n engl j med 365;23 nejm.org december 8, 2011 2237 developed during the study, only 44 had macrolide resistance (41%) (P<0.001). In my opinion, this is the correct method to evaluate resistance to macrolides because it compares the subjects of the two groups in whom colonization developed during the study period and not the whole group, as suggested by Hahn. Adequate data on resistance from the whole group were not available for comparison. I strongly believe that macrolide resistance is an issue and should be taken into account, especially when treating a subsequent acute exacerbation of COPD.