Intervention Review

You have free access to this content

Inhaled corticosteroids for subacute and chronic cough in adults

  1. Kate J Johnstone1,*,
  2. Anne B Chang2,3,4,
  3. Kwun M Fong1,5,
  4. Rayleen V Bowman1,5,
  5. Ian A Yang1,5

Editorial Group: Cochrane Airways Group

Published Online: 28 MAR 2013

Assessed as up-to-date: 13 DEC 2012

DOI: 10.1002/14651858.CD009305.pub2


How to Cite

Johnstone KJ, Chang AB, Fong KM, Bowman RV, Yang IA. Inhaled corticosteroids for subacute and chronic cough in adults. Cochrane Database of Systematic Reviews 2013, Issue 3. Art. No.: CD009305. DOI: 10.1002/14651858.CD009305.pub2.

Author Information

  1. 1

    The University of Queensland, School of Medicine, Brisbane, Queensland, Australia

  2. 2

    Charles Darwin University, Menzies School of Health Research, Casuarina, Northern Territories, Australia

  3. 3

    Royal Children's Hospital, Queensland Children's Respiratory Centre, Brisbane, Queensland, Australia

  4. 4

    The University of Queensland, Queensland Children's Medical Research Institute, Brisbane, Australia

  5. 5

    The Prince Charles Hospital, Thoracic Medicine Program, Brisbane, Queensland, Australia

*Kate J Johnstone, School of Medicine, The University of Queensland, The Prince Charles Hospital, Rode Rd, Brisbane, Queensland, 4032, Australia. kate.johnstone@uqconnect.edu.au.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 28 MAR 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Boulet 1994

MethodsDesign: cross-over, no washout, first period data available.

Randomisation: yes, method not reported.

Blinding: double blind.

Withdrawals: none.


ParticipantsSetting: single-centre study, Laval Hospital clinic (Canada).

Number screened: not reported.

Number eligible: 19.

Number randomised: 14.

Number in treatment group: 14 (cross-over).

Number in control group: 14 (cross-over).

Number of withdrawals: 0.

Number completing trial: 14.

Sex: 4 M, 15 F (eligible population).

Age: 25 58 years (eligible population).

Cough duration: mean 3.8 years.

Inclusion criteria: dry cough for > 4 weeks, normal airway response to methacholine (PC20, the provacative concentration of methacholine inducing a 20% decrease in FEV1 being 20 mg/mL or more), normal chest examination and radiograph.

Exclusion criteria: use of a medication known to induce chronic cough (e.g. angiotensin-converting enzyme (ACE) inhibitors, beta-blockers), inhaled or oral steroid intake within the preceding 3 months, evidence of respiratory infection in the previous 4 weeks, smoking within the preceding 2 years, past or present history of asthma, chronic bronchitis, any other chest or systemic disease.

Baseline characteristics of treatment:control groups: comparable (cross-over).


InterventionsICS: beclomethasone dipropionate 500 μg, 4 times daily (2000 μg/day).

Control: placebo.

Administration method: MDI with Aerochamber.

Treatment duration: 4 weeks, no washout.

Co-interventions: none.


OutcomesSymptom score.

Other respiratory symptoms.

BHR (methacholine challenge, citric acid challenge).


NotesMain outcome: Mean daily cough scores not significantly different between the two treatment groups.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomised, method not reported.

Allocation concealment (selection bias)Unclear riskMethod not reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo drop-outs.

Selective reporting (reporting bias)Low riskNumber eligible reported, all outcomes reported.

Other biasLow riskFirst period data available.

Chaudhuri 2004

MethodsDesign: cross-over, 2 weeks washout between treatments.

Randomisation: yes, computer-generated.

Blinding: double blind.

Withdrawals: stated.


ParticipantsSetting: participants recruited from the community (general practitioner referrals and newspaper advertisement) and hospital respiratory clinics (Scotland).

Number screened: 120.

Number eligible: 93.

Number randomised: 93.

Number in treatment group: N/A (cross-over).

Number in control group: N/A (cross-over).

Number of withdrawals: 5.

Number completing trial: 88.

Sex: 32 M, 57 F (of 89 with sex recorded).

Age (years): mean 59.0 (SD 12.7).

Cough duration (years): mean 16.2 (SD 16.1).

Inclusion criteria: adults with cough for > 1 year.

Exclusion criteria: evidence of any other lung disease on the basis of history, clinical examination, chest radiography, and spirometry, treatment with inhaled or oral corticosteroids within 3 weeks of inclusion, URTI within 6 weeks of inclusion, ACE inhibitor treatment, smoking within the past year

Baseline characteristics of treatment:control groups: comparable (cross-over).


InterventionsICS: fluticasone 500 μg twice daily (1000 μg/day).

Control: placebo.

Administration method: DPI (Accuhaler).

Treatment duration: 2 weeks with 2 weeks washout.

Co-interventions: none.


OutcomesVAS (cough severity).

Sputum total and differential cell counts.

Eosinophilic cationic protein (ECP).

Myeloperoxidase (MPO).

Prostaglandin E2 (PGE2).

Leukotriene B4 (LTB4).

Cys-leukotrienes (Cys-LT).

Interleukin-8 (IL-8).

Tumour necrosis factor-alpha (TNF-α).

Exhaled nitric oxide (eNO).

Carbon monoxide (CO).


NotesMain outcome: cough severity and sputum ECP levels were modestly reduced by ICS. Other sputum biomarker levels were unaltered.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation.

Allocation concealment (selection bias)Low riskRandomisation code withheld from investigators until completion of the study.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind, study medication packed by central pharmacy and was identical in appearance and taste.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskFive withdrawals with reasons reported, treatment received by drop-outs not reported.

Selective reporting (reporting bias)Low riskNumber screened and eligible reported, reasons for withdrawal reported, all outcomes reported.

Other biasLow riskNo carry-over effect observed despite cross-over trial design with 2 weeks washout.

Evald 1989

MethodsDesign: cross-over, no washout, first period data available.

Randomisation: method not reported.

Blinding: double blind.

Withdrawals: stated.


ParticipantsSetting: Bispebjerg Hospital chest clinic (Denmark).

Number screened: not reported.

Number eligible: not reported.

Number randomised: 40.

Number in treatment group: N/A (cross-over).

Number in control group: N/A (cross-over).

Number of withdrawals: 13 (7 run-in, 2 first period, 4 second period).

Number completing trial: 31 first period, 27 second period.

Sex: 7 M, 24 F.

Age: 15-64 years.

Cough duration: 3 < 15 days; 6 > 1 month; 10 > 3 months; 9 > 1 year 3 > 5 years.

Inclusion criteria: daily dry cough of at least 1 hour duration in more than half of the last 30 days.

Exclusion criteria: obstructive lung function (ratio between the FEV1 and FVC no more than 70%), significant reversibility after bronchodilating treatment (increase in FEV1 30 min after three inhalations of salbutamol 300 μg and three inhalations ipratropium 60 μg not exceeding 20% or 500 mL), diurnal variation above 20% in morning and evening PEF during home monitoring for 1 week (calculated at visit 2 after the run-in period), treatment with anti-asthmatic drugs, abnormal chest X-ray, recent respiratory infection, other chest disease, pregnancy.

Baseline characteristics of treatment:control groups: comparable (cross-over).


InterventionsICS: beclomethasone dipropionate 50 μg 4 puffs twice daily (400 μg/day).

Control: placebo.

Administration method: MDI.

Treatment duration: 2 weeks, no washout.

Co-interventions: none.


OutcomesPatient's subjective effect of treatment (score).

Degree of cough (including number of days with cough, number of cough attacks a day).

Duration of cough attacks estimated for the whole day.

Spirometry.

PEF.

Number of awakenings each night because of cough.

Duration of insomnia at night because of cough.


NotesMain outcome: no significant treatment effect found for any of the measured variables.

Baseline investigations: spirometry, bronchodilator reversibility, bloods, skin prick test, bronchial provocation test.

Run-in period: 1 week


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomised, method not reported.

Allocation concealment (selection bias)Unclear riskMethod not reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Unclear risk2 withdrawals during period one, 4 during period two; treatment received by drop-outs not reported.

Selective reporting (reporting bias)Unclear riskNumber screened and eligible not reported, scores for participants' subjective effect of treatment not reported, reasons for withdrawal reported.

Other biasLow riskFirst period results described separately; run-in period.

Pizzichini 1999

MethodsDesign: parallel group.

Randomisation: yes, computer-generated.

Blinding: double blind.

Withdrawals: stated.


ParticipantsSetting: Firestone Regional Chest and Allergy Clinic and recruited by advertisement (Canada).

Number screened: 84.

Number eligible: 50.

Number randomised: 50.

Number in treatment group: 21 (completed).

Number in control group: 23 (completed).

Number of withdrawals (treatment:control): 6 (4:2).

Number completing trial (treatment:control): 44 (21:23).

Sex: 16 M, 28 F (completed).

Age: 20-75 years.

Cough duration: 6-19.2 years (completed).

Inclusion criteria: daily bothersome cough for at least 1 year with no other respiratory symptoms, no evidence of asthma (FEV1 > 70% predicted, FEV1/FVC > 70%, little response to a bronchodilator), normal methacholine airway responsiveness (a provocation concentration of methacholine to cause a fall in FEV1 of 20% (PC20) of > 8 mg/mL). People with symptomatic gastroesophageal reflux not improved with treatment and PND who had been previously investigated and treated, and who had no sinusitis on sinus X-rays were included.

Exclusion criteria: smokers, ex-smokers of 6 months or less, indication of respiratory infection in previous month, history of chronic bronchitis, radiological evidence of chest disease, other recognised condition or drugs to account for the cough, cardiovascular or renal disease requiring regular medication, pregnant, received corticosteroids within the month.

Baseline characteristics of treatment:control groups: comparable.


InterventionsICS: budesonide 400 μg twice daily (800 μg/day).

Control: placebo.

Administration method: DPI (Turbuhaler).

Treatment duration: 2 weeks.

Co-interventions: none.


OutcomesQuestionnaire of cough frequency, cough discomfort (9-point Likert scale).

VAS (cough discomfort in the previous two days).

Spirometry.

Sputum total and differential cell counts, ECP, IL-8, fibrinogen, albumin, substance P.

Local side effects (structured questionnaire, oropharyngeal inspection).


NotesMain outcome: treatment did not affect cough or sputum measurements, perhaps because the cause was not associated with sputum eosinophilia.

Baseline investigations: bloods, allergy skin prick tests.

Study period followed by 2 weeks of open label budesonide.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation.

Allocation concealment (selection bias)Low riskRandomisation generated off-site, concealed from investigators, administered by research nurse.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind, identical placebo.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Low riskSimilar withdrawal rates for treatment and control groups.

Selective reporting (reporting bias)Unclear riskSeveral outcomes not reported (change in cough questionnaire, BHR, spirometry, albumin), number screened and eligible reported, reasons for withdrawals stated.

Other biasLow riskParallel group trial.

Ponsioen 2005

MethodsDesign: parallel group.

Randomisation: yes, computer-generated.

Blinding: double blind.

Withdrawals: 2.


ParticipantsSetting: community-based primary healthcare centre (6 practices; The Netherlands).

Number screened: 162.

Number eligible: 135.

Number randomised:135.

Total study population (published data):

Number in treatment group: 67 randomised (GSK report), 65 analysed.

Number in control group: 68 randomised (GSK report), 68 analysed.

Number of withdrawals (treatment:control): 2 (2:0).

Number completing trial (treatment:control): 133 (65:68).

Sex: 47 M, 86 F (analysed population).

Age (years): mean 47.0 (SD 10.1) treatment, mean 43.4 (SD 11.2) control.

Cough duration (weeks): mean 4.2 (SD 2.5) treatment, mean 5 (SD 3.7) control; 31 acute cough (< 3 weeks), 89 subacute cough (3-8 weeks), 13 chronic cough (8-17 weeks) (analysed population).

Subacute and chronic cough participants only (unpublished data):

Number in treatment group: 52.

Number in control group: 50.

Number of withdrawals (treatment:control): 0.

Number completing trial (treatment:control): 102 (52:50).

Sex: 34 M, 68 F.

Age (years): mean 46.7 (SD 10.5) treatment, mean 44.5 (SD 11.2) control.

Cough duration (weeks): mean 4.8 (SD 2.5) treatment, mean 6.1 (SD 3.7) control.

Inclusion criteria: aged 18–65 years with cough of ≥ 2 weeks duration. Participants completed a daily diary card for cough (score 0 = absent, 1 = mild, 2 = moderate, 3 = severe) and other LRT symptoms regarding the previous day and night. Only people with a night score of ≥1 point and a combined day plus night score of ≥3 points were included.

Exclusion criteria: history of asthma; incidences of self-reported wheeze, pharmacy data indicating asthma-like symptoms or variability in lung function in the previous year; current treatment that might influence the cough; FEV1 < 60% predicted; any concurrent airway disease (e.g. pneumonia, cancer, tuberculosis, tonsillitis, sinusitis); uncontrolled systemic disease or pregnancy; people previously randomised for the study.

Baseline characteristics of treatment:control groups: comparable.


InterventionsICS: fluticasone propionate 500 μg twice daily (1000 μg/day).

Control: placebo.

Administration method: MDI via Volumatic spacer.

Treatment duration: 2 weeks.

Co-interventions: no rescue medication allowed, concurrent medication for lower respiratory tract symptoms was registered in the diary.


OutcomesSymptom score (cough, sputum production, wheezing, shortness of breath and chest tightness).

Perception of whether coughing had strongly improved, improved, not changed or increased.

Spirometry (FEV1, FVC).

BHR (histamine challenge).

Number of awakenings at night.

Number of days off work.

Requirement for additional medication after the treatment period.

Hoarseness.

Other adverse events.

(Number cigarettes smoked).


NotesMain outcome: among the total study population, cough score decreased significantly more in the treatment group, however a favourable effect was only detectable in non-smokers. ICS was not effective after exclusion of participants with acute cough. Allergy, FEV1 and BHR at baseline did not predict the efficacy of ICS.

Baseline investigations: bloods (CRP, Phadiatop).


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation.

Allocation concealment (selection bias)Low riskStudy medication provided by GlaxoSmithKline (GSK) according to randomisation list.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind, matching placebo inhaler (GSK report).

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Low riskOnly 2 withdrawals (treatment group).

Selective reporting (reporting bias)Unclear riskFEF25-75% reported in results but is not described as an outcome in methods section, number screened and eligible reported, reasons for withdrawals stated.

Other biasLow riskCompliance measured; parallel group trial.

Pornsuriyasak 2005

MethodsDesign: parallel group.

Randomisation: yes, method not reported.

Blinding: double blind.

Withdrawals: stated.


ParticipantsSetting: 1200-bed university hospital (Thailand).

Number screened: not reported.

Number eligible: not reported.

Number randomised: 30.

Number in treatment group: 15.

Number in control group: 15.

Number of withdrawals (treatment:control): 4 (1:3).

Number completing trial (treatment:control): 26 (14:12).

Sex: 6 M, 24 F.

Age (years): mean 40.6 (SD 11.8) treatment, mean 38.8 (SD 13.0) control.

Cough duration (weeks): mean 5.93 (SD 1.94) treatment, mean 4.66 (SD 2.05) control.

Inclusion criteria: consenting, non-smoking adults with persistent post-URTI cough of > 3 weeks duration; aged > 15 years; normal physical examination; normal CXR; spirometry at baseline: FVC ≥ 80% predicted, FEV1 ≥ 80% predicted, and FEV1/FVC (≥ 70% predicted); sinusitis treated appropriately by an Ear, Nose, Throat physician before entry into the study, if the patient had physical signs and/or an abnormal sinus radiography. People with a negative methacholine challenge test were randomised, however three people with a mildly positive bronchial provocation test were included.

Exclusion criteria: medical history suggesting asthma; symptoms suggesting gastroesophageal reflux prior to or during post-URTI coughing bouts; history of taking medication that induces coughing; contraindication to methacholine challenge testing. Initial treatment with beta2 agonists (inhaled and oral), theophyllines, corticosteroids (oral), and inhaled anticholinergics were terminated at least 1 week prior to entry into the study.

Baseline characteristics of treatment:control groups: comparable.


InterventionsICS: budesonide 100 μg, 4 puffs twice daily (800 μg daily).

Control: placebo, 4 puffs twice daily.

Administration method: DPI.

Treatment duration: 4 weeks.

Co-interventions: other medications were allowed.


OutcomesSymptom score (frequency of cough, frequency of coughing bouts, symptoms associated with cough, night-time cough).

BHR (methacholine challenge test).

Spirometry (FEV1, FVC, FEF 25%-75%).

Frequency of taking medications to relieve cough.

Number of medications to relieve cough.


NotesMain outcome: ICS ineffective in treating persistent post-URTI cough in previously healthy individuals.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandomised, method not reported.

Allocation concealment (selection bias)Unclear riskMethod not reported.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Low riskSimilar attrition rates for treatment and control groups.

Selective reporting (reporting bias)Unclear riskNumber of participants screened and eligible not reported, reasons for withdrawals stated, all outcomes reported.

Other biasUnclear riskPost-URTI cough not defined; parallel group trial.

Ribeiro 2007

MethodsDesign: parallel group.

Randomisation: yes, random number table.

Blinding: double blind.

Withdrawals: none.


ParticipantsSetting: Outpatient Respiratory Clinic of the Hospital São Paulo, general practitioners and hospital clinics (Brazil).

Number screened: 147.

Number eligible: 64.

Number randomised: 64.

Number in treatment group: 44.

Number in control group: 20.

Number of withdrawals (treatment:control): 0 (0:0).

Number completing trial (treatment:control): 64 (44:20).

Sex: 22 M, 42 F.

Age (years): mean 46.4 (SD 17.4) treatment, mean 50.1 (SD 18.1) control.

Cough duration (weeks): mean 48.2 (SD 99.6) treatment, mean 36.7 (SD 45.5) control.

Inclusion criteria: cough for at least 8 weeks with normal chest radiograph, plain sinus radiographs in four positions, and ears, nose and throat examination.

Exclusion criteria: previous gastroesophageal reflux disease diagnosis, positive 24-hour oesophageal pH measurement, concurrent respiratory tract infections, and a history or medical diagnosis of asthma, chronic obstructive pulmonary disease, or chronic rhinosinusitis (PND syndrome) and evidence of airflow limitation with a FEV1/FVC of ≤ 70%, no use of medications for cough in the 4 weeks leading up to entry into the study.

Baseline characteristics of treatment:control groups: PD20 (provocation dose causing a decline in FEV1 of 20%) greater in treatment group (5.35 mg/mL ± 3.2 mg/mL versus 4.56 mg/mL ± 3.7 mg/mL; P value 0.01), otherwise comparable.


InterventionsICS: chlorofluorocarbon-beclomethasone 250 μg, 2 puffs 3 times daily (1500 μg/day).

Control: placebo, 2 puffs 3 times daily.

Administration method: MDI.

Treatment duration: 2 weeks.

Co-interventions: none.


OutcomesSymptom diary: frequency (throughout the day), severity (on arising and throughout the day), duration of coughing (on arising and throughout the day), sleep interruption (throughout the night).

VAS.

Adverse events reported by participants.


NotesMain outcome: ICS provided an excellent response in a subgroup of participants with chronic cough that did not correlate with atopy or airway hyper-responsiveness.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskSimple randomisation in an unbalanced design using a random number table at a ratio of 2:1 (treatment versus control).

Allocation concealment (selection bias)Low riskRandomisation performed by an outside observer.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind, packaging of the study and placebo was identical in appearance and taste and identically marked.

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind.

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo withdrawals.

Selective reporting (reporting bias)Unclear riskAdverse effects not reported as an outcome of interest in methods section but is reported in results, number screened and eligible reported, all outcomes reported.

Other biasLow riskParallel group trial.

Rytilä 2008

MethodsDesign: parallel group.

Randomisation: yes, computer-generated.

Blinding: double blind.

Withdrawals: stated.


ParticipantsSetting: 23 study centres in Finland, Sweden, Norway, Greece, Hungary, UK and Canada; participants were referred to specialists, who were the investigators in the study.

Number screened: 229.

Number eligible: not reported.

Number randomised: 144.

Number in treatment group: 71.

Number in control group: 73.

Number of withdrawals (treatment:control): 23 (10:13).

Number completing trial (treatment:control): 121 (61:60).

Sex: 42 M, 99 F (of 141 with sex recorded).

Age: 20-67 years.

Cough duration: not reported, inclusion criteria ≥ 2 months.

Inclusion criteria: FEV1 ≥ 80% predicted; cough (with or without sputum production) plus at least one additional symptom from chest tightness, wheezing, shortness of breath, or exercise-induced cough or wheezing for ≥ 2 months but < 2 years; average symptom score of ≥1 (scale 0–3) for cough and for sputum production during 7 days of the run-in period (1-2 weeks).

Exclusion criteria: physician-diagnosed asthma; ≥12% increase in absolute FEV1 during reversibility testing at screening; average daily morning/evening peak expiratory flow (PEF) variability ≥ 20% for the week prior to baseline; history of smoking within 12 months prior to screening or a smoking history > 10 pack-years; evidence of chronic obstructive pulmonary disease, chronic cough due to PND, asthma, chronic bronchitis, sinusitis or gastro-oesophageal reflux (careful medical history and radiographs of the chest and paranasal sinuses were obtained); an URTI within 4 weeks prior to screening. People with symptoms of allergic/nonallergic rhinitis were treated with nasal corticosteroids and/or antihistamines before the study; such treatment could not be changed during the study.

Baseline characteristics of treatment:control groups: comparable.


InterventionsICS: mometasone furoate 400 μg daily.

Control: placebo.

Administration method: DPI (Twisthaler).

Treatment duration: 8 weeks.

Co-interventions: salbutamol inhaler could be used as a reliever medication, no other medications were allowed.


OutcomesSymptom scores: cough, sputum production, wheeze, shortness of breath, chest tightness and exercise-induced cough/wheeze.

BHR (histamine or methacholine).

PEF.

Requirement for supplemental salbutamol use.

Sputum eosinophils, ECP.

Adverse events reported by patient.


NotesMain outcomes: ICS improved total morning symptom scores but not total evening symptom scores. ICS improved all individual symptom scores, although this was not always statistically significant. ICS improved morning and evening PEF and reduced sputum eosinophils and ECP.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomisation

Allocation concealment (selection bias)Low riskRandomisation code maintained in sealed envelope

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDouble blind, identical-looking placebo inhaler

Blinding of outcome assessment (detection bias)
All outcomes
Low riskDouble blind

Incomplete outcome data (attrition bias)
All outcomes
Low riskSimilar withdrawal rates for treatment and control groups; reasons for withdrawals reported

Selective reporting (reporting bias)Unclear riskSome outcomes not reported (BHR at 8 weeks, individual symptom scores other than wheeze and cough), PEF monitoring not part of study design section but is reported in methods, number eligible not reported, number screened reported, reasons for withdrawals stated

Other biasLow risk

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Brightling 2000Open label uncontrolled trial.

Cheriyan 1994Retrospective uncontrolled trial.

Fujimoto 2003Prospective uncontrolled trial of bronchodilator, antiallergic and inhaled or oral glucocorticoid therapy.

Gillissen 2007Inclusion criteria of post-infectious cough of between 3 and 14 days duration (i.e. acute cough).

Han 2009Randomised trial of inhaled fluticasone versus oral codeine plus levodropropizine.

Matsuoka 2010Retrospective uncontrolled study.

Park 2007Uncontrolled trial assessing roles of the capsaicin cough sensitivity test, methacholine bronchial provocation test and induced sputum test in evaluation of chronic nonproductive cough.

Rytilä 2000Inclusion criteria of at least 2 of 6 respiratory symptoms (cough, chest tightness with wheezing, shortness of breath, sputum production, wheezing or cough at exercise, and disturbed sleep), not necessarily always including cough.

Stankovic 2010Open-label, non-randomised trial of inhaled corticosteroid and β2 agonist versus oral β2 agonist.

Stankovik 2004Open-label, non-randomised, uncontrolled trial of inhaled β2 agonist followed by ICS treatment.

Wei 2011Prospective observational study of inhaled and oral bronchodilator therapy versus inhaled budesonide and oral bronchodilator therapy for cough-variant asthma.

Xu 2011Comparison of 4, 8 and 16 weeks inhaled budesonide therapy for eosinophilic bronchitis; no placebo comparison.

 
Characteristics of studies awaiting assessment [ordered by study ID]
Tagaya 2009

MethodsDesign: parallel group.

Randomisation: yes, method not reported.

Blinding: not reported.

Withdrawals: not reported.

ParticipantsSetting: not reported (Japan).

Number screened: not reported.

Number eligible: not reported.

Number randomised: 25.

Number in treatment group: not reported.

Number in control group: not reported.

Number of withdrawals (treatment:control): not reported.

Number completing trial (treatment:control): not reported.

Sex: not reported.

Age range: not reported.

Cough duration: not reported.

Inclusion criteria: CVA.

Exclusion criteria: not reported.

Baseline characteristics of treatment:control groups: not reported.

InterventionsICS: salmeterol/fluticasone propionate combination 50/100 μg, once daily (100 μg/day).

Control: salmeterol 50 μg twice daily - different dose of salmeterol.

Administration method: not reported.

Treatment duration: 12 weeks.

Co-interventions: long acting beta-adrenoreceptor agonist (LABA).

OutcomesCough score.

FEV1.

PEF.

Sputum eosinophils, ECP.

NotesMain outcome: maintenance therapy with SFC improved cough symptoms, pulmonary function and airway inflammation. Discontinuation caused worsening of the disease.

Reported as abstract. Contacted Dr Tagaya but unpublished data not available. Awaiting publication of full paper.

Tagaya 2011

MethodsDesign: parallel group.

Randomisation: yes, method not reported.

Blinding: not reported.

Withdrawals: not reported.

ParticipantsSetting: multicentre trial (Japan).

Number screened: not reported.

Number eligible: not reported.

Number randomised: 27.

Number in treatment group: 14.

Number in control group: 13.

Number of withdrawals (treatment:control): not reported.

Number completing trial (treatment:control): not reported.

Sex: not reported.

Age range: not reported.

Cough duration: not reported.

Inclusion criteria: CVA according to Japanese cough guidelines.

Exclusion criteria: not reported.

Baseline characteristics of treatment:control groups: comparable.

InterventionsICS: budesonide/formoterol combination 160/4.5 μg twice daily (total budesonide dose 320 μg/day).

Control: salmeterol 50 μg twice daily.

Administration method: DPI.

Treatment duration: 8 weeks.

Co-interventions: LABA, supplemental procaterol (SABA).

OutcomesCough symptom score.

Cough and sputum assessment questionnaire (CASA-Q).

FEV1.

PEF.

Supplemental use of inhaled procaterol (SABA).

Sputum eosinophils, ECP.

NotesMain outcome: treatment decreased cough symptom scores, CASA-Q scores, diurnal variation of PEF, eosinophil counts and ECP.

Reported as abstract. Contacted Dr Tagaya but unpublished data not available. Awaiting publication of full paper.

 
Comparison 1. ICS versus Placebo

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

 1 Proportion of participants with clinical cure or significant improvement (> 70% reduction in cough severity measure) at follow up (clinical success)3Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

    1.1 Cough score
3Odds Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    1.2 VAS
1Odds Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

 2 Proportion of participants with clinical cure or > 50% reduction in cough severity measure at follow up4Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 3 Proportion of participants with clinical cure at follow up4Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 4 Mean change in cough score5346Std. Mean Difference (IV, Fixed, 95% CI)-0.34 [-0.56, -0.13]

 5 Mean change in cough measures on VAS2Std. Mean Difference (Fixed, 95% CI)Totals not selected

 6 Mean change in VAS after 2 weeks by final diagnosis [cm]1Mean Difference (Fixed, 95% CI)Totals not selected

    6.1 PNDS
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    6.2 GORD
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    6.3 CVA
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    6.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    6.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

 7 Mean change in morning and evening cough score1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    7.1 Morning cough score 4 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    7.2 Morning cough score 8 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    7.3 Evening cough score 4 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    7.4 Evening cough score 8 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 8 Mean change in morning and evening total symptom score1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    8.1 Morning total symptom score 4 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    8.2 Morning total symptom score 8 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    8.3 Evening total symptom score 4 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    8.4 Evening total symptom score 8 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 9 Mean change in cough frequency1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    9.1 After 1 week
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    9.2 After 2 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 10 Mean change in cough severity1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    10.1 After 1 week
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    10.2 After 2 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 11 Mean change in cough time of day1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    11.1 After 1 week
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    11.2 After 2 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 12 Proportion without BHR after treatment194Odds Ratio (M-H, Fixed, 95% CI)1.28 [0.38, 4.34]

    12.1 Baseline BHR
134Odds Ratio (M-H, Fixed, 95% CI)1.27 [0.33, 4.93]

    12.2 No baseline BHR
160Odds Ratio (M-H, Fixed, 95% CI)1.32 [0.08, 22.15]

 13 Mean change in BHR1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    13.1 Measured by methacholine challenge
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    13.2 Measured by citric acid challenge
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 14 Change in FEV11Mean Difference (IV, Fixed, 95% CI)Totals not selected

    14.1 Non-smokers
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    14.2 Smokers
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 15 Proportion requiring additional medication1132Odds Ratio (M-H, Fixed, 95% CI)0.46 [0.23, 0.91]

    15.1 Non-smokers
184Odds Ratio (M-H, Fixed, 95% CI)0.31 [0.13, 0.76]

    15.2 Smokers
148Odds Ratio (M-H, Fixed, 95% CI)0.85 [0.27, 2.63]

 16 Mean change in sleep interruption1Mean Difference (IV, Fixed, 95% CI)Totals not selected

    16.1 After 1 week
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

    16.2 After 2 weeks
1Mean Difference (IV, Fixed, 95% CI)0.0 [0.0, 0.0]

 17 Mean change in sputum ECP [ng/mL]1Mean Difference (Fixed, 95% CI)-396.0 [-791.99, -0.01]

 18 Mean change in sputum ECP by final diagnosis [ng/mL]1Mean Difference (Fixed, 95% CI)Totals not selected

    18.1 PNDS
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    18.2 GORD
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    18.3 CVA
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    18.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    18.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

 19 Change in sputum total cells [×106]1Mean Difference (Fixed, 95% CI)Subtotals only

    19.1 PNDS
1Mean Difference (Fixed, 95% CI)-1.5 [-11.95, 8.95]

    19.2 GORD
1Mean Difference (Fixed, 95% CI)-4.6 [-12.30, 3.10]

    19.3 CVA
1Mean Difference (Fixed, 95% CI)3.4 [-7.05, 13.85]

    19.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    19.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-2.8 [-15.20, 9.60]

    19.6 All causes
1Mean Difference (Fixed, 95% CI)-1.0 [-8.55, 6.55]

 20 Change in sputum neutrophils [%]1Mean Difference (Fixed, 95% CI)Subtotals only

    20.1 PNDS
1Mean Difference (Fixed, 95% CI)-1.2 [-18.60, 16.20]

    20.2 GORD
1Mean Difference (Fixed, 95% CI)-13.8 [-29.75, 2.15]

    20.3 CVA
1Mean Difference (Fixed, 95% CI)0.9 [-19.90, 21.70]

    20.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)0.6 [-20.20, 21.40]

    20.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)5.3 [-14.20, 24.80]

    20.6 All causes
1Mean Difference (Fixed, 95% CI)1.1 [-7.50, 9.70]

 21 Change in sputum eosinophils [%]1Mean Difference (Fixed, 95% CI)Subtotals only

    21.1 PNDS
1Mean Difference (Fixed, 95% CI)0.0 [-2.05, 2.05]

    21.2 GORD
1Mean Difference (Fixed, 95% CI)-0.1 [0.00, 1.80]

    21.3 CVA
1Mean Difference (Fixed, 95% CI)-4.6 [-7.10, -2.10]

    21.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)1.5 [-0.95, 3.95]

    21.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)0.1 [-2.20, 2.40]

    21.6 All causes
1Mean Difference (Fixed, 95% CI)-0.7 [-1.75, 0.35]

 22 Change in sputum lymphocytes [%]1Mean Difference (Fixed, 95% CI)Subtotals only

    22.1 PNDS
1Mean Difference (Fixed, 95% CI)0.0 [-0.35, 0.35]

    22.2 GORD
1Mean Difference (Fixed, 95% CI)0.1 [-0.25, 0.45]

    22.3 CVA
1Mean Difference (Fixed, 95% CI)-0.3 [-0.70, 0.10]

    22.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-0.1 [-0.50, 0.30]

    22.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-0.1 [-0.50, 0.30]

    22.6 All causes
1Mean Difference (Fixed, 95% CI)-0.1 [-0.25, 0.05]

 23 Change in sputum MPO [µg/mL]1Mean Difference (Fixed, 95% CI)Subtotals only

    23.1 PNDS
1Mean Difference (Fixed, 95% CI)-34.0 [-113.30, 45.30]

    23.2 GORD
1Mean Difference (Fixed, 95% CI)-17.7 [-97.05, 61.65]

    23.3 CVA
1Mean Difference (Fixed, 95% CI)70.0 [-98.30, 238.30]

    23.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)133.5 [27.05, 239.95]

    23.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)9.1 [-128.30, 146.50]

    23.6 All causes
1Mean Difference (Fixed, 95% CI)10.7 [-29.60, 51.00]

 24 Change in sputum PGE2 [ng/mL]1Mean Difference (Fixed, 95% CI)-1.82 [-7.21, 3.56]

    24.1 PNDS
1Mean Difference (Fixed, 95% CI)4.9 [-10.30, 20.10]

    24.2 GORD
1Mean Difference (Fixed, 95% CI)-4.7 [-20.85, 11.45]

    24.3 CVA
1Mean Difference (Fixed, 95% CI)12.1 [-10.70, 34.90]

    24.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-11.6 [-30.25, 7.05]

    24.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-9.8 [-32.65, 13.05]

    24.6 All causes
1Mean Difference (Fixed, 95% CI)-1.9 [-9.05, 5.25]

 25 Change in sputum LTB4 [ng/mL]1Mean Difference (Fixed, 95% CI)Subtotals only

    25.1 PNDS
1Mean Difference (Fixed, 95% CI)5.1 [-32.85, 43.05]

    25.2 GORD
1Mean Difference (Fixed, 95% CI)1.70 [-36.30, 39.70]

    25.3 CVA
1Mean Difference (Fixed, 95% CI)25.6 [-18.25, 69.45]

    25.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-15.7 [-49.65, 18.25]

    25.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-9.2 [-53.05, 34.65]

    25.6 All causes
1Mean Difference (Fixed, 95% CI)2.3 [-13.20, 17.80]

 26 Change in sputum Cys-LT [ng/mL]1Mean Difference (Fixed, 95% CI)Subtotals only

    26.1 PNDS
1Mean Difference (Fixed, 95% CI)0.1 [-1.50, 1.70]

    26.2 GORD
1Mean Difference (Fixed, 95% CI)-0.2 [-1.85, 1.45]

    26.3 CVA
1Mean Difference (Fixed, 95% CI)-0.6 [-2.85, 1.65]

    26.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-1.4 [-3.25, 0.45]

    26.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)0.3 [-2.30, 2.90]

    26.6 All causes
1Mean Difference (Fixed, 95% CI)-0.4 [-1.20, 0.40]

 27 Change in sputum IL-8 [ng/mL]1Mean Difference (Fixed, 95% CI)Subtotals only

    27.1 PNDS
1Mean Difference (Fixed, 95% CI)-3.5 [-61.95, 54.95]

    27.2 GORD
1Mean Difference (Fixed, 95% CI)-27.0 [-88.95, 34.95]

    27.3 CVA
1Mean Difference (Fixed, 95% CI)1.5 [-86.10, 89.10]

    27.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-74.7 [-146.30, -3.10]

    27.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-20.9 [-122.10, 80.30]

    27.6 All causes
1Mean Difference (Fixed, 95% CI)-21.5 [-48.25, 5.25]

 28 Change in sputum TNF-α [ng/mL]1Mean Difference (Fixed, 95% CI)Subtotals only

    28.1 PNDS
1Mean Difference (Fixed, 95% CI)-3.9 [-17.70, 9.90]

    28.2 GORD
1Mean Difference (Fixed, 95% CI)3.6 [-8.40, 15.60]

    28.3 CVA
1Mean Difference (Fixed, 95% CI)0.0 [-13.80, 13.80]

    28.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)0.4 [-10.30, 11.10]

    28.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)0.0 [0.0, 0.0]

    28.6 All causes
1Mean Difference (Fixed, 95% CI)0.3 [-4.85, 5.45]

 29 Change in eNO [ppb]1Mean Difference (Fixed, 95% CI)Subtotals only

    29.1 PNDS
1Mean Difference (Fixed, 95% CI)0.0 [0.00, 2.00]

    29.2 GORD
1Mean Difference (Fixed, 95% CI)-3.1 [-5.75, -0.45]

    29.3 CVA
1Mean Difference (Fixed, 95% CI)-3.3 [-6.45, -0.15]

    29.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-0.90 [-4.50, 2.70]

    29.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-1.4 [-4.85, 2.05]

    29.6 All causes
1Mean Difference (Fixed, 95% CI)-2.1 [-3.56, -0.64]

 30 Change in exhaled CO [ppm]1Mean Difference (Fixed, 95% CI)Subtotals only

    30.1 PNDS
1Mean Difference (Fixed, 95% CI)-0.1 [-0.60, 0.40]

    30.2 GORD
1Mean Difference (Fixed, 95% CI)-0.3 [1.00, 0.40]

    30.3 CVA
1Mean Difference (Fixed, 95% CI)-0.3 [-1.10, 0.50]

    30.4 Bronchiectasis
1Mean Difference (Fixed, 95% CI)-0.7 [-1.60, 0.20]

    30.5 Idiopathic cough
1Mean Difference (Fixed, 95% CI)-0.1 [-0.95, 0.75]

    30.6 All causes
1Mean Difference (Fixed, 95% CI)-0.34 [-0.66, -0.02]

 31 Proportion with adverse effects4381Odds Ratio (M-H, Fixed, 95% CI)1.67 [0.92, 3.04]

 32 Proportion with specific adverse effects2Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

    32.1 Hoarseness
2Odds Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    32.2 Sore throat
2Odds Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

    32.3 Oral candidiasis
2Odds Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]

 33 Proportion with severe adverse effects3Odds Ratio (M-H, Fixed, 95% CI)Totals not selected

 
Summary of findings for the main comparison. Inhaled corticosteroids (ICS) compared to placebo for adults with subacute and chronic cough

Inhaled corticosteroids (ICS) compared to placebo for adults with subacute and chronic cough

Patient or population: adults with subacute and chronic cough
Settings: all
Intervention: ICS
Comparison: placebo

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

PlaceboICS

Primary outcome

Proportion of participants who achieved clinical cure or significant improvement (> 70% reduction in cough severity measure) at follow up (clinical success)
Symptomatic cough severity measure as assessed by the patient
Follow up: 2 to 4 weeks
See commentSee commentNot estimable180
(3 studies)
⊕⊕⊝⊝
low1,2
Meta-analysis not appropriate; heterogeneity explained by differences in study design and outcomes.

Secondary outcomes

Proportion of participants who achieved clinical cure or > 50% reduction in cough severity measure at follow up
Symptomatic cough severity measure as assessed by the patient
Followup: 2 to 4 weeks
See commentSee commentNot estimable230
(4 studies)
⊕⊕⊝⊝
low2,3
Meta-analysis not appropriate; heterogeneity explained by differences in study design and outcomes.

Proportion of participants with clinical cure at follow upSee commentSee commentNot estimable320
(4 studies)
⊕⊕⊝⊝
low2,4
Meta-analysis not appropriate; heterogeneity explained by differences in study design and outcomes.

Mean change in cough score
Symptomatic cough severity measure as assessed by the patient
Follow up: 2 to 8 weeks
The mean difference in cough severity measure in the intervention groups was
0.34 standard deviations lower
(0.56 to 0.13 lower),
SMD -0.34 (-0.56 to -0.13)346
(5 studies)
⊕⊕⊝⊝
low5,6

Proportion with adverse effects of treatment
Follow up: mean 2 to 8 weeks
116 per 1000180 per 1000
(108 to 285)
OR 1.67
(0.92 to 3.04)
381
(4 studies)
⊕⊕⊕⊝
moderate7

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; OR: Odds ratio

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

 1 Unclear risk of selection bias (Boulet 1994); unclear risk of reporting bias (Ponsioen 2005; Ribeiro 2007).
2 Dichotomous outcome data based on less than 300 events.
3 Unclear risk of selection bias (Boulet 1994).
4 Unclear risk of selection bias (Boulet 1994); unclear risk of reporting bias (Ponsioen 2005; Ribeiro 2007; Rytilä 2008).
5 Unclear risk of selection bias (Boulet 1994; Pornsuriyasak 2005) and other bias (Pornsuriyasak 2005).
6 Continuous outcome data based on total population size less than 400.
7 Wide 95% CI.