This is not the most recent version of the article. View current version (9 OCT 2014)

Intervention Review

You have free access to this content

Antibiotics for community acquired pneumonia in adult outpatients

  1. Lise M Bjerre1,*,
  2. Theo JM Verheij2,
  3. Michael M Kochen1

Editorial Group: Cochrane Acute Respiratory Infections Group

Published Online: 7 OCT 2009

Assessed as up-to-date: 6 JUL 2009

DOI: 10.1002/14651858.CD002109.pub3


How to Cite

Bjerre LM, Verheij TJM, Kochen MM. Antibiotics for community acquired pneumonia in adult outpatients. Cochrane Database of Systematic Reviews 2009, Issue 4. Art. No.: CD002109. DOI: 10.1002/14651858.CD002109.pub3.

Author Information

  1. 1

    University of Göttingen, Department of General Practice/Family Medicine, Göttingen, Germany

  2. 2

    University Medical Center Utrecht, Julius Center for General Practice and Patient-Oriented Research, Utrecht, Netherlands

*Lise M Bjerre, Department of General Practice/Family Medicine, University of Göttingen, Humboldtallee 38, Göttingen, D-37073, Germany. lbjerre@hotmail.com. lbjerre@gwdg.de.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 7 OCT 2009

SEARCH

This is not the most recent version of the article. View current version (09 OCT 2014)

 
Characteristics of included studies [ordered by study ID]
Anderson 1991

MethodsDate and duration of study: not specified. Follow-up: 6 to 8 weeks. Patients were included from 57 general practitioners in the UK. Double-blind, double-dummy technique, intention-to-treat results provided


ParticipantsPatients with CAP older than 18 years. CAP diagnosis confirmed by 3 of the following features: pyrexia, dyspnoea, tachypnoea, rales, localized reduced breath sounds and cough. Diagnosis of CAP was later confirmed radiographically. Total: n = 208. Evaluable for efficacy: n = 108 (exclusion usually due to failure to confirm initial Dx on CXR), n = 64 (clarithromycin), n = 44 (erythromycin). Exclusion criteria clear


InterventionsClarithromycin 250 mg bid for 14 days or erythromycin 500 mg qid for 14 days, each given at least 1 hour before or 2 hours after meals, mean treatment duration: 13 days (clarithromycin), or 10 days (erythromycin). Compliance assessment: tablet count


OutcomesPrimary outcome: Clinical response at 2 weeks (test-of-cure visit): 98% (clarithromycin), 91% (erythromycin). Treatment-related adverse events: 16% (clarithromycin) versus 33% (erythromycin), p = 0.004, mainly gastrointestinal side effects


NotesThree of 5 authors from Abbott Laboratories, source of funding not specified


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?UnclearRandomization method not specified

Allocation concealment?Unclearbecause randomization method not specified

Blinding?
All outcomes
YesDouble-blind, double-dummy technique

Incomplete outcome data addressed?
All outcomes
YesDetailed list of reasons for exclusion from efficacy analyses, with number of patients affected

Free of selective reporting?YesPre-specified outcomes appear to be fully reported

Free of other bias?UnclearThree of 5 authors from Abbott Laboratories, source of funding not specified

Chien 1993

MethodsDate and duration of study: January 1989 to June 1990. Follow-up: 4-6 weeks. Multicenter study (15 centers of the Canada-Sweden Clarithromycin-Pneumonia Study Group, 11 in Canada, 4 in Sweden). Double-blind, double-dummy technique, no intention-to-treat results provided


ParticipantsAmbulatory patients older than 12 years with CAP. N = 268 all patients, after exclusions 173 "evaluable patients": n = 92 (clarithromycin), n = 81 (erythromycin). Patients with mild or moderate infection. Drop-outs: 35% (due to less than minimum therapy, premature discontinuation, unavailable for follow-up, misdiagnosis, inadequate data collection, concomitant medication, underlying condition). Exclusion criteria clear


InterventionsClarithromycin: 250 mg every 12 hours, or erythromycin stearate: 500 mg every 6 hours. Mean treatment duration not specified (minimum duration: 7 days, intended duration: 7 to 14 days). Compliance assessment: tablet count


OutcomesPrimary outcome: Clinical success (cure and improvement) 97% clarithromycin, 96% erythromycin. Treatment-related adverse events: 31% clarithromycin, 59% erythromycin (p < 0.001)


NotesResearch supported by Abbott Laboratories, Chicago


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?UnclearRandomization method not specified

Allocation concealment?YesQuote: "The randomization was blinded to both the investigators and the patients."

Blinding?
All outcomes
YesDouble-blind, double-dummy technique

Incomplete outcome data addressed?
All outcomes
YesDetailed list of reasons for exclusion from efficacy analyses, with number of patients affected

Free of selective reporting?YesPre-specified outcomes appear to be fully reported

Free of other bias?UnclearResearch supported by Abbott Laboratories, Chicago

D'Ignazio 2005

MethodsDate and duration of study: April 2003 to April 2004. Double-blind, double-dummy, non-inferiority trial, patients were recruited from 56 centers worldwide (Canada, Chile, India, Lithuania, Mexico, Peru, Russia, United States).


Participants427 outpatients, 18 years or older, 423 patients received the study medication. Patients were eligible for enrollment if they had a clinical diagnosis of mild to moderate CAP (signs and symptoms) and radiographic evidence of new pulmonary infiltrate. Patients also had to have a Fine Mortaliy risk class of I, II or III (< 90 points). Exclusion criteria are clearly listed


InterventionsSingle, 2 g dose of azithromycin microspheres versus levofloxacin 500 mg once daily for 7 days


OutcomesPrimary endpoint: clinical response in the clinical per-protocol population, on the basis of signs and symptoms of CAP, assessed at test of cure visit (TOC; day 13 to 21): azithro 89.7%, levo 93,7%, non significant difference. Treatment-related adverse effect: azi 19.9%, levo 12.3%, p = 0.0063, mainly gastrointestinal side effects


NotesStudy funded by Pfizer Inc., 3 of 5 authors are employees of Pfizer


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?UnclearRandomization method not specified

Allocation concealment?Unclearbecause randomization method not specified

Blinding?
All outcomes
YesDouble-blind, double-dummy technique

Incomplete outcome data addressed?
All outcomes
YesDetailed list of reasons for exclusion from efficacy analyses, with number of patients affected

Free of selective reporting?YesPre-specified outcomes appear to be fully reported

Free of other bias?UnclearStudy funded by Pfizer Inc., 3 of 5 authors are employees of Pfizer

Drehobl 2005

MethodsDate and duration of study: not specified. Double-blind, double-dummy, phase III trial, conducted in 58 outpatient centers worldwide (United States, Canada, Argentina, Russia, India, Estonia, Lithuania)


Participants501 outpatients, 16 years or older. Patients were eligible for enrollment if they had a clinical diagnosis of mild to moderate CAP (signs and symptoms) and radiographic evidence of new pulmonary infiltrate. Patients also had to have a modified Fine risk score of I or II (< 70 points).


InterventionsSingle 2 g dose of azithromycin microspheres administered as an oral suspension versus extended-release clarithromycin administered orally as two 500 mg capsules once daily for 7 days


OutcomesPrimary endpoint: clinical response in the clinical per protocol population, on the basis of signs and symptoms of CAP, assessed at test of cure visit (TOC; day 14 to 21): azi 91.8% versus clari 94.7% (non-significant difference). Treatment-related adverse effects: azi 26.3% versus clari 24.6% (non-significant difference), mainly gastrointestinal side effects


NotesStudy funded by Pfizer Inc., 2 of 4 authors are employees of Pfizer


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?YesQuote: "Subjects were randomized according to computer-generated pseudorandom code using the method of permutated blocks, balanced within investigational site"

Allocation concealment?YesQuote: " Randomization numbers were assigned by a central Web/telephone computer-based telerandomization system"

Blinding?
All outcomes
YesDouble-blind, double-dummy technique

Incomplete outcome data addressed?
All outcomes
YesDetailed list of reasons for exclusion from efficacy analyses, with number of patients affected

Free of selective reporting?YesPre-specified outcomes appear to be fully reported

Free of other bias?UnclearStudy sponsored by Pfizer

Kohno 2003

MethodsDate and duration of study: December 2000 to June 2001. Double-blind, double-dummy, randomized phase III trial, conducted in 117 centers in Japan


Participants270 patients (mix of in- and outpatients), aged 16 to 80 years. 123 outpatients were included. Exclusion criteria are clearly stated and patient selection flow chart is provided


InterventionsTelithromycin 600 mg (= 2 x 300 mg) once daily after breakfast versus levofloxacin 100 mg three times daily for 7 days


OutcomesResults are reported separately for in- and outpatients. Primary endpoint: clinical response in the clinical per protocol population, on the basis of signs and symptoms of CAP, assessed at end of treatment (EOT; 7 days after treatment initiation) teli 95.7% versus clari 96.3% (non-significant difference). Treatment-related adverse effects:teli 33.6% versus levo 33.9% (non-significant difference), mostly gastrointestinal side effects


NotesStudy report is in Japanese, but extensive figures and tables are provided in English, so that most of the necessary information could be extracted. Missing information was kindly provided by the main author, Prof. Shigeru Kohno, MD, PhD, of Nagasaki University, Japan. Funding provided by Astellas (Fujisawa) Pharmaceutical Co. Ltd


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?UnclearText in Japanese; tables and figures in English

Allocation concealment?UnclearText in Japanese; tables and figures in English

Blinding?
All outcomes
YesDouble-blind, double-dummy technique

Incomplete outcome data addressed?
All outcomes
YesDetailed flow chart of reasons for exclusion from efficacy analyses, with number of patients affected

Free of selective reporting?UnclearText in Japanese; tables and figures in English

Free of other bias?UnclearStudy sponsored by Fujisawa

Mathers Dunbar 2004

MethodsDate and duration of study: May 1998 to Sept. 1999. Double-blind, double-dummy, parallel-group clinical trial conducted at 54 centers in 4 countries (United States, Canada, Argentiina and Chile)


Participants493 outpatients aged 18 and over. Patients were eligible for enrollment if they had a clinical diagnosis of acute CAP (signs and symptoms) and chest radiographic findings supporting the clinical diagnosis of bacterial pneumonia


InterventionsTelithromycin 800 mg (= 2 x 400mg capsules) once daily in the morning versus clarithromycin 500 mg (= 2 x 250 mg capsules) twice daily for 10 days


OutcomesPrimary endpoint: clinical response in the clinically assessable per protocol population, on the basis of signs and symptoms of CAP, assessed at test of cure visit (TOC; day 17 to 24): teli 88.3% versus clari 88.5% (non-significant difference). Treatment-related adverse effects: teli 38.5% versus clari 27.9%, mostly gastrointestinal side effects


NotesFunded by an unrestricted educational grant from Aventis Pharmaceuticals


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?YesQuote: "Patients... were randomized according to a schedule generated by the sponsor for each center. The schedule linked sequential treatment assignment number to treatment codes allocated at random."

Allocation concealment?YesQuote: " Randomization schedules were held by the sponsor".

Blinding?
All outcomes
YesDouble-blind, double-dummy technique. Quote: "Treatment blinding was ensured by encapsulation of both active study medication and placebo within identical opaque capsules."

Incomplete outcome data addressed?
All outcomes
YesDetailed list of reasons for exclusion from efficacy analyses, with number of patients affected

Free of selective reporting?YesPre-specified outcomes appear to be fully reported

Free of other bias?UnclearFunded by an unrestricted educational grant from Aventis Pharmaceuticals

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Balgos 1999Mix of in- and outpatients (unspecified proportions)
Mix of diagnoses (chronic bronchitis and CAP), only about 50% with CAP; results reported separately
Comparison of two dosage regimens of the same drug (amoxycillin/clavulanic acid 875/125 mg bid versus 500/125 mg tid)

Ball 1994Review of a series of unblinded trials
Mix of patients of different studies

Balmes 1991Mix of diagnoses (acute bronchitis and pneumonia)
Only 8/110 (7%) patients had a diagnosis of pneumonia
Mix of in- and out-patients (unspecified proportions)

Bantz 1987Mix of diagnoses (bronchitis and pneumonia)
Only 15/108 (14%) patients had a diagnosis of pneumonia
Data not reported separately

Biermann 1988No chest X-ray for every patient with suspected pneumonia

Bonvehi 2003Focus on pneumococcal CAP; positive sputum cultures used as a selection criterion

Carbon 1999Mix of in- and out-patients, data not reported separately; authors did not respond, or separate data could not be provided by authors

Chodosh 1991Only bacteriologically evaluable patients were included

Dark 1991Mix of diagnoses (bronchitis and pneumonia)
Only 23/272 (8%) patients had a diagnosis of pneumonia
Mix of in- and out-patients (unspecified proportions)

Dautzenberg 1992Open-label study

De Cock 1988Mix of diagnoses (acute bronchitis, acute superinfection of chronic bronchitis, pneumonia) and results reported separately.
Only 42/198 (21%) patients had a diagnosis of pneumonia (results not reported separately for each diagnostic group)

Dean 2006Open-label study

Donowitz 1997Focus on bacterial pneumonia (acute onset, purulent sputum, Gram stain and chest X-ray consistent with bacterial pneumonia were required for inclusion into the study)

File 1997Mix of in- and out-patients, data not reported separately, authors did not respond, or separate data could not be provided by authors

File 2001Mix of in- and out-patients, data not reported separately, authors did not respond, or separate data could not be provided by authors

File 2004Focus on bacterial pneumonia

Fogarty 1999Mix of in- and out-patients, data not reported separately, authors did not respond, or separate data could not be provided by authors

Fogarty 2002Focus on bacterial pneumonia. Serologic evidence of M. pneumoniae or C. pneumoniae was an exclusion criterion; serologic testing was systematically carried out pre-treatment as well as twice after treatment

Fong 1995Too small (n < 30)

Gotfried 2002Culture confirmation of bacterial CAP required

Gris 1996Too small (n < 30)

Hagberg 2002Mix of in and out-patients, data not reported separately, authors did not respond, or separate data could not be provided by authors

Higuera 1996Open-label study

Hoeffken 2001Patients were recruited as outpatients, however, 30% were hospitalized in the course of the study, data for in- and outpatients are not reported separately, and hospitalization was not considered treatment failure

Hoepelman 1993Mix of diagnoses (infective exacerbation of chronic obstructive pulmonary disease, purulent bronchitis and pneumonia) and results not reported separately
Only 9/99 (9%) patients had pneumonia

Pneumonia not necessarily community acquired

Hoepelman 1998Mix of diagnoses (infective exacerbation of chronic obstructive pulmonary disease, purulent bronchitis and pneumonia) and results not reported separately; only 3 of 144 patients had CAP

Kammer 1991Only bacteriologically evaluable patients were included

Kiani 1990Mix of in and out-patients, data not reported separately
Mix of diagnoses (acute bronchitis, pneumonia, acute exacerbation of chronic bronchopulmonary infection) but data reported separately
Only 10/110 (9%) patients had a diagnosis of pneumonia

Kinasewitz 1991Focus on bacterial pneumonia (purulent sputum and leucocytosis were required for inclusion into study, patients without sputum pathogens were excluded from efficacy analysis, etc.)

Lacny 1972Mix of diagnoses (infection of soft tissue, infection of the upper respiratory tract, otitis, skin infection, conjunctivitis, pneumonia) and results not reported separately
Only 2/121 (2.6%) patients had a diagnosis of pneumonia
Completely institutionalized population
32.5% patients were younger than 16 years

Laurent 1996Mix of diagnoses (acute bronchitis, acute infectious exacerbations of chronic brochitis, or pneumonia) but results reported separately
Only 43/204 (21%) patients had a diagnosis of pneumonia
Mix of in- and outpatients (unclear in which proportions)

Leophonte 2004Focus on pneumococcal pneumonia

Liipo 1994Dirithromycin withdrawn from market

Lode 1995Only inpatients (personal communication H. Lode), contrary to study report (Patients reportely treated as either in- or outpatients)

Lode 1998Combines patients from four other RCTs, including only those patients with S. pneumoniae pneumonia confirmed by blood culture, i.e. highly select sub-group, generalizability questionable. The data from the four studies can be obtained from the original reports (one of which is Örtqvist 1996, already excluded from this review because open-label)

Lode 2004aGatifloxacin withdrawn from market

Lode 2004bGatifloxacin withdrawn from market

Macfarlane 1996Single-blind study, no chest X-ray required

Moola 1999Grepafloxacin withdrawn from market

Müller 1992Only bacteriologically evaluable patients were included

NAPSG 1997Fleroxacin withdrawn from market

Neu 1993Mix of diagnoses (acute bacterial exacerbation of chronic bronchitis or asthmatic bronchitis and bacterial pneumonia) and data not reported separately
Unclear whether in- or outpatients
Only 43/213 (20%) had a diagnosis of CAP

O'Doherty 1997Grepafloxacin withdrawn from worldwide market in October 1999 due to risk of severe arrhythmias

O'Doherty 1998Open-label study

Pavie 2005Descriptive study, no RCT

Petitpretz 2001Focus on pneumococcal pneumonia

Peugeot 1991Open-label study

Pullman 2003Excluded because trovafloxacin no longer available for outpatients due to severe hepatotoxicity. Study recruitment was terminated because FDA restricted use of trovafloxacin to hospitalized patients with severe life- or limb-threatening infections (1999)

Rahav 2004Open-label study

Ramirez 1999Sparfloxacin withdrawn from market due to increased risk of severe phototoxicity and rash, as well as rhabdomyolysis

Rayman 1996Comparison of different dosage regimen of the same drug.
Included patients with either CAP or acute exacerbation of chronic bronchitis (i.e. mixed indications)
Groups were not similar at baseline (more smokers and more failures of previous treatment in the "twice daily" group, for which more adverse reactions were reported)

Saito 2008Mix of in- and out-patients, separate data for outpatients not available.

Note: Article in Japanese. However, tables and figures in English; additional information obtained through personal communication with one of the authors (Kohno S)

Salvarezza 1998Open-label study

Schleupner 1988Mix of diagnoses (bronchitis, pneumonia)
Only 34/61 (56%) patients had a diagnosis of pneumonia
Mix of in- and out-patients (unspecified proportions)

Siquier 2006Focus on pneumococcal pneumonia

Sokol 2002Excluded because trovafloxacin no longer indicated for outpatients, later removed from market due to severe hepatotoxicity. Study recruitment was terminated because FDA restricted use of trovafloxacin to hospitalized patients with severe life- or limb-threatening infections (1999)

Sopena 2004Open-label study

Tellier 2004Mix of in- and out-patients, data not reported separately, authors did not respond, or separate data could not be provided by authors

Tilyard 1992Too small (n < 30)

Torres 2003Mix of in- and out-patients, data not reported separately, authors did not respond, or separate data could not be provided by authors

Trémolières 1998Trovafloxacin withdrawn from market due to severe hepatotoxicity

Trémolières 2005Focus on bacterial pneumonia

van Zyl 2002Focus on bacterial pneumonia; purulent sputum sample required; patients with serologic evidence of Mycoplasma pneumoniae or Chlamydophila pneumoniae were excluded

Zuck 1990Open-label, unblinded study comparing oral administration twice daily versus intramuscular administration once daily. Included only high-risk patients; unclear whether treated as in- or outpatients

Örtqvist 1996Patients had CAP but were treated exclusively as inpatients

 
Comparison 1. Erythromycin versus clarythromycin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Clinical success2280Odds Ratio (M-H, Fixed, 95% CI)2.27 [0.66, 7.80]

 2 Bacteriological success257Odds Ratio (M-H, Fixed, 95% CI)0.28 [0.03, 2.57]

 3 Radiological success2276Odds Ratio (M-H, Fixed, 95% CI)0.91 [0.33, 2.49]

 4 Drug-related adverse events2476Odds Ratio (M-H, Fixed, 95% CI)0.30 [0.20, 0.46]

 
Comparison 2. Azithromycin microspheres versus levofloxacin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Clinical success1363Odds Ratio (M-H, Fixed, 95% CI)0.59 [0.27, 1.26]

 2 Bacteriological success1237Odds Ratio (M-H, Fixed, 95% CI)0.81 [0.32, 2.02]

 3 Drug-related adverse events1423Odds Ratio (M-H, Fixed, 95% CI)1.78 [1.04, 3.03]

 
Comparison 3. Azithromycin microspheres versus clarithromycin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Clinical success1411Odds Ratio (M-H, Fixed, 95% CI)0.69 [0.31, 1.55]

 2 Bacteriological success1303Odds Ratio (M-H, Fixed, 95% CI)1.17 [0.52, 2.61]

 3 Drug-related adverse events1499Odds Ratio (M-H, Fixed, 95% CI)1.09 [0.73, 1.64]

 
Comparison 4. Telithromycin versus clarithromycin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Clinical success1318Odds Ratio (M-H, Fixed, 95% CI)0.98 [0.49, 1.95]

 2 Bacteriological success162Odds Ratio (M-H, Fixed, 95% CI)0.24 [0.03, 2.29]

 3 Drug-related adverse effects1443Odds Ratio (M-H, Fixed, 95% CI)1.61 [1.08, 2.40]

 
Comparison 5. Telithromycin versus levofloxacin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Clinical success1123Odds Ratio (M-H, Fixed, 95% CI)0.85 [0.14, 5.25]

 2 Bacteriological success186Odds Ratio (M-H, Fixed, 95% CI)0.03 [0.00, 0.60]

 3 Clinical efficacy against H. influenzae146Odds Ratio (M-H, Fixed, 95% CI)4.62 [0.38, 55.51]

 4 Drug-related adverse effects1240Odds Ratio (M-H, Fixed, 95% CI)0.99 [0.58, 1.68]