In the May 2013 issue of CHEST, Wedzicha et al1 conclude in the abstract section of their article that the reduction in severe exacerbations leading to hospitalization or death associated with roflumilast was similar between subgroups defined as infrequent and frequent exacerbators. This is misleading, as it implies there is a reduction in severe exacerbations associated with roflumilast. The datasets published in this article report the number of patients who experienced infrequent exacerbations (zero or one) or frequent exacerbations (two or more) with roflumilast vs placebo. Although listed as a study end point, there are no data presented illustrating the number of patients experiencing severe exacerbations to support the claim that there is a reduction in severe exacerbations associated with roflumilast. The authors state that in patients defined as frequent exacerbators in the year prior to study, 3.4% in the roflumilast and 6.5% in the placebo group had two or more severe exacerbations after 1 year.1 This approached, but did not reach, statistical difference (P = .0516).1 They do not, however, include severe exacerbation data in the placebo and roflumilast groups at year 0 to support there being a reduction in severe exacerbations at 1 year. From the data presented, all that can be concluded is that there is no difference in the number of frequent exacerbator phenotype patients who experience two or more severe exacerbations while given roflumilast when compared with placebo.