Figure 3.

Time to onset of action of study medication at baseline (day 0; similar plots were obtained at weeks 2, 6 and 12). Data based on Kaplan-Meier survival analysis (full analysis set). Onset of action was defined as the first time point post-dose at which FEV₁ was at least 12% greater than the pre-dose value. Analysis of time to onset of action using the multiple failures time model showed superiority of fluticasone/formoterol over fluticasone/salmeterol (hazard ratio: 1.64; 95% CI: 1.28, 2.10; p < 0.001). On day 0, more than twice as many patients had a bronchodilatory response that met the onset of action criterion in the fluticasone/formoterol group than in the fluticasone/salmeterol group at 5 minutes post-dose (39% versus 14%, respectively). In addition, a larger proportion of patients in the fluticasone/formoterol group met the onset of action criteria within 2 hours (post-dose), than in the fluticasone/salmeterol group (77% versus 64%, respectively; day 0). This trend continued over the 12-week treatment period, with consistently more patients achieving the predefined onset of action response in the fluticasone/formoterol group than in the fluticasone/salmeterol group at each post-dose time point (5–120 minutes). CI, confidence interval; FEV₁, forced expiratory volume in the first second.