Comparison of pneumonia incidence between long-acting muscarinic antagonist and inhaled corticosteroid plus long-acting beta agonist in patients with COPD : A retrospective cohort study in a real-world setting

Abstract Background Pneumonia is an infectious disease commonly associated with chronic obstructive pulmonary disease (COPD). However, few studies have directly compared the incidence of pneumonia in patients on common COPD treatments such as long-acting muscarinic antagonists (LAMA) with those on inhaled corticosteroids and long-acting β2-agonist (ICS/LABA). Moreover, risk factors for pneumonia in COPD are still unclear. We compared the incidence of pneumonia in COPD patients on LAMA and those on ICS/LABA and explored the risk factors associated with pneumonia. Methods This nationwide cohort study used Korean National Health Insurance claim data from January 2002 to April 2016. Patients who received COPD medication, either LAMA or ICS/LABA, with the COPD diagnostic code, were selected. We enrolled patients with good compliance (medication possession ratio ≥ 80%). The primary outcome was pneumonia in COPD patients initiating LAMA or ICS/LABA. We investigated the risk factors associated with pneumonia, including the sub-types of ICS treatments. The observation period was a minimum of 12 months and a maximum of 136 months. Results After propensity score matching, the incidence rate per 1,000 person-years of pneumonia was 93.96 for LAMA (n = 1,003) and 136.42 for ICS/LABA (n = 1,003) patients (p < 0.001). The adjusted hazard ratio (HR) for pneumonia in patients on fluticasone/LABA was 1.496 (95% confidence interval [CI] 1.204–1.859) compared with LAMA (p < 0.001). In multivariable analysis, a history of pneumonia was a risk factor associated with pneumonia (HR, 2.123; 95% CI, 1.580–2.852; p < 0.001). Conclusions The incidence of pneumonia was higher in COPD patients on ICS/LABA compared with those on LAMA. It is recommended that ICS use be avoided in COPD patients with high pneumonia risk.