Roflumilast is given as an add-on to inhalation medication in patients with chronic obstructive pulmonary disease (COPD) and chronic bronchitis. Animal experiments have documented deleterious effects of roflumilast in bacterial infections, but trials have not reported the risk of bacterial infections in patients. The objective of this study is to determine, among outpatients with severe COPD in a two-year follow-up period, the risk of hospitalization-requiring pneumonia, severe acute exacerbation in COPD (AECOPD-hosp), and death. Patients with COPD using roflumilast (roflumilast users) were compared to a propensity score-matched COPD control group not using roflumilast (non-roflumilast users). Roflumilast users had an increased 2-year risk of hospitalization-requiring pneumonia (HR 1.5, 95% CI 1.3 to 1.8, p-value <0.0001) compared to controls, and of AECOPD-Hosp (hazard ratio(HR) 1.6, 95%, confidence interval (CI) 1.5 to 1.8, p-value <0.0001) and. When adding an active comparator (theophylline) as a matching variable, the signal was largely unchanged. In conclusion, roflumilast was associated with an increased number of hospitalizations for pneumonia and for AECOPD. Since trials have not reported risks of bacterial complications and data regarding severe exacerbations in roflumilast users are sparse and diverging, these data are concerning. Trials focused on the risk of pneumonia, AECOPD, and other bacterial infections in roflumilast users are needed urgently.