Inhaled corticosteroids for stable chronic obstructive pulmonary disease
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To determine the effect of inhaled corticosteroids in stable patients with non‐asthmatic chronic obstructive pulmonary disease in terms of objective and subjective outcomes:
1. Spirometry (FEV1, FVC or VC) and peak flow
2. Exacerbations of COPD
3. 6 minute walk distance
4. Quality of life and symptom scores
5. Use of rescue bronchodilators
Background
Inhaled corticosteroids (ICS) have proven benefit in the treatment of airway inflammation in asthma but there is no consensus about their effectiveness in non‐asthmatic COPD patients. Major international guidelines for COPD mention selective use of regular ICS in patients with responsiveness to oral corticosteroids (ATS 1995; Siafakas 1995; BTS 1997). However the role of ICS in COPD patients in general has not been established.
The bronchi of patients with COPD show a predominantly neutrophilic infiltration, in studies of induced sputum (Keatings et al. 1996), bronchoalveolar lavage (Martin et al. 1985) and bronchial mucosal biopsy (Ollerenshaw 1992). Upregulation of adhesion molecules occurs in COPD (Di Stefano 1994). In addition, a mononuclear cell infiltrate is often present in the submucosa of the bronchi (Saetta 1993). The potential site of action of ICS has not been defined, but may involve lymphocytes and inflammatory cytokines.
Studies of regular ICS in COPD have been reviewed (van Schayck 1996; Barnes 1998) and one meta‐analysis of 3 long‐term studies of the effect of ICS on decline in lung function has been performed (van Grunsven 1999). To date, there has not been a published meta‐analysis of all relevant clinical studies of efficacy of ICS in COPD. Considering the known adverse effects of oral and laryngeal candidiasis, increased risk for cataracts and potential hypothalamus‐pituitary‐adrenal axis suppression in higher doses, there is an urgent need to carefully assess the current medical literature of the effects of regular inhaled corticosteroids in patients with COPD. Cost‐effectiveness is another important consideration.
Objectives
To determine the effect of inhaled corticosteroids in stable patients with non‐asthmatic chronic obstructive pulmonary disease in terms of objective and subjective outcomes:
1. Spirometry (FEV1, FVC or VC) and peak flow
2. Exacerbations of COPD
3. 6 minute walk distance
4. Quality of life and symptom scores
5. Use of rescue bronchodilators
Methods
Criteria for considering studies for this review
Types of studies
All published and unpublished randomised controlled trials of regular inhaled corticosteroids will be considered. Placebo‐controlled trials and trials comparing different forms of inhaled corticosteroid will be included. Only trials with random allocation and double‐blinding will be included. Trials analysed on an intention‐to‐treat basis are preferred. All trials will be assessed for major quality criteria including proper randomisation, degree of follow‐up and blinding. The scoring method of Jadad 1996 will be used to assess quality.
Types of participants
All adults of either gender, regardless of smoking history, with COPD defined as symptomatic, progressive chronic airflow obstruction (FEV1<80% of predicted) who are in a clinically stable state without recent exacerbation, hospitalisation or need for antibiotics or oral/parenteral corticosteroids. Patients will not have clinical features of asthma. Acute bronchodilator reversibility to short term beta2‐agonists will not be excluded.
Types of interventions
Regular inhaled corticosteroids (beclomethasone, budesonide or fluticasone) administered by any inhalation device including metered dose inhaler, dry powder inhaler, spacer devices and nebuliser.
Types of outcome measures
The studies will be reviewed in terms of the following outcome measures:
1. Spirometry (FEV1, FVC or VC) and peak flow
2. Exacerbations of COPD
3. 6 minute walk distance
4. Quality of life and symptom scores
5. Use of rescue bronchodilators
Search methods for identification of studies
Randomised controlled trials of inhaled corticosteroids of adults with COPD from 1966 to October 2000 will be examined and combined. A list of all potentially relevant trials will be obtained from MEDLINE, CINAHL, EMBASE and the Cochrane Airways Group RCT COPD database, using the search strategy: inhaled AND (corticosteroid OR cortico‐steroid OR beclomethasone OR Aerobec OR Beclazone OR Becotide OR Becloforte OR budesonide OR Pulmicort OR fluticasone OR Flixotide OR QVAR OR Zonivent OR Filair OR Asmabec OR Becodisk). The bibliographies of each relevant RCT will be searched for additional studies that may contain RCTs.
Data collection and analysis
The titles and, where available, abstracts of all trials retrieved by the search strategy will be assessed independently for relevance by 2 reviewers. All relevant or potentially relevant articles will then be retrieved in full. These articles will be categorised as relevant (meet the inclusion criteria for considering studies) or not relevant (do not meet the inclusion criteria for considering studies). Disagreements about relevance will be resolved by consensus. All relevant trials will be assessed independently for quality by 2 reviewers, using the scoring method of Jadad 1996. Disagreements about quality assessment will be resolved by consensus.
Data will be extracted independently from included studies by 2 reviewers. Any missing data in the publication will be sought from the authors by correspondence. The planned analysis will be inhaled corticosteroid versus placebo, although comparisons of different forms of inhaled corticosteroid with each other will be performed if such trials are found. The treatment periods will be stratified into less than 4 weeks and greater than 4 weeks to 6 months treatment duration.
Statistical considerations:
All trials will be combined using Review Manager 4.04. Proportional outcomes, such as proportion who improved, will be summarised using an odds ratio: the Mantel‐Haenszel test will be used for combined estimates of the overall odds ratio and its variance. Tests for heterogeneity will be performed using Review Manager; if this appears important, a random effects model will also be applied.
