Hospitalization for chronic obstructive pulmonary disease and pneumonia: association with the dose of inhaled corticosteroids. A nation-wide cohort study of 52 100 outpatients.

OBJECTIVES: International guidelines only advocate the use of inhaled corticosteroids (ICSs) in patients with chronic obstructive pulmonary disease (COPD) experiencing recurring exacerbations and eosinophilic inflammation. However, ICSs are commonly used in patients with COPD and without exacerbations and signs of eosinophilic inflammation, thus possibly increasing the risk of hospitalization for pneumonia. Thus, we aimed to determine the risk of hospitalization for pneumonia associated with increasing cumulated ICS doses among patients with COPD to establish whether there is dose dependency. METHODS: A retrospective cohort study included all patients with COPD treated at a respiratory outpatient clinic in Denmark. The patients were divided into four groups based on their average daily ICS exposure. The dose-response relationship was investigated using a multivariable Cox proportional hazard regression analysis. RESULTS: In total, 52 100 patients were included, who were divided into the no-use (n = 15 755), low-dose (n = 12 050), moderate-dose (n = 12 488), and high-dose (n = 11 807) groups. ICS use was strongly associated with hospitalization for pneumonia (hazard ratio [HR], 1.3; CI, 1.2-1.3) (ICS vs. no ICS). The risk of hospitalization for pneumonia increased with every dosing group step: low dose: HR, 1.1 (CI, 1.0-1.2); moderate dose: HR, 1.2 (CI, 1.1-1.3), and high dose: HR, 1.5 (CI, 1.4-1.6); "no use" was the reference. Sensitivity analyses confirmed these findings. CONCLUSIONS: In the dose-response relationship analysis, ICS dose were associated with a substantially increased risk of hospitalization for pneumonia of up to 50%. Our data support that ICSs should be administered at the lowest possible dose and only to patients with COPD who have a documented need.