of two antibiotics. Comparisons involving

intravenous drugs are usually carried out in a hospital setting. However, as this might

occasionally be performed in an ambulatory setting, we did not a priori exclude stud-

ies dealing with intravenous drug applications.

Trials allowing concurrent use of other medications such as antitussives, anti-

pyretics, bronchodilators, or mucolytics were included if they allowed equal access to

such medications for patients in both arms of the trial.

4.1.5 Types of outcome measures

When available, the following outcomes measures were documented in each

of the selected studies:

1. Clinical response: improvement of signs and symptoms. Where possible, dura-

tion of clinical signs and symptoms were used as outcome measures. We used a

clinical definition of cure as the primary outcome since radiographic resolution lags

behind clinical improvement (Macfarlane et al. 1984).

2. Radiologic response: resolution or improvement of a new finding on chest x-ray

after antibiotic therapy

3. Bacteriologic response: negative sputum culture in patients previously found to

have causative pathogens in their sputum.

4. Frequency of hospitalization.

5. Mortality.
 31

4.2 Search strategy for identification of studies

The Cochrane Acute Respiratory Infections Group's trial register, The Coch-

rane Library, EMBASE and MEDLINE (1966-December 31st 2001) were searched

using the following terms: COMMUNITY-ACQUIRED INFECTION, PNEUMONIA,

RESPIRATORY TRACT INFECTION, ANTIBIOTICS. Studies were also identified by

checking the bibliographies of studies and review articles retrieved, and if necessary

by contacting the first or corresponding authors of included studies. This search

strategy yielded a total of 1417 references, some of which were double entries, due

to the overlapping content of databases.

4.3 Selection process

Titles and abstracts of the identified citations were screened to exclude trials

that clearly did not meet the inclusion criteria of the review. If it was felt that a trial

might possibly meet the inclusion criteria, the full paper was obtained for further

study. The most common reason for exclusion was that studies were conducted ex-

clusively in hospitalized patients. Articles having passed this initial screen (34) were

then reviewed independently by two reviewers (myself (LMB) and TJMV) to deter-

mine whether they met the inclusion criteria of the review. The selection process is

shown graphically in Fig. 1.

 32
Figure 1: Study selection process

Potential Studies (n=1417)

Figure 1: Exclude
Studyonselection process
the basis of abstract Exclude n= 1383

RCT? Exclude n=2 (Ball 1994, Lode & 1998)

No
Yes
Yes
Same drug, different dose? Exclude n=2 (Balgos & 1999, Rayman & 1996)

No
Study Population Inpatients only Exclude n=1 (Lacny 1972)

Mix of in- and

Out- Results re- Exclude n=9 (Balmes &
outpatients
patients ported No 1991, Dark 1991, Kiani &
only Yes separately? 1990, Laurent 1996, Liipo
(retain outpatients only) & 1994, Lode & 1995,
NAPSG 1997, Örtqvist &
1996, Schleupner & 1988)
Diagnoses
(CAP vs Others) Results re-
Mix of diagnoses ported No
separately? Exclude n=5 (Bantz &
1987, DeCock and
CAP only
Poels 1988, Hoepel-
man & 1993, Hoepel-
Yes man & 1998, Neu and
(retain CAP only) Chick 1993)

Diagnosis confirmed by No
chest-X-ray in all patients? Exclude n=1 (Biermann & 1988)

Yes

Bacterial sample required for Yes Exclude n=3 (Chodosh 1991, Kammer
inclusion into study? and Ress 1991, Müller and Wettich 1992)

No
No
Total study size >30? Exclude n=3 (Fong & 1995, Gris 1996,
Tilyard and Dovey 1992)
Yes

Patients and investigators No Exclude n=5 (Dautzenberg & 1992, Hi-

blinded? guera & 1996, O’Doherty & 1998, Peu-
geot & 1991, Salvarezza & 1998)
Yes but retain for effectiveness (open-
label studies) analysis
Selected studies for efficacy
(primary) analysis: (n=3)
Anderson & 1991
Chien & 1993
Ramirez & 1999 & = et al.
 33

Studies could be excluded for any one of the following reasons: if they were

not truly randomized, if they only compared two doses of the same substance, if the

results were not reported separately for inpatients and outpatients, or if the indication

for treatment consisted of a mix of diagnoses (most commonly: acute bronchitis, ex-

acerbation of chronic bronchitis, and pneumonia) and the results were not reported

separately for each diagnostic group. Studies including only bacteriologically evalu-

able patients were excluded, because these patients are not necessarily typical of

the spectrum encountered in primary practice. In order to avoid that patients with

bronchitis or other non-pneumonia lower respiratory tract infections be included, we

excluded studies that did not confirm the diagnosis of CAP by chest x-ray. Finally,

studies were excluded if the total number of patients was less than 30, because be-

low this limit, the estimate of a binomial parameter (in this case, the proportion of pa-

tients cured or improved) becomes unstable (Armitage and Berry 1994). Further-

more, when randomized controlled trials are too small, one can no longer safely as-

sume that all potential confounders (both documented and undocumented) have

been controlled for by being distributed equally between the two treatment groups

(Rothman and Greenland 1998). For the primary analysis, we included only blinded

RCT conducted in outpatients with CAP. Trials in which allocation to treatment or

control group was not concealed (open-label studies) that otherwise fulfilled all other

inclusion and exclusion criteria were retained for a sub-group analysis concerning the

effect of drug schedule on compliance (effectiveness analysis) (see Fig.2 and Sec-

tion 5.2.2).
 Cefixime Clarithromycin
Anderson & 1991
Chien & 1993
Salvarezza & 1998
Ramirez & 1999

Erythromycin
Roxithromycin
Sparfloxacin
O’Doherty & 1998

Dautzenberg & 1992

Peugeot & 1991

Amoxicillin- Azithromycin Ofloxacin

clavulanate

Higuera & 1996

Cefuroxime
: Randomized controlled trials

Figure 2: Antibiotic comparisons : Open-label studies

: Beta-lactams
Figure 2: Antibiotic comparisons : Quinolones / Gyrase inhibitors

: Macrolides

& : et al.
 35

4.4 Data extraction

The following data were extracted from each study whenever possible:

- description of participants (outpatients over 12 years)

- description of pathogens identified and their anti-microbial resistance profiles

- description of intervention

- description of control therapy

- total number of participants in each arm of the trial

- study setting

- mean duration of symptoms in each arm of the trial

- clinical, radiographic and bacteriologic cure rates in each arm of the trial

- proportion of patients admitted to hospital in each arm of the trial

- mortality rates in each arm of the trial

- study sponsor

The studies were assessed independently by 2 reviewers (LMB and TJMV)

and disagreements were resolved by discussion and consensus. There were no ir-

reconcilable disagreements. Reviewers were not blinded to the identity and affiliation

of the study authors.

4.5 Analyses

For dichotomous outcome data, an estimate of the relative risk with approxi-

mate 95% confidence intervals was calculated. This was done using the Cochrane

Collaboration's Review Manager software, version 4.1.

 36

5. Results

5.1 Description of studies

The characteristics of the studies included in the primary analysis as well as in

the effectiveness analysis are shown in Table 2. All studies except one (Peugeot et

al. 1991) acknowledged the sponsorship of a corporate sponsor. Two of the open-

label studies (Dautzenberg et al. 1992, Peugeot et al. 1991) included multiple diag-

noses but provided separate data for CAP patients, so it was possible to include

these studies into the review. Remarkably, four of the five open-label studies used

different administration schedules for the drugs being tested within each study – for

example, twice daily compared to four times daily. The effect of these differences on

compliance will be examined (see Section 5.2.2).

Table 2: Drug administration schedules and other study characteristics
Study Drug Regimen Sponsor Multiple
Dose Frequency Duration Diagnoses

Double-blind studies
Anderson & 1991 Clarithromycin 250 mg bid 14 d
Erythromycin 500 mg qid 14 d Abbott no
Chien & 1993 Clarithromycin 250 mg bid 7-14 d
Erythromycin 500 mg qid 7-14 d Abbott no
Ramirez & 1999 Clarithromycin 250 mg bid 10 d
Sparfloxacin 200 mg* qd 10 d Rhône-Poulenc Rorer no

Open-label studies
Dautzenberg & Amoxi-Clav 500 mg + 125 mg tid 14 d
1992 Roussel yes
Roxithromycin 150 mg bid 14 d
Higuera & 1996 § Amoxi-Clav 500 mg + 125 mg tid 10 d
Cefuroxime 500 mg bid 10 d Glaxo Wellcome no
O’Doherty & 1998 Azithromycin 500 mg qd 3d
Clarithromycin 250 mg bid 10 d Pfizer no
Peugeot & 1991 Erythromycin 400 mg qid 10 d
Ofloxacin 400 mg bid 10 d none declared yes
Salvarezza & 1998 Cefixime 400 mg qd 8-10 d
Roxithromycin 300 mg qd 8-10 d Hoechst Marion Roussel no
* except Day 1: 400 mg loading dose
§ Investigators blinded to administration schedules, patients not blinded & = et al.

Abbreviations:
qd = once daily bid = twice daily tid = three times a day qid = four times a day
 38

5.1.1 Number of trials and trial size

Three randomized controlled trials involving a total of 622 patients aged 12

years and older diagnosed with community-acquired pneumonia were included in the

primary analysis (Anderson et al. 1991, Chien et al. 1993, Ramirez et al. 1999). Five

open-label randomized, but unblinded trials including a total of 405 patients and

meeting all other inclusion/exclusion criteria were retained for sub-group analyses of

effectiveness and compliance (Dautzenberg et al. 1992, Higuera et al. 1996,

O’Doherty et al. 1998, Peugeot et al. 1991, Salvarezza et al. 1998). The trials in the

primary analysis included varying numbers of patients, the largest having 342 pa-

tients (Ramirez et al. 1999), the smallest 107 (Anderson et al. 1991). The median

trial size in the primary analysis was 173 patients, the mean size 207; in the effec-

tiveness analysis, the median size was 60 and the mean size 81.

5.1.2 Diagnoses

All three trials in the primary analysis exclusively enrolled patients with com-

munity-acquired pneumonia. Two of the open-label trials (Dautzenberg et al. 1992,

Peugeot et al. 1991) also included patients with other diagnoses, usually acute bron-

chitis or acute exacerbation of chronic bronchitis, but reported results separately for

each diagnostic group, so that it was possible to extract data separately for pneumo-

nia patients.
 39

5.1.3 Diagnostic criteria

In all trials, the diagnosis of community-acquired pneumonia was based on

clinical signs and symptoms as well as confirmation by radiographic findings in all

patients. The signs and symptoms used as diagnostic criteria included combinations

of the following: fever, chills, recent onset of productive cough, pleuritic chest pain,

dyspnoea, pyrexia, tachypnoea, dullness to percussion, egophony, rales, localized

reduced breath sounds and bronchial breath sounds.

5.1.4 Out- vs Inpatients

In all trials, patients were treated exclusively as outpatients.

5.1.5 Patient inclusion and exclusion criteria

Two trials (Anderson et al. 1991, Ramirez et al. 1999) in the primary analysis

included only adult patients, one (Chien et al. 1993) also included adolescents (12

years of age and older). In the open-label trials, two trials reported including patients

12 years of age and older (Higuera et al. 1996, O’Doherty et al. 1998) and one trial

(Dautzenberg et al. 1992), patients as old as 90. Only one of the studies used older

age (>75 years) as an exclusion criterion (O’Doherty et al. 1998). Overall, the trials

excluded patients with conditions that could have affected the treatment or interfered

with follow-up. Exclusion criteria were reported in more or less detail in the various

study reports. The most common criteria reported were: pregnancy and lactation,

women not using adequate contraception (usually oral contraceptives or a barrier

method), history of allergic reaction to the study drugs, recent treatment with or con-
 40

comitant use of an antimicrobial agent, concurrent medication with ergotamine, cyc-

losporin, antacids (except H2-antagonists) or digitalis, conditions affecting GI absorp-

tion, severe renal or hepatic impairment, terminal illness or conditions precluding

study completion, infectious mononucleosis, HIV/AIDS, and prior participation in the

study.

5.1.6 Antibiotics

The trials varied with respect to the antibiotics studied (see Fig. 2). Two trials

in the primary analysis (Anderson et al. 1991, Chien et al. 1993) studied the same

antibiotic pair (clarithromycin and erythromycin). The other trial (Ramirez et al. 1999)

studied a different antibiotic pair, namely clarithromycin and sparfloxacin. Antibiotic

pairs studied in open-label trials were: Dautzenberg et al. 1992: amoxiclav vs

roxithromycin; Higuera et al. 1996: amoxiclav vs cefuroxime; O'Doherty et al. 1998:

azithromycin vs clarithromycin; Peugeot et al. 1991: erythromycin vs ofloxacin; Salva-

rezza et al. 1998: cefixime vs roxithromycin (see Fig. 2).

5.1.7 Methodological quality of included studies

All three trials included in the primary analysis were randomized, double-blind

studies comparing two antibiotics. The extent of reporting was variable between

studies. None of the studies clearly stated the randomization method used. None of

the articles reported any test of effectiveness of the blinding procedures used. Com-

pliance with treatment was assessed by pill count in two studies (Anderson et al.

1991, Chien et al. 1993); neither reported any difference in the number of pills re-
 41

maining between the two groups, however in the Chien et al. (1993) study, forty pa-

tients were excluded because they received "less than the minimum therapy" (7

days) and these patients were distributed unevenly across the two groups (10 in the

clarithromycin group and 30 in the erythromycin group). The third study (Ramirez et

al. 1999) reports having assessed patient compliance but does not state how. Re-

garding co-interventions with other medications, most studies excluded patients

whose co-medication included certain drugs such as other antibiotics, chemothera-

peutics or antiretrovirals. Only one study (Chien et al. 1993) reported how many pa-

tients were excluded because of forbidden co-medication. Withdrawals were re-

ported by all studies with varying degree of detail as to the reasons for withdrawal.

The number of patients lost to follow-up was reported in all three studies. Losses to

follow-up appeared to be minor, amounting to a maximum of 10% of the initially ran-

domized patients.

5.2 Study results

5.2.1 Efficacy analysis (“primary analysis”)

The success rates for each of the treatment arms of the three trials are shown

in Table 3. "Success" was defined as cure or improvement, be it clinical, bacterio-

logical or radiological, as assessed at a predefined follow-up visit that took place be-

tween 7 and 14 days after initiation of therapy. None of the clinical, bacterial or radio-

logical success rates differed significantly among treatment arms within each of the

studies, nor did they achieve clinical significance when the results of the two studies

comparing clarythromycin with erythromycin (Anderson et al. 1991 and Chien et al.

1993) were pooled together.

 42

To assess this, we pooled the outcomes of the two studies (Anderson et al.

1991 and Chien et al. 1993) and calculated relative “risks” (RR) of success, be it

clinical, bacteriological or radiological. These analyses were performed using the

Cochrane collaboration’s Review Manager Software Version 4.1, and illustrated

graphically using the MetaView software, a subset of the Review Manager package

(see Fig. 3, 4 and 5).

 43

Table 3: Clinical, bacteriological and radiological cure rates

Study Drug Clinical success Bacteriological Radiological suc-

success cess
n cured / n cured / n cured /
% N total % N isolated % N x-rayed

Double-blind studies

Anderson & 1991 Clarithromycin 98% 63/64 89% 8/9 90% 55/61
Erythromycin 91% 39/43 100% 5/5 90% 38/42
Chien & 1993 Clarithromycin 97% 89/92 88% 23/26 96% 88/92
Erythromycin 96% 78/81 100% 17/17 96% 78/81
Ramirez & 1999 Clarithromycin 83% 145/175 91% 74/81 Not reported sepa-
Sparfloxacin 80% 133/167 97% 64/66 rately

Open-label studies

Dautzenberg & Amoxi-Clav 63% 10/16 No specimen taken Not reported sepa-
1992 Roxithromycin 93% 14/15 rately
Higuera & 1996 Amoxi-Clav 100% 55/55 93% 37/40 Not reported sepa-
Cefuroxime 96% 49/51 94% 32/34 rately
O’Doherty & 1998 Azithromycin 94% 83/88 97% 31/32 Not reported sepa-
Clarithromycin 95% 84/88 91% 32/35 rately
Peugeot & 1991 Erythromycin 100% 13/13 Not reported sepa- 92% 12/13*
Ofloxacin 100% 19/19 rately 94% 18/19*
Salvarezza & Cefixime 94% 28/30 95% 20/21 97% 28/29
1998 Roxithromycin 100% 30/30 100% 19/19 90% 27/30
* one patient per group not followed up by x-ray

Definitions of success:
clinical success = cure or improvement
radiological success = resolution or improvement
bacteriological success = eradication of a previously identified pathological strain

& = et al.
Figure 3: Clinical success (pooled results)

ycin

Figure 4: Bacteriological success (pooled results)

ycin
 45

Figure 5: Radiological success (pooled results)

ycin
 46

It can be seen in figures 3 to 5 that all RR are close to the null value of 1, and

all confidence intervals include or abut on 1, thereby indicating no significant differ-

ence in the respective success rates of treatment with clarythromycin and erythromy-

cin. In the case of bacteriological success, there appears to be a tendency favouring

erythromycin, however this does not achieve statistical significance.

5.2.1.1 Comparisons across antibiotic groups

The only comparison across antibiotic groups is provided by the study by Ra-

mirez et al. (1999), whereby a macrolide (clarithromycin) and a quinolone (sparflox-

acin) were compared. Again, there was no significant difference in clinical or bacte-

riological success (see Table 3); radiological outcomes were not reported separately

for the two treatment arms.

5.2.1.2 Side-effects

There were, however, significant differences in the occurrence of side-effects

attributed to the study drug in the two studies comparing clarithromycin with erythro-

mycin (Anderson et al. 1991 and Chien et al. 1993) (Table 4). In both cases, there

were significantly more side-effects in the erythromycin group, the majority being gas-

trointestinal side-effects. This was not, however, reflected in the rate of side-effects

leading to withdrawal from the study, which was not significantly different across

treatment arms.
Table 4: Reasons for exclusion from efficacy analyses and drug-related side-effects
Study Drug Reasons for exclusion from efficacy
Regimen analyses Adverse events
Less than
minimum ther- Leading to with-
apy Loss to follow-up Total (drug-related) drawals
Fre- N / total N N / total N N / total N N / total N
Dose quency Duration excluded % excluded % enrolled % enrolled %

Double-blind studies
Anderson & Clarithromycin 250 mg bid 14 d 4/32 13% 0/32 0% 15/96* 16% 4/96 4%
1991 Erythromycin 500 mg qid 14 d 16/68 24% 1/68 1% 37/112* 33% 21/112 19%
Chien & Clarithromycin 250 mg bid 7-14 d 20/41 49% 6/41 15% 34/133* 26% 6/133 5%
1993 Erythromycin 500 mg qid 7-14 d 43/54 80% 5/54 9% 76/135* 56% 37/135 27%
Ramirez & Clarithromycin 250 mg bid 10 d 10/48 $ 21% 7/48 15% 47/175 27% 10/175 6%
1999 Sparfloxacin 200 mg* qd 10 d 10/42 $ 24% 6/42 14% 42/167 25% 14/167 8%

Open-label studies
Dautzenberg Amoxi-Clav 500 mg
& 1992 ! + 125 mg tid 14 d Not reported 13/477 3% (both 67/242* 28% 21/242 9%
Roxithromycin 150 mg bid 14 d Not reported groups combined) 21/235* 9% 3/235 1%
Higuera & Amoxi-Clav 500 mg tid 10 d 56#/162 35% Not reported 6/78 8% Not reported
1996 § + 125 mg (both groups
Cefuroxime 500 mg bid 10 d combined) Not reported 3/84 4% Not reported
O’Doherty & Azithromycin 500 mg qd 3d 1/101 1% Not reported 14/101 14% 0/101 0%
1998 Clarithromycin 250 mg bid 10 d 4/102 4% Not reported 13/102 13% 2/102 2%
Peugeot & Erythromycin 400 mg qid 10 d Not reported Not reported 4/28 14% 0/28 0%
1991 ! Ofloxacin 400 mg bid 10 d Not reported Not reported 8/28 29% 2/28 7%
Salvarezza Cefixime 400 mg qd 8-10 d Not reported Not reported 0/30 0% 0/30 0%
& 1998 Roxithromycin 300 mg qd 8-10 d Not reported Not reported 0/30 0% 0/30 0%
 48

Footnotes:

* = significant difference between treatment groups of a given study

! = results are for all patients in study (all diagnoses combined), results for pneumonia were not reported separately
§ = investigators blinded, patients not blinded
$ = includes "protocol deviations" (not further specified)
# = "clinically unevaluable" (less than 3 days of therapy, resistant pathogen in pre-treatment culture, other protocol deviation)

& = et al.
 49

5.2.1.3 Bacterial pathogens

Various bacterial pathogens were identified with varying frequency across

studies (see Table 5). The proportion of samples yielding an identifiable pathogen

ranged from 19% (Anderson et al. 1991) to 43% (Ramirez et al. 1999). Anderson et

al. (1991) reported on a majority of cases being positive for Haemophilus influenzae

(62% of positive cultures) with S. pneumoniae (18%) being second most common,

whereas Chien et al. (1993) predominantly identified Streptococcus pneumoniae as

the causative organism in 56% of cultures, with H. influenzae taking second place at

40%. On the contrary, Haemophilus parainfluenzae (31%) was the most commonly

identified pathogen in the study by Ramirez et al. (1999), with H. influenzae (23%) in

second place.
Table 5: Eradication rates and proportional importance of bacterial pathogens

Bacterial isolates: Pathogens as % of all isolates

Study n isolated/Total N Gram-negative aerobes Gram-positive aerobes
sampled H. influenzae H. parainfluenzae H. parahaemolyticus M. catarrhalis S. Pneumoniae Staph. Aureus

Double-blind studies
Anderson &1991 39/208 19% 24/39 62% 5/39 13% ****** ****** 3/39 8% 7/39 18% ****** ******
Chien & 1993 43/173 25% 17/43 40% 2/43 5% ****** ****** ****** ****** 24/43 56% ****** ******
Ramirez & 1999 147/342 43% 34/147 23% 45/147 31% 8/147 5% ****** ****** 18/147 12% 5/147 3%

Open-label studies
Dautzenberg & 1992 ***** Not tested ******
Higuera & 1996 § 97/162 60% 17/97 18% ******* ******* ******* 37/97 38% *******
O’Doherty & 1998 66/203 33% 34/67 51% ******* ******* 9/67 13% 6/67 9% 2/67 3%
Peugeot & 1991 6/32 19% 3/6 50% ******* ******* ******* 3/6 50% *******
Salvarezza & 1998 37/58 64% 3/37 8% ******* ******* 3/37 8% 26/37 70% *******

§ = investigators blinded, patients not blinded

& = et al.
 51

5.2.1.4 Serologically identified pathogens

Only two of the three studies carried out serologic tests to identify putative

pathogens (Anderson et al. 1991 and Chien et al. 1993; see Table 6). In both these

studies, the most frequently identified pathogen was Mycoplasma pneumoniae, which

represented 69% (Anderson et al. 1991) and 74% (Chien et al. 1993) of positive se-

rology results. Chlamydia pneumoniae accounted for the remainder with 38%

(Anderson et al. 1991) and 26% (Chien et al. 1993) respectively. There were no

samples positive for Legionella pneumoniae or for Chlamydia psittaci in either study.

Table 6: Proportional importance of serologically identified pathogens

Study Serology: positive serologic tests
n positive / Mycoplasma Chlamydia Legionella Chlamydia
total N tested pneumoniae pneumoniae pneumoniae psittaci

Double-blind studies
Anderson & 1991 16/208° 8% 11/16 69% 6/16 38% ******* *******
Chien & 1993 27/173 16% 20/27 74% 7/27 26% ******* *******
Ramirez & 1999 ***** Not tested ******

Open-label studies
Dautzenberg & 1992 ***** Not tested ******
Higuera & 1996 ***** Not tested ******
O’Doherty & 1998 7/203 3% 4/7 57% 1/7 14% 2/7 29% *******
Peugeot & 1991 2/32 6% ******* 2/2 100% ******* *******
Salvarezza & 1998 10/60 17% 5/10 50% ******* ******* 5/10 50%

° = one patient was positive for both M. pneumoniae and C. pneumoniae

& = et al.
 52

5.2.2 Effectiveness analysis (open-label studies)

All of the trials in the effectiveness analysis were randomized open-label or

single-blind studies comparing two antibiotics. In one of the studies (Higuera et al.

1996), the investigator were blinded to the treatment insofar as the drugs were dis-

pensed by a pharmacist or study coordinator and the patients were instructed not to

discuss the frequency of dosing with the investigator. All studies except one (Salva-

rezza et al. 1998) compared different dosage regimen (see Table 2), whereby one

drug was given at a higher frequency (twice vs four times daily; Peugeot et al. 1991);

in some cases, one of the drugs was also given over a longer time period (ten vs

three days; O’Doherty et al. 1998). Only two of the studies (Salvarezza et al. 1998,

Higuera et al. 1996) clearly stated the randomization method used. Compliance with

treatment was assessed by pill count as well as urine testing in one of the studies

(Higuera et al. 1996). Most studies excluded patients who received co-medication

with certain drugs, such as other antibiotics, chemotherapeutics or anti-retrovirals.

Only one study (O’Doherty et al. 1998) reported how many patients were excluded

because of forbidden co-medication. One other study (Higuera et al. 1996) reported

that patients were excluded on these grounds but did not report how many patients

were thereby excluded. None of the studies reported losses to follow-up; therefore, it

is not possible to quantify the extent of the problem. In most studies, withdrawals

were reported along with the reasons for withdrawal.

5.2.2.1 Effect of administration schedule on adherence to treatment

It would be reasonable to assume that a higher frequency and longer duration

of drug administration would impose a burden on patients that could result in de-
 53

creased compliance. In order to test this hypothesis, I extracted data on losses to

follow-up and patients receiving “less than minimum therapy” for all open-label stud-

ies (see Table 4). Unfortunately, four studies (Higuera et al. 1996, O’Doherty et al.

1998, Peugeot et al. 1991, Salvarezza et al. 1998) did not report on losses to follow-

up and the only study that did (Dautzenberg et al. 1992) pooled the results of both

treatment groups together, making it impossible to determine whether there was any

differential loss to follow-up across treatment groups. Likewise, three studies

(Dautzenberg et al. 1992, Peugeot et al. 1991, Salvarezza et al. 1998) did not report

how many patients received “less than minimum therapy”. One of the studies that did

(Higuera et al. 1996) reported combined results for both treatment groups. Finally,

O’Doherty et al. (1998) reported results separately for both treatment groups, and

these were not significantly different from each other.

Therefore, due to inadequate reporting, it was not possible to draw any con-

clusions about the effect of administration schedules on compliance.

 54

6. Discussion

6.1 Evidence

The overwhelming feature of this systematic review is the utter paucity of rele-

vant studies that could be identified and included in the review. Given this current

state of affairs, it is not possible to make solid evidence-based recommendations for

the treatment of community-acquired pneumonia in ambulatory outpatients. One im-

portant reason for this lack of evidence is that a large number of the trials originally

identified were conducted in hospitalized patients and were therefore excluded be-

cause they were not directly relevant to the treatment of ambulatory patients.

Another remarkable feature of the collected studies is the incompleteness of

the reporting. This further complicated matters by making it difficult or impossible to

compare and combine the available results in a meaningful way. Finally, the hetero-

geneity of the antibiotic pairs studied precludes pooling study results quantitatively.

6.2 Methods

It could be argued that the inclusion/exclusion criteria for this review were too

strict and that this is the reason why so few studies were retained. I do, however,

believe that the criteria I applied were necessary in order to validly address the ques-

tion of the efficacy of treatment of CAP in ambulatory patients. In particular, it could

be argued that the decision to exclude studies based on size is not desirable, since

one aim of the review is to pool results and that each study therefore would contrib-

ute some information. I felt, however, that this criterion was necessary to exclude

studies where the number of patients with pneumonia was so small that randomiza-
 55

tion could no longer be expected to achieve a balanced distribution of confounders,

both known and unknown, across study groups. Finally, arguing retrospectively, it

can be seen that dropping size as an exclusion criteria would not have made much

difference to the results obtained, as the three excluded studies would have contrib-

uted a total of 23 (Fong et al. 1995), 6 (Gris 1996) and 8 (Tilyard and Dovey 1992)

patients respectively, thereby only increasing the size of the primary analysis group

by 5.9 % (from 622 to 659). Furthermore, the study by Fong et al. (1995) being an

open-label study, it would only have been included in the effectiveness analysis.

As for the requirement that the diagnosis of CAP be confirmed by a chest ra-

diograph, I felt that this was necessary to avoid diagnostic misclassification, which

could have led to inclusion of patients with bronchitis into the review.

6.3 Implications for practice

Currently available evidence from RCT is insufficient to make truly evidence-

based recommendations for the treatment of community-acquired pneumonia. Indi-

vidual study results do not reveal significant differences in efficacy between various

antibiotics and antibiotic groups. Until better evidence becomes available, practitio-

ners should favour shorter course therapies with lower drug costs whenever possible

as a means of enhancing compliance and reducing expenses, while still taking into

account the regional resistance profiles prevalent in their area of practice as well as

the characteristics of each patient.

National guidelines may provide additional decision-making support, but clini-

cians should keep in mind that current guideline recommendations for the treatment

of CAP in ambulatory patients are based on very scanty evidence.

 56

6.4 Implications for research

Multi-drug, multi-drug-group double-blind comparisons using similar admini-

stration schedules and carried out in the ambulatory setting are needed to provide

the evidence necessary for practice recommendations if these are to be applicable to

outpatient treatment. Study conditions should ensure that diagnosis and manage-

ment of patients with community-acquired pneumonia are as similar as possible to

real practice, while still ensuring that the study question is addressed in a valid way.

Whether a study of this extent and character would find the necessary funding is,

however, doubtful.
 57

7. Summary

BACKGROUND

Community-acquired pneumonia (CAP) is a common condition representing a

significant disease burden for the community, particularly for the elderly. It also con-

tributes significantly to health care expenditures and antibiotic use, which is associ-

ated with well-known problems of resistance. Although many studies have been pub-

lished concerning CAP and its treatment, there is no concise summary of the avail-

able evidence and only few guidelines that can help clinicians in choosing the most

appropriate antibiotic

OBJECTIVES

The goal of this study was to summarize the evidence currently available from

randomized controlled trials (RCT) concerning the efficacy of antiobiotic treatment of

CAP in ambulatory adolescent and adult patients and to formulate evidence-based

practice recommendations.

SEARCH STRATEGY

The Cochrane Acute Respiratory Infections Group's trial register, The Coch-

rane Library, MEDLINE and EMBASE were searched using the following terms:

COMMUNITY-ACQUIRED INFECTION, PNEUMONIA, RESPIRATORY TRACT IN-

FECTION, and ANTIBIOTICS up to and including December 31st 2001. Studies were

also identified by checking the bibliographies of the articles retrieved.

 58

SELECTION CRITERIA

All randomized controlled trials in which one or more antibiotics were tested for

the treatment of CAP in ambulatory adolescent or adult patients were considered for

inclusion. Studies testing one or more antibiotics and reporting the diagnostic criteria

used in selecting patients as well as the clinical outcomes achieved were included.

No language restrictions were applied.

DATA COLLECTION & ANALYSIS

Study reports were assessed and data extracted using predefined criteria. Au-

thors of studies were contacted as needed to resolve any ambiguities in the study

reports. The data were analyzed using the Cochrane Collaboration's Review Man-

ager 4.1 Software.

MAIN RESULTS

Three randomized controlled trials involving a total of 622 patients aged 12

years and older diagnosed with community-acquired pneumonia were included. The

quality of the studies and of the reporting was variable. A variety of clinical, radio-

logical and bacteriological diagnostic criteria and outcomes were reported. Overall

there was no significant difference in the efficacy or effectiveness of the various anti-

biotics under study. However, the available evidence is too incomplete to form a

solid basis for making evidence-based recommendations.

Five open-label studies were retained in order to carry out an effectiveness

analysis to assess the impact of different drug regimen on patient compliance. Unfor-
 59

tunately, the reporting of losses to follow-up was so incomplete that it precluded

drawing any conclusions about this question.

CONCLUSIONS

Currently available evidence from randomized controlled trials is insufficient to

make evidence-based recommendations for the treatment of community-acquired

pneumonia. Pooling of study data was limited by the very low number of studies, by

their heterogeneity and by incomplete reporting of study results. Individual study re-

sults do not reveal significant differences in efficacy between various antibiotics and

antibiotic groups. Multi-drug comparisons using similar administration schedules are

needed to provide the evidence necessary for practice recommendations. Until bet-

ter evidence becomes available, practitioners should favour shorter course therapies

with lower drug costs whenever possible as a means of enhancing compliance and

reducing expenses, while still taking into account the regional resistance profiles

prevalent in their area of practice as well as the clinical course of each patient.
 60

8. Zusammenfassung (Summary in German)

EINLEITUNG

Die ambulant erworbene Pneumonie (AEP) ist eine häufige, oft schwere Er-

krankung von großer Bedeutung für die Bevölkerung, die insbesondere für ältere Pa-

tienten kompliziert verlaufen kann. Die übliche Behandlung der AEP besteht in der

Gabe von Antibiotika. Dies trägt u.a. zur Zunahme von Antibiotikaresistenzen bei.

Obwohl bereits viele Studien zur Behandlung der AEP veröffentlicht wurden, gibt es

keine aktuelle Zusammenfassung der vorhandenen Evidenz und nur einzelne Leitli-

nien, die den behandelnden Arzt bei der Wahl eines Antibiotikums unterstützen.

ZIEL

Ziel dieser Studie war es, die vorhandene Evidenz randomisierter, kontrollier-

ter Studien (RCT) bzgl. der AEP Behandlung bei ambulanten Patienten zusammen-

zufassen und ggf. eine evidenz-basierte Leitlinie/Therapieempfehlung zu formulieren.

METHODEN

Eine umfassende elektronische und manuelle Literatursuche wurde durchge-

führt mit dem Ziel, alle RCT zu identifizieren, in denen zwei oder mehr Antibiotika für

die Behandlung der AEP geprüft wurden. Die Cochrane Library, MEDLINE und EM-

BASE (bis einschließlich 31.12.2001) wurden mit folgenden Suchbegriffen durch-

sucht: COMMUNITY-ACQUIRED INFECTION, PNEUMONIA, RESPIRATORY

TRACT INFECTION und ANTIBIOTICS. Die Literaturverzeichnisse der ausgewählten

Studien wurden auf weitere relevante Artikel durchsucht. Alle RCT, die ein oder meh-

rere Antibiotika prüften und deren diagnostische Kriterien und Outcomes dargestellt
 61

waren, wurden eingeschlossen. Es wurden keine Spracheinschränkungen verwen-

det. Die jeweiligen Artikel wurden begutachtet und deren Inhalt mit der Software „Re-

view Manager“ (Version 4.1) in standardisierter Form extrahiert.

ERGEBNISSE

Es wurden drei Studien gefunden, die die Einschlusskriterien erfüllen. Die

Studien umfassten insgesamt 612 Patienten mit einer ambulant erworbenen Pneu-

monie. Verschiedene klinische, bakteriologische und radiologische Endpunkte wur-

den berichtet. Die vorhandene Evidenz wurde zusammengefasst und beurteilt. Ins-

gesamt gab es keinen wesentlichen Unterschied zwischen den verschiedenen Prä-

paraten bezogen auf die Wirksamkeit oder damit verbundene unerwünschte Arznei-

mittelwirkungen. Jedoch ist die vorhandene Evidenz zu unvollständig, um eine ein-

heitliche evidenz-basierte Leitlinie oder Therapieempfehlungen formulieren zu kön-

nen.

Fünf randomisierte Studien ohne Verblindung wurden in einer Effektivitätsana-

lyse beurteilt, um den Einfluss verschiedener Therapiepläne auf die Patientencompli-

ance zu evaluieren. Leider wurden Therapieabbrüche so unvollständig berichtet,

dass es unmöglich war, Schlussfolgerungen aus diesen Daten zu ziehen.

SCHLUSSFOLGERUNGEN

Die zur Zeit vorhandene Evidenz von randomisierten kontrollierten Studien

ambulanter Patienten reicht nicht aus, um evidenz-basierte Therapieempfehlungen

zu formulieren. Die Zusammenfassung der Studienergebnisse wurde durch die nied-

rige Studienzahl, die Heterogenität der Studien und die Unvollständigkeit der Studi-

enberichte stark eingeschränkt. Einzelne Studienergebnisse zeigen keinen signifikan-

 62

ten Unterschied in der Wirksamkeit der verschiedenen Antibiotika. Randomisierte

kontrollierte Blindstudien, in denen mehrere Antibiotika aus verschiedenen Gruppen

miteinander verglichen werden, sind erforderlich, um evidenz-basierte Therapieemp-

fehlungen formulieren zu können. Bis solche Evidenz zur Verfügung steht, sollten

behandelnde Ärzte kurze Therapien mit geringen Kosten bevorzugen. Dabei sollten

die Charakteristika der einzelnen Patienten und, soweit vorhanden, regionale Resi-

stenzprofile berücksichtigt werden.

 63

9. Appendices

9.1 List of abbreviations

CAP Community-acquired pneumonia

COPD Chronic obstructive pulmonary disease

DRSP Drug-resistant Streptococcus pneumoniae

ICU Intensive care unit

LRT Lower respiratory tract

RCT Randomized controlled trial

RR Relative risk

qd once daily

bid twice daily

tid three times a day

qid four times a day

po by mouth

iv intravenous

MMK Prof. Michael M. Kochen

TJMV Prof. Theo V.M. Verheij

LMB Lise M. Bjerre

 64

9.2 List of figures

Figure 1: Study selection process.............................................................................32

Figure 2: Antibiotic comparisons ..............................................................................35
Figure 3: Clinical success (pooled results) ...............................................................44
Figure 4: Bacteriological success (pooled results)....................................................44
Figure 5: Radiological success (pooled results) ......................................................45
 65

9.3 List of tables

Table 1: Characteristics and recommendations of clinical practice guidelines for the

treatment of CAP .......................................................................................26
Table 2: Drug administration schedules and other study characteristics .................37
Table 3: Clinical, bacteriological and radiological cure rates ...................................43
Table 4: Reasons for exclusion from efficacy analyses and drug-related side-effects.
.................................................................................................................47
Table 5: Eradication rates and proportional importance of bacterial pathogens ......50
Table 6: Proportional importance of serologically identified pathogens ...................51
 66

10. References
Adams P, Hendershot G, Marano M (1996): Current estimates from the National
Health Interview Survey 1996. Vital Health Stat 10, 200

American Thoracic Society (2001): Guidelines for the management of adults with
community-acquired pneumonia. Am J Respir Crit Care Med 163, 1730-54

Anderson G, Esmonde TS, Coles S, Macklin J, Carnegie C (1991): A comparative

safety and efficacy study of clarithromycin and erythromycin stearate in community-
acquired pneumonia. J Antimicrob Chemother 27 Suppl. A, 117-124

Armitage P, Berry G: Statistical methods in medical research, 3rd edition; Blackwell

Scientific Publications, Oxford 1994

Balgos AA, Rodriguez-Gomez G, Nasnas R, Mahasur AA, Margono BP, Tinoco-

Favila JC, Sansores-Martinez RH, Hassan M, Beppo O, Wongsa A, Cukier A, Vargas
F (1999): Efficacy of twice-daily amoxycillin/clavulanate in lower respiratory tract in-
fections. Int J Clin Pract 53, 325-330

Ball P (1994): Efficacy and safety of cefprozil versus other beta-lactam antibiotics in
the treatment of lower respiratory tract infections. Eur J Clin Microbiol Infect Dis 13,
851-6

Balmes P, Clerc G, Dupont B, Labram C, Pariente P, Poirier R (1991): Comparative

study of azithromycin and amoxicillin/clavulanic acid in the treatment of lower respira-
tory tract infections. Eur J Clin Microbiol Infect Dis 10, 437-9

Bantz PM, Grote J, Peters-Haertel W, Stahmann J, Timm J, Kasten R, Bruck H

(1987): Low-dose ciprofloxacin in respiratory tract infections: a randomized compari-
son with doxycycline in general practice. Am J Med 82 Suppl 4A, 208-10

Biermann C, Loken A, Riise R (1988): Comparison of spiramycin and doxycycline in

the treatment of lower respiratory infections in general practice. J Antimicrob Chemo-
ther 22 Suppl. B, 155-8

British Thoracic Society Research Committee and Public Health Laboratory Service
(1992): The aetiology, management and outcome of severe community-acquired
pneumonia on the intensive care unit. Respir Med 86, 7-13

British Thoracic Society (2001): British Thoracic Society Guidelines for the manage-
ment of adults with community-acquired pneumonia. Thorax 56 Suppl. IV, iv1-iv64

Canadian Community-Acquired Pneumonia Working Group (2000): Canadian guide-

lines for the initial management of community-acquired pneumonia: an evidence-
based update by the Canadian Infectious Diseases Society and the Canadian Tho-
racic Society. Clin Infect Dis 31, 383-421
 67

Chien SM, Pichotta P, Siepman N, Chan CK and the Canada-Sweden Clarithromy-

cin-Pneumonia Group (1993): Treatment of Community-Acquired Pneumonia: A mul-
ticenter, double-blind, randomized study comparing clarithromycin with erythromycin.
Chest 103, 697-701

Chodosh S (1991): Temafloxacin compared with ciprofloxacin in mild to moderate

lower respiratory tract infections in ambulatory patients. Chest 100, 1497-1502

Clark M, Horton R (2001): Bringing it all together: Lancet-Cochrane collaborate on

systematic reviews. Cochrane News 22, 5

Cochrane AL, Chapman PF, Oldham PD (1951): Observers’ errors in taking medical
histories. Lancet 1951, 1, 1007-8

Dark D (1991): Multicenter evaluation of azithromycin and cefaclor in acute lower

respiratory tract infections. Am J Med 91 Suppl 3A, 31S-35S

Dautzenberg B, Scheimberg A, Brambilla C, Camus P, Godard P, Guerin JC, Le-

marie E, Rezvani Y, Rosenbaum M, Tuchais E, Wallaert B (1992): Comparison of two
oral antibiotics, roxithromycin and amoxicillin plus clavulanic acid, in lower respiratory
tract infections. Diagn Microbiol Infect Dis 15, 85S-89S

De Cock L, Poels R (1988): Comparison of spiramycin with erythromycin for lower

respiratory tract infections. J Antimicrob Chemother 22 Suppl B, 159-63

Dusch H, Täuber MG (2001): Ambulant erworbene Pneumonie – Erreger und Abklä-

rung. Ther Umsch 58, 575-81

Fine MJ, Smith MA, Carson CA (1996): Prognosis and outcomes of patients with
community-acquired pneumonia. A meta-analysis. JAMA 275,134-41

Fine MJ, Auble TE, Yealy DM (1997): A prediction rule to identify low-risk patients
with community-acquired pneumonia. NEJM 336, 243-50

Fletcher CM (1964): The problem of observer variation in medical diagnosis with

special reference to chest diseases. Method Inform Med 3, 98-103

Fong IW, Laforge J, Dubois J, Small D, Grossman R, Zakhari R and the Canadian
Bronchitis Study Group (1995): Clarithromycin versus cefaclor in lower respiratory
tract infections. Clin Invest Med 18, 131-8

Foy HM, Cooney MK, McMahan R (1973): Viral and mycoplasmal pneumonia in a
prepaid medical care group during an eight-year period. Am J Epidemiol 97, 93-102

Gillissen A, Ewig S (2000 a): Diagnostik der ambulant erworbenen Pneumonie. Med
Klin 95, 101-2

Gillissen A, Ewig S (2000 b): Therapie der ambulant erworbenen Pneumonie. Med
Klin 95, 169-70
 68

Gris P (1996): Once-daily, 3-day azithromycin versus a three-times-daily, 10-day

course of co-amoxiclav in the treatment of adults with lower respriatory tract infec-
tions: results of a randomized, double-blind comparative study. J Antimicrob Chemo-
ther 37 Suppl. C, 93-101

Henney JE (1999): Quick test for pneumonia. JAMA 282, 1218

Higuera F, Hidalgo H, Feris J, Giguere G, Collins JJ (1996): Comparison of oral cefu-

roxime axetil and oral amoxycillin/clavulanate in the treatment of community-acquired
pneumonia. J Antimicrob Chemother 37, 555-64

Hoepelman IM, Sips AP, van Helmond JLM, van Barneveld PWC, Neve AJ, Zwinkels
M, Rozenberg-Arska M, Verhoef J (1993): A single-blind comparison of three-day
azithromycin and ten-day co-amoxiclav treatment of acute lower respiratory tract in-
fections. J Antimicrob Chemother 31 Suppl. E, 147-52

Hoepelman IM, Möllers MJ, van Schie MH, Greefhorst APM, Schlösser NJJ, Sin-
ninghe Damsté EJ, van de Mossdijk CNF, Dalinghaus WH, Eland ME, Mol SJM,
Rozenberg-Arska M (1998): 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 Antimi-
crob Agents 9, 141-6

Infectious Diseases Society of America (2000): Practice guidelines for the manage-
ment of community-acquired pneumonia in adults. Clin Infect Dis 31, 347-82

Jokinen C, Heiskanen L, Juvonen H (1993): Incidence of community-acquired pneu-

monia in the population of four municipalities in eastern Finland. Am J Epidemiol 137,
977-88

Kammer RB, Ress R (1991): Randomized comparative study of ceftibuten versus

cefaclor in the treatment of acute lower respiratory tract infections. Diagn Microbiol
Infect Dis 14, 101-5

Keeley D (2002): Guidelines for managing community-acquired pneumonia in adults:

Many recommendations remain based on the lowest grade of evidence. BMJ 324,
436-7

Kiani R, Coulson L, Johnson D, Hammershaimb L (1990): Comparison of once-daily

and twice-daily cefixime regimens with amoxicillin in the treatment of acute lower res-
piratory tract infections. Curr Ther Res 48, 841-52

Lacny J (1972): A comparison of oral therapy with clindamycin and penicillin G in

common gram-positive infections. Curr Ther Res 14, 26-30

Laurent K (1996): Efficacy, safety and tolerability of azithromycin versus roxithromy-

cin in the treatment of acute lower respiratory tract infections. J Antimicrob Chemo-
ther 37 Suppl. C, 115-24
 69

Lave JR, Lin CC, Fine MJ (1999): The cost of treating patients with community-
acquired pneumonia. Semin Respir Crit Care Med 20, 189-98

Liippo K, Tala E, Puolijoki H, Brückner OJ, Rodrig J, Smits JPH (1994): A compara-
tive study of dirithromycin and erythromycin in bacterial pneumonia. J Infect 28, 131-
9

Lode H, Garau J, Grassi C, Hosie J, Huchon G, Legakis N, Segev S, Wijnands G

(1995): Treatment of community-acquired pneumonia: a randomized comparison of
sparfloxacin, amoxycillin-clavulanic acid and erythromycin. Eur Respir J 8, 1999-2007

Lode H, Aubier M, Portier H, Örtqvist A and the Sparfloxacin Study Group (1998):
Sparfloxacin as alternative treatment to standard therapy for community-acquired
bacteremic pneumococcal pneumonia. Clin Microbiol Infect 4, 135-43

Macfarlane JT, Miller AC, Smith WHR et al. (1984): Comparative radiographic fea-
tures of community-acquired legionnaires' disease, pneumococcal pneumonia, my-
coplasma pneumonia and psittacosis. Thorax 39, 28-33

Marrie TJ (1990): Epidemiology of community-acquired pneumonia in the elderly.

Semin Respir Infect 5, 260-8

Marrie TJ: Community-acquired pneumonia; in: Clinical evidence, Vol.5; BMJ

Publishing Group, London 2001, 1040-49

Minogue MF, Coley CM, Fine MJ (1998): Patients hospitalized after initial outpatient
treatment for community-acquired pneumonia. Ann Emerg Med 31, 376-80

MMWR (1995): Pneumonia and influenza death rates: United States, 1979-1994.
Morb Mortal Wkly Rep 44, 535-537

Müller O, Wettich K (1992): Comparison of loracarbef (LY 163892) versus amoxicillin

in the treatment of bronchopneumonia and lobar pneumonia. Infection 20, 176-82

NAPSG = The Nordic Atypical Pneumonia Study Group (1997): Atypical pneumonia
in the Nordic countries: aetiology and clinical results of a trial comparing fleroxacin
and doxycyclin. J Antimicrob Chemother 39, 499-508

Neu HC, Chick TW (1993): Efficacy and safety of clarithromycin compared to ce-
fixime as outpatient treatment of lower respiratory tract infections. Chest 104, 1393-9

O'Doherty B, Muller O and the Azithromycin Study Group (1998): Randomized, multi-
centre study of the efficacy and tolerance of Azithromycin versus clarithromycin in the
treatment of adults with mild to moderate community-acquired pneumonia. Eur J Clin
Microbiol Infect Dis 17, 828-33

Örtqvist A, Valtonen M, Cars O, Wahl M, Saikku P, Jean C and the Scandinavian

Sparfloxacin Study Group (1996): Oral empiric treatment of community-acquired
pneumonia: a multicenter, double-blind, randomized study comparing sparfloxacin
with roxithromycin. Chest 110, 1499-1506
 70

Peugeot RL, Lipsky BA, Hooton TM, Pecoraro RE (1991): Treatment of lower respira-
tory infections in outpatients with ofloxacin compared with erythromycin. Drugs Exptl
Clin Res 17, 253-7

Potter J, Stott DJ, Roberts MA (1997): Influenza vaccination of health care workers
in long-term care hospitals reduces the mortality of elderly patients. J Infect Dis 175,
1-6

Ramirez J, Unowsky J, Talbot GH, Zhang H, Townsend L (1999): Sparfloxacin vs

clarithromycin in the treatment of community-acquired pneumonia. Clin Therapeutics
21, 103-117

Rayman J, Krchnavý M, Balciar P, Duchon J, Fortuník J, Faith L, Manduchova H,

Drgona L, Koren P, Kraszko I, Krcméry V Jr and the Slovak Multipractice Study
Group (1996): Ofloxacin once daily versus twice daily in community-acquired pneu-
monia and acute exacerbation of chronic bronchitis. Chemotherapy 42, 227-30

Rosseau S, Suttorp N (2000): Ambulant erworbene Infektionen der Atemwege und

des Lungenparenchyms. Internist 41, 1158-69

Rothman KJ, Greenland S: Modern epidemiology. 2nd edition; Lippincott-Raven,

Philadelphia, PA 1998

Ruef C (2001): Ambulant erworbene Pneumonie – Risikoenschätzungen, empirische

und gezielte Therapie. Ther Umsch 58, 582-7

Salvarezza CR, Mingrone H, Fachinelli H, Kijanczuk S (1998): Comparison of

roxithromycin with cefixime in the treatment of adults with community-acquired
pneumonia. J Antimicrob Chemother 41 Suppl. B, 75-80

Schilling RSF, Hughes JPW, Dingwall-Fordyce I (1955): Disagreement between ob-

servers in an epidemiological study of respiratory disease. BMJ 230, 55-68

Schleupner CJ, Anthony WC, Tan J, File TM, Lifland P, Craig W, Vogelman B (1988):
Blinded comparison of cefuroxime to cefaclor for lower respiratory tract infections.
Arch Intern Med 148, 343-8

Spiteri MA, Cook DG, Clarke SW (1988): Reliability of eliciting physical signs in ex-
amination of the chest. Lancet 1(8590), 873-5

Tilyard MW, Dovey SM (1992): A randomized double-blind controlled trial of

roxithromycin and cefaclor in the treatment of acute lower respiratory tract infections
in general practice. Diagn Microbiol Infect Dis 15, 97S-101S

Torres A, Serra-Batlles J, Ferrer A (1991): Severe community-acquired pneumonia.

Epidemiology and prognostic factors. Am Rev Respir Dis 144, 312-8

Verheij TJM, Kochen MM, Hoepelman IM, Lammers JW, Macfarlane J, Woodhead M
(2001): Antibiotics for community-acquired pneumonia in adult outpatients. The
Cochrane Library 4, 1-7
 71

Woodhead MA, Macfarlane JT, Rodgers FG (1985): Aetiology and outcome of severe
community-acquired pneumonia. J Infect 10, 204-10

www.cochrane.de Accessed 4.6.2002

www.cochranelibrary.net Accessed 4.6.2002

www.cochrane.org Accessed 4.6.2002

 72

Acknowledgements

I wish to express my gratitude to my supervisor and chairman, Prof. Dr. med.

M. M. Kochen, for suggesting the topic of this dissertation and for supporting me in

many ways throughout the completion of this project.

My gratitude also extends to Priv. Doz. Dr. Wolfgang Himmel for guiding me

with good humour through the intricacies of German academia and helping me sur-

mount the various hurdles of the thesis submission process as smoothly as possible.

My thanks to Prof. Theo JM Verheij of the Julius Center for General Practice

and Patient-Oriented Research, University Medical Center Utrecht, Netherlands, who

was the other Cochrane reviewer for this study, for helpful and thought-provoking

discussions.

Many thanks to Dr. Frederike Behn (Hamburg) for precious help with parts of

the data extraction process.

Finally, I also wish to extend my thanks to all my colleagues in the Department

of Family Medicine here at Göttingen University for stimulating discussions, helpful

suggestions, good company and strong support throughout.

 73

Curriculum Vitae

I, Lise Marie Bjerre, was born on May 23rd, 1969 in Montreal, Quebec, Canada,

the first-born child of Poul Bjerre, MSEng, MBA, electrical engineer, and of Dr. Carmen

Couillard Bjerre, BA, MSW, PhD, social worker.

From 1975 to 1981, I attended grade school in my home-town of Brossard, a

suburb of Montreal, and, from 1981 to 1986, Jean-de-la-Mennais high school in the

nearby town of LaPrairie, from which I graduated in June of 1986 with a Diploma of

Secondary Studies (Diplôme d’études secondaires; DÉS). I attended Champlain Col-

lege (St-Lambert Campus) from 1986 to 1989, obtaining a Diploma of Collegial Stud-

ies in Pure and Applied Sciences (Diplôme d’études collégiales; DÉC).

In 1989, I was admitted to Concordia University (Montreal), where I obtained a

Bachelor of Science with Honours in Biology (BSc Hon.). In addition, I completed the

Science College Program, a multidisciplinary program for gifted science students.

After graduating from Concordia University in 1993, I pursued a Master of Science in

Epidemiology (MSc) at McGill University (Montreal) under the supervision of Prof. Olli

S. Miettinen, which I obtained in 1995. From 1995 to 1999, I attended McGill Univer-

sity Medical school (Montreal), from which I graduated with the degree of Doctor of

Medicine, Master of Surgery (MDCM) in May of 1999.

In the fall of 1999, I got married and moved to my husband’s native Germany.

Since the 15th of November 1999, I have been working as a research fellow and fam-

ily medicine resident in the Department of Family Medicine of Göttingen University. I

completed my training as a resident intern (“Arzt im Praktikum”), obtaining the Ger-

 74

man certification as a physician (“Approbation”) on April 20th 2001. I continued my

clinical training as a resident until the 31st of December 2001. Since then I have

taken a one-and-a-half year leave of absence from clinical work in order to complete

a research project (PhD thesis, Department of Epidemiology and Biostatistics, McGill

University) in collaboration with Prof. Jacques LeLorier (Université de Montréal) and

Prof. Michal Abrahamowicz (McGill University), for which we successfully obtained

funding. I also continue to work on a part-time basis in the Department of Family

Medicine. I plan on completing my residency training in Family Medicine starting in

July 2003, after finishing the PhD research project.

Göttingen, November 2002

 75

Lebenslauf
Ich, Lise Marie Bjerre, wurde am 23. Mai 1969 als erstes Kind des Diplomingenieurs

Poul Bjerre und seiner Ehefrau, der Sozialarbeiterin Dr. Carmen Couillard Bjerre, in

Montreal (Kanada) geboren.

Von 1975 bis 1981 besuchte ich die Grundschule in meiner Heimatstadt Brossard,

einem Vorort von Montreal. Nach dem Wechsel auf die Jean-de-la-Mennais-

Hochschule in der Nachbarstadt LaPrairie konnte ich 1986 die Abiturprüfung ablegen

(Diplôme d’études secondaires; DÉS). Von 1986 bis 1989 lernte ich am Champlain

College (St Lambert Campus), an dem ich das Kollegdiplom in Natur- und ange-

wandten Wissenschaften erlangte (Diplôme d’études collégiales en sciences pures et

appliquées; DÉC).

Von 1989 bis zum erfolgreichen Abschluss des „Bachelor of Science with Honours in

Biology“ (BSc Hon.) im Jahr 1993 studierte ich an der Universität Concordia (Montre-

al). Parallel dazu habe ich den Studiengang des “Science College Program”, einem

multidisziplinären Programm für begabte Studenten der Naturwissenschaften, erfolg-

reich absolviert. Anschließend absolvierte ich von 1993 bis 1995 ein Aufbaustudium

im Bereich Epidemiologie an der McGill University (Montreal) unter Betreuung von

Prof. Olli S. Miettinen; den Abschluss „Master of Science in Epidemiology“ (MSc) er-

langte ich 1995. Von 1995 bis 1999 studierte ich an der medizinischen Hochschule

der McGill Universität (Montreal) und beendete erfolgreich das Medizinstudium mit

der Erlangung des Titels „Doctor of Medicine, Master of Surgery“ (MDCM) im Mai

1999.
 76

Im Herbst 1999 heiratete ich und emigrierte mit meinem deutschen Mann, Dr. rer.

nat. Christoph Läer, Dipl. Epid., MRPharmS, in die Bundesrepublik Deutschland.

Seit dem 15. November 1999 bin ich als Ärztin im Praktikum (ÄiP) und anschließend

dazu als wissenschaftliche Mitarbeiterin und Weiterbildungsassistentin an der Uni-

versität Göttingen, Abteilung für Allgemeinmedizin, tätig. Ab 1.1.2001 habe ich paral-

lel dazu als ÄiP und danach als Assistenzärztin im St. Martini Krankenhaus, Abtei-

lung Innere Medizin (Duderstadt), meine Weiterbildung fortgesetzt. Die deutsche

Approbation als Ärztin erhielt ich am 20.4.2001.

Meine Tätigkeit als Assistenzärztin setzte ich bis zum 31. Dezember 2001 fort. Seit-

dem habe ich meine Weiterbildung für anderthalb Jahre vorübergehend unterbro-

chen, um mich mit einem Forschungsvorhaben zu beschäftigen (PhD thesis, De-

partment of Epidemiology and Biostatistics, McGill University), das ich in Zusammen-

arbeit mit Prof. Jacques LeLorier (Université de Montréal) und Prof. Michal Abraha-

mowicz (McGill University) erfolgreich beantragt habe. Weiterhin setze ich meine

Forschungstätigkeit in der hiesigen Abteilung für Allgemeinmedizin auf Teilzeitbasis

fort. Ich habe vor, meine Facharztausbildung im Bereich Allgemeinmedizin nach

dem Abschluss des PhD-Projekts (etwa Mitte 2003) abzuschließen.

Göttingen, im November 2002