OBJECTIVES:  Roflumilast (ROF) is an oral phosphodiesterase-4 inhibitor indicated for maintenance treatment of severe COPD associated with chronic bronchitis in adult patients with a history of frequent exacerbations as add on to bronchodilator treatment. The REACT study (NCT01329029) showed that ROF reduces exacerbations and hospital admissions when added to triple therapy of inhaled glucocorticosteroid (ICS), long-acting β2-agonist (LABA), and long-acting muscarinic antagonist (LAMA). The objective of this analysis was to estimate the cost-effectiveness of ROF+ICS+LABA+LAMA vs ICS+LABA+LAMA in managing severe and very severe COPD patients in a Swedish healthcare setting incorporating efficacy outcomes from REACT. METHODS:  A Markov model developed in TreeAge 2009 with an Excel interface was used to simulate life-long costs and outcomes (QALYs). The model had monthly cycles and three health states: severe COPD, very severe COPD and death, with irreversible transitions to very severe COPD. All input parameters on costs and epidemiology were obtained from Swedish sources. Clinical effectiveness was based on results from clinical trials. Analyses adopted a societal perspective and included lost productivity using a human capital approach. Transition probabilities, state utilities, exacerbation disutilities and hospital mortality were based on published literature. Background mortality was determined through risk of death in the general population and standardised mortality ratios from the literature. Costs and outcomes were discounted at 3%. Uncertainty was investigated through one-way and probabilistic sensitivity analyses. RESULTS: