Intervention Review

You have free access to this content

Corticosteroids for pneumonia

  1. Yuanjing Chen1,
  2. Ka Li2,
  3. Hongshan Pu3,
  4. Taixiang Wu4,*

Editorial Group: Cochrane Acute Respiratory Infections Group

Published Online: 16 MAR 2011

Assessed as up-to-date: 31 DEC 2010

DOI: 10.1002/14651858.CD007720.pub2


How to Cite

Chen Y, Li K, Pu H, Wu T. Corticosteroids for pneumonia. Cochrane Database of Systematic Reviews 2011, Issue 3. Art. No.: CD007720. DOI: 10.1002/14651858.CD007720.pub2.

Author Information

  1. 1

    West China Hospital, Sichuan University, Department of Anesthesiology, Chengdu, Sichuan, China

  2. 2

    West China Hospital, Sichuan University, Department of Surgery III, Chengdu, Sichuan, China

  3. 3

    West China Hospital, Sichuan University, Chengdu, Sichuan, China

  4. 4

    West China Hospital, Sichuan University, Chinese Cochrane Centre, Chinese Clinical Trial Registry, Chinese Evidence-Based Medicine Centre, INCLEN Resource and Training Centre, Chengdu, Sichuan, China

*Taixiang Wu, Chinese Cochrane Centre, Chinese Clinical Trial Registry, Chinese Evidence-Based Medicine Centre, INCLEN Resource and Training Centre, West China Hospital, Sichuan University, No. 37, Guo Xue Xiang, Chengdu, Sichuan, 610041, China. txwutx@hotmail.com.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 16 MAR 2011

SEARCH

 
Characteristics of included studies [ordered by study ID]
Cao 2007

MethodsDesign: randomized controlled trial

Method of randomisation: adequate. Randomisation numbers were generated by computer

Concealment of allocation: adequate. The randomisation numbers were sealed in envelopes

Outcome assessor blinding: unclear

Withdrawals/drop-outs: all participants accounted for


ParticipantsCountry: China

Eligible: not clearly stated in the article

Randomised: 120 (30 in control group, 30 in budesonide (Pulmicort) group, 30 in clarityne (loratadine tablets) group, 30 in thymosin or Transfer Factor injection)

Completed: 120

Age: from 3 months to 14 years old. The detail of the age proportion in each group was not stated in the article

Sex: 76 males and 44 females in the study, but the detail of the sex proportion in each group was not stated in the article

Diagnosis: not clearly stated in the article


InterventionsControl group: azithromycin or erythromycin

Budesonide (Pulmicort): azithromycin or erythromycin as control group plus budesonide (Pulmicort) inhalation 250 to 500 g/d for 7 days

Clarityne group: azithromycin or erythromycin as control group plus oral clarityne (loratadine tablets) for 1 month

Thymosin or Transfer Factor injection: thymosin 5 mg or Transfer Factor 1 U per time, twice a week for 1 month


OutcomesThe period until clinical symptoms resolved

The frequency of recurrent respiratory tract infections

The level of serum eosinophil cationic protein (ECP), eosinophils (EOS), interleukin-4 (IL- 4) and total immunoglobulin E (T- IgE)


NotesIn the article there was no description about the randomisation method. We telephoned the author and were told that the allocation sequence was generated by computer and the randomized numbers were sealed in envelopes. The assignment of the treatment for each participant was done by opening the envelopes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomisation method was not mentioned in the article

Allocation concealment (selection bias)High riskNot mentioned in the article

Blinding (performance bias and detection bias)
All outcomes
High riskNot stated clearly in the text. We contacted the author who told us it was single-blinded because the doctor knew the assignment of the treatment

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo missing data in the study

Selective reporting (reporting bias)Unclear riskNot described

Other biasHigh riskNot clearly stated in the article

Confalonieri 2005

MethodsDesign: randomized, double-blind, placebo-controlled trial

Method of randomisation: adequate. Randomisation schemes were generated in blocks of 10 for each participating site by a central randomisation centre

Concealment of allocation: adequate. The randomisation assignment was given to the recruiting centre in sealed envelopes

Outcome assessor blinding: yes

Withdrawals/drop-outs: all participants accounted for, apart from 2 participants who met the exclusion criteria after randomisation


ParticipantsEligible: assessed 121 for eligibility, 52 met eligibility criteria and 48 participants consented to the study

Randomised: 48 (24 in hydrocortisone infusion group, 24 in placebo group)

Completed: 46 (23 in hydrocortisone infusion group, 23 in placebo group)

Age: hydrocortisone infusion group: 60.4 (± 17.3), placebo group: 66.6 (± 14.7)

Sex (male): hydrocortisone infusion group: 15, placebo group: 17

Diagnosis: clinical and radiographic evidence of pneumonia


InterventionsHydrocortisone infusion group: hydrocortisone was given as an intravenous 200 mg loading bolus followed by an infusion (hydrocortisone 240 mg in 500 cc 0.9% saline) at a rate of 10 mg/hour for 7 days and protocol-guided antibiotic treatment

Placebo group: protocol-guided antibiotic treatment plus placebo


OutcomesImprovement in PaO2:FiO2 (PaO2:FiO2 300 or 100 increase from study entry)

Multiple organ dysfunction syndrome (MODS) score by study day 8

Development of delayed septic shock

Duration of mechanical ventilation

Length of ICU and hospital stay

Survival until hospital discharge and to 60 days after discharge


NotesRandomisation schemes were generated in blocks of 10 for each participating site by a central randomisation centre. The randomisation assignment was given to the recruiting centre in sealed envelopes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomisation schemes were generated in blocks of 10 for each participating site by a central randomisation centre

Allocation concealment (selection bias)Low riskThe randomisation assignment was provided to the recruiting centre in sealed envelopes

Blinding (performance bias and detection bias)
All outcomes
Low riskThe participants did not know whether they were given hydrocortisone or placebo

Incomplete outcome data (attrition bias)
All outcomes
Low riskAll participants accounted for, except 2 participants who met the exclusion criteria after randomisation

Selective reporting (reporting bias)Unclear riskNot described

Other biasUnclear riskNot clearly stated in article

Marik 1993

MethodsDesign: prospective, randomized, placebo-controlled trial

Method of randomisation: adequate; participants were randomized by a random number generator to receive either hydrocortisone or placebo

Concealment of allocation: unclear

Outcome assessor blinding: yes

Withdrawals/drop-outs: all participants accounted for


ParticipantsEligible: not stated

Randomised: 30 (14 in hydrocortisone group, 16 in placebo group)

Completed: 30 (14 in hydrocortisone group, 16 in placebo group)

Age: hydrocortisone group: 40.6 ± 14.7, placebo group: 31.7 ± 12.8

Sex: not stated in the article

Diagnosis: the diagnosis of pneumonia was based on clinical signs and symptoms of lower respiratory tract infection


InterventionsThe participants received 10 mg/kg of hydrocortisone or placebo intravenously 30 minutes prior to starting antibiotic therapy

All participants were treated with cefotaxime 1 g intravenously every 6 hours


OutcomesThe level of Tumour Necrosis Factor-α in both groups

The length of stay in the ICU

APACHE score

Mortality


NotesParticipants were randomized by a random number generator to receive either hydrocortisone or placebo


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskParticipants were randomized by a random number generator to receive either hydrocortisone or placebo

Allocation concealment (selection bias)High riskIt was not clearly stated in the text

Blinding (performance bias and detection bias)
All outcomes
Low riskThe participants did not know the drug they were given

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo missing data in the trial

Selective reporting (reporting bias)Unclear riskNot reported

Other biasUnclear riskNot described

McHardy 1972

MethodsThe study was carried out in the respiratory wards of the City Hospital, Edinburgh, New South Wales, Australia. Parallel design


Participants126 participants were included and no children under 12 years of age. The criteria for inclusion were radiological evidence of pneumonia or clinical evidence of pneumonia if no pre-treatment chest radiograph was available. Participants with pneumonia were excluded if they were either classed as "desperately ill" and judged to be at risk of dying within 24 hours, if they were known to be hypersensitive to penicillin or ampicillin. Patients with diabetes mellitus or symptoms of recent peptic ulceration were included in the ampicillin dosage trial but excluded from the random allocation of prednisolone

The participants were divided into four groups, 43 in group 1, 20 in group 2, 43 in group 3 and 20 in group 4


InterventionsPariticipants in group 1 were administered ampicillin 1 g daily; ampicillin 1 g plus prednisolone 20 mg daily in group 2; ampicillin 2 g daily in group 3 and ampicillin 2 g plus prednisolone 20 mg daily in group 4


Outcomes1. Death
2. Duration of treatment
3. Change of treatment
4. Resolution of temperature
5. Clearance of pathogens from sputum or laryngeal swabs
6. Maximum radiological clearance


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskThe method of generating the allocation sequence was not described in detail

Allocation concealment (selection bias)High riskAllocation concealment was not mentioned

Blinding (performance bias and detection bias)
All outcomes
High riskNot mentioned

Incomplete outcome data (attrition bias)
All outcomes
Low risk

Mikami 2007

MethodsDesign: prospective, randomized, open-label, controlled trial

Method of randomisation: unclear

Concealment of allocation: adequate

Outcome assessor blinding: yes

Withdrawals/drop-outs: all participants were accounted for


ParticipantsEligible: 110 hospitalized participants with adult CAP; 60 met eligibility criteria and 31 patients consented to the study

Randomised: 31

Completed: 31 (15 in steroid group, 16 in control group)

Age: steroid group: 75.9 (± 16.0), control group: 68.4 (± 22.8)

Sex (male): steroid group: 73.3%, control group: 75.0%

Diagnosis: the diagnosis of CAP was based on clinical signs and symptoms of lower respiratory tract infections. Radiographic abnormalities consistent with infection were neither pre-existing nor caused by any other previous conditions


InterventionsSteroid group: 40 mg of prednisolone intravenously for 3 days plus intravenous antibiotics within 8 hours of hospital arrival and modified based on culture results

Control group: intravenous antibiotics within 8 hours of hospital arrival and modified based on culture results


OutcomesLength of hospital stay

Duration of IV antibiotic treatment and time required to stabilise vital signs


NotesThe investigators were not actively involved in the treatment of the participants and the physicians who took care of the participants were not informed of the outcome parameters. Decisions on the time to discharge or the time to finish intravenous antibiotic therapy were made with objective data and were dependent on the condition of the participant

Sources of the drugs not stated


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskMentioned randomisation in the article, but did not describe the method

Allocation concealment (selection bias)High riskDid not mention this in the article

Blinding (performance bias and detection bias)
All outcomes
High riskAn open-label study

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo missing data

Selective reporting (reporting bias)Unclear riskNot reported

Other biasUnclear riskNot described

van Woensel 2003

MethodsDesign: prospective, multicentre, randomized, double-blind, placebo-controlled trial

Method of randomisation: adequate

Concealment of allocation: adequate

Outcome assessor blinding: yes

Withdrawals/drop-outs: 3 participants dropout


ParticipantsEligible: 155 eligible patients. Not randomized (n = 70) including not approached for participation (n = 29), more than 24 hours mechanical ventilation (n = 12), corticosteroids before randomisation (n = 11), and no permission from parents (n = 18)

Randomised: 85 (39 in steroid group, 46 in control group)

Completed: 82 (37 in steroid group including 18 bronchiolitis and 17 pneumonia, 45 in control group including 21 bronchiolitis and 22 pneumonia)

Age: steroid group: 5.9 (0.9), control group: 9.8 (1.6)

Sex: steroid group male: 30, control group: 27

Diagnosis: the diagnosis was based on clinical signs and symptoms of lower respiratory tract infection. The presence of RSV infection was confirmed by direct immunofluorescence assay (Imagen, Dako, Denmark) of a nasopharyngeal aspirate during the first 24 hours after hospital admission


InterventionsThe trial medication was intravenous dexamethasone (0.15 mg/kg 6-hourly for 48 hours) or placebo and had to be started within 24 hours after the start of mechanical ventilation


OutcomesDuration of mechanical ventilation

Length of stay in the PICU and in the hospital

Duration of supplemental oxygen


NotesParticipants were allocated to dexamethasone or placebo by computerized block randomisation in groups of 10, centrally performed by an independent pharmacist. Each participating centre received a randomisation list and the trial medication preparation protocol. The trial medication was prepared in advance in the pharmacy centre where the concealed randomisation list was kept until the study was completed. Each enrolled participant received the trial medication with the next number in sequence


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskParticipants were allocated to dexamethasone or placebo by computerized block randomisation in groups of 10, centrally performed by an independent pharmacist. Each participating centre received a randomisation list and the trial medication preparation protocol

Allocation concealment (selection bias)Low riskThe trial medication was prepared in advance in the pharmacy centre where the concealed randomisation list was kept until the study was completed

Blinding (performance bias and detection bias)
All outcomes
Low riskEach enrolled participant received the trial medication with the next number in sequence

Incomplete outcome data (attrition bias)
All outcomes
Low riskAfter enrolment 3 participants were withdrawn: 1 (dexamethasone group) because she died 4 days after admission due to severe cerebral oedema in combination with refractory seizures already existing before inclusion, and 2 because their medication vials (1 containing dexamethasone and the other placebo) were accidentally interchanged

Selective reporting (reporting bias)Unclear riskNo report

Other biasUnclear riskNot described

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Bai 2003The author refused to explain the details of the study

Cai 2007We contacted the trial author who told us it was a retrospective study

Chen 2006We contacted the trial author who told us it was a retrospective study

Chen 2006bWe contacted the trial author who told us it was a retrospective study

Chen 2007The study mainly worked on the effect of ambroxol. We contacted the trial author who told us it was a retrospective study

Chu 2009We contacted the trial author who told us that there was no blinding in the study and the participants were allocated by the date of administration

Dai 2009We contacted the trial author who told us it was a retrospective study

Fang 2003We contacted the trial author who told us it was a retrospective study

Feng 2001We contacted the trial author who told us it was a retrospective study

Finch 2002The treatment drug did not include corticosteroids

Guan 2004We contacted the author but she refused to give details of the study

Han 2003We contacted the trial author who told us it was a retrospective study

Han 2007We contacted the trial author who told us it was a retrospective study

He 2008A prospective study. The participants were allocated casually by the doctor and there was no blinding in the study

Hong 2005We contacted the trial author who told us it was a retrospective study

Huang 2008We contacted the author who told us the study was retrospective and prospective. In the prospective part, the participants were allocated by the doctor according to the date of administration

Li 2006We contacted the trial author who told us it was a retrospective study

Marks 1972Hydrocarbon pneumonia is one of the most significant complications after ingestion of hydrocarbon distillates. Study did not meet our inclusion criteria

Mer 1998An uncontrolled, retrospective, prospective study

Song 2005We contacted the trial author who told us it was a retrospective study

Song 2008Mainly compared the effect of Mucosolvan in the study

Sun 2006We contacted the trial author who told us it was a retrospective study

Tao 2008We contacted the trial author who told us it was a retrospective study

Wang 2007We contacted the trial author who told us it was a retrospective study

Wong 2006In the study the investigator compared the effect of corticosteroids plus Mucosolvan; this did not meet our intervention inclusion criteria

Wu 2007We contacted the trial author who told us it was a retrospective study

Wu 2009We could not contact the author as he had changed workplace, so we could not determine whether the study was a randomized controlled trial

Xie 2005We contacted the trial author who told us it was a retrospective study

Zhang 2004We contacted the trial author who told us it was a retrospective study

Zhao 2008We contacted the author who told us it was a retrospective study. In addition, there were some other interventions in the corticosteroid group

Zheng 2008We contacted the trial author who told us it was a retrospective study

 
Characteristics of studies awaiting assessment [ordered by study ID]
Braun 1986

MethodsWe are trying to retrieve the article

Participants

Interventions

Outcomes

Notes

Chen 2003

MethodsWe could not find the author as she had changed workplace

Participants

Interventions

Outcomes

Notes

Confalnonier 2006

MethodsWe are trying to retrieve the article

Participants

Interventions

Outcomes

Notes

Dambrava 2006

MethodsWe are trying to retrieve the article

Participants

Interventions

Outcomes

Notes

Hatakeyama 1995

MethodsWe are trying to retrieve the article

Participants

Interventions

Outcomes

Notes

Lee 1980

MethodsWe are trying to retrieve the article

Participants

Interventions

Outcomes

Notes

Li 2005

MethodsWe are trying to contact the author to identify the randomisation method used

Participants

Interventions

Outcomes

Notes

Li 2007

MethodsThe author was abroad, so we could not contact her

Participants

Interventions

Outcomes

Notes

Liu 2008

MethodsWe are trying to contact the author to identify the randomisation method used

Participants

Interventions

Outcomes

Notes

Lv 1991

MethodsWe are trying to contact the author to identify the randomisation method used

Participants

Interventions

Outcomes

Notes

Ma 2006

MethodsThe author was abroad, so we could not contact her

Participants

Interventions

Outcomes

Notes

Soboleva 2000

MethodsWe are trying to retrieve the article

Participants

Interventions

Outcomes

Notes

Wang 2008

MethodsWe are trying to contact the author to identify the randomisation method used

Participants

Interventions

Outcomes

Notes

 
Comparison 1. Hydrocortisone versus placebo

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

 1 Mortality276Peto Odds Ratio (Peto, Fixed, 95% CI)0.26 [0.05, 1.37]

 2 PaO2:FIO2improvement ≥100 from study entry146Risk Ratio (M-H, Fixed, 95% CI)2.5 [1.40, 4.47]

 3 PaO2:FIO2146Mean Difference (IV, Fixed, 95% CI)95.00 [45.17, 144.83]

 4 Chest radiograph score146Mean Difference (IV, Fixed, 95% CI)-1.5 [-2.10, -0.90]

 5 Improvement in chest radiograph score from day 1 to day 8146Risk Ratio (M-H, Fixed, 95% CI)4.2 [1.91, 9.21]

 6 On mechanical ventilation276Risk Ratio (M-H, Fixed, 95% CI)0.43 [0.22, 0.85]

 7 Length of stay in the ICU130Mean Difference (IV, Fixed, 95% CI)-0.30 [-3.81, 3.21]

 
Comparison 2. Prednisolone plus antibiotics versus antibiotics alone

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

 1 Death1126Peto Odds Ratio (Peto, Fixed, 95% CI)1.41 [0.39, 5.04]

    1.1 Ampicillin (1g daily) vs ampicillin (1g daily) with prednisolone (20mg daily)
163Peto Odds Ratio (Peto, Fixed, 95% CI)2.72 [0.56, 13.22]

    1.2 Ampicillin (2g daily) vs ampicillin (2g daily) with prednisolone (20mg daily)
163Peto Odds Ratio (Peto, Fixed, 95% CI)0.41 [0.05, 3.57]

 2 Participants require to change the treatment1126Odds Ratio (M-H, Fixed, 95% CI)0.84 [0.33, 2.14]

    2.1 Ampicillin (1g daily) vs ampicillin (1g daily) plus prednisolone
163Odds Ratio (M-H, Fixed, 95% CI)1.46 [0.27, 7.96]

    2.2 Ampicillin (2g daily) vs ampicillin (2g daily) plus prednisolone
163Odds Ratio (M-H, Fixed, 95% CI)0.64 [0.20, 2.01]

 3 Duration of treatment (number of participants to be treated more than two weeks)1126Odds Ratio (M-H, Fixed, 95% CI)2.0 [0.53, 7.53]

    3.1 Ampicillin (1g) vs ampicillin (1g) plus prednisolone (20mg)
163Odds Ratio (M-H, Fixed, 95% CI)3.08 [0.35, 27.48]

    3.2 Ampicillin (2g) vs ampicillin (2g) plus prednisolone (20mg)
163Odds Ratio (M-H, Fixed, 95% CI)1.46 [0.27, 7.96]

 4 Time to normalisation of spO2 (days)131Mean Difference (IV, Fixed, 95% CI)-2.10 [-4.43, 0.23]

 5 Length of hospital stay (days)131Mean Difference (IV, Fixed, 95% CI)-4.20 [-10.14, 1.74]

 6 Duration of IV antibiotics (days)131Mean Difference (IV, Fixed, 95% CI)-3.80 [-6.94, -0.66]

 7 Time to normalisation of body temperature (days)131Mean Difference (IV, Fixed, 95% CI)-3.6 [-6.28, -0.92]

 8 Time to normalisation of respiratory rate (days)131Mean Difference (IV, Fixed, 95% CI)-3.40 [-5.52, -1.28]

 9 Time to normalisation of C reactive protein (days)131Mean Difference (IV, Fixed, 95% CI)-4.1 [-7.15, -1.05]

 10 Time to normalisation of WBC (days)131Mean Difference (IV, Fixed, 95% CI)2.00 [-0.61, 4.61]

 
Comparison 3. Budesonide (Pulmicort) plus antibiotics versus antibiotics only

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

 1 Time to clinical symptoms disappearing160Mean Difference (IV, Fixed, 95% CI)-8.0 [-9.25, -6.75]

 2 Rate of relapse160Mean Difference (IV, Fixed, 95% CI)-2.43 [-2.79, -2.07]

 
Comparison 4. Intravenous dexamethasone versus placebo

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

 1 Duration of mechanical ventilation141Mean Difference (IV, Fixed, 95% CI)0.80 [0.15, 1.45]

 2 Duration of supplemental oxygen141Mean Difference (IV, Fixed, 95% CI)1.70 [0.75, 2.65]

 3 Length of stay in the paediatric intensive care unit141Mean Difference (IV, Fixed, 95% CI)0.20 [-0.50, 0.90]

 4 Length of stay in the hospital141Mean Difference (IV, Fixed, 95% CI)2.60 [1.43, 3.77]