Primary

• To examine the effectiveness of oral versus parenteral antibiotics for severe pneumonia in children aged 2 to 59 months.

Secondary

• To examine the safety of oral versus parenteral antibiotics for severe pneumonia in children aged 2 to 59 months.

Secondary

• To examine the odds of death when treated with oral antibiotics as compared to parenteral antibiotics for severe pneumonia in children aged 2 to 59 months.

Secondary

• To examine the odds of adverse effects of therapy in children when treated with oral antibiotics as compared to parenteral antibiotics for severe pneumonia in children aged 2 to 59 months.

Databases searched:

• MEDLINE (Ovid MEDLINE(R) 1946 to May Week 4 2012, Ovid MEDLINE(R) In-Process & Other Non-Indexed Citations May 30, 2012, Ovid MEDLINE(R) Daily Update May 30, 2012).

• EMBASE (1974 to 2012 May 30).

• Cochrane Central Register of Controlled Trials (CENTRAL) (EBM Reviews - Cochrane Central Register of Controlled Trials May 2012).

• Global Health (1910 to April 2012).

• Clinical Trials.gov (May 31st 2012).

• WHOLIS (World Health Organization Library Information System) (May 31st 2012).

• LILACS (Literatura Latino-Americana e do Caribe em Ciências da Saúde) (May 31st 2012).

• AIM (African Index Medicus) (May 31st 2012).

• WPRIM (Western Pacific Region Index Medicus) (May 31st 2012).

• IMSEAR (Index Medicus for South-East Asia Region) (May 31st 2012).

• IMEMR (Index Medicus for the Eastern Mediterranean Region) (May 31st 2012).

• OpenSIGLE (Open System for Information on Grey Literature in Europe) (May 31st 2012).

Reference lists from included studies will be hand searched.

Abstracts from the most recent conference proceedings will be hand searched for:

• American Academy of Pediatrics.

• Royal College of Paediatrics and Child Health (UK).

• European Society of Pediatric Infectious Disease.

• World Society of Pediatric Infectious Disease.

• European Academy of Paediatric Societies.

• International Congress of Pediatrics (International Pediatric Association).

Relevant specialist societies/associations will also be contacted to inquire of any unpublished literature in the area.

Inclusion criteria:

• Controlled trials with concurrent controls.

• Published or unpublished trials.

• An English translation of the trial is available.

Exclusion criteria:

• No controls.

• Historical controls.

Pneumonia is the leading global cause of mortality in children under five years old, responsible for an estimated 1.6 million deaths in 2008. Of these deaths, 70% occurred in sub-Saharan Africa and South East Asia alone.

The World Health Organization (WHO) guidelines for treatment of pneumonia in children under five years depend on disease severity, with current advice to treat non-severe pneumonia with oral antibiotics and severe or very severe pneumonia with parenteral antibiotics.

Treating severe pneumonia thus includes the cost of parenteral antibiotics and the direct costs of nursing care in hospital. In addition, indirect costs to the family including travel for treatment and loss of income whilst caring for the child are substantial. Referral to hospital from some rural areas can also be problematic and may not always lead to hospitalisation.

In 2006, the Cochrane Collaboration reviewed the available evidence on this subject, but only identified two studies suitable for inclusion. Whilst the evidence from these studies supported the equivalence of the two antibiotic regimes, no meta-analysis was performed in view of the significant heterogeneity.

Inclusion criteria. At least 80% of study participants must fit the inclusion criteria for the study to be included:

• Children aged between 2 and 59 months.

• Pneumonia classified by the authors as severe.

• Pneumonia was likely to have been community acquired.

Exclusion criteria:

• Pneumonia not described by the author as severe.

• Pneumonia described by the authors as very severe.

Inclusion criteria:

• Exclusively oral course of antibiotics (any antibiotic, any dose, any duration).

• Studies that use community case management as part of the intervention will be included.

The comparator is standard therapy. This will depend on the setting of the study (hospital/community) and include:

1) Exclusively parenteral antibiotics (any antibiotic, any dose, any duration), or

2) Combined parenteral and oral antibiotics i.e.

• one dose of oral antibiotics whilst urgently referring for parenteral antibiotics (in a community setting), or

• parenteral antibiotics until improved followed by oral antibiotics (in a hospital setting).

3) The chance to receive IV antibiotics as part of standard therapy for severe pneumonia depending on clinical judgement.

Primary outcomes

Effectiveness: odds of treatment failure as defined by the authors.

Odds ratio will be used. If available the unadjusted OR will be used.

Secondary outcomes

Effectiveness:

• Odds of clinical deterioration as defined by the authors.

• Odds of death.

• Odds of relapse of severe pneumonia.

• Odds of developing empyema.

Safety:

• Odds of developing a rash.

• Odds of diarrhoea.

• Odds of injection site reaction.

• Odds of author-defined adverse drug reactions.

Odds ratio will be used. If available the unadjusted OR will be used.

Two data extractors will collect the data into a pre-designed data collection sheet (available on request). Any discrepancies will be discussed and, if a consensus cannot be reached, a third data collector will be asked for their opinion.

Bias will be assessed using the Cochrane Collaboration Risk of Bias Tool.

Both qualitative and quantitative syntheses are planned.

Non-randomised studies will not be included in the quantitative synthesis but will be discussed in the qualitative synthesis.

Identified trials are likely to be “equivalence” studies. After expert statistical advice, the decision has been made to consider the estimated treatment effects of the studies instead of considering their hypotheses. Thus a standard meta-analysis approach can be justified, as used for trials examining for superiority. Methods to be used will be:

• Fixed effects model.

• Generic inverse variance method.

• Odds ratio – if available, the unadjusted odds ratio will be used. If unavailable, the adjusted odds ratio will be used.

• Cluster trials will be discussed with expert statistician to ensure appropriate analysis has taken place. If not, the methods outlined in the Cochrane Handbook will be used to calculate the effective sample size of the trial and the subsequent OR.

A priori sub-groups for analysis:

World Health Organization Criteria for Severe Pneumonia:

• Trials using WHO criteria for inclusion into the study or not.

Antibiotics:

• Control group uses only parenteral antibiotics.

• Control group uses mixed parenteral antibiotics.

• Intervention oral antibiotic stratified by type and dose (high dose or low dose as per the British National Formulary for Children).

Community case management:

• Community case management used in intervention group or not.

Country Income:

• High or high/middle income countries (as per World Bank country list).

• Low or low/middle income countries (as per World Bank country list).

Radiology:

• X-ray used as part of diagnostic criteria or not.

Publication will be sought in an appropriate peer reviewed journal.

Olivia Swann

West Lodge,

Main Road,

Westfield

West Lothian

EH48 3DF

livvyswann@doctors.org.uk

Unversity of Manchester

Dr Olivia Swann, University of Manchester
Dr Louisa Pollock, Imperial College London
Mr Andrew Vail, University of Manchester
Professor Harry Campbell, Univeristy of Edinburgh
Mr Ewen Harrison, Univerisity of Edinburgh

22 June 2012

03 September 2012

This review is being undertaken as a dissertation for a Masters of Research in Public Health through the University of Manchester. There are no sources of funding.

Professor Harry Campbell is the first author on one of the papers to be discussed in this review. He has not been involved in data collection, analysis or assessment of bias. His role in this review has been as a clinical expert only.

English

England

Subject indexing assigned by CRD

Administration, Oral; Anti-Bacterial Agents; Child, Preschool; Humans; Infant; Infusions, Parenteral; Pneumonia

13 July 2012

13 July 2012