Rescue medication use as a patient-reported outcome in COPD: A systematic review and regression analysis

Abstract

Background: Reducing rescue medication use is a guideline-defined goal of asthma treatment, however, little is known about the validity of rescue medicine use as a marker of symptoms in chronic obstructive pulmonary disease (COPD). To improve patient outcomes, greater insight is needed into the relationship between rescue medication use and alternative COPD outcomes. Methods: A systematic search of electronic databases (Embase®, MEDLINE® and Cochrane CENTRAL) was conducted from database start to 26 May, 2015. Studies of bronchodilator therapy with a duration of ≥24weeks were included if they reported either mean change from baseline (CFB) in rescue medication use in puffs/day or % rescue-free days (%RFD), and at least one other COPD endpoint. Correlation and meta-regression analyses were undertaken to test the association between rescue medication use and other COPD outcomes using weighted means (weights proportional to the sample size of the treatment group) and unweighted means (equal weight for each treatment group). Each association was assessed at 6months and study end. Results: Forty-six studies involving 46,531 patients provided mean data from 145 treatment groups for evaluation. Changes in both measures of rescue medication use were correlated with changes in trough forced expiratory volume in one second ([FEV1]; Pearson correlation coefficients |r|≥0.63; p<0.0001) and with St George's Respiratory Questionnaire (SGRQ) score (|r|≥0.70; p<0.0001) at study end. Change in rescue medication use in puffs/day during the study correlated with annualized rates of moderate/severe exacerbations at 6months and study end (both r=0.66; p≤0.0028). CFB in puffs/day was not well correlated with Transition Dyspnoea Index (TDI), but %RFD did correlate with TDI score at 6months and study end (both r=0.69; p<0.0001). The values for CFB in puffs/day corresponding to the proposed minimal clinically important differences for trough FEV1 and SGRQ score were -1.3 and -0.6 puffs/day, respectively. A -1.0 puffs/day CFB in rescue use corresponded to a change of 0.26 events/patient-year in moderate/severe exacerbations. Conclusion: This analysis provides clear evidence of associations at a patient group level between rescue medication use and other clinically important COPD outcomes.