The Kaplan–Meier analyses of the time to the first exacerbation show that the benefit with tiotropium as compared with salmeterol became evident as early as approximately 1 month after
the initiation of treatment and was maintained over the entire 1-year study period. Thus, it appears to be unlikely that the difference in favor of tiotropium was due to early discontinuation of treatment among patients in the salmeterol group who did not have a response to that drug.