Association of Inhaled Corticosteroids With All-Cause Mortality Risk in Patients With COPD: A Meta-analysis of 60 Randomized Controlled Trials

Background: Inhaled corticosteroids (ICSs) have been used widely in the maintenance therapy of COPD. However, whether inhaled therapy containing ICSs can reduce the all-cause mortality risk and the possible benefited patient subgroups is unclear.

Research Question: Does inhaled therapy containing ICSs reduce the all-cause mortality risk in patients with COPD compared with other inhaled therapies not containing ICSs?

Study Design And Methods: We searched PubMed, Cochrane Library, Embase, and ClinicalTrials.gov for relevant randomized clinical trials (RCTs). Pooled results were calculated using Peto ORs with corresponding 95% CIs.

Results: Sixty RCTs enrolling 103,034 patients were analyzed. Inhaled therapy containing ICSs (Peto OR, 0.90; 95% CI, 0.84-0.97), especially triple therapy (Peto OR, 0.73; 95% CI, 0.59-0.91), was associated with a reduction in the all-cause mortality risk among patients with COPD when compared with inhaled therapy without ICSs. Subgroup analyses revealed that treatment duration of > 6 months (Peto OR, 0.90; 95% CI, 0.83-0.97), medium-dose ICSs (Peto OR, 0.71; 95% CI, 0.56-0.91), low-dose ICSs (Peto OR, 0.88; 95% CI, 0.79-0.97), and budesonide (Peto OR, 0.75; 95% CI, 0.59-0.94) were involved in this association. The predictors of this association included eosinophil counts of ≥ 200/μL or percentage of ≥ 2%, documented history of ≥ 2 moderate and severe exacerbations in the previous year, Global Initiative for Chronic Obstructive Lung Disease stages III or IV, age younger than 65 years, and BMI of ≥ 25 kg/m, among which eosinophil counts of ≥ 200/μL (Peto OR, 0.58; 95% CI, 0.36-0.95) were the strongest predictor.

Interpretation: Inhaled therapy containing ICSs, especially triple therapy, of longer than 6 months was associated with a reduction in the all-cause mortality risk in patients with COPD. The predictors of this association included medication factors and patient characteristics, among which eosinophil counts of ≥ 200/μL were the strongest predictor.

Trial Registry: PROSPERO; No.: CRD42022304725; URL: https://www.crd.york.ac.uk/prospero/.